Trailing-Edge
-
PDP-10 Archives
-
BB-LW55A-BM_1988
-
7-documentation/link.doc
There are 32 other files named link.doc in the archive. Click here to see a list.
TOPS-20
LINK Reference Manual
AA-4183D-TM
This document describes LINK-20, the linking
loader for TOPS-20.
This document revises the document of the
same name, Order No. AA-4183C-TM, published
April 1982 and its update, Order No.
AD-413C-T1, published March 1983.
OPERATING SYSTEM AND VERSION: TOPS-20, Version 4.1, 5.1
SOFTWARE VERSION: LINK-20, Version 6.0
1
First Printing, January 1976
Revised, January 1978
April 1982
Updated, March 1983
Revised, May 1985
The information in this document is subject to change without notice
and should not be construed as a commitment by Digital Equipment
Corporation. Digital Equipment Corporation assumes no responsibility
for any errors that may appear in this document.
The software described in this document is furnished under a license
and may be used or copied only in accordance with the terms of such
license.
Digital Equipment Corporation assumes no responsibility for the use or
reliability of its software on equipment that is not supplied by
DIGITAL.
Copyright C 1976, 1985 by Digital Equipment Corporation
The postage prepaid READER'S COMMENTS form on the last page of this
document requests your critical evaluation to assist us in preparing
future documentation.
The following are trademarks of Digital Equipment Corporation.
DIGITAL TOPS-10 MASSBUS
DEC DECtape OMNIBUS
PDP DIBOL OS/8
DECUS EDUSYSTEM PHA
UNIBUS FLIP CHIP RSTS
COMPUTER LABS FOCAL RSX
COMTEX INDAC TYPESET-8
DDT LAB-8 TYPESET-10
DECCOMM TOPS-20 TYPESET-11
2
CONTENTS
CHAPTER 1 INTRODUCTION TO LINK
1.1 INPUT TO LINK . . . . . . . . . . . . . . . . . . 1-2
1.1.1 Object Modules . . . . . . . . . . . . . . . . . 1-2
1.1.2 Commands to LINK . . . . . . . . . . . . . . . . 1-3
1.1.3 Libraries And Searches . . . . . . . . . . . . . 1-3
1.2 OUTPUT FROM LINK . . . . . . . . . . . . . . . . . 1-4
1.2.1 Executable Program . . . . . . . . . . . . . . . 1-4
1.2.2 Messages . . . . . . . . . . . . . . . . . . . . 1-5
1.2.3 Special Files . . . . . . . . . . . . . . . . . 1-5
1.3 LINK'S OVERLAY FACILITY . . . . . . . . . . . . . 1-5
1.4 LINK AND EXTENDED ADDRESSING . . . . . . . . . . . 1-5
CHAPTER 2 USING LINK AUTOMATICALLY
2.1 COMMAND FORMATS . . . . . . . . . . . . . . . . . 2-1
2.2 COMMAND SWITCHES . . . . . . . . . . . . . . . . . 2-2
2.3 EXAMPLE OF USING LINK AUTOMATICALLY . . . . . . . 2-4
CHAPTER 3 USING LINK DIRECTLY
3.1 RUNNING AND EXITING LINK . . . . . . . . . . . . . 3-1
3.2 COMMAND FORMATS . . . . . . . . . . . . . . . . . 3-2
3.2.1 LINK File Specification Defaults . . . . . . . . 3-3
3.2.2 Logical Names . . . . . . . . . . . . . . . . . 3-4
3.2.3 Translating Directories . . . . . . . . . . . . 3-4
3.3 LINK SWITCHES . . . . . . . . . . . . . . . . . . 3-4
3.3.1 Switch Abbreviation . . . . . . . . . . . . . . 3-4
3.3.2 Switch Arguments . . . . . . . . . . . . . . . . 3-5
3.3.3 Switch Placement . . . . . . . . . . . . . . . . 3-6
3.3.4 Alphabetical Listing of LINK Switches . . . . . 3-8
/ARSIZE . . . . . . . . . . . . . . . . . . . . 3-9
/COMMON . . . . . . . . . . . . . . . . . . . 3-10
/CONTENTS . . . . . . . . . . . . . . . . . . 3-11
/COUNTERS . . . . . . . . . . . . . . . . . . 3-13
/CPU . . . . . . . . . . . . . . . . . . . . . 3-15
/DDEBUG . . . . . . . . . . . . . . . . . . . 3-16
/DEBUG . . . . . . . . . . . . . . . . . . . . 3-17
/DEFAULT . . . . . . . . . . . . . . . . . . . 3-19
/DEFINE . . . . . . . . . . . . . . . . . . . 3-20
/ENTRY . . . . . . . . . . . . . . . . . . . . 3-21
/ERRORLEVEL . . . . . . . . . . . . . . . . . 3-22
/EXCLUDE . . . . . . . . . . . . . . . . . . . 3-23
/EXECUTE . . . . . . . . . . . . . . . . . . . 3-24
/FRECOR . . . . . . . . . . . . . . . . . . . 3-25
/GO . . . . . . . . . . . . . . . . . . . . . 3-26
3
/EXIT . . . . . . . . . . . . . . . . . . . . 3-27
/HASHSIZE . . . . . . . . . . . . . . . . . . 3-28
/HELP . . . . . . . . . . . . . . . . . . . . 3-29
/INCLUDE . . . . . . . . . . . . . . . . . . . 3-30
/LIMIT . . . . . . . . . . . . . . . . . . . . 3-32
/LINK . . . . . . . . . . . . . . . . . . . . 3-35
/LOCALS . . . . . . . . . . . . . . . . . . . 3-36
/LOG . . . . . . . . . . . . . . . . . . . . . 3-37
/LOGLEVEL . . . . . . . . . . . . . . . . . . 3-38
/MAP . . . . . . . . . . . . . . . . . . . . . 3-39
/MAXNODE . . . . . . . . . . . . . . . . . . . 3-40
/MESSAGE . . . . . . . . . . . . . . . . . . . 3-41
/MISSING . . . . . . . . . . . . . . . . . . . 3-43
/NEWPAGE . . . . . . . . . . . . . . . . . . . 3-44
/NODE . . . . . . . . . . . . . . . . . . . . 3-45
/NOENTRY . . . . . . . . . . . . . . . . . . . 3-47
/NOINCLUDE . . . . . . . . . . . . . . . . . . 3-48
/NOINITIAL . . . . . . . . . . . . . . . . . . 3-49
/NOJOBDAT . . . . . . . . . . . . . . . . . . 3-50
/NOLOCAL . . . . . . . . . . . . . . . . . . . 3-51
/NOREQUEST . . . . . . . . . . . . . . . . . . 3-52
/NOSEARCH . . . . . . . . . . . . . . . . . . 3-53
/NOSTART . . . . . . . . . . . . . . . . . . . 3-54
/NOSYMBOL . . . . . . . . . . . . . . . . . . 3-55
/NOSYSLIB . . . . . . . . . . . . . . . . . . 3-56
/NOUSERLIB . . . . . . . . . . . . . . . . . . 3-57
/ONLY . . . . . . . . . . . . . . . . . . . . 3-58
/OPTION . . . . . . . . . . . . . . . . . . . 3-59
/OTSEGMENT . . . . . . . . . . . . . . . . . . 3-60
/OVERLAY . . . . . . . . . . . . . . . . . . . 3-62
/PATCHSIZE . . . . . . . . . . . . . . . . . . 3-64
/PLOT . . . . . . . . . . . . . . . . . . . . 3-65
/PLTTYP . . . . . . . . . . . . . . . . . . . 3-67
/PSCOMMON . . . . . . . . . . . . . . . . . . 3-68
/PVBLOCK . . . . . . . . . . . . . . . . . . . 3-69
/PVDATA . . . . . . . . . . . . . . . . . . . 3-71
/REDIRECT . . . . . . . . . . . . . . . . . . 3-74
/REQUEST . . . . . . . . . . . . . . . . . . . 3-75
/REQUIRE . . . . . . . . . . . . . . . . . . . 3-76
/RUN . . . . . . . . . . . . . . . . . . . . . 3-77
/RUNAME . . . . . . . . . . . . . . . . . . . 3-78
/RUNOFFSET . . . . . . . . . . . . . . . . . . 3-79
/SAVE . . . . . . . . . . . . . . . . . . . . 3-80
/SEARCH . . . . . . . . . . . . . . . . . . . 3-81
/SEGMENT . . . . . . . . . . . . . . . . . . . 3-83
/SET . . . . . . . . . . . . . . . . . . . . . 3-85
/SEVERITY . . . . . . . . . . . . . . . . . . 3-86
/SPACE . . . . . . . . . . . . . . . . . . . . 3-87
/START . . . . . . . . . . . . . . . . . . . 3-88
/SUPPRESS . . . . . . . . . . . . . . . . . . 3-89
/SYFILE . . . . . . . . . . . . . . . . . . . 3-91
/SYMSEG . . . . . . . . . . . . . . . . . . . 3-92
4
/SYSLIB . . . . . . . . . . . . . . . . . . . 3-93
/TEST . . . . . . . . . . . . . . . . . . . . 3-94
/UNDEFINED . . . . . . . . . . . . . . . . . . 3-96
/UPTO . . . . . . . . . . . . . . . . . . . . 3-97
/USERLIB . . . . . . . . . . . . . . . . . . . 3-98
/VALUE . . . . . . . . . . . . . . . . . . . . 3-99
/VERBOSITY . . . . . . . . . . . . . . . . . . 3-100
/VERSION . . . . . . . . . . . . . . . . . . . 3-102
3.5 EXAMPLES USING LINK DIRECTLY . . . . . . . . . . 3-103
CHAPTER 4 OUTPUT FROM LINK
4.1 THE EXECUTABLE PROGRAM . . . . . . . . . . . . . . 4-1
4.2 OUTPUT FILES . . . . . . . . . . . . . . . . . . . 4-2
4.2.1 Sharable Save Files . . . . . . . . . . . . . . 4-2
4.2.1.1 Format of Sharable Save Files . . . . . . . . 4-2
4.2.2 LOG Files . . . . . . . . . . . . . . . . . . . 4-6
4.2.3 Map files . . . . . . . . . . . . . . . . . . . 4-7
4.2.4 Symbol Files . . . . . . . . . . . . . . . . . . 4-7
4.3 MESSAGES . . . . . . . . . . . . . . . . . . . . . 4-7
CHAPTER 5 OVERLAYS
5.1 OVERLAY STRUCTURES . . . . . . . . . . . . . . . . 5-1
5.1.1 Defining Overlay Structures . . . . . . . . . . 5-2
5.1.2 An Overlay Example . . . . . . . . . . . . . . . 5-4
5.1.2.1 Source Files . . . . . . . . . . . . . . . . . 5-5
5.1.2.2 Source File Compilation . . . . . . . . . . . 5-9
5.1.2.3 Interactive use of LINK . . . . . . . . . . . 5-9
5.1.2.4 TEST.LOG . . . . . . . . . . . . . . . . . . 5-11
5.1.2.5 TEST.MAP . . . . . . . . . . . . . . . . . . 5-12
5.1.2.6 Tree Diagram . . . . . . . . . . . . . . . . 5-20
5.1.2.7 Executable File . . . . . . . . . . . . . . 5-21
5.2 WRITABLE OVERLAYS . . . . . . . . . . . . . . . 5-22
5.2.1 Writable Overlay Syntax . . . . . . . . . . . 5-22
5.2.2 Writable Overlay Error Messages . . . . . . . 5-22
5.3 RELOCATABLE OVERLAYS . . . . . . . . . . . . . . 5-22
5.3.1 Relocatable Overlay Syntax . . . . . . . . . . 5-23
5.3.2 Relocatable Overlay Messages . . . . . . . . . 5-23
5.4 RESTRICTIONS ON OVERLAYS . . . . . . . . . . . . 5-23
5.4.1 Restrictions on Absolute Overlays . . . . . . 5-24
5.4.2 Restrictions on Relocatable Overlays . . . . . 5-25
5.4.3 Restrictions on FORTRAN Overlays . . . . . . . 5-25
5.5 SIZE OF OVERLAY PROGRAMS . . . . . . . . . . . . 5-26
5.6 DEBUGGING OVERLAID PROGRAMS . . . . . . . . . . 5-26
5.7 THE OVERLAY HANDLER . . . . . . . . . . . . . . 5-26
5.7.1 Calls to the Overlay Handler . . . . . . . . . 5-27
5.7.2 Overlay Handler Subroutines . . . . . . . . . 5-28
5.7.3 Overlay Handler Messages . . . . . . . . . . . 5-35
5.7.4 The FUNCT. Subroutine . . . . . . . . . . . . 5-40
5
5.8 THE OVERLAY (OVL) FILE . . . . . . . . . . . . . 5-47
5.8.1 The Directory Block . . . . . . . . . . . . . 5-48
5.8.2 The Link Number Table . . . . . . . . . . . . 5-49
5.8.3 The Link Name Table . . . . . . . . . . . . . 5-49
5.8.4 The Overlay Link . . . . . . . . . . . . . . . 5-50
CHAPTER 6 PSECTs
6.1 LOADING PROGRAMS WITH PSECTs . . . . . . . . . . . 6-1
6.2 PSECT ATTRIBUTES . . . . . . . . . . . . . . . . . 6-3
6.2.1 CONCATENATED and OVERLAID . . . . . . . . . . . 6-4
6.2.2 RONLY and RWRITE . . . . . . . . . . . . . . . . 6-5
CHAPTER 7 PDVs
7.1 PDV FORMAT . . . . . . . . . . . . . . . . . . . . 7-3
7.2 THE PDV STATIC MEMORY MAP . . . . . . . . . . . . 7-5
7.3 SYMBOL TABLE VECTOR . . . . . . . . . . . . . . . 7-7
APPENDIX A REL BLOCKS
Block Type 0 (Ignored) . . . . . . . . . . . . . . A-4
Block Type 1 (Code) . . . . . . . . . . . . . . . A-5
Block Type 2 (Symbols) . . . . . . . . . . . . . . A-6
Block Type 3 (HISEG) . . . . . . . . . . . . . . A-11
Block Type 4 (Entry) . . . . . . . . . . . . . . A-12
Block Type 5 (End) . . . . . . . . . . . . . . . A-13
Block Type 6 (Name) . . . . . . . . . . . . . . A-14
Block Type 7 (Start) . . . . . . . . . . . . . . A-15
Block Type 10 (Internal Request) . . . . . . . . A-16
Block Type 11 (Polish) . . . . . . . . . . . . . A-18
Block Type 12 (Chain) . . . . . . . . . . . . . A-24
Block Type 14 (Index) . . . . . . . . . . . . . A-31
Block Type 15 (ALGOL) . . . . . . . . . . . . . A-33
Block Type 16 (Request Load) . . . . . . . . . . A-34
Block Type 17 (Request Library) . . . . . . . . A-35
Block Type 20 (Common) . . . . . . . . . . . . . A-36
Block Type 21 (Sparse Data) . . . . . . . . . . A-37
Block Type 22 (PSECT Origin) . . . . . . . . . . A-38
Block Type 23 (PSECT End Block) . . . . . . . . A-39
Block Type 24 (PSECT Header Block) . . . . . . . A-40
Block Type 37 (COBOL Symbols) . . . . . . . . . A-41
Block Type 100 (.ASSIGN) . . . . . . . . . . . . A-42
Block Type 776 (Symbol File) . . . . . . . . . . A-43
Block Type 777 (Universal File) . . . . . . . . A-44
Block Type 1000 (Ignored) . . . . . . . . . . . A-45
Block Type 1001 (Entry) . . . . . . . . . . . . A-46
Block Type 1002 (Long Entry) . . . . . . . . . . A-47
Block Type 1003 (Long Title) . . . . . . . . . . A-48
6
Block Type 1004 (Byte Initialization) . . . . . A-51
Block Types 1010 - 1037 (Code Blocks) . . . . . A-52
Blocks 1010 - 1017 (Right Relocation) . . . . . A-53
Block Types 1020-1027 (Left/Right Relocation
Blocks) . . . . . . . . . . . . . . . . . . . . A-55
Block Types 1030 - 1037 (Thirty-bit Relocation
Blocks) . . . . . . . . . . . . . . . . . . . . A-57
Block Type 1042 (Request Load for SFDs) . . . . A-59
Block Type 1043 (Request Library for SFDs) . . . A-60
Block Type 1044 (ALGOL Symbols) . . . . . . . . A-61
Block Type 1045 (Writable Links) . . . . . . . . A-62
Block Type 1050 (Long PSECT Name Block) . . . . A-64
Block Type 1051 (Set Current PSECT) . . . . . . A-66
Block Type 1052 (PSECT End) . . . . . . . . . . A-67
Block Type 1060 (Trace Block Data) . . . . . . . A-68
Block Type 1070 (Long Symbol Names) . . . . . . A-69
Block Type 1072 (Long Polish Block) . . . . . . A-74
Block Type 1074 (Long Common Name) . . . . . . . A-78
Block types 1120-1127 (Argument Descriptor
Blocks) . . . . . . . . . . . . . . . . . . . . A-79
Block Type 1130 (Coercion Block) . . . . . . . . A-84
Block Type 1131 (TWOSEG Redirection Block) . . . A-87
Block Type 1140 (PL/1 debugger information) . . A-88
Block Type 1160 (Extended Sparse Data
Initialization Block) . . . . . . . . . . . . . A-89
Block Type Greater Than 3777 (ASCIZ) . . . . . . A-92
APPENDIX B LINK MESSAGES
B.1 DESCRIPTION OF MESSAGES . . . . . . . . . . . . . B-1
B.1.1 Message Levels . . . . . . . . . . . . . . . . B-1
B.1.2 Message Severity . . . . . . . . . . . . . . . . B-2
B.1.3 Message Length . . . . . . . . . . . . . . . . . B-2
B.1.4 Message Conventions . . . . . . . . . . . . . . B-3
B.2 LIST OF MESSAGES . . . . . . . . . . . . . . . . . B-4
B.3 REASON EXPLANATION . . . . . . . . . . . . . . . B-35
APPENDIX C JOB DATA AREA LOCATIONS SET BY LINK
C.1 JOB DATA AREA . . . . . . . . . . . . . . . . . . C-1
C.1.1 Vestigial Job Data Area . . . . . . . . . . . . C-2
GLOSSARY
INDEX
7
FIGURES
5-1 Example of an Overlay Structure . . . . . . . . . 5-2
TABLES
1-1 Summary of LINK Switches . . . . . . . . . . . . . 1-6
2-1 Switches for TOPS-20 Commands . . . . . . . . . . 2-2
5-1 Summary of LINK's Overlay-Related Switches . . . . 5-2
B-1 Severity Codes . . . . . . . . . . . . . . . . . . B-2
B-2 Special Message Segments . . . . . . . . . . . . . B-4
8
PREFACE
This manual is the reference document for LINK, the TOPS-20 linking
loader.
Document Structure
Chapter 1 provides a general introduction to LINK, a discussion of
libraries, library searches, and extended addressing. Chapter 1 also
contains a summary of LINK switches.
Chapter 2 describes automatic use of LINK through one of the TOPS-20
DEBUG, EXECUTE, or LOAD commands. This chapter is sufficient for most
loading tasks.
Chapter 3 describes running and exiting LINK, LINK's command format,
and LINK switches. This chapter contains the alphabetical listing of
LINK switches and an example of using LINK directly.
Chapter 4 describes output from LINK: executable programs, most
output files, and LINK messages. Included are descriptions of the
internal format of save (EXE) files.
Chapter 5 discusses overlays, including overlay structures,
overlay-related output files, the overlay handler and its messages,
and the FUNCT. subroutine. This chapter has an extensive example of
an overlay load. Many of the elements of this example are of interest
outside the context of overlays.
Chapter 6 discusses PSECTs (Program SECTions), loading programs with
PSECTS, and PSECT attributes.
Chapter 7 discusses PDVs (Program Data Vectors), their format, memory
map, and symbol table vector.
Appendix A gives a technical description of REL (RELocatable) Blocks,
the main input to LINK.
Appendix B lists all LINK messages except the OVERLAY handler runtime
messages that are contained in Chapter 5.
Appendix C describes JOBDAT, the Job Data Area.
A Glossary follows the appendixes.
9
Reference Material
TOPS-20 User's Guide
TOPS-20 Commands Reference Manual
MACRO Assembler Reference Manual
FORTRAN Language Manual
COBOL-68 Language Manual
COBOL-74 Language Manual
DECsystem-10/DECSYSTEM-20 Processor Reference Manual
TOPS-20 Monitor Calls Reference Manual
Conventions
<ESC> indicates that you are to enter an ESCAPE key.
<RET> indicates that you are to enter a carriage-return.
bold is used in reference section of Chapter 3 to show
italic defaults and Digital supported software.
Loading Map
To load most programs,
1. Become familiar with the concepts in the TOPS-20 User's Guide
and TOPS-20 Commands Reference Manual.
2. Refer to Chapter 2 for a review of the DEBUG, EXECUTE, and
LOAD commands that automatically use LINK to load programs.
To load programs that use overlays,
1. Refer to Chapter 5.
2. Refer to the /ARSIZE, /LINK, /MAXNODE, /NODE, /OVERLAY,
/PLOT, and /SPACE switches in Chapter 3.
To load programs that use PSECTs,
1. Refer to Chapter 6.
2. Refer to the /SET and /LIMIT switches in Chapter 3.
To load extended addressing programs,
10
1. Become familiar with the concepts in
the DECsystem-10/DECSYSTEM-20 Processor Reference Manual
the Chapter 8 of the TOPS-20 Monitor Calls User's Guide
TOPS-20 Monitor Calls Reference Manual
2. Refer to Section 3.7.
3. Refer to Section 6.2 on loading PSECTs.
11
1-1
CHAPTER 1
INTRODUCTION TO LINK
LINK is TOPS-20's linking loader. It merges independently compiled or
assembled modules into a single executable program.
This merging process requires LINK to:
1. Convert relocatable addresses to absolute addresses, and bind
program segments and PSECTs (Program SECTions) to addresses.
Relocatable addresses are addresses within a module that are
specified as an offset from the first location in that
module.
Absolute addresses are fixed locations in your address space.
Program segments define the program as a single-segment or
two segment program.
PSECTs are programmer- or system-defined regions.
2. Resolve global symbol references using chained fixups, Polish
fixups and library searches.
Global symbols are defined in one module and used in other
modules. LINK resolves the references to global symbols by
storing the reference until LINK loads the module that
contains the global symbol's definition. After LINK loads
that module, it "fixes up" the location in memory where LINK
stored the reference.
Chained fixups are a list of locations that require the
global symbols' definition. One location points to the next
location that requires the global symbol's definition.
Polish fixups use an algorithm to find the locations that
require the global symbol's definition.
1-1
FIELD TEST
3. Produce an executable program, a JOBDAT area or a PDV
(Program Data Vector) and a debugger symbol table.
An executable program is the memory image of the program.
A JOBDAT or PDV contain information about the program such as
its version number.
A debugger symbol table contains entries and values for the
symbols defined or used in a program.
1.1 INPUT TO LINK
LINK accepts three kinds of input:
o Object modules that are the main input to LINK, and are
output by a language translator.
o LINK commands
o Libraries that contain object modules for specific languages
and applications.
1.1.1 Object Modules
Object modules contain machine language that corresponds to the source
program, and are contained in relocatable binary files. A relocatable
binary file can contain many modules or one module, and is formatted
into REL (RELocatable) Blocks. LINK recognizes REL Blocks and handles
them appropriately. The format of each REL Block Type is described in
Appendix A. The default file type for a relocatable binary file is
REL, and throughout this manual relocatable binary files are referred
to as REL files.
As their name implies, REL files contain relocatable code or
addresses. LINK converts relocatable addresses to absolute addresses
by loading relocatable code at an arbitrary memory address, and adding
a constant to each address referenced in the program.
In the process of converting addresses, LINK binds program segments,
and PSECTS to addresses.
A program that uses program segments can be a single-segment or
two-segment program. A single segment program uses one relocation
counter for LINK to use in the loading process. A two-segment program
uses two relocation counters, and is divided into a high-segment or
low-segment. The LINK /ONLY, /REDIRECT, /SET, and /SEGMENT switches
1-2
FIELD TEST
can be used to manipulate segments during loading. A relocation
counter is an address counter that LINK uses while loading relocatable
code.
A program that uses PSECTs can have one or multiple PSECTs. Each
PSECT in a program has its own relocation counter. For example, if a
program has 5 PSECTs, it also has 5 relocation counters, one for each
PSECT. Refer to Chapter 6 for information on PSECTs.
Besides relocating and loading your object modules, LINK resolves
values for local, global and entry name symbols. Local symbols are
those that are defined and referenced only within a module. Global
symbols are those that are defined in one module and used in others.
Entry name symbols are special global symbols in a module that contain
an entry point name for other modules.
1.1.2 Commands to LINK
LINK commands control and modify the loading process. Commands
consist of file specifications and switches. LINK commands are
summarized at the end of this chapter and are detailed in Chapter 3.
1.1.3 Libraries And Searches
Libraries are files that contain object modules that may be needed to
resolve global symbol references. LINK only loads the object modules
that contain entries for the global symbols to be resolved. There are
two kinds of libraries: system or user.
System libraries are available to all users for searching. Most
language translators also have libraries associated with them.
Translators generate calls for subroutines or functions in their
corresponding libraries. Library searches find and load the necessary
modules.
LINK searches system libraries before it finishes loading a program.
LINK determines which system library to load from the object module
that contains translator information for the program. LINK performs
system library searches when it finds a LINK /GO or /LINK switch. For
example, if you load a FORTRAN-compiled module, LINK searches the
system FORTRAN library SYS:FORLIB.REL when a /GO or /LINK switch is
processed. This search resolves requests for FORTRAN-defined
subroutines and functions in FORLIB.REL.
The /SYSLIB and /NOSYSLIB switches allow you to manipulate system
library searches.
The /SYSLIB switch requires LINK to search specified system libraries
1-3
FIELD TEST
no matter what kind of modules were loaded. The /NOSYSLIB switch
prevents searching of specified system libraries. Using these two
switches, you can select the order for searching system libraries.
User libraries are libraries that you create for LINK to search.
The /USERLIB and /NOUSERLIB switches allow you to specify user library
searches. /USERLIB specifies that a user library must be searched
before the corresponding system library. For example, using the
switch MYFORL/USERLIB:FORTRAN requires LINK to search MYFORL.REL
before searching FORLIB. The /NOUSERLIB switch suspends the effect of
a /USERLIB switch.
The /SEARCH and /NOSEARCH switches allow you to specify how an input
file is to be loaded. /SEARCH specifies that LINK search input files
as libraries, and load only those modules that resolve a global
symbol. /NOSEARCH specifies that LINK is to load all modules from
each input file. /NOSEARCH is the default.
Using combinations of these switches gives you precise control of
library searches. See Chapter 3 for information on these switches.
1.2 OUTPUT FROM LINK
LINK outputs an executable program in memory, messages, and special
files.
1.2.1 Executable Program
The main output from LINK is an executable program. In this
executable program, all addresses are resolved to absolute memory
locations, and all symbols (including subroutine calls) are resolved
to absolute values or addresses.
The executable program can be executed immediately or saved as a
sharable save file. A sharable save file has a default file type of
EXE, and is referred to to as an EXE file. LINK automatically creates
an EXE file when you use /SAVE with LINK.
You can also execute the executable program under the control of a
debugger using the /DEBUG or /TEST switch. LINK outputs a debugger
symbol table for this purpose. A debugger symbol table contains the
program's symbols and their definitions.
LINK also generates a JOBDAT or a PDV for a program. Both JOBDAT and
PDV contain information about the program such as debugger symbol
table pointers, version numbers, and memory use. Refer to Appendix C
for JOBDAT information and Chapter 7 for PDV information.
1-4
FIELD TEST
1.2.2 Messages
During its processing, LINK generates messages that are output to your
terminal or a log file. Some of these give information about LINK's
operation. Some warn about possible problems. Some identify errors.
LINK messages are described in Appendix B.
1.2.3 Special Files
At your option, LINK can generate several special files:
o a map file that contains information about where the program
was loaded, and what symbols are used in the program modules.
o a log file that contains a record of the messages that LINK
returned during the linking and loading process.
o a symbol file that contains a symbol table for the program
that was loaded.
o a plotter file that contains a tree diagram of an overlay
program.
o an overlay file that contains information about the overlay
links.
LINK's output files are described in Chapter 4.
1.3 LINK'S OVERLAY FACILITY
If a program is too large to execute in one piece, you can use LINK's
overlay facility to define an overlay structure for the program. To
do this, you define a tree structure for the program's modules. Then
at execution time, only part of the tree is in memory at one time.
This reduces the amount of memory needed for execution. See Chapter 5
for a discussion on overlays.
1.4 LINK AND EXTENDED ADDRESSING
The KL Model B processor is capable of using an address space
consisting of 32 sections, each containing 512 pages. As of TOPS-20
Version 5, programs have been able to reference this expanded address
space. For information on using extended addressing with a
programming language, consult the documentation for that language.
Use PSECTs to load a program into a nonzero section. See Section 6.1
1-5
FIELD TEST
for information on loading PSECTs.
When loading a program that uses extended addressing, pay particular
attention to the use of 18-bit and 30-bit addresses. If a program
uses 30-bit addresses and you reference the 30-bit addresses as a
18-bit address, LINK truncates the 30-bit address and notifies you
with the following message:
%LNKFTH Fullword value [symbol] truncated to halfword
Refer to Appendix B for more information about this message.
LINK issues this warning if the truncation results in the loss of a
section number.
While writing an extended addressing program, keep the following
restrictions in mind:
o Programs that use overlays cannot use nonzero sections.
o LINK does not set up JOBDAT for a program loaded in any
nonzero section. LINK stores information about an extended
addressing program in a PDV. PDVs are described in Chapter
7. JOBDAT is described in Appendix C.
o Programs should not store executable code into locations 0
through 17 of nonzero sections. However, you can store data
that is not executed in these locations. If data is stored
in locations 0 through 17, use global addresses to reference
the locations. If local addresses are used, the locations
are referenced as ACs (accumulators). In section one,
locations 0 through 17 are ACs.
Table 1-1: Summary of LINK Switches
+--------------------------------------------------------------------+
| |
| Switch Description |
+--------------------------------------------------------------------+
| |
| /ARSIZE Sets the size of the overlay handler's table of |
| multiply-defined global symbols. |
+--------------------------------------------------------------------+
| |
| /COMMON Allocates words of labeled COMMON storage for |
| FORTRAN and FORTRAN-compatible programs. |
+--------------------------------------------------------------------+
1-6
FIELD TEST
Table 1-1 (cont.)
+--------------------------------------------------------------------+
| |
| Switch Description |
+--------------------------------------------------------------------+
| |
| /CONTENTS Specifies the symbol types to be included in the |
| map file if the file is generated. |
+--------------------------------------------------------------------+
| |
| /COUNTERS Displays relocation counters on the terminal. |
+--------------------------------------------------------------------+
| |
| /DDEBUG Specifies a default debugger to be loaded if the |
| /DEBUG or /TEST switch appears without an argument. |
+--------------------------------------------------------------------+
| |
| /DEBUG Requests loading of a debugger and sets the start |
| address for execution at the start address of the |
| debugger. |
+--------------------------------------------------------------------+
| |
| /DEFAULT Changes default specifications for input or output |
| files. |
+--------------------------------------------------------------------+
| |
| /DEFINE Assigns a decimal value to a symbol. |
+--------------------------------------------------------------------+
| |
| /ENTRY Displays all entry name symbols that have been |
| loaded on the terminal. |
+--------------------------------------------------------------------+
| |
| /ERRORLEVEL Suppresses terminal display of LINK messages. |
+--------------------------------------------------------------------+
| |
| /EXCLUDE Prevents loading of the specified modules from the |
| current file. |
+--------------------------------------------------------------------+
| |
| /EXECUTE Directs LINK to execute the loaded program |
| beginning at the program's start address. |
+--------------------------------------------------------------------+
| |
| /EXIT Exits LINK. |
+--------------------------------------------------------------------+
| |
| /FRECOR Requires LINK to maintain a minimum amount of free |
| memory after any expansions. |
+--------------------------------------------------------------------+
1-7
FIELD TEST
Table 1-1 (cont.)
+--------------------------------------------------------------------+
| |
| Switch Description |
+--------------------------------------------------------------------+
| |
| /GO Ends loading after the current file and exits LINK. |
+--------------------------------------------------------------------+
| |
| /HASHSIZE Gives a minimum for the initial size of the global |
| symbol table. |
+--------------------------------------------------------------------+
| |
| /HELP Displays information about LINK on the terminal. |
+--------------------------------------------------------------------+
| |
| /INCLUDE Specifies modules to be loaded regardless of any |
| global requests for them. |
+--------------------------------------------------------------------+
| |
| /LIMIT Specifies an upper bound for a PSECT. |
+--------------------------------------------------------------------+
| |
| /LINK Closes an overlay link. |
+--------------------------------------------------------------------+
| |
| /LOCALS Includes local symbols from a module in the symbol |
| table. |
+--------------------------------------------------------------------+
| |
| /LOG Specifies a file specification for the log file. |
+--------------------------------------------------------------------+
| |
| /LOGLEVEL Suppresses logging of LINK messages. |
+--------------------------------------------------------------------+
| |
| /MAP Specifies a file specification for the map output |
| file. |
+--------------------------------------------------------------------+
| |
| /MAXNODE Specifies the number of links to be defined when |
| the overlayed program requires more than 256 links. |
+--------------------------------------------------------------------+
| |
| /MESSAGE Displays messages on the terminal in the format |
| specified by keyword. |
+--------------------------------------------------------------------+
1-8
FIELD TEST
Table 1-1 (cont.)
+--------------------------------------------------------------------+
| |
| Switch Description |
+--------------------------------------------------------------------+
| |
| /MISSING Displays on the terminal the number of modules |
| requested with the /INCLUDE switch that have not |
| yet been loaded. |
+--------------------------------------------------------------------+
| |
| /NEWPAGE Sets the relocation counter to the first word of |
| the next page. |
+--------------------------------------------------------------------+
| |
| /NODE Opens an overlay link. |
+--------------------------------------------------------------------+
| |
| /NOENTRY Deletes entry name symbols from LINK's overhead |
| tables when loading overlays. |
+--------------------------------------------------------------------+
| |
| /NOINCLUDE Clears requests for modules that were specified in |
| a previous /INCLUDE. |
+--------------------------------------------------------------------+
| |
| /NOINITIAL Prevents loading of LINK's initial global symbol |
| table (JOBDAT). |
+--------------------------------------------------------------------+
| |
| /NOJOBDAT Prevents LINK from filling in JOBDAT's or vestigial |
| JOBDAT's address space, and creates a PDV. |
+--------------------------------------------------------------------+
| |
| /NOLOCAL Suspends the effect of a preceding /LOCALS switch |
| so that local symbol tables are not loaded with |
| their modules. |
+--------------------------------------------------------------------+
| |
| /NOREQUEST Deletes references to overlay links from LINK's |
| overhead tables when loading overlay programs. |
+--------------------------------------------------------------------+
| |
| /NOSEARCH Suspends the effect of a previous /SEARCH switch. |
+--------------------------------------------------------------------+
| |
| /NOSTART Directs LINK to disregard any start addresses found |
| after the /NOSTART switch. |
+--------------------------------------------------------------------+
1-9
FIELD TEST
Table 1-1 (cont.)
+--------------------------------------------------------------------+
| |
| Switch Description |
+--------------------------------------------------------------------+
| |
| /NOSYMBOL Prevents construction of user symbol tables. |
+--------------------------------------------------------------------+
| |
| /NOSYSLIB Prevents automatic searching of system libraries. |
+--------------------------------------------------------------------+
| |
| /NOUSERLIB Discontinues automatic searching of files at each |
| /LINK or /GO switch. |
+--------------------------------------------------------------------+
| |
| /ONLY Directs LINK to load the specified segment of |
| two-segment modules. |
+--------------------------------------------------------------------+
| |
| /OPTION Reads the SWITCH.INI file to determine switch |
| defaults for LINK. |
+--------------------------------------------------------------------+
| |
| /OTSEGMENT Specifies the time and manner of loading the |
| object-time system. |
+--------------------------------------------------------------------+
| |
| /OVERLAY Initiates construction of an overlay structure. |
+--------------------------------------------------------------------+
| |
| /PATCHSIZE Allocates words of storage to precede the symbol |
| table. |
+--------------------------------------------------------------------+
| |
| /PLOT Directs LINK to output a tree diagram of an overlay |
| structure. |
+--------------------------------------------------------------------+
| |
| /PLTTYP Specifies the type of plot file to be generated by |
| the /PLOT switch. |
+--------------------------------------------------------------------+
| |
| /PSCOMMON Specifies where LINK is to load COMMON blocks. |
+--------------------------------------------------------------------+
| |
| /PVBLOCK Requests a PDV (Program Data Vector) from LINK and |
| gives you control over where the vector goes. |
+--------------------------------------------------------------------+
1-10
FIELD TEST
Table 1-1 (cont.)
+--------------------------------------------------------------------+
| |
| Switch Description |
+--------------------------------------------------------------------+
| |
| /PVDATA Changes the contents of a PDV block. |
+--------------------------------------------------------------------+
| |
| /REDIRECT Loads two-segment formatted REL files as part of a |
| program using PSECTs. |
+--------------------------------------------------------------------+
| |
| /REQUEST Displays external overlay link references on the |
| terminal. |
+--------------------------------------------------------------------+
| |
| /REQUIRE Generates global requests for the specified |
| symbols. |
+--------------------------------------------------------------------+
| |
| /RUN Directs LINK to run a program after it is loaded. |
+--------------------------------------------------------------------+
| |
| /RUNAME Assigns a job name for execution of your program. |
+--------------------------------------------------------------------+
| |
| /RUNOFFSET Runs the program specified in a /RUN switch with an |
| offset. |
+--------------------------------------------------------------------+
| |
| /SAVE Directs LINK to create an EXE file. |
+--------------------------------------------------------------------+
| |
| /SEARCH Directs LINK to search the input files and load |
| only the modules whose entry point names resolves a |
| global symbol reference. |
+--------------------------------------------------------------------+
| |
| /SEGMENT Specifies whether the high segment or the low |
| segment of a two-segment program is to be used for |
| loading the following modules. |
+--------------------------------------------------------------------+
| |
| /SET Sets the origin of a PSECT, or sets the .HIGH. or |
| .LOW. relocation counter. |
+--------------------------------------------------------------------+
| |
| /SEVERITY Specifies that messages of a certain severity level |
| terminate the load. |
+--------------------------------------------------------------------+
1-11
FIELD TEST
Table 1-1 (cont.)
+--------------------------------------------------------------------+
| |
| Switch Description |
+--------------------------------------------------------------------+
| |
| /SPACE Specifies that n words of memory follow the current |
| link at execution time. |
+--------------------------------------------------------------------+
| |
| /START Specifies the start address for the loaded program. |
+--------------------------------------------------------------------+
| |
| /SUPPRESS Suppresses a previously defined global symbol. |
+--------------------------------------------------------------------+
| |
| /SYFILE Requests LINK to output a symbol file to the given |
| filespec, and sets the /SYMSEG:DEFAULT switch. |
+--------------------------------------------------------------------+
| |
| /SYMSEG Allows you to specify where the symbol table is to |
| be placed. |
+--------------------------------------------------------------------+
| |
| /SYSLIB Forces searching of one or more system libraries, |
| immediately after you end the command line. |
+--------------------------------------------------------------------+
| |
| /TEST Loads a debugger. |
+--------------------------------------------------------------------+
| |
| /UNDEFINED Displays undefined global symbols on the terminal. |
+--------------------------------------------------------------------+
| |
| /UPTO Sets an upper limit to which the symbol table can |
| expand. |
+--------------------------------------------------------------------+
| |
| /USERLIB Directs LINK to search a user library. |
+--------------------------------------------------------------------+
| |
| /VALUE Displays global symbol values on the terminal. |
+--------------------------------------------------------------------+
| |
| /VERBOSITY Specifies the length of LINK messages. |
+--------------------------------------------------------------------+
| |
| /VERSION Allows you to specify a version number. |
+--------------------------------------------------------------------+
1-12
CHAPTER 2
USING LINK AUTOMATICALLY
The TOPS-20 LOAD, EXECUTE, and DEBUG commands invoke LINK
automatically. Each of these commands uses a simple command string
that the system converts into more complicated LINK commands.
This discussion of the LOAD, EXECUTE, and DEBUG commands does not
attempt to describe them completely. Only those switches applying
directly to loading are discussed here. For a full discussion, see
the TOPS-20 Commands Reference Manual.
Each of these commands invoke LINK:
o The LOAD command uses LINK to load your object modules into
memory. LOAD compiles source files first if necessary, but
does not execute the program.
o The EXECUTE command uses LINK to load your program, and then
executes the loaded program. Before loading, your source
files are compiled, if necessary.
o The DEBUG command loads your program into memory along with a
debugger. The program is then executed under the control of
the debugger. The debugger that is loaded depends on the
type of program being loaded. See the /TEST switch for a
list of languages. TOPS-20 uses the file type to determine
the language in which the program is written. Therefore, it
is highly recommended that you use standard file types when
naming your programs' files. Standard file types are listed
in the TOPS-20 Commands Reference Manual.
2.1 COMMAND FORMATS
The LOAD, EXECUTE, and DEBUG commands follow the same format. Each
can accept a list of input file specifications and switches. The
format for these commands is:
2-1
FIELD TEST
@command/switches input-filespec/switches, input-filespec/switches,...
Where:
command is LOAD, EXECUTE, or DEBUG.
input-spec is the file specification of the program you want
to load. This input file specification accepts a
6-character device name, 39-character directory
name, a 6-character filename and a 3-character
extension.
switches are any of the valid switches for the command.
Table 2-1 contains switches that apply to LINK.
If you separate the input file specifications with commas, each source
file is compiled into a separate object (relocatable) file. For
example,
@LOAD PROGA,PROGB,PROGC
compile into separate object modules, PROGA.REL, PROGB.REL and
PROGC.REL.
If you separate the input file specifications with plus signs, they
will be compiled into a single object file.
@LOAD PROGD+PROGE+LIBRAY
compile into a single object module, LIBRAY.REL
Section 2.3 shows additional examples of using LINK automatically.
2.2 COMMAND SWITCHES
You can use switches with the LOAD, EXECUTE, and DEBUG commands to
control the program's loading. Table 2-1 briefly describes some of
the command switches that apply to LINK. Refer to the TOPS-20
Commands Reference Manual for complete descriptions of the switches
for these commands.
Table 2-1: Switches for TOPS-20 Commands
+-------------------------------------------------------------------+
| |
| Switch Meaning |
+-------------------------------------------------------------------+
2-2
FIELD TEST
Table 2-1 (cont.)
+-------------------------------------------------------------------+
| |
| Switch Meaning |
+-------------------------------------------------------------------+
| |
| /COMPILE Forces compilation of source files even if a |
| sufficiently recent REL file exists. |
+-------------------------------------------------------------------+
| |
| /DDT Loads DDT. This supersedes the default debugger |
| selection that is based on the file type of the |
| first file in the command string. |
+-------------------------------------------------------------------+
| |
| /MAP Produces a map file at the end of loading. |
+-------------------------------------------------------------------+
| |
| /NOCOMPILE Compiles source files only if their REL files are |
| older than the source files. /NOCOMPILE is the |
| default. |
+-------------------------------------------------------------------+
| |
| /NOSEARCH Suspends the effect of an earlier global /SEARCH |
| switch. This is the default action. |
+-------------------------------------------------------------------+
| |
| /NOSYMBOLS Prevents loading of symbol tables with their |
| modules. |
+-------------------------------------------------------------------+
| |
| /SEARCH Loads only the modules from the specified library |
| file that satisfy global references in the program. |
+-------------------------------------------------------------------+
You can use any LINK program switches with LOAD, EXECUTE, or DEBUG by
using a special switch format. This format requires that you use a
percent sign (%) instead of the usual slash (/), and that the entire
switch specification be enclosed in double quotation marks ("). For
example, you can pass the /SYMSEG:HIGH switch to LINK by using the
command:
@EXECUTE MYPROG %"SYMSEG:HIGH"<RET>
If you give more than one switch in this format, succeeding switches
within the quotation marks must have the usual slashes:
@EXECUTE MYPROG %"SYMSEG:HIGH/COUNTERS"<RET>
LINK program switches are described in Chapter 3.
2-3
FIELD TEST
2.3 EXAMPLE OF USING LINK AUTOMATICALLY
For this example, the following program, named MYPROG.FOR, is used:
TYPE 10
10 FORMAT (' This is written by MYPROG')
STOP
END
The following example shows an interactive execution of the program
using the EXECUTE command:
@EXE<ESC>CUTE (FROM) MYPROG.FOR<RET>
FORTRAN: MYPROG
MAIN.
LINK: Loading
[LNKXCT MYPROG execution]
This is written by MYPROG
CPU time 0.21 Elapsed time 1.31
The following example shows how to load a program for debugging using
the DEBUG command:
@DEBUG MYPROG.FOR<RET>
FORTRAN: MYPROG
MAIN.
LINK: Loading
[LNKDEB FORDDT execution]
STARTING FORTRAN DDT
>>START<RET>
This is written by MYPROG
CPU time 0.17 Elapsed time 0.46
2-4
CHAPTER 3
USING LINK DIRECTLY
You can choose to use the TOPS-20 LOAD, EXECUTE, and DEBUG commands to
load a program, or you can use LINK directly. Using LINK directly is
useful when you have a large or complicated program, are loading
overlays, or are using PSECTs because you can better control the
loading process. For example, in the case of a program that uses
PSECTs, you can use LINK switches to specify the origins of the
PSECTs. In the case of an overlay program, you can use LINK switches
to define the program's tree structure.
Before you load a program with LINK, you must have compiled or
assembled all required object modules.
3.1 RUNNING AND EXITING LINK
To run LINK, type LINK after the TOPS-20 system prompt. LINK returns
an asterisk (*) and waits for you to enter commands. For example,
@LINK<RET>
*
To exit LINK, use the /GO switch.
The /GO switch finishes loading the program and exits. /GO exits to
TOPS-20 command level, the loaded program for execution, or to a
debugger. /GO passes control to the loaded program if you specified
/EXECUTE, and to a debugger if you specified /DEBUG. For example, in
the following /GO passes control to the loaded program.
@LINK<RET>
*MYPROG/EXECUTE/GO
[LNKXCT MYPROG execution]
This is written by MYPROG
CPU time 0.21 Elapsed time 0.82
If LINK encounters a fatal error, LINK prints an error message and
3-1
FIELD TEST
exits to TOPS-20.
3.2 COMMAND FORMATS
A LINK command can contain file specifications and switches. The
general format of a LINK command is:
*/switch input-filespec /switch output-filespec /switch
where:
switch is a LINK switch. LINK switches are described in
Section 3.3. Section 3.3 also discusses switch
arguments, abbreviations, and placement.
filespec consists of 1- to 6-character device name, 1- to
6-character filename, a 1- to 3-character file
type, and a PPN (project-programmer number).
You only need to supply a filename as LINK
supplies the missing parts of the file
specifications from its defaults. These defaults
are listed in Section 3.2. You can use a logical
name to specify a device name, and the TOPS-20
TRANSLATE command to determine a PPN. See
Sections 3.2.2 and 3.2.3 for additional
information on logical names and TRANSLATE
respectively.
You can continue a command, include a comment, or use indirect command
files with LINK.
By typing a hyphen before entering a carriage return, you can continue
a command onto the next line. LINK continues the line by prompting
with a hash sign (#). For example,
@LINK<RET>
*MYPROG,MYMAP/MAP/CONTENTS:ALL-<RET>
#/ERRORLEVEL:0/LOG/LOGLEVEL:5<RET>
*
You can include a comment in a command by beginning the comment with a
semicolon. LINK does not process the text after the semicolon.
Commenting commands can be useful in indirect command files. For
example,
/CONTENTS:(LOCALS,UNDEFINED) ;Include local and undefined
;symbols in my map file.
PROMAP/MAP:ERROR ;Give me a map file named PROMAP
;when LINK encounters an error.
3-2
FIELD TEST
PROGAM
/GO ;Load my program
These commands and comments when placed in an indirect command file
provide you with a description of how you want a program loaded.
To have LINK read commands from an indirect command file, prefix the
command file specification with an at-sign (@) and enter it after the
LINK prompt. For example,
@LINK<RET>
*@LNKPRG
LINK reads the commands in LNKPRG and processes them. LNKPRG can have
a .CCL or .CMD file type. Both are default file types for indirect
command files. LINK first looks for a .CCL file type and then a .CMD
file type.
3.2.1 LINK File Specification Defaults
For input files, the defaults are:
device logical name (DSK:)
type REL
directory the connected directory. If you are using a PPN,
refer to Section 3.3.3 for information on translating
directories.
For output files, the defaults are:
device logical name (DSK:)
filename name of last module with start address or, if none,
then nnnLNK where nnn is your job number in decimal,
with any leading zeros.
type one or all of the following depending on the switches
you use.
/EXECUTE executable file EXE
/LOG log file LOG
/MAP map file MAP
/PLOT plotter file PLT
/SYFILE symbol file SYM
directory the connected directory. If you are using a PPN,
refer to Section 3.3.3 for information on translating
directories.
3-3
FIELD TEST
You can change these defaults by using the /DEFAULT switch. See
Section 3.3.4.
3.2.2 Logical Names
To use a logical name:
1. Give the TOPS-20 DEFINE command to define a logical name. A
logical name cannot be longer than six characters.
2. Use the logical name as the device name whenever giving the
file specification.
See the TOPS-20 User's Guide for more information.
3.2.3 Translating Directories
To find a corresponding PPN:
1. Use the TOPS-20 TRANSLATE command to find the corresponding
PPN for the given directory name.
2. Include the PPN enclosed in square brackets ([]) at the end
of the file specification in the format:
device:filename.type[PPN]
3.3 LINK SWITCHES
LINK switches control and modify the loading process. This section
describes switch abbreviations, arguments, and placements followed by
an alphabetical listing of the LINK switches.
3.3.1 Switch Abbreviation
You can abbreviate switches to save typing. However, when
abbreviating a switch, you must must include enough characters to make
the abbreviation unique for the switch. For example, both /NOSYMBOL
and /NOSYSLIB begin with the characters /NOSY, therefore you cannot
abbreviate the /NOSYMBOL switch to /NOSY. You can abbreviate
/NOSYMBOL to /NOSYM because /NOSYM is a unique set of characters for
/NOSYMBOL.
The following switches can be abbreviated to a single letter by
3-4
FIELD TEST
default:
/D for /DEBUG
/E for /EXECUTE
/G for /GO
/H for /HELP
/L for /LOCALS
/M for /MAP
/N for /NOLOCAL
/S for /SEARCH
/T for /TEST
/U for /UNDEFINE
/V for /VERSION
Use the full switch name when placing switches in batch or indirect
command files to avoid incorrect or ambiguous abbreviations.
3.3.2 Switch Arguments
Many LINK switches require arguments. An argument can be a value, a
symbol name, a module name, an output file specification, a PSECT
name, a keyword, or an address. The specific requirements of each
switch are discussed in the FORMAT section of the switch's
description.
A value can be specified in decimal or octal. The default is provided
in the FORMAT information. If the value can be specified in octal,
this is noted in the OPTIONAL NOTATIONS section of the switch
description. To specify an octal value, type a hash sign (#) before
the octal number. For example, /ARSIZE:39 can be specified in octal
as /ARSIZE:#47.
A symbol name, a module name, or a PSECT name consists of from 1 to 72
SIXBIT-compatible ASCII characters. You cannot embed spaces in any of
the names. If the name is not Radix-50 compatible, you must enclose
it in double quotes. For example, NOT^RADIX would be specified as
"NOT^RADIX" in a LINK command. If you enclose a name in quotes, be
aware that how you enter the name (all in UPPERCASE, all in lowercase,
or in Mixed cAse) affect the interpretation of the symbol. For
example,
"DIFFERENT" "different" "Different"
all define individual symbols. However, if you do not enclose a name
in quotes and enter the name in mixed case, the name defines the same
symbol. For example,
SAME same Same
all define the symbol same and are equivalent to the quoted name
3-5
FIELD TEST
"SAME".
An overlay link name consists of from 1- to 6-Radix-50 characters that
are compatible with ASCII characters.
A keyword is a unique argument that is associated with the switch.
For example, /DEBUG's keywords identify the debugger to be loaded with
the program.
An address is an octal number.
An output file specification specifies a file that LINK is to create.
See Section 3.2 for format and default information on output file
specifications. The following switches produce an output file:
/DEFAULT /LOG
/MAP /OVERLAY
/PLOT /NOUSERLIB
/SAVE /SYFILE
/USERLIB
When specifying output file specifications with these switches, append
the switch to the output file specification in the format:
outputfilespec/switch
For example,
*LOGFIL/LOG
directs LINK to write LINK messages into the output file LOGFIL.LOG.
The other file specifications in a LINK command specify input files.
For example, the following command tells LINK to use an input file
called MYREL.REL to generate a saved output file called MYEXE.EXE.
/SAVE specifies MYEXE as the output filename, and .EXE is the default
file type:
*MYREL,MYEXE/SAVE/GO
3.3.3 Switch Placement
The /LOCALS, /NOLOCAL, /NOSTART, /SEGMENT, /INCLUDE, /ONLY, /SEARCH,
/START and /NOSEARCH switches can be local or global switches. A
local switch applies to a single file specification and is appended to
that file specification in the command. The switch is appended before
the comma in a list of file specifications. For example, in the
following command, /SEARCH is used as a local switch to load FILE2 in
search mode. /SEARCH is appended to FILE2 before the comma.
3-6
FIELD TEST
*FILE1,FILE2/SEARCH,FILE3<RET>
A global switch applies to all the files in a command and is disabled
at the end of the command when you enter a carriage-return. A global
switch is not appended to a file specification, and can be placed
after the LINK prompt (*), or after a filename. A global switch is
appended after the comma in a list of file specifications. For
example, in the following commands /SEARCH is a global switch used to
load all the files in search mode. In the first command, /SEARCH is
placed after the LINK prompt. In the second, note that /SEARCH is
placed after the comma following the first file specification.
*/SEARCH FILE4,FILE5,FILE6<RET>
*FILE4,/SEARCH FILE5,FILE6<RET>
A global switch applies to all the files entered after the asterisk
and before the carriage-return, unless it is overridden by another
switch. If this second switch is a global switch, it persists for all
the files that follow in the command. For example, in the following
command, the /NOSEARCH disables the /SEARCH switch for files FILE10
and FILE11.
*FILE7,/SEARCH FILE8, FILE9,/NOSEARCH FILE10, FILE11<RET>
If the second switch is a local switch, it overrides the first switch
only for the file to which it is appended. For example, in the
following command string, the global /SEARCH switch is overridden by
the local /NOSEARCH switch, but only for FILE2. FILE1 and FILE3 are
loaded in search mode.
*/SEARCH FILE1,FILE2/NOSEARCH,FILE3
NOTE
When the /GO switch is present in the same command as
a global switch, /GO causes the effects of the global
switch to go beyond the command and apply to any
modules loaded during library searches. For example,
*/SEGMENT:LOW FILEA,FILEB,FILEC/60<RET>
causes the input files on this line and files loaded
during a library search to be loaded in the program's
low-segment. However, defaults for certain languages
can override the global switches that you specify.
3-7
FIELD TEST
3.3.4 Alphabetical Listing of LINK Switches
In this section the following information is shown, if appropriate,
for each switch:
FORMAT
FUNCTION
EXAMPLES
OPTIONAL NOTATIONS
RELATED SWITCHES
3-8
FIELD TEST
/ARSIZE
FORMAT /ARSIZE:n
n is a positive decimal integer.
FUNCTION Sets the size of the overlay handler's table of
multiply-defined global symbols. Use this switch if you
have received LNKARL, LNKTMA, and LNKABT messages in a
previous attempt to load your program. These messages
give instructions for the argument to the /ARSIZE switch.
EXAMPLES @LINK<RET>
*/ARSIZE:39<RET>
*
Allocates 39 words for the multiply-defined global symbol
table in each link of an overlay structure.
OPTIONAL You can specify the table size in octal.
NOTATIONS
3-9
FIELD TEST
/COMMON
FORMAT /COMMON:name:n
name is a common name. The symbol name rules described in
Section 3.3.2 apply to common name.
n is a positive decimal integer.
FUNCTION Allocates n words of labeled COMMON storage for FORTRAN
and FORTRAN-compatible programs. The COMMON label is a
name which becomes defined as a global symbol.
For unlabeled COMMON storage, use .COMM. as the name, or
simply omit the name.
You cannot expand a given COMMON area during loading. If
your program modules define a given COMMON area to have
different sizes, the module giving the largest definition
must be loaded first. If the /COMMON switch gives the
largest definition, it must precede the loading of the
modules.
EXAMPLES @LINK<RET>
*/COMMON:ALPHA:1000<RET>
*
Creates a labeled COMMON area of 1000 words.
@LINK<RET>
*/COMMON:.COMM.:1000<RET>
*
Creates an unlabeled COMMON area of 1000 words.
@LINK<RET>
*/COMMON::1000<RET>
*
Creates an unlabeled COMMON area of 1000 words.
OPTIONAL You can specify the number of words in octal.
NOTATIONS
RELATED /PSCOMMON
SWITCHES
3-10
FIELD TEST
/CONTENTS
FORMAT /CONTENTS:(keyword,...,keyword)
FUNCTION specifies the symbol types to be included in the map file
if the file is generated. To generate the map file, use
the /MAP switch.
The keywords ALL, NONE, and DEFAULT reset all symbol
types. Otherwise, using the /CONTENTS switch resets only
those symbol types specified by keywords. In the
following list of keywords, the defaults are shown in
bold:
Keyword Description
ABSOLUTE Include absolute symbols.
ALL Include all symbols.
COMMON Include COMMON symbols.
DEFAULT Reset to LINK's defaults.
ENTRY Include entry-name symbols.
GLOBAL Include global symbols.
LOCALS Include local symbols. The local
symbols cannot be included in the map
file unless the /LOCALS switch is also
given.
NOABSOLUTE Exclude absolute symbols.
NOCOMMON Exclude COMMON symbols.
NOENTRY Exclude entry-name symbols.
NOGLOBAL Exclude global symbols.
NOLOCAL Exclude local symbols.
NONE Exclude all symbols.
NORELOCATABLE Exclude relocatable symbols.
NOUNDEFINED Exclude undefined symbols.
NOZERO Exclude symbols in zero-length
programs. (a zero-length program
contains no code or data; it contains
only symbol definitions, for example,
JOBDAT.)
RELOCATABLE Include relocatable symbols.
UNDEFINED Include undefined symbols.
ZERO Include symbols in zero-length
programs.
Only those symbols that satisfy all conditions in the
keyword list will appear in the MAP file. For example, if
both the NOGLOBAL and RELOCATABLE settings are in force,
all global symbols are excluded regardless of their
relocatability.
3-11
FIELD TEST
EXAMPLES @LINK<RET>
*/CONTENTS:(NOCOMMON,NOENTRY)<RET>
*
Excludes COMMON and entry-name symbols.
@LINK<RET>
*/CONTENTS:ALL<RET>
*
Includes all symbols.
OPTIONAL You can omit the parentheses if you give only one keyword.
NOTATIONS
RELATED /MAP
SWITCHES
3-12
FIELD TEST
/COUNTERS
FORMAT /COUNTERS
FUNCTION Displays information about relocation counters on the
terminal. A relocation counter is an address counter that
LINK uses while loading relocatable code.
/COUNTERS returns the name, initial value, current value,
and limit value of each counter. /COUNTERS first prints a
header:
Reloc. ctr. initial value current value limit value
Reloc. ctr. gives the name of relocation counter.
initial value is the origin of the relocation counter.
current value is the address of the next free location
after the relocation counter has been
loaded.
limit value is an upper bound that you set using
/LIMIT or that LINK sets by default for
the relocation counter. This upper
bound defines a point the relocation
counter should not load beyond. If
/LIMIT is used and the counter loads
beyond this bound, LINK returns
messages. See /LIMIT for more
information.
/COUNTERS may be used to determine the size of overlays
when loading large programs that might be too large for
the allocated memory space. Refer to Section 5.4 for more
information.
You can also use /COUNTERS to determine the size of PSECTs
when loading extended addressing programs or programs that
use PSECTs to conserve memory space. Refer to Chapter 6.
EXAMPLES The following examples illustrate the various /COUNTERS
displays.
The following display results from loading a module that
does not contain code.
@LINK<RET>
*/COUNTERS<RET>
[LNKRLC No relocation counters set]
3-13
FIELD TEST
The following display results from loading only absolute
code.
@LINK<RET>
*ABCODE/COUNTERS<RET>
[LNKRLC No relocation counters set
Absolute code loaded]
*
The following display results from loading only PSECT
code.
@LINK<RET>
*PSCODE/COUNTERS<RET>
[LNKRLC Reloc. ctr. initial value current value limit value
PSCODE 20 25 1000000]
*
The following display results from loading code that
contains both absolute and PSECT code.
@LINK<RET>
*MIXED/COUNTERS<RET>
[LNKRLC Reloc. ctr. initial value current value limit value
PSECTA 1400000 1400001 4000000
PSECTB 2500000 2500001 4000000
PSECTC 3600000 3600001 4000000
Absolute code loaded]
*
The following display results from loading two-segmented
formatted code.
@LINK<RET>
*TWOPRT/COUNTERS<RET>
[LNKRLC Reloc. ctr. initial value current value limit value
.LOW. 0 1642 1000000
.HIGH. 400000 400753 1000000]
*
RELATED /NEWPAGE, /SET, /LIMIT
SWITCHES
3-14
FIELD TEST
/CPU
FORMAT /CPU:(keyword,keyword)
Keyword: KA10
KI10
KL10
KS10
FUNCTION Overrides LINK's handling of the processor information
found in the REL files being loaded. (See the description
of Block Type 6 in Appendix A). Ordinarily, LINK prints a
warning if all REL files being loaded together do not have
identical CPU types. This switch can be used either to
make LINK flag certain modules built for a specific CPU
type (by specifying all but that CPU type as keywords to
/CPU) or to suppress LINK's warning message (by specifying
all the CPU types associated with the REL files being
loaded).
EXAMPLES @LINK<RET>
*/CPU:KI10<RET>
*
Causes LINK to issue the %LNKCCD message if any modules
with the KL10 CPU type are encountered.
OPTIONAL You can omit the parentheses if you specify only one CPU.
NOTATIONS
3-15
FIELD TEST
/DDEBUG
FORMAT /DDEBUG:keyword
FUNCTION Specifies a default debugger to be loaded if the /DEBUG or
/TEST switch appears without an argument.
The keywords and the debuggers they specify are listed
below. Only those shown in bold are supported by DIGITAL.
Keyword Debugger
ALGDDT ALGDDT
ALGOL ALGDDT
COBDDT COBDDT
COBOL COBDDT
DDT DDT
FAIL SDDT (SAIL debugger)
FORDDT FORDDT
FORTRAN FORDDT
MACRO DDT
PASCAL PASDDT
PASDDT PASDDT
SAIL SDDT (SAIL debugger)
SDDT SDDT (SAIL debugger)
SIMDDT SIMDDT
SIMULA SIMDDT
EXAMPLES @LINK<RET>
*/DDEBUG:FORTRAN<RET>
*
Specifies FORDDT as the default debugger for the /DEBUG or
/TEST switch.
RELATED /DEBUG, /TEST
SWITCHES
3-16
FIELD TEST
/DEBUG
FORMAT /DEBUG:keyword
FUNCTION Requests loading of a debugger and sets the start address
for execution at the start address of the debugger. The
/DEBUG switch also sets the /EXECUTE switch because it is
assumed that the program is to be executed. The /GO
switch is still required to end loading and begin
execution.
The /DEBUG switch turns on the /LOCALS switch for the
remainder of the load. You can override this by using the
/NOLOCAL switch, but the override lasts only during
processing of the current command.
Local symbols for the debugger itself are never loaded.
If debugging overlaid programs, you must specify /DEBUG
when loading the root node. (Refer to Section 5.4 for
more information.)
The keywords and the programs they load are listed below.
Only those shown in bold supported by DIGITAL.
Keyword Debugger
ALGDDT ALGDDT
ALGOL ALGDDT
COBDDT COBDDT
COBOL COBDDT
DDT DDT
FAIL SDDT (SAIL debugger)
FORDDT FORDDT
FORTRAN FORDDT
MACRO DDT
PASCAL PASDDT
PASDDT PASDDT
SAIL SDDT (SAIL debugger)
SDDT SDDT (SAIL debugger)
SIMDDT SIMDDT
SIMULA SIMDDT
If you give no keyword with /DEBUG, the default is either
DDT or the debugger specified by the /DDEBUG switch.
3-17
FIELD TEST
EXAMPLES @LINK<RET>
*/DEBUG:FORDDT<RET>
*
Loads FORDDT, sets the /EXECUTE switch, and specifies that
FORDDT control execution.
OPTIONAL You can abbreviate /DEBUG to /D.
NOTATIONS
RELATED /DDEBUG, /TEST
SWITCHES
3-18
FIELD TEST
/DEFAULT
FORMAT filespec/DEFAULT:keyword
FUNCTION Changes default specifications for input or output files.
The defaults specified remain in effect until changed by
another /DEFAULT switch.
The keywords allowed are:
INPUT Specifies the defaults for input file
specifications.
OUTPUT Specifies the defaults for output file
specifications.
For input files, the initial defaults are:
device DSK:
file type REL
directory Your connected directory
For output files, the initial defaults are:
device DSK:
filename Name of main program
directory Your connected directory
EXAMPLES @LINK<RET>
*/DEFAULT:INPUT .BIN<RET>
*
Resets input file default extension to BIN.
@LINK<RET>
*/DEFAULT:OUTPUT MYSTR:<RET>
*
Resets output file default device to MYSTR:.
OPTIONAL If you omit the keyword, INPUT is assumed.
NOTATIONS
3-19
FIELD TEST
/DEFINE
FORMAT /DEFINE:(symbol:value,...,symbol:value)
symbol is a global symbol.
value is a decimal number.
FUNCTION Assigns a decimal value to a symbol. This assignment
causes the symbols to be global symbols. To get a list of
any undefined symbols, use the /UNDEFINED switch.
Defining an already defined symbol with /DEFINE generates
an error message.
EXAMPLES */UNDEFINED<RET>
[LNKUGS 2 UNDEFINED GLOBAL SYMBOLS]
ALPHA 400123
BETA 402017
*/DEFINE:(ALPHA:591,BETA:1)<RET>
*
Gives the decimal values 591 and 1 to ALPHA and BETA,
respectively.
OPTIONAL You can give the value in octal by typing a hash sign (#)
NOTATIONS before the value. You can omit the parentheses if you
define only one symbol.
RELATED /UNDEFINED, /VALUE
SWITCHES
3-20
FIELD TEST
/ENTRY
FORMAT /ENTRY
FUNCTION Displays all entry name symbols that have been loaded on
the terminal. Each entry name symbol is defined by an
ENTRY statement (MACRO, FORTRAN, or BLISS), a FUNCTION
statement (FORTRAN), a SUBROUTINE statement (FORTRAN, or
COBOL), or a PROCEDURE declaration (ALGOL, or PASCAL).
If you are using the overlay facility, /ENTRY requests
only the entry name symbols for the current overlay link.
EXAMPLES @LINK<RET>
/ENTRY<RET>
[LNKLSS LIBRARY SEARCH SYMBOLS (ENTRY POINTS)]
SQRT. 3456
*
RELATED /NOENTRY
SWITCHES
3-21
FIELD TEST
/ERRORLEVEL
FORMAT /ERRORLEVEL:n
n is a positive decimal integer.
FUNCTION Suppresses the terminal display of LINK messages with
message level less than n, where n is a decimal number
between 0 and 31 inclusive. You cannot suppress level 31
messages. LINK's default is /ERRORLEVEL:8.
See Appendix B for the level of each LINK message.
EXAMPLES @LINK<RET>
*/ERRORLEVEL:15<RET>
*
Suppresses all messages less than level 15.
@LINK<RET>
*/ERRORLEVEL:0<RET>
*
Permits typeout of all messages.
RELATED /VERBOSITY, /MESSAGE
SWITCHES
3-22
FIELD TEST
/EXCLUDE
FORMAT /EXCLUDE:(module,...,module)
FUNCTION Prevents loading of the specified modules from the current
file even if they are required to resolve global symbol
references. You can use the /EXCLUDE switch for any of
the following purposes:
o If a library has several modules with the same search
symbols, you can select the module you want by
excluding the others.
o If several modules contain the same symbol, you can
select the module with the symbol definition you want
by excluding the others. This also prevents modules
from giving multiple definitions of a symbol.
o In defining an overlay structure, you can delay
loading of a module until a later overlay link by
excluding the module.
EXAMPLES @LINK<RET>
*/SEARCH LIBFIL.REL/EXCLUDE:(MOD1,MOD2)<RET>
*
Searches LIBFIL as a library but prevents loading of MOD1
and MOD2 even if they resolve global symbol references.
OPTIONAL You can omit the parentheses if you specify only one
NOTATIONS module.
RELATED /INCLUDE, /NOINCLUDE
SWITCHES
3-23
FIELD TEST
/EXECUTE
FORMAT /EXECUTE
FUNCTION Directs LINK to execute the loaded program beginning at
the program's start address. LINK continues loading until
a /GO switch is found.
Use either the /EXECUTE or /DEBUG switches to load and
execute a program.
EXAMPLES @LINK<RET>
*/EXECUTE<RET>
*PROGRM/GO
OPTIONAL You can abbreviate /EXECUTE to /E.
NOTATIONS
RELATED /DEBUG, /GO, /RUN, /TEST
SWITCHES
3-24
FIELD TEST
/FRECOR
FORMAT /FRECOR:nK
n is a positive decimal integer.
K is the representation for 1024.
FUNCTION Requires LINK to maintain a minimum amount of free memory
after any expansions. LINK's default free memory is 4K.
If you use the /FRECOR:nK switch, LINK computes n times
1024 words and maintains the resulting number of words of
free memory, if possible.
If the modules to be loaded are quite large, a larger
amount of free memory avoids some moving of areas.
LINK has ten areas that may be expanded during loading:
1. ALGOL symbol information (AS). +
2. Bound global symbols (BG).
3. Dynamic area (DY).
4. Fixup area (FX).
5. Global symbol tables (GS).
6. User's high segment code (HC). +
7. User's low segment code (LC). +
8. Local symbol tables (LS). +
9. Relocation tables (RT).
10. Argument typechecking (TP). +
Areas that can overflow are marked with a +. Each of
these areas has a lower bound, an actual upper bound, and
a maximum upper bound. LINK normally maintains space
between the actual and maximum upper bounds for each area.
The total of these ten spaces is at least the space given
by the /FRECOR switch, if possible.
LINK recovers free memory by concatenating these ten free
areas. When all this recovered space is used and one or
more of the ten areas overflows then free memory is no
longer maintained because the free memory has been used.
EXAMPLES @LINK<RET>
*/FRECOR:7K<RET>
*
Maintains 7K of free memory, if possible.
OPTIONAL You can specify the free memory in octal using a hash sign
NOTATIONS (#). For example, /FRECOR:#11K.
3-25
FIELD TEST
/GO
FORMAT /GO
FUNCTION Ends loading after the current file and exits LINK. LINK
then performs any required library searches, generates any
required output files, and does one of the following:
o Begins execution at the start address of the loaded
program if you used /EXECUTE.
o Begins execution at the start address of the debugger
if you used /DEBUG.
o Exits to the monitor if you used no execution switch.
EXAMPLES @LINK<RET>
*MYPROG/EXECUTE/GO<RET>
[LNKXCT MYPROG execution]
Begins execution of the loaded program at its start
address.
@LINK<RET>
*MYPROG/DEBUG/GO<RET>
[LNKDEB DDT execution]
Begins execution of the loaded program at the start
address of DDT.
OPTIONAL You can abbreviate /GO to /G.
NOTATIONS
RELATED /DEBUG, /EXECUTE, /RUN
SWITCHES
3-26
FIELD TEST
/EXIT
FORMAT /EXIT
FUNCTION Exits LINK. LINK does not finish loading the program. To
finish loading the program and to exit LINK, use /GO.
EXAMPLES @LINK
*MYPROG<RET>
*/EXIT
@
3-27
FIELD TEST
/HASHSIZE
FORMAT /HASHSIZE:n
n is a positive decimal integer.
FUNCTION Gives a minimum for the initial size of the global symbol
table. LINK selects a prime number larger than n for the
initial size.
LINK automatically expands a symbol table. However, if
you know that you need a large global symbol table, you
can save time and space by allocating space for it with
/HASHSIZE. You should give a hash size at least 10
percent larger than the number of global symbols in the
table.
If LINK gives the message [LNKRGS Rehashing Global Symbol
Table] during a load, you should use the /HASHSIZE switch
for future loads of the same program. The minimum hash
size for loading a program appears in the header lines of
the map file.
The default hash size is a LINK assembly parameter
(initially 251 decimal).
EXAMPLES @LINK<RET>
/HASHSIZE:1000<RET>
*
Sets the hash size to the prime number 1021.
3-28
FIELD TEST
/HELP
FORMAT /HELP:keyword
DEFAULTS /HELP:TEXT
FUNCTION Displays information about LINK on the terminal. The
keywords are:
Keyword Description
SWITCHES prints a list of valid switches.
TEXT prints a more detailed description of
command formats and switches.
OPTIONAL You can abbreviate /HELP to /H.
NOTATIONS
3-29
FIELD TEST
/INCLUDE
FORMAT /INCLUDE:(module,...,module)
FUNCTION Requires the loading of the specified modules whether or
not there are any global requests for them. For example,
*/INCLUDE:(MODU6,MODU9)<RET>
requires that modules MODU6 and MODU9 be loaded.
/INCLUDE can be used as a local or global switch. If
/INCLUDE is used as a local switch the request for the
module is cleared after the file is loaded. For example,
*FILE8/INCLUDE:(MODU7,MODU8)<RET>
loads module MODU7 and MODU8 from FILE8.
If /INCLUDE is used as a global switch, the request for
the modules are not cleared until the modules are loaded.
For example,
*/INCLUDE:(MODU2,MODU3,MODU4) FILEA,FILEB,FILEC<RET>
requires loading of MODU2, MODU3, and MODU4. If these
modules do not exist in FILEA, FILEB, and FILEC, /INCLUDE
request continues until they are loaded.
If you used /SEARCH specifying library search mode and
specify /INCLUDE, the /INCLUDE switch causes the modules
specified in the switch to be loaded in addition to
modules that are loaded to satisfy global symbol
references. For example,
*LIBARY/SEARCH/INCLUDE:(DATMOD,NAMMOD,NUMMOD)<RET>
Causes LIBARY to be searched and DATMOD, NAMMOD, and
NUMMOD to be loaded along with other modules that are
needed to resolve global symbol references.
You can use /INCLUDE in an overlay load to force a module
to be loaded in an ancestor link common to successor links
that reference that module. This makes the module
available to all links that are successors to its link.
EXAMPLES @LINK<RET>
*/SEARCH LIB1/INCLUDE:(MOD1,MOD2)<RET>
*
Searches LIB1 and loads MOD1 and MOD2 even if they are not
needed to resolve global symbol references.
3-30
FIELD TEST
OPTIONAL You can omit the parentheses if you specify only one
NOTATIONS module.
RELATED /EXCLUDE, /NOINCLUDE, /MISSING
SWITCHES
3-31
FIELD TEST
/LIMIT
FORMAT /LIMIT:psect:origin
psect is a PSECT name.
origin is a thirty-bit octal address or a defined symbol.
FUNCTION specifies an upper bound for a PSECT. If the PSECT loads
beyond this bound, LINK returns error messages.
Before using /LIMIT, the PSECT name must be defined with
either the /SET switch or in one of the modules already
loaded.
The upper bound can be expressed in either numeric or
symbolic form. This address is one greater than the
highest location which may be loaded in the PSECT and need
not be in the same section as the PSECT origin.
If the PSECT grows beyond the address specified in the
/LIMIT switch, LINK sends a warning to your terminal, but
continues to process input files and to load code. The
warning message takes the following form:
%LNKPEL PSECT <psect> exceeded limit of <address>
Although LINK does continue to process input files and
load code, the program is incomplete and should NOT be
used. LINK produces the following fatal error:
?LNKCFS Chained fixups have been suppressed
Chained fixups are a method that LINK uses to resolve
symbol references.
Using /LIMIT to define an upper bound prevents unintended
PSECT overlaps. PSECT overlaps can cause LINK to loop and
produces other unpredictable behavior, because LINK uses
the unresolved address relocation chains in the user
program being built.
EXAMPLES In the following example, the LRGPRO and BIGPRO programs
each contain two PSECTs, BIG and GRAND. LRGPRO is loaded
first using /COUNTERS to check the origin and the current
value of the PSECTs. The PSECT origin is found by looking
under the initial value column and the current value is
found by looking under the current value of the /COUNTERS
output.
@LINK<RET>
*/SET:BIG:1000<RET>
3-32
FIELD TEST
*/SET:GRAND:5400<RET>
*LRGPRO<RET>
*/COUNTERS<RET>
[LNKRLC Reloc. ctr. initial value current value limit value
.LOW. 0 140 1000000
BIG 1000 5100 1000000
GRAND 5400 10500 1000000
Absolute code loaded]
/COUNTERS shows that the current value for PSECT BIG and
the initial value for PSECT GRAND are close together in
memory. The current value for BIG is 5100 and the PSECT
origin for GRAND is 5400. The /LIMIT switch can now be
used to restrict PSECT BIG upper bound to PSECT GRAND's
initial value using:
*/LIMIT:BIG:GRAND<RET>
This prevents an unintended overlap when PSECT BIG from
BIGPRO is loaded because the upper bound for BIG is set to
GRAND's origin. Now when BIGPRO is loaded, if PSECT BIG
exceeds the limit, LINK produces the LNKPEL warning
indicating that the upper bound needs to be increased.
The /COUNTERS switch shows a new current value greater
than 5400. Notice that the limit set with the /LIMIT
switch is shown in the LIMIT VALUE column.
*BIGPRO<RET>
%LNKPEL PSECT BIG exceeded limit of 5100
detected in module .MAIN from file BIGPRO.REL
*/COUNTERS<RET>
[LNKRLC Reloc. ctr. initial value current value limit value
.LOW. 0 140 1000000
BIG 1000 6300 5400
GRAND 5400 10500 1000000]
Absolute code loaded]
/GO continues loading the program, and LINK produces the
POV warning message and the CFS fatal error message.
*/GO<RET>
%LNKPOV PSECTs BIG and GRAND overlap from address 5400 to 6300
?LNKCFS chained fixups have been suppressed
To avoid the overlap, move the PSECTs farther apart in
memory. In this example, GRAND's origin is reset from
5400 to 7000.
@LINK<RET>
*/SET:BIG:1000<RET>
*/SET:GRAND:7000<RET>
*/LRGPRO<RET>
3-33
FIELD TEST
*/LIMIT:BIG:GRAND<RET>
*BIGPRO<RET>
*/GO<RET>
@
3-34
FIELD TEST
/LINK
FORMAT /LINK:name
name is up to 6 Radix-50 compatible characters.
FUNCTION Closes an overlay link. /LINK directs LINK to give the
specified overlay link name to the current memory image
and outputs the memory image to the overlay file. LINK
first performs any required library searches and assigns a
number to the overlay link.
For a discussion of overlay structures, see Chapter 5.
The current memory image has all modules loaded since the
beginning of the load or since the last /LINK switch.
EXAMPLES @LINK<RET>
*SPEXP/LINK:ALPHA<RET>
*
Loads module SPEXP and outputs the memory image to the
overlay file as an overlay link named ALPHA.
OPTIONAL If you omit the overlay link name, LINK uses the overlay
NOTATIONS link's assigned number.
RELATED /NODE
SWITCH
3-35
FIELD TEST
/LOCALS
FORMAT /LOCALS
FUNCTION Includes local symbols from a module in the symbol table.
LINK does not need these tables, but you may want them for
debugging.
/LOCALS and /NOLOCAL may be used as either local or global
switches. If the switch is appended to a file
specification, it applies only to that file; if it is not
appended to a file specification, it applies to all the
files that follow in the same command line.
EXAMPLES @LINK<RET>
*/LOCALS ALPHA,BETA/NOLOCAL,CAPPA,/NOLOCAL DELTA<RET>
*
Loads ALPHA with local symbols, BETA without local
symbols, CAPPA with local symbols, and DELTA without local
symbols.
OPTIONAL You can abbreviate /LOCALS to /L.
NOTATIONS
RELATED /NOLOCAL, /SYMSEG
SWITCHES
3-36
FIELD TEST
/LOG
FORMAT logfilespec/LOG
FUNCTION Specifies a file specification for the log file (see
Section 4.2.2). Any LINK messages output before the /LOG
switch is encountered are not entered in the log file.
Any messages output after the LOG file is closed are also
not entered in the file.
EXAMPLES @LINK<RET>
*LOGFIL/LOG<RET>
*
Specifies the file DSK:LOGFIL.LOG in the user's directory.
@LINK<RET>
*TTY:/LOG<RET>
*
Directs log messages to the user's terminal.
OPTIONAL You can omit all or part of the logfilespec.
NOTATIONS The defaults are:
device DSK:
filename name of main program
file type LOG
directory your connected directory
You can change the defaults using the /DEFAULT switch.
RELATED /LOGLEVEL
SWITCHES
3-37
FIELD TEST
/LOGLEVEL
FORMAT /LOGLEVEL:n
n is a positive decimal integer.
FUNCTION Suppresses logging of LINK messages with message level
less than n, where n is a decimal number between 0 and 31
inclusive. You cannot suppress level 31 messages.
See Appendix B for the level of each LINK message.
The default is /LOGLEVEL:8.
EXAMPLES @LINK<RET>
*/LOGLEVEL:0<RET>
*
Logs all messages.
RELATED /ERRORLEVEL, /LOG
SWITCHES
3-38
FIELD TEST
/MAP
FORMAT mapfilespec/MAP:keyword
FUNCTION Specifies a file specification for the map output file
(see Section 4.2). The contents of the file are
determined by the /CONTENTS switch or its defaults.
Keyword Description
END Produces a map file at the end of
the load. This is the default if
you omit the keyword.
ERROR Produces a map file if a fatal
error occurs. Any modules loaded
after this switch do not appear in
the log. To ensure that a MAP file
is generated, specify this switch
before the loading of REL files.
NOW Produces a map file immediately.
Library searches are not performed
unless forced.
EXAMPLES @LINK<RET>
*MAPFIL/MAP:END<RET>
*
Generates a map in the file DSK:MAPFIL.MAP in your disk
area at the end of loading.
OPTIONAL You can omit all or part of the mapfilespec.
NOTATIONS The defaults are:
device DSK:
filename name of main program
file type MAP
directory user's connected directory
You can change the defaults using the /DEFAULT switch.
You can abbreviate /MAP to /M.
RELATED /CONTENTS
SWITCHES
3-39
FIELD TEST
/MAXNODE
FORMAT /MAXNODE:n
n is a positive decimal integer.
FUNCTION Specifies the number of overlay links to be defined when
the overlayed program requires more than 256 overlay
links. LINK allocates extra space in the OVL file for
certain fixed-length tables based on the number of overlay
links specified with this switch.
Note that this switch must be placed after the /OVERLAY
switch and it must precede the first /NODE switch in the
set of commands to LINK.
EXAMPLES @LINK<RET>
*TEST/OVERLAY/MAXNODE:500<RET>
*
Reserves space for 500 defined overlay links. See Chapter
5 for a discussion on overlays.
RELATED /OVERLAY
SWITCHES
3-40
FIELD TEST
/MESSAGE
FORMAT /MESSAGE:keyword
FUNCTION Displays messages on the terminal in the format specified
by keyword. Both /MESSAGE and /VERBOSITY display messages
on the terminal, but use different keywords and return
slightly different output.
Keyword Description
PREFIX displays the message code.
NOPREFIX prevents the display of the message
code.
FIRST displays the short message.
NOFIRST prevents the display of the short
message.
CONTINUATION displays the longer message.
NOCONTINUATION prevents the display of the longer
message.
EXAMPLES In all the following examples, an incorrect file
specification (KSD:MISTKE) is entered. In the following
example, /NOPREFIX prevents the display of the message
code.
@LINK<RET>
*/MESSAGE:NOPREFIX<RET>
*KSD:MISTKE<RET>
% Non-existent device KSD:
[ Please retype the incorrect parts of the file
specification ]
*
In the next example, a short message is returned.
@LINK<RET>
*/MESSAGE:FIRST<RET>
*KSD:MISTKE<RET>
%LNKNED Non-existent device KSD:
[ Please retype the incorrect parts of the file
specification]
In this example, a long message is returned.
3-41
FIELD TEST
@LINK<RET>
*/MESSAGE:CONTINUATION<RET>
*KSD:MISTKE<RET>
%LNKNED Non-existent device KSD:
You gave a device that does not exist on this system.
Correct your input files and reload.
[ Please retype the incorrect parts of the file specification]
RELATED /ERRORLEVEL, /VERBOSITY
SWITCHES
3-42
FIELD TEST
/MISSING
FORMAT /MISSING
FUNCTION Displays on the terminal the number of modules requested
with the /INCLUDE switch that have not yet been loaded.
EXAMPLES In the following example, MYPROG is loaded, library
searches are requested, MOD1 and MOD2 are to be loaded
during the search, and LIB1 is to be loaded.
@LINK<RET>
*MYPROG<RET>
*/SEARCH/INCLUDE:(MOD1,MOD2) LIB1<RET>
Next, /MISSING is used to check if both modules were
loaded.
*/MISSING<RET>
[LNKIMM 1 included module missing]
Now, LIB2 is loaded, MOD2 is to be included, and /MISSING
is used to check if MOD2 was loaded.
*LIB2/INCLUDE:(MOD2)<RET>
*/MISSING<RET>
[LNKIMM no included modules missing]
*
RELATED /EXCLUDE, /INCLUDE, /UNDEFINED
SWITCHES
3-43
FIELD TEST
/NEWPAGE
FORMAT /NEWPAGE:keyword
FUNCTION Sets the relocation counter to the first word of the next
page. If the counter is already at a new page, this
switch is ignored.
Keyword Description
LOW Resets the low-segment counter to new page.
If you omit the keyword, this is the
default.
HIGH Resets the high-segment counter to new
page.
EXAMPLES @LINK<RET>
*/NEWPAGE:HIGH<RET>
*SUBR1<RET>
*/NEWPAGE:LOW<RET>
*SUBR2<RET>
*
Sets the high-segment counter to a new page, loads SUBR1,
sets the low-segment counter to a new page, and loads
SUBR2. Note that SUBR1 and SUBR2 are not necessarily
loaded into the high and low segments respectively; the
/NEWPAGE switch sets a counter, but does not force the
next loaded module into the specified segment.
RELATED /SET, /COUNTERS
SWITCHES
3-44
FIELD TEST
/NODE
FORMAT /NODE:argument
FUNCTION Opens an overlay link. /NODE places LINK's relocation
counter at the end of a previously defined overlay link in
an overlay structure, which becomes the immediate ancestor
to the next overlay link defined. (For a discussion of
overlay structures, see Chapter 5.)
The /NODE switch must precede any modules to be placed in
the new overlay link.
Argument Description
name is a name given with a previous
/LINK switch. LINK places the
relocation counter at the end of
the specified overlay link.
negative number is a negative number (-n). LINK
backs up n overlay links along the
current path.
positive number is a positive number n or 0. LINK
begins further loading at the end
of overlay link number n. You can
use 0 to begin loading at the root
link.
NOTE
It is recommended that you
use an overlay link name
(or 0 for the root link)
rather than a nonzero
number. This is because a
change in commands defining
an overlay may change some
of the overlay link
numbers.
EXAMPLES In the following example, /NODE opens overlay link FATHER,
SON1 and SON2 are loaded, and /LINK closes the overlay
link.
@LINK<RET>
*/NODE:FATHER<RET>
*SON1<RET>
*SON2<RET>
*/LINK:FATHER<RET>
*
3-45
FIELD TEST
See Chapter 5 for more examples.
RELATED /LINK, /OVERLAY, /PLOT
SWITCHES
3-46
FIELD TEST
/NOENTRY
FORMAT /NOENTRY:(symbol,symbol,...)
symbol is an entry name symbol.
FUNCTION Deletes entry name symbols from LINK's overhead tables
when loading overlays, thereby saving space at run time.
If you know that execution of the current load does not
reference certain entry points, you can use /NOENTRY to
delete them.
/NOENTRY differs from /NOREQUEST in that /NOREQUEST
deletes requests for symbols, while /NOENTRY deletes
symbols that might be requested.
EXAMPLES @LINK<RET>
*/ENTRY<RET>
[LNKLSS LIBRARY SEARCH SYMBOLS (ENTRY POINTS)]
SQRT. 3456
*/NOENTRY:(SQRT.)<RET>
*/ENTRY<RET>
*
Deletes SQRT. so that it cannot be used to fulfill a
symbol request.
OPTIONAL You can omit the parentheses if only one symbol is given.
NOTATIONS
RELATED /ENTRY, /EXCLUDE, /NOEXCLUDE, /INCLUDE, /NOINCLUDE,
SWITCHES /MISSING, /REQUEST, /NOREQUEST
3-47
FIELD TEST
/NOINCLUDE
FORMAT /NOINCLUDE
FUNCTION Clears requests for modules that were specified in a
previous /INCLUDE.
EXAMPLES @LINK<RET>
*LIB1/INCLUDE:(MOD1,MOD3)<RET>
*/NOINCLUDE<RET>
*
Loads MOD1 and MOD3 from LIB1. However, if the modules
are not found immediately, LINK stops searching.
RELATED /INCLUDE, /EXCLUDE, /MISSING
SWITCHES
3-48
FIELD TEST
/NOINITIAL
FORMAT /NOINITIAL
FUNCTION Prevents loading of LINK's initial global symbol table
(JOBDAT). The /NOINITIAL switch cannot operate after the
first file specification because JOBDAT is already loaded.
The initial global symbol table contains the JBxxx symbols
described in Appendix C.
The /NOINITIAL switch is commonly used for:
o Loading LINK itself (to get the latest copy of
JOBDAT).
o Loading a private copy of JOBDAT (to alter if
necessary).
o Building an EXE file that eventually runs in executive
mode (for example, a monitor or bootstrap loader).
o Building a TOPS-20 native program which does not use a
JOBDAT area.
EXAMPLES @LINK<RET>
*/NOINITIAL<RET>
*
RELATED /NOJOBDAT
SWITCHES
3-49
FIELD TEST
/NOJOBDAT
FORMAT /NOJOBDAT
FUNCTION Prevents LINK from filling in JOBDAT's or vestigial
JOBDAT's address space and causes LINK to create a PDV.
JOBDAT contains program information such as debugger
symbol table location and version number. Vestigial
JOBDAT is used in two segment programs.
LINK creates either a JOBDAT or a PDV (Program Data
Vector) when loading a program. LINK creates a JOBDAT for
most section zero programs. LINK creates a PDV:
o for extended addressing programs.
o if a PSECT has an origin below 140.
o if you specify a PDV using /PVBLOCK or /PVDATA.
o if you specify /NOJOBDAT.
/NOJOBDAT does not suppress symbols.
EXAMPLES @LINK<RET>
*PROGAM/NOJOBDAT<RET>
*
Loads PROGAM and keeps LINK from filling in JOBDAT
information during loading.
RELATED /NOINITIAL
SWITCHES
3-50
FIELD TEST
/NOLOCAL
FORMAT /NOLOCAL
FUNCTION Suspends the effect of a preceding /LOCALS switch so that
local symbol tables are not loaded with their modules.
The /LOCALS and /NOLOCAL switches may be used as local or
global switches. If the switch is appended to a file
specification, it applies only to that file; if it is not
appended to a file specification, it applies to all the
following files in the same command.
This switch is useful if you need to conserve memory
space, because local symbols are loaded into the low
segment by default.
EXAMPLES @LINK<RET>
*/LOCALS MODA,MODB/NOLOCAL,MODC,/NOLOCAL MODD<RET>
*
Loads MODA with local symbols, MODB without local symbols,
MODC with local symbols, and MODD without local symbols.
OPTIONAL Abbreviate /NOLOCAL to /N.
NOTATIONS
RELATED /LOCALS
SWITCHES
3-51
FIELD TEST
/NOREQUEST
FORMAT /NOREQUEST:(symbol,symbol,...)
FUNCTION Deletes references to overlay links from LINK's overhead
tables when loading overlay programs. If you know that
the execution of the current load does not require certain
overlay links, you can use /NOREQUEST to delete references
to them.
/NOREQUEST differs from /NOENTRY in that /NOENTRY deletes
symbols that might be requested, while /NOREQUEST deletes
the requests for them.
EXAMPLES @LINK<RET>
*/REQUEST<RET>
[LNKRER REQUEST EXTERNAL REFERENCES]
ROUTN.
SQRT.
*/NOREQUEST:(ROUTN.,SQRT.)
*/REQUEST
*
Deletes references to ROUTN. and SQRT.
OPTIONAL You can omit the parentheses if only one symbol is given.
NOTATIONS
RELATED /NOENTRY
SWITCH
3-52
FIELD TEST
/NOSEARCH
FORMAT /NOSEARCH
FUNCTION Suspends the effect of a previous /SEARCH switch. Files
named between a /SEARCH and the next /NOSEARCH are
searched as libraries, so that modules are loaded only to
resolve global references.
The /SEARCH and /NOSEARCH switches may be used either
locally or globally. If the switch is appended to a file
specification, it applies only to that file; if it is not
appended to a file specification, it applies to all
following files in the same command string.
EXAMPLES @LINK<RET>
*/SEARCH ALPHA,BETA/NOSEARCH,CAPPA,/NOSEARCH DELTA<RET>
*
Searches ALPHA, loads BETA, searches CAPPA, and loads
DELTA.
RELATED /SEARCH
SWITCHES
3-53
FIELD TEST
/NOSTART
FORMAT /NOSTART
FUNCTION Directs LINK to disregard any start addresses found after
the /NOSTART switch. Normally LINK keeps the most recent
start address found, overwriting any previously found.
The /NOSTART switch prevents this replacement.
EXAMPLES @LINK<RET>
MAIN1,/NOSTART MAIN2,MAIN3<RET>
*
Directs LINK to save the start address from MAIN1 instead
of replacing it with other start addresses from MAIN2 and
MAIN3.
RELATED /START
SWITCHES
3-54
FIELD TEST
/NOSYMBOL
FORMAT /NOSYMBOL
FUNCTION Prevents construction of user symbol tables. Symbols are
then not available for the map file, but the header for
the file can still be generated by the /MAP switch.
The /NOSYMBOL switch prevents writing an ALGOL SYM file if
it would otherwise have been written.
If you do not need the map file or symbols, you can free
some memory space using the /NOSYMBOL switch.
EXAMPLES @LINK<RET>
*/NOSYMBOL<RET>
*
3-55
FIELD TEST
/NOSYSLIB
FORMAT /NOSYSLIB:(keyword,...,keyword)
FUNCTION Prevents automatic search of the system libraries named as
keywords. LINK searches system libraries at the end of
loading to satisfy unresolved global references. The
/NOSYSLIB switch prevents this search.
The /NOSYSLIB switch can also be used to terminate
searching of libraries that were specified in a previous
/SYSLIB switch. When you specify searching of a library
with /SYSLIB, that library continues to be searched for
every module you load. You can use /NOSYSLIB to specify
libraries that should not be searched. Refer to /SYSLIB
for more information.
The keywords and the libraries they specify are listed
below. Only those shown in bold specify libraries
supported by DIGITAL.
Keyword Library
ANY all
ALGOL ALGLIB
BCPL BCPLIB
COBOL LIBOL or C74LIB
FORTRAN FORLIB
NELIAC LIBNEL
PASCAL PASLIB
SAIL SAILIB
SIMULA SIMLIB
EXAMPLES @LINK<RET>
*/NOSYSLIB:(ALGOL,COBOL)<RET>
*
Prevents search of the system libraries ALGLIB and LIBOL.
OPTIONAL If you omit keyword it defaults to ANY.
NOTATIONS You can omit parentheses if only one keyword is given.
RELATED /SYSLIB
SWITCHES
3-56
FIELD TEST
/NOUSERLIB
FORMAT /NOUSERLIB
FUNCTION Discontinues automatic searching of the user libraries at
each /LINK or /GO switch. You can specify which library
to stop searching by appending /USERLIB to a file
specification. For example,
*MYLIB/NOUSERLIB
stops searching MYLIB.
For programs that use overlays, if you need a file
searched for some overlay links but not others, you can
use the /USERLIB and /NOUSERLIB switches to enable and
disable automatic searching of the file.
EXAMPLES @LINK<RET>
*FILE/OVERLAY<RET>
*FILE4/USERLIB:FORTRAN<RET>
*LINKA/LINK:ROOT<RET>
*/NODE:ROOT<RET>
*LINKB/LINK:ONE<RET>
*/NODE:ROOT<RET>
*LINKA,LINKB/NOUSERLIB<RET>
*
OPTIONAL If you omit the filespec, LINK discontinues search of all
NOTATIONS user libraries.
RELATED /USERLIB
SWITCHES
3-57
FIELD TEST
/ONLY
FORMAT /ONLY:keyword
FUNCTION Directs LINK to load the specified segments of two-segment
modules. The keywords are:
Keyword Segments
HIGH high
LOW low
BOTH both
If you are loading a module using /ONLY:HIGH or /ONLY:LOW,
it cannot contain chained fixups that cross segments.
/ONLY and /SEGMENT are incompatible switches.
EXAMPLES @LINK<RET>
*/ONLY:HIGH MOD1,MOD2<RET>
*MOD3/ONLY:BOTH<RET>
*
Loads high segment for MOD1 and MOD2; loads both segments
for MOD3.
RELATED /SEGMENT
SWITCHES
3-58
FIELD TEST
/OPTION
FORMAT /OPTION:name
FUNCTION Reads the SWITCH.INI file to determine the specified
switch defaults for LINK. /NOOPTION ignores the
SWITCH.INI file.
A SWITCH.INI contains switches for LINK to use, and allows
you to override system defaults. For example, the
following line when placed in SWITCH.INI changes the
default for input files from DSK: to MYSTR:.
LINK/DEFAULT:INPUT MYSTR:
The name argument specifies which line to read in
SWITCH.INI file. For example,
/OPTION:LINK
reads the LINK line in the SWITCH.INI file.
EXAMPLES: @LINK<RET>
*/OPTION:LINK<RET>
*INPUT<RET>
*
3-59
FIELD TEST
/OTSEGMENT
FORMAT /OTSEGMENT:keyword
FUNCTION Specifies the time and manner of loading the object-time
system. An object-time system is a collection of modules
that is called by compiled code for a particular language
in order to perform various utility functions such as I/O
and trap-handling. For example, FOROTS is the object-time
system for FORTRAN. For more information on object-time
systems for the language you are using, refer to the
appropriate language manual.
Keyword Description
DEFAULT Suspends the effect of a previous
/OTSEGMENT:SHARABLE or
/OTSEGMENT:NONSHARABLE switch.
NONSHARABLE Loads the object-time system into the
executable program at load time. Both
the user program and the object-time
system may have code in both the high
and low segments.
SHARABLE Binds the object-time system at
execution time. The user program is in
the low segment and the object-time
system is in the high segment.
LINK's default action is to bind the object-time system at
execution time, except in the following cases. If any of
the following cases occur, a non-sharable object-time
system is loaded as part of the program.
o You specify /OTSEGMENT:NONSHARABLE.
o You have loaded any code into the high segment.
o You have specified /SYMSEG:HIGH.
o Your low segment is too big for the sharable
object-time systems to fit, and your program is loaded
in section 0.
EXAMPLES @LINK<RET>
*MYPROG/SYSLIB/OTSEGMENT:NONSHAR<RET>
*
Loads a non-sharable copy of the object-time system as
part of your program.
3-60
FIELD TEST
RELATED /SEGMENT
SWITCHES
3-61
FIELD TEST
/OVERLAY
FORMAT /OVERLAY:(keyword,...,keyword)
FUNCTION Initiates construction of an overlay structure. For a
discussion of overlay structures, see Chapter 5.
You can append the /OVERLAY switch to a file specification
in the form:
filespec/OVERLAY:(keyword,...keyword)
The keywords and their meanings are listed below. The
default settings are shown in bold
Keyword Description
ABSOLUTE Specifies that links are absolute.
This is the default situation when
overlays are loaded. The inverse
situation is to use
/OVERLAY:RELOCATABLE. Relocatable
overlays are described in Chapter
5.
LOGFILE Outputs runtime overlay messages to
your terminal.
NOLOGFILE Suppresses output of runtime
overlay messages.
NOWARNING Suppresses overlay warning
messages.
RELOCATABLE Specifies that links are
relocatable.
TREE Specifies that the overlay has a
tree structure.
WARNING Outputs overlay warning messages to
user terminal.
WRITABLE Specifies that the links are
writable. Refer to Chapter 5 for
more information.
EXAMPLES @LINK<RET>
*/OVERLAY:(TREE,LOGFILE)
Specifies that an overlay is to be constructed, and that
3-62
FIELD TEST
runtime messages are to be printed on the terminal.
OPTIONAL You can omit the parentheses if only one keyword is given.
NOTATIONS
RELATED /LINK, /NODE, /PLOT
SWITCHES
3-63
FIELD TEST
/PATCHSIZE
FORMAT /PATCHSIZE:n
n is a positive decimal integer.
FUNCTION Allocates n words of storage to precede the symbol table.
The allocated storage is in the same segment (high or low)
as the symbol table. The default is /PATCHSIZE:64.
The storage allocated is available for patching or for
defining new symbols with DDT, and is identified by the
global symbol "PAT.."
EXAMPLES @LINK<RET>
*/SYMSEG:HIGH/PATCHSIZE:200<RET>
*
Loads the symbol table in the high segment after
allocating 200 words between the last loaded module and
the symbol table.
OPTIONAL You can specify the patchsize in octal.
NOTATIONS
RELATED /SYMSEG
SWITCHES
3-64
FIELD TEST
/PLOT
FORMAT filespec/PLOT
FUNCTION Directs LINK to output a tree diagram of your overlay
structure. You can have the diagram formatted for a
plotter (by default) or for a line printer (by giving the
device as LPT:).
Each box in the diagram shows a link number, its name (if
you gave one with the /LINK switch), and its relationship
to other links (as defined by your commands).
The /PLOT switch cannot precede the /OVERLAY switch.
EXAMPLES @LINK<RET>
*OVLPRO/OVERLAY:(TREE,LOGFILE)<RET>
*LPT:OVLPRO/PLOT<RET>
*
requests a tree diagram for the overlay. The diagram is
formatted for the line printer because LPT: is specified.
OPTIONAL LINK has default settings for the size of the overlay
NOTATIONS diagram and the increment for drawing lines. You can
override these by giving the /PLOT switch in the form:
filespec/PLOT:(LEAVES:value,INCHES:value,STEPS:value)
where the values for each parameter define:
INCHES Width of diagram in inches. The defaults are
INCHES:29 for plotter and INCHES:12 for line
printer.
LEAVES Number of links without successors that can
appear in one row. The defaults are
LEAVES:16 for plotter and LEAVES:8 for line
printer.
STEPS Increments per inch for drawing lines. The
defaults are STEPS:100 for plotter and
STEPS:20 for line printer.
For line printer diagrams, you cannot give INCHES or
LEAVES different from the defaults. The STEPS parameter
should be between 10 and 25.
For plotter diagrams, you should give INCHES and LEAVES in
a ratio of about 2 to 1. For example, INCHES:40 and
LEAVES:20.
3-65
FIELD TEST
If LINK cannot design the diagram on one page, it
automatically designs subtrees for diagrams on more pages.
RELATED /LINK, /NODE, /OVERLAY
SWITCHES
3-66
FIELD TEST
/PLTTYP
FORMAT /PLTTYP:keyword
FUNCTION Specifies the type of plot file to be generated by the
/PLOT switch.
KEYWORDS DEFAULT Generate output for a printer only if the device
is a printer or terminal.
PLOTTER Generate output for a plotter.
PRINTER Generate output for a printer.
EXAMPLES @LINK
*OVLPRO/OVERLAY
*OVLPRO/PLOT/PLTTYP:PRINTER
Causes all output from the /PLOT switch to be in line
printer format.
RELATED /PLOT
SWITCHES
3-67
FIELD TEST
/PSCOMMON
FORMAT /PSCOMMON:psect:common
FUNCTION Specifies where LINK is to load COMMON blocks. This
switch causes the FORTRAN common specified by the argument
common to be loaded into the PSECT specified in the
argument psect. Use the /PSCOMMON switch before loading
the specified common and before declaring the common's
size with the /COMMON switch.
/PSCOMMON only affects common blocks defined with the
/COMMON switch. If the common block is created by a REL
block, /PSCOMMON is ignored, and the PSECT specified by
the REL file is used.
EXAMPLES In the following example, /SET defines the SECTA PSECT's
origin, /PSCOMMON specifies that SECTA is loaded into
COMABC, and /COMMON defines the common size.
@LINK<RET>
*/SET:SECTA:3000000
*/PSCOMMON:SECTA:COMABC
*/COMMON:COMABC:10000
*PROG
*
RELATED /COMMON
SWITCHES
3-68
FIELD TEST
/PVBLOCK
FORMAT /PVBLOCK:keyword
FUNCTION Requests a PDV (Program Data Vector) from LINK and gives
you control over where the vector goes.
NOTE
PDVs are useful only with TOPS-20 Version 5 and
later monitors. Earlier monitors ignore PDVs.
Refer to Chapter 7 for more information about
PDVs.
Keyword Description
DEFAULT disables the previous /PVBLOCK:HIGH
or /PVBLOCK:LOW, and restores the
default.
LINK supplies a PDV by default:
o for an extended addressing
program,
o if /PVDATA is specified
o if /NOJOBDAT is specified.
o if a PSECT's start address is
below 140.
The default PDV is placed at the
end of the low segment even for
programs that use PSECTs.
HIGH places the PDV at the end of the
high segment.
LOW places the PDV at the end of the
low segment.
NONE prevents loading of the PDV.
PSECT:name places the PDV at the end of the
named PSECT.
EXAMPLES In the following example, PVTEST is loaded and a PDV is
requested. The PDV is placed at the end of the low
segment as indicated by the LOW keyword.
@LINK<RET>
3-69
FIELD TEST
*PVTEST<RET>
*/PVBLOCK:LOW/GO<RET>
Next, the TOPS-20 INFORMATION VERSION command is used to
display the program and PDV name.
@INFO VERSION<RET>
.
.
.
Program is PVBLCK
PDVs: Program name PVTEST, version
@
OPTIONAL If you specify /PVBLOCK with no keyword, DEFAULT is the
NOTATION default.
RELATED /PVDATA, /NOJOBDAT
SWITCHES
3-70
FIELD TEST
/PVDATA
FORMAT /PVDATA:keyword:value
FUNCTION Changes the contents of a PDV block. The /PVDATA switch
also allocates storage for the PDV. If the storage to be
allocated conflicts with any PSECT, LINK issues a message
with the severity level of 16, and does not write the PDV
information into the executable program or sharable save
(EXE) file. Refer to Chapter 7 for more information on
PDVs.
NOTE
PDVs are useful only with TOPS-20 Version 5 and
later monitors. Earlier monitors ignore PDVs.
Refer to Chapter 7 for more information about
PDVs.
Keyword Description
NAME is an optionally quoted ASCII
string that identifies the PDV.
This string is not converted to
uppercase. If you do not quote the
program name, you can use only
alphanumerics and the period (.),
dollar sign ($), and percent sign
(%) characters to specify program
name.
VERSION is a hash sign (#) followed by an
octal value, a version number that
starts with a number, or a global
symbol.
MEMORY is the address of the user supplied
memory map in octal, or a
previously defined global symbol
that contains the address of a user
supplied memory map. The user
supplied memory map suppresses the
map generated by LINK.
PROGRAM is the address of a
program-specific data block in
octal value, or a previously
defined global symbol that contains
the address of a program-specific
data block.
3-71
FIELD TEST
CBLOCK is the address of a
customer-defined block in octal, or
a previously defined global symbol
that contains the address of a
customer-defined block.
EXPORT is the address of a block of
information defined by a program in
octal, or a previously defined
global symbol that contains the
address of a block of information
defined by a program.
EXAMPLES In the following example, a PDV name and version number
are defined. /PVDATA:NAME:T.S$T% defines the T.S$T% as
the PDV name, and /PVDATA:VERSION:1A(3) defines 1.1(3) as
the version number.
@LINK<RET>
*/PVDATA:NAME:T.S$T%<RET>
*/PVDATA:VERSION:1A(3)<RET>
*PVTEST/GO<RET>
Next, the TOPS-20 INFORMATION VERSION command is used to
display the PDV name and version.
@INFO VERSION<RET>
.
.
.
Program is PVBLCK
PDVs: Program name T.S$T%, version 1.1(3)
@
The next example also specifies a PDV name and version,
but in a different format.
@LINK<RET>
*/PVDATA:NAME:"TST@ "<RET>
*/PVDATA:VERSION: 101000003<RET>
*PVTEST/GO<RET>
@INFO VERSION<RET>
.
.
.
Program is PVBLCK
PDVs: Program name TST@ , version 1.1(3)
@
RELATED /PVBLOCK
3-72
FIELD TEST
SWITCHES
3-73
FIELD TEST
/REDIRECT
FORMAT /REDIRECT:Lowpsect:Highpsect
FUNCTION Loads two-segmented formatted REL files as part of a
program using PSECTs. The argument Lowpsect is the name
of the PSECT to receive the low segment code and Highpsect
is the name of the PSECT to receive the high segment code.
You must redirect both segments, you cannot redirect just
the high or the low segment.
EXAMPLES The following example loads a two-segment program
(TWOPRT), and displays the low- and high-segment values
using /COUNTERS.
@LINK<RET>
*TWOPRT<RET>
*/COUNTERS<RET>
[LNKRLC Reloc. ctr. initial value current value limit
value
.LOW. 0 1642 1000000
.HIGH. 400000 400753 1000000]
*
Next, PSECT origins are set for PSHI and PSLO, .LOW. is
redirected into PSLO, .HIGH. is redirected into PSHI, and
/COUNTERS is used to display PSHI and PSLO values.
@LINK<RET>
*/SET:PSHI:400010<RET>
*/SET:PSLO:3500<RET>
*/REDIRECT:PSLO:PSHI<RET>
*TWOSEG<RET>
*/COUNTERS<RET>
[LNKRLC Reloc. ctr. initial value current value limit
value
PSHI 400010 400753 1000000
PSLO 3500 5202 1000000]
3-74
FIELD TEST
/REQUEST
FORMAT /REQUEST
FUNCTION Displays external global symbol references on the
terminal. Use /REQUEST to determine if a module uses a
global symbol that is loaded in another module.
If you use /REQUEST to get the names of external
references, you can then either delete the references with
the /NOREQUEST switch, or load the referenced modules.
EXAMPLES @LINK<RET>
*/REQUEST<RET>
[LNKRER REQUEST EXTERNAL REFERENCES]
ROUTN.
SQRT.
*/NOREQUEST:ROUTN.<RET>
*/SEARCH LIB1<RET>
*
Obtains the external references ROUTN. and SQRT.; deletes
the request for ROUTN.; searches the file LIB1 for a
module containing the entry point SQRT.
RELATED /NOREQUEST
SWITCHES
3-75
FIELD TEST
/REQUIRE
FORMAT /REQUIRE:(symbol,...,symbol)
FUNCTION Generates global requests for the specified symbols. LINK
uses these symbols as library search symbols (entry
points).
/REQUIRE differs from /INCLUDE in that /INCLUDE requests a
module by name, while /REQUIRE requests an entry name
symbol. Thus you can use /REQUIRE to specify a function
(for example, SQRT.) even if you do not know the module
name.
You can use /REQUIRE to load a module into an overlay link
common to all overlay links that reference the module.
Note that the global requests generated by the /REQUIRE
switch do not use the standard FORTRAN calling sequence,
and are therefore not visible to the /REQUEST switch.
EXAMPLES @LINK<RET>
*/UNDEFINED<RET>
[LNKUGS NO UNDEFINED GLOBAL SYMBOLS]
*/REQUIRE:(ROUTN.,SQRT.)<RET>
*/UNDEFINED<RET>
[LNKUGS 2 UNDEFINED GLOBAL SYMBOLS]
SQRT. 0
ROUTN. 0]
*
OPTIONAL You can omit the parentheses if only one symbol is given.
NOTATIONS
RELATED /SEARCH, /NOSEARCH
SWITCHES
3-76
FIELD TEST
/RUN
FORMAT /RUN:file
FUNCTION Directs LINK to run the program after it is loaded. LINK
ignores this switch if you used /EXECUTE or /DEBUG.
EXAMPLES: @LINK<RET>
*/RUN:MYPROG<RET>
*MYPROG<RET>
*
3-77
FIELD TEST
/RUNAME
FORMAT /RUNAME:name
FUNCTION Assigns a job name for execution of your program. This
job name is used in the SYSTAT display.
EXAMPLES @LINK<RET>
*/RUNAME:LNKDEV<RET>
*PROGRM
*
Assigns the name LNKDEV for program execution regardless
of the file names that are loaded.
3-78
FIELD TEST
/RUNOFFSET
FORMAT /RUNOFFSET:n
FUNCTION Runs the program specified in a /RUN switch with an offset
of n. If you omit the switch the default is 0. If you
omit the n argument, the default is 1.
EXAMPLES In the following example, MACRO is run with an offset of 1
after MCPROG is loaded.
@LINK<RET>
*MCPROG<RET>
*/RUN:MACRO/RUNOFFSET:1/GO<RET>
3-79
FIELD TEST
/SAVE
FORMAT filespec/SAVE
FUNCTION Directs LINK to create a sharable save file using the
specified file specification. If you omit the file type,
LINK uses EXE.
To run the sharable save file with the TOPS-20 RUN
command, the file must have an EXE file type.
EXAMPLES @LINK<RET>
*MYPROG<RET>
*DSKZ:RUNPRO.EXE/SAVE/GO<RET>
@
Directs LINK to save the loaded version of MYPROG as
RUNPRO.EXE on DSKZ:.
3-80
FIELD TEST
/SEARCH
FORMAT /SEARCH
FUNCTION Directs LINK to search the input files and load only the
modules whose entry point names resolves a global symbol
request.
The /NOSEARCH discontinues library searching of input
files, and loads all the modules in the input files.
Although /NOSEARCH is used, LINK continues to search
system libraries unless you used the /NOSYSLIB, and
searches user libraries if you used the /USERLIB switch.
For example, the FILE9, FILE8, and FILE7 input files are
to be searched as libraries and only the modules that
resolve global symbols are to be loaded, but all the
modules in FILE1, FILE2, and FILE3 are to be loaded.
@LINK<RET>
*/SEARCH FILE9,FILE8,FILE7<RET>
*/NOSEARCH FILE1,FILE2,FILE3<RET>
The /SEARCH and /NOSEARCH switches can be used as local or
global switches. See Section 3.3.3 for a discussion of
switch placement.
Note that search requests in .TEXT blocks may be processed
in the reverse order of entered /SEARCH switches. Keep
this in mind when specifying the order in which modules
are to be searched. See Block types greater than 3777 for
more information.
EXAMPLES @LINK<RET>
*/SEARCH MODA,MODB/NOSEARCH,MODC,/NOSEARCH MODD<RET>
*
Searches MODA, loads MODB, searches MODC, and loads MODD.
This examples illustrates the use of /SEARCH and /NOSEARCH
as local and global switches. The /SEARCH switch is a
global switch because it is entered after the asterisk,
and applies to MODA and MOBC. The first /NOSEARCH switch
is a local switch because it is appended to MODB. Notice
its placement before the comma. The second /NOSEARCH
switch is a global switch and applies to the remainder of
the line.
OPTIONAL You can abbreviate /SEARCH to /S.
NOTATIONS
RELATED /NOSEARCH
3-81
FIELD TEST
SWITCHES
3-82
FIELD TEST
/SEGMENT
FORMAT /SEGMENT:keyword
FUNCTION Specifies whether the high segment or the low segment of a
two-segment program is to be used for loading the modules
that follow. If /SEGMENT is omitted, LINK follows the
specifications in the program.
LINK does not follow the programs specifications for
FORTRAN object code, but loads both segments into the low
segment. LINK does not load FORTRAN object code into the
low segment if:
o You used the /OTSEGMENT:NONSHARABLE switch.
o You used the /SEGMENT:HIGH switch to load code into
the high segment.
o You used the /SEGMENT:DEFAULT switch to load code into
both segments.
o A high segment already exists.
The keywords for the /SEGMENT switch are:
Keyword Description
DEFAULT Suspends effect of /SEGMENT:LOW or
/SEGMENT:HIGH.
HIGH Loads into high segment.
LOW Loads into low segment.
NONE Same as DEFAULT.
If the switch is appended to a file specification, it
applies only to that file; if it is not appended to a file
specification, it applies to all following files in the
same command string.
/SEGMENT and /ONLY are incompatible.
EXAMPLES @LINK<RET>
*/SEGMENT:LOW MOD1,MOD2,/SEGMENT:HIGH MOD3<RET>
*
Loads MOD1 and MOD2 into the low segment; loads MOD3 into
the high segment.
RELATED /OTSEGMENT
3-83
FIELD TEST
SWITCHES
3-84
FIELD TEST
/SET
FORMAT /SET:name:address
name is .HIGH., .LOW., or a PSECT name.
address is a thirty-bit octal address or a defined symbol.
FUNCTION Sets the origin of a PSECT, or sets the .HIGH. or .LOW.
relocation counter.
For setting the origin of a PSECT, name is the name of the
PSECT, and address is a virtual memory address. The /SET
switch must precede the modules that make up the specified
PSECT. The /SET switch is not needed if the REL files
already contain the PSECT origin information.
EXAMPLES @LINK<RET>
*/SET:MEMLOCA:200000<RET>
*
Specifies that the PSECT named MEMLOCA is to be loaded
with an origin of address 200000.
*/SET:.HIGH.:400000<RET>
*
Sets the high segment relocation counter .HIGH. to the
address 400000. Note that saying /SET:.HIGH. causes a
high segment to appear and a vestigial JOBDAT area to be
built.
RELATED /COUNTERS, /LIMIT
SWITCHES
3-85
FIELD TEST
/SEVERITY
FORMAT /SEVERITY:n
n is a positive decimal integer.
FUNCTION Specifies that errors of severity level greater than or
equal to n terminate the load, where n is a decimal number
between 0 and 31 inclusive. Level 31 errors always
terminate the load.
The defaults are /SEVERITY:28 for timesharing jobs, and
/SEVERITY:16 for batch jobs. See Appendix B for a list of
severity codes.
EXAMPLES @LINK<RET>
*/SEVERITY:29<RET>
*
Specifies that level 29 errors and above are fatal.
3-86
FIELD TEST
/SPACE
FORMAT /SPACE:n
n is a positive decimal integer.
FUNCTION Specifies that n words of memory follow the current
overlay link at execution time. This memory allocation
does not increase the size of the overlay file, but it
does increase the size of the program at run time.
The /SPACE switch is used to allocate space for use by the
object time system. The OTS uses this space for I/O
buffers, and as scratch space in FORTRAN and heap space in
ALGOL.
You should place the /SPACE switch before the first /LINK
switch, to ensure allocation for the root link. It is
possible to allocate space after one or more overlays are
linked. This might be useful if an overlay has unusual
storage requirements: such as buffers for a file which is
open only while that overlay is resident, or a large local
matrix.
To allocate space between overlays, use /SPACE when
loading the overlay that use this file or matrix. LINK
allows one /SPACE switch for the root node, and one for
each overlay.
The default amount of memory allocated, if you do not
specify /SPACE, is 2000 for the root link and 0 (zero) for
other overlay links.
If the space allocated for a relocatable link is too
small, the overlay handler can relocate it. If the space
allocated for an absolute link is too small, a fatal error
occurs.
EXAMPLES @LINK<RET>
*/OVERLAY<RET>
*TEST/SPACE:90/LINK:MAIN<RET>
Allocates 90 words of memory to follow the root link for
the program. See Chapter 5 for a discussion on overlay.
OPTIONAL You can specify the number of words in octal.
NOTATIONS
3-87
FIELD TEST
/START
FORMAT /START:symbol
/START:address
/START
symbol is thirty-bit octal address or a symbol.
FUNCTION Specifies the start address for the loaded program, and
prevents replacement by any start addresses found in later
modules.
The /START switch without an argument disable a previously
given /NOSTART switch.
LINK starts a program using a TOPS-10 style entry vector
if the entry length vector is zero.
LINK starts a program using a TOPS-20 style entry vector:
o if the program contains nonzero sections and the
length of the entry vector is not specified
o if the length of the entry vector is 1
o if /NOJOBDAT equals 1
See the TOPS-20 Monitor Calls Reference Manual for more
information on entry vectors.
EXAMPLES @LINK<RET>
*MAIN1/START:ENTRY1,MAIN2,MAIN3<RET>
*
Defines the start address as ENTRY1 in MAIN1, and prevents
replacement of this start address by any others found in
MAIN2 or MAIN3.
OPTIONAL You can specify the start address in octal.
NOTATIONS
RELATED /NOSTART
SWITCHES
3-88
FIELD TEST
/SUPPRESS
FORMAT /SUPPRESS:symbol
Where symbol is a previously defined global symbol.
FUNCTION Suppresses a previously defined global symbol. If the
symbol is unknown, this switch has no effect. Use this
switch if the same global symbol is defined in two modules
and you wish to suppress one of the definitions.
LINK suppresses a defined global symbol by setting its
definition to undefined in the global symbol table. LINK
does not remove the symbol definition from the symbol
table. As a result, the symbol table built for debugging
contains both the old and new values of the symbol.
Since LINK sets the symbol to undefined in the symbol
table, it expects that a subsequent module will be loaded
that contains a global definition for the symbol. If the
symbol is not defined later, LINK issues the Undefined
Global Symbol (LNKUGS) error.
EXAMPLES In the following example, the ENTPTR symbol is used in
both the TEST and TEST2 programs. First, TEST is loaded,
and the value of ENTPTR is show using the /VALUE switch.
@LINK<RET>
*TEST<RET>
*/VALUE:ENTPTR<RET>
[LNKVAL Symbol ENTPTR 140 defined]
Next, ENTPTR's value is suppressed using /SUPPRESS and its
current value is shown. Note that the value is now
undefined.
*/SUPPRESS:ENTPTR<RET>
*/VALUE:ENTPTR<RET>
[LNKVAL Symbol ENTPTR 0 undefined]
Finally TEST2 is loaded and the value is shown again.
*TEST2<RET>
*/VALUE:ENTPTR<RET>
[LNKVAL Symbol ENTPTR 200 defined]
In the next example, TEST and TEST2 are loaded, but ENTPTR
is not suppressed after TEST is loaded. In this example,
LINK issues the Multiply-defined global symbol warning.
@LINK<RET>
TEST<RET>
3-89
FIELD TEST
TEST2<RET>
%LNKMDS Multiply-defined global symbol ENTPTR
Detected in module .MAIN from file TEST2.REL.1
Defined value = 140, this value = 200
*
3-90
FIELD TEST
/SYFILE
FORMAT filespec/SYFILE:keyword
FUNCTION Requests LINK to output a symbol file to the given file
specification, and sets the /SYMSEG:DEFAULT switch. If
you previously specified /NOSYM, the /SYFILE switch has no
effect.
The symbol file contains global symbols sorted for DDT,
and has a SYM file type. If you used the /LOCALS switch,
this file also contains local symbols, module names, and
module lengths.
Keyword Description
ALGOL Requests symbols in ALGOL's format.
RADIX-50 Requests symbols in Radix-50
format.
TRIPLET Requests symbols in triplet format.
See Chapter 4 for symbol table formats.
EXAMPLES @LINK<RET>
*SYMBOL/SYFILE<RET>
*
Creates a symbol file called SYMBOL with the symbols in
Radix-50 format and a file type of .SYM.
OPTIONAL If you omit the keyword, Radix-50 is assumed.
NOTATIONS
3-91
FIELD TEST
/SYMSEG
FORMAT /SYMSEG:keyword
FUNCTION Allows you to specify where the symbol table is to be
placed.
Keyword Description
DEFAULT Places the symbol table in the low
segment, except for overlayed
programs, in which case symbols are
not loaded by default.
HIGH Places the symbol table in the high
segment.
LOW Places the symbol table in the low
segment.
NONE Prevents loading of the symbol
table.
PSECT:name Places the symbol table at the end
of the PSECT after allocating the
patch space.
EXAMPLES @LINK<RET>
*/SYMSEG:LOW<RET>
*
Places the symbol table in the program low segment.
RELATED /LOCALS, /NOLOCALS
SWITCHES
3-92
FIELD TEST
/SYSLIB
FORMAT /SYSLIB:keyword
FUNCTION Forces searching of one or more system libraries
immediately after you enter a carriage-return ending the
command.
LINK automatically searches a system library, by default,
if code from the corresponding translator has been loaded.
This search is performed after all the modules of the
program are loaded. /SYSLIB forces this search to take
place immediately rather than after all the program
modules are loaded.
After you specify a library with /SYSLIB, the library you
specified is searched every time you load a module, until
you use /NOSYSLIB to end searching of that library.
The keywords and the libraries they specify are listed
below. Keywords shown in bold indicate libraries
supported by DIGITAL.
Keyword Library
ANY all
ALGOL ALGLIB
BCPL BCPLIB
COBOL LIBOL or C74LIB
FORTRAN FORLIB
NELIAC LIBNEL
PASCAL PASLIB
SAIL SAILIB
SIMULA SIMLIB
EXAMPLES @LINK<RET>
*MAIN/SYSLIB:COBOL<RET>
*SUB1<RET>
*SUB2/NOSYSLIB<RET>
*
OPTIONAL You can omit the keyword. LINK searches all libraries
NOTATIONS for which corresponding code has been loaded.
RELATED /NOSYSLIB
SWITCHES
3-93
FIELD TEST
/TEST
FORMAT /TEST:keyword
FUNCTION Loads the debugger indicated by keyword. Unlike the
/DEBUG switch, /TEST causes execution to begin in the
loaded program (not in the debugging module). This switch
is useful if you expect the program to run successfully,
but want the debugger available in case the program has
errors.
The /TEST switch turns on the /LOCALS switch for the
remainder of the load. You can override this by using the
/NOLOCAL switch, but the override lasts only during
processing of the current command.
Local symbols for the debugging module itself are never
loaded.
The keywords and the programs they load are listed below.
Only those shown in bold are supported by DIGITAL.
Keyword Debugger
ALGDDT ALGDDT
ALGOL ALGDDT
COBDDT COBDDT
COBOL COBDDT
DDT DDT
FAIL SDDT (SAIL debugger)
FORDDT FORDDT
FORTRAN FORDDT
MACRO DDT
PASCAL PASDDT
PASDDT PASDDT
SAIL SDDT (SAIL debugger)
SDDT SDDT (SAIL debugger)
SIMDDT SIMDDT
SIMULA SIMDDT
EXAMPLES @LINK<RET>
*MYPROG/TEST:FORTRAN<RET>
*
Loads MYPROG and FORDDT.
OPTIONAL If you give no keyword with /TEST, the default is either
NOTATIONS DDT or the debugger specified by the /DDEBUG switch.
You can abbreviate /TEST to /T.
RELATED /DDEBUG, /DEBUG
3-94
FIELD TEST
SWITCHES
3-95
FIELD TEST
/UNDEFINED
FORMAT /UNDEFINED
FUNCTION Displays undefined global symbols on the terminal. An
undefined global symbol is one that LINK has not yet
resolved. You can use /UNDEFINED to get a list of
undefined symbols, and then define them with the /DEFINE
switch.
EXAMPLES @LINK<RET>
*/UNDEFINED<RET>
[LNKUGS 2 undefined global symbols]
ALPHA 400123
IOTA 402017
*/DEFINE:(ALPHA:591,IOTA:1)<RET>
*
Displays two undefined global symbols (ALPHA and IOTA),
and defines a decimal value of 591 to ALPHA and a decimal
value of 1 to IOTA.
OPTIONAL You can abbreviate /UNDEFINE to /U.
NOTATIONS
RELATED /DEFINE, /VALUE
SWITCHES
3-96
FIELD TEST
/UPTO
FORMAT /UPTO:address
address is a thirty-bit octal address or a defined symbol.
FUNCTION Sets an upper limit to which the symbol table can expand.
EXAMPLES @LINK<RET>
*/UPTO:550000<RET>
*FORPRO<RET>
Defines a 550,000 upper limit for the FORTRAN symbol
table. This switch overrides the default upper bound for
the FORTOS symbol table. This might be used if FOROTS
begins above 400000.
RELATED /SYMSEG
SWITCH
3-97
FIELD TEST
/USERLIB
FORMAT filespec/USERLIB:(keyword,...,keyword)
FUNCTION Directs LINK to search the user library given in the file
specification before searching system libraries. The
keyword indicates that the given library is to be searched
only if code from the corresponding translator was loaded.
Keywords and their meanings are given below. Only those
shown in bold are supported by DIGITAL.
Keyword Library
ALGOL ALGOL
ANY This library
BCPL BCPL
COBOL COBOL
FORTRAN FORTRAN
NELIAC NELIAC
PASCAL PASCAL
SAIL SAIL
SIMULA SIMULA
EXAMPLES @LINK<RET>
*MYFORL/USERLIB:FORTRAN<RET>
*
Directs LINK to search the user library MYFORL (before
searching FORLIB) if any FORTRAN-compiled code is loaded.
OPTIONAL You can omit the parentheses if only one keyword is given.
NOTATIONS
RELATED /NOUSERLIB, /SYSLIB
SWITCHES
3-98
FIELD TEST
/VALUE
FORMAT /VALUE:(symbol,symbol,...)
FUNCTION Displays the values of each specified global symbol on the
terminal. LINK issues the LNKVAL message, giving the
symbol, the symbol's current value, and symbol's status.
Status is one of the following:
defined The symbol and its value are known.
undefined The symbol is known, but has no
value.
common The symbol is known and is defined
as COMMON.
unknown The symbol is not in the symbol
table.
EXAMPLES @LINK<RET>
*TEST<RET>
*SPEXP<RET>
*SPEX2<RET>
*/VALUE:(SPEX2,DPEXP,X2,X)<RET>
[LNKVAL SPEX2 460 DEFINED]
[LNKVAL DPEXP 221 UNDEFINED]
[LNKVAL X2 324 COMMON, LENGTH 1 (DECIMAL)]
[LNKVAL X UNKNOWN]
*
OPTIONAL You can omit the parentheses if only one
NOTATIONS symbol is given.
3-99
FIELD TEST
/VERBOSITY
FORMAT /VERBOSITY:keyword
FUNCTION Specifies the length of LINK messages. This switch is
similar to /MESSAGE. Both switches determine message
length, but use different keywords and output.
Keyword Description
SHORT Output only the 6-letter code.
MEDIUM Output the 6-letter code and the
medium-length message (usually one
line or less).
LONG Output the 6-letter code, the
medium-length message, and the long
message (usually several lines).
For a few messages no long message exists; in these cases
the LONG specification is ignored.
EXAMPLES In all the examples that follow an incorrect file
specification (ABC:WRONG) is used. In the following
example, only the 6-letter code and instructions for
correcting the error is returned.
@LINK<RET>
*/VERBOSITY:SHORT
*ABC:WRONG<RET>
%LNKNED
[ Please retype the incorrect parts of the file
specification]
*
In the next example, a medium length message is returned.
@LINK<RET>
*/VERBOSITY:MEDIUM<RET>
*ABC:WRONG<RET>
%LNKNED Non-existent device ABC:
[ Please retype the incorrect parts of the file specification]
*
In this example, a longer explanation is returned.
@LINK<RET>
/VERBOSITY:LONG<RET>
3-100
FIELD TEST
*ABC:WRONG<RET>
%LNKNED Non-existent device ABC:
You gave a device that does not exist on this system.
Correct your input files and reload.
[ Please retype the incorrect parts of the file specification]
*
RELATED /ERRORLEVEL, /MESSAGE
SWITCHES
3-101
FIELD TEST
/VERSION
FORMAT /VERSION:ic(j)-k
i = an octal number between 0 and 777 inclusive.
c = one or two alphabetic characters.
j = an octal number between 0 and 777777 inclusive.
k = an octal number between 0 and 7 inclusive.
FUNCTION Allows you to specify a version number. /VERSION changes
the value of .JBVER (location 137 in JOBDAT) and .JBHVR in
the vestigial job data area.
If the switch is appended to an input specification, or
with no specification, the version number is entered in
.JBVER and .JBHVR (location 4 in the vestigial job data
area).
There are four parts to the version arguments, given as i,
c, j, and k above. The first number (i) gives the major
version number. The character (c) gives the minor
version. The second number (j) gives the edit number.
The last number (k), which must be preceded by a hyphen
(-), shows which group last modified the file (0 = DIGITAL
development, 1 = other DIGITAL personnel, 2-7 = customer
use).
EXAMPLES @LINK<RET>
*/VERSION:3A(461)-0<RET>
*
Sets the version so that the major version is 3, the minor
version is A, the edit number is 461, and the last group
to modify the file was DIGITAL development.
OPTIONAL You can abbreviate /VERSION to /V.
NOTATIONS
3-102
FIELD TEST
3.5 EXAMPLES USING LINK DIRECTLY
For the following examples, the loaded program is a FORTRAN program
called MYPROG that writes the following:
This is written by MYPROG.
The following example shows an interactive execution of the program
using a LINK command. After running LINK, the command calls for
MYPROG to be loaded. Then the string MYLIB/USERLIB requests searching
of the library DSK:MYLIB.REL at the end of loading. The /NOSYSLIB
switch prevents searching the default system library. Finally the
/EXECUTE switch directs LINK to execute the loaded program, and the
/GO switch tells LINK that there are no more commands.
@LINK<RET>
*MYPROG,MYLIB/USERLIB/NOSYSLIB/EXECUTE/GO<RET>
[LNKXCT MYPROG execution]
This is written by MYPROG
CPU time 0.21 Elapsed time 1.31
@
The example below shows how to use LINK to load the program exactly as
above, except that the program is executed under the control of a
debugger (FORDDT for FORTRAN programs):
@LINK<RET>
*/DEBUG:FORDDT MYPROG,MYLIB/USERLIB/NOSYSLIB/GO<RET>
[LNKDEB FORDDT execution]
STARTING FORTRAN DDT
>>START<RET>
This is written by MYPROG
CPU time 0.17 Elapsed time 0.46
@
3-103
4-1
CHAPTER 4
OUTPUT FROM LINK
The primary output from LINK is the executable program formed from
input modules and switches. During its processing, LINK gives errors,
warnings, and informational messages. At your option, LINK can
generate any of several files.
4.1 THE EXECUTABLE PROGRAM
The executable program that LINK generates consists mostly of data and
machine instructions from your object modules. In the executable
program, all relocatable addresses have been resolved to absolute
addresses, and the values of all global references have been resolved.
You have several options for loading the program, depending on the
purpose of the load. Those options are:
o Execute the program. To do this, include the /EXECUTE switch
any place before the /GO switch. LINK passes control to your
program for execution.
o Execute the program under the control of a debugger. To do
this, use the /DEBUG switch before the first input file
specification.
o Execute the program and debug it after execution. To do
this, use the /TEST and /EXECUTE switches before the first
input file specification. After execution, type DEBUG to the
system to enter the debugger.
o Save the executable as a sharable save file or an EXE file.
To do this, use the /SAVE switch. See Section 4.2.
4-1
FIELD TEST
4.2 OUTPUT FILES
At your option, LINK can produce any of the following output files:
o Sharable save file.
o Log file.
o Map file.
o Symbol file.
o Plotter file (see Section 5.1).
o Overlay file (see Section 5.1).
4.2.1 Sharable Save Files
The sharable file, save file or EXE file, is a copy of the completed
executable program generated by LINK. You can create a sharable save
file by supplying the /SAVE switch before the /GO switch when you are
loading the program with direct commands to LINK. The sharable save
file retains the same file name as the source program, with a file
type of EXE.
Alternatively, you can type the file specification, followed by /SAVE,
and the sharable save file is written to the file you specified. If
you load the program with the executable LOAD TOPS-20 command, you may
then save the sharable save file by typing the TOPS-20 SAVE command.
You can run the sharable save file later, without running LINK, by
using the TOPS-20 RUN command, or the two TOPS-20 GET and START
commands. The following section describes the internal format of the
sharable save file.
4.2.1.1 Format of Sharable Save Files - A sharable save file is
divided into two main areas: the directory area contains information
about the structure of the file, and the data area contains the data
of the file.
4-2
FIELD TEST
The following diagram illustrates the general format of a sharable
save file:
Directory ========================
Area: ! Directory Section !
! !
! !
------------------------
! Entry Vector Section !
------------------------
! PDV !
! Section !
------------------------
! Terminating Section !
========================
Data Area: ! Data Section !
! !
! !
! !
! !
! !
! !
! !
! !
========================
NOTE
The PDV area is useful only with TOPS-20 version 5 and
later monitors. Earlier monitors ignore this area.
See Chapter 7 for more information on PDVs.
The directory area of the sharable save file has four distinct
sections: the directory section, the entry vector section, the PDV
section, and the terminating section. The size of the directory area
depends on the access characteristics of the pages in the data area of
the save file.
Each of the sections in the directory area begins with a header word
containing its identifier code in the left half and its length in the
right half. Each section is described in the following paragraphs.
The directory section is the first of the four sections and describes
groups of contiguous pages that have identical access. The length of
this section varies according to the number of groups that can be
generated from the data portion of the save file. The more data pages
that can be combined into a single group, the fewer groups required,
and the smaller the directory section.
The format of the directory section is as follows:
4-3
FIELD TEST
0 8 9 17 18 35
!=======================================================!
! Identifier code ! Number of words !
! 1776 ! (including this word) !
! ! in directory section !
!=======================================================!
! Access ! Page number in file, or 0 if group !
! bits ! of pages is all zero !
!=======================================================!
! Repeat ! Page number in the process !
! count ! !
!=======================================================!
.
.
.
!=======================================================!
! Access bits ! Page number in the file !
!=======================================================!
! Repeat count ! Page number in the process !
!=======================================================!
PSECT attributes are used to set the access bits. Refer to the
description of Block Type 24 in Appendix A. The bits currently
defined in the directory section are:
B1 The process pages in this group are sharable
B2 The process pages in this group are writable
The remaining access bits in the directory section are zero.
The repeat count is the number (minus 1) of consecutive pages in the
group described by the word pair. Pages are considered to be in a
group when the following three conditions are met:
1. The pages are contiguous.
2. The pages have the same access.
3. The pages are allocated but not loaded.
A group of all zero pages is indicated by a file page number of 0.
The word pairs are repeated for each group of pages in the address
space.
The entry vector section follows the directory section and points to
the first word of the entry vector and gives the length of the vector.
0 17 18 35
!=======================================================!
4-4
FIELD TEST
! Identifier code ! Number of words !
! 1775 ! (including this word) !
! ! in entry vector section !
!=======================================================!
! 254000 !
!=======================================================!
! Starting Address !
!=======================================================!
This format is the default. However, if you make special provisions
in your program, the format becomes the following. (Refer to the
description of Block Type 7 in Appendix A and the description of the
SFRKV% JSYS in the TOPS-20 Monitor Calls Reference Manual for further
information.)
0 17 18 35
!=======================================================!
! Identifier code ! Number of words !
! 1775 ! (including this word) !
! ! in entry vector section !
!=======================================================!
! Number of words in entry vector !
!=======================================================!
! Address of entry vector !
!=======================================================!
The data for this section is the address of the entry vector.
4-5
FIELD TEST
The PDV section may follow the entry vector section and contains the
addresses at which the PDVs begin (PDVAs). This section is optional
and only appears for an extended addressing program, if the program
declared a PDV, or if /NOJOBDAT is used causing LINK to create a PDV.
The format of the PDV section is as follows:
0 17 18 35
!=======================================================!
! Identifier code ! Number of words !
! 1774 ! (including this word) !
! ! in data vector section !
!=======================================================!
! Address of PDV 1 !
!=======================================================!
! Address of PDV 2 !
!=======================================================!
.
.
.
!=======================================================!
! Address of PDV n !
!=======================================================!
For information on the format of a PDV, see Chapter 7.
The terminating section, called the end section, always immediately
precedes the data section. The format of the terminating section is
the following:
!=======================================================!
! Identifier code ! !
! 1777 ! 1 !
!=======================================================!
The data area follows the terminating section, beginning at the next
page boundary.
4.2.2 LOG Files
A LOG file, .LOG, is generated if you use the /LOG switch. LINK then
writes most of its messages into the specified file. You can control
the kinds of messages entered in the LOG file by using the /LOGLEVEL
switch. For an example of a LOG file, see Section 5.1.
4-6
FIELD TEST
4.2.3 Map files
The map file, .MAP, is generated if you use the /MAP switch. LINK
constructs a symbol map in this file. The kinds of symbols included
depends on your use of the /CONTENTS, /LOCALS, /NOLOCALS, /NOINITIAL,
and /NOSYMBOLS switches. For an example of a map file, see Section
5.1. For a list of /MAP options, refer to Chapter 3.
4.2.4 Symbol Files
The symbol file, SYM file, is generated if you use the /SYFILE switch.
This file contains all global symbols, module names, and module
lengths, and, if you used the /LOCALS switch, all local symbols.
The format of the symbol file is defined using /SYFILE:ALGOL,
/SYFILE:RADIX-50, or /SYFILE:TRIPLET.
/SYSFILE:ALGOL creates a symbol file where the first word of the table
is XWD 1044, count. The remaining words are copied out of Type 1044
REL blocks. If an ALGOL main program has been loaded, then
/SYFILE:ALGOL becomes the default.
/SYFILE:RADIX-50 creates a symbol file where the first word of the
table is negative. Each symbol requires two words in the table: the
first is the symbol name in Radix-50 format; the second is the symbol
value. This is the default if just /SYFILE is used.
/SYFILE:TRIPLET creates internal format used by LINK.
4.3 MESSAGES
During its processing, LINK issues messages about what it is doing,
and about errors or possible errors it finds. LINK also responds to
query switches such as /COUNTERS, /ENTRY, /MISSING, /REQUEST, and
/UNDEFINED.
Each LINK message has an assigned level and an assigned severity. See
Appendix B for the level and severity of each message.
The level of a message determines whether the message is output to
your terminal, the log file, or both. You can control this output by
using the /ERRORLEVEL switch for the terminal and the /LOGLEVEL switch
for the log file. LINK's defaults are /ERRORLEVEL:8 and /LOGLEVEL:8.
4-7
FIELD TEST
Responses to query switches and messages that require you to do
something immediately are never output to the LOG file. For example,
if you use the /UNDEFINE switch, LINK responds with the LNKUGS
message; this message is output to the terminal but not to the log
file.
The severity of a message determines whether LINK considers the
message fatal (that is, whether the job is terminated). You can set
the fatal severity with the /SEVERITY switch. The default severities
are 24 for interactive jobs and 16 for batch jobs.
For both terminal messages and log file entries, LINK can issue short,
medium, or long messages, depending on your use of the /VERBOSITY
switch. For /VERBOSITY:SHORT, LINK gives only a 6-letter code; for
/VERBOSITY:MEDIUM, LINK gives the code and a medium-length message;
for /VERBOSITY:LONG, LINK gives the code, a medium-length message, and
a long message. The /MESSAGE switch can also be used to specify
message length. See Chapter 3 for more information on /MESSAGE and
the other switches discussed in this section.
Appendix B gives each 6-letter message code, its level and severity,
and its medium-length and long messages.
4-8
CHAPTER 5
OVERLAYS
If your loaded program is too large to execute in one piece, you may
be able to define an overlay structure for it. This permits the
system to execute the program with only some parts at a time in your
virtual address space. The overlay handler removes and reads in parts
of the program, according to the overlay structure.
NOTE
You only need an overlay structure if your program is
too large for your virtual address space. If the
program can fit in your virtual space, you should not
define an overlay structure for it; the monitor's page
swapping facility is faster than overlay execution.
5.1 OVERLAY STRUCTURES
An overlay program has a tree structure. (The tree is usually
pictured upside down.) The tree is made up of links, each containing
one or more program modules. These links are connected by paths.
Using LINK switches, you define each link and each path.
At the top of the (upside down) tree is the root link, which must
contain the main program. First-level links are below the root link;
each first-level link is connected to the root link by one path.
Second-level links are below the first-level links, and each is
connected by a path to exactly one first-level link. A link at level
n is connected by a path to exactly one link at level n-1.
Notice that a link can have more than one downward path (to successor
links), but only one upward path (to predecessor links).
Figure 5-1 shows a diagram of an overlay structure with 5 links. The
root link is TEST; the first-level links are LEFT and RIGHT; the
second-level links are LEFT1 and LEFT2.
5-1
FIELD TEST
Art work number: MRS-2595-83
Figure 5-1: Example of an Overlay Structure
Defining an overlay structure allows your program to execute in a
smaller space. This is because the code in a given link is allowed to
make reference to memory only in links along a direct upward or
downward path.
In the structure in Figure 5-1, the link LEFT can reference memory in
itself, in the root link (TEST), or in its successor links LEFT1 and
LEFT2. More generally, a link can reference memory in any link that
is vertically connected to it.
Referencing memory in any other link is not allowed. For example, a
path from LEFT1 to LEFT2 is not a direct upward or downward path.
Because of this restriction on memory references, only one complete
vertical path (at most) is required in the virtual address space at
any one time. The remaining links can be stored on disk while they
are not needed.
5.1.1 Defining Overlay Structures
LINK has a family of overlay-related switches. These switches are
summarized in Table 5-1 below and described in detail in Section
3.2.2. In addition to the overlay-related switches, you can use other
LINK switches such as /ERRORLEVEL or /LOG when loading overlays.
Table 5-1: Summary of LINK's Overlay-Related Switches
Switch Description
5-2
FIELD TEST
Table 5-1 (cont.)
Switch Description
/ARSIZE Sets the size of the overlay handler's table of
multiply-defined global symbols.
/LINK Closes an overlay link.
/MAXNODE Specifies the number of links to be defined when the
overlayed program requires more than 256 links.
/NODE Opens an overlay link.
/NOREQUEST Deletes references to links from LINK's overhead tables
when loading overlay programs.
/OVERLAY Initiates construction of an overlay structure.
/PLOT Directs LINK to output a tree diagram of your overlay
structure.
/PLTTYP Specifies the type of plot file to be generated by the
/PLOT switch.
The following example shows commands for defining the overlay
diagrammed in Figure 5-1. Each command is followed by an explanation.
*TEST/LOG/LOGLEVEL:2
Defines the log file for the overlay. TEST/LOG specifies that the
file is named TEST.LOG. The /LOGLEVEL:2 switch directs that messages
of level 2 and above be entered in the log file.
*/ERRORLEVEL:5
Directs LINK to return messages of level 5 and above to the terminal.
*TEST/OVERLAY
Tells LINK that an overlay structure is to be defined, and that the
file for the overlay is to be TEST.OVL.
*TEST/MAP
Defines the to contain symbol maps for each link.
*LPT:TEST/PLOT
Directs that a tree diagram of the overlay links be printed on the
line printer.
5-3
FIELD TEST
*OVL0,OVL1/LINK:TEST
Loads the OVL0.REL and OVL1.REL files into the root link. The
/LINK:TEST switch tells LINK that no more modules are to be in the
root link, and that the link name is TEST.
*/NODE:TEST OVL2/LINK:LEFT
Defines a link named LEFT. /NODE:TEST tells LINK that the link being
defined (LEFT) is to be an immediate successor to TEST, the root link.
OVL2/LINK:LEFT loads the OVL2.REL file, ends the link, and names it
LEFT.
*/NODE:LEFT OVL5/LINK:LEFT1
Defines a link named LEFT1. /NODE:LEFT tells LINK that the link being
defined (LINK1) is an immediate successor to LEFT. OVL5/LINK:LEFT1
loads the OVL5.REL file, ends the link, and names it LEFT1.
*/NODE:LEFT OVL6/LINK:LEFT2
Defines another immediate successor to LEFT named LEFT2.
*/NODE:TEST OVL3,OVL4/LINK:RIGHT
Defines the last link, RIGHT. /NODE:TEST defines RIGHT as an
immediate successor to TEST, loads the OVL3.REL and OVL4.REL files,
and names the link RIGHT.
*TEST/SAVE
Directs LINK to create the saved file TEST.EXE.
*/EXECUTE/GO
Specifies that the loaded program (TEST) is to be executed, and that
all commands to LINK are completed.
This process also produced an executable file TEST.EXE, which can be
run using the RUN system command. However, to run the program, the
file TEST.OVL must be present, because it provides the code for the
links.
5.1.2 An Overlay Example
The following pages show samples of the files used in the previous
example.
5-4
FIELD TEST
5.1.2.1 Source Files - Copies of the FORTRAN source files used in the
overlay are displayed on the terminal using the TOPS-20 TYPE command.
@TYPE OVL0.FOR<RET>
CTHIS SOFTWARE IS FURNISHED UNDER A LICENSE AND MAY ONLY BE USED
C OR COPIED IN ACCORDANCE WITH THE TERMS OF SUCH LICENSE.
C
CCOPYRIGHT (C) DIGITAL EQUIPMENT CORPORATION 1982, 1983
C Simple overlay tests
C OVL0
C / \
C OVL2A OVL4
C / \ \
C OVL5 OVL6 OVL1
C \ /
C OVL2B OVL3
IMPLICIT DOUBLE PRECISION (D)
IMPLICIT COMPLEX (C)
IMPLICIT INTEGER (A)
TYPE 1
1 FORMAT ('1','Execution begins in main program OVL0')
TYPE 11
11 FORMAT (1X,'OVL0 CALLS OVL2A')
IVAR=0
CALL OVL2A(IVAR)
TYPE 2
2 FORMAT (/1X,'RETURN TO OVL0')
IF (IVAR.NE.1) TYPE 100
100 FORMAT(' ?Error. Value not returned from OVL2A')
TYPE 21
21 FORMAT (1X,'OVL0 CALLS OVL4')
NUMB=0
CALL OVL4(NUMB)
TYPE 2
IF (IVAR.NE.1) TYPE 200
200 FORMAT(' ?Error. Value not returned from NUMB')
TYPE 3
3 FORMAT (/1X,'Execution ends in main program OVL0'//)
STOP
END
5-5
FIELD TEST
@TYPE OVL1.FOR
SUBROUTINE OVL1(IVAR)
C THIS SOFTWARE IS FURNISHED UNDER A LICENSE AND MAY ONLY BE USED
C OR COPIED IN ACCORDANCE WITH THE TERMS OF SUCH LICENSE.
C
C COPYRIGHT (C) DIGITAL EQUIPMENT CORPORATION 1982, 1983
IMPLICIT DOUBLE PRECISION (D)
IMPLICIT COMPLEX (C)
IMPLICIT INTEGER (A)
TYPE 1
1 FORMAT (30x,' OVL1 CALLS OVL3'/)
NUM=0
CALL OVL3(NUM)
TYPE 2
2 FORMAT (30x,' RETURN TO OVL1'/)
IF (NUM.NE.1) TYPE 100
100 FORMAT(' ?Error. Variable not returned from OVL3.')
IVAR=1
RETURN
END
@TYPE OVL2.FOR
SUBROUTINE OVL2A(ARG)
C THIS SOFTWARE IS FURNISHED UNDER A LICENSE AND MAY ONLY BE USED
C OR COPIED IN ACCORDANCE WITH THE TERMS OF SUCH LICENSE.
C
C COPYRIGHT (C) DIGITAL EQUIPMENT CORPORATION 1982, 1983
IMPLICIT DOUBLE PRECISION (D)
IMPLICIT COMPLEX (C)
IMPLICIT INTEGER (A)
INTEGER OVL6
TYPE 1
1 FORMAT (1X,' OVL2A CALLS OVL5'/)
ARG=0
CALL OVL5(ARG)
TYPE 2
2 FORMAT (1X,' RETURN TO OVL2A'/)
IF (ARG.NE.1) TYPE 100
100 FORMAT(' ?Error. Value not returned from OVL5')
5-6
FIELD TEST
TYPE 3
3 FORMAT (1X,' OVL2A CALLS OVL6'/)
NUM=0
NUM=OVL6(NUM)
TYPE 2
IF (NUM.NE.1) TYPE 300
300 FORMAT(' ?Error. Value not returned from OVL6')
ARG=1 !Return value
RETURN
END
SUBROUTINE OVL2B
IMPLICIT DOUBLE PRECISION (D)
IMPLICIT COMPLEX (C)
IMPLICIT INTEGER (A)
TYPE 1
1 FORMAT (/1X,' OVL2B DOESN''T CALL ANYTHING')
RETURN
END
@TYPE OVL3.FOR
SUBROUTINE OVL3(IVAR)
C THIS SOFTWARE IS FURNISHED UNDER A LICENSE AND MAY ONLY BE USED
C OR COPIED IN ACCORDANCE WITH THE TERMS OF SUCH LICENSE.
C
C COPYRIGHT (C) DIGITAL EQUIPMENT CORPORATION 1982, 1983
IMPLICIT DOUBLE PRECISION (D)
IMPLICIT COMPLEX (C)
IMPLICIT INTEGER (A)
IVAR=1 !Got here!
TYPE 1
1 FORMAT (30x,' OVL3 DOESN''T CALL ANYTHING'/)
RETURN
END
@TYPE OVL4.FOR
SUBROUTINE OVL4(ARGU)
C THIS SOFTWARE IS FURNISHED UNDER A LICENSE AND MAY ONLY BE USED
5-7
FIELD TEST
C OR COPIED IN ACCORDANCE WITH THE TERMS OF SUCH LICENSE.
C
C COPYRIGHT (C) DIGITAL EQUIPMENT CORPORATION 1982, 1982
IMPLICIT DOUBLE PRECISION (D)
IMPLICIT COMPLEX (C)
IMPLICIT INTEGER (A)
TYPE 1
1 FORMAT (30x,' OVL4 CALLS OVL1')
NUM=0
CALL OVL1(NUM)
IF (NUM.NE.1) TYPE 100
100 FORMAT(' ?Error. Value not returned from OVL1')
TYPE 2
2 FORMAT (30x,' RETURN TO OVL4')
ARGU=1
RETURN
END
@TYPE OVL5.FOR
SUBROUTINE OVL5(NUM)
C THIS SOFTWARE IS FURNISHED UNDER A LICENSE AND MAY ONLY BE USED
C OR COPIED IN ACCORDANCE WITH THE TERMS OF SUCH LICENSE.
C
C COPYRIGHT (C) DIGITAL EQUIPMENT CORPORATION 1982, 1983
IMPLICIT DOUBLE PRECISION (D)
IMPLICIT COMPLEX (C)
IMPLICIT INTEGER (A)
COMMON C
C=(1,1)
C=C**C
TYPE 1
1 FORMAT (30x,' OVL5 CALLS CEXP3.'/)
NUM=1
RETURN
END
@TYPE OVL6.FOR
INTEGER FUNCTION OVL6(ARG)
5-8
FIELD TEST
C THIS SOFTWARE IS FURNISHED UNDER A LICENSE AND MAY ONLY BE USED
C OR COPIED IN ACCORDANCE WITH THE TERMS OF SUCH LICENSE.
C
C COPYRIGHT (C) DIGITAL EQUIPMENT CORPORATION 1982, 1983
IMPLICIT DOUBLE PRECISION (D)
IMPLICIT COMPLEX (C)
IMPLICIT INTEGER (A)
TYPE 1
1 FORMAT (30x,' OVL6 CALLS OVL2B'/)
NUM=0
CALL OVL2B(NUM)
TYPE 2
2 FORMAT (30x,' RETURN TO OVL6'/)
IF (NUM .NE. 0) TYPE 100
100 FORMAT(' ?Error. Value not returned from OVL2A')
OVL6=1
RETURN
END
5.1.2.2 Source File Compilation - Source files are compiled using the
TOPS-20 COMPILE command.
@compile ovl0,ovl1,ovl2,ovl3,ovl4,ovl5,ovl6
FORTRAN: OVL0
OVL0
FORTRAN: OVL1
OVL1
FORTRAN: OVL2
OVL2A
OVL2B
FORTRAN: OVL3
OVL3
FORTRAN: OVL4
OVL4
FORTRAN: OVL5
OVL5
FORTRAN: OVL6
OVL6
5.1.2.3 Interactive use of LINK - LINK is run and used to define and
execute the overlay.
@LINK
TEST/LOG/LOGLEVEL:5
/ERRORLEVEL:5
5-9
FIELD TEST
TEST/OVERLAY
TEST/MAP
LPT:TEST/PLOT
OVL0,OVL1/LINK:TEST
[LNKLMN Loading module OVL0 from file OVL0.REL.1]
[LNKLMN Loading module OVL1 from file OVL1.REL.1]
[LNKLMN Loading module OVRLAY from file SYS:OVRLAY.REL.10]
[LNKLMN Loading module FORHST from file SYS:FORLIB.REL.1]
[LNKLMN Loading module FORSHR from file SYS:FORLIB.REL.1]
[LNKLMN Loading module FORGET from file SYS:FORLIB.REL.1]
[LNKLMN Loading module FDBDUM from file SYS:FORLIB.REL.1]
[LNKLMN Loading module KSORT from file SYS:FORLIB.REL.1]
[LNKLMN Loading module FORMSL from file SYS:FORLIB.REL.1]
[LNKLMN Loading module FORPSE from file SYS:FORLIB.REL.1]
[LNKELN End of link number 0 name TEST]
NODE:TEST OVL2/LINK:LEFT
[LNKLMN Loading module OVL2A from file OVL2.REL.1]
[LNKLMN Loading module OVL2B from file OVL2.REL.1]
[LNKELN End of link number 1 name LEFT]
/NODE:LEFT OVL5/LINK:LEFT1
[LNKLMN Loading module OVL5 from file OVL5.REL.1]
[LNKLMN Loading module CEXP3. from file SYS:FORLIB.REL.1]
[LNKLMN Loading module CDLOG from file SYS:FORLIB.REL.1]
[LNKLMN Loading module CGLOG from file SYS:FORLIB.REL.1]
[LNKLMN Loading module CGLOG from file SYS:FORLIB.REL.1]
[LNKLMN Loading module GATAN. from file SYS:FORLIB.REL.1]
[LNKLMN Loading module GLOG. from file SYS:FORLIB.REL.1]
[LNKLMN Loading module DATAN. from file SYS:FORLIB.REL.1]
[LNKLMN Loading module DEXP. from file SYS:FORLIB.REL.1]
[LNKLMN Loading module DLOG. from file SYS:FORLIB.REL.1]
[LNKLMN Loading module DSIN. from file SYS:FORLIB.REL.1]
[LNKLMN Loading module DFL.4 from file SYS:FORLIB.REL.1]
[LNKLMN Loading module EXP2. from file SYS:FORLIB.REL.1]
[LNKLMN Loading module MTHMSG from file SYS:FORLIB.REL.1]
[LNKELN End of link number 2 name LEFT1]
/NODE:LEFT OVL6/LINK:LEFT2
[LNKLMN Loading module OVL6 from file OVL6.REL.1]
[LNKELN End of link number 3 name LEFT2]
/NODE:TEST OVL3,OVL4/LINK:RIGHT
[LNKLMN Loading module OVL3 from file OVL3.REL.1]
[LNKLMN Loading module OVL4 from file OVL4.REL.1]
[LNKELN End of link number 4 name RIGHT]
TEST/SAVE
/EXECUTE/GO
[LNKXCT OVL0 execution]
Execution begins in main program OVL0
OVL0 CALLS OVL2A
OVL2A CALLS OVL5
OVL5 CALLS CEXP3.
5-10
FIELD TEST
RETURN TO OVL2A
OVL2A CALLS OVL6
OVL6 CALLS OVL2B
OVL2B DOESN'T CALL ANYTHING
RETURN TO OVL6
RETURN TO OVL2A
RETURN TO OVL0
OVL0 CALLS OVL4
OVL4 CALLS OVL1
OVL1 CALLS OVL3
OVL3 DOESN'T CALL ANYTHING
RETURN TO OVL1
RETURN TO OVL4
RETURN TO OVL0
Execution ends in main program OVL0
CPU time 0.85 Elapsed time 11.85
5.1.2.4 TEST.LOG - LINK generated the following TEST.LOG file. It
shows the log messages issued during the load.
15:30:48 6 1 LMN Loading module OVL0 from file OVL0.REL.1
15:30:49 6 1 LMN Loading module OVL1 from file OVL1.REL.1
15:30:50 6 1 LMN Loading module OVRLAY from file SYS:OVRLAY.REL.10
15:30:52 6 1 LMN Loading module FORHST from file SYS:FORLIB.REL.1
15:30:52 6 1 LMN Loading module FORSHR from file SYS:FORLIB.REL.1
15:30:52 6 1 LMN Loading module FORGET from file SYS:FORLIB.REL.1
15:30:53 6 1 LMN Loading module FDBDUM from file SYS:FORLIB.REL.1
15:30:53 6 1 LMN Loading module KSORT from file SYS:FORLIB.REL.1
15:30:54 6 1 LMN Loading module FORMSL from file SYS:FORLIB.REL.1
15:30:54 6 1 LMN Loading module FORPSE from file SYS:FORLIB.REL.1
15:31:23 7 1 ELN End of link number 0 name TEST
15:31:42 6 1 LMN Loading module OVL2A from file OVL2.REL.1
15:31:42 6 1 LMN Loading module OVL2B from file OVL2.REL.1
15:31:48 7 1 ELN End of link number 1 name LEFT
5-11
FIELD TEST
15:32:04 6 1 LMN Loading module OVL5 from file OVL5.REL.1
15:32:06 6 1 LMN Loading module CEXP3. from file SYS:FORLIB.REL.1
15:32:06 6 1 LMN Loading module CDLOG from file SYS:FORLIB.REL.1
15:32:07 6 1 LMN Loading module CGLOG from file SYS:FORLIB.REL.1
15:32:07 6 1 LMN Loading module CGLOG from file SYS:FORLIB.REL.1
15:32:07 6 1 LMN Loading module GATAN. from file SYS:FORLIB.REL.1
15:32:08 6 1 LMN Loading module GLOG. from file SYS:FORLIB.REL.1
15:32:09 6 1 LMN Loading module DATAN. from file SYS:FORLIB.REL.1
15:32:09 6 1 LMN Loading module DEXP. from file SYS:FORLIB.REL.1
15:32:09 6 1 LMN Loading module DLOG. from file SYS:FORLIB.REL.1
15:32:09 6 1 LMN Loading module DSIN. from file SYS:FORLIB.REL.1
15:32:10 6 1 LMN Loading module DFL.4 from file SYS:FORLIB.REL.1
15:32:10 6 1 LMN Loading module EXP2. from file SYS:FORLIB.REL.1
15:32:10 6 1 LMN Loading module MTHMSG from file SYS:FORLIB.REL.1
15:32:12 7 1 ELN End of link number 2 name LEFT1
15:32:36 6 1 LMN Loading module OVL6 from file OVL6.REL.1
15:32:37 7 1 ELN End of link number 3 name LEFT2
15:32:55 6 1 LMN Loading module OVL3 from file OVL3.REL.1
15:32:56 6 1 LMN Loading module OVL4 from file OVL4.REL.1
15:32:56 7 1 ELN End of link number 4 name RIGHT
5.1.2.5 TEST.MAP - LINK generated the following TEST.MAP file. It
shows symbol maps for the overlay.
LINK symbol map of OVL0 page 1
Produced by LINK version 6(2353) on 11-Feb-85 at 11:53:18
Overlay no. 0 name TEST
Overlay is absolute
Low segment starts at 0 ends at 10243 length 10244 = 9P
Control Block address is 10202, length 32 (octal), 26. (decimal)
348 words free in Low segment
188 Global symbols loaded, therefore min. hash size is 209
Start address is 235, located in program OVL0
*************
JOBDAT-INITIAL-SYMBOLS
%JOBDT 43200000454 Global Absolute Suppressed .JB41 41 Global Absolute
.JBAPR 125 Global Absolute .JBBLT 45 Global Absolute
.JBCHN 131 Global Absolute .JBCNI 126 Global Absolute
.JBCOR 133 Global Absolute .JBCST 136 Global Absolute
.JBDA 140 Global Absolute .JBDDT 74 Global Absolute
.JBEDV 112 Global Absolute .JBERR 42 Global Absolute
.JBFF 121 Global Absolute .JBH41 1 Global Absolute Suppressed
.JBHCR 2 Global Absolute Suppressed .JBHDA 10 Global Absolute Suppressed
.JBHGA 7 Global Absolute Suppressed .JBHGH 400000 Global Absolute Suppressed
.JBHNM 5 Global Absolute Suppressed .JBHRL 115 Global Absolute
5-12
FIELD TEST
.JBHRN 3 Global Absolute Suppressed .JBHSA 0 Global Absolute Suppressed
.JBHSM 6 Global Absolute Suppressed .JBHVR 4 Global Absolute Suppressed
.JBINT 134 Global Absolute .JBOPC 130 Global Absolute
.JBOPS 135 Global Absolute .JBOVL 131 Global Absolute
.JBPFH 123 Global Absolute .JBPFI 74 Global Absolute
.JBREL 44 Global Absolute .JBREN 124 Global Absolute
.JBSA 120 Global Absolute .JBSYM 116 Global Absolute
.JBTPC 127 Global Absolute .JBUSY 117 Global Absolute
.JBUUO 40 Global Absolute .JBVER 137 Global Absolute
*************
OVL0 from DSK:OVL0.REL created by FORTRAN on 11-Feb-85 at 11:52:27
Low segment starts at 140 ends at 234 length 75 (octal), 61. (decimal)
High segment starts at 235 ends at 345 length 111 (octal), 73. (decimal)
MAIN. 235 Global Relocatable OVL0 235 Entry Relocatable
*************
OVL1 from DSK:OVL1.REL created by FORTRAN on 11-Feb-85 at 11:52:28
Low segment starts at 346 ends at 377 length 32 (octal), 26. (decimal)
High segment starts at 400 ends at 442 length 43 (octal), 35. (decimal)
OVL1 401 Entry Relocatable FOROT% 400010 Global Absolute
*************
OVRLAY from SYS:OVRLAY.REL created by MACRO on 25-Jan-85 at 5:25:39
Low segment starts at 4252 ends at 5245 length 774 (octal), 508. (decimal)
High segment starts at 443 ends at 4150 length 3506 (octal), 1862. (decimal)
LINK symbol map of OVL0 page 2
OVRLAY
BOUT% 104000000051 Global Absolute CLOSF% 104000000022 Global Absolute
CLROV. 2501 Entry Relocatable ERJMP 320700000000 Global Absolute
ERSTR% 104000000011 Global Absolute GCVEC% 104000000300 Global Absolute
GETOV. 2376 Entry Relocatable GJ%FOU 400000000000 Global Absolute Suppressed
GJ%OLD 100000000000 Global Absolute Suppressed GJ%SHT 1000000 Global Absolute Suppressed
GJ%TMP 10000000000 Global Absolute Suppressed GTJFN% 104000000020 Global Absolute
HALTF% 104000000170 Global Absolute INIOV. 2354 Entry Relocatable
JFNS% 104000000030 Global Absolute JS%DEV 300000000000 Global Absolute Suppressed
JS%DIR 70000000000 Global Absolute Suppressed JS%NAM 7000000000 Global Absolute Suppressed
JS%PAF 1 Global Absolute Suppressed LOGOV. 2520 Entry Relocatable
OF%BSZ 770000000000 Global Absolute Suppressed OF%RD 200000 Global Absolute Suppressed
OF%WR 100000 Global Absolute Suppressed OPENF% 104000000021 Global Absolute
PA%EX 20000000000 Global Absolute Suppressed PA%PRV 200000000 Global Absolute Suppressed
PBOUT% 104000000074 Global Absolute REMOV. 2412 Entry Relocatable
RMAP% 104000000061 Global Absolute RPACS% 104000000057 Global Absolute
RUNOV. 2430 Entry Relocatable RUNTM% 104000000015 Global Absolute
SAVOV. 2454 Entry Relocatable SFPTR% 104000000027 Global Absolute
5-13
FIELD TEST
SIN% 104000000052 Global Absolute SOUT% 104000000053 Global Absolute
%OVRLA 501000212 Global Absolute Suppressed .FHJOB 777773 Global Absolute Suppressed
.FHSLF 400000 Global Absolute Suppressed .GJEXT 5 Global Absolute Suppressed
.GJGEN 0 Global Absolute Suppressed .JSAOF 1 Global Absolute Suppressed
.NULIO 377777 Global Absolute Suppressed .OVRLA 4273 Entry Relocatable
.OVRLO 4314 Global Relocatable .OVRLU 3254 Entry Relocatable
.OVRWA 4313 Global Relocatable
*************
FORHST from SYS:FORLIB.REL created by MACRO on 25-Jan-85 at 5:24:29
Zero length module
*************
FORSHR from SYS:FORLIB.REL created by MACRO on 25-Jan-85 at 5:24:29
Zero length module
*************
FORGET from SYS:FORLIB.REL created by MACRO on 25-Jan-85 at 5:24:29
Low segment starts at 5725 ends at 7134 length 1210 (octal), 648. (decimal)
High segment starts at 5246 ends at 5724 length 457 (octal), 303. (decimal)
ABORT. 5603 Entry Relocatable
ALCHN. 5557 Entry Relocatable ALCOR. 5553 Entry Relocatable
CFIBF% 104000000100 Global Absolute CHRPT. 5725 Global Relocatable
CLOSE. 5515 Entry Relocatable DEC. 5533 Entry Relocatable
DECHN. 5561 Entry Relocatable DECOR. 5555 Entry Relocatable
ENC. 5531 Entry Relocatable ERROR. 5511 Entry Relocatable
ESOUT% 104000000313 Global Absolute EXIT. 5551 Entry Relocatable
EXIT1. 5517 Entry Relocatable FDBMS. 5567 Entry Relocatable
FIN. 5543 Entry Relocatable FIND. 5547 Entry Relocatable
FORER. 5511 Entry Relocatable FOROP. 5573 Entry Relocatable
FOROT$ 5610 Entry Relocatable FOROT. 5246 Entry Relocatable
FUNCT. 5565 Entry Relocatable GET% 104000000200 Global Absolute
LINK symbol map of OVL0 page 3
FORGET
GEVEC% 104000000205 Global Absolute GJ%PHY 10000000 Global Absolute Suppressed
GT%NOV 40000 Global Absolute Suppressed IFI. 5575 Entry Relocatable
IFO. 5577 Entry Relocatable IN. 5521 Entry Relocatable
INIT. 5507 Entry Relocatable INQ. 5571 Entry Relocatable
IOLST. 5541 Entry Relocatable MTHER. 5601 Entry Relocatable
MTOP. 5545 Entry Relocatable NLI. 5535 Entry Relocatable
NLO. 5537 Entry Relocatable OPEN. 5513 Entry Relocatable
OUT. 5523 Entry Relocatable PA%PEX 10000000000 Global Absolute Suppressed
PM%CNT 400000000000 Global Absolute Suppressed PM%RD 100000000000 Global Absolute Suppressed
PM%RWX 160000000000 Global Absolute Suppressed PM%WR 40000000000 Global Absolute Suppressed
PMAP% 104000000056 Global Absolute PSOUT% 104000000076 Global Absolute
RD%JFN 4000000000 Global Absolute Suppressed RD%TOP 200000000000 Global Absolute Suppressed
5-14
FIELD TEST
RF%LNG 400000000000 Global Absolute Suppressed RFSTS% 104000000156 Global Absolute
RTB. 5525 Entry Relocatable SEVEC% 104000000204 Global Absolute
TEXTI% 104000000524 Global Absolute TRACE. 5563 Entry Relocatable
WTB. 5527 Entry Relocatable .PRIIN 100 Global Absolute Suppressed
.PRIOU 101 Global Absolute Suppressed .RFSFL 4 Global Absolute Suppressed
*************
FDBDUM from SYS:FORLIB.REL created by MACRO on 25-Jan-85 at 5:24:29
High segment starts at 7135 ends at 7135 length 1 (octal), 1. (decimal)
FDBMS% 7135 Global Relocatable KDBMS. 7135 Global Relocatable
%DBSTP 7135 Entry Relocatable
*************
KSORT from SYS:FORLIB.REL created by MACRO on 25-Jan-85 at 5:24:29
Low segment starts at 7152 ends at 7154 length 3 (octal), 3. (decimal)
High segment starts at 7136 ends at 7151 length 14 (octal), 12. (decimal)
KSORT. 7136 Entry Relocatable
%PASRT 7154 Global Relocatable %SRTAD 7152 Global Relocatable
*************
FORMSL from SYS:FORLIB.REL created by MACRO on 25-Jan-85 at 5:24:29
High segment starts at 7155 ends at 7714 length 540 (octal), 352. (decimal)
F.AQS 7267 Entry Relocatable F.CFS 7360 Entry Relocatable
F.CGP 7333 Entry Relocatable F.CGS 7367 Entry Relocatable
F.CLE 7233 Entry Relocatable F.CPP 7376 Entry Relocatable
F.CWU 7223 Entry Relocatable F.ICE 7242 Entry Relocatable
F.IDU 7173 Entry Relocatable F.IJE 7424 Entry Relocatable
F.IOE 7155 Entry Relocatable F.IPN 7405 Entry Relocatable
F.MXD 7342 Entry Relocatable F.NCA 7260 Entry Relocatable
F.NCS 7251 Entry Relocatable F.NOF 7213 Entry Relocatable
F.NOR 7164 Entry Relocatable F.NSS 7351 Entry Relocatable
F.SNH 7415 Entry Relocatable F.SRE 7311 Entry Relocatable
F.SSE 7276 Entry Relocatable F.TMA 7324 Entry Relocatable
F.UNO 7203 Entry Relocatable
*************
LINK symbol map of OVL0 page 4
FORPSE from SYS:FORLIB.REL created by MACRO on 25-Jan-85 at 5:24:29
Low segment starts at 10121 ends at 10201 length 61 (octal), 49. (decimal)
High segment starts at 7715 ends at 10120 length 204 (octal), 132. (decimal)
PAUS. 7716 Entry Relocatable
RFMOD% 104000000107 Global Absolute SFMOD% 104000000110 Global Absolute
STOP. 7721 Entry Relocatable TT%OSP 400000000000 Global Absolute Suppressed
5-15
FIELD TEST
*************
Index to LINK symbol map of OVL0 page 5
Name Page Name Page Name Page Name Page
FDBDUM 3 FORMSL 3 KSORT 3 OVL1 1
FORGET 2 FORPSE 4 OVL0 1 OVRLAY 1
FORHST 2 FORSHR 2
LINK symbol map of OVL0 #1 page 6
Overlay no. 1 name LEFT
Overlay is absolute
Low segment starts at 14244 ends at 14501 length 236 = 1P
Control Block address is 14440, length 30 (octal), 24. (decimal)
Path is 0
190 words free in Low segment
6 Global symbols loaded, therefore min. hash size is 7
*************
OVL2A from DSK:OVL2.REL created by FORTRAN on 11-Feb-85 at 11:52:30
Low segment starts at 14244 ends at 14315 length 52 (octal), 42. (decimal)
High segment starts at 14316 ends at 14412 length 75 (octal), 61. (decimal)
OVL2A 14317 Entry Relocatable
*************
OVL2B from DSK:OVL2.REL created by FORTRAN on 11-Feb-85 at 11:52:30
Low segment starts at 14413 ends at 14425 length 13 (octal), 11. (decimal)
High segment starts at 14426 ends at 14437 length 12 (octal), 10. (decimal)
OVL2B 14427 Entry Relocatable
*************
LINK symbol map of OVL0 #2 page 7
Overlay no. 2 name LEFT1
Overlay is absolute
Low segment starts at 14502 ends at 22675 length 6174 = 7P
Control Block address is 22510, length 20 (octal), 16. (decimal)
Path is 0, 1
66 words free in Low segment
53 Global symbols loaded, therefore min. hash size is 59
*************
5-16
FIELD TEST
OVL5 from DSK:OVL5.REL created by FORTRAN on 11-Feb-85 at 11:52:36
Low segment starts at 14502 ends at 14517 length 16 (octal), 14. (decimal)
High segment starts at 14520 ends at 14547 length 30 (octal), 24. (decimal)
.COMM. 14502 Common length 2. OVL5 14521 Entry Relocatable
*************
CEXP3. from SYS:FORLIB.REL created by MACRO on 25-Jan-85 at 5:24:29
Low segment starts at 16062 ends at 16144 length 63 (octal), 51. (decimal)
High segment starts at 14550 ends at 16061 length 1312 (octal), 714. (decimal)
CEXP2. 14551 Entry Relocatable CEXP3. 14655 Entry Relocatable
*************
CDLOG from SYS:FORLIB.REL created by MACRO on 25-Jan-85 at 5:24:29
Low segment starts at 16515 ends at 16544 length 30 (octal), 24. (decimal)
High segment starts at 16145 ends at 16514 length 350 (octal), 232. (decimal)
CDLOG 16146 Entry Relocatable
*************
CGLOG from SYS:FORLIB.REL created by MACRO on 25-Jan-85 at 5:24:29
Zero length module
*************
CGLOG from SYS:FORLIB.REL created by MACRO on 25-Jan-85 at 5:24:29
Low segment starts at 17112 ends at 17145 length 34 (octal), 28. (decimal)
High segment starts at 16545 ends at 17111 length 345 (octal), 229. (decimal)
CGLOG. 16546 Entry Relocatable
*************
GATAN. from SYS:FORLIB.REL created by MACRO on 25-Jan-85 at 5:24:29
Low segment starts at 17715 ends at 17723 length 7 (octal), 7. (decimal)
High segment starts at 17146 ends at 17714 length 547 (octal), 359. (decimal)
GATAN. 17147 Entry Relocatable GATN2. 17242 Entry Relocatable
*************
LINK symbol map of OVL0 #2 page 8
GLOG. from SYS:FORLIB.REL created by MACRO on 25-Jan-85 at 5:24:29
Low segment starts at 20075 ends at 20102 length 6 (octal), 6. (decimal)
High segment starts at 17724 ends at 20074 length 151 (octal), 105. (decimal)
5-17
FIELD TEST
GLG10. 17725 Entry Relocatable GLOG. 17730 Entry Relocatable
*************
DATAN. from SYS:FORLIB.REL created by MACRO on 25-Jan-85 at 5:24:29
Low segment starts at 20652 ends at 20660 length 7 (octal), 7. (decimal)
High segment starts at 20103 ends at 20651 length 547 (octal), 359. (decimal)
DATAN. 20104 Entry Relocatable DATN2. 20177 Entry Relocatable
*************
DEXP. from SYS:FORLIB.REL created by MACRO on 25-Jan-85 at 5:24:29
Low segment starts at 21042 ends at 21044 length 3 (octal), 3. (decimal)
High segment starts at 20661 ends at 21041 length 161 (octal), 113. (decimal)
DEXP. 20662 Entry Relocatable
*************
DLOG. from SYS:FORLIB.REL created by MACRO on 25-Jan-85 at 5:24:29
Low segment starts at 21217 ends at 21222 length 4 (octal), 4. (decimal)
High segment starts at 21045 ends at 21216 length 152 (octal), 106. (decimal)
DLG10. 21046 Entry Relocatable
DLOG. 21051 Entry Relocatable
*************
DSIN. from SYS:FORLIB.REL created by MACRO on 25-Jan-85 at 5:24:29
Low segment starts at 21450 ends at 21454 length 5 (octal), 5. (decimal)
High segment starts at 21223 ends at 21447 length 225 (octal), 149. (decimal)
DCOS. 21224 Entry Relocatable
DSIN. 21236 Entry Relocatable
*************
DFL.4 from SYS:FORLIB.REL created by MACRO on 25-Jan-85 at 5:24:29
High segment starts at 21455 ends at 21464 length 10 (octal), 8. (decimal)
DFL.4 21456 Entry Relocatable
*************
EXP2. from SYS:FORLIB.REL created by MACRO on 25-Jan-85 at 5:24:29
Low segment starts at 21606 ends at 21610 length 3 (octal), 3. (decimal)
High segment starts at 21465 ends at 21605 length 121 (octal), 81. (decimal)
EXP2. 21466 Entry Relocatable
*************
LINK symbol map of OVL0 #2 page 9
5-18
FIELD TEST
MTHMSG from SYS:FORLIB.REL created by MACRO on 25-Jan-85 at 5:24:29
High segment starts at 21611 ends at 22507 length 677 (octal), 447. (decimal)
L.888 22106 Entry Relocatable
L.999 22077 Entry Relocatable L.AII 21762 Entry Relocatable
L.AIR 21753 Entry Relocatable L.AIZ 21735 Entry Relocatable
L.AOI 22043 Entry Relocatable L.ARZ 21744 Entry Relocatable
L.ATI 21710 Entry Relocatable L.ATZ 21701 Entry Relocatable
L.AZM 21771 Entry Relocatable L.BAZ 22052 Entry Relocatable
L.BPI 22016 Entry Relocatable L.BPO 21663 Entry Relocatable
L.BPU 21672 Entry Relocatable L.IPO 21645 Entry Relocatable
L.IPU 21627 Entry Relocatable L.MFX 22124 Entry Relocatable
L.NAA 22025 Entry Relocatable L.NNA 22061 Entry Relocatable
L.NOR 22115 Entry Relocatable L.ROV 21611 Entry Relocatable
L.RPO 21654 Entry Relocatable L.RPU 21636 Entry Relocatable
L.RTI 21717 Entry Relocatable L.RTR 21726 Entry Relocatable
L.RUN 21620 Entry Relocatable L.ZCI 22007 Entry Relocatable
L.ZIZ 22034 Entry Relocatable L.ZNI 22070 Entry Relocatable
L.ZZZ 22000 Entry Relocatable
*************
Index to LINK symbol map of OVL0 #2 page 10
Name Page Name Page Name Page Name Page
CDLOG 7 DATAN. 8 DSIN. 8 GLOG. 8
CEXP3. 7 DEXP. 8 EXP2. 8 MTHMSG 9
CGLOG 7 DFL.4 8 GATAN. 7 OVL5 7
CGLOG 7 DLOG. 8
LINK symbol map of OVL0 #3 page 11
Overlay no. 3 name LEFT2
Overlay is absolute
Low segment starts at 14502 ends at 14666 length 165 = 1P
Control Block address is 14637, length 20 (octal), 16. (decimal)
Path is 0, 1
73 words free in Low segment
4 Global symbols loaded, therefore min. hash size is 5
*************
OVL6 from DSK:OVL6.REL created by FORTRAN on 11-Feb-85 at 11:52:37
Low segment starts at 14502 ends at 14550 length 47 (octal), 39. (decimal)
High segment starts at 14551 ends at 14636 length 66 (octal), 54. (decimal)
OVL6 14552 Entry Relocatable
*************
5-19
FIELD TEST
LINK symbol map of OVL0 #4 page 12
Overlay no. 4 name RIGHT
Overlay is absolute
Low segment starts at 14244 ends at 14426 length 163 = 1P
Control Block address is 14375, length 22 (octal), 18. (decimal)
Path is 0
233 words free in Low segment
5 Global symbols loaded, therefore min. hash size is 6
*************
OVL3 from DSK:OVL3.REL created by FORTRAN on 11-Feb-85 at 11:52:32
Low segment starts at 14244 ends at 14257 length 14 (octal), 12. (decimal)
High segment starts at 14260 ends at 14277 length 20 (octal), 16. (decimal)
OVL3 14261 Entry Relocatable
*************
OVL4 from DSK:OVL4.REL created by FORTRAN on 11-Feb-85 at 11:52:34
Low segment starts at 14300 ends at 14331 length 32 (octal), 26. (decimal)
High segment starts at 14332 ends at 14374 length 43 (octal), 35. (decimal)
OVL4 14333 Entry Relocatable
*************
Index to overlay numbers of OVL0 page 13
Overlay Page Overlay Page Overlay Page Overlay Page
#0 5 #2 10 #3 11 #4 12
#1 6
Index to overlay names of OVL0
Name Page Name Page Name Page Name Page
LEFT 6 LEFT2 11 RIGHT 12 TEST 5
LEFT1 10
[End of LINK map of OVL0]
5.1.2.6 Tree Diagram - The tree diagram requested by the
LPT:TEST/PLOT switch.
5-20
FIELD TEST
5.1.2.7 Executable File - The process also produced an executable
file TEST.EXE, which can be run using the RUN system command.
However, to run the program, the file TEST.OVL must be present,
because it provides the code for the links.
5-21
FIELD TEST
5.2 WRITABLE OVERLAYS
Ordinarily each overlay link built by LINK is copied by the overlay
handler from the OVL file to the address space at runtime. The
contents of any locations that have been modified are lost each time
the overlay link is copied from the OVL file. This can be prevented
by the use of writable overlays.
If a link is specified as writable, the overlay handler copies that
link to a temporary file on disk before overwriting it. Later, when
the copied link is needed, the overlay handler retrieves the link from
the temporary file rather than the OVL file. In this way, any
modified values are preserved. Because writable overlays involve more
file I/O, they are slower than the default (nonwritable) overlays and
should only be used when the program structure and storage
requirements demand dynamic storage in overlay links.
To specify that an overlay is writable, use the FORTRAN SAVE statement
in the program, and specify /OVERLAY:WRITABLE when loading the program
with LINK.
5.2.1 Writable Overlay Syntax
To build a writable overlay, specify the keyword WRITABLE with the
/OVERLAY switch in the LINK command:
filespec/OVERLAY:WRITABLE
5.2.2 Writable Overlay Error Messages
The overlay handler must write and update a temporary file. In
addition to the error messages associated with all overlays, there are
two additional error messages for writable overlays:
? OVLCWF Cannot write file [filename]: [reason]
? OVLCUF Cannot update file [filename]: [reason]
If either of these messages appears, you should check for disk quota
violations or other conditions that could prevent the overlay handler
from writing a temporary file.
5.3 RELOCATABLE OVERLAYS
LINK ordinarily allocates 2000 extra words at the end of the root link
5-22
FIELD TEST
and no extra space at the end of each subsequent link. This is
adequate for programs with static storage requirements. If a link
requires extra storage at run-time, you can use the /SPACE switch to
make the necessary allowances for the program's requirements. The
/SPACE switch allows you to specify the number of words to be
allocated after the current link is loaded.
However, there are programs whose dynamic run-time storage
requirements are unpredictable. For example, a program's run-time
storage requirements may vary according to the program's input. For
this class of programs, relocatable overlays can be useful.
For relocatable overlays LINK places extra relocation information in
the OVL file, permitting overlay links to be relocated at runtime.
The overlay handler, using the FUNCT. subroutine, can determine where
the link will fit in the address space and resolve relocatable
addresses within the link. This extra processing causes relocatable
overlays to run slower than nonrelocatable overlays. Relocatable
overlays should only be used when you cannot determine the dynamic
storage requirements of a program.
5.3.1 Relocatable Overlay Syntax
To build a relocatable overlay, specify the RELOCATABLE keyword to the
/OVERLAY switch in the LINK command:
filespec/OVERLAY:RELOCATABLE
5.3.2 Relocatable Overlay Messages
If /OVERLAY:(LOGFILE,RELOCATABLE) is specified during the loading of a
program, informational messages of the following form are sent to the
your terminal:
%OVLRLL Relocating link [linkname] at [address]
5.4 RESTRICTIONS ON OVERLAYS
The following restrictions apply to all overlaid programs:
o Overlaid programs cannot be run execute-only.
o Overlaid programs cannot use PDVs.
o PSECTed programs cannot be overlaid.
5-23
FIELD TEST
o Overlaid programs with large buffer requirements must use the
/SPACE switch. If an %OVLMAN (Memory not available) error is
encountered, the program should be reloaded using the /SPACE
switch with each link.
o If the program uses more than 256 links, use the /MAXNODE
switch to specify the number of links necessary for the
program. LINK allocates extra space in the the OVL file for
tables that require it, based on the number of links you
specify.
o If the program uses argument checking, the callee argument
checking block must be seen in the same overlay node as the
definition of the global symbol of the same name.
o If the program uses character fixup, the argument block and
the character descriptor must be in the same segment (low or
high), and they must both be either relocatable or absolute.
5.4.1 Restrictions on Absolute Overlays
The following restrictions apply to absolute overlaid programs:
1. Any intermediate results stored in non-root links are lost as
soon as the links are overlaid. Do not expect to retain a
value stored in a non-root link unless /OVERLAY:WRITABLE has
been specified.
2. Certain forms of global, inter-overlay references are not
recommended because you cannot be sure that the necessary
modules will be in memory at the right time. Some of these
references are:
o Additive fixups, in the form ABC##+XYZ where ABC is in
another overlay.
o Left-hand fixups, in the form XWD ABC##,XYZ, where ABC is
in another overlay.
o Fullword fixups, in the form EXP ABC##, where ABC is in
another overlay.
o Similarly, MOVEI 1,ABC##, where ABC is in a different
overlay, should not be used, because the necessary module
may not be in memory.
In fact, the only predictable inter-overlay global reference
is one that brings the necessary module into memory, such as
PUSHJ P,ABC##.
5-24
FIELD TEST
5.4.2 Restrictions on Relocatable Overlays
The following restriction applies to relocatable overlays:
o Complex expressions involving relocatable symbols are not
relocated properly in a relocatable overlay. No standard
DIGITAL compiler produces such expressions. MACRO
programmers should avoid using them in modules that are to be
loaded as part of an overlaid program. Any expression that
causes MACRO to generate a Polish fixup block is not properly
relocated at runtime. The following are examples of such
complex expressions:
MOVEI 1,A##+B##+C##
A,,0
5.4.3 Restrictions on FORTRAN Overlays
The following restrictions apply to FORTRAN programs that are written
with associate variables and using the overlay facility.
o If the associate variable is declared in a subroutine, that
subroutine must be loaded in the root link of the overlay
structure. Accessing a file opened with an associate
variable changes the value of the specified variable. If
this variable is in a nonresident overlay link when the
access is made, program execution produces unpredictable
results. Moreover, the value of the variable is reset to
zero each time its overlay link is removed from memory. Only
variables declared in routines that are loaded into the root
link are always resident. However, variables declared in
COMMON and in the root link are always resident, and may be
safely used as associate variables.
o If you place COMMON in a writable overlay, be sure that all
references to the variables in that COMMON are in the same
overlay or its successors.
o A FORTRAN ASSIGN statement may be used in a relocatable
overlay. If the ASSIGN is made in a subroutine, the value of
the assigned variable may be preserved from one call of that
subroutine to the next. However, the overlay containing that
subroutine could then be replaced in memory by a different
overlay. If the overlay containing the subroutine is
relocated differently when brought back into memory, any
subsequent GOTO may fail.
5-25
FIELD TEST
5.5 SIZE OF OVERLAY PROGRAMS
Although most programs have a consistent size, the size of an overlay
program depends on which overlays are in memory. This can be
ascertained by using the /COUNTER switch when linking the program. To
do this, place /COUNTER after the /LINK switch for the overlay of
which you want to know the size, but before the next /NODE switch.
For example,
*OVLAY/LINK
*/COUNTER
*/NODE:TOP
This gives you the size of the program when the overlay is actually
loaded into memory. The display includes all routines loaded from the
runtime libraries. This allows you to determine which overlay is the
largest, and whether the program can be loaded without restructuring.
5.6 DEBUGGING OVERLAID PROGRAMS
COBDDT and ALGDDT can be used to debug overlay programs, but FORDDT
cannot. To use DDT with an overlaid program, the program should be
loaded using /SYMSEG:LOW, with local symbols for the desired modules.
To set breakpoints in an overlay, put a subroutine in the root node,
and call the subroutine from the overlay. Such a subroutine need
consist only of a SUBROUTINE statement, a RETURN, and an END. The
breakpoint can be set at this subroutine before the program starts
running.
RESET. in FOROTS removes the symbol table when a FORTRAN program
starts running. The symbol table will return after the first overlay
is called. If you need the symbols for debugging the root link,
insert a CALL INIOVL at the beginning of the main program (refer to
Section 5.7.1 for more information). This call will reinstall the
symbol table. LINK builds a separate symbol table for each overlay,
so that all the symbols known to DDT are for modules that are
currently in memory. Note that it is not possible to single-step
through RESET. The DDT commands $X and $$X do not work. Set a
breakpoint after RESET. if you are debugging a root link, and use the
DDT command $G.
5.7 THE OVERLAY HANDLER
LINK's overlay handler is the program that supervises execution of
overlay structures defined by LINK switches.
The overlay handler is in the file SYS:OVRLAY.REL. Some installations
5-26
FIELD TEST
will install LINK Version 5 or later without the overlay handler that
was shipped with it. To find the version of the overlay handler, type
the following:
@LINK<RET>
*SYS:OVRLAY<RET>
*/VALUE:%OVRLA<RET>
[LNKVAL Symbol %OVRLA 501000210 defined]
*
The left halfword of $OVRLA contains the version number of the overlay
handler, and should be 501, corresponding to the Version 6 of LINK.
The right halfword is the edit number, and should be 000210 if field
image, or greater if edits have been installed.
When you load an overlay structure, the overlay handler is loaded into
the root link of the structure. From there it can supervise
overlaying operations, because the root link is always in your virtual
address space during execution. During execution, when a link not in
memory is called, the overlay handler brings in the link, possibly
overlaying one or more links already in memory. The overlay handler
consists of self-modifying code and data, and two 128-word buffers.
One of these buffers, IDXBFR, contains a 128-word section of the link
number index table. This allows 256 links to be directly referenced
at any one time. The second buffer, INBFR, contains the preambles and
relocation tables, if required, of the individual links.
There are two ways of overlaying links during execution:
1. A call to a link not in memory implicitly calls the overlay
handler to overlay one or more links with the required links.
This action of the overlay handler is transparent to the
user.
2. An explicit call to one of several entry points in the
overlay handler can cause one or more links to be overlaid.
These entry points and calls to them are discussed in the
sections below.
5.7.1 Calls to the Overlay Handler
Overlays can be used transparently, or they can be explicitly called
from the program. Such calls are made to one of the entry points in
the overlay handler.
The overlay handler has five entry points that are available for calls
from user programs. To call the overlay handler from a MACRO program,
you must use the standard calling sequence, which is:
MOVEI 16,arglst
5-27
FIELD TEST
PUSHJ 17,entry-name
Where arglst is the address of the first argument in the argument
list, and entry-name is the entry-point name.
The argument list must be of the form:
-n,,0 ;n is number of arguments
arglst: Z code,addr1 ;For first argument
.
.
.
Z code,addrn ;For nth argument
Where addr... is the address of the argument.
The legal values of "code" are 2 (for a link number), 17 (for an ASCIZ
string), and 15 (for a character string descriptor).
For each word of the argument list, the code indicates the type of
argument. The code occupies the AC field, bits 9 through 12. The
address gives the location of the argument; it can be indirect and
indexed.
To call the overlay handler from a FORTRAN program, the call must be
of the form:
CALL subroutine (arglst)
Where subroutine is the name of the desired subroutine, and arglst is
a list of arguments separated by commas.
5.7.2 Overlay Handler Subroutines
Each of the seven callable subroutines in the overlay handler has an
entry name symbol for use with MACRO, and a subroutine name for use
with FORTRAN, as follows:
5-28
FIELD TEST
MACRO Entry FORTRAN Subroutine
Name Symbol Subroutine Function
CLROV. CLROVL Specifies a non-writable overlay.
GETOV. GETOVL Brings specified links into memory.
INIOV. INIOVL Specifies the file from which the overlay
program is read, if the load time
specification is to be overridden.
LOGOV. LOGOVL Specifies or closes the file in which
runtime messages from the overlay handler
are written.
REMOV. REMOVL Removes specified links from memory.
RUNOV. RUNOVL Moves into memory a specified link and
begins execution at its start address.
SAVOV. SAVOVL Specifies a writable overlay.
Declaring a Non-Writable Link (CLROV.)
You can declare an overlay link to be non-writable, using the CLROV.
entry point. This does not immediately affect the program, but waits
until the link is about to be overlaid or read in. If the link is
already non-writable, this entry point has no effect.
5-29
FIELD TEST
Example
MOVEI 16,arglst
PUSHJ 17,CLROV.
-n,,0 ;n is number of arguments
arglst: Z 17,addr1 ;for first ASCIZ linkname
.
.
.
Z 17,addrn ;for nth ASCIZ linkname
OR
-n,,0 ;n is number of arguments
arglst: Z 2,addr1 ;for first link number
.
.
.
Z 2,addrn ;for nth link number
Where addr... is the address of the argument.
Getting a Specific Path (GETOV.)
The subroutine to bring a specific path into core can be used to make
sure that a particular path is used when otherwise the overlay handler
might have a choice of paths. It is illegal to specify a path that
overlays the calling link.
To call the subroutine from a FORTRAN program, use:
CALL GETOVL (linkname,...,linkname)
where each linkname is the ASCIZ name of a link in the desired path.
To call the subroutine from a MACRO program, use the standard FORTRAN
calling sequence:
MOVEI 16,arglst
PUSHJ 17,GETOV.
The argument list has one word for each link required to be in the
path.
Example
-n,,0 ;n is number of arguments
arglst: Z 17,addr1
.
.
.
5-30
FIELD TEST
Z 17,addrn
OR
-n,,0 ;n is number of arguments
arglst: Z 2,addr1
.
.
.
Z 2,addrn
Where addr... is the address of the argument.
Initializing an Overlay (INIOV.)
The overlay initializing subroutine specifies a file from which the
overlay program is read. This subroutine is used to override the file
specified at load time. The file specified to INIOV. can have any
valid specification, but it must be in the correct format for an
overlay (OVL) file.
To call the subroutine from a FORTRAN program, use:
CALL INIOVL ('filespec')
where `filespec' is a literal constant that can give a 39-character
device, a 39-character filename, a 39-character file type, and a
39-character directory.
To call the subroutine from a MACRO program, use the standard FORTRAN
calling sequence:
MOVEI 16,arglst
PUSHJ 17,INIOV.
The argument list is of the form:
-1,,0
arglst: Z 17,address of ASCIZ filespec
where filespec is an ASCIZ string (ASCII ending with nulls) that can
give a 39-character device, a 39-character filename, a 39-character
file type, and a 39-character directory.
NOTE
If you call INIOV. with no arguments, it initiates the
overlay handler and reads in the symbols for the root
link, using the overlay file specified at load time.
This can be useful for debugging the root link before
any successor links have been read in, because symbols
are not normally available until the first link comes
5-31
FIELD TEST
into memory.
Specifying an Overlay Log File (LOGOV.)
You can specify an output file for runtime messages from the overlay
handler. These messages are listed in Section 5.5. The log file
entry includes the elapsed run time since the first call to the
overlay handler.
To call this subroutine from a FORTRAN program, use:
CALL LOGOVL ('filespec')
where `filespec' is a literal constant that can give a device, a
filename, a file type, and a PPN.
To close the file, use
CALL LOGOVL (0)
You can omit the (0) argument to close the file.
To call the subroutine from a MACRO program, use the standard FORTRAN
calling sequence:
MOVEI 16,arglst
PUSHJ 17,LOGOV.
The argument list is of the form:
-1,,0
arglst: Z 17,address of ASCIZ filespec
Where filespec is an ASCIZ string that can give a device, a filename,
a file type, and a PPN.
To close the log file, the argument list is:
-1,,0
arglst: Z 17,address of word containing zero
Removing Specific Links from Memory (REMOV.)
The subroutine to remove specific links from memory, once they are no
longer required, can be used to reduce core image size for faster
execution. Specifying removal of the calling link causes an error.
To call the subroutine from a FORTRAN program, use:
CALL REMOVL (linkname,...,linkname)
5-32
FIELD TEST
Where each linkname is the ASCIZ name of a link to be removed from
memory.
To call the subroutine from a MACRO program, use the standard FORTRAN
calling sequence:
MOVEI 16,arglst
PUSHJ 17,REMOV.
The argument list has one word for each link to be removed.
Example
-n,,0 ;n is number of arguments
arglst: Z 17,addr1
.
.
.
Z 17,addrn
OR
-n,,0 ;n is number of arguments
arglst: Z 2,addr1
.
.
.
Z 2,addrn
Where addr... is the address of the argument.
Running a Specific Link (RUNOV.)
The subroutine for running a specific link allows you to transfer
program execution to the start address of a particular link. (An
error occurs if the link has no start address.) If the link is not
already in memory, it and its path are brought in.
You can use this subroutine to overlay the calling link, because the
next instruction executed is the start address of the named link;
therefore, there is no automatic return to the calling link.
NOTE
The FORTRAN compiler does not generate start addresses
for subroutines. FORTRAN main programs cannot be
loaded into non-root links. Therefore, to use RUNOVL
to transfer control to a FORTRAN subroutine in a
non-root link, you must use the /START switch at load
time to define a start address for the link.
5-33
FIELD TEST
To call the subroutine RUNOVL from a FORTRAN program, use:
CALL RUNOVL (linkname)
Where linkname is the ASCIZ name of the link to be run.
To call the subroutine from a MACRO program, use the standard FORTRAN
calling sequence:
MOVEI 16,arglst
PUSHJ 17,RUNOV.
The argument list is of the form:
-1,,0
arglst: Z 17,address of ASCIZ linkname
OR
-1,,0
arglst: Z 2,address of link number
Declaring A Writable Link (SAVOV.)
You can dynamically declare an overlay link to be writable by calling
SAVOV. This does not affect the current state of the code
immediately, but waits until the link is about to be overlaid. If the
link is already writable, this symbol has no effect.
Example
MOVEI 16,arglst
PUSHJ 17,SAVOV.
-n,,0 ;n is number of arguments
arglst: Z 17,addr1 ;for first ASCIZ linkname
.
.
.
Z 17,addrn ;for nth ASCIZ linkname
OR
-n,,0 ;n is number of arguments
arglist: Z 2,addr1 ;for first link number
.
.
.
Z 2, addrn ;for nth link number
5-34
FIELD TEST
Where addr... is the address of the argument.
If called with no arguments, SAVOV. only initializes the temporary
file.
5.7.3 Overlay Handler Messages
This section lists all of the overlay handler's messages. LINK
messages, which have the LNK prefix, are given in Appendix B.
For each overlay handler message, the last three letters of the
six-letter code, the severity, and the text of the message are given
in boldface. Then, in lightface type, comes an explanation of the
message.
When a message is issued, the three letters are suffixed to the
letters OVL, forming a 6-letter code of the form OVLxxx. The
explanation of the message is printed only if you use the /OVERLAY:LOG
switch.
The severity of a message determines whether the job is terminated
when the message is issued. Level 31 messages terminate program
execution. Level 8 messages are warnings: they do not terminate
execution, but the error may affect the execution of the program.
Level 1 messages are informational and are printed on the terminal
only if you specified /OVERLAY:LOGFILE.
5-35
FIELD TEST
Code Sev Message and Explanation
ARC 31 Attempt to remove caller from link [name or number]
The named link attempted to remove the link that called
it. This error occurs when the call to the REMOV.
subroutine requests removal of the calling link.
ARL 8 Ambiguous request in link number [number] for [symbol],
using link number [number]
More than one successor link satisfies a call from a
predecessor link, and none of these successors is in
memory. Since all their paths are of equal length, the
overlay handler has selected an arbitrary link.
CDL 31 Cannot delete link [name or number], FUNCT. return status
[number]
This is an internal LINK error, and is not expected to
occur. If it does, please notify your Software
Specialist, or send a Software Performance Report (SPR) to
DIGITAL.
Return status is one of the following:
1 Core already deallocated
3 Illegal argument passed to FUNCT. module
CGM 31 Cannot get memory from OTS, FUNCT. return status [octal]
The system does not have enough free memory to load the
link. The status is returned from the object-time system
and depends on the particular FUNCT. function the overlay
handler used. See Section 5.7.4 for the FUNCT. function
codes and status messages.
CRF 31 Cannot read file [file] [reason]
An error occurred when reading the overlay file. The file
was closed after the last successful read operation.
CSM 31 Cannot shrink memory, FUNCT. return status [octal]
A request to the object-time system to reduce memory, if
possible, failed. This error is not expected to occur.
If it does, please notify your Software Specialist or send
a Software Performance Report (SPR) to DIGITAL.
CUF 31 Cannot update file [file] [reason]
An error occurred when updating the TMP file into which
5-36
FIELD TEST
non-resident writable overlay links are written.
CWF 31 Cannot write file [file] [reason]
An error occurred when creating the TMP file used to store
non-resident writable overlay links.
DLN 1 Deleting link [name or number] after [hh:mm:ss]
The named link has been removed from memory as a result of
a call to the REMOV. subroutine. The time given is
elapsed time since the first call to the overlay handler.
This message is output only to the overlay log file, if
any.
IAT 31 Illegal argument type on call to [subroutine]
A user call to the named overlay handler subroutine gave
an illegal type of argument.
IEF 31 Input error for file [file] [reason]
An error occurred while reading the OVL or TMP file.
ILN 31 Illegal link number [number]
A user call to one of the overlay handler subroutines gave
an illegal link number as an argument.
IMP 31 Impossible error condition at PC=[address]
This is an internal error caused by monitor call error
returns that should not occur. This message is issued
instead of the HALT message. This error is not expected
to occur. If it does, please notify your Software
Specialist or send a Software Performance Report (SPR) to
DIGITAL.
IPE 31 Input positioning error for file [file] [reason]
An error occurred while reading the OVL or TMP file.
IVN 8 Inconsistent version numbers
The OVL and EXE files found were not created at the same
time, and may not be compatible.
LNM 31 Link number [decimal] not in memory
A call to the REMOV. subroutine has removed the named
link from memory. It must be restored by a call to GETOV.
or RUNOV.
5-37
FIELD TEST
MAN 31 Memory not available for absolute [link], FUNCT. return
status [octal]
There is not enough room for the overlay handler to load
the specified link into the part of memory the link was
built for. Two options are available: a) Use the /SPACE
switch at load time to reserve more space for the link, or
b) Build a relocatable overlay using the RELOCATABLE
option to the /OVERLAY switch at load time.
MEF 31 Memory expansion failed, FUNCT. return status [octal]
The overlay handler was unable to get free space from the
memory manager. Restructure your overlay so that the
minimum number of links are in memory at any time.
NMS 8 Not enough memory to load symbols, FUNCT. return status
[octal]
There was not enough free space available to load symbols
into memory.
NRS 31 No relocation table for symbols
A relocation table was not included for the symbol table.
It is possible that LINK failed to load the relocation
table because there wasn't enough room in memory.
NSA 31 No start address for link [name or number]
A user call to the RUNOV. subroutine requests execution to
continue at the start address of the named link, but that
link has no start address.
NSD 31 No such device for [file]
An invalid device was specified.
OEF 31 Output error for file [file] [reason]
An error occurred when writing the overlay file. The file
was closed after the last successful write operation.
OPE 31 Output positioning error for file [file] [reason]
An error occurred while writing the TMP file used to hold
non-resident writable overlay links.
OPP 31 Overlay handler in private page
The overlay handler has been loaded into a non-sharable
page of the program. Your program should not write into
5-38
FIELD TEST
the pages occupied by the overlay handler. Ask LINK for a
map of your program if there is doubt. If the program is
not writing into these pages, this error may reflect an
internal LINK error. This error is not expected to occur.
If it does, please notify your Software Specialist or send
a Software Performance Report (SPR) to DIGITAL.
RLL 1 Relocating link [name or number] at [address]
The named relocatable link has been loaded at the given
address. This message is output only to the overlay log
file.
RLN 1 Reading in link [name or number] after [time]
The named link has been loaded. The time given is elapsed
time since the first call to the overlay handler. This
message is output only to the overlay log file.
RMP 31 RMAP JSYS failed
This is an internal error and is not expected to occur.
If it does, please notify your Software Specialist or send
a Software Performance Report (SPR) to DIGITAL.
RPA 31 RPACS JSYS failed
This is an internal error and is not expected to occur.
If it does, please notify your Software Specialist or send
a Software Performance Report (SPR) to DIGITAL.
STS 8 OTS reserved space too small
The object-time system does not have space for its minimum
number of buffers. Reload, using the /SPACE switch for
the root link with an argument greater than 2000 (octal).
ULN 31 Unknown link name [name]
A call to one of the overlay handler subroutines gave an
invalid link name as an argument. Correct the call.
USC 8 Undefined subroutine [name] called from [address]
A required subroutine was not loaded. The instruction at
the given program counter address calls for an undefined
subroutine. Correct the call or load the required
subroutine.
WLN 1 Writing [link] after [time]
The overlay handler is writing out a writable overlay
5-39
FIELD TEST
link.
5.7.4 The FUNCT. Subroutine
Each DIGITAL-supplied object-time system has a subroutine that the
overlay handler uses for memory management, I/O, and message handling.
This subroutine has a single entry point, FUNCT., and is called by the
sequence:
MOVEI 16,arglst
PUSHJ 17,FUNCT.
The format of the argument list is:
-<n+3>,,0
arglst: Z 2,address of integer function code
Z 2,address for error code on return
Z 2,address for status code on return
Z code,address of first argument
.
.
.
Z code,address of nth argument
Where function code is one of the function codes described below;
error code is a 3-letter ASCII mnemonic output by the object-time
system (after ?, %, or [); and status (on return) contains one of the
following values:
-1 Function not implemented
0 Successful return
n Number of the error message
Most object-time systems allocate separate space for their own use and
for the use of the overlay handler. This minimizes the possibility
that the overlay handler requests space that the object-time system is
already using.
The permitted function code arguments, their names, and their meanings
are:
Code Name Function
0 ILL Illegal function; returns -1 status.
1 GAD Get a specific segment of memory.
2 COR Get a given amount of memory from anywhere in the
space allocated to the overlay handler.
5-40
FIELD TEST
3 RAD Return a specific segment of memory.
4 GCH Get an I/O channel.
5 RCH Return an I/O channel.
6 GOT Get memory from the space allocated to the
object-time system.
7 ROT Return memory to the object-time system.
10 RNT Get the initial runtime, in milliseconds, from the
object-time system.
11 IFS Get the initial runtime file specification of the
program being run.
12 CBC Cut back core (if possible) to reduce job size.
13 RRS Read retain status (DBMS)
14 WRS Write retain status (DBMS)
15 GPG Get pages
16 RPG Return pages
17 GPSI Get TOPS-20 PSI channel
20 RPSI Return TOPS-20 PSI channel
21 MPG Get contiguous set of pages
All FUNCT. codes are reserved to DIGITAL.
The following subsections describe each function of the
FUNCT. subroutine (except the reserved functions).
ILL (0) Function
This function is illegal. The argument list is ignored, and the
status returned is -1.
GAD (1) Function
The GAD function gets memory from a specific address in the space
allocated to the overlay handler. The argument list points to:
arg 1 Address of requested memory
arg 2 Size of requested allocation (in words)
5-41
FIELD TEST
A call to GAD with arg 2 equal to -1 requests all available memory.
On return, the status is one of the following:
0 Successful allocation
1 Not enough memory available
2 Memory not available at specified address
3 Illegal arguments (address + size > 256K)
COR (2) Function
The COR function gets memory from any available space allocated to the
overlay handler. The arguments are:
arg 1 Undefined (address of allocated memory on return)
arg 2 Size of requested allocation
On return, the status is:
0 Core allocated
1 Not enough memory available
3 Illegal argument (size > 256K)
RAD (3) Function
The RAD function returns the memory starting at the specified address
to the overlay handler. The arguments are:
arg 1 Address of memory to be returned
arg 2 Size of memory to be returned (in words)
On return, the status is one of the following:
0 Successful return of memory
1 Memory cannot be returned
3 Illegal argument (address or size > 256K)
GCH (4) Function
Returns a status of 1. The channel is not available.
RCH (5) Function
Returns a status of 1. The channel is not available.
GCH (4) Function
5-42
FIELD TEST
The GCH function gets an input/output channel. The arguments are:
arg 1 Undefined (channel number allocated on return)
arg 2 Ignored
On return, the status is one of the following:
0 Successful channel allocation
1 No channels available
RCH (5) Function
The RCH function returns an input/output channel. Its arguments are:
arg 1 Number of channel to be returned
arg 2 Ignored
On return, the status is one of the following:
0 Channel released
1 Channel number invalid for user
GOT (6) Function
The GOT function gets memory from the space allocated to the
object-time system. Its arguments are:
arg 1 Undefined (address of allocated memory on return)
arg 2 Size of memory requested
On return, the status is one of the following:
0 Successful allocation
1 Not enough memory available
3 Illegal argument (size > 256K)
ROT (7) Function
The ROT function returns memory to the object-time system. Its
arguments are:
arg 1 Address of memory to be returned
arg 2 Size of memory to be returned (in words)
On return, the status is one of the following:
0 Successful return of memory
1 Memory cannot be returned
3 Illegal argument (address or size > 256K)
5-43
FIELD TEST
RNT (10) Function
The RNT function returns the initial runtime, in milliseconds, from
the object-time system. (At the beginning of the program, the
object-time system will have executed a RUNTIM UUO; the result is the
time returned by RNT.) Its arguments are:
arg 1 Undefined (contains initial runtime on return)
arg 2 Ignored
On return, the runtime is in arg 1, and the status is 0. The status
is 0.
IFS (11) Function
Always returns a value of -1. This function is not implemented.
CBC (12) Function
The CBC function cuts back memory if possible, which reduces the size
of the job. It uses no arguments, and the returned status is 0.
RRS (13) Function ( Reserved for DBMS )
Returns ARG1 = 0. On return, the status is always 0.
WRS (14) Function ( Reserved for DBMS )
Returns ARG1 = 0. On return, the status is always 0.
GPG (15) Function
The GPG function is used to fetch a page. The arguments are:
arg2: size to be allocated, in words
On return,
arg1 = address of allocated memory, on page boundary
and the status is one of the following:
0 if allocated OK
1 if not enough memory
3 if argument error
RPG (16) Function
5-44
FIELD TEST
The RPG function is used to return pages. The arguments are:
arg1: address (a word)
arg2: size (in words)
On return, the status is:
0 if deallocated OK
1 if wasn't allocated
3 if argument error
GPSI (17)
The GPSI function can be used to get a PSI channel for programs
running in a TOPS-20 environment. This entry point provides only
controlled access to the PSI tables. GPSI arranges that the tables
exist and that SIR and EIR have been done but does not do AIC or any
other JSYS necessary to set up the channel (ATI or MTOPR, for
example).
The arguments are:
arg1: channel number,
or -1 to allocate any user-assignable channel
arg2: level number
arg3: address of interrupt routine
On return, arg1 contains the channel number allocated (if -1 was
originally specified). On return, the status is:
0 if OK
1 if channel was already assigned
2 if no free channels
3 if argument error
NOTE
This function is used by TOPS-20 programs. It is a
reserved function in the TOPS-10 environment.
RPSI (20) Function
This entry point provides only controlled access to the PSI tables.
It does not do DIC or any other JSYS necessary to release a channel.
It just clears the level and interrupt address fields in CHNTAB.
This function accepts the following argument:
arg1: channel number
On return the status is one of the following:
0 if OK
1 if channel wasn't in use
5-45
FIELD TEST
3 if argument error
NOTE
This function is used by TOPS-20 programs. It is a
reserved function in the TOPS-10 environment.
MPG (21) Function
This function gets a contiguous set of pages. The pages requested are
always allocated from the section FOROTS is in. The user cannot
depend upon this call to either create or destroy the pages.
arg 1 first page number to allocate. The page number must be
in the range 0 to 777.
arg 2 number of pages to allocate
On return, the status is one of the following:
0 successful deallocation of all given pages
1 one or more pages was not allocated through MPG
3 illegal argument (bad page number count)
5-46
FIELD TEST
5.8 THE OVERLAY (OVL) FILE
This section contains diagrams of the contents of the overlay file
output by LINK as a result of the /OVERLAY switch. The following
diagram shows the overall scheme of the file:
Scheme of the Overlay (OVL) File
|=======================================================|
| |
| Directory Block |
| |
|=======================================================|
| |
| Link Number Table |
| |
|=======================================================|
| |
| Link Name Table |
| |
|=======================================================|
| |
| Writable Link Flags Table |
| |
|=======================================================|
| |
| Link |
| |
|-------------------------------------------------------|
.
.
.
|-------------------------------------------------------|
| |
| Link |
| |
|=======================================================|
5-47
FIELD TEST
5.8.1 The Directory Block
The following diagram shows the contents of the Directory Block:
Directory Block
|=======================================================|
.DIHDR: | 0 (Reserved) | Length of Directory Block |
|-------------------------------------------------------|
.DIRGN: | 0 (Reserved) |
|-------------------------------------------------------|
.DIVER: | Version Number of Corresponding EXE file |
|-------------------------------------------------------|
.DILPT: | -(Size of Link No. Table) |Link Number Table Block No.|
|-------------------------------------------------------|
.DINPT: |-(Size of Link Name Table) | Link Name Table Block No. |
|-------------------------------------------------------|
.DIWPT: |-(Size of Writable Flg Tbl)| Writable Flg Tbl Block No |
|-------------------------------------------------------|
.DIFLG: | Flags |
|-------------------------------------------------------|
| 0 (Reserved) |
|=======================================================|
In the fourth word above, the size of the Link Number Table (in words)
is half the number of links (rounded upward); the Link Number Table
Block No. is the number of the 128-word disk block containing the Link
Number Table. (There are four disk blocks per disk page.)
In the fifth word above, the size of the Link Name Table (in words) is
twice the number of links; the Link Name Table Block No. is the number
of the 128-word disk block containing the Link Name Table.
The table defined by the .DIWPT word above consists of a string of
two-bit bytes. The first bit, OW.WRT, indicates whether the
corresponding overlay link is writable. This bit is set under the
control of a REL block of type 1045 (writable links). The second bit,
OW.PAG, indicates whether the corresponding overlay link is currently
paged into the runtime overlay temporary file. This is strictly a
run-time flag and should be zero in the overlay file. This flag is
defined in the overlay file to allow the overlay handler to set up its
flag table with a single read operation.
The .DIFLG word in the directory block contains a single bit flag (bit
0). If this bit is set, the overlay file contains at least one
writable overlay. This information is also contained in the Writable
Link Table. However, by having the information available in the
directory block the overlay handler can determine if any links are
writable without scanning the Writable Link Table. All other bits in
the .DIFLG word are reserved and must be zero.
5-48
FIELD TEST
NOTE
If you request both writable and relocatable overlays,
only halfwords known to be relocatable at load time
will be correctly relocated when the link is
refetched.
5.8.2 The Link Number Table
The following diagram shows the contents of the Link Number Table:
Link Number Table
|=======================================================|
| Pointer to Link 0 | Pointer to Link 1 |
|-------------------------------------------------------|
| Pointer to Link 2 | Pointer to Link 3 |
|-------------------------------------------------------|
.
.
.
|-------------------------------------------------------|
| Pointer to Link n-1 | Pointer to Link n |
|=======================================================|
Each pointer is a disk block number. Any unused words in the last
disk block of the Link Number Table are zeros.
5.8.3 The Link Name Table
The following diagram shows the contents of the Link Name Table:
Link Name Table
|=======================================================|
| Link Number |
|-------------------------------------------------------|
| SIXBIT Link Name |
|=======================================================|
.
.
.
|=======================================================|
| Link Number |
5-49
FIELD TEST
|-------------------------------------------------------|
| SIXBIT Link Name |
|=======================================================|
Any unused words in the last disk block of the Link Name Table are
zeros.
5.8.4 The Overlay Link
The following diagram shows the overall scheme of each overlay link in
the overlay file:
Scheme of an Overlay Link
|=======================================================|
| |
| Preamble |
| |
|=======================================================|
| |
| Code for Link |
| |
|=======================================================|
| |
| Link Control Section |
| |
|=======================================================|
| |
| EXTTAB |
| |
|=======================================================|
| |
| INTTAB |
| |
|=======================================================|
| |
| Relocation Table |
| |
|=======================================================|
| |
| Other Relocation Tables |
| |
|=======================================================|
5-50
FIELD TEST
The Preamble
The following diagram shows the contents of the preamble for an
overlay link:
Preamble
|=======================================================|
| 0 (Reserved) | Length of Preamble |
|-------------------------------------------------------|
| 0 (Reserved) | 0 (Reserved) |
|-------------------------------------------------------|
| 0 (Reserved) | Link Number |
|-------------------------------------------------------|
| SIXBIT Link Name |
|-------------------------------------------------------|
|Pointer to List of Bound Links Starting with Root Link |
|-------------------------------------------------------|
| Pointer to List of Bound Links Ending with Root Link |
|-------------------------------------------------------|
| Equivalence Pointer |
|-------------------------------------------------------|
| Address of Control Section |
|-------------------------------------------------------|
| Flags |
|-------------------------------------------------------|
| Absolute Address at Which Link Loaded |
|-------------------------------------------------------|
| Length of Link (Code through INTTAB) |
|-------------------------------------------------------|
| Disk Block Number of Start of Link Code |
|-------------------------------------------------------|
| 0 (Reserved) |
|-------------------------------------------------------|
| Disk Block Number of Relocation Table |
|-------------------------------------------------------|
| Disk Block Number of Other Relocation Tables |
|-------------------------------------------------------|
| 0 (Reserved) |
|-------------------------------------------------------|
| Disk Block Number of Radix-50 Symbols |
|-------------------------------------------------------|
|Block Number of Relocation Tables for Radix-50 Symbols |
|-------------------------------------------------------|
| Next Free Memory Location for Next Link |
|=======================================================|
5-51
FIELD TEST
Code for the Link
The code for each link consists of a core image that was constructed
from the REL files placed in the link. This core image contains the
code and data for the link.
The Control Section
The following diagram shows the contents of the Control Section:
Control Section
|=======================================================|
| 0 (Reserved) | Length of Header |
|-------------------------------------------------------|
| 0 (Reserved) | 0 (Reserved) |
|-------------------------------------------------------|
| 0 (Reserved) | Link Number |
|-------------------------------------------------------|
| SIXBIT Link Name |
|-------------------------------------------------------|
| Ptr to Ancestor in Core | Ptr to Successor in Core |
|-------------------------------------------------------|
| -(Length of Symbol Table) | Address of Symbol Table |
|-------------------------------------------------------|
| 0 (Reserved) | Start Address for Link |
|-------------------------------------------------------|
|Memory Needed to Load Link | First Address in Link |
|-------------------------------------------------------|
| -(Length of EXTTAB) | Pointer to EXTTAB |
|-------------------------------------------------------|
| -(Length of INTTAB) | Pointer to INTTAB |
|-------------------------------------------------------|
| Address of Symbols on Disk |
|-------------------------------------------------------|
| Relocation Address |
|-------------------------------------------------------|
| Copy of Block Number for Code |
|-------------------------------------------------------|
|-(Length of Radix-50 SymTab)|Blk No. of Radix-50 SymTab|
|=======================================================|
Ptr = Pointer
5-52
FIELD TEST
The EXTTAB Table
The following diagram shows the contents of the EXTTAB table:
EXTTAB
|=======================================================|
| JSP 1,.OVRLA |
|-------------------------------------------------------|
| Flags |Address of Callee's INTTAB |
|-------------------------------------------------------|
| Callee's Link Number |Ptr to Callee's Control Sec|
|-------------------------------------------------------|
| Backward Pointer | Forward Pointer |
|=======================================================|
.
.
.
|=======================================================|
| JSP 1,.OVRLA |
|-------------------------------------------------------|
| Flags |Address of Callee's INTTAB |
|-------------------------------------------------------|
| Callee's Link Number |Ptr to Callee's Control Sec|
|-------------------------------------------------------|
| Backward Pointer | Forward Pointer |
|=======================================================|
Ptr = Pointer
The flags in the left half of the second word have the following
meanings:
Bit Meaning (if bit is on)
0 Module is in core.
1 Module is in more than one link.
2 Relocatable link is already relocated.
5-53
FIELD TEST
The INTTAB Table
The following diagram shows the contents of the INTTAB table:
INTTAB
|=======================================================|
| 0 (Reserved) | Address of Entry Point |
|-------------------------------------------------------|
| 0 (Reserved) | Forward Pointer |
|=======================================================|
.
.
.
|=======================================================|
| 0 (Reserved) | Address of Entry Point |
|-------------------------------------------------------|
| 0 (Reserved) | Forward Pointer |
|=======================================================|
5-54
FIELD TEST
The Relocation Table
The following diagram shows the contents of the Relocation Table:
Relocation Table
|=======================================================|
| Relocation Word |
|-------------------------------------------------------|
.
.
.
|-------------------------------------------------------|
| Relocation Word |
|=======================================================|
The Relocation Table contains one bit for each halfword of the link.
If the bit is on, the halfword is relocatable; if it is off, the
halfword is not relocatable.
The first word contains the relocation bits for the first 22 (octal)
words of the link; the second word contains the relocation bits for
the next 22 (octal) words; and so forth for all words in the link.
This table exists only when relocatable overlays are requested with
the /OVERLAY:RELOCATABLE switch.
5-55
FIELD TEST
The Other Relocation Tables
The following diagram shows the contents of the Other Relocation
Tables:
Other Relocation Tables
|=======================================================|
| Number of Words Following for This Link |
|-------------------------------------------------------|
| Link Number | Planned Load Address |
|-------------------------------------------------------|
| Relocation Halfword | Ptr to Words of Code |
|-------------------------------------------------------|
.
.
.
|-------------------------------------------------------|
| Relocation Halfword | Ptr to Words of Code |
|=======================================================|
.
.
.
|=======================================================|
| Number of Words Following for This Link |
|-------------------------------------------------------|
| Link Number | Planned Load Address |
|-------------------------------------------------------|
| Relocation Halfword | Ptr to Words of Code |
|-------------------------------------------------------|
.
.
.
|-------------------------------------------------------|
| Relocation Halfword | Ptr to Words of Code |
|=======================================================|
Ptr = Pointer
This table exists only when relocatable overlays have been requested
with the OVERLAY/RELOCATABLE switch. The Other Relocation Tables are
used to hold internal LINK references.
5-56
CHAPTER 6
PSECTs
PSECTs (Program SECTions) are programmer- or system-defined regions of
code and data that LINK relocates in memory. PSECTs are used for two
reasons:
o To load extended addressing programs.
o To structure a program's memory space.
6.1 LOADING PROGRAMS WITH PSECTs
When loading programs with PSECTs (including extended addressing
programs), you must specify the origin of the PSECT. LINK then uses
this PSECT origin to store the data in the PSECTs.
To specify a PSECT origin, include the origin in the source program or
use the LINK /SET switch. See the appropriate language manual for
including the origin in the source program and Chapter 3 for the /SET
switch.
Defining an upper bound is also important when loading PSECTs. The
LINK /LIMIT switch defines an upper bound for a PSECT. If the PSECT
loads to this bound, LINK returns a warning and an error message.
Despite these messages, LINK continues to process input files and to
load code. The warning is:
%LNKPEL PSECT [psect] exceeded limit of [address]
Although LINK does continue to process input files and load code, the
program is incomplete and should not be used. LINK does produce the
following fatal error:
?LNKCFS Chained fixups have been suppressed
Chained fixups are a method that LINK uses to resolve symbol
references. See the Glossary for a definition of chained fixups.
6-1
FIELD TEST
Using /LIMIT to define an upper bound prevents unintended PSECT
overlaps. PSECT overlaps can cause LINK to loop and produce other
unpredictable behavior.
For example, the LRGPRO and BIGPRO modules each contain two PSECTs,
BIG and GRAND. LRGPRO is loaded and /COUNTERS is used to check PSECT
origins and current values. PSECT origins are found by looking under
the initial value column and PSECT current values are found by looking
under the current value column of the /COUNTERS output. The upper
bound is found by looking under the limit value column.
@LINK<RET>
*/SET:BIG:1000<RET>
*/SET:GRAND:5400<RET>
*LRGPRO<RET>
*/COUNTERS<RET>
[LNKRLC Reloc. ctr. initial value current value limit value
.LOW. 0 140 1000000
BIG 1000 5100 1000000
GRAND 5400 10500 1000000
Absolute code loaded]
*
/COUNTERS shows that the current value for PSECT BIG and the initial
value for PSECT GRAND are close together in memory. The current value
for BIG is 5100 and the PSECT origin for GRAND is 5400. The /LIMIT
switch can now be used to restrict PSECT BIG's current value to PSECT
GRAND's initial value using the following:
*/LIMIT:BIG:GRAND<RET>
/LIMIT prevents an unintended overlap because it causes LINK to issue
a warning if the current value for BIG exceeds GRAND's origin. The
warning is:
%LNKPEL PSECT [psect] exceeded limit of [address]
The warning message indicates that the PSECTs overlapped, and that BIG
and GRAND need to be farther apart in memory. The /COUNTERS switch
shows a new current value greater than 5400. Notice that the limit
set with the /LIMIT switch is shown in the limit value column.
*BIGPRO<RET>
%LNKPEL PSECT BIG exceeded limit of 5400
detected in module .MAIN from file BIGPRO.REL.1
*/COUNTERS<RET>
[LNKRLC Reloc. ctr. initial value current value limit value
.LOW. 0 140 1000000
BIG 1000 6300 5400
GRAND 5400 10500 1000000]
Absolute code loaded]
6-2
FIELD TEST
/GO continues loading the program, and LINK issues a warning and fatal
error message. The warning is:
%LNKPOV Psects [psect] and [psect] overlap from address [address] to
[address]
The fatal error message is:
?LNKCFS chained fixups have been suppressed
For example,
*/GO<RET>
%LNKPOV Psects BIG and GRAND overlap from address 5400 to 6300
?LNKCFS chained fixups have been suppressed
@
Now, LINK is re-run and the PSECTs are moved farther apart in memory.
In this example, GRAND's origin is reset from 5400 to 7000.
@LINK<RET>
*/SET:BIG:1000<RET>
*/SET:GRAND:7000<RET>
*LRGPRO<RET>
*/COUNTERS<RET>
[LNKRLC Reloc. ctr. initial value current value limit value
.LOW. 0 140 1000000
BIG 1000 5100 1000000
GRAND 7000 10500 1000000]
*/LIMIT:BIG:GRAND<RET>
*BIGPRO<RET>
*/COUNTERS<RET>
[LNKRLC Reloc. ctr. initial value current value limit value
.LOW. 0 140 1000000
BIG 1000 6300 5400
GRAND 7000 10500 1000000]
*/GO<RET>
@
6.2 PSECT ATTRIBUTES
PSECT attributes specify how LINK stores a PSECT in memory, and the
page access of the PSECT.
The CONCATENATED or OVERLAID attribute specifies how LINK stores
PSECTs, and the RONLY and RWRITE attributes specify the page access.
6-3
FIELD TEST
6.2.1 CONCATENATED and OVERLAID
LINK uses the CONCATENATED or OVERLAID attributes when loading PSECTs
into memory. These attributes are specified when the PSECT is defined
in the source program, and are contained in REL Blocks 24 and 1050.
See Appendix A for information on these blocks. If the attribute is
not specified, LINK uses CONCATENATED.
The following example illustrates how PSECTs are stored in memory. In
this example, modules MAINKO and MAINKC contain three PSECTs, ALPHA,
BETA, and GAMMA. There is an additional module named SUBMD1. The
ALPHA and BETA PSECTs have the CONCATENATE attribute. The GAMMA
PSECT, which is a data PSECT declared in each module, has attributes
defined differently in each module to illustrate the differences in
memory storage. GAMMA has the OVERLAID attribute defined in MAINKO
and the CONCATENATED attribute defined in MAINKC.
First, LINK is run and the origin is set for PSECTs ALPHA, BETA, and
GAMMA.
@LINK<RET>
*/SET:ALPHA:3000/SET:BETA:5000/SET:GAMMA:7000<RET>
Next, MAINKO is loaded with GAMMA defined as OVERLAID, and /COUNTERS
is used to display the initial, current, and limit values.
*MAINKO ;OVERLAID GAMMA<RET>
*/COUNTERS<RET>
[LNKRLC Reloc. ctr. initial value current value limit value
ALPHA 3000 3017 1000000
BETA 5000 5011 1000000
GAMMA 7000 7025 1000000]
Now, SUBMD1 is loaded, /COUNTERS is used, and /GO is used to load the
modules and exit LINK. Notice that the current values for ALPHA and
BETA have increased, and that the current value for GAMMA remains the
same.
*SUBMD1<RET>
*/COUNTERS<RET>
[LNKRLC Reloc. ctr. initial value current value limit value
ALPHA 3000 3033 1000000
BETA 5000 5041 1000000
GAMMA 7000 7025 1000000]
*/GO<RET>
@
In the following example, LINK is run and the origin is set for PSECTs
ALPHA, BETA, and GAMMA.
@LINK<RET>
*/SET:ALPHA:3000/SET:BETA:5000/SET:GAMMA:7000<RET>
6-4
FIELD TEST
Now, MAINKC is loaded with GAMMA defined as CONCATENATE, and /COUNTERS
is used.
*MAINKC ;CONCATENATED GAMMA<RET>
*/COUNTERS<RET>
[LNKRLC Reloc. ctr. initial value current value limit value
ALPHA 3000 3017 1000000
BETA 5000 5011 1000000
GAMMA 7000 7025 1000000]
*
Next, SUBMD1 is loaded, /COUNTERS is used, and /GO is used to load the
modules and exit LINK. Notice that all current values have increased.
*SUBMD1<RET>
*/COUNTERS<RET>
[LNKRLC Reloc. ctr. initial value current value limit value
ALPHA 3000 3033 1000000
BETA 5000 5041 1000000
GAMMA 7000 7035 1000000]
*/GO<RET>
@
6.2.2 RONLY and RWRITE
The RONLY and RWRITE attributes set the page access for PSECTs. RONLY
defines the page access to be read-only and RWRITE defines the page
access to be read/write. These attributes are set in the source
program. If the attribute is not set, LINK uses RWRITE.
For example, the following MACRO statement defines read/write access
for the RED PSECT.
.PSECT RED/RWRITE,1000
1000 is the PSECT's origin.
LINK sets the page access to read-only if a memory page contains part
or all of a read-only PSECT even if a read/write PSECT exists on the
same page.
6-5
7-1
CHAPTER 7
PDVs
LINK stores information about the program that it is loading into a
PDV or JOBDAT. A PDV (Program Data Vector) is a block of information.
JOBDAT is the job data area.
LINK stores information in a PDV:
o for extended addressing programs.
o if a PSECT has an origin below 140.
o if you specify a PDV using /PVBLOCK or /PVDATA.
o if /NOJOBDAT is used.
LINK stores information in JOBDAT for most section zero programs.
LINK stores such information as debugger symbol table pointers,
version numbers, and memory use.
Although LINK builds no more than one PDV per load, a program can use
the TOPS-20 PDVOP% JSYS at runtime to build new PDVs or to find PDVs
associated with object time systems and utilities merged into the
address space. LINK cannot build a PDV for a program that uses
overlays.
At link time, you can use the /PVBLOCK and /PVDATA switches to request
a PDV, control its memory location, or change its contents. See
Chapter 3 for information on these switches.
Each PDV contains location .PVCST reserved for customer use. A
typical use for this location might be to point to the address of a
transfer vector that contains entry points to a utility package. Any
program that wants to call the utility can then use the PDVOP% JSYS to
find the address of the transfer vector. This eliminates the need to
build the transfer vector address of the utility package into a
program.
7-1
FIELD TEST
For more information about the PDVOP% JSYS, refer to the TOPS-20
Monitor Calls Reference Manual.
7-2
FIELD TEST
7.1 PDV FORMAT
The format of a PDV follows:
Word Symbol Meaning
0 .PVCNT is the length of the PDV including this word.
1 .PVNAM points to the name in ASCIZ of this PDV. The default
PDV name is nnnLNK, where nnn is the job name.
The following rules apply to the assignment of PDV
names. If these rules are followed, conflicting PDV
names can be avoided:
1. PDV names assigned by DIGITAL will contain the
percent character (%). PDV names assigned by
users should not contain the percent character.
2. All PDV names containing the period character (.)
are reserved to DIGITAL for future use.
3. The dollar sign character ($) is reserved for
special use: PDV names of the form
string1$string2 are reserved for the special
class of use named by string1. Rules 1 and 2
still apply in this case.
As a general guideline, PDV names should be as
specific as possible to avoid name conflicts and
confusion.
If bit 0 is 1, this word contains a section-local
address. If bit 0 is 0, this word contains a global
address. A section-local address is an 18-bit
address. A global address is a 30-bit address.
2 .PVEXP points to a counted vector of exported information.
A counted vector is a contiguous block of information
where the length of the block is contained in the
first word. A PDV is an example of a counted vector.
Exported information is information that is available
for use by other programs.
By convention the remainder of the block is defined
by the program. If a program class exists, the
program class defines the remainder of the program.
See the third rule in the PDV naming rules above for
program class.
7-3
FIELD TEST
If bit 0 is 1, this word contains a section-local
address. If bit 0 is 0, this word contains a global
address. A section-local address is an 18-bit
address. A global address is a 30-bit address.
3 .PVREE is a word reserved for DIGITAL and must be zero.
4 .PVVER is the program's version number. See .JBVER in
Appendix C for more information.
5 .PVMEM points to the static memory map for this program.
See Section 7.2 for a description of the memory map.
If bit 0 is 1, this word contains a section-local
address. If bit 0 is 0, this word contains a global
address. A section-local address is an 18-bit
address. A global address is a 30-bit address.
6 .PVSYM points to the symbol table vector for this program.
See Section 7.3 for a description of this vector.
If bit 0 is 1, this word contains a section-local
address. If bit 0 is 0, this word contains a global
address. A section-local address is an 18-bit
address. A global address is a 30-bit address.
7 .PVCTM is the time that your program was compiled in TOPS-20
format.
10 .PVCVR is the version number of the main program's compiler.
See .JBVER in Appendix C for more information.
11 .PVLTM is the time that the program was loaded in TOPS-20
format.
12 .PVLVR is LINK's version number.
13 .PVMON points to a monitor data block. (Reserved for
DIGITAL.)
14 .PVPRG points to a program data block.
If bit 0 is 1, this word contains a section-local
address. If bit 0 is 0, this word contains a global
address. A section-local address is an 18-bit
address. A global address is a 30-bit address.
7-4
FIELD TEST
15 .PVCST points to a customer-defined data block.
If bit 0 is 1, this word contains a section-local
address. If bit 0 is 0, this word contains a global
address. A section-local address is an 18-bit
address. A global address is a 30-bit address.
7.2 THE PDV STATIC MEMORY MAP
LINK creates the static memory map along with the PDV. This static
memory map can be used as a resource in dynamic memory allocation. In
addition to containing information about allocated pages in use, this
memory map contains information about pages that are allocated but
zero.
In the PDV, the .PVMEM (5) word points to the program's memory map.
The memory map is a variable length table of variable length
subtables.
=================================== ^
| Memory Map Table | |
|---------------------------------| ^ |
| | | Variable |
| Subtable | | |
| | | Lengths |
|---------------------------------| V |
. |
. |
. |
|---------------------------------| ^ |
| | | |
| Subtable | | |
| | | |
|---------------------------------| V |
=================================== V
A subtable defines the beginning and end of a block of contiguous
words of memory with the same read/write attributes.
The addresses in the subtable are ordered from lowest to highest with
the lowest block of memory entered into the subtable first and the
highest entered into the subtable last.
7-5
FIELD TEST
The format of the memory map is illustrated below.
-- 0 17|18 35
| |=========================================
| | Length Of Table In Words |
| -- ------------------------------------------
| | | Access Bits | Subtable Length "n" |
|Sub- | ------------------------------------------
|table | | Lowest Address Of Block |
| n | ------------------------------------------
Memory | | | Highest Address Of Block |
Map | | ------------------------------------------
| | / /
| | / /
| -- ------------------------------------------
| .
| .
| .
| -- ------------------------------------------
| | | Access Bits | Subtable Length "n" |
|Sub- | ------------------------------------------
|table | | Lowest Address Of Block |
| n | ------------------------------------------
| | | Highest Address Of Block |
| | ------------------------------------------
| | / /
| -- / /
-- ==========================================
Memory Map
Word Symbol Meaning
0 .MMLEN is the length in words of the entire memory map
table, including this word.
Subtable
Word Symbol Meaning
0 .MMDAT is the first subtable word and contains two
half-words: MM%ACC and MM%LEN.
MM%ACC contains the block access bits. The access
bits define the block attributes, namely,
whether this block is read-only or
read/write.
7-6
FIELD TEST
These bits are set according to how the
PSECT attributes are set in the source
program. LINK defaults read and write
attributes if the source program does not
explicitly set attributes. All blocks in
the memory map implicitly have read and
execute access. If bit 3, MM%WR, is set,
the block is writable. Undefined access
bits are reserved for DIGITAL and must be
zero.
MM%LEN contains the subtable length including the
first word. The length of a subtable is
normally 3, but can be greater if
additional information is stored after
.MMHGH. This area after .MMHGH is reserved
for DIGITAL.
To obtain the address of the next subtable,
add the length of the current subtable to
the address of the current subtable. For
example, if the length of the current
subtable is 5 and its address is 15672, the
address of the subtable following is 15677.
1 .MMLOW is the address of the first word of a contiguous
block of memory with the same read/write or read-only
attributes.
If bit 0 is 1, this word contains a section-local
address. If bit 0 is 0, this word contains a global
address. A section-local address is an 18-bit
address. A global address is a 30-bit address.
2 .MMHGH is the address of the last word of the block.
If bit 0 is 1, this word contains a section-local
address. If bit 0 is 0, this word contains a global
address. A section-local address is an 18-bit
address. A global address is a 30-bit address.
7.3 SYMBOL TABLE VECTOR
A symbol table vector is a pointer to the symbol tables of a program.
There is one symbol table vector, and an undefined and defined symbol
table per program.
The .PVSYM (6) word in the PDV points to the symbol table vector. If
there is no symbol table (for example, when you use /NOSYM at link
time), the .PVSYM word is 0. If LINK builds the symbol table vector,
7-7
FIELD TEST
it always contains the defined and undefined symbol table pointers.
The symbol table vector contains subtables that point to each symbol
table and gives their length and type. Each subtable is three words
long, although only the first two words are currently used. The
format of a symbol table vector is illustrated below.
0 5|6 35
-- ===================================
Symbol | | Vector Length In Words |
Table | -- -----------------------------------
Vector | | | Type | Symbol Table Length |
| Sub- | -----------------------------------
| Table| | Symbol Table Pointer |
| n | -----------------------------------
| | | Reserved For DIGITAL, Must Be 0 |
| -- -----------------------------------
| .
| .
| .
|
| Sub- -- -----------------------------------
| Table| | Type | Symbol Table Length |
| n | -----------------------------------
| | | Symbol Table Pointer |
| | -----------------------------------
| | | Reserved For DIGITAL, Must Be 0 |
-- -- ===================================
Symbol Table
Word Symbol Meaning
0 .STLEN defines the length in words of the symbol table
vector including this word.
Subtable
Word Symbol Meaning
0 .STDAT is the first subtable word and contains two fields:
ST%TYP and ST%LEN.
ST%TYP is a 6-bit field that contains the symbol
table type.
The types are:
7-8
FIELD TEST
Code Name Type
0 .UNDFD Undefined
1 .R50D Radix-50 defined symbols
2 .R50U Radix-50 undefined symbols
3-37 Reserved for DIGITAL
40-77 Reserved for customers
ST%LEN is a 30-bit field that contains the length in
words of the particular symbol table.
1 .STPTR is the lowest word in the table.
If bit 0 is 1, this word contains a section-local
address. If bit 0 is 0, this word contains a global
address. A section-local address is an 18-bit
address. A global address is a 30-bit address.
The Reserved word is reserved for DIGITAL and must be zero.
7-9
A-1
APPENDIX A
REL BLOCKS
REL (RELocatable) Blocks are the main input to LINK and contain
information that LINK uses to load a program.
This appendix describes each type of REL Block and gives its format.
Terms used throughout this discussion are defined as follows:
Header Word: a fullword that contains the REL Block Type in its
left half and a short count or long count in its
right half.
Short Count: a halfword that contains the length of the REL
Block, excluding relocation words. The short
count appears before each group of 18 decimal, or
22 octal words and excludes the header word.
Long Count: a halfword that contains the length of the REL
Block, including all words in the block except the
header word itself.
Relocation Word: a fullword that contains the relocation bits for
up to 18 decimal or 22 octal following words.
Each relocation bit is either 1, indicating a
relocatable halfword, or 0, indicating a
nonrelocatable halfword.
The first two relocation bits give the
relocatability of the left and right halves,
respectively, of the next following word; the next
two bits give the relocatability of the two halves
of the second following word; and so forth for all
bits in the word, except any unused bits, which
are zero.
If a REL Block has relocation words, the first one
follows the header word. If more than 18
(decimal) data words follow this relocation word,
the next word (after the 18 words) is another
A-1
FIELD TEST
relocation word. Thus, a REL Block that has
relocation words will have one for each 18 words
of data that it contains. If the REL Block does
not contain an integral multiple of 18 words, the
last relocation word has unused bits.
NOTE
A block with a zero short count does not
include a relocation word.
Data Word: Any word other than a header word or a relocation
word.
MBZ: Must Be Zero.
NOTE
All numbers in this appendix are octal unless
specifically noted as decimal.
The diagram below shows a REL Block having a short count of 7, and a
relocation word.
|=======================================================|
| Block Type | 7 |
|-------------------------------------------------------|
| Relocation Word |
|-------------------------------------------------------|
| Data Word 1 |
|-------------------------------------------------------|
| Data Word 2 |
|-------------------------------------------------------|
| Data Word 3 |
|-------------------------------------------------------|
| Data Word 4 |
|-------------------------------------------------------|
| Data Word 5 |
|-------------------------------------------------------|
| Data Word 6 |
|-------------------------------------------------------|
| Data Word 7 |
|=======================================================|
A-2
FIELD TEST
The diagram below shows a REL Block having a short count of 31 (octal)
and two relocation words.
|=======================================================|
| Block Type | 31 |
|-------------------------------------------------------|
| Relocation Word |
|-------------------------------------------------------|
| Data Word 1 |
|-------------------------------------------------------|
.
.
.
|-------------------------------------------------------|
| Data Word 22 |
|-------------------------------------------------------|
| Relocation Word |
|-------------------------------------------------------|
| Data Word 23 |
|-------------------------------------------------------|
.
.
.
|-------------------------------------------------------|
| Data Word 31 |
|=======================================================|
REL Block Types must be numbered in the range 0 to 777777. The
following list shows which numbers are reserved for DIGITAL, and which
for customers:
Type Numbers Use
0 - 37 Reserved for DIGITAL
40 - 77 Reserved for customers
100 - 401 Reserved for DIGITAL
402 - 577 Reserved for customers
600 - 677 Reserved for customer files
700 - 777 Reserved for DIGITAL files
1000 - 1777 Reserved for DIGITAL
2000 - 3777 Reserved for customers
4000 - 777777 Reserved for ASCII text
A-3
FIELD TEST
Block Type 0 (Ignored)
|=======================================================|
| 0 | Short Count |
|-------------------------------------------------------|
| Optional Relocation Word |
|-------------------------------------------------------|
| Data Word |
|-------------------------------------------------------|
.
.
.
|-------------------------------------------------------|
| Data Word |
|=======================================================|
Block Type 0 is ignored by LINK.
If the short count is 0, then no relocation word follows, and the
block consists of only one word. This is how LINK bypasses zero words
in a REL file.
A-4
FIELD TEST
Block Type 1 (Code)
|=======================================================|
| 1 | Short Count |
|-------------------------------------------------------|
| Relocation Word |
|-------------------------------------------------------|
| 0 | Address Word |
|-------------------------------------------------------|
.
.
.
|-------------------------------------------------------|
| Data Word |
|=======================================================|
Block Type 1 contains data and code. The address is where LINK is to
load the data. This address can be relocatable or absolute, depending
on the value of bit 1 of the relocation word. LINK loads the
remaining data words beginning at that address.
If a symbol is used to specify the start address, the following format
of Block Type 1 is used:
|=======================================================|
| 1 | Short Count |
|-------------------------------------------------------|
| Relocation Word |
|-------------------------------------------------------|
| Symbol |
|-------------------------------------------------------|
| Offset |
|-------------------------------------------------------|
.
.
.
|-------------------------------------------------------|
| Data Word |
|=======================================================|
In this alternate format, the first four bits of the first data word
(Symbol) are 1100 (binary), and the word is assumed to be a Radix-50
symbol of type 60. The load address is calculated by adding the value
of the global symbol to the offset given in the following word. The
third and following data words are loaded beginning at the resulting
address. The global symbol must be defined when the Type 1 Block is
found.
A-5
FIELD TEST
Block Type 2 (Symbols)
|=======================================================|
| 2 | Short Count |
|-------------------------------------------------------|
| Relocation Word |
|-------------------------------------------------------|
|Code | Radix-50 Symbol |
|-------------------------------------------------------|
| Second Word of Pair |
|-------------------------------------------------------|
.
.
.
|-------------------------------------------------------|
|Code | Radix-50 Symbol |
|-------------------------------------------------------|
| Second Word of Pair |
|=======================================================|
The first word of each pair has a code in bits 0 to 3 and a Radix-50
symbol in bits 4 to 35 (decimal). The contents of the second word of
a pair depends on the given code. The octal codes and their meanings
are:
Code Meaning
00 This code is illegal in a symbol block.
04 The given symbol is a global definition. Its value, contained
in the second word of the pair, is available to other programs.
10 The given symbol is a local definition, and its value is
contained in the second word of the pair. If the symbol is
followed by one of the special pairs or by a Polish REL Block
(as explained below, under code 24), the symbol is considered a
partially defined local symbol. Otherwise, it is considered
fully defined.
14 The given symbol is a block name (from a translator that uses
block structure). The second word of the pair contains the
block level. The symbol is considered local; if local symbols
are loaded, the value of the block name is entered in the
symbol table as its block level.
24 The given symbol is a global definition. However, it is only
partially defined at this time, and LINK cannot yet use its
value. If the symbol is defined in terms of another symbol,
then the next entry in the REL file must be a word pair in a
Block Type 2 as follows:
|=======================================================|
A-6
FIELD TEST
| 60 | Other Symbol |
|-------------------------------------------------------|
| 50 | This Symbol |
|=======================================================|
In this format, code 50 indicates that the right half of the
word depends on the other symbol.
If the partially defined symbol is defined in terms of a Polish
expression, then the next entry in the REL file must be Block
Type 11 (Polish), whose store operator gives this symbol as the
symbol to be fixed up. A fixup resolves the symbol. The store
operator must be -4 or -6.
30 The given symbol is a global definition. However, it is only
partially defined at this time, and LINK cannot yet use its
value. If the symbol is defined in terms of another symbol,
then the next entry in the REL file must be a word pair in a
Block Type 2 as follows:
|=======================================================|
| 60 | Other Symbol |
|-------------------------------------------------------|
| 70 | This Symbol |
|=======================================================|
In this format, code 70 indicates that the left half of the
word depends on the other symbol.
If the partially defined symbol is defined in terms of a Polish
expression, then the next entry in the REL file must be Block
Type 11 (Polish), whose store operator gives this symbol as the
symbol to be fixed up. The store operator must be -5.
34 The given symbol is a global definition. However, it is only
partially defined at this time, and LINK cannot yet use its
value. If the symbol is defined in terms of another symbol,
then the next entry in the REL file must be a word pair in a
Block Type 2 as follows:
|=======================================================|
| 60 | Other Symbol |
|-------------------------------------------------------|
| 50 | This Symbol |
|-------------------------------------------------------|
| 60 | Other Symbol |
|-------------------------------------------------------|
| 70 | This Symbol |
|=======================================================|
This format indicates that both halves of the word depend on
the other symbol.
A-7
FIELD TEST
44 The given symbol is a global definition exactly as in code 04,
except that DDT does not output the symbol.
50 The given symbol is a local symbol exactly as in code 10,
except that DDT does not output the symbol.
60 The given symbol is a global request. LINK's handling of the
symbol depends on the value of the code in the first four bits
of the second word of the pair. These codes and their meanings
are:
00 The right half of the word gives the address of the
first word in a chain of requests for the global memory
address. In each request, the right half of the word
gives the address of the next request. The chain ends
when the address is 0.
40 The right half of the word contains an address. The
right half of the value of the requested symbol is added
to the right half of this word.
50 The rest of the word contains a Radix-50 symbol whose
value depends on the requested global symbol. (If the
given Radix-50 symbol is not the one defined in the
previous word pair, then this word is ignored.) When
the value of the requested symbol is resolved, it is
added to the right half of the value of the Radix-50
symbol.
60 The right half of the word contains an address. The
right half of the value of the requested symbol is added
to the left half of this word.
70 The rest of the word contains a Radix-50 symbol whose
value depends on the requested global symbol. (If the
given Radix-50 symbol is not the one defined in the
previous word pair, then this word is ignored.) When
the value of the requested global symbol is resolved, it
is added to the left half of the value of the Radix-50
symbol.
64 The given symbol is a global definition exactly as in code 24,
except that DDT does not output the symbol.
70 The given symbol is partially defined, where the left half is
deferred, as in code 30, except that DDT does not output the
symbol.
74 The given symbol is partially defined, where the right half is
deferred, as in code 34, except that DDT does not output the
symbol.
A-8
FIELD TEST
Symbols are placed in the symbol table in the order that LINK finds
them. However, DDT expects to find the symbols in a specific order.
For a non-block-structured program, that order is:
Program Name
Symbols for Program
For a block-structured program whose structure is:
Begin Block 1 (same as program name)
Begin Block 2
End Block 2
Begin Block 3
Begin Block 4
End Block 4
End Block 3
End Block 1
the order is:
Program Name (Block 1)
Block Name 2
Symbols for Block 2
Block Name 4
Symbols for Block 4
Block Name 3
Symbols for Block 3
Block Name 1
Symbols for Block 1
This ordering follows the rule that the name and symbols for each
block must occur in the symbol table in the order of the block endings
in the program.
The following fixup rules apply to this block:
1. Only one fixup by a Type 2, 10, 11, 12, 15, 1070, 1072, or
1120 Block is allowed for a given word. (There can be
separate fixups for the left and right halves of the same
word.)
2. Fixups are not necessarily performed in the order LINK finds
them.
3. Chained halfword fixups cannot cross section boundaries.
Also, they cannot fixup a location which resolves to word
zero of a section unless it is the only address in the chain.
4. Chained fixups must be in strict descending address order.
A-9
FIELD TEST
5. A location must contain data before the location can be fixed
up.
A-10
FIELD TEST
Block Type 3 (HISEG)
|=======================================================|
| 3 | Short Count |
|-------------------------------------------------------|
| Relocation Word |
|-------------------------------------------------------|
|High-Segment Program Break | High-Segment Origin |
|-------------------------------------------------------|
|(Low-Segment Program Break)| (Low-Segment Origin) |
|=======================================================|
Block Type 3 tells LINK that code is to be loaded into the high
segment.
The Short Count is either 1 or 2.
If the left half of the first data word is 0, subsequent Type 1 blocks
found are assumed to have been produced by the MACRO pseudo-op HISEG.
This usage is not recommended. It means that the addresses in the
blocks are relative to 0, but are to be placed in the program high
segment. The right half of the first data word is the beginning of
the high segment (usually 400000).
If the left half of the first data word is nonzero (the preferred
usage), subsequent Type 1 blocks found are assumed to have been
produced by the MACRO pseudo-op TWOSEG.
The right half is interpreted as the beginning of the high segment,
and the left half is the high-segment break; the high-segment length
is the difference of the left and right halves.
(One-pass translators that cannot calculate the high-segment break
should set the left half equal to the right half.)
If the second word appears in the HISEG block, its left half shows the
low-segment program break, and its right half shows the low-segment
origin (usually 0).
A-11
FIELD TEST
Block Type 4 (Entry)
|=======================================================|
| 4 | Short Count |
|-------------------------------------------------------|
| Relocation Word (Zero) |
|-------------------------------------------------------|
| Radix-50 Symbol |
|-------------------------------------------------------|
.
.
.
|-------------------------------------------------------|
| Radix-50 Symbol |
|=======================================================|
Block Type 4 lists the entry name symbols for a program module. If a
Type 4 block appears in a module, it must be the first block in the
module. A library file contains a Type 4 block for each of its
modules.
When LINK is in library search mode, the symbols in the block are
compared to the current list of global requests for the load. If one
or more matches occur, the module is loaded.
The 4 high order bits of the symbol are 0 and are ignored.
A-12
FIELD TEST
Block Type 5 (End)
|=======================================================|
| 5 | Short Count |
|-------------------------------------------------------|
| Relocation Word |
|-------------------------------------------------------|
| First Data Word |
|-------------------------------------------------------|
| (Second Data Word) |
|=======================================================|
Block Type 5 ends a program module. A Block Type 6 (Name) must be
encountered earlier in the module than the Type 5 block.
The Short Count is 1 or 2.
If the module contains a two-segment program, the first data word is
the high-segment break and the second data word is the low-segment
break. If the module contains a one-segment program, the first data
word is the program break and the second data word is the absolute
break. If the count is 1, then the second word is assumed to be 0.
If the module contains a program that uses PSECTs, Block Type 5 has no
effect except if the /REDIRECT switch was used. If /REDIRECT was
used, the break information is used to set the highest location to be
loaded for the PSECT specified in /REDIRECT.
Each PRGEND pseudo-op in a MACRO program generates a Type 5 REL block.
Therefore, a REL file may contain more than one pair of Type 6 and
Type 5 blocks.
A library REL file has a Type 5 block at the end of each of its
modules.
A-13
FIELD TEST
Block Type 6 (Name)
|=======================================================|
| 6 | Short Count |
|-------------------------------------------------------|
| Relocation Word |
|-------------------------------------------------------|
| Radix-50 Symbol |
|-------------------------------------------------------|
| (CPU) | (Compiler) | (Length of Blank Common) |
|-------------------------------------------------------|
Block Type 6 contains the program name, and must precede any Type 2
blocks. (A module should begin with a Type 6 or 1003 block and end
with a Type 5 block.)
The Short Count is 1 or 2.
The first data word is the program name in Radix-50 format; this name
cannot be blanks. The second data word is optional; if it appears, it
contains CPU codes in bits 0 to 5, a compiler code in bits 6 to 17
(decimal), and the length of the program's blank COMMON in the right
halfword.
The CPU codes specify processors for program execution as:
Bit 2 KS10
Bit 3 KL10
Bit 4 KI10
Bit 5 KA10
If none of these bits are on, then any of the processors can be used
for execution.
The compiler code specifies the compiler that produced the REL file.
The defined codes are:
0 Unknown 10 FORTRAN 20 BLISS-36
1 Not used 11 MACRO 21 BASIC
2 COBOL-68 12 FAIL 22 SITGO
3 ALGOL 13 BCPL 23 (Reserved)
4 NELIAC 14 MIDAS 24 PASCAL
5 PL/I 15 SIMULA 25 JOVIAL
6 BLISS 16 COBOL-7 26 ADA
7 SAIL 17 COBOL
A-14
FIELD TEST
Block Type 7 (Start)
|=======================================================|
| 7 | Short Count |
|-------------------------------------------------------|
| Relocation Word |
|-------------------------------------------------------|
| Start Address |
|-------------------------------------------------------|
|(60) | (Optional Radix-50 Symbol) |
|=======================================================|
Block Type 7 contains the start address for program execution. LINK
uses the start address in the last such block processed by the load,
unless /START or /NOSTART switches specify otherwise.
Short Count is 2.
If the Optional Radix-50 Symbol word is present, it must be a Radix-50
symbol with the code 60. LINK forms the start address by adding the
value of the symbol to the value in the right half of the preceding
word (Start Address).
LINK defaults a TOPS-10 style entry vector if the entry length vector
is zero.
LINK defaults a TOPS-20 style entry vector:
o if the program contains nonzero sections and the length of
the entry vector is not specified
o if the length of the entry vector is 1
o if /NOJOBDAT equals 1
A-15
FIELD TEST
Block Type 10 (Internal Request)
|=======================================================|
| 10 | Short Count |
|-------------------------------------------------------|
| Relocation Word |
|-------------------------------------------------------|
| Pointer to Last Request | Value |
|-------------------------------------------------------|
.
.
.
|-------------------------------------------------------|
| Pointer to Last Request | Value |
|=======================================================|
Block Type 10 is generated by one-pass translators to resolve requests
caused by forward references to internal symbols. The MACRO assembler
also generates Type 10 blocks to resolve requests for labels defined
in literals; a separate chain is required for each PSECT in a program
that contains PSECTs.
Each data word contains one request for an internal symbol. The left
half is the address of the last request for a given symbol. The right
half is the value of the symbol. The right half of the last request
contains the address of the next-to-last request, and so on, until a
zero right half is found. (This is exactly analogous to Radix-50 code
60 with second-word code 00 in a Block Type 2.)
If a data word contains -1, then the following word contains a request
for the left (rather than right) half of the specified word. In this
case, the left half of the word being fixed up contains the address of
the next-to-last left half request, and so on, until a zero left half
is found. (This is a left half chain analogous to the right half
chain described above.)
The following fixup rules apply to this block:
1. Only one fixup by a Type 2, 10, 11, 12, 15, 1070, 1072, or
1120 Block is allowed for a given field. (There can be
separate fixups for the left and right halves of the same
word.)
2. Fixups are not necessarily performed in the order LINK finds
them.
3. These are halfword fixups and cannot cross section
boundaries. They wrap within a section. Also, halfword
fixups cannot fixup a location which resolves to word zero of
a section unless it is the only address in the chain.
A-16
FIELD TEST
4. Chained fixups must be in strict descending address order.
5. A location must contain data before the location can be fixed
up.
A-17
FIELD TEST
Block Type 11 (Polish)
|=======================================================|
| 11 | Short Count |
|-------------------------------------------------------|
| Relocation Word |
|-------------------------------------------------------|
| Data Halfword | Data Halfword |
|-------------------------------------------------------|
.
.
.
|-------------------------------------------------------|
| Data Halfword | Data Halfword |
|=======================================================|
Block Type 11 defines Polish fixups for operations on relocatable
values or external symbols. Only one store operator code can appear
in a Block Type 11; this store operator code can be either a symbol
fixup code or a chained fixup code. The store operator code appears
at the end of the block.
The following fixup rules apply to this block:
1. Only one fixup by a Type 2, 10, 11, 15, 1070, 1072, or 1120
Block is allowed for a given field. (There can be separate
fixups for the left and right halves of the same word.)
2. Fixups are not necessarily performed in the order LINK finds
them.
3. Chained halfword fixups cannot cross section boundaries.
Also, they cannot fixup a location which resolves to word
zero of a section unless it is the only address in the chain.
4. Chained fixups must be in strict descending address order.
5. A location must contain data before the location can be fixed
up.
The data words of a Type 11 block form one Polish string of halfwords.
Each halfword contains one of the following:
1. A symbol fixup store operator code.
A symbol fixup defines the value to be stored in the value
field of the symbol table for the given symbol. A symbol
fixup store operator code is followed by two or four data
halfwords.
2. A chained fixup store operator code.
A-18
FIELD TEST
A chained fixup takes a relocatable address whose corrected
virtual address is the location for storing or chaining. A
chained fixup store operator code is followed by one data
halfword.
3. A data type code.
Data type code 0 is followed by a data halfword; a data type
code 1 or 2 is followed by two data halfwords.
4. An arithmetic or logical operator code.
5. A PSECT index code.
This code defines a PSECT index to be used for calculating
the relocated addresses that appear in this block. PSECT
indexes are needed only for PSECTed programs.
A global PSECT index is associated with a Block Type 11.
This index appears as the first halfword after the relocation
word, and it defines the PSECT for the store address or store
symbol. Any addresses for a different PSECT must be preceded
by a different PSECT index.
Thus, a relocatable data halfword in a different PSECT must
appear in one of the following formats:
|-------------------------------------------------------|
| 400nnn | (operator code) |
|-------------------------------------------------------|
| (operands) |
|-------------------------------------------------------|
OR
|-------------------------------------------------------|
| . . . | 400nnn |
|-------------------------------------------------------|
| (operator code) | (operands) |
|-------------------------------------------------------|
where the different PSECT index is nnn+1.
Any relocatable address that does not have an explicit
preceding PSECT index code preceding its data type code is
assumed to be in the same PSECT as the store address for the
block. The current PSECT may be set by a previous REL Block
type.
6. A halfword of data (preceded by a data type 0 halfword) or
two halfwords of data (preceded by a data type 1 or 2
halfword).
A-19
FIELD TEST
A sequence of halfwords containing a data type code 0 and a
data halfword can begin in either half of a word.
The codes and their meanings are:
.LINK fixup codes
-10 does a store into the link that was specified if the address
is negative. Stores the address as a link-end. Links the
result of the Polish Block.
-7 Fullword replacement. No chaining is done.
Symbol Fixup Store Operator Codes:
-6 Fullword symbol fixup. The following one or two words
contain the Radix-50 symbol(s) (with their 4-bit codes).
The first is the symbol to be fixed up, and the second is
the block name for a block-structured program (0 or
nonexistent for other programs).
-5 Left half symbol fixup. The following one or two words
contain the Radix-50 symbols. The first is the symbol to be
fixed up, and the second is the block name for a
block-structured program (0 or nonexistent for other
programs).
-4 Right half symbol fixup. The following one or two words
contain the Radix-50 symbols. The first is the symbol to be
fixed up, and the second is the block name for a
block-structured program (0 or nonexistent for other
programs).
Chained Fixup Store Operator Codes:
-3 Fullword chained fixup. The halfword following points to
the first element in the chain. The entire word pointed to
is replaced, and the old right half points to the next
fullword.
-2 Left half chained fixup. The halfword following points to
the first element in the chain.
-1 Right half chained fixup. The halfword following points to
the first element in the chain.
Data Type Codes:
0 The next halfword is an operand.
1 The next two halfwords form a fullword operand.
A-20
FIELD TEST
2 The next two halfwords form a Radix-50 symbol that is a
global request. The operand is the value of the symbol.
Arithmetic and Logical Operator Codes:
3 Add.
4 Subtract. The first operand is subtracted by the second.
5 Multiply.
6 Divide. The first operand is divided by the second.
7 Logical AND.
10 Logical OR.
11 Logical shift. The first operand is shifted by the second.
A positive second operand causes a shift to the left. A
negative operand causes a shift to the right.
12 Logical XOR.
13 Logical NOT (one's complement).
14 Arithmetic negation (two's complement).
15 Count leading zeros (like JFFO instruction). Refer to the
MACRO Assembler Reference Manual for information about the
^L operand, which this code implements.
16 Remainder. The first operand is REM by the second.
17 Magnitude.
20 Maximum.
21 Minimum.
22 Comparison. Returns 0 if the two operands are different; -1
if they are equal.
23 Used to resolve the links in a chain. The address of the
specified link is used. See -10.
24 Symbol definition test. Returns 0 if the operand (a
Radix-50 symbol) is unknown; 1 if it is known but undefined;
-1 if it is known and defined.
25 Skip N words of Polish.
A-21
FIELD TEST
26 Skip the remainder of the REL module if the argument is
nonzero, otherwise return the argument. Undefined symbols
are not allowed with this store operator in data type 2
operands.
27 Return contents of location N. No fixups may be done on
location N.
PSECT Index Codes:
400nnn PSECT index nnn, where nnn is a 3-digit octal integer.
For an example of a Type 11 block, the MACRO statements
EXTERN B
A: EXP <A*B+A>
Generate (assuming that A has a relocatable value of zero):
|=======================================================|
| 11 | 6 |
|-------------------------------------------------------|
|00|01|00|00|10|10| 0 |
|-------------------------------------------------------|
| 3 (Add) | 5 (Multiply) |
|-------------------------------------------------------|
| 0 (Halfword Operand Next) | 0 (Relocatable) |
|-------------------------------------------------------|
|2 (Fullword Radix-50 Next) | 1st Half of Radix-50 B |
|-------------------------------------------------------|
| 2nd Half of Radix-50 B | 0 (Halfword Operand Next) |
|-------------------------------------------------------|
| 0 (Relocatable) | -3 ( Chained Fixup Next) |
|-------------------------------------------------------|
| 0 (Chain Starts at 0') | . . . |
|=======================================================|
The first word contains the block type (11) and the short count (6).
The second word is the relocation word; it shows that the following
halfwords are to be relocated: right half of second following word,
left half of fifth following word, left half of sixth following word.
The next word shows that the two operations to be performed are
addition and multiplication; because this is in Polish prefix format,
the multiplication is to be performed on the first two operands first,
then addition is performed on the product and the third operand.
The next two halfwords define the first operand. The first halfword
is a data type code 0, showing that the operand is a single halfword;
the next halfword is the operand (relocatable 0).
The next three halfwords define the second operand. The first of
A-22
FIELD TEST
these halfwords contains a data type code 2, showing that the operand
is two halfwords containing a Radix-50 symbol with code 60. The next
two halfwords give the symbol (B).
The next two halfwords define the third operand. The first of these
halfwords contains a data type code 0, showing that the operand is a
single halfword; the next halfword gives the value of the operand
(relocatable 0).
The next two halfwords give the store operator for the block. The
first of these halfwords contains the chained fixup store operator
code -3, showing that a fullword chained fixup is required; the next
halfword contains the operand (relocatable 0), showing that the chain
starts at relocatable zero.
The last halfword is irrelevant, and should be zero. If it is not,
LINK issues the LNKJPB error message.
A-23
FIELD TEST
Block Type 12 (Chain)
|=======================================================|
| 12 | Short Count |
|-------------------------------------------------------|
| Relocation Word |
|-------------------------------------------------------|
| Chain Number |
|-------------------------------------------------------|
| Chain Address | Store Address |
|-------------------------------------------------------|
.
.
.
|-------------------------------------------------------|
| Chain Number |
|-------------------------------------------------------|
| Chain Address | Store Address |
|=======================================================|
Block Type 12 chains together data structures from separately compiled
modules. (The MACRO pseudo-ops .LINK and .LNKEND generate Type 12
blocks.) Block Type 12 allows linked lists that have entries in
separately compiled modules to be constructed so that new entries can
be added to one module without editing or recompiling any other
module.
The data words in a Type 12 block are paired. The first word of each
pair contains a chain number between 1 and 100 (octal). (The chain
number is negative if the pair was generated by a .LNKEND pseudo-op.)
The second word contains a store address in the right half, and a
chain address in the left half. The store address points to the
location where LINK will place the chain address of the last entry
encountered for the current chain. The first entry in a chain has a
zero in the word pointed to by the store address.
A MACRO statement of the form:
.LINK chain-number,store-address,chain-address
generates a word pair in a Type 12 block as shown above. A MACRO
statement of the form:
.LINK chain-number,store-address
generates a word pair in a Type 12 block with a 0 for the chain
address field in the REL block. A MACRO statement of the form:
.LNKEND chain-number,store-address
generates a word pair in a Type 12 Block with a 0 for the chain
address and a negative chain number.
A-24
FIELD TEST
As LINK processes a load, it performs a separate chaining for each
different chain number found; thus a word pair in a Type 12 block is
related to all other word pairs having the same chain number (even in
other loaded modules). Type 12 pairs having different chain numbers
(even in the same module), are not related.
The following fixup rules apply to this block:
1. Only one fixup by a Type 2, 10, 11, 12, 15, 1070, 1072, or
1120 Block is allowed for a given field. (There can be
separate fixups for the left and right halves of the same
word.)
2. Fixups are not necessarily performed in the order LINK finds
them.
3. Chain numbers above 100 (octal) are reserved by DIGITAL.
4. These are halfword fixups and cannot cross section
boundaries. They wrap within a section. Also, halfword
fixups cannot fixup a location which resolves to word zero of
a section unless it is the only address in the chain.
5. A location must contain data before the location can be fixed
up.
To show how the chains are formed, we will take some pairs from
different programs having the same chain number (1 in the example).
The following four programs contain .LINK or .LNKEND pseudo-ops for
the chain numbered 1. After each program, the word pair generated in
the Type 12 block appears.
NOTE
When LINK stores an address resulting from a Type 12
REL Block, only the right half of the receiving
location is written. You can safely store another
value in the left half; it will not be overwritten.
A-25
FIELD TEST
Example
TITLE MOD0
.
.
.
TAG0: BLOCK 1
.
.
.
.LNKEND 1,TAG0
.
.
.
END
|=======================================================|
| -1 |
|-------------------------------------------------------|
| 0 | Value of TAG0 |
|=======================================================|
TITLE MOD1
.
.
.
TAG1: BLOCK 1
.
.
.
.LINK 1,TAG1
.
.
.
END
|=======================================================|
| 1 |
|-------------------------------------------------------|
| 0 | Value of TAG1 |
|=======================================================|
TITLE MOD2
.
.
.
TAG2: BLOCK 1
.
.
.
.LINK 1,TAG2
.
A-26
FIELD TEST
.
.
END
A-27
FIELD TEST
|=======================================================|
| 1 |
|-------------------------------------------------------|
| 0 | Value of TAG2 |
|=======================================================|
TITLE MOD3
.
.
.
TAG3: BLOCK 1
.
.
.
TAG33: BLOCK 1
.
.
.
.LINK 1,TAG33,TAG3
.
.
.
END
|=======================================================|
| 1 |
|-------------------------------------------------------|
| Value of TAG3 | Value of TAG33 |
|=======================================================|
A-28
FIELD TEST
Suppose we load MOD0 first. The .LNKEND statement for MOD0 generates
a negative chain number. LINK sees the negative chain number (-1) and
recognizes this as the result of a .LNKEND statement for chain number
1. LINK remembers the store address (value of TAG0) as the base of
the chain.
Next we load MOD1. The .LINK statement for MOD1 does not use the
third argument, so the chain address is 0. LINK sees that this is the
first entry for chain number 1. Because it is the first entry, LINK
places a 0 in the store address (value of TAG1). LINK then remembers
the value of TAG1 for use in the next chain entry. (If the chain
address is 0, as it is in MOD1, LINK remembers the store address; if
the chain address is nonzero, LINK remembers the chain address.)
Next we load MOD3. The .LINK statement in MOD3 uses a third argument
(TAG3), therefore, the value of TAG3 is used as the chain address.
LINK places its remembered address (value of TAG1) in the store
address (value of TAG33). Because the chain address (value of TAG3)
is nonzero, LINK remembers it for the next entry.
Finally we load MOD2. Like MOD1, the .LINK statement for MOD2 does
not take a third argument, and thus the chain address is 0. LINK
places the remembered address (value of TAG3) in the store address
(value of TAG2). Because the chain address is 0, LINK remembers the
store address (value of TAG2).
At the end of loading, LINK places the last remembered address (value
of TAG2) at the address (value of TAG0) given by the .LNKEND statement
in MOD0.
The results of the chaining can be seen in the following diagram of
the loaded core image:
MOD0 MOD2
----------------- ----------------
| | | |
| | | |
TAG0: |Value of TAG2 | TAG2:|Value of TAG3 |
| | | |
| | | |
| | | |
----------------- ----------------
MOD3 MOD1
----------------- ----------------
| | | |
| | | |
TAG3: | | TAG1:| 0 |
| | | |
TAG33: |Value of TAG1 | | |
| | | |
----------------- ----------------
A-29
FIELD TEST
Note that the order of loading for modules with .LINK
statements is critical. (A module containing a .LNKEND
statement can be loaded any time;
its treatment is not affected by the order of loading.)
For example, if we load the four programs in the order
MOD2, MOD3, MOD0, MOD1, we get a different resulting
core image:
MOD0 MOD1
----------------- ----------------
| | | |
| | | |
TAG0: |Value of TAG1 | TAG1:|Value of TAG3 |
| | | |
| | | |
| | | |
----------------- ----------------
MOD3 MOD2
----------------- ----------------
| | | |
| | | |
TAG3: | | TAG2:| 0 |
| | | |
TAG33: |Value of TAG2 | | |
| | | |
----------------- ----------------
A-30
FIELD TEST
Block Type 14 (Index)
|=======================================================|
| 14 | 177 |
|-------------------------------------------------------|
| Sub-Block |
| |
|-------------------------------------------------------|
.
.
.
|-------------------------------------------------------|
| |
| Sub-Block |
| |
|-------------------------------------------------------|
| -1 | Ptr To Nxt Rel Blk Typ 14 |
|=======================================================|
Each sub-block is of the form:
|=======================================================|
| Index-Version Number | Count of Symbols |
|-------------------------------------------------------|
| Radix-50 Symbol |
|-------------------------------------------------------|
.
.
.
|-------------------------------------------------------|
| Radix-50 Symbol |
|-------------------------------------------------------|
| Pointer to Module Containing Entry Symbols |
|=======================================================|
Block Type 14 contains a list of all entry points in a library
produced by MAKLIB. The block contains 177 (octal) data words (with
no relocation words); if the index requires more entries, additional
Type 14 blocks are used. If 177 data words are not needed, zero words
pad the block to a length of 177. -1 indicates the end of the
sub-block information.
The Type 14 block consists of a header word, a number of sub-blocks,
and a trailer word containing the disk block address of the next Type
14 block, if any. Each disk block is 128 words.
Each sub-block is like a Type 4 block, with three differences:
A-31
FIELD TEST
1. The sub-block has no relocation words.
2. The last word of the sub-block points to the module that
contains the entry points listed in the sub-block. The right
half of the pointer has the disk block number of the module
within the file; the left half has the number of words (in
that block) that precede the module. If there is no next
block, then the word after the last sub-block is -1.
3. The index-version number is used so that old blocks can still
be loaded, even if the format changes in the future.
A-32
FIELD TEST
Block Type 15 (ALGOL)
|=======================================================|
| 15 | Short Count |
|-------------------------------------------------------|
| Relocation Word |
|-------------------------------------------------------|
| Load Address | Length |
|-------------------------------------------------------|
| Chain Address | Offset |
|-------------------------------------------------------|
.
.
.
|-------------------------------------------------------|
| Chain Address | Offset |
|=======================================================|
Block Type 15 is used to build the special ALGOL OWN block.
The first data word contains the length of the module's OWN block in
the right half, and the desired load address for the current OWN block
in the left half. Each following word contains an offset for the
start of the OWN block in the right half, and the address of a
standard righthalf chain of requests for that word of the OWN block in
the left half.
When LINK sees a REL Block Type 15, it allocates a block of the
requested size at the requested address. The length of the block is
then placed in the left half of the first word, and the address of the
last OWN block seen is placed in the right half. If this is the first
OWN block seen, 0 is stored in the right half of the first word.
The remaining data words are then processed by adding the address of
the first word of the OWN block to each offset, and then storing the
resulting value in all the locations chained together, starting with
the chain address.
At the end of loading, LINK checks to see if the symbol %OWN is
undefined. If it is undefined, then it is defined to be the address
of the last OWN block seen. In addition, if LINK is creating an ALGOL
symbol file, the file specification of the symbol file is stored in
the first OWN block loaded. This file specification must use the
standard TOPS-10 format below:
device:name.type[project-programmer number]
A-33
FIELD TEST
Block Type 16 (Request Load)
|=======================================================|
| 16 | Short Count |
|-------------------------------------------------------|
| Relocation Word (Zero) |
|-------------------------------------------------------|
| SIXBIT Filename |
|-------------------------------------------------------|
| Project-Programmer Number |
|-------------------------------------------------------|
| SIXBIT Device |
|-------------------------------------------------------|
.
.
.
|-------------------------------------------------------|
| SIXBIT Filename |
|-------------------------------------------------------|
| Project-Programmer Number |
|-------------------------------------------------------|
| SIXBIT Device |
|=======================================================|
Block Type 16 contains a list of files to be loaded. The data words
are arranged in triplets; each triplet contains information for one
file: file name, project-programmer number, and device. The file
type is assumed to be .REL.
LINK saves the specifications for the files to be loaded, discarding
duplicates. At the end of loading, LINK loads all specified files
immediately before beginning library searches.
The MACRO pseudo-op .REQUIRE generates a Type 16 REL Block.
A-34
FIELD TEST
Block Type 17 (Request Library)
|=======================================================|
| 17 | Short Count |
|-------------------------------------------------------|
| Relocation Word (Zero) |
|-------------------------------------------------------|
| SIXBIT Filename |
|-------------------------------------------------------|
| Project-Programmer Number |
|-------------------------------------------------------|
| SIXBIT Device |
|-------------------------------------------------------|
.
.
.
|-------------------------------------------------------|
| SIXBIT Filename |
|-------------------------------------------------------|
| Project-Programmer Number |
|-------------------------------------------------------|
| SIXBIT Device |
|=======================================================|
Block Type 17 is identical to Block Type 16 except that the specified
files are loaded in library search mode. The specified files are
searched after loading files given in Type 16 blocks, but before
searching system or user libraries.
The MACRO pseudo-op .REQUEST generates a Type 17 REL Block.
A-35
FIELD TEST
Block Type 20 (Common)
|=======================================================|
| 20 | Short Count |
|-------------------------------------------------------|
| Relocation Word (Zero) |
|-------------------------------------------------------|
| Radix-50 Symbol |
|-------------------------------------------------------|
| Length of Labeled Common Block |
|-------------------------------------------------------|
.
.
.
|-------------------------------------------------------|
| Radix-50 Symbol |
|-------------------------------------------------------|
| Length of Labeled Common Block |
|=======================================================|
Block Type 20 allocates labeled COMMON areas. The label for unlabeled
COMMON is ".COMM.". If a Block Type 20 appears in a REL file, it must
appear before any other block that causes code to be loaded or storage
to be allocated in the core image.
LINK allocates the specified COMMON into the current PSECT only when
allocating COMMON areas in a program that contains PSECTs. The
current PSECT is defined as the PSECT specified in a previous Type 22
PSECT Origin Block or Type 1051 Set PSECT Block.
If the program does not use PSECTs, the COMMON areas are allocated
into the HISEG or LOSEG as specified with the /SEGMENT switch.
The data words are arranged in pairs. The first word of each pair
contains a COMMON name in Radix-50 format (the four-bit code field
must contain 60). The second contains the length of the area to be
allocated.
For each COMMON entry found, LINK first determines whether the COMMON
area is already allocated. If not, LINK allocates it. If the area
has been allocated, the allocated area must be at least as large as
the current requested allocation.
COMMON blocks can be referenced from other block types as standard
globally defined symbols. However, a COMMON block must be initially
allocated by Block Type 20, Block Type 1074, Block Type 6 (for blank
COMMON), or by the /COMMON switch to LINK. Any attempt to initially
define a COMMON block with a standard global symbol definition causes
the LNKSNC error when the redefining Block Type 20 is later seen.
A-36
FIELD TEST
Block Type 21 (Sparse Data)
|=======================================================|
| 21 | Short Count |
|-------------------------------------------------------|
| Relocation Word (Zero) |
|-------------------------------------------------------|
| |
| Sub-Block |
| |
|-------------------------------------------------------|
.
.
.
|-------------------------------------------------------|
| |
| Sub-Block |
| |
|=======================================================|
Each sub-block is of the form:
|=======================================================|
| Long Count | Address |
|-------------------------------------------------------|
| Data Word |
|-------------------------------------------------------|
.
.
.
|-------------------------------------------------------|
| Data Word |
|=======================================================|
Block Type 21 contains data to be loaded sparsely in a large area.
The first word of each sub-block contains the long count for the
sub-block in the left half, and the address for loading the data words
in the right half.
If the first four bits of the first data word of each sub-block are
1100 (binary) then the word is assumed to be a previously defined
Radix-50 symbol of type 60; in this case the left half of the second
word is the sub-block count, and the right half plus the value of the
symbol is the load address.
A-37
FIELD TEST
Block Type 22 (PSECT Origin)
|=======================================================|
| 22 | Short Count |
|-------------------------------------------------------|
| Relocation Word |
|-------------------------------------------------------|
| (SIXBIT PSECT Name) or (PSECT Index) |
|-------------------------------------------------------|
| PSECT Origin |
|=======================================================|
Block Type 22 contains the PSECT origin (base address).
Block Type 22 tells LINK to set the value of the relocation counter to
the value of the counter associated with the given PSECT name. All
following TWOSEG REL blocks are relocated with respect to this PSECT
until the next Block Type 22, 23, or 1051 is found.
When data or code is being loaded into this PSECT, all relocatable
addresses are relocated for the PSECT counter.
MACRO generates a Block Type 22 for each .PSECT and .ENDPS pseudo-op
it processes. These Type 22 blocks are interleaved with the other
blocks to indicate PSECT changes. A Type 22 block is also generated
at the beginning of each symbol table to show to which PSECT the table
refers.
A-38
FIELD TEST
Block Type 23 (PSECT End Block)
|=======================================================|
| 23 | Short Count |
|-------------------------------------------------------|
| Relocation Word |
|-------------------------------------------------------|
| PSECT Index |
|-------------------------------------------------------|
| PSECT Break |
|=======================================================|
Block Type 23 contains information about a PSECT.
The PSECT index uniquely identifies the PSECT within the module being
loaded. The Type 24 block assigns the index.
The PSECT break gives the length of the PSECT. This break is
interpreted as being relative to the PSECT's origin in the current
module.
A-39
FIELD TEST
Block Type 24 (PSECT Header Block)
|=======================================================|
| 24 | Short Count |
|-------------------------------------------------------|
| Relocation Word |
|-------------------------------------------------------|
| PSECT Name (SIXBIT) |
|-------------------------------------------------------|
| Attributes | PSECT Index |
|-------------------------------------------------------|
| PSECT Origin (optional) |
|=======================================================|
Block Type 24 contains information concerning a specified PSECT. The
first word contains the block type number and the number of words
associated with the block. The second word contains the relocation
information. The third word contains the PSECT name in SIXBIT. The
fourth word is the PSECT origin specified for this module.
Bit Interpretation MACRO .PSECT Keyword
11 PSECT is all within one
section. This is the default.
12 PSECT is in a nonzero
section.
13 PSECT is page-aligned. PALIGNED
14 Concatenate parts of CONCATENATE
PSECTs seen in distinct modules.
15 Overlay parts of PSECTs OVERLAY
seen in distinct modules.
16 Read-only RONLY
17 Read and write RWRITE
LINK must find a Type 24 or 1050 block for a PSECT before it finds the
index for that PSECT. (MACRO generates a complete set of Type 24
blocks for all PSECTs in a module before generating Type 2 (Symbol
Table) Blocks and Type 11 (POLISH) Blocks.)
A-40
FIELD TEST
Block Type 37 (COBOL Symbols)
|=======================================================|
| 37 | Short Count |
|-------------------------------------------------------|
| Relocation Word |
|-------------------------------------------------------|
| Data Word |
|-------------------------------------------------------|
.
.
.
|-------------------------------------------------------|
| Data Word |
|=======================================================|
Block Type 37 contains a debugging symbol table for COBDDT, the COBOL
debugging program. If local symbols are being loaded, the table is
loaded.
If a REL file contains a Block Type 37, it must appear after all other
blocks that cause code to be loaded or storage to be allocated in the
core image.
This block is in the same format as the Type 1 REL Block.
A-41
FIELD TEST
Block Type 100 (.ASSIGN)
|=======================================================|
| 100 | Short Count |
|-------------------------------------------------------|
| Relocation Word |
|-------------------------------------------------------|
|Code | Radix-50 Symbol 1 |
|-------------------------------------------------------|
|Code | Radix-50 Symbol 2 |
|-------------------------------------------------------|
| Offset |
|=======================================================|
Block Type 100 defines Symbol 1 (in the diagram above) as a new global
symbol with the current value of Symbol 2, and then increases the
value of Symbol 2 by the value of the given offset.
LINK ignores Code in Symbol 2.
NOTE
Symbol 2 must be completely defined when the Block
Type 100 is found.
The MACRO pseudo-op .ASSIGN generates a Type 100 REL Block.
A-42
FIELD TEST
Block Type 776 (Symbol File)
|=======================================================|
| 776 | Long Count |
|-------------------------------------------------------|
| .JBSYM-Style Symbol Table Pointer |
|-------------------------------------------------------|
| .JBUSY-Style Symbol Table Pointer |
|-------------------------------------------------------|
| Data Word |
|-------------------------------------------------------|
.
.
.
|-------------------------------------------------------|
| Data Word |
|=======================================================|
Block Type 776 must begin in the first word of the file, if it occurs
at all. This block type shows that the file is a Radix-50 symbol
file.
The data words form a Radix-50 symbol table for DDT in the same format
as the table loaded for the switches /LOCALS/SYMSEG or the switch
/DEBUG.
A-43
FIELD TEST
Block Type 777 (Universal File)
|=======================================================|
| 777 | Long Count |
|-------------------------------------------------------|
| Data Word |
|-------------------------------------------------------|
.
.
.
|-------------------------------------------------------|
| Data Word |
|=======================================================|
Block Type 777 is included in a universal (UNV) file that is produced
by MACRO so that LINK will recognize when a UNV file is being loaded
inadvertently. When a Block Type 777 is encountered, LINK produces a
?LNKUNS error.
A-44
FIELD TEST
Block Type 1000 (Ignored)
|=======================================================|
| 1000 | Long Count |
|-------------------------------------------------------|
| Data Word |
|-------------------------------------------------------|
.
.
.
|-------------------------------------------------------|
| Data Word |
|=======================================================|
Block Type 1000 is ignored by LINK.
A-45
FIELD TEST
Block Type 1001 (Entry)
|=======================================================|
| 1001 | Long Count |
|-------------------------------------------------------|
| Symbol |
|-------------------------------------------------------|
.
.
.
|-------------------------------------------------------|
| Symbol |
|=======================================================|
Block type 1001 is used to declare symbolic entry points. Each word
contains one SIXBIT symbol. This block is similar in function to
block type 4.
A-46
FIELD TEST
Block Type 1002 (Long Entry)
|=======================================================|
| 1002 | Long Count |
|-------------------------------------------------------|
| Symbol Name |
|-------------------------------------------------------|
.
.
.
|-------------------------------------------------------|
| Symbol Name |
|=======================================================|
Block type 1002 is used to declare a symbolic entry point with a long
name in SIXBIT. The count reflects the symbol length in words.
A-47
FIELD TEST
Block Type 1003 (Long Title)
=======================================================
| 1003 | Long Count |
|-----------------------------------------------------|
| 1 | Count of Title words |
|-----------------------------------------------------|
| Program Title |
|-----------------------------------------------------|
| Additional Program Title |
|-----------------------------------------------------|
| Additional Program Title |
|-----------------------------------------------------|
.
.
.
|-----------------------------------------------------|
| | Count of |
| 2 | ASCII Comment Words |
|-----------------------------------------------------|
| More Comment Words |
|-----------------------------------------------------|
| More Comment Words |
|-----------------------------------------------------|
.
.
.
|-----------------------------------------------------|
| 3 | Count Of Compiler Words |
|-----------------------------------------------------|
| Compiler Code | CPU Bits |
|-----------------------------------------------------|
| Compiler Name (in ASCII) |
|-----------------------------------------------------|
| Additional Compiler Name |
|-----------------------------------------------------|
| Additional Compiler Name |
|-----------------------------------------------------|
.
.
.
------------------------------------------------------
| 4 | 0 |
|----------------------------------------------------|
| Compile Date and Time |
|----------------------------------------------------|
| Compiler Version Number |
|----------------------------------------------------|
| 5 | 0 |
|----------------------------------------------------|
| Device Name |
|----------------------------------------------------|
A-48
FIELD TEST
| UFD or 0 for TOPS-20 |
|----------------------------------------------------|
| 6 | 0 |
|----------------------------------------------------|
| TOPS-10 File Name |
|----------------------------------------------------|
| File Extension | 0 |
|----------------------------------------------------|
| 7 | number of SFDs |
|----------------------------------------------------|
| SIXBIT SFD 1 |
|----------------------------------------------------|
| SIXBIT SFD 2 |
|----------------------------------------------------|
.
.
.
|----------------------------------------------------|
| 10 | Count of TOPS-20 File |
| | Spec Words (in ASCII) |
|----------------------------------------------------|
| TOPS-20 File Spec |
|----------------------------------------------------|
| TOPS-20 File Spec |
|----------------------------------------------------|
.
.
.
|----------------------------------------------------|
| 11 | 0 |
|----------------------------------------------------|
| Source Version Number |
|----------------------------------------------------|
| Date and Time |
|====================================================|
Block type 1003 is used to declare long title symbols in SIXBIT and to
furnish other information about the source module. This Block Type
contains the information that LINK prints in the map file.
Block type 1003 consists of sub-blocks 1 through 11 (octal). The
Title sub-block must be the first sub-block specified and cannot be
omitted. You can omit other sub-blocks, but the sub-blocks must
remain in numerical order.
The Program Title is a one word title from 1- to 72-SIXBIT characters
long. You can specify a title of 0, and LINK defaults to .MAIN, but
you may want to enter a more specific title.
For the compiler code and the CPU code, refer to the explanation of
Block Type 6, where these codes are listed.
A-49
FIELD TEST
In sub-block 5, the Device name is where the source file resides and
the UFD and SFD words are 0 for TOPS-20.
In sub-block 10, the TOPS-20 file specification must be specified in
the following format:
<directory.subdirectories>filename.filetype.version number;attributes
This specification identifies the source file. LINK outputs this file
specification to the map file in the order you enter it.
The Time and Date are in TOPS-10 format. The date is derived from a
code. That is given by the following formula:
code = 31[12(year-1964)+(month-1)]+(day-1)
You can obtain the current day, month, and year using the formulas:
day = mod(code,31)+1
month = mod(code/31,12)+1
year = (code/372)+1964
The Time is the time in milliseconds that has elasped since midnight.
See the TOPS-10 Monitor Calls Reference Manual for additional
information on date and time.
A-50
FIELD TEST
Block Type 1004 (Byte Initialization)
|======================================================|
| 1004 | Long Count |
|------------------------------------------------------|
| Relocation Word |
|------------------------------------------------------|
| Byte Count |
|------------------------------------------------------|
| Byte Pointer |
|------------------------------------------------------|
| Byte String |
|------------------------------------------------------|
.
.
.
The above Block Type 1004 format is used to move a character string
into static storage. This format uses old style relocation.
The byte count is the number of bytes in the string. The byte pointer
is relocated and used to initialize a string in the user's program.
A second format for Block Type 1004 follows:
|=======================================================|
| 1004 | Long Count |
|-------------------------------------------------------|
| Relocation Word |
|-------------------------------------------------------|
| Global Symbol |
|-------------------------------------------------------|
| Byte Count |
|-------------------------------------------------------|
| Byte Pointer |
|-------------------------------------------------------|
| Byte String |
|-------------------------------------------------------|
.
.
.
In this format, the global symbol (in SIXBIT) is used to relocate the
byte pointer. The symbol must be defined when this REL block is
encountered.
A-51
FIELD TEST
Block Types 1010 - 1037 (Code Blocks)
Block types 1010 through 1037 are similar in function to blocks of
Type 1. They contain code and data to be loaded. These blocks also
contain relocation bytes that permit inclusion of PSECT indexes local
to the module. For PSECTed programs with many inter-PSECT references
this permits a substantial decrease in the size of the REL files. The
number of PSECTs that can be encoded in this manner is limited by the
size of the relocation byte. A set of parallel code blocks differing
only in the size of the relocation byte permits the compiler or
assembler to select the most space efficient representation according
to the number of PSECTs referenced in a given load module.
This set of blocks is divided by the type of relocation:
Right Relocation Block types 1010 - 1017
Left/Right Relocation Block types 1020 - 1027
Thirty-bit Relocation Block types 1030 - 1037
A-52
FIELD TEST
Blocks 1010 - 1017 (Right Relocation)
|=======================================================|
| 1010 | Long Count |
|-------------------------------------------------------|
| b1 | b2 | b3 | . . . | bi |
|-------------------------------------------------------|
| Beginning Address |
|-------------------------------------------------------|
| Data 1 |
|-------------------------------------------------------|
| Data 2 |
|-------------------------------------------------------|
.
.
.
|-------------------------------------------------------|
| Data (i-1) |
|=======================================================|
Block Types 1010 - 1017 are identical in function. They differ only
in the size and number of relocation bytes. Each relocation byte
applies to the right half of the corresponding data word.
Long Count is the length of the REL block, including all
words in the block except the Header word.
b1,b2...bi are the relocation bytes.
Each relocation byte contains a PSECT index
number. A PSECT index must be explicitly
specified. There is no default for PSECT
index. A zero byte means no relocation
(absolute data). All PSECT index numbers
must reference predefined PSECTs. In the
table below, I-value is the maximum number of
PSECT indexes that can be referenced in a
field.
Size I-value Block Type
2 18 1010
3 12 1011
6 6 1012
9 4 1013
18 2 1014
A size of 2 allows 3 PSECTs; a size of 3
allows 7 (2**3-1) PSECTs, etc.
A-53
FIELD TEST
Beginning Addr is the address where the block of code is to
be loaded. This address is relocated with
respect to the 30-bit address for the PSECT
in "b1". It is not necessary to declare the
current PSECT with a block of Type 22.
Data1...Data(i-1) are the words to be loaded. The right halves
of these words are relocated with respect to
the various PSECTs that are specified by the
corresponding relocation bytes, b2,b3,...bi.
A-54
FIELD TEST
Block Types 1020-1027 (Left/Right Relocation Blocks)
|=======================================================|
| 1020 | Long Count |
|-------------------------------------------------------|
| L1 | R1 | L2 | R2 | . . . | Li | Ri |
|-------------------------------------------------------|
| Beginning Address |
|-------------------------------------------------------|
| Data 1 |
|-------------------------------------------------------|
| Data 2 |
|-------------------------------------------------------|
.
.
.
|-------------------------------------------------------|
| Data (i-1) |
|=======================================================|
Block Types 1020 - 1027 are identical in function. They differ only
in the size and number of relocation bytes. Each pair of bytes
applies to the left and right halves, respectively, of the
corresponding data word.
Long Count is the length of the REL block, including all
words except the Header word.
L1,R1 are the relocation byte pairs for the left
and right halves respectively. In the table
below, I-value is the maximum number of PSECT
indexes that can be referenced in a field.
Size I-Value Block Type
2 9 1020
3 6 1021
6 3 1022
9 2 1023
(Block Types 1024-1027 are reserved)
Polish blocks must be used to do left
relocation if there are more than (2**9)-1
(decimal 511) PSECTs local to the module.
Beginning Addr is the address of the block of code to be
loaded. This address is relocated with
respect to the 30-bit address for the PSECT
in "R1". "L1" must be zero.
A-55
FIELD TEST
Data1,..Data(i-1) is the block of code to be loaded, whose left
and right halves are relocated with respect
to the various PSECTs as specified by the
corresponding byte pairs. The L2 index
relocates the left half of data word 1 and R2
relocates the right half of data word 1.
Note that these blocks contain 2 bytes for
each data word as compared to one byte for
Block Types 1010 - 1017.
A-56
FIELD TEST
Block Types 1030 - 1037 (Thirty-bit Relocation Blocks)
|=======================================================|
| 1030 | Long Count |
|-------------------------------------------------------|
| b1 | b2 | . . . | bi |
|-------------------------------------------------------|
| Beginning Address |
|-------------------------------------------------------|
| Data 1 |
|-------------------------------------------------------|
| Data 2 |
|-------------------------------------------------------|
.
.
.
|-------------------------------------------------------|
| Data (i-1) |
|=======================================================|
Block Types 1030 - 1037 are identical in function. They differ only
in the size and number of relocation bytes. Each relocation byte
applies to the entire 30-bit address field of the corresponding data
word.
103x is the Block Type number
Long Count is the length of the REL block, including all
words in the block except the Header word.
b1,b2..bi are the relocation bytes.
Each relocation byte contains a PSECT index
number. A zero byte means no relocation
(absolute data). All PSECT index numbers
must reference predefined PSECTs. In the
table below, I-value is the maximum number of
PSECT indexes that can be referenced in a
field.
Size I-Value Block Type Maximum No. of
PSECTs
2 18 1030 3
3 12 1031 7
6 6 1032 63
9 4 1033 511
18 2 1034 More than 511
(Block Types 1035 - 1037 are reserved)
Beginning Addr is the address where the block of code is to
A-57
FIELD TEST
be loaded. This address is relocated with
respect to the PSECT in "b1". It is not
necessary to declare the current PSECT with a
block of Type 22.
Data1...Data(i-1) are the words to be loaded. The 30-bit
address field of these words is relocated
with respect to the various PSECTs that are
specified by the corresponding relocation
bytes, b2,b3,...bi.
A-58
FIELD TEST
Block Type 1042 (Request Load for SFDs)
|=======================================================|
| 1042 | Long Count |
|-------------------------------------------------------|
| Device |
|-------------------------------------------------------|
| SIXBIT Filename |
|-------------------------------------------------------|
| File Extension | Directory Count |
|-------------------------------------------------------|
| Project-Programmer Number |
|-------------------------------------------------------|
| SFD1 |
|-------------------------------------------------------|
| SFD2 |
|-------------------------------------------------------|
.
.
.
Block Type 1042 contains a list of files to be loaded. It is similar
to blocks of Type 16, but it supplies TOPS-10 sub-file directories for
the files being requested. The first three data words (device, file
name, and extension) are required. The right half of the third word
(directory count) specifies the number of directory levels that are
included. For example, the directory [27,5434,SFD1,SFD2] would have a
directory count of 3.
LINK saves the specifications for the files to be loaded, discarding
duplicates. LINK loads all specified files at the end of loading, and
immediately before beginning library searches.
A-59
FIELD TEST
Block Type 1043 (Request Library for SFDs)
|=======================================================|
| 1043 | Long Count |
|-------------------------------------------------------|
| Device |
|-------------------------------------------------------|
| SIXBIT Filename |
|-------------------------------------------------------|
| File Extension | Directory Count |
|-------------------------------------------------------|
| Project-Programmer Number |
|-------------------------------------------------------|
| SFD1 |
|-------------------------------------------------------|
| SFD2 |
|-------------------------------------------------------|
.
.
.
Block Type 1043 specifies the files to be searched as libraries. It
is similar to Type 17 Blocks, except that it provides TOPS-10 sub-file
directories. The first three data words (device, file name, and
extension) are required. The right half of the third word (directory
count) specifies the number of directory levels that are included.
For example, the directory [27,5434,SFD1,SFD2] would have a directory
count of 3.
The specified files are searched after requested files are loaded, but
before user and system libraries are searched.
A-60
FIELD TEST
Block Type 1044 (ALGOL Symbols)
|=======================================================|
| 1044 | Long Count |
|-------------------------------------------------------|
| Data Word |
|-------------------------------------------------------|
.
.
.
|-------------------------------------------------------|
| Data Word |
|=======================================================|
Block Type 1044 contains a debugging symbol table for ALGDDT, the
ALGOL debugging program.
If an ALGOL main program has been loaded, or if you have used the
/SYFILE:ALGOL switch, LINK writes the data words into a SYM file. In
addition, if any Type 15 (ALGOL OWN) REL blocks have been seen, LINK
stores the file specification of the file into the first OWN block
loaded.
NOTE
If you have specified the /NOSYMBOLS switch, or if you
have specified the /SYFILE switch with an argument
other than ALGOL, then LINK ignores any Type 1044
blocks found.
A-61
FIELD TEST
Block Type 1045 (Writable Links)
|=======================================================|
| 1045 | Long Count |
|-------------------------------------------------------|
| Flags |
|-------------------------------------------------------|
| Symbol |
|-------------------------------------------------------|
| Symbol |
|-------------------------------------------------------|
.
.
.
|-------------------------------------------------------|
| Symbol |
|=======================================================|
Block type 1045 declares as writable either the link containing the
current module or the links containing the definitions of the
specified symbols or both. This block type must follow any common
block declarations (Types 20 or 6) in a module.
The flag word indicates which links are writable. If bit one is set
then the link containing the current module and the links containing
the definitions of the specified symbols are writable. If bit one of
the flag word is not set then the link containing the current module
is not writable, but the links containing the specified symbols are
writable. All unused flag bits are reserved and should be zero.
Any symbols specified in a block of Type 1045 must be defined in the
path of links leading from the root link to the current link. A
module cannot declare a parallel or inferior link to be writable.
If the symbol name contains six or fewer characters it is represented
in a single word, left justified, with the following format:
|=======================================================|
| SIXBIT Symbol Name |
|=======================================================|
A-62
FIELD TEST
If the symbol name contains more than six characters it is represented
in the following format:
Bits
0 5 6 29 30 35
|=======================================================|
| 0 | Reserved (0) | Long Count |
|-------------------------------------------------------|
| Word 1 of SIXBIT Symbol Name |
|-------------------------------------------------------|
| Word 2 of SIXBIT Symbol Name |
|-------------------------------------------------------|
.
.
.
|-------------------------------------------------------|
| Word (Long Count -1) of SIXBIT Symbol Name |
|=======================================================|
The first six bits of a long symbol are always 0. This distinguishes
a long symbol name from a single word symbol name. N is the length of
the symbol name including the header word. The remaining words
contain the symbol name in SIXBIT, six characters to a word, left
justified.
A-63
FIELD TEST
Block Type 1050 (Long PSECT Name Block)
0 17 18 35
|=======================================================|
| 1050 | Long Count |
|-------------------------------------------------------|
| Reserved for DIGITAL, MBZ | Index |
|-------------------------------------------------------|
| SIXBIT Symbol Name |
|-------------------------------------------------------|
| Attributes |
|-------------------------------------------------------|
| Origin (Optional) |
|=======================================================|
where SIXBIT Symbol Name may be either a word of up to six SIXBIT
characters, or the following block.
0 5 6 29 30 35
|=======================================================|
| 0 | Reserved for DIGITAL, MBZ | Long Count |
|-------------------------------------------------------|
| First word of SIXBIT Symbol Name |
|-------------------------------------------------------|
| Second word of SIXBIT Symbol Name |
|-------------------------------------------------------|
|-------------------------------------------------------|
| Nth word of SIXBIT Symbol Name |
|=======================================================|
Block Type 1050 creates a PSECT with the given name, if none currently
exists. It also assigns a unique index number to the PSECT. This
index is binding only in the current module. LINK clears PSECT
indexes at the end of each module. PSECT indexes in any given module
must be declared in consecutive order starting at index "1".
Blocks Type 1050 also assigns attributes to a PSECT and specifies the
PSECT's origin address. The attributes that can be assigned are:
Bit Description
11 PSECT is confined to one section. If this bit is set,
LINK gives an error if the PSECT overflows. You can
set Bit 11 or Bit 12, but not both. Bit 11 is the
default. There is no equivalent MACRO .PSECT Keyword.
12 PSECT is in a nonzero section. If this bit is set,
LINK gives a warning if the PSECT is placed in section
zero. There is no equivalent MACRO .PSECT Keyword.
A-64
FIELD TEST
13 PSECT is PAGE-ALIGNED. PALIGNED is the equivalent
MACRO .PSECT Keyword.
14 CONCATENATE parts of PSECTs seen in distinct modules.
You can set Bit 14 (CONCATENATED) or Bit 15 (OVERLAID),
but not both. The CONCATENATE and OVERLAID (listed
below) attributes are mutually exclusive. These
attributes also span modules; so if one module sets an
attribute and a later module sets a mutually exclusive
attribute, LINK issues the warning:
%LNKCOE Both CONCATENATE and OVERLAY attributes
specified for psect [name].
If neither is set, CONCATENATED is the default, and a
warning message is not returned if subsequent pieces of
the PSECT are marked OVERLAID.
CONCATENATED is the equivalent MACRO .PSECT Keyword.
15 OVERLAID parts of PSECTs seen in distinct modules.
OVERLAID is the equivalent MACRO .PSECT Keyword.
16 This PSECT must be READ-ONLY.
You can set Bit 16 (READ-ONLY) or Bit 17 (WRITABLE),
but not both. The READ-ONLY and WRITABLE (listed
below) attributes are mutually exclusive. These
attributes also span modules; so if one module sets an
attribute and a later module sets a mutually exclusive
attribute, LINK issues the warning:
%LNKRWA Both READ-ONLY and WRITABLE attributes
specified for psect [name]. If neither is set,
WRITABLE is the default, and a warning message is not
returned if subsequent pieces of the PSECT are marked
READ-ONLY.
RONLY is the equivalent MACRO .PSECT Keyword.
17 This PSECT must be WRITABLE. RWRITE is the equivalent
MACRO .PSECT Keyword.
All other bits in the Attributes word must be 0.
The origin specified in this block is absolute.
At least one block type 1050 (or the related block 24) is required for
each PSECT being loaded, and this block must be loaded prior to any
other blocks that reference its PSECT (that is, use the unique index
number).
A-65
FIELD TEST
Block Type 1051 (Set Current PSECT)
|=======================================================|
| 1051 | Long Count |
|-------------------------------------------------------|
| Reserved For DIGITAL, MBZ | Index |
|=======================================================|
Block type 1051 resets the "current PSECT" against which LINK
relocates subsequent REL blocks if no PSECT is explicitly specified.
A-66
FIELD TEST
Block Type 1052 (PSECT End)
|=======================================================|
| 1052 | Long Count |
|-------------------------------------------------------|
| MBZ | PSECT Index Number |
|-------------------------------------------------------|
| PSECT Break |
|-------------------------------------------------------|
.
.
.
|-------------------------------------------------------|
| MBZ | PSECT Index Number |
|-------------------------------------------------------|
| PSECT Break |
|=======================================================|
Block Type 1052 allocates additional space for a given PSECT. This
space is located between the last address in the PSECT containing data
and the address given by the PSECT break. A block of Type 1052 can
contain more than one pair of PSECT indexes and breaks.
A module must contain a block of Type 24 (PSECT Name) or Type 1050
(Long PSECT Name) with the given PSECT index before a block of Type
1052 is generated. If a given PSECT has more than one block 1052 in a
single module, the block with the largest break address is used.
The break is interpreted as being relative to the PSECT's origin in
the current module.
A-67
FIELD TEST
Block Type 1060 (Trace Block Data)
|=======================================================|
| 1060 | Long Count |
|-------------------------------------------------------|
| SIXBIT Edit Name |
|-------------------------------------------------------|
| Active Code | Last Changer |
|-------------------------------------------------------|
| Creator Code | 15-Bit Date Created |
|-------------------------------------------------------|
| Installer Code | 15-Bit Date Installed |
|-------------------------------------------------------|
| Reserved |
|-------------------------------------------------------|
| Edit Count | PCO Group Count |
|-------------------------------------------------------|
| |
\ /
Associated Edit Names And Codes
/ \
| |
|-------------------------------------------------------|
| |
\ /
Program Change Order Groups
/ \
| |
|=======================================================|
Block Type 1060 contains data used by the MAKLIB program. LINK
ignores this block type.
A-68
FIELD TEST
Block Type 1070 (Long Symbol Names)
|====================================================================|
| 1070 ! Long Count |
|--------------------------------------------------------------------|
| Code ! 0 ! N !P! R ! V ! 0 |
|--------------------------------------------------------------------|
| Left PSECT index ! Right PSECT index |
|--------------------------------------------------------------------|
| Value |
|--------------------------------------------------------------------|
| Name |
|--------------------------------------------------------------------|
\ \
\ N additional name words \
\ \
|--------------------------------------------------------------------|
\ \
\ V additional value words \
\ \
|====================================================================|
This block defines a long symbol. A symbol defined with this block
can:
o be output to the DDT symbol table. Symbols longer than 6
characters are truncated when output to the DDT symbol table.
o be output to LINK MAP if requested.
o have its value relocated as specified.
o resolve global requests.
The Long Symbol Name Block is divided into two sections, the basic and
the extension sections.
The basic section consists of four words: the flag word, an optional
PSECT index word, the value word, and name word.
The Flags word contains information about the type of symbol, the
length of the symbol name, and relocation. The optional word defines
the PSECT index. The Value word contains the symbol's value. The
Name word contains the symbol's name.
If the name or the value cannot fit in a single word, the block
contains an extension section that consists of as many words as are
necessary to accommodate the symbol name and the value. The length of
the symbol name and value is stored in the Flag word and determines
how many words are allocated for the long symbol name in the extension
section. The maximum size for the symbol is 72 characters. In the
case of a short symbol name only the basic section is used.
A-69
FIELD TEST
The following pages provide detailed information on the block. For
each word, the field, bits, and description is given.
Field Bits
Header Word
Description
Block Type 0-17 1070
Block Length 18-35 Number of words used in this block
Flag Word
Description
Code 0-8 A nine-bit code field:
bit 0 Must Be Zero
000 Program name
100 Local symbol definitions
110 Suppressed to DDT
120 MAP only
200 Global symbols completely
defined by one word
202 Undefined
203 Right fixup
204 Left fixup
205 Right and left fixups
206 30-bit fixup
207 Fullword fixup
210-217 Suppress to DDT
220-227 MAP only
240 to Global symbol request for
247 chain fixup
240 Ignored (No fixup)
241 Undefined
242 Undefined
243 RH fixup
244 LH fixup
245 Undefined
246 30-bit fixup
247 Fullword fixup
250 to Global request for
A-70
FIELD TEST
257 additive fixups (the value of
x has the same meaning as in
0-7 above)
260 to Global request for additive
267 symbol fixups (the value of x
has the same meaning as in 0-7
above)
300 Block names
A-71
FIELD TEST
NOTE
All symbols that require a fixup for their definition
must have the fixup block immediately following the
entry.
Field Bits Description
Flag Word (Continued)
0 9-10 Must Be Zero
N--Name length 11-17 If not zero, extended name field of
length n words is used, so that the
name occupies N+1 words.
P--PSECT Flag 18 If bit 18=0, relocate with respect to
the current PSECT. No PSECT numbers
are needed.
If bit 18=1, relocate with respect to
the PSECT specified in the next word.
R--Relocation Type 19-21 3-bit relocation type field.
0 Absolute
1 Right half
2 Left half
3 Both halves
4 30-bit
5 Fullword
V--Value field 22-28 Number of additional value words if
value is a long symbol.
0 29-35 Not used
PSECT Indexes
PSECT Indexes Exists only if bit 18 equals 1 in the
Flag word. Contains Left and Right
PSECT numbers. Bit 0 and bit 18 of
this word are zeros.
Value
Value Word Contains the symbol value, it may be
relocated as specified by the
relocation type and the PSECT numbers
provided. Contains a symbol for 26x
A-72
FIELD TEST
codes.
Name
Name Word Contains the symbol name in SIXBIT.
N Additional Name Words
Additional name field Optional. It exists only if N > 0.
It contains the additional characters
when a long symbol name is used.
V Additional Value Words
Additional value field Optional. It exist only if the V
field is greater than 0. This field
contains the additional characters
when a long symbol name is being
resolved. The first word contains
the length of the extended field.
The following fixup rules apply to this block:
1. Only one fixup by a Type 2, 10, 11, 12, 15, 1070, 1072, or
1120 Block is allowed for a given field. (There can be
separate fixups for the left and right halves of the same
word.)
2. Fixups are not necessarily performed in the order LINK finds
them.
3. A location must contain data before the location can be fixed
up.
A-73
FIELD TEST
Block Type 1072 (Long Polish Block)
|=======================================================|
| 1072 | Long Count |
|-------------------------------------------------------|
| Half-Word Polish String |
|-------------------------------------------------------|
.
.
.
|=======================================================|
Long Polish Blocks of type 1072 define Polish fixups for operations on
relocatable long external symbols. This Block Type is interpreted as
a string of 18-bit operators and operands. The block is in Polish
prefix format, with the store operator at the end of the block. Each
halfword can contain one of the following:
o A halfword code in which the first 9 bits contain the data
length (when applicable) and the second 9 bits contain the
code telling LINK how to interpret the data that follows.
o A halfword data or a part of a larger data packet to be
interpreted by LINK as indicated by the code that immediately
precedes it.
o A PSECT index of the format 400000+N. The PSECT index field
of a long Polish block causes LINK to relocate addresses
against the PSECT number specified in the "N" of the PSECT
index 400000+N.
o A Polish operator.
NOTE
Operations are performed in the order in which they
are encountered.
A-74
FIELD TEST
CODE DEFINITIONS
Data Packet Codes
Category Code Description
Operand xxxyyy next "xxx+1" halfwords contain data of
type "yyy"
000000 halfword - absolute
001000 fullword - absolute
000001 halfword - relocatable
001001 fullword - relocatable
000010 fullword symbol name in Radix-50
xxx010 xxx+1 halfwords of symbol name in
SIXBIT
NOTE
You cannot store a symbol in a single halfword. You
must place the symbol in the first halfword and fill
the second halfword with zeroes.
Polish Operator Codes
Category Code Description
Operator
000100 Add
000101 Subtract
000102 Multiply
000103 Divide
000104 Logical AND
000105 Logical OR
000106 Logical shift
000107 Logical XOR
000110 One's complement (not)
000111 Two's complement (negative)
000112 Count leading zeros
000113 Remainder
000114 Magnitude
000115 Maximum
000116 Minimum
000117 Equal relation
000120 Link
000121 Defined
000122-00177 Reserved
A-75
FIELD TEST
Store Operator Codes
Store Operator xxx=0 or 1
For xxx=0 Next two halfwords contain a Radix-50
symbol to be resolved.
xxx777-xxx770 Chained fixup with relocatable
addresses. Next xxx+1 halfwords
contain the start address of the
chain.
000777 Right half chained fixup with
relocatable address. Next halfword
contains a relocatable address.
000776 Left half chained fixup with
relocatable address. Next halfword
contains a relocatable address.
000775 30-bit chained fixup with relocatable
address. Next halfword contains a
relocatable address.
000774 Fullword chained fixup with
relocatable address. Next halfword
contains a relocatable address.
001777 Right half chained fixup with
relocatable address. Next fullword
contains a relocatable address.
001776 Left half chained fixup with
relocatable address. Next fullword
contains a relocatable address.
001775 30-bit chained fixup with relocatable
address. Next fullword contains a
relocatable address.
001774 Fullword chained fixup with
relocatable address. Next halfword
contains a relocatable address.
xxx767-xxx764 Chained fixups with absolute
addresses.
000767-000764 Chained fixup with absolute address.
Next halfword contains an absolute
address.
001767-001764 Chained fixup with absolute fullword
A-76
FIELD TEST
address. Next two halfwords contain
absolute address.
xxx757-xxx754 Symbol fixup. For 1<=xxx<=377 the
next xxx+1 halfwords contain a SIXBIT
symbol name to be resolved.
xxx757 Right half symbol fixup.
xxx756 Left half symbol fixup.
xxx755 30-bit symbol fixup.
xxx754 Fullword symbol fixup.
xxx747-xxx700 Not defined
PSECT index 4000000+N PSECT index for PSECT N.
The following fixup rules apply to this block:
1. Only one fixup by a Type 2, 10, 11, 12, 15, 1070, 1072, or
1120 Block is allowed for a given field. (There can be
separate fixups for the left and right halves of the same
word.)
2. Fixups are not necessarily performed in the order LINK finds
them.
3. A location must contain data before the location can be fixed
up.
A-77
FIELD TEST
Block Type 1074 (Long Common Name)
|=======================================================|
| 1074 | Long Count |
|-------------------------------------------------------|
| PSECT Index | Symbol Length |
|-------------------------------------------------------|
| Common Block Length |
|-------------------------------------------------------|
| Symbol |
| (More Symbol) |
==========================================================
Block Type 1074 defines a long COMMON name.
A-78
FIELD TEST
Block types 1120-1127 (Argument Descriptor Blocks)
|=======================================================|
| 1120 - 1127 | Long Count |
|-------------------------------------------------------|
| N-Bit Byte Relocation Information |
|-------------------------------------------------------|
| Argument Block Address or 0 |
|-------------------------------------------------------|
| Associated Call Address or 0 |
|-------------------------------------------------------|
| Loading Address or 0 |
|-------------------------------------------------------|
| Length of Function Name (in bytes) |
|-------------------------------------------------------|
| Function Name (ASCIZ) |
|-------------------------------------------------------|
.
.
.
|-------------------------------------------------------|
| Flag Bits | Argument Count |
|-------------------------------------------------------|
| First Argument's Primary Descriptor |
|-------------------------------------------------------|
| First Argument's Secondary Descriptor |
|-------------------------------------------------------|
| Second Argument's Primary Descriptor |
|-------------------------------------------------------|
| Second Argument's Secondary Descriptor |
|-------------------------------------------------------|
.
.
.
|-------------------------------------------------------|
| nth Argument's Primary Descriptor |
|-------------------------------------------------------|
| nth Argument's Secondary Descriptor |
|=======================================================|
A block of this type is generated for the argument list to each
subroutine call. The subroutine entry point also specifies one block
with this format, though for the callee the argument block address is
zero. If a descriptor block is associated with an argument list it
must always follow the loading of the argument list.
The associated call address is used by LINK in diagnostic error
messages and its value is determined by the compiler. The argument
block address is nonzero if the descriptor block is associated with a
call. In this case the argument block address points to the base of
the argument block.
A-79
FIELD TEST
The argument block address, associated call address and the loading
address are all relocatable.
The argument descriptors in these type blocks describe the properties
of each formal (in the case of an entry point) or actual (in the case
of a call). In either case the name of the associated routine is
specified as a byte count followed by an ASCIZ string. Each primary
description is optionally followed by a secondary descriptor.
There are five flag bits in the Descriptor Block:
Bit Usage
0 If bit 0 is 1 then a difference between the actual
number of arguments and the expected number of
arguments is flagged as a warning at load time. If
bit 0 is 0 no action is taken.
1 If bit 1 is 1 then the block is associated with a
function call. If bit 1 is 0 then the block is
associated with the function definition.
2 If bit 2 is 1 then the descriptor block is loaded
into user memory at the loading address. This bit
is ignored.
3 If bit 3 is 1 then the callee returns a value and
the value's descriptor is the last descriptor
specified.
4 If bit 4 is 1, and the caller expects a return
value, which is not provided by the called function,
or if the called function unexpectedly returns a
value, then LINK will issue an error. The severity
of the error is controlled by the coercion block.
A-80
FIELD TEST
The format for the argument descriptors is as follows:
Bit Usage
0 (Reserved)
1 No update. In a caller block the argument is a
literal, constant, or expression. In a callee block
the argument won't be modified.
2-4 Passing mechanism
000 - pass by address
001 - pass by descriptor
010 - pass immediate value
Others - reserved
5 Compile-time constant
6-11 Argument type code (see below)
12-18 (Reserved)
19-26 (Reserved)
27-35 Number of secondary descriptors
The argument type codes are as follows:
Type-Code Usage
0 No use
1 FORTRAN logical
2 Integer
3 (Reserved)
4 Real
5 (Reserved)
6 36-bit string
7 Alternate return (label)
10 Double real
11 Double integer
12 Double octal
13 G-floating real
14 Complex
15 COBOL format byte string descriptor (for constant
strings), or FORTRAN character for a program
compiled with FORTRAN /NOEXTEND switch
16 BASIC shared string descriptor
17 ASCIZ string
A-81
FIELD TEST
20 Seven-bit ASCII string
21 FORTRAN character, one-word global byte pointer for
a program compiled with /EXTEND
A-82
FIELD TEST
Secondary descriptors are used to convey information about the length
of a data object passed as an argument and (in the case of the
callee's argument descriptor block) whether or not a mismatched length
is permissible. Secondary descriptors have the following format:
Bit Pos Usage
0-2 (For callee only) Defines the permissible relationships
between formal and actual lengths. The values are:
000 - Any relationships are allowed
001 - Lengths must be equal
010 - Actual < formal
011 - Actual <= formal
100 - Actual > formal
101 - Actual >= formal
110 - Reserved
111 - Reserved
3-5 Length of argument (in words)
A-83
FIELD TEST
Block Type 1130 (Coercion Block)
|=======================================================|
| 1130 | Long Count |
|-------------------------------------------------------|
| Field Code | Action |
|-------------------------------------------------------|
| Formal Attribute | Actual Attribute |
|-------------------------------------------------------|
| Field Code | Action |
|-------------------------------------------------------|
| Formal Attribute | Actual Attribute |
|-------------------------------------------------------|
.
.
.
|-------------------------------------------------------|
| Field Code | Action |
|-------------------------------------------------------|
| Formal Attribute | Actual Attribute |
|=======================================================|
Block Type 1130 specifies which data type associations are permissible
and what action LINK should take if an illegal type association is
attempted. It may also specify actions to be taken by LINK to modify
an actual parameter.
The Coercion Block must be placed before any instance of the
caller/callee descriptor block in the REL file. If more than one
coercion block is seen during a load, the last block seen is used for
type checking.
If the description block and command strings are not in the same
section, no error message is given.
When a caller's argument descriptor block is compared to the
descriptor block provided by the callee, LINK first checks bit 0 and
the argument counts of the descriptor block. If bit 0 is set and the
argument counts differ, a warning is given. However, if a byte
description is not word-aligned, no warning is given.
Next LINK compares the argument descriptors. The particular
formal/actual pair is looked up in the internal table LINK builds
using the information in the coercion block. The item field code
designates which field of the argument descriptor is being checked.
The field codes are defined as follows:
Field Code Condition
0 Check update
1 Check passing mechanism
2 Check argument type code
A-84
3 Check if compile-time constant
4 Check number of arguments
5 Check for return value
6 Check length of argument
If the fields of the formal/actual pair do not match, LINK searches
the internal table set up by the coercion block. If the table does
not specify an action to take in the event of such a mismatch, LINK
issues an informational message. If the formal/actual pair differs in
more than one field then LINK takes the most severe action specified
for the mismatches.
If an actual/formal pair differ and no coercion block has been seen,
LINK ignores the difference. If the caller has specified a descriptor
block but the subroutine has not, or if the subroutine has specified a
descriptor and the caller has not, LINK does not flag the condition as
an error and does not take any special action.
If LINK finds an entry in its internal table for a particular
actual/formal mismatch, it uses the action code found in the entry to
select one of the following five possible responses:
Code (18 Bits) Action
0 Informational message
1 Warning
2 Error
3 Reserved for the specific conversion of
static descriptor pointers (in the argument
list) into addresses. The descriptor
pointers are supplied by FORTRAN blocks of
types 112x.
NOTE
The actual conversion process
involves the following actions:
o If byte descriptor's P field is
not word-aligned, issue a warning
and continue.
o Pick up word address of start of
string.
o If the string is not in the same
section as the argument block,
nonfatal error and continue.
A-85
FIELD TEST
o Put the address of the string
into the associated argument
block in place of the address of
the string descriptor.
4 Suppress the message.
5-777776 Reserved
777777 Fatal error
These messages can be displayed or suppressed. Refer to the
descriptions of the /ERRORLEVEL and /LOGLEVEL switches.
A-86
FIELD TEST
Block Type 1131 (TWOSEG Redirection Block)
|=====================================================|
| 1131 | Long Count |
|-----------------------------------------------------|
| Name of PSECT for low segment, or 0 |
|-----------------------------------------------------|
| Name of PSECT for high segment, or 0 |
|=====================================================|
where each PSECT name has the form:
|=====================================================|
| SIXBIT Symbol Name |
|=====================================================|
or
0 5 6 17 18 29 30 35
|=======================================================|
| 0 | Reserved for DIGITAL, MBZ | Long Count |
|-------------------------------------------------------|
| First word of SIXBIT Symbol Name |
|-------------------------------------------------------|
| Second word of SIXBIT Symbol Name |
|-------------------------------------------------------|
|-------------------------------------------------------|
| Nth word of SIXBIT Symbol Name |
|=======================================================|
Block Type 1131 permits TWOSEG REL modules to be loaded into PSECTs by
a compiler. You must redirect both the high and the low segment, you
cannot redirect one or the other. Also, you cannot redirect both the
high and low segment into the same PSECT.
This block does not affect the current module, but all subsequent
modules to be loaded.
A-87
FIELD TEST
Block Type 1140 (PL/1 debugger information)
|=====================================================|
| 1140 | Long Count |
|-----------------------------------------------------|
| Data Word |
|-----------------------------------------------------|
.
.
.
|-----------------------------------------------------|
| Data Word |
|=====================================================|
Block Type 1140 is ignored by LINK.
A-88
FIELD TEST
Block Type 1160 (Extended Sparse Data Initialization Block)
=======================================================
| 1160 | Long Count |
|-----------------------------------------------------|
| R | F | B | P | 0 | SYMLEN | PSECT |
|-----------------------------------------------------|
| Symbol (SYMLEN words) |
|-----------------------------------------------------|
| S | Origin Address |
|-----------------------------------------------------|
| Repetition Count if R=1 |
|-----------------------------------------------------|
| Fill Count if F=1 |
|-----------------------------------------------------|
| Fill Byte if F=1 |
|-----------------------------------------------------|
| Byte Count if B=1 |
|-----------------------------------------------------|
| Data Bytes |
=======================================================
Block Type 1160 supports the loading of data into different PSECTs and
sections. This REL Block allow separate program units to load data
into different bytes in the same word of memory at different times
during the loading process.
Block Type 1160 fields are described below.
Field Name Position Description
R Bit 0 is a 1-bit field. If R is
one, the Repetition Count word
exists. If R is zero, the
Repetition Count is assumed to
be 1.
F Bit 1 is a 1-bit field. If F is
one, the Fill Count and Fill
Byte words exist. If F is
zero, no fill is used.
B Bit 2 is a 1-bit field. If B is
one, the Byte Count word
exits. If B is zero, one Data
Byte is assumed.
P Bits 3-8 is a 6-bit field. This is the
position within the word where
the first byte is to stored.
Unused Bit 9 is an unused bit that must be
A-89
FIELD TEST
zero.
SYMLEN Bits 10-17 is an 8-bit field. SYMLEN is
the length in words of the
global symbol to be used to
calculate the address to store
the byte string. If SYMLEN is
zero, there is no global
symbol. The value of the
symbol is added to the origin
address. The symbol must be
completely defined before this
addition occurs.
PSECT Bits 18-35 is an 18-bit field. PSECT is
the PSECT to relocate the
Origin Address against. The
relocation is 30-bit. If
PSECT is zero, the Origin
Address is absolute.
Symbol Bits 0-35 is a SIXBIT symbol name of the
length specified in SYMLEN.
The value of this symbol is
added to the Origin Address.
The symbol must be defined
when the block is seen, or a
fatal error occurs.
S Bits 0-5 is a 6-bit field. S is the
size of the data bytes.
Origin Address Bits 6-35 is a 30-bit field. Origin
Address is the address where
LINK begins to store Data
Bytes.
Repetition Count Bits 0-35 is a 36-bit field. If flag
bit R is one, Repetition Count
exists and contains the number
of times to repeat the data
store. The Data Bytes are
stored and the fill operation
is performed as many times as
specified in the Repetition
Count.
Fill Count Bits 0-35 is a 36-bit field. If flag
bit F is one, Fill Count
exists and specifies how many
times to store the Fill Byte
after storing Data Bytes.
A-90
FIELD TEST
Fill Byte Bits 0-35 is a 36-bit field. If flag
bit F is one, Fill Byte exists
and contains the right
justified value to be used in
the fill operation.
Byte Count Bits 0-35 is a 36-bit field. If flag
bit B is one, Byte Count exits
and specifies the number of
Data Bytes to be stored.
Data Bytes Bits 0-35 are the data to be stored, of
the length specified by the
Byte Count, or 1 if flag bit B
is not set. This data is
stored left-justified, packed
as many to a word as will fit
without overlapping a word
boundary.
A-91
FIELD TEST
Block Type Greater Than 3777 (ASCIZ)
|=======================================================|
| ASCII | ASCII | ASCII | ASCII | ASCII | 0 |
|-------------------------------------------------------|
| ASCII | ASCII | ASCII | ASCII | ASCII | 0 |
|-------------------------------------------------------|
.
.
.
|-------------------------------------------------------|
| ASCII | ASCII | ASCII | ASCII | 0 | 0 |
|=======================================================|
When LINK reads a number larger than 3777 in the left half of a REL
Block header word, the block is assumed to contain ASCIZ text. If the
module containing the text is being loaded, LINK reads the ASCII
characters as if they were a command string, input from the user's
terminal.
LINK reads the string as five 7-bit ASCII characters per word; bit 35
of each word is ignored. The string and the block end when the first
null ASCII character (000) is found in the fifth 7-bit byte of a word
(bits 28-34).
After loading the current REL file, LINK processes text statements in
the reverse order in which they are encountered -- from the end to the
beginning of the module. For example, the first, second, and third
statements from the beginning of a module are processed third, second,
and first. As a result, search requests in .TEXT blocks may be
processed in the reverse order of entered /SEARCH switches. Keep this
in mind when specifying the order the modules are to be searched.
A-92
APPENDIX B
LINK MESSAGES
This appendix lists all of LINK's messages except the messages from
the overlay handler. Overlay handler messages have the OVL prefix and
appear in Chapter 5. Section B.1 describes the format of messages and
Section B.2 lists LINK's messages.
B.1 DESCRIPTION OF MESSAGES
For each message, the last three letters of the 6-letter code, the
level, the severity, and its medium-length message are given in bold.
The long message follows, and appears beneath the medium message.
CODE LEV SEV MEDIUM MESSAGE
LONG MESSAGE
When a message is issued, the three letters are appended to the
letters LNK, forming a 6-letter code of the form LNKxxx. For example,
EXS is the 3-letter code for EXIT segment. EXS is appended to LNK to
form the 6-letter code LNKEXS.
B.1.1 Message Levels
The level of a message determines whether it is returned to the
terminal, the log file, or both. You can use /ERRORLEVEL to control
message output to the terminal and /LOGLEVEL to control message output
to the log file. Both switches suppress messages with a level of 8 or
below. For some messages an asterisk (*) is given for the level or
severity. This means that the value is variable, and depends on the
conditions that generated the message.
B-1
FIELD TEST
B.1.2 Message Severity
The severity of a message determines whether the load is terminated
when the message is issued. Table B-1 lists the severity codes used
in LINK, with their meanings. Severity codes are decimal. The
/SEVERITY switch provides a means for lowering the severity that is
considered fatal. For example, if the severity is set at 28 (the
default for timesharing jobs) and you want to terminate the load if
you receive a warning message, you can lower the severity to 16. This
causes LINK to terminate the load if a message with a severity of 16
is issued.
The severity also determines the first character on the message line
output to the terminal. This character can then be detected by the
batch system. For all informational messages, the character is a
square bracket ([). Informational messages also end in a square
bracket (]). Warnings use percent sign (%), and fatal errors use
question mark (?).
Table B-1: Severity Codes
Decimal Meaning
Code
1-7 Informational; messages of this severity generally
indicate LINK's progress through the load.
8-15 Warning; LINK is able to recover by itself and continue
the load.
16 Warning if timesharing, but fatal and stops the load if
running under batch.
20 Fatal; LINK can only partially recover and continue the
load. The loaded program may be incorrect. Undefined
symbols cause this action.
24 This is for file access errors. Under batch, this is
fatal and stops the load. Under timesharing, this is a
warning, and LINK prompts for the correct file
specification if possible.
31 Always fatal; LINK stops the load.
B.1.3 Message Length
The /VERBOSITY switch determines whether the medium-length and long
B-2
FIELD TEST
messages are issued. If you use /VERBOSITY:SHORT, only the 6-letter
code is issued. For example,
[LNKEXS]
If you use /VERBOSITY:MEDIUM, the medium-length message is also
issued. For example,
[LNKEXS EXIT segment]
If you use /VERBOSITY:LONG, the code, medium-length message, and long
message are issued. For example,
[LNKEXS EXIT segment
LINK is in the last stages of loading your program
(for example, creating .EXE and symbol files,
preparing execution if requested).]
The /MESSAGE switch also specifies message length.
B.1.4 Message Conventions
{} indicate optional portions of the medium-length
message, and are printed only in appropriate
circumstances.
The JSYS error text optional message is a monitor call
message. Refer to the TOPS-20 Monitor Calls Reference
Manual for a description of these messages.
[] contain information pertinent to the particular error.
This portion of the medium-length messages is filled in
at runtime. Table B-2 describes each of these
bracketed quantities.
file identifies the module or file where the error occurred.
Whenever possible, LINK attempts to indicate the module
and file associated with an error. This information
represents the module currently being processed by
LINK, and may not always be the actual module
containing the error. For instance, if LINK detects a
multiply-defined symbol, either value may be the
incorrect one. In this case, LINK reports only the
last and latest redefinition and the module containing
it.
B-3
FIELD TEST
Table B-2: Special Message Segments
Segment Description
[area] The name of one of LINK's internal memory
management areas. See the /FRECOR switch in
Chapter 3 for a list of these areas.
[date] The date when LINK is running.
[decimal] A decimal number.
[device] A device name.
[file] A file specification.
[label] An internal label in LINK.
[memory] A memory size, such as 17P.
[name] The name of the loaded program or a node in an
overlaid program.
[octal] An octal number.
[reason] The reason for a file access failure, one of the
messages shown in Section B.3.
[switch] The name of a switch associated with the error.
[symbol] The name of a symbol, such as a subroutine or
common block name.
[type] The type or attribute associated with a symbol.
B.2 LIST OF MESSAGES
Code Lev Sev Message
ABT 31 31 Load aborted due to %LNKTMA errors, max.
/ARSIZE: needed was [decimal]
You loaded programs containing more ambiguous
subroutine requests than can fit in the tables
of one or more overlay links. You received a
LNKARL message for each ambiguous request, and a
B-4
FIELD TEST
LNKTMA message for each link with too many
requests. You can solve this problem by using
the /ARSIZE switch just before each /LINK switch
to expand the tables separately.
AIC 31 31 Attempt to increase size of {blank common}
{common [symbol]} from [decimal] to [decimal]
{Detected in module [symbol] from file [file]}
FORTRAN common areas cannot be expanded once
defined. Either load the module with the
largest definition first, or use the /COMMON:
switch to reserve the needed space.
AMM + + Argument mismatch in argument [decimal] in call
to routine [symbol] called from module [symbol]
at location [octal]
The caller supplied argument does not match the
argument expected by the callee.
[1]
AMP 8 8 ALGOL main program not loaded
You loaded ALGOL procedures, but no main
program. The missing start address and
undefined symbols will cause termination of
execution.
ANM 31 31 Address not in memory
LINK expected a particular user address to be in
memory, but it is not there. This is an
internal LINK error. This message is not
expected to occur. If it does, please notify
your Software Specialist or send a Software
Performance Report (SPR) to DIGITAL.
ARL 8 8 Ambiguous request in link [decimal] {name
[name]} for [symbol] defined in links [decimal],
[decimal], ...
More than one successor link can satisfy a call
from a predecessor link. The predecessor link
requested an entry point that is contained in
---------------
[1] + The level and severity of this message is determined by a
compiler-generated coercion block. A coercion block specifies
which data type associations are permissible and what action
LINK should take if an illegal type association is attempted.
See Block Type 1130 in Appendix A for more information.
B-5
FIELD TEST
two or more of its successors. You should
revise your overlay structure to remove the
ambiguity.
If you execute the current load, one of the following
occurs when the ambiguous call is executed:
o If only one module satisfying the request is
in memory, that module will be called.
o If two or more modules satisfying the
request are in memory, the one with the most
links in memory will be called.
o If no modules satisfying the request are in
memory, the one with the most links in
memory will be called.
If a module cannot be selected by the methods 2
or 3 above, an arbitrarily selected module will
be called.
AZW 31 31 Allocating zero words
LINK's memory manager was called with a request
for 0 words. This is an internal LINK error.
This message is not expected to occur. If it
does, please notify your Software Specialist or
send a Software Performance Report (SPR) to
DIGITAL.
CBO 31 31 Cannot build overlays outside section zero
You have tried to build an overlay structure for
a program that is either too large to fit in
section zero or which loads outside section zero
by default. Check your LINK commands and also
the assembly or compiler commands used to
produce the program modules.
CCD 31 31 CPU conflict
{Detected in module [symbol] from file [file]}
You have loaded modules compiled with
conflicting CPU specifications, such as loading
a MACRO program compiled with the statement
.DIRECTIVE KL10 and another compiled with
.DIRECTIVE KI10. Recompile the affected modules
with compatible CPU specifications.
CCE 8 8 Character constant not word aligned in call to
routine [routine] called from module [module] at
B-6
FIELD TEST
location [address]
Link has detected a character constant that did
not begin on a word boundary. This error is
most likely the result of a language translator
error. This error is not expected to occur. If
it does, please notify your Software Specialist
or send a Software Performance Report
CCS 31 31 Cannot create section [octal]
{Detected in module [symbol] from file [file]}
LINK is unable to create the specified section.
This could be because your system does not have
extended addressing hardware, or because there
are insufficient resources to create a section.
CFS 31 31 Chained fixups have been suppressed
The specified PSECT grew beyond the address
specified in the /LIMIT switch. The program is
probably incorrect. Use the /MAP or /COUNTER
switch to check for accidental PSECT overlaps.
Refer to Section 3.2.2 for more information
about the /LIMIT switch.
CLD 31 28 Cannot load DDT
{JSYS error text}
DDT could not be loaded into memory with your
program. The second line of the error message
is the last TOPS-20 process error and indicates
why the error occurred.> ;[2304]
CLF 1 1 Closing log file, continuing on file [file]
You have changed the log file specification.
The old log file is closed; further log entries
are written in the new log file.
CMC 31 31 Cannot mix COBOL-68 and COBOL-74 compiled code
{Detected in module [symbol] from file [file]}
You cannot use COBOL-68 and COBOL-74 files in
the same load. Compile all COBOL programs with
the same compiler and reload.
CMF 31 31 COBOL module must be loaded first
{Detected in module [symbol] from file [file]}
You are loading a mixture of COBOL-compiled
B-7
FIELD TEST
files and other files. Load one of the
COBOL-compiled files first.
CMP 31 28 Common [symbol] declared in multiple psects
{Detected in module [symbol] from file [file]}
You have loaded a module which specifies that
the named common block must be loaded in a PSECT
which is not compatible with the psect in which
it was originally loaded. Compile the module
with the common in the same PSECT as the
original.
CMX 8 8 Cannot mix GFloating FORTRAN compiled code with
FORTRAN compiled code
You cannot load modules produced by FORTRAN with
modules produced by GFLoating FORTRAN. Compile
all FORTRAN modules the same way, then reload.
CNW 31 31 Code not yet written at [label]
You attempted to use an unimplemented feature.
This is an internal LINK error. This message is
not expected to occur. If it does, please
notify your Software Specialist or send a
Software Performance Report (SPR) to DIGITAL.
COE 8 8 Both CONCATENATE and OVERLAY attributes
specified for psect [ name ]
One of the modules you have already loaded
explicitly sets an attribute for the named PSECT
which conflicts with the declaration of PSECT
attributes in the current module. Check the
compiler switches or assembly language
directives that were used in the generation of
these modules.
COF + + Cannot open file [file]
{JSYS error text}
LINK cannot open the specified file for input
CPU 31 31 Module incompatible with specified CPU
{Detected in module [symbol] from file [file]}
The module you are attempting to load does not
contain a .DIRECTIVE for any of the CPUs you
specified with the /CPU switch. Recompile the
module with the proper .DIRECTIVE, or use a
different /CPU switch.
B-8
FIELD TEST
CRS 1 1 Creating section [octal]
LINK prints this informational message when a
module is loaded into a new section. The
message is printed only if you have specified
/ERROR:0.
CSF 1 1 Creating saved file
LINK is generating your executable (.EXE) file.
CSP 31 28 Cannot setup UDDT symbol table pointers
{JSYS error text}
An error occurred while writing UDDT's symbol
table pointers. Symbols may not be available
for use now. This is probably because the
symbol table pointers were on a write-protected
memory page.
DEB 31 1 [name] execution
LINK is beginning program execution at the named
debugger.
DLT 31 1 Execution deleted
Although you have asked for program execution,
LINK cannot proceed due to earlier fatal
compiler or LINK errors. Your program is left
in memory or in an executable file.
DNA 31 28 DDT not available
{JSYS error text}
SYS:UDDT.EXE could not be found.
DRC 8 8 Decreasing relocation counter [symbol] from
[octal] to [octal]
{Detected in module [symbol] from file [file]}
You are using the /SET switch to reduce the
value of an already defined relocation counter.
Unless you know exactly where each module is
loaded, code may be overwritten.
DSC 31 31 Data store to common [symbol] not in link number
[decimal]
{Detected in module [symbol] from file [file]}
You loaded a FORTRAN-compiled module with DATA
statement assignments to a common area. The
B-9
FIELD TEST
common area is already defined in an ancestor
link. Restructure the load so that the DATA
statements are loaded in the same link as the
common area to which they refer.
DSL 31 * Data store to location [octal] not in link
number [decimal]
{Detected in module [symbol] from file [file]}
You have a data store for an absolute location
outside the specified link. Load the module
into the root link.
NOTE
If the location is less than 140, this
message has level 8 and severity 8.
DUZ 31 31 Decreasing undefined symbol count below zero
LINK's undefined symbol count has become
negative. This message is not expected to
occur. If it does, please notify your Software
Specialist or send a Software Performance Report
(SPR) to DIGITAL.
EAS 31 31 Error creating area AS overflow file [file]
{JSYS error text}
LINK could not make the ALGOL symbol table on
disk. You could be over your disk quota, or the
disk could be full or have errors.
ECE 31 31 Error creating EXE file [file]
{JSYS error text}
LINK could not write the saved file on disk.
You could be over your disk quota, or the disk
could be full or have errors.
EIF 31 31 Error for input file [file]
{JSYS error text}
A read error has occurred on the input file.
Use of the file is terminated and the file is
released.
ELF 1 1 End of log file
LINK has finished writing your log file. The
file is closed.
B-10
FIELD TEST
ELN 1 1 End of link number [decimal] {name [name]}
The link is loaded.
ELS 31 31 Error creating area LS overflow file [file]
{JSYS error text}
LINK could not write your local symbol table on
the disk. You could be over your disk quota, or
the disk could be full or have errors.
EMS 1 1 End of MAP segment
The map file is completed and closed.
EOE 31 31 EXE file output error [file]
{JSYS error text}
LINK could not write the saved file on the disk.
EOI 31 31 Error on input [file]
An error has been detected while reading the
named file.
EOO 31 31 Error on output [file]
An error has been detected while writing the
named file.
EOV 31 31 Error creating overlay file [reason] [file]
LINK could not write the overlay file on the
disk.
ETP 31 31 Error creating area TP overflow file {[reason]}
[file]
{JSYS error text}
LINK could not make the typechecking area on the
disk. You could be over your disk quota, or the
disk could be full or have errors.
EXS 1 1 EXIT segment
LINK is in the last stages of loading your
program (for example, creating .EXE and symbol
files, preparing for execution if requested).
FCF 1 1 Final code fixups
B-11
FIELD TEST
LINK is reading one or both segment overflow
files backwards to perform any needed code
fixups. This may cause considerable disk
overhead, but occurs only if your program is too
big for memory.
FEE * * ENTER error (octal) [reason] [file]
One of the following conditions occurred:
1. The specified file name was illegal.
2. When updating a file, the specified file
name did not match the file being updated.
3. The RENAME monitor call following a LOOKUP
monitor call failed.
FIN 1 1 LINK finished
LINK is finished. Control is passed to the
monitor, or to the loaded program for execution.
FLE * * LOOKUP error (octal) [reason] [file]
One of the following conditions occurred:
1. The specified file name was illegal.
2. When updating a file, the specified file
name did not match the file being updated.
3. The RENAME monitor call following a LOOKUP
monitor call failed.
FRE * * RENAME error [reason] [file]
One of the following conditions occurred:
1. The specified file name was illegal.
2. When updating a file, the specified file
name did not match the file being updated.
3. The RENAME monitor call following a LOOKUP
monitor call failed.
FSN 31 31 FUNCT. subroutine not loaded
B-12
FIELD TEST
During final processing of your root link, LINK
found that the FUNCT. subroutine was not
loaded. This would cause an infinite recursion
if your program were executed. The FUNCT.
subroutine is requested by the overlay handler,
and is usually loaded from a default system
library. Either you prevented searching of
system libraries, or you did not load a main
program from an overlay-supporting compiler into
the root link.
FTH 15 15 Fullword value [symbol] truncated to halfword
This message is printed when a symbol that has a
value greater than 777777 is used to resolve a
halfword reference. This warning message helps
you to be sure that global addresses are used
properly throughout the modules in a load.
HCL 31 31 High segment code not allowed in an overlay link
{Detected in module [symbol] from file [file]}
You have attempted to load high segment code
into an overlay link other than the root. Any
high segment code in an overlaid program must be
in the root.
HSL 31 31 Attempt to set high segment origin too low
{Detected in module [symbol] from file [file]}
You have set the high-segment counter to a page
containing low-segment code. Reload, using the
/SET:.HIGH.:n switch, or (for MACRO programs)
reassemble after changing your TWOSEG pseudo-op.
HTL 31 31 Symbol hash table too large
Your symbol hash table is larger than the
maximum LINK can generate (about 50P). This
table size is an assembly parameter. This
message is not expected to occur. If it does,
please notify your Software Specialist or send a
Software Performance Report (SPR) to DIGITAL.
IAS 31 31 Error reading area AS from file [file]
{JSYS error text}
An error occurred while reading in the ALGOL
symbol table.
ICB 8 8 Invalid chain REL block (type 12) link number
[octal]
B-13
FIELD TEST
{Detected in module [symbol] from file [file]}
REL block type 12 (Chain), generated by the
MACRO pseudo-op .LINK and .LNKEND, must contain
a number from 1 to 100 (octal) in its first
word. The link word is ignored.
IDM 31 31 Illegal data mode for device [device]
You specified an illegal combination of device
and data mode (for example, terminal and dump
mode). Specify a legal device.
IHC 31 31 Error reading area HC
{JSYS error text}
An error occurred while reading in your
high-segment code.
ILC 31 31 Error reading area LC
{JSYS error text}
An error occurred while reading in your
low-segment code.
ILS 31 31 Error reading area LS from file [file]
{JSYS error text}
An error occurred while reading in your local
symbol table.
IMA 8 8 Incremental maps not yet available
The INCREMENTAL keyword for the /MAP switch is
not implemented. The switch is ignored.
IMI 31 31 Insufficient memory to initialize LINK
LINK needs more memory than is available.
IMM * 1 [Decimal] included modules missing {from file
[file]}
You have requested with the /INCLUDE switch that
the named modules (if any) be loaded. Specify
files containing these modules.
INS 31 31 I/O data block not set
LINK attempted a monitor call (for example,
LOOKUP, ENTER) for a channel that is not set up.
B-14
FIELD TEST
This is an internal LINK error. This message is
not expected to occur. If it does, please
notify your Software Specialist or send a
Software Performance Report (SPR) to DIGITAL.
IOV 31 31 Input error for overlay file [file]
An error occurred when reading the overlay file.
IPO 31 31 Invalid Polish operator [octal]
{Detected in module [symbol] from file [file]}
You are attempting to load a file containing an
invalid REL Block Type 11 (Polish). This
message is not expected to occur. If it does,
please notify your Software Specialist or send a
Software Performance Report (SPR) to DIGITAL.
IPX 31 31 Invalid PSECT index {for PSECT [symbol]}
{Detected in module [symbol] from file [file]}
A REL block contains a reference to a
nonexistent PSECT. This error is probably
caused by a fault in the language translator
used for the program. This error is not
expected to occur. If it does, please notify
your Software Specialist or send a Software
Performance Report (SPR) to DIGITAL.
IRB 31 31 Illegal REL block type [octal]
{Detected in module [symbol] from file [file]}
The file is not in the proper binary format. It
may have been generated by a translator that
LINK does not recognize, or it may be an ASCII
or .EXE file.
IRC 31 31 Illegal relocation counter
{Detected in module [symbol] from file [file]}
One of the new style 1000+ block types has an
illegal relocation counter. This message is not
expected to occur. If it does, please notify
your Software Specialist or send a Software
Performance Report (SPR) to DIGITAL.
IRR 8 8 Illegal request/require block
{Detected in module [symbol] from file [file]}
One of the REL block types 1042 or 1043 is in
the wrong format. This message is not expected
to occur. If it does, please notify your
B-15
FIELD TEST
Software Specialist or send a Software
Performance Report (SPR) to DIGITAL.
ISM 31 31 Incomplete symbol in store operator in Polish
block (type 11 or 1072)
{Detected in module [symbol] from file [file]}
The specified module contains an incorrectly
formatted Polish Fixup Block (Type 11). The
store operator specifies a symbol fixup, but the
block ends before the symbol is fully specified.
This error is probably caused by a fault in the
language translator used for the program. This
error is not expected to occur. If it does,
please notify your Software Specialist or send a
Software Performance Report (SPR) to DIGITAL.
ISN 31 31 Illegal symbol name [symbol]
{Detected in module [symbol] from file [file]}
The LINK symbol table routine was called with
the blank symbol. This error can be caused by a
fault in the language translator used for the
program. This message is not expected to occur.
If it does, please notify your Software
Specialist or send a Software Performance Report
(SPR) to DIGITAL.
ISP 31 31 Incorrect symbol pointer
There is an error in the global symbol table.
This is an internal LINK error. This message is
not expected to occur. If it does, please
notify your Software Specialist or send a
Software Performance Report (SPR) to DIGITAL.
ISS 8 8 Insufficient space for symbol table after PSECT
[symbol] -- table truncated
There is insufficient address space for the
symbol table between the named PSECT and the
next higher one or the end of the address space.
Restructure your PSECT layout to allow
sufficient room for the symbol table, or use
/UPTO to allow more room.
IST 31 31 Inconsistency in switch table
LINK has found errors in the switch table passed
from the SCAN module. This is an internal
error. This message is not expected to occur.
If it does, please notify your Software
B-16
FIELD TEST
Specialist or send a Software Performance Report
(SPR) to DIGITAL.
ITB 31 31 Invalid text in ASCII block from file [file]
LINK has failed to complete the processing of an
ASCII text REL block from the named file. This
is an internal error. This message is not
expected to occur. If it does, please notify
your Software Specialist or send a Software
Performance Report (SPR) to DIGITAL.
ITP 31 31 Error reading area TP {Status [octal]} from file
[file]
{JSYS error text}
An error occurred while reading in the
typechecking area.
IUU 31 Illegal user UUO at PC [octal]
LINK's user UUO (LUUO) handler has detected an
illegal UUO. This is an internal error. This
message is not expected to occur. If it does,
please notify your Software Specialist or send a
Software Performance Report (SPR) to DIGITAL.
IVC 31 31 Index validation check failed at address [octal]
The range checking of LINK's internal tables and
arrays failed. The address given is the point
in a LINK segment at which failure occurred.
This is an internal error. This message is not
expected to occur. If it does, please notify
your Software Specialist or send a Software
Performance Report (SPR) to DIGITAL.
JPB 8 8 Junk at end of Polish block
{Detected in module [symbol] from file [file]}
The specified module contains an incorrectly
formatted Polish Fixup Block (Type 11). Either
the last unused halfword (if it exists) is
nonzero, or there are extra halfwords following
all valid data.
LDS 1 1 LOAD segment
The LINK module LNKLOD is beginning its
processing.
LFB 1 1 LINK log file begun on [date]
B-17
FIELD TEST
LINK is creating your log file as a result of
defining the logical name LOG:.
LFC 1 1 Log file continuation
LINK is continuing your log file as a result of
the /LOG switch.
LFI 1 1 Log file initialization
LINK is beginning your log file as a result of
the /LOG switch.
LII 8 1 Library index inconsistent, continuing
A REL Block Type 14 (Index) for a MAKLIB
generated library file is inconsistent. The
library is searched, but the index is ignored.
LIN 1 1 LINK initialization
LINK is beginning its processing by initializing
its internal tables and variables.
LMM + + Length mismatch for argument [decimal] in call
to routine [symbol] called from module [symbol]
at location [octal]
The length of the argument passed by the caller
does not match what the called routine expects
it to be.
[2]
LMN 6 1 Loading module [symbol] from file [file]
LINK is loading the named module.
LNA 8 8 Link name [name] already assigned to link number
[decimal]
You used this name for another link. Specify a
different name for this link.
LNL 8 8 Link number [decimal] not loaded
---------------
[2] + The level and severity of this message is determined by a
compiler-generated coercion block. A coercion block specifies
which data type associations are permissible and what action
LINK should take if an illegal type association is attempted.
See Block Type 1130 in Appendix A for more information.
B-18
FIELD TEST
The link with this number has not yet been
loaded. The /NODE switch is ignored. If you
have used link numbers instead of link names
with the /NODE switch, you may have confused the
link numbers. To avoid this, use link names.
LNM 31 31 Link number [decimal] not in memory
LINK cannot find the named link in memory. This
is an internal error. This message is not
expected to occur. If it does, please notify
your Software Specialist or send a Software
Performance Report (SPR) to DIGITAL.
LNN 8 8 Link name [name] not assigned
The name you gave with the /NODE switch is not
the name of any loaded link. The switch is
ignored.
LNS 31 8 Low segment data base not same size
The length of LINK's low segment differs from
the length stored in the current LINK high
segment. This occurs if some but not all of
LINK's .EXE files have been updated after
rebuilding LINK from sources. Update all of
LINK's .EXE files.
LSM 8 8 /LINK switch missing while loading link number
[decimal] -- assumed
Your use of the /NODE switch shows that you want
to begin a new overlay link, but the current
link is not yet completely loaded. LINK assumes
a /LINK switch immediately preceding the /NODE
switch, and loads the link (without a link
name).
LSS 31 1 {No} Library search symbols (entry points)
{[symbol] [octal]
The listed symbols and their values (if any) are
those that are library search entry points.
MDS 8 8 Multiply-defined global symbol [symbol]
{Detected in module [symbol] from file [file]}
Defined value = [octal], this value = [octal]
The named module contains a new definition of an
already defined global symbol. The old
definition is used. Make the definitions
B-19
FIELD TEST
consistent and reload.
MEF 31 31 Memory expansion failed
LINK cannot expand memory further. All
permitted overflows to disk have been tried, but
your program is still too large for available
memory. A probable cause is a large global
symbol table, which cannot be overflowed to
disk. It may be necessary to restructure your
program, or use overlays, to alleviate this
problem.
MMF 31 31 Memory manager failure
The internal memory manager in LINK has failed a
consistency check. This error is not expected
to occur. If it does, please notify your
Software Specialist or send a Software
Performance Report (SPR) to DIGITAL.
MOV 1 1 Moving low segment to expand area [area]
LINK is rearranging its low segment to make more
room for the specified area. Area is one of the
following:
AS ALGOL symbol table
BG bound global symbols
DY dynamic free memory
FX fixup area
GS global symbol table
HC your high-segment code
LC your low-segment code
LS local symbol tables
RT relocation tables
MPS 1 1 MAP segment
The LINK module LNKMAP is writing a map file.
MPT 31 31 Mixed PSECT and TWOSEG code in same module
{Detected in module [symbol] from file [file]}
This module contains both PSECT code and TWOSEG
code. LINK cannot load such a module. Change
the source code to use PSECTs .HIGH. and .LOW.
as the high and low segments, and remove the
TWOSEG or HISEG pseudo-ops.
MRN 1 1 Multiple regions not yet implemented
B-20
FIELD TEST
The REGION keyword for the /OVERLAY switch is
not implemented. The argument is ignored.
MSN 8 8 Map sorting not yet implemented
Alphabetical or numerical sorting of the map
file is not implemented. The symbols in the map
file appear in the order they are found in the
REL files.
NAP 31 31 No store address in polish block (type 11 or
1072)
{Detected in module [symbol] from file [file]}
The specified module contains an incorrectly
formatted polish fixup block (type 11 or 1072).
The store operator specifies a memory fixup, but
the block ends before the address is specified.
This error is probably caused by a fault in the
language translator used for the program. This
error is not expected to occur. If it does,
please notify your Software Specialist or send a
Software Performance Report (SPR) to DIGITAL.
NBR 31 31 Attempt to position to node before the root
The argument you gave for the /NODE switch would
indicate a link before the root link. (For
example, from a position after the third link in
a path, you cannot give /NODE:-4.)
NEB 8 8 No end block seen
{Detected in module [symbol] from file [file]}
No REL Block Type 5 (End) was found in the named
module. This will happen if LINK finds two Type
6 blocks (Name) without an intervening end, or
if an end-of-file is found before the end block
is seen. LINK simulates the missing end block.
However, fatal messages usually follow this,
because this condition usually indicates a bad
REL file.
NED 31 24 Non-existent device [device]
You gave a device that does not exist on this
system. Correct your input files and reload.
NFS 31 28 No free section for XDDT
There is no free nonzero section to load
SYS:XDDT.EXE into. XDDT can only be loaded into
B-21
FIELD TEST
a section which is unused by your program. You
must reconfigure your program to be able to use
XDDT.
NHN 31 31 No high segment in nonzero section
You have attempted to load high segment code
into a program that is being loaded into a
nonzero section. Programs with high segments
must be in section zero.
NPS 8 8 Non-existent PSECT [symbol] specified for symbol
table
You have specified the name of a PSECT after
which LINK should append the symbol table, but
no PSECT with that name was loaded. Load the
named PSECT or specify an existing PSECT for the
symbols.
NSA 31 1 No start address
Your program does not have a starting address.
This can happen if you neglect to load a main
program. Program execution, if requested, will
be suppressed unless you specified debugger
execution.
NSM 31 31 /NODE switch missing after /LINK switch
You used the /LINK switch, which indicates that
you want to begin a new overlay link, but you
have not specified a /NODE switch to tell LINK
where to put the new overlay link.
NSO 31 31 No store operator in Polish block (type 11)
{Detected in module [symbol] from file [file]}
The specified module contains an incorrectly
formatted Polish Fixup Block (Type 11). Either
the block does not have a store operator, or
LINK was not able to detect it due to the
block's invalid format. This error is probably
caused by a fault in the language translator
used for the program. This error is not
expected to occur. If it does, please notify
your Software Specialist or send a Software
Performance Report (SPR) to DIGITAL.
NVR + + No value returned by routine [symbol] called
from module [symbol] at location [octal]
B-22
FIELD TEST
The called routine does not return a value,
however the caller expected a returned value.
[3]
OAS 31 31 Error writing area AS from file [file]
{JSYS error text}
An error occurred while writing out the ALGOL
symbol table.
OEL 8 8 Output error on log file, file closed, load
continuing {[file]}
An error has occurred on the output file. The
output file is closed at the end of the last
data successfully output.
OEM 8 8 Output error on map file, file closed, load
continuing [file]
An error has occurred on the output file. The
output file is closed at the end of the last
data successfully output.
OES 8 8 Output error on symbol file, file closed, load
continuing [file]
An error has occurred on the output file. The
output file is closed at the end of the last
data successfully output.
OFD 31 31 OPEN failure for device [device]
An OPEN or INIT monitor call for the specified
device failed. The device may be under another
user's control.>
OFN 31 31 Old FORTRAN (F40) module not available
{Detected in module [symbol] from file [file]}
The standard released version of LINK does not
support F40 code.
OFS 31 31 Overlay file must be created on a file structure
---------------
[3] + The level and severity of this message is determined by a
compiler-generated coercion block. A coercion block specifies
which data type associations are permissible and what action
LINK should take if an illegal type association is attempted.
See Block Type 1130 in Appendix A for more information.
B-23
FIELD TEST
Specify a disk device for the overlay file.
OHC 31 31 Error writing area HC
{JSYS error text}
An error occurred while writing out your
high-segment code.
OHN 31 31 Overlay handler not loaded
Internal symbols in the overlay handler could
not be referenced. If you are using your own
overlay handler, this is a user error; if not,
it is an internal error and is not expected to
occur. If it does, please notify your Software
Specialist or send a Software Performance Report
(SPR) to DIGITAL.
OLC 31 31 Error writing area LC
{JSYS error text}
An error occurred while writing out your
low-segment code.
OLS 31 31 Error writing area LS from file [file]
{JSYS error text}
An error occurred while writing out your local
symbol table.
OMB 31 31 /OVERLAY switch must be first
The /OVERLAY switch must appear before you can
use any of the following switches: /ARSIZE,
/LINK, /NODE, /PLOT, /SPACE. (It is sufficient
that the /OVERLAY switch appear on the same line
as the first of these switches you use.)
ONS 8 1 Overlays not supported in this version of LINK
LINK handles overlays with its LNKOV1 and LNKOV2
modules. Your installation has substituted
dummy versions of these. You should request
that your installation rebuild LINK with the
real LNKOV1 and LNKOV2 modules.
OOV 31 31 Output error for overlay file [file]
An error has occurred while writing the overlay
file.
OS2 1 1 Overlay segment phase 2
B-24
FIELD TEST
LINK's module LNKOV2 is writing your overlay
file.
OSL 8 8 Overlaid program symbols must be in low segment
You have specified /SYMSEG:HIGH or /SYMSEG:PSECT
when loading an overlay structure. Specify
/SYMSEG:LOW or /SYMSEG:DEFAULT.
OTP 31 31 Error writing area TP {Status [octal]} from file
[file]
{JSYS error text}
An error occurred while writing out the
typechecking area.
PAS 1 1 Area AS overflowing to disk
The load is too large to fit into the allowed
memory and the ALGOL symbol table is being moved
to disk.
PBI 8 8 Program break [octal] invalid {Detected in
module [symbol] from file [file]}
The highest address allocated in the named
module is greater than 512P. This is usually
caused by dimensioning large arrays. Modify
your programs or load list to reduce the size of
the load.
PCL 8 8 Program too complex to load, saving as file
[file]
Your program is too complex to load into memory
for one of the following reasons:
1. There are page gaps between psects (except
below the high segment).
2. There are psects above the origin of the
high segment.
3. Your program will not fit in memory along
with LINK's final placement code.
4. Your program's low segment has the read-only
attribute, or the high segment has the read
and write attribute.
LINK has saved your program as an EXE file on
disk and cleared your user memory. You can use
B-25
FIELD TEST
a GET or RUN command to load the EXE file.
PCX 8 1 Program too complex to load and execute, will
run from file [file]
Your program is too complex to load into memory
for one of the following reasons:
1. There are page gaps between psects (except
below the high segment).
2. There are psects above the origin of the
high segment.
3. Your program will not fit in memory along
with LINK's final placement code.
4. Your program's low segment has the read-only
attribute, or the high segment has the read
and write attribute.
LINK will save your program as an EXE file on
disk and automatically run it, but the EXE file
will not be deleted.
PEF 31 8 Premature end of file from file [file]
LINK found an end-of-file inside a REL block
(that is, the word count for the block extended
beyond the end-of-file). This error may be
caused by a fault in the language translator
used for the program.
PEL 31 8 PSECT [symbol] exceeded limit of [octal]
The specified PSECT grew beyond the address
specified in the /LIMIT switch. The program is
probably incorrect. Use the /MAP or /COUNTER
switch to check for accidental PSECT overlaps.
Refer to Section 3.2.2 for more information
about the /LIMIT switch.
PHC 1 1 Area HC overflowing to disk
The load is too large to fit into the allowed
memory and your high-segment code is being moved
to disk.
PLC 1 1 Area LC overflowing to disk
The load is too large to fit into the allowed
memory and your low-segment code is being moved
B-26
FIELD TEST
to disk.
PLS 1 1 Area LS overflowing to disk
The load is too large to fit into the allowed
memory and your local symbol tables are being
moved to disk.
PMA + + Possible modification of argument [decimal] in
call to routine [symbol] called from module
[symbol] at location [octal]
The caller has specified that the argument
should not be modified. The called routine
contains code which may modify this argument.
In some cases this message will occur although
the argument is not actually modified by the
routine.
[4]
PNO 8 8 Program Data Vectors not allowed in overlay
links
Program data vectors cannot be loaded as part of
an overlay program. The load continues, but no
program data vector will be provided.
POT 1 1 Plotting overlay tree
LINK is creating your overlay tree file.
POV 8 8 PSECTs [symbol] and [symbol] overlap from
address [octal] to [octal]
The named PSECTs overlap each other in the
indicated range of addresses. If you do not
expect this message, restructure your PSECT
origins with the /SET switch.
PTL 31 31 Program too long
{Detected in module [symbol] from file [file]}
Your program extends beyond location 777777,
which is the highest location that LINK is
capable of loading. You may be able to make
---------------
[4] + The level and severity of this message is determined by a
compiler-generated coercion block. A coercion block specifies
which data type associations are permissible and what action
LINK should take if an illegal type association is attempted.
See Block Type 1130 in Appendix A for more information.
B-27
FIELD TEST
your program fit by moving PSECT origins,
lowering the high-segment origin, loading into a
single segment, reducing the size of arrays in
your program, or using the overlay facility.
RBS 31 31 REL block type [octal] too short
{Detected in module [symbol] from file [file]}
The REL block is inconsistent. This may be
caused by incorrect output from a translator
(for example, missing argument for an end
block). Recompile the module and reload.
RED 1 1 Reducing low segment to [memory]
LINK is reclaiming memory by deleting its
internal tables.
RER * 1 {No} Request external references (inter-link
entry points)
{[symbol] [octal]}
The listed symbols and their values (if any)
represent subroutine entry points in the current
link.
RGS 1 1 Rehashing global symbol table from [decimal] to
[decimal]
LINK is expanding the global symbol table either
to a prime number larger than your /HASHSIZE
switch requested, or by about 50 percent. You
can speed up future loads of this program by
setting /HASHSIZE this large at the beginning of
the load.
RLC 31 1 Reloc ctr. initial value current value
{[symbol] [octal] [octal]}
The listed symbols and values represent the
current placement of PSECTs in your address
space.
RUM 31 31 Returning unavailable memory
LINK attempted to return memory to the memory
manager, but the specified memory was not
previously allocated. This is an internal
error. This message is not expected to occur.
If it does, please notify your Software
Specialist or send a Software Performance Report
(SPR) to DIGITAL.
B-28
FIELD TEST
RWA 8 8 Both READ-ONLY and WRITEABLE attributes
specified for psect [ name ]
One of the modules you have already loaded
explicitly sets an attribute for the named PSECT
which conflicts with the declaration of PSECT
attributes in the current module. Check the
compiler switches or assembly language
directives that were used in the generation of
these modules.
SFU 8 8 Symbol table fouled up
There are errors in the local symbol table.
Loading continues, but any maps you request will
not contain control section lengths. This is an
internal error. This message is not expected to
occur. If it does, please notify your Software
Specialist or send a Software Performance Report
(SPR) to DIGITAL.
SIF 31 31 Symbol insert failure, non-zero hole found
LINK's hashing algorithms failed; they are
trying to write a new symbol over an old one.
You may be able to load your files in a
different order. This is an internal error.
This message is not expected to occur. If it
does, please notify your Software Specialist or
send a Software Performance Report (SPR) to
DIGITAL.
SMP 8 8 SIMULA main program not loaded
You loaded some SIMULA procedures or classes,
but no main program. Missing start address and
undefined symbols will terminate execution.
SNC 31 31 Symbol [symbol] already defined, but not as
common
{Detected in module [symbol] from file [file]}
You defined a FORTRAN common area with the same
name as a non-common symbol. You must indicate
which definition you want. If you want the
common definition, load the common area first.
SNL 1 1 Scanning new command line
LINK is ready to process the next command line.
SNP 8 8 Subroutine [symbol] in link number [decimal] not
B-29
FIELD TEST
on path for call from link number [decimal]
{name [name]}
The named subroutine is in a different path than
the calling link. Redefine your overlay
structure so that the subroutine is in the
correct path.
SNS 31 31 SITGO not supported
{Detected in module [symbol] from file [file]}
LINK does not support the REL file format
produced by the SITGO compiler. Load your
program by using SITGO.
SOE 31 31 Saved file output error [file]
An error occurred in outputting the EXE file.
SPF 31 8 Splice fork failed {,saving on file [file]}
LINK was unable to replace itself with your
program using the extended splice fork JSYS.
The probable cause is your monitor not having
the extended splice fork code installed. LINK
will save your program if it has not already
been saved, and will get or run it from the
resulting EXE file.
SRB 8 8 Attempt to set relocation counter [symbol] below
initial value of [octal]
{Detected in module [symbol] from file [file]}
You cannot use the /SET switch to set the named
relocation counter below its initial value. The
attempt is ignored.
SRP 31 31 /SET: switch required for PSECT [symbol]
{Detected in module [symbol] from file [file]}
Relocatable PSECTs are not implemented; you must
specify an explicit absolute origin with the
/SET switch for the named PSECT.
SSN 8 8 Symbol table sorting not yet implemented
Alphabetical or numerical sorting of the symbol
table is not implemented. The symbols appear in
the order they are found.
SST 1 1 Sorting symbol table
B-30
FIELD TEST
LINK is rearranging the symbol table, and if
required, is converting the symbols from the new
to old format as indicated on the /SYMSEG,
/SYFILE, or /DEBUG switch.
STC 1 1 Symbol table completed
The symbol table has been sorted and moved
according to the /SYMSEG, /SYFILE, or /DEBUG
switch.
STL 31 31 Symbol too long
A symbol specified in a rel block is longer than
the maximum allowed by LINK.
SUP 1 1 Loading suppressed
During the compilation process, errors occurred
and a nonzero number was entered in JBERR. LINK
does not attempt to load the REL files that
result from this compile because they may be
incorrect. Some compilers do not produce REL
files if errors are encountered.
T13 31 31 LVAR REL block (type 13) not implemented
{Detected in module [symbol] from file [file]}
REL Block Type 13 (LVAR) is obsolete. Use the
MACRO pseudo-op TWOSEG.
TDS 8 8 Too late to delete initial symbols
LINK has already loaded the initial symbol
table. To prevent this loading, place the
/NOINITIAL switch before the first file
specification.
TMA 31 8 Too many ambiguous requests in link [decimal]
{name [name]}, use /ARSIZE:[decimal]
{Detected in module [symbol] from file [file]}
You have more ambiguous subroutine requests
(indicated by LNKARL messages) than will fit in
the table for this link. Continue loading.
Your load will abort at the end with a LNKABT
message; if you have loaded all modules, the
message will give the size of the needed /ARSIZE
switch for a reload.
TML 31 31 Too many links, use /MAXNODE
B-31
FIELD TEST
You have specified more overlay links than were
allowed by the current value for the /MAXNODE
switch. Reload your program with a larger
/MAXNODE value.
TMM + + Type mismatch seen for argument [decimal] in
call to routine [symbol] called from module
[symbol] at location [octal]
The data type of the argument passed by the
caller does not match what the called routine
expects.
[5]
TTF 8 8 Too many titles found
In producing the index for a map file, LINK
found more program names than there are
programs. The symbol table is in error. This
is an internal error. This message is not
expected to occur. If it does, please notify
your Software Specialist or send a Software
Performance Report (SPR) to DIGITAL.
UAR 8 8 Undefined assign for [symbol]
{Detected in module [symbol] from file [file]}
The named symbol was referenced in a REL Block
Type 100 (ASSIGN), but the symbol is undefined.
This is generated with the MACRO pseudo-op
.ASSIGN. The assignment is ignored. You should
load a module that defines the symbol.
UCB 8 8 Unknown COMMON [symbol] referenced
A reference was made to a common block, however
the block does not exist.
UIE 31 31 Unexpected internal error during processing
An error occurred while processing a previous
error. This message is not expected to occur.
If it does, please notify your Software
Specialist or send a Software Performance Report
(SPR) to DIGITAL.
---------------
[5] + The level and severity of this message is determined by a
compiler-generated coercion block. A coercion block specifies
which data type associations are permissible and what action
LINK should take if an illegal type association is attempted.
See Block Type 1130 in Appendix A for more information.
B-32
FIELD TEST
UGS * 1 {No} Undefined global symbols {[symbol] [octal]
The listed symbols and their values (if any)
represent symbols not yet defined by any module.
Each value is the first address in a chain of
references for the associated symbol.
If this message resulted automatically at the
end of loading, this is a user error. In this
case, the load will continue, leaving references
to these symbols unresolved.
UMF 31 31 Unexpected monitor call failure for [JSYS] at PC
[octal]
{JSYS error text}
A monitor call has unexpectedly failed. The
second line of the error message is the last
TOPS-20 process error and indicates why the
error occurred. This may be either a system
problem or a LINK internal error, depending on
the message.
UNS 31 31 Universal file REL block (type 777) not
supported from file [file]
Extraction of symbols from a MACRO universal
file is not implemented.
URC 31 1 Unknown Radix-50 symbol code [octal] [symbol]
{Detected in module [symbol] from file [file]}
In a REL Block Type 2 (Symbols), the first 4
bits of each word pair contain the Radix-50
symbol code. LINK found one or more invalid
codes in the block. This error can be caused by
a fault in the language translator used for the
program.
URV + + Unexpected return value in call to routine
[symbol] called from module [symbol] at location
[octal]
The called routine returns a value which was not
expected by the caller.
[6]
USA 8 8 Undefined start address [symbol]
You gave an undefined global symbol as the start
address. Load a module that defines the symbol.
B-33
FIELD TEST
USB 8 8 Undefined symbol in byte array (type 1004) block
{Detected in module [symbol] from file [file]}
LINK has detected an undefined global symbol in
a type 1004 REL block. This global symbol is
used to relocate a byte pointer and must be
defined before the 1004 block that uses it is
seen. This error is most likely the result of
an error in the language translator used to
generate the REL file.
USC 31 8 Undefined subroutine [symbol] called from link
number [decimal] {name [name]}
The named link contains a call for a subroutine
you have not loaded. If the subroutine is
required for execution, you must reload,
including the required module in the link.
USD 31 31 Undefined symbol [ symbol ] used in loading code
or data blocks
You have loaded a module that is loading code or
data at a symbolic address and the symbol is
currently unknown to LINK. Check the order in
which you are loading modules.
USI 8 16 Undefined symbol [symbol] illegal in switch
[switch]
You have specified an undefined symbol to a
switch that can only take a defined symbol or a
number. Specify the correct switch value.
UUA 8 8 Undefined /UPTO: address [symbol]
You gave the named symbol as an argument to the
/UPTO switch, but the symbol was never defined.
Load a module that defines the symbol, or change
your argument to the /UPTO switch.
VAL 31 1 Symbol [symbol] [octal] [type]
LINK has printed the specified symbol, its value
and its attributes as requested.
---------------
[6] + The level and severity of this message is determined by a
compiler-generated coercion block. A coercion block specifies
which data type associations are permissible and what action
LINK should take if an illegal type association is attempted.
See Block Type 1130 in Appendix A for more information.
B-34
FIELD TEST
WNA + + Wrong number of arguments in call to routine
[symbol] called from module [symbol] at location
[octal]
The number of arguments in the routine call is
not the number of arguments expected by the
called routine.
XCT 31 1 [Name] execution
LINK is beginning execution of your program.
ZSV 8 8 Zero switch value illegal
You omitted required arguments for a switch (for
example, /REQUIRE with no symbols). Respecify
the switch.
B.3 REASON EXPLANATION
The [reason] message segment of the LNKEOV, LNKETP, LNKFEE, LNKFLE,
and LNKFRE messages may contain one of the following codes.
Code Description
0 (LNKFEE and LNKFRE only)
One of the following conditions occurred:
1. The specified file name was illegal.
2. When updating a file, the specified file name
did not match the file to be updated.
3. The RENAME monitor call following a LOOKUP
monitor call failed.
0 (LNKFLE only)
The named file was not found. Specify an existing
file.
1 The named directory does not exist on the named
file structure, or the project-programmer number
given was incorrect.
2 You do not have the sufficient access privileges
to use the named file.
B-35
FIELD TEST
3 Another job is currently modifying the named file.
Try accessing the file later.
4 The named file already exists, or a different file
was specified on the ENTER monitor call following
a LOOKUP monitor call.
6 One of the following conditions occurred:
1. A transmission, device or data error occurred
while attempting to read the directory or the
RIB of the named file.
2. A hardware-detected device or data error was
detected while reading the named directory's
RIB or data block.
3. A software-detected data inconsistency error
was detected while reading the named
directory's or file's RIB.
7 The named file is not saved a file. This message
can never occur and is included only for
completeness of the LOOKUP, ENTER and RENAME error
codes. This message is not expected to occur. If
it does, please notify your Software Specialist or
send a Software Performance Report (SPR) to
DIGITAL.
14 You have exceeded the quota of the named
directory, or the entire capacity of the file
structure, Delete some files, or specify a
directory or structure with sufficient space.
15 The named device is write-locked. Specify a
write-enabled device or ask the operator to
write-enable the named device.
24 A LOOKUP and ENTER monitor call was performed on
generic device DSK: and the search list is empty.
30 A LOOKUP and ENTER monitor call was given to
update a file, but the file cannot be updated for
some reason For example, another user is
superseding it or the file was deleted between the
time of the LOOKUP and ENTER.
42 This message indicates that a LOOKUP, ENTER, or
RENAME error occurred that was larger in number
than the errors LINK knows about. This message is
not expected to occur. If it does, please notify
your Software Specialist or send a Software
B-36
FIELD TEST
Performance Report (SPR) to DIGITAL.
B-37
C-1
APPENDIX C
JOB DATA AREA LOCATIONS SET BY LINK
LINK sets a number of locations between 40 and 140 (octal) in the
user's program. These locations are known as the Job Data Area
(commonly abbreviated to JOBDAT). They are used by many languages and
programs. In addition, two segment programs have a Vestigial Job Data
Area of eight words following the high segment origin.
The /NOJOBDAT switch, described in Section 3.2.2, keeps LINK from
filling in JOBDAT.
C.1 JOB DATA AREA
Address Symbolic Use
41 .JB41 HALT if not specified otherwise.
Executes by LUUOs.
42 .JBERR Right: Number of errors during loading.
74 .JBDDT Left: Highest location occupied by DDT.
Right: Start address of DDT if loaded.
115 .JBHRL Left: High segment length.
Right: Highest address in high segment.
116 .JBSYM Left: Negative length of symbol table.
Right: Address of table.
117 .JBUSY Left: Negative length of undefined symbol
table.
Right: Address of undefined symbol table.
120 .JBSA Left: First free location in low segment.
Right: Start address of program.
C-1
FIELD TEST
Address Symbolic Use
121 .JBFF Right: First free location in low
segment.
124 .JBREN Right: Reenter address of program.
131 .JBOVL Address of header block for the root
link in an overlaid program.
133 .JBCOR Left: Highest location of low segment loaded
with data.
137 .JBVER Version Program version number (in octal) and
number: flags, in the format shown below.
Bits Meaning
0-2 Modifier flag:
Flag Meaning
0 DIGITAL development group
last modified the program.
1 Other DIGITAL employees
last modified the program.
2-4 A customer last modified
the program.
5-7 A customer's user last
modified the program.
3-11 DIGITAL's latest major revision
number, usually incremented by 1
for each release.
12-17 DIGITAL's minor revision number,
which is usually 0, unless the
program has been modified since
the last release.
18-35 The edit number, increased by 1
after each edit to the program.
This value is never reset.
C.1.1 Vestigial Job Data Area
Offset Symbolic Use
0 .JBHSA Copy of .JBSA.
1 .JBH41 Copy of .JB41.
C-2
FIELD TEST
2 .JBHCR Copy of .JBCOR.
3 .JBHRH LH: left half of .JBHRL.
RH: right half of .JBREN.
4 .JBHVR Copy of .JBVER.
5 .JBHNM Program Name.
6 .JBHSM High segment symbol table, if any.
7 .JBHGA High segment origin page in bits 9-17.
C-3
8-1
GLOSSARY
absolute address is a fixed location in user virtual address
space which cannot be relocatable by the
software.
ASCII is the American Standard Code for Information
Interchange. A 7-bit code in which textual
information is recorded. ASCII code can
represent 128 distinct characters. These
characters are the upper and lower case
letters, numbers, common punctuation marks,
and special control characters.
assembler translates assembly language into machine
language.
relocatable contains machine language code corresponding
binary file to your source program.
chained fixup is a linked list of locations that require a
global symbol definition. The left-half,
right-half, or whole words points to the next
location in the list that requires the
definition.
compiler translates high-level language into machine
language.
counted vector is a region of memory whose size is in the
first word of a block.
debugger is a program that helps you locate
programming errors or bugs.
entry name symbol is a symbol in a module that contains an
entry point name for other modules.
executable program is the form of a program that is ready to be
executed by the computer.
FIELD TEST
fixup is a process LINK uses to resolve global
symbol references. LINK stores the reference
until it loads the module that contains the
global symbol's definition. After LINK loads
that module, it places the definition where
the reference is stored.
global symbol is a symbol defined in one module that can be
referenced by other modules.
JOBDAT contains information about a program in
locations 20 decimal to 137 decimal such as
its debugger symbol table pointer, version
numbers, and memory use.
language translator is a compiler or assembler that translates
source code into machine executable format.
library is a file that contains object modules that
may be needed by programs to resolve global
symbol references.
object module contains a source program's relocatable code
in machine-readable format.
PDV stands for Program Data Vector and contains
information about a program such as its
debugger symbol table pointer, version
number, and memory use. A PDV is used
usually in place of JOBDAT for an extended
addressing program.
Polish chained fixups is a process that uses an algorithm to
resolve global symbols.
PSECTs stands for Program SECTions and are
programmer or system defined regions of code
or data that LINK relocates together in
memory.
Radix-50 is a highly compressed code used to record
textual information. The characters
available in Radix-50 are uppercase letters,
0-9, ., %, $, and a space. A space is equal
to 0, and cannot be used in names.
relocatable address is an address within a module that is
specified as an offset from the first
location in that module.
REL blocks are output from a language translator that
make up and contain the information that LINK
FIELD TEST
uses to load a program.
REL file is the file produced by compilers or
assemblers that contains the REL blocks used
by LINK to load a program.
SIXBIT is a 6-bit code in which textual information
is recorded. It is a compressed form of the
ASCII character set, therefore not all of the
characters in ASCII are available in SIXBIT,
notably the nonprinting characters and lower
case letters are omitted.
symbol table contains entries and values for each symbol
defined or used within a program.
sharable save file is the executable program that has been
stored in a file using the LINK /SAVE switch
or the TOPS-20 SAVE command. This file has
an .EXE file extension.
3
INDEX
Abbreviating LINK switches, 3-4 loading, 3-17, 3-94
Abbreviating switches, 3-4 specifying, 3-16
Absolute address, 1-2 Debugging overlaid programs, 5-26
Address Declaring non-writable links,
absolute, 1-2 5-29
relocatable, 1-2 Declaring writable links, 5-34
virtual, 1-2 Default
Allocating changing file specifications,
FORTRAN COMMON storage, 3-10 3-19
Allocating memory /DEFAULT switch, 3-19
for overlays, 3-87 Defaults
/ARSIZE switch, 3-9 file specifications, 3-3
specifying, 3-59
Blocks /DEFINE switch, 3-20
REL, 1-2, A-1 Defining
Building PDVs, 7-1 global symbols, 3-20
logical names, 3-4
Calls to overlay handler, 5-27 Deleting
CCL file, 3-3 entry name symbols, 3-47
Changing PDVs contents, 3-71 Deleting overlay links, 3-52
Clearing modules, 3-48 Displayiing
Closing overlay links, 3-35 needed modules, 3-43
CLROV., 5-29 Displaying
CLROVL, 5-29 entry name symbols, 3-21
CMD file, 3-3 external global symbol
Command switches references, 3-75
TOPS-20, 2-3 global symbol values, 3-99
Commenting LINK commands, 3-2 relocation counters, 3-13
/COMMON switch, 3-10 undefined global symbols, 3-96
CONCATENATE Displaying messages, 3-41, 3-100
PSECTs attributes, 6-4
Conserving memory space, 3-51 Ending loading, 3-26
Constructing overlays, 3-62 Entry name
/CONTENTS switch, 3-11 symbols, 1-3
Continuing LINK commands, 3-2 Entry name symbols
Core image file, 4-1 deleting, 3-47
/COUNTERS switch, 3-13 displaying, 3-21
/CPU switch, 3-15 Entry points overlay handler,
CPU type 5-27
specifying, 3-15 /ENTRY switch, 3-21
Creating /ERRORLEVEL switch, 3-22
EXE files, 3-80 /EXCLUDE switch, 3-23
sharable save files, 3-80 EXE file, 1-4
Creating PDVs, 3-50 Executable program, 1-4, 4-1
EXECUTE
Data word, A-2 TOPS-20 command, 2-1
/DDEBUG switch, 3-16 /EXECUTE switch, 3-1, 3-24
DEBUG Executing a loaded program, 3-24
Index-1
TOPS-20 command, 2-1 /EXIT switch, 3-27
/DEBUG switch, 1-4, 3-1, 3-17 Exiting LINK, 3-1, 3-26
Debuggers EXTTAB table, 5-53
Index-2
File when LINK creates, 7-1
core image, 4-1
EXE, 1-4 Levels
indirect command, 3-3 message, 4-7
log, 1-5, 4-6 Libraries
map, 1-5, 4-7 searching, 3-81
overlay, 1-5 Library, 1-3
plotter, 1-5 searching, 1-3
REL, 1-2 system, 1-3
sharable save, 1-4 user, 1-4
symbol, 1-5, 3-91, 4-7 /LIMIT LINK switch, 6-2
File specification defaults /LIMIT switch, 3-32
changing, 3-19 LINK, 1-1
Forcing modules loading, 3-30 input, 1-2
Forcing system library searching, messages, 1-5
3-93 output, 1-4
Format LINK commands
link overlay, 5-50 commenting, 3-2
overlay file, 5-47 continuing, 3-2
PDV, 7-3 file specification defaults,
FORTRAN COMMON 3-3
allocating, 3-10 format, 3-2
/FRECOR switch, 3-25 in indirect command files, 3-3
Free memory LINK messages, 4-7, B-1
maintaining, 3-25 description, B-1
FUNCT. subroutine, 5-23, 5-40 level, B-1
severity, B-2
GETOV., 5-30 Link name table format, 5-49
GETOVL, 5-30 LINK number table format, 5-49
Global LINK switches, 3-7 Link overlay code, 5-52
Global symbol Link overlay format, 5-50
displaying values, 3-99 /LINK switch, 3-35
Global symbols, 1-3 LINK switches, 3-4
defining, 3-20 abbreviating, 3-4
displaying undefined, 3-96 arguments, 3-5
suppressing, 3-89 format for use with TOPS-20
/GO switch, 3-1, 3-26 commands, 2-3
placement, 3-6
/HASHSIZE switch, 3-28 LOAD
Header word, A-1 TOPS-20 command, 2-1
/HELP switch, 3-29 Loading
debuggers, 3-17, 3-94
IDXBFR, 5-27 ending, 3-26
INBFR, 5-27 FORTRAN into PSECTs COMMONS,
/INCLUDE switch, 3-30 3-68
Including local symbols, 3-36 object-time systems, 3-60
Indirect command file, 3-3 PSECTs, 6-1
INIOV., 5-31 segments, 3-58
INIOVL, 5-31 two-segment code using PSECTs,
INTTAB table, 5-54 3-74
Local LINK switches, 3-6
Job data area, C-1 Local symbols, 1-3
Index-3
Job names including, 3-36
specifying, 3-78 /LOCALS switch, 3-36
JOBDAT, 1-4, C-1 Log file, 1-5, 4-6
Index-4
specifying, 3-37 /NOREQUEST switch, 3-52
Log files /NOSEARCH switch, 1-4, 3-53
overlay, 5-32 /NOSTART switch, 3-54
/LOG switch, 3-37 /NOSYMBOL switch, 3-55
Logical names /NOSYSLIB switch, 1-4, 3-56
defining, 3-4 /NOUSERLIB switch, 1-4, 3-57
/LOGLEVEL switch, 3-38 Number of overlay links
LOGOV., 5-32 specifying, 3-40
LOGOVL, 5-32
Long count, A-1 Object modules, 1-2
Object-time systems
Maintaining free memory, 3-25 loading, 3-60
Map file, 1-5, 4-7 /ONLY switch, 1-2, 3-58
resetting symbol types, 3-11 Opening overlay links, 3-45
specifying symbol types, 3-11 /OPTION switch, 3-59
/MAP switch, 3-39 Origin
/MAXNODE switch, 3-40 PSECTs, 6-1
MBZ, A-2 /OTSEGMENT switch, 3-60
Memory Output file specification
allocating specifying, 3-6
for overlays, 3-87 OVERLAID
Memory map PSECTs attributes, 6-4
format of PDV, 7-6 Overlaid programs
length, 7-6 debugging, 5-26
PDV, 7-5 Overlay
Memory space file, 1-5
conserving, 3-51 Overlay file format, 5-47
Message levels, 4-7 Overlay handler, 5-26
Message severity, 4-8 calls to, 5-27
Message severity codes, B-2 entry points, 5-27
/MESSAGE switch, 3-41 Overlay handler messages, 5-35
Messages Overlay link names
Displaying, 3-100 specifying, 3-6
LINK, 4-7 Overlay link paths, 5-1
overlay handler, 5-35 Overlay link preamble, 5-51
suppressing, 3-22 Overlay links
suppressing logging, 3-38 closing, 3-35
/MISSING switch, 3-43 deleting, 3-52
Module names naming, 3-35
specifying, 3-5 opening, 3-45
Modules predecessor, 5-1
clearing, 3-48 successor, 5-1
forcing loading, 3-30 Overlay log files, 5-32
preventing loading, 3-23 Overlay program size, 5-26
Overlay structure, 5-1
Naming overlay links, 3-35 /OVERLAY switch, 3-62
/NEWPAGE switch, 3-44 Overlay switches, 5-2
/NODE switch, 3-45 Overlaying links, 5-27
/NOENTRY switch, 3-47 Overlays, 1-5
/NOINCLUDE switch, 3-48 constructing, 3-62
/NOINITIAL switch, 3-49 restrictions, 5-23
/NOJOBDAT switch, 3-50 writable, 5-22
Index-5
/NOLOCAL switch, 3-51 Overlays relocatable, 5-23
Non-writable links
declaring, 5-29 Page access
Index-6
setting, 6-5 origin, 6-1
/PATCHSIZE switch, 3-64 preventing unintended overlaps,
Paths 6-2
overlay link, 5-1 specifying upper bounds, 3-32
PDV, 1-4, 7-1 upper bound, 6-2
building, 7-1 /PVBLOCK switch, 3-69
format, 7-3 /PVDATA switch, 3-71
memory map, 7-5
format, 7-6 /REDIRECT switch, 1-3, 3-74
when LINK creates, 7-1 REL
PDVOP% JSYS, 7-1 blocks, 1-2
PDVs file, 1-2
changing contents, 3-71 REL blocks, A-1
requesting, 3-69 Relocatable
Plot file overlays, 5-23
specifying, 3-67 Relocatable address, 1-2
/PLOT switch, 3-65 Relocatable binary file
Plotter file, 1-5 See REL file.
/PLTTYP switch, 3-67 Relocation counter
Predecessor links, 5-1 definition, 1-3
Predecessor overlay links, 5-1 Relocation counters
Preventing displaying, 3-13
automatic system library for PSECTed programs, 1-3
searching, 3-56 for segmented programs, 1-2
JOBDAT, 3-50 setting, 3-44, 3-85
module loading, 3-23 Relocation table, 5-55
user symbol tables, 3-55 Relocation word, A-1
Preventing JOBDAT, 3-49 REMOV., 5-32
Program Removing links, 5-32
executable, 1-4 REMOVL, 5-32
executing, 3-24 /REQUEST switch, 3-75
single-segment, 1-2 Requesting
specifying termination, 3-86 symbols, 3-76
two-segment, 1-2 Requesting PDVs, 3-69
Program Data Vector /REQUIRE switch, 3-76
See PDV. Resetting symbol types for the
Program Data Vectors map file, 3-11
See PDVs Restrictions overlays, 5-23
Program size RONLY
overlay, 5-26 PSECTs attributes, 6-5
/PSCOMMON switch, 3-68 Root link, 5-1
PSECT bounds, 6-1 /RUN switch, 3-77
PSECT names /RUNAME switch, 3-78
specifying, 3-5 Running LINK, 3-1
PSECT origins Running links, 5-33
setting, 3-85 /RUNOFFSET switch, 3-79
PSECTs, 6-1 RUNOV., 5-33
attributes, 6-3 RUNOVL, 5-33
CONCATENATE, 6-4 RWRITE
OVERLAID, 6-4 PSECTs attributes, 6-5
RONLY, 6-5
RWRITE, 6-5 Save file format, 4-2
Index-7
loading, 6-1 /SAVE switch, 1-4, 3-80
loading two-segment code into, SAVOV., 5-34
3-74 SAVOVL, 5-34
Index-8
/SEARCH switch, 1-4, 3-81 Successor links, 5-1
Searching libraries, 3-81 successor overlay links, 5-1
Searching user libraries, 3-98 /SUPPRESS switch, 3-89
/SEGMENT switch, 3-83 Suppressing
Segments global symbols, 3-89
loading, 3-58 messages display, 3-22
specifying, 3-83 Suppressing logging messages,
/SET LINK switch, 6-1 3-38
/SET switch, 3-85 /SYFILE switch, 3-91
Setting global symbol table size, Symbol file, 1-5, 4-7
3-28 specifying, 3-91
Setting page access, 6-5 Symbol names
Setting PSECT origins, 3-85 specifying, 3-5
Setting relocation counters, 3-44, Symbol table limits
3-85 setting, 3-97
Setting symbol table limits, 3-97 Symbol table vector, 7-7
Severity format, 7-8
message, 4-8 Symbol tables
Severity codes specifying location, 3-92
message, B-2 Symbols
/SEVERITY switch, 3-86 entry name, 1-3
Sharable save file, 1-4 global, 1-3
Short count, A-1 local, 1-3
Single-segment program, 1-2 /SYMSEG switch, 3-92
Size /SYSLIB switch, 1-4, 3-93
setting global symbol table, System libraries
3-28 forcing searching, 3-93
/SPACE switch, 3-87 preventing automatic searching,
Specifying 3-56
CPU type, 3-15 System library, 1-3
debuggers, 3-16
job names, 3-78 /TEST switch, 1-4, 3-94
log file, 3-37 TOPS-20 command switches, 2-3
module names, 3-5 Translating directories, 3-4
number of overlay links, 3-40 Tree diagram
output file specifications, 3-6 outputting, 3-65
overlay link names, 3-6 Tree structure, 5-1
plot file, 3-67 Two-segment program, 1-2
program termination, 3-86
PSECT names, 3-5 /UNDEFINED switch, 3-96
PSECTs upper bounds, 3-32 Upper bound for PSECTs, 6-2
segments, 3-83 /UPTO switch, 3-97
start addresses, 3-88 User libraries, 1-4
symbol names, 3-5 searching, 3-98
symbol types for the map file, stopping automatic searching,
3-11 3-57
version numbers, 3-102 User symbol tables
Specifying values, 3-5 preventing, 3-55
Start addresses /USERLIB switch, 1-4, 3-98
specifying program, 3-88
/START switch program, 3-88 /VALUE switch, 3-99
Stopping automatic user library Values
Index-9
searching, 3-57 specifying, 3-5
Subroutine /VERBOSITY switch, 3-100
FUNCT., 5-23, 5-40 Version numbers
Index-10
specifying, 3-102 Writable links
/VERSION switch, 3-102 declaring, 5-34
Virtual address, 1-2 Writable overlays, 5-22
Index-11