GC24 5073 2_DOS_System_Programmers_Guide_Sep71 2 DOS System Programmers Guide Sep71

GC24-5073-2_DOS_System_Programmers_Guide_Sep71 GC24-5073-2_DOS_System_Programmers_Guide_Sep71

User Manual: GC24-5073-2_DOS_System_Programmers_Guide_Sep71

Open the PDF directly: View PDF .
Page Count: 298

Download
Open PDF In Browser	View PDF

File Number S360-50
Order No. GC24-5073-2

Systems Reference Library

DO'S System Programmer's Guide
This reference publication is intended primarily
for the system programmer who is involved in
making decisions relating to the components of the
installation's supervisor, file organization and
program design. To form a single publication,
this manual brings together and expands upon
information from many sources. Major topics
discussed are:
1.

Supervisor Planning Concepts

Data Management

Program Design

Debugging Aids

For each major section, the Preface lists the most
closely related publications. For a complete list
of available publications, see the IBM System/360
and System/310 Bibliography, GA22-6822.

DOS

Third Edition (September 1971)
This publication was formerly titled IBM System/360 Disk
operating System: System Programmer's Guide. Although
titles of some DOS publications (including this one) have
been simplified, the change does not affect the contents of
the publications.
This edition replaces and obsoletes GC24-5073-1. Changes are
continually made to the specifications herein; before using
this publication in connection with the operation of IBM
systems, consult the latest System/360 and System/370 SRL
Newsletter, GN20-0360, for the editions that are applicable
and current.
summary of Amendments
For a list of changes made in this edition, see page 3.
Changes or additions to the text and illustrations are
indicated by a vertical line to the left of the change.
Requests for copies of IBM publications should be made to
your IBM representative or to the IBM branch office serving
your locality.

A form for readers' comments is provided at the back of this
publication. If the form has been removed. comments may be
addressed to IBM Laboratory, Publications Dept., P.O. Box 24,
Uithoorn, Netherlands.

@ Copyright International Business Machines Corporation 1969, 1970, 1971

Summary of Amendments
for GC24-S073-2

NEW FUNCTIONS
This edition documents support for these
system control and service functions:

• Directory Service Display (DSERV)
• Linkage Editor (LNKEDT)

• Independent Directory Read-In Area
(IDRA)
NEW DEVICES
• On-Line Test Executive Programs
(OLTEP)

This edition also documents support for
these new devices:

• Job Accounting Interface
• IBM 1255 Magnetic Character Reader
• Data Set Security
• IBM 2319 Disk Storage
• ISAM Track Hold
• IBM 3210 Console Printer-Keyboard
• Private Core Image Library (PCIL)
• IBM 3211 Printer
• Label Cylinder Display (LSERV)
• IBM 3215 Console Printer-Keyboard
• Recovery Management Support (RMS) for
System/370, which consists of Machine
Check Analysis and Recording (MCAR)
and Channel Check Handler (eCH)
• Reliability Data Extractor (RDE)
function for System/370
• RETAIN/370 for. system/370

• IBM 3420 Magnetic Tape Unit

MISCELLANEOUS CHANGES
Rewritten sections: The section on hard
waits is expanded and rewritten. The
section on COBOL ANS replaces the COBOL D
section. A glossary is included.

PROGRAM ENHANCEMENTS
This edition also documents the
enhancements to these programs:
• Error Statistics by Tape ,Volume
(ESTV)
• Environmental Recording, Editing, and
Printing (EREP)
• DOS Stand-Alone Dump Generator
(DUMPGEN)
• Problem Determination Serviceability
Aids (PDAID)
• Forced End-of-Volume for Disk macro
(FEOVD)

ORGANIZATION OF PUBLICATION
The manual has been reorganized and
reformatted.
Information in Appendixes A
- L in the previous edition has been
incorporated into the appropriate
sections in this edition. The total
publication has a General Contents, a
Figure list, and an Index. Each section
has a section Outline and a separate
figure list. Figure numbers throughout
the publication are in the form: Figure
1.3, where 1 is the section number and 3
is the figure number within that section.
A tab in the upper right-hand corner of
the beginning of each section identifies
the section by number.

DOS System Programmer's Guide

Preface

This publication is divided into the
following four major sections:

Data Management
DOS Data Management Concepts, GC24-3427.

supervisor Planning Concepts

Data Management

Program Design

Debugging Aids

DOS Supervisor and I/O Macros, GC24-5037.
DOS DASD Labels, GC24-5072.

The supervisor Planning Concepts section
describes system residence organization,
some supervisor tables, optional supervisor
facilities at a conceptual level, and
presents guidelines for their
implementation. The Data Management
section explains the concepts of file
organization and data manipulation at the
physical and logical IOCS levels. The
Program Design section contains suggestions
for efficient programming. The topics
discussed are link editing functions,
overlay structure, self-relocating
programs, checkpoint/restart facility, the
3211 Printer support, and macro writing.
The Debugging Aids section contains aids
for debugging problem programs written in
assembler language, PL/I, FORTRAN, COBOL,
and RPG.

Tape Labels for BPS, BOS, TOS and DOS,
GC24-5070.
IBM System/360 Principles of Operation.
GA22-6821.
IBM System/370 Principles of Operation.
GA22-7000.
Program Design
DOS System Control and Service,
GC24-5036.
DOS Supervisor and I/O Macros, GC24-5037.
IBM System/360 Disk and Tape Operating
systems, Assembler Language, GC24-3414.
Debugging Aids

Multitasking and link editing examples,
and physical IOCS, STXIT support,
self-relocating, and language translator
sample programs are included.
Closely related publications by section
follow.
Note: Although titles of some DOS
publications have been simplified, the
change does not affect the contents of the
publications.
Supervisor Planning Concepts
DOS System Generation, GC24-5033.
DOS System Control and service,
GC24-5036.
DOS Operating Guide, GC24-5022.
DOS Messages,

GC24-5074~

DOS OLTEP, GC24-5086.
IBM System/360 Disk and Tape Operating
systems, COBOL Programmer's Guide,
GC24-5025.
IBM Svstem/360 Disk and Tape Operating
systems, PL/I Programmer's Guide,
GC24-9005.
IBM System/360 Disk Operating System,
FORTRAN IV Programmer's Guide, GC28-6397.
IBM System/360 Disk and Tape Operating
systems, Report Program Generator
Specifications. GC26-3570.
IBM System/360 Disk and Tape Operating
systems, Assembler Language, GC24-3414.
DOS Messages, GC24-5074.
For further information concerning terms
referenced in this publication, see the IBM
Data Processing Glossary, GC20-1699.
Preface

DOS System Programmer's Guide

Contents

Introduction

• • • • • • • • • • • • • • • 11

Section 1: Supervisor Planning Concepts • • • 13
System Residence Organization
• • 15
SUPVR Macro • • • •
• • 18
FOPT Macro • •
• • 56
ALLOC Macro • • • •
• • 84
84
IOTAB Macro
DVCGEN Macro •
• • 87

Section 2: Data Management • • • • • • • • 91
Concepts of Data Management • • • • • • 93
Physical Input/Output Control System
(PIOCS)
• • • • • • • • • • • • • • • • 99
Logical Input/Output Control system
(LIOCS)
• • • • • • • • • • • • • • • .111

Section 3: Program Design •
Link Editing • • • • • •
Self-Relocating Programs
Linkage Editor Examples
Checkpoint/Restart • • •
IBM 3211 Printer Support
Macro Writing • • • • •

•
• .139
•
.141
•
• .149
• • • •
.154
•
• • • • 166
.173
•
•
.175

Section 4: Debugging Aids • • •
• .187
Gathering Documentation
.189
Wait States • • • • • • • •
.202
Debugging Assembler Programs
.204
Debugging COBOL Programs·.
• • • • 204
Debugging FORTRAN Programs •
.234
Debugging RPG Programs • •
• .243
Debugging PL/I Programs • • • •
.255

Glossary

• • • • • • • • • • • • • • • • 283

Index • • • • • • • • • • • • • • • • • • 289

Contents

;. OOS system Programmer's Guide

Figures

Figure 1.1. system Residence
Organization • • • • • • • • • • • • • •
Figure 1.2. MPS storage Map
•••••
Figure 1.3. supervisor Calls (Part 1
of 2)
• • • • • • • • • •
• •
Figure 1.4. Processing and I/O
Requests Relationship • • • • •
•
Figure 1.5. First Example of
Processing and I/O Requests with I/O
Interrupts
••••••••••
• •
Figure 1.6. Second Example of
Processing and I/O Requests with I/O
Interrupts
••••••••••
•
Figure 1.7. system Communications
Region (Part 1 of 5) • • • • • • • • • •
Figure 1.8 Background Communications
Region Extension (BGXTNSN)
• •
Figure 1.9. Tables for MICR DTF
Addresses and Pointers • • • • •
• •
Figure 1.10. Example of Multitasking
Priorities
••••••••••••
Figure 1.11. Event Control Block
(ECB)
• • • • • • • • •
•
Figure 1.12. Resource Control Block
(RCB)
• • • • • • • •
•
Figure 1.13. Option Tables
•
Figure 1.14. I/O Table
Interrelationship
• • • • •
Figure 1.15. Tape Error Block
••
Figure 1.16. TEBV Table Showing
status Block and Error Blocks
Figure 1.17. Console Buffering Table
and Work Areas
•••••••••
•
Figure 1.18. Example Using CCHAIN
Support
• • • • • • • • • • • • • •
Figure 1.19. Processing of STXIT
Condi ti ons
••••••••••••••
Figure 1.20. STXIT Sample Program • • •
Figure 1.21. Logical Unit Block (LUB)
Table
• • • • • • • • • • • • • •
Figure 1.22. Job Information Block
(JIB) Table
• • • • • • • • • • •
Figure 1.23. DASD File Protect Logic
Flow • • • •
• • • • • • • • • • •
Figure 1.24. Disk Information Block
(DIB) Table
• • • • • • • • • • •
Figure 1.25. Example of Dedicated
peIL in a Multiple Partition
Environment • • • • • • • • • • • • • •
Figure 1.26. Example of PCIL in a
Single Partition Environment
•••••
Figure 1.27. Identification of SYSLNK
Files
• • • • • • • • • • • • • •
Figure 1.28. Job Accounting Table
•
Figure 1.29. CHANQ, LUBID, REQID,
LUBDSP, and TKREQID Tables
•
Figure 1.30. Physical Unit Block
(PUB) Table • • • • • • • • • • •
•
Figure 1.31. Device Type Codes
•
Figure 2.1. Fixed Length Unblocked
Record Format • • • • • • • • • • • • •

17
19
21
24
25
26
30
35
37
38
41
41
57
59
60
61
64
65
67
69
71
72
74
75
78
79
80
83
85
89
90'
94

Figure 2.2. Fixed Length Blocked
Record Format • • • • • • •
• 95
Figure 2.3. Variable Length Unblocked
Record Format • • • • • • • • • • • • • 97
Figure 2.4. Variable Length Blocked
Record Format • • • • • • • • • • • • • 98
Figure 2.5. I/O Operation Using PIOCS 101
Figure 2.6. Command Control Block
(CCB) (Part 1 of 3)
• • • • • • • 104
Figure 2.7. Format of the CCW
.106
Figure 2.8. Flowchart for EXCP Coding
Example
• • • • •
•• • •
.109
Figure 2.9. Physical IOCS Sample
Program
• • • • • • • • • • •
.110
Figure 2.10. Retrieving a Record
Using Logical IOCS • • • • • • • •
.112
Figure 2.11. Generated Name Structure
for Logic Modules (Part 1 of 2)
.116
Figure 2.12. DOS Relocatable Library
Module Name Prefixes
.118
Figure 2.13. DTF and Module Macro
Relationships
• • • •
.118
Figure 2.14. Coding Example of DTF and
Module Relationship
• • • • • • • 119
Figure 2.15. Logical IOCS Imperative
Macros and DTFs • • • • • • • • • • • • 120
Figure 2.16. Logical IOCS Imperative
Macros and Devices
• • • • • '. • 121
Figure 2.17. Direct Access Address
Chaining • • • • • • • • • • • • • • • .127
Figure 2.18. DASD Address Formats • • • 128
Figure 2.19. Example of Track Index
.130
Figure 2.20. Example of Track Index
after Addition to File. • • • • •
.131
Figure 2.21. File Organization on a
Disk/Tape System • • • • • • • • • • • • 135
Figure 2.22. Indexed Sequential
Versus sequential File Organization • • 137
Figure 3.1. Linkage Editor System Flow 142
Figure 3.2. Module Dependency • • • • • 148
Figure 3.3. Overlay Tree Structure • • 148
Figure 3.4. Relocating Address
Constants in a Calling List
.151
Figure 3.5. Self-Relocating Sample
Program
• .. • • • • • • • • .153
Figure 3.6. Using Checkpoint Facility
on Disk
• • • • • • • • • • • • • • 169
Figure 3.7. Procedure for Building
Tape Repositioning Tables • • • • • • • 170
Figure 3.8. Procedure for Building
DASD Operator Verification Table
.172
Figure 3.9. Format of the Checkpoint
Header/Trailer Records
••••
.172
Figure 3.10. 3211 Error Status
Indicator Bits in the CCB • • • •
.175
Figure 3.11. Operand Field Formats • • 176
Figure 3.12. Keyword Macro
Instruction • • • • • • • • • •
.176
Figure 3.13. Positional Macro
Instruction • • '. • • • • • • •
• 1 77
Figure 3.14. Macro Instruction with
Prototype • • • • • • • • • • • • • • • 177
Figures

Figure 3.15. Mixed Macro Instruction .177
Figure 3.16. Mixed-Mode Definition • • 177
Figure 3.17. Sublist Illustration
• • 179
Figure 3.18. Format of Globals and
Locals
• • • • • • • • •
• • .180
Figure 3.19. Format of SETA
Instruction • • • • • • • • •
• .180
Figure 3.20. Format of SETB
Instruction • • • • • • • • •
• .180
Figure 3.21. Format of SETC
Instruction • • • • • • • • •
• • • 181
Figure 3.22. Concatenation and
Generated Coding
• • • • • • • • • • .182
Figure 3.23. Conditional Branch
Instruction • • • • • • • • • •
• .182
Figure 3.24. Unconditional Branch
Instruction • • • • • • • • • • • • • .183
Figure 3.25. Assembly No Operation
Instruction • • • • • • • • •
• .183
Figure 3.26. Macro Definition Exit
Instruction • • • • • • • • • •
• .184
Figure 3.27. MNOTE Instruction • • • • 184
• • .184
Figure 3.28. Sample MNOTE
Figure 3.29. Sample MSG Macro
• .185
Figure 3.30. sample MSG Coding
.186
Figure 4.1. SDR Communications Region
(Part 1 of 2) • • • • • • • • • • • • .190
Figure 4.2. Machine Check Recording
and Recovery (MCRR) Linkage Table
.192
Figure 4.3. RMS Linkage Area (RASLINK) 193

DOS ,system Programmer's Guide

Figure 4.4. First Part of Program
Information Block (PIB) Table
• • 195
Figure 4.5. PIB Flag Expansions • • • • 196
Figure 4.6. Second Part of Program
Information Block (PIB) Table • • • • • 197
Figure 4.7. Causes for Message 05041
(Cancel Code X' 21') . . . . . . . . . . 2 0 0
Figure 4.8. Low Core Error Bytes • • • 203
Figure 4.9. COBOL Sample Program
(Part 1 of 27)
............207
Figure 4.10. FORTRAN Sample Program
(Part 1 of 8)
• • • • • • • • • • • • • 235
Figure 4.11. Using RPG Pointers to DTF 243
Figure 4.12. Halt Indicator (HO)
Analysis Aid
•••••••••
.244
Figure 4.13. RPG Sample Program (Part
1 of 10)
• • • • • • • • • • • • • • • 245
Figure 4.14. PL/I Program Structure • • 255
Figure 4.15. PL/I Storage Areas
• • • 256
Figure 4.16. Entry Point Table
.257
Figure 4.17. Object Time Core Usage
.258
Figure 4.18. Library Work Space • • • • 259
Figure 4.19. Communications Area
Switches
• • • • • • • • • • • • • • • 260
Figure 4.20. Dummy DSA and DSA Layout 260
Figure 4.21. Block Description
.261
Figure 4.22. DSA Chaining • • • • • • • 262
Figure 4.23. PL/I Consecutive File
DTF-A Appendage • • • • • • • • • • • • 263
Figure 4.24. PL/I Regional File DTF-A .263
Figure 4.25. PL/I sample Program
(Part 1 of 18) • • • • • • • • • • • • .265

Introduction

As a system progranuner~ you make decisions
involving the components of your
installation's supervisor, file
organization, program design, and so forth.
From time to time, you may be called upon
to advise other programmers concerning DOS.
In order to assist you in this task, this
publication brings together and expands
upon information from many sources. It is
divided into four major sections:
supervisor Planning Concepts, Data
Management, Program Design, and Debugging
Aids.
The Supervisor Planning Concepts section
is of interest to the person(s) responsible
for tailoring the IBM-supplied Disk
Operating system to meet the needs of the
installation. This section describes in
detail optional supervisor facilities
available under DOS. Using the information
presented in this section together with the
information on main storage requirements
and implementation procedures found in the
DOS system Generation listed in the
Preface, the system progranuner can decide
whether or not to include a particular
facility within the installation's
supervisor. Guidelines for implementing
these facilities at system generation time
are also presented.
The Data Management section is of
interest to the person(s) responsible for
choosing the type of file organization best
suited for an application. This section
discusses data management concepts, the
advantages and disadvantages of each type
of file organization (sequential, direct
access and indexed sequential) and criteria
for choosing the best file organization and
retrieval method.
In addition, data manipulation is
described at both the physical and logical
IOCS levels. Detailed information for
coding at the physical IOCS level is
included. This section also defines the

macros for implementing logical IOCS and
describes the interrelationships of the DTF
and logic module generation macros.
The Program Design section is of
interest to the person(s) responsible for
program design and imp1ementation. The
Disk Operating System offers the programmer
a great deal of flexibility in the
generation of his system and in its
operation. This section discusses
effective use of the linkage editor and the
checkpoint/restart facility. In addition,
system programming techniques such as macro
coding, overlay structure and
self-relocating programs are discussed.
The IBM 3211 Printer support is also
discussed.
The Debugging Aids section is of
interest "to both the application and system
programmer. This section describes system
action on a cancel condition, gives
register conventions for following program
flow, describes the types of documentation
to be gathered for debugging purposes and
the action to be taken when a hard wait or
program loop is encountered. Problem
determination aids are briefly described.
Aids for debugging problem programs written
in assembler language, COBOL, FORTRAN, PL/I
and RPG are given.

r-----------------------------------------,I
Note:

In case of difference between thel
conventions given in this manual I
for control program functions andl
those appearing in IBM-supplied I
DOS component publications (such I
as guides for language
I
translators, sorts, utilities,
I
specifications manuals, etc).
I
observe the specific restrictions I
of the component.
I

_________________________________________ JI

Introduction

•

DOS System Programmer's Guide

Section 1:

Supervisor Planning Concepts

Section Outline
System Residence Organization • • • •
IPL (Cylinder 0, Track 0, Records
1 and 2) • • • • • • • • • • • • •
Volume Label (Cylinder 0, Track 0,
Record 3)
•
System Directory (Cylinder 0,
Track 1) • • • • • • • • •
System Work Area (Cylinder 0,
Tracks 2, 3 and 4) • • • •
Transient Directory (Cylinder 0,
Track 5) • • • • • • • • • • • • •
Open Directory (Cylinder 0, Track

• 15

• 15
• 15

15
• 15
• 15

6) • • • • • • • • • • • •

Library Routine Directory
(Cylinder 0, Track 7)
Foreground Program Directory
(Cylinder 0, Track 8)
••••
Phase Directory (Cylinder 0, Track
9) •
• • • • • • • • •
Core Image Library Directory
Core Image Library • • • • • •
Relocatable Library Directory
Relocatable Library
• • • • •
Source Statement Library Directory
Source statement Library • •
Label Information Cylinder •
SUPVR Macro
• • • •
Multiprogramming Support (MPS)
Partitioning • •
•
Control Method • • • • •
Task Selection • • • •
•
System Considerations
• • • •
Concurrent Peripheral Operation
(CPO)
••••••••••
Operational Considerations • • • •
programming Considerations • • • •
American National Standard Code for
Information Interchange (ASCII)
Support • • • • • • • • • • •
System Considerations • • • •
System Generation Guidelines • • •
Magnetic Ink Character Recognition
Support (MICR) • • • • • • •
system Considerations
Asynchronous Processing (AP)
system Considerations
Multitasking Macro Usage •
Subtask Initiation and Normal
Termination • • • • • • • • • • •
ATTACH Macro Considerations • • •
DETACH Macro Considerations
•
Resource Protection
• •
ENQ Macro Considerations • •

15
15
16
• 16
16
• 16
• 16
• 18
• 18
• 18
18
• 18
18
• 20
20
• 20
• 27
27
• 27
• 28
28
• 28
• 29
36
• 36

38
• 38
39
•
•
•
•
•

39
39
40
41
42

DEQ Macro Considerations •
• •
Intertask Communication
•
POST Macro Considerations • • • • •
Summary of Multitasking
Considerations • • •
•
Multitasking Examples
• •
ATTACH Macro Example • •
• •
DETACH Macro Example •
•
ENQ/DEQ and RCB Macros Examples • •
POST Macro Example • •
WAITM Macro Example • • • • • • • •
STXIT AB Macro Example • • • • • • •
Track Hold and Reentrant Modules
Example
• • • • • • • • • • •
FOPT Macro • • • • • • • • • • • • • • •
Operator Communications Support (OC) •
Operational Considerations •
•
Interval Timer Support (IT)
•••• •
Program Check support (PC) • •
Tape Error Recording • • • • •
•
Seek separate Support (SKSEP)
•
Physical Transient OVerlap Support
(PTO)
••••••••••••••••
Console Buffering Support (CBF)
•
Independent Directory Read-In Area
(IDRA) • • • • • • • • • • • • • • • •
System Generation Considerations • •
Command Chaining support (CCHAIN)
••
Track Hold support (TRKHLD)
•••••
supervisor Considerations • • • • •
LIOCS Considerations • • • • • • • •
Abnormal Termination Support (AB)
STXIT Macro Considerations •
•
Multiple Wait support (WAITM)
•
Multiple Wait Considerations
•
DASD File Protect Support (DASDFP) • •
Mode of Operation • • • • • • • • •
System Files On Disk Support (SYSFIL)
Considerations When Using System
Files on Disk • • • • • • •
•
Private Core Image Libraries • •
•
system considerations • • • •
•
System Generation Guidelines • • • •
Job Accounting Interface Support •
Programming Considerations •
•
System Generation Guidelines • • • •
ALLOC Macro • • • • • • •
•
System Considerations • • •
• •
IOTAB Macro • • • • • • •
• • • •
Job Information Block (JIB)
•••••
Channel Queue (CHANQ) • • • •
•
Determining Size of the CHANQ
•
DVCGEN Macro • • • • • • • • • • • • • •
Channel Rescheduling Considerations •
Section 1: supervisor Planning Concepts

42
42
42

43
45
45

41
48
51
52
53
55

56
56
56
56
58
58
62
62

63
64
64
64
65
65
66
67
67
70
70
70
70
74
75
76
76
81
81
81
82
84
84
84
84
84
84
87
87
13

Section Figures

Figure 1.1. System Residence
Organization • • • • • • • • • • • • • •
Figure 1.2. MPS storage Map
•••••
Figure 1.3. supervisor Calls (Part 1
of 2)
• • • • • • • • • • •
•
Figure 1.4. Processing and I/O
Requests Relationship • • • • • •
•
Figure 1.5. First Example of
Processing and I/O Requests with I/O
Interrupts
••••••••••
Figure 1.6. Second Example of
Processing and I/O Requests with I/O
Interrupts
••••••••••
•
Figure 1.7. System Communications
Region (Part 1 of 5) • • • • • • • • • •
Figure 1.8 Background Communications
Region Extension (BGXTNSN)
••••••
Figure 1.9. Tables for MICR DTF
Addresses and Pointers • • • • • • • • •
Figure 1.10. Example of Multitasking
Priorities
••••••••••••••
Figure 1.11. Event Control Block
(ECB)
• • • • • • • • •
Figure 1.12. Resource Control Block
(RCB)
• • • • • • • •
•
Figure 1.13. Option Tables • • • • • •
Figure 1.14. I/O Table
Interrelationship • • • • • •
• •
Figure 1.15. Tape Error Block
••
Figure 1.16. TEBV Table Showing
Status Block and Error Blocks
• • •
Figure 1.17. Console Buffering Table
and Work Areas
•••••••••
•
Figure 1.18. Example Using CCHAIN
Support
• • • • • • • • • • • • • •
Figure 1.19. Processing of STXIT
Conditions
••••••••••••••
Figure 1.20. STXIT Sample Program • • •
Figure 1.21. Logical Unit Block (LUB)
Table
• • • • • • • • • • • • • •
Figure 1.22. Job Information Block
(JIB) Table
• • • • • • • • • • •
Figure 1.23. DASO File Protect Logic
Flow • • • •
• • • • • • • • • • •
Figure 1.24. Disk Information Block
(OIB) Table
• • • • • • • • • • •
Figure 1.25. Example of Dedicated
PCIL in a Multiple Partition
Environment • • • • • • • • • •
• •
Figure 1.26. Example of PCIL in a
Single Partition Environment
•
Figure 1.27. Identification ofSYSLNK
Files
• • • • • • • • • • • • • •
Figure 1.28. Job Accounting Table • • •
Figure 1.29. CRANQ, LUBID, REQID,
LUBDSP, and TKREQID Tables
• •
Figure 1.30. Physical Unit Block
(PUB) Table • • • • • • • • • •
•
Figure 1.31. Device Type Codes
••

DOS System Programmer's Guide

17
19
21
24
25
26
30
35
37
38
41
41
57
59
60
61
64
65
67
69
71
72

74
75
78
79
80
83
85
89
90

To assist you in planning a supervisor
tailored to meet the nee9s of your
installation, this section presents the
organization of a disk resident system as
received from Program Information
Department (PID), describes optional
supervisor facilities at a conceptual
level, and presents guidelines for their
implementation. Those optional facilities
requiring special consideration are
described in the order in which they appear
as parameters within the supervisor
generation macros. For information on the
other optional supervisor facilities, see
the DOS system Generation listed in the
Preface.

System Residence Organization
Figure 1.1 describes the organization of
the DOS disk resident system. The disk
resident system may be on an IBM 2311 Disk
Storage Drive, an IBM 2314 Direct Access
Storage Facility, or an IBM 2319 Disk
Storage Facility. The organization of the
disk pack is as follows:

system Work Area (Cylinder 0, Tracks 2, 3
and 4)
This 3-track area is reserved as a work
area for the librarian programs and linkage
editor. The format of the records in the
librarian area depends on the program using
the area at a specific time.

Transient Directory (CYlinder 0, Track 5)
This single track directory contains
entries for the A- and B-transient routines
that are located in the core image library.
The entries in this directory are taken
from the core image library directory.
The core image library phases referenced
in this directory have phase names prefixed
by $$A (A-transients) or $$B
(B-transients). This directory has a
maximum capacity of 144 entries for the
2311, or 270 entries for the 2314/2319.
Track format is identical to the core image
library directory.

IPL (Cylinder 0, Track 0, Records 1 and 2)
Open Directory (Cylinder 0, Track 6)
This area contains the IPL bootstrap
program.

Volume Label (Cylinder 0, Track 0, Record

11
The volume label contains the address of
the Volume Table of Contents (VTOC)
established when the pack was initialized.

System Directory (Cylinder 0, Track 1)
This directory consists of five records
that make up the system master directory,
Records 1 through 4 are 80 bytes in length.
Record 1 contains information describing
the core image library and directory.
Records 2 and 3 contain the starting
address of the relocatable library
directory and the source statement library
directory, respectively. Record 4 is not
used. Record 5 is the IPL loader program
($$A$IPL2).

This single track directory contains
entries for the LIOCS open phases located
in the core image library. The entries in
this directory are taken from the core
image library directory. The core image
library phases referenced in this directory
have phase names prefixed by the characters
$$BO. This directory has a maximum
capacity of 144 entries for the 2311, or
270 entries for the 2314/2319.

Library Routine Directory (CYlinder 0,
Track 7)
This single track directory contains
entries for frequently used core image
library phases, such as job control,
linkage editor, and so forth. The entries
in this directory are taken from the core
image library directory. The core image
library phases that are placed in this
directory have phase names prefixed by a $
(for example, $LNKEDT). This entry has a
maximum capacity of 144 entries for the
2311, or 270 entries for the 2314/2319.
section 1: supervisor Planning Concepts

Foreground Program Directory (Cylinder 0,
Track 8)

• Phase name
• Loading address

This single track directory contains
entries for the foreground program phases
located in the core image library. The
entries in this directory are taken from
the core image library directory. The core
image library phases referenced in this
directory have phase names prefixed by the
characters FGP. This directory has a
maximum capacity of 144 entries for the
2311, or 270 entries for the 2314/2319.

• Number of blocks
• Entry point
• Starting disk address in the core image
library
• Length of last block

Core Image Library
Phase Directory (cylinder 0, Track 9)
This single track directory contains
entries for the phases of the current BG
problem program. The entries in this
directory are constructed by job control
before each job step is executed in the
background partition only. They are taken
from the core image library directory.
The phase naming conventions that permit
the use of the phase directory are:
1.

All program names must be unique in
the first four characters.

The first four characters of the name
of each phase of a program must be
identical to the first four characters
of the program name. All eight
characters of the first phase name
must be identical to the program name.
Example: WXVZPROG
WXVZPROG - phase 1
WXVZPH1 - phase 2
WXVZPH2 - phase 3

The maximum capacity of this directory is
144 entries for the 2311 or 270 entries for
the 2314/2319.

The core image library consists of five or
more tracks, depending on the allocation
specified by the user. For the 2311, each
track contains two blocks with a maximum
capacity of 1728 bytes. For the 2314/2319',
each track contains four blocks with a
maximum capacity of 1688 bytes~ The number
of programs (phases) and the size of each
program to be contained in the core image
library dictates the number of cylinders
that must be allocated. Each program
starts with a new block and only the last
block of a program can contain less than
1728 bytes of data for the 2311 or 1688
bytes of data for the 2314/2319.
Note: A phase is an overlay of a
mUltiphase program or a complete program if
not multiphase.

Relocatable Library Directory
This directory consists of one or more
tracks, depending on the allocation
specified by the user. It contains two
types of information:
1.

System directory information for the
relocatable directory and library.
This information occupies the first
five entries of the first record in
the relocatable directory.

An entry that describes each module in

Core Image Library Directory
This directory consists of one or more
tracks, depending on the allocation
specified by the user. It contains one
entry for each of the phases in the core
image library.

the relocatable library and contains:
a.

Module name

b.
Note: A phase is an overlay of a
multiphase program or a complete program if
not multiphase.

Total number of text-record blocks
required to contain this module

Starting disk address of the first
text-record of this module

Each directory entry describes one phase
in the core image library and contains:

Change level identification.

DOS System Programmer's Guide

STARTING DISK ADDRESS
COMPONENT

NO.

IPL Bootstrap Record 1 ($A$IPLl)

IPL Bootstrap Record 2 ($A$IPLA)

NUMBER
OF TRACKS
(A Ilocat ion)

R=REQUIRED
O=OPTIONAL
R
R

1
Volume Lobel

User Volume Label

00
00
00
00

01
01
01
01

1
2
3
4

R
R
R
R

IPL Retrieval Program ($$A$IPL2)

System Work Area (Librarian Area)

Transient Directory ($$A and $SB Transients)

Open Directory (SSBO)

Library Routine Directory ($ Phasenames)

Foreground Program Directory (FGP)

Phose Directory (For Problem Program Phases)

Core Image Library Directory

Core Image Library

End of CI Directory
Y +1
X

Relocatable Library Directory

End of CI Library
Z+l
00

Relocatable Library

End of RL Directory
Y+l
X

Source Statement Libra ry Directory

End of RL Library
00
Z+l

Source Statement Library

End of SS Directory
Y+1
X

Volume Area File Definition Storage Area

End of SS Library
00
Z+l

User Area

End of Volume Area
00
Z+2

System
Directory
2

Record
Record
Record
Record

1
2
3
4

01 for 2311 00for2311
00 for 2314/ 1a for 2314/
2319
2319

2311:10
2314/2319:20
*

*Allocation Dependent On User Requirements
X=Ending CC of the Preceding Directory
Y = Ending HH of the Preceding Directory
Z = Ending CC of the Preceding Library

Figure 1.1.

System Residence Organization

section 1: Supervisor Planning Concepts

Relocatable Library
The relocatable library consists of five or
more tracks, depending on the allocation
specified by the user. The number of
modules and the size of each module to be
contained in this library dictates the
number of tracks that must be allocated.
Each allocated track contains 9 blocks
(2311) or 16 blocks (2314/2319), and each
block has a fixed length of 322 bytes.
Each module starts with a new block but not
necessarily a new track.

'(This directory consists of one or more
/ tracks, depending on the allocation
specified by the user. It contains two
types of information:

SUPVR Macro

System directory information for the
source statement directory and
library. This information occupies
the first five entries of the first
record in the source statement
directory.
An entry that describes each book (see
Note 1) in the source statement
library and contains:
a.

A sublibrary prefix: any
alphameric character, $, i, or
except A and C (see Note 2).

Book name.

Starting disk address of the first
block of this book.

Total number of blocks required to
contain this book in the source
statement library.

Change level identification.

Note 1: A book is a sequence of source
language statements, in compressed card
image format, accessed by a single name.
Note 2: A and C are reserved for assembler
and COBOL, respectively.

Source Statement Library
The source statement library consists of
five or more tracks, depending on the
allocation specified by the user. The
number of books and the size of each book
to be contained in this library dictates
18

DOS System Programmer's Guide

Label Information Cylinder
The label information cylinder (10 tracks
for 2311 or 20 tracks for 2314/2319)
contains background and foreground user and
standard label information.

Source Statement Library Directory

the number of tracks that must be
allocated. Each track contains 16 blocks
(2311) or 27 blocks (2314/2319), and each
block has a fixed length of 160 bytes.
Each book starts with a new block but not
necessarily a new track. Each book in the
source statement library contains
compressed card images of the source
language input to the assembler or language
translators. A compressed card image can
overflow from one block to another.

MULTIPROGRAMMING SUPPORT (MPS)
Multiprogramming is the ability to run
multiple programs concurrently, provision
for which must be included in the DOS
supervisor at system generation time. Each
program resides in a different area of main
storage called a partition. The three
problem program partitions are designated
background (BG), foreground 1 (Fl), and
foreground 2 (F2).
The background partition must be at
least 10K because job control runs in the
background partition and requires 10K bytes
of main storage. However, 14K allows
faster assemblies and linkage editing.
The remaining main storage is divided
between the two foreground partitions. To
satisfy the requirements for the storage
protect special feature, these partitions
must begin and end on 2K boundaries.
Because the MPS supervisor requires a
minimum of 8K and the background partition
requires a minimum of 10K, MPS will not
function on systems with less than 24K of
main storage (see Figure 1.2).
MPS operates under the principle that in
most commercial installations, the CPU is
heavily I/O bound. Much of the CPU running
time is spent waiting for a printer, a
reader or a punch to complete a previous
operation before the subsequent one can be
started. With MPS, when a partition
becomes I/O bound (that is, it cannot
continue until the completion of some I/O
operation), a task selection routine in the
MPS supervisor attempts to give CPU control
to the next partition that is ready to run.

Supervisor Nucleus

I/O

Tables and Information Blocks ,.--------.,

80 Bytes (112 with Floating Point)

Background Program Area
10K Minimum

F2 Save Area
88 Bytes
(120 with Floating Point)
Foreground 2 Program Area
Multiple of 2K

F1 Save Area@- 88 Bytes
(120 with Floating POint)
Foreground 1 Program Area
Multiple of 2K

General Registers 9 through 8
64 Bytes

Figure 1.2.

General Registers 9 through 8
64 Bytes

MPS Storage Map
Section 1: Supervisor Planning Concepts

The partition to which control is given
is determined by a priority system. Fl has
highest priority followed by F2 and BG. A
partition cannot be interrupted by one of
lower priority. When an Fl program is
ready to run, the task selection routine
seizes control from a program of lower
priority. One of the disadvantages in such
an arrangement is the possibility of a high
priority program never relinquishing
control to other partitions. This can
happen if the high priority program has few
I/O requests. In general, the program with
the largest number of I/O operations or
wait time should reside in the highest
priority partition. A compute (CPU) bound
program should reside in a low priority
partition (for example, BG).
Because the channel scheduler attempts
to keep devices and channels busy
asynchronously, it is an advantage for all
programs to use SUfficient I/O G~eas to
keep the channel queue stacked with
requests. A lower priority partition is
more likely to require multiple I/O areas
to achieve this than is a higher priority
partition.

Partitioning
Multiprogramming has been defined as a
technique whereby two or more programs may
operate concurrently, sharing system
resources between them. The DOS
multiprogramming support requires that such
programs be co-resident in main storage.
To achieve this, storage is divided into 3
sections termed partitions (BG, Fl, and
F2). Each partition is capable of holding
a distinct problem program. When a
partition contains a program that is in
course of execution, the partition is said
to be active. When a partition does not
contain such a program, or is not
physically present in the system, then it
is said to be inactive.
Each partition is of fixed physical size
and is defined by fixed boundaries. For
this reason, DOS multiprogramming is termed

DOS System Programmer's Guide

fixed partition multiprogramming,.
Partition boundaries may be altered in any
manner when all partitions are inactive
(minimum background size is 10K). When any
partitions are active. boundaries can be
altered only if the lower limit of each
active partition is unchanged and the size
of each active partition is not reduced.
Each partition is allotted its own
unique storage protection key; for this
reason the size of any partition must be an
integral multiple of 2K. Thus, a program
contained in anyone partition can read but
not alter the contents of core locations
contained in any other partition.

Control Method
Multiprogramming increases system
efficiency by making better use of the
available system resources than is possible
in a single partition batched job
environment.
Multiprogramming support is written in
such a way that the central processing unit
does not enter the wait state if useful
processing can be performed in any
partition. Multiprogramming allows the
input and output functions of one program
to overlap with the processing functions of
other programs.

Task Selection
A program yields control by issuing a
supervisor call instruction to pass control
to the supervisor routines. The SVC
instruction contains a code that indicates
its purpose. The most numerically
significant of these codes in a
multiprogramming context are those
associated with input and output
operations; therefore, only these codes are
described. A complete list of DOS
supervisor calls is given in Figure 1.3.

SVC
Function

Macro Supported
Dec.

Hex.

EXCP

Execute channel programs.

FETCH

1
2
3

Fetch any phase.
Fetch a logical transient (B-transient).
Fetch or return from a physical transient (A - transient).

LOAD

Load any phase.

MVCOM

Modify supervisor communications region.

CANCEL

Cancel a problem program or task.

WAIT

Wait for a CCB or TECB.

Transfer control to the problem program from a logical transient (B - transient.)

Return to a logical transient (B - transient) from the problem program after an
SVC 8.

10*

Set timer interval.

11
12

B
C

Return from a logical transient (B - transient).
Logical AND (Reset) to second job control byte (displacement 57 in communications region).
Logical OR (Set) to second job control byte (displacement 57 in communications
region) .

Cancel job and go to job control for end of job step.

Same as SVC 0 except ignored if CHANQ table is full.

STXIT (PC)

16*

Provide supervisor with I inkage to userls PC routine for program check interrupts.

EXIT (PC)

17*

Return from userls PC routine.

STXIT (IT)

18*

Provide supervisor with Iinkage to userls IT routine for interval timer interrupts.

EXIT (IT)

19*

Return from userls IT routine.

STXIT {OC}

20*

Provide supervisor with Iinkage to userls OC routine for external or attention
interrupts (operator communications).

EXIT {OC}

21*

Return from userls OC routine.

22*

23*

The first SVC 22 seizes the system for the issuing program by disabling multiprogram operation. The second SVC 22 releases the system {enables multi program
operat ion} •
Load phase header. Phase load address is stored at userls address.

24*

Provide supervisor with linkage to user·ls TECB and set timer interval.

25*

26*
27*

1A
1B

Issue HALT I/O on a teleprocessing device, or HALT I/O on any device if issued
by OLTEP.
Val idate address limits.
Specia I H 10 on teleprocessing devices.

LBRET

SETIME

EOJ

SETIME

(Primari Iy used by ERP.)

* = optional
Figure 1.3.

Supervisor Calls (Part 1 of 2)

section 1: supervisor Planning Concepts

SVC
Macro Supported

Hex.

28*

Return from user's stacker select routine (MICR type devices only).

29*

Provide return from multiple wait macros WAITF and WAITM (except MICR type
devices) •

QWAIT

30*

Wait for a QTAM element.

QPOST

31*

Post a QTAM element.

32
33
34

20
21
22

(Reserved)
Reserved for internal macro COMRG •
Reserved for internal macro G ETIME .

HOLD

35*

Hold a track for use by the requesting task onl y.

FREE

36*

Free a track held by the task issuing the FREE.

STXIT (AB)

37*

Provide supervisor with I inkage to user's AB routine for abnormal termination of
a task.

ATTACH

38*

Initialize a subtask and establish its priority.

DETACH

39*

Perform normal termination of a subtask. It includes calling the FREE routine to
free any tracks held by the subtask.

POST

40*

Inform the system of the termination of an event and ready any waiting tasks.

DEQ

' 41*

Inform the system that a previously enqueued resource is now avai lable.

ENQ

42*

Prevent tasks from simultaneous manipulation of a shared data area (resource).

43*
44*
45*
46*
47*
48
49
50
51*

2B
2C
20
2E
2F
30
31
32
33

Provide supervisor support for external creation and updating of SDR records.
Provide supervisor support for external creation of OBR records.
Provide emulator interface.
Provide OlTEP with the facility to operate in supervisory state.
Provide return from wait multiple WAITF for MICR type device.
(Reserved)
(Reserved)
Reserved for LlOCS error recovery.
Return phase length at OlTEP request.

EXIT (MR)

= optional

Figure 1.3.

Function

Dec.

Supervisor Calls (Part 2 of 2)

DOS system Programmer's Guide

SVC Code 0 (EXCP)
This code requests the supervisor to
initiate an input or output operation. The
address of a Command Control Block (CCB)
located in the requesting program is also
passed to the supervisor via register 1.
This block contains information that
describes the precise nature of the
operation to be performed.
When the supervisor receives the EXCP
request but is unable to initiate the
required operation, it places the request
in a queue for later action. The traffic
bit in the relevant CCB is then set 0 to
whether or not an operation can be started
when requested. The supervisor then
returns control to the p~ogram that
requested the operation.
SVC Code 7 (WAIT)
This code informs the supervisor that the
program is unable to proceed further until
a previously requested operation has been
completed, and that the operation is still
in progress because the relevant traffic
bit is still set to O. The program is
placed in the wait state. Note that the
instructions immediately preceding the
supervisor call instruction form a test of
the traffic bit so that, if the traffic bit
is set to 1, the supervisor call is
bypassed. When the SVC 7 is recognized,
the program return address is reduced so
that the instructions generated by the WAIT
macro will be issued.

When the supervisor recognizes an SVC 7
interrupt, it records that the program is
in the wait state. The supervisor then
gives control to the next partition of the
highest priority that is ready to run. If
such a partition does not exist, the
supervisor causes the central processing
unit to enter the wait state until an
interrupt occurs.
A program loses control to the
supervisor whenever an interrupt occurs.
Only the input/output interrupts are
described in the following text because
they are the most significant interrupts in
a multiprogramming environment.

When an input/output interrupt occurs,
the supervisor identifies the operation in
question and checks whether it has been
satisfactorily performed. If it has, the
traffic bit in the related CCB is set to 1;
the owning partition is made ready to run,
that is, if it was in a wait state it is
removed from that state; and any further
input/output operation pending for this
channel or device is initiated. The task
selection routine in the supervisor then
gives control to the next partition of the
highest system priority that is ready to
run.
The following three figures show the
relationship between six different
operations while operating under MPS. All
three partitions are active.
Figure 1.4 assumes that programs were
initiated at some prior point in time in
all three partitions. The first part of
Figure 1.4 shows that foreground 1 has
control of the CPU and is processing. At
some point in time, the program in F1
returns control of the CPU to the
supervisor by issuing an I/O wait (SVC 7).
The supervisor goes through a task
selection process, determines that
foreground 2 is ready to run and gives CPU
control to it. F2 processes its program
until it requires I/O, and then returns
control of the CPU to the supervisor by
issuing an I/O wait (SVC 7).
The supervisor goes through a task
selection process, determines that F1 is
not ready to run but the background is
ready to run, and gives CPU control to the
background. BG processes its program until
it requires I/O and then returns control of
the CPU to the supervisor by issuing an I/O
wait (SVC 7).
The supervisor goes through a task
selection process, determines that no
partitions are ready to run, and gives
control of the CPU to a task called All
Bound. This task loads a PSW that puts the
system in the wait state with all
interrupts enabled. Note that no I/O
interrupts have occurred.

section 1: Supervisor Planning Concepts

.Foreground 1 (Fl)

Processing

Waiting for I/o ~ompletion

~
I/O Wait* :
I

I
I

Foreground 2 (F2)

Waiting for I/O Completion

Background (BG)
letion
I

I/O Wait* I
I

Task
ISelectio

Supervisor

Task
Selection

5
Active State

NOTES

Inactive State

* I/O Wait means an SVC 7 is issued (WAIT Macro)

Figure 1.4.

All
Bound

System in Wait State

I/O Interrupts

Task
Selection

Processing and I/O Requests Relationship

DOS system Programmer's Guide

Foreground 1 (Fl)

Waiting for I/O

I
I

Foreground 2 (F2)

Waiting for I/O

Background (BG)

Waiting for I/O

Supervisor

Waiting to be Selected

I
I

Waiting to be Selected

I
Posting Traffic Bit and Performing Task Selection
I
I
I
I
I
I
I
I
I
I
I
I
I
I
Fl
BG
F2

J
All Bound
System in Wait State

I/O Interrupts

I/O INTERRUPT SEQUENCE
1.
2.
3.

Figure 1.S.

Flls I/O Interrupt Occurs
BG IS I/O I nterrupt Occurs
F2 1s I/O Interrupt Occurs

First Example of Processing and I/O Requests with I/O Interrupts

Figure 1.5 starts with all partitions
waiting for I/O, and the system in the wait
state with interrupts enabled.
When the first I/O interrupt occurs, the
supervisor I/O interrupt routine gets
control of the CPU, and turns on the
traffic bit in theCCB associated with the
device causing the I/O interrupt (Fl
partition). The supervisor task selection
routine then gets control, determines that
Fl is through waiting, and gives CPU
control to it. Fl starts processing its
program.
An I/O interrupt occurs and control
passes to the supervisor I/O interrupt
routine. It turns on the traffic bit in
the CCB associated with the device causing
the I/O interrupt (BG partition). The
supervisor task selection routine gets
control, determines that Fl is still the

highest priority program ready to run, and
gives CPU control to it. Note that BG is
still waiting, but now is ready to run
because its I/O wait is complete.
While Fl is processing, another I/O
interrupt occurs and control passes to the
supervisor I/O interrupt routine. It turns
on the traffic bit in the CCB associated
with the device causing the I/O interrupt
(F2 partition). The supervisor task
selection routine then gets control of the
CPU, determines that Fl is still the
highest priority program that is ready to
run, and gives CPU control to it. Note
that both BG and F2 are now ready to run,
but control of the CPU is retained by the
highest priority partition (Fl) until it
issues an SVC or until one of the four
other system interrupts occurs.

Section 1: Supervisor Planning Concepts

2
I

I
I

I
Waiting for I/O

Foreground 1 (F 1 )

Processing

I
I
I
I

Waiting for I/O

Foreground 2 (F2)

Waiting to be Reselected

I
Waiting to be Reselected

I
Waiting for I/O

Background (BG)

Supervisor

\ lot

Posting Traffic Bit and Performing Task Selection
I
I
System in Wa it State

All Bound

I
I

I/O Interrupt

Jt-------'~~----I\~I\----------i5
I/O INTERRUPT SEQUENCE

1. BG IS I/O Interrupt Occurs
2. F2 1s I/O Interrupt Occurs
3. Flls I/O Interrupt Occurs

Figure 1.6.

Second Example of Processing and I/O Requests with I/O Interrupts

Figures 1.4 and 1.5 show a simplified
version of what actually occurs. In actual
practice, the process time for F1 would be
longer than F2 or BG, and the sequence of
I/O interrupts would probably be staggered.
Figure 1.6 starts with all partitions
waiting for I/O, and the system in the wait
state with interrupts enabled.
When the first I/O interrupt occurs, the
supervisor I/O interrupt routine gets
control of the CPU, and turns on the
traffic bit in the CCB associated with the
device causing the I/O interrupt (BG
partition). The supervisor task selection
routine gets control, determines that the
BG partition is the highest priority task
ready to run, and gives CPU control to it.
The BG partition starts processing with the
instruction 6 bytes before the I/O wait
(SVC 7).
26

DOS system Programmer's Guide

When the second I/O interrupt occurs,
the supervisor I/O interrupt routine gets
control of the CPU, and turns on the
traffic bit in the CCB associated with the
device causing the I/O interrupt (F2
partition). The supervisor task selection
routine gets control of the CPU, determines
that the F2 partition is now the highest
priority partition ready to run, and gives
CPU control to it. The F2 partition starts
processing with the instruction 6 bytes
before the I/O wait (SVC 7).
When the third I/O interrupt occurs, the
supervisor I/O interrupt routines gets
control of the CPU, and turns on the
traffic bit in the CCB associated with the
device causing the I/O interrupt (F1
partition). The supervisor task selection
routine gets control of the CPU, determines
that the F1 partition is now the highest

priority partition ready to run, and gives
CPU control to it.
The foreground 1 partition remains in
control of the CPU until it issues an SVC
or until one of the four other system
interrupts occurs.

system Considerations
Any program that is process-bound
(compute-bound) completely suppresses any
event from taking place in any partition of
lower priority, regardless of the
utilization of channels and devices. Thus,
if the program in the foreground 1
partition is process-bound,
multiprogrammdng cannot take place because
a partition of lower priority cannot get
control. In general, a foreground program
should never be process-bound in a
multiprogramming environment.
Because the slowest devices are those
that are associated with unit recordequipment (e.g., card readers, card
punches, line printers and terminals), unit
record and teleprocessing devices in a
multiprogramming environment should be
associated with the partitions of highest
system priority.
In general, the slowest input/output
devices should run with maximum efficiency,
because loss of efficiency on such devices
is more serious in terms of system
throughput than on faster devices.
Note: I/O bound programs with the slowest
system devices should be allocated to the
highest priority partitions.
The efficiency of a low priority program
depends on the quantity and frequency of
processing time made available to it. The
most advantageous high priority program (in
terms of overall system efficiency)
combines lengthy input and output
operations with a minimum of processing.
Every time an input/output interrupt
occurs, it must be interrogated for all
partitions of higher priority than the
partition to which it relates. Therefore,
a reduction in the number of such
interrupts (which may be achieved by
increasing blocking factors of tape and
disk files), promotes greater
multiprogramming efficiency because it
lowers system overhead. Increased blocking
factors, however, result in increased core
storage requirements.

Another important system consideration
is the use of two input or output areas in
connection with one file (double
buffering>. Double buffering increases
computer utilization by allowing the
overlap of input/output operations with
processing. Multiprogramming has the same
purpose. In double buffering, input/output
operations are overlapped with processing
relating to the same program; in
multiprog~amming, input/output operations
are overlapped with processing relating to
a different program. Therefore, the more
efficiently double buffering operates in
relation to a given foreground program, the
less time will be freed for use by the
background program.
In a multiprogramming environment, it
may prove more efficient in terms of total
system throughput to dispense with double
buffering for programs operating in
foreground partitions. In most cases, main
storage can be better applied to the
accommodation of greater block sizes than
to the accommodation of double input/output
areas.

Concurrent Peripheral Operation (CPO)
The CPO (Concurrent Peripheral Operation)
concept is the application of
multiprogramming techniques to the basic
utility operations of card-to-tape,
tape-to-punch, tape-to-printer, and their
disk equivalents. Such operations are
performed in a separate storage partition.
A typical CPO program is a file-to-file
utility containing a minimum amount of
processing. Blocking and deblocking
operations and some data selection can be
performed, but this is all. Thus, CPO fits
the specifications for an efficient high
priority program. A CPO program is
normally associated with a unit record
device and the greater part of its
input/output time is likely to utilize
subchannels of the multiplexor channel.
Therefore, channel contention with other
concurrent programs is greatly reduced.
CPO is one of the most efficient practical
applications of the multiprogramming
technique.

Operational Considerations
Two methods used to initialize and operate
programs in the foreground partitions are:
1.

Batch Job Foreground (BJF): This
method uses the job control program.

Section 1: Supervisor Planning Concepts

The foreground partition(s)
essentially operates like the
background partition (i.e., batched
processing automatic job-to-job
transition). The foreground
partition(s) must be a minimum of 10K.
2.

Single Program Initiator (SPI): This
method uses the IBM 1052
Printer-Keyboard for System/360 or the
IBM 3210 or 3215 Console
Printer-Keyboards for System/370 to
initialize single programs in the
foreground partition(s). When the
single program is terminated (either
under its own control by issuing an
EOJ, DUMP or CANCEL macro, or through
operator action, a program error or
certain I/O failures), a terminating
message is printed on the console and
the foreground partition becomes
inactive. To run the next single
program in the foreground partition,
the operator must again initialize it
from the 1052, 3210, or 3215 console.
The major advantage of SPI is that the
foreground partition size can be a
minimum of 2K bytes.

Programming Considerations
The output of the language compilers can be
link-edited to run in any partition,
provided enough core storage is available.
In a system that supports both
batched-job foreground and private core
image libraries (see Private Core Image
Libraries), the function of compiling can
be run in the foreground partitions as well
as the background partition, providing
enough core storage is available in which
to execute the compiler. The linkage
editor can execute in any batched-job
partition. A private core image library is
required when executing the linkage editor
in a foreground partition.

system Generation Guidelines
The multiprogramming facility is specified
at system generation time by the
MPS= keyword parameter in the SUPVR macro.

MPS={~~}
.BJFJ
NO

This option automatically creates
a background partition consisting
of all available main storage
above the supervisor and negates
DOS system Programmer's Guide

any and all multiple partition
operation. NO is the assumed
value.
YES

When yes is specified, one, two,
or three problem program
partitions may exist. The
foreground partitions may only be
initialized using the single
program initiator (SPI).

BJF

Under this option, one, two, or
three problem program partitions
may exist and the foreground
partition(s) may be controlled
either by job control or the
single program initiator (SPI).

The parameter YES or BJF must be
specified if the supervisor is to be
generated to support QTAM.

AMERICAN NATIONAL STANDARD CODE FOR
INFORMATION INTERCHANGE (ASCII) SUPPORT
In addition to processing EBCDIC data
files, DOS can process magnetic tape files
written in ASCII (American National
Standard Code for Information Interchange),
a 128-character, 7-bit code. The
high-order bit in the System/360 8-bit
environment is zero. ASCII tape files may
be either unlabeled or labeled according to
the specifications of the American National
Standards Institute, Inc. (ANSI).

System Considerations
ASCII tape files may be processed in either
a foreground or background partition.
Because internal processing of ASCII files
is performed in EBCDIC, the data is
translated at I/O time. Input files
containing ASCII data are translated to
EBCDIC as soon as the record is read into
the I/O area. Output files described as
ASCII are translated from EBCDIC to ASCII
just prior to writing the record.
Two translate tables (providing for the
conversion from ASCII to EBCDIC and from
EBCDIC to ASCII) are generated at system
generation time by specifying the ASCII=YES
parameter in the SUPVR macro. These tables
are located immediately before the Seek
Address Blocks in the supervisor. The
address of the ASCII-EBCDIC translate table
is in locations 44-47 (decimal) of the
communications region extension. Add 256
(decimal) to this table to get the address
of the EBCDIC-ASCII translate table. The
address of the communications region

extension is in bytes 136-139 (decimal) of
the communications region. Figures 1.7 and
1.8 show the system and background
communications regions.

When ASCII=NO is specified, or if
the parameter is omitted or
incorrectly specified, the
translate tables are not
generated. The system will then
process EBCDIC files only.

system Generation Guidelines
YES
To generate supervisor support for handling
ASCII tape files, ASCII=YES must be
specified in the SUPVR macro.

ASCII={~~S}

When YES is specified, the two
translate tables are generated in
the supervisor. The address of
the first table (ASCII-EBCDIC) is
inserted in each communication
region extension to enable
accessibility by any task in any
partition.

Section 1: Supervisor Planning Concepts

COMREG*
Displacement
hexadecimal
Displacement
decimal

UPSI
Byte

Job Name

36
40
Address of
End Address
Uppermost
of Last Phase
Byte of Phase
Fetched or
with Highest
Loaded
Ending Address

Problem Program Use

32
Highest
Storage
Address
of the
Partition

XXXXXXXX

XXX X

Address of Address of
PPBEG
EOSSP

Date

XXXXXXXX
D i spl acemen t
hexadecimal
Displacement
decimal

XXXXXXXXXXX

XXXX

r46

PIK
(PI D)

End of
Storage
Address

Machine
Confg.
Byte

System
Confg.
Byte

XXXX

Standard
Language
Translator
I/O
Options

Dump,
Log and
ASCII
Options

Job
Control
Byte

Linkage Language Job
Control Translator Duration
Control
Indicator
Byte
Byte
Byte
X

Label
Area
Length
XX

Disk
Address
Address of of
Label
FOCL
Cylinder

...
Job Control Switches

Displacement
hexadecimal
Displacement
decimal

Address
of
LUB

Line
Count
for
SYSLST

System Date

LlOCS
C6mm.
Bytes

Address of
1st Part of
PIB Table

ID Number
of Last
Checkpoint

XXXXXXXXX

Displacement
hexadecimal
Displacement
decimal

Address
of
PUB

Address
of
FAVP

Address Address Address Address
of
of
of
of
TEB
FICL
NICL
JIB

100

102

104

106

108

110

Length of LUB
ID Queue =
No. of Channel
Queue Entries

Address of
Disk
Information
Block (DIB)

Address of
PC Option
Table less
8 bytes

Address of
Error
Recovery
Block

Address of
IT Option
Table less
8 bytes

Address of
OC Option
Table less
8 bytes

Key of
Program
with Timer
Support

Address of
the LUBID
Queue

112

·124

126

128

132

134

Address of
BG Comm.
Region

Option
Indicator

System
Config·
uration
Byte 2

Pointer to
Comm.
Region
Extension

XXXX

Supervisor Constants

Address of
2nd
Part of
PIB Table

Address of
MICR DTF
Table
(PDTABB)

Address of
QTAM
Vector
Table

XXXXXXXXXXXX

XXXX

Displacement values illustrated can be used to access the listing and/or the key that follows the figure.
The key offers more detailed information about each area when necessary.

* The address of the communications region is in fixed location X'14' - X'17'.

Figure 1.7.

XX
Displacement
hexadecimal
Displacement
decimal

System Communications Region (Part 1 of 5)

DOS System Programmer's Guide

135

136

Logical
Transient
Key
XX

Key to Communications Region Displacements:
MM/DD/YY or DD/MM/YY obtained from the job control date statement. Format controlled by COMREG + 53
(System Configuration Byte, date convention bit 0).
Address of the problem program area.
Address of the beginning of the problem program area. Y (EOSSP)=Y (PPBEG) if the storage protection option has not
been selected. Y (EOSSP) equals the first main storage location with a storage protection key of 1, if storage protection
is supported.
User area. If seek separation option is specified, bytes 12 and 13 are used at IPL time for the address of the seek
address block.
User program switch indicator.
Job name set by the job control program from information found in the job statement.
Address of the uppermost byte of the problem program area as determined by the IPL program (Clear storage routine
determines the address, ENDRD routine of $$A$IPL2 stores it.), or the address of the uppermost byte of the partition
as determined during processing of the ALLOC statement.
Address of the uppermost byte of the last phase of the problem program fetched or loaded. The initial va lue (as shown)
is overlaid by the first fetch or load to the problem program area.
Highest ending main-storage address of the phase among all the phases having the same first four characters as the
operand on the EXEC statement. For the background partition only, job control builds a phase directory of these
phases. The address value may be incorrect if the program loads any of these phases above its Iink-edited origin
address. If the EXEC statement has no operand, job control places in this location the ending address of the program
just link-edited.
Length of the problem program label area.
Program Interrupt Key - PIK (if asynchronous processing is not supported): Value is equal to the displacement from the
start of the PIB table to the PI B for the task.
OR
Partition Identifier - PID (if asynchronous processing is supported): Value is hex 10, 20, or 30 to identify the partition
in which a maintask or a subtask is running. (See the communications region extension, displacement 18, for the PIK
in an asynchronous processing supervisor.)
First byte - always zero.
Second byte - contains the key of the program that was last enabled for interrupts, or the partition identifier in
an AP supervisor.
Task
*AII Bound
BG
*F2
*F1
Attn Rtn
Quiesce I/O
Supervisor

PIK (PID) Value
X'OO'
X'lO'
X'20'
X'30'
X'40'
X'50'
X'60'

*These tasks do not exist in a non - MPS supervisor.

Logical end of main storage address.

Figure 1.7.

system Communications Region (Part 2 of 5)

)
section 1: supervisor Planning Concepts

Key to Communications Region Displacements:
Machine Configuration Byte (Values set at supervisor generation time.)
Bit 0:

2:
3:
4:
5:
6:
7:

1 = Storage protect feature

o = No storage protect feature
1 = Decimal feature
No decimal feature
1 = Floating-point feature
o = No floating-point feature
1 = Physical transient overlap option
o = No physical transient overlap option
1 = Timer feature
o = No timer feature
1 = Channel switching device
0= No channel switching device
1 = Burst mode on multiplex channel support
0= No burst mode on multiplex channel support
Reserved

System Configuration Byte
Bit 0:

1:
2:

3:
4:
5:
6:
7:

6: ~~~~~~}

(Date convention bit set at generation time by STDJC)

1 = Multiprogramming environment
= Batch job environment
1 = DASD file-protect supported
o = No file-protect support for DASD
1 = DASD SYSIN - SYSOUT
0= No DASD SYSIN - SYSOUT
1 = Teleprocessing
o = No teleprocessing
1 = Batch job in foreground
0= No BJF
1 = Asynchronous processing
0= No AP
1 = Track Hold
o = No Track Hold

This byte contains the standard language translator I/O options (set by the STDJC macro).

Bit 0:
1:
2:
3:
4:
5:
6:
7:

DECK option
LIST option
LlSTX option
SYM option
XREF option
ERRS option
CHARSET option
Reserved

1 = yes, output object modules on SYSPCH
1 = yes, output source module listings and diagnostics on SYSLST
1 = yes, output hexadecimal object module listings on SYSLST (compilers only)
1 = yes, output symbol tables on SYSLST/SYSPCH
1 = yes, output symbolic cross reference list on SYSLST
1 = yes, output diagnostics on SYSLST (compilers only)
1 = 48, input on SYSIPT is 48 or 60 character set

This byte contains the standard supervisor options for abnormal EOJ and control statement display, and the indicator
for the presence of the ASCII-EBCDIC and EBCD IC-ASCII translation tables.

Bit 0:
1:
2:
3:
4-6:
7:

Figure 1.7.

Always on
DUMP option
Reserved
LOG option
Reserved
ASCII option

= yes,

dump registers and storage on SYSLST

= yes, I ist all control statements on SYSLST

= yes,

ASCII supported

system Communications Region (Part 3 of 5)

DOS System Programmer's Guide

Key to Communications Region Displacement:
Job control byte
Bit 0:

1:
2:
3:
4:
5:
6:
7:

1 = Job Accounting
Interface (JA) not supported
0= Job Accounting
Interface (JA) is supported
1 = Return to caller on LlOCS disk open failure
0= Do not return to caller on LlOCS disk open failure
1 = Job control input from SYSRDR
0= Job control input from SYSLOG
1 = Job control output on SYSLOG
0= Job control output not on SYSLOG
1 = Cancel job
o = Do not cancel job
1 = Pause at end-of-job step
0= No pause at end-of-job step
1 = SYSLOG is not a 1052
o = SYSLOG is a 1052
1 = SYSLOG is assigned to the same device as SYSLST
0= SYSLOG is not assigned to the same device as SYSLST

linkage control byte
Bit 0:
1:
2:
3:
4:
5:
6:
7:

1 = SYSLN K open for output

o = SYSLN K not open for output
1 = $ or FG program phase deleted, renamed, or cataloged (flag bit for $MAINEOJ)
1 = Allow EXEC
o = Suppress EXEC
1 = Catalog linkage editor output
0= Do not catalog linkage editor output
1 = Supervisor has been updated
o = Supervisor has not been updated
1 = Executing in AUTOTEST mode
o = Not executing in AUTOTEST mode
1 = R,eallocate or condense in progress
1 = Fetch $MAINEOJ at end of job to update system directory
0= Do not fetch $MAINEOJ at end of job for update

Language processor control byte. This is a set of switches used to specify nonstandard language translator options.
The switches within the byte are controlled by job control OPTION statements and when set to 1, override standard
options. The format of this byte is identical to the stnadard option byte (displacement 54) with one exception:
Bit 7 in this byte is used to indicate to LlOCS that the rewind a'nd unload option has been specified.
Job duration indicator byte
Bit 0:

1:
2:
3:

4:
5:
6:
7:

Figure 1.7.

1 = Within a job condition
job condition
1 = Dump on an abnormal end-of-job condition
0= No dump on abnormal EOJ
1 = Pause at EOJ step }
'.
.
Set by Attention Routine for Job Control
O = N 0 pause a t EOJ
1 = Job control output on SYSLST
o = Output not on SYSLST
1 = Job is being run out of sequence with a temporary assignment for SYSRDR
o = Conditions for 1 setting not met
1 = PCIL is being condensed
o = PCIL is not being condensed
Reserved
1 = Batch command just issued
o = Condition for 1 setting did not occur

o = Outside a

system Communications Region (Part 4 of 5)

Section 1: Supervisor Planning Concepts

Key to Communications Region Displacements:

Binary disk address of the volume label area (label cyl inder).

o
o
G
G

G
G
G

As illustrated (Figures for information blocks, I/O tables, and pointers
begin at Figure 1.14 which refers to more detailed Figures).

Set to the value nn specified in the LINES = nn parameter of the STDJC macro.
The format of the system date contained within this field is determined by the IPL program from information supplied
in the date convention byte (displacement 53). Bytes 85 - 87 contain the day count.
Bytes reserved for use by LlOCS. Transient dump programs insert a key to indicate to the LlOCS end-of-volume
routine, $$BCMT07, that it was called by a B-transient.
Address of the first part of the program information block (PIB) table. (See Figures 4.4, and 4.5.)
ID number of the last checkpoint. Temporary indicator of file protected DASD. Used at IPL time, when
DASDFP is specified.
Length of the LUBID queue (in bytes). This equals the number of channel queue entries. It can also be used to
access the REQID, LUBDSP, and TKREQID queues: (See Figure 1.29.)
Address of disk I/O position data. This is the starting address of the disk information block{DIB)table (See Figure I .24).
Address of the beginning of the error recovery block. The error recovery block contains addresses of error recovery
exits, error recovery queue information that can be used by physical transients routines, and defines storage for the
error queue entries
As illustrated (See Figure 1.13).
Key of the program (BG, F2, or FI) that has timer support.
As illustrated (See Figure I .29).
Logical Transient Key (LTK) contains the same value as the PIK (PID) (Displacement 46) when the logical transient is
requested. When the transient area is not in use, LTK is equal to zero. The SVC 2 routine sets the LTK. The
SVC II routine resets the LTK.
Supervisor constants:
DOLLARBO
SSKADR
LTAREA

(4 bytes) = C'$$BO'
(5 bytes) = XL5'O'
(3 bytes) = Adcon of LTSVPT, logical transient save pointer

Address of second part of program information block (PIB) table (See Figure 4.6).
Address of PDT ABB, table of DTF addresses for MICR support (See Figure I .9).
Address of OT AM vector table (IJLQTT AD).
Address of background communications region.
Option Indicator Byte
Bit 0:

I = MCRR indicated for OBR writer
No MCRR indicated for OBR writer
I = EU interface acti ve
o = EU interface not active
I ~ Teleprocessing request
o = No teleprocessing request
I = Supervisor support for only 9-track tape
0= Supervisor does not support 9-track tape exclusively
Reserved
J = RETAIN/370 support generated
0= RETAIN/370 support not generated
Reserved

I:
2:
3:
4:
5:
6-7:

System Configuration Byte 2
Bit 0: I = PCIL supported
0= PCIL not supported
I -7: Reserved
Pointer to communications region extension (See Figure 1.8).

Figure 1.7.

System Communications Region (Part 5 of 5)

DOS System Programmer's Guide

BGXTNSN (See Note)

o (Hexadecimal

4
Displacement)
(Decimal
4
Displacement)
CE Table
Address

XXXX

Difference
Track Hold
Between 1st
Table Address and 2nd Part
(THTABAD)
of PIB Table
(PIBDtFF)

24 (Hexadecimal 28
Displacement)
40
36 (Decimal
Displacement)
OLTEP
Linkage
Address
XXXX

XXXX

RMS
Linkage
Address
(RASLlNK)
XXXX

ID of
Task
Owning
LTA
(LID)

AB Termination Table
Address -8
(ABPTR)

XXXX

ID of
Task
Running
(PIK)

Task Requester ID
Tab Ie Address
(TKIDPTR)

XXXX

ASCII-EBCDIC
Translation
Table Address
XXXX

Address
Used by
QTAM
(MVCFLD)
XXXX

SDR
Communications
Address
(SDRTABLE)
XXXX

TEBV
Table
Address
(TEBYTAB)
XXXX

JAI Common
JAI Partition &SYSPARM
(Reserved) Table Address Table Address Field
(ACCTCOMN) (ACCTxx)
Address
XXX X

XXXX

Key to displacements:

CE Table Address.

~ Track Hold Table Address (THTABAD).
~ Difference between addresses of first part of PIB table and second part of

G
G

PIB table (PIBDIFF).

Abnormal Termination Table Address (minus 8) (ABPTR).
Identification (LID) of the task owning the Logical Transient Area. Contciins same value as PIK (displacement 18) when LTAis in use.
Contains zero when LTA is not in use.
Program Interrupt Key (PIK) if asynchronous processing is supported. Value is equal to the displacement of the start of the PIB table
to the PIB of the main task or subtask being selected (running).
First byte - zero
Second byte - contains the displacement into the PIB table for a maintask or a subtask.
Maintask - PIK value is hex 10, 20, or 30.
Subtask - PIK value is hex 70, 80, 90,
FO.

~ Task Requester ID Table Address (TKIDPTR).

~ MVCFLD address used by QTAM.
~ Statistical Data Recorder Tcible Address (SDRTABLE).
~ Tape Error Blocks by Volume Tobie Address (TEBVTAB).
~ Pointer to OLTEP Linkage Addresses
~ RMS Linkage Area Address (RASLlNK)

ASCII-EBCDIC Translation Table Address.

~ (Reserved)

JAI Common Table Address (ACCTCOMN)

~ JAI Partition Table Address (ACCTxx; where xx = BG,
~Address of&SYSPARM Field.
Note:

F2, or Fl).

If communications regions are generated for the foreground partitions, the labels in those extensions will be F2XTNSN
and F1XTNSN. The extensions, wherever used, are generated by the COMMNEX macro. Following the background
extension (and immediately preceding the MCRR Linkage Table) is a six-byte area. The first four bytes are the address of
the background save area (BGSAV), and the last two bytes are the value 4,096, used to restore base registers.

Figure 1.8

Background Communications Region Extension (BGXTNSN)
section 1: Supervisor Planning Concepts

MAGNETIC INK CHARACTER RECOGNITION SUPPORT
(MICR)
A Magnetic Character Reader, such as an IBM
1255, 1259, 1412, or 1419, can be attached
to a channel for reading
magnetically-inscribed data on checks and
other banking documents. They differ
mainly in document reading rates. The 1255
reads at speeds as high as 500
six-inch-Iong documents per minute on its
Model 1, and 750 documents per minute on
its Models 2 and 3; the 1259 reads at
speeds as high as 600 six-inch-Iong
documents per minute; the 1412, at speeds
as high as 950 documents per minute; the
1419, at speeds as high as 1,600 per
minute. Specific speeds depend on document
length as well as on the program.

instructions are provided. For multiple
MICR processing, READ, CHECK, and WAITF
macro instructions allow processing to
continue as long as one of the files has
documents ready for processing. Figure 1.9
shows the tables for MICR DTF addresses and
pointers.

System Generation Guidelines
To specify supervisor support for MICR
devices, the MICR= parameter must be
included in the SUPVR macro.
1.

If both 1412s and 1419s are present,
specify MICR=1419. If 1255s or 1259s
are to be supported, also specify
MICR=1419. MICR=1419D indicates Dual
Address Adapter 1419s.

1419 support gives 1255/1259/1270/1275
capability. The 1270/1275 are optical
reader/sorters.

If 1255/1259/1270/1275/1412/1419s are
attached to the multiplexor channel,
the PIOCS parameter BMPX=YES is not
supported.

If MICR support is required on a Model
65, specify MODEL=65 in the CONFG
macro.

For MICR support on selector channel,
specify MRSLCH=YES in the PIOCS macro.

system Considerations
The DOS supervisor support allows operation
of Magnetic Ink Character Recognition
(MICR) devices in either a foreground or
background area. An extension to the DOS
supervisor monitors, by means of external
interrupts, the reading of documents into a
user-supplied I/O area (document buffer
area). All MIC~ documents must be accessed
through logical IOCS. Logical IOCS gives
you the next sequential document and
automatically engages and disengages the
devices, as necessary, to provide a
continuous stream of input. Detected error
conditions and information are passed to
you in each document buffer.
The magnetic character readers are
unique in that documents must be read at a
rate dictated by the device rather than by
the program. To ensure time for necessary
processing (including determination of
pocket selection) a MICR device generates
an external interrupt at read completion of
each MICR document. The supervisor gives
highest priority to external interrupt
processing.
In an MPS system with MICR document
processing, any partition (background or
foreground) can use MICR devices. For
programs with one MICR device, GET macro

DOS System Programmer's Guide

MICR processing requires at least two I/O
channels. If MICR devices are attached to
the multiplexor channel, no burst mode
devices are supported on the multiplexor
channel. MICRs should be attached as the
highest priority devices on the multiplexor
channel. Single addressing 1412s or 1419s
are supported on any selector channel, but
device performance is maintained only if a
selector channel is dedicated to a single
MICR device. Also note that the Dual
Address 1419 is not attachable to selector
channels.
In addition, MICR processing requires
either the direct control feature or the
external interrupt feature.

The table of DTF addresses (PDTABB) contains six 8- byte entries; one for each external line of the
direct control feature on the system.
PDTABB
Byte ---+- 0

PDSTAT + 1
PDSTAT + 1
PDSTAT + 1
PDSTAT+l
PDSTAT + 1
PDSTAT+l,

NI
NI
NI
NI
NI
NI

0
8
16
24
32
40

X'FE '
X'FD '
X'FB '
X ' F7 1
X'EF '
X'DF '

.9-

...!:
~

s::
3:

au::

DTF address for MICR:
Device on line 7
Device on line 6
Device on line 5
Device on line 4
Device on line 3
Device on line 2

Background =10
Foreground 2 = 20
Foreground 1 = 30
• Bytes 0- 3 - - Contain an lAND I instruction that is executed in main I ine coding to turn off the
external line status after its detection.
PDSTAT + 1 contains one or more of the following interrupt codes:
PSW Interrupt
Code Bit

External
Interrupt Cause

Interrupt Code
(PSW Bits 26- 31)*
nnnnnnnl
nnnnnnln
nnnnnlnn
nnnnlnnn
nnnlnnnn
nnlnnnnn

31
30
29
28
27
26

External
External
External
External
External
External

signal
signal
siRnal
signal
signal
signal

7
6
5
4
3
2

• Byte 4- - Contains the flag of the partition containing the DTF.
• Bytes 5- 7 - - Contain the address of the DTF table.
Table of pointers (PDTABA) to DTF addresses associated with the external interrupt line. The table
contains the status in descend ing order from Bit 31 to Bit 26 of the external old PSW.
PDTABA
Byte~

0
8
16
24
32
40
48
56

00
00
00
00
00
00
00
00

6
I

08
08
08
08
08
08
08
08

00
00
00
00
00
00
00
00

10
10
10
10
10
10
10
10

00
00
00
00
00
00
00
00

08
08
08
08
08
08
08
08

00
00
00
00
00
00
00
00

18
20
18
28
18
20
18

*n =other external - interrupt cond itions.
Bytes 126 and 127 (X'7E'- '7F') of the communications region contain the address of these
tables. Label PDTABB identifies the first byte of the first table.

Figure 1.9.

Tables for MICR DTF Addresses and Pointers

Section 1: supervisor Planning Concepts

ASYNCHRONOUS PROCESSING (AP)
The asynchronous processing function, also
known as multitasking, provides greater use
of system resources at the partition level.
Multitasking provides the ability to
execute more than one program in a
partition, that is, the ability to do
multiprogramming within a partition (or in
all three partitions) of the DOS system.
Just as multiprogramming between partitions
can increase the system throughput,
multitasking can increase overlap of I/O
activity and computer processing for a
given job.
To perform multiprogramming within a
partition, the program must consist of a
main program (main task) and one or more
subprograms (subtasks).
Because multitasking is a logical
extension of the current task selection
mechanism, a maximum of nine subtasks can
exist in the system at any given time.
These nine subtasks can all reside in one
partition, or can be spread among the three
available partitions. A total of 12 tasks
(a task can be considered either a main
task or a subtask) can be executed
concurrently in the system.
The subtasks share the same partition
their associated main task. The main
task initiates (attaches) execution of the
subtasks. The ability of the main task to
attach subtasks minimizes operator
intervention. Storage within the partition
may be allocated to the main task and its
associated subtasks in any way desired by
the user. Subtasks have the same storage
protect key as the main task.
~ith

When subtasks are attached to a given
partition, they retain the priority of that
partition. Priorities are also established
within the partitions. The priority within
a partition is determined by the order in
which a subtask is initiated. The first
subtask to be attached has the highest
priority, and as each subsequent subtask is
attached, it has the next highest priority,
followed by the main task which has the
lowest priority. When a subtask is
attached, it receives control from the
system before control is returned to the
main task. See Figure 1.10 for an example
describing priority structure in a
multitasking environment. If the Fl
partition has two attached subtasks, the F2
partition has four attached subtasks, and
the BG partition has three attached
subtasks (a maximum of nine subtasks),
their priority would be as shown (with 1
being the highest priority and 12 the
lowest priority).
38

DOS System Programmer's Guide

A subtask can operate independently of
its main task and has its own save area for
registers. The subtask can communicate
with other subtasks and main task via a set
of macro instructions (see Intertask
Communication).

r----------------------T------------------,
Partition
I
Priority
I

~----------------------+------------------~

I Fl
I
I

Subtask 1
Subtask 2
Main Task

I F2
I
I
I
I

Subtask 1
Subtask 2
Subtask 3
Subtask 4
Main task

I
I
I

1
2

I
I
I
I
I

4
5
6
7
8

I
I
I

~----------------------+------------------~

I
I
I
I
I

~----------------------+------------------~

I BG
Subtask 1
I
Subtask 2
Subtask 3
I
IL _____________________
Main Task
Figure 1.10.

I
9
I
I
10
I
I
11
I
JI
I __________________
12

Example of Multitasking
Priorities

When a subtask is no longer required, it
can be detached from the system. The
subtask can either detach itself or be
detached by its main task. When one or
more subtasks are detached, subtasks with
lower priorities receive the next highest
priority. If a detached subtask is later
attached, it becomes the lowest priority
subtask in the partition, but it still has
higher priority than the main task.

system Considerations
Under DOS there are additional optional
components and specifications that greatly
enhance multitasking operations.
Track Hold: The track hold facility
prevents two independent subroutines in the
same partition or in two different
partitions from simultaneously trying to
update the same record or write a new
record on the same track when processing
DTFDA, DTFIS, and DTFSD files. When this
facility is used, a second routine
requesting an I/O operation on a track
being held must wait for that track to be
freed by the first routine. Because track
hold is implemented by programming rather
than hardware, all routines processing the
same DASD files must use this facility to
ensure proper protection. This facility
can be used without specifying AP=YES and
is specified at system generation time.
See discussion of TRKHLD parameter under
FOPT Macro for more information.

Multiple Wait: Under DOS, a number of
independent logical IOCS operations
(requiring explicit waiting for completion)
can be initiated before waiting for the
completion of any particular operation.
Once all logical IOCS operations have been
initiated, you must determine the sequence
in which you will wait for their
completion. Once you wait for a particular
operation, you no longer have control, even
though one of the remaining operations
completes before the one on which the wait
occurs and useful processing could have
been done. This can be avoided at the
physical IOCS level, but requires some
additional coding effort on your part.
The multiple wait facility allows you to
wait asynchronously for anyone of a number
of I/O operations to complete at either the
logical or physical IOCS level for the
above situation. This facility provides
increased I/O overlap processing and is
specified at system generation time. See
discussion of WAITM parameter under FOPT
Macro for more information.
Abnormal Termination: Under DOS, your
program is canceled when certain error
conditions occur. In many cases, it is
desirable to perform certain termination
functions (e.g., close files) to minimize
any problems that may occur. The abnormal
termination facility allows for these
situations via a user exit. This function
is specified at system generation time and
implemented via the AB operand of the STXIT
macro.
When the supervisor determines that the
task has been abnormally terminated,
control passes to the task's abnormal
termination routine. In this routine, you
may close data files (such as an indexed
sequential ADD) or perform other operations
that are necessary to minimize any possible
damage. Abnormal termination exits can be
established for both main tasks and
subtasks or, if desired, subtasks can share
the coding of their main task's abnormal
termination routine. It is strongly
suggested that in the shared abnormal
termination routine no I/O be performed.
If I/O is attempted and causes cancelation,
all tasks in the partition are canceled.
Any abnormal termination within the
abnormal termination routine causes the
task (or job if in the main task) to be
canceled without regard to an abnormal
termination exit. This facility of sharing
an abnormal termination routine can be used
even if the multitasking function (AP=YES)
is not used. See AB parameter under FOPT
Macro for more explanation.
Reentrant Modules: Reentrant modules for
CDMOD, DAMOD, ISMOD, MTMOD, PRMOD, SDMOD,
and DIMOD allow a module to be shared by

the same device type DTFs in a multitasking
environment. For example, one PRMOD can
support several subtasks using multiple
printers within a partition. One DAMOD can
support several subtasks within a
partition.

MultitaskinqMacro Usage
Although these four functions (asynchronous
processing, track hold, mUltiple wait and
abnormal termination) can be used
independently, they are discussed under the
heading where they are most frequently
used. The multitasking macros are designed
to handle three basic situations: subtask
initiation and normal termination"
resource protection and intertask
communication. See Supervisor and I/O
Macros listed in the Preface for a
description of the macro formats·. Some
examples using the multitasking macros are
included.

Subtask Initiation and Normal Termination
Subtask initiation can only be performed by
a main task that issues an ATTACH macro
instruction. Normal subtask termination
can be performed by either a main task or a
subtask that issues a DETACH macro
instruction.

ATTACH Macro Considerations
Only a main task can attach subtasks. A
maximum of nine subtasks can be attached in
the system at any given time. They can all
reside in one partition or be spread among
the three partitions in any combination.
If a main task attempts to attach a tenth
subtask to the system, a supervisor Event
Control Block (SPVECB) is unposted
(SPVECB+2, bit 0, set to zero), the address
of the ECB is stored in general register 1
of the main task, and bit 0 of register 1
is set to 1 giving the register a negative
value. The main task can test register 1
for a negative value and, if found, wait on
register 1 until one of the nine subtasks
is detached. Figure 1.11 illustrates the
ECB.
When a subtask is successfully attached,
it has a higher priority than its main
task. Therefore, control is passed to the
subtask before it is returned to the main
task. In addition, the registers of the
subtask contain the same values as the
Section 1: Supervisor Planning Concepts

registers of the main task (both the
general registers and floating point
registers if specified), with the following
two exceptions:

If a main task is canceled, all subtasks
in that partition are canceled.
DETACH Macro Considerations

Register 1 of the subtask contains the
address of the save area for the main
task.
Register 0 of the main task contains
the address of the byte immediately
following the save area of the subtask
(save area+96 if no floating point
registers, or save area+128, if
floating point registers).

The passing of the main task's registers
to its attached subtask(s) is worth noting,
because the subtask(s) can be under control
of .and use the main task's base register
without initializing it. In addition:
1.

The subtask ID (a value from X'70'to
X'FO') is stored in the subtask's save
area (save area+88, if no floating
pOint registers or save area+120, if
floating point registers)

The address of the subtask's entry
point is stored in the save area (save
area+13),and

Byte 2 bits 0 and 1 of the subtask's
ECB are set to 0 (unposted).

You should store the subtask name in the
first eight bytes of the save area to be
used for subtask identification when
messages are printed on SYSLOG.
In certain instances, a routine to be
attached may not be in main storage. In
this case, the entry point could be the
label of a FETCH or LOAD routine that
fetches the desired routine into storage.
The following conditions cause
cancelation of a main task (or possibly a
subtask).
1.

A main task has not issued the ATTACH
macro (issued by another subtask).

The subtask save area is not aligned
on a doubleword boundary.

The save area of the subtask being
attached is not within the partition.

The entry point of the subtask itself
is not within the partition.

The ABSAVE save area, if any, is not
within the partition.

The Event Control Block (ECB) of the
subtask is not within the partition.

DOS system Programmer's Guide

A main task can detach any subtask (within
its partition), but a subtask can only
detach itself. In addition, a subtask can
be detached by issuing the CANCEL, EOJ and
DUMP macros. If a subtask is detached, all
pending I/O operations are completed before
the DETACH operation is completed. In
addition, any tracks being held by the
subtaskjare freed.
If the subtask has an ECB, the ECB is
posted (ECB+2, bit 0 set to one) and any
tasks waiting on the ECB are removed from
wait state. The task with the highest
priority then gains control. The
supervisor ECB is also posted (SPVECB+2,
bit 0 set to one) and any main task waiting
on it is removed from wait state.
Although a main task can detach a
subtask, it is generally more desirable for
a subtask to detach itself. The entire
system could be put into wait state if two
(or all three) main tasks attempted to
attach more than the nine subtasks allowed
by the system. The following two examples
show what could happen if the main task is
allowed to detach its subtasks.
Example 1: All partitions attempt to
attach five subtasks apiece. Each set of
subtasks is independent and processing is
such that each main task has a chance to
attach a subtask before anyone main task
has attached all its subtasks. The entire
system could then be placed in wait state,
because the main task is not able to get to
the routine to detach a subtask when it has
successfully completed (it is waiting to
attach another subtask).
Example 2: Placing one or two of the three
partitions in wait state is another
situation that could occur, if all
partitions attempted to attach five
subtasks each. The F1 partition may attach
all five of the subtasks. F2 partition may
attach four of its five subtasks, and the
BG partition may not attach any of its
subtasks and, therefore, be in wait state.
The main task of the F2 partition would
also be in wait state because it still has
one more subtask to attach. These subtasks
remaining to be attached have to wait until
one or more of the attached subtasks are
detached. In addition, if the five
subtasks in the F2 partition are dependent
upon each other, the entire F2 partition
could also be in wait state. Thus, only
the tasks in the F1 partition may be
executing.

The following conditions cause
cancelation of a main task or a subtask:
1.

The main
pass the
area (if
the save

The main task detaches and the subtask
is already canceling or canceled.

a resource protected by an enqueue in
one partition is not protected in
another partition because an RCB is
generated in its own partition and
cannot be accessed by other
partitions. Figures 1.11 and 1.12
show the ECB and RCB.

task detaches and does not
address of the subtask save
a subtask detaches and passes
area address, it is ignored).

The limits of the save area specified
in the DETACH macro do not reside in
the partition of the main task. If
the main task is canceled, all
subtasks within the partition are
canceled.

The subtask 10 stored in the save area
is not a valid subtask 10 (hexadecimal

The second technique can only be
applied to DTFDA, DTFIS, and DTFSD
DASD files or those files you created
using physical IOCS (EXCP/WAIT
macros). This is the track hold
facility previously discussed. In
contrast to the first technique, track
hold applies across partitions.

70-FO).

In the last case, a check cannot be made
if the value has been altered to that of
another subtask 10. In this case, it would
be possible for the wrong subtask to be
detached because this is the only way the
system can locate the task being detached.
(This is also the reason why the main task
must specify the save area of the subtask.)
In addition, the system sets the invalid 10
to binary zeros.

Resource Protection
The resource may be a file, an I/O device,
a DTF, a work area or I/O area, or a set of
non-reentrant code, etc. In general, it is
anything that has the possibility of being
shared by two or more tasks. A means of
protection has been provided so that two
tasks sharing the same resource don't
access the resource at the same time.

Figure 1.11.

Event Control Block (ECB)

Figure 1.12.

Resource Control Block (RCB)

Resource protection can be accomplished
in one of two ways, depending upon the
resource to be protected.
1.

The first technique applies to the
types of resources just stated and
requires three macro instructions
(RCB, ENQ, DEQ). The RCB macro
generates a Resource Control Block
(RCB) that is associated with the
resource to be protected, but is not
necessarily a part of that resource.
When a resource is to be protected, an
ENQ macro must be issued to enqueue an
RCB. This places a hold on the RCB
associated with the resource until the
enqueuing task releases the RCB by
issuing a DEQ macro to dequeue the
RCB. These resource protection macros
apply only within a partition and not
across partition boundaries. That is,

Section 1: Supervisor Planning Concepts

ENQ Macro Considerations
A resource can only be protected within a
partition.
Every subtask that enqueues a resource
must have an ECB specified in its ATTACH
macro (issued by the main task) and that
ECB should not be used for any other
purpose while a resource is enqueued. The
address of the ECB is stored in the RCB. A
main task does not require an ECB and has
no means of specifying that it has an ECB
(the address of which could be stored in
the RCB). When a main task enqueues a
resource, the ECB address field of the RCB
is set to binary zeros which identifies the
resource owner as being the main task.
A task requesting use of a resource is
either enqueued and executed or put into
wait state if the resource has already been
enqueued by another task (byte 0 of the RCB
contains binary ones). In the latter case,
the flag byte is turned on in the RCB (byte
4, bit 0 is set to one) and the ECB address
of the current resource owner is stored in
general register 1 of the task placed into
wait state.
The following conditions can cause
cancelation of a task:
1.

A subtask does not have an ECB.

A previous owner of a resource
terminated without having dequeued the
resource. (If the main task
terminated, the entire partition was
terminated.)

A task issued two consecutive ENQs for
the same RCB without having issued an
intervening DEQ.

The limits of the RCB specified in the
ENQ do not reside in the partition of
the enqueuing task.

The following conditions cause
cancelation of a task:
1.

The RCB has the wait bit on in its
flag byte (byte 4) and no waiting task
which has been enqueued can be found
for the RCB. This could be caused by
the flag byte being inadvertently
altered while a resource was enqueued.

A subtask does not have an ECB.

The limits of the RCB specified in the
DEQ do not reside in the partition of
the dequeuing task.

Intertask Communication
In certain situations, tasks may be
dependent on other tasks within a
partition. In these cases, macro
instructions (POST, WAIT and WAlTH) have
been provided to permit synchronization of
tasks for intertask communication. To use
these macros, each task must have an Event
Control Block (ECB) associated with it.
When a particular task is performing a
function and other tasks are waiting for
its completion, it can indicate completion
via the ECB by issuing a POST macro
instruction. The tasks waiting for the
completion may indicate this by issuing the
WAIT macro instruction, designating the ECB
on which the waiting is to be done. If a
task is waiting for completion of a number
of events (in which the order of completion
is of no importance), it can issue the
multiple wait macro instruction (WAITM)
designating a number of ECBs.

POST Macro Considerations

DEQ Macro Considerations
A resource can only be dequeued within a
partition. Only the current owner of a
resource can dequeue that resource.
When an RCB is dequeued, byte 0 is reset
to binary zeros, and all tasks waiting for
that resource are removed-from wait state.
The task with the highest priority obtains
control. If no other tasks are waiting for
the RCB, control returns to the dequeuing
task.
42

DOS System Programmer's Guide

The tasks removed from wait state are those
placed in wait state by ENQ, WAIT, or
WAITM. When the SAVE parameter is omitted
in the POST macro instruction, all tasks
waiting on the specified ECB are removed
from wait state. By specifying a SAVE
parameter, only the task identified by the
save area is removed from wait state. This
parameter can be used for synchronizing the
order in which tasks are to receive
control. The synchronization technique
prevents the priority task within a
partition from gaining control.
Be careful with this technique when the
ECB to be posted is the ECB specified in
the ATTACH macro and ENQ/DEQ macros are
used, because DEQ also removes all waiting
tasks from wait state. When the posting

task dequeues, all tasks waiting for the
RCB are removed from wait state.

from each other. The ENQ, DEQ, and RCB
macros offer protection of resources, but
only if all subtasks enqueue and dequeue
before using the resource. They do not
protect against inadvertent coding errors.
In addition, the user must be careful when
using ENQ/DEQ. If a higher priority
subtask dequeues an RCB and does not enter
wait state before enqueuing the RCB again,
a lower priority task does not gain control
of that RCB on which it has also enqueued.
Thus, the lower priority task has to wait
until the higher priority task terminates,
before it can enqueue the RCB.

A similar situation exists if the
posting task dequeues before posting. The
DEQ removes all tasks waiting for the RCB
from wait state. Then, if it issues a POST
to a particular task, the POST acts as a
NOP because task selection gives control to
the highest priority task ready to use the
cpu. Although the task being posted is
removed from wait state, the posting task
is still active. If the posting task
issues another POST to the same or another
ECB, all other tasks waiting on the posted
ECB are removed from wait state. To avoid
this situation, use a second ECB when
synchronizing tasks. It is your
responsibility to reset the wait bit in the
second ECB that is to be posted (MVI
ECB+2,X'OO'), so that tasks testing that
ECB can be put in wait state.

Access To Communications Region: Only one
communications region exists in a
batched-job partition. Therefore, i t is
likely that only one task per partition has
meaningful access to it.
System Logical Units: Only one set of
system logical units exist per partition
(SYSLST, SYSRDR, etc). Therefore,
interspersed usage by several independent
tasks is not practical, although, if either
the resource protection facilities or the
intertask communication macros are
employed, it can be done.

If a task associated with the specified
save area cannot be found, the post
operation is ignored and control passes to
the highest priority task that is ready to
run.
A task can be canceled if the ECB
specified in the POST macro instruction
does not reside within the partition.

Operator Intervention: While operator
intervention is minimized for subtask
initiation, SYSLOG will probably be used by
all tasks within all partitions. The
additional number of messages possible on
one SYSLOG could possibly increase the
responsibility of the operator and require
more careful operation than in the past.

Summary of Multitasking Considerations
Maximum Number of Tasks: A maximum of 9
subtasks can be attached to the system.
They can all reside in one partition or be
spread among the three partitions. Thus, a
total of 12 tasks can be executed
concurrently in the system.
Subtask Priority: Each subtask must be
initiated by a main task. A subtask has a
higher priority than its main task.
Subtask priority within a partition is
determined by the order of attachment. The
first subtask attachment has the highest
priority in the partition, the next subtask
has the second highest priority in the
partition, etc. Of course, the priority of
the partitions remains the same (i.e., Fl,
F2, and BG).
If the highest priority
subtask is terminated and later reattached,
it will then be the lowest priority subtask
within the partition, but still higher in
priority than its main task, or any
subsequent subtask that may be attached
within that partition.
Storage Protection: Because subtasks are
subprograms within a partition, they have
the sam~ storage protect key as the main
task. Therefore, the main task and its
subtasks do not have storage protection

STXIT Macro Usage: Subtasks may only
provide their own AB and PC routines via
the STXIT macro. IT and OC operations must
be performed via the main task. An AB exit
is not taken for a task when it is already
in its AB routine (prevents looping on
abnormal termination condition). The task
is canceled.
Checkpoint Consideration:
can issue checkpoints.

Only main tasks

Track Hold Facility: Files being shared on
DASD are not protected unless the HOLD
option is specified by the various users
(only applies to DTFDA, DTFIS, DTFSD, and
DTFPH files).
Register Usage: Although a subtask has the
register values (2-15) of its main task
upon being attached, the registers cannot
be used for passing information between
tasks once attached.
(It is possible for a
task to access the register save areas of
other tasks within its partition.)
Process Bound Tasks Considerations:
Because subtasks are executed in priority
order, a process-bound task can degrade
Section 1: supervisor Planning Concepts

performance of lower priority tasks, or in
extreme cases, even prevent execution of
lOwer priority tasks until it has
terminated.

as the cancel routines.) Therefore, when
the abnormal termination routine is
entered, you should dequeue all RCBs that
the subtask could have enqueued.

Task Synchronization: Task synchronization
is normally performed by POST, WAIT, and
WAITM macros. This can also be done by
ATTACH/DETACH or ENQ/DEQ providing you are
careful, particularly when intermixing
POST, ENQ/DEQ, and ATTACH/DETACH macros.
While POST may be used to free one waiting
task, DEQ and DETACH can free all tasks
waiting on the ECB posted, if the ECB is
the same one specified in the ATTACH.

If the subtask issues a DEQ for an RCB
on which it has not enqueued, the DEQ is
ignored and the supervisor returns to the
subtask's abnormal termination routine.

Resource Protection: The POST, WAIT, and
WAITM macros are also used for resource
protection, providing you are careful in
your synchronization techniques.
Resource contention: The problem of
resource contention cannot be over
emphasized in this system. It has already
been pointed out that you can interlock two
or more tasks, or even put the system into
wait state when two or more partitions are
concurrently attaching more than 9 subtasks
or when two or more tasks (or partitions)
are contending for the same sets of tracks
while using track hold. In addition, a
similar problem can exist when two or more
tasks within a partition are enqueuing and
dequeuing on the same set of RCBs. For
example, if task A enqueues RCB 1 and task
B enqueues RCB 2, task A is put in wait
state when it attempts to enqueue RCB 2.
The same is true when task B attempts to
enqueue RCB 1. Neither task is able to get
out of wait state to release the resource
it has enqueued.
This problem can be avoided by having
each task, which shares common resources
with other tasks, enqueue on the same
resources in order. For example, task A
enqueues on RCB 1 and then task B enqueues
on RCB 1 (instead of RCB 2 first); task B
goes into wait state. Task A can now
enqueue on RCB 2 without entering wait
state. When task A dequeues RCB 1, task B
has the chance of enqueuing RCB 1
(providing task A does not enqueue RCB 1
again, before task B has a chance to
reattempt its ENQ).
Another possibility of task interlock is
for two tasks to wait on ECBs, with each
task assuming that the other task will post
the ECB on which it is waiting.
Subtask Cancelation: While the cancelation
of a subtask frees tracks being held, and
posts the subtask's ECB (as. specified in
the ATTACH macro) it does not dequeue any
RCBs enqueued by that subtask.
(Cancelation of a subtask executes the
DETACH routin~s of the supervisor as well
44

DOS system Programmer's Guide

Wait Considerations: Although tasks can
wait on ECBs to be posted by other tasks or
on the ECBs of other tasks (in their own
partition), they cannot wait on the CCB of
another task when that task has initiated
the I/O operation. This does not mean that
two tasks cannot share a CCB as a resource.
It only means that the system identifies
the CCB with the task dOing the I/O
operation. Therefore, only that task can
be removed from wait state by the system.
Any other task waiting on another task's
CCB can only be removed from wait state by
having the task that started the I/O
operation issue a POST to the CCB. In this
case, the CCB would function like an ECB.
Also, note that the task doing the I/O
operation must issue the POST macro after
the WAIT macro rather than before the WAIT
macro. Otherwise it would never enter the
wait state or determine when the I/O
operation is completed.
Abnormal Termination: In all abnormal
termination conditions where an exit is
taken to an abnormal termination routine,
the register values are stored in the
ABSAVE save area before the appropriate
error code is stored in the low-order byte
of register O. To have this value
available when looking at a storage dump,
you should store (STC or ST) register 0 in
another save area upon entry into the
abnormal termination routine. You will
find that the SVC code shown in the wOS04I
ILLEGAL SVC-.,•• w message along with the
error codes in register 0 will be helpful
in tracing program errors. See Debugging
Aids section for add1tional information on
abnormal termination codes.
System Generation Guidelines
The multitasking facility is provided at
system generation time by specifying AP=YES
in the SUPVR macro. When AP=YES is
specified, MPS=YES and WAITM=YES are
implied. To implement the other facilities
related to multitasking, the following
additional specifications are required at
system generation time:
• For abnormal termination support,
AB=YES must be specified in the FOPT
macro. AP=YES in the SUPVR macro is
not required to utilize this function,.

• For multiple wait support, WAITM=YES
must be specified in the FOPT macro.
Although the multiple wait function can
be used without specifying AP=YES,
AP=YES cannot be used without
specifying WAITM=YES.

MULTITASKING EXAMPLES
ATTACH Macro Example
The normal procedure for attaching subtasks is as follows:
1.
2.

MAINTASK

BALR
USING

•

STXIT
3.
4.
5.
6.
7.
8.
9.
10.

11.
12.
13.
14.
15.
16.

MVC

ATSTl

ATTACH
LTR
BNM
WAIT
B

ATST10K

SUBTASKl
MTABENO

BCTR
ST

•
•

BALR
USING
ST
STC
C

•
•
EQU
•
•

2,0

*,2
AB,MTABEND,MTSAVE
SUB1SAV(8),SUB1NAME
Initialize subtask 1 save area
SUBTASK1,SAVE=ST1SAV,ECB=ST1ECB,ABSAVE=STlABSV
1,1
Test if ATTACH is successful
ATST10K
BR if successful
(1)
WAIT to retry ATTACH
ATSTl
BR to retry
0,0
Get end of subtask 1 save area
0,ST1SVEND
Store ending address of subtask 1 save
area
3,0

*,3

1,MTSVAR
O,ABSVCOOE
1,=A(ST1ABSV)
ST1ABENO

Store address of main task save area
Save ABTERM code
Test if subtask 1 ABTERM
BR if YES

17.

STlABENO

18.
19.

ST1SAV

OS
OC

00
160'0'

20.
21.
22.

ST1ABSV
ST1ECB
MTSVAR

OC
OC
OC

90'0'
FlO'
F'O'

23.

ST1SVENO

F'O'

24.
25.
26.

SUB1NAME
ABSVCOOE
MTSAVE

DC
OC
OS

C'SUBTASK1'
X'O'
90

Align on doubleword boundary
Subtask 1 save area
with floating point
registers
Subtask 1 AB save area
Subtask 1 ECB
Address of main task
save area
Ending address of
subtask 1 save area
Subtask 1 name
Main task save area
used by STXIT

)
Section 1: supervisor Planning Concepts

Explanation for ATTACH Macro: Statement 3 initializes the subtask save area with the
name of the subtask which is used for messages for subtask identification when messages
are written on SYSLOG.
statement 4 is the ATTACH of the subtask. SUBTASK1 is the entry point of the subtask,
ST1SAV is the save area for the subtask, ST1ECB is its ECB, and ST1ABSV is the ABTERM
save area for the subtask. In this case, the subtask is using the main task's abnormal
termination routine.
statements 5 and 6 test for a successful ATTACH. If the ATTACH was not successful
(nine subtasks already attached), the main task waits until another subtask is detached
and retries the ATTACH. If the ATTACH was successful, the main task stores the ending
address of the subtask's save area for later reference, if necessary. The main task can
then continue to do other processing.
statement 11 is the entry point to the subtask. In this example, the subtask and the
main task use different base registers. This may not be necessary, depending on program
design. The subtask could have omitted the BALR and USING statements because
addressability is available through the main task register (register 2). The values in
the main task registers are passed to the task. Therefore, register 2 wou1d still be
initialized.
Statement 13 saves the address of the main task's save area for reference by the
subtask (if it is necessary for the subtask to name the main task in the POST macro
instruction). Statement 14 stores the ABTERM code when the abnormal termination routine
is entered. This routine is shared by both the main task and subtask 1. Statements 15
and 16 determine which task abnormally terminated (ABTERM save area of the task in error
is stored in register 1). Statement 18 aligns the save areas on a doubleword boundary.
Statement 21 is the user-coded ECB for the subtask.

c
46

DOS System Programmer's Guide

DETACH Macro Example
A subtask may detach itself or be detached by the main task.
1. MAINTASK
BALR
2, 0
2.
USING
*,2

ATSTl

•
•
ATTACH
•

ATST2

ATTACH ST2,SAVE=ST2SAV,ECB=ST2ECB

•

•
•
DETACH
•

ST1,SAVE=ST1SAV,ECB=ST1ECB

•

SAVE=ST1SAV

STl

1,MTSVAR1

7.
8.

ST2

B
ST

ST1+4
1,MTSVAR2

9.
10.

*DETACH

•
•

Detach subtask 1

Subtask 2
DETACH

Explanation for DETACH Macro: The main task attaches two subtasks. When subtask 1
completes processing, i t indicates this to the main task. The main task then detaches
subtask 1 by issuing a DETACH macro and specifying the save area for subtask 1 (statement
5). When subtask 2 completes its processing, it detaches itself (statement 10). Note
that an operand was not specified when subtask 2 detached itself, and that the comment
was placed in a comments card (statement 9). The comment would have acted as an operand,
resulting in an error.

Section 1: Supervisor Planning Concepts

ENQ/DEQ and RCB Macros Examples
EXAMPLE 1: When two subtasks share the same resource within a partition, they can use
the resource protection macros as follows:
1.
2.

MAINTASK

SUBTASKl

SBTASKlA

5.
6.

7.
8.

SUBTASK2

9.
10.
11.

SBTASK2A

•
•
EQU
•
•

ENQ
BAL
DEQ

•
•
B
•
•
EQU
•
•

ENQ
BAL
DEQ

•

•
•
•

12.
13.

BALR
USING

RCBl

2,0
*,2
*
RCBl
4,WRITEDTA
RCBl

Protect resource.
Write a record.
Release resource.

SBTASK1A
*
RCBl
4,WRITEDTA
RCBl

Protect resource.
Write a record.
Release resource.

SBTASK2A

RCB

Resource control block for WRITEDTA

Explanation for Example 1: Both subtask 1 and subtask 2 are sharing the same file using
a common subroutine. The subroutine is not reentrant, and the file cannot use track
hold. Therefore, i t is necessary for each subtask to enqueue on the RCB associated with
the resource and dequeue when the resource can be released for a waiting subtask.

DOS system Programmer's Guide

EXAMPLE 2: In the following example, two subtasks are sharing a common processing
routine defined in the first subtask.
1.

MTASK

2.
3.

STASK1

•
•
ATTACH
•
•
ATTACH
•
•
ENQ

•

•
•
•
DEQ
•
•
EQU
•
•
B
•
•

STASK1,SAVE=SAVE1,ECB=ECB1
STASK2,SAVE=SAVE2,ECB=ECB2
RCBA

STASK2

8.
9. RCBA
10. TOTAL

RCB
DS

Protect resource TOTAL.
Used by STASKl and STASK2.

Process TOTAL

START 0

RCBA

Release resource TOTAL.

*
STASK1

Process TOTAL
RCB for resource TOTAL
Shared resource

Explanation for Example 2: The resource (TOTAL) in STASK1 is protected by the resource
control block named RCBA. The protection remains in effect only if every segment of
coding within the partition referring to TOTAL issues the ENQ macro before executing that
selection of coding and subsequently dequeues that resource with the DEQ macro. This is
effectively accomplished by branching to the same physical set of code.
Note that the coding from statements 4-6 does not necessarily have to be reentrant,
but you should ensure that values for constants associated with the subroutine do not
have to be retained from one reference to the next, whenever the resource is used. If
so, these values should be saved with the appropriate subtask and then later restored.

Section 1: Supervisor Planning Concepts

EXAMPLE 3: In this example, the subtasks again share the same resource, but use
different subroutines for processing that resource.
1.

MTASK

START 0

•
•
ATTACH
•

ATTACH ST2,SAVE=SAVE2,ECB=ECB2

•

•
•
EQU
•

STASKl

•

ENQ

5.
6.

*Update

•
•
RESRCA
•

•

DEQ

7.
8.

STASK2

•

Updat.e

12.

•
•
RES RCA
•

•

11.

13.

•
•
EQU
•
ENQ

10.

ST1,SAVE=SAVE1,ECB=ECBl

DEQ
RCBA
RESRCA

•
•

*
ReBA

Protect resource RES RCA
Process using RESRCA

ReBA

Release resource RES RCA

*
ReBA

Protect resource RES RCA
Process using RESRCA

RCBA

RCB
OS or DTF

Release resource RES RCA
RCB for resource RESRCA
Shared resource

Explanation for Example 3: RESRCA can be simply an area in main storage or a file
defined by a declarative macro. In either case, RESRCA is protected from subtask 2 while
subtask 1 is operating on it. Thus, if all tasks enqueue and dequeue each reference to
RESRCA, then RESRCA is protected during the time it takes to process instructions from
that task's ENQ instruction to its DEQ instruction. This is readily apparent if RESRCA
is in main storage. However, if it is a file, the record being operated upon is
protected while in main storage, but it is not necessarily protected on the external
storage device.
If the file is on DASD, the HOLD function should be utilized, if possible. In any
such situation, the priorities of subtasks must be considered for proper operation.

DOS System Programmer's Guide

POST Macro Example
The POST macro can be used by one task to inform another task of the completion of some
event, or it can release a number of tasks from wait state.
1.
2.

MAINTASK

BALR
USING

•
•
ATTACH
•
•
ATTACH
•
•
ATTACH
•
•

3.
4.
5.
6.
7.

2,0
*,2
ST1,SAVE=AREA1,ECB=ECB1
ST2,SAVE=AREA2,ECB=ECB2
ST3,SAVE=AREA3,ECB=ECB3

ECBl
WAIT
DETACH SAVE=AREAl

Wait for completion of subtask 1
Detach subtask 1

•

8.
9.

•

ST1

12.
13.
14.

ST1EOJ

15.

ST2

16.

ST2A

•
•

POST
WAIT

17.
18.
19.

ST3

20.

ST3A

21.

•
•
EQU
•
•
EQU
•
•
POST
•
•
B
•
•
EQU
•
•
EQU

•

•
•
•
B
•
•

POST

22.

)

•
•

1,MTSVAR

WAITM ECB2,ECB3
2(1),X'7F'
NI

10.
11.

23.
24.
25.
26.
27.

EOJ
ST

MTSVAR
ECB1A
ECBl
ECB2
ECB3

DC
DC
DC
DC
DC

O,MTSVAR
ECB1,SAVE=(O)
ECB1A

Store address of main task save area
Wait for subtask 2 or subtask 3
Turn off WAIT bit
Get address of main task save area
POST ECB for main task
WAIT to be detached

*
*
ECB2

POST ECB for subtask 1

ST2A
*
*
ECB3
ST3A
F'O'
F'O'
F'O'
F'O'
F'O'

Save area address for main task
Dummy ECB for subtask 1
ECBs for subtasks

section 1: supervisor Planning Concepts

Explanation for POST Macro: Subtask 1 (ST1) is dependent on input from subtask 2 (ST2)
and subtask 3 (ST3). Therefore, it issues a WAITM on the ECBs for those subtasks.
Notice that statement 11 resets the wait bit (set to 0) in the ECB that satisfies the
wait condition. This ensures that the wait bit is off before reissuing the WAITM.
(Subtask 1 is the highest priority task and, therefore, would gain control before subtask
2 and subtask 3. The result is that the WAITM is always satisfied from the first
operation.) Initially, subtask 1 is placed in wait state by the WAITM. Control is then
passed to subtask 2 and then to subtask 3. When either of the two subtasks has the
necessary data for subtask 1, it posts its ECB that removes subtask 1 from wait state.
when subtask 1 finishes its processing, i t posts its ECB causing the main task to be
taken out of wait state which then detaches subtask 1.

WAITM Macro Example
A task issuing the WAITM macro should ensure that if an event has the possibility of not
occurring (perhaps the task posting the event was terminated), the waiting task should
allow an eventual outlet. This outlet, as shown in the following example, can also wait
on the terminating ECB of the task that was to perform the preferred event.
1.
2.

MAINTASK

•
•
ATTACH
•
•

3.
4.
5.
6.
7.
8.

11.
12.
13.
14.
15.

2,0
*,2
ST1,SAVE=SAVE1,ECB=ECBl

WAITM ECB2,ECBl
2(1),X'7F'
NI
4(1)
B
PREVENT
TEVENT

10.

BALR
USING

STl

•
•

EQU

•

•
EQU
•
•
EOJ
•
•
EQU
•
•

ECBl
ECB2

POST
DC
B
DC
B

Wait for preferred or secondary event
Turn off WAIT bit
BR to branch in vector table

Main task preferred event

Main task secondary event

Main task end of job
*
ECB2
F'O'
TEVENT
F'O'
PREVENT

POST completion of preferred event.
ECB for secondary event
Vector BR to secondary event
ECB for preferred event
Vector BR to preferred event

Explanation for WAITM Macro: In this example, the WAITM macro contains a preferred event
as the first operand and a secondary event as the second operand. The preferred event is
the posting of ECB2 after subtask 1 completes its calculation. If subtask 1 terminated
before its calculation is completed, the supervisor posts the ATTACH macro ECB of subtask
1, ECB1, and the secondary event can satisfy the WAITM macro. In either event, after the
WAITM macro is satisfied, the address of the posted ECB is contained in register 1. This
address can select a problem program routine.
In this particular case, a branch instruction points to a table containing a.list of
ECBs with corresponding branch instructions to the routine to be given control when the
ECB is posted. This table can easily be expanded to include up to a maximum of 16 ECBs
(due to the WAITM format used).

DOS System Programmer's Guide

STXIT AB Macro Example
The STXIT AB macro instruction establishes linkage to an abnormal termination routine for
either a main task or a subtask.
1.
2.
3.

MAINTASK

4.
5.
6.
7.
8.
9.

MTABEND

BALR
USING
STXIT

•
•
ATTACH
•
•
ATTACH
•
•
ATTACH
•
•
STC
C
BE

2,0
*,2
AB,MTABEND,MTABSV

Initialize AB exit

ST1, SAVE=SAVE1, ECB=ECB1,ABSAVE=ST1ABSV
ST2,SAVE=SAVE2,ECB=ECB2
ST3, SAVE=SAVE3, ECB=ECB3
O,MTABCODE
1, A (ST1ABSV)
ST1ABEND

Save AB code
Test if ST1ABTERM
BR if YES

•

10.
11.

•
CANCEL
STlABEND

12.
13.

ST1

14.

ST2

15.

ST2ABEND

16.

17.

ST3

18.

ST3ABEND

CANCEL for main task cancels all tasks

EQU

•
•
CANCEL
•
•
EQU
•

•

STXIT

•
•
STC
•
•
CANCEL
•
•
•
STXIT
•

•

STC

•
•

19.

CANCEL

20.
21.
22.
23.
24.
25.
26.
27.

OC
DS
OC
OS
OS
DC
DC
OC

*
CANCEL ALL for subtask 1

ALL

*
AB,ST2ABEND,ST2ABSV

Initialize subtask 2 AB exit

O,ST2ABCOD

Save AB code
CANCEL for subtask 2

AB,ST3ABEND,ST2ABSV

Initialize subtask 3 AB exit

0,ST3ABCOO

Save AB code

OD
9D'0'
90'0'
9D'0'
90'0'
X'O'
X'O'
X'O'

Align on doubleword boundary
Main task AB save area
Subtask 1 AB save area
Subtask 2 AB save area
Subtask 3 AB save area
Save area for AB codes

•
•

MTABSV
ST1ABSV
ST2ABSV
ST3ABSV
MTABCODE
ST2ABCOO
ST3ABCOD

section 1: Supervisor Planning Concepts

Explanation for STXIT Macro: statement 3 establishes linkage to the abnormal termination
routine for the main task. statement 4 attaches subtask 1 and indicates to the
supervisor that it will use the main task's abnormal termination routine by specifying
the ABSAVE parameter. Note that the main task's abnormal termination routine tests for a
main task or subtask 1 abnormal termination by comparing the address in register 1 to the
address of subtask l's AB save area.
When the main task or subtask 1 cancels (CANCEL ALL), the entire partition is
canceled. Subtasks 2 and 3 initialize their own abnormal termination exits because they
use their own abnormal termination routines. When either subtask 2 or subtask 3 cancels,
only that subtask is terminated; the other tasks within the partition continue
processing.

DOS System Programmer's Guide

Track Hold and Reentrant Modules Example
Although track hold applies across partitions, this example only shows two subtasks
sharing the same DA file and the same DA modules. It is possible that a similar set of
routines could be executing in a second partition also sharing the file with the first
partition, but that partition would have its own DA module.
1.

MAINTASK

START

•
•

ATTACH ST1,SAVE=AREA1,ECB=ECB1

•

3.
4.

ST1

5.
6.

•
ATTACH
•
•
OPEN
•
•
LA

READ

•
•
WAITF
•
•

7.
8.
9.
10.

WRITE
WAITF
FREE

11.

DAFILE1

12.

ST2

•
•
DTFDA
•
•
OPEN
•
•

ST2,SAVE=AREA2,ECB=ECB2
DAFILE1

OPEN DA master file

13,DASAVE1
DAFlLE1,KEY

Initialize register 13 with DA save area
Read and hold record

DAFILE1
DAFILE1,KEY
DAFILE1
DAFILEl

Write updated record
Release track

HOLD=YES,RDONLY=YES, •••
DAFlLE2

OPEN DA master file.
Initialize register 13 with DA save area
Read and hold record
From DA master file

13.
14.
15.

WAITF

13,DASAVE2
DAFlLE2,KEY
DAFILE2

16.
17.

WRITE
WAITF

DAFILE2.KEY
DAFILE2

Write updated record

DAFILE2

Release track

LA
READ

•
•

18.
19.

DAFILE2

20.
21.
22.

DASAVE1
DASAVE2

•
•
FREE
•
•
DTFDA
•
•
DAMOD
•
•
DS
OS

HOLD=YES.RDONLY=YES ••••
HOLD=YES.RDONLY=YES, •••
8D'0'
8D' 0'

Save areas used by DAMOD
when shared and reentrant

Section 1: Supervisor Planning Concepts

Explanation for TRACK HOLD and REENTRANT Modules: Because both subtasks share the same
file, HOLD=YES and RDONLY=YES must be specified in both DTFs and in the DAMOD. In
addition, before any I/O operation is issued (READ, WRITE, or WAITF), register 13 must
contain the address of a unique save area to store the registers used by the module.
Register 13 is not altered between I/O operations executed by a given subtask, and
therefore, only needs to be initialized once. If other reentrant access methods were
used by the subtask, register 13 would have to be initialized for each LIOCS function.

FOPT Macro
OPERATOR COMMUNICATIONS SUPPORT (OC)
Operator Communications (OC) refers to the
processing of an external interrupt by a
problem program. In a multitasking
environment, only the main task can
communicate via the OC linkage. By
specifying OC=YES, a table (OC option
table) is generated within the supervisor
(see Figure 1.13). When the problem
program issues the STXIT macro, the address
of its external interrupt routine is moved
to the OC option table. The user's routine
is terminated by issuing the EXIT macro.
When OC=YES is specified, support is
available to all partitions. Figure 1.20
illustrates a sample program using this
support.
The Tape Compare Utility program
requires this support. OC=YES is also
required if emulator program operator
services are to be requested through the
INTERRUPT key.

Operational Considerations
To cause an external interrupt for the
background partition, the operator presses
the INTERRUPT key on the CPU console. To
cause an external interrupt for the

DOS system Programmer's Guide

foreground partitions, the operator
the REQUEST key on the 1052 console
reply to the ATTN routine statement
FOR COMMUNICATIONS', types 'MSG Fl'
F2'.

presses
and. in
'READY
or'MSG

INTERVAL TIMER SUPPORT (IT)
This parameter generates programming
support for the hardware timer feature.
The timer consists of two parts. The first
part keeps track of the time-of-day and is
used to time-stamp system time. The second
part of the timer allows a problem program
to set a time interval (via SETIME macro).
By using the STXIT, EXIT, and TECB macros,
a specific routine within the problem
program is entered when this time interval
elapses. In a multitasking environment,
only the main task can set a Timer Event
Control Block (TECB).
The interval timer is in addition to and
separate from the time-of-day support
generated by the specification of the
TIMER= parameter of the CONFG macro. When
interval timer support is specified (i.e.,
IT=BG, Fl, or F2), TIMER=YES is assumed for
the CONFG macro. Support is only available
to one partition at a tim~_as defined at
system generation time. The TIMER command
can change the assignment from one
partition to another after the supervisor
has been generated. QTAM requires IT=Fl.

PC Option Table and OC Option Table:

::11-----:-----1-1- :-------II
1-1- - - - word -~---+I-,----- word -----+II

No STXIT given: a = 0
STXIT issued: a = address of the user program check (operator communications) routine
STXIT issued when the user routine is already in use: a = complement of user program check (operatar communications)
routine address

No STXIT given: b = 0
STXIT issued: address of the user save area
IT Option Table

11------ word -----+--I- - - - word -----il
No TECB or STXIT issued: c = 0
TECB issued: c = addr.ess of the timer event control block
STXIT issued: c = address of the user interval timer routine
STXIT issued when user routine is already in use: c = complement of the user interval time routine address

0
No TECB or STXIT issued: d = 0
TECB issued: d = complement of the TECB address
STXIT issued: d = address of the user save area
AB Option Tabl~

I
1

word

f
word

I
1

No STXIT given: e = 0
STXIT issued and rtnaddr parameter passed: e=address of entry point of user's abnormal termination routine. If AP
(asynchronous processing) .is supported, the main task and subtasks may have the same or different AB routines.
When a subtask is attached after a STXIT AB macro has been issued by the main task, the subtask will receive
the AB routine address specified by the main task only if the ATTACH macro for that subtask has the ABSAVE
parameter specified. The subtask can override this by issuing its own STXIT AB macro.

No STXIT given or no save area parameter passed: f = 0
STXIT issued and save area parameter passed: f=address of a 72- byte save area used by the supervisol" to store the
old PSW and general registers 0-15.
Each table address (less 8 bytes) is found in the communications region at the byte locations specified below. The labels
shown identify the first byte of the corresponding table.
Table
PC
IT
OC
AB

Bytes in COM REG

Label
-

100- 101 (X'64'- '65')
102-103 (X'66'- '67')
104- 105 (X'68'- '69')
12-13 (X'OC'- 'OD')
of extension

PCTAB
ITTAB
OCTAB
ABTAB

Figure 1.13.

Option Tables
Section 1: supervisor Planning Concepts

PROGRAM CHECK SUPPORT (PC)
Program Check (PC) support generates a PC
table within the supervisor (see Figure
1.13). The PC table contains the address
of a user program check routine. This
address is placed in the table via the
STXIT macro issued by the problem program.
If the STXIT PC linkage is established and
a program check within this program occurs,
the supervisor gives control to the user's
routine instead of canceling the job being
run in this partition. The support is
extremely advantageous when using LIOCS
(e.g., files can be closed before job
termination).
If a program check occurs in
a routine being executed from the logical
transient area (LTA), only the task
associated with that routine is abnormally
terminated.
In a multitasking environment each
subtask and main task may have its own PC
routine. A PC routine can be shared by
more than one task within a partition.
This can be done by issuing a STXIT macro
in each task with the same routine address
but with separate save areas. To
successfully share the same PC routine it
must be reentrant (capable of being used
concurrently by two or more tasks).
Figure 1.20 shows a sample program using
this support. Refer to supervisor and I/O
Macros listed in the Preface for fUrther
information on problem program macro
formats and their use.

TAPE ERROR RECORDING
The three options and their system
generation specifications for tape error
recording are:

TEB: The system generates the number of
tape unit error blocks specified by n. TEB
support automatically collects and writes
magnetic tape unit status on SYSLOG at the
end of every job utilizing these units.
The n must be at least equal to the number
of tape units and/or tape cartridge readers
attached to the system.
(TEB is the only
one of the three options supporting the
tape cartridge reader.) Additional TEBs
should be specified for possible future
expansion.
ESTV: The system generates a TEBV table
with a status block and the number of error
blocks specified by the n in
TEBV=(DASD,n) I (SYSLOG,n). ESTV transients
format and write the error records on DASD
or SYSLOG each time a particular volume is
ended by CLOSE, EOVi EOJ, or abnormal
termination.
For TEBV=(DASD,n), the ESTV recorder
file (ESTVFLE) must be formatted and opened
for the collection of statistics using
ESTVFMT, the ESTV format program. SYSREC
is the system logical unit used to collect
ESTV statistics. The information written
on ESTVFLE may be retrieved by executing
the ESTV dump program, ESTVUT.
For TEBV=(SYSLOG,n), ESTV transients
format and write the data on SYSLOG and do
not access any intermediate storage device.
EVA: The system generates a TEBV table and
prints a message on SYSLOG when the
temporary read error threshold (r) or
temporary write error threshold (w) has
been exceeded on a currently accessed tape
volume. The number of SIOs is also
included in the message. EVA can be used
for both labeled and unlabeled tape
volumes. Only the first four bytes of the
status block portion of the TEBV table are
generated if TEBV is not specified also.
The status block is followed by the number
of error blocks specified by the n
parameter.

• Tape Error Block (TEB) by unit,
TEB=n.
• Error statistics by Tape Volume (ESTV),
TEBV=(DASD,n) or TEBV=(SYSLOG,n).

system Generation Guidelines

• Error Volume Analysis (EVA),
EVA=(r,w,n).

The FOPT macro checks the options in the
following order: TEB, TEBV, EVA. An
invalid specification for any option (n is
outside the range 0-254) results in NO
being assumed for that option. If TEB has
been validly specified, the TEBV n will
take on the value of the TEB n. If either
TEB or TEBV has been validly specified, the
EVA n will take on the value of the
previously set n. If any n does not have
the same value as an earlier valid n, an
MNOTE is issued.

Anyone or any combination of these options
may be selected. If more than one is
included, the n must be the same for each.
A TEB table is generated for the TEB
option~
A TEBV table is generated for the
ESTV and EVA options. Status information
contained in these tables is shown in
Figure 1.15 and 1.16. Figure 1.14 shows
the I/O interrelationship.

DOS System Programmer's Guide

NICL

Sy.em

BG
F2
FI

FICL

~
n

++-

n
n

LUB

JIB

~.-y

~r-

n = number of un its
i = index displacement
within LUB table

~~
~r-

r-r-

FAVP

(~
II'

TEBV

'--'-h

FOCL

PUB

I I I I I I I I I I I I 1
111111111111

I I I I I I I I I I L.I J
TEB

I I I II I

l1li
~r'::: ;:...-

L-::

::::;?

~~r
CHANQ
!>.

If'

FLPTR

...

KEY.

THFLPTR

r 1 1 1 1 1 1 1

(

I 1 I 1 1 I 1

SAB

TKHDTAB

1 I

l
LUB REQ
10
10

LUB TKREQ
DSP 10

~~~~

~;:--

If'

t.-::::;:;--

Jir=:?

i.-:~

NICL (Number in Class)

The first byte contains the number of system class units. The second, third, and fourth bytes contain the number of
programmer class units (BG, F2, Fl) (Figure 1.21).

FICL (First in Class)

The first byte points to the first system class unit in the LUB table. (Always the first LUB table entry.) The second byte
points to the first programmer class unit in the LUB table BG area. The third points to the first programmer class unit in
the LUB table F2 area. The fourth points to the first programmer class unit in the LUB table Fl area (Figure 1.21).

LUB (Logical Unit Block) Table

The first byte points to a PUB table entry (if the logical unit is assigned) or contains X'FF'. The second byte points to a
JIB table entry or contains X'FF' (Figure 1.21).

PUB (Physical Unit Block) Table

The first two bytes contain the channel and unit address of the physical device; the third a CHANQ pointer; the fourth
a TEB pointer; the fifth device type codes; the sixth a device characteristic code or a SAB pointer; the seventh the
channel scheduler flag; and the eighth has the job control flag (Figure 1 .30).

FOCL (First on Channel List)

The first byte points to the first PUB (highest priority) on channel zero. The next byte points to the first PUB (highest
priority) on channel one, etc. A hexadecimal FF indicates the associated channel is not supported.

TEB (Tape Error Block by Unit)

One TEB is built for each tape unit at supervisor generation time if tape error statistics by unit are required (Figure 1.15).

TEBV (Tape Error Block by Volume)

One TEBV is built for each tape unit at supervisor generation time if tape error statistics by volume or error volume
analysis are required (Figure 1 .16).

FAVP (First Available Pointer)

A one - byte pointer to the next available JI B entry.

JI B (Job Information Block)

The first two bytes contain extent or LUB information. The third contains ownership and JIB flags. The fourth contains
JIB chaining information (Figure 1.22).

CHANQ (Channel Queue) Table

The first byte contains the chain field (a pointer to the next in queue). The last three bytes contain the CCB address
(Figure 1.29).

LUBID (LUB Identification)

A ane - byte pointer to the LUB making the I/O request.

REQID (Requestor Identification)

A one - byte pointer to the program containing the CCB (Figure 1.29).

LUBDSP (LUB Displacement)

A one - byte value equal to the absolute LUB number (CCB byte 7).

FLPTR (Free List Pointer)

A one - byte pointer to the next free entry in the channel queue (Figure 1.29).

SAB (Seek Address Block)

A four - byte (BCCH) address that is the current disk address of the device plus a fifth byte that contains a Track Hold
Table pointer or X'FF'. If the Track Hold function is not supported, the fifth byte contains X'OO'.

TKHDTAB (Track Hold Table)

The first byte contains a pointer to the next available entry (or X'FF'); bytes 2 -4 have CCB address of the requesting
task; bytes 5 - 10 have disk address (BBCCHH) of track being held; byte 11 has key of owning track; and byte 12 has
two uses: bit 0 = 1 means a task is waiting for the track, and bits 4 - 7 count the number of holds on the track. Note
that the number of holds is one greater than the value v~ bits 4-7 of the last byte.

THFLPTR (Track Hold Free List Pointer)

A one - byte pointer to the next free entry in the Track Hold Table.

TKREQID (Track Requestor Identification)

A one - byte pointer to the PI B of the task requesting I/O.

Figure 1.1'1.

I/O Table Interrelationship

Section 1: Supervisor Planning Concepts

TEB Table
TEBTAB

TEB 1
TEB 2

TEB 4
Byte
TEB 5
TEB 6

1
2
3
4
5

Error recovery retry count.
Permanent read data check error count.
Number of times the read data check error routine is entered.
Number of times the write data check error routine is entered.
Write skip (erase gap) count.
Noise record count.

One TEB is generated for each 2400 series or 3420 magnetic tape or 2495 Tape Cartridge Reader unit if the FOPT macro contains the TEB = n parameter. Job control resets each TEB at normal or abnormal End-of-Job. An unused TEB contains
HEX'FFOOOOOOOOOO'. A TEB is referenced from byte 3 of a magnetic tape unit PUB.
Bytes 70 and 71 (X'46'-'47') of the communications region contain the address of the TEB table entry. Label TEBTAB
identifies the first byte of the table.

Figure 1.15.

Tape Error Blpck

DOS system Programmer's Guide

Decimal
Displacement

Label

Byte
Length

TEBVTAB

Description
Symbolic address of the TEBV Table

(TEBV Status Block portion of TEBV Table - Note 1)

o
1
r.---

2
3

4 ---

---~--6

7
11
12
17

------

TEBLEN
TSBLEN
EVARTH
EVAWTH
TEBsTAr -

1
1
1
1

-1".! _

Length of TEBV Error Block (for each Error Block generated)
Length of TEBV Status Block (4, 6, or 22 bytes - see Note 1)
EVA Read Error Threshold
EVA Write Error Threshold
-DASD ESTVFil;Stat;;s - - - - - - - - - - - - - - -

- -

- - - ---

!..E~~C _ _ _

~IY ~b~I"y~~e_C~~e!..

TEBDEV
UPXTNT
TEBRPT
NXTESR

Data Set Device Code
Disk Address of Upper Extent of Data Set (cchh)
Number of Records per Track
Disk Address of Next Available Space for Data Record (cchhr)
P~inter19 .E,.ST'yF.!:E la!?.elJ!l '{J~C ~c~rL _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

1
4
1
5
~ST'y~B.!: _ _ 5 __

_ _ _ _ _ _ _ _ _ _ _ _ _ _ ...., _ _ _ _ _ _ _ _ _ _ _ _ _

(TEBV Error Block Portion of TEBV Table - Note 2)
22

TEBV

1
1
1

25
26
27
28

29
30

32
34
1----------

1
1
2
6

Status Indicator (giving status of posting and writing error conditions)
Usage Indicator (X'OO'=TEBV Error Block in use; X'FF'=Error Block generated but not serving any tape unit)
Retry Counter
Permanent Read Errors
Temporary Read Errors
Temporary Write Errors
Erase Gaps
Noise Blocks
Permanent Write Errors
Clearer Actions
Number of Start I/Os
I-V~I~m~ S~~I Nu~b..:..r ~~~

JPl _______ _

- - --- -

- ------

(Begin repeating bytes 22 - 39 for second TEBV Error Block)
I
I
I

Note 1:

The TEBV Table is composed of one Status Block and (n) Error Blocks and is addressed symbolically by TEBVTAB. The table is
generated if EVA and/or ESTV are included in the system.
The size of the TEBV Status Block is determined by supervisor options in the FOPT macro at generation time:
• When EVA is chosen without ESTV, the TEBV Status Block is four bytes long (bytes 0-3), followed by TEBV Error Blocks,
so that bytes 4 - 21 are omitted.
• When ESTV output is on SYSLOG, the TEBV Status Block is six bytes long (bytes 0 - 5), followed by TEBV Error Blocks,
so that bytes 6 - 21 are omitted.
• When ESTV output is on DASD, the TEBV Status Block is 22 bytes long (bytes 0 - 21), followed by TEBV Error Blocks.

Note 2:

The number of TEBV Error Blocks generated corresponds to the (n) parameter in the FOPTmacro for TEB, TEBV, or EVA options.
A TEBV Error Block always contains 18 bytes, as shown in bytes 22-39 of this figure. The TEBV Table, therefore, is composed
of one TEBV Status Block (with the length dependent upon supervisor generation options as described in Note 1), followed by
(n) number of 18 - byte TEBV Error Blocks.

Figure 1.16.

TEBV Table Showing Status Block and Error Blocks

Section 1: Supervisor Planning Concepts

SEEK SEPARATE SUPPORT (SKSEP)
When DASD devices are command chained, they
monopolize the channel until channel end
for the device associated with the last CCW
in the chain is reached. This means that
the channel is unavailable for use by other
partitions requiring I/O operation of other
DASD devices on the channel.
The seek separation feature was designed
to improve the performance of systems
running under DOS. This feature enables
the supervisor to separate a seek from its
associated read or write so that the seek
can be separately scheduled. This means
that multiple seeks can be issued to
devices on a channel, and the reads and
writes scheduled as the seeks are
completed. As this is a supervisor
function, it automatically applies to
programs written at any language level
and/or operated in any of the three
partitions. As the number of devices on a
channel increases, the benefits derived
from this feature increase. The
implementation of this feature is such that
when a seek has been issued to a device,
the arm cannot again be shifted until the
I/O operation that initiated the seek has
been completed. In other words, arm
stealing has been prevented. After seek is
issued, the channel is available for
scheduling any other I/O operations. In a
multiprogramming environment, this feature
is particularly important when the
different partitions have a mix of
input/output requests for a single channel
with multiple direct access devices.
The seek separation capability is
provided by unchaining the user's seek and
by posting the seek separation bit (X'10'
of the 13th byte of the user's CCB) and the
"Wait for Device End" bit in the user's
CCB. When START I/O is performed on the
new unchained seek, channel end is
immediately sent back from the control
unit, thereby freeing the channel for
performing seeks to other units on the
channel. When the device end interrupt
occurs at the completion of a seek, the
seek separation and ·Wait for Device End"
bits are turned off, the seek is rechained
to the remaining CCWs for the I/O request,
and the device is not dequeued. A START
I/O is now performed on this unit that has
its arm already positioned at the correct
cylinder.
Specifying this feature generates a Seek
Address Block (SAB) within the supervisor
(refer to Figure 1.14). Each DASD device
has an entry in the SAB containing the
current disk address for that device. Each
time the user performs a seek, his seek
address is compared with the entry for that
62

DOS System programmer's Guide

device in theSAB. If they are equal, no
seek separation is performed. If they are
not equal, the seek address replaces the
current entry in theSAB, and seek
separation is performed. Each user's DASD
ccw chain must start with a long seek
(X'07') in order to use this feature.

System Generation Guidelines
Specifying SKSEP=YES indicates support for
all DASD devices specified by the DVCGEN
macro at system generation time. N is the
number of DASD devices to be supported and
cannot be less than the number of DASD
devices specified at system generation.

PHYSICAL TRANSIENT OVERLAP SUPPORT (PTO)
Physical Transient Overlap (PTO) support
allows the system to:
• Overlap I/O operations in the error
recovery routines (ERP) with problem
program processing.
• Overlap I/O time required to fetch user
programs and system transient routines
into main storage with problem program
processing.
The PTO option allows other tasks to be
selected when an I/O operation is being
performed during one of the following:
• Fetching a phase (SVC 1)
• Fetching a logical ($$B) transient (SVC
2)

• Fetching a physical ($$A) transient
(SVC 3)
• Loading a phase (SVC 4)
• Loading a phase header (SVC 23)
• Modifying the system communications
region via MVCOM macro (SVC 5)
• Fetching of the ATTN routine
• performing ERP I/O operations
For example, if an error occurs in the
background partition while reading a tape
record, an error recovery routine is called
into the physical transient area to reread
the record up to 100 times. Each reread
operation requires time to read the record
itself plus time to backspace the record
before reading it again. Without PTO, the

entire 100 rereads by the error recovery
routine would not be overlapped with other
processing and other tasks would not
receive control during this interval. with
PTO, I/O time is overlapped and available
to the foreground partitions for processing
until another error recovery routine or
system transient routine is required.
Another advantage of PTO involves fetch
I/O time. Fetching requires one or more
searches of the library directory on
SYSRES, reading of the directory into the
physical transient area, and searching for
and reading one or more records of the
program or transient into the proper area
of main storage. This involves several
disk I/O operations. PTO allows the fetch
I/O time to be overlapped with processing
by any partition requiring CPU time (in
priority sequence) until another fetch or
other use of the physical transient area is
required.

Data chaining and command chaining are
not used.

The CCB associated with the operations
does not request any sense information
(CCB+12 set to X'20') and/or does not
have either "accept unrecoverable I/O
error", "post at device end", or "user
error routine" bits on (CCB+2 set to
X'15' or any combination of these).
In other words, no error checking is
required.

The CCW must have either an X'Ol' or
X'09 1 command code (i.e., only a WRITE
operation).

For user messages, the CCW and the
message must reside within the
partition issuing the write operation.

If these conditions are not met, the
console operation is handled in a normal
manner.
CONSOLE BUFFERING SUPPORT (CBF)
Previously, a system component could Qold
up the system and degrade throughput
performance whenever an I/O operation was
issued to the console typewriter (1052),
followed by a WAIT. The console buffering
option (CBF) alleviates this situation by
queuing write operations to a 1052, 3210,
or 3215 and returning control immediately
to the routine that issued the write with
the WAIT bit posted. The user's WAIT is
satisfied immediately, rather than at
actual I/O completion. Each message is
assigned a buffer containing the message
itself and control blocks for I/O interrupt
handling (Figure 1.17). The system writes
the message from the buffer on the 1052,
3210, or 3215 as soon as possible while
user-processing continues.
The CBF facility is used only for write
operations which require no user error
handling. Only those messages that meet
the following criteria are -buffered:
1.

Messages cannot exceed 80 characters.

You have the option of specifying the
number of buffers required for queuing
messages, but the total may not exceed 50.
If a buffer is not available, your WAIT is
not satisfied until a buffer is freed. At
least one buffer should be specified for
each partition or task issuing messages so
that buffers are available and the task can
continue processing while the message is
being printed. Job control often issues
several console messages in succession;
therefore several additional buffers should
be specified for partitions that use job
control frequently.
Another factor to be considered in the
selection of the CBF option is extra
channel usage. Specification of the CBF
option and selection of the CBANQ default
(6 channel queue entries), results in the
number of buffers specified being added to
the CHANQ default. However, when both the
CBF and CHANQ options are specified, it is
recommended that the number of CHANQ
entries desired be increased by the number
of buffers specified. Otherwise, the
number of entries generated in the channel
queue will be less than desired.

)
Section 1: Supervisor Planning Concepts

CBTAB

xx
CBINUSE

7 8
CBCCW

,,/"

CBCCB

20 21
23 24
CBPREX

103
CBDATA

3
r---C-B-N-E-X-T.......( /
Po i nts to next
available buffer

CBTABND

The Buffer Table is a 104 multiplied - by- n byte area of main storage, where n is 1 -50. Each buffer entry in the table is 104 bytes.
CBNEXT is a halfword constant that points to the next available buffer entry. It is initialized with the address of CBTAB and is
incremented by 104 every time a buffer is used, so that it points to the next entry. When its val ue becomes greater than CBTABND,
it is reinitialized with the value CBTAB.
CBINUSE is a one- byte counter that contains the number of entries currently in use. It is incremented whenever a buffer is used and
is decremented at dequeue time when the buffer becomes free.
Each buffer entry contains the fol lowing fields:
Displacement:

a CCW moved

from requestor core. The data address portion of the CCW is modified to
point to the data portion of the buffer.

8 CCB moved from requestor core. The CCW address in the CCB is modified to point to
the CCW in the buffer.
21 Prefix. SYSLOG ID moved from problem program PIB. The prefix is printed with the
data to identify BG, F1, or F2 partitions.
24 Data moved from requestor core.

Figure 1.17.

I INDEPENDENT

Console Buffering Table and Work Areas
DIRECTORY READ- IN AREA (IDRA)

An independent directory read-in area
(IDRA) can be included if the option is
specified at system generation time. The
funcion of this area is to enhance
performance by reducing contention for the
physical transient area (PTA) on fetches'
from the core image library. During the
time the physical transient area is busy,
fetches from other partitions can be
processed.

system Generation Considerations

The YES option causes generation of the
IDRA area in the supervisor. IDRA requires
MPS=YES or BJF, and PTO=YES.
If IDRA=NO, the core image library
directories are loaded and scanned in the
64

DOS System Programmer's Guide

PTA. If IDRA=YES, the core image library
directories are loaded and scanned in the
IDRA. The IDRA is used for all supervisor
calls that require reading of directories.

COMMAND CHAINING SUPPORT (CCHAIN)
Command chaining support (CCBAIN=YES)
allows the DOS error routines to retry ~n
I/O operation starting with the last ccw
executed rather than at the beginning of
the chain. Under normal error recovery
procedures, the entire CCW chain is
reexecuted. To utilize the CCHAIN support,
you must indicate in the CCB macro
instruction that you want the error
procedures to retry the I/O operation
starting with the CCW in error by setting
byte 3, bit 7 on (third operand of
CCB=X·0001'). In addition, bytes 8-11 of
the CCB must be initialized with the
address of the first CCW in the chain each
time the chain is executed. See Figure
1.18 for an example using CCHAIN support.

r------------------------------------------------------------------------------,
BEGIN
START 0

READCCB
READ

BALR

4,0

USING

*,4

1, READ

1,READCCB+8

EXCP

READCCB

WAIT

READCCB

•
•

•
•
BCT
•
•

Initializes Bytes 8-11 of CCB
with address of first CCW in
chain each time chain is executed.

3,DO

CCB

SYS004,READ, X'0001 1

CCW

2,DATA,X'60',100

CCW

2,DATA+l00,X'60',100

Indicates Command Chain Retry.

CCW
2,DATA+200,X'20',100
- - - -__________________________________________________________________________
J
Figure 1.18.

Example Using CCHAIN Support

CCHAIN=YES must be specified if data or
command chaining of IBM 2495 Tape Cartridge
Reader input is performed.

• DTFSD - updating without work files
• DTFSD - other functions with work files
• DTFDA - all functions

TRACK HOLD SUPPORT (TRKHLD)
The track hold feature allows tasks within
the same partition or tasks outside of a
partition to-share DASD data files when
using the direct access method or the
sequential method of file organization.
This track protection facility can be used
with or without multitasking to provide
protection between partitions. By
definition, track protection means a DASD
track that is being modified by some task
in main storage is prevented from being
accessed by another task in that partition
or in any other partition.
Any programs using DTFSD/SDMODXX,
DTFIS/ISMOD, or DTFDA/DAMOD can use the
track protection macros when performing the
following functions.
• DTFIS - all functions except LOAD
• DTFSD - updating with work files

The actual holding of a track is a
combination of supervisor (PIOCS) and data
management (LIOCS). Track hold involves
the actual request that a track be held and
the request to free a track. The hold
request is generated if you specify
parameters in the DTFs and logic modules
for the program(s) involved.

Supervisor Considerations
At system generation time, you must specify
the number of unique tracks that can be
held at anyone time (255 maximum and 10
minimum). A table for the number of
entries specified is built into the
supervisor for storing the necessary
information required for track protection.
The maximum number of holds (without
intervening frees) that a task can issue
for a given track is 16. If more than 16
holds are attempted, the task is canceled.
Section 1: Supervisor Planning Concepts

If the DASD channel program does not
start with a long seek, track hold is not
implemented.
If multiple track search operations are
issued, only the first specified track is
held but it is not necessarily the same
track on which the search is satisfied. In
this case, the track on which the search is
satisfied is not protected.
If a task requests a track already held
by another task, the requesting task is
placed in wait state (until the track
becomes available). In addition, if the
task requests a hold on the 256th unique
track or exceeds the limit specified at
system generation time, the requesting task
is placed in wait state until a previously
held track is freed. In either of these
cases, especially in the first situation
which has a greater chance of occurring
than the latter, make sure that the
following conditions do not occur.
1.

If more than one track is being held,
it is possible for a task to put the
entire system into wait state. This
occurs if one task is waiting for a
track that is already held by another
task. In the first case, if task A
holds track 1 on drive 1 and attempts
to hold track 2 on drive 2 while task
B is holding it, task A is placed in
wait state until task B frees that
track. Then, if task B attempts to
hold track 1 on drive 1, it enters the
wait state until task A frees that
track. Because task A is already in
wait state, neither task ever regains
control. If they are the only two
tasks running in the system, the
entire system is in wait state.
A similar situation can occur if two
or more tasks attempt to hold more
tracks than the maximum number
allowed. For example, assume a
maximum of 20 unique tracks can be
held by the system. Task A issues 15
DASD I/O operations and task B issues
15 DASD I/O operations (altogether
requiring track hold for 30 unique
tracks). If both task A and task B
are alternately issuing DASD I/O
operations, then when each task issues
its 11th track hold (at this point
both task A and task B have held the
maximum of 20 unique tracks), it is
placed into wait state. Again,
neither will regain control, and if
they are the only two tasks running in
the system, the entire system is in
wait state.

In both examples, if the operator is not
aware of what has occurred, he may decide
to cancel the tasks. If the lockout is
66

DOS System Programmer's Guide

between two subtasks within a partition,
the partition is canceled. If the lockout
is between two partitions, one partition is
canceled, its tracks are freed and the
second partition is removed from wait
state.
When a task issues a free for a
particular track, one of two conditions
occurs:
1.

If a task has more than one hold on
the track (16 maximum), a counter
associated with that track is
decreased by one, the task retains its
hold on the track and control is
returned to the task issuing the free.

If the task has only one hold on the
track, the track is'-freed by removing
its entry from the supervisor table.
Any tasks waiting for the track or an
available entry in the table are
removed from wait state. The task
with the highest priority gains
control of that track or puts a hold
on another track.

Track hold only occurs for DOS-supported
DASDs. Otherwise, the I/O operation is
issued, but no track-hold function is
involved. (This is true for any device,
including non-DASDs.)
The following conditions cause
cancelation of a task:
1.

A free was issued to a non-DASD type
or a nonsupported DASD,

Bolds were not issued to that track,

A task attempts to free a track held
by another task.

In the last case, when the main task
detaches a subtask, any track held by that
subtask is freed.
LIOCS Considerations
Track hold is only supported for DTFIS,
DTFSD and DTFDA files. Support of the
track hold feature requires specification
of a HOLD parameter in both the
DTFIS/DTFDA/DTFSD and the ISMOD/DAMOD/SDMOD
macros. In addition, a FREE macro must be
issued for all DTFDA or DTFIS files and for
DTFSD work files without update.
A maximum of 16 holds can be applied to
one track. A free issued for a track with
more than one hold decrements the track
hold count, but does not release the track.
A free issued for a track not held,
terminates the task.

The track hold functions can be used in
five specific situations:
1.

Using DTFSD for updating with work
files (via the WRITE macro).

Using DTFSD for updating without work
files (via the PUT macro).

Using DTFSD with work files, when no
WRITE is issued for updating.

For all DTFDA files.

For all DTFIS files except for LOAD.

information on the format and use of the
STXIT macro.

STXIT Macro Considerations

In the first two situations, the tracks
being held are freed automatically by the
system. For the last three situations, the
task must issue the FREE macro instruction
for each hold placed on a track by the READ
macro. It should be noted that for DTFDA,
a FREE macro need not be given wben only
WRITE macros are issued. If BOLD=YES and
ERREXT=YES, you must issue tbe ERET macro
to return to the ISAM module to free any
held tracks.

If an IT (Interval Timer) condition occurs
while executing the OC (Operator
Communication) or PC (program Check)
routines, these routines should not use the
same save area (see Figure 1.19).

r-----------------------------------------,I

r-------------~-----------------------,I
IRoutine Being I Condition Occurring
II
I Processed
r-----,._---~---,._- I
I
I AB I IT , OC I PC I
~------------+------+------+------+----~

I
I
I

lAB

I
I IT
I

T
S

10C

L
____________ L-_____
IPC
S

I
I
I
I
I
:

I
I

I
I
I
I
I

H :

I
B

I
I
I
I
I

IbEf:

I I
H

I
I
I

I ______
H
I ______
B I ____
T JI

System Generation Guidelines
Ef -

Error message issued in foreground
program and control returns to
interrupted OC routine.

B -

Condition honored. When
processing of new routine
completes, control returns to
interrupted routine.

Condition ignored for all
partitions.

Interrupt ignored in a background
partition.

TRKBLD=n
The maximum number of tracks that can be
held at one time is 255. The default is 10
if n is an invalid parameter (nonnumeric or
outside the range 1-255). MPS=YES or BJF
must be specified if TRKHLD=n. If a task
attempts to hold an additional track after
the maximum has been reached, the task goes
into the wait state until a previously held
track has been freed.

S -

ABNORMAL TERMINATION SUPPORT (AB)
Abnormal termination exits are available
for main tasks and/or subtasks, allowing
you to gain control before an abnormal
condition removes the task from the system.
For example, in the abnormal termination
routine, you can close your files. This
function is provided by the AB operand of
the STXIT macro. See Supervisor and I/O
Macros listed in the Preface for detailed

Execution of the routine being
processed is suspended and control
is transferred to the AB routine.

T -

Job abnormally terminated. If AB
routine is present and it has not
been interrupted by itself, its
exit is taken; otherwise, a system
L--______________________________________
_
abnormal termination occurs.
Figure 1.19.

Processing of STXIT
Conditions

Section 1: Supervisor Planning Concepts

when the time elapses. The SETIME,
STXIT, and EXIT macros are used for
this. However, only the main task of
the partition owning the timer can
issue these instructions.

When a supervisor is not generated to
handle the requested facility, the task is
abnormally terminated if a STXIT macro is
used. This also applies to a program that
requests the timer interrupt and has not
been allocated the timer.
If an AB or PC condition occurs and
linkage has not been established for a main
task abnormal termination routine, the
partition is abnormally terminated.
However, if the AB or PC condition occurs
in a subtask without exit linkage
established, only the subtask is
terminated.
Only one main task at a time can use the
STXIT IT macro. The partition of the main
task is specified at system generation time
but can be changed by the operator with the
TIMER command. There are two distinct
methods for using the STXIT IT macro. Only
' one method can be used at a time.
1.

The first method allows the main task
to set the timer and enter a routine

DOS System Programmer's Guide

The second method allows a given
routine to be performed at timed
intervals. The time set is a real
time interval, and is not stopped or
adjusted when the task using the timer
does not have control. The SETIME,
TECB, and WAIT macros are used.
However, if multitasking, only one
task within the partition can use the
method and the WAIT on the TECB must
appear in that task. Consideration
should be given to the priority of the
task assigned to process interval
timer interrupts. Subtasks can use
this method.

Figure 1.20 contains a coding example
using STXIT support.

PAGE
LOC

OBJECT CODE

ADDR1 ADDR2

STMT

SOURCE STATEMENT

1 STXEXAM

000000
000000 0520
000002

00005A 9263 2058

~005A

00005E 41FO 2014

00016

23
21
33
39
43
49
50
51
52

FOOS CL 3-5 10130169

CHK

START 0
BALR ?oIO
USING .',2
PRINT NDGEN
STXfT PC,PRGeK,PCSAVE
STXIT OC,COMMUN,OCSAVE
SETIME 10,TIMEBLK
EXCP CCB1
WAIT CCB1
WAIT TIMEBLK
EXCP tCB2
WAIT CCB2
MVI
*,99

PRGCK

B
EXCP

GOAGAIN
CCB3

PRINT PROGRAM CHECK MESSAGE

WAIT
EXIT
EXCP

CCB3
PC
CCB4

RETURN FROM PROGRAM CHECK ROUTINE
PRINT EXTERNAL INTERRUPT MESSAGE

WAIT
STXIT
STXIT
EXIT
EOJ
OS
OS
TECB
CCB
CeB
CCB

CCB4
PC
DC
OC

DEACTIVATE PROGRAM CHECK EXIT
DEACTIVATE EXTERNAL INTERRUPT EXIT
RETURN FROM EXTERNAL INTERRUPT RTN

2
3
4
5
11
17 GOAGAIN

*LOOP

Acn,VATE US ERS PROG CHECK RTN
ACTIVATE USERS OPER.COMM. RTN
SET INTERVAL TIMER TO 10 SECONDS
PRINT TIMER BEGINNING MESSAGE
WAIT FOR TIMER INTERRUPT
PRINT TIMEOUT MESS_GE
THIS WILL CAUSE A PROG CHECK
NEXT TIME THROUGH THIS RTN

THE

(See ~below)

56
62
65 COMMUN
(See~

69
15
80
85
88
91
92
93
96
107
118
129
140
141
142
143
144

00009C
0000E4

000110
000118
000180
000188

0900019000000064
090001F400000064
0900025800000064
090002BCOOOOO064

000190 E3C8C540E3C904C5
0001F4 E3C8 C9 E240 03C905
000258 E3C8C9E240D3C9D5
0002SC
000000
000320
000324
000328
00032C
000330
000334
000338

Cl E240C14009C5E2

0000009C
00000062
000000E4
00000018
OOOOOOOA
00012COO
0000012C

below)

----- -- ---- - - -- -

DONE
PCSAVE
OCSAVE
TIMEBLK
CCB1
CCB2
CCB3
CCB4
CCW1
CCW2
eCW3
CCW4
MSG1

cca

CCW
CCW
CCW
CCW
DC

145 MSG2

146 MSG3

141 MSG4

148
149
150
151
152
153
154
155

END

CL12
CL 12
SYSOOO,CCW1
SYSOOO,CCW2
SYSOOO fCC W3
SYSOOO,CCW4
9,MSG1,X'00',100
9,MSG2,X·OQ',100
9,MSG3,X'00' ,100
9,MSG4,X~00',100

CL100'THE TIME HAS JUST BEEN SET TO A VALUE OF TEN SECONO
DS,THIS LINE lS PRINTING WHILE THE TIMER IS RUNNING.'
CL100'THIS LINE HAS PRINTED AS A RESULT OF A TIMER INTERX
RUPT IE. THE TIMER VALUE HAS GONE THROUGH ZERO.'
CL100'THIS LINE HAS PRINTED DUE TO THE FORCED PROGRAM CHX
ECK AT LABEL *CHK*.'
CllOO'AS A RESULT OF THE INTERR"UPT KEY. THIS PROGRAM HAS X
GONE TO NORMAL END OF JOB;'
STXEXAM
=A IPCSAVE)
=AIPRGCK)
=A(OCSAVE)
=A ICOMMUN)
=A(lO) ,
sF '16800'
=AITIMEBLK)

- -=
:::::=:
-::-: ..:::::
- - - - - - - - - - - - - - - - - - - - - -- - - - - - --- - -- - - - - - - - - - - -- ---;..

PAGE
LOC
00033C
000340
000344
000348

OBJECT C(10E
00000130
00000140
00000150
00000160

[

AOOR1 ADDR2

-----

STMT

SOURCE STATEMENT
=A(CCB11
=AICCB2)
=A(CCB31
=A(CCB4}

156
151
158
159

FOOS CL3-5 10130169

:=;..-

Note: If LlOCS or other CSECTS that might alter the base register are used, then' a base register must be establ ished as the first step
in the interrupt routine. For example:

BALR

3,0

USING *,3

Figure 1.20.

STXIT Sample Program

Section 1: Supervisor Planning Concepts

second most preferable event. The user may
use this to his advantage in task
synchronization. The system checks the
ECBs in the order written in the WAITM
macro.

MULTIPLE WAIT SUPPORT (WAITM)
The multiple wait facility allows a task to
be put into the wait state until one of a
number of events occur before proceeding.
Control passes to a task when one of the
ECBs specified in the WAITM macro has been
posted. The multiple wait facility can be
used without specifying AP=YES. However,
AP=YES cannot be used without WAITM=YES
being specified.
An Event Control Block (ECB) is a task
communications block written by the problem
programmer (see Figure 1.11). When a
subtask is successfully attached, byte 2
bits 0 and 1 are set to o. When the
subtask terminates, the supervisor posts
(sets a bit to 1) byte 2 bit 0 of the ECB
after processing of the abnormal
termination routine. In addition, byte 2
bit 1 is posted if the subtask abnormally
terminates; that is, if task termination is
not the result of issuing the CANCEL,
DETACH, DUMP, or EOJ macro instructions.
See supervisor and I/O Macros listed in the
Preface for more detailed information.

DASD FILE PROTECT SUPPORT (DASDFP)

The DASD file protect facility provides
read/write file protection for DASDs. All
seeks issued by a problem program are
monitored by the supervisor to ensure that
the seeks do not stray outside of limits
that were validated at file open time.
(This could occur during the loading of a
direct access file.) DASDFP functions on a
cylinder basis for the 2311, 2314, and
2319, and on a strip basis on the 2321.
DASDFP routines are not executed when the
current PSW has a storage protect key of
zero. This occurs for supervisor, job
control, and transient routines ($$A, $$B,
$$R).

Mode of Operation
Multiple Wait Considerations
When control passes to the waiting task,
its register 1 points to the address of the
ECB that had byte 2, bit 0 set.
The user may use CCBs or TECBs in place
of ECBs. This is possible because the
posting of byte 2, bit 0 occurs upon their
event completions. BTAM/QTAM ECBs, QTAM
control blocks, and RCBs cannot be waited
on because their format would never satisfy
a WAITM, i.e., byte 2, bit 0 will never be
posted. When waiting on only one ECB, the
WAIT macro can be used.
When using WAITM, a provision should be
made for ultimate outlet if none of the
events being waited on occur. For example,
an abnormal ending of all the tasks on
which a multiple wait was dependent would
never allow the multiple wait to be
satisfied.
The user may specify a preferred event.
Consider the f~llowin9 example:
WAITM

ECBA, ECBB, ECBC.

ECBA is considered the preferred event
because it appears first on the list. For
the same reason, ECBB is considered the

DOS system Programmer's Guide

Job control or the foreground initiator
reads the DLBL/EXTENT (or VOL/DLAB/XTENT)
control commands or statements from the
input job stream. This information is
reformatted and written on the label
information cylinder on SYSRES as one or
more DASD records. Each of these DASD
records is preceded by a key created from
the filename (DTF name) entry in the
DLBL/VOL command or statement.
When the problem program opens the file
(DTF), the OPEN transient routine extracts
an a-byte alphameric constant from the
generated DTF table. This constant is
called the filename, and is an exact
replica of the DTF name.
The OPEN transient routine searches the
label information cylinder on SYSRES to
find a key equal to this DTF filename
constant. When it finds a match, it reads
the data portion of this job control DASD
record into the label save area in main
storage.
The extent information is extracted from
this job control DASD record, and is stored
in the JIB table. A 2311/2314/2319 extent
requires one JIB entry; a 2321 extent
requires two JIB entries (see Figures 1.21
and 1.22).

(I)

.
I-'

Number in
Class List
(NICL)

LUB Table

First
Class List
(FICL)

LUB Table for any Partition
--,-

II.)

I-'

\ SYSRDR

CDiSYSPCH

SYSLST

(')

S»

.....

• • • • • • • • • • • 112
Main Storage Requirements
• • 113
Modular/Tabular System • • • • • • • • 113
DTF (Define the File) Macros • • • • 113
MOD (Module Generation) Macros • • • 114
Reentrant Modules • • • • • • • • • 118
Interrelationships of the DTF and
Module Macro Instructions
• • 118
Imperative Macros • • • • • • • • • .119
Sequential File Organization.
• .122
Card Files • • • •
• • • • • • 122
Console Typewriter •
• • • • • .122
Line Printers
• • 122
Paper Tape Files • •
• • • 122
Magnetic Tape Files
• • 122
sequential Disk
• • 123
Direct Access Method (DAM) File
Organization • • • • • • •
• .124
Random Addressing Techniques • • • • 124
• • .125
Random Addressing Formula
Prime Number Division
• .126
Synonym Records • • • • •
• .127
DASD Address Specification For
Read/Write Operations
• • 128
Relative Track Addressing
• • • 128
Indexed sequential File Management
System (ISFMS) • • • • • • • • • • • .129
Loading an Indexed Sequential File .131
Adding Records to the File • • • • • 132
Random Retrieval • • • • • • • • • • 132
Sequential Retrieval • • • • • • • • 132
Choosing the Right File Organization
and Retrieval Method • • • • • • • • • 133
Criteria. • •
• • • • 133
Sample Files.
• .136
Summary
• • • •
.138

Section 2:

Da~a

Management

Section Figures

Figure 2.1. Fixed Length Unblocked
Record Format
• • • • • • • • • • • 94
Figure 2.2. Fixed Length Blocked
Record Format
• • • • •
• 95
Figure 2.3. Variable Length Unblocked
Record Format • • • • • • • • • • • • • 97
Figure 2.4. Variable Length Blocked
Record Format • • • • • • • • • • • • • 98
Figure 2.5. I/O Operation Using PIOCS 101
Figure 2.6. Command Control Block
(CCB) (Part 1 of 3)
• • • • • • • 104
Figure 2.7. Format of the CCW
.106
Figure 2.8. Flowchart for EXCP Coding
Example
• • • • •
• • • •
.109
Figure 2.9. Physical IOCS Sample
Program
• • • • • • • • • • •
.110
Figure 2.10. Retrieving a Record
Using Logical IOCS • • • • • • • • • • • 112
Figure 2.11. Generated Name Structure
for Logic Modules (Part 1 of 2)
.116
Figure 2.12. DOS Relocatable Library
Module Name Prefixes
.118
Figure 2.13. DTF and Module Macro
Relationships
• • • •
.118
Figure 2.14. Coding Example of DTF and
Module Relationship • • • • • • • • • • 119
Figure 2.15. Logical IOCS Imperative
Macros and DTFs • • • • • • • • • • • • 120
Figure 2.16. Logical IOCS Imperative
Macros and Devices
• • • • • • • 121
Figure 2.17. Direct Access Address
Chaining • • • • • • • • • • • • • • • • 127
Figure 2.18. DASD Address Formats • • • 128
Figure 2.19. Example of Track Index
.130
Figure 2.20. Example of Track Index
after Addition to File. • • • • •
.131
Figure 2.21. File Organization on a
Disk/Tape System • • • • • • • • • • • .135
Figure 2.22. Indexed Sequential
Versus Sequential File Organization • .137

DOS System programmer's Guide

Concepts of Data Management
This section is written for the reader who
seeks a general understanding of the
functions and concepts of data structures
and storage media. It explains the basic
concepts of data organization and defines
some of the terms that will be encountered
in subsequent sections of this and other
IBM publications.
The experienced programmer may find this
section of interest, but i t is written
primarily for the reader whose experience
with tape and disk files is limited.

DATA FILES AND RECORDS
Data files stored on such media as paper,
cards, tapes, or disk storage devices, are
encountered in practically every business
activity. These files provide the basis
for most manual, mechanical, and electronic
data processing. Data files are composed
of individual records ranging from a few
records up to thousands or millions of
records.
A record can be defined as a collection
of information comprised of alphameric
and/or nonalphameric characters related to
a common identifier. The common identifier
is known as a record's control field or
key. Usually one of the prime information
elements (fields) present within a record
is used to identify the record. For
example, man number could be used as the
key or identifier for a payroll record, and
policy number could be the key of an
insurance policy file.

The size or length of records varies
from file to file because the size can
range from a single character up to
thousands of characters.
A data field is a sequence of one or
more characters treated as a processing
unit of information. A single record
usually includes one or more logical data
fields. An individual data field is
normally identified by its location within
a record.
The logical structure of records and of
fields within records has become
increasingly important since the advent of
computers and high-speed recording media
such as magnetic tapes and disks. This
logical structure is strongly affected by
whether a record is fixed or variable
length.
Data records of an ASCII tape file may
include a block prefix and/or padding
characters. These options are in addition
to the fields contained in data records
written in EBCDIC mode. When present, the
block prefix is the first field of a
physical record, and it may be 0-99 bytes
long. DOS can use this field to check the
length of the physical record only with
ASCII variable length records. For fixed
length and undefined records, DOS ignores
the block prefix on input and does not
restore this field on output.
The length of an ASCII physical record
includes the block prefix field and the
number of padding characters. Padding
characters ensure that all blocks conform
with the required length. DOS accepts
these padding characters (corresponding to
EBCDIC X'SF') on ASCII input but does not
perform any padding operation on output.

)
section 2: Data Management

Fixed lenqth,unblocked (Figure 2.1): Each
logical record is the same length as the
physical record.
For ASCII fixed length, unblocked
records, the block prefix and padding
characters are optional. Regardless of the

IRG

= Inter-Record Gap

presence of these optional fields. ASCII
fixed length with only one logical record
per physical record is considered to be
unblocked. The number of padding
characters must always be less than the
size of the logical record.

TAPE RECORDS

Physical
Record

IRG = Inter-Record Gap
BLPR = Block Prefix (optional)
PADDI NG = Padding Characters (optional)

= Gap,

ASCII TAPE RECORDS

Data Area

Area Separator

Data Record

DISK RECORDS - WITH KEY AREA

Data Record

G = Gap, Area Separator

Figure 2.1.

Data Record

DISK RECORDS - WITHOUT KEY AREA

Fixed Length Unblocked Record Format

c
94

DOS System

P~ogrammer's

Guide

Each physical record in ASCII fixed
length, blocked format may contain a block
prefix and/or padding characters. The
number of padding characters must be less
than the size of the logical record.
Physical records are deblocked until the
number of bytes remaining in the physical
record is less than the specified logical
record length.

Fixed length, blocked (Figure 2.2):
Blocked records are usually considered to
be two or more logical records within one
physical record. The number of records in
each block (blocking.factor) is usually
kept constant. For example, the
illustrations show blocked records with a
blocking factor of 3 meaning there are
three logical records within each block
(physical record).

logical
Record
1

IRG

Physical Record

= Inter -

logical
Record
2

rPhysical Record

logical
Record

r-PhYSiCal Record 3 - - 1

logical
Record
7

'""".1----- Physical

Record l - - - - I..~I

logical
Record
1

logical
Record
2

'"""1.~--PhYSiCal

Record 2 - - - - 1..~1

logical
Record

log i ca I Record
Data Area
1

logical Record
Data Area
2

Area Separator

Figure 2.2.

logical
Record

logical
Record
7

Logical
Record

logical
Record

'"""14. . . . . - - - - - - Physical Record 2 ------~
.. I

logical Record
Data Area

log i ca I Record
Data Area

DISK RECORDS- WITH KEY AREAS

Area Separator

log i ca I Record
Data Area
1

= Gap,

logical Record
Data Area

Physical Record 1-------.....,~~

logical
Record
10

ASCII TAPE RECORDS

I'""". f - - - - - - - - - - Physical Record 1 - - - - - - - - - - - 1.... 11

= Gap,

logical
Record

TAPE RECORDS

Record Gap

IRG = Inter- Record Gap
BlPR = Block Prefix (optional)
PADDING = Padding Characters (optional)

Logical
Record

logical Record
Data Area
2

,..I.f-----~---="-'---Physical Record 2-------~.1

logical Record
Data Area

Logical Record
Data Area

log ica I Record
Data Area
5

logical Record
Data Area
6

DISK RECORDS- WITHOUT KEY AREAS

Fixed Length Blocked Record Format

)
Section 2: Data Management

Variable length, unblocked (Figure 2.3):
Each physical reco~d contains one logical
record that can vary in length. Each
record must contain both a block length
field (BL) and a record length field (RL)
giving the size of the block and the size
of the logical record respectively. The
first two characters (XX) of the block
length field (BL) specify the actual block
length in 16-bit binary form. The la~t two
characters (indicated by bb) are binary
zero. For variable length unblocked
records, BL specifies the logical record
length plus 4 bytes (the size of BL) •.

The first four bytes following the block
length field must contain the record length
field (RL). The first two bytes (XX)
specify the length of the logical record
including the bytes used for RL field
itself. The remaining two bytes (bb) are
binary zero.

For ASCII variable length, unblocked
records, each physical record may contain a
block prefix. If the block prefix is four
bytes long, it may contain the length of
the physical record that DOS can use for
checking purposes. Each logical record
must contain a record length field (dddd)
giving the size of the logical record in
unpacked decimal format. If a padding
character is found in the first position of
a record length field, all remaining bytes
in that block are bypassed, and the next
logical record is retrieved from the next
block.
Variable length, blocked (Figure 2.4): One
or more logical records are contained
within each physical record. The first
four bytes (block length field) of each

physical record (block) specifies the total
number of bytes in the block. The first
two bytes (XX) specify the length of the
block (including the four bytes for the
block length field itself). The remaining
two bytes (bb) are blank. The size of each
logical record must be placed in a record
length field (RL). The RL must be the
first four bytes of the logical record.
The first two bytes (XX) of RL specify the
length of the logical record including the
bytes used for the RL field. The remaining
two bytes (bb) are binary zero.

For ASCII variable length, blocked
records, each physical record may contain a
block prefix. For checking purposes, the
block prefix may contain the length of the
physical record (in unpacked decimal
format) if the block prefix is four bytes
long. The length of each logical record,
also in unpacked decimal format (dddd),
must be placed in the first four bytes of
the logical record. If a padding character
is found in the first position of a record
length field, all remaining bytes in the
block are bypassed, and the next logical
record is retrieved from the next block.
Undefined: When file records do not
conform to any of the four previous
formats, they are classified as undefined.
For example, any variable length record not
conforming to IBM's variable length format
is considered undefined.

ASCII undefined records may include a
block prefix and/or·padding characters.
DOS will not attempt to distinguish padding
characters from the physical record. The
entire physical record is passed on to the
user.

(
96

DOS System Programmer's Guide

Ifr-

Phys;col Record

Log ical Record ~
Data

Data

Record 1

Data

7
RL

Record 3

Record 2

= 80

103
RL

= 100

= 50

53 11
I

BL = Block Length } 'm b'mary ha Ifwo rd ( 16 b')
It f ormat, p I us two bl an k bytes
RL = Reco rd Lengt h
IRG = Inter- Record Gap
G = Gap
TAPE RECORDS

Physical R e c o r d - l

1"-- Logical Record-l I
Data

Data

Record 1

Data
Record 3

Record 2

IRG = Inter- Record Gap
BLPR = Block Prefix (optional)
PADDING = Padding Characters (optional)
DDDD = length of the physical record if BLPR is 4 bytes long (unpacked decimal)
dddd = length of logical record in unpacked decimal format
RL = Record Length
ASCII TAPE RECORDS

r--

---=I

Physical Record

Logical Record

Data

Record

Data

Record 2

= 80

= 100

Record 3
103 I
I
I

7
RL

= 50

BL = Block Length }
RL = Record Length
in binary halfword (16 bit) format I plus two binary zero bytes
IRG = Inter- Record Gap
G = Gap
DISK RECORDS

Figure 2.3.

Variable Length Unblocked Record Format

section 2: Data Management

Logical Record

Physical Record

---1

Data

Record 1

Record 2

Record 3

xxbb

3 14
BL = 234:

Record 4

xxbb

83 :84 87
RL = 100
1

RL = 80

Data

1831184187
1
RL = 50

,
,

233'

RL = 56

BL = Block Length }.
•
.
RL = Record Length
m bmary halfword (16 bit) format, plus two blank bytes.
IR G = Inter-record gap
G =Gap

TAPE RECORDS

. - - - - - - - - - - - - - - Physical Record ------------~
t - - - Log i co I Record

Data

Data
RL

Record 1

Data
RL

Record 2

dddd

Record 3

dddd

IRG = Inter-Record Gap
BLPR = Block Prefix (optional)
PADDING = Padding Characters (optional)
DODD = Length of the physical record if BLPR is 4 bytes long (unpacked decimal)
dddd = Length of logical record in unpacked decimal format
Rl = Record Length
ASCII TAPE RECORDS

.-----------..,..---Ir

Physical Record
Logical Record-l
Data

= 80

--------------I·~I

Data

xxbb

83184
1
I

RL = 100

183 184 187
1
RL = 50
I

BL -= R
Block dlength
ipb
t
L
h }.m b'mary ha Ifword (16 b'It ) r
format,
us two 'mary zero b yes.
RL - ecor engt
IR G = Inter-record gap
G =Gap
DISK RECORDS

Figure 2.4.

Variable Length Blocked Record Format

DOS System Programmer's Guide

233 1
1

BLOCKING RECORDS
The length of individual data records
varies with the type of data and the
application requiring such data. The
format design of a data record is very
significant to the efficient use of the
various storage media available on the
system/360 and System/370. One important
I element in the design of data records
involves what is commonly called blocking
and deblocking. Input/output units
(storage media) are relatively inefficient
when used to store short blocks of
information. TO increase the efficiency of
input/output units, data records are
assembled into blocks of records with size
convenient and efficient for processing.
Each physical record on either tape or disk
requires interrecord gaps. These gaps are
blank areas that distinguish beginning and
ending points of a record. If records are
blocked before loading onto a tape or disk,
many of these gaps can be eliminated. One
of the most important advantages of blocked
records is the increased channel/CPU
processing time overlap. The average
number of reads required to locate a record
can usually be reduced by increasing the
blocking factor (number of records per
block). The greater the blocking factor,
the greater the chance that the next record
required will be in the same block. This
is an important consideration when
designing jobs that involve file searching
either on tape or disk. It is particularly
important when using disk storage
techniques that develop overflow records.
overflow records occur when there are more
items' assigned to a disk track than can be
stored on that track.
Blocked records normally require more
main storage than unblocked records because
main storage has to contain the block of
records being read or written on a storage
device. Also, more main storage is
required to hold blocking and deblocking
program instructions. The LIOCS macro
instructions are designed to handle the
blocking and deblocking of records so that
the user need only design the most
efficient blocking factor for his
particular data file and equipment
specifications.

LOGICAL FILE VS PHYSICAL UNIT
A logical file consists of one record for
each item of a group (i.e., an inventory
file would contain one record for each
inventory item). A physical unites) is
used to store a logical file.
For example,
the IBM 2400 Magnetic Tape Unit, the IBM

3420 Magnetic Tape Unit, the IBM
2311/2314/2319/2321 Direct Access storage
Devices (DASD) and the IBM 2540 Card Read
Punch can be considered physical units when
data records punched into cards are being
read into the system.

A logical file may be of such a size
that it requires several reels of tape or
disk packs to contain it. Such a file
would be referred to as a multivolume file.
(Each reel of tape, disk pack or each cell
within the 2321 is considered to be a
volume.) If a small file does not require
an entire reel of tape or disk pack to
contain it, the remaining space on the
volume could be utilized by another
complete or partial file.
This volume
would be referred to as a multifile volume.

DATA MANIPULATION
Data manipUlation with DOS is implemented
at two distinct levels. The first level,
physical IOCS requires extensive knowledge
of System/360 and/or System/370
input/output devices, as well as a detailed
understanding of the basic assembler
language. The second level, logical IOCS,
uses a series of IBM-supplied macros to
construct and process data files. Logical
IOCS requires a minimum knowledge of the
hardware I/O devices and is easily
implemented within the problem program by
the coding of macros. This system is also
used by most of the DOS high-level
languages to control I/O operations.

Physical Input/Output Control System
(PIOeS)
Physical IOCS consists of input/output
(I/O) routines that handle the actual
transfer of data records between external
storage devices (cards, tape, disk, etc.)
and main storage. Program routines
incorporated in the channel scheduler
portion of the supervisor handle the
following functions of PIOCS:
• Building a schedule of I/O operations
for all devices on the system (CHANQ
table).
• Starting the actual I/O operations on a
device (SIO).
• Scheduling the start of all I/O
operations and monitoring all events
assQciated with I/O.
section 2: Data Management

• performing error recovery procedures
(ERP) for all DOS supported devices,
when necessary.
A user's problem program normally uses
logical IOCS for file processing. Logical
IOCS, in turn, uses physical IOCS to
perform actual data transfers. There are
occasions, however, when a user may need to
bypass the logical IOCS routines to perform
a particular I/O operation. (Physical IOCS
is implemented at the assembler language
level.) Three macro instructions are
provided to allow the user to communicate
directly with physical Ioes.
CCB

This macro instruction creates a
command control block. This is
also considered an event control
block (ECB). The CCB contains
the user's information about
special considerations and/or
options that he has chosen for
this I/O operation. It is
comprised of constant statements
(OS) that are used in two-way
communications with the
supervisor.

EXCP

This macro instruction is
converted to the proper SVC
instruction (SVC 0) to request
execution of a channel program.
It supplies the location of the
corresponding CCB to the
supervisor.
The EXCP macro instruction
provides more freedom in
controlling devices than the
logical IOCS macros, yet retains
many of the operational
advantages of the Disk Operating
system. DOS provides scheduling
and queuing of I/O requests,
efficient use of channels and
devices, data protection,
interruption procedures, and
error recognition and retry. To
use physical IOCS, however, the
programmer needs detailed
knowledge of device control and
system functions. He must supply
his own channel programs, using
the COW (channel command word)
assembler instruction statement
(See Channel Command Word
section).

WAIT

100

This macro instruction tests CCB
byte 2, bit 0 (traffic bit) to
determine when an I/O operation
has been completed. If the
operation is not completed, the
supervisor gets control until
physical IOCS within the
supervisor sets the traffic bit
to indicate completion of the
DOS System Programmer's Guide

operation. The WAIT macro should
always be issued for each
requested I/O operation.
Whenever physical IOCS macro
are used, the programmer must
construct the channel command words (CCW)
for his input/output operations. He uses
the assembler COW statement to do this.
See Channel Command Word for detailed
information on the cow. Figure 2.5 and the
following text shows how an I/O operation
can be traced through physical IOCS.
instruction~

A request is made to physical IOCS to
start an I/O operation by means of the EXCP
macro instruction in the problem program.
From information in the CCB, physical IOCS
determines the channel for which the
request was made and places the request on
a queue for that device. If the channel(s)
or device is not busy, the I/O is started
and control returns to the problem program.
If the channel is busy, control returns to
the problem program, but the I/O request
waits in the channel queue. When the
request reaches the top of the channel
queue, the I/O is started.
Control returns to the program
requesting the I/O unless there was an
error condition detected on the START I/O
(SIO) instruction. The problem program
normally continues processing until it
requires that the requested I/O operation
be complete (either the information being
read into main storage is needed or the
output area must be freed on an output
operation). At this time, the WAIT macro
should be issued. The WAIT macro causes
the now waiting task to be removed from
task selection until the proper interrupt
is processed for this device by the
supervisor.
Any problem program that is running will
be interrupted when the I/O operation is
complete (all data transferred to or from
main storage and the external device and no
permanent errors have been detected). At
this point, the request is removed from the
channel queue and normal task selection
resumes.
If an error was detected that could not
be corrected by the device error routines,
the problem program or the computer
operator would be notified via a message on
SYSLOG. User error routines can be
notified via the CCB to handle conditions
such as wrong length record.
Physical IOCS always attempts to perform
its function so that the time for executing
an I/O operation is overlapped with the I/O
operations on other channels and also
allows the I/O operations to be overlapped
with processing.

SUPERVISOR

CHANNEL
INTERFACE

PROBLEM PROGRAM
CHANNEL
SCHEDULE
ROUTINE

••
. ("
•• \

/
I

·."-··

. "

START
I/O

\\
\

.',', \
•

•

,~
" \ ....... ,

r·---

I
TEST

I
I

READCCB

L-_ --~

Figure 2.5.

I/O INTERRUPT
ROUTINE
•

•

,,/', .-----,

•

/ , .

C~B; SYSIPT, RDCCW

•

CCW
OS

••

I'
I

.. - - - ---,~_~ ____

:
RDCCW
INAREA

WAIT READCCB' ,

I/O DEVICE

I/O REQUEST
TABLE (CHANQ)

..J

I
2,INAREA, X'20',80
Cl80

I/O Operation Using

Ploes

Section 2: Data Management

101

Command Control Block (CCB)

error does exist. The only,additional
information that the supervisor
provides under this circumstance is
the CSW status information posted in
bytes 4 and 5 of the user CCB. If the
user specifies the fourth operand
(sense Address), the sense information
is present in the sense area that the
user has specified (if an error
condition existed at the end of the
I/O operation).

The CCB establishes communication between
the problem program and physical IOCS. The
CCB is two doublewords in length with eight
major fields, as shown in Figure 2.6. The
optional 8 bytes are generated if the user
requests that a sense operation be
performed, on an I/O error. Data
transferred from the device to main storage
during a sense operation provides
information concerning both unusual
conditions detected in the last operation
and the status of the device. All data in
the CCB is in the hexadecimal format. The
eight fields of the CCB are listed and
described as follows:
1.

Count Field (Bytes 0,1): Contains the
residual count from the channel status
word (CSW), that is stored by PIOCS
when the CCB is removed from the
queue.. The residual count in
conjunction with the original count
specified in the last ccw used,
indicates the number of bytes
transferred to or from the area
designated by the Ccw. When an input
operation is terminated, the
difference between the original count
in the CCW and the CSW is equal to the
number of bytes transferred to main
storage. For an output operation, the
difference is equal to the number of
bytes transferred to the I/O device.

Symbolic Unit Number (Bytes 6,7):
Contains the 2-byte hexadecimal
representation of SYSnnn (symbolic
unit). This value represents the
location of the logical unit in the
LUB table (see Figure 1.19) and is
placed in the CCB by the problem
programmer's specification of the
symbolic unit to be used for this I/O
operation. The symbolic unit is
converted to a 2-byte hexadecimal
representation by the CCB macro. The
first byte indicates the class of the
symbolic unit (system class=X'OO',
programmer class=X'Ol'). The second
byte of the representation indicates
the number of the unit within one of
the previously mentioned classes. An
example of such a conversion can be
illustrated with the symbolic unit
SYS007, which converts to X'0107'.

Byte 8: Contains the length of the
block prefix. This length is
X'00'-X'63' for ASCII input tapes and
X'OO' or X'04' for ASCII output tapes.
For EBCDIC tape files, this byte is
always X'OO'.

CCW Address (Bytes 9-11): Contains
the address of the CCW that is
associated with this CCB. This
address is placed in the CCB by the
specification of the second operand of
the CCB macro. In the case of chained

Transmission Information (Bytes 2 and
Used for two-way communication
between the supervisor and the problem
program. Each bit within these two
bytes represents either a condition
that was detected by the supervisor
and posted to the user, or a user
option to be communicated to the
supervisor. Refer to Figure 2.6. The
designation pr.pr. indicates those
bits that the problem programmer may
set and are not reset by the
supervisor for each I/O request.
PIOCS indicates those bits that the
supervisor is capable of
setting/resetting just before each I/O
operation associated with this CCB.

DOS System Programmer's Guide

CSW Status Bits (Bytes 4,5): Contains
the CSW status information that is
stored by PIOCS before control returns
to the problem program.
Note: The particular bits that are
turned on in bytes 2 through 5
indicate the conditions that were
detected by PIOCS and/or the problem
program by the specification of the
third operand of the CCB macro.

1L:

The user options may be initialized
at assembly time by specifying the
third operand of the CCB macro. (See
supervisor and I/O Macro listed in the
Preface for detailed information for
coding the CCB). A third operand with
a value X'OlOO' instructs PIOCS to
return to the user after each I/O
operation for this CCB whether or not
any errors have occurred. It is then
up to the user to handle all aspects
of error recovery and/or retry, even
to the point of determining if an
102

Note: Bytes 2 through 5 are ANDed off
(mask setting is X'lF050000') by PIOCS
when the CCB is placed in the queue.
Communication bits that were set on by
the problem program are left on
because an AND instruction is used by
PIOCS for resetting bytes 2 through 5.

CCWs, this operand specifies the
address of the first ccw.
7.

Byte 12: Contains information used by
physical IOCS that must not be
modified by the user.

CCW Address in CSW (Bytes 13-15):
Contains the CCW address from the CSW
stored by PIOCS before control returns
to the problem program. A CCB that
has been queued by PIOCS to service a
problem program I/O request cannot be
used for a second problem program I/O
request until the first request has
been completed.

Optional Sense CCW (Bytes 16-23):
Bytes 16-23 are appended to the CCB by
the CCB macro expansion when the user
wants sense information on

unrecoverable I/O errors returned. If
the user specifies a sense address by
coding the fourth operand of the CCB,
an 8-byte CCW for reading sense
information is generated as t~e last
field of the CCB. The name field
(sense address) of the area that the
user supplies must have an assigned
length attribute of at least one byte.
Physical IOCS uses this length
attribute in the ccw to determine the
number of bytes of sense information
the user wants at this sense address.
For more detailed information
concerning this sense information see
the Principles of Operation listed in
the Preface. For detailed information
regarding the actual coding of the
physical IOCS macros refer to
Supervisor and I/O Macros listed in
the Preface.

section 2: Data Management

103

...-------r-------,r---------,-------r--------., - - - - - -_._,

~
.....

s::

Bytes-.\O

r. 5

Yes
Yes

No
No

Document feeding stopped.

Channel data check or
Busout check.
Yes

1 User Error Routine

Data check in DASD count Field.
Data check - 1285, 1287, or 1288.
MICR - SCU not operational.
3211 Print Check (equipment check).

X'80'

I/O requested and not
completed.

I/O error passed back due to
program option or operator
option.

1017/1018 data checks.
Return any DASD data checks.
51

Mask for
Test Under
Mask
Instruc ti on

Unrecoverable I/O Error

4 1 2671 data check.

I/O Completed. Normally set
at Channel End. Set at Device
End if bit 5 is ON.

On Values
for Third
Operand in
CCB Macro

X'80'

No
No
No
X'40'

X'20'

Figure 2.6.

command Control Block (CCB)

(Part 2 of 3)

Section 2: Data Management

105

Condition Indicated
Bit

Byte

Tape read data check (2400 - series or
2495); 2540 or 2520 punch equipment
check; or any DASD data check.
1017/1018 data check.
1285, 1287, or 1288 equipment check.
3211 data check (print check).

Operation was unsuccessful.
Byte 2, bit 2 is also ON.
Byte 3, bit 0 is OFF.
Yes
Yes
Yes

Questionable Condition.

Card: Unusual command
sequence (2540) . DASD:
No record found.
1285/1287/1288: Document
jam or torn tape.
Yes

Nonrecovery
UCSB parity check (command retry).

..
5

No
Set the questionable
condition bit ON and
return to user.

Verify error for DASD or Carriage
Channel 9 overflow

Yes. (Set ON when Channel 9
is reached only if Byte 2, bit 5
is ON).
Yes

Yes

Retry begins at last CCW
executed.

Retry begins at first CCW
of channel program.

Command Chain Retry

X'JO'

X'08'

Retry command if no record
found condition occurs (disk) .

Mask for
Test Under
Mask
Instruction

No
No
No

No record found condition

1287 document mode -late stacker
select.
1288 End - of - Page (EOP).

o (OFF)

1 (ON)

On Values
for Third
Operand in
CCB Macro

X '0004'

X'04'

X'02'

X'OOOI'

X'OI'

User Option Bits. Set in CCB macro. Physical 10CS sets the other bits OFF at EXCP time and ON when the condition specified occurs.
I/O program check, command reject, or tape equipment check always termi nates the program.
For System/360, the user must handle all error or exceptional conditions except Channel Control Check, Interface Control Check,
I/O Program Check, and I/O Protection Check. For System/370, the user may handle Channel Control Checks and Interface Control
Checks. The occurrence of a channel data check, unit check, or chaining check causes a byte 2, bit X'20' of the CCB to turn on,
and completion posting and dequeuing to occur .1/0 program and protection checks always cause program termination. Incorrect
length and unit exception are treated as normal conditions (posted with completion). Also, the user must request device end posting
(CCB byte 2, bit X'04') in order to obtain errors after channel end.
Error correction feature for 1018 is not supported by physicallOCS. When a 1018 data check occurs and CCB byte 2, bit X'02' is on,
control returns directly to the user with CCB byte 3, bit X'10' turned on.
A line position error can occur as a result of an equipment check, data check, or FCB parity check.

Figure 2.6.

Command Control Block (CCB) (Part 3 of 3)

Channel Command Word (CCW)

Command Code

To aid the programmer in using physical
IOCS, an assembler instruction statement,
CCW (channel command word), is provided.
This ccw assembler instruction statement is
a convenient means to define and generate
the a-byte channel command words needed for
the channel program. See Figure 2.7 for
the format of the CCW assembler
instruction.

The ccw command code (1 byte) may be
expressed as a decimal number, hexadecimal
representation, or as a symbol that has
been equated to the proper hexadecimal or
decimal value (e.g., 19, X'13', PR, where
PR EQU 19 respectively). The assembler
moves, or converts and moves, the command
code to the first byte of the machine
language CCW it is generating. Each I/O
device has a specific set of commands to
which it will respond; any other commands
to that device are rejected.

r------T---------~-----------------------,

IName I Operation I Operand
I
~------+---------+------------------------~
I Any
ICCW
IFour operands separated I
I Symbol I
Iby commas specifying the I
lor Not I
Icommand code, data ad- I
ILUsed
dress, flags, and count. I
______ I _________ I ______________________
~

Figure 2.7.
106

Format of the CCW

DOS System Programmer's Guide

~-~

Data Address
This field must contain a data address
unless the CCW command is a control command
(for example, skip to channel 1 for a
printer). Normally this field is expressed

as a symbol for ease of program relocation
and reflects the address of the
input/output area.

The third field of the CCW mnemonic is used
to communicate special considerations to
the channel regarding this CCW. The value
of the flag byte may be expressed as a
decimal number, hexadecimal character or as
a symbol that has been equated to the
proper decimal or hexadecimal value.
Although it is referred to as the flag
byte, only the five high order bits (bits
32-36 of the resultant CCW) represent
individual flags.
The three low order bits
must be zero in their final hexadecimal
form. The function of each flag bit is as
follows:
Bit 32 (High Order Bit of the Flaq Field):
This is the chain data (CD) flag.
Chaining
refers to a series of CCWs in contiguous
(consecutive) doubleword storage locations
that are linked to each other forming a
chain.
Data chaining permits the reading or
writing of an I/O record from different
areas of main storage.
If the CD flag bit is set to
on), the data address and count
sequential CCW are also used in
data record.
Both CCWs pertain
I/O record.

1 (CD flag
in the next
storing a
to the same

When data chaining, the command bytes of
successive CCWs are ignored (unless it is a
transfer in channel (TIC», but the field
must contain a valid command. only the
first CCW's command byte is used. It is
important to note that data chaining
pertains to only one I/O operation or data
record. This could be one punch card or
one tape record or one line of print and so
forth.
Example: TO read columns 1-30 of a punch
card into storage beginning at location ONE
and columns 31-S0 into storage beginning at
location TWO, two CCWs could be used. The
first CCW would have its CD flag bit set to
1 as shown:
CCW1

CCW
CCW

2,ONE,X'SO',30
2,TWO,X'00',50

CCW1 causes thirty bytes to be read into
storage starting at location ONE. The
second CCW causes the next fifty bytes of
the card to be read into storage starting
at location TWO.
Note: The ability to data chain is
dependent upon device and channel speeds.

J &00
Q (oO

o I) 10
r\ til'"

Bit 33: Bit position 33 of the CCW is the
command chaining (CC) flag. This bit when
set to one causes the next sequential CCW
to initiate another operation on the same
I/O device.
For example, i t is possible to read two
cards into main storage as a result of one
EXCP macro instruction. Two CCWs with the
CC flag set in the first CCW are required.
CCW1

CCW
CCW

2,DATA,X'40',SO
2,DATA+SO,X'OO',SO

In this example, CCWl initiates a read
command. The first card is read into
storage locations DATA through DATA+79.
Because the CC flag bit in the first CCW
is set to one, the second CCW is used to
initiate another card read operation. The
next card is read into storage locations
DATA+80 through DATA+159.
Note that when data chaining, only one
I/O operation occurs. The data from the
one I/O record is placed in main storage
under control of two or more CCWs.
When command chaining, each CCW controls
a different I/O operation. The commands
that are chained do not need to be the
same. For instance, it is possible using
one EXCP macro instruction to do a
write-backspace-read combination with a
magnetic tape unit by command chaining with
three CCWs in one EXCP macro instruction.
When command chaining, an I/O interrupt
cannot occur at the end of each command,
but can occur when the last command is
executed.
When command chaining, each command
processes one physical record. If the byte
count of the record does not agree with the
count field of the CCW, an incorrect length
indication results.
Bit 34: Bit position 34 of the CCW is the
suppress incorrect length indication (SLI)
flag. The SLI bit may be turned on to
prevent an indication to channel of an
unequal compare between the byte count
specified in the CCW and the actual bytes
read from the record.
Example of Incorrect Length: If you want
to read columns 1-50 of a card into main
storage starting at location BUFFER and
bypass the remaining 30 columns, the
following CCW can be used:
CCW1

CCW

2,BUFFER,X'20',50

An incorrect length indication results
because one CCW causes the device to
process one complete record and the device
section 2: Data Management

107

could not be stopped on the 51st byte.
When an incorrect length indication occurs,
the status bit in the CCB is set for
testing by the programmer.
If the SLI flag bit is on, it suppresses
the incorrect length indication for each
data record.
For example, if you want to
read columns 1-40 of a card into main
storage starting at location BUFFER1 and
columns 41-60 into main storage starting at
location BUFFER2, the following two CCws
could be used:
CCW1
CCW2

CCW
CCW

2,BUFFER1,X'SO',40
0,BUFFER2,X'20',20

CCW1 causes 40 bytes to be read into main
storage starting at location BUFFER1. CCW2
causes the next 20 bytes of the card to be
read into storage starting at location
BUFFER2. A breakdown of the flag bytes
would be:
CCW1

Data Chaining

CCW2

Suppress incorrect length indication

Note that the SLI flag is not required for
the first CCW. The incorrect length
indication is relevant to an entire data
record and not to a given ccw. The SLI
flag is present in the second CCW because
it is the last CCW for the record and the
total value of the two CCW count fields
does not equal the actual record length.
Bit 35: Bit position 35 of the CCW is the
SKIP flag.
The SKIP flag suppresses the
transfer of information into main storage.
It can be used together with the data
chaining feature to read selected portions
of an I/O record into main storage. To do
this, the CD flag is used along with bit 35
of the CCW. For example, if columns 51-SO
of a punch card are to be read into
locations DATA through DATA+29, the card
reader reads all SO columns of information
and attempts to transfer SO bytes of data
into main storage. The channel needs 2
CCws.
CCW1

CCW
CCW

2,DATA,X'90',50
2,DATA,X'00',30

CCW1 has both the CD and the SKIP flag bits
on.
Because the skip flag bit is on, CCW1 is
used to suppress the transfer of data into

lOS

DOS System Programmer's Guide

main storage. The count field causes 50
bytes to be bypassed.
Because the CD flag bit is on, the
second CCW is fetched after the first 50
bytes have been skipped. The second CCW
causes bytes 51 through SO to be read into
storage starting at location DATA.
Bit 36: Bit position 36 of the CCW is the
program control interrupt (PCI) flag.
CPU
normally receives an I/O interrupt at the
end of the I/O operation. However, bit
position 36 of the CCW can be used to
signal an I/O interrupt before the end of
the operation.
When a CCW is fetched that has its PCI
flag bit on, CPU receives an I/O interrupt
as soon as it can accept it. In this way,
CPU is notified of the progress of an I/O
operation. This notification could be used
to initiate the processing of the data
records that have been read into main
storage by the preceding CCws in the chain.
I/O interrupts normally occur at the end
of the operation. An interrupt caused by
PCI would not affect the I/O operation.
The interrupt is taken by the CPU and the
I/O operation continues in the same manner
as if the interrupt had not occurred.
The PCI-caused interrupt occurs as soon
as possible after start of execution of the
command containing the PCI flag. The
occurrence of the interrupt may be delayed,
depending on the model of system/360 and/or
the particular device that it is operating.
A more detailed explanation of the CCW
flags can be found in Principles of
Operation listed in the front of this
manual.

The count field gives the total number of
bytes in the storage area (the physical
size of the records being read or written).
The count field specifies any number of
bytes up to 65,535. Except for a CCW
specifying a transfer in channel, the count
field cannot contain the value zero.
To illustrate the use of physical IOCS,
Figures 2.S and 2.9 show a sample program
with a related flowchart,.

Figure 2.8.

Flowchart for EXCP Coding Example

Section 2: Data Management

109

PAGE
LOC

OBJECT CODE

AOOR1 AOOR2

000000
OOOOOQ
000002 0'540
600002 4130 0006

00006

OOI'lOtA 0204 4002 410B 00002 0010B

1)0005C
00('1060
000064
000068

4361)
4166
4260
4630

4006
0001
4006
4034

00006
00001
00006
00034

00006F 0700

01)00110
1)001)81l
I)OOOCO
Doooe8
00':'01)('
00(01)7
000(1)8
00001'0
O,,00E8
OOOOFO
OOOOFIl
000100
00010('1
000101'1
OOr:'12C

0700000040000006
2900000760000004
0800ooB800000000
0600010020000050
00000000000000
E50503F1
0('100000000
0700000040000006
3100000260000005
080000E800000000
OE0001002000008C

1'101'11 A9 00000000000000
0001110 0900018C6000000C
000188 -1 0 00010020000050

LOC

0001CO
0001C8
00011)0
000108

VTOCCCW CCW
LONG SEEK ,SEEK,CC ,6
SEEK TO THE VTOC CYL INOER/HEAD
SEARCHIO CCW
SRCHIDE,SEEK+2,CC+SLI,5
SEARCH 10 EQUAL FOR VTOC RCD
CCW
TIC,SEARCHIO,O,O
TIC BACK TO SEARCH
CCW
REAOKOTII,VTOCRCO,SLI,L'VTOCRCO READ KEY/OATil FROM VTOC
VTOCRCO OS
OC1140
VOLUME
OS
OCL80
KEY
OS
CL44
DATIl
OS
CL96
VOLHEAO OC
C'VOLUME LABEL'
KEYHEAO DC
C'VTOC KEY'
DATAHEAO OC
C'VTOC DATA'
********************************************************************
* PRINT VOLUME HEADER AND SPACE ONE
* PRINT VOLUME LABEL AND SPACE THREE
* ********************************************************************

090'1019860000008
190001006000002C
0900(11 A060000009
1900012C2(1(100060

(1)1)0~1

OOOO!'F

1)00009
01'1001

000080
000040
OOOO:?O
0('0010
OOOOn.'!
000000
00000(170
0001)0090
(1)('(1')080
onn('Hl(' 40

STMT
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
15<1
16(1

SOURCE STIITEMENT

110

Physical

Ioes

0030
0020
00005
0070
0080
0090
0100
0110
0120
0140
0180
0190
0210
0220

0270

0330
0340
0350
03600370
0380
0390

- - - - - - - - - - - - - - - - PAGE
- - - -2
FOOS CL3-5 10/31/69

* PRINT KEY OF VTOC RECORD
* PRINT DATA HEADING
* PR INT DATA POll TI ON OF VTOC RECORD
* ********************************************************************
VTOCPCCW CCW
PRINTSP1 ,KEYHEAO,CC+SLl ,L' KEYHEAD
CCW
PR I "lTSP3 ,KE Y,CC+S LI ,L' KEY
CCW
PRINTSP1,DATAHEAO,CC+SLI,L'OATAHEAO
CCW
PRINTSP3,OATA,SLl,L'OATA
* ********************************************************************
* ,SAMPLE OASO CHANNH COMMAND EQUATES
* ********************************************************************
LO"lGSEEK EQU
07
SRCHKEYE EQU
41
TIC
EQU
08
REAOOATA EOU
06
SRCHIOE EQU
49
READKOTA EQU
14
* ********************************************************************
* SAMPLE 1403 OR 1443 PRINTER COMMAND EQUATES
* ********************************************************************
PRINTSP1 EOU
09
PRINTSP3 EQU
25
* ********************************************************************
*
FLAG OPEIIANOS
* ** ADO SYMBOLS TOGETHER FOR COMBINATIONS SUCH AS CO+SL 1*************
* ********************************************************************
CO
EOU
128 CHAI N DATA
CC
EQU
64
COMMAND CHAIN
SLI
~QU
32
SUPPRESS LENGTH INDICATION
SKIP
EQU
16
SKIP PORTIO"lS OF THE RECORD
PCI
EQU
8
PROGRAM CONTROLLED INTERRUPT
ENOCCWS EOU
0
END ,OF CCW LI ST
END
=A (VOLREAOI
=A (VOLPRT)
=A(VTOCREAOI
=AIVTnCPRT)

[

Figure 2.9.

0010

120 VOLPTCCW CCW
PRINTSP1,VOLHEAO,CC+SLI ,L'VOLHEIID
121
CCW
PRINTSP3,VOLUME,SLl ,L'VOLUME
122 * ****************************************.****************************
123 * PRINT VTnc KEY HEIIOING

AOOR1 AOOR2

000007
1'100029
001'(01)8
0000')6

01)1)11'(1
oonlE4
OnC11'8
(\''''11'r

STAR T 0
PR I NT NOGEN
BALR 4,0
U5I-NG ~,4
.lA
3,6
EXCP VOLREAO
READ THE VOLUME LABEL
WAIT VOLREAO
WAIT FOR THE. READ TO COMPLETE
MVC SEEK+2(51,VOLUME+ll
SET UP SEEK/SEARCH ARGUMENT
EXCP VOLPRT
PRINT THE VOLUME LABEL
21
WAIT VOLPRT
WAIT FOR PRINT TO COMPLETE
27 REAOVTOC EXCP VTOCREAD
READ THE VTOC
31
WAIT VTOCREAO
WAIT FOR THE READ TO COMPLETE
37 PRTVTOC EXCP VTOCPRT
PRINT A VTOC RECORD, KEY AND DATA
41
WAIT VTOCPRT
WAIT FOR PRINT TO COMPLETE
47
IC
6,SEEK+6
UPDATE THE RECORD COUNT IN THE
48
LA
6,1(61
SEEK/SEARCH ARGUMENT
49
STC
6,SEEK+6
STORE UPDATED 10 INTO SEEK/SEARCH ARG.
50
BCT
3,REAOVTOC
51 EOJ
EOJ
54
CNOP 0,4
SYS004,VOLCCW
55 VOLREAO CCB
66 VTOCREAO CCB
SYS004, VTOCCCW
SYS006, VOLPTCCW
77 VOLPRT
CCB
88 VTOCPRT CCB
SY S006, VTOC PCCW
LONGSEEK,SEEK,CC,6
SEEK TO CYLINDER 0, .HEAO 0
99 VOLCCW
CCW
SRCHKEYE ,KEYSCH,CC+SLl ,4
SEARCH KEY EQUAL (VOL11
100 SEARCHKY CCW
101
CCW
TIC,SEARCHKY,O,O
TIC BACK TO SEARCH
REAOOATII,VOLUME,SLl ,L'VOLUME REIID VOL LABEL DATA FIELD
102
CCW
103 SEEK
DC
X' 00000000000000'
104 KEYSCH
DC
CL4'VOLl'

- - - - -- - - - -- - - - - - - - - - - -

OBJt=CT CODE

FOOS CL3-5 10/31/69

SOURCE STATEMENT

1 01 SKR EAO
2
3
4
5
6 REAOVOL
10
16
17 PRTVOL

105
106
107
108
109
110
111
112
113
114
115
116
117
118
119

')onll~C E506D3E404C54003
00019A E5E306C34002C5EI!
0"01 AO E5E31l6040 C4C1 E3

-- - - - - - - -

SHIT

Sample Program

DOS System Programmer's Guide

0460

Logical Input/Output Control System
(LIOCS)

LOGICAL IOCS PROCESSING METHODS

Logical Ioes (LIOCS) provides data
management functions required to locate and
access logical records in a file (just as
the problem programmer would). LIOCS uses
physical IOCS to accomplish actual data
transfer and device control. The data
management functions include:

The IBM Disk Operating System provides
logical IOCS routines to process records in
sequential order, in random order by the
Direct Access Method (DAM), or randomly and
sequentially by the Indexed Sequential File
Management System (ISFMS). Sequential
processing applies to all files on serial
I/O devices (such as card reader, tape,
printer, etc), and to records on IBM 2311,
2314, and 2319 disk or 2321 data cell
devices when they are processed serially.
The types of processing performed by DAM
and ISFMS apply only to files of Direct
Access storage Device (DASD) records.

• Organization (blocking and deblocking)
of logical records.
• Control (switching) of I/O areas when
more than one area is used.
• Handling of the open and close
requirements, and end-of-file and
end-of-volume conditions.
• Resolving symbolic references to
physical I/O devices.
• Translating data in the I/O areas from
ASCII to EBCDIC (on input) and from
EBCDIC to ASCII (on output).
with IBM-supplied macro instructions,
logical IOCS eliminates the repetitive
detail coding required for standardized
input and output routines. These logical
IOCS macros, referred to as imperative
macros, supply the facilities for reading,
writing, blocking and deblocking records,
file labeling, and error checking. To make
use of these facilities, logical IOCS
imposes certain requirements that must be
handled by the problem programmer. These
include descriptive entries in other
IBM-supplied macros, called declarative
macros, to specify the characteristics of a
file that is to be processed by logical
IOCS. The IBM assembler program uses the
descriptive information when processing the
macro statements to tailor the logical IOCS
routines for the specific application.
Logical IOCS handles data transferred to
or from I/O devices as logical files of
data. When logical IOCS determines that an
I/O area contains no logical record (or
records, in the case of blocked records)
needing processing, it issues a physical
IOCS macro to execute the actual data
transfer. Figure 2.10 shows the
relationship between logical and physical
IOCS for a logical IOCS imperative GET
macro issued to an input file when one I/O
area is used.
Logical IOCS is a generative system that
uses the capabilities of a macro language.
It is this generative nature that provides
the problem programmer with the ability to
include only those routines needed for a
specific job or job step.

Logical disk files can be data set
secured. A data-secured file cannot be
accessed accidentally by problem programs.
The data set security function provides a
record on the system log each time a
data-secured file is opened. It is the
user's responsibility to insure data
security.
In addition, logical disk files can have
the track hold option. When track hold is
specified in the DTF, a track that is being
modified by a task in one partition cannot
be concurrently accessed by a task or
subtask in the same or another partition
provided that all programs competing for a
track specify the track hold option. Any
program that does not use the track hold
option will not be denied access to a
track, and can modify a track that is being
held by another program. Therefore, all
programs accessing the same file should
have the track hold option.

Sequential Processing
Sequential processing reads, writes and
processes successive records in a logical
file. For example, card records are
processed in the order the cards are fed.
Tape records are processed starting with
the first record after a header label and
continuing through the records to the
trailer label. DASD records are processed
starting with a beginning DASD address and
continuing in order through the records on
successive tracks and cylinders to the
ending address.
A sequential file on DASD is contained
within one or more sets of limits called
extents that are specified by the user in
job control EXTENT/XTENT statements or
commands. If the logical file consists of
more than one set of limits, logical IOCS
automatically processes each set as
required by the user. The records within
Section 2: Data Management

111

Problem Program Partition

Supervisor

I
Problem
Program

Logical lacs

Physical lacs
(Channel
Scheduler)

I/o
Device

I
Issue GET

Provide record
(deblock) and
return to the
Problem Program,
I

or
If I/O required,
issue EXCP macro

I
I
I
I
I
I

Determine channel
and:

and
a) If channel is not
WAIT mtcro

Return to the
Problem Program
when I/O is
comple~e, or, if an
Next
instruction
after GET
macro
expansion

Figure 2.10.

ASCII input file,
translate data to
EBCDIC in I/O area
and return to
problem. program.

~busy, start I/O

b) If channel is busy,
place in channel
queue and return
to LlOCS.
I

Start Device

I/O Complete

When I/O'is complete,
return to LlOCS via
Jnterrupt Rou ti ne •

Retrieving a Record Using Logical IOCS

each set must be adjacent and contained
within one volume (disk pack or data cell).
The sets are not required to be adjacent or
on the same volume. sequential processing
of a DASD file written by the Direct Access
Method can be performed if the physical
structure of a sequential file is present.

record key, if records contain key areas,
or the record identifier (ID) that is in
the count area of each DASD record.
Logical IOCS seeks the specified track and
searches for the record on that track, or
on the succeeding tracks in the cylinder.

Direct Access Method (DAM)

Indexed Sequential File Management System
(ISFMS)

The Direct Access Method (DAM) processes
records contained on IBM 2311, 2314, 2319,
or 2321 DASD devices that are usually
organized in a random manner.
Logical IOCS locates a DASD record for
processing by referring to a record
location reference supplied by the problem
program. The location reference consists
of two parts: a track reference and a
record reference. The track reference
specifies the track (or the first of
multiple tracks) to be searched for the
record. The record reference may be the
112

DOS system Programmer's Guide

DASD records contained within an indexed
sequential file can be processed in a
random order or in sequential order by
control information. Both orders use the
control information of the records (such as
employee number, part number, etc), that is
available in the key area of each DASD
record. Any record stored at any location
in the logical file can be processed by the
random me~hod. The user supplies ISFMS
with the key (control information) of the
desired record. ISFMS searches for the
record and makes it available for
processing.

In sequential processing, ISFMS makes a
series of records available. The records
are available one after the other in order
by the control information (key) in the
records. The user specifies the first
record to be processed. ISFMS retrieves
the succeeding records (on demand) from the
logical file, in key order, until the
problem program terminates the operation.
ISFMS creates an organized file and then
adds to, reads from, and updates records in
that file. The file is organized from
records that are presorted by control
information. As the DASD records are
loaded, ISFMS constructs indexes for the
logical files. The indexes permit
individual records to be found in
subsequent processing operations. The
indexes are created in such a way that
records can be retrieved randomly or
sequentially. If records are added to the
file at a later date, ISFMS updates the
indexes to reflect the new records.

MAIN STORAGE REQUIREMENTS
Logical IOCS routines are generated as part
of the problem program. Imperative macro
expansions, that serve as linkages to the
logical or physical IOCS routines, are
generated in-line at the point the macro is
used in the problem program. The open,
close, EOF/EOV, and other special purpose
routines are called into the B-transient
(logical transient) area as required. The
physical IOCS routines used by logical IOCS
are generated as part of the supervisor
program.

MODULAR/TABULAR SYSTEM
The terms modular and tabular mean that the
system uses tables in conjunction with data
handling modules to implement its
functions.
The modular/tabular system has the
following advantages:
• Saves assembly time by allowing the
data handling modules to be generated
separately and stored in the
relocatable library for subsequent use.
• Uses one module for many files if the
device types are the same and the files
are similar.
The major advantage is the ability to use
one logical IOCS module with many different
files. It not only saves a large portion

of main storage but also makes the system
more versatile.
The modular/tabular combination for a
specific file is generated by two macros:
the file definition macro in the form
DTFxx, and the module generation macro in
the form xxMOD.
The file definition macros describe the
logical file, indicate the type of
processing to be used for the file, and
specify main storage areas (work area, I/O
area, etc.) for the file,. A number of
file definition macros define the files
processed by logical IOCS, and one defines
files processed by physical IOCS (DTFPB).
The file to be processed determines the
type of file definition macro to use.
The module generation macros generate
the data handling logic modules. These
modules contain generalized routines needed
to perform the functions of the logical.
IOCS imperative macros. The generalized
routines in the logic modules are altered
and made more specific through various
parameters (specified by the problem
programmer) included in the xxMOD macro
statements. It is possible, therefore, to
generate many variations of a particular
type of logic module, each specifically
suited to the need of the problem
progranuner.

DTF (Define the File) Ma.cros
Whenever logical IOCS imperative macro
instructions are used in a problem program
to control the transfer of records in a
file, that file must be defined by a
declarative DTF macro instruction. The DTF
macro instruction describes (through
various parameters specified by the problem
programmer) the characteristics of the
logical file, indicates the type of
processing for the file, and specifies the
main storage areas and routines. Detailed
descriptions of the logical IOCS file
definition (DTF) macros and their
parameters appear in Supervisor and I/O
Macros listed in the Preface.
In general, the IBM-supplied file
definition (DTF) macros are device
oriented. In addition, three macros,
DTFSR, DTFBG and DTFEN are supported by the
Disk Operating system to provide upward
compatibility from the IBM System/360 Basic
Operating System (SK system). A brief
description follows for each of the DTF
macros available to users of the IBM Disk
Operating system.
Section 2: Data Management

113

DTFCD: Define The File for a Card Device.
To define a file associated with the
records on a card unit.

DTFCN: Define The File for a CoNsole. To
define a file associated with the console
(1052, 3210, 3215) printer.
DTFDA:

Define The File for Direct Access
To define a file when DASD (Direct
Access Storage Device) records are to be
processed by the Direct Access Method.

I method.

DTFDI: Define The File for Device
Independent system files. To define files
assigned to the device independent system
logical units SYSRDR, SYSIPT, SYSPCH, and
SYSLST.
DTFIS: Define The File for Indexed
sequential file management system. To
define a file organized and processed by
the indexed sequential file management
system.
DTFMR: Define The File for Magnetic
Recognition. To define a file associated
with a magnetic ink character recognition
(MICR) device.
DTFMT: Define The File for Magnetic Tape.
To define a file associated with a magnetic
tape device.

When a DTFPH macroinstruction is
encountered at assembly time, the assembler
builds a DTF table that includes only the
parameters needed for the OPEN, CLOSE(R),
FEOV, and FEOVD routines. The OPEN,
CLOSE(R), FEOV, andFEOVD macro expansions
call the open and close routines into the
supervisor B-transient area at object time.
DTFPR: Define The File for a PRinter. To
define a file associated with a printer
device.
DTFPT: Define The File for Paper Tape. To
define a file associated with a paper tape
device.
DTFSD: Define The File for Sequential
DASD. To define sequential files on a
direct access storage device (DASD).
DTFSR: Define The File in a SeRial type
file device. To define a file for
sequential processing of records on any
IOCS supported I/O device.
The DOS DTFSR macro definition accepts
either the BOS or BPS DTFSR macro as valid
input. After determining the device type
reqUired, the DOS DTFSR macro calls the
appropriate DOS DTF macro from the source
statement library. The DTF macro called by
the DOS DTFSR then sets up a DTF t~ble in
the usual manner.

DTFOR: Define The File for an Optical
Reader. To define a file associated with
an optical character reader device.

The DOS DTFSR macro definition only
allows upward compatibility and should not
be used as a statement in the user's DOS
source deck.

DTFPH: Define The File for processing by
PHysical IOCS. To define a magnetic tape
or DASD file with standard labels that is
to be processed by physical IOCS when the
OPEN and CLOSE macros are used for label
processing. DTFPH parameters define the
magnetic tape and DASD files. No other
files processed by physical IOCS require
definition.

DTFBG: Define The File for
BeGin-definition. Must be punched with
DTFBG in the operation field and DISK in
the operand field. The name field is left
blank. DTFBG is included in DOS to provide
compatibility with the BOS DTFSR macro
instruction.

Only the following logical loeS
functions can be performed for files
defined by a DTFPH macro:

DTFEN: Define The Field ENd. To show
there are no more DTF source statements to
process. The DOS DTFEN macro definition
allows upward compatibility for BOS and BPS
users.

• Check the header labels on input files,
and close these files when requested.
MOD (Module Generation) Macros
• For ASCII tape files only:
Translate labels to EBCDIC for input
files and check them. Create labels
and translate them to ASCII for output
files.
• Create header labels on output files,
and create trailer labels when the file
is closed.
• Force end-of-volume on an output file
when requested.
114

DOS System Programmer's Guide

Each DTF (except DTFCN, DTFPH andDTFSR) is
linked to a logical IOCS module generated
by an xxMOD macro instruction. These
modules provide the necessary instructions
to perform the input/output functions
required by the problem program. For
example, the module can read or write data,
test for unusual input/output conditions,
block or deblock records, or place logical
records in a work area.

Some of the module functions are
provided on a selective basis, according to
the parameters specified in the xxMOD macro
instruction. The problem programmer has
the option of selecting (or omitting) some
of these functions according to the
requirements of his program. Omitting some
of these functions results in smaller maip
storage requirements for a particular
module.
There are two options for MOD macros.
The user can:
1.

Insert the MOD macro instruction with
its file parameters in the problem
program source deck. In this case,
the logic module is assembled in-line
with the problem program.

Choose at system generation time to
generate the logic modules needed for
his file formats and system
configuration. To do this, source
decks using macro parameters to
describe the file attributes are
punched for each MOD macro statement.
The logic module macro definition
generates its own unique name, or the
user can name the module in the name
field of the MOD macro statement. The
user name overrides the name the macro
definition normally generates.

For each type of xxMOD macro, the problem
programmer can generate many logic modules
by issuing the macro with varying
parameters for each required module. The

logic modules can be cataloged in the
relocatable library. The CATALR control
cards are automatically generated when the
module is assembled.
At assembly time, the assembler produces
an EXTRN (External Symbol) card for every
V-type constant (or EXTRN statement), in
the user program. The assembler expansion
of the DTF statement produces an EXTRN card
with the name of the logic module needed to
support the parameters that were specified
in the DTF macro. The IBM-generated module
names indicate the type of file and the
support that each is capable of supplying
for the DTF. Refer to Figure 2.11 for ~
breakdown of these names. Because of the
descriptive nature of the IBM standard
names, the programmer should be careful
when specifying his own names for the logic
modules or overriding the IBM standard
names. At the time this program is
link-edited, the linkage editor resolves
these EXTRN symbols (AUTOLINK). If the
program is not to be executed immediately,
the linkage editor catalogs the program
into the core image library.
Figure 2.12 gives the module name
prefixes used in the IBM-supplied programs.
Figure 2.13 shows the relationship of the
DTFxx and the xxMOD macros. The DTFCN
macro is unique in that it generates its
own logic module and combines it with the
DTF table. The logic module for DTFCN is
always punched in the object deck, along
with the DTF table (A of Figure 2.13).

Section 2: Data Management

115

Logic
Module

Prefix

CDMOD

IJC

PRMOD

IJD

IJE

PTMOD

MTMOD
(GET/
PUT)

IJF

MTMOD
(Work
Files)

IJE

4th
Character

7th
Character

6th
Character

8th
Character

F RECFORM=FIXUN B
V RECFORM=VARUNB
U RECFORM=UNDEF

A
Y
C
Z

CTLCHR=ASA
CTLCHR=YES
CONTROL=YES
Neither CTLCHR nor
CONTROL specified

B RDONL Y=YES and
TYPEFLE=CMBN 0
C TYPEFLE=CM8ND
H RDONL Y=YES and
TYPEFLE=lNPUT
I TYPEFLE=INPUT
N RDONLY=YES and
TYPEFLE=0UTPUT
o TYPEFLE=0UTPUT

Z Neither WORKA nor
IOAREA2 specified
W WORKA=YES
I IOAREA2=YES
8 80th WORKA=YES and
IOAREA2=YES
Z WORKA=YES not
specified (CM8ND
file only)

F RECFORM=FIXUN 8
V RECFORM=VARUN 8
U RECFORM=UN DEF

A
Y
C
S
Z

CTLCHR=ASA
CTLCHR=YES
CONTROL=YES
STLlST=YES
Neither CTLCHR nor
CONTROL nor
STLIST specified

C = 8 if ERROPT=YES
(ERROPT=name in
DTFPR) and
PRINTOV=YES
= P if PRINTOV=YES
and ERROPT is not
specified (ERROPT=
RETRY or is omitted
in DTFPR)
= E if ERROPT=YES
(ERROPT=name in
DTFPR)
= Z if neither ERROPT
(ERROPT=RETRY or
is omitted in
DTFPR) nor
PRINTOV is
specified

I IOAREA2=YES
V RDON LY=YES and
Z IOAREA2 not specified
WORKA=YES
W WORKA=YES
Y RDONL Y=YES
Z Neither RDONLY nor
WORK A specified

S SCAN=YES
Z SCAN not specified

T TRANS=YES and
SCAN not specified
Z TRANS not specified

F RECFORM=FIXUNB
and SCAN=YES
U RECFORM=UNDEF
and SCAN=YES
Z SCAN not specified
and/or DEVICE=I 018

I DEVICE=1017
2 DEVICE=IOJ8
Z DEVICE=2671 or not
specified

8 READ=8ACK
F RECFORM=F.IXUN8
or FIX8LK
Z READ =FORWARD or
N RECFORM=UNDEF
not specified
(ASCII files)
V RECFORM=VARUN8
or VAR8LK (E8CDIC
mode)
U RECFORM=UNDEF
(E8CDIC mode)
X RECFORM=FIXUNB or
FIXBLK (ASCII files)
R RECFORM=VARUN8 or
VAR8LK (ASCII mode)
S RECFORM=SPN UN 8
or SPN8LK (spanned
records)

C CKPTREC=YES
Z CKPTREC not
specified

W WORKA=YES
Z WORKA not specified

M ERREXT=YES and
RDONLY=YES
N ERREXT=YES
Y RDONLY=YES
Z Neither ERREXT' nor
RDONLY specified

W TYPEFLE=WORK

N NOTEPNT=YES
S NOTEPNT=POINTS
Z NOTEPNT not
specified

M ERREXT=YES and
RDONLY=YES
N ERREXT=YES
Y RDONL Y=YES
Z Neither ERREXT nor
RDONLY specified

Figure 2.11.

116

5th
Character

E ERROPT=YES
Z ERROPT not specified

DEVICE=2540
DEVICE=1442
DEVICE=2520
DEVICE=2501
DEVICE=2540 and
CRDERR=RETRY
5 DEVICE=2520 and
CRDERR=RETRY
0
I
2
3
4

IJC

IJD

IJE

Generated Name Structure for Logic Modules (Part 1 of 2)

DOS System Programmer's Guide

Subsetting/
Supersetting
+ Permitted
* Not Permitted

IJF

*** * *
FA880
VYC I I
U+HW2
CIZ3
ZN 4
0 5

* *+* *
F AB I V
VYEZ W
US+ Y
+ P Z
CZ
Z

* * * *
ZZZZZ
ZTZZ
S ZF Z
SZUZ
ZZZI
Z TZ I
S ZF I
S ZU I
S ZZ 2
ZT Z2
* + ++ +
F 8 C WM
NZZZY
+
R
N
U
X
Z
+
S
V

+
+ +
IJF WE N Z M
ZS Y
+
Z
N
Z

Logic
Module

4th
Character

Prefix

SDMOD
(GET/
PUT)

IJG

SDMOD
(Work
Files)

IJG

ISMOD

IJH

6th
Character

5th
Character

7th
Choracter

8th
Character

U SDMODxU
I SDMODxl
SDMODxO

C ERROPT=YES and
ERREXT=YES
E ERROPT=YES
Z Neither ERROPT nor
ERREXT specified

M TRUNCS=YES and
FEOVD=YES
T TRUNCS=YES
W FEOVD=YES
Z Neither TRUNCS nor
FEOVD specified

B CONTROL=YES and
RDONLY=YES
C CONTROL=YES
Y RDONLY=YES
Z Neither RDONLY nor
CONTROL specified

+ * + * +
IJG C U C T B
FIE Z Y
+OZ
+
R
C
U
Z
+
P
Q
V
+
P
5
V

T HOLD=YES
W HOLD=YES not
specified

C ERROPT=YES and
ERREXT=YES
E ERROPT=YES
Z Neither ERROPT nor
ERREXT specified

N NOTEPNT=YES
R NOTEPNT=POINTRW
Z NOTEPNT not
specified

C CONTROL=YES
Z CONTROL not
specified

T RDONLY=YES and
UPDATE=YES
U UPDATE=YES
Y RDONLY=YES
Z Neither RDONLY nor
UPDATE specified

+ + + + +
IJG T C NC T
WERZY
+
ZZ
U
Z

A RECFORM=BOTH and
IOROUT=ADD or
ADDRTR
B RECFORM=FIXBLK
and IOROUT=ADD
or ADDRTR
U RECFORM=FIXUNB
and IOROUT=ADD or
ADDRTR
Z RECFORM not
specified (lOROUT
specifies LOAD or
RETRVE)

A
I
L
R

IOROUT=ADDRTR
IOROUT=ADD
IOROUT=LOAD
IOROUT=RETRVE

B TYPEFLE=RANSEQ
G IOAREA2=YES and
TYPEFLE=SEQNTL or
IOROUT=LOAD
R TYPEFLE=RAN DOM
5 TYPEFLE=SEQNTL
Z Neither TYPEFLE nor
IOAREA2 specified
(IOROUT=ADD or
LOAD)

B CORINDX=YES and
HOLD=YES
C DORINDX=YES
HOLD=YES
Z Neither CORINDX
nor HOLD is specified

F CORDAT A=YES,
ERREXT=YES and
RDONLY=YES
G CORDAT A=YES and
ERREXT=YES
0 CORDAT A=YES and
RDONLY=YES
P CORDATA=YES
5 ERREXT=YES and
RDONLY=YES
T ERREXT=YES
Y RDONLY=YES
Z Neither CORDATA
nor RDONLY nor
ERREXT specified

+ + + + +
IJH AABBF
B I ROO
Z + + + +
+ ABC 5
A R 5 Z Y
+
U * +
ZLG G
5
P
+
+
T
G
Z
Z

I IDLOC=YES
E IDLOC=YES and
FEOVD=YES
R FEOVD=YES
Z Neither FEOVD nor
IDLOC specified

H ERREXT=YES and
RELTRK=YES
P ERREXT=YES
R RELTRK=YES
Z Neither ERREXT nor
RELTRK specified

W HOLD=YES and
RDONLY=YES
X HOLD=YES
Y RDONLY=YES
Z Neither HOLD nor
RDONLY specified

+++++
IJI BAEHW
F ZIP X
+
ZZZ
+++
5
EHW
V
R RY
ZZZ

B TYPEFLE=OUTPUT

I IOAREA2=YES
Z IOAREA2 not specified

C RDONLY=YES
D RDONLY not
specified

+ + *
IJJ FCBI C
IZD

C SDMODFx specifies
HOLD=YES
F SDMODFx does not
specify HOLD=YES
R SDMODUx specifies
HOLD=YES
U SDMODUx does not
specify HOLD=YES
P SDMODVx specifies
HOLD=YES (spanned
records)
Q SDMODVx does not
specify HOLD=YES
(spanned records)
5 SDMODVx specifies
HOLD=YES
V SDMODVx does not
specify HOLD=YES

DAMOD

IJI

F RECFORM=FIXUNB
B RECFORM=U NDEF
(handles both UNDEF
and FIXUNB)
5 RECFORM=SPNUNB
V RECFORM=V ARUN B

A AFTER=YES
Z AFTER not specified

DIMOD

IJJ

F Fixed unblocked
record format

C ASA and System/360
control character
support for printers
and punches

F
X
U
D

C CONTROL=YES
Z CONTROL not
specified

I IOAREA2=YES
W WORKA=YES
B IOAREA2=YES and
WORKA=YES
Z Neither IOAREA2 nor
WORKA specified

T Device is in tope mode
D Device is in document
mode

ORMOD

MRMOD

IJM

IJU

Subsetting/
Supersetting
+ Permitted
* Not Permitted

RECFORM=FIXUNB
RECFORM=FIXBLK
RECFORM=UNDEF
RECFORM=UNDEF
and BLKFAC=YES

5 5 ing Ie address adopter
D Duo I address adapter

Figure 2.11.

TYPEFLE=INPUT

* + * *
IJM DCBDZ
F ZIT
W
U
X
Z

*
IJU DZZZZ
S

Generated Name Structure for Logic Modules (Part 2 of 2)

section 2: Data Management

117

Interrelationships of the DTF and Module
Macro Instructions

IHD

{

:~~

IJD
IJE
DATA
IJF
MANAGEMENT
IJG
IJH
IJI
IJJ
IJK
DATA
IJL
MANAGEMENT -IJM
IJN
IJO
IJP
IJQ
IJR
IJS
DATA
IJT
MANAGEMENT -IJU
IJV
IJW
IJX
IJY
IJZ
IKL
ILA
ILB
ILC
ILF
ILH

I Figure

2.12.

S/360 COBOL processing subroutines
System Service and System Control
Card logic
Printer logic
Paper tape logic
Magnetic tape logic
Sequential DASD logic
Indexed Sequential DASD logic
Direct Access DASD logic
Device independent logic
PL/I processing subroutines
Teleprocessing routines
Optical Reader logic
Audio Response Vocabulary File Utility
Disk Sort/Merge
Tape Sort/Merge
D-Ievel Assembler
RPG compiler
S/360 COBOL compiler
Basic FORTRAN, and FORTRAN subroutines
Magnetic Readers logic
Autotest
Utilities
PL/I compiler
F-Ievel Assembler
OLTEP
COBOL LCP
ANS COBOL compiler
ANS COBOL processing subroutines
CE Serviceability Programs
F-Ievel FORTRAN IV and FORTRAN IV subroutines
Tape and Disk Sort/Merge

The DTFCD, DTFDA, DTFDI, DTFIS, DTFMR,
DTFMT, DTFOR, DTFPR, DTFPT, and DTFSD
declarative macros are similar in that they
each generate a DTF table that references
an IOCS logic module (refer to B of Figure
2.13). The first 20 bytes of each table
have the same format, that is, a command
control block (CCB) and a logic module
address. The remainder of each table is
tailored to the particular device and file
type.
The description that follows is general
and includes all of the DTF types included
in B of Figure 2.13.

DTF Table and
Logic Module

DOS Relocatable Library
Module Name Prefixes

Reentrant Modules
A reentrant module is a logic module that
can be used asynchronously, or shared by
more than one file. The RDONLY=YES
parameter in the module generation macro
generates a reentrant logic module. The
RDONLY (read only) parameter implies that
the generated logic module is never
mOdified in any way regardless of the
processing requirements of any file(s)
using the module. To provide this feature,
unique save areas external to the logic
module are established, one for each task
using the module. Each save area must be
72 bytes and doubleword aligned. Before a
logic module is entered or an imperative
macro is issued to the file, the task must
provide the address of its unique save area
in register 13.

Partial
DTF Table

118

DOS System Programmer's Guide

CDMOD

DTFDA

DAMOD

DTFDI

DIMOD

DTFIS

ISMOD

DTFMR

MRMOD

DTFMT

MTMOD

DTFOR

ORMOD

DTFPR

PRMOD

DTFPT

PTMOD

DTFSD

SDMOD

DTF Tables

Logic Modules

DTFEN
DTFSR
DTFBG

I
8K BOS
Compatibility

Figure 2.13.
Reentrant modules include: CDMOD,
DAMOD, DIMOD, ISMOD, MTMOD, and SDMOD.

DTFCD

DTF and Module Macro
Relationships

When one of these DTF macro instructions
is encountered at assembly time, the
assembler builds a DTF table tailored to
the DTF parameters. The table contains:

DISK

DTFDA

• Device CCB (Figure 2.6).
DC V(IJIFZIZZ)

• A V-type statement used by the Linkage
Editor to resolve the linkage to the
logic module associated with this DTF.
• Logic indicators, that is, one I/O
area, two I/O areas, device type, etc.
• Addresses of all of the areas and
controls used by this device (except
work files).
Regardless of the method of assembling
logic modules and DTF tables (with the main
program or separately), a symbolic linkage
results between the DTF table and the logic
module. The Linkage Editor resolves these
linkages at edit time.
To accomplish the linkage between the
DTF table and the logic module, the
assembler generates a V-type address
constant in the DTF of a named CSECT in the
logic module. To resolve this linkage, the
linkage symbols (module names) must be
identical.
Figure 2.14 shows the relationship of
the program, the DTF and the logic module.
It also shows a DTFDA macro with a filename
of DISK. The assumed parameters have
generated a request for a DAMOD named
IJIFZIZZ. Based on this name, the linkage
editor was able to locate the module. The
read statement generated coding to load the
address of the DTF table into register 1.
This gives the program access to the DAMOD
address, and the program branches to the
required routine within the module.

IJIFZIZZ

CSECT

READ DISK, KEY
*L 1,=A(DISK)
*L 15,16(1)

*BAL 14,24(15)

Get DTF table address.
Get logic module address
located 16 bytes from
start of DTF table.
Branch to read routine in
logic module.

* Macro generated coding
Figure 2.14.

Coding Example of DTF and
Module Relationship

generates an in-line expansion that links
the instruction to the DTF table (thus the
logic module) for the specified file. For
an operand, the imperative macro
instruction must always contain the
filename in the DTFxx macro describing the
file.

IMPERATIVE MACROS
The problem programmer issues imperative
logical IOCS macro instructions to initiate
such functions as opening a file, making
records available for processing, writing
records that have been processed,
controlling physical device operations,
etc.

Figure 2.15 summarizes the macro
instructions provided by IBM for logical
IOCS. Figure 2.16 further defines the
general function of each of the macro
instructions and indicates the devices with
which they are used. The Supervisor and
I/O Macros listed in the Preface gives a
detailed explanation of each of the
imperative macros.

For each imperative macro issued by the
problem programmer, the assembler program

Section 2: Data Management

119

MACROS
CHECK
CLOSE(R)
CNTRL
DISEN
DSPLY
ENDFL
ERET
EXETL
FEOV
FREE
GET
LBRET
LITE
NOTE
OPEN(R)
POINTR
POINTS
POINTW
PRTOV
PUT
RDLNE
READ
RELSE
RESCN
SEOV
SETFL
SETL
TRUNC
WAITF
WRITE

DTFCD DTFCN DTFDA DTFDI
X
X

X
X

DTFIS

DTFMR DTFMT
X
X
X
X
X
X

DTFOR DTFPH
X
X

X
X

DTFPT
X

DTFSD
X
X
X

DTFSR
X
X

X
X
X

X
Xl
X
X

X
X
X

X
X

X
X
X

X
X

X
X
Xl
Xl
Xl

Figure 2.15.

X
X

X
X
Xl
Xl
Xl

X
X

X
X
X
X

X
X
X
X
X

X
X

1. Work files only.

120

DTFPR

Logical Ioes Imperative Macros and DTFs

DOS System Programmer's Guide

TYPE OF PROCESSING WITH LOGICAL IOCS
Indexed Sequential File
Management System

Sequential

Macro Instruction
"0

... c
a.. 0
N...o

>..

l/"')

Initial ize

Process

Set Mode

Non Data
Operations

Work Files
for DASD
and
Magnetic
Tape

Complete

Notes:

l/"')

....

OPEN(R)
LBRET 1
GET
PUT
READ
WRITE
CHECK
RELSE 5
TRUNC"
WAITF
RDLNE
RESCN
DSPLY
SETFL
ENDFL
SETL
ESETL
CNTRL
CHNG 7
PRTOV
DISEN
LITE
ERET
READ
WRITE
CHECK
NOTE
POINTR
POINTW
POINTS
CLOSE(R)
FEOV
FEOVD
FREE
LBRET 1
SEOV

Q)
l/"')"'O

co
N

C
Q)

~....c:

x
X

(")

X
X
X
X
X

X
3

x
x
x
x
x

x
x

'"
:G

U
U

~
u..
"'0

-l

~a..

x
X

c
a......c:
co U
~
c:
o ::>

~a..

~ §

Qj
a.

C
o

l/"')

1'-."'0

co
N

o . . o ...

x
x
x

~
o

x
x

'"
~
o

~
E

Q)
~

"'0
"'0

"'0

X
X
X

X
X

x
x

X
X

X
X
X
X
X
X
X
X
X

Section 2: Data Management

121

X
X
X

X
X
X
X
X
X

X
X

1. Applies only if DTFSR, DTFMT, DTFDA, or DTF PH LABADDR or XTNTXIT is specified.
2. In the 2520 or 2540, GET normally reads cards in the read feed. If TYPEFLE =CMBND is specified, GET reads cards at the
punch - feed - read station.
3. PUT rewrites on input DASD record if UPDATE is specified.
4. In the 1442, 2520, or 2540, PUT punches an input card with additional information if TYPEFLE = CMBND is specified.
5. Applies only to blocked input records.
6. Applies only to blocked output records.
7. Provided only for upward compatibility for BPS and BOS.
8. Workfi les on Iy •
9. Applicable to 1419 and 1275 with the Pocket Light Feature.

Figure 2.16.

Logical IOCS Imperative Macros and Devices

SEQUENTIAL FILE ORGANIZATION

Line Printers

Card Files

The DTFPR/PRMOD access method provides the
ability to print a record by issuing a PUT
macro instruction. The record to be
printed can be presented to the access
method via a work area (WORKA) or can be
placed in the buffer area. Two buffer
areas (IOAREA1 and IOAREA2) can be
specified to allow overlap processing. In
this case, a work area or a general
register must be used to indicate the
proper buffer area. Three types of
printer-form control are provided by the
access method:

The DTFCD/CDMOD macros provide the user
with the ability to read (GET) a record
from a card reader or punch (PUT) a record
on a card punch (up to 80 characters in
both cases). The user has the option of
specifying one or two I/O areas (IOAREA1
and IOAREA2), also called buffer areas, to
hold the data transferred to main storage
from the device, or vice versa. The second
buffer area allows the user to overlap
processing while the following record is
read or the previous record is punched.
The user can process the data read (or data
to be punched) in the buffer area if only
one area is specified. If two buffer areas
are specified, a register (IOREG) can be
specified to point to the start of the data
field (leftmost position) in the current
buffer area, or the data can be placed into
a work area (WORKA) for processing (or
punching).
In the latter case, DTFCD/CDMOD
transfers the record from the buffer area
to the work area or vice versa. The
capability for reading a card record and
punching information into the same card
(CMBND) is also provided if the user has an
IBM 1442, and IBM 2520, or an IBM 2540 with
the punch-feed-read (PFR) feature. Many
data processing installations use the
stacker selection capabilities made
available by the following three methods:

The user may issue a CNTRL macro
instruction after a GET or before a
PUT to select the desired stacker,

First character control (CTLCHR) may
be used, where the first character of
a record may be an ASA (American
standard Association) or System/360
control character for stacker
selection, and

The selection for a given file may be
specified (SSELECT) when the DTF table
is generated.

• CNTRL macro instruction for line
spacing or page skipping.
• PRTOV (printer overflow) macro
instruction for page skipping or
exiting to a user-supplied routine
(indicated in the macro instruction)
that can perform certain end-of-page
and/or start-of-page functions.
• First character control (CTLCHR) that
can be used where the first character
of a record may be an ASA or system/360
control character for line spacing or
page skipping.

Paper Tape Files

The DTFPT/PTMOD access method provides the
ability to retrieve a data record from a
supported Paper Tape device. Two buffer
areas (IOAREA1 and IOAREA2) can be used for
overlap processing. In this case, a
general register must be specified to point
to the record in the buffer area currently
being used. This access method also
handles shifted code for figure shift
(FTRANS and SCAN) and/or letter shift
(LTRANS and SCAN), or nonshifted code to be
translated into system/360 and System/370
code (TRANS). The user must supply the
various translation tables needed.

Console Typewriter

Magnetic Tape Files

The DTFCN macro allows reading and writing
of a record from or to an IBM 1052
Printer-Keyboard or an IBM 3210 or 3215
Console Printer-Keyboard by issuing a GET
or PUT macro instruction. The record may
be processed in the buffer area or in a
work area (WORKA). Only one buffer area
can be used.with this access method. This
file does not require the use of OPEN and
CLOSE routines.

The DTFMT/MTMOD access method provides the
ability to create or retrieve magnetic tape
records in sequential order. The file is
created by indicating via the DTFMT
parameters that it is an output file
(TYPEFLE) and then issuing PUT macro
instructions.
If the records are to be
retrieved, an input file (TYPEFLE) is
indicated and a GET macro instruction is
issued. In either case, the records may be

122

DOS System Programmer's Guide

processed in a work area (WORKA) or in the
buffer area by using a general register
(IOREG). When an input file is to be
processed, the access method can also
support the read-backward feature for
magnetic tape units.

Sequential Disk
Records can be created or retrieved and
updated from a direct access device by
using the DTFSD/SDMOD access method. The
file can be created by specifying an output
file (TYPEFLE) and issuing PUT macro
instructions. If the records are to be
retrieved, an input file is indicated
(TYPEFLE) and GET macro instructions are
issued. It is also possible to update
those records in the same location on DASD
that were retrieved by a GET macro
instruction. In this case, a PUT macro
instruction must be issued for the file
after the GET for the record to be updated
and preceding the GET for the next record.
The access method also provides double
buffering (IOAREA1 and IOAREA2) ability for
overlap processing. The user can either
process the record in a work area (WORKA)
or use a general register to point to the
record in the current buffer area. Another
macro instruction allows the user to skip
the remaining records in an input block
(RELSE). This can be useful if records are
grouped by specific categories. An inverse
function of RELSE, allows the programmer to
truncate or write short blocks of records
(TRUNC) for an output file. The CNTRL
macro instruction may also be used to seek
the track address of the next record to be
processed.
In the case of a data cell,
CNTRL can restore a strip if the user knows
that processing on it has been completed.

When an ASCII file is to be processed,
each record read into the I/O area is
translated from ASCII to EBCDIC. When an
ASCII file is to be created, the record in
the output area is translated from EBCDIC
to ASCII just before i t is written on the
tape.
Other optional features that greatly
enhance the DTFMT/MTMOD access method are:
• Specifying two buffer areas for overlap
processing capabilities (IOAREA1 and
IOAREA2).
• Alternate tape switching between two
tape units.
• Bypassing of checkpoint records on
input files (CKPTREC). An ASCII file
must not contain checkpoint records.
Other macro instructions allow the user
to rewind, rewind and unload, execute
various other magnetic tape device
functions (CNTRL), or release (SKIP) the
remaining records in an input block
(RELSE). This can be useful if records are
grouped by specific categories. An inverse
function of RELSE allows the programmer to
truncate or write short blocks of records
(TRUNC) for an output file.
In addition to GET/PUT functions,
DTFMT/MTMOD provides the highly useful
feature of issuing READ/WRITE macro
instructions ~o create a file or retrieve
records from a work file (WORK). OVerlap
processing can take place while the I/O
operation is being performed. The user
awaits the READ/WRITE operation by issuing
a CHECK macro instruction that ensures
completion of the operation. The
particular features that enhance this
facility are the NOTE and POINT macro
instructions. By issuing a NOTE macro
instruction, the location of the data block
in the file can then be obtained. POINT
macro instructions provide the ability to
reposition to a given block in the file.
POINTR can position the tape to the block
indicated, POINTW can position the tape
after the block indicated, and POINTS can
position the tape to the beginning of the
file.

DTFSD/SDMOD also provides the
READIWRITE, CHECK, NOTE, and POINT macro
instructions described earlier under
magnetic tape files, although an IBM
2311/2314/2319/2321 DASD is used as an I/O
device in this case. NOTE/POINT uses
cylinder, track, and record identification
for noting and locating blocks in the file.
Also, if a NOTE follows a WRITE, the unused
space on a track can be returned when the
file is being created or when records are
to be written in the count, key, and data
format of DASD. If a POINTR or POINTW is
issued before a READ or WRITE (UPDATE), the
READ or WRITE macro instruction processes
the block indicated. If a formatting WRITE
(write count, key, and data) is issued, the
WRITE macro instruction writes a new record
after the block indicated.
Another feature available with this
access method that can be used in either of
the two mentioned processing modes, is the
split cylinder mode. This mode allows two
or more files to share the same cylinder.
Each file occupies the same track positions
through the range of assigned cylinders.
The technique has the advantage of
minimizing access-arm movement when cross
referencing among two or more files that
perform similar functions.
The use of this
Section 2: Data Management

123

facility is indicated in one of the job
control statements (XTEN'I'/EXTENT) needed at
execution time of the user's program. See
System Control and system service listed in
the Preface for detailed information on the
job control statements.

whether the I/O operation was
successful to the problem program in
the field specified by DTFDA ERRBYTE.
Three other options, enhancing the
capabilities of the access method, are:
1.

The ability to utilize the multiple
track search facility of a DASD
(SRCHM).

The ability to return the record
location (the first five bytes of the
record identification, namely CCHHR)
when retrieving a record by its key
(IDLOC). These features facilitate
skip-sequential processing. For
example, by issuing a READ (KEY) macro
instruction and starting at the
beginning of the file, the direct
access device searches multiple tracks
until the record is found.
The record
location is returned and is used as a
reference point in starting the search
for the next record to be retrieved.
In this manner of processing, only the
required records are read, whereas in
the normal sequential mode of
processing, all records are read even
if they are not all processed.

The ability to seek ahead by issuing a
CNTRL macro instruction (or restore a
strip to an IBM 2321 Data Cell) and
then continue normal processing or
issue other I/O device macro
instructions.

DIRECT ACCESS METHOD (DAM) FILE
ORGANIZATION
Great flexibility in reading or writing a
record from or to a direct access device is
achieved by the DTFDA/DAMOD access method.
with the WRITE (AFTER) macro instruction,
the user may create a file in any manner
desired. For example, a part number or
control field can be converted to a DASD
address using a randomizing algorithm, and
the record can be written at that disk
address. Also, the user can create a
sequential file with keys (control fields)
to be processed later in a skip-sequential
fashion.
In skip-sequential processing, a file
that has been sorted on a key control
information field is created in a
sequential manner. Records are retrieved
by scanning or searching the file using the
key of each record desired. In this way,
only these records are retrieved and the
job throughput is improved when a large
percentage of the file is processed.
Record retrieval is accomplished by
issuing a READ macro instruction. Two
types of READ macro instructions may be
issued for record retrieval. The user may
read the record by simply supplying the
track and record location (ID), or by
supplying the track location along with the
record key (KEY) to be used for record
retrieval. Record updating is performed by
two corresponding WRITE macro instructions
(ID and KEY). They work in the same way as
the READ macro instructions.
In addition to the READ/WRITE macro
instructions, there is a WAITF macro
instruction. This macro instruction is
issued after each READ/WRITE. The WAITF
macro instruction serves a two-fold
purpose:
1.

It allows overlap processing or
issuing of another I/O device macro
instruction (within the limitations of
the system configuration) while the
present I/O operation is taking place,
and

It communicates termination of the
READ/WRITE operation with which it was
associated, and returns any
error/status information indicating

124

DOS System Programmer's Guide

When creating a file using the WRITE
(AFTER) macro instruction, the access
method automatically maintains the amount
of space available on each track on which
the records are stored along with the
address of the last record on each track
(provided the DASD has been properly
initialized). This access method does not
provide blocking and deblocking facilities.
They must be performed by the user.

Random Addressing Techniques
In addition to the specification of the
DTFDA and DAMOD macros, the problem
programmer must supply the DASD address of
the record to be read or written before
issuing each READ/WRITE imperative macro.
The following discussion presents various
methods for determining a DASD address from
a record control field.
File addressing involves the file of
records that must be stored and retrieved
in a data processing system, and the direct
access storage device itself. The data
records that must be stored in a direct

Random Addressing Formula

access storage device are usually
identified by a control field, such as part
number, and employee number. Normally, the
numbers or characters in the control field
are unevenly distributed. For example, a
seven-positioned control field may be used
to identify 25,000 items in a parts master
file. However, with a seven-position
number, it is possible to identify ten
million items. In this example, only 0.25%
of the available numbers are used.
The direct access storage devices, on
the other hand, are usually composed of
physical locations that are identified by
an evenly distributed set of numbers. The
addressing problem converts an unevenly
distributed set of numbers to an evenly
distributed sequential set of numbers
within the address limits of the direct
access device. Many addressing techniques
have been developed to accomplish this
task. In choosing a technique for address
conversion, it is important to remember
that an ideal distribution of control
fields is a completely uniform one.
Uniform distribution means that the
difference between any pair of successive
control fields taken in ascending order is
constant.
The worst distribution of control fields
is a random one. There is no way to
transfer from random keys to addresses with
better than random distribution. In
practice, purely random control field sets
and completely uniform ones are rare. A
data file is likely to have control fields
that distribute in groups or clusters of
irregular length and separation. This kind
of grouping of numbers introduces a degree
of uniformity. The irregular length and
separation of the number groups implies a
degree of randomness. A well chosen
conversion technique produces an address
set that reflects both elements and has a
distribution intermediate between random
and uniform. TO be ideal for use in direct
access storage devices, the conversion
technique should produce a unique storage
address for every record in a file. This
is seldom possible. Most control-field
conversion routines result in assigning
some address to more than one record.
These duplicate addresses are sometimes
referred to as synonyms. The selected
conversion routine should convert the
control fields (keys) of the records in a
data file to a series of addresses with a
minimum number of synonyms and within the
desired storage address range. The
following sections discuss briefly the most
successful conversion routine, followed by
a discussion of synonym handling.

The simplest method of file organization is
that in which a unique DASD address is
obtained from the control data of each
record. This is referred to as the random
addressing method. If the control numbers
of a set of data records in a file are
consecutive numbers without gaps, they may
be converted to DASD addresses by simple
arithmetic. For example, if the account
numbers for a customer file run from 10000
to 17563 (7564 account numbers), and ten
account records can be stored on each disk
track, 757 tracks are needed. By
subtracting 10,000 from an account number
and then dividing by 10, a numeric address
in the range 000 to 756 is obtained. To
place this file on a 2311/2314/2319 disk
drive, starting at track address 1200
(cylinder 120 head 0), a constant 1200 is
added to the quotient and a constant (1) is
added to the remainder. This constant (1)
is required because record zero (RO) of
each track is reserved to facilitate the
handling of defective recording areas that
may occur during the life of the disk pack.
Using this approach, a record containing
the data for account number 16349 would be
stored at track reference 1834 in
record-reference ten, calculated as
follows:
16349 - 10000
6349 • 10

634 + 1200

= 6349

634 with remainder 9

= 1834 = track

9 (remainder + 1)

reference

= 10 = record

reference

When processing this file randomly, any
record can be found with a single seek.
When it is possible to process
sequentially, only one seek is needed per
cylinder. Record retrieval time is thus at
a minimum. This is an optimum situation,
and it rarely occurs in actual practice.
Normally, the control data of a file of
records can seldom be used directly as DASD
addresses. If a file does not have control
fields that can be used directly as DASD
addresses, it is sometimes possible to
preassign addresses. For example. the item
number 513XP could become 513XP-13472, that
could then be converted to a track and
record reference.

Section 2: Data Management

125

Prime Number Division
If the control fields of a file of records
are not consecutive or contain numerous
unused numbers, as is usually the case, the
random addressing technique under the topic
Random Addressing Formula makes inefficient
use of storage locations. All possible
numbers are assigned locations, and those
numbers not used leave empty record areas
in the storage unit. Files established
with control numbers composed of coded
information usually have a much higher
potential range of items than is required
for storage. To handle this situation,
initial conversion is made on the control
numbers to reduce the range to a practical
size. This conversion is often referred to
as randomizing.
Randomizing generally refers to the
techniques developed to convert a set of
control numbers with numerous unused
numbers to a tightly packed set, to result
in very few unused storage areas. There
are many techniques used for this
conversion of numbers, a few being:
folding, extracting, squaring, and radix
transformation. One method, sometimes
called prime number division* or divide
remainder, is adaptable and usually
satisfactory for converting a file of
numbers.
To illustrate the prime number division
technique, suppose the customer file in the
example under the topic Random Addressing
Formula used a coded control number of ten
digits. The first three could be a
geographical code (branch office number),
the next two could describe the nature of
the business, the next one could be a
size-of-customer code, and the final four
could be sequentially assigned within
class.
Thus, account number 139 457 0307 would
be the 307th account assigned branch office
139. It would belong to a customer-of-size
code 7 in industry class 45. Because this
ten-digit number cannot be used efficiently
to describe 7564 accounts, it is converted
by dividing by the closest prime number to
the number of storage locations available.
Assume 10,000 locations available, then
divide by 9973. The remainder serves as
the control number and a technique similar
to the example under the topic Random
Addressing Formula calculates a track and
record reference.

* A prime number is a number divisible only
by itself, or one.
126

DOS System Programmer's Guide

1394570307
of 5825.

9973

= 139834

with a remainder

This remainder is operated on as in the
example, under the topic Random Addressing
Formula, assuming 10 records per track.
5825

582 with a remainder of 5.

To load the file starting at track address
1100, add 1100 to 582 for a sum of 1682.
The track reference for this record is
1682, and the record is the sixth record on
the track (remainder of 5 + 1 = 6).
TO summarize prime number division:
1.

Select a divisor equal to or greater
than the number of records to be
stored (10 to 20% greater is
recommended). The best divisors are
primes. Even numbers or multiples of
five should never be used - divisors
must end in 1, 3, 7, or 9. The
divisor is called the range.

Divide the control number by the range
and use the remainder to generate the
track address.

Prime number division always works; that
is, it always converts control numbers into
the desired range because in division, the
remainder is always less than the divisor
and the highest valued remainder is the
divisor-I. Dividing any number, no matter
what size, by a desired range, always
produces remainders in the desired range.
Using a prime number as the divisor usually
results in relatively few duplicate
remainders, and therefore relatively few
address synonyms.
A prime number is not always the best
choice of divisor for a given set of keys.
Also, i t is not necessarily true that all
primes produce equally good results.
However, primes avoid serious
maldistribution and may be safely used with
little analysis of the control field set of
the data files.

Synonym Records
The transformation of record control fields
to direct access storage device addresses
usually produces some synonym records.
The file organization used with a data
file employing indirect addressing
(addresses converted by a random addressing
formula) must be able to accommodate the
synonyms or duplicate addresses.
The first consideration in organizing
the file is sometimes called the packing
factor. The number of synonyms produced by
a random addressing conversion routine can
be reduced by assigning more DASD storage
space than is actually required by the
file.
The percentage of the file area
actually used for record is called the
packing factor. The packing factor for an
efficiently organized file can vary from
65% to 95~. A packing factor of 80%
usually proves to be a good starting point.
After all efforts have been made to design
a file conversion technique with few
synonyms, an approach to handling the
remaining synonyms must be chosen.

and another read command is issued using
this address. The process is repeated
until the desired record is found.
Record Reference

Overflow
Address Field

CC HH R

73 06 1

CC HH R

74 02

74 02 3
74 09 2
74 09 2
69 06

69 06 4

74 35 6

Blank

CC = cylinder
HH = head
R = record

One such technique is often referred to
as the chaining method. As each record is
read into the computer for loading into a
direct storage access device, its control
field or key is converted to a physical
address. These converted addresses are
called home addresses.

Figure 2.17.

Note: The home addresses discussed here
are-not directly related to the track home
address used to control the "physical
operation of the DASD. These home
addresses are related to the record
identifier (ID) associated with DAM.

Two other techniques that solve the
synonym problem are similar in concept to
the chaining method but do not require a
chaining field to be present in each
record.

The first record converted to a
particular address is stored in the home
address location. The additional records
converted to this address are stored in
synonym locations. The address of the
first synonym location is stored in the
home address location. The address of the
second synonym location is stored in the
first synonym location, etc. Chaining
requires that in the home address and all
synonym locations, space be reserved for
the address of the next location or link in
the chain (Figure 2.17).
Retrieval of records is accomplished by
converting the control information (record
key) to the home address. The record in
the home address location is read into main
storage, and its control information is
compared to that of the record being
sought. If the control fields are not
equal, the address of the first overflow
record is extracted from the home record,

Direct Access Address
Chaining

Preassigned tracks synonym technique.
When there is not room to store a
record in its home location, a
specific preassigned synonym track is
used. The synonym track(s) should be
defined to be in the same cylinder
containing the home location to reduce
the number of seeks required to locate
a record.

Consecutive spill synonym technique.
When a synonym occurs, a sequential
search is made starting at the next
record within the cylinder until an
empty record storage location is
found.
If the last track in a given
cylinder overflows, a return is made
to the first track. This technique
does not require the use of a chaining
field, nor is a seek required to
locate the synonym record (must be on
same cylinder).
Section 2: Data Management

127

DASD Address Specification For Read/Write
Operations
The direct access method requires DASD
addresses for all read/write operations.
These addresses may be supplied in one of
two ways: as an actual physical address
(MBBCCHHR) or as a relative track address
(see Relative Track Addressing).
The actual physical address can be
specified as an 8-byte binary address (see
Figure 2.18) in the form MBBCCHHR. These
8-byte addresses are used either as the
starting point for a search on record key
or as the actual address for a READ or
WRITE ID. When searching for a record key,
the programmer may specify that the search
be only within the specified track, or from
track to track, starting at the address
given and continuing either until the
record is found or until the end of the
cylinder is reached.

---,

Volume
Number
(M)
Bytes_
May __
Contain

I
I

0
0-221 :

1 0
1

0
0

1
0-221 I

Address Specified by
SEEKADR=Name

Figure 2.18.

3
1
r

I
1

4
I

: 0-199 1 0
I

I
I

I
I
1
1
1 0-9 10-19. 1 0-9 I 0-4 10-19 10-2551 2321
1
I
I
1
r
I
r
1
I
10 - 19 10 - 255 1 2314
I
0 10-1991 0
0
1
J
1
I

0-221 :

1
Record:
(R) r
I
1
1 6
7
I
---I
1
1 0-9 10- 255 1 2311
1
I

Head
(HH)

Cylinder
(CC)

Cell
(BB)

DASD Address Formats

DOS System Programmer's Guide

Relative Track Addressing
The required DASD addresses may also be
given as a relative address. This address
is then converted by IOCS to the actual
physical address (see Figure 2.18).
Relative track addressing is more
convenient to use than the actual physical
address for the following reasons:
1.

The data in the file appears to be one
logically-continuous area, although it
may be physically noncontiguous.

The user needs to know only the
relative position of the data within
the file; its actual physical address
is not required.

ReqUired for Record
Reference by I D

For certain types of operations, the
system can be requested to return the ID
(CCHHR) of the record read or written or of
the next record following the one read or
written. The programmer can place these in
the 8-byte address field to either READ or
WRITE a new record or to update the one
read. For example, to delete a record from
a random file with keys, the programmer can
randomize the record key to a starting
location, search on key to read the record,
and then use the ID returned to write a
blank or zeroed record (key and data) back
into the same location. The descriptions
of the READ and WRITE macro instructions
explain when the ID can be returned and
whether the ID returned is that of the same
or of the next record. See Data Management
Concepts listed in the Preface for a
description of these macros.
128

When the ID returned is that of the next
record, the system obtains the ID by
chaining to a read-count command. This
command skips to the next track if the
record read or written was the last
currently on the track. The system does
not read the next ID if the end of a
cylinder is reached. In this case, it adds
one to the CC portion of the previous ID,
forces the HH portion to 0, and forces R to
1 for a 2311, 2314, or 2319 file. For a
2321 file, it adds one to the high order H,
forces the low order H to zero, and forces
R to one. An overflow from the high order
H increases the low order C by one, forces
both Hs to zero, and forces R to one.
Subsequent overflows of address locations
cause increases in the next higher
positions of the addresses. (It is the
user's responsibility to check the validity
of the address returned in IDLOC.)

The relative address may be specified by
the user in either of two/ formats:
hexadecimal (in the form TTTR), or zoned
decimal (in the form TTTTTTTTRR).
In both the hexadecimal and zoned
decimal format, the Ts represent the track
number relative to the start of the data
file, and the Rs represent the record
number on that track. The hexadecimal
format requires 4 bytes, while the zoned
decimal format require~ 10 bytes. Relative
track addressing is implemented through the
DTFDA macro. Parameters in this macro .
specify the number of extents in the file.
the form of relative addressing used
(hexadecimal or zoned decimal), and other
required information. For specific
information on the implementation of
relative track addressing, see the
Supervisor and I/O macros listed in the
Preface.

INDEXED SEQUENTIAL FILE MANAGEMENT SYSTEM
(ISFMS)
The DTFIS/ISMOD facility is both an access
method and a DASD file organization
technique. Facilities are provided for you
to create a file, add new records in any
order to a previously created file, and
retrieve all records in the file either
randomiy or sequentially.
The track hold facility is also provided
for protecting DASD tracks that are
currently being processed. Track hold
prevents two or more programs from updating
the same record at the same time provided
all the programs use the facility.
The file is created in ascending
sequential order from the input that has
been previously sorted on the record keys.
As the file is created, an index
hierarchy is developed. The ~owest level
is called the track index and occupies the
first track (in the case of an IBM 2321,
one or possibly more tracks) of each
cylinder that is contained in the file area
called the prime data area. This index
includes a pair of index entries for each
track of the cylinder containing the user's
data records. The first entry of each pair
indexes the highest record key on the
appropriate track being referenced. The
second entry locates overflow records that
can occur from that track when new records
(additions> are added to the file.

The second level of index, called a
cylinder index, is generated on a DASD area
separate from the prime data area. An
entry is made in the cylinder index for the
highest record key of each cylinder in the
prime data area, and each entry points to
the track index on the appropriate
cylinder.
A third, optional level of index, the
master index, is generated in the same area
as the cylinder index and precedes it. The
master index has an entry for the highest
key on each track within the cylinder index
area. For a small file, this level of
index is generally not needed because
searching a cylinder index of two or three
tracks is as fast as searching the master
index and then the cylinder index.
An index entry is composed of a key and
an address. The track index is composed of
two types of entries, a normal entry and
overflow entry for each prime data track
within that cylinder. In the example
illustrated in Figure 2.19, the normal
entry indicates that the highest key on
track 1 is 8, and the address is at the
beginning of track 1. The overflow entry
indicates the same key after loading as the
normal entry. The address of hexadecimal
Fs indicates no entries for this track in
the overflow area.

Note: If the track index does not occupy
all of track 0, track 0 also contains data
records.

Section 2: Data Management

129

IHighest Key on Cylinder I

Cylinder Index
Only First Entry Shown

(
Track Index
Entries for Only
Track 1 Shown

Track Index Address

Data

_-----N-O-R-M-~'-E-N-T-R-Y----____.,~
I
'I 1

Highest Key on Track

Address of First
Record on Track
Track 1
Record a

OVERFLOW_EN_TR_Y_ _ _ _ __

H.ighest Key Associated
. . With Track

1
Address of Lowest Recor I
. in Overflow Chain
.
FF

~-------------------------------------V ----------------------------------)

Figure 2.19.

Example of Track Index

If record 7 is added to the file as
illustrated in Figure 2.20, record 8 would
be forced into the overflow area, and
record 7 would be the highest record on
track 1. In this case, the key of normal
entry is changed to 7, and the address of
the overflow entry indicates the location
of record 8, which is on track 8 record 1.
No other changes to the indexes are
required.
The cylinder overflow control record
(COCR), that is maintained in the data
portion of record zero, would indicate that
no records were in the cylinder overflow
area after loading. After the addition of
record 7, it would be updated to indicate
that the last record in the cylinder
overflow area was on track 8, record 1.
The cylinder index indicates that record
32 is the highest key on the cylinder, and
its address points to the track index. The
cylinder index has its own extents that
must be defined at job control time. These
130

DOS System Programmer's Guide

extents must be outside the limits of any
data extents. (Note that at least two sets
of extent information must be defined at
job control time. These would define a
data extent and a cylinder index extent.)
If the file exceeds one disk pack,
additional data extents are provided, and
at the user's option, an independent
overflow area extent may be defined. The
cylinder index extent must be on-line when
the file is processed. It may be on the
same pack as the data file, or it may be on
a separate pack.
Because the file is sequential with a
hierarchy of indexes, it is called an
indexed sequential file. The indexes
provide direct reference to records,
allowing their random retrieval with a
minimum of search time. The sequential
order of the data records, coupled with the
ability to reference overflow records in
sequence via the track index, provides
sequential retrieval capability. The
indexed sequential method consists of four

Cylinder Index

IHighest Key on Cyl inder ,

Track Index Address

Data

Points to Track Index

NORMAL ENTRY
(

OVERFLOW ENTRY

r--------------~'-------------_
Address of First
\

_____T_r_ac_k_ln_d_e_x_---"

I Highest Key on Track I

Points to Data Track

Record on Track

----------------~~----------------Address of Lowest Recor~

(Highest Key Associated
with Track

I I

Track 1
Record 0

Iin Overflow Chain

Track 8
Record 1

~----------------------------------y ---------------------------------)

o
1

2
Prime Data

8
Cylinder Overflow

Figure 2.20.

Example of Track Index after Addition to File

basic functions that provide capabilities
for creating a file, adding new records to
it, and retrieving the records. A
description of each function follows.

(IOAREAL) and then written in the prime
data area. The appropriate index entries
are also made. After the user has
presented all the data records needed to
create the file, an ENDFL macro instruction
(end file load mode) is issued to terminate
the load.

Loading an Indexed Sequential File
A file and its associated indexes are
created by the load function.
In addition
to the DTFIS macro instruction specifying
IOROUT=LOAD, there are three other macro
instructions used to create the file. The
first one to be issued is a SETFL macro
instruction (set file load mode) that does
the initializing needed for file creation.
When the WRITE (NEWKEY) macro instruction
is issued, the key and data record placed
by the user in a work area (WORKL) are
moved to an output block in the buffer area

Facility is provided for protecting DASD
tracks that are currently being accessed.
For ISAM. the hold applies to both data
records and index records. Because track
hold cannot be performed on a LOAD file,
BOLD=YES cannot be specified when
IOROUT=LOAD.
Facility is also provided to extend the
file by adding new records higher in
collating sequence than the current last
(high) record in the file. Using the same
user program, this can be done by
specifying ISE in the DLBL/DLAB job control
Section 2: Data Management

131

statement (required at execution time of
the user's program) to indicate a file
extension is to take place.
The same macro instructions that load a
file originally can be used to extend the
file. If it is necessary to increase the
size of the file to contain records with
keys higher than the last key on the
original file, the records can be loaded at
the end of the file. A job control
XTENT/EXTENT statement extends the upper
limit of the prime data area of the file so
that the new records can be loaded into the
file.
(Overflow area is not required
because the file is merely extended further
on the DASD.)

Adding Records to the File
The add function provides the ability to
insert new records in the file.
In
addition to the DTFIS macro instruction
specifying IOROUT=ADD, there are two other
macros used for adding records to a file.
The first is a WRITE (NEWKEY), that
initiates the addition process (searching
the indexes, etc.) and returns control to
the user to allow overlap processing. To
complete the addition operation, a WAITF
macro instruction must be issued. This not
only ensures that all necessary I/O
operations have been completed but also
returns status conditions indicating any
abnormal operation. When additions are
made to the file, the user presents the key
and data record to be added in a work area
(WORKL). A search is made through the
index structure to determine where the
record is to be inserted in the file. The
record is either inserted in key sequence
in the prime data area, or placed in the
overflow area by use of a chaining
technique that maintains the proper
sequence of the file.
Two overflow area options that may be
used in any combination are provided. One
option allows the user to specify that one
or more tracks be reserved at the end of
each cylinder to store overflow records
(CYLOFL). The second option allows an
independent overflow area separate from the
prime data area to be reserved for storing
overflow records. The first option has the
advantage of reducing access time for the
retrieval of overflow records associated
with a given cylinder~ The second option
has the advantage of utilizing DASD space
more efficiently.

132

DOS System Programmer's Guide

Random Retrieval
The random retrieval function is used for
random retrieval and updating of records.
In addition to the DTFIS macro specifying
either IOROUT=RETRVE or IOROUT=ADDRTR and
TYPEFLE=RANDOM, there are three other
macros for randomly retrieving and updating
records. The first of three macro
instructions to be issued for this purpose
is a READ (KEY). The READ macro
instruction performs a search of the
indexes, using the key of the requested
record (KEYARG) provided by the user as the
search argument. While the I/O operations
that perform this function are taking
place, control returns to the user to allow
overlap processing. To complete the READ
function and receive the record, the user
must issue a WAITF macro instruction. If
the operation is completed successfully,
this macro instruction either places the
record into a work area (WORKR) or points
to the starting location (leftmost
position) of the record within the buffer
area by using a general register (IOREG).
If the operation was not successful,
indications of the resulting abnormal
conditions are given in the filenameC
status byte in the DTF.
If the user wants to update and return
the record to the file, he must issue a
WRITE (KEY) macro instruction. The WRITE
follows the READ of the record to be
updated and precedes the READ for the next
record. Again, processing can overlap
execution of the WRITE instruction. To
complete the operation, the user must issue
a WAITF macro instruction.
Sequential Retrieval
The sequential retrieval function makes it
possible to sequentially retrieve and
update records. In addition to the DTFIS
macro instruction specifying either
IOROUT=RETRVE or IOROUT=ADDRTR and
TYPEFLE=SEQNTL, there are four other macros
for sequentially retrieving and updating
records. The first of four macro
instructions to be issued for this purpose
is the SETL (set lower limit), that locates
the starting point where retrieval begins.
This macro instruction provides four
methods of starting retrieval:
• From the beginning of the file (BOF)
• At any record location in the prime
data area (ID)
• with any record in the file by
supplying the key of the desired
starting record (KEY)

• With the first record of a group of
records in the same class by supplying
the generic key for that class; a class
being any group of records that
contains identical control information
in the first few high-order bytes of
the record keys (GKEY, a key equal to
or lower than the first record of the
desired group)
After the SETL has been successfully
executed, the user can issue GET macro
instructions to retrieve each record in the
file in key sequence. The record can be
placed in a work area (WORKS), or a general
register can be used to point to the
starting location of the record in the
buffer area.
If the user wants to update
the record and return i t to the file, he
must issue a PUT macro instruction. This
POT must follow the GET of the record to be
updated and precede the GET of the next
record. After all of the desired records
have been processed, the sequential
retrieval function is terminated by issuing
an ESETL macro instruction. This process
of issuing the SETL. GET. PUT, and ESETL
macro instructions can be repeated as many
times as desired. By combining the macro
instructions of the sequential retrieval
and load functions, the file can be
reorganized. The user can retrieve the
current file in its proper sequence (both
prime data area records and the associated
overflow records) and recreate the file in
a new prime data area.

CHOOSING THE RIGHT FILE ORGANIZATION AND
RETRIEVAL METHOD
The flexibility of a disk system lends
itself to several different file
organization and processing methods. It is
important, therefore, to analyze each file
and the program(s) that process i t to
ensUre that the chosen method constitutes
the optimum solution with respect to the
data processing requirements of the
installation.
In ma~y cases, the type of organization
and processing best suited to a file is
immediately evident. However, some
applications may require additional study,
because of their complexity, their unusual
processing requirements, or because of the
wide range of processing programs that use
a file. This is an important aspect of
planning for a data processing system.
Decisions in this area may affect system
configuration requirements and should be
made before programming begins. The
general level of efficiency of the data
processing installation may be affected.

There are no absolute rules for the
resolution of an uncertain situation
regarding the organization and retrieval
alternatives. However, there are several
criteria that may provide an indication of
the optimum solution.

Criteria
The following items form a basis for a
decision concerning the organization of a
file.
File Activity: Activity refers to the
number of records in a file for which there
are transactions. This is usually
expressed as a percentage.
For example. 10% activity in an
inventory file means that, during some
specific period, there are transactions to
be posted to 10% of the records contained
in the inventory file. As the activity
increases, sequential processing becomes
more efficient. Sequential processing
implies either sequential or indexed
sequential organization. Activity implies
batch processing. This means that.
transactions do not need to be posted the
moment they occur. In fact, the time that
may lag between the occurrence and the post
may vary from a few hours to weeks or even
months. depending on the application.
Although the activity of a file is
measured over time, there are applications
where transactions cannot be batched. An
example would be an on-line inventory file
where the transactions would have to be
handled as they occur.
Low activity may justify random instead
of sequential retrieval.
Another important consideration involves
the level of activity when sequential
processing becomes more efficient than
random processing. In order to make this
evaluation, you must know:
• the record length and the blocking
factor
• the average number of additions to the
file
• whether the input is sorted.
If the Indexed sequential File Management
System is being evaluated for random or
sequential processing, you must also know:
• the average time required for access to
the cylinder index.
Section 2: Data Management

133

• which overflow options were chosen.
• if the resident cylinder index facility
has been selected, and if so, what
percentage of the cylinder index can be
main storage resident.
During random processing with ISFMS i t is
advantageous to presort the transactions if
the resident cylinder index facility has
been implemented.
File Volatility: Volatility refers to the
number of additions to and deletions from a
file.
First, consider the effect of making
additions to or deletions from a
sequentially organized file. Two files
must be defined, and the operation must be
handled as i t would be with tape, reading
from the input file and writing to the
output file.
With high volatility (that
is, many additions and deletions), indexed
sequential organization provides a
practical solution. One of the advantages
of an indexed sequential file is that
additions and deletions can be handled
without copying the file. However, as the
number of additions increases, the
efficiency of processing an indexed
sequential file decreases. Additions cause
records to be placed in overflow areas.
Retrieval of·these records in the collating
sequence of the file requires more time
than simply retrieving the records from
contiguous tracks of the original file in
their physical sequence. This is due to
the additional access-arm movement required
to read from the separate overflow area(s)
and additional reads even when cylinder
overflow is used.
With relatively few additions, the
decrease in efficiency is minimal.
However, there is a point at which it
becomes advisable to reorganize the file.
Reorganization means building a new indexed
sequential file from the old one and, in
the process, physically excluding all
records that are tagged for deletion. In
the same operation, all records in the
overflow area are merged into the main
file. At this point, the cycle begins
again. Processing with the reorganized
file is highly efficient. As additions and
deletions occur, this high level of
efficiency gradually diminishes, until we
reach a point where reorganization again
becomes advisable.
Additions to a direct access file do not
necessitate the creation of a new file as
they do for sequential organization.
However, as the DA file extents fill up,
the randomizing algorithm and its
corresponding synonym processor is more
heavily taxed.
134

DOS System Programmer's Guide

Therefore, the following considerations
should be made:
• At what point is it most advantageous
to reorganize an indexed sequential
file?
• What level of volatility excludes
indexed sequential organization as a
practical method of file organization?
• What are the operational considerations
for each method of organization?
Many variables must be considered to answer
these questions. It is impossible to
provide direct answers except in terms of a
specific file and a well defined
application.
File Size: The user must consider the fact
that his on-line capacity is limited.
Three important file organization
considerations are affected by the size of
a file:
1.

An indexed sequential file must be
entirely on-line for any type of
processing.

A sequential file on disk may be
written on any number of packs, that
are then mounted and processed
consecutively. Each disk pack may be
mounted as needed but the manual
intervention that is required is time
consuming.

A direct access file must be entirely
on-line for.any type of processing.

The fact that an indexed sequential or
direct access file must be entirely on-line
whenever i t is to be processed imposes
obvious physical restrictions on maximum
file size. We have talked about the
necessity of periodically reorganizing an
indexed-sequential file. The user must
also consider how this requirement affects
file size. For the purpose of file
reorganization, two files must be defined:
the file to be reorganized and the newly
created file.
If the user's installation
does not include tape units, and if the
user does not want to punch his entire file
into cards, the on-line disk capacity must
be sufficient to accommodate both these
files.
If the installation includes tape
units, the reorganization can be
accomplished in two steps. The first
program sequentially retrieves records from
the indexed sequential file that is to be
reorganized and creates an output file of
these records on tape. The second program
uses this tape file as input and writes the
reorganized (output) file on disk. Thus,
the maximum size of an indexed sequential
file can be doubled by using magnetic tape

as an intermediate storage medium. (An
alternative would be a series of disk
volumes forming a sequential disk file and
sequentially sharing the same drive.)
Figure 2.21 illustrates organization on a
disk-tape system.

FILEl

FILE 2

ORIGINAL
INDEXED SEQUENTIAL
FILE

First consider a DOS system with two
disk drives. In this case, a sequential
file may be contained on any number of
packs. When a processing program reaches
the record with the highest control
information of the disk packs on-line, it
stops. The operator can then mount the
next pack on the same drive and the
operation can be resumed. When a
multivolume file is assigned to a disk
drive, all subsequent volumes (disk packs)
of the file must be mounted on the same
drive.
Most sequential files are altered
periodically by additions and deletions.
In this case, two drives should be used
with multivolume files; one for the input
file volumes and one for the output file
volumes. Each time the end of the
available capacity of a pack on either
drive is reached, the system issues a
message on the console to alert the
operator that it is time to change disk
packs.

SEQUENTIAL
RETRIEVAL
RUN 1

INDEXED SEQUENTIAL
LOAD PROGRAM
RUN 2

REORGANIZED
INDEXED SEQUENTIAL
FILE

Figure 2.21.

sequential organization it can be organized
sequentially.

File Organization on a
Disk/Tape system

It is possible and frequently desirable
to divide a large indexed sequential file
into several smaller files. Although this
approach reduces the flexibility of
processing the records, it provides many
significant advantages for applications
with a low activity rate.
In many cases, if a disk file exceeds
the size limitations imposed by indexed

consider the one-drive system. In this
environment, multivolume sequential files,
that are retrieved only, or retrieved and
updated, can be processed on the single
drive. If records are to be added and
deleted, two files are involved: one input
and one output file. Both files must be
completely contained in one volume and the
area available to each of the two files
must be large enough to accommodate any net
increase in file size brought about by
subsequent additions and deletions. The
single-drive system does not readily
accommodate multivolume sequential files to
which additions and deletions must be made.
This limitation can be circumvented by
dividing a large-file into smaller
segments, each of which can be defined as a
complete file and accommodated within the
available portion of a single pack. This
area is less than half a pack. The user
must treat these segments consistently as
separate files, going through the normal
cycle of job-to-job transition and program
retrieval as each volume is mounted. This
approach, while solving one problem, may
create others. If totals or statistics
must be computed when processing the entire
file, a means must be devised for carrying
results forward from one volume (file
segment) and one program to the next. This
can be accomplished but it introduces an
additional programming complexity with
which the user of the two-drive system need
not be concerned.
section 2: Data Management

135

Response Time: One of the important
advantages of a computer system with a
direct access storage device is the ability
to answer inquiries. Not all applications
require the use of an inquiry capability.
In some data processing installations,
there are no inquiry applications at all.
Where it is required, response time to an
inquiry is a critical consideration. The
less critical the response time, the
greater the choice of organization and
retrieval options.

The example for the sequential file is a
typical procedure for sequential processing
equipment. If this job were run
frequently, the system design
considerations would probably preclude the
use of a sequential file organization.
The considerations previously proposed
establish criteria for choosing the
organization and retrieval method for a
file. In the subsequent paragraphs, these
criteria are applied to a number of sample
files.

The user should consider the following:
• Can the answer to an inquiry wait until
the next batched, sequential updating
of the relevant file? If it can, then
inquiries become an additional
transaction type and are processed
sequentially with all other
transactions against the file. File
organization, in this case, could be
either sequential or indexed
sequential. If the response provided
by this method is not fast enough,
random access is required.
• Can the answer wait until the end of
the present computer run? If so, the
relevant file is mounted at the
completion of the current job; the
inquiry program is loaded; and the file
is processed to produce the required
answers. Obviously, the time delay
involved here varies considerably
depending on the job that is in
progress when the inquiry arrives.
Random Retrieval Consideration: Many files
that could be organized sequentially are
organized as indexed sequential files to
facilitate system design. It is often
possible to reduce the number of peripheral
operations by using random retrieval from
an indexed sequential file. This is true,
for example, of files that have fields
which are used in several jobs.
As an example, assume that invoice
summary cards are to be listed in the
sequence of invoices. Further assume that
the cards do not contain customer names,
but that these names are required in the
listing. Customer names may be obtained
from the customer master file by random
retrieval (if that file is organized as an
indexed sequential file) or by sequential
retrieval. Note that, in the latter case,
the invoice summary cards must first be
sorted into customer number sequence.
Figure 2.22 illustrates the two solutions
and shows the additional steps required if
the customer master file is organized
sequentially.

136

DOS System Programmer's Guide

Sample Files
This part illustrates the choosing of file
organization and retrieval methods for some
typical sample files.
The characteristics of the sample file
are chosen arbitrarily. Different
characteristics could be attributed to
files of similar functions. The examples
are furnished to demonstrate the
application of the criteria we have just
discussed to certain specific file
characteristics; and to show, under these
circumstances, the optimum organization and
retrieval methods.
Example 1
File:

table file

Characteristics: the file is stable and
requires few changes and infrequent
additions and deletions. When alterations
are required, the source card file is
altered and the file is reloaded onto disk.
Normal processing involves retrieval only.
Organization:

sequential

Example 2
File:

payroll file

Characteristics: the file has generally
low volatility; and a relatively low level
of additions and deletions. However, there
is a high activity rate. Processing for
each pay period involves updating of a high
percentage of the year-to-date master
payroll information. Batching of
transactions, (time cards, changes, etc.)
is normal. Fast response to inquiries is
not required.
Organization:

sequential

RUN 1
RANDOM
RETRIEVAL

RUN 1
SORT CARDS
BY
CUSTOMER NO.

RUN 2
SEQUENTIAL
RETRIEVAL

RUN 3
SORT INVOICES
TO
ORIGINAL SEQUENCE

RUN 4
PRINT
REPORT

Figure 2.22.

Indexed Sequential Versus Sequential File Organization

Section 2: Data Management

137

Example 3
File:

wholesale inventory file

Characteristics: the file has moderate
volatility and moderate activity. Normal
transactions may be batched for posting
once or twice a day. Recurring stock
status, activity, and reorder reports are
sequential. Response to inquiries
concerning availability and stock level is
required within one hour.
Organization:

indexed sequential

Example 4
File:

upon many factors. These factors must be
analyzed for each file in anyone
particular application. Often, more than
one organization scheme can be considered
for the same file. In one application,
records could be processed purely at
random; in another, the same records could
be processed in sequence by various control
fields. For example, records within a file
might be processed at random during an
updating run and sequentially within
certain groups such as branch office or due
date when producing reports or billing. A
file such as this would be analyzed to
determine whether it should be organized:
1.

Randomly, thus keeping process time at
a minimum during one run but
destroying the advantage of the
sequential nature of the other.

Sequentially, thus minimizing the time
required to produce reports but
increasing updating time.

Randomly for updating, and then sorted
into sequence for reports.

on-line inventory, parts

Characteristics: the file has a low
volatility but a high activity.
Transactions are processed as they are
received. Responses to inquiries
concerning availability and stock level is
required within 2 minutes. Recurring stock
status, activity, and reorder reports are
sequential but are only produced bimonthly.
organization:

random
The decision would depend on the nature of
the file. other considerations might be:

Example 5
1.

Characteristics: the file has low
volatility and low activity. Transactions
are combined in batches for daily posting.
Billing is cyclic. statements are written
throughout the month by sequentially
retrieving records from the file between
specific limits. Inquiries are processed
twice daily.

Can transactions be batched and sorted
before processing, or must they be
processed as they occur?

Is the activity distributed throughout
the file in such a manner as to
warrant passing the entire file when
updating?

Organization:

Would the processing time saved by
sorting warrant the time and effort
required?

File:

accounts receivable file

indexed sequential

summary
The method of organization best suited to a
particular file of disk records depends

138

DOS System Programmer's Guide

Questions of this kind apply to each file
in an installation. In choosing
organization methods, the over-all
processing objectives of the system must be
kept in mind at all times.

Section 3:

Program Design

Section Outline
Link Editing • • • • • • • •
• • .141
System Flow • • • • • •
.141
Symbolic Units Required • • • •
.143
Linkage Editor Control Statements • • 143
ACTION Statement • •
• • • • .143
PHASE Statement
• • • • 144
INCLUDE statement • • • • • • • • • 145
ENTRY Statement • • • • •
.146
LBLTYP Job Control Statement •
.146
Summary of Considerations for LINK
and CATAL Options • • • • • • • • • • 146
Linkage Editor Program Considerations 147
Program Overlay Structures.
• .147
Overlay Tree Structure • • • • • • • 147
Overlay Tree Design
.147
Overlay Communication
• • • .149
Self-Relocating Programs • • • • • • • • 149
Rules for Writing Self-Relocating
Programs. • • • • • • • • • • •
.149
Advantages of Self-Relocating
Programs • • • • • • • • • • • •
.150
Disadvantages of Self-Relocating
Programs • • • • • • • • • • • •
.150
Programming Techniques • • • • •
.150
Linkage Editor Examples • • • • •
.154
Link Edit-and-Execute-Example • • • • 154
Explanation for Link Edit and
Execute • • • • • • • • • • •
• 154
Catalog to Core Image Library Example 156
Explanation for Catalog to Core
Image Library
• • • • • • • .156
Execute Linkage Editor in Foreground
and Catalog to Private Core Image
Library Example • • • • • • • • • • • 158
Explanation for Catalog to Private
Core Image Library • • • • • • • • • 158
compile and Execute Example • • • • • 160
Explanation for Compile and Execute 160
Catalog for Phase Overlay Example • • 162
Explanation for Catalog for Phase
Overlay • • • • • • • • • • •
.162
Submodular Structure Example • • • • • 163
Explanation for Submodular
structure • • • • • • • • •
• .163

self-Relocating and Multiple Link
Edits Example • • • • • • • • • • • • 165
Explanation for self-Relocating and
Multiple Link Edits • • • • •
.165
Checkpoint/Restart. • • • • • • •
.166
Problem Program Responsibilities
.166
Use of CHKPT Macro. • • • • •
.166
CHKPT Macro
• • • • • •
.167
Notes For DASD and MICR Files
.168
Checkpoint File • • • •
.168
Checkpoints On Tape
• • • • • 168
Checkpoints On Disk
• • • • • 168
Repositioning I/O Files
• • 170
Repositioning Magnetic Tape
.170
DASD Operator Verification Table • • 171
Bypassing Embedded Checkpoint
Records on Tape with Physical IOCS • • 172
Bypassing Checkpoint Records on Tape
with Logical IOCS • • • • • • •
.172
Restarting Checkpointed Programs
.173
RSTRT Statement
.173
IBM 3211 Printer support. • •
.173
System Considerations
.173
Error Recovery Techniques
.174
Macro Writing • • • •
• • • • • • • 175
Macro Instruction • • • • • •
.176
Macro Definition. • • • • • •
.177
Elements of the Macro Definition • • 178
Attributes. • •
.178
Sublist Notation. •
• • • • • • • 179
Variable Symbols. •
• • • • 179
Symbolic Parameter
.179
SET Symbols
• • • • 180
system Variables.
• • • • • • • 181
Concatenation
• • • • • • • 181
Sequence symbols • •
• • • • • • 182
Conditional Assembly Instructions • • 182
AIF--Conditional Branch • • • • • • 182
AGO--Unconditional Branch • • • • • 183
ACTR--Conditional Assembly Loop
• • 183
Counter • • • • • • • • • •
.183
ANOP--Assembly No Operation
Extended Capabilities • • • • • • • • 183
.183
MEXIT -- Macro Definition Exit
MNOTE Statement • • • • • • • • • • 184

Section 3: Program Design

139

Section Figures

Figure 3.1. Linkage Editor System Flow 142
Figure 3.2. Module Dependency • • • • • 148
Figure 3.3. Overlay Tree Structure •• 148
Figure 3.4. Relocating Address
Constants in a Calling List
.151
Figure 3.5. self-Relocating Sample
Program
• • • • • • • • • • • 153
Figure 3.6. Using Checkpoint Facility
on Disk
• • • • • • • • • • • • • .169
Figure 3.7. Procedure for Building
Tape Repositioning Tables • • • • • • • 170
Figure 3.8. Procedure for Building
.172
DASD Operator Verification Table
Figure 3.9. Format of the Checkpoint
Header/Trailer Records
••••
.172
Figure 3.10. 3211 Error Status
Indicator Bits in the CCB
•• 175
Figure 3.11. Operand Field Formats •• 176
Figure 3.12. Keyword Macro
Instruction • • • • • • • • • •
.176
Figure 3.13. positional Macro
Instruction • • • • • • • • • • • • • • 177
Figure 3.14. Macro Instruction with
Prototype • • • • • • • • • • • • • • • 177
Figure 3.15. Mixed Macro Instruction .177
Figure 3.-16. Mixed-Mode Definition • .177
Figure 3.17. Sublist Illustration
•• 179
Figure 3.18. Format of Globals and
.180
Locals
• • • • • • • •
Figure 3.19. Format of SETA
.180
Instruction • • • • • • • •
Figure 3.20. Format of SETB
• • • .180
Instruction • • • • • • • •
Figure 3.21. Format of SETC
Instruction • • • • • • • •
.181
Figure 3.22. Concatenation and
.182
Generated Coding
•••••••
Figure 3.23. Conditional Branch
Instruction • • • • • • • • • • • • • • 182
Figure 3.24. Unconditional Branch
Instruction • • • • • • • • • • • • • • 183
Figure 3.25. Assembly No Operation
Instruction • • • • • • • • •
• .183
Figure 3.26. Macro Definition Exit
.184
Instruction • • • • • • • • • • •
•• 184
Figure 3.27. MNOTE Instruction
Figure 3.28. Sample MNOTE
.184
Figure 3.29. Sample MSG Macro
.185
.186
Figure 3.30. Sample MSG Coding

140

DOS system Programmer's Guide

The following are included in this section
which presents techniques for the effective
use of some DOS facilities:
• Link Editing. Includes the linkage
editor control statements and overlay
structures, design and communication.
In addition, examples illustrating
module dependency and overlay tree
structure are included.
• self-Relocating Program. Explains how
to write a self-relocating program. It
presents the rules, advantages,
disadvantages, and techniques for
writing self-relocating code.
• Checkpoint/Restart. Includes the
programmer's responsibilities when a
checkpoint is taken during problem
program execution. This also explains
the CHKPT macro for checkpointing a
program and describes checkpoint
considerations for tape and disk, as
well as for repositioning of files and
restarting checkpointed programs.

of the control program. On the other hand,
if a program is edited and executed
immediately without cataloging, the linkage
editor is required again the next time the
program is to be executed. Cataloging is a
system design decision based on such
factors as frequency of use and space
available in the core image library.
In a system having a minimum of 32K
positions of main storage, batched-job
foreground multiprogramming, and private
core image library support, the linkage
editor can execute in any partition. A
private core image library must be assigned
when executing the linkage editor in a
foreground partition. When executing the
linkage editor in the background partition,
if a private core image library is not
assigned, the default is to the system core
image library. Without the two options
specified the linkage editor can execute in
the background partition only.

SYSTEM FLOW
• IBM 3211 Printer support. Includes the
error recovery techniques.
• Macro Writing. Includes the macro
language and its effective use. Also
included are examples illustrating how
to write and use macros.

Link Editing
Link editing provides the user with the
capability of combining separately
assembled or compiled program sections or
subprograms. To make this possible, the
output of the language translators must be
processed. At first, the separate program
sections are in relocatable form; that is,
the address constants are identified for
later modification to absolute execution
time values. The linkage editor (LNKEDT)
links and relocates separate program
sections into a single phase that can be
loaded by the control program and then
executed.
Every relocatable program must be
processed by the linkage editor before it
can be executed. Once a program is edited,
it can be executed immediately, cataloged
as a permanent entry in the core image
library, or both cataloged and executed
immediately. When a program is cataloged
in the core image library, the linkage
editor is no longer required for that
program, because it can be loaded directly
from the resident pack by the system loader

Figure 3.1 shows the system flow for the
linkage editor program. Before the linkage
editor program is executed, job control
must perform these functions:
• Process the OPTION statement. OPTION
LINK or CATAL turns on the control
program switches which cause job
control to open the SYSLNK file.
Unless these switches are on, the
linkage editor control statements are
invalid.
• Copy the linkage editor control
statements onto SYSLNK. The linkage
editor control statements are ACTION,
PHASE, INCLUDE, and ENTRY. The ACTION,
PHASE, and ENTRY statements are copied
directly on SYSLNK. There are two
forms of the INCLUDE statement.
INCLUDE statements with no operand are
not copied but cause the data (object
module) on SYSIPT to be written until
the end-of-data (/*) occurs. If the
object module to be linkage edited is
cataloged in the relocatable library,
the INCLUDE statement must have the
name of the module as an operand. The
format of the INCLUDE statement is
copied, but the module is not.
• Write an ENTRY statement with a blank
operand if the job stream does not
already contain one. When the EXEC
LNKEDT statement is encountered, an
ENTRY statement is created to ensure
termination of the link edit input.

Section 3: Program Design

141

SYSRD~
~

II EXEC
1-------'

LNKEDT

Object module
blNCLUDE
and/or

bPHASE
//OPTION
CATAL (or
LINK)

/ / JOB

System loader brings in job control.

Job control reads and interprets job control statements.

l
l

SYSRES
Core Image Library

Job control

Linkage editor

Program phase
Relocatable Library

Job control causes linkage editor (LE) control statements and
relocatable object modules to be placed on SYSLNKi object
modules may be either on SYSIPT or in the relocatable library.

Object Module

When job control reads EXEC LNKEDT, it causes the system
loader to bring in the I inkage editor program.

Using SYSLNK (input) and SYSOOl (work file) the linkage
editor develops executable program •

•

If OPTION CATAL is specified, linkage editor output is placed
permanently in the core image library (the core image directory
is updated). If OPTION LINK is specified, linkage editor output is placed temporari Iy in the core image I ibrary (the core image
directory is not updated), and the next linkage editor output overlays the last temporarily placed linkage editor output.
Note: If a private core image library is assigned, output is placed in
the private I ibrary rather than in the system core image library
on SYSRES. Output can be placed in the system core image
library only when executing the linkage editor in the background partition.

Linkage editor
statements
Object modules

SYSOOl

If a program just link-edited is to be executed, the EXEC statement
causes the system loader to fetch it from core image I ibrary and
begin execution.

Figure 3.1.

142

SYSLNK

Linkage Editor System Flow

DOS System Programmer's Guide

Work file

• Inform the system loader to load the
linkage editor program. The linkage
editor program uses the data on SYSLNK
as input. "It handles the relocatable
modules as directed by the PHASE and
INCLUDE statements to develop
executable program phases. Some of the
linkage editor program functions are:
1.

SYMBOLIC UNITS REQUIRED
The symbolic units required by the linkage
editor are basically a subset of those
needed by the language translators:
SYSIPT

Module input

Extracting the modules named in
INCLUDE statements from the
relocatable library. If, in
extracting a module, another
INCLUDE statement occurs, this
module is also retrieved. The
nesting of modules is possible up
to a depth of five (a level of
six).

SYSLST

Programmer messages and
listings

SYSLOG

Operator messages

SYSRDR

Control statement input (via
job control)

SYSLNK

Input to the linkage editor

Constructing composite
dictionaries for ESD and RLD data,
to resolve all linkages between
different control sections.

SYSOOl

Work file

Relocating each control section as
necessary within a phase.

Assigning the entire phase to a
contiguous area in main storage.

Modifying all relocatable address
constants to contain the relocated
value of their symbols.

Searching the relocatable library
for a cataloged object module with
the same name as each unresolved
external reference. The automatic
library lookup feature (AUTOLINK)
is particularly useful for
retrieving IOCS modules. It may
be suppressed.

Note that SYSRDR and SYSIPT may contain
input for the linkage editor. This input
is written on SYSLNK by job control.
If output from the linkage editor is to
be placed in a private core image library,
the following symbolic unit is required:
SYSCLB

Building the core image directory
phase headers and cataloging to
the core image library, if CATAL
is specified.

If a phase by the same name was
cataloged previously, the old phase is
deleted and the new one is cataloged.
Deletion removes the item from the
directory, but it does not release the
space in the library until a condense
function occurs.
If a private core image library is
assigned, the linkage editor output is
placed in the private rather than the
system core image library, permanently if
OPTION CATAL is specified or temporarily if
OPTION LINK is specified (see Figure 3.1,
pOint 6). In this case, the library need
not .be on SYSRES as shown in Figure 3.1.
Output may be placed in the system core
image library (either permanently or
temporarily) only when executing the
linkage editor in the background partition.

the private core image library
may be assigned anywhere in
the job stream but must be
before the // EXEC LNKEDT
statement.

LINKAGE EDITOR CONTROL STATEMENTS
Position 1 must be blank on linkage editor
control statements. Otherwise, they follow
the same format as job control commands.
Refer to DOS system Control and System
Service listed in the Preface for a
detailed explanation of the control
statements.

ACTION Statement

ACTION

CLEAR
MAP
NOMAP
NOAUTO
CANCEL
BG
Fl
F2

This statement specifies linkage editor
options.
It is not required, but if used,
i t must appear as the first linkage editor
statement in the input stream. If multiple
operands are required, they may be placed
on separate ACTION statements or on one
section 3: Program DeSign

143

ACTION statement separated by·commas. The
ACTION statement is effective only for the
next linkage editor execution. The
parameters have the following meanings:
CLEAR

set the unused area of the core
image library to binary zeros.
The linkage editor clears from the
next available entry (taken from
the core image directory) to the
end of the core image library.
Because this is time-consuming,
use i t only if the areas defined
by DS statements must be filled
with zeros.

operands are used in a non-MPS
environment, they are ignored.
If none of these operands is
present, the program will be linkedited to execute in the partition
in which the link edit function is
taking place, unless otherwise
specified on the PHASE statement.
An example of the use of the
ACTION F1 statement follows.
Assume a 64K machine with:
8K supervisor
24K background area

MAP

Write main storage map and error
diagnostics on SYSLST. Whenever
SYSLST is assigned, MAP is
automatic unless NOMAP is
specified.

NOMAP

Nullify MAP action.

NOAUTO

Suppress AUTOLINK function for the
entire program, not just one
phase.

CANCEL

Cancel the job if the content of
the linkage editor input is in
error. See messages 21001 to
21701 in the DOS Messages listed
in the Preface.

F1
F2

Causes the end of supervisor
address used in linkage editor
calculations to be set to the
beginning of the partition
specified, plus the length of the
label area and of the save area.
The end of supervisor address in
the communications region is not
changed.
The BG, Fl and F2 operands
provide the capability of link
editing a program to execute in a
partition other than that in which
the link edit function is taking
place. Programs that have a phase
origin of S (or * for the first
phase of a program) can be
origined to the specified
partition by use of the operands.
Use of the ACTION BG statement
is possible only in a system
supporting the batched-job
foreground and private core image
library options when the linkage
editor is executing in a
foreground partition. .
Use of the ACTION F1 (or F2)
statement in a mUltiprogramming
environment requires that the
partition be allocated. If these

144

DOS System Programmer's Guide

16K foreground 2 area
16K foreground 1 area.
When executing the linkage
editor in the background the
statement PHASE PHASE1,S causes
PHASE1 to be origined at 8K (the
end of the supervisor area). The
sequence
ACTION F1
PHASE PHASE1,S
causes PHASE1 to be origined at
48K (the beginning of the
foreground 1 area) plus the length
of the foreground save area.
When executing the linkage
editor in foreground 2, the
statement PHASE PHASE1,S causes
PHASE1 to be origined at 32K (the
beginning of the foreground 2
area) plus the length of the
foreground save area. The
sequence
ACTION BG
PHASE PHASE1,S
causes PHASE1 to be origined at
the end of the supervisor area.
PHASE Statement
PHASE name,origin[,NOAUTOl
A program phase is the section of a program
that is loaded by the system loader as a
single overlay with a single FETCH or LOAD.
The input for building a single phase
consists of the text from one or more
complete control sections. Therefore,
programs may consist of many phases, or a
phase may consist of many subprograms or
control sections.

The PHASE statement provides the linkage
editor with the phase name and an origin
point for the phase. The phase name
catalogs the phase in the core image
library and retrieves it for execution.
Job control uses the phase name to
construct a single track phase directory
before each job step is executed. This
phase directory is built for all background
program executions. If executing in the
foreground, this phase directory is built
under link, edit, and go conditions. The
entries to this directory are taken from
the core image directory for any phase
where the first four characters of the name
are identical to those in the name
specified in the EXEC statement. The
directory entry contains such information
as loading address, entry point, and
starting disk address in the core image
library. The separate phase directory
permits faster retrieval of the phases.

partition is allocated,
ACTION F1 or F2 has the same
effect as F+address.

symbolic name of the phase,
consisting of 1 to 8 alphameric
characters. The first 4
characters of a multi phase program
should be the same to achieve
maximum retrieval efficiency.

origin

Specification of the load address
of the phase. The load address
can be one of six forms:

*[~relocation].
This is the
most frequently used format
and specifies an origin point
for a phase at the next
available· core location.

symbol [(phase)]
[±relocationJ. This format
specifies an origin point for
a phase at the same point as
previously defined symbol
(for overlays).

A detailed explanation of the
origin parameter is given in the
DOS System Control and system
Service listed in the Preface.
Also refer to Link Editing
Examples.

The entries in the operand field
represent the following:
name

NOAUTO

Suppress the AUTOLINK function for
this phase only.

INCLUDE Statement

S [+relocationJ. If link
editing in the background,
the origin point is at the
end of the supervisor, the
save areas, and the area
assigned to the COMMON pool,
if any. If link editing in
the foreground, the beginning
address of the partition is
substituted for the end of
supervisor address.
ROOT. This phase is
designated as the root phase,
which remains in main storage
throughout execution. Its
location is the same as with
format s.

+displacement. The origin
point is set at a specified
location; +0 must be used for
any self-relocating program.

F+address. This format is
used to begin the program at
the start of a foreground
partition when link editing
in the background and the
foreground partition is not
allocated. If the foreground

INCLUDE [modulenamel [,enamelist)]
This statement specifies that an object
module is to be included for editing by the
linkage editor. The system assumes the
location of the module as follows:
1.

Both operands missing. The object
module is on SYSIPT; it is copied onto
SYSLNK.

Modulename given. The object module
is cataloged in the relocatable
library under the same name.

Second operand only given. The object
module is in the input stream on
SYSLNK. The parameters represent the
following:

modulename Use this parameter only if the
module is cataloged in the
relocatable library. It
consists of 1 to 8 alphameric
characters and must be the same
as the name used when the module
was cataloged.
(namelist) This parameter provides the
ability to select particular
control sections from a given
module. It is expressed as
(csname1, csname2, ••• ).
Section 3: Program Design

145

labeled tapes are processed. This same
area is used by nonsequential DASD
files with fewer extents, and by
labeled tape files.

ENTRY statement
ENTRY [entrypoint1
The ENTRY statement signals the end of
program input to the linkage editor. The
entrypoint operand indicates the transfer
address for the first phase as follows:

If omitted, the first significant
address in an END record encountered
during the generation of the first
phase is used; or, if no such operand
is found, the load address of the
first phase is used.
If given, it must be the name of a
CSECT or a label definition defined in
the first phase.

The ENTRY statement can be completely
omitted because job control automatically
writes an ENTRY statement with a blank
operand when it encounters the EXEC LNKEDT
statement.

The LBLTYP statement is not required if
only unlabeled tape files and/or sequential
DASD files are being processed. Only one
LBLTYP statement is submitted. The
placement of the statement in the job
stream varies as shown:
• Non-self-relocating
(background/foreground). Immediately
preceding the EXEC LNKEDT statement at
linkage editor time.
• self-relocating
(background/foreground).
EXEC for the program.

Ahead of the

Examples of various linkage editor
functions follow.

SUMMARY OF CONSIDERATIONS FOR LINK AND
CATAL OPTIONS
LBLTYP Job Control Statement
/ / LBLTYP

SYSLNK must be assigned, or LINK and
CATAL options are ignored (switches
are not set). If executing the
linkage editor in a foreground
partition, a private core image
library (SYSCLB) must be assigned.
This is possible only in a system
supporting the batched-job foreground
and private core image library
options.

Unless the switches are set by the
LINK or CATAL option, the linkage
editor control statements are ignored.

The CATAL option sets the LINK and
CATAL switches.

When the LINK switches are set, the
output of the language tr;anslators is
placed on SYSLNK.

LINK and CATAL switches are turned off
by:

{TAPE (nn)}
NSD (nn)

The label storage records for standard
labeled tape files and nonsequential DASD
files (direct access, indexed sequential,
or DTFPH with all packs mounted) are
brought into the label save area of the
partition containing the processing
program. Therefore, main storage must be
reserved by the user whenever such files
are processed. Because this area is used
during OPEN for one file at a time, the
total area needed is that required by the
largest file.
Main storage reservation is accomplished
by the LBLTYP statement. The amount of
main storage reserved is governed by the
operand TAPE or NSD as follows:
• TAPE reserves 80 bytes of main storage.
This format is used when standard
labeled tape files and n2 nonsequential
DASD files are processed. nn is
ignored by job control. This same
80-byte area is used by all labeled
tape files.
• NSD reserves 84 bytes plus 20 bytes per
extent. The number of extents is
specified in the nn parameter for the
nonsequential file that has the largest
number of extents. This format is used
when nonsequential DASD files are
processed, regardless of whether

146

DOS System Programmer's Guide

/& or JOB statement

An error during compilation.

Completion of cataloging (update of
transient, library routine, and
foreground program directories, and
system status report) occurs when the
/& statement is read by job control.

If a successful link edit has not
occurred, cataloging does not take
place.

If multiple linkage editor job steps
are set up as one job, keep these
points in mind:
a.

It is not possible to CATAL into
the core image library with
// OPTION CATAL and then have
another linkage editor job step
with // OPTION LINK in the same
job. Operator message lSlnD
(STATEMENT OUT OF SEQUENCE)
results.
If a compilation is being
performed, the link switches may
be turned off by an error. When
cataloging to the core image
library, therefore, it is
advisable to handle multiple job
steps as separate jobs (each with
/i) to be sure that the cataloging
operation is finished on the /&.
SYSLNK extents are reset each time
before Job Control writes a new
series of link edit control
statements onto SYSLNK. Core
image library directory and
subdirectory are updated at /& if
the CATAL option is included.

Overlay Tree Structure
To place a program in an overlay structure,
the programmer should be familiar with the
following terms:
Phase: A phase is the smallest functional
unit (one or more control sections) that
can be loaded as one logical entity during
the execution of the program. A phase can
contain up to 524,288 bytes of text. The
root phase (first phase) remains in main
storage throughout execution.
Tree: A tree is the graphic representation
that shows how phases can use main storage
at different times. It does not imply the
order of execution, although the root phase
is the first to receive control.
The design of an overlay program
requires the organization of the control
sections of the program in an overlay tree
structure. The tree structure is developed
considering:
1.

The amount of available main storage.

The frequency of use of each control
section.

LINKAGE EDITOR PROGRAM CONSIDERATIONS

The dependencies between control
sections.

The linkage editor program consists of
eight phases. The efficiency of link
editing programs into the core image
library depends upon the amount of main
storage allocated to the partition where
the linkage editor is executing. If a
minimum of 14K of main storage is available
several phases of the linkage editor
program are contained within main storage
during linkage editor processing. If less
than 14K is available, only one linkage
editor processing phase may be resident in
the partition at one time. Some time is
lost fetching the additional phases as they
are needed.

The manner in which control should
pass within a path, from one path to
another, and from one region to
another.

PROGRAM OVERLAY STRUCTURES
Overlay is a programming technique that
minimizes the main storage requirements of
a program. To use overlay, the programmer
should be familiar with two related
techniques:
1.

Organizing the program as an overlay
structure.

Communicating with the control program
during execution through FETCH and
LOAD macros.

When the overlay tree structure for a
program is determined, PHASE statements are
prepared which segment the program in that
manner.

Overlay Tree Design
To begin constructing an overlay tree, the
programmer should select those modules that
receive control at the beginning of
execution plus those that should always
remain in main storage; these form the root
phase. The rest of the tree can be
developed by determining the dependency of
the remaining phases and how they can use
the same main storage locations at
different times during execution.
Module dependency is determined by the
requirements of a control section or module
for a given routine in another control
section. A module depends upon a control
section to which it branches or whose data
i t must process. The required control
section must be in main storage before
Section 3: Program Design

147

execution can continue beyond a given point
in the program. Figure 3.2 illustrates how
modules depend on each other, and the paths
that result from these dependencies.
The module containing control sections A
and B can be used to form the root phase.
The module containing control sections C
and 0 can use the same main storage as the
module containing control sections J and K.
Phases that use the same main storage area
can overlay each other during execution.
The module containing control section E can
use the same main storage as the module
containing control sections F and G. The
module containing control section H can use
the same main storage as the module
containing control section I. The module
containing control section L can use the
same main storage as the module containing
control sections M and N.

determined by 0, plus the length of all
phases in the path. For example, the
origin of phases 3 and 4 is equal to 0 +
5,000 (the length of phase 2) + 6,000 (the
length of the root phase).

Root
Phase 1
(6000)

Phase 2
:' A
(5000)!S··

..Ic

,.,.,.,.....,.,.................,.,.,.,.,.,.,.,.,.,...,.".,.....,.......

P~;~;O~
10
.' ,.,.,.,.".,.,.,.,............... "
;,
.: E

P~~~~O~
,·e
IIGF

Phase 7
(6000)

..... '" ................................,.,.,.,.,.,.,.,.,.,.,.,..".",..".".,.,.,.,.,.,.".".,•..,.,.,.,

P~~~~O~
li"K' P~;~~O~
"""""'.,.,."'., ,.,.,.,.,.,. .,.,.,. . ,\.,"""""""',. .,.,.,.,.,.,.,.,. . .,.,.,.•

Phase 5
Ph
6 . . :
(7000) ,,, ..,.. t,,..,, ..mw.'
L
•. I , . . . . .
:

(;~~O)i

Path 1

Path 2

Path 3

Path 4

Path 5

liM

[t. . . . ..

Longest Path (Phases 1, 2, 4, 5) =21,000 bytes

A
B

Figure 3.3.

Overlay Tree Structure

J
K

M
N

Note:

A through N are Control Sections.

Figure 3.2.

Module Dependency

Figure 3.3 shows the resulting overlay
tree structure. The longest path in this
structure is formed by phases 1, 2, 4, and
5, because the program requires 21,000
bytes when they are in main storage. Thus,
the minimum main storage requirement for
the program is 21,000 bytes. The program
would require 46,000 bytes ·if not put in an
overlay structure. The linkage editor
assigns the relocatable origin of the root
phase (the origin of the program) at O.
The relative origin of each phase is
148

DOS System Programmer's Guide

When a phase is in main storage, all
phases in its path are in main storage.
(Each time a phase is loaded, all phases in
its path are also loaded if they are not
already in main storage.) In Figure 3.3,
when phase 4 is in main storage, phases 1
and 2 are also in main storage. This does
not imply that phase 5 or 6 is in main
storage because neither phase is in the
path of segment 4.
The position of the phases in an overlay
tree structure does not necessarily imply
the order in which they are executed. A
phase can be loaded and overlaid as many
times as required by the logic of the
program. If a phase is modified during
execution, that modification remains only
until the phase is overlaid. However, a
phase cannot be overlaid by itself.
Phases that can be in main storage
simultaneously are considered to be
inclusive. Phases not in the same path are
considered to be exclusive: they cannot be
in main storage simultaneously.
Phases on which two or more exclusive
phases are dependent are called common
phases. A phase common to two other phases
is part of each. In Figure 3.3, phase 1 is
common to phases 8 and 9, but not to phase
2.

overlay Communication
The programmer must know how his program
can communicate with the control program
during execution. Two ways in which he can
overlay phases in his program are:
1.

By a LOAD macro instruction, that
loads the named phase into main
storage and returns control to the
calling phase.

By a FETCH macro instruction, that
loads the specified phase into main
storage and passes control to the
entry address of the fetched phase.

Some of the advantages are:
1.

The LOAD macro allows the programmer
to load his phases in violation of the
tree structure that was defined at
linkage editor time.

The LOAD macro allows the programmer
to load phases from other programs.

However, these responsibilities are
associated with the use of these macros:
1.

The programmer must keep track of
which phases are in main storage.

The programmer must be aware of
overlay structure and he must know tne
phase names at compilation or assembly
time.

This causes the linkage editor to assume
that the program is loaded at core location
zero, and to compute all absolute addresses
from the beginning of the phase. The job
control EXEC function recognizes a zero
phase address and adjusts the origin
address to compensate for the current
partition boundary save area and label area
(if any). It then gives control to the
updated entry address of the phase.

RULES FOR WRITING SELF-RELOCATING PROGRAMS
In general, if a problem program is written
to be self-relocating, the following rules
must be adhered to:
1.

The supervisor must support
multiprogramming (that is, MPS=YES or
BJF must be specified as a parameter
in the SUPVR macro at system
generation time).

The PHASE card must specify an origin
of +0.

The program must relocate all address
constants used in the program.
Whenever possible, use the LA
instruction to load an address in a
register instead of using an A-type
address constant. For example,
Instead of using:
USING
BALR

phase that is loaded into main storage
by the FETCH macro is always relocated
unless the self-relocating option was
specified at linkage editor time. A phase
that is loaded into main storage by the
LOAD macro is relocated relative to the
same structure as when it was linkage
edited; i.e., LOADs can be made outside the
defined tree, but it is ·not possible for
the system to ensure that references
outside the tree will be valid.
A

Self-Relocating Programs
A system supporting multiprogramming has
the capability of executing self-relocating
programs. A self-relocating program is one
that can be executed at any location in
main storage. Writing a self-relocating
program is an efficient coding technique
because self-relocating programs are
link edited only once for execution in
any partition. When linkage editing, use
OPTION CATAL and a PHASE card such as:
PHASE

Phasename,+O

BCTR
BCTR
LA

•

L
EOF

•
•
EOJ

AEOF

•
DC

*,12
12,0
12,0(12)
12,0
12,0
1,EOF
1,AEOF
10,AEOF

•

A (EOF)

Use:
USING
BALR
LA

BCTR
BCTR

•
•
LA
•
EOF

*,12
12,0
12,0(12)
12,0
12,0
10,EOF

•

EOJ
Section 3: Program Design

149

If logical IOCS is used, the program
must use the OPENR and CLOSER macros
to open and close files.

If physical IOCS is used, the program
must relocate all CCW address fields.

Register notation must be used when
issuing an imperative macro (I/O, I/O
control, and supervisor
communication). Register notation
utilizes less main storage and permits
faster execution.

The following rules apply to multimodule
programs.
7.

The relocation factor should be
calculated and stored in a register
for future use. For register economy,
the base register can hold the
relocation factor.

The calling program is responsible for
relocating all address constants in
the calling list(s). See Figure 3.4
for an example of calling program
relocating the address constants in a
calling list.

ADVANTAGES OF SELF-RELOCATING PROGRAMS
self-relocating programs have the ability
to run in anyone of the three problem
program partitions without needing linkage
editing again. The program can also be
loaded anywhere within a partition. The
restriction of specific partition
allocations need not be adhered to with a
self-relocating program because it
relocates itself.

For example:
USING
BALR
LA

BCTR
BCTR

*,12
12,0
12,0(12)
12,0
12,0

DISADVANTAGES OF SELF-RELOCATING PROGRAMS
self-relocating programs are slightly more
time consuming to write and they usually
require slightly more main storage.

When branching to an external address,
use one of the following techniques:
PROGRAMMING TECHNIQUES
a.

L
BAL

15,=V(EXTERNAL)
14,0(12,15)

15,=V(EXTERNAL)
15,12
14,15

b.
AR

BALR

where register 12 is the base register
containing the relocation factor.

150

DOS System Programmer's Guide

A self-relocating program is capable of
proper execution, regardless of where it is
loaded. DTFDI should be used to resolve
device differences between partitions. A
self-relocating program must also adjust
all of its own absolute addresses to point
to the proper address. This must be done
after the program is loaded, and before the
absolute addresses are used.

r------------------------------------------------------------------------------,
// JOB A
// OPTION LINK
/ / EXEC ASSEMBLY
CSECTl
START 0
USING *,12
BALR 12,0
LA
12,0(12)
BCTR 12,0
BCTR 12,0

•
•

LA
LA
LA
LA

STM
LA
L

CALL
EQU
EOJ
SAVEAREA DC

LIST

Use load point value as the base to
find the load point value.

1,A
2,B
3,C
Modify the CALL address constant list.
4,D
1,4,LIST
13, SAVEAREA
15,=V(EXTERNAL)
15,12
Adjust CALL address by relocation
factor.
(15),(A,B,C,D)
*-16
For address constants (4 bytes each).
9D'0'

END

/* *

/ / EXEC. ASSEMBLY
CSECT2
START 0
ENTRY EXTERNAL
EXTERNAL SAVE (14,12)
USING *,12
BALR 12,0

Establish new base

•

RETURN(14,12)
END

/* *
_
IL/_____________________________________________________________________________
/ EXEC LNKEDT
Figure 3.4.

Relocating Address Constants in a Calling List

Within these self-relocating programs,
some macros generate self-relocating code.
For example, the MPS utility macros are
self-relocating (that is, they modify all
of their own address constants to their
proper values before using them). OPENR
and CLOSER macros are designed to be used
in self-relocating programs. OPENR and
CLOSER can be used in place of OPEN and
CLOSE, and adjust all of the address
constants in the DTFs opened and closed.
OPENR and CLOSER can be used in any program
because the OPENR macro computes the amount
of relocation. If relocation is 0, the
standard open is executed. In addition,
all of the module generation (xxMOD) macros
are self-relocating.
The addresses of all address constants
containing relocatable values are listed in
the relocation dictionary in the assembly
listing. This dictionary includes both
those address constants that are modified
by self-relocating macros, and those that

are not. The address constants not
mOdified by self~relocating macros must be
modified by some other technique. After
the program has been linkage edited with a
phase origin of +0, the contents of each
address constant is the displacement from
the beginning of the phase to the address
pointed to by that address constant.
The following techniques place relocated
absolute addresses in address constants.
These techniques are required only when the
LA instruction 'cannot be used.
Technique 1
Named A-type address constants:

•

ADCON

•

4,ADCONAME
4,ADCON
A (ADCONAME)
Section 3: Program Design

151

Technique 2

If CCW list is static during program
execution:

A-type address constants in the literal
pool:

•

LA
LA
ST

•
•

•

ST
MVI

4, IOAREA
4,TAPECCW
TAPECCW,l

CCW

1,IOAREA,X'20',100

3,=A(ADCONAME)
4, ADCONAME
4,0(3)
TAPECCW

LTORG
=A(ADCONAME)

•
•

•

CLI00

Technique 3
A-type address constants with a specified
length of three bytes, and a nonzero value
in the adjacent left byte (as in CCWs):

Technique 4
Named V-type or A-type address constants:

If CCW list dynamically changes during
program execution:

•

TAPECCW
IOAREA
2.

IC
LA
ST
STC

3,TAPECCW
4,IOAREA
4,TAPECCW
3,TAPECCW

ccw

1,IOAREA,X'20',100

•

,CLI00

•

USING *,12
BALR 12,0
12,0(12)
LA
BCTR 12,0
BCTR 12,0 Reg. 12 contains relocation factor

•

L
ALR

•
•

TAPECCW

CCW

IOAREA

152

•

11,TAPECCW
11,12
11,TAPECCW
1,IOAREA,X'20',100
CLI00

DOS System Programmer's Guide

•
•

L
AR
ST

•

ADCONAST DC
ADCON
DC

3,ADCONAST
3,ADCONAST

Determine
Relocation
factor

4,ADCON
4,3 Add Relocation factor
4,ADCON
A(*)
V(NAME)

Note that the load point of the phase is
not synonymous with the relocation factor
as developed in register 3 (technique 4).
If the load point of the phase is taken
from register 0 (or calculated by a BALR
and subtracting 2) immediately after the
phase is loaded, it may be added to address
constants with varying results. If the
phase was linkage edited with an origin of
+0, the correct results are obtained. If
the phase was linkage edited with an origin
of * or S, incorrect results are obtained
because, both the linkage editor and the
program itself have added the load point to
all address constants. See Figure 3.5 for
an example of a self-relocating program.

SOURCE STATEMENTS
REPRO
+0 ORIGIN IMPLIES SELF-RELOCATION
PHASE EXAMPLE,+O
PRINT NOGEN
PROGRAM
START 0
BALR 12,0
US I NG ~~, 12
ROUTINE TO RELOCATE ADDRESS CONSTANTS
LA
1,PRINTCCW
RELOCATE CCW ADDRESS
ST
1,PRINTCCB+8
IN CCB FOR PRINTER
LA
1,TAPECCW
RELOCATE CCW ADDRESS
ST
1,TAPECCB+8
IN CCB FOR INPUT TAPE
IC
2,PRINTCCW
SAVE PRINT CCW OP CODE
LA
1,OUTAREA
RELOCATE OUTPUT AREA ADDRESS
ST
1,PRINTCCW
IN PRINTER CCW
RESTORE PRINT CCW OP CODE
STC 2,PRINTCCW
RELOCATE INPUT AREA ADDRESS
LA
1,.INAREA
IN TAPE CCW
ST
1,TAPECCW
MVI TAPECCW,READ
SET TAPE CCW OP CODE TO READ
~
MAIN ROUTINE ... READ TAPE AND PRINT RECORDS
GET CCB ADDRESS
READTAPE
LA
1,TAPECCB
EXCP (1)
READ ONE RECORD FROM TAPE
WAIT (1)
WAIT FOR I/O COMPLETION
GET ADDRESS OF TAPE EOF ROUTINE
LA
10,EOFTAPE
BAL 14,CHECK
GO TO UNIT EXCEPTION SUBROUTINE
MVC OUTAREA(10),INAREA
EDIT RECORD
MVC OUTAREA+15(70),INAREA+10
IN
MVC OUTAREA+90(20),INAREA+80
OUTPUT AREA
GET CCB ADDRESS
LA
1,PRINTCCB
EXCP (1)
PRINT EDITED RECORD
WAIT (1)
WAIT FOR I/O COMPLETION
LA
10, CHA12
GET ADDRESS OF CHAN 12 ROUTINE
BAL 14,CHECK
GO TO UNIT EXCEPTION SUBROUTINE
B
READTAPE
TM
4(1),1
CHECK
CHECK FOR UNIT EXEC. IN CCB
BCR 1,10
YES-GO TO PROPER ROUTINE
BR
14
NO-RETURN TO MAINLINE
CHA12
MVI PRINTCCW,SKIPT01
SET SEEK TO CHAN 1 OP CODE
EXCP (1)
SEEK TO CHAN 1 IMMEDIATELY
WAIT (1)
WAIT FOR I/O COMPLETION
MVI PRINTCCW,PRINT
SET PRINTER OP CODE TO WRITE
BR
14
RETURN TO MAINLINE
EOFTAPE
EOJ
END OF JOB
CNOP 0,4
ALIGN CCB'S TO FULL WORD
PRINTCCB
CCB SYS004,PRINTCCW,X'0400'
TAPECCB
CCB SYS001,TAPECCW
PRINTCCW
TAPECCW
OUTAREA
INAREA
SLI
READ
PRINT
SKIPT01

Figure 3.5.

CCW
CCW
DC
DC
EQU
EQU
EQU
EQU
END

PRINT,OUTAREA,SLI,L'OUTAREA
READ,INAREA,SLI,L'INAREA
CL110"
CL100"
X'20'
2
9
X'8B'
PROGRAM

Self-Relocating Sample Program

)
Section 3: Program Design

153

Linkage Editor Examples
LINK EDIT-AND-EXECUTE-EXAMPLE
/ / JOB LINKEXEC

1.
2.
3.

LINK EDIT AND EXECUTE IN BACKGROUND. SINGLE PHASE, SINGLE OBJECT MODULE
RELOCATABLE MODULE NOT CATALOGED, BACKGROUND PROGRAM
NONSEQUENTIAL DASD & LABELED TAPE FILES TO BE PROCESSED
ASSGN SYSLNK,X'190'
/ / OPTION LINK
PHASE PROGA, *
INCLUDE

*
*
*
//

4.
5.
6.

Relocatable object deck
/*
/ / LBLTYP NSD(2)
/ / EXEC LNKEDT

Any job statements required for execution such as ASSGN or label statements.

7.
8.

/ / EXEC

Data input as required.

/*
/&

*
*
*
*
*

1.
2.
3.

TO CATALOG AND EXECUTE, CHANGE STATEMENT 2 TO // OPTION CATAL.
TO CATALOG ONLY, CHANGE STATEMENT 2 TO / / OPTION CATAL AND
REMOVE ALL STATEMENTS FOLLOWING LNKEDT EXCEPT /&
TO USE MODULE FROM RELOCATABLE LIBRARY, CHANGE STATEMENT 3
TO INCLUDE MODULES AND REMOVE ALL STATEMENTS UP TO / / LBLTYP.

Explanation for Link Edit and Execute
This example illustrates the basic concept
of linkage editing and executing by using a
single phase that is constructed from a
single relocatable object deck contained in
punched cards. The program is executed in
the background partition. Labeled tape and
nonsequential DASD files are to be
processed when the phase is executed. No
more than two extents are used by any DASD
file.
Statement 1: No assignments are necessary,
because the system units required for
linkage editing are in the assumed
configuration. However, an ASSGN for
SYSLNK is included to illustrate its
position relative to the OPTION statement
in case assignment is required.
statement 2: The OPTION LINK statement
sets switches to indicate that a linkage
editor operation is to be performed. If
SYSLNK has not been assigned, the statement
is ignored. Linkage editor control
statements are not accepted unless the
154

DOS System Programmer's Guide

OPTION statement is processed. Because
option is LINK, not CATAL, only link
editing is performed; cataloging to the
core image library does not occur.
Statement 3.: The PHASE statement is copied
on SYSLNK, because position 1 is blank and
the LINK switch is on. The operands are
not examined until SYSLNK becomes input to
the linkage editor program.
When the PHASE statement is processed by
the linkage editor, only one phase is
constructed, because only one PHASE
statement is submitted for the entire
LNKEDT. The name of this phase is PROGA,
as specified in the first operand. The
second operand indicates the origin point
for the phase. Because an
has been used,
the phase begins in the next main storage
location available, with forced doubleword
alignment. Because this is the first and
only phase, it is located at the end of the
supervisor plus length of the label save
area (reserved by LBLTYP) plus length of
any area assigned to the COMMON pool (as
designated by a CM entry in the relocatable
module).

A relocation factor, either plus or
minus, is used with the *, such as *+1024.
This causes the origin point of the phase
to be set relative to the * by the amount
of the relocation term. This term can be
expressed as:
X'hhhhhh'
dddddddd
nK

1 to 6 hexadecimal digits
1 to e decimal digits
where K = 1024

*+1024 uses the second format and adds 1024
bytes to the origin location. +lK or
+X'400' gives the same result as +1024.
Statement 4: The INCLUDE statement has no
operands, so the system reads the records
from SYSIPT and writes them on SYSLNK until
SYSIPT has an end-of-data (/*) record. The
data on SYSIPT is expected to be the object
module in card image format that is used in
this linkage editor operation. If the
output of the language translator (SYSPCH)
is placed on 2311/2314/2319 instead of
cards, it cannot be used directly as SYSIPT
in a linkage editor operation because the
records contain a stacker select code in
position 1. SYSPCH must be converted to an
SO-position card image record.
Statement 5: The LBLTYP statement causes a
computation of the number of bytes that are
required for label storage data in the
program to be linkage edited. In this
example, 124 bytes are reserved (S4 +
[2x20]). The calculation is saved by job
control and passed on first to the linkage
editor and later to LIOCS.
Statement 6: The EXEC LNKEDT writes an
ENTRY statement with no operand on SYSLNK
and causes the system loader to bring in
the linkage editor program.
Using the data just placed on SYSLNK as
input, the linkage editor develops
executable code. The output is placed in
the next available space of the core image
library (immediately after the last
cataloged phase). This is true regardless
of whether the program is cataloged or not.
Cataloging causes the updating of the
directory to reflect a new ending point for
the library.
If cataloging does not occur,
the next program that is linkage edited
overlays it. For this reason, a linkage
edited program that is not cataloged is
said to be placed in the temporary area of
the core image library. Also, a program
that is linkage edited without cataloging

must be linkage edited whenever i t is used.
No ACTION options are specified.
Therefore, in resolving the external
references, the system makes use of the
AUTOLINK feature. Error diagnostics and a
main storage map are written on SYSLST,
because SYSLST is assigned.
Statement 7: Because the program is not
cataloged, it must be executed immediately.
Any pertinent job control statements are
entered at this point.
Statement S: An EXEC statement with no
operand indicates that the phase to be
executed was just linkage edited.
Therefore, no search of the core image
directory is required, and the system
loader brings the program into main storage
from the temporary area and transfers
control to its entry point.
In this
example, the entry point is either the
address specified in the END record, or the
phase load address if the END address is
omitted, because the automatic ENTRY
statement is in effect.
This example can be modified to
illustrate the following:
1.

Catalog and execute. To cause this
phase to be cataloged rather than
merely linkage edited, change the
OPTION (statement 2) from LINK to
CATAL. The core image library
directory still refers to the old
version of the program. It is not
updated until /& has been read.

Catalog only. To catalog only, change
the OPTION (statement 2) from LINK to
CATAL and remOve all data following
the EXEC LNKEDT (statement 6) up to
the /& statement.

Catalog Object module in relocatable
library. The name to catalog the
object module into the relocatable
library must be added to the INCLUDE
statement. If the name is RELOCA, the
statement becomes INCLUDE RELOCA. The
relocatable object deck and /*
statement are removed. This form of
the INCLUDE statement is written on
SYSLNK when it is read by job control.
The linkage editor retrieves the
object module when it encounters the
INCLUDE statement because it uses
SYSLNK for input.

Section 3: Program Design

155

CATALOG TO CORE IMAGE LIBRARY EXAMPLE
/ / JOB CATALCIL

1.
2.
3.
4.

LINK EDIT AND CATALOG TO CORE IMAGE LIBRARY
SINGLE PHASE, MULTIPLE OBJECT MODULES, FOREGROUND PROGRAM
MIXTURE OF CATALOGED AND UNCATALOGED RELOCATABLE MODULES
LABELED TAPE FILES AND SEQUENTIAL DASD FILES TO BE PROCESSED
/ / ASSGN SYSLNK,X'190'
/ / OPTION CATAL
PHASE PROGB,F+32168
INCLUDE

*
*
*
*

Relocatable object deck
/*

INCLUDE SUBRX
INCLUDE SUBRY
INCLUDE
Relocatable object deck
/*

5.
6.
7.

// LBLTYP TAPE
// EXEC LNKEDT
/&

Explanation for Catalog to Core Image
Library
This example illustrates the cataloging of
a single phase composed of multiple
relocatable object modules. These modules
are located in the input stream and in the
relocatable library. Labeled tape files
and sequential DASD files are processed
when the phase is executed. The program is
executed in a foreground partition. Assume
that the foreground partition begins at
location 32168.
Statement 1: The SYSLNK assignment
indicates the relationship to the OPTION
statement, although it is not required
because of the assumed configuration.

phase, F plus the location address in the
foreground partition must be specified.
A program may be linkage edited to any
address that 7 falls within a foreground
partition. The load address does not have
to coincide with the partition address.
However, the program must be of such a size
that it can reside in the available core
defined from the load address to the end of
the partition.
The address may be expressed in one of
three forms:
X'hhhhhh'
dddddddd
nnnnK

A hexadecimal number of 4
to 6 digits
A decimal number of 5 to
8 digits
Where K
1024 and n is 2
to 4 digits

statement 2: The OPTION CATAL statement
sets the LINK switches, as well as a CATAL
switch. If SYSLNK is not assigned, the
statement is ignored. The linkage editor
control statements are not accepted unless
the OPTION statement is processed. Linkage
editing and cataloging to the core image
library will occur.

+X'8000', +32168, and +32K are equivalent.
The actual origin point of the phase is
adjusted upward from the address.
specification to allow for the partition
save area, and the label information
CLBLTYP statement reservation>.

Statement 3: Only one PHASE is
constructed.
It is cataloged to the core
image library and retrieved by the name
PROGB. Because this is to be a foreground

Statement 4: Four modules make up this
phase. The first and last are not
cataloged in the relocatable library;
therefore the object decks must be on

156

DOS System Programmer's Guide

SYSIPT, and each must be followed by the
end-ef-data recerd (/*>. SUBRX and SUBRY
are cataloged previously to' the relocatable
library by these names. Jeb centrel puts
the uncataloged modules on SYSLNK in place
ef their INCLUDE statements. Jeb control
copies the INCLUDE statements for the
cataloged modules.
Statement 5: The LBLTYP statement has the
operand TAPE, rather than NSD because
labeled tapes and sequential DASD files are
precessed when the phase is executed.
Eighty bytes are reserved ahead of the
actual phase for label infermatien. LBLTYP
NSD is also. satisfactory because it
generates a minimum of 104 bytes and tapes
require only 80.
Statement 6: The EXEC LNKEDT statement
causes the system loader to' bring in the
linkage editor program. SYSLNK new becemes
input to' the linkage editer. It centains
the fellewing:
PHASE PROGB,F+32768
First uncataleged relecatable deck
INCLUDE SUBRX
INCLUDE SUBRY
Second uncataloged relocatable deck
ENTRY
The medules are linkage edited so that they
occupy contigueus areas in main storage in
the sequence in which they appear in the
input stream. When the linkage editing is
completed, cataloging to' the core image
library eccurs because ef the CATAL option.
The core image directery is checked to' make
sure the new phase entries fit.
If net,
the job is canceled. The directory is
scanned fer any match to' a phase being

cataleged. A match is deleted frem the
directery. The system directery is updated
to' reflect the changes. Jeb contrel is
brought into main sterage.
Statement 7: Because CATAL was specified,
a special routine is executed when the /&
centrol statement is read by job centrel.
This reutine updates the transient,
library-routine, and foreground-pregram
directeries. A system status repert is
printed to reflect the usage and available
space in each ef the libraries and
directeries. These eperations de not occur
in a LINK situation. The /& resets the
CATAL eption, that is, it turns eff the
LINK and CATAL switches.
The example can be medified to
illustrate a cataleg-and-execute eperatien
by inserting the fellowing data between the
EXEC LNKEDT and /& statements:
1.

Any job centrol statements required
fer executien er PROGB

A // EXEC statement

Any card reader input for PROGB

Note that the actual update ef the
directories and the system status report
are delayed until completien ef the
execution of PROGB, when /& is read.
Frem
a system design standpeint this is net
desirable because ef possible eperatienal
problems. Making the execution ef PROGB a
separate job avoids any difficulties. All
cere image library directories are updated
at /&. This is time censuming and sheuld
net be done fer each module cataloged.

Section 3: Program Design

157

EXECUTE LINKAGE EDITOR IN FOREGROUND AND CATALOG TO PRIVATE CORE IMAGE LIBRARY tXAMPLE

1.
2.
3.
4.
5.

// JOB CATLCIL
* LINK EDIT AND CATALOG TO PRIVATE CORE IMAGE LIBRARY
* LINKAGE EDITOR EXECUTING IN FOREGROUND
* SINGLE PHASE, MULTIPLE OBJECT MODULES, FOREGROUND PROGRAM
* MIXTURE OF CATALOGED AND UNCATALOGED RELOCATABLE MODULES
* LABELED TAPE FILES AND SEQUENTIAL DASD FILES TO BE PROCESSED
ASSGN SYSCLB,X'191'
// ASSGN SYSLNK,X'190'
// OPTION CATAL
PHASE PROGB,S
INCLUDE
Relocatable object deck

INCLUDE SUBRX
INCLUDE SUBRY
INCLUDE
Relocatable object deck
6.
7.

// LBLTYP TAPE
/ / EXEC LNKEDT

Explanation for Catalog to Private Core
Image Library

the OPTION statement is processed. Linkage
editing and cataloging to the core image
library will occur.

This exampl.,a_illustrates the execution of
the linkage editor in a foreground
partition and the cataloging of a phase to
a private core image library. This
function is possible only in a system
supporting the batched-job foreground and
private core image library options. The
phase being cataloged is the same as that
in the previous example where the link edit
was executed in the background.

Statement 4: Only one PHASE is
constructed. It is cataloged to the
private core image library and retrieved by
the name PROGB. An origin point of S
origins PROGB at the starting address of
the foreground partition, plus the length
of the save areas and the area aSSigned to
the COMMON pool, if any.

Statement 1: The assignment of a private
library is accomplished by the ASSGN SYSCLB
statement. The label for SYSCLB must be
stored on PARSTD or STDLABEL cylinder, or,
if the DLBL statement is included in the
job stream, i t must follow the ASSGN SYSCLB
statement.
Statement 2: The SYSLNK assignment
indicates the relationship to the OPTION
statement, although it is not required
because of the assumed configuration.
statement 3: The OPTION CATAL statement
sets the LINK switches, as well as a CATAL
switch. If SYSLNK is not assigned, the
statement is ignored. The linkage editor
control statements are not accepted unless
158

DOS System Programmer's Guide

Statement 5: Four modules make up this
phase. The first and last are not
cataloged in the relocatable library;
therefore, the object decks must be on
SYSIPT, and each must be followed by the
end-of-data record (/*). SUBRX and SUBRY
are cataloged previously to the relocatable
library by those names. Job control puts
the uncataloged modules on SYSLNK in place
of their INCLUDE statements. Job control
copies the INCLUDE statements for the
cataloged modules.
Statement 6: The LBLTYP statement has the
operand TAPE, rather than NSD because
labeled tapes and sequential DASD files are
processed when the phase is executed.
Eighty bytes are reserved ahead of actual
phase for label information. LBLTYP NSD is
also satisfactory because it generates a

minimum of 104 bytes and tapes require only
80.

Statement 7: The EXEC LNKEDT statement
causes the system loader to bring in the
linkage editor program. SYSLNK now becomes
input to the linkage editor. It contains
the following:
PHASE PROGB,S
First uncataloged relocatable deck
INCLUDE SUBRX
Second uncataloged relocatable deck
The modules are link-edited so that they
occupy contiguous areas in main storage in
the sequence in which they appear in the
input stream. When the linkage editing is
completed, cataloging to the private core
image library occurs because of the CATAL
option. The private core image directory
is checked to make sure the new phase
entries fit. If not, the job is canceled.
The directory is scanned for any match to a
phase being cataloged. A match is deleted
from the directory. The system directory
is updated to reflect the changes. Job
control is brought into main storage.
Statement 8: Because CATAL was specified,
a special routine is executed when the /&
control statement is read by job control.
This routine updates the transient,

library-routine, and foreground-program
directories for the private core image
library. A system status report is printed
to reflect the usage and available space in
each of the libraries and directories •.
These operations do not occur in a LINK
situation. The /& resets the CATAL option,
that is, it turns off the LINK and CATAL
switches.
The example can be modified to
illustrate a catalog-and-execute operation
by inserting the following data between the
EXEC LNKEDT and /& statements:
1.

Any job control statements required
for execution or PROGB

A // EXEC statement

Any card reader input for PROGB

Note that the actual update of the
directories and the system status report·
are delayed until completion of the
execution of PROGB, when /& is read. From
a system design standpoint this is not
desirable because of possible operational
problems. Making the execution of PROGB a
separate job avoids any difficulties. All
core image library directories are updated
at /&. This is time consuming and should
not be done for each module cataloged.

Section 3: Program Design

159

COMPILE AND EXECUTE EXAMPLE

1.
2.
3.

// JOB COMPEXEC
* COMPILE OR ASSEMBLE, LINK EDIT AND EXECUTE
* SINGLE PHASE, MULTIPLE OBJECT MODULES, BACKGROUND PROGRAM
* SEQUENTIAL DASD FILES TO BE PROCESSED
* INPUT TO LINKAGE EDITOR FROM LANGUAGE TRANSLATOR, RELOCATABLE LIBRARY AND
// ASSGN SYSLNK,X'190'
// OPTION LINK
PHASE PROGA,S
// EXEC COBOL

SYSIPT~

COBOL source statements
/*

INCLUDE SUBRX
INCLUDE
Relocatable object module

6•

ENTRY BEGIN1
/ / EXEC LNKEDT
Any job control statements required for PROGA execution.

// EXEC
Any input data required for PROGA execution.
/*

Explanation for compile and Execute
The language translators provide the option
of placing their output on SYSLNK rather
than SYSPCH. Because the linkage editor
uses SYSLNK for input, a program can be
assembled or compiled, then linkage edited
and executed. This operation, known as
assemble/compile and execute, is
illustrated by this example.
All three sources of object module input
to the linkage editor are used: SYSIPT,
the relocatable library, and the output
from a language translator. It is assumed
that the phase is executed in the
background partition, and that only
sequential DASD files or unlabeled tape
files are processed.
statement 1: The SYSLNK assignment is
given to illustrate the relationship to the
OPTION statement, although it is not
required because of the assumed
configuration.
160

DOS System Programmer's Guide

Statement 2: Because SYSLNK is assigned,
the OPTION LINK statement sets the link
indicator switches.
Statement 3: The PHASE statement,must
always precede the relocatable modules to
which it applies; therefore, it is written
on SYSLNK first for later use by the
linkage editor. S is the origin ,point,
that is, the phase originate~ with the
first doubl~word at the end of the
supervisor plus length of the label save
area (as defined ~ LBLTYP) plus length of
the area assigned to the COMMON pool (if
any). This gives the same effect as *
gives for a single phase or the first
phase. As with the *, the S may be used
with a relocation factor, for example,
S+1024. The factor must always be
positive, because a negative factor could
cause the origin point to overlay the
supervisor.
Statement 4: The appropriate language
translator is called (in this case, COBOL).
The normal rules for compiling are
followed; the source deck must be on the
unit assigned to SYSIPT and the /* defines

the end of the source data. Because the
LINK switches are set, the output of the
language translator is written on SYSLNK.
Except for PL/I, FORTRAN (F) and the
ASSEMBLER (F) and 14K variant, the DECK
option is ignored when SYSLNK is used.
Statement 5: The INCLUDE SUBRX statement
is written on SYSLNK. The linkage editor
retrieves the named module from the
relocatable library. Because the operand
is blank, the next INCLUDE statement
signifies that the relocatable module is on
SYSIPT. The data on SYSIPT is copied on
SYSLNK until the /* statement.
Statement 6: The ENTRY statement is
written on SYSLNK as the last linkage
control statement. The symbol BEGIN1 must
be the name of a CSECT or a label
definition defined in the first phase. The
address of BEGIN1 becomes the transfer
address for the first phase of the program.
The ENTRY statement is used because the
user wishes to provide a specific entry
point rather than to use the point
specified in the END record or the load
address of the phase. The ENTRY statement
affects the first, or onlYr-phase.
Statement 7: No LBLTYP statement is
required, because only sequential DASD

files are to be processed. The rest of the
statements follow the same pattern as
discussed in the Linkage Editing and
Execute example. The input from SYSLNK to
the linkage editor is:
PHASE PROGA, S
Relocatable module produced by COBOL
compilation
INCLUDE SUBRX
Relocatable module from SYSIPT
ENTRY BEGIN1
If an error is detected during
compilation of a source program, the LINK
option is suppressed. Under these
circumstances the EXEC LNKEDT and EXEC
statements are ignored in this example.
This LINK option suppression should be kept
in mind if a series of programs is to be
compiled and cataloged as a single job.
Failure of one job step would cause failure
of all succeeding steps. Remember that an
OPTION LINK cannot be given if OPTION CATAL
is in effect, because message lS1nD
(STATEMENT OUT OF SEQUENCE) results. This
is an error in instruction to the system
because CATAL has functions that must be
performed when the next /& statement is
read. Therefore, the CATAL switch must
remain on, and linkage editing only cannot
be performed.

Section 3: Program Design

161

CATALOG FOR PHASE OVERLAY EXAMPLE

1.
2.
3.
4.
5.

//
*
*
*
//
//

JOB MULTPHAS
LINK EDIT AND CATALOG TO CORE IMAGE LIBRARY
MULTIPLE PHASES, MULTIPLE OBJECT MODULES, BACKGROUND PROGRAM
NO LABELED TAPE OR NONSEQUENTIAL DASD FILES TO BE PROCESSED
ASSGN SYSLNK,X'190'
OPTION CATAL
PHASE PHASEA,ROOT
INCLUDE MOD1
PHASE PHASEB,*
INCLUDE MOD2
PHASE PHASEC, PHASEB
INCLUDE MOD3
/ / EXEC LNKEDT
/&

Explanation for Catalog for Phase Overlay
Sometimes i t is not possible to bring an
entire program into main storage because it
requires more bytes of main storage than
are available. To solve this problem the
program can be broken into separate phases
that can be brought into main storage as
required, overlaying all or part of another
phase if desired. The linkage editing of
three cataloged relocatable modules into
phases for overlay is illustrated by this
example.
Statement 1: The OPTION CATAL sets the
switches so that the phases can be linkage
edited and cataloged into the core image
library.
Statement 2: PHASEA is considered the ROOT
phase, that is, it is always resident in
main storage during program execution. The
origin point is the first doubleword
address after the supervisor plus length of
the label save area (if any) plus length of
the area assigned to the COMMON pool (if

162

DOS System Programmer's Guide

any). Only the first phase statement is
permitted to specify ROOT.
Statement 3: The * in the PHASE card for
PHASEB causes MOD2 to be 110kage edited at
the end of PHASEA.
Statement 4: Because PHASEB is specified
as the load address of PHASEC, i t is
linkage edited into the same address as
PHASEB. The symbol that designates the
origin point may be a previously defined
phase name as in this example, a previously
defined control section, or a previously
defined external label. A plus or minus
relocation factor may be used (for example,
PHASE2+100).
statement 5: The EXEC LNKEDT causes all
three phases to be linkage edited and
cataloged. When the phases are executed,
the ROOT phase normally is loaded by the
system loader. The other phases would
probably be brought into main storage by
means of the FETCH or LOAD macro issued by
the calling phase.

SUBMODULAR STRUCTURE EXAMPLE
//
•
•
•
//
//
1.
2.
3.
4.

JOB SUBMOD
LINK EDIT AND CATALOG TO CORE IMAGE LIBRARY
MULTIPLE PHASES, ONE OBJECT MODULE (SUBMODULAR STRUCTURE)
BACKGROUND PROGRAM, NO LABEL STORAGE RESERVATION
ASSGN SYSLNK,X'190'
OPl'IONCATAL
PHASE PBASE1,ROOT
INCLUDE , (CSECT1,CSECT3)
PHASE PBASE2,.
INCLUDE , (CSECT2,CSECTS)
PHASE PHASE3, PHASE2
INCLUDE , (CSECT4,CSECT6)
INCLUDE
Relocatable object deck

// EXEC LNKEDT
/f,

Explanation for Submodular Structure
Several relocatable modules are structured
into several phases. In this example, a
single object module is broken into several
phases. The object module is composed of
CSECT1-CSECT6. It is structured into three
phases with overlay. The module is not
cataloged in the relocatable library. Only
the PHASE AND INCLUDE statements are
discussed.
statement 1: The INCLUDE statement tells
the linkage editor to place CSECT1 and
CSECT3 into PHASE1. The sequence in which
the CSECTs are linkage edited is determined
by the sequence in the input module rather
than the sequence in the INCLUDE statement.
(CSECT3,CSECT1) would give the same result
as (CSECT1,CSECT3). The sequence can be
controlled by issuing separate INCLUDE
statements. For example, INCLUDE ,(CSECT3)
followed by INCLUDE , (CSECT1) causes CSECT3
to be linkage edited before CSECT1,
regardless of the sequence in the object
module.
Note that the first operand is missing
in the INCLUDE statement, as indicated by
the leading comma. This format of the
INCLUDE statement searches the next
succeeding object module on SYSLNK to
locate the named CSECTs. See Statement 4.

Statement 2: The INCLUDE statement causes
CSECT2 and CSECTS to be used for PHASE2.
This phase is located at the end of PHASE1.
Statement 3: PHASE3 is made up of CSECT4
and CSECT6 and overlays PHASE2 because its
origin point is at the same address as
PHASE2.
Statement 4: This INCLUDE statement with a
blank operand is required to write the
object module that follows in the card
reader onto SYSLNK, to satisfy the INCLUDE
statements with a blank first operand.
With the sequence of statements shown in
the example, the PHASE and INCLUDE
statements are read from SYSRDR. However.
it is permissible to read PHASE and INCLUDE
statements from SYSIPT. To do this.
Statement 4 (INCLUDE blank) is placed ahead
of Statement 1. The INCLUDE with the blank
operand directs job control to read the
following data (which includes the PHASE.
INCLUDE~ and then the object module) on
SYSLNK from SYSIPT to the /* statement. If
SYSRDR and SYSIPT are separate devices.
take care to place the PHASE and INCLUDE
statements on the correct device.

PHASE1 is located at the end of the
supervisor plus length of the label save
area, and the COMMON area (if any).
Section 3: Program Design

163

PHASE and INCLUDE statements can also be
in the relocatable library. If the object
module is in the relocatable library under
the name MODi, the following changes are
made:
1.

Remove Statements 1 through 3, and add
module name to Statement 4.
// JOB SUBMOD
// OPTION CATAL
INCLUDE MODl
/ / EXEC LNKEDT

When the relocatable module is
cataloged to the library, precede it
with the following statements:
PHASE PHASE1, *
INCLUDE MODi, (CSECT1,CSECT3)
PHASE PHASE2, *
INCLUDE MOD1,(CSECT2,CSECTS)
PHASE PHASE3,PHASE2
INCLUDE MOD1,(CSECT4,CSECT6)
Relocatable object deck

164

DOS system programmer's Guide

This form of the INCLUDE statement causes
the linkage editor to search the module
that follows the last INCLUDE statement in
the library for the required control
sections.

SELF-RELOCATING AND MULTIPLE LINK EDITS EXAMPLE
//

*
//

1.
//

JOB MULTCATL
SEVERAL LINK EDITS AS A SINGLE JOB
OPTION CATAL
PHASE PROG1,+0
INCLUDE PROO1
EXEC LNKEDT
PHASE PROG2, *
INCLUDE
Relocatable object module

/*
//

//
/&
//
//

EXEC LNKEDT
PHASE PROG3,*
INCLUDE PROO 3
EXEC LNKEDT
JOB LINKGO
OPTION LINK
PHASE PROG4, *
INCLUDE
Relocatable object module

// EXEC LNKEDT
Any job control statements required for PROG4 execution.
/ / EXEC

Any input data required for PROG4.
/*

PHASE PROG5,*
INCLUDE PROG5
// EXEC LNKEDT
Any job control statements required for PROG5 execution.
// ,EXEC
Any input data required for PROG5.
/*
/&

Explanation for Self-Relocating and
Multiple Link Edits
The linkage editing reqUirements for a
self-relocating program and the combining
of several cataloging or link-and-execute
job steps into a single job are illustrated
in this example. Use discretion in
deciding how many steps should be combined
before a /&, because a failure in one step
can cause successive steps to be bypassed
unnecessarily.
Statement 1: The +0 displacement as the
origin part for PROG1 designates this
program as self-relocating. Once this
program is cataloged to the core image
library, it may be executed in any
partition. It can be initiated into the
background partition by job control, if the
supervisor was generated with

multiprogramming capabilities. A non-MPS
supervisor will give message OP77I
(CANCELED DUE TO INVALID ADDRESS).
However, the program can be loaded by using
the LOAD macro in a calling phase.
Statement 2: PROG2 and PROG3 are linkage
edited and cataloged as job steps within
the job MULTCATL. Note that OPTION CATAL
holds for these steps.
Statement 3: A new job is initiated
because the succeeding job steps are linked
and executed without cataloging.
(OPTION
LINK cannot be issued when OPTION CATAL is
in effect.) Note that OPTION LINK need not
be reissued before the next job step.

Section 3: Program Design

165

Checkpoint/Restart
When a background program or a hatched-job
foreground program is expected to run for
an extended period of time, provision may
be made for taking checkpoint records
periodically during the run. The records
contain the status of the job and system at
the time the records were written. Thus,
they provide a means of restarting at some
midway point rather than at the beginning
of the entire job, if processing must be
terminated for any reason before the normal
end of job.
For example, some malfunction, such as a
power failure, may occur and cause such an
interruption. If checkpoint records are
written periodically, operation can be
restarted with a set of checkpoint records
written before the interruption. These
records contain everything needed to
reinitialize the system when processing is
restarted.
Any programmer logical unit
(SYSOOO-SYSmax) assigned to tape, or the
2311, 2314, or 2319 can be used for
recording checkpoints if the proper file
definitions are made and the correct label
statements are submitted. Checkpoints must
not be taken on ASCII tape files.
The Disk Operating system includes
routines to take checkpoint records and to
restart a job at a given checkpoint. The
checkpoint and restart routines are
included in the core image library when the
system is generated. The checkpoint
routine is executed in the logical
transient area and is called in response to
a CHKPT macro instruction in the problem
program. The restart routine is called by
job control when it reads a RSTRT control
statement. When a program is restarted,
the user must reissue any STXIT macro
instructions that are desired because the
STXIT linkages established before the
checkpoint was taken are destroyed.
Checkpoint/restart does not save or restore
floating point registers.
(If needed,
these registers should be stored in the
problem program area before issuing CHKPT
macro, and restored in a user restart
routine.)
Only background programs or batched-job
foreground programs may be checkpointed.
Checkpoint records are written on a 2311,
2314, or 2319 DASD or on magnetic tape.
Each checkpoint is uniquely identified.
When it is restarted, the RSTRT control
statement specifies which checkpoint is to
be loaded. If multireel files are being
used, the operator must be aware of which
reels were being processed when the
checkpoint was taken.
166

DOS System Programmer's Guide

Multitasking users should only issue the
CHKPT macro in the main task with no
subtasks attached. In addition, no tracks
on any DASD should be in the hold state. A
multitasking abnormal termination routine
should not contain a CHKPT macro.
Checkpoints should be taken while a program
is running successfully, not while i t is
canceling. Checkpointed programs must be
restarted in the same partition in which
they were checkpointed. Multiple jobs
maybe checkpointed on the same tape.
Checkpoint records written by previous
versions of the system are not acceptable
to the current version of the system.
However, if they are embedded on magnetic
tape, they are bypassed by the current
version.
It is possible to increase partition
allocation between the time the checkpoint
is taken and the time the program is
restarted, if the starting address of the
partition remains unchanged.

PROBLEM PROGRAM RESPONSIBILITIES
Use of CHKPT Macro
Any partition, except a foreground
partition in a single program mode, can
issue the CHKPT macro successfully. If
multitasking, only the main task can
successfully checkpoint. CHKPT is ignored
when issued by a subtask, a foreground
partition in single program mode, or in any
of the following additional conditions:
1.

The device on which the checkpoint
records are to be written is not a
magnetic tape or a disk pack.
(The
device must be a 2311, 2314, or 2319
disk if the filename operand is
present. )

End of reel is detected while writing
the checkpoint on tape.

The area on disk is not large enough
for a single checkpoint.

The macro is issued by a
teleprocessing program that has any
I/O operation(s) pending on a
teleprocessing device.

The user-specified end address is
greater than the end of the problem
program area.

The CHKPT macro is issued before the
disk checkpoint file is opened.

Any of the required DTFPH parameters
for the disk checkpoint file contain
errors.

CHKPT Macro

If a subtask is attached in the
partition being checkpointed.

I Name I Operation I Operand

If any DASD track for the partition
being checkpointed is in the HOLD
state.

Note: Checkpoint records are not
permitted on ASCII tape files.
If a checkpoint is ignored, control
returns to the user with binary zeros in
register o. Otherwise, register 0 contains
the appropriate checkpoint number (in
unpacked decimal).
Checkpoints are usually taken after a
specified period of time has elapsed, or
after a certain volume of input is
processed. When multitasking, use the
following as a guide for selecting a
method:
1.

The multitasking operation requires
checkpoints to be taken on a time
interval basis. Therefore, at main
task execution time, a STXIT macro
establishes linkage for an interval
timer interrupt. In the main task
interval timer routine, the problem
program issues WAIT macros to wait for
the detachment of each subtask in the
partition, and then takes the
checkpoint. If the main task must
take an immediate checkpoint, the
interval timer routine in the main
task must first detach all subtasks,
disregarding current processing,
before it can successfully issue the
CHKPT macro.
The multitasking operation requires
checkpoints to be taken on a volume
basis. Therefore, the main task
attaches the subtasks necessary to
perform the job, and then issues WAIT
macros to wait for each subtask in the
partition to detach. Each subtask
keeps a count on the unit of work to
be performed and detaches when it is
finished. When all subtasks are
detached, the main task can take the
checkpoint.

After the checkpoint is taken, the main
task can then either attach more, or the
same, subtasks to continue processing.

r----T---------T--------------------------,I

~----+---------+--------------------------~
InamelCHKPT
1 sysnnn,{restart address} I

I
I

(r1)

I[,{
end address}] r, {tpointer:tll
I
(r2)
L (r3) fJI

I
I
I
I
I
I
I r, {dPointer}] r, {Filename}] I
L
__________________________
JI
I ____ I _________ I L
(r4)
L
(rS)
~

SYSnnn specifies the logical unit on which
the checkpoint information is stored. It
must be a magnetic tape or a disk pack.
(See Checkpoint File.)
Restart address (or r1) specifies a
symbolic name of the problem program
statement (or register containing the
address) at which execution is to restart
if processing must be continued later.
End address (or r2) is a symbolic name (or
register containing the address) of the
uppermost byte of the problem program area
required for restart. This address must
follow the logic modules being included
from the relocatable library.
If this operand is omitted, all of main
storage allocated to the partition are
checkpointed.
This operand provides two advantages:
1.

Less time and space is required for
recording the checkpoint record set.

If a program using 24K of storage is
being run in a larger system and only
24K is checkpointed, that program can
be restarted, either on a 24K system
or as a 24K partition in a
multiprogramming system.

Tpointer (or r3) is the symbolic name of an
eight-byte field contained in the problem
program area.
(See Repositioning Magnetic
Tape. )
Dpointer (or r4) is the symbolic name of a
DASD operator verification table that the
user can set up in his own area of main
storage.
(See DASD Operator Verification
Table. )
Filename (or rS) is used only for
checkpoint records on disk. It is the name
of the associated DTFPH macro.
(see
Checkpoint on Disk.) Special register
notation cannot be used with any of these
operands.
Section 3: Program Design

167

Information That Is and Is Not Saved: When
the CHKPT macro is issued, the following
information is saved:
• Information for the restart and other
supervisor or job control routines.
• The general registers.
• Bytes 8-10 and 12-45 of the
communication region.

CHECKPOINT FILE
The checkpoint information must be written
on a disk pack or a magnetic tape (either
7- or 9-track, EBCDIC only). The 7-track
tape can be in either data conversion or
translation mode; however, the magnetic
tape unit must have the data conversion
feature. On 7-track tapes, the 20-byte
checkpoint header and trailer labels are
written in the mode of the tape (Figure
3.9). The data records are written in data
convert mode, odd parity.

• The problem program area (see End
Address Operand).
Checkpoints On Tape
• All DASD file protection extents
attached to logical units belonging to
the checkpointed program.
The following information is not saved:
• The floating point registers.
(If
needed, these registers should be
stored in the problem program area
before issuing CHKPT, and restored in a
user restart routine.)
• Any linkage to user routines set by the
STXIT macro.
(If needed, STXIT should
be used in user's restart routine.>

The programmer can either establish a
separate file for checkpoints or embed the
checkpoint records in an output data file
(EBCDIC only). When the data file is read
at a later time using logical IOCS, the
checkpoint records are automaticallY
bypassed. If physical IOCS is used, the
user must program to bypass the checkpoint
record sets.
If a separate magnetic tape checkpoint
file with standard labels is maintained,
the labels should be either checked by an
OPEN routine or bypassed by a MTC command
before the first checkpoint is taken.

• Any timer values set by the SETIME
macro.
(If needed, SETIME should be
used in a user's restart routine.)
Checkpoints On Disk
• The program mask in problem program
PSW.
(If other than all zeros is
desired, the mask should be reset in
user's restart routine.)

NOTES FOR DASD AND MICR FILES
DASD system input or output files (SYSIPT,
SYSLST, etc.) must be reopened at restart
time. In the user's restart routine, the
programmer must be able to identify the
last record processed before checkpoint.
Magnetic Ink Character Reader (MICR)
files require the DTFMR supervisor linkages
to be initiated at restart time. This can
be accomplished by reopening the MICR file
in the user's restart routine. Because the
OPEN macro clears the docum~nt buffer, the
problem program must disengage the device
and process all follow up documents in the
document buffer before taking each
checkpoint.
168

DOS System Programmer's Guide

If checkpoints are written on disk, the
following must be observed:
1.

Define area of disk to be used by
writing a DTFPH macro and using a
DLBL, EXTENT label set.

The number of tracks required is
computed as follows:
n 1 +
[

X + L
30
20

+ .£
z

]

where:
n = the number of sets of checkpoint
records to be retained.
(When the
defined extent is full, the first
set of checkpoint records is
overlaid.)
c

The number of bytes to be
checkpointed in the user's problem
program up to the end address
specified in the CHKPT macro
operand.

x = The number of disk extents
including nonoverlapping
split-cylinder extents. If
split-cylinder extents overlap on
the same cylinder, the number of
extents counted is the one used by
the program.
(This number is zero
if DASD file-protect is not used.)

Open the area on disk by issuing an
OPEN to the DTFPH.

Issue a CHKPT macro that points to the
DTFPH to be used.

When restarting checkpointed jobs, the
DTFPH filename is specified in the
RSTRT job control card.

Each program can use a common
checkpoint file or define a separate
one. If a common file is used, only
the last program using the file can be
restarted.

y = same as preceding for 2321.
z = 3625, if checkpoint records are
written on a 2311. 7249, if
checkpoint records are written on
a 2314/2319.
For each division, the remainder is
rounded to the next highest whole
number before multiplying by n.

See Figure 3.6 for an example using the
checkpointed facility on disk.

r----------T-----------T-------------------------------------------------------,
PAYROLL
START
I
•

CHKPDSK

•
DTFPH

DEVICE=2311,MOUNTED=SINGLE,TYPEFLE=OUTPUT

•
•
OPEN

CHKPDSK

•

•
CHKPT

SYS004,RSTRT,END"DVER,CHKPDSK

•

I
II
I
II
I
II
I
II

•
END
~----------~-----------~-------------------------------------------------------~

I
I / / JOB CHKPT
I
I ASSGN SYS004,X'190'
I
I ASSGN SYSOOO,X'180'
I
I ASSGN SYS001,X'181'
I
I ASSGN SYS002,X'182'
I
I // DLBL CHKPDSK,'CHECKPOINT FILE'",
I
I // EXTENT SYS004,DOS-II",1900,89
I
IL ______________________________________________________________________________
/ / EXEC PAYROLL
J
Figure 3.6.

Using Checkpoint Facility on Disk

Section 3: Program Design

169

REPOSITIONING I/O FILES
The I/O files used by the checkpointed
program must be repositioned on restart to
the next record that the user wants to read
or write. The checkpoint facility does not
provide aids for repositioning unit record
files. The programmer must establish his
own repositioning aids and communicate
these to the operator, when necessary.
Some suggested ways are:
1.

The fourth operand of the CBKPT macro
pOints to two V-type address constants that
the user specifies in his coding. The
order of these constants is important.
1.

The first constant points to a table
containing the filenames of all the
logical IOCS magnetic tape files that
are to be repositioned.

The second constant points to a table
containing repositioning information
for physical IOCS magnetic tape files
that are to be repositioned.

If the first, second, or both
constants are zero, no tapes processed
by logical, physical, or both types of
IOCS, respectively, are repositioned.

Take checkpoints at a logical breaking
point in the data, such as paper tape
end~of-reel.

Switch card stackers after each
checkpoint.

Print information at the time of
checkpoint to identify the record in
process.

Issue checkpoints on operator demand.

Operand

Name

User sequential DASD input, output, or work
files require no repositioning.
When updating DASD records in an
existing file, the programmer must be able
to identify the last record updated at the
time of the checkpoint in case he needs to
restart. This can be done in various ways:
1.

Create a history file to record all
updates by dumping an image of the
direct access record on tape as soon
as it has been read. When a restart
is initiated, these records can be
used to rewrite the file and establish
the status that existed when the
corresponding checkpoint was taken.
When this is completed, normal restart
procedures are accomplished and
reprocessing begins.
Create a field in updated records to
identify the last transaction record
that updated it. This field can be
compared with each transaction at
restart time.

Repositioning Magnetic Tape
Checkpoint provides some aid in
repositioning magnetic tape files at
restart. Files can be repositioned to the
record following the last record processed
at checkpoint.
The following discussion presents the
procedure in correlation with Figure 3.7.

110

DOS System Programmer's Guide

SYSOOx,(rl)" POINTER

CHKPT

,.r----------------------~
POINTER

r - - - - - V(lOGICL)
V(PHYSCL)
I

r---------'
I

LOGICL

CNOP

r-----J
2,4

H 'n' number of entries
as follows.

V(filename 1)
V(filename2)

•
•
•

symbolic DTF
name of each
tape file to
be repositioned
at restart

V{filenamen}
r - - - - - - - - - - ..J

,.
I

PHYSCL

H'n' number of entries
as follows.
3H
•
•
•

r- - - - - - - - - - - - _J

six bytes (3 halfwords)
for each tape fi Ie that
is to be repositioned
at restart.

I
I

,
I

filename 1

DTfxx

Figure 3.1.

Procedure for Building Tape
Repositioning Tables

If the tables are contained in the same
CSECT as the CHKPT macro, the constants may
be defined as A-type constants. The user
must build the tables discussed. Each
filename in the logical IOCS table points
to the corresponding DTF table where IOCS
maintains repositioning information. The
user should note the following:
1.

Magnetic tapes with nonstandard labels
should be repositioned past the labels
at restart time (presumably the labels
are followed by a tapemark so that
foreward space file may be used).
If a tape that is to be repositioned
is processed with nonstandard labels
and is read backwards, the user must
keep a physical IOCS repositioning
table, because the physical record
count kept by loes will be incorrect.
The physical record count must be the
number of forward reads necessary for
restart to position the tape.

Restart does not rewind magnetic tapes
when repositioning them.

A multifile reel should be
prepositioned to the beginning of the
desired file.

The correct volume of a multivolume
file must be mounted for restart.

For tapes with a standard VOL label,
restart writes the file serial number
and volume sequence number on SYSLOG,
and gives the operator the opportunity
to verify that the correct reel is
mounted.

IOCS can completely reposition files
on system logical units (SYSIPT,
SYSLST, etc), if the tape is not
shared with any other program and if
the user keeps a physical IOCS
repositioning table. However, if a
system logical unit file is shared
with other programs, a problem exists.
output produced after the checkpoint
is duplicated at restart. Input
records must be reconstructed from the
checkpoint, or the user restart
routine must find the last record
processed before checkpoint.

The entries in the physical IOCS table are
as follows:
First halfword: hexadecimal representation
of the symbolic unit number of the tape
(copy from CCB bytes 6 and 7).
Second halfword: number of files within
the tape in binary notation. That is, the
number of tapemarks between the beginning
of tape and the position at checkpoint.
Third halfword: number, in binary
notation, of physical records between the
preceding tapemark and the position at
checkpoint.

DASD operator Verification Table
If the Dpointer operand in the CHKPT macro
is used, the user can build a table in his
own area of main storage to provide the
symbolic unit number and the bin (cell)
number of each DASD file used by his
program. At restart, the volume sequence
number of these files is printed on SYSLOG,
and the operator can verify them.
The entries in the DASD operator
verification table must consist of the
following two halfwords, in the order
stated:
1.

The symbolic unit in hexadecimal
notation copied from CCB bytes 6 and
7.

The bin (cell) number in hexadecimal
notation is always zero, except for a
2321, in which case the bin number
varies with the cell (0-9) being
verified.

There must be one table entry for each
DASD unit to be verified by the operator.

section 3: Program Design

171

See Figure 3.8 for the procedure for
building a DASD operator verification
table.

r-----T-------~---------------------------l

I Bytes I Contents

~-----+-----------------------------------~
0-11

.:.>:~.:~.:.:.!.;.:.:.:.:.;.:~.:.:.:.;.:.~:.;.:.:.:.:.:.','

:~·:'.~"'h;~·:·:~~~·;·:·:·:·:·:~·:·:·}:·:·:·;·:·:~·:·:';';.;.:.;

CHKPT

SYSOOx, RSTRT, END" DVER, CHKPDSK
I

r----------------------~
I

I
I
I

r---------------~
I

t
DVER

CNOP

2, 4

H'n' number of entries
as follows:
2H
•
•

,.. _________ J

•

4 bytes (2 halfwords)
are required for each
DASD so that the
operator can verify
each volume sequence
number at restart time.

CHKPDSK

Figure 3.8.

DTFPH

Procedure for Building DASD
Operator Verification Table

BYPASSING EMBEDDED CHECKPOINT RECORDS ON
TAPE WITH PHYSICAL IOCS
The checkpoint information saved is written
as a set of magnetic tape records
consisting of a 20-byte header record, as
many core-image records as required to save
the necessary parts of main storage, and a
20-byte trailer record identical to the
header. See Figure 3.9 for the format of
header and trailer record.
If checkpoint sets are embedded in a
file being read with physical IOCS, they
must be recognized and bypassed. On any
mode input tape, checkpoint sets may be
identified by the first 12 bytes of the
header or trailer records. Note that when
reading backwards, the checkpoint header
occupies the 20 low-order bytes of the
input area.

172

14-15 The total number, in unpacked
hexadecimal, of records following
the header.
16-19 The serial number of the check
L _____ ~ ___________________________________
J
point.

Figure 3.9.

DOS System Programmer's Guide

Format of the Checkpoint
Header/Trailer Records

When bypassing checkpoint sets, three
methods are possible:
1.

Go into a read loop (forward or
backward) until the checkpoint trailer
(header if backward) is encountered.

Extract the count from bytes 12-13 of
the header (or trailer if backwards),
add 2 to this, and forward-space (or
backspace) that number of records.
Read commands could also be used.

Extract bytes 14-15 of the header (or
trailer if backwards), pack and
convert the field to binary, and
forward-space (or backspace) that
number of records. Read commands
could also be used.

/// CHKPT //

12-13 The number, in binary, of core
image records following the header •

Operand

Operation

Name

When bypassing checkpoint sets on
7-track tapes in translate mode, only
method 3 can be used and only forward-space
(or backspace) record commands (not reads)
can be used. Reads would create data
checks.

BYPASSING CHECKPOINT RECORDS ON TAPE WITH
LOGICAL IOCS
When a tape input file contains checkpoint
records interspersed among the data
records, the DTFMT macro parameter
CKPTREC=YES is required. When this
parameter is specified, logical IOCS
bypasses the embedded checkpoint records.

RESTARTING CHECKPOINTED PROGRAMS
Job control prepares the system for
restarting from a checkpoint by loading the
restart program that repositions tape
units, reinitializes the communication
region, and stores the information from the
RSTRT statement. The restart program
handles the actual restarting of the
problem program.

When a checkpoint is taken, the
completed checkpoint is noted on SYSLOG.
Restarting can be done from any checkpoint
record. not just the last. The job name
specified in the JOB statement must be
identical to the job name used when the
checkpoint was taken. The proper 1/0
device assignments must precede the RSTRT
control statement.
Assignment of input/output devices to
symbolic unit names may vary from the
initial assignment. Assignments are made
for restarting jobs in the same manner as
assignments are made for normal jobs.

RSTRT Statement
The restart facility allows the programmer
to continue execution of an interrupted job
at a point other than the beginning. The
procedure is to submit a group of job
control statements including a restart
(RSTRT) statement. The control statements
necessary to restart a job from a
checkpoint are:
1.

JOB statement specifying the same job
name used when the checkpoint was
taken.

ASSGN statements for assigning I/O
devices to the symbolic unit names.

RSTRT statement specifying the unit
that contains the checkpoints and the
checkpoint ID number taken from the
message printed when the checkpoint
was taken. The format of the RSTRT
statement is:

IBM 3211 Printer Support
The addition of tapeless forms control and
improvements in the use of the loadable
print character buffer required special
programming support for the IBM 3211
Printer. SYSBUFLD is the service program
that loads the Universal Character Set
Buffer (UCSB) and the Forms Control Buffer
(FCB) with buffer load programs for the
3211 printer. SYSBUFLD is self-relocating.
requires 2K of main storage. and is
executed as a job step under BJF or SPI.
It is initiated by the command:
// EXEC SYSBUFLD
$$BUFLDR is another 3211 program, called
by IPL to load the UCSB and FCB buffer
loads from the core image library.

// RSTRT SYsxxx,nnnn,filename
SYSxxx

nnnn

filename

Symbolic unit name of the device
on which the checkpoint records
are stored. This unit must have
been previously assigned.
Identification of the checkpoint
record to be used for restarting.
This serial number is four
characters and corresponds to the
checkpoint identification used
when the checkpoint was taken.
The serial number is supplied by
the checkpoint routine.
symbolic name of the 2311 or 2314
disk checkpoint file to be used
for restarting. It must be
identical to the filename of the
DTFPH used to describe the disk
checkpoint file and the fifth
parameter of the CHKPT macro
instruction. This operand
applies only when specifying a
2311 or 2314 disk as the
checkpoint file.

system Considerations
When a 3211 PUB is encountered, the IPL
program calls the buffer load transient,
$$BUFLDR. $$BUFLDR in turn calls $$BUCB to
load the UCSB. and $$BFCB to load the FCB.
As supplied by IBM, these two phases
contain:
$$BFCB the configuration for a 66-line
page (at 6 lines per inch). with
56 lines available for printing.
a channel 1 for line 1, and a
channel 12 for line 56.
$$BUCB the character configuration for
the All train.
Also, $$BUFLDR sets the UCSB for folding
and suppressing data checks.
SYSBUFLD can be used to change the UCSB
or FCB configurations, or to reload these
buffers if a hardware failure occurs.
Section 3: Program Design

173

Only one FCB load is supplied in the DOS
system. Additional FCB loads can be
created using the procedure in the DOS
system Control and Service listed in the
Preface. FCB loads can be either cataloged
in the core image library or can be card
images. Any of your FCB loads can be
cataloged in the core image library as
$$BFCB to enable the IPL program to load
it. If a phase name is not specified
during an FCB load, SYSBUFLD loads the
buffer from SYSIPT.
All UCSB programs must be cataloged in
the core image library. They are loaded by
specifying them in the SYSBUFLD control
card.
The four rema1n1ng standard train
configurations (Gll, Hl1, P11, or T11) are
in the relocatable library under the name
IJBTRxll, where x is G, H, P, or T. These
can be cataloged in the CIL to be loaded
according to the train configuration.
Non-standard UCSB loads can be created
following the procedure in the DOS System
Control and Service listed in the Preface.
Any of these standard or nonstandard loads
can be cataloged as $$BUCB to enable the
IPL program to load that particular train
configuration.

174

DOS System Programmer's Guide

Error Recovery Techniques
The simplest error recovery technique for
the 3211 printer is specifying ERROPT=RETRY
in the DTFPR, which sets CCB byte 2, bit 5
(PRINTOV=YES also sets this bit). This
causes one automatic retry of the
equipment-check/command-retry error.
A more comprehensive technique is
specifying ERROPT=YES in the PRMOD and
ERROPT=name in the DTFPR, which sets eeB
byte 2, bits 5 and 6. These bits indicate
linkage to your error recovery routine
named in the DTFPR, and provide automatic
retry of the equipment-check/command-retry
error.
Return from your error recovery routine
is by register 14. Both registers 14 and
15 must be saved if LIoes is used during
error recovery.
PIoes users can provide linkage to error
recovery by testing the applicable bits in
the eeB. See Figure 3.10 for the error
indicators.
If PIoes or LIoes is used, the sense
information is not available to the user.
The sense command, issued by the DOS error
recovery routines, clears the 3211 sense
information.

CCB Byte

Bit

UCSB Parity Check - Line Complete:
There has been a parity error in at least one position of the UCSB. All characters in the I ine have been
printed and line spacing has taken place. The position in error has been cleared, so errors will not be
encountered during subsequent scans of the UCSB.
Printer speed is degraded as a result of this error. Full printing speed cannot be regained until the UCSB
is reloaded. You need a routine to reload the UCSB during the job step, or use SYSBUFLD to reload the
UCSB before the next job step.

Equipment Check/Print Check:
This is a hardware error that has resulted in an incomplete I ine, but I ine spacing has taken place. Check
for possible line position or print quality errors. The line in error must be either accepted or the page
reprinted (a user - written routine is needed).

Equipment Check/Print Quality:
This indicates a hardware error has occurred that caused light or blurred printing. Line spacing has
taken place. Check for possible print check or line position errors.
The line in error must be accepted or the page reprinted (a user-written routine is needed).

Line Position Error:
This can be a hardware error, a parity error in the FCB, or the result of a skip to a channel code not in
the FeB. Check for possible print check or print quality errors.
There is no way to tell where the carriage is positioned relative to the FBC; physical repositioning, as
well as reloading of the FCB, may be necessary.

Data Check/Print Check:
An unprintable character has been sent to the printer. The appl icable position (s) in the print line are
bl ank and the paper has been spaced.
The line in error must be either accepted or the page reprinted (a user-written routine is needed).
Note: Check or reload the UCSB to ensure that the correct load has been used.

USCB Parity Check/Command Retry:
There has been a parity error in at least one position of the UCSB. The appl icable position(s) in the
print line is blank, but the paper has not been spaced. Subsequent attempts to print the applicable
character will result in a data check/print check.
The UCSB can be reloaded (a user - written routine is needed), and the applicable command (s)
reissued.

I Figure

3.10.

Error

3211 Error Status Indicator Bits in the CCB

Macro Writing
The macro-definition language discussed
here provides a systematic means by which
the oos/360 assembler language programmer
can develop macro instructions, thereby
expanding the set of machine-oriented
instructions that serve as the basis of the
assembler language. This enables the
programmer to reduce programming effort and
shorten the assembler language source
programs. With the aid of the macro
language, any sequence of statements can be
summarized into a single macro definition.
Once written, this definition can be stored
and referred to at any time, thus supplying
the programmer with precoded routines. The
programmer only writes a single statement,
a macro instruction, to access the macro

definition and retain access to all machine
facilities.
systematic use of macro instructions
simplifies the coding of programs, reduces
the frequency of programming errors, and
encourages the use of carefully
standardized sequences of assembler
language statements for routine functions.
There are two classes of macros in the
Disk Operating System: system macros,
which are IBM-written macros supplied with
the system, and user macros, which are
defined by the user. The user macros may
be included in the source program and/or
may be entered into the source statement
library.
Section 3: Program Design

175

The source statement library contains
both user and system macro definitions.
This library, which can be a part of system
Residence (SYSRES) or a private library,
eliminates the need for including
definitions in the source module.

MACRO INSTRUCTION
The macro instruction statement is in
assembler statement format. Symbols are
used as a shorthand method of representing
rules, definitions, etc. These macro
instructions result in a one-for-one
assembler statement.
The name field of the macro instruction
may contain a symbol that is not defined
unless a symbolic parameter appears in the
name field of the prototype and the same
parameter appears in the name field of the
generated model statement (see The Macro
Definition) •
The operation field contains the
mnemonic operation code of the macro

instruction and has to be the same as the
mnemonic operation code in the source
program or in the source statement library.
The placement and order of the operands
in the macro instruction statement is
determined by the placement and order of
the symbolic parameters in the operand
field of the prototype statement. The
operand field contains from 0-200 entries,
separated by commas (entries are commonly
referred to as parameters). Any
combination of up to 255 characters may be
used as a macro instruction operand if the
rules concerning apostrophes, parentheses,
equal signs, ampersands, commas and blanks
are observed. These are described in the
Tape Operating Systems Assembler Language
publication listed in the front of this
manual.
The operand maybe written in a format
different than that used for assembler
language statements. The alternate format
described here allows the programmer to
write an operand on each line and allows
the interspersing of operands and comments
in the statement. Figure 3.11 illustrates
the operand formats.

r------T---------~--------------------------------------T-----------,

I Name I OperationlOperand Comments
I Col. 72
I
~------+----------+--------------------------------------+-----------~
I
I
I
I
I
I NAMEl I OPl
I OPERAND1, OPERAND2,
I
x
I
I
I
IOPERAND3 THIS IS THE NORMAL FORMAT
I
I

INAME2 I OP2
IOPERAND1, THIS IS THE
I
x
I
I
I
I OPERAND 2 , OPERAND3 ALTERNATE FORMAT
I
I
I
I
I
I
I
INAMEJ I OP3
IOPERAND1, THIS IS A COMBINATION
I
x
I
I
I
IOPERAND2,OPERAND3,OPERAND4,
I
x
I
IL______ I _________ I ______________________________________
OPERANDS OF BOTH FORMATS
I ___________ JI
~

Figure 3.11.

Operand Field Formats

When a program is written in the Disk
Operating system macro language (an
extension of DOS assembler language), one
of three macro instruction formats can be
used: keyword, pOSitional, or mixed.
Figure 3.12 shows the typical form of a
keyword macro instruction to be used with a
keyword macro definition.

176

DOS System Programmer's Guide

r----------T--------~--------------------,

I Name

I Operation I Operand
I
~----------+---------+~-------------------~
IA symbol, Imnemonic IZero to 100 or 200 I
I sequence I operation I operands, separated I
Isymbol, orlcode
Iby commas.
I
I _________
not used -LI ________-LI ____________________JI
L

Figure 3.12.

Keyword Macro Instruction

Each operand consists of a keyword
immediately followed by an optional value.
Nested keywords are not permitted. A
keyword consists of one to seven letters
and digits, the first of which must be a
letter. The operands of a keyword macro
instruction may be written in any order.
If an operand is omitted, the comma that
would have separated it from the next
operand need not be written.
The following are valid keyword macro
instruction operands:
A4=F'6041'
DUPE4=MEMBER

r----y---------T--------------------------,I

I Name I Operation I Operand

~----+---------+--------------------------~

I
IEXAMPLE liA,&B,&C,&D,&E,iF
I
__________________________JI
IL____ I _________
EXAMPLE 117,*+4"AREA,FIELD(6)
~

Figure 3.14.

Macro Instruction with
Prototype

Mixed-mode macro instruction operands
are a combination of both positional and
keyword operands. Certain operand entries
(positional) must be written in a fixed
order: other operand entries (keyword) can
be specified in any order. Figure 3.15
illustrates the mixed-mode macro
instruction.

so=
r---------~---------T--------------------,

I Name
The following are invalid keyword macro
instruction operands:

I Operation I Operand

~----------+---------+--------------------~

&X4.P3=O(1,4)

Keyword does not begin
with a letter.

IA symbol, Imnemonic Izero to 100 or 200 I
I sequence I operation I operands, separated I
I symbol or I code
I by commas.
I
L
_________
Inot
used -LI ________ -LI ____________________ JI

CARDAREA=B+l

Keyword is more than seven
characters.

Figure 3.15.

=(TO(S),(AFTER»

No keyword.

.Thetypical macro instruction is
.
positional unless otherwise indicated. The
positional macro instruction operands are
written in a fixed order. Figure 3.13
illustrates the positional macro
instruction.

r----------y---------T--------------------,
I Operation I Operand
I

I Name

~----------+---------+--------------------~

IA symbol, Imnemonic Izero to 100 or 200 I
I sequence loperationloperands, separated I
I symbol, or I code
I by commas.
I
L
I __________
not used I _________ I ____________ _______ JI
~

Figure 3.13.

Positional Macro Instruction

The positional operands, if omi~ted from
the macro instruction but appear in the
prototype, are replaced by the comma that
would have separated them from the next
operand. If the last operand is omitted
from a macro instruction, then the comma(s)
separating the last operand from the
previous operand may be omitted.
Figure 3.14 shows a macro instruction
preceded by its corresponding prototype
statement. The third and sixth operands of
the macro instruction corresponding to the
third and sixth operands of the prototype
statement are omitted in this example.

Mixed Macro Instruction

The operand consists of two parts. The
first part corresponds to the positional
operands. and is written in the same way
that the operand entry of a positional
macro instruction is written. The second
part of the operand corresponds to the
keyword operands. This part is written in
the same way that the operand entry of a
keyword macro instruction is written.
Figure 3.16 illustrates these facilities.

r----T---------T--------------------------,I

I Name I Operation I Operand

~----+---------+--------------------------~

I MACRO
I
I
I&N IMOVE
I iTY,&P,iR,iTO=,iF=
I
liN IST&TY
liR,SAVE
I
I
I LiTY
I iR, &P , iF
I
I
IST&TY
liR,iP,iTO
I
_________ liR,SAVE
__________________________ JI
IL____ IL&TY
~

Figure 3.16.

Mixed-Mode Definition

MACRO DEFINITION
A macro instruction cannot be assembled
unless a macro definition is made available
to the assembler. A macro definition is a
set of statements that provide the
assembler with:
1.

The name entry, the mnemonic operation
code, and the form of the macro
instruction operand, and
Section 3: Program Design

177

The sequence of statements that the
assembler uses when the macro
instruction appears in the source
program.

Assembler Language publication listed
in the front of this manual.
2.

Elements of the Macro Definition
Every macro definition is made up of four
elements: header statement, prototype
statement, model statement, and a
macro-definition trailer statement.
The macro instruction header and trailer
statements denote the beginning and end of
a macro definition respectively. The
header statement's name field contains
blanks, the operation field contains the
word MACRO, and the operand field contains
blanks. The header statement is the first
statement of the macro definition. The
trailer statement follows the same general
outline as the header statement; only the
operation field, that contains the word
MEND, is different. The trailer statement
must be present to denote the end of the
macro definition.

3 and 4. scaling (S') and Integer (I'):
Scaling and integer attributes are
provided for symbols that name fixed
point, floating point, and decimal DC
or OS statement. The programmer may
refer to the length, scaling, and
integer attributes in the operand
field of a SETA instruction, or in
arithmetic relations in the operand
fields ofSETB or AIF instructions.
5.

A prototype is defined as an original
model on which something is patterned. The
prototype statement of a macro definition
specifies in the operation column, the name
of the macro and the format for the operand
of all macro instructions that make use of
this definition. This prototype statement
must be the second statement in the macro
definition.
Like the operand of the macro
instruction, the prototype statement may be
written in a form different from that used
for machine or assembler instructions. The
alternate form is described under The Macro
Language.

ATTRIBUTES
6.
The assembler assigns attributes to macro
instruction operands and to symbols in the
program. These attributes may be referred
to in conditional assembly instructions
under Conditional Assembly Statements.
There are six kinds of attributes:
1.

178

Type (T'): The type attribute of a
macro instruction is a letter. The
type attribute may be referred to in
the operand of a SETC instruction, or
in character relations in the operands
of SETB or AIF instruction. The
letters used with the type attribute
and their meanings can be found in the
DOS System Programmer's Guide

Length (L'): The length attribute of
a symbol (or ofa macro instruction
operand that is a symbol) is the
length of the specified operand.
Reference to the length attribute of a
variable symbol is illegal except for
symbolic parameters in SETA, SETB, and
AIF statements. Reference must not be
made to the length attributes of
symbols whose type attributes are the
letters M, N, 0, T, or U.

Count Attribute (K'): The programmer
may refer to the count attribute of
macro instruction operands only. The
count attribute is a value equal to
the number of characters in the macro
instruction operand after substituting
for variable symbols excluding commas.
If the operand is a sublist, the count
attribute includes the beginning and
ending parentheses and the commas
within the sublist. The count
attribute of an omitted operand is
zero. The count attribute is one of
the following:
a.

the operand field of a SETA
instruction, or

in arithmetic relations, the
operand field of SETB or AIF
instructions which are part of a
macro definition can be referenced
by the user.

Number Attribute (N'): The number
attribute of macro instruction
operands, only, is referenced. The
number attribute is a value equal to
the number of operands in an operand
sublist. The number of operands in an
operand sublist is equal to one, plus
the number of commas that indicate the
end of an operand sublist. If the
macro instruction operand is not a
sublist, the number attribute is one.
If the macro instruction operand is
omitted, the number attribute is zero.
Reference may be made to the number
attribute in the operand field of a
SETA instruction, or in arithmetic
relations in the operand fields of

SETB and AIF instructions that are
part of a macro definition.

Name field entry which can contain
blanks, a symbol, a symbolic parameter
or a sequence symbol.

The operation entry may contain any
machine instruction, conditional
instruction, assembler instruction, or
symbolic parameter except COPY, END,
ICTL, ISEQ and PRINT; or it may
contain a variable symbol, depending
on the statement.

The operand entry may contain ordinary
symbols or variable symbols. After
substitution, the operand must not be
greater than 121 characters. Model
statements must follow the rules for
paired apostrophes, ampersands, and
blanks, as macro instruction operands.

The comments field entry which can
contain descriptive items of
information about the program is
inserted after the operand. All 256
valid characters, including blanks,
may be used in writing a comment. The
entry cannot extend beyond the end
column (normally column 11) and a
blank must separate i t from the
operand.

SUBLIST NOTATION
A sublist is one or more operands,
separated by commas and enclosed in paired
parentheses. An operand of a macro
instruction may be a sublist. The entire
sublist, including all operands, commas and
parentheses is considered one macro
instruction operand. Each operand entry
within the parentheses is called a sublist
member.
Sublists provide the user with a
convenient way to refer to a collection of
macro instruction operands as a single
operand, or a single operand in a
collection of operands. For the accessing
of individual members, the left parenthesis
of the sublist notation must immediately
follow the last character of the symbolic
parameter. A period should not be placed
between the left parenthesis and the last
character of the symbolic parameter (Figure
3.17).
r----~--------T-----------------,

I NAME I OPERATION I OPERAND
I
~----+---------+-----------------~
HEADER
I
I MACRO
I
I
Prototype I
IADDNUM
I&NUM,®,&AREA I
Model
I
IL
I®,&NUM(l)
I
Model
I
IA
I®,&NUM(2)
I
Model
I
IA
I®,&NUM(3)
I
Model
liST
I®,&AREA
I
Trailer I
I MEND
I
I
~----+---------+-----------------~

Figure 3.11.

The three types of variable symbols are
symbolic parameters, SET symbols and system
variables.

MACRO
I
IADDNUM
I (A,B,C),6,SUM
I
Generated I
IL
16,A
I
Generated I
IA
16,B
I
Generated I
IA
16,C
I
Generated LI ____ 1ST
_________ 16,SUM
_________________ JI
~

VARIABLE SYMBOLS

Symbolic Parameter

Sublist Illustration

The operand of the macro instruction
that corresponds to symbolic parameter &NUM
is a sublist. One of the operands in the
sublist is referred to in the operand entry
of three of the model statements.
Model statements are the macro
definition statements from which the
desired sequences of machine instructions
and certain assembler instructions are
generated. Zero or more model statements
may follow the prototype statement. A
model statement consists of one to four
entries:

The symbolic parameter consists of an
ampersand (first character) followed by one
to seven letters and/or numbers, the first
of which must be a letter. Symbolic
parameters appear in prototype and model
statements. They are assigned values by
the programmer when he writes a macro
instruction. The programmer should not use
&SYS as the first four characters of a
symbolic parameter.
Example:
Valid

Invalid

&LOOP2

&2BAC (first character
after & not letter)

'READER

&AREA2456 (too long)
section 3: Program Design

119

SET Symbols
SET symbols follow the same rules for
structure as symbolic parameters. SET
symbols differ from symbolic parameters in
three ways:
1.

their position in an assembler
language source program,

how they are assigned values, and

how their assigned values can be
changed.

character data. These operands are called
local SETA, SETB, or SETC symbols. If a
local instruction is part of a macro
definition, it must immediately follow the
prototype statement and any global
instructions or another LCLA, LCLB, or
LCLC. The local definition of these
operands indicates that each SET symbol is
defined only within the defining macro.
The GBLA or LCLA, GBLB or LCLB, GBLC or
LCLC operands are assigned initial values
of 0 (X'FO'), 0 (X'OO'), and null character
(no hexadecimal number), respectively.
The SETA, SETB, and SETC symbols are
assigned the initial values of 0, 0, and
null character value, respectively.

Defining SET Symbols
A SET symbol must be defined by the
programmer before it can be used. It is
defined by appearing as an operand of a
global or local instruction (GBLA, GBLB,
GBLC, LCLA, LCLB, LCLC). Figure 3.18 shows
the typical format of global and local
instructions.

r----------T---------T--------------------,
I Name
I Operation I Operand
I
~----------+---------+--------------------~
Not used; IGBLA,
lOne or more variable
must not IGBLB,
Isymbols used as
be present GBLC, or ISET symbols and
I separated by commas.
I(If more than one
Imacro defines one
LCLA,
Iglobal SET symbol
LCLB,
land these macros are
LCLC
lassembled together,
Ithe global SET
Isymbol value is set
Iby the first macro,
Ibut is not altered
Iby subsequent
L__________ _________ Imacros.)
____________________ J
~

Figure 3.18.

Format of Globals and Locals

A global instruction (GBLA, GBLB, GBLC)
defines one or more operands as names
associated with arithmetic, binary, or
character data. These operands are called
global SETA, SETB, or SETC symbols. If a
global instruction is part of a macro
definition, i t must immediately follow the
prototype statement or another GBLA, GBLB,
or GBLC. The global definition of these
operands indicates that each SET symbol is
defined both inside and outside the
defining macro and enables communication
between discrete macros.
A local instruction (LCLA, LCLB, LCLC)
defines one or more operands as names
associated with arithmetic, binary, or
180

DOS System Programmer's Guide

The SETA instruction in the operand
entry is evaluated as a signed 32-bit
arithmetic value that is assigned to the
SETA symbol in the name entry. Figure 3.19
shows the format of the SETA instruction.
r------T--------~------------------------,

I Name I Operation I Operand
I
~------+---------+------------------------~
I A SETA I SETA
lone term, or
I
I symbol I
Ian arithmetic
I
I
I
lexpression, not
I
Iless than -23~ nor
I
I
I
than +23~-1.
I ______ I _________ Igreater
L
________________________
JI
~

Figure 3.19.

Format of SETA Instruction

The expression may consist of one term
or an arithmetic combination of terms, the
minimum and maximum values of which are
_23~ and +23~-1, respectively.
The
arithmetic value assigned to a SETA symbol
is substituted for the SETA symbol when it
is used in an arithmetic expres~ion.
The SETB instruction may assign the
binary value 0 or 1 to a SETB symbol.
Figure 3.20 illustrates the format of this
instruction.

r------T---------T------------------------,
I Operation I Operand
I

I Name

~------+---------+------------------------~

IA SETBISETB
IA 0 or a 1, (0) or (1), I
I symbol I
lor a logical expression I
L
________________________
JI
I ______ I _________ Iwithin
parentheses
~

Figure 3.20.

Format of SETB Instruction

The operand may contain a 0 or a 1 or a
logical expression enclosed in parentheses.
No explicit binary zeros or ones are
allowed in parentheses other than in the
form (O) or (1). A logical expression is
evaluated to determine if it is true or
false. The SETB symbol in the name entry
is then assigned the binary value 1 or 0

corresponding to true or false,
respectively.
The following are valid operand fields
of SETB instructions:
(&AREA+2 GT 29)
(T'&T02 EQ 'C')
The following are invalid:
&B (not enclosed in parentheses)
(T'&P12 EQ IF' &B) two terms in succession
The SETC instruction assigns a character
value to a SETC symbol. Figure 3.21 shows
the format of a SETC instruction.

r------T---------T------------------------,
IOperationloperand
I

I Name

~------+---------+------------------------~

I A SETC I SErC
lOne operand of
I
I symbol I
I the type attribute,
I
I
I
I character expression
I
I
I
lor a substring
I
I
I
I notation. A SETA
I
I
I
I symbol may appear
I
IL______ I _________ I here.
________________________ JI
~

Figure 3.21.

Format of SETC Instruction

The character value assigned to a SETC
symbol may be a type attribute. If the
type attribute is used, it must appear
alone in the operand field.
A character expression usually appears
in the operand field. A character
expression consists of any combination of
characters enclosed in apostrophes. The
maximum length of a character expression is
127 characters. The character value
enclosed in apostrophes in the operand
field is assigned to the SETC symbol in the
name entry. The maximum length character
value that can be assigned to a SETC symbol
is eight characters. If a value greater
than 8 is specified, the leftmost 8
characters are used.

system Variables

~ I

system variable symbols are assigned values
automatically by the assembler. There are
four system variable symbols:
&SYSNDX,
&SYSPARM, &SYSECT, and &SYSLIST. System
variable symbols may be used in the name,
operation, and operand entries of
statements in macro definitions, but not in
statements outside of macro definitions
with the exception of &SYSPARM. They may
not be defined as symbolic parameters or

SET symbols, nor may they be assigned
values by SETA, SETB, and SErC
instructions.
The &SYSNDX symbol is assigned the
four-digit nwUber 0001 for the first macro
instruction processed by the assembler, and
it is incremented by one for each
subsequent inner and outer macro
instruction processed. &SYSNDX may be
combined with other characters to create
unique names for statements generated from
the same model statement. The &SYSECT
symbol carries a character value that is
the name of the last START, CSECT, or DSECT
statement encountered before the expansion
of the USING macro.
The &SYSPARM is specified in theSTDJC
macro at system generation time.
&SYSPARM
allows the user to control conditional
assembly flow and source code generated
through the use of the parameter specified
in the job control OPTION statement.
&SYSPARM acts as a global SETC, except its
value is set by the job control OPTION
statement.
If no named CSECT, DSECT, or START
statements occur before a macro
instruction, &SYSECT is assigned a null
character value for that macro instruction.
The &SYSLIST symbol (not available in
keyword macro definitions) is the symbol
reference for the entire macro instruction
operand field. This symbol refers to the
nth macro instruction operand. If the nth
operand is a sublist, then &SYSLIST(n,m)
may refer to the mth operand in the
sublist, where nand m may be any
arithmetic expressions allowed in the
operand field of a SETA statement.

CONCATENATION
Concatenation is defined as a linking
together in a series or chain; a process of
linking or joining together in a sequence,
with a specified order.
If a symbolic parameter in a model
statement is immediately preceded or
followed by other characters or another
symbolic parameter, the characters that
correspond to the symbolic parameter are
combined, in the order given in the
generated statement, with the other
characters or the characters that
correspond to the other symbolic parameter.
This process is called concatenation. When
a symbolic parameter is concatenated with
any following character value, the extent
of the symbol must be defined (delimited).
If the first character of the following
character value is not a recognized
Section 3: Program Design

181

delimiter, a special delimiter character (a
period '.'), must be used when the first
character is a letter, digit, left
parenthesis or a period. A period is
optional when the first character is an
ampersand (&). See Figure 3.22.

Vary the number and sequence of
generated statements.

These conditional assembly instructions
give true flexibility to the macro
definition language.

r----~---------T----------------,

IName I Operation I Operand
I
~-----+---------+----------------~
Header
I
I MACRO
I
I
Prototype I &NAME I MOVE
I&P,&S,&R1,&R2
I
Model
I&NAMEIST
I&R1,&S.(&R2)
I
Model
I
IL
I&R1,&P.B
I
Model
liST
I&R1,&P.A
I
Trailer I
IL
I&R1,&S.(&R2)
I
Macro

I MEND

There are 13 conditional assembly
instructions: LCLA, LCLB, LCLC, GBLA,
GBLB, GBLC, SETA, SETB, and SETC, that are
discussed under SET Symbols, and AlF, AGO,
ACTR, and ANOP that are discussed in this
section.

~-----+---------+----------------~
IHERE IMOVE
IFIELD,SAVE,2,4 I

GeneratedlHERE 1ST
12,SAVE(4)
I
Generated I
IL
12,FIELDB
I
Generated I
1ST
12,FIELDA
I
Generated IL_____ IL
_________ 12,SAVE(4)
________________ JI
Figure 3.22.

Concatenation and Generated
Coding

SEQUENCE SYMBOLS
The name entry of a statement may contain a
sequence symbol that provides the
programmer with the ability to vary the
sequence in which the assembler processes
statements. These symbols are never
variables. They name a branch point in the
definition and consist of a period followed
by a letter and seven letters and/or
digits.
A sequence symbol in the operand entry
of an AIF or AGO statement (see Conditional
Assembly Statements) references the
statement named by the sequence symbol. It
can be used in the name entry of any
statement that does not contain a symbol or
SET symbol, except a prototype statement,
or a MACRO, LeLA, LCLB, LCLe, GBLA, GBLB,
GBLC, ACTR, ICTL, ISEQ, or COpy
instruction. For example, • READER, • LOOP2,
.A23456 and .X4F2 are valid sequence
symbols.

AIF--Conditional Branch
The AIF instruction alters conditionally
the sequence in which source program
statements are processed by the assembler.
The conditional branch is located within or
outside of the macro definition. If the
logical expression in the operand field is
true, the macro generator branches to the
sequence symbol following the logical
expression. Figure 3.23 illustrates the
typical form of this instruction.

r----------y---------y--------------------,

I Name
I Operation I Operand
I
~----------+---------+--------------------~
IA sequencelAIF
IA logical expression I
Isymbol or I
lenclosed in paren- I
Inot used I
Itheses, immediately I
I
I
Ifollowed by a
I
L
____________________
JI
I __________ I ________ Isequence
symbol.
~

Figure 3.23.

Conditional Branch
Instruction

The following are valid operands of AIF
instructions:
(&AREA+X'2D' GT 29).READER
(&NAME+FlVE GT 2).POSSIBLE
The following are invalid operands of
AIF instructions:

CONDITIONAL ASSEMBLY INSTRUCTIONS
The conditional assembly instructions allow
the programmer to:
1.

182

Define and assign values to SET
symbols that can be used to vary parts
of generated statements and
DOS System Programmer's Guide

(T'&ABC NE T'&XYZ)

No sequence symbol

(T'&ABC NE T'&XYZ).X4F2
Blank between
logical expression
and sequence symbol

.X4F2

No logical expression

AGO--Unconditional Branch
The AGO instruction causes an unconditional
branch to the sequence symbol in the
operand. Figure 3.24 illustrates the
typical form of this instruction.

r----------T---------T--------------------,
I Name
I Operation I Operand
I
~---------_+---------+--------------------f
IA sequencelAGO
IA sequence symbol
I
Isymbol or I
I
I
IL__________
not used I _________ I ____________________ JI
~

Figure 3.24.

to be branched to. Figure 3.25 illustrates
the typical form of this instruction.

r----------T---------T--------------------,
I Operation I operand ,
I

I Name

~----------+---------+-------------------.~

IA sequencelANOP
INot used, must not I
___________________
JI
IL__________
symbol
I _________ Ibe
present.
~

Figure 3.25.

Assembly No Operation
Instruction

Unconditional Branch
Instruction

The statement named by the sequence
symbol in the operand is the next statement
processed by the assembler. The statement
named by the sequence symbol may precede or
follow the AGO instruction.

If the programmer wants to use an AIF or
AGO instruction and has already entered a
symbol or variable symbol in the name entry
of the statement to which he wishes to
branch, he cannot place a sequence symbol
in the name entry. An ANOP instruction can
be placed before that instruction, and then
branched to. This has the same effect as
branching to the statement immediately
after the ANOP statement.

ACTR--Conditional Assembly LOop Counter
The ACTR limits the number of AGO and AIF
branches executed within a macro
definition. When used, the ACTR must
appear after the globals and locals symbol
definition statements and before any other
type of model statement. The ACTR
instruction assigns a maximum count to the
number of AGO and AIF branches executed
within the macro definition. When the
count reaches zero, an END card is
generated. If the count is zero before
decrementing, the assembler takes one of
two actions:
1.

If a macro definition is being
processed, the processing of it and
any nested macros above it is
terminated, and the next statement in
the main portion of the program is
processed.
If the main portion of the program is
being processed, conditional assembly
is terminated, and the portion of the
program generated so far is assembled.
If an ACTR statement is not given, the
assumed value of the counter is 150.

EXTENDED CAPABILITIES
The macro language provides additional
features that allows the system to:
1.

Terminate processing of macro
definition

Generate error messages

Define global SET symbols (discussed
in the section SET symbols)

Use system variable symbols (discussed
under System Variable Symbols)

prepare keyword and mixed-mode macro
definitions and write keyword and
mixed-mode macro instructions
(discussed under Macro Instruct~on
Formats)

MEXIT -- Macro Definition Exit
ANOP--Assembly No Operation
The ANOP instruction facilitates branching
to a statement that has a symbol or
variable symbol in the name field. The
ANOP instruction causes no operation and is
inserted immediately before the statement

The MEXIT instruction allows exit from the
macro definition at various points in the
definition. However, when this instruction
terminates the macro definition, it does
not signify the physical -end of the
definition. Figure 3.26 illustrates the
typical form of the instruction.
Section 3: Program Design

183

r----------T---------T--------------------,
I Operation I Operand
I

I Name

~----------+---------+--------------------~
IA sequencelMEXIT
INot used, must not I
Isymbol or I
Ibe present.
I
__________
IL not
used I _________ I ______ _____________ JI
~

Figure 3.26.

Macro Definition Exit
Instruction

MNOTE statement
The MNOTE instruction may generate a
message and indicate the level of severity
of the error. The severity code is for the
programmer's information only and is not
used by the DOS assembler or co~trol
program. This instruction requests a
message to be printed on the output
listing. Figure 3.21 illustrates the
typical form of this instruction.
r----------T---------~-------------------,

I Name
I Operation I Operand
I
~----------+---------+--------------------~
IA sequencelMNOTE
Isee examples in text I
I symbol or I
I
I
__________
ILnot
used I _________ I ____________________ JI
Figure 3.21.

184

severity code, 'message'

, 'message'

'message'

MEXIT should not be confused with MEND.
MEND indicates the physical end of the
macro definition. MEND must be the last
statement of every macro definition,
including those that contain one or more
MEXIT instructions.

The operand entry of the MNOTE
instruction may be written in one of the
following forms:

MNOTE Instruction

DOS System Programmer's Guide

For two and three, the severity code is
assumed as one.
The MNOTE statement appears in the
listing with a statement number at the
point where it was generated. Because the
message portion of the MNOTE operand is
enclosed in apostrophes, two apostrophes
must be used to represent a single
apostrophe. Two ampersands must be used to
represent a single ampersand that is not
part of a variable symbol (see Figure
3.28).
r----T------~----------------------------,

I NamelOpera-loperand

I
Ition I
I
~----+------+-----------------------------~
L
______ 11,'ERROR**--NOT
______ ______________________
I ____ IMNOTE
RECOVERABLE' JI
~

Figure 3.28.

Sample MNOTE

Figures 3.29 and 3.30 are examples of
macro writing and usage. Figure 3.29
defines the MSG ~acro. There are no
locally defined symbols. All are globals.
Figure 3.30 is an example using the MSG
macro to write an appropria'te message
pertaining to a particular step of the
routine. The generated coding of the MSG
macro is indicated by a '+' after the
statement number.

PAGE
STMT

SOURCE STATEMENT

2 &NAME
3
4
5
6
7
8
9

10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32

.BADMSG
.SETREP
&REPGLB
.INORDER
&REPGLB
.BEGIN
..
&NAME
.C
&MSGLTH

&NAME.A
.D
.B

Figure 3.29.

MACRO
MSG
GBLA
GBLC
GBLB
AIF
AIF
AIF
MNOTE
MEXIT
MNOTE
MEXIT
ANOP
SETB
AGO
ANOP
SETB
ANOP
MSG MACRO

DOS CL3-4 06/09/69
MSG00010
MSG00020

&A~&B

&MSGLTH
&LATBR

PROTOTYPE

&REPGLB~&MSGGLB

('&A' EQ ").BADMSG
('&B' EQ 'REPLY').SETREP
('&B' EQ ").INORDER
1~'INVALID REPLY OPERAND~ I QUIT'
1~'NO

MESSAGE

CODED~

MSG00030
MSG00040
MSG00050
MSG00060
MSG00070
MSG00080
MSG00090

I QUIT'
MSG00110
MSG00120
MSG00130
MSG00140
MSG00150
MSG00160
MSG00170

.BEGIN

o
CHANGE LEVEL 2-0

15~=V(MSGRTN)

AIF
BAL
ANOP
SETA
DC
DC
AIF
DC
ANOP
DS
MEXIT
BAL
AGO
MEND

(&REPGLB).B
14~4(15)
NON-REPLY HANDLER
K'&A-2
FL1'&MSGLTH.' MESSAGE LENGTH
C&A MESSAGE
(NOT &REPGLB).D
C' , ANSWER BYTE
OH

RESTORE BOUNDARY

14~0(15)

REPLY HANDLER

Sample MSG Macro

Section 3: Program Design

185

STMT

SOURCE STATEMENT

DOS CL3-4

1176 ~
1177 xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
117g ~
1179 ~
THE FOLLOWING ROUTINE CHECKS THE RETURN CODE AFTER A
1180 ~
READ,WRITE,CONTROL COMMAND IS GIVEN
1181 ~
REGISTER EIGHT CONTAINS THE ADDRESS OF ROUTINE WHICH
1182 ~
JUST GAVE THE READ OR WRITE COMMAND. IF THE COMPLETION
1183 ~
IS BAD REG 8 IS USED TO TURN THIS DEVICE OFF CNO MORE
1184 ~
OPERATIONS WILL OCCUR ON THIS DEVICE),THE OPERATOR MUST
1185 ~
RELOAD THE PROGRAM INORDER TO RESTART THIS DEVICE
1186 ~
1187
1188 ~

1190 CKCONDC LTR
RF,RF
CK CONDITION CODE
1191
BNZ
CONTCK
1192
BR
R2
EXIT BACK CK GOOD
1193 CONTCK
MSG
'ERROR OCCURRED ON CKING RETURN CODE'
1194+x
MSG MACRO CHANGE LEVEL 2-0
1195+CONTCK
L
15,=VCMSGRTN)
1196+
BAL
14,4CI5) NON-REPLY HANDLER
1197+
DC
FLl'35' MESSAGE LENGTH
1198+
DC
C'ERROR OCCURRED ON CKING RETURN CODE' MESSAGE
1199+
DC
OH RESTORE BOUNDARY
1200
LR
R3,RF
SWITCH REGS
1201
BAL
R2,FORMAT
GO FORMAT CONDITION CODE
1202
STC
R3,ERRMSG+28
INSERT CHAR
1203
LR
R3 J RF
1204
BAL
R2,FORMATI
GO FORMAT SECOND CHAR
1205
STC
R3,ERRMSG+27
INSERT SECOND CHAR
1206 ERRMSG
MSG
'CONDITION CODE =
1207+ x
MSG MACRO CHANGE LEVEL 2-0
1208+ERRMSG
L
15,=VCMSGRTN)
1209+
BAl
14,4CI5) NON-REPLY HANDLER
1210+
DC
FLl'21' MESSAGE LENGTH
, MESSAGE
1211+
DC
C'CONDITION CODE =
1212+
DS
OH RESTORE BOUNDARY
1213 X
MVC
MSG2+27(7),OPERATIN
1214 MSG2
MSG
'LAST OPERATION 1215+ x
MSG MACRO CHANGE LEVEL 2-0
1216+MSG2
L
15,=VCMSGRTN)
BAL
14,4CI5) NON-REPLY HANDLER
1217+
1218+
DC
FLl'36' MESSAGE LENGTH
1219+
DC
C'LAST OPERATION , MESSAGE
1220+
DS
OH RESTORE BOUNDARY
1221 PDUMP
PDUMP START END
1222+;: 360N-CL-453 PDUMP' CHANGE LEVEL 3-0
Figure 3.30.

186

Sample MSG Coding

DOS System programmer's Guide

Section 4:

Debugging Aids

Section Outline
Gathering Documentation
.189
System Action Under Cancel
• • 198
Wait States • • • • • • • •
• .202
50ft Waits • • • • • • • • •
• .202
Hard Waits. • • • • • • • •
• .202
Debugging Assembler Programs
• .204
Debugging COBOL Programs • • • • • • • • 204
How to Use a Dump
.205
Locating a DTF • • • • •
• .205
Locating Data • • • • •
• • • 206
Debugging FORTRAN Programs •
.234
Debugging RPG Programs. •
.243
Halt Analysis • • • • • • • • • • • • 243
Causes of a Halt Zero Condition • • • 243
Debugging PL/I Programs • • • •
.255
Summary of PL/I Debugging Aids • • • • 263
Handling Compile Time Aborts. •
.264

Section 4: Debugging Aids

187

Section Figures

Figure 4.1. SDR Communications Region
(Part 1 of 2)
• • • • • • • • • • • • .190
Figure 4.2. Machine Check Recording
and Recovery (MCRR) Linkage Table
.192
Figure 4.3. RMS Linkage Area (RASLINK) 193
Figure 4.4. First Part of Program
Information Block (PIB) Table
• .195
Figure 4.5. PIB Flag Expansions • • • • 196
Figure 4.6. Second Part of Program
Information Block (PIB) Table • • • • • 197
Figure 4.7. Causes for Message OS04I
(Cancel Code X'21') • • • • • • • • • • 200
Figure 4.8. Low Core E.rror Bytes
.203
Figure 4.9. COBOL Sample Program
(Part 1 of 27)
............207
Figure 4.10. FORTRAN Sample Program
(Part 1 of 8)
• • • • • • • • • • • • • 235
Figure 4.11. Using RPG Pointers to DTF 243
Figure 4.12. Halt Indicator (HO)
Analysis Aid
•••••••••
.244
Figure 4.13. RPG Sample Program (Part
1 of 10)
• • • • • • • • • • • • • • • 245
Figure 4.14. PL/I Program Structure • • 255
Figure 4.15. PL/I Storage Areas
• • • 256
Figure 4.16. Entry Point Table
• • 257
Figure 4.17. Object Time Core Usage
.258
Figure 4.18. Library Work Space • • • • 259
Figure 4.19. Communications Area
switches
• • • • • • • • • • • • • • • 260
Figure 4.20. Dummy DSA and DSA Layout 260
Figure 4.21. Block Description
.261
Figure 4.22. DSA Chaining • • • • • • • 262
Figure 4.23. PL/I Consecutive File
DTF-A Appendage • • • • • • • • • • • • 263
Figure 4.24. PL/I Regional File DTF-A .263
Figure 4.25. PL/I Sample Program
(Part 1 of 18) • • • • • •
• • • • .265

188

DOS system Programmer's Guide

This section presents debugging aids
considered helpful to both the application
and system programmer. Information
contained in this section includes:
• System action on all cancel conditions
• Register conventions for following
program flow
• When a storage print is useful
• The types of documentation used in
locating program problems
• The action taken when a hard wait or
unending loop is encountered
• The importance of low-core messages and
system error messages as an aid in
determining a starting point for
approaching a programming problem.
Sample programs in COBOL, PL/I, FORTRAN,
and RPG, together with their respective
linkage editor maps to show how to locate
programs and partition save areas in main
storage are included.

Gathering Documentation
This is an explanation of the types of
documentation useful in debugging problem
programs. All SYSLOG and SYSLST error
messages, program listings, the supervisor
listing, linkage editor map, and a core
dump should be gathered.
A system dump of main storage should be
available, but if the system is in a hard
wait or an unending loop, a stand-alone
(self-loading) dump will have to be taken.

The DOS stand-alone dump generator,
DUMPGEN, produces a stand-alone dump
program tailored to system requirements.
The dump can either be a conventional dump
program or a formatting dump program. A
DSERV of the core image library directory
may be helpful when program checks occur in
the logical transient area. The directory
can be displayed by an alphamerically
sorted listing of the directory entries, or
a listing of the entries in the order they
appear in the directory.
The label cylinder display program
(LSERV) can also be used for error
analysis. LSERV displays the TLBL and the
DLBL and EXTENT information contained on
the SYSRES label cylinder.
Information
about secured data files is not displayed.
DOS I/O error logging, MCRR (Machine
Check Recording and Recovery), RMS
(Recovery Management Support), and the DUMP
option of job control are additional
facilities for error analysis. The RMS
consists of two functions:
MCAR (Machine
Check Analysis and Recording), and CCH
(Channel Check Handler).
Other facilities for error analysis are
EREP (Environmental Recording, Editing, and
Printing Program), ESTVUT and ESTVFMT
(Error Statistics by Tape Volume Utility
programs). EREP edits and prints data that
has been stored in the recorder file
(SYSREC) by the I/O error logging and/or
MCRR and/or MCAR/CCH functions. For the
IBM system/370, EREP creates and maintains
a history tape, and, if specified at system
generation time, RDE (Reliability Data
Extractor) of OBR/MCAR/CCH and, if
specified by the ROD command, IPL/EOD (End
of Day) data. Figures 4.1, 4.2, and 4.3
show the SDR communications region, the
MCRR linkage table, and the RMS linkage
area.

Section 4: Debugging Aids

189

0
SDK
Flags

(SDRTABLE)
32

10 11

3 4
Number
of SDR
Records

Partition
ID

First SDR ID
BBCCHHR

Address of
SDR
Accumulator

First OBR ID
BBCCHHR

Address of
SDR Unit
Switches

Mask
Bytes

71
List Save Area

103 104

95 96
Test Under
Mask Table

SDR 1 Work Area

SDR2 Work Area

107 108

155 156
Area Modified by A - Transients

111 112

Test Under
Mask
Instructi on

Temporary
Work Area

135 136

118

31
Last 0 BR ID
BBCCHHR

Current OBR ID
BBCCHHR

Reserved

75 76

24 25

43 44

39 40

35 36

17 18

163 164

Branch
Instruction

117

SDR Queue
Save Area

F'65536'

159 160

SDR Error
Message
Save Area

115 116

167

OBR/SDR
Flag Byte
Address

168

250
Data Area for OBR/SDR Records

Key to SDR Communications Region Displacements:

SDR Flags:
Bit 0:
1:
2:
3:

Key of OBR
RDE option
Initial IPL time
RF option = NO, recording is suppressed

Bit 4:
5:
6:
7:

RF option = CREA TE
RF option = YES
Error while recording
Recorder fi Ie ready

Set and tested by Job Control ..

Set by EREP transient SSBSDRUP to identify the partition making the call for EREP recording.
Settings: X'10'
X '20'
X'30'
X'Ol'
X '00'

if EREP is running in BG.
if EREP is running in F2.
if EREP is running in Fl.
with one of the above if recorder file is ready.
with one of the above if recorder fi Ie is not ready.

Initial number of SDR records specified. If SDR record count is not specified, the file is formatted for OBR records only
(SJOBCTLM, see IPL and Job Control PLM, GY24-5086).
Disk address of first SDR record.
Disk address of first OBR record.
Disk address of current OBR record.
Disk address of last OBR record.

Figure 4.1.

190

SDR Communications Region (Part 1 of 2)

DOS system Programmer's Guide

Key to SDR Communications Region Displacements:

Address of SDR accumulator area which contains half - byte coutlTers and accumulated error conditions.

Address of SDR unit switches.
SDR switch byte (1 for each PUB):
X'80' - Update operations complete
X'40' - Counters on external file overflowed
X'20' - I/O error during write
X'08' - SDR update half - byte counters routine required
X'04' - Update SDR record routine required
Other - Reserved
When entry contains X '01000000', indicates MCRR, no SDR supported.

Reserved.

SDR1 register save area.

Mask formats for interpretive error accumulator, SDR1:
X'FF' X'FE' X'FD' X'FC' Other -

End of update
Bypass counter
Set up 'OR' condition to previous counter
Ignore list item
Test bit in error queue

Used by the interpretive error accumulator routine to process list passed by OBR/SDR A -transient.

Used by the interpretive error accumulator routine.

11041 Used by the interpretive error accumulator routine for address alignment.

Executed by the interpretive error accumulator routine.

Loop counter for the SDR counter update.

§]

Save area for pointers to entries in the SDR error queue.

Work area where half byte error counters are unpacked and updated.

list of devices passed to the SDR processor from $$AN ERAD .

11561 Used by SDR/OBR recorder phases to pass error message displacements and disk error addresses in event of
1160

I Entry point from OBR/SDR A - transients.

0'1

error.

Branches to label SDRMM.

11641 Pointer into the OBR/SDR unit switches. Status posted by recorder phases. (See byte 36).

OBR and SDR records formatted by the recorder phases.

Figure 4.1.

SDR Communications Region (Part 2 of 2)

Section 4: Debugging Aids

191

MCRRPSW1 (See Note)

o (Hexadecimal
o (Decimal

Displacement)

MCRR: PSW
I Reentrant
Address of
I MCRR Routine

I
I

XXXXXXXX

16
MCRR

PSW

Address of

I MCRR Routine
I

XXXXXXXX

Address of
Channel
Failure
Routine

Address of
Machine
Check
Routine

XXXX

Key to displacement:
Machine Check Recording and Recovery PSW. Loaded to enable machine check interrupts. Second word (displacement 4-7)
contains reentrant address (MCRETURN) to MCRR routine.
Machine Check Recording and Recovery PSW. Loaded to enable machine check interrupts. Second word (displacement 12-15)
contains initial address (MCRRRTN) of the MCRR routine.
Address of channel failure routine (MACHEKl).
Address of machine check routine (MACHEK).
Note: MCRRPSW1 is the label of the first byte of the MCRR Linkage Table.

Figure 4.2.

192

Machine Check Recording and Recovery (MCRR) Linkage Table

DOS System Programmer's Guide

RASLINK

o (Decimal Displacement)

CPUID

RASDMC

RASFLAGS

MCFLAGS

RASMODEL

RASTABA

RASBASE

CPU ID field

Damaged
Channel
byte

RAS flag
byte

Machine
Check
flags

CPU
Model

RAS Table
(RASTAB)
address

Base add ress
for RAS
Monitor

XXXXXXXX

XXXX

Key to RAS Linkage Area displacements:
CPU ID field.
Address of damaged channel, or X 'FF' if no channel damaged.
RAS Flag byte:

bit

flag

description

6
7

X'80'
X'40'
X'20'
X'lO'
X'08'
X'04'
X'02'
X'Ol'

RAS active
RAS SIO flag
RTA in control
RAS I/O delayed
Channel check on error SIO
Reserved
Channel check ~n SIO
I/O active for SIO

bit

flag

description

X'04'
X'02'
X'Ol'

Reserved
Hard machine check
All machine records built
All channel check records built

1
2
3
4

Machine Check Flags:

0- 4
5
6
7

G
I Figure

Largest CPU Model.
Address of RAS Table (RASTAB).
Address used for base register iii RAS Monitor Program.

4.3.

RMS Linkage Area (RASLINK)

Section 4: Debugging Aids

193

Further documentation can be obtained by
executing the PDAID program, which records
(traces) certain events and either writes
·chem on the I/O device specified or
maintains them in the CE area (or alternate
address area). The program can trace:
1.

fetching or loading of programs or
phases (Fetch/Load Trace).

input/output activity (I/O Trace).

supervisor calls (GSVC Trace).

QTAM input/output activity (QTAM
Trace).

The On-Line Test Executive Program
(OLTEP), together with the On-Line Tests
(OLTS), make up the On-Line Test System,
which tests I/O devices with minimum
interference to other programs running on
the system. RETAIN/370 is an OLTEP
function that allows the OLTEP programs to
be executed on the system/370 from a remote
location. RETAIN/370 is yet another
problem determination tool. See the DOS
OLTEP listed in the Preface for a detailed
description.
The facilities mentioned form the DOS
problem determination aids. Problem
determination is a process or a procedure
for<· determining the cause of an error. The
DOS Messages listed in the Preface
recommends a specific procedure to follow
when an error condition occurs; the DOS
system Control and service gives a detailed
explanation of problem determination aids.
Error messages are very important and
can supply useful information in
determining where to start looking for the

194

DOS System Programmer's Guide

trouble (i.e., in what partition the
failure tock place). A message number or
code is supplied to give further
information about the error. The program
check message gives the location of the
failing operation code and the condition
code from the program status word (PSW).
This gives you a starting point for
reconstructing the cause of the error.
The program listing is an extremely
useful tool in determining if the error
condition was caused by a logic error or a
particular condition that you had not
considered when the program was written.
The supervisor (SUPVR) listing allows
you to check facts at the time of failure
and to determine if the error indications
were valid. The listing and the main
storage dump allow you to locate the
Program Interrupt Key (PIK) to determine
the task in control of the system at the
time of failure.
By locating the Program Information
Block (PIB) table, you can easily locate
programs in main storage. See Figures 4.4,
4.5, and 4.6 for a description of the PIB
table. Using the information in the PIB
table, you can check the cancel code and
find the partition save area address. The
partition save area supplies. you with such
useful information as the PSW and register
values that you can use to locate the last
instruction executed.

The system communications region (Figure
1.7) within the supervisor contains the
address of the PIB table, and other useful
information for determining the nature of
the error.

PIB TABLE
Byte
Number

All
Bound
PIB

= 16

Byte
Length

Problem
Program
PIB
(Note 1)

Cancel
Code

Attention
PIB

NOP
SYSLOG ID
Instruct(BG, F2, or Fl)
ion
(CR)

SYSLOG ID
(AR)

Branch
Code
(BC)

Partition Save Area

Active=Address of
Save Area
Inactive = Remainder of
BC Instruction

Scratch
Byte
X'OQ'

Cancel
Code

Quiesce
PIB

Cancel
Code

Supervisor
PIB

Subtask
PIB for AP
Note 3)

Cancel
Code

Switch
Byte
See F

General Exit Routine

NOP
SYSLOG ID
Instruc(BG, F2, or Fl) tion

Address of the
Save Area

Logical Transient Bucket
(contains save
area address)

Address of
the Logical
Transient

nnel PU
X'OO'

Address of
SYSRES PUB

X'04'

X'08'

Length of Error
Queue Entry

Number
Core
Blocks
(Note 2)

X'OC'

X'lO'

X'14'

X'18'

Constants to Clear Bytes
2 -5 of CCB
X'IF'

X'05'

X'OO'

PIB
Assign
Flag
See D

User
LUB
Index

of
LUBs

Flag
Byte
See C

Note 1: Three problem program PIBs are built in this sequence when the MPS or BJF feature is selected as a generation option:

Background PI B
Foreground 2 PIB

Foreground 1 PIB
When a batch-only environment is established at generation time, the All Baund and Foreground PIBs are excluded from the table, and only
one (BG) problem program PIB is built. However, the X'20' bytes that F2 and Fl PIBs normally occupy (between PISSG and PIBAR) are
filled with 32 bytes of DIBs data.
Note 2:

Number is in multiples of 2K for F2 and Fl. BG is always 10K (X' OA').

Note 3: Total of nine subtask PISs are generated, and only when AP is specified at generation time.

* See Figure 4.5 for flag byte expansions A, B, C, D, E and F.
Bytes 90 and 91 (X'5A'- '5B') of the communications region contain the address of the first part of the PIB Table. Label PIBTAB identifies the first
byte of the table.

E'igure 4.4.

First Part of Program Information Block (PIB) Table

Section 4: Debugging Aids

195

[ ; ] Supervisor, Quiesce, and ALL Bound PIB Flags:
Bit 0:
1-4
5
6
7

1=
0=
1=
1=
0=
1=
0=

Always one
Always zero
Always one
Active
Inactive
Active
Inactive

Note: IfPTO=YES is specified, Bit 6 is a one in the
Quiesce I/o PIB when attached by the supervisor. Otherwise it is always zero.
Problem Program PIB Flag (First Byte in PIB):

PIB Assign Flag
X '80' = SYSRES DASD fi Ie protect inhibited (allow write
operation on SYSRES)
X'40' = Channel appendage exit allowed (BTAM)
X'20' = Cancel in progress (used in terminator function)
X' 10' = Cancel control (set on a foreground cancel)
X '08' = Hold-Release flag for foreground assignments
X '07' = Supervisor or Attention routine PIB assign flag setting
X'04' = Background program PIB assign flag setting
X' 02' = Foreground 1 program PI B assign flag setti ng
X 'a I' = Foreground 2 program PIB assign flag setting
Attention PIB Flag
Bit 0:

Bit 0:

1 = Registers stored
o = Registers not stored
0 = Always zero
1-3
4
1 == QTAM Wait active
0== QTAM Wait inactive
5
0 = Normal execution
1 == Program has 'seized the system
6
1 == Unbound
0= SVC 2-bound CB-transient in progress)
7
1 = Unbound
0= SVC 7-bound (waiting for an I/O interrupt)
X'80' indicates the program is not present in the system
X'87' indicates the program is PTO bound
X'89' indicates the program is IDRA bound
Problem Program PIB Flag (Last Byte in PIB):
Bit 0:
1:
2:
3-4:
5:
6:
7:

1 == Batched Job in Foreground

a == No BJF

Cancel in LTA and Device not Assigned
1 == /& on SYSIN if DASD '
a = No /& on SYSIN
Reserved
1 = Task is cancelled
a == Task not cancelling
1 = Subtask (s) attached
a = No subtasks attached
1 = In AB Routine
a -= Not in AB Routine

Figure 4.5.

196

PIB Flag Expansions

DOS System Programmer's Guide

1-5
6
7

1 == Registers stored

a = Registers not stored

a = Always
1=
a=
1=
a=

zero
Attention routine active
Attention routine SVC 2-bound
Active
SVC 7-bound

X'80' indicates the attention routine is not present in the system.
X'89' indicates the program is IDRA bound
Attention PIB Switch Byte
Bit 0-2: Reserved
3: 1 == PTAFTCH (Fetch $$ANERRY, Z, or 0) Switch ON
0= PTAFTCH (Fetch $$ANERRY, Z, or 0) Switch OFF
4: 1 = Detach Logical Attention Routine ($$BA TTNA)
Switch ON
0= Detach Logical Attention Routine ($$BATTNA)
Switch OFF
5: 1 == Physical Attention Recall Switch ON
a = Physical Attention Recall Switch OFF
6: 1 = Attention Request Switch ON
a = Attention Request Switch OFF
7: 1 = External Interrupt Request Switch ON
a = External Interrupt Request Switch OFF

Byte
Number

= 16
Byte
Length

All Bound
PIB

Background
PIB

FG2
PIB

System
LUB Index

(Note 1)

FG1
PIB

Address of Termination
ECB, if any, or F'O'

System
LUB Index

(Note 1)
Attention
PIB

Quiesce
PIB

I/o

Supervisor
PIB

ECB Address for

Subtask
PIB
(Note 3)
Note 1. Generated only if MPS is specified.
Note 2 •. Always background communications region except when MPS

= BJF.

Note 3. Total of nine subtasks generated, and only when AP is specified.
1
Note 4. Will be filled in with halfword indicating the relative priority of task in the system (range H'41 to H'15 , the lower the
number the higher the priority).
Bytes 124 and 125 (X'7C-'7D') of the communications region contain the address of the second part of the PIB table. Label
PIB2AD identifies the first byte of the table. The second part of PIB table comes before the first part in storage allocation.

Figure 4.6.

Second Part of Program Information Block (PIB) Table

Section 4: Debugging Aids

197

are detached and EOJ is taken after the
dump is complete.

SYSTEM ACTION UNDER CANCEL
The following lists all cancel codes and
their message prefixes. Some do not appear
in a foreground PIB, such as the X'FF' code
(supervisor catalog failure). This type of
function can be performed only in the
background partition. The linkage editor
and system maintenance functions must also
be performed in the background area.
Byte one of the PIB'table contains a
cancel code stored by the system any time a
cancel condition is encountered. The PIB
table can be located by displaying on the
console the communication region address
(located at X'16'~X'17') plus the
displacement of aX' SA' and X'5B'. This is
the address of the first part of the PIB
table. Remember each entry is 16 decimal
(X'10') bytes in length. Each byte of the
PIB is numbered starting with 0 and
continuing through 15. The layout of the
PLB table and the communications region can
be found in Figures 4.4-4.6 and Figure 1.7.
Cancel Code (Hexadecimal):
Message Code:

Cancel Code (Hexadecimal):
Message Code:

Description, Action or Condition: This is
normal end of job (EOJ). Cancel code X'10'
is posted in byte 1 of the PIB for the
program issuing the SVC 14. The next time
the' canceled program is selected on general
exit, an SVC 2 is taken to call in a
B-transient program, which, in turn, ca~s
job control to perform the end-of-job step.

Description, Action or Condition: This is
caused by the operator responding to an I/O
error message with the cancel option on the
1052.

Cancel Code (Hexadecimal):
Message Code:

Message Code:

OS021

Description, Action or Condition: This is
caused by the main task in a partition
issuing the CANCEL macro without detaching
all subtasks running under its control._

OP731

Message Code:

None

Description, Action or Condition: This is
caused by the main task issuing the DUMP
macro with subtasks attached. It allows
the dump to take place without the error
cancel message being printed. All subtasks

198

DOS System Programmer's Guide

OP821

Description, Action or Condition:
caused by a channel failure.

Cancel Code (Hexadecimal):

This is

05141

Description, Action or Condition: This is
caused by a subtask issuing the CANCEL ALL
macro. This causes all other subtasks to
be detached and canceled. The main task is
canceled, and a dump of the supervisor and
partition involved results.

Cancel Code (Hexadecimal):
Cancel Code (Hexadecimal):

Description, Action or Condition: This is
caused by an I/O error that cannot be
handled by the program (task), thus causing
the program to be canceled. If the DUMP
option is specified at system generation
time, a dump of the supervisor and the
partition in which the program was running
will be taken.

Message Code:
Cancel Code (Hexadecimal):

OP741

Cancel Code (Hexadecimal):

None

Message Code:

05121

Description, Action or Condition: This is
caused when the main task terminates before
all subtasks have been detached. This
indicates the subtasks were canceled before
they came to a normal EOJ. The subtasks
are detached, and the complete partition is
canceled.

Cancel Code (Hexadecimal):
Message Code:

program or task, the program or partition
is canceled. If issued by a subtask, the
subtask only is canceled.

OS13I

Description. Action or Condition: This is
caused by the combination of one task
issuing an enqueue for a resource, and
another task issuing a dequeue for that
same resource. As a result, the previous
owner cannot be identified because register
o in the s~ve area has been modified'.
Cancel Code (Hesadecimal):
Message Code:

Cancel Code (Hexadecimal):
Message Code:

OSOlI

1F
Description. Action or Condition: This is
a result of an operator entering CANCEL
from the 1052.

OP81I

Description. Action or Condition:
caused bya CPU failure.

This is
Cancel Code (Hexadecimal):

Cancel Code (Hexadecimal):
Message Code:

OS03I or OS11I

Description. Action or Condition: This is
caused by a program check interrupt. The
program is canceled by the system. The
user may supply a PC or AB routine to
handle this condition via the STXIT macro.
This code is also used when a routine in
the transient area is canceled due to a
program check in the task or subtask using
it.

Cancel Code (Hexadecimal):
Message Code:

Message Code:

Cancel Code (Hexadecimal):

OS05I or OS06I

Description. Action or Condition: This is
caused by the issuing of a FETCH (SVC 1) or
a LOAD (SVC 4) macro whose phase name
cannot be found. This cancel code is also
used when a logical transient is canceled.

OP77I

Description, Action or Condition: This is
a result of attempting to load a problem
program phase at an address outside main
storage or outside the requester's area
(background or foreground). This condition
also occurs:
1.

if the program requires more main
storage than is allocated to the
partition where the program is to run
or

if an improper address is detected on
an SVC interrupt (i.e., CCW address in
CCB is invalid).

Cancel Code (Hexadecimal):
Message Code:

Cancel Code (Hexadecimal):

OS02I

Description, Action or Condition: This is
caused by a program, task or subtask
issuing a CANCEL m~cro. If issued by a

OP71I

Description. Action or Condition: This is
a result of a program issuing an I/O
request for a logical unit that is not
assigned to a device. If a dump is taken,
general register 1 contains the address of
the CCB. If the CCB is unavailable, the
logical unit message contains SYSxxx.

Cancel Code (Hexadecimal):
Message Code:

Message Code:

OS04I or OS09I

Description. Action or Condition: This can
be caused by many user errors. See Figure
4.7 for a list of the causes.

Message Code:

OP70I

Description, Action or Condition: This is
a result of a program issuing an I/O
request for a logical unit for which there
is no logical unit block (LOB) entry
(invalid LOB code in CCB). If a dump is
taken, general register 1 contains the
address of the CCB.

section 4: Debugging Aids

199

The complete text for message OS041 is:
ILLEGAL SVC - HEX LOCATION nnnnnn - SVC
CODE nn

10.

When nn is 29: A DEQ is issued by a task that did not
ENQ the resource. (This is valid in an AB routine.)

11.

When nn is 2A: A subtask (without an ECB = parameter)
has issued an ENQ macro, or

where nn is in hexadecimal notation.
A subtask has issued an ENQ macro to a resource that
has not been dequeued by another task that has been
terminated, or

This message results from the following causes:
1.

When nn is 02: The phase name given does not start
with $$B, or
For LlOCS, macros called in invalid sequence. As a
result, an SVC 8 is issued after an SVC 2 before an
SVC 9 has been issued to free the transient area, or
For other conditions, the user specified a temporary
exit (SVC 8) for a logical transient. In the temporary
exit routine, another routine is called (by an SVC 2)
before an SVC 9 is issued to free the transi ent area.

12.

When nn is 2D: Emulator execution was attempted,
but the EU parameter of the SUPVR macro was
omitted or incorrectly specified during system
generation.

13.

When nn is 32: For LlOCS:
a.

An imperative macro (such as WRITE or PUT) was
issued to a module that does not contain the
requested function, or

MVCOM macro was issued by a foreground program,
operating under single program initiation.

A PUT was issued for an ISAM retrieve module
without a preceding GET, or

When nn is OA, 12, 13, or 18: The interval timer was
not allocated to this partition, or

An inva Iid ASA fi rst character for the printer
was used, or

The supervisor was generated without the timer option.

A wrong length record indication occurred while
processing 1287 documents when
RECFORM= UNDEF, or

The 1287 program erroneously contained a CCW{s)
with the SLI flag bit ·OFF', or

When nn is 05: The ·to· range specified in the MVCOM
macro is invalid, or

When nn is OB: The call was not given by a logical
transient routine.

When nn is 16,17, or 1A: The caller did not have a
PSW key of zero. This is applicable only in a
multiprogramming system.

A task has issued two ENQ macros to the same resource
without an intervening DEQ.

When nn is 23: More than 16 holds have been issued
for the same track.

When nn i.s 24: Free a non - DASD or a track that is
not held.

When nn is 26: A subtask issued attach, or the save
area is not on a doubleword boundary.

When nn is 27: A main task issued detach without
SA VE = parameter, or

For COBOL, a wrong length record was detected in
the object program.
14.

When nn is any other value: The supervisor function
requested by the operand of the SVC is not defined
for the supervisor being used.

A main task issued detach, but the ID of the subtask in
the save area passed is not valid, or
If a main task attempts to detach an already terminating
subtask.

I Figure

200

4.7.

Causes for Message OS04I (Cancel Code X'2i')

DOS System Programmer's Guide

Cancel Code (Hexadecimal):
Message Code:

None

Description, . Action or Condition:
cancel in progress)

Cancel Code

(Hexadec~al):

Message Code:

Description, Action or Condition: This
occurred when an IBM-supplied component
failed to post a valid cancel code.
(QTAM

OP72I

Description, Action or Condition: This is
a result of a program ignoring the reading
of the /& statement on SYSRDR or SYSIPT.

Cancel Code (Hexadecimal):
Message Code:

OP75I

Description, Action or Condition: This is
a result of the number of pending I/O
errors exceeding supervisor capacity.
Cancel Code

(Hexadec~al):

Message Code:

Cancel Code (Hexadecimal):

OP79I

Description, Action or Condition: This
occurs when a DASD command chain in a
file-protected environment does not start
with a command code of X'07'. This code
indicates a long seek and must be the first
command in the chain.

Cancel Code (Hexadecimal):
Message Code:

OP84I

Description, Action or Condition: This is
caused by an unrecoverable I/O error during
a FETCH of a non-$ phase, thus resulting in
the job being canceled.

Cancel Code (Hexadecimal):
Message Code:

The linkage editor map can be a great
help in locating programs and subroutines
that are included in the programs at object
time. Common areas, load address,
relocation factors, low-core and high-core
addresses are also shown. In addition, the
PHASE card is displayed to show where the
phase was' loaded (i.e., directly following
the supervisor or at some other location).
This map is also helpful when working with
multiphase programs.

OP76I

Description, Action or Condition: This is
caused by DASD file-protect limits being
exceeded or by an incorrect record
reference for system files on disk. It
will also be posted for unrecoverable I/O
errors on tape.

Message Code:

All of these cancel codes cancel the
program, task, or subtask when they occur.
If multitasking is being used and a main
task is canceled, all of the subtasks
attached are detached and canceled as a
result of the main task being canceled,
with the exception of cancel code X'23'.
If a dump option was specified at system
generation time, the contents of the
supervisor and the partition in which the
cancel condition occurred is written on
SYSLST.

OP78I

The system dump of main storage used
with these items allows the programmer to
relate all the information he has gathered
to the contents of main storage at the time
the error occurred. By using the dump and
the listing, the programmer can see how his
program appeared in main storage at the
time of the error. By using the values
found in the PIK and PIB table in the dump,
he 'can see partition save areas, registers,
and instructions to determine what actually
caused the error.
There are times when a system dump is
not available to the programmer, such as
hard waits and unending loops. When one of
these conditions occurs, the only way to
get a dump of main storage is to use a
stand-alone dump. Remember that the
address of the communication region (COMRG)
is lost when a stand-alone dump is taken.
Therefore, bytes X'16'-X'17' should be
displayed before taking a dump of main
storage to ensure that the programmer has
the correct communication region address to
use when he is analyzing the dump. If
bytes X'16'-X'17' are not displayed, the
communications region start address can
still be found by scanning the dump for the
date in the form MM/DD/YY or DD/MM/YY (this
indicates the start of COMRG). Although
the register values in a stand-alone dump
(register print area of the dump) may not
be valid, the partition save area values
most likely will be valid.
Section 4: Debugging Aids

201

Wait States
The system is said to be in a wait state
when the ·wait· light is continuously lit
and the ·system· light is off. wait states
are divided into hard waits and soft waits.
If the system is in a hard wait, the
wait bit in the current PSW (bit 14) is set
to one and the system mask is set to zeros,
thus disabling all interrupts. Because no
interrupts are allowed, a PSW swap cannot
occur and the system mUst be re-IPLed to
continue processing.
A soft wait occurs when the DOS
supervisor finds no in-core programs ready
to run and loads a PSW with the wait bit
set to one and the system mask set to all
ones. The first interrupt returns control
to the supervisor and processing may
continue.
A wait can easily be determined as hard
or soft by causing an interrupt. If the
system responds with some action, the wait
is softi if not, the wait is hard. The
most convenient way for the operator to
cause an interrupt is to press the 1052,
3210, or 3215 request key. If the wait is
soft, the attention routine responds with
the ·READY FOR COMMUNICATIONS· message.

attention of the customer engineer as a
possible hardware failure.) If this
technique does not end the wait, take a
stand-alone dump to find what the system
was waiting for.

HARD WAITS
The DOS supervisor loads a hard-wait PSW
when a failure occurs that puts th~
integrity of the control program or system
data in doubt. The supervisor attempts to
place a message in low core bytes 0-4.
Figure 4.8 shows the explanation for each
error.
If a hard wait occurs, it is imperative
that this message be retrieved and
recorded. Effective diagnosis is extremely
difficult if this step is neglected.
If byte one of main storage contains an
S (X'E2'), the following information can be
obtained easily:
Check byte X'13' for, a X'OP'. This
indicates either a channel control check or
an interface control check. Bytes
X'3A'-X'3B'contain the device address. If
byte X'13' does n9t contain a X'OP', a
machine check must ha~e occurred.

SOFT WAITS

Byte one may have a W. If a W (X'E6')
is found, a hard stop on SYSRES is
indicated.

If the system is in a continuous soft wait,
it is waiting for an interrupt to signal
the completion of an event. Although the
expected interrupt may be from the timer or
external interrupt key, a missing
device-end caused by hardware is the most
frequent cause. The operator can make each
device not-ready, then ready, to generate a
device-end interrupt from each address.
The system light flashes briefly as the
supervisor examines and discards interrupts
for which it was not waiting. The
interrupt from the device waited for causes
normal processing to continue. (The
occurrence should be brought to the

If the CPU d~tects an error in its own
circuitry, or (in theSystem/360, model 50
or smaller) in the channel or interface
control circuits, it forces a machine check
interrupt. The system places anS in byte
1 and enters a hard,wait. The S is a
request to run the SEREP (system
Environmental Recording, Editing, and
Printing) dump to format and display the
contents of the CPU's hardware registers
and log-out area for use by the customer
engineer. (A SEREP dump configured for the
system should be available to th~ operator.
A copy can be obtained from the customer
engineer responsible for the CPU.)

202

DOS System Programmer's Guide

Byte 0

Byte 1

Byte 2

Byte 3

Explanation

SYSTEM/360 SEREP Codes:

X'OO'

X'E2'

Not used

Machine check. Load SEREP. Re -IPL system.

X'Ol'

X'E2'

Reserved

Channel failure: interface or channel control check. Load SEREP. Re -IPL system.

SYSTEM/370 SEREP Codes:
X'Cl'

X'E2'

A, I, S*

Not used

Unrecoverable machine check.

X'C2'

X'E2'

Not used

Unrecoverable channel failure during RMS f",tch.

X'C3'

X'E2'

A, I, S*

Not used

Channel failure on SYSLOG when RMS message scheduled.

X'C4'

X'E2'

A, I, S*

Not used

Reserved (should not occur)

X'C5'

X'E2'

A, I, S*

Not used

Channel failure: ERPI Bs exhausted.

X'C6'

X'E2'

A, I, S*

Not used

Channel failure; two channels damaged or a damaged channel situation occurred while
RMS was executing an I/o operation.

X'C7'

X'E2'

A, I, S*

Not used

Channel failure; system reset was presented by a channel.

X'C8'

X'E2'

A, I, S*

Not used

Channel failure; system codes in ECSW are invalid.

X'C9'

X'E2'

A, I, S*

Not used

Channel failure; channel address invalid.

SYSTEM/360 and SYSTEM/370 WAIT Codes:
X'03'

X'E6'

Channel

Unit

DOS unrecoverable disk error during program fetch. The first six sense bytes are
placed in hex bytes 5-A. Re-IPL system.

X'04'

X'E6'

Not used

Cancel condition has occurred while performing a Supervisor fundtion (not a Supervisor
detected problem -program error). Normally a Program Check while in Supervisor State.
This condition also occurs if a fetch has been issued for and IBM - suppl ied transient
which is not in the system core image library. IBM-supplied $$A, $$B, and $$R
transients cannot be placed in a private core image library. Take a stand -alone dump;
the name of the transient involved is in the first 8 bytes of the appropriate transient
area. Place the transient in the system core image library. Re -I PL system.

X'05'

X'E6'

Channel

Unit

I/o Error Queue has overflowed as the result of an I/O error on a program fetch
channel program. Re -IPL system.

X,06'

Not used

Reserved (shou Id not occ ur) •

X'07'

X'E6'

Channel

Unit

IPL I/O error. Channel can unit indicate whether SYSRES or communication device.
Re -IPL system.

Channel

Unit

Error recovery messages. Refer to OP messages in DOS Messages, found in Preface.

X'08' to
X'60'

* Note: A (X'Cl ') = SYSREC error recording unsuccessful.
I (X 'C9') = SYSREC error recording incomplete.
S (X'E2') = SYSREC error recording successful.

Figure 4.8.

Low Core Error Bytes

Section 4: Debugging Aids

203

If a program check interrupt occurs
while the DOS supervisor is in control of
the system, the integrity of the control
program itself is in doubt. system
response is to put a message of 04W
(X'04E6') in bytes 0 and 1 and enter a hard
wait. Note that many programs may run in
the supervisor state and hence cause this
type of hard wait. These programs include
BTMOD (in its channel appendage routine),
SPOOLing programs such as POWER (which
alters the address of the SVC new PSW to
point to the POWER partition), most $$A and
some $$B transients.
After the 04W message has been noted. a
stand-alone dump should be taken. The
first diagnostic steps, as with any program
check. are to locate the failing
instruction and determine the program in
error. Use the supervisor assembly listing
to determine if the program check address
(location X'2D'-X'2F') falls within the
supervisor nucleus (address less than label
NUCEND), within the logical transient area
(label LTA to LTA+X'4BO'). the physical
transient area (label PTA to PTA+X'22S'),
or outside the supervisor (address greater
than label PPBEG).
In the first case. use the supervisor
listing to find what routine was being
executed and what function the supervisor
was attempting to perform. Use the I/O old
PSW to find the device involved in the last
I/O interrupt received and the SVC old PSW
for the last SVC executed.
In the second and third cases, the name
of the transient involved appears at the
beginning of the transient area involved.
Use the DOS system Generation listed in the
Preface to find the function of that
transient.
When the program check address is
outside the supervisor, find the partition
it falls in and use the program
documentation to locate the failure.
If W is not present in location 1,
record the communications region address
contained in locations X'16'-X'17' and take
a stand-alone dump. Check the PIK(located
at displacement X'2E' in the communications
region) to determine the task in control.
Then, locate and examine the PIB table
entry for the task in control.
The All Bound PIB is usually active,
indicating an I/O interrupt or event has
not occurred. The program or task save
areas indicate the device(s) or resource(s)
being waited on. The PIB supplies
information such as the cancel code and the
address of the partition save area. The
save area contains the PSW. The
instruction address portion of the PSW
204

DOS System Programmer's Guide

should be pointing to the last executed
instruction. The register values can also
be helpful at this point.
Register 14 is used as a standard return
from an IOCS module. Register 15 contains
the address of the IOCS module. . Register 1
points to an ECB, RCB. CCB. or last phase
or transient fetched or loaded.
Note: Certain unusual hardware and
software failures can cause the system to
halt processing with both the system light
and the wait light on continuously. This
indicates the current PSW has its wait bit
set on.but'the CPU is operating
(processing microprogram instructions). If
possible. the system should be left in this
state until a customer engineer has
arrived. A stand-alone dump can show the
I/O operations in progress. No low-core
message will be found. and a re-IPL is
necessary to continue processing_ .
By gathering all of the proper
documentation and using some of the aids
given in this section. most errors should
be resolved without too much difficulty.

Debugging Assembler Programs
The proper documentation and careful
interpretation of system messages is needed
to ensure that the programmer understands
the diagnostics provided by the system.
Internal pointers found in the system
communications region and PIB table allow
the programmer to analyze main storage
dumps. to locate programs and save areas.
and to determine the cause of the error.
The linkage editor map shows where
programs should be located in main storage.
where overlays are loaded. and whether the
program is relocatable or assembled for
operation in only one partition. Remember
that all relocatable programs are assembled
with a load address of zero in the Disk
Operating system.
See Figures 1.7 and 4.4-4.6 for the
format of the system communications region
and the PIB table and Figure 1.2 for the
format of the partition save areas.

Debugging COBOL Programs
Debugging information for the COBOL
programmer includes an example of a program
named TESTRUN, consisting of a source
statement listing. Data Division map.
Procedure Division map. diagnostic

messages, linkage editor map, and an
abnormal termination dump. The IBM Disk
Operating system FUll American National
standard COBOL Compiler is used for the
compilation job step. Figure 4.9 contains
the program output in its entirety.
The Data Division map provides the
internal name generated by the compiler for
data names and file names defined in the
program. This internal name is the same as
that used in the object code listing. An
address is provided for each name, in the
form of a base and a displacement.
The Procedure Division map is a listing
of the object code. Compiler-generated
card numbers identify the COBOL statement
in the source deck containing the verb that
corresponds to the object code. The object
code listing also contains the relative
address of the object code instruction. In
addition to the object code, a Task Global
Table (TGT), a Program Global Table (PGT),
a literal pool and register assignments are
provided when a Procedure Division map is
requested. The TGT is used to record and
save information needed during the
execution of the object program. The PGT
contains literals and the addresses of
procedure names and generated procedure
names referenced by Procedure Division
instructions.
The linkage editor map contains the load
address of the program and lists the names
and locations of COBOL subroutines in main
storage.

HOW TO USE A DUMP
When a job is abnormally terminated due to
a serious error in the problem program, a
message is written on SYSLST that indicates
the:
1.

type of interrupt; for example, a
program check

hexadecimal address of the instruction
that caused the interrupt

condition code

reason for the interrupt: for example,
a data exception.

The instruction address can be compared
to the Procedure Division map, where a
relative address is provided for each
statement. The load address of the module
(which can be obtained from the map of main
storage generated by the Linkage Editor)
must be subtracted from the instruction
address to obtain the relative instruction

address as shown in the Procedure Division
map. If the interrupt occurred within the
COBOL program, the programmer can use the
error address and the Procedure Division
map to locate the specific statement in the
program that caused the dump to the taken.
Examination of the statement and the fields
associated with it may produce information
as to the specific nature of the error.
Figure 4.9 illustrates a dump
a data exception. Invalid data,
data that does not correspond to
is placed in the numeric field B
result of redefinition.

caused by
that is,
its usage,
as a

• The program interrupt occurred at
hexadecimal location 0039Bc. This is
indicated in theSYSLST message printed
just before the dump.
• The linkage editor map indicates that
the program was loaded into address
0032A0. This is determined by
examining the load point of the control
section TESTRUN. TESTRUN is the name
assigned to the program module by the
source coding: PROGRAM-ID. TESTRUN.
• The specific instructin which caused
the dump is located by subtracting the
load address from the interrupt address
(that is, subtracting 32A0 from 39BC).
The result, 71C, is the relative
interrupt address and can be found in
the object code listing. In this case,
the instruction in question is AP (add
decimal).
'
• The left-hand column of the object code
listing gives the compiler-generated
.~.~ card number associated with the
instruction. It is card 69. As seen
in the source listing, card 6' contains
the COMPUTE statement.
LOCATING A DTF
One or more DTF's are generated by the
compiler for each file opened in the COBOL
program. All information about that file
is found within the DTF or in the fields
preceding the DTF. A particular DTF may be
located in a system dump as follows:
1.

Determine the order of the DTF address
cells in the TGT from the DTF numbers
shown for each file name in the Data
Division map.
'Note: Since the order is the same as
the FD's (File Description) in the
Data Division, the order can be
determined from the source program if
the Data Division map is not
requested.
section 4: Debugging Aids

205

Find the relative starting address of
the block of DTF cells from the TGT.

Calculate the absolute starting
address of the block by adding the
hexadecimal relocation factor for the
beginning of the object module as
given in the linkage editor map.

Allowing one fullword per DTF cell,
count off the cells from the starting
address found in Step 3, using the
order determined in step 1 to locate
the desired DTF cell.

LOCATING DATA

If more than one DTF is generated for
a file, the above procedure should be
followed using the PGT and the SUBDTF
cells rather than the TGT and the
DTFADDR cells. The order in which
multiple DTF's appear in main storage
is dependent on the OPEN option as
follows:

The location assigned to a given data name
may similarly be found by using the BL
(Base Locator) number and displacement
given for that entry in the Data Division
map, and then locating the appro~riate
fullword BL cell in the TGT. The sum of
the displacement and the contents of the
cell give the relative address of the
desired area. This can then be converted
to an absolute address as described for
locating a DTF.

(a) INPUT
(b) OUTPUT
(c) I-O or INPUT REVERSED

206

DOS System Programmer's Guide

There can be two or three SUBDTF's for
each file"with multiple OPEN options.
The Data Division map sould be used to
determine the file and the number of
the SUBDTF CELL assigned to it.

...."IJ

(1)

// JOB DT~CHK
// OPTION NODECK,LINK,LIST,LISTX,SYM,ERRS
// EXEC FCOBOL

11.49.30

•

g
en

I-'
~

]

.....
~

rt
~

oH\

.....

en
~

....rt

g
~

'='
~

g
....

~
~

....g,
)I

CI1
N

.....

CBL QUOTE, SEQ
00001
000010
00002
000020
00003
000030
00004
000040
00005
000050
00006
000060
00007
000070
00008
000080
00009
000090
00010
000100
00011
000110
00012
000120
00013
000130
00014
000140
00015
000150
00016
000160
00017
000170
00018
000180
00019
000190
00020
000200
00021
000210
00022
000220
00023
000230
00024
000240
00025
000250
00026
000255
00027
000260
00028
000270
00029
000280
00030
000290
00031
000300
00032
000310
00033
000320
00034
000330
00035
000340
00036
000350
00037
000360
00038
000370
00039
000380
00040
000390
00041
000400
00042
000410
00043
000420
00044
000430
00045
000440
00046
000450
00047
000460
00048
000470
00049
000480
00050
000490
00051
000500
00052
000510
00053
000520
00054
000530
00055
000534
00056
000535

IBM DOS AMERIC~N N~TIONAL STAND~D COBOL

IDENTIFIC~TION DIVISION.
PROGRAM-ID. TESTRUN.
AUTHOR. PROGR~MMER N~ME.
INSTALL~TION. NE~ YORK PROGRAMMING CENTER.
DATE-~RITTEN.
FEBRU~RY 4, 1971
D~rE-COMPILED. 07/23/71
REM~RKS.
THIS PROGRAM HAS BEEN WRITTEN ~S A SAMPLE PROGRAM FOR
COBOL USERS. IT CREATES AN OUTPUT FILE ~ND READS IT BACK AS
INPUT.

ENVIRONMENT DIVISION.
CONFIGURATION SECTION.
SOURCE-COMPUTER. IBM-360-H50 •
OBJECT-COMPUTER. IBM-360-H50.
INPUT-OUTPUT SECTION.
FILE-CONTROL.
SELECT FILE-1 ASSIGN TO SYS008-UT-2400-S.
SELECT FILE-2 ASSIGN TO SYS008-UT-2400-S.
D~T~ DIVISION.
FILE SECTION.
FD FILE-1
LABEL RECORDS ~RE OMITTED
BLOCK CONTAINS 5 RECORDS
RECORDING MODE IS F
RECORD CONTAINS 20 CHARACTERS
D~TA RECORD IS RECORD-1.
01 RECORD-1.
05 FIELD-~ PIC X(20).
FD FILE-2
LABEL RECORDS ARE OMITTED
BLOCK CONTAINS 5 RECORDS
RECORD CONT~INS 20 CH~R~CTERS
RECORDING MODE IS F
DATA RECORD IS RECORD-2.
01 RECORD- 2.
05 FIELD-~ PIC X(20).
~ORKING-STORAGE SECTION •
01 FILLER.
02 COUNT PIC S99 COMP SYNC.
02 ALPH~BET PIC X(26) V~LUE IS "ABCDEFGHIJKLMNOPQRSTUVWXYZ".
02 ALPHA REDEFINES ALPHABET PIC X OCCURS 26 TIMES.
02 NUMBR PIC S99 COMP SYNC.
02 DEPENDENTS PIC X(26) VALUE "01234012340123401234012340".
02 DEPEND REDEFINES DEPENDENTS PIC X OCCURS 26 TIMES.
01 ~ORK-RECORD.
05 NAME-FIELD PIC X.
05 FILLER PIC X.
05 RECORD-NO PIC 9999 •
05 FILLER PIC X VALUE IS SPACE.
05 LOC~TION PIC AA~ V~LUE IS "NYC".
05 FILLER PIC X VALUE IS SPACE.
05 NO-OF-DEPENDENTS PIC XX •
05 FILLER PIC X(7) VALUE IS SPACES.
01 RECORD~.
02 A PICTURE S9(4) VALUE 1234.

CBF CL3-3 07/23/71

to.)

cgo

8en
en
'<
en
rt
(D

..,.
~

4::

0
txI
0
1:"1

11
0
11

i
.
en
(D

~...,
(D
~

Cil
~

11
0
IQ
11

..,.
(D

SlI

~
to.)

to.)

....a

I I

2
00057
00058
00059
00060
00061
00062
00063
00064
00065
00066
00067
00068
00069
00070
00071
00072
00073
00074
00075
00076
00077
00078
00079
00080
00081
00082
00083
00084

000536
000540
000550
000560
000570
000580
000590
000600
000610
000620
000630
000640
000645
000650
000660
000670
000680
000690
000700
000710
000720
000730
000740
000750
000760
000770
000780
000790

02 B REDEFINES A PICTURE S9(7) COMPUTATIONAL-3.
PROCEDURE DIVISION •
BEGIN. READY TR~CE.
NOTE rHAT THE FOLLOWING OPENS THE OUTPUT FILE TO BE CREATED
AND INITIALIZES COUNTERS.
STEP-l. OPEN OUTPUT FILE-l. MOVE ZERO TO COUNT, NUMBR.
NOTE THAT THE FOLLOWING CREATES INTERNALLY THE RECORDS TO BE
CONTAINED IN THE FILE, WRITES THEM ON rAPE, AND DISPLAYS
THEM ON THE CONSOLE.
STEP-2. ADD 1 TO COUNT, NUMBR. MOVE ALPHA (COUNT) TO
NAME-FIELD.
COMPUTE B = B + 1.
MOVE DEPEND (COUNT) TO NO-OF-DEPENDENTS.
MOVE NUMBR TO RECORD-NO •
STEP-3. DISPLAY WORK-RECORD UPON CONSOLE. WRITE RECORD-l FROM
WORK-RECORD.
STEP-4. PERFORM STEP-2 THRU STEP-3 UNTIL COUNT IS EQUAL TO 26.
NOTE THAT THE FOLLOWING CLOSES THE OUTPUT FILE AND REOPENS
IT AS INPUT.
STEP-5. CLOSE FILE-l. OPEN INPUT FILE-2.
NOTE THAT THE FOLLOWING READS BACK THE FILE AND SINGLES
OUT EMPLOYEES WITH NO DEPENDENTS.
STEP-6. READ FILE-2 RECORD INTO WORK-RECORD AT END GO TO STEP-8.
STEP-7. IF NO-OF-DEPENDENTS IS EQUAL TO ·0· MOVE "Z" TO
NO-OF-DEPENDENTS. EXHIBIT NAMED WORK-RECORD. GO TO STEP-6.
STEP-8. CLOSE FILE-2.
STOP RUN.

"':I

I.Q

I I

Ii
CD

..
.c::

0
bj
0

t-t

CIl
$lI

~~
CD

I.Q

~
$lI

~
W

.....J

en
CD
n
rt

g
.c::

t::7

~
~
I.Q

:::J

I.Q
~

INTRNL NAME

LVL SOURCE NAME

DNM=1-148
DNM=1-178
DNM=1-199
DNM=1-216
DNM=1-246
DNM=1-267
DNM=1-287
DNM=1-306
DNM=1-321
DNM=1-339
DNM=1-357
DNM=1-372
DNM=1-392
DNM=1-408
DNM=1-432
DNM=1-452
DNM=1-471
DNM=1-490
DNM=2-000
DNM=2-018
DNM=2-037
DNM=2-063
DNM=2-082
DNM=2-102
DNM=2-113

FD
01
02
FD
01
02
01
02
02
02
02
02
02
01
02
02
02
02
02
02
02
02
01
02
02

FILE-l
RECORD-l
FIELD-A
FILE-2
RECORD-2
FIELD-A
FILLER
COUNT
ALPHABET
ALPHA
NUMBR
DEPENDENTS
DEPEND
WORK-RECORD
NAME-FIELD
FILLER
RECORD-NO
FILLER
LOCATION
FILLER
NO-OF-DEPENDENTS
FILLER
RECORDA
A
B

BASE
DTF=Ol
BL=l
BL=l
DTF=02
BL=2
BL=2
BL=3
BL=3
BL=3
BL=3
BL=3
BL=3
BL=3
BL=3
BL=3
BL=3
BL=3
BL=3
.BL=3
BL=3
BL=3
BL=3
BL=3
BL=3
BL=3

DISPL

000
000
000
000
000
000
002
002
01C
OlE
OlE
038
038
039
03A
03E
03F
042
043
045
050
050
050

INTRNL NAME
DNM=1-148
DNM=1-178
DNM=1-199
DNM=1-216
DNM=1-246
DNM=1-267
DNM=1-287
DNM=1-306
DNM=1-321
DNM=1-339
DNM=1-357
DNM=1-372
DNM=1-392
DNM=1.-408
DNM=1-432
DNM=1-452
DNM=1-471
DNM=1-490
DNM=2-000
DNM=2-018
DNM=2-037
DNM=2-063
DNM=2-082
DNM=2-102
DNM=2-113

DEFINITION

DS OCL20
DS 20C
DS
OS
OS
DS
OS
OS
OS
OS
DS
OS
DS
OS
DS
DS
DS
DS
OS
OS
OS
OS
OS

OCL20
20C
OCL56
lH
26C
lC
lH
26C
lC
OCL20
lC
lC
4C
lC
3C
lC
2C
7C
OCL4
4C
4P

·USAGE
DTFMT
GROUP
DISP
DTFMT
GROUP
DISP
GROUP
COMP
DISP
DISP
COMP
DISP
DISP
GROUP
DISP
DISP
DISP-NM
DISP
DISP
DISP
DISP
DISP
GROUP
DISP-NM
COMP-3

M
F
F

....

N
0

CI'l

'<
en
rt
CD

...."'iI

I.Q

s::
,f::

.§

.
en
t1

en
~

..g
t1
0

I.Q

....s::
CD

003E8

TGT
SAllE AREA
SWITCH
TALLY
SORT SAVE
ENTRY-SAVE
SORT CORE SIZE
NSTD-REELS
SORT RET
WORKING CELLS
SORT FILE SIZE
SORT MODE SIZE
PGT-VN TBL
TGT-VN TBL
SORTAB AODRESS
LENGTH OF IlN TBL
LNGTH OF SORTAB
PGM ID
A(INIT1)
UPS I SWITCHES
OVERFLOW CELLS
BL CELLS
DTFADR CELLS
TEMP STORAGE
TEMP STORAGE-2
TEMP STORAGE-3
TEMP STORAGE-4
BLL CELLS
VLC CELLS
SBL CELLS
INDEX CELLS
SUBADR CELLS
ONCTL CELLS
PFMCTL CELLS
PFMSAV CELLS
VN CELLS
SAllE AREA =2
XSASW CELLS
XSA·CELLS
PARAM CELLS
RPTSAV ARE-A
CHECKPT CTR
IOPTR CELLS

;CD

MEMORY MAP

0
t"'I

..g
t1
0

I.Q

I. I

t1
CD

t1
~

~
t1
rt'

,f::

-..J

003E8
00430
00434
00438
0043C
00440
00444
00446
00448
00578
0057C
00580
00584
00588
0058C
0058E
00590
00598
0059C
005A4
005A4
005BO
005B8
005CO
OOSCO
005CO
OOSCO
005C4
005C4
005C4
005C4
005CC
005CC
005CC
00500
00504
00504
00504
005D4
00508
00508
00508

LITERAL POOL (HEXI
00618 \ LIT+O)
00630 (LIT+24)

00000001
5BSBC2C6

1COOO01A
C3D4E4D3

DISPLAY LITERALS (BCD)
00640 (LTL+40)

, WORK-RECORD'

SB5BC2D6
FOE90000

D7CSDS40

cooooooo

5BSBC2C3

D3D6E2C5

....tlIj

I.Q

s::
~
.c::

PGT

OVERFLOW CELLS
VIRTUAL CELLS
PROCEDURE NAME CELLS
GENERATED NAME CELLS
SUBDTF ADDRESS CELLS
VNI CELLS
LITERALS
DISPLAY LITERALS

(')

~...,

REG 6
REG 7
REG 8

j
a

60
63

o
.....

CD
0

....rt

63
67

::s
.c::
II

t:;
CD

s::

I.Q
I.Q

....
::s

I.Q

....

)II

......
N

005EO
005EO
005EC
00600
00610
00610
00618
00640

005EO

I I

BL =3
BL =1
BL =2
00064C
00064C
000650
000652
000655
00065C
000660
000664
000666
000669
000610
000614
000678
00061A
00061C
000680
000684
000688
00068A
00068E
000692
000694
000698
00069c
0006AO
0006A4
0006AA
0006BO
0006BO
0006B4
0006B6
0006B9
0006cO
0006C4
0006C8
0006CC
0006D2

START
58 FO C 004
05 1F
000140
04F6F0404040
96 40 D 048
58 FO C 004
05 1F
000140
04F6F3404040
41 10 C 040
58 00 D 1C8
18 40
05 FO
50 00 F 008
45 00 F OOC
00000000
OA 02
41 00 D 1C8
58 FO C 008
05 EF
58 10 D 1C8
96 10 1 020
50 20 D 1BC
58 10 D 1BC
D2 01 6 000 C 038
D2 01 6 01C C 038

SVC

PN=Ol
58 FO C 004
05 1F
000140
04F6F7404040
48 30 C 03A
4A 30 6 000
4E 30 D 1DO
D7 05 D 1DO D 1DO
94 OF D 1D6

EQU
L
BALR
DC
DC
01
L
BALR
DC
DC
LA
L
LR
BALR
ST
BAL
DC
LA
L
BALR
L
01
ST
L
MVC
MVC
EQU
L
BALR
DC
DC
LH
AH
CVD

XC
NI

15,004(0,12)
1,15
x' 000140'
X'04F6F0404040'
048 (13), x' 40'
15,004 (0,12)
1,15
X'OOO140'
X'04F6F3404040'
1,040(0,12)
O,lC8(O,13)
4,0
15,0
0,008(0,15)
0,00C(0,15)
X'OOOOOOOO'
2
0,lC8(0,13)
15,008(0,12)
14,15
l,lC8(O,13)
020(1),X'10'
2,lBC(Q,13)
7, 1BC(0, 13)
000(2,61,038(12)
01C(2,6),038(12)

*
15,004
(0,12)
1,15
X'OOO140'
X'04F6F7404040'
3, 03A(0, 12)
3,000(0,6)
3,lDO(0,13)
1DO(6,13I,lDO(13)
1D6(13), x' OF'

v (ILBDDSPO)

SWT+O
V

ttl

0
t:-t

I.Q

CIl
DI

11
0
11

i.
11
(J)

Q.,

CIl

'<
en

I I

~CD

...,

"0
CD

ttl

11
0

69
70

I.Q

11
DI

-is
:+

'"0
~

...a

0006D6
0006DA
0006DE
0006E2
0006E6
0006EA
0006FO
0006F4
0006F8
0006FC
000700
000704
000708
00070A
00070E
000712
00b716
00071C
000722
000726
00072A
00072E
000730
000734
000738
00073C
000742
000746
00074A
00074E
000754
000758
00075C
00075E
000761
000768
00076C
00076E
000770
000771
000774
000778
00077A
00077C
000782
000786
000788
00078C
000790
000794
000798
00079C
00079E
00079E
0007A2
0007A.4
0007A7

4F 30 D lDO
40 30 6 000
48 30 C 03A
4A 30 6 01C
4E 30 D lDO
D7 05 D lDO D
94 OF D lD6
4F 30 D lDO
40 30 6 01C
41 40 6 002
48 20 6 000
4C 20 C 03A
lA 42
5B 40 C 038
50 40 D lDC
58 EO D lDC
D2 00 6 038 E
FA 30 6 050 C
41 40 6 OlE
48 20 6 000
4C 20 C 03A
lA 42
5B 40 C 038
50 40 D lEO
58 EO D lEO
D2 00 6 043 E
92 40 6 044
48 30 6 01C
4E 30 D lDO
F3 31 6 03A D
96 FO 6 03D
58 FO C 004
05 IF
000140
04F7F2404040
58 FO C 004
05 IF
0002
00
000014
ODOO01C4
0038
FFFF
D2 13 7 000 6
58 10 D lC8
18 41
58 FO 1 010
45 EO F OOC
50 20 D lBC
58 70 D lBC
58 10 D lE8
07 Fl

lDO

000
03C

000

lD6

038

PN=02
58 FO C 004
05 IF
000140
04F7F4404040

CVB
srH
LH
AH
CVD
XC
N1
CVB
8TH
LA
LH
MH
AR
S
ST
L
MVC
AP
LA
LH
MH
AR
S
ST
L
MVC
MV1
LH
CVD
UNPK
01
L
BALR
DC
DC
L
BALR
DC
DC
DC
DC
DC
DC
MVC
L
LR
L
BAL
ST
L
L
BCR
EQU
L
BALR
DC
DC

3,lDO(0,13)
3,000(0,6)
3,03ACO,12)
3,01C
2,13,050 (11)
14,054(0,13)
15,14
048 (13) , x, 20'
6,004(0,0)
1,00C(0,12)
7,038(0,12)
7,0
5,0
4,000(0,1>
4,11
4,000(0,1>
1,6,000(5)
8, 1BC(0, 13)
7, 1CF'
....
Q.a

...
N

JOB

DTACHK

ACTION
LIST
LIST
LIST
LIST
LIST
LIST
LIST

07/23/71

TAKEN
MAP'
AUTOLINK
IJFFBZZN
AUTOLINK
ILBDDSPO
INCLUDE IJJCPDl
AUTOLINK
ILBDIMLO
AUTOLINK
ILBDMNSO
AurOLINK
ILBDSAEO
ENTRY

DISK LINKAGE EDITOR DIAGNOSTIC OF INPUT

0001

IV
IV
0

..,."III
I..Q

s::

8en
en
~
en
rt
(1)

a
ttl
11
0

I
.
(I)

en
..,.s::
Q,

(I)

01/23/71

PHASE

XFR-AD

LOCORE

HICORE

DSK-AD

PHA.8E***

0032AO

004ADB

63 01 2

ESD TYPE

LABEL

LOADED

REL-FR

CSECT

TESTRUN

0032AO

CSECT
* ENTRY
* ENTRY
* ENTRY

IJFFBZZN
IJFFZZZN
IJFFBZZZ
IJFFZZZZ

003C50
003C50
003C50
003C50

003C50

CSECT
ENTRY

ILBDSAEO
ILBDSAE1

0049FO
004A06

0049FO

CSECT

I LBDMNS 0

0049E8

CSECT
* ENTRY
* ENTRY
* ENTRY

ILBDDSPO
ILBDDSP1
ILBDDSP2
ILBDDSP3

0041B8
004708
0047AO
004958

0041B8

(1)

.f=

n
t»

t"4
CI)

(I)

ttl
11

;

ttl
SlI

11
rt

.f=

H\
IV

...J
..,

CSECT

ILBDIMLO

004990

CSECT
ENTRY
* ENTRY

IJJCPD1
IJJCPD1N
IJJCPD3

003FCO
003FCO
003FCO

003FCO

o
o

en
en

:>i

z(!lU

en iii
en><

..:t:iiI
"

C"'ll'-

V:>>D

I Figure

4. 9.

COBOL Sample Program (Part 15 of 27)

Section 4: Debugging Aids

221

I Figure

222

4.9.

COBOL Sample Program (Part 16 of 27)

DOS System Programmer's Guide

..,.
I'Ij

~CD

..
01::

07/23/11

OTACBK

\.Q

11.52.21

GR 0-1
GR 8-F
FP REG
COMREG

00003850 00003960 00000001 00000001
.000035B8 00003BE2 000032AO 000032AO
00000000 00000000 00000000 00000000
BG ADOR IS 000208

0000338A 50003C12 00003388 00003550
00003880 00003688 0000338A 000041B8
00000000 00000000 00000000 00000000

000000
000020
000040
000060
000080
OOOOAO
OOOOCO
OOOOEO
000100
000120
000140
000160
000180
0001AO
0001CO
0001EO
000200
000220
000240
000260
000280
0002AO
0002CO
0002EO
000300
000320
000340
000360
000380
0003AO
0003CO
0003EO
000400
000420
000440
000460
000480
0004AO
0004CO
0004EO
000500
000520
000540
000560
000580
0005AO
0005CO
0005EO
000600
000620
000640
000660

00000000
FF050001
00002F28
00040000
00000000
0146940F
06B006BO
AOOC4110
010E9104
C0440118
4570B218
C0441BAA
A0009140
4BA00262
90309890
41100030
00003000
C4E3C1C3
A8A07COO
00002044
01001F98
00001F2C
0000289C
48A00236
410CBOOO
48E001C8
02E04190
48661000
1B234140
00CA4230
960C1004
03CC4284
95600237
4121000F
92FF0409
20004810
B60E0501
00234180
02740200
4780B62E
02CC9507
02004144
91406004
02034000
48603000
50184180
95035018
9C000184
00454110
58600048
47700654
47700500

00000000
5B5BC2C5
F15A1COO
00000000
00050003
41900156
00504510
00C695C1
A0004BAO
A0009680
BC104510
023141AA
68009058
90108200
01FOBF50
FF050001
32A03000
00004ADB
25102514
1F140020
00000294
00000000
00002A9C
58BOA004
B000410C
BF5801FO
002341BO
BCAC4860
1A234220
950B1001
41455000
5000BF14
04204111
00C61B33
1007AOOE
03B88930
BE5A4720
00014148
42205000
9560C09C
581002CC
9120100C
41800558
30029198
01694060
06FF91FO
02020049
01F94130
05E29550
41F0089A
41F0065C
91401038

PAGE

I:J:'

g
en

\.Q
Ii

rt
~
...,J

...,J

en
CD

..,.~

g
01::

\.Q
\.Q

..,.

..,.lJII
ft
N
N
W

00000000
40002E06
OCOOOOOO
00000336
00000000
B41B41AO
06B006BO
00EA9260
A0004180
94F9103B
01F842BO
0006BEEE
A0014110
41AACOOO
01F08200
41FOB166
80001048
C8024040
00C62111
OOOOOOOC
20000000
00000000
00003228
4AA00262
07F99601
0201BF50
01B69500
07F6181F
03904130
04094820
07F91858
000001F9
47800366
4510BCAC
02000440
02544338
0022BB2c
04~9110

04B14000
47F004EC
60004170
00000503
41800560
700195FF
95003000
08409101
41000634
41100,5F2
060C91AF
4A60BOOC
45100B84
45700B84

00000000
FF150001
00001038
00040000
00000000
C0544510
06B041BB
A00195E2
010E9203
01011058
00E148BO
000C43A1
BB640201
02010016
00389284
96030039
FOF161F2
0007AFFF
21182269
22E21E4E
00000000
00000000
100020CE
9180AOOO
A00048BO
E00094FO
00234180
1B664121
00151B23
02364322
41430002
95FF04B1
0502A005
05021009
A0004123
10004930
4780B8E4
30064180
50104000
5880BF8C
05660202
10014000
94BF6004
30024710
47800592
100C4110
9560C09C
4032COB4
00440789
50600040
9120800F
4B800262

00000000
E00039C2
00000000
0000141A
00000003
OB8418A8
001141BB
A0024180
008F9281
10589283
02C847FO
000142AO
01F09008
A0009898
COA40201
82000038
F361F1F1
00004ADB
226AOOOO
1EF41F04
02080000
00000000
000010C8
47100306
02C841CB
BF510213
03F49526
000F4570
41B00392
C0031A23
43540000
07891BOO
02814770
023941BO
000B0500
BE3847BO
0501BE4C
B2384180
92FF4000
44000CB8
02C06001
41800566
91101002
03C64280
9F006000
0500940F
4180061C
96803006
0201003A
4032COB4
41100094
4818COOO

00000208 FF050000
0601F440 FF050184
01BB0994 00040000
OOOOOBBC 00040000
06B006BO 06B041BB
4180B2CE 41FOOOOA
01464180 01569640
A0024180 00C69561
0262481A C00049AO
A0014400 04080188
BC700205 BEEEBEF5
C0444400 A0045890
68209060 68409068
01F04400 A0045890
9890BF58 820001FO
40002E06 00001000
00000000 00000000
00000010 0001FFFF
25183CFO F7F2F3F7
214C0010 5B5BC206
00000000 000025AC
00000000 00000000
923801C9 909001FO
9018B030 48B002C8
B00095FF AOOF0189
01F99090 01F09220
00BE4860 00221A66
BE5C1B22 43201001
04094320 100141FO
41F00454 412000C8
1A444A40 BE5495FF
95FF04B1 4180B238
00004910 BE3A41BO
43301001 95011006
47BOOOCA 04031002
00034A30 024891FO
04AE4930 BE144710
80001A44 4A40BE54
4260500C 92035018
41700566 020202CO
1B444340 30054C40
41100560 91051002
96401002 96141002
30060119 43203000
05E29550 30044180
500C4180 05000300
1009940F 07030300
OBC69106 00454110
30044110 06200202
950150184120065C
91203006 41100094
47100500 96F006FF

00000000
80002EOO
OF0014BA
00000204
00134570
06B006BO
A0019120
A0024180
021641AA
941FA001
OC05BEEE
A0044220
68609010
A0049818
9680AOOO
00002000
00000000
F815EC01
F1F2FOF4
00130001
00000044
00000000
4190086C
41CBBOOO
90EOBF6C
01C94590
481002CA
4130001F
04584120
96801002
40004110
48600236
04201B66
4110039E
COB04182
30044180
04AE950F
18584A50
91F03004
10095860
BE485A40
41100560
9601100C
4322C090
080C9504
C09C3004
0048500C
OE9C913F
00491009
9560C09C
9560C09C
95003003

•••• •••••••••• B

$$BEOJ4 ••••••••
7 •••••••••••••••
••••••••••••••• M

••••••••• O••••••
• &. • •• • •• ••

••••••• - ••• S ••••

••••• 9 •• P •••••••
••••• 8 ••• X••• B.0

• •••

.................
.F.A••••• F./ ••••
•••••• K •••• 5 ••••

••• • ••••• K •• 0 ••
•• ••• • •• K •••••••

.•••••
.....o••••••••••
. -. . .. ..

••••• 0 •••••••• K.

• O. & ••••••• 0 ••••

••••• 0 ••••••••••

•••••••• 01/23/11
OTACBK
•• Gl •• F ••••••••••
••••••••• S ••• 4 ••

•••••••••••• 8 •• J
••••••. 072311204
••••• <•• $$BO ••••

• •••••••••••••• B

••••••. r ••.

0 ••••

••••••••••• B ••••

••••• 9 ••••••• B••
••• BK •• &•••••• K.
••••••••••••• 4 ••
••••• 6 ••••••••••

••• 0.9 ••• 0 ••• r ••

. ...... - ....... .

• .. -. * .••••••.•.
••••• R ••••• 0 ••••

•• ••• R ••••••••••

• •••• 9 ••••••••••
• •••••• 9 ••••••••

..•••••
&....0 •••••
....B••••
...
& •••••••••••• - ••

.- •••••• N•••••••
• ••••••• N •••••••
••• RK ••••••••• N.
•• N •••••••• UN •• <

• F ••••••••••••••
•••••••• M•••••••

••••••••••• 0 ••••
••• • •• • ••

• •••• &

• .K •••• &• • • •

•• & •• -& ••• & •• 0 ••

• •••• 0 ••••••••••

• - •••••• K•••••• -

•••• - ••••• K ••• -.
•••••• N••••••••

.... - .. - .....
K ••••••••••• F ••
· - ............ -.
-

&. • •• • •••••••••
• • &•••••• - ••••• Gl
••••••• 2 •••• ~ ••
•• •••• •• •••• K •••
• - ••• - •• &-.

•• • • ••• •• < •••

. ...... - ...... .
-. S. &••••••••

••• 0 &••••• L •••••
K ••••••••• L ••• &.

.9 ••• F ••••••••••
.S. & •••••• K•••••
.0 •••• & ••••
• O. *-••••••••• - ••
••••••• O••••••

*.-..

..,.I7J

N
N

8til

..
~

~rt

~....

11
\Q

til

CD
11

(1)

en
(j)

..,.s=

a
i

Pol
CD

t-A

111

...a

....-...

OTACHK
000680
0006AO
0006CO
0006EO
000700
000720
000740
000760
000780
0007AO
0007CO
0007EO
000800
000820
000840
000860
000880
0008AO
0008CO
0008EO
000900
000920
000940
000960
000980
000911_0
0009CO
0009EO
OOOAOO
OOOA2Q
000A40
000A60
000A80
OOOAAO
OOOACO
OOOAEO
OOOBOO
000B20
000B40
00-0B60
000B80
OOOBAO
OOOBCO
OOOBEO
OOOCOO
000C20
000C40
000C60
000C80
OOOCAO
OOOCCO
OOOCEO
000000
000020
000040
000060
000D80
OOOOAO

47700500
BE3E0202
4870BOAE
4770009A
07704700
47800098
47F00728
00024A70
95025018
41770018
00080204
80024770
47F00604
30050203
9501500C
00480202
lB224320
0044B020
OC1495FF
910A0241
95000044
94713006
47100948
01529601
47F009A8
94BF1002
91103006
474009E8
4832COB4
30060117
07F195FF
47200AC6
02624878
92018006
70014270
58770000
45700B84
96803006
02000B50
070647FO
47FOOA1E
OBAA910F
OBB88200
47800C08
96101002
OCF49212
OCEC95FF
COAE4060
OC839204
00450774
C09C4780
9130500C
50800048
002647FO
20009000
500C4770
100C4780
02CC9640

PAGE

07/23/71

96FOB47B
00491009
4070400A
02074000
08045876
45400758
41706008
024C9130
478001AO
41870018
BE427000
00985876
940FB47B
BEOC1008
47800500
70091009
003A4322
47800A42
30024110
07769500
07890602
07F99477
92003003
800047FO
944B1002
47F00A24
478009E8
47F009CC
02000All
41100580
30024780
0300C09C

~0009140

06600500
60009501
05017000
9620800F
91041002
003A4832
OA004842
48800276
500C4780
OBB80729
440004F8
94BF1002
OC839104
30024780
BE5AD715
600A9690
91020C83
00029560
478000BC
02021009
00F0940E
20004760
006E9520
00A69101
60044190

45700B84
58600048
96F00703
60009508
00001B66
05008000
47F006F6
700247AO
4120079C
01F6D506
95025018
00180507
47F00506
5870BEOC
9101100C
47F0050A
C0904832
05010044
OC149139
30040776
00410041
30069104
96801002
09A894FE
440004F8
92FF500C
4810BE5A
95003000
003A0510
lB554350
09E8D201
30049560
70384710
80058011
50184720
BEE04780
47FOOB10
4780093C
COB695FF
COB44832
91F0500C
OB944180
02010044
588002CC
92040C83
30064780
OC529110
60006000
30064186
4780000E
30044780
47F00096
004947FO
00449506
00424580
60004770
600C4710
01529287

9180800C
4870BOAC
91406004
60084770
41870003
70004740
95FF7003
014C01F4
4188C044
8000800A
474007F8
80027000
95015018
95047000
47100500
419001B6
COB40500
B0224780
00454710
9610C08C
47700A42
30064780
9101100C
100C91FO
586002CC
43703002
49307008
41700AOO
41330008
30024145
10000046
C09C4170
OAC64560
47700ABA
OB185870
OB100501
96203006
96043006
30000789
COB64570
47800B9A
006047FO
BEE20202
95078000
91090045
OC5294FO
100C4710
02076000
000C4080
4180BO~8

B73E9562
96201002
OC329120
00444780
OE449102
00B295FF
00F09130
C09447FO

47100604
4010400A
4710060C
07445870
43605000
00980500
47800098
4B800262
9120800F
47200098
4340BE46
47700098
4720081C
50100048
45100B84
91060045
3001003B
OA429102
OC084160
9603C084
91040044
09E80601
4710096A
30044770
96406004
D2003002
41700908
95FF3002
95013000
50001A44
02021000
OAC69101
077C4160
02008005
00404B70
7000C096
96401002
95003000
05609582
OB7A4133
4380500C
OB940700
0041B009
47700C08
47800C42
Oc839120
OCEC4860
00404030
B04A4140
45700E54
30044180
47F00802
100C4780
OE549220
600A4710
02414780
00454780
09E80205

lB444340
50400048
940FB47B
60085074
89600001
80007001
41670002
4878COOO
47100094
02020049
41440001
47F00804
91203006
4110B038
0201B035
47700EA2
078747FO
00444780
08F89180
41900152
47700926
10040044
45700B84
09844370
91401002
40000200
9602700A
07899198
47200B66
4A40BE54
00410601
100C4710
80060500
80124160
BOOC0502
47700B18
96011004
47700AOO
30010187
00081934
4188C044
07000700
9560C09C
944B1002
92040C83
100C4110
BE5A4960
600895FF
004A9102
47F00026
B73E07F4
91011003
OCAC4181
600B4180
00A69502
00B29560
00A694FE
0005600C

05E24C40
95065018
47F00500
00109208
4A600254
47000804
lB774376
416007AA
41F007A4
10095860
4240BE46
4400010C
47100094
47800862
80024170
94F00039
OB3A4570
08CA9180
00440786
91100044
95500044
94733006
198A4700
BF900670
478009AO
400004Bl
4B70COAE
30064780
95FF3002
4A500214
10040044
OAc64570
80068010
80034370
70010239
9560C09C
91270044
07F90200
41330008
072995FF
07F79180
00E201E2
47700C08
91401002
58800040
OC429101
BE5847BO
30024770
00444710
0300C09C
91020C83
41800C32
00104570
B0485080
BE504770
C09C4770
100C4400
910600c7

BE404A40
477006CC
95076000
40104700
95FF6001
4540014C
00018970
48710060
41770018
00485816
0504BE42
41100098
0200B038
47F0085E
B0285010
4170089A
OA280501
30064110
96040044
47100910
47700AOO
9180100C
09624190
4270BF90
96101002
427004Bl
4970BE4C
OA2407F9
01879198
58104000
95035018
OB844B80
47700ABA
60004170
47200B18
47700BOC
47700938
OB53003B
95013000
30020181
100C4110
02010000
9101100C
47800C04
12884770
10024710
OC084A60
OC829602
OCA49130
30049540
47100802
4B80BOOC
OE5447FO
00489COO
006E9502
00A69101
04F85860
970100C7

••••• 0 ••••••••••
•• K •••••• - ••••••

•• 0 ••• •••• K • • - ••• - •••
• • • • • • • N ••••••

• 0 •••• - •• O. 6 ••••

••• M• • • • S< •
• & •••• & •••••

••••••••• 0 •••• -.
•••• -.& . . . . . . . .
.-& •• - ••• - •••• -.
..N ••••••••••• <

•• & •••••••••••••

...
...... - .....
••••••••• O••••••

••••••••• 6N •••••
•• K ••••••• &• • • 8

•••• K •••••• - ••••
• •• • • •• •• N •••

•••••••••• N •••••
• O. Me •••• O. o •• &.
•• K •••••••••••••
• • &•••••••••••••

••••• 0 ••••••••••

••••• <••••••• <.4

•••••••••••• K •••
& •••••••••••• 0 ••

• ••• K ••••••••• &.

• .K •••••• 0 ••••••
• ••••••••••••• N.
•••••••• N•••••••

••••••• 0 •••••• N.
••• -. 8 ••••••••••

•••••• 0 •••••••••

••••••••• & ••••••

••••• 9 ••••••••••

.YO •••••••••••••

••••••• 0 ••••••• 0
• 0 ••••••••• 8. - ••
••••• 0 •••• &•••••

K•••• K••••••••

• •••••• Y ••••••••

• Y.O ••••••••••
•••• K ••••.•••••••
•• • • • • • • • • • 6 ••••
•. 7 ••••••• YK •••••
•• • FL •••••• - ••••
••••••• ••••• F.-

••••• -N •••••••••
•••• - ••• & •••••••
•••• N ••••••••• N.

• •• Q •••••••••••

••••••••••••••• 9
&••••••• &••••
K ••••• 0 ••••••• &.

• F ••••••• F ••••••
.01. - •• N•••••••••
K •••••• - •••• - •••
• ••• N •••••••••

••••••••• O••••••

·............ .

••••••••• 9K •••••
K •••••••••••••••

.0. 0 ••••••••••••
.0 ••••••• 0& •••••
• ••• &•••••••• -. 0
•••••••• K •••• SK.

•• & •••••• 7 ••••••
• •.••••••• S.SK •••
••• R.- ••••••••••

••••••• 8 ••••••••

.4 ••••••• ~ ••••••

••• - ••• - ••••• Q.••

- •• P. -. -. K. -.

••••• O•• L ••••••
••••• K. - ••••••••
•• & •••••• 0 ••••••
& ••• K •••••• 0 ••••

••••••• 4 ••••••• K

...... . ....... .

•••• - ••••• & •••••

••• O. O••••••••••
-

• O. K••••••••••••
• •••••••••••••• 0
-

.............. .

&••••••• - •••••••

• •.•••••• - •••• 0 ••
••• - •••••••••• 0

••••••••••••• 8. • YK ••• - •••• G ••• G

s::

Ii
CD

..-=
'"
()

0
tI'
0
t""
CIl

ttl

u::l
Ii
SlI

ttl
SlI

'"0

"->
-...J

m
CD

0
rt'

0
ts

-=
tj

t:r

s::

u::l
u::l

u::l
)t
~.

"->
"->

07/23/71

OTACHK

u::l

II
•

OOODCO
OOOOEO
OOOEOO
000E20
000E40
000E60
000E80
OOOEAO
OOOECO
OOOEEO
OOOFOO
000F20
000F40
000F60
000F80
OOOFAO
OOOFCO
OOOFEO
001000
001020
001040
001060
001080
0010AO
0010cO
0010EO
001100
001120
001140
001160
001180
OOllAO
OOllCO
OOllEO
001200
001220
001240
001260
001280
0012AO
OO12cO
0012EO
001300
001320
001340
001360
001380
0013AO
0013CO
0013EO
001400
001420
001440
001460
001480
0014AO
0014CO
0014EO

02010002
600C0774
00F04570
OE3496FO
946F3006
42200048
47700E9E
003AD201
BE749550
9118F006
4780B25C
4890BOFE
955B9000
OF764144
4580B08A
41EOE014
B1244770
88700008
BE24BE25
B02A4880
9286BOEO
B07C4190
5840B120
OEOC95FF
4710BOC6
BE034780
000007F8
800011E4
01565860
6009B101
4770B194
4770B22A
41FF0008
45400EF4
477000BE
A005C080
50800024
47100156
01569101
41900156
013307F9
BE644800
04084770
95600237
10020719
802C5800
00140600
A0009703
OEF01864
10021321
88700008
4170B3E8
023847BO
41110004
01F09228
B5FC4880
01F04590
01524700

30009204
91020C83
OE5C47FO
OE1F4830
1B224320
4820003A
91390045
OOOOOBBA
F0044780
077E95FF
94BFA001
D2079000
47800F6A
01004144
18E94180
15E84740
OF041840
12004780
9200BE25
B0944080
4580B08A
C0004A90
9200BE50
04B14780
OA0748FO
BOF84370
80000058
00000000
BF141266
07F74400
9604AOOF
D200B1F5
50FOBFA4
9140A001
9680AOOO
4400B218
9502BE50
D501A004
00214710
4160BE6C
9601COA4
BOFE1B66
B3081B66
4770B33E
94FEAOOO
800C50EO
60391000
008F4200
47FOOEF4
9012BF14
5E700054
18310500
00C64570
47FOB44C
01C94590
02364188
02E094FO
B4FE9856

000092E6
471000FO
OEOA4580
BE740257
003A4322
9C002000
07779102
920F0073
OECE92C1
F0024770
9202BE50
10001B44
05029000
01004144
9168955C
OF8A47FO
02024001
OFF41810
0201BOE6
BOE61821
5810BOEO
B1004899
07F45080
BOA245EO
COAC95FF
BE244177
48FOCOAC
4570BC70
4750B152
04084780
47FOB194
02379500
1B664111
47100152
92000277
07F70500
47700152
B27E4780
00BE9203
940FOE1F
0201003A
45400EF4
45400EF4
5080BF54
47F00152
800C4570
07F991FO
002007F9
4570B424
58200050
50700054
02370273
BCA407F9
94FEAOOO
02E09530
C0444570
00199140
00500203

PAGE
00018200
96021003
OE1E47FO
BE6CBE82
C09D4570
47400E7C
00440717
82000BB8
BE6BOA03
OEE007FE
92600237
05039000
BE304780
01004144
E00841FO
OF800503
E0109200
1B871A08
B0961B99
4570BCA8
41101000
00000610
020445EO
B10441FO
F0024770
00014270
0704F002
1B004540
186947FO
B1AA4111
5B5BC2C4
00934780
00005510
95000277
92FFBFAO
02370277
9180AOOO
01565880
00939501
07FF9200
30004580
18B09200
50009030
9180F002
02030024
BC7047FO
00214770
91FOO021
18105870
50300050
88300008
078747FO
58610000
47FOB238
02374720
B55647FO
001B4780
0050BEE4

OBB89541
02070040
08B29104
4400B2C8
OA2807F8
90002000
91080044
0501BE4C
9284C094
48600236
4260B101
02784780
OF669510
01004144
01564780
B120B124
BE5007FF
47FOOFF8
4390E013
4880BOE6
4A10BOE6
49190008
OE089190
BE0092FO
B00645EO
02700201
BE1C0704
OEF05860
B15A5860
00005510
E4040740
OEF44570
023847BO
47700156
97030080
478000BE
47800152
00240680
BE6B4720
013358AO
OE4447FO
009347FB
47F00156
07199284
BFC44570
B2149204
00BE1200
477000BE
A0045000
5520BEE4
88200008
00BE91FO
95001000
5810A004
00BC9500
0.OBC47FO
B40A9601
8A500008

30044770
60004580
600A4780
4870BE5A
18284320
47200E70
47170004
BE5A4720
47F00156
4570BC70
41201007
OF6E0501
B1014700
05015040
OFAE0507
477000CO
5880E014
18081817
0201B094
06201A28
12994720
4740B07C
F0064710
04930AOO
B1045880
027FB056
BE1~F002

902C1266
BF184170
023847BO
95000093
BC7048FO
00C60502
92100277
41700156
94FOAOOO
5080BF54
06805890
B20E9284
01F44580
093892F1
00084860
58800024
A0009200
BC7058EO
B38247FO
47AOB39E
4570BC70
70304121
4700B40C
5E300054
00214770
4770B468
50601030
02374780
00BC47FO
C0849601
1B655E60

000E94F3
OE1E47FO
00F04180
4B70COAE
02360922
47100032
47FOOE68
OE0092C4
48FOCOAC
9287AOOO
4570BCAC
00309000
OF625840
BE240207
9000E008
47FOOFCE
88800008
0202BE25
E01A4880
4570BCA8
B01A9500
50190024
BOA29180
92200493
02049203
0204BE23
4400B10C
4780B13E
01525000
00C60507
47800EF4
C09290F1
10000030
47FOB210
0702C08D
58800024
47F00152
A0045080
COA4078F
00EA41AO
BE6B4570
02364570
4B80BOOE
009347FO
BFA45880
B37E9206
03000021
1B444400
00034570
92FOB4E3
1F325030
00BE4111
91806002
07F94700
00BC4140
00BC9218
C08C9180
BEE85060

600C9501
08024400
09E847FO
4070BE5A
50800048
91060045
18264020
BE6B48FO
8000BE39
95000093
5040B120
47800F72
BE3447FO
BOEOBOFO
47800FBC
0503B120
5870E010
E0170200
B0964690
5010BOEO
B1014780
5890A004
F0064710
9180BE02
00939120
027C4160
07FEOOOO
18064170
600C0300
1000B18C
95000277
BFC0180F
477000BE
44000408
A0050702
06800680
9140A001
900C47FO
9209BE6A
COA49280
BC704110
BC704400
50800024
01569180
C08C9001
B3825860
02379706
04084770
BCA4947F
13721B22
005407F9
00005510
4710B470
06049090
BEF44570
01C99090
001B4780
00549200

K•••••••••• W••••
- •••••••••• 0 ••••
.0 ••• *.0 ••••••• 0
••• 0 •••••• K•••••

••••••••••• 3- •••
K•• - •••••• O.K ••
•••• - •••• O••• y.O

...
··...............
·
...............
............
·...............
• •• H ••••••••

••• 8 •••••••• &•••

.iil

•• K•••••••••••••
••• SO •••••• A., ••
•• 0 ••••• 0 •••••••

.~.*

......... -..

••••••••• 0 ••••
N•• < ••••••• D., .0
• •••• O••• O••••••

· - ..............

• - •••••••••• & ••
•••••• N•••••••••
••• 0
•••••• & •• K•••• 0
•••••• N•••••••••
• •••••••• 0 •• N•••

·..........
·• ...............
•••• Z••••• * ••• 0
•••• K••••••• N•••
• $ •••••• N•••••••

••••• y •••• O•• N.
••••• M. K•
••••••••• 4 ••••••
••• •••• • K•• W••••

................

• iil ••••••••••••••
••• ••• • 4& •••••
••• ••• •• • • • ••• • 0
••• F ••• O•••• O•••
• •••• 8 ••••••••••
• •• 8 ••••• 0 •• P. O.

• • • •• • iil& •••••••
• Q •• 0 ••••••• 0 •••
••• 0 ••••••••••••
• O••••••••••••••
•• K••••• K•••• iil.•• P ••• 0 •••••••••
• O. - ••••••••••••
••• - •••••• &. -. L •
••••••••• FN •••••
OMP ••••••• 4 .....
.4 ••••• 0 ••• 1 ••••
• •••• FN •••••••••
••••••••• 0 ••••••
•••••••• P ••••• P •

.....

..... •• w••••••
··...............

• •• u •••••••••••
••• - ••••• &••••• 0
- •••• 7 ••••••••••
••••••••• O•• $$BD
•• •• K•• 5 ••••••••
• ••• &0 ••••••••••
.4.

·.. ........................
••••••••• 7N •••••
&•••••••••••••••
•••• N•••••••••••

.0.8 •••• K••••• K•
•••• K•••••••••••
• •• w•••••••• &•••

• •• w••••••••••••

·••••
...............
&•••• O••• ..

•••••••••• &•••• 0
··...............
.
... .......... ••••• 4 ••••••••••
• •• 9 •••• K•••••••
.• .....
. ........
........ .4• 4&••••
•••
O•••••••••• &•••
··........
., • •••••••••• R ••

• • • 0 ••• 1. , ••••••
-

• - •••••• &••••• 0.
••••••••• O•• K•••
•••••• &•••••••• 0
•• 0.- •••• 9.0 ••••
••••••••••• 9. 0 ••
.0 ••• 0.4 ••••••••
••••••••••• && •• &
• ••••••• &•••••••
••• y •• N•••••••• O
••••• F ••••• 9./ ••
••••• 0.< ••••• 0 ••
• 0 ••• I ••••••••••
• •••••••••••••• 0
• O. •• • • • • •• • •••
•• • •••••• &K •• &. U

• •••••••••• 0 ••••
• O••••••••••••••
••• M••• O••• O••• • ••••••• L•••••••

·• ...............
• &•••••••••••••
••• U••••• O.T ••••
•••••••••• &•••• 9
••• 0 ••••••••••••

......... . .....
·..........
-

•••• &- ••• 9 ••• M••
.4 ••
••• 0 ••• 0 ••••• 1 ••

·• ...............
&••••• -. Y&- ••••

tv
tv

....

ITJ

8CI:I

CI:I

rt
CD
lEI

.'"
n

I'd
11

g
g

CI:I

~....

I'd
11

CD
11

....~

~
~

I-h

......

001500
001520,
001540
001560
001580
0015AO
0015CO
0015EO
001600
001620
001640
001660
001680
0016AO
0016CO
0016EO
001700
01}172 0
001740
001760
001780
0017AO
0017CO
0017EO
001800
001820
001840
001860
001880
001'8AO
0018CO
0018EO
001900
001920
001940
001960
001980
0019AO
0019CO
0019EO
001ADO
001A20
001A40
001A60
001A80
001AAO
001ACO
001AEO
001BOO
001B20
001B40
001B60
001B80
001BAO
001BCO
001BEO
001COO
001C20

Dl'ACHK

07/23/71

B4E3940F B0014830
4140BF14 4570B556
4780B54C 96805002
02764787 00089528
4570BC70 5893C048
47FOB5C6 4570B426
9010602C 10355034
12554740 B5C64121
18538850 00011A45
BF949503 BF904780
BF945870 BF845970
41880001 4280BF90
70085080 400092FF
70075886 000095FF
92007001 45700B84
700847FO 04E80202
95096000 4780B70E
05016006 BF910725
30034177 00014270
58550000 02036012
B7C69101 600C4710
07779102 60000717
60050717 07F49102
91041003 4710B822
45700E54 47F00032
OE12950A 30030724
BOOC5070 B0909208
BE1B47BO B8CA9513
500C4780 B8A29210
B8A69200 B0854180
00320203 6012BE11
47F003F4 92000022
BB249507 40004770
4780B958 0200BB29
910F700B 471000BE
41700000 42750004
70044000 0200700A
4C70BB2A 5A700298
B2389224 BB200203
BB2F4780-B9EE4140
43203005 4C20BE48
47FOOOBE 0500700A
07F99140 BB2F4710
70009180 700B9200
4780BA9C 43407000
B9630200 B9632004
43430000 4A400262
BB2C4770 BAB49200
40007004 47850004
47FOBAF6 00000000
00474570 BCA41810
000407F9 9680AOOO
5840C08C 47F~BB94
58320000 12334780
02024000 20010703
20040200 2004A001
60004860 02368860
BF14BF14 47F00152

02724183 C0449510
47F00152 9607C084
96018000 0707BF14
01C94780 B5809502
02076000 90080223
1B444570 B5F44570
000047FO 01524570
00474570 BCA49001
98564000 07F79056
B23847FO 04829056
BF7C4740 B6525870
020F7008 10004180
40000207 70006000
B6B34770,B6B29200
5870BF84 4B70BF88
BF801009 5860BF8C
95016000 077591CO
58760000 41770000
30039220 B05C9102
50024170 B82C9108
B7F09140 600C0717
9180600C 07149102
60004710 B7F64550
96081003 47FOOOFO
47FOOE12 4550B8B6
47FOOE12 95013003
B0904550 B8B29101
BE1B47BO B8CA5870
B0859506 50184770
B0684320 600047FO
07F59620 10030600
0300BB2E 30049560
048C5844 00001B55
50044870 BB284550
4340700B 41440001
4C70BB2A 5A700298
023747FO 048C95FF
0200B963 700047FO
BB241002 4540B80C
BB3047FO B9FC0202
5A200200 95FF2004
02374770 00BE910F
BA7C4340 B9630200
700A9200 700B4710
02007000 B9630200
92FF2004 47FOBA62
95804000 4780BAB4
500A9601 40004190
94BFBB2F 95FF7000
OOOOOOOC FF230000
18204570 BCA44840
9101AOOF 4710BC28
9563A001 4780010C
010C5852 00040207
20002000 0703402C
9200A001 9601AOOF
00014A60 02700707
94FEAOOF 48200236

PAGE
800147BO
9601C084
BF1447FO
BE504770
6008902C
BC700203
B4264140
400007F9
BF944550
BF944550
BF781887
70004080
43860007
B6B30211
02017015
5080BF8C
60040775
4A760006
600A4710
60050717
9108600C
600C4710
B8B20501
9180600C
95FF3003
072447FO
BE104780
BE184177
B8A64570
00FE9240
10036000
BB2E07F4
43503005
BAF647FO
4240700B
0200B963
B9634780
B9769680
02031002
BB251009
478000BE
700B4780
B963BB29
BAB095FF
B9632004
4130BEF7
91014000
015247FO
07850200
00000000
02368840
9550A001
584A0004
50004008
402C0200
48600236
60006000
88200001

01529856
96FOB001
01521255
B5800500
021B602C
900C6004
00014570
4A400022
B6E69856
B6E69856
4A80BF88
70129200
06804280
70188000
80029240
18764A70
9115io02
59700238
B76AD202
9110600C
4710B80C
S8369160
BE186012
0714950A
0784910F
OE124400
B87A9562
00015070
OB849240
B05C4180
47FOOOFO
4770048C
4C50BE48
B98E0500
47F0048C
70005010
B9B60200
700B94FE
BB249223
5840BB24
43720004
BA424360
5860BB24
70004780
42420004
41330001
4710BAB4
BAB44C70
BB297000
00001831
00015A40
4780010C
48200236
02235008
402FA001
88600001
05010272
5A2002AO

BF141266
47F00152
07079601
B1010273
90105050
0223902C
B5F41330
4840425B
BF9447FO
BF9447FO
5080BF84
70119601
B6B34188
41807015
70179680
BOOC5970
07759120
072547F5
BE191009
07179108
4170B822
60004770
4770B7F6
30030724
30034780
00F05870
30044780
BE184570
B0859180
B0584570
4140B8F4
0202BB25
5A500200
700A0237
95FFB963
700092FF
7000B963
A00092FF
BB204770
58440000
9640BB2F
700B0660
02006000
BAB007F9
47FOBA62
95003000
58504004
BB2A5A70
5070BB24
12004780
02A05004
44000408
88200001
402C021B
943F402F
4A60026C
02364770
0200A001

4740B53C
95808000
80004930
47870004
900Cl135
6008021B
4780B5C6
48300236
04829856
04E85850
4380BF90
70144180
00044280
40807002
10024110
02380725
100C0775
00084370
5850BE18
60004710
9124600C
B8CE9101
9680600A
4180B050
B7FC47FO
00404B70
B8669509
OA2891FO
800C4780
OE5447FO
92FF0022
10095840
95FF5004
4770B9B2
4780B9B6
70000205
4870B962
002247FO
00BE9180
1B771B22
4550BAF6
4260700B
70004240
95FF7000
02007000
07891B44
0501500A
02980505
4870BB28
BB4E4121
00005034
4770BB88
5A2002AO
502C4010
02034030
07076000
BC240707
20049640

• T •••••••• ,••••••
••••••• 0 ••••••

••• <•• , ••••• P •••

• •••••••• I ••••••
•• •• •• •• K. - ••• K.

• O.F ••••••••• 4 ••
•• - ••• ' •••• 0 ••••
• ••

• F ••••••••••

••• , •••••••• 7 ••
• •••••••••• O••••
• •••••••• iii. • •••
• ••••••• K •••••••
•• , • • • • • K ••• -.

• •••• O••• O••••••
••• 0 ••••• P ••••••
• ••••• N •••••••••
- ••• K. - ••• " ••••
•• K ••• -.K ••• -.K.

•••

• •••• 4 ••••• F
~. 9 • • • • • $ ••••
• •• ,.W ••••• 0 ••••

••• ,.W••••• O.Y.,
........
,...... .
•• K ••••••••••••

,.............. .

K.. •••• • •••••••

·.- ....... - .... .
• .'. O. YK •••••• - ••

N. - •••••••••••••

.••..•••.• ... * •.

•••• K. -. , •••••••

.F •• - •••• O. - •••• 4 •• - •••• 6.'
••••••••••••• 0.0

.... ..... .....

••••• 0 ••• 0 ••• ' ••

.. ,........ ,... .
,........... ,.. .
•••• •••• • 0 ••••••

• •••••••••• 5 ••••
- ••••• K •••••• , ••

.. - ........ ... -.

- ..... .

.. - •........... ,
•• N ••• - •••• 6 •• -.

• •••••••••••••• 0
••• 0 ••••• O. •• • •

•••••• , •••••••• 0

•••••••••••• - •• 0

.* ••..•.••• 0

•• K.- •••• 5 •••• 0.
• O. 4 •••• L •••••• -

•• - •• 0.0 • • 4 ••••

• ••• K ••• , •••••• ,

•••••• •• <•••••••
••

••• 4 •••• K ••••••

·,.. <, ... ,.... ,.

.6.0 •• N•••••••••
• ••• 0 ••••••••••
K ••••• , ••••••• K •

• K •••••• 0 ••••

• ••••• K •••••••••

·..... K...... ..

• •••••••••••••• 0
K •••••••••••••••

•••• <•••••••••••

.. . ... . .. . .. '.6

<••••••• K•••••• O
•• ••••• • •• 0 •• K.

• 0 •• N •••••••••••

.9 •••••• IiI • • • K.

••••• - •• K.- ••••

•• K ••••• K.
•• K •••••••••• 0 ••

••• 7 ••••••••••••

• •••••••••• 9 ••••
••••••••• O•• K•••

·....... ,. .. ...

•••••••• ,

• O. 6 ••••••••••••

··..,.............
... .. ,... ,..

••• 9 ••••••••••••
• ••• O••••••••••
•••••••••••••• K.
K. • •• P ••••• P.
•• K •••••••••••••

- •• - ••• - ••• - •• P.
••••• 0 ••••••••••

• N. ,.

••• O•• <••••••• N.
•• K••••• ' •••••••

,. .K. , • • K. ,.
.K • • • • • • • K.
• - ••• - ••• - •• P.- •

-.-.N ••••••••• P.
•••••••• K ••••••

....
~

OTACBK

I.Q

..
-=-

\£)

g
en

~.....

m
I'd

I.Q

11
SlI

~
tv

....a

en
(1)

....~
o

::s

-=~

~
~

I.Q
I.Q

....
::s

I.Q

....)I
01

tv
tv

....a

001C40
001C60
001C80
001CAO
001CCO
001CEO
001000
001020
001040
001060
001080
0010AO
0010CO
0010EO
001EOO
001E20
001E40
001E60
001E80
001EAO
001ECO
001EEO
OOiFOO
001F40
001F60
001F80
001FAO
001FCO
001FEO
002000
002020
002040
002060
002080
0020AO
0020CO
0020EO
002100
002120
002140
002160
002180
0021AO
0021CO
0021EO
002200
002220
002240
002260
002280
0022AO
0022CO
0022EO
002300
002320
002340
002360
002380

A0015832
021B4010
90246000
A00007F7
472000C6
COB20985
4700B502
OC000800
08002F20
16001E18
IFOOI085
IFOOIE3C
080010CO
86002B50
00008000
00000000
00180000
12C40156
00090000
00000000
00000000
615C44FO
00000000
00000000
00001000
00030000
FF050007
00002B50
00000005
00000000
2720001F
02080000
240B1480
023FOAOO
OCI01418
22682268
134C1420
00BE1432
00002F08
OAOOOOOO
FFFFFFFF
OOOCFFOO
001FFFOO
0133FFOO
0137FFOO
0183FFOO
0192FFOO
0282FFOO
FFOOOOOO
00060000
OOOEOOOO
00160000
OOFFOOOO
00000050
00001900
00000000
00000051
00001300

07/23/71
00044400
302C0223
90586020
48600236
4910BE3A
19684770
4110BE28
00000000
60000001
20000004
60000001
40000001
00000000
60000308
OC000006
13000101
00000000
5B5BCI05
00000000
--SAME-00000000
FFOOOOOO

04084770
402C3008
90606040
41220000
474000C6
00C607F7
9283AOOO
01001035
04001E78
07001E16
lAOOlE10
08003B38
07001E21
9200027C
000010C8
5B5BC2Cl
00050000
C5090900
00000000

PAGE
BC565840
47FOOI0C
90686060
59200238
IB881851
9510AOOl
9103A008
60000003
20000006
40000006
50000005
lE3C1E30
40000006
20000005
000010FO
E3E305Cl
lE6COE12
000010FO
00000000

C08C47FO
9180AOOO
9070D-213
472000C6
4350COB2
47B0017E
4770B166
080020FO
13001EI0
39001E18
08001090
070010BO
3io01E23
86003008
0000Q400
C6C70700
OOFOOOFE
OE000005
00000000

BCSA584A
07175890
901001FO
12660787
09851968
92E9BE6B
9280AOOO
60C2C740
20000001
60000004
030806CO
40000006
40000005
60000168
00000000
01000001
211C12AE
22100400
00000000

00040207
A0044880
02079008
41110000
477000C6
9110A008
9200A008
C3001038
lAOOlE10
08001070
07003B51
IFOOIE3C
08001000
OOOOOOOQ
00001DC8
00FE32AO
lEC41E56
002000C8
00000000

40083000
01C890EO
80009680
59100238
18524350
4710B20E
47F00156
60000001
20000005
00000000
60000006
40000001
0008000A
1E003000
FF006000
10401080
000012B2
00000000
00000000

00000000 00000000
OOFFOOOO 01008484

00000000 00000000 00000000 00002F20
80808083 85605040 30201000 00000000

00000000 00000000
00003000 900010AE
00680000 00000001
00003228 00002054
00002F08 00000020
00000000 00000006
00000000 00000000
00000000 00000000
47FOOIE8 00000000
1907BOOO 0507BOOO
47FOBCEC 00003228
47FOOICA 22100016
22682268 22682268
137AOOCE03E815D2
144COOBE OOBEOOBE
04000000 05000000
FFOOOOOO 03030500
FFFFFFFF FFFFFFFF
OOOOFFOO 210000FC
0130FFOO 620000FB
0134FFOO 620400F8
0180FFOO 50930190
0184FFOO 50C300C4
0193FFOO 600BOOFB
0283FFOO 50C300CO
FF020000 0005FFFF
00080000 00090000
00100000 00110000
00180000 00190000
00000000 00000000
00001900 00000000
00000000 00000000
00000050 00000000
00001900 000001F4
00000000 00000000

00040007
00001EOO
00005000
00000208
00002B08
00000000
02080000
27·B00034
OOOOOOFF
012AOAOO
07002B50
IF050000
11281166
15A815CE
01B601B6
06000000
00000000
00000618
OOOEFFOO
0131FFOO
0135FFOO
0181FFOO
0190FFOO
0194FFOO
FFOOOOOO
80FFOOOO
OOOAOOOO
00120000
001AOOOO
00000000
00000000
00000079
00001900
00000000
00000050

•••••••••••• ~..

• •• O•••••• K.

K • • • • K • • • • • 0 ••

- ••••••••••••• B ••

••• 7. - ••••••••••

• •• F ••••••••••••

.. - ... - .. -- .. --

...
F..... . F....
••••••••• F. 7 ••••
.. ..........- ...
.... - .......... .
·....
....- .......
....... ,......

•• K•••• OK •••••••
• i ••••••••• F ••• ,

• •••• Z••••••••••
••••••••••••• 0 ••
••• O-BG C••• - •••
• •••• Q•••••• - •••

••• i- •• Q ••• lil ••••

. ... - .......... .

•••••••• $$BATTNA

••••••••••• B•• -.
FGP ••••••••••
• O••••••• O ••••••

· .0 •..••.... B••• 0
.O •• $$ANERR •••• O

••••••••••• R••••

/ •• 0 ••••••••••••

••••• -i ••••••••

--SAME-~

00000000
00002000
000024A8
40002E06
00000007
000002E4
00000000
00000000
8500E207
8000C6F2
8000CI09
8400E207
22682268
I1FC1372
OOBEOOBE
030045F8
OBOOOOOO
030305FF
110000FC
000000F8
620300F8
620700F8
50C300CO
600AOOFC
50C300CO
00000000
00070000
OOOFOOOO
00170000
00000000
00000000
000001F4
00000000
00000000
00000700

400010BC
00002B50
00002140
00000001
00000008
00000000
00000000
0000000-0
0310C2C7
8000C6Fl
84006150
21782178
128212F6
lSA41502
180819C2
07000000
00000000
FFFFFFFF
420000FC
620100F8
620500F8
50C300CO
600800FC
600COOF8
00000000
00030000
OOOBOOOO
00130000
001BOOOO
00000000
00000000
00000000
00000000

00000000
00002308
00000000
80002C4C
00001FAO
00000000
00000000
00000000
19001F98
19070800
47FOOEBO
220821A8
120600C2
15A8138A
IB360000
08000000
FFFFFFFF
0009FFOO
OOOFFFOO
0132FFOO
0136FFOO
0182FFOO
0191FFOO
0281FFOO
FFOOOOOO
00000265
OOOCOOOO
00140000
001COOOO
00000000
00000051
00001300
00000000
OOOOOOO~ 00000079
00000000 00001900

00002054
0000250F
00000000
90002CFE
00002B6A
00000000
00000000
00000000
OA0032AO
05070800
50000408
22382238
1314135A
13AC13C8
02004478
09000000
FFFFFFFF
000000F8
420000F8
620200F8
620600F8
50C300C4
600900FC
50C300CO
00000000
28FFOOOO
00000000
00150000
00100000
00000000
00000000
00000700
00000000
00000000
00000000

•••••.• i ••• Q••••
• • i •••••••••••••

••••••••••• < ••••
••• i ••••• ~ ••••••
• •••••• U ••••••••

• •• Q••••••••••••

·• •••••
..........
.... .
BG ••••••••
'

• ••••• SP. O. Y••••
••••• ~ F2 ••••••••
• ••••• AR. O••••••
•••••• SP.O ••••••

• <••••••••••• Y. K
••••••••• <••••••

• ••••• Fl •• Q ••• Q.
• •• i •• /i. O•• i •••
••••••• 6.0.B ••••

••••••• K••••••• R
• •••• Q. B ••••••••

••••••• 8 ••••••••
••••••••••••••• 8
••••••••••••••• 8

••••••• 8 ••••••• 8
••••••• 8 ••••••• 8
••••••• 8 •••• i •••
•••• iC •••••• iC.O
• ~ •• - ••••••• - •• 8
•••• iC •••••• iC ••

••••••• 8 ••••••• 8
••••••• 8 ••••••• 8
• ••• iC •••••• iC.O

... .- ....... -.. .
•••• - •• 8 •••• iC ••

••• i ••••••••••••

••••••• 4 ••••••••
••••••••••• i ••••

••• ~ ••••••••••• 4
••• i ••••••••••••

"IiI

IV
IV

~(I)

8en

.r=
\Q

CIl

t+
(I)
S

"tI

I.Q

til

~....

(I)

It!

(j)

I.Q

0,
(I)

~
:+
IV
IV

...,J

...-..

PAGE

07/23171

DTACHK

I.Q

0023AO
0023CO
0023EO
002400
002420
002440
002460
002480
002500
002520
002540
002560
002580
0025AO
0025cO
0025EO
002600
002620
002640
002660
002680
0026AO
002720
002740
002760
002780
0027AO
0027CO
0027EO
002800
002820'
'002840
002860
002880
0028AO
0028CO
0028EO
002900
002920
002980
0029AO
0029CO
0029EO
002AOO
002A20
002A40
002A60
002A80
002AAO
002ACO
002AEO
002B4"0

00000000
00000079
00008000
40C3D6D4
00000000
09002455
E3E3D54B
00000000
00000000
00FF1504
FFFFFFFF
FFFF13FF
FFFFFFFF
FFFFFFFF
00000000
00000000
08000000
00000000
06000000
10000000
00000000
00000000
FOF761F2
0007D7FF
21782269
232A1E4E
00000000
00000000
0007FFFF
F7F2F3F7
00000000
000025AC
00000000
00000000
372D2E2F
5B6c507D
C4C5C6C7
84858687
3F3F3F3F
3F3F3F3F
lA091A7F
lA0A171B
lA1A1A1A
lA1A1A1A
64656667
75767778
444546.47
55565758
FFOOOOOO
OOOOOOFF
00000000
00000000

00000000 00000051
00000000 00001300
08000004 000023D8
D4E4D5C9 C3C1E3C9
--SAME-20000015 00008000
40F040FO C34BOOOO
--SAME-00000000 00000000
13FFFFFF FFFFFFFF
FFFFFFFF FFFFFFFF
FFFFFFFF FFFFFFFF
FFFFFFFF OOFFFFFF
FFFFFFFFFFFF0014
03000000 00000000
00000000 06000000
00000000 00000000
OBOOOOOO 00000000
00000000 OEOOOOOO
00000000 00000000
13000000 00000000
--SAME-F361F7Fl 32A03000
00003Fl0 00003Fl0
226AOOOO 25102514
1EF41F04 1F140020
02080000 0000283C
00000000 D5D640D5
F875ECDl 8000EC50
F1F2FOF4 00002044
00000000 00000000
00000044 00001F2C
00000000 0000289C
00000000 00000000
1605250B OCODOEOF
4D5D5C4E 6B604B61
C8C9D1D2 D3D4D5D6
88899192 93949596
--SAME-3F3F3F3F 3F3F3F3F
lA1A1AOB OCODOEOF
lA1A1A1A lA050607
lA1A5B2E 3C282B21
1A1A7C2C 255F3E3F
68691A1A lA1A1A1A
797A1A1A lA1A1A1A
48491A1A lAiA1A1A'
595A1A1A lA1A1A1A
OOFFOOOO OOOOFFOO
00000000 FFOOOOOO
--SAME-00000000 00000000

LBLTYP
--BG-003220
003240
003260

HEX LENGTH

00000000
00000700
000023EO
D6D5E24B

00001900
00000000
F1C9FOFO
00000000

000001F4
00000000
C14040D9
00000000

00000000
090023ED
C5C1C4E8
00000000

00000000
20000023
40C6D6D9
00000000

08000004
OOOOOOO~

00002440 00002448 F1C3FOFO C14040Cl
00000000 00000000 00000000 00000000

00000000
FFFFFFFF
FFFFFFFF
FFFFFFFF
13FFFFFF
01000000
00000000
00000000
09000000
00000000
00000000
11000000
00000000

000B2A3F
14FF15FF
FFFFFFFF
FFFFFFFF
FFFFFFFF
00000000
04000000
00000000
00000000
OCOQOOOO
000000.00
00000000
FFOOOOOO

OB140AOA
14FF15FF
FFFFFFFF
FFFFFFFF
FFFFFFFF
00000000
00000000
07000000
00000000
00000000
OFOOOOOO
00000000
00000000

01FF01FF
FFFF03FF
FFFFFFFF
FFFFFFFF.
FFFFFFFF
02000000
00000000
00000000
OAOOOOOO
00000000
00000000
12000000
00000000

02FF12FF
FFFF100l
FFFFOOFF
FFFFFFFF
FFFFFFFF
00000000
05000000
00000000
00000000
ODOOOOOO
00000000
00000000
00000000

00000000
00000020
25183CFO
214C0020
00000000
C1D4c540
00002171
OOOOOOOC
20000000
00000000
00000000
00000000
10111213
FOF1F2F3
D7D8D9E2
979899A2

00000000
0007FFFF
F1F2F3F7
00000000
FOF761F2
0007FFFF
21782269
23721E4E
00000000
-00000000
000.025AC
00000000
3C3D3226
F4F5F6F7
E3E4E5E6
A3A4ASA6

00000000
F875ECD1
F1F2FOF4
00000000
F361F7Fl
00003Fl0
226AOOOO
lEF41F04
02080000
00000000
00000044
00000000
18193F27
F8F97A5E
E7E8E94A
A7A8A9CO

D5D640D5
8000EC50
00002044
00000000
32A03000
00003Fl0
25102514
lF140020
0000286C
00000000
00001F2C
0000289C
lC1D1E1F
4C7E6E6F
E05ASF6D
6ADOA107

C1D4C540
00002171
OOOOOOOC
20000000
00000000
00000030
25183CFO
214C0030
00000000
00000000
00000000
00010203
404F7F7B
7CC1C2C3
79818283
'3F3F3F3F

3F3F3F3F
10111213
lA1A161A
261A1A1A
lA1A1A1A
lA6A6B6C
lA1A1A1A
7D4A4B4C
30313233
OOOOOOFF
OOFFOOOO

3F3F3F3F
lA1A081A
1A1A1A04
1A1A1A1A
1A1A1A1A
6D6E6F70
1A1A1A1A
4D4E4F50
34353637
00000000
OOOOFFOO

3F3F3F3F
18191A1A
lA1A1AlA
1A1A5D24
lA603A23
71721A1A
lA1A1A1A
51521A1A
383.91A1A
FFOOOOOO
OOOOOOFF

3F3F3F3F
1C1D1E1F
1415lA1A
2A293B5E
40273D22
lA1A1A1A
lA1AlAlA
lA1A1A1A
1A1AlA1A
OOFFOOOO
00000000

00010203
1A1A1A1A
201A1A1A
2D2F1A1A
1A616263
1A7E7374
7B414243
5C1AS354
00001300
OOOOFFOO
FFFFOOOO

.................
• ••••• ••••• Q••••

••••••• 4 ••••••••
1I00A

READY FOR

COMMUNICATIONS.
•••.•••• 1COOA

TTN. 0 OC •••••••

07/23/71 ••••••••
•• P •••••••••••••
••••••••••••••• 0

••••• 4 ••••••• <••

•••••••• NO NAME
•••• 8 •• J ••• i •.••
72371204 ••••••••
• a'e • • • • • • • • • • • • •

07/23/71.: ••••••
•••••••• NO NAME
•••• 8 •• J ••• , ••••
72371204 ••••••••

$., ...•.

,-./0123
DEFGHIJKLMNOPQRS

••••••••••••••• 0
• •••• 4 ••••••• < ••

456789 •• <••• aABC
TUVWXYZ •••••••••

·••••••
..............
.
$ •••••••••
.... - ...... ..
·............. , . .................
•••••••••••••.•• < .. ,..•.....•...
•••••• i •••••••••

00000000

0000

D5064005 C1D4C540
000032AO 00003880 00003688 0000338A
00000001 0000338A 50003C12 00003388

FF150007 E00039C2 00003BE2 000032AO
000041B8 00003850 00003960 00000001
00003550 000035B8 000050B1 00C2DF17

·.............. . -... ,......
NO NAME

•••••• • B•••

s ....

•••.•.• i •.. - .••.

•••••••• i •••••••

'~.B

I'Ij

1-1'

OTACHK

I.Q

c::

t1
CD

.:::

0
t:I:I
0
I:"i

SlI

~.....
CD

"'C
t1

a
I.Q
t1
SlI

"'C
SlI
t1
IT

'"W
a

'"....,
..,J

CD
0
IT
1-1'

a
::s

.:::
t7

c::

I.Q
I.Q

1-1'

::s

1-1.

0..

'"
ID

003280
0032AO
0032CO
0032EO
003300
003320
003340
003360
003380
0033AO
0033CO
0033EO
003400
003420
003440
003460
003480
0034AO
0034CO
0034EO
0036CO
0036EO
003700
003780
0037AO
0037cO
0037EO
003800
003820
003840
003860
003880
0038AO
0038cO
0038EO
003900
003920
003940
003960
003980
0039AO
0039CO
0039EO
003AOO
003A20
003A40
003A60
003A80
003AAO
003ACO
003AEO
003BOO
003B20
003B40
003B60
003B80
003BAO
003BCO

00000000
05F00700
FFFFFFDC
00005218
00000000
00000000
47FOFOAC
000032AO
E3C5E2E3
E6E7E8E9
el000000
01010014
00003430
00000000
00000014
00000000
10003C50
41EOEOOl
00640063
00000000
00000000
00000000
00000000
00000000
00003550
000032AO
00010000
00000000
00000000
00000000
00000000
000049E8
00003A5E
5B5BC2D6
E6D609D2
58FOC004
4500FOOC
5870D1BC
4830C03A
C03A4A30
48206000
C03e~140

E0009240
4004F7F2
60385810
C004051F
60004930
00014004
4500FOOC
5000F008
58FOC004
41FOC028
58FOC018
D500C058
412001EC
FFFF5810
18011840
C0500A02

07/23/11

PAGE

--SAME-900EFOOA 47FOF082 000032AO
000032AO 00003238 00004DB6
00000208 00000000 00000000

000032AO 0007AFFF 000032AO 0007AF24
OA0407F1 000032BO 000032BO 00004218
00000000 00000000 00000000 00000000

--SA~£--

58COFOC6
98CEF03A
000032AO
D9E4D540
0001FOFl
000040D5
00000000
00000000
86BCF018
000035B3
00000000
1168E2E8
58201044
00000000
--SANE-00000000
00000000
--SAME-00003958
000032AO
000032AO
00003958
--SAME-00000000
00003550
0000338A
000041B8
00003A72
07C50540
60D9C5C3
051FOOOl
000033F8
02016000
4A306000
601C4E30
4C20C03A
601E4820
60444830
40404000
01C81841
00014004
C03E58FO
F7F74040
00000000
4500Fooe
051F0001
02021025
07FF5810
60430772
58FOC004
C01407Fl
4110C048
OAOEOAOE

58EOCOOO
90ECDOOC
00003880
0001C1C2
F2F3F4FO
E8C34000
00000000
10003C50
~lEOEOOl

00640063
00000000
E2FOFOF8
020035B8
00004A06

58DOFOCA
185D989F'
00003688
C3C4C5C6
F1F2F3F4
00404040
00000000
1160E2E8
58201044
00000000
04000000
40400272
00000064
000049FO

9500EOOO
FOBA9110
000038EC
C7C8C9Dl
FOF1F2F3
40404040
10000000
E2FOFOF8
010034E8
00000000
00008200
00000000
00003620
00000000

4770FOA2
D0480719
00003BC8
D2D30405
F4FOF1F2
00000000
04000000
40400162
20000064
000049FO
00000108
20000000
00000000
00000000

9610D048
07FF0700
C3D6C2C6
0607D809
F3F4FOFl
F1F2F3C4
00009200
10000000
00003550
01010014
000034A8
00000000
00000014
00000000

92FFEOOO
00003BE2
FOFOFOF1
E2E3E4E5
F2F3F4FO
00000000
00000108
04000000
00003550
00000000
00000000
86BCF018
00000000
00000000

00000000 00000000
000033F8 00003550

7000004B 00000000 00000000 000038EC
000032AO 000033F8 50003C12 00000000

00003550
000033F8
00003880
00000000

01000000
50003C12
00003960
00000000

70003934
00003388
000041B8
00000000

000041B8
00003550
00003850
00000000

00003850
000035B8
00004708
00000000

00003960
00003BE2
00003550
00000000

00000000
000035B8
00000000
00004990
00003B26
5B5BC2C3
0609C400
4004F6F3
OA024100
C038D201
4E300100
0100D705
1A425B40
60004C20
601C4E30
58FOC004
58F01010
F7F44040
C024078F
40005810
OA025800
00000000
4004F8FO
F00158FO
C01C07F1
95406044
051F8001
58FOC004
070005FO
50D05008

00000000
00003388
00000000
00003950
00003B58
D3D6E2C5
58FOC004
40404000
D1C858FO
601CC038
07050100
01DOD100
C0385040
e03A1A42
D1DOF331
051FOO02
45EOFOOC
40005800
5810COOC
D1C894EF
D1C84110
OA024100
40404000
101045EO
58FOC004
07720200
1000000B
051FOO01
5000F008
5050D004

00000000
000033F8
00003A3E
00003A3E
00003A3E
5B5BC2C6
051FOOOl
4l10C040
C00805EF
58FOC004
Dl00940F
940FD1D6
D1DC58EO
5B40C038
603AD1D6
00000014
5020D1BC
D1E85000
07F15800
10201801
C0500A02
D1CC58FO
5810D1CC
F0085020
051F0001
6043C059
OC000060
4004F8F3
4500FOOC
5820COOO

00000000
00003470
00000000
00003AEO
000032AO
C304E4D3
4004F6FO
5800D1C8
5810D1C8
051F0001
D1D64F30
4F30D1DO
D1DCD200
504001EO
96F0603D
000001C4
587001BC
D1E45800
D1E45000
18404110
4l10C040
C00805EF
58FOC028
01C05880
4004F8Fl
92406044
00000000
40404000
00000000
95002000

000032AO
00000000
00000000
00003B2C
00000001
FOE90000
40404000
184005FO
96101020
4004F6F7
D1D04030
4030601C
6038EOOO
58EOD1EO
58FCC004
0038FFFF
5810D1E8
C0205000
D1E858FO
C04805FO
5800D1CC
5810D1CC
91201010
D1COD213
40404000
5810C05C
00140DOO
58l0D1CC
OA025800
077992FF

00000000
0000001C
00000000
00003B88
lCOOO01A
COOOOOOO
9640D048
5000F008
5020D1BC
40404000
60004830
41406002
K&306050
D2006043
051FOO01
D2137000
07F158FO
D1E84830
C004051F
5000F008
184005FO
96101020
071F1841
60388000
5820C02C
5010D1EC
01C40038
94EF1020
01CC4110
20009610

• 0 •••• 0 •• 00 •••••
·...............
................ ••• 1 ••••••••••••
· ............... ................
•••••• OF •••••• O.
.00 ••• 0 •••••••••

................

TESTRUN •• ABCOEF
WXYZ •• 0123401234
A••••• NYC ••

• ••••• 0 •••••••••
0 •••••••••••••• S
••••••• HCOBFOO01
GHIJKLMNOPQRSTUV
0123401234012340
•••• 123D ••••

·............... ·...............
..........
••••••••••• &.-SY
•••••• 0 •••••••••

·...............

S008
• •• Y••••••• &••• &.
••••••• 0 ••••••••

••• &•• SYS008

• ••••••••••••• O•

................ ................
................ ................

• •••••••••••••• 0

.................

................

·•••••••
...............
8& •••••••

• •••••••••• &••••
••• &••••••• 8& •••

••••••••••• &••• • •••••• &••••••• S
••••••• &••••••• &

••••••••••• 8 ••• &

...............

·................. ................
................
·••...............
• • • • • &••••••••
••• 8 ••••••••••••
·•••...............
·
...............
Y••••••••••• & ................
................
·$$BOPEN
...............
$$BCLOSE $$BFCMULOZ ••••••
WORK-RECORD •• 0 ••
• 0 ••••••• 63
•• 0 •••• 8 •••• JH. 0
•• J.K.- ••• K.- •••
•••••• - ••• J.P.J.
•••• - ••• J.P.J.J.
•• -.< ••••• $ •• &

••• - ••• -. <•••••

••• - ••• - ••• J.3.
.72
•• 0 ••••••
- ••• JH ••• 0 •••• O.
.74
- •••••• 0 ••••••••
.77
••• JH ••
•• 0 ••••••••• JH ••
&. 0 ••• 0 •••••••••
• 0 •••••• .80
.O •• K••• O•• O••••
• 0 ••••••••• 1. 0 ••
N••• - •••• - ••• K.
• .J •• 0 ••••••••••
••••••• 1. 0 ••••••
•••••••••• 0&. O.
• &•••••• &. &. && ••

·.

.... .60
...
•. JH •

.0&. O.
•••••. JH •••• &.J.
• 0 ••••••• 67
J ••• JO •• J.
..JO •. J • • - •• -.
J ••• J.K. - ••••• -&
$ •• & J ••• J. K. -.
-.JO.O- •• O••••••
••••••• D•••• K•••
&.J ••• J ••• JY.l.0
JY&.Ju •••• &.JY ••
• 1 •• JU&. JY. 0 ••••
••••• 0&. O.
• &••••• ..J•• .0
J •• 0 •••••• J •••••
•• J •• 0 ••••••••••
O.&.J ••• J.K.- •••
••••• 81
- •••• - •••• *&.J.
••• - ••••••••• D••
.83
••• J •••••
• .0 ••••••••• J •••

...

.....

·...............

"->

CIl

C/)

..,.I'Zj
~

"d
11
0

0
1:4

CIl
III

i
.

en
..,.s::

C/l

I'd
11

0,
(I)

11
III

,..
I'd

III
11

"->

-.J

003BEO
003COO
003C20
003C40
003C60
003C80
003CAO
003CCO
003CEO
003DOO
003D20
003D40
003D60
003D80
003DAO
003DCO
003DEO
003EOO
003E20
003E40
003E60
003E80
003EAO
003ECO
003EEO
003FOO
003F20
003F40
003F60
003F80
003FCO
003FEO
004000
004020
004040
004060
004080
0040AO
0040CO
0040EO
004100
004120
004140
004160
004180
0041AO
0041CO
0041EO
004200
004220
004240
004260
004280
0042AO
0042CO
0042EO
004300
004320

D04850EO
D0484160
418001BC
D1C45810
41FOF21A
103C4510
F0369011
4510F260
OA014200
07FE9104
101541EO
OA009180
5BB01038
48B01000
4170F140
96801015
OA0001F7
F1B89108
5810104e
F1F6940F
OA320000
OA3298Fl
F35805EE
47EOFOE8
41EOF296
F2609120
104e58EO
41BOF2FA
102B4400
00000000
5B5BC2C5
F1CC9104
102B4710
4780F068
94BF102C
91801002
07F31831
1030024F
4710FllE
102C4710
1050D500
19434770
10364720
1052F1C8
OA021812
F9C3F14E
C4E207FO
416D007C
4710F08A
9140D048
F126D200
58440000
00608950
4A510002
4710F116
DOFC4580
015e4190
F53C9200

PAGE

01/23/11

DTACHK

D05405FO
00044110
4110D1CF
01BC5880
41FOF052
FOCA9871
F32c58CO
91801002
10384800
10154710
FODE819A
10024110
41BBOOOO
58A01048
41E0105C
948Fl005
189B4B90
1015071E
41110001
10029108
OA320000
E12447FO
98F1EOF8
OA0047FO
OA029801
10154710
F35C94EF
58E01024
103050EO
--SAME-09D9E305
1010071E
F13C4140
4640F05A
4530F1A8
4710FOAA
4110FOOO
10483000
9106102A
F1B69108
10401048
F1124202
FOB20200
4770FOEB
47FOF132
E5E660FO
F3F390Ee
5810F7BC
lBEE43El
4780F1FA
00601004
4A410008
00088850
48550000
12714780
F2004700
007C1989
F53C5010

9120D048
COOC4110
05105800
D1C058EO
41FOF260
F32C4400
10445BCO
4110F01E
10000AOO
FOC2D703
F01B41FO
FOFCOA01
91041015
10AA12BB
91201002
58B01038
10545090
941Fl015
5010F358
10154110
OA329110
E03098Fl
9070F32e
FOE89610
F3509110
Fl0441FO
102401FE
07FE960B
102847FO

41EOF016
C0380610
80001EOB
D05407FE
47FOF114
10349101
10404180
OA010201
91801002
104Cl04C
FOE2B69A
91011004
4110F194
4780F154
4110F1AA
41BBOOOO
10444BBO
58EEOOOO
5810F354
F21A9512
10024180
E1249070
5870F328
10249001
10244710
Fll0910C
91011015
10249108
F2865B5B

5800B048
05505840
50008000
41FOF088
C9D1C6C6
1015018E
F0369180
103El050
4710FOA6
07FED703
F01B5090
4710F20C
50B01044
96401005
41E01058
D2021039
10504ABO
5010F328
41FOF1EA
10144770
F24E4110
F32C5870
91801015
F35050EO
F2BE907C
101547BO
4710Fl04
10154710
C2c3C5D6

9820B050
10001E4B
87861000
47FOFOB2
C2E9E9D5
OA099011
10024710
41FOF036
OA074000
10441044
104407F1
58E01028
4AB01052
12BB4740
941Fl015
104150BO
100041FO
981DF32C
D203F358
F20A58EO
F35405EE
F3289680
4710FOE8
F35C94F7
F32e5870
F2D2D703
5070F328
F28A9120
E5Fl0000

58E00054
50401000
0203D1E8
41FOF046
F3F89011
F32C4510
F06COA01
91801002
1000D200
07FE989B
91201015
44001030
4BB01000
F154948F
91401005
10409120
F12C9541
91041015
10449110
10684110
47FOEOOA
101547FO
96801015
10241801
F3289108
10441044
987CF32C
10154710
00000000

01FE9620
81165000
C0305860
47FOF02A
F32c9620
FOCA41FO
40C0103E
4110F092
1038101E
10449104
4710FOF2
50E01028
50B0104C
10051BBA
4710F1AA
10150181
EOO04170
4180F1E4
10034780
F35405EE
41FOE012
F1745810
91281015
4110F31E
10154710
0703104C
91401002
F28658EO
00000000

41FOF02A
OA099026
FOE85846
4940F1CA
4144F1E9
OA079110
OA021813
4530F1A8
4750F132
102A4710
4770F192
103e8800
10401049
OA009180
00000004
400BCBE3
010896FF
9400Dl04
00011AE1
lB554351
9640D060
910F1000
000847FO
91020105
F12647FO
000247FO
47BOF162
015C4110

41FOF02A
Flce5860
00000640
47BOF078
02006000
10034780
02001040
9108102A
58401010
F13258EO
18534800
00084620
D203104C
10024710
00003550
8B030141
F7C90201
91801000
41EEOO03
00014141
43800060
4740F35C
FOFOD200
071E1957
F1A25950
F1924110
95408000
F5400AOO

01011052
100895FF
18304304
920B6000
40000600
FOCC0501
10380200
4780FOFC
41440000
10409826
10344120
F1584144
103C4340
F1B40A07
000032AO
lB130B61
D1581000
41110002
91011001
00021B75
4480F7CA
91201000
D0601004
47COF1A8
F7BC4720
00785830
4770F162
91801002

10529026
101E4180
00004140
4530F1A8
60001028
1052F1C8
10381030
92016000
05014000
F1CC9620
00031843
00014242
10404144
07F31821
0000395A
5COOOOOO
4580F200
4710F426
4710F064
06704450
4A410006
4710F006
96500060
910F1000
Fl0C91FF
F7BC1B37
06804630
4710F182

F1CC4530
F03E4A60
00094334
9140102C
9108102C
4770FOC2
024Fl098
4530F1A8
F1F34170
102C07FE
4342103C
103C1835
00014240
48020006
00000004
47FOFOOE
47000004
94000105
41EEOOOl
F1CA9601
91101000
02030060
43800060
4710Fl16
F7C94710
4180007B
F14E5890
OA075810

F13C9826
104A9103
F1DF1930
4780F088
4780FOE8
96F01049
10485830
91011004
F1329102
02001040
43321044
0503103C
1050D501
41101016
000048EO
C9D3C2C4
47FOF030
95401000
50EOOl14
010441FO
4710FOAC
10005850
4480F7CA
9101D15A
Fl0C9012
1A835090
015C5030
015C58FO

•• &•••• 0 •••••• 0.

... - ............

•• J ••• J •••••••••
JO •• J ••• J •••••••
.02 •• 00 •• 02-.01.
•••• 0 ••• 3 •••••••
0 ••• 3 ••••• $ ••
•• 2- •••••• 0 ••• K.

................

•••••••• OBP •• <.<
•••• O••• O•• OOS ••
•••••••• 0 •••••••
•••••••••••••• 1.
•••••••••••••• 1.
•• 1 ••• * •••••• 1.

·•••...............
1 •••••• &•••••
1 •••••••••••••••
•• • <•••• &.3 ••• 3.
16 •••• ~ ••••• 2 •••

·•••...............
1 ••• 0 ••• 1 ••••
3 •••• 1.8 •• 3 ••• 3.
•• OY••• OOY ••••••
• .2 ••••• 3& ••••••
2- •••••• 1 •• 01 •••
• <•• 3* ••••••••••
•• 2 •••••••••••••
• ••••• &•••• 02. $$

$$BERRTN.OO •• OO.
1 ••••••••••• 1 •• •••• 1 •• OY •••••
•• 0 •• 0 •• 1 ••• O.
•••••• 1 ••• 1ZK.-.
•••••• 0 •••••••••
.3 •••• 0 ••••• K••
•• K••••••• 1 •••••
• .1 •••••• &1 ••
•••• 1 ••••••• 1 •••
• &N •• •••• 1 •••••
•••• 1 •••••••••••
•••• O.K •••• K•• <
• .1H •• OY ••••••••
••••• 01 •••••••• &
gel. vw-o •• T ••••
OSP033 •• J ••• 71K.
••• iii •• 7 ••• J •••••
•• 0 •••••••••••••
•••• 1 •••••••••
1.K •• - •••
- .............. 3*
• -. &••• &••• OOOK.

...

•••••••••• J •••••

•• 1 ••••• 1 •• 01 •• &
•••• 2 •••••• 01 •••
• * ••• @•••• 1 ••
5 ••• 5.& •• *•• 5 ••

••••••• &••••••••
• &. • ••• & •••• &.
&••••••• K.JY~ •• .00 •• 00 •• 00 •• 00.
1JFFBZZN38 •• 3 •••
•••••••• 3 ••• ·0 •• 0
0 ••••••• 0 •••
••• &.00 ••••••• O.
• .·0 ••• ••• K•••••
•• P •••••••••••••

0.& •••• 1 •••••. 02
..2 ••••••••• &•••
&••••••••••• &•• <
• ••••• 1 •••••••
......... •••• 1 •
K••••• &••
• &••••• 01 •••••••
&.3 ••• 3 ••••••. 1U
.01.K.3 •••••••••
• ••• 2 ••••••• 3 •••
2 ••• 3 •••• 0 ••• 0 ••
3 ••• 3 •••••• 01. ~.
• ••••. OY ••••••••
3&&.3*.1 •••••• 3.
2 •• @3 ••• 3 •••••••
• ••• 2KP ••••• P •• <
•• 1.&.3 •• @3 ••
•••• 2 ••••••• 2 •••
BCEOV1 ••••••••••

P ••••••• 1 ••• 1 •••
•••••••• 0 •• - ••••
....... •••• 1 •••
•• - ••• 1 •• •••• O.
.O.-••••••••. OY
O. N••• lH •• OB. 0 ••
• .K ••••• K•••••••
•• 0 ••• - ••• 1 •••••
• ••• N. .13 •• 1 •••
• •• 1 ••••••• K••

................

1 ••••••••••• N•••
• &N.
1'•••• 3 ••••••••••

...

.....

·...............

••• /* •••• 00. 1LBO
J ••••• 2 •••••• 00.
• ••••• 4 ••• J ••
•••••• 0 ••••• &•••
••••••• &1 ••• J •• 0
•• 7 ••••••••••• O.
•••••• OOK •• - ••••
.- ••• &.- ••• - •• 7.
•• 1 ••••••• 1 ••• J.
7 ••• 1 ••• 71 •• 1 •••
•••• 7 ••••••••• &•
• .1 ••••• 1 •••• *&.
• ••••• 1 •••••• * ••

....I7J
~

..
~

~....
CD

a
i

~
~
N

c.n

oH\

....,

en
CD

g.
....
o

1:1

t:r

s::

07/23/71

orACHK

004340
004360
004380
0043A.O
0043CO
0043EO
004400
004420
004440
004460
004480
0044AO
0044CO
0044EO
004500
004520
004540
004560
004580
0045AO
0045CO
0045EO
004600
004620
004640
004660
004680
0046AO
0046CO
0046EO
004700
004720
004740
004760
004780
0047AO
0047CO
0047EO
004800
004820
004840
004860
004880
0048AO
004940
004960
004980
0049AO
0049CO
0049EO
004A.OO
004A20
004A40
004A60
004A80
004AAO
004ACO
004AEO

010C416D
18851288
F126D200
18E141EE
4710F2E2
80034770
9101F7AO
014447FO
4870F600
F2B258Fl
01594710
0201F600
4110F700
F2845861
F5389209
4770F37C
40005800
00605000
00FC47FO
00690065
96F04000
01059400
41900020
96040105
4710F4C4
12774780
96020105
F4EC1885
47FOF4A8
47FOF1E8
000046FO
C2C304E3
00000000
07004742
0500480F
00008400
00000000
00000000
3100470C
3100470C
01004898
00000000
00000000
40404040
40404040
0000480F
40FF1800
00605810
00044A30
98150060
lB0047FO
4710F022
4780F08A
F09E4780
96015001
4110FOC4
00000000
00000000

007C5870
47COF1BA
60004000
000250EO
91020159
F2444810
4710F26E
F2844870
91030159
001045EF
F20A4166
F6040203
4500F314
002C4166
F5385810
lB000201
00604EOO
F7C04El0
F3FC1B88
F1540060
41440001
F7C91233
lEAA0690
lEAA4730
47FOF4C8
F51A9240
45EOF08A
06809240
02006001
4690F44C
00000000
FOF70000
OOOOFFOO
40000006
60000079
08000002
00000800
00000000
40000005
40000005
20000050
--SAME-00000000
--SAME-40404040
0000488F
00000000
00140A21
10481B55
07FE0008
F0184100
91204002
18350501
F0804403
47FOF080
OA021BOO
00000000
--SAME--

F7BC9101
06804480
94FE015A
01149601
4710F33E
F6000610
9601F7AO
F6000670
47EOF2A2
000C98E2
00014177
F7BOF7B8
000047AO
00004170
015C47F8
00604000
0060F395
00684EOO
43810001
0060F384
lB554851
4780F43C
4730F458
F4909101
45EOF08A
60004166
195747CO
60004480
600089AO
47FOF440
00008000
00000000
00000000
31004744
31004744
000047F8
002020F3
19000020
08004800
08004820
00000000

PAGE
01044710
F1CA4188
9101Dl05
010447FO
0203F7A4
lB171211
90E2D144
91048003
5860F7A4
01445860
000147FO
5810F7BO
OA0298E2
00485810
00080200
48000060
00680062
0070F384
07090060
00600061
00021B45
94000106
47FOF512
01064710
4140F7C6
00010670
F4EA1277
F50C45EO
000147FO
00000000
OCOOOO03
00000800
00000000
40000005
40000005
00000000
2400488E
00000000
20000001
20000001
00000000

F1FA9812
00011A68
071E4111
F12698EC
F7A80201
4770F244
4110F700
4780F284
58200118
F7A40670
F3364170
9101F7Al
01444870
015C47F8
60016000
47FOF38E
47FOF3FC
0060006B
00604150
96F00068
9601015A
47FOF440
91101000
F49047FO
41500003
47FOF51A
4770F4E4
F1B61255
F4A44111
090046FO
00004778
002010F1
13000000
08004768
08004788
00003FCO
80000000
00000051
100047EC
lE004830
00000000

00FC47FO
lB781B58
000A0501
010807Fl
F600F602
91040105
4500F264
4160F728
02076048
06709240
004847FO
4710F322
F60047FO
00044110
91021000
95041001
501000FC
96F00068
006A1B58
41400071
47FOFOFO
96010106
4780F47E
F4BC1EAA
45EOFOFO
4111000A
45EOF126
4780F504
000A0501
00000001
00004780
2600480F
00000000
20000001
20000001
320007C3
00000000
47000000
A0000008
30000059
00000000

FOF01887
4780Fl00
1000F7C4
5010015C
0203F7BO
4770F336
00004708
47FOF336
20009104
60004470
F3360203
9601F7Al
F2A29104
00784111
4710F3CE
4770F39C
02070060
F3840069
06804480
91304000
02005000
58A20000
910Fl000
4730F4C4
4111000A
47FOF51A
47FOF400
47FOF400
1000F7C4
00000000
00003FCO
80000000
00000079
0100480F
lE004798
C8C4E3C6
00000000
0700470A
0500488E
01004848
00000000

47FOF1AA
lA4847FO
4770F08A
91030159
F7B49104
5810F7BO
OA0298E2
90E20144
80034710
F3569103
F7A4F7AC
90E20144
80034780
00045010
95021001
02030060
40009801
00735810
F420F384
4740F40C
40009201
41220004
4770F47E
91010106
45EOF08A
47FOF4A4
188747FO
94010105
4770F52C
00000004
32005B5B
00000000
47000000
20000079
30000081
40400000
OOOOFFOO
40000006
60000051
20000050
00000000

00790079 0050E540

40404040 40404040 40404040 40404040

40404040
00004708
5B5BC206
18305823
43530000
FFOOOOOO
000105FO
071E5811
FOA23000
00064403
96015000
OA060000
00000000

40404040
000047AO
47FOFOOE
4B20F056
4A50104C
05F047FO
41440000
00004B50
41330002
92F05000
58FOFOA4
00000000
5B5BC2C3

40404040
00004708
07C5D540
00001832
89500001
00000000
9005FOA8
00044154
4780F04C
000A1852
9805FOA8
000005FO
00000000

40404040
00000064
C903C2C4
1B444343
89400001
FOOEC9D3
91104014
FOCC5855
47FOF03A
02065001
07FF8900
00000000
06C2C509

01000000
3B9ACAOO
C90403FO
00074830
1A540200
C2C4E2Cl
4780F022
00004155
5050FOA4
50001200
00184B40
00000000
00040008

0000480F
FFFF407E
F3F39015
102E8830
20005000
C5FOF3F3
91084015
00001255
D5033006
4780F07C
FOCE1604
00000000
00000000

••••• lil •• 7 ••• J •••
•••••• 1 ••••• 1 •••
1.K.- • ••• J ••• J.
•••••• &••••• J •• O
•• 2S •• J ••• 3.K. 7.
•••• 2 ••• 6 •••••••
•• 7 ••• 2 ••• 7 •• SJ.
J •• 02 ••• 6 •••••••
•• 6 ••• J ••• 2 •• -7.
2 •• 1 ••••••• SJ •• J ••• 2 •••••••••• 0
K.6.6MK.7.7 ••• 7.
•• 7~ •• 3 •••••••• S
2 •• / ••••••••••••
5 ••• 5 •••• *. 8 •• K.
•• 3lil •• K•• - •••• •••• - ••• -3 •••••
.- •• 7 ••••••••• 3.
••• 03 ••••••• P •• •••• 1 •• -. -3 •• -. /
• 0 •••••••••••••
J ••• 71 •••• 4 ••• J.
•••••••••• 4 •• 05.
•• J ••••• 4 ••• J •••
•• 4D. 04H •• 0 .. 7F
•••• 5 •• •• J ••• 0 ••••• 4 •••
4 •••••• - ••• 5 •••
.04. K. -. - •••••• 0
.01Y •• 4<.04 ••••
••• 0 ••••••••••••
BCMT07 ••••••••••

...........

.. ...

................ ................

.... ....... ... ................
...... .. - ............. ... . ...............
• ••••. PCHO'IF
• •••••••••• 8 ••••
................
••••••••••• 3 ••••
................. ............ ...
.... ........... ........................ - .....
...............
• •••••• &••••••••

••••••••••••••• &

................

••••••••••••• &V

.. . .. .. .. .. .. .. ..

.. .. .. .. .. .. ..

••••••• $$BOPEN

.- .......................
• •••••••••••• &••

...... - .......................

• •• 00 •••••• 0 •• O.
•• 0 ••• .........
•• O••• N.O ••••• O<
0 ••• 0 •••••••••••
• • &•• 00 ••• &••• O.
•• 00 ••••••••••• 0

................

................
.00. 1LB01ML033 ••
•• 0 •••••••••••••
• &. <. •••• K••• &.
.0.00.ILBDSAE033
• •• 0 •••
••• &0 •••••••••••
• •••• 00. &&0. N•••
.0&.K.&.& ••••. Olil
• 00 •• ; ••••• 0 •••

......

................
$$BCOBER••••••••

~
~

....

....IJI'g,
en

N
W

...

1 •••••• 000 ••• 01.
• ••••••••• 1 •••• 0
•••••• N••• 70 •• O.
1 ••• J •• 1& •• * •• J •
7. K. 6. 6KK. 7. 7 •••
• .2 ••• J ••• 3 ••• 7.
•• 7 ••• 2 •••••••• S
• • 2 •• - 7 •• 03 •• SJ.
•••• K. - •••••••••
7 •••••• - ••• 3 •••
3 •••••• 03. K. 7. 7.
..7 ••• 3 ••• 7 •• SJ.
J ••• 6 •• 02 •••••••
• *.8 •••••••••• &.
-. - ••••••• 3 •••••
.03 ••••••• 3.K •• .03.& ••• K•• .- ••• 0 •• 3 •••••••
• - •••••••••• 4.3.
4.
• 0 ••••••••
•• J •• OOOK.&.
.04 •• J •••••••••
•••••• 4 ••••••• 4 •
4 •• 04 ••••• 40 •• J •
• &•••• 00 •••••• O•
.05 •••••• 05 •• 04.
•• 4U •• 1 •• 04 •••• 0
1 ••••• 5 •• 04 ••• J.
4 ••••• N••• 70 •• 5.
• •• 0 ••••••••••••
•••••••••••••. $$
• •• 1 ••••••••••••

- - - - - -

-------~-------

0
0
0

0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0

0
0
0
0
0
0
0
0

,,.,....r-,r-N

0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0

:.::
11:1

~
Fo4

0
0
0
0
0
0
0
0
0
0

I'l

r..
~
r--

Figure 4.9.

232

COBOL Sample Program (Part 26 of 27)

DOS System Programmer's Guide

I Figure

4.9.

COBOL Sample Program (Part 27 of 27)

Section 4: Debugging Aids

233

Debugging FORTRAN Programs
When debugging FORTRAN programs, gather the
program listing, linkage editor map, main
storage dump, and any system messages. The
linkage editor map is the most useful
documentation. Using this map together
wi th the main storage dump, the programmer
can locate such information as DTF tables,
I/O areas, logic modules and FORTRAN
subroutines used at object time.
A sample FORTRAN program, linkage editor
map, and storage map are included to show
the relationship between the linkage editor
map and the contents of main storage, and
to indicate where the DTF tables and I/O
areas are located (see Figure 4.10). The
DTF tables are built as they are opened,
starting from 'the end of main storage or
the end of the partition in which the
program is loaded. The I/O areas directly
follow the end of the problem program. The
FORTRAN program itself builds only one I/O
area.

234

DOS System Programmer's Guide

User-written code is always the first
item in main storage followed by the object
time FORTRAN subroutines to perform the
functions required by the problem program.
Only one logic module is accessed by all
FORTRAN I/O units (except console
typewriter) for each FORTRAN program. The
console typewriter uses its own DTFCN that
includes a logic module to handle all
console typewriter functions. Each FORTRAN
program includes an I/O unit table to
convert a FORTRAN unit to a DOS symbolic
unit.
FORTRAN uses the STXIT macro to handle
its own program checks. It also uses the
normal DOS register conventions. In
addition, register 13 is the base register
for the FORTRAN in-line code and points to
the register save area. Register 15 is the
base register for FORTRAN object time
subprograms.

'1iI

c::

(1)

.
.
-'=

II JOB ANYNAME
II OPTION DUMP,LINK
II EXEC ffORTRAN

~
~

~
en

~...
(1)

I'd
11

I'd

PI
t1

c+
~

en
(1)

g
-'=
00

t='

c::

IQ
IQ

..,.:PI
Q,
C/)

tv
W
U'I

FTC LIS TX, NAME= SAHP2

•

Refer to hex '18' in LISTX Map.

MAP

YES

LOAD

DECK

LIST

YES -

Output A Listing

LIS T)(

YE S -

Output Object Module Listing

E8CDIC

t-.,)

....
~

8CIl

(1)

Ct:l

(1)

....
•

"t1

~
~

~....

"t1

~
(1)

(1)

....c::

0003
0001t
0005
0006
0007
C008
0009
0010
0011
0012
0013

a
i
.
t;

0001
0002

DOS FORTRAN IV 360N-FO-1t19 3-0

CIl

III

0014

(1)

Internal
Statement
Number
assigned by

"t1
III

FORTRAN

Compiler.

t-.,)

SAMP2

OUE

11/11/68

TIME

C PRIME NU~BER GENERATOR
WRITE (3,1)
1 FOR~AT ('IFOLLOwING IS A LIST OF PRIME NUMBERS FROM 2 TO 1000'1
119X,lH2/19X,lH3)
DO It I z 5,lOOO,2
KaSQRTCFLOAT(I))
DO 2 J=3,K,2
IF C~OD(I,J) .EQ. 0) GO TO 4
2 CONTINUE
WRITE (l,l) I
3 FOR~AT (120)
4 CONTINUE
WRITE (3,5)
5 FORMAT (I THIS IS THE END OF THE PROGRAM')
STOP
END

00.00.06

PAGE 0001

....
I'Ij

DOS FORTRAN IV 360N-FO-419 3-0

(1)

SAMP2

DATE

11/11/6B

00.00.06

TIME:

PAGE 0002

-'=

•~

.
o

SCALAR MAP
SYMBOL
I

LOCATION
A4

SYMBOL

LOCATION
A8

SYMBOL

LOCATION

SUBPROGRAMS CALLEO
LOCATION
SYMBOL
LOCATION
SYMBOL
SYMBOL
LOCATION
SYMBOL
LOCATION
SQRT
Displacement to an adcon containing the address of subprograms. Check in

LOCH ION

SYM80L

Locate these values in storage print.
P. P. begin = 2800 Variable = 2800 + DSP.

SYMBOL
IBCOM=

storage- print and use Linkage Editor Map.

!2:

CI'l

SYMBOL

~.....

LOCATION

SYMBOL
K

I'Ij

o
~
~

SYMBOL

LOCATION
AO

fORMAT STATEMENT MAP
LOCATION
SYMBUL
108

LOCATION

SYMBOL

LOCATION

SYMBOL

LOCATION

10C

(1)

It!

~
~
S

I'tI

til

(I)

....~
o
::s

-'=

'='
(I)
tr
~

IQ
IQ

!:S.
IQ
~

....
Q,

...a

Displacement to symbols used on FORMAT STATEMENTS. Locate the beginning of
Symbol 1 in dump and cross-reference to LISTX Map. Check the LISTX Map for the
internal statement numbers 2, 9, and 12. They are not there because a Format Source
Statement causes a constant to be assembled rather than machine instructions.

co
t:I
0

'<
en

...."IiI

~
~

11
0

11
III

~(1)

.
.
~

"IiI

~
~

~
Z

III

a
10

(1)

Cil

....SJ,.
~

(1)

TIME

11/11/68

00.00.06

PAGE 000)

(1)

'tj

DATE

SAMP2

DOS FORTRAN IV 360N-FO-419 )-0

"tj

11
III

' tj

III

rt
~

I-h

Hex

lOCHION
000000
000002
000006
\ OOOOOA
eOOOOE
000012
000014
000018
00001C
000020
000024
000028
000114
000178
00011C
000180
000184
000186
ooa18A
00018C
00018E
0001CJ2
000194
000198
00019C
aOOlAO
OOOlA4
OeOlA8
OOOlAC
0001BO
0001B4
0001B8
0001BA
0001RE
0001C2
0001C6
0001CS
0001CA
0001CE
000100
000104
000108
0001De
0001DE
0001E2
0001E4
0001E6
OOOlEA
0001EE
000lF2
OOOlF4
OOOlF6
OOOlFA

LABEL

STA NUM

OP
8AlR

LM
L

Internal
Statement
Number

l!L.

.Ag

lA
ST
BCR
DC

uction
rated
nternal
ment 1.

oocoooee

J!J..

~g~A~~

DC
DC

00000000
00000000
00000000
13,4(0.13)

DC
l
L

12(13).255

ST
LPR
ST

15,128(0,13)
14,4(0.15)
00000003
000000e4
15,128(0,13)
14,16(0.15)
0,31610.13)
0,116(0.13)

STD

0.28BIO,13'

LTDR

2,2
14,0
11.610.141
0.0

BAlR
BC
lCO~

STE
LA
L

BALR
Be

SDR
LER
AW

BUR
BC

LeR

Program Initialization
Example: IBCOM will issue STXIT macro so FORTRAN
can handle its own program checks.

I BCOM=

Subprogram address being loaded.

Name of a subprogram (I/O)

IBeOM=

Decimal

Refer to internal statement 3.
Variable I is initialized to 5.
Look at storage print. Problem program
area begins at 2800. R13 = 282C
316 = 13C in hex and 116 =1!.in hex.

•

LTR

IBCOM=

J.£,

13,4
14.64(0,15)

1.0
1.28010.(3)
0,276(0,13)
0,260(0.13)
0,0
14.0
1l.6(0.l't)
0,0
0,320(0,13)
1 , 136« 0, 13 )
15,132(0,13)
14.15
0,4(0,0)
2,2
2,0
2.29210,13)
2.284(0.13)

2,12.28(13)
15,14
15,128(0,13)

l, DC
~Al
'1 DC
~ ~Al

Address constant labels are prefixed with an . A'.

14,12'0.13J

BCR

L
LR

Establish Addr.essability
for Program

Program Name
No. of chars.
A20
that can be used
A36
in a CALL Stmt.
A52

MVI

lR
SAL
Inst
Gen
for
Stat

BCD OPERAND

LM
A36

OPERAND
15,0
2,3,34(15)
13,30(0.15)
15,2(0,15)
15,4(0,13)
15,2

seRT
\

282C
282C
+ 13C
-1!..
2968 - Initialization
28AO - Location
Value
ofI
The easy way to locate variables is to use the Storage Map.
J-------SUbprogra.m address being loaded.

Name of a subprogram (square root).

INTERNAL ADCONS

....
I'lr:j

(D
~

...
0

•
~

~
~

en
III

~~
(D

~
11

III

to
III

U'I

en
(D

....~
g
~
00

'='

....::s

....>'
0-

m
to.,)

\C)

DOS FORTtAN IV 360N-F ~-479 ]-0
OOOlFt
000200
000204
000208
00020e
000210
000214
00021800021e
OOOllE
000222
000226
000228
00022e
000230
000234
000238
00023t
000240
000244
000248
00024t
000250
000254
0002')8
0002lie
000260
000264
000268
00016t
000210
000214
000218
00021e
000280
000284
000288
00028C
000290

J.Z.Q.
b

SAHPZ
O.llOI O. U)
0,32410.13.
0,12410,13)

O,1l610,ln

SROA
0
H

0,321,0)
0,124(0,13)
0.12410,13)
1,11610,11)
1,1
1,16010, U.
14.9610.11'
8.14
O,l2'tIO,13)

L
BXLE
L
BAL
DC
DC

L
BAL
DC
SAl
10

L
LA
lA
BXlE

1111 1/68

ST
L
ST

LtR
C
l
BCR
7

DATE

BAl

l.ln4'n.11I

2,210,0)
3,12010, U)
0.2.0(1)
15,12810,13)
14,410,15)
OOCOOO03
00000108
15,12810,13.
14,810,15.
04500074
14.1610.151
0,11610.13)
1,10010,1))
Z,210.0)
3,100010,0.
O. Z.OI 1)
15,lZ810.13)
14,410.15)
00000001
0000010C
14,16(0,15)
15,IZ810,1))
14,52(0.15)
05404040
40FO

DC
DC
aAL
L
13
SAL
DC
DC
END
T~TAL MEMORY REQUIREMENTS 000291 BYTES
I
HIGHEST SEVERITY LEVEL OF ERRORS FOR THIS MODULE WAS D

TIME

00.00.06

PAGE 0004

J
J
J
I

4
J

LZ!l

Internal Adcon used in this case fer
the limits of a DO LOOP.

18COH=

IBCOM:

.1fIBCOH=

IBeOM=

Hex number of bytes.

'"
~

0
CIl
CIl

'<
CIl
rt
CD

to
0

...
...

\.Q

i...

CIl

(j)

..,.s::
Q,

..,.
~

~CD

11/11/68

.
.

PHASE
PHAse...

XFR-AD
002800

LOCORE
002800

HICORE
005FF7

DSK-AD
1A 06

The program name is contained at dis-t
placement of X'IS' in LIST X Map.

A member of the FORTRAN System
Library that performs object time
implementation of I/O source statements and handles all abnormal or
normal termination of FORTRAN
object time programs.

~
~

2:
CIl

~I-'

ESD TYPE

LABEL

LOADED

REL-FR

eSECT

SAHP2

002/00

002800

... eSEe'
ENTRY
• ENTRY
ENTRY
ENTRY
ENTRY
1-.
ENTRY

ILFIBeOM
IBCOH=
OPSYS
INTSW
PDPAR
DUHPSW=
IJTlNTSW

Refer to FORTRAN storage map for address
002A98 002A98
002A98 -------------------ofsubprogram.
003750
0038At
0036e4
00365A
0038A1

[eSEeT

ILFSSQRT
SQRT

005E48
005Ett8

005E48]

.[eSEeT
EN

ILFF INT
SAVERR

004AC8
004FA8

004AC8

eSECT
ENTRY
ENTRY
ENTRY
ENTRY
ENTRY
EI'HRY
ENTRY

lLFADCDN
IlFFCVEO
ILFFCVLO
ILFFCVIO
[lFFeVeO
IlFFeVAO
ILFFeVlO
INT6SW

0039FO
004406
003C82
003FD4
0046FO
003RF2
003B4C
004A80

0039FO

CSECT
ENTRY

IlFFIOCS
ILFFBORG
ILFBFORG
IJSYSlO
AOIOCR=

005098
005AFC
005AF8
005B08
00,)A54

005098

ILFU~TAB

005EF8

IJJCPD1N
IJJCPD3

005e88
005C88

005C88

IJ2L0005

005C48

005098

A member of the FORTRAN System
Library used to calculate square roots.

...I'd

~
~

A member of the FORTRAN System Library processes
arithmetic-type and speCification program interrupts
and continues EXECUTION. It also fixes any boundary
alie:nment error that occurs.

...PIrt

A member of the FORTRAN System library
that performs INPUT and OUTPUT data
conversion.

A member of the FORTRAN System Library that
performs object time I/O data management.

A member of the FORTRAN System Library that
contains a Unit Table that is used to convert from
a FORTRAN unit to a DOS symbolic unit.

I/O LOGIC MODULE used by all I/O units except
console.

A control section in ILFFlOCS that is the logic
module for the console.

~:~:~

•

ENTRY

I rt= CSEC T

'-r CSEeENTRY
T
eSECT

...:::::r-

Problem Program begin.

Refer to FORTRAN storage map for address
of subprogram.

........
I-z;I
~.

t.Q

OWERT

PAGE

11111/«»8

.
.
,f::

tor.!

~
~
2:
CIl

SlJ

~....
CD

ttf

2Ii
SlJ

t tf

SlJ
Ii
rt'

0""'"

IX)

CIl
CD

g.
~.

o
:::s

,f::

~
~

t.Q
t.Q

:::s

t.Q

....Q,
til
tv
,f::
~

(OR 0- 1
(OR 8-F
~P ReG
1>00000
vOOO)O
000060
0000'10
OOOOCO
OOOOFO
uOO120
0001.,0
000180
00011\0
0001EO
000210
000240
000210
0002AO
000200
000300
000330
000360
000390
0003CO
0003fO
000420
000450
000480
0004BO
0004EO
000510
000540
000570
'0005AO
(J00500
000600
000630
000660
000690
0006CO
0006FO
000720
000150
000780
0001BO
0001EO
000810
000840
000870
0008AO
000800
000900
000930
000960

00001«»(4 00004FAO 00002986
OA0401FI 00002810
00000000 00000000
00000000 00000000 00000000
5H~8C2C5 06DIF140 FFO~OOOE
00040000 0000020E 00040000
00000000 00010005 00010680
0«»8006BO 068006BO 41BR0017
9S~2A002 4780090C 48A0023A
AOOl4510 001A07F8 42ROOOE5
42A0020F 414A88AC 4400A004
90709898 90108200 01C84400
9680AOOO 41100030 41FOOCA2
FIF161Fl F161F6F8 28002800
OOOOFFfF E8401E90 AOA07EOO
1448148E 1574158C 15940010
10000000 00000000 92380190
5880A004 90188030 48800270
000094FO 85B9D20F 85COOIC8
418001C8 45A90064 92200190
181F4860 849CIR22 43201007
00C84230 04634820 020E4322
41455000 lA444144 893895FF
9560020F 47800330 0502A005
10064770 035E92FF 04630200
49308488 47B00378 89300003
40005010 40P092FF 40004220
30029198 30060779 43203000
50164780 05200202 00491009
90988050 18984860 300045A9
4A608424 50600040 4032B91C
00480202 10091009 47F0048C
41100722 457006AO 91020044
47700580 9610B8F4 960388E(
417005~2 95500044 47700678
9~503004 4180076A 92003003
100C91FO 30044770 06484310
95003000 41700678 95FF3002
95fF3002 07879198 30060771
02021000 00410601 10040044
47100700 96203006 96401002
07F90200 07390038 0200073F
4832891E 45700160 41330008
4880024E 91F0500B 41800796
02010044 B5600202 0041B421
080694FO 08379120 100(4710
088C4A60 B9164060 849A0715
00444710 084E9130 00454170
95408904 41800918 95603004
96201002 47F00588 9120100C
45800984 9102600A 411008E(
09009701 09000201 00023000
02070040 60004580 095E41FO
091496FO 095F4830 84840257
000040~8
41~00000

00001974
00002810
00000000
00000000
80000000
00000FI2
41880071
41BR0050
49A0024E
48800270
5890A004
A0045890
96020019
00000000
00871997
19640010
909COIC8
41C88000
07f9909C
45900290
4130001A
68681A23
40004110
02594770
0403AOOO
4A300220
50004260
43228905
03000048
005C9898
47F00514
4190018A
411001(8
41900134
94713006
96801002
B7FC0670
01899198
41100490
95015016
96011004
003A4832
19340729
43805008
92040831
07F69101
60006000
08009102
4180819A
47800856
9502500B
92040000
05884400
84AC84C2

0000FF74
0000282C
8EIA1000
00000000
00002120
00000000
45700128
45100128
41AAB8AC
47F08JIA
4220AOOO
A004Q818
82000018
00000000
198E19EF
5R5BC2D6
47F0028C
07F99601
01C841FO
41~0018A

18234140
95081001
038(4284
03E04111
41230008
91F03004
500B9203
48601000
50089(00
90509106
9501~00B

91060045
91390045
91100044
07F99411
9l01100(
4210R7F(
30064180
IB5~4350

47200100
91210044
891E4110
95FF3002
4188R8A(
91090045
10024710
02076000
08374780
95623004
41810010
418008F8
92E60001
09304570
44000086

FFfFFF8C 00002800
0000282C 6200290f
00000000 48170F25
000001EO FF070000
08400000 0000211A
000009F8 00040000
41A088BC 4510078~
41800138 9640~001
071894F9 02180701
4510B31A 0205B56F.
9140AOOI 411001BO
9010989C 01C88200
00000000 FF050001
00000000 D8E6C50Q
19FOIA04 lA221A26
00020000 0100180r.
5890008C 45A90060
A0004880 027041CR
03089101 002341BO
95000023 41800384
03504130 00101823
41F0040A 472000C6
000001F9 95FF043R
00004910 B48A41BO
05001007 AOOE4780
47808076 41800001
501691FO 30044180
95003000 418004A?
000E4170 04044037
00454770 09E4913F
41800486 9101100C
471009EA 94F0003Q
4770078C91800044
41100~CC 95000044
30069104 3006478~
4710062E 45100780
41F00648 92FF5008
069C01F9 4832891C
30024145 50001A44
58100040 4B70B424
477005F4 96803006
05740560 950(3001
078147FO 06961871
01F19180 100C4110
478007F6 92040831
08A095FF 3002478r.
00404030 600895FF
08A84186 000C4080
4180B19A 47F008DO
45100994 47F008CO
95FF0219 47800BFB
820009EO 94F3600(
099C47FO 094A4580
4870849A 4B708916

00002798
00000000
20000000
00000000
00000000
00000278
18484190
9120AOOC
02380238
85150C05
02070lC8
00389284
400020A8
E3404040
1A2A38Fl
IP680000
4190054C
0OO09SFF
03089500
95230023
47800352
96801002
07891800
03E01H66
00C80403
41488000
B09647FO
9F006000
891(9680
004~4170

41100486
18224320
418005BO
417005(2
06600601
198A4100
43703002
02000689
41448938
58710000
91041002
07819500
43701003
01A6910f
58800040
08069110
30024110
B3Dl4180
91020837
47F00930
41900134

9501600(
Oq~f.41FO

4070B49A

Ff050r07
749017FF
00000000
01384180
411000E8
9680AOOI
856EB8AC
90086800
B90COZ07
00001000
OOOOFFFF
FIFIFIF6
oeoooooo
4840020E
AOOF0789
00994780
418000B(
lA234220
960CI004
5000B50(
4121000F
10028918
lA444144
04165880
07694060
300601F9
050891AF
45100180
003A4322
96040044
01F90602
10040044
06264190
02003002
003AO~70

4A50024C
05011000
47800~F8

30004133
4C70830E
50084780
12884770
100C4110
08369602
83D84510
47100588
940E0044
928788FC
47100800
05789104
946F3006

40007048
007FE8S7
00000000
ODBC47FO
9250A001
9283AOOO
1BAA0006
90586820
01C8B568
Ou002000

FF150001
00040000
00000000
000806BO
9281AOOO
440003C8
856E0012
90606840
9~90B~CO

80000RSC
00005F41
F8F3F1F6 000018AC
01EOOOOO 00001AC8
44A00234 ~180AOOO
90D0!35DO 481J0G19C
03085890 008CD201
4B600022 1A664866
04634320 100147FO
07F91858 41430002
95FF043B 4780003A
4570B356 18334330
41822000 4870022C
89381858 4A50024C
87F895FF 30024170
04(49120 30064110
473007AE 47F004F2
00440789 40600034
0201B305 80024110
89054832 891C0500
91040219 477005BO
00410041 477006~A
94133006 9180100C
01349601 800047FO
40000200 4000043B
41330008 95013000
58104000 07F7D201
855E4180 06F80501
96043006 95003000
00080706 47F00678
4A100226 92007000
07904180 006041FO
07E69212 08319104
08A04860 B49A4960
08379204 600A9b90
099447FO 08C00300
91010039 47800902
95060044 47800994
47F00660 02050005
91020831 47100930
600~4180 09304180
1B224320 003A4122
0000~F47

62000002
OFOOOF54
00002508
06B006RO
920300Q5
078rl941F
43410007
90686860
820001e8
00000000
00000010
OOOOOOOK
00002188
47100282
0207B5A8
0090020E

R51807Fb
040E4720
43540000
4860020E
10079501
43387000
0200043B
03864280
009A9504
5880008C
58600048
83C85070
30010036
95003004
91040044
4710060C
064894F~

42700438
4720074C
10000046
100088FE
47700618
4842B91C
47F0060C
07900729
30064180
849847BO
30069102
B9041004
47FOO09C
9220600B
600C9706
96021003
066047FO
R9054570

I'..)

c..
en

.r::

001
002
003
004
005
006
001
008
009
010
011
012
013
OH
015
016
011
018
019
020
021
022
023
024
025
026
021
028
029
030
03.1
032
033
030\
035
036
031
038
039
040
041
042
043
044

00
01
01
01
01
01
01
01
01
01
01
01
01
01
01
01
01
01
01
01
01
01
01
01
01
01
01
01
01
01
02
02
02
02
02
02
02
02
02
02
02
02
02
02
02

000
010
020
010
020
030
040
050
060
010
080
090
010
020
010
020
030
040
050
060
010
080
090
100
110
120
130

lito

150
160
010
020
030
040
050
060
010
080
090
100
110
120
130
140
150

03122161

RPGSPI

H
IPE F 80 80
FINPUT
V 132 132
FOUTPUT 0
AA
1 l01
IINPUT
I
I
I
I
I
I
I
I
AOO
INVAMT
C
01
ADO
INVAMT
01
C
IP
OOUTPUT H 201
OF
OR
0
0
0
0
H
1P
0
OF
OR
0
0
0
0
1P
H 2
0
OR
OF
0
0
0
0
0
01
0 2
0
0
0
0
0
0
0
0
D
Ll
2
0
0
0
lR
T 2
0
0
0

REA040 SYSIPT
PRINTERSYSLST
29 NAME
3l0MONTH
HODAY
]it
380lNVNO
39 430CUSTNOLl
44 450STA TE
46 480CITY
14 8021NVAMT
TOTAL
12
GRPTOT 12
8
30
32

TOTAL
GRPTOT

A C C 0 U N T S R'
53 '
11 ' E C E I V A B L E R E '
88 'G 1ST E R'
25 'CUSTOMER'
80 'LOCATION
109 'INVOICE DATE
42
46
19
108
CUSTNOZ
NAME
STAre Z
CITY I
INVNO l
MONTH Z
DAY
I
INVAMT

'NUMBER
'NAME'
CITY
' STATE
DAY
' 140

23
53
59
61
19
90
96
109 'S

GRPTOT B 109 'S
110 'e'

TOTAL

109 'S
111 'ee'

INVOICE'
INVOICE'
CUSTOMER '
NUMBER'
AMOUNT'

PAGE 0001
RPGOOI
RPG002
RPG003
RPG004
RPG005
RPG006
RPGOOl
RPC;008
RPG009
RPG010
RPGOll
RPG012
RPG013
RPG014
RPG015
RPG016
RPGOI1
RPG018
RPG019
RPG020
RPG021
RPG022
RPG023
RPG024
RPG025
RPG026
RPG021
RPG028
RPG029
RPG030
RPG031
RPG032
RPG033
RPG034
RPG035
RPG036
RPG031
RPG038
RPG039
RPG040
RPG041
RPG042
RPG043
RPG044
RPG045

IV
~

0
0

en
en
'<

rt
CD

a
It:I

t'Ij

...,c:

.
~

t-to

RESULTiNG

Cj)

ADDRESS RI

11
PI

toO

Cj)

c:
...,OJ

RPGSPI

03122167

SYM80L

::0

-1t:I

DOS/360*RPC*V2.LO

•

11
0

SAMPLE PROGRAM LISTING (CONTINUED)

a
~

I'd
11
0

11
PI

"tI

INDICATORS

000011 Of
000085 HO
00008C H7
FielD NAMES

ADDRESS RI

000014 IP
000086 HI
000080 H8

000015 LR
000087 H2
00008E H9

000016 00
000088 H3

000017 01
000089 H4

00007A LO
00008A H5

000078 Ll
000088 H6

ADDRESS FIelD

000121
000143

000139
00014!»

000138
0001H

NAME
STATE

ADDRESS LITERAL

DAY
INVAMT

ADDRESS FielD

ADDRESS FIELD

000130
000148

000140
00014F

INVNO
TOTAL

ADORE SS LITERAL

ADDRESS LITERAL

000168
000196

000183

INVOICE
NUM8ER

0001C3
0001F7

000153
AC C 0 U NT S
00018E CUSTOMER"
LITERALS
ADDRESS LITERAL

0001AC INVOICE DATE
0001Df
STATE
CITY
Ll TERALS

MONTH
CIfY

LITERALS

t-to

PAGE 0002

TA8LES

ADDRESS LITERAL

ADDRESS LITERAL
00020C

E C E I V A8 L E R E
LOCATION
INVOICE

--,--/.--

NUM8ER
MO
DAY

ADDRESS LITERAL
CUSTOMER
AMOUNT

000108

NAME

ADDRESS LITERAL

000217

000218

MEMORY MAP
INPUT/OUTPUT INTERCEPT
TABLE IINPUT AND OUTPUT)
DETER"INE RECORD TYPE
DATA SPECIFICATION
GET INPUT RECORD
DETAIL CALCULATIONS
TOTAL CALCULATIONS
DETAIL LINES
TOTAL LINES
INPUT/OUTPUT REQUEST BLOCKS POINTER
LOCATION OF DTF TABLE POINTERS

GIS T E R

000220
00021C
000464
000248
00078C
000904
000958
000AD2
00096C
00126C
000018

CUSTNO
GRPTOT

...."IJ

~CD

SAMPLE PROGRAM LISTING (CONTINUED)

~
I

W
I

(jl

I-'
CD

ttl
Ii

I.Q

Ii
SlI

ttl
SlI
Ii
rt

OOS/3bO-RPG-V2.l0
INPUT/OUTPUT INTERFACE ROUTINES
WORK AREA POINTER
OVERflOW RYPASS
CONTROL lEVEl
TlBLEIlSSEMBLE 4)
TEST lONE 18CO)
OVERFLOW LINES
LINKAGE PROGRAM
PROGRAM LENGTH

001699

lEND OF COMPILATION'

til

..,.~

g
~

t:I

0"
~

\Q
\Q

..,.
ts

..,.

en
N

..,J

RPGSPl

03122167
OOOOEO
001574
OOOACA
00063C
OOOBAC
0012AC
0009EA
0013E4

PAGE 0003

"->

.c:-

I'Zj

1-/-

t1
0

en
en

.
.
~

!:tI

Ci:l

I.Q

S1I

11
0
11
S1I

Ci:l

1-/-

Pol

C\)

SAMPLE PROGRAM LINKAGE EDITOR MAP

.c:-

~I-'

JOB

RPGSPI

03/ ..~/61

AtTW'" rAKlN

11
0

1I S T
LJ ST
LJ Sf

I.Q

DISK LINKAGE EDITOR DIAGNOSTIC OF INPUT

MAP

INCLUDE
INCLUDE

RPGOOl14
RPGOO1l5

IJCFC 110
IJDFYPll

FlURY

S1I

...."0
S1I

.c:-

OU2l1"1

PHASE

XfR-AD

lIJCORE

HICORE

DSK-AD

ESD TYPE

LABel

LOADED

REl-fR

PIIA Sf •••

0033£4

OOZOOO

00H8R

ZI) 6 1

CSECT

RPGOOI

002000

CSECT
ENTRY

IJCfCIlO
IJCFlllO

0036AO

OO~6I\0

CSECT
ENTRY

IJDFYPll
I JOfYlZZ

00311 0
003110

003110

....

""11

~CD

SAMPLE PROGRAM'S OUTPUT

.
.
.p
~

~
CIl

~....
CD

I'd
11

11
~

ACCOUNTS

~
~

CUSTOMER
NUMBER

CUSTOMER NAME

10112

AMALGAMATED CORP

R E C E I V A8 L
LOCATION
CITY
STA TE
H

REG I S r E R
INVOICE
NUMBER
11603

INVOICE
AMOUNT

I NVO ICE DA TE
DAY
MO
11

38.

The chain forward and back addresses
that allow the programmer to follow
the program flow from procedure to
procedure.

A dummy DSA is built to indicate the
beginning of a chain of DSAs when backward
chaining. Figure 4.20 shows the dummy DSA.
Figure 4.22 gives an example of DSA
chaining.

260

DOS System Programmer's Guide

Word
1

Content
Flags *

AL3(Block Description)

Chain Back Address

Chain Forward Address

Return Address (R 14)

Entry Point (R15)

Save Area For

Registers '0 - 12

Invocation Count

Regi ster 0 of Call ing Block
D,ynamic Storage
For
AUTOMATIC Data

* Flag Byte
XIOO I - Dummy DSA
X 1011 or X 1811 - ON Entries
X 103 1 or X 183 1 - No ON Entries
X I05 1 - Library Work Space

Figure 4.20.

Dummy DSA and DSA Layout

Block Description
Offset to
First ON
Entry

Static
Mask

Dynamic
Prefix

0123.45·67
Fixed Overflow
Size Error
Not Used
Size
"'"'""'--....".-- Conversion
"-----it-..,..... Fixed Overflow
I:...-...;.;,--........-~- Overflow
'"------.........:-..:;;;........ Underflow
Zero Divide
&....;,.i"--o"':;'-

1..,,----__. . . . .-.;;. . . . ON Entry Word
r-----~----~~----~--------~

ON Code

Fi Ie Address

FI'Ggs
01234567

Pointer to LABEL
ON Unit is a GO TO
ON Unit is SYSTEM

I:...-_---:l~ ON Statement is Not Executed
............_......:O~ Last ON Entry in This Block
LABEL
Word 1
Word 2·

Base Address
Invocation Ceunt

ON Code

X'Ol'
X'02'
X'03'
X'04'

X'05'
X'Ol>'
X'09'
X'OA'
X'OB'

X'OC'
X'OD'
X'OE'

Fi<}ure 4.21.

Offset Tol:.abel of
GO TO

Meaning

PrO!1lram Interrupt

Overflow
Underflow
Zero Divide
Fixed C>verfl'ow
Size
Conversion
Error
Endfile
Endpage
Transmit
Key

Exponent Overflow
Exponent Underflow
Floating Point Divide
()ecimal/Fixed Point Overflow

~ecord

Block DescriptieD

Section 4: Debu99in9 Aids

261

MAIN •. PROCEDURE OPTIONS (MAIN), .

SUB1 •• PROCEDURE, •

CALL SUB1
-I- - .

L
BALR

r--'--

END, •
L
LM
BR

15,=V(SUBl)
14,15

13,4(13)
14,12,12(13)
14 To STOP Routine

:'I~----------------~
Static Storage

~------------~~.~

IAL3(Block Description)

T
I-

f T-I

I I I

I II

I I

I-

I
I

_...J

Figure 4.22.

---

PROLOGUE

-RETURN
13,4(13)
L
LM
14,12,12(13)
BR
14

I I
I

I
__ -1

DSA Chaining

DOS System Programmer's Guide

DSA SAVLAST
~
I

I
I

Lr ,..-

Chain FOlWard
A(Next Available Core)
Return Register 14

I
I I

--T...J
I
I
I

Entry Register 15

I·~

I
I
h

Static Storage

I
Flags \AL3(Block Description) J I
I
Chain Back
A(Call ing DSA)

When debugging PL/I problem programs,
the programmer must know how to locate the
DTF tables. In PL/I, the DTF table has an
appendage that precedes the DTF proper and
contains a pointer to the DTF (Figure
4.23). The appendage can easily be found
because i t is listed as the linkage editor
map by filename. Figure 4.24 shows the
format of the DTF used by PL/I. A sample
program, linkage editor map and core dump
are included in Figure 4.25 to aid the
programmer in debugging his PL/I problem
program.

262

PROLOGUE

I L

Return Register 14
Entry Register 15

I
IL __
I I_

USING *,15
14,12,12(13)

I- .... STM

DSA SAVSUB1

Chain Back
A(DUMDSA}
Chain FOlWard
A(Next Available Core}

Static Storage

DSA SAVMAIN
Flags

RETURN
13,4(13}
L
14,12,12(13}
LM
BR
14

USING *,15
STM
14,12,12(13}

CALL LAST
15, = V(LAST)
L
.... BALR 14,15

DUMDSA

LAST •• PROCEDURE,.

_-1

I
I
I

Flags AL3(Block Description' -J
Chain Back
A(Calling DSA)
Chain FOlWard
A(NextAvailable Core)
Return Register 14
Entry Reg ister 15

Word
Contents

Word
Open Mask

1
2

Flag 1

Flag 2

Flag 1

Chain Address

Flag 2

Maintenance Word

Buffer Address

Record Length

TFLl
Flag
Byte 1

Record Length

Maintenance Word

Record Add ress

0008-

AL3(Disk Address Routine)

7
Print
Files

Current Line

TFL2
Flag
Byte 2

TXRR
Communications
Byte

Stream

First Time

Bit

TOPM
Open
Mask

File Closed

Input

Clear Open Mask

Consecutive

Transm it Error

Output

Regional

Wrong Length
Record

Update

00 - Fixed
01 - Variable
11 - Undefined

Sti II to Write

Update Next
Must Be a Read

Direct

Still to Wait For
A Regional File

File Has Keys

Allow Page Size

Sequential

File Has No
Keys

Backwards

Unbuffered

EOF

No Buffers

Not End of Chain
Address Table

Buffered

System File

Print File

Communications Byte

Relative Pointer
Page Size

Chain Address

AL3(DTF - T)

AL3(DTF-T)

Communications
Byte

Open Mask

Number of Records/Track - Regional 1

9
10

12
13

Key

Bin

Head

Number of Tracks

Note: Meaning if Bit = 1

Figure 4.23.

PL/I Consecutive File DTF-A
Appendage
Figure 4.24.

PL/I Regional File DTF-A

SUMMARY OF PL/I DEBUGGING AIDS
1.

Always chain back from the library
work space or DSA to find the active
DSAs.

The DSA (main) can be located by using
the value contained in the COMREG at a
decimal displacement 40-43. Adjust
this value to a doubleword boundary.

If the PL/I error signaled is a
program check, use the PSW and
registers located in IJKZWSI. This
area has a PSW followed by the
registers stored RO through R15.

The PL/I control program intercepts
and tries to handle all errors except
machine checks. The control program
issues a PL/I error message, and a
main storage dump is executed from the
control program. Following the dump,
PL/I closes all files. If a printed
output is produced at the time of
error, PL/I prints the last item after
the dump is taken. This gives the
programmer a ,starting point to the
correct area of the failure.

PL/I program check error message codes
11 through lE are the same as the
interrupt codes 01 -through OE on the
IBM S/360 Reference Data card,
GX20-1103.
Section 4: Debugging Aids

263

PL/I error messages point to the
approximate location where an error
occurred.

The name of the last PLII transient
fetched is at the label 'CIJKS'. To
find the labels CIJKS and DISPL (item
9), use register 13 and the chain back
fields to locate the dummy DSA
(identified by its first byte being
zero). Scan backward approximately
four fullwords and locate X'OA04'
(load instruction). Following the
load is X'47Flxxxx', which is labeled
DISPL. Immediately following this is
CIJKS.
(PLII (D) transients are of
the format 'IJKSxx'
DISPL and CIJKS
are in the routine IJKSZCA.)

14.

The following language aids are useful
in debugging both source and object
problems.
a.

Dyndump

Display

Display using the reply option (to
halt a program)

Null labels

Signal statement (to force dwnp)

Note: In multiphase programs, the first
four characters of the phase names must be
identical and unique to the program. If
they are not, either the DSA is overlaid or
the DSA may not fit into the available
core.

The branch instruction that activates
a transient just fetched is at the
label 'DISPL'.
HANDLING COMPILE TIME ABORTS

10.

11.

To locate the current file, look in
'WCFL' in the library work space.
This contains a request code and
pointer to a DTF-A. To find WCFL, use
the label IJKZWSI in the linkage
editor map. Start in the core dump at
that location and scan backward until
you locate a full word beginning with
X'FE'. The fullword immediately
preceding this is the current file
address, labeled WCFL. (WCFL is in
the routine IJKSZCA.)

The following pointers within the control
program indicate how far the compiler
progressed before the abort condition
occurred:
1.

The KSAVEl area contains return
registers in the following order:

The first word of any DTF-A points to
a logical IOCS DTF.

14 - points to last active routine

R 15
R 0

12.

The linkage editor map has a CSECT for
each DTF-A.

R 1
R 2

13.

PL/I register usage is as follows:

To locate KSAVE1, add X'Da' to the
contents of register 12.

264

The instruction flow base
registers in a procedure are 13
(11 and 10) for dynamic storage
and 12 for static storage.

The data accessing base registers
in a procedure are 13 (11 and 10)
for dynamic storage and 12 for
static storage.

The instruction flow base
registers in a PLII subroutine are
15 and 12.

The data accessing base registers
in a PLII subroutine are 13 and
11.

The parameter passing registers
are 0 through 5. (Register 1 is
the m9st commonly used.)

DOS System Programmer's Guide

The K5PH area (a bytes) contains the
name of the phase now in storage. The
phase name is constructed in the
following manner:
PL/lxxxx
The las.t four bytes xxxx contain
the actual phase identifier such as
D75. To locate K5PH, add X'2a4' to
the contents of register 12.

Exceptions: 000, DOS, 010. During these
three phases, the phase name can be located
by adding X'10a' to the contents of
register 12. If K5PH contains phase C95,
the actual phase may be either C95 or Dll.
4.

IJKZWSI is valid only if a PSW has
been stored there.

Ii
CD

.
.

DOS PL/I COMPILER

360N-PL-464 CL2-0

CHAiN

06/01../66

EXAMPLE OF THREE EXTERNAL PROCEOURES PASSING PARAMETER *1

"d
t"'

CIl

~....
CD

6
1

"d
AI

10
11

t:I
CD

IQ
IQ

::s

IV
0\
U1

6
9

g.
g

CIl
CD

EXAMPLE OF THREE EXTERNAL PROCEDURES PASSING PARAMETER */
/- THIS 15 THE MAIN PROCEDURE IT PASSES (CTR) TO PROC SUBI *1
STAHT..
PROCEOURE OPTIONS (MAIN),.
DECLARE
INCARD FILE INPUT RECORD ENVIRONMENT (F(7~) MEDIUM
(SVSIPT,2540»,
OTCARD FILE OUTPUT RECORD ENVIRONMENT (F(80) MEDIUM
(SVSPCH,2540»,
CARDIN CHARACTER (15),
I CARDOUT,
2 FIRST15 CHARACTER (15),
2 SECCTR PICTURE '9QQ99',.
DECLARE CTR PICTURE 'Q9999',.
OPEN..
OPEN FILE (INCARD),.
OPN..
OPEN FILE (OTCARO),.
C..
CALL SURIlCTR),.
READ..
REAU FILE ([NCARD) INTO (CAROI~),.
ONE..
ON ENDFIL~ (INCARD) GO TO END,.
FIR..
FIRST15-CAROIN,.
SEQ..
SEQCTR=CTR,.
WRITE..
WRITE FILE (OrCARO) FROM «(ARDOUr),.
CC..
CALL REPEAT (CTR),.
GOTO..
GO TO READ,.
END..
CLOSE FILE lINCARO),.
ENOl..
CLOSE FILE (OTCARD),.
END,.
/*

12
13
14
15
16

PAGE 001

II.)
~
~

8CIl
CIl

'<
en

rt
In

ttl

11
0

"zj

11
In

II.)

S Y

U'I

ttl

t'"I
t-t

CIl
QI

In
11

en
en

ttl

B 0 L

STAHr
INCARD
OTCARO
CARDIN
CARDOUT
fIR S T15
SEQCTft
erR
UPEN
llPN

0102
0100
0101
0106
0105
0104
0103
0101
0108
0109

~EAD

010~

rlR

010B
010e
0100

ttl

~EIJ

OlOE

ft~(TE

11
rt

OIOF
0110

~OTO

Olll

II.)

t""O
ENOl
SU81
f{EPEAT

0112
0113
0114
0115

s:::

360N-PL-464 CL2-0

Ct1A I;"

PAGE 002

06/06/66

.r=

i
In

DOS PL/I COMPILER

O~E

DOS Pl/I COMPILER

TAB L E
00
0.1
01
01
01
01
01
01
01
01
01
01
01
01
01
01
01
01
01
01
01
01

0
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1

ENTRY

STRUCT.
STRUCT.
STRUCT.

ENTRY
ENTRY

360N-PL-464 CL2-0

LISTING
ARITHM» DECIMAL FLOAT
FILE
FILE
STRING ALIGNED CHAR.
1
PACKED
2 STRING
CHAR •
2 PICTURE OECI~AL FIXED
PICTURE OECIMAl FIXED
LABEL
CONST.
LABEL
CaNST.
lAREl
CaNST.
LABEL
CONST.
CaNST.
LABEL
LABEL
CaNST.
LAREL
CONS T.
LABEL
CONST.
LABEL
eONST.
LABEL
CaNST.
LABEL
CaNST.
LAREL
CONST.
ARITHM. DECIMAL FLOAT
ARITHM. DECIMAL flOAT

L.HAIN

INTERNAL NAME

OfFSET

TYPE

MODULE OFfSET

0102
0105
0106
0107
0114
0115

0034
0148
00f4
013F
0058
0064

STATIC
AUTOMATIC

000114

AUTO~ATlC

AUTOMATIC
STATIC
STATIC

000138
000144

[:'XT
EXT
EXT

6
75
75
5,0
5,0

AUTO~

I~T

AUT OM. INT
[NT
INT
AUTOM. INT
INT
INT
INT
INT
INT
I N,T
(NT
(NT
(NT

6
6

06/06/66
OFFSET TABLE

INT
(NT
INT
EXT
EXT

PAGE 003

1-"

~t1
(I)

. I
.

DOS PLI I COMPILER

CHAIN

360N-PL.-"6" Cl2-0

05/02161

PAGE 00 ..

t-.,)

LaC.

OBJECT coDe

000000
000002
000002
00000"
000008
OOOOOC
000010
000012
00001 ..
000018
00001A
OOOOIE
000020
000021
00002 ..
000028
00002e
00C030
COOO 32
000038
00003(,
000040
0000 .. 2
0000 .. 6
oaOO .. A
00C04e
OOOOSO
00005 ..
0000S6
00005A
0OO05E
000060
000064
000068
00006e
000010
000012
OOC016
00001A
000080
000084
000088
00008C
000092
000098
00009C
OOOOAO
0000A4
OOOOA8
OOCOAA
OOOOAE

OSFO

LABEL

OPe

OPERANDS

BAL •
L' 0102'

BEGIN
BCR
BAL
DC
L
LR
LR
L
BALR
LA
BALR
DC
DC

F,O
PRCt.OGUE
OF BLOCK
01
0,0
E,X'OOA'CF)
ACN'FFFF') A (STATIC STORAGE)
C,X'OOO'CEt
9.F
REG 9 ALSO USED AS BASE REG
3,1
NO MEANING IN A 'MAIN' PROCEDURE
DSA I NITfAl/ ZE
F,N'OOII' =V (IJKSZCM)
E,O
E,X'OOE'IEt
1,f
X' 01' DSA 'FLAG BYTE'
AU.CN'01l6t) AL3 (~LOCK DESCRIPTION)
LENGTH OF OSA OF BLOCK 01
ENTRY POINT MQVE
I • X' 044 ' CC) =A (E NTRY POI NT TABLE)
F.N'0016'
=V (IJKSZlM)
E,F
IN If NE PROLOGUE
X· 050' C04,oJ, X '000' C3) NOT USEFUL IN A 'MAl N' PROCEDURE
OPEN FILE (INCARD), .
I ,X' 0 10' CC) A \FllE NAME)
f,N'0018'
=V (JJKTOPM) 'OPEN'
E,F
OPEN FILE (OTCARD),.
1,X'011t' CC) A (FILE NAME)
F ,N' 00 18' =V (JJKTOPM) 'OPEN'
E,F
CALL SUB1 (CTR),.
E,N'OI07' REG 14 = ADDRESS OF 'CTR'
E,E,X'OS8'(0)
0,0
F,N'0114' =V (SUB1)
I,X'058'CO)
E,F
READ FILE (I NCARD:l INTO (CARDIN)
1,N'0106' REG 1 = ADDRESS OF 'CARDIN'
1.X'020'CC) STORE 1 TO 'WORKAREA ADDRESS' OF 'CONTROL BLOCK'
I,X'018'((.) REG 1 = ADDRESS OF 'CONTROL BLOCK'
F,N'0038' =V (JJKTCBM) 'TRANSMITTER'
E.F
ON ENDFILE (INCARD) GOTO END,.
2. N' 0116' REG 2 = ADDRESS OF BLOCK DESCRIPTION
2,X'000' (2) INCREMENT REG 2 TO POINT AT 'ON' ENTRY
X'060' (02,C) ,X'OItS' (0) MOVE DSA INVOCATION COUNT TO STATIC STORAGE
5,X.OSC'CC) lA 5, o-A (STAHT)
5,X'004' (2) STORE REG 5 TO 'ON' ENTRY
X'004' (2),X'S8' RESTORE FLAG BYTE IN 'ON' ENTRY
FIRST 75'~ CARDIN, .
N'OIOS' (4B),N'010o'
SEQCTR:· CTR, .
N'010S'+4BCOS),N'0101'
WRITE
FILE (OTCARD) FROM (CARDOUT),.
I,N'OIOS' REG 1 = ADDRESS OF 'CARDOUT'
l,X'02e' (C) STORE REG 1 TO 'CONTROL BLOCK'
1,X'024' (C) REG 1 = ADDRESS OF 'CONTROL BLOCK'
f,N'003a' =V (JJKTCBM) 'TRANSMITTER'
E,F
CALL REPEAT (CTR), .
E,N'0101' REG 14::: ADDRESS OF 'CTR'
E,E,X'OSS' (D)

t-I

I-'

(I)

I.Q

SlI
t1

rt
W

en
(I)

n
rt
1-"

0
t:I

~
00

(I)

s::
I.Q
I.Q

1-"

t:S

I.Q

>'
1-"

0..

t-.,)

-.J

STAHT..
0100
.. SEO FOOA
OOOOOOEO
58CO EOOO
189F
1831
58FO COS ..
OSEO
.. lEO EOOE
OSIF
01
000110
00000198
.. 110 eO .. 4
5afO COSO

DC
LA
L
BALR

aSH
0203 OOSO 3000
4110 COlO
OPEN •• L'010S'
S8FO CO .. C
OPN ..

l'0109'

c..

.
L'010A'

LA
L
BALR
LA
L
BAlR
LA

l"OIOB'

LR
L
LA
BALR
LA

OSH
.. 110 C014
S8FO
OSEF
41EO o13F
90EE 0058
lBOo
58fO.COS8
.. 110 0058
OSEF
.. 11000F..
5010 C020
.. 110 C018
S8FO COItO

co .. e

READ ..

STM

OSH
4120
4A20
0201
4150
S050
9288
024A
0204
4110
SOlO
Itll0
58FO
OSEF
41EO
90EE

C030
ONE ..
2000
C060 0048
case
2004
2004
01lt8 oOflt FIR..
0193 D13FSEQ..
0148
WRITE ..
e02C
C021t
C040
013F
0058

14VC

cc..

L'010C'

LA
L
BAlR
LA
AH

MVe
LA

L'0100'
L'OIOE'
l'010F'

ST
MVI
MVC
MVC
LA

l'0110'

LA
L
BAlR
LA

STM

N
G\

....tsj

t:1

•N

g
'<
CD

(1-

J
t'it

(I)

OOS Pl/l COMPILER

·to
t-4

I
t.t!S

....
(I)

lOC.

000082
000084
000088
00008C
00008E
0000e2
0000C6
OOOOCA
OOOOce

000000
00'00'04

ti)t

000006
00 00.0 A
OO:()0 O'E

OOO:OEO

OOOOEO

11
CD

360N-Pl-464 Cl2-0

05/02161

CHAIN

PAGE 005

(I)

.,
Q

-l
to

Q
H\

....

OO(),OEC
OOOOofO

OO'()Ofl
OOOO'f4
OOOOfS
000,OF8
OOOOF9
OOOOFC
000 "00

0!OOI'()'4

000105
0{)0'l08
000"OC
OO'OllO

oe,0114
00'9118
000119
eO'OlIC

O,etH20

LA8El

OPe

OPERANDS

1800

0,0

S8FO C064
4110 0058
OSEF
41FO 90SE
4110 COlO
S8FO C048
O.5·EF
4110 C014

L
LA
8'ALR

F,N'OIIS' =V (REPEAT)
I,X'OSS' 'D) REGl POINTS AT A (CTR)

F,N'OIOB'
1 ,X' 0 10' '·C) =A (lNCARD)
F ,N'0019'
=V (IJKTCLM) 'CLOSE'

GO TO READ
CLOSE FILE (lNCARD),.

I.X'014"C)
F,N'0019'
E.f
O,X'004'CO) 'REG 13=A(DUMMY DSA)

CLOSE FILE (OTCARD),.

08JECT CODE

~fO

GOTO ••
END .•

L'Ol1l'
l '01,12'

END 1 ••

L'OII3'

C048

00,0;"28

REG 14

TERMI NATION

A (STOP ROUTINE)

F.E

S TA T r.e

s reR.G.E
X'Oe0030004000S000'

~'()0007000'

L' 0,00,6'

ODe

L'OllO'

000000

80
000'000
5·8
0000&0
0000004'8
0000000,0
23
000000
0000:0'05.0
000,00000
0008'OaF8
0000009·2
O.A
00090'9
80000000
00·000000
8F
000000
00000000

OOCOOOOO

o,a-ol,O

E,C,X·'OO.C' (0)

00000000
00000000
0000000'0
0000000·2
OOOOOlleO

OCil013C
000140
000144
O(Wl48

l'FFFF'
L'0004'

00003,,00'040005000
60007'0,00
00000:000

IHl0130

oa0138

LM
8'CR

E,F

'END eF Bl.eeK

000l2t
0'()0134

&ALR
LA
L
8A'lR
l

OS£F
5,800 0004
98.eC oooe
07'FE

a.Q.O·ll4

00'·0--124
oeO'125

LA
L

E,F

Doooeooo

L "0 l'1E'
l ' 011,F'

L'012'0'

a.e
DC

DC
DC
ac
DC
ue

~'&O·

Vl3'CN'OleO'»

x'8'0'

VL3tN'0101')

x • S8 '

READ I NiO

Vl.nN'OH)O'» FILE ADDRESS
X'0000004B' DATA LENGTH
X' 00'000000'
ADDRESS OF WORKAREA
,)('23'

V~3(N'0101"

x'oooaoooo'

De
L'O 11,6'
L'Ol().2"

X' 0·0000000'

ADDRESS INCREMENTS USED WHEN DATA OR
BRANCHING IS FURTHER THAN 4K FROM BASE REG

x'OOOOOOSO'

X' 0(10800F8' . BLOCK DESCRI·PTION

A , N' (».J 02' ) ,
X'S,.' 'ON'ENrRY
aev,uc N' ~H00 ' )

Be
BC

x'sooooeoo'

L'0038'

ac'

ViN' 0038' »

L'OOOS'
L'SQ3l'

IilC

x'aF' EN'TRYPOINT TA8LE BIT 0 =1 MEANS 'LAST ENTRY'

L' 001,9'

l ' 0018·'
L'0016'
L'OOll'
l ' 0 114'
L'(H1C'

&C
Be
DC
DC
DC

VL3cN'a'03P)

'.HN' 00·1 9'
V Ht'00'18'
VI'N' 0,016 '
VI N' 00 11 '
v 'N "0 114'
AIN" 0 l02 '

)
)
»

)
J
)

x 'ooooooeo'

2 BYTES - INVOCATION COUNT NEXT 2 BYTES 'GOTO' ADDR tN 'ON' ENTRY

VIN'Ql1S')

0000000000000000

ll"a llS'
L'eDOI'

X 'OOOOOO·0.oeo.oooooo' ,

FFFFFfFFFFFFfFFF

L'~H)02'

X 'FFFFFFFFFFfFFFFF'

I':I:j

""~

DOS PLII COMPILER

360N-PL-464 CL2-0

ChAIN

06/06/66

PAGE 006

(;HAIN

06/06/66

PAGE 001

06/06/66

PAGE 008

.
~

LOC.

OBJECT CODE

LABEL

OPe

OPERANDS

000158
OOO15C

OOfEFFFF

L'OOO"
LIOOOB'

DC
DC
END

X'OOFEFFFFI
X'OC'

•

~
I-t
en

I-'

DOS PLII COMPILER

360N-Pl-464 CLl-O

I'd

11
0

SYMBOL

INCARD
I JCFllZO
IJKTXCf
OTClRO
I JCFlOl4
STAHT
I NCAItO
OTCARD

SUB1
REPEAT

TYPE

ESIO

AOOR

LENGTH

0001
0002
0003
0004
0005
0001
0002
0003
0004
0005
0006
0001
0008
0009
OOOA
0008
OOOC

000000

000098

000000

OOOOFO

000000

000160

PI
11
rt

1"1\

....

(X)

IJKSICA
IJKSICM
I JKSllM
IJKTOPM
IJKTCLM
IJKSYSA
IJKTC8M

m
(1)

l:S

t:I

~
~

....l:S

""g,
en

IV
0\
\D

ER
SO
ER
SO
ER
ER
ER
ER
Elt

ER
ER
fit

ER
ER
ER

DOS PL/I COMPILER

o("t

""o

fit

8LOCK
01

LENGTH OF OSA
0198

ESIO

360N-PL-464 CL2-0
BLOCKTABLE

EXTERNAL SYM80l TABLE

CHAIN

to.)

o"'"'

g
CJl

'rt

I.Q

~
"",

-=
•to.)

"",

EXAMPLE OF THREE EXTERNAL PROCEDURES PASSING P~R~~ETlR ./
/_ THIS PROCEDURE ACCEPTS (CTR)
ZEROES lION THE FIRST ENTRY -I
I. THEN PASSES IT TO PROCEDURE 'lAST' FOR !HCREMENTING ~I
/. ON SUBSEQUENT ENTRIES THIS PROCEDURE ONLY PASSES {erRl TO 'lAST ·1
SUBl.. PROCEDURE (CTR),.
DECLARE
CTR PICTURE '999QQ',.
CTR=O,.
CT ••
CALL LAST (CTR),.
Cl ••
RETURN,.
RET ••
ENTRY (CTR), •
REPEAT ••
CALL LAST (CTRt,.
CC ••
RETURN,.
R••
END,.
END ••

1
2

I.Q

PAGE 001

06/06/66

CHAIN

3bON-PL-464 Cl2-0

EXAMPLE OF THREE EXTERNAL PROCEDURES PASSING PARAMETER ./

•'-"

I.
CD

~.....

DOS PL/I COMPILER

)

5
6

"tI

8
'I

DOS PL/I COMPILER

S Y MB 0 L

0100
0101
0102
0"103
0104
010')
0106
0107
0108
0109

~UBI

loft

CTR
CT
CL
KEf
;{EPEAT

(X)

CC
j(

END
LAST

DOS PL/I COMPILER

00
01
01
01
01
01
01
01
01
01

TAB L E

LISTING

0 ENTRY
1

AR ITHM.
PICTURE
LABEL
LABEl
LABEl
AR ITHM.
lASH

DECIMAL FLOA T
DECIMAL FIXED
CONST.
CONST.
CONST.
DECIMAL flOAT
CONST.

LABEl

CONSf.

LABEl

CONST.
61NARY

1
1

0 ENTRY
1

1
1
1 ENTRY

3bON-PL-4b4 CL2-0

AR ITHM.

ChAIN

INTERNAL NAME

OFFSET

TYPE

MODUL E OFF SE T

0100
0101
0105
0109

0014

STAT IC
AUTOMATIC
STATIC
STATIC

OOOODe

00~8

0018

0010

OOOOEO
OOODE8

'PAGE

06/06/66

CHAIN

360N-PL-464 Cl2-0

F T XED

'>.0

002

EXT
PARAM. iNT
INT

INT
INT
FXT
fNT

INT
INT
1.5

06/06/66

OFF SE T TA BLE

['XT

PAGE

003

...."zJ

DOS PL/I COMPILER

(1)

•

N
U'I

ttl

....\:
en

~....
(1)

ttl

r+
....a
0

en
(1)
n

r+
....

g
~

(1)

....

en
N

....a
~

LOC.
000000
000000
000004
000008
OOOOOC
000010
000012
000014
000018
OOOOU
OOOOIE
000020
000021
000024
COO028
oOO02E
000014
000038
00001E
000041
000046
00004'
UOO04t
000050
0000'.)4
0000S6
00005A
OOOOSE
000060
000064
000068
oa006C
000010
000014
000018
00001'
DOODlE
COO080
000082
000086
.000088
000089
00008C
000090
000096
0OO09C
OOOOAO
'JOOOA4
OOOOA8
OOOOU
OOOOAE

08JECT COOE
90EC OOOC
4ljEO FOOC
000000C8
58CO EOOO
189F
1831
58FO COIC
05EO
UEO EOOE
051F
03
000008
00000100
0203 0058
0203 0050
'.)860 0058
0204 6000
58'.)0 0058
41EO 5000
90EE 0060
1800
58FO C020
4110 0060
05EF
5800 0004
98EC DOOC
07FE
47F0 909C
90EC OOOC
45EO FOOe
000000C8
S8CO EOOO
S890 C014
1831
S8FO CalC
0700
OSEO
4lEO EOOE
051F
03
000008
00000100
0203 0058
0203 0050
5860 0()58
41EO 6coo
90EE 0060
1800
58FO C020
4110 0060

360N-PL-464 CL2-0

CHA

LA8EL

OPe

OPERANDS

L'OIOO'

BEGIN
STH
BAL
DC
L
LR
LR
L
BALR
LA
BALR
DC

OF 8LOCK 01
e,c,X'OOC'(O)
E,X'OOC'(F)
A(N'FFFF')
e,X'OOO'(E)
9,F

DC
3000
]004
L'OlO2'
CO 30
L'0103'

DC
MVC
MVC
l
MVC
L
LA
SUI
l~

lO, 0104'
l'OU2'
l'Ol05'

L
LA
BALR
L
LM
BCR
BC
STH

BAL
L'01l4'

1000
3004
L'Ol06'

DC
L
L
LR
L
BCR
8ALR
LA
86lR
DC
DC
DC
folVC
MVC
L
LA
STM
LR
L

3.1
F,N'0012'
E,O
E,X'OOE'(E)
1, F
X'D]'
AL3eN'010A')
LENGTH OF OSA OF BLOCK 01
N'0101'(04),X'OOO'I3)
X'050'(04,01,X'004'(3)
b,N'OlOl'
X'OOO'(05,6),X'030'CC)
5,~'010I'

E,X'OOO'(5)
E,E,X'060'IOI
0,0
F,N'Ol09'
1,)('060'ID)
E,F
D~X'004' (0)

E,C,X'OOC' (01
F,E
F,N'Ollt'
E,C,X'OOC'CD)
E,X'OOC'Cfl
AIN'FFFF')
C,X'OOO'(E)
9,X'024'CC)
3,1
F,N'0012'
0,0
EtO
E,X'OOE'(E)
l,F
K'03'
AL1CN'OlOA' )
LENGTH OF OSA OF BLOCK 01
N'0101'(04t,X'000'(31
X'050'(04,0),X'004'(3)
6,N'0101'
E,X'OOO'(6)
E,E,X'060' (D)
0,0
F,N'0109'
I,X'060'CO)

06/06/b6

PAGE 004

N
...,
N

'<
en
rt
(D

ttl

11
0

III

G'l

....dg,
tD

....IlIj

DOS Pl/l COMPILER

360N-PL-464 CL2-0

CHAIN

C6/06/66

PAGE 005

06/06/66

PAGE 006

06/06/66

PAGE 007

.
.c::

LOC.

OBJECT CODE

til

000082
000084
000088
OOOOBC
OOOOBE
0000C2
0000C6
0000C8

05EF
S800
98EC
01FE
S80t)
98EC
01fE

0000C8

0000300040005000
60007000
00000000
000COOF8
00000000
00000064

U'I

,
•

t-I

~....
ttl

OOOOO~

000008
OOOOOC
OOOOEO
OOOOE4
0000E4
00OGE8
oooeEC
OGOOFO
UOOOFB
000100

III

rt
co
0
t"h

LABEL

0004
OOOC

.l' 0101'

0004
OOOC

l'0108'

L'FFFF'
L'0004'
l'0006'
l'OlOA'
L'OIOO'
L'0105'
L'OOOS'
L'OOIZ'
L'Ol09'
L'Olll'
l'OOOI'
L'ooel'
L'OOOA'
L'OOOS'
L'OlOC'

00000000
00000000
00000000
0000000000000000
FFFFFFF~FFFFFFFF

OOFHFFF

ooalO~

000105

FOFOFOFOFO

OP •

OPERANDS

BALR
L
LM
BCR

E,F
D,X'004' (0)
E,C,X'OOC'CO)
FtE
D, )(1004 ' (0 )
l
LM
E,C,X'OOC'CO)
F,E
BCR
END OF BLOCK
STATIC STORAGE
DC
X'0000100040005000'
DC
X'60001000'
DC
X'OOOOOOOO'
DC
X'OOOCOOFS'
DC
AIN'OlOO')
DC
ACH'QI05')
X' ,
DC
DC
VCN'OOIZ')
DC
VlN'0109')
DC
ACH'OIOO' »
DC
X'OOOOOOOOOOOOOOOO'
DC
X'FFFFFFFFFFFFFfFF'
DC
X'OOFEFFFF'
DC
X'OC'
DC
X'FOFOFOFOFO'
END

DOS PL/I COMP~LER
SYMBOL
SUBl
REPEA'
LAST
I JKSlCN

TYPE

ESID

ADDR

LENGTH

0001

000000
000064

000110

LD
ER
ER

0002
000]

DOS Pl/l COMPILER
BLOCK
01

]60N-PL-~64

LENGTH OF OSA
0100

CL2-0
ESID

ChAIN
EXTER~AL

SYMBOL TABLE

0001

360N-PL-464 Cl2-0
BlOCKTABLE

CHAIN

....
t'Zj

DOS Pl/f COMPILER

~
~
1

.'"

EXAMPLE Of THREE

360N-PL-464 CL2-0

EXT~RNAl

CHAIN

06/06/66

PAGE 001

PROCEOURES PASSING PARAMETER */

'*'a THIS PROCEDURE
OF THREE EXTERNAL PROCEDURES PASSING PARAMETER -I
INCREMENTS FIELD (CTRI AND PETURNS IT VALUE -I

IJ1

EXA~PLE

LAST..

I'd

~
H

)

CT ••
RET ••
END ••

PROCEDURE (CTR),.
DECLARE
crR PICTURE 199999 1 ,.
CTR=CTR+l,.
RETURN ••
END ••

I'd

DOS PL/I COMPILER

160N-PL-464 CL2-0

CHAIN

06/06/66

PAGE 002

ttl

~
ID

TAB L E

S Y MB 0 L
0100
0101
0102
0103
0104

LAST
crR
CT
RET
END

00 0 ENTRY

01
01
01
01

1
1
1

LIS TIN G
AR ITHM.
PICTURE
LABEL
LABEl
LABEL

BINARY FIXED
DECIMAL FIXED
CONST ..
CONST.
CONST.

15
5.0

EXT
PARAM. INT
INT
INT
INT

....

DOS PL/I COMPILER

en
~

....
rt

g
~
00

t:I

\.Q
\.Q

....

\.Q

....Qt
)I

...,

360N-~L-464

CL2-0

CHAIN

INTERNAL NAME

OFfSET

TYPE

MODULE OFFSET

0100
0101

0014
0058

STATIC
AUTOMATIC

000091t

06/06/66

OFFSET TABLE

PAGE 003

I\J

.....J

.1=

t::I

CIl

en
'<
en
rt
CD
S

"'Ij
~

\,Q

c::

.
VI

...
CIl
III

III

.
en
CD

Cil

CL2-0

CHAIN

Lot.

OBJECT CODE

LABEL

OPe

L'OIOO'

BEGIN OF BLOCK 01
E,C,X'OOC'IO)
E,X'OOC'IF)
AIN'FFFFt"J
e,X'OOO'IE)
9,F
3,1
F,N'OOl2'
E,O
E,X'OOE'IE)
I,F
X'03'
AL31N'Ol05"
LENGTH OF OSA OF BLOCK 01
N'OlOl'C04',X'OOO'(3)
X'050'(04,D),X'004'C3)
6,N'tHOI'
X'OEO'13,O),X'000'(S,6)
X'060'CO,O),X'03S'Cl,C)
X'060'CO,D),X'OEO'C},n)
X'OE8'(4,01,X'060'(0,DI
N'OI05'+2,X'F9!
X'070'C},O),X'OE8'(4,O)
N'OlOS'+2,X'FS'
6,N'OI01'
X'000'(5,6),X'070'C3,0)
01
X'004'(6),X'FO'
L
O,X'004'COI
lM
E,C,X'OOC'WI
BeR
F,E
l
D,X'004'CO)
lM
E,C,X'OOC'CO)
F,E
BeR
END OF BLOCK

OPERANDS

I\J

\,Q

36ON-PL-~64

11
0

DOS PL/I COMPILER

~I-'
CD

11
0
\,Q
11

I'd
III

rt
~

000000
000000
OOOOO~

000008
COOOOC
000010
000012
00001~

000018
OOOOlA
OOOOIE
000020
000021
00002~

000028
OOOD2E
00003~

000038
00003E
OOOO~~
OOOO~A

000050
000054
000054
00005E
000062
000068
00006C
000010
00007"
000016
OOOOlA
00001E
000080
GOO080
COO08C
000090
000094
000098
000098
0OOO9C
UOOOAO
OOOOA8
000080
000084
000085

90EC OOOC
FOOC
00000080
5BCO EeOO
189F
1831
58FO C018
05EO
'tIEO EOOE
051F
03
000090
00000110
0203 0058
0203 0050
5860 0058
F224 DOEO
F8FO 0060
FAF2 0060
F83F DOE8
92F9 C012
F823 0070
92F8 C012
5860 0058
F342 6000
96FO 6004
5800 0004
98fC DOOC
01FE
5800 0004
98fC DOOC
01FE

STM

~5EO

3000
3004
L'OlOZ'
6000
C035
DOEO
0060
DOES
0070

000030004000S000
60007000
00000000
000800f8
00000000

L'OIO)'
l'O104'

L'FFFF'
L'OOO4'

00000000

L'0006'
L'OlOS'
l'OlOO'
l'OOOS'
l'OOIZ'

0000000000000000
FfFfFFfFFFfFFFFF
OOFEFFFF
OC
IC

L'OOOl'
L'oe02'
L'OOOA'
L'OOOS'
L'OI07'

BAl
DC
L
lR
lR
l
BALR
lA
BAlR
DC
DC
DC
MVC
MVC
l
PACK
lAP
AP
lAP
MVI
lAP
MVI
l
UNPK

STATIC STORAGE
X'0000300040005000'
X'60007000'
X'OOOOOOOO'
X'OO0800fB'
A(N'OIOO' )

DC
DC
DC
DC
DC

OS
De

DC
DC

X, ,

V(N'OOlZ')
ClOOO4
X'OOCOOOOOOOOOOOOC'
X'FFFFFFFFFFfFFfFF •
J('OOFEFFfF'

X'OC'

X'lC'

END

06/06/66

PAGE 004

..,.
~

~CD

.
.

DUS Pl/l

COMPllE~

160N~Pl-4b~

Cl2-0

CHA I ~~

Ob/Ob/b6

PAGE 005

06/06/66

PAGE 006

SVM80L

TYPE

ESIC

AOOR

LE~CTH

0001
0002

000000

000088

LAST
·1 JKSlCN

ESJQ

EXTERNAL SYMBOL TASlE

t-I

~....
CD

DOS Pl/l COMpILER

BLOCK

LENGTH Of OSA

360N-Pl-464 el2-0

ChAIN

8l0CKTABlE

0110

"d
S»
11
rt

~
~

I'tI

en
CD

~
..,.

g
\Q
\Q

..,.

c.
en
N

o,.J

JOB

COMPILE

06/06/66

ACTION TAKE"
MAP
PHASE CHAIN,S
LIST
AUTOllNK
IJCFlllO
AUTOllNK
IJCFl014
LIST
AUTOllNK
IJKSYSA·
LIST
LIST
AUTOllNK
IJJCP1N
AUTOllNK
IJKSlCA
LIST
AUTOllNK
IJKSILM
liST
AUTOllNK
IJKTCBM
LIST
AUTOLINK
IJKTXCF
LIST
ENTRY
LIST

LI ST

DISK LINKAGE EDITOR DIAGNOSTIC OF INPUT

I\,)

CI.l
CI.l

'<
en

..,.
!ozj

I.Q

s::::

LOCORE

HICORE

DSK-AD

ESD TYPE

LABEL

LOADED

REL-FR

002190

002000

00301F

30 1 1

CSECT

INCARD

002000

U'I

CSECT

IJCflllO

002488

0024B8

IJKTXCF
IJKTXCR
IJKTXCW

003008
003030
003042

003008

I'd

CSECT
ENTRY
• ENTRY
•

1-1

eSECT

OTCARO

0020AO

CSECT

IJCFlOl4

002508

....ro

CSECT

STAHr

002190

ttl

CSECT
ENTRY

SUBl
REPEAT

0022FO
002354

0022FO

.
.
~

I\,)

a
ttl

11
0

I.Q

11
~

PHASE

XFR-AD

CI.l

11
0

06/06/66

CHAIN

(j)

I.Q

s::::

CSECT

IJKSlLH

0020AO

002040

CSECT

IJKSYSA

002580

0025BO

..,.
p"

I'd

CSECI

IJKTC8H

0020EO

11
rt

CSECT

LAST

002400

~
I\,)

eSECT
ENTRY
ENTRY
ENTRY
ENTRY
• ENTRY
• ENTRY
ENTRY
• ENTRY
ENTRY
ENTRY

I JKSICN
IJKSlCA
IJKSlCH
IJKTOPH
IJKTCLH
IJKSlCI
IJKSlCP
IJKSZCS
IJKSlCT
IJKlWSA
IJKlWSI

002158
002988
002990
00298A
0029C6
0027E6
00291A
002816
0029AC
002A20
002844

002758

CSECT
• ENTRY

IJJCP1N
IJJCP3

002680
002680

002680

t'I:j
~.

~t;
(1)

.
.

FLOATING POINT REGISTERS

I\J
U'1

I'd

I-'
CD

I'd

~
~

en
CD
C'l

rt
~.

t:S

.r=
~

c::

\Q
\Q

t:S

)I
~.

0..

I\J
..,J
..,J

o
0000000000000000

.r=

REGO
REGS
40-CSW

00003218
00002000
KEY-DO

1E0024l0
00002400

ADDR-002040

EXTERN~L
PSW COrtTENTS
r 1 ELD
FORMAT-OLD 18
BIT-lll11111
SYSTEM MASK
PI{OTEC TlON KEY HEX-O
AMWP
BIT-alaI
PHERRUPT CODE HEX-OOOO
INSTR LENGTH
DEC-O
CUNDITION CODE DEC-O
PROGRAM MASK
BIT-OOOO
HEX-OOOOOO
INSTR ADDRESS

4C-UNUSED-OOOOOODO

000022A8
00002000

2
0000000000000000

00003218
00002000

STATUS-0000I00000000000

00002270
00002480
COUNT~OOOOOO

INTE~RUPT
SUPERviSOR CALL
-NEW 58
-OLD 20
-NEW 60
-00000000 -11111111 -00000000
-0
-0
-0
-0101
-0100
-0100
-0000
-0000
-0000
-0
-0
-1
-0
-0
-0
-l1ll
-0000
-1110
-DOOGCD
-0024CA
·-000B02

50-TIMER-FE9912FF

4
0000000000000000

000022ce
00003318

OOOO.3lBF

00002428

4R-CAW

PROGRAM CHECK
-NEW b8
-OLD 28
-11111111 -00000000
-0
-0
-0100
-0101
-0000
-0007
-3
-0
-2
-0
-0000
-1110
-00OC94
-0021t8A

54-UNUSED-00FFB342

KEy-aD

6
0000000000000000
00.0010DO
000021'>8

ADDR-OO?018

MACHINE: CHECK
-OLD )0
-NEW 70
-01011011 -00000000
-0
-5
-1011
-0000
-(2C5
-0000
-3
-0
-0
-1
-0110
-0000
-OlF340
-0008F4

INPUT/OUTPUT
-NEW 18
-OLD 38
-11111111 -00000000
-0
-0
-0100
-0101
-0000
-oooc
-0
-2
-0
-0
-0000
-1110
-000206
-D024CA

~
....,

"iI
..,0
~
t1

8en

(I)

en
~

,f::

~
....

til

~....

~
..,0
(I)

_. ___ ._

~ .________
. ___

A(BUFFER)
00002050

BUF~NGTH
START
OF OTF oeOOOOOI
TABLE
004 048
0008000

002020

00002038

00002040

00002488

02810202

00003008

02002050

20000048

002040

47000000

47000048

00000000

00002050
BUFFER
C940('104

40e140C4

CIE)CI40

C3CID9C4

5C5C5C5C

5C5C5(.5C

5C5C5(.01

Cl\34780

002060 TO THE NEXT LINE ADDRESS CONTAINS 5C5e5C5e
t\1

t\1
t1

~___ ._

(I)

REMAIN
IN

OTF-A
'INCARO'
002000

002080
0020AO

5C5C5C5C
5C5C5C5C
OTF-A 'OTCARO'
22002088
01002000

5C5C5C5C
45100050

10000000

00002140

00500050

00008400

04000002

0020CO

000020EO

00000000

00002508

04904141

00002140

40404040

01004100

00000040

0020EO

41002140

20000050

010020FO

20000050

40404040

5(.5(.5C5C

5('5C5C5C

5C5C5C5C

5C5('5C5C

58eOEOOO

(I)

t\1

002100 TO THE NEXT LINE ADDRESS CONTAINS 40404040

002140

002160 TO THE NEXT LINE ADDRESS CONTAINS 5C5('5C5('

002180
0021AO

0021CO

t\1
Sl.I

,f::

C940CI04

40C140C4

CIE3(.140

(.3(.109(.4

5e5('5('5C

5C5('5CFO

FOFOFOF1

STAHT
105F00100

45EOFOOA

00002210

189F1831

58FOC054

05E041EO

EOOE051F

010022AO

00000198

4110(,044

58FOC050

05EF0203

00503000

4110COI0

58FO('04('

05EF4110

C01458FO

C04C05EF

41E0013F

5C5C5('5C

0021EO

90EE0058

180058FO

(,0584110

005805EF

411000F4

5010C020

4110(.018

58FOC040

002200

05EF4120

C0304A20

20000201

C0600048

4150C05C

50502004

92882004

024A0148

002220

00F40204

0193D13F

41100148

5010(,02(,

4110e024

58FOC040

05EF4'EO

013F90EE

002240

00581800

58FOC064

41100058

05EF41FO

905E4110

eOl058FO

C04805EF

4110(,014

002260

58FOC.048

05EF5800

000498EC

000C01FE

00003000

40005000

60007000

00000000

002280

80002000

800020AO

58002000

00000048

00003174

230020AO

00000050

000031C8

0022AO

0008F8f8

00002192

OA002000

880022CC

000020EO

8F0025BO

00002ge6

0000298A

0022CO

00002DAO

00002990

000022FO

00002192

00000000

FFFFFFFF

OOFEFFFF

oeoooooo

000100CO
SUB 1
190ECOOOC

00002354

0022EO

45EOFooe

00002388

58COEOOO

002300

189F 1831

58FOC01C

05E041EO

EOOE051F

030023C8

00000100

02030058

30000203

002320

00503C04

58600058

02046000

C0305850

005841EO

500090E.E

DOt- H800

58fOC020

OU2340

41100060

05Ef5800

000498EC

OOOC01FE

47f0909C

REPEAT
I90ECOOOC

45EOFOOC

00002388

002360

58eoECOO

'>890C024

183158fO

C01C0700

05E041EO

EOO€051F

030023C8

00000100

002380

02030058

30000203

00503004

58600058

41E06000

90EED060

lA0058FO

e0204110

0023AO

006005EF

~800C004

98EeDooe

07FE5800

000498EC

000CQ7FE

00003000

40005000

0023CO

600070CO

00000000

000CF8F8

000022FO

00002354

00002758

00002400

000022fO

0023EO

OCCOCCOO
LAST
90f:COCOC

00000000

FFFFFFfF

FFFfFFFF

OOFEFFFF

OCFOFOFO

fOF002FF

03fFOQfF

45EOFOOC

00002480

58COEOOO

189F 1831

58FOCOl8

0!JE:041EO

EOOE051f

03002490

00000110

02030058

30000203

00503004

586000')8

F224DOEO

6000f8fO

002400
002420

....
I'I.j

002440

0060C035

FAF20060

00EOF83F

00E8D060

92f9C012

F8230070

·00E892F8

C0125860

000498E(.

')00,"07FE

00002158

00000",')0

.(1)

002460

0058F342

60000070

96F06Q04

58000004

98E(.000C

07FE5800

002480

00003000

1t0005000

60001000

00000000

0008F8F8

00002ltOO

,f::

I\J
U'I

•

~
~

IJCFZIZO

0024AO

OOOOOCOO

00000000

fFFffFFF

OOFEfFFF

OC lCOOOO

I0A320000

OA320000

91011004

4180F02C

91401002
58EOF04('

0024CO

1t7FOFOI0

OA3200oo

OA009180

10024710

foltOA07

0024EO

1t780FOI0

50EOFOttC

58EOI020

0501F04A

E0004770

F04lt58EO

101C01FE

lJCFZOZ4

002500

07FE615C

00002f4~

IOA320000

OA320000

OAl20000

50EOFOAlt

OA009180

10024710

002'i20

FOICOAo7

91101003

4710F040

D2001030

10280201

1036102E

58EOI028

024FI038

002540

EOO.058EO

FOAItOlFE

41El0030

50EOI008

OA009180

10024710

F0540A07

18E141l0

002560

F0780AOO

9180F07A

4710F066

OA07181E

ItlE 10028

50EOI008

47FOfOl0

00000000

002580

00000000

00000004

00002590

00000000

09002598

00000000

CIl

FFFFFFFF

(I)

t'd

a
11

F4Flf5FO

C94009C5

00000000

81010078

00000000

002SAO

E3D9E800

00000000

0025CO

00002600

00000000

07000700

00008400

08000003

000025F8

00000000

t'd
I»

0025EO

03002b80

3100C9Dl

E2E8E2D3

E2400040

90002600

47000000

01002600

20000079

002600 TO THE NEXT LINE ADDRESS CONTAINS 401t04040

002660

401t0404o

40404040

40000000

00000000

002680

00000000

0026AO

00091803

U'I

1t0404040

40404040

00000000

47FOFOOC

9023FOOO

91031015

4770F066

58301028

06304120

1t3030~00

1t332FOCB

19034780

F0364620

F0284920

FOCA47BO

F0469208

I"tI
~

CIl

OOOOOOOB

A0002500

IJKSYSA

0026CO

10284530

FOA09180

10154780

F056947F

10154530

FOA04122

FOCOOZOO

10282000

0026EO

06001028

101EOAOO

91041015

4780F07A

4530FOA2

92011028

OA004530

FOA29101

002700

10044.780

F09A9104

10154710

F09A9108

101F4710

FOAE58EO

10189823

F00007FE

002720

OA009180

10024710

FOACOA07

07F31821

48020006

41101016

47FOF09A

002740

OBCBE38B

0301411B

13080004

f9C3Fl4E

E5E660FO

4010C128

OA021812
IJKSZCN
905CF408

58COF2B4

(I)

002760

58500008

50050004

18Ao1805

5A510004

5550C29C

41BOC01E

50500008

0200C28A

....rt

002780

C2884850

('29A1255

4780(,038

485A0048

41550001

4050('29A

40500048

0203004C

0027AO

A0145851

00005050

00000200

50025003

91030000

4710(,078

4A550000

96405004

47FOC064

985CC408

07FE41FO

COSE985C

C40892~3

9680C3EE

5800C428

C3EFltll0

,f::

0027CO

91805004

t:J

OOHEO

U000051F

4710(,078
41550008
IIJKSZCI
210~5FO
58('OF224

t:T

s:=

002800

(,2689202

C2684780

COE892fl

C2959200

C28F47FO

02330014
C3F49500
IIJKSZCS
C27A~OEC
000C58CO

\Q
\Q

....

002820

C2B8947F

C3EE947f

(.39492FE

C1191850

91055COO

4740COfO

58550004

47FOCOD8

~
\Q

41FOC006

58450000

9'>061COO

4720C122

02COC109

10000COO

CI09C118

(I)

F I Fb9200

002840

920EC 119

002860

91004002

4710C122

947FC3EE

982C001C

01FEOAll

20408010

040818b5

91036000

oe2880

4780(,198

471C(.152

58460000

4A440000

D5004000

lC0041RO

C1SA9180

40()441l 0

....QI
I\J

~
tX)

I1:j

.."

s::

0028AO

C1524144

t:1
0

(1)

0028CO

CIl

11
~

"ttl

CIl

'<
(J)

~
H

CIl

m
(1)

C124950A

10004740

C174950F

4001C3A1

4770C142

91404004

4710C142

91104004

4110C198

91084004

4780C 110

947FC394

58140004

47FOCICA

95081000

4170CU8

9640C'j'~4

47FOC110

002900

18619101

C28A92F4

C2954110

CIBA92F2

C2959200

C28F47FO

C27

002920

FOf21850

91055000

4180C22A

4740CIEO

58550004

47FOC1CC

05011004

50484770

002940

C 1081805

58610000

50600038

4A610006

50600010

58550000

02005002

500.39180

O(J2960

C3EE4170

C21498EC

947fC3EE

C3F 10011

5000C4,6

OA114510

COC62200

0233C3f4
IIJKSZCM
.9208F082

00140202

002980

000C01FF
IIJKSZCA
.9207f08A

184C58CO

F07C4700

C24C4510

0029AO

C08E2300

9200C243

47FOC438

F03F9200

f03941FO

F02

~JKTCLM

0029CO

F02045FO

F012F3

F0279204

0029EO

41fl0000

58C901D2

0020400

F0140A04
PROG MASK
OEOOOOOO.
SSA2
05000000

en
s::

10004720

CI740502

to
0

(1)

000447FO

0028EO

(1)

(J)

.."

C1385866

c.n

(1)

to
11
0

000847FO

002A20

45FOFOOC
,IJKSZCT
05F092F3

00002844

000027E6

00002758

OOO~~

00000000

00002840
SAVE
00000000

J!iK.:zCP
OEC
000C58CO

~JKTOPM

2FO

F0339200

00002868

00002830

00000000

002A40 TO THE NEXT LINE ADDRESS CONTAINS 00000000

11
rt

002A60

00000000

SSAI
05000000

00003080

00000000

0000ZF40

002A80

00002488

00003080

00002018

00002050

00002000

00000048

40002180

58002000

002440

00000048

000Q3174

40002192

00002000

00002A20

000020EO

00000000

002ACO TO THE NEXr LINE ADDRESS CONTAINS 00000000
SCALE FACTOR~
002AEO
00000000
00000000
000000 0
FORMAT
ENTRY POINT TABLE
002800 TO THE NEXT LINE ADDRESS CONTAINS 00000000

CURRENT
FILE
230020AO

FILE CHAIN
ANCHOR
FEC020AO

0'1

H\
~

tX)

100~~0~O

002820

00000000

002840

002880

AOO02580
REG 5
000022CC
REG 13
00000000

002BAO

00899001

58800088

91808000

4780COB4

91808008

4780CO'3£

'H088004

4180C03E

002BCO

96048004

48100016

02008019

104£:9140

80044180

C04A9407

80000600

80009000

0028EO

94098009

91428008

4710COBC

58108GCO

D7011000

100058?0

80104110

Cl'100202

002COO

C0798001

4500C01C

0000}088

OA025QlO

80109680

800991'11

80094710

(;09A0202

0()2C20

80050080

02020080

D089947F

80009180

90004199

OC0447~O

COOE5f:lDO

000,,98EC

O()2C40

DOOC07FE

94FB8004

47FOC09E

91208000

4710C05C

58108000

41110000

95101014

.002e60

4770COSC

0208C144

10C00203

1000(140

CH801015

4710CCFO

9207C 138

45f:OClIE

002C80

91101015

923FC138

4180CllO

9227C138

45EOC 11E

91011C"4

7l0C 114

92 3 7C 138

Ou2CAO

4SEOCllE

02081000

C14447FO

C0824120

C1385021

00080AOO

91801002

4.71 OC 132

002860

00000000
00000000
,IJKZWSI - INTERRUPT SAVE
ooooooooPSWoooooQQQI
REG 7
REG 6
0000318F
00001000
REG 15
REG 14
00000000
00000000

00000000
AREA REG 0
00000000
REG 8
00002000

DECIMAL INTEGER. .
00000000
a a OJH)(LQ 0

OOOCCOOO

00000000
REG 1
00000000
REG 9
00002400
START OF 65
18C11880

00000000
REG 2
00000000
REG 10
00002000
DOUBLE WORD
41D040RO

00000000
COO02S3C
REG 3
REG 4
00000000
00000000
REG 11
REG 12
00007000
QOC02480
PVI TRANSIENT AREA
50R00004
18950202

tzj
~.

\,Q

002eeo

Ol0701FE

00000000

01000000

20000001

00000400

00000100

00000138

00000000

OOleEO

.s858e2D6

07C50540

00000000

OOO(JOOOO

(1)

N
U1

•

t:"'

'en"
t-I

~....

002000 TO THE NEXT LINE IDDRESS CONTAINS 00000000
002080
002010
0020CO
0020EO

00145820

50109180

50004780

C0484510

COIE5858

00001000

05400202

90E~000C

58~OF038

41]040E4

43210000

89200019

88200011

1A210202

20011001

91801000

UllOO04

4780Fooe

98E4000C

07FE~510

C1080501

00002120

~170C078

18CF5880

C2185000

80544100

80509868

10005060

8008949F

600943FO

18Cl9834

,IJKSZLM

IIJKTCBM

90£COOOC

002EOO

600044FO

C20C4710

C03258FO

C21e051F

65009140

80085820

60104110

C1309180

002£20

60094770

COAE9111

800841CO

C0744970

60164700

C05E4810

60169620

60095810

(1)

60109130

60044180

C06E4lAA

0001tl898

45EOC1EA

58106000

41110000

58FlOOIO

002E40
002£60

910C6004

4780C08E

lB0045EF

00044840

601641EO

C09E90fC

OOOC47FF

00oC91l0

1'1

002E80

80084780

e0819608

600041FO

C1C6947F

60094830

600A4133

00049610

60094970

002EAO

60014180

C0024740

COC£4870

600A9620

60099130

60044740

CI004070

60169101

002ECO

80084110

COF45020

80980202

80018099

47FOC1C6

189818A2

45EOCIEI

47FOCIEO

\,Q

S»

1'1

rt
~

002EEO

41970004

19934100

C1115810

60004111

000058F1

001045EF

00148990

00105090

20008098

41220004

47FOCOOE

947F6009

91026009

4780C144

58FOC21C

41110000

4840600A

41EOC160

90ECOOOC

47FFOO08

002FOO

80980203

002F20

051FOAOO

58106000

58F10010

002F40

50206010

94F76tlOO

91306004

4780C182

02018098

20004840

80984840

C2204122

OU2F60

00044040

60169101

80084110

C19C5020

80980202

10018099

47FOC184

19744780

002F80

C1IC4100

C1I81814

96206009

18921818

45EOClEA

58FOC21C

91406009

4780CIC6

002FIO

4110C222

05EF58FO

C21C9130

60044710

C1E09120

60094180

elE04110

C221t05EF

002FCO

58000004

98ECOOOC

01FEltIFO

0100197F

4700e204

187F02FF

10009000

1I9FUAF

..J
H\
~

en
(1)
n

002FEO

41FOC1EE

06104470

003000

00040COO

OEOOOOOO

003020

000t01FE

58FOE06£

91008008

02001000

90000000

00002120

00002816

,05E05880

E0729140

80084710

E01A5850

80089602

500998Et

980COO14

051FOAOO

05E05880

E04A5850

80089640

500947fO

05E09101

50084780

EOIE9130

50044190

~940CI04

40C140C4

C tE3C 140_

5C5C5C'iC

5cse5C5C

5C5C5C5C

C21007FE

,IJKTXCF

0
t:I
~

98ECOOOC

t:I
(1)
tr
~

\Q
\,Q

t:I

en
N

REG 15

00002354

REG 7

0030AO
0030CO
0030EO TO THE NEXT LINE ACORESS CONTAINS 00000000
003160

00000000

OOOOCOOO

00000000

003180

C3~4__ ~L5C5t5'

_ 5C5C5C5'

_5C5C5C~_C

5CSC5C5C

I CARDIN

tv
co
tv

t-:j
.....
I.Q

s:::

t:I

C/)

CIl

'<
en

(t)

.
~

ttl

I.Q
t;

III

~(t)

(j)

s:::
.....
0..
(t)

5C5C5C5C

0031CO

CTR
FOFOfOF2.

SCSC5C5C

oooooeoo

SCSC5C5C
CARDOUT
C940CI04

5C5C5C5C

5CSC5CSC

SC5C5CljC

SC 5CSC5C

~TR
5C5C5C_
Q

40C140C4

CIE3Cl40

C3CI09C4

S("SC5C5C

5CSC5C5C

U'I

003lEO TO THE NEXT LINE ACORESS CONTAINS 5C5C5C5C
003200

5C5C5C5C

5C5C5CSC

5C5C5C5C

5C5C5CFO

fOFOFOFl

DSA 'SUB l'
030023(.'3

00C03080

003220

00003318

1E0023A4

00002400

00003218

000022A8

'00003008

00002270

III

OU3240

000022CC

000031BF

00001000

00002000

000022FO

00002000

00002388

.....

OU3260

00020000

00003080

00000000

00003lBF

00000000

00003lBF

00000000

ttl

003280 TO THE NEXT LINE ACORESS CONTAINS 00000000

(t)

0031AO

ttl

~
H

C/)

a
"0
(t)

I.Q
t;

003300

00000000

oocooono

DSA 'LAST '
03002490

00003218

003320

00003428

00000000

ttl
III

003340 TO THE NEXT LINE ACORESS CONTA1NS 00000000

003360

00030000

00003218

00000000

00003lBF

00000000

OU3380

00000000

0000002C

00002COO

00000000

I-h

OU33AO TO THE NEXT LINE AOORESS CONTAINS 00000000
0033EO

00000000

COOOIFOO

00000000

003400

0000002C

00000000

003420 TO THE NEXT LINE ADDRESS CONTAINS 00000000
007FEO

00000000

Glossary

For a more complete list of data processing
terms, refer to IBM Data processing
Techniques, A Data Processing Glossary,
GC20-1699.
access method: Any of" the data management
techniques (sequential, indexed sequential,
or direct) available to the user for
transferring data between main storage and
an input/output device.
ASCII (American National Standard Code for
Information Interchange): A 128-character,
7-bit code. The high-order bit in the
system/360 8-bit environment is zero.
asynchronous: Without regular time
relationship. The user's programs run
asynchronously with the I/O interrupts.
BTAM's channel appendage routine runs
synchronously with the I/O interrupts.
background program: In multiprogramming,
the background program is the program with
lowest priority. Background programs
execute from a stacked job input.
batched job: Programs that execute from a
stacked job input. Batched jobs run under
the control of job control.
block:
r:--To group records physically for the
purpose of conserving storage space or
increasing the efficiency of access or
processing.
2.

A physical record on tape or DASD.

BTAM (Basic Telecommunications Access
Method): A basic access method that
permits a READ/wRITE communication with
remote devices.
catalog: To enter a phase, module, or book
into one of the system libraries.
CCB:

See Command Control Block.

CCH:

See Channel Check Handler.

Channel Check Handler (CCH): A feature
that assesses system/370 channel errors to
determine if the system can continue
operations.
channel program: One or more Channel
Command Words (CCWs) that control(s) a
specific sequence of channel operations.
Execution of the specific sequence is
initiated by a single start I/O
instruction.

checkpoint record: A record containing the
status of the job and of the system at the
time the checkpoint routine writes the
record. This record provides the necessary
information for restarting a job without
returning to the beginning of the job.
checkpoint/restart: A means of restarting
execution of a program at some point other
that the beginning. When a checkpoint
macro instruction is issued in a problem
program, checkpoint records are created.
These records contain the status of the
program and the machine. When it is
necessary to restart a program at a point
other than the beginning, the restart
procedure uses the checkpoint records to
reinitialize the system.
checkpoint routine: A routine that records
information for a checkpoint.
Command Control Block (CCB): A 16-byte
field required for each channel program
executed by physical IOCS. This field is
used for communication between physical
IOCS and the problem program.
communications region: An area of the
supervisor set aside for interprogram and
intraprogram communication. It contains
information useful to both the supervisor
and the problem program.
control program: A group of programs that
provides functions such as the handling of
input/output operations, error detection
and recovery, program loading, and
communication between the program and the
operator. IPL, supervisor, and job control
make up the control program in the Disk and
Tape Operating systems.
core storage:

See main storage.

data file: A collection of related data
records organized in a specific manner.
For example, a payroll file (one record for
each employee, showing his rate of pay,
deductions, etc) or an inventory file (one
record for each inventory item, showing the
cost, selling price, number in stock, etc).
data protection: A safeguard inVOked to
prevent the loss or destruction of customer
data.
data set security: A feature that provides
protection for disk files. A data secured
file cannot be accidentally accessed by a
problem program.
Glossary

283

Disk Operating System (DOS): A disk
resident programmdng system that provides
operating system capabilities for 16K and
larger IBM System/360 and IBM system/310.
DOS:

See Disk Operating system.

DOS Volume Statistics: A facility that
monitors and records the number of
temporary read and write errors on
currently accessed tape volumes. This
facility has two options, Error statistics
by Tape Volume (ESTV) and Error Volume
Analysis (EVA).
OTF (define the file) macro instruction: A
macro instruction that describes the
characteristics of a logical input/output
file, indicates the type of processing for
the file, and specifies the main storage
area and routines to process the file. To
do this, use the appropriate entries in the
keyword operands associated with the DTF
macro instruction.
Environmental Recording, Editing, and
Printing (EREP): A program that edits and
prints data that has been stored on the
system recorder file.

To transfer control to·the system
loader.

Fetch/Load Trace (F/L Trace): A program
that records information about phases and
transients as they are called from the core
image libra'ry.
file:

See data file.

fixed length record: A record having the
same length as all other records with which
it is logically or physically associated.
F/L Trace:

See Fetch/Load Trace.

foreground initiation: A set of system
routines to process operator commands for
initiation of a foreground program.
foreground program: In multiprogramming,
foreground programs are the highest
priority programs. Foreground programs may
be executed from a job stack or in an SPI
environment.

EREP: See Environmental Recording,
EdIting, and Printing.

Forms Control.Buffer (FCB): The buffer in
the IBM 3811 Printer Control Unit that
stores carriage control information for the
IBM 3211 Printer.

Error Statistics by Tape Volume (ESTV):
One of the two options of the DOS volume
Statistics. With ESTV support, the system
collects data on tape errors by volume for
any tape volumes used by the system.

'Generalized supervisor Calls' Trace" (GSVC
Trace): A program that records SVC
interrupts as they occur. Allor a
selected group of SVcs can be traced.

Error Volume Analysis (EVA): One of the
two options of the DOS Volume statistics.
With this option, the system issues a
message to the operator when a number of
temporary read or write errors (specified
by the user at system generation time) has
been exceeded on a currently accessed tape
volume.
ESTV:
EVA:

See Error Statistics by Tape Volume.
See Error Volume Analysis.

extent: The physical locations on
Input/Output devices occupied by or
reserved for a particular file.
fetch:
r:--To bring a program phase into main
storage from a core image library for
immediate execution.
2.

284

The routine that retrieves requested
phases and loads them into main storage
(see system loader).
The name of a macro instruction (FETCH)
used to tranfer control to the system
loader.
DOS System Programmer's Guide

GSVC Trace: See Generalized supervisor
Calls Trace.
lORA:
Area.

See Independent Directory Read-in

Independent Directory Read-in Area (IDRA):
A resident area, created by supervisor
option, into which the system reads core
image library directories for fetch and
load operations. Using IDRA frees the
physical transient area to perform error
recovery procedures.
Initial Program Load (IPL): The
initialization procedure that causes Disk
and Tape Operating Systems to commence
operation.
interrupt: A break in the normal sequence
of instruction execution. It causes an
automatic transfer to a preset storage
location where appropriate action is taken.
IOCS (Input/Output Control system): A
group of macro instruction routines
provided by IBM for handling the transfer
of data between main storage and external
storage devices.

I/O (inpUt/output) error logging: The
process of recording OBR and SDR records or
the system recorder file.
I/O Trace (Input/Output Trace): A program
that records I/O device activity for all or
a selected group of I/O devices.
IPL:

See Initial Program Load.

data files.
LSERV (label cylinder display): A program
that formats a listing of the label
cylinder located on SYSRES. LSERV can run
in and partition any outputs the list on
SYSLST, which may be assigned to disk,
tape, or printer.

Job Accounting Interface: A program that
accumulates accounting information for each
job step to: charge usage of the system,
help plan new applications, and help
supervise system operation more
efficiently.

Machine Check Analysis and Recording
(MCAR): A feature that records System/370
machine check interrupt error information
on the system recorder file and then
attempts to recover from the interrupt.

Job Accounting Table: An area in the
supervisor where accounting information is
accumulated for the user.

main storage: All addressable storage from
which instructions can be executed or from
which data can be loaded directly into
registers.

job control: A program that is called into
storage to prepare each job or job step to
be run. Some of its functions are to
assign I/O devices to certain symbolic
names, set switches for program use, log
(or print) job control statements, and
fetch the first program phase of each job
step.
job step: The execution of a single
processing program.
~:

1024.

language translators: A general term for
any assembler, compiler, or other routine
that accepts statements in one language and
produces equivalent machine language
instructions. For example, Assembler,
COBOL, etc are language translators.
librarian: The set of programs that
maintains, services, and organizes the
system libraries.

MPS:

multiplexer channel: A channel designed to
operate with a number of I/O devices
simultaneously on a byte basis. That is,
several I/O devices can be transferring
records over the multiplexer channel, time
sharing it on a byte basis.
multiprogramming system: A system that
controls more than one program
simultaneously by interleaving their
execution.
multitask operation: Multiprogramming;
called multitask operation to express not
only concurrent execution of one or more
programs in a partition, but also of a
single reenterable program used by many
tasks.
OBR:
OLTEP:

library: An organized collection of
programs, source statements, or object
modules maintained on the system resident
device. Three libraries are used by the
Disk and Tape operating systems: core
image library, source statement library,
and relocatable library.
linkage editor: A system service program
that edits the output of language
translators and produces executable program
phases. It relocates programs or program
sections and links together separately
assembled (or compiled) sections.
load: TO fetch, i.e., to read a phase into
main storage returning control to the
calling phase.
lo.gical loeS: A comprehensive set of macro
instruction routines provided to handle
creation, retrieval, and maintenance of

See multiprogramming system.

See outboard recorder.
See On-Line Test Executive Program.

On-Line Test Executive Program (OLTEP):
The control program of the on-line test
system. OLTEP is the interface between the
on-line test and the operating system.
outboard recorder (OBR): A feature that
records pertinent data on the system
recorder file when an unrecoverable I/O
error occurs.
overlap: To do something at the same time
that something else is being done; for
example, to perform input/output operations
while instructions are being executed by
the central processing unit.
overlay: A program segment (phase) that is
loaded into main storage. It replaces all
or part of a previously retrieved section.
PCIL:

See Private Core Image Library.
Glossary

285

PDAID:

See Problem Determination Aids.

phase: The smallest complete unit that can
be referenced in a core image library.
Each program overlay is a complete phase.
If the program has no overlays, the program
itself is a complete phase.
physical IOCS: Macro instructions and
supervisor routines (Channel Scheduler)
that schedule and supervise the execution
of channel programs. Physical IOCS
controls the actual transfer of records
between the external storage medium and
main storage, and provides I/O device error
recovery.
Private Core Image Library (peIL): A file
reference in the same manner and for the
same purpose as the system core image
library, but distinct from the system core
image library. PCIL increases available
core image library space to enable
compiling, linkage editing, and executing
in the foreground partition, when a private
core image library is assigned to that
foreground partition.
private library: A core image,
relocatable, or source statement library
that is separate and distinct from the
system library.
problem determination: A procedure or
process (provided by IBM) that the user can
follow after an error message to determine
the cause of the error.
Problem Determination Aids (PDAID):
Programs that trace a specified event when
it occurs during the operation of a
program. The traces provided are: QTAM
Trace, I/O Trace, F/L Trace, and GSVC
Trace.
QTAM Trace: A program that records certain
supervisor and I/O activities on tape or in
main storage.
RDE:

See Reliability Data Extractor.

record: A general term for any unit of
data that is distinct from all others when
considered in a particular context.
o

Recovery Management Support (RMS): A
feature for System/370 that consists of the
MCAR (Machine Check Analysis and Recording)
and CCH (Channel Check Handler) functions.
RMS gathers information about System/370
hardware reliability and attempts certain
error recovery operations. RMS is a part
of the entire reliability, availability,
and serviceability support for System/370.
Reliability Data Extractor (ROE): A
function that provides hardware reliability
data that is analyzed by IBM.
286

DOS System Programmer's Guide

,relocatable: A module or control section
whose address constants can be modified to
compensate for a change in origin.
,restart:
RMS:

See checkpoint/restart.

See Recovery Management Support.

SDR (statistical data recorder): A feature
'that records the cumulative error status of
an I/O device on the system recorder file.
,selector channel: A channel designed to
operate with only one I/O device at a time.
Once the I/O device is selected, a complete
record is transferred one byte at a time.
self-relocating: A programmed routine that
is loaded at any doubleword boundary and
can adjust its address values so as to be
executed at that location.
self-relocating program: A program that is
able to run in any area of storage by
having an initialization routine to modify
all address constants at object time.
Single Program Initiator (SPI): Under DOS,
a program that is called into storage to
perform job control type functions for
foreground programs not executing in batch
job mode.
SORTED DSERV: A program that gives you an
alphamerically sorted listing of any or all
of the library directories.
SPI:

See Single Program Initiator.

stand-alone dump: A program that displays
the contents of main storage from a minimum
of 8K bytes to a maximum of 16384K bytes.
It helps to determine the cause of an
error.
supervisor: A component of the control
program. It consists of routines to
control the functions of program loading,
machine interruptions, external
interruptions, operator communications and
physical loes requests and interruptions.
The supervisor alone operates in the
privileged (supervisor) state. It coexists
in main storage with' problem programs.
system residence: The external storage
space allocated for storing the basic
operating system. It refers to an on-line
tape reel or disk pack that contains the
necessary programs required for executing a
job on the data processing system.
task selection: The supervisor mechanism
for determining which program should gain
control of CPU processing.

telecommunications: A general term
expressing data transmission between remote
locations.

supervisor) into the transient area when
needed for execution.

teleprocessing: A term associated with IBM
telecommunications systems expressing data
transmission between a computer and remote
devices.

UCS:
UCSB:

trace:
r:--To record a series of events as they
occur.

undefined record: A record having an
unspecified or unknown length.

The record of a series of events.

track hold: A function for protecting DASD
tracks that are currenbly being processed.
When track hold is specified in the DTF, a
track that is being modified by a task in
one partition cannot be concurrently
accessed by a task or subtask in another
partition.
transient area: This is a main storage
area (within the supervisor area) used for
temporary storage of transient routines.
transient routines: These self-relocating
routines are permanently stored on the
system residence device and loaded (by the

See Universal Character Set.
See Universal Character Set Buffer.

Universal Character Set: A printer feature
that permits the use of a variety of
character arrays.
Universal Character Set Buffer (UCSB):
buffer in a printer control unit that
stores the code equivalents of the
characters on an interchangeable print
chain or train cartridge.

variable length record: A record having a
length independent of the length of other
records with which it is logically or
physically associated. (Contrasted with
fixed length record.) It contains fields
specifying physical and logical record
lengths.

Glossary

287

288

DOS System Programmer's Guide

Index
Indexes to systems reference library
manuals are consolidated in the publication
DOS Master Index, GC24-5063. For
additional information about any subject
listed below, refer to other publications
for the same subject in the Master Index.

AB (abnormal termination) support
description of
67
multitasking
39,44
option table
57
ACTION statement
141,143
ACTR instruction
183
ADCON (see relocating address constants)
adding records to indexed sequential file
132
AGO instruction
183
AIF instruction
182
ALLOC macro
84
Am. Natl. Std. Code for Info. Interchange
(see ASCII)
ANOP instruction
183
AP (asynchronous processing) (see
multitasking)
ASCII
generation guidelines
29
support
28
system considerations
28
assembler program debugging aids
204
asynchronous processing (AP) (see
multitasking)
ATTACH macro
cancel conditions
40
considerations
39
example
45
use of
39
attributes, macro writing
178
background (BG) communications region
extension
35
BJF (batch job foreground)
27
blocking records
99
bypassing checkpoint records
172
cancel codes, description of
198
cancel, system action
198
card (DTFCD/CDMOD) files
122
catalog CI library example
156
catalog for phase overlay example
catalog private CI library example
causes for message OS041
200
CBF (console buffering) support
channel considerations
63
description of
63
table
64

162
158

CC (command chaining) (see CCW)
CCB (command control block)
CCW address
102
CCW address in CSW
103
count field
102
102
CSW status bits
definition of
100
format of
104
optional sense CCW
103
symbolic unit number
102
transmission information
102
3211 error indicators
173
CCH (channel check handler)
189
CCHAIN (command chaining) support
64
CCW (channel command word)
command chaining (CC)
107
106
command code
count
108,111
data address
106
data chaining (CD)
107
flags
107
format of
106
PCI
108
SKIP
108
SLI
107
CD (chain data) (see CCW)
chaining (DAM)
127
channel check handler (CCH)
189
channel command word (see CCW)
channel queue (see CHANQ)
channel rescheduling considerations
87
CHANQ (channel queue)
format of
59,85
size of
84
table
85
checkpoint file
99
checkpoint/restart
bypassing checkpoint records
172
checkpoint file
168
checkpoint on disk
168
checkpoint on tape
168
CHKPT macro, format of
167
CHKPT macro, use of
166
DASD and MICR files
168
DASD operator verification table
172
multitasking considerations
166,167
problem program responsibilities
166
repositing magnetic tape
170
repositioning I/O files
170
restarting checkpointed programs
173
RSTRT statement
173
checkpoints
168
on disk
168
on tape
CHKPT macro
format of
167
use of
166
choosing file organization and retrieval
method
133
choosing file organization
activity
133
criteria
133
random retrieval need
136
response time
136

Index

289

choosing file organization (CONT.)
size
134
volatility
134
COBOL program debugging aids
121,204-206
COBOL sample program
207
COCR (cylinder overflow control record)
130
command chaining support
64
command control block (see CCB)
COMMON pool
154,160
communication between tasks
42
communications region
BG extension
35
SDR
190
system
30
compile and execute example
160
concatenation
181
concurrent peripheral operation (CPO)
27
conditional assembly instructions
ACTR
183
AGO
183
AIF
182
ANOP
183
consecutive spill overflow technique (DAM)
127
console buffering (CBF) table
64
console buffering (see CBF support)
console typewriter (DTFCN)
122
CPO (concurrent peripheral operation)
27
cylinder index (ISFMS)
129
cylinder overflow area (ISFMS)
132
cylinder overflow control record (ISFMS)
130
DAM (see direct access method)
DASD address specification, READ/WRITE
(DAM)
128
DASD file protect support (see DASDFP
support)
DASDFP (DASD file protect) support
mode of operation
70
SYSGEN
73
data file definition
93
data management concepts
91-138
data manipulation
99
debugging aids
167-282
assembler
204
COBOL
204
FORTRAN
234
gathering documentation
189
PL/I
255,263
RPG
243
system action under CANCEL
198
wait states
202
declarative macros
111,113
DEQ macro
cancel condition
42
considerations
42
use of
42
DETACH macro
cancel condition
41
considerations
40
example
47

290

DETACH macro (CONT.)
use of
40
device type codes
90
DIB (disk information block)
definition of
74
format of
75
direct access method
file organization
124
prime number division
126
random addressing formula
125
random addressing techniques
124
synonym records
127
disk information block (see DIB)
divide remainder (see prime number
division)
DOS stand-alone dump generator (DUMPGEN)
189
double buffering
27
DTF (define the file) macros
DTFBG
114
DTFCD
114
DTFCN
114
DTFDA
114
DTFDI
114
DTFEN
114
DTFIS
114
DTFMR
114
DTFMT
114
DTFOR
114
DTFPH
114
DTFPR
114
DTFPT
114
DTFSD
114
DTFSR
114
DUMPGEN (stand-alone dump generator)
189
duplicate addresses, DAM, (see synonyms)
DVCGEN macro
channel rescheduling considerations
87
system generation guidelines
88
ECB (event control block)
39,41
end of day (EOD)
189
ENQ macro
cancel conditions
42
considerations
42
use of
42
ENQ/DEQ and RCB macros example
48
ENTRY statement
141,146
environmental recording, editing, and
printing (EREP)
189
EOD (end of day)
189
EREP (environmental recording, editing, and
printing)
189
error logging
189
error recovery techniques for IBM 3211
printer
174
error statistics by tape volume (ESTV)
58
error volume analysis (EVA)
58
ESTV (error statistics by tape volume)
27
ESTV utilities (ESTVFMT,ESTVUT)
58
EVA (error volume analysis)
58
event control block (see ECB)

DOS System Programmer's Guide
.~

EXCP (execute channel program)
23,100
extended capabilities
MEXIT
183
MNOTE
184
extending an indexed sequential file
132
external interrupt
56
F/L (Fetch/Load) traqe
194
FAVP (first available pointer) table
59
FETCH macro, use of
149
FICL (first in class) table
59
file definition (DTF) macros
113
file
activity
133
definition
93,99
need for random retrieval
136
organization
122-133
processing methods
113
response time consideration
136
size
134
volatility
134
fixed length records
ASCII
93
blocked
95
unblocked
94
FLPTR (free list pointer)
59
FOCL (first on channel list) table
59
FOPT macro
abnormal termination (AB)
67
64
command chaining (CCHAIN)
console buffering (CBF)
63
DASD file protect (DASDFP)
70
ESTV
58
EVA
58
Independent Directory Read-In Area
(IDRA)
64
interval timer (IT)
56
job accounting interface
81
multiple wait (WAITM)
70
operator communications (OC)
56
physical transient overlap (PTO)
62
private core image libraries (PCIL)
76
program check (PC)
58
seek separate (SKSEP)
62
system files on disk (SYSFIL)
74
system generation guidelines
58
tape error block (TEB)
58
tape error recording
58
track hold (TRKHLD)
65
FORTRAN program debugging aids
234
FORTRAN sample program
234
global SET symbols
180
GSVC (Generalized Supervisor Calls) trace
194

halt analysis, RPG
243
header labels, checkpoint
168,172
header statement, macro
178,179

home addresses (DAM)
I/O
I/O
I/O
IBM

127

(Input/Output) trace
194
error logging
189
table interrelationship
59
3211 Printer support
error recove~y techniques
174
system considerations
173

IDRA
definition
64
system generation guidelines
64
imperative macros
111,119
INCLUDE statement
141,145
Independent Directory Read-In Area (IDRA)
64
independent overflow area (ISFMS)
132
index (ISFMS)
cylinder
129
master
129
track
129
indexed sequential file management system
add function
129,132
COCR
130
129
cylinder
load
131
load extend
131
master index
129
prime data
129
random
132
sequential
132
track index
129
interrelationship of DTF and module macro
instructions
118
interrupts
23
42
intertask communication
interval timer support
concept
56
option table
57
IOCS (input/output control system)
logical
111
physical
99
IOTAB macro
channel queue (CHANQ)
84
84
job information block (JIB)
ISFMS (see indexed sequential file
management system)
IT (interval timer) support
concept
56
option table
57
JIB (job information block)
definition of
84
format of
59,72
job accounting interface support
prog. considerations
82
table
83
job information block (see JIB)

label cylinder display (LSERV)
189
LBLTYP statement
146
line printers (DTFPR/PRMOD)
122

Index

291

link edit and execute ~xample
154
link editing, examples of
154
link editing
ACTION statement
143
control statements
143
definition of
141
ENTRY statement
146
INCLUDE statement
145
LBLTYP statement
146
OPTION LINK and CATAL
147
overlay structures
147
PHASE statement
144
private core image libraries
141
program considerations
146
program name
141
symbolic units required
143
system flow
141
linkage editor examples
catalog CI library
156
catalog phase overlay
162
catalog private CI library
158
compile and execute
160
link edit and execute
154
link editing
165
submodular structure
162,163
LIOCS (see logical input/output control
system)
LOAD macro, use of
149
loading an indexed sequential file
131
local SET symbols
180
logic module name structure
116
logical file definition
99
Logical Input/Output Control System (LIOCS)
devices
121
logical input/output control system
DAM
124
DTF macros
113
file organization
133
function
111
imperative macros
119
ISFMS
129
macros and devices
121
macros and DTFS
120
MOD macros
114
modular/tabular
113
processing methods
111
sequential file
122
storage required
113
logical IOCS (see logical input/output
control system)
low core error bytes
203
LSERV (label cylinder display)
189
LUB (logical unit block) table
59,71
LUBDSP (LUB displacement) table
59,85
LUBID (LUB identification) table
59,85
machine check analysis and recording (MCAR)
189
machine check recording and recovery (MCRR)
189
macro definition
177
macro instruction
keyword
176

292

DOS System Programmer's Guide

macro instruction (CONT.)
mixed
177
positioning
177
macro writing
attributes
178
concatenation
181
conditional assembly
182
definition
177
extended capabilities
183
header statement
178,179
macro instruction
176
model statement
178,179
prototype statement
178,179
sample MSG coding
186
sample MSG macro
185
sequence symbols
182
sublist notation
179
trailer statement
178,179
variable symbols
179
macros
declarative
111
imperative
111
magnetic ink character re~ognition (see
MICR)
magnetic tape (DTFMT/MTMOD) files
122
main storage requir~ments (LIOCS)
113
master index (ISFMS)
129
MCAR (machine check analysis and recordin~)
189
MCAR linkage area
193
MCRR (machine check recording and recovery)
189
MCRR linkage table
192
MEXIT instruction
183
MICR (magnetic ink character recognition)
SYSGEN
36
table
37
MNOTE instruction
184
MOD (module generation) macros
114
model statement
178,179
modular/tabular system, definition of
113
module generation (MOD) macros
114
MPS, multiprogramming support (see
multiprogramming)
multiple link edits example
165
multiple wait facility (see WAITM)
multiprogramming (MPS)
concepts
18
control method
20
CPO concept
27
operational considerations
27
partitioning
20
programming considerations
28
storage map
19
system considerations
27
system generation guidelines
28
task selection
20
multitasking considerations for checkpoint
166,167
multitasking examples
45
ATTACH macro
47
DETACH macro
47
ENQ/DEQ and RCB
48
POST macro
51

multitasking examples (CONT.)
STXIT AB macro
53
track hold
55
WAITM macro
52
multitasking
abnormal termination
39,44
access to communication region
43
checkpoint consideration
43
definition of
38
examples
45
intertask communication
42
macro usage
39
maximum tasks
43
multiple wait
39
operator intervention
43
priorities
38,43
reentrant modules
39
register usage
43
resource contention
44
resource protection
41,43
storage protection
43
STXIT macro usage
43
subtask cancelation
44
subtask initiation
39
subtask termination
39
system generation guidelines
44
system logical units
43
track hold
38,43
wait considerations
44
NICL (number in class) table

OC (operator communications) support
concept
56
considerations
56
option table
57
OLTEP (On-Line Test Executive Program)
194
OLTEP
OLTS
194
RETAIN/370
194
OLTS (On-Line Tests)
194
On-Line Test Executive Program (OLTEP)
194
On-Line Tests (OLTS)
194
OPTION statement
141,147
option tables; IT,PC,OC,AB
57
organization
file
122-133
system residence
15
overflow area (ISFMS)
cylinder
132
independent
132
overflow record handling (DAM)
chaining
127
127
consecutive spill
127
preassigned tracks
overflow records (ISFMS)
130,132
overlay
147
communication
149
techniques
148
tree design
147

overlay (CONT.)
tree structure

147

packing factor (DAM)
127
paper tape (DTFPT/PTMOD) files
122
partition size specification
84
partition
definition of
18,20
types of
20
PC (program check) support
concept
58
option table
57
PCI (program controlled interrupt)
141
PCIL (private core image libraries)
76
PDAID (problem determination aids)
194
PHASE statement
141,144
phase, definition of
147
Physical I/O Control System (PIOCS)
CCB
100,102
CCW
106
EXCP
100
sample program
110
WAIT
100
physical IOCS sample program
110
physical transient overlap support
62
physical unit block (see PUB)
physical unit, definition of
99
PIB (program information block) table
195
PIOCS (see physical input/output control
system)
PL/I program debugging aids
255
PL/I program
compile time aborts
264
debugging aids
263,264
debugging summary
263
sample
265
POST macro
23
cancel conditions
43
example
51
use of
42
preassigned tracks overflow technique (DAM)
127
prime data area (ISFMS)
129
prime number _division
,126
prime number, definition of
126
Printer, IBM 3211
173
priority system
multiprogramming
20
multitasking
38
private core image libraries
link edit considerations
141,144
sysgen guidelines
81
system considerations
76
problem determination aids
DUMPGEN
189
EREP
189
LSERV
189
MCAR/CCH
189
MCRR
189
OLTEP
194
PDArD
194
RMS
189
processing methods
direct access method
112

Index

293

processing methods (CaNT.)
ISFMS
112
sequential
111
program check support
concept
58
option table
57
program design
139-186
program overlay structure
147
prototype statement, macro
178
PTa (physical transient overlap) support
62
PUB (physical unit block)
format of
59,89
use of
87
QTAM trace

194

random addressing
formula
125
techniques
124
random retrieval (ISFMS)
132
randomizing, definition of
126
RCB (resource control block)
41,43
RCB and ENQ/DEQ macros example
48
ROE (reliability data extractor)
189
RDONLY parameter
118
record definition
93
record types
fixed length, blocked
95
fixed length, unblocked
94
undefined
96
variable length, blocked
96
variable length, unblocked
96
recovery management support (RMS)
189
reentrant modules
39,118
relative track addressing
128
reliability data extractor (RDE)
189
relocatable module name prefixes
118
relocating address constants
151
repositioning I/O files
DASD
171
magnetic tape
170
REQID (requestor identification) table
85
rescheduling multiplexor channel
87
rescheduling selector channel
87
resource contention
44
resource control block (see RCB)
resource protection
41,43
restarting checkpointed programs
173
RETAIN/370
194
RMS (recovery management support)
189
RPG program debugging aids
243
halt analysis
243
RPG sample program
245
RSTRT statement
173
SAB (seek address block)
definition of
62
format of
59

294

DOS System Programmer's Guide

sample file organizations
136
sample programs
COBOL
207
FORTRAN
234
physical laCS
110
PL/I
265
RPG
245
self-relocating
153
STXIT
69
SDR communications region
190
seek address block (see SAB)
seek separate support
concept
62
system generation guidelines
62
self-relocating programs
advantages
150
disadvantages
150
programming techniques
150
rules for writing
149
sample program
153
self-relocating
example
165
sample program
153
sequence symbols
182
sequential DASD
123
sequential files
DASO
123
magnetic tape
122
unit record
122
111
sequential processing method
sequential retrieval (ISFMS)
132
SET symbols
defining of
180
global
180
local
180
versus symbolic parameters
180
single program initiator
28
SKSEP (see seek separate support)
SLI (suppress incorrect length indication)
(see CCW)
28
SPI (single program initiator)
STXIT AB macro example
53
STXIT macro
AB parameter
57,67
IT parameter
56,57
OC parameter
56,57
PC parameter
57,58
sample program
69
use of
43,67
STXIT sample program
69,157
sublist notation
179
submodular structure example
163
subtask
cancelation
44
definition of
38
initiation
39
priority
38,43
termination
39
supervisor calls (SVC), list of
21
supervisor calls
list
21
o (EXCP)
23
23
7 O"lAIT)
supervisor communications region
218

supervisor planning concepts
13-90
superv1sor tables
AB option
57
BG communications region
35
CBF
64
CHANQ
59
device type codes
90
DIB
75
ECB
41
FAVP
59
FICL
59
FLPTR
59
FOCL
59
IT option
57
JIB
59,72
job accounting
83
LUB
59,71
LUBDSP
59,85
LUBID
59,85
MCAR linkage area
193
MCRR linkage table
192
MICR
37
NICL
59
OC option
57
PC option
57
PIB
195
PUB
59,89
RCB
41
REQID
59,85
SAB
59
SDR communications
190
SVC list
21
system communications
30
TEB
59,60
TEBV
59,61
THFLPTR
59
TKHDTAB
59
TKREQID
59,85
SUPVR macro
MICR
36
multiprogramming
18
multitasking
38
SVC 0 (EXCP)
23
SVC 7 (WAIT)
23
symbolic parameter
179
synonyms
definition of
125
techniques for handling
127
SYSBUFLD
173
SYSFIL (system files on disk) support
concept of
74
considerations
75
sysgen
76
SYSRES (system residence) organization
15
system communications region
30
system files on disk support (see SYSFIL
support)
system residence organization
15
system variable symbols
181
tape error block (see TEB)
tape error recording
ESTV
58

tape error recording (CONT.)
EVA

TEB
58
task selection
20
TEB (tape error block)
concept of
58
format of
59,60
TEBV (to specify ESTV)
58
TEBV table
format of
59,61
THFLPTR ( track hold free list pointer)
59
TKHDTAB (track hold table) table
59
TKREQID (track requestor identification)
table
59,85
trace
F/L trace
194
GSVC trace
194
I/O trace
194
QTAM trace
194
track hold
129
track hold example
55
track hold facility (see TRKHLD)
track index (ISFMS)
definition of
129
normal entry
129
overflow entry
130
trailer labels, checkpoint
168,172
trailer statement
178,179
tree, definition of
147
TRKHLD (track hold) facility
LIOCS considerations
65,66
supervisor considerations
65
system generation
67
types of file organization
direct access method
124
ISFMS
129
sequential DASD
123
sequential tape
122
sequential unit record
122
undefined records
96
unit record files
card
122
console typewriter
line printers
122
paper tape
122

122

variable length records
ASCII
96
blocked
96
unblocked
96
variable symbols
SET
180
symbolic parameter
179
system variable
181
WAIT macro
multitasking considerations
opera tion
23
use of
70
wait states
hard waits
202'

Index

295

GC24-5073-2

wait states (CONT.)
low core error bytes
203
soft waits
202
WAITM (multiple wait) facility
concept of
70
considerations
70
multitasking
39
WAITM macro example
52
3211, IBM Printer

173

llrnoo
Internation! Business Machines COl')loratioll
Data Processing Division
tt!! Westchester Avenue, White Plains, New York 10894
[U.S.A. only]
IBM World Trade Corperatioll
821 United Nations Plaza, New Yurk, New Yark 10017
[International]

DOS System Programmer's Guide

READER'S
COMMENT
FORM

GC24-S073-2

This sheet is for comments and suggestions about this manual. We would appreciate your
views, favorable or unfavorable, in order to aid us in improving this publication. This form
will be sent directly to the author's department. Please include your name and address if
you wish a reply. Contact your IBM branch office for answers to technical questions about
the system or when requesting additional publications. Thank you.
Name

How did you use this manual?

Address

As a reference source

As a classroom text
As a self-study text
What is your occupation?

Your comments· and suggestions:

* We would especially appreciate your comments on any of the following topics:
Clarity of the text
Organization of the text

Accuracy
Cross-references

Index
Tables

Illustrations
Examples

Appearance
Printing

Paper
Binding

GC24-5073-2

YOUR COMMENTS, PLEASE .••

This manual is part of a library that serves as a reference source for systems analysts,
programmers and operators of IBM systems. Your answers to the questions on the back of this
form, together with your comments, will help us produce better publications for your use. Each
reply will be carefully reviewed by the persons responsible for writing and publishing this
material. All comments and suggestions become the property of IBM.

n
c

-I

»
r-

o
Z

-I

::t:

Please note: Requests for copies of publications and for assistance in utilizing your IBM system should ,be directed to your.IBM representative or to the IBM sales office serving your
locality.

Fold

.......................................................................................................................·:
··
FIRST CLASS

PERMIT NO. 1359
WHITE PLAINS. N. Y.

BUSINESS REPLY MAIL
NO POSTAGE STAMP NECESSARY IF MAILED IN THE UNITED STATES

POSTAGE WILL BE PAID BY .•.

I BM Corporation
112 East Post Road
White Plains, N.Y. 1060.1

iil

Attention: Department 813 U

3
3

.........................................................................................................................

Fold

International Business Machines Corporation
Data Processing Division
1133 Westchester Avenue, White Plains, New York 10604
[U.S.A. only]
IBM World Trade Corporation
821 United Nations Plaza, New York, New York 10017
[International]

Fold

.,
CD

Source Exif Data:

File Type                       : PDF
File Type Extension             : pdf
MIME Type                       : application/pdf
PDF Version                     : 1.3
Linearized                      : No
XMP Toolkit                     : Adobe XMP Core 4.2.1-c043 52.372728, 2009/01/18-15:56:37
Create Date                     : 2012:06:01 10:28:04-08:00
Modify Date                     : 2012:06:01 12:52:47-07:00
Metadata Date                   : 2012:06:01 12:52:47-07:00
Producer                        : Adobe Acrobat 9.51 Paper Capture Plug-in
Format                          : application/pdf
Document ID                     : uuid:3fee1d4e-997d-4ffb-b0e3-866fde654eb6
Instance ID                     : uuid:9dc065c7-d914-426e-bc10-3c70b2272417
Page Layout                     : SinglePage
Page Mode                       : UseNone
Page Count                      : 298

EXIF Metadata provided by EXIF.tools

GC24 5073 2_DOS_System_Programmers_Guide_Sep71 2 DOS System Programmers Guide Sep71

GC24-5073-2_DOS_System_Programmers_Guide_Sep71 GC24-5073-2_DOS_System_Programmers_Guide_Sep71

Navigation menu

Versions of this User Manual:

Views

Navigation