Digital PDFs

XX-60DA3-1C

June 1972

140 pages

Original

5.4MB

view download

OCR Version

5.1MB

view download

Document:	02 1973AsmRef macro
Order Number:	XX-60DA3-1C
Revision:	0
Pages:	140
Original Filename:	02_1973AsmRef_macro.pdf

OCR Text

-197DEC-10~AMZC-D

MACRO-10 ASSEMBLER

PROGRAMMER’S REFERENCE MANUAL

DIGITAL

EQUIPMENT

CORPORATION

MAYNARD, MASSACHUSETTS

MACRO

-198-

The

(:)

The

Edition April 1967
Printing October 1967
Edition (Rev) August 1968
Edition (Rev) June 1969
Edition (Rev) October 1969
Edition (Rev) August 1970
Edition (Rev) April 1972

1967, 1968, 1969, 1970, 1971, 1972 by

Digital

material

purposes

l1st
2nd
3rd
4th
5th
6th
7th

and

this

subject
to

following

Corporation,

are

manual

for

information

change without

trademarks

Maynard,

DEC
FLIP

Equipment Corporation

Digital

Equipment

Massachusetts:

PDP
CHIP

DIGITAL

FOCAL
COMPUTER

notice.

LAB

-199-

MACRO

CONTENTS

CHAPTER

INTRODUCTION

205

1.1

MACRO-10 LANGUAGE - STATEMENTS

206

1.2

INSTRUCTION WORD FORMATS

206

1.2.1

Primary Instruction Format

207

1.2.2

Input/Output

208

1.3

COMMUNICATION WITH MONITORS

209

1.4

OPERATING PROCEDURES

209

1.5

MACRO

209

1.5.1

Symbols

209

1.5.2

Labels

210

Instruction

Format

STATEMENTS

1.5.3

Symbolic Addresses

210

1.5.4

Operators

211

1.5.5

Symbolic Operators

211

1.5.6

Operands

212

1.5.7

Symbolic Operands-

212

1.5.8

Comments

213

1.6

STATEMENT

PROCESSING

213

1.6.1

Order of Statement Evaluation

1.6.2

Order

214
214

1.7

USER-DEFINED SYMBOLS

215

1.7.1

Direct Assignment Statements

215

1.7.2

chél and Global Symbols

216

1.7.3

Deleted Symbols

217

1.8

NUMBERS

218

1.8.1

Arithmetic and Logical Operations

1.8.2

Evaiuating Expressions

219

1.8:3

Numeric Terms

220

1.8.4

Binary

1.8.5

Left Arrow Shifting

1.8.6

Floating Point Decimal Numbers

Expression

Evaluation

Shifting

221

222

1.8.7

Fixed Point Decifial Numbers

222

1.9

ADDRESS

223

1.9.1

Setting and Referencing

ASSIGNMENTS

Counter

the Location

224

1.9.2

Indirect Addressing

224

1.9.3

Indexing

224

1.10

LITERALS

225

Version 47

June

iii

1972

-200-

MACRO

MACRO-10 ASSEMBLER
STATEMENTS - PSEUDO-OPS

CHAPTER 2

227

RELOCATABLE OR ABSOLUTE

2.1

ADDRESS MODE:

2.1.1

Relocation Before Execution - PHASE
and

DEPHASE

227

229

Statements

2.2

NAMING PROGRAMS

230

2.2.1

Program Subtitles

231

2.3

PROGRAM ORIGIN

231

2.3.1

HISEG Statements

- The HISEG Pseudo-Op

232

Statement

2.3.2

TWOSEG Statements

232

2.4

ENTERING DATA

233

2.4.1

RADIX Statements

233

2.4.2

Entering Data Under the Prevailing Radix

234

2.4.3

DEC and OCT Statements

234

2.4.4

Changing the Local Radix for a Single

235

Term

Numeric

2.4.5

50 Statement
RADIX

236

2.4.6

EXP Statement

236

2.4.7

Z Statement

236

2.5

INPUT DATA WORD FORMATTING

236

2.5.1

BYTE Statement

236

2.5.2

POINT Statement - Handling Bytes

237

Formatting I/0

239

2.5.3
2.5.4

IOWD Statement:
Transfer

Entering Two Half-Words

XWD Statement:
of

2.5.5

Words

239

Data

240

Text Input
ASCIZ,

240

and SIXBIT Statement

2.5.5.1

ASCII,

2.5.6

Reserving Storage

241

2.5.6.1

Reserving a Single Location

242

2.5.7

VAR Statements

243

2.5.8

BLOCK Statements

243

2.5.9

END Statements

243

2.5.10

LIT Statements

244

2.5.11

Multi-Program Assembly

244

2.5.12

PASS2 Statements

245

2.5.13

PURGE Statements

245

2.5.14

XPUNGE Statements

245

Linking Subroutines

246

2.5.15

2.5.15.1 EXTERN Statements

246

2.5.15.2 INTERN Statements

247

2.5.15.3 ENTRY Statements

247

June

Version 47
iv

1972

-2012.6

SUPPRESSION OF SYMBOLS

2.6.1

SUPPRESS SYMBOL Statement

248

2.6.2

ASUPPRESS Statement

248

2.6.3

Listing Control Statements

249

2.7

CONDITIONAL ASSEMBLY

252

2.8

ASSEMBLER CONTROL STATEMENTS

253

2.8.1

REPEAT Statements

253

2.8.2

OPDEF

2.8.3

SYN

Statements

254

Statements

255

2.8.4

Extended

2.9

MULTI-FILE ASSEMBLY

257

2.9.1

UNIVERSAL

257

2.9.2

" SEARCH Name

Instruction

256

Statements

Name

258

CHAPTER

MACRO

MACROS
DEFINITION

259

MACROS

259

MACRO CALLS

260

MACRO FORMAT

261

CREATED

SYMBOLS

262

CONCATENATION
DEFAULT

263

ARGUMENTS

264

INDEFINITE REPEAT
NESTING

ASCII

CHAPTER

AND

265

REDEFINITION

266

Interpretation

268

ERROR DETECTION

269

4.1

SINGLE-LETTER

4.2

ERROR MESSAGES

275

4.2.1

LOOKUP- Errors

277

4.2.2

MACRO

I/O

ERROR

Error

CODES

269

Messages

278

279

RELOCATION

ASSEMBLY

283

OUTPUT

6.1

ASSEMBLY

6.2

BINARY

6.2.1

Relocatable Binary Programs

LISTING

PROGRAM

283

OUTPUT

284

- LINK

284

Format

Version

June

1972

MACRO

CHAPTER

APPENDIX

-2026 .2:1.1

LINK

6 .2.2

Absolute

Formats

6 .2.2.1

RIMIOB

6 .2.2.2

RIM10

for

Binary

the

Block

Types

285

Programs

288

Format

288

Format

289

6 .2.2.3

RIM

Format

290

6 .2.2.4

END

Statements

290

EXAMPLES

293

CODES,

PSEUDO-OPS,

307

MONITOR

I/0 COMMANDS

PROGRAMMING

AND
A.l

ASSEMBLER PSEUDO-OPS

AND MONITOR CO

307

COMMANDS

A.2

APPENDIX

MACHINE

MNEMONICS

SUMMARY

OCTAL

CODES

309

311

PSEUDO-OPS

B.1l

PSEUDO-OPS

311

B.1l.1

Conditional Assembly Statements

313

APPENDIX

SUMMARY

CHARACTER

STORAGE

RADIX

INTERPRETATIONS

315

319

ALLOCATION

TEXT

323

CODES

REPRESENTATION

SUMMARY
DEFINING

Version

AND

RULES

325

327

FOR

CALLING

MACROS

G.1

ASSEMBLER

INTERPRETATION

G.2

CHARACTER

HANDLING

G.2.1

Blanks

G.2.2

Brackets

327

G.2.3

Parentheses

328

G.2.4

Commas

328

327
327

327

G.2.5

Semicolons

328

G.2.6

Carriage

328

Return

G.2.7

Back-Slash

328

G.3

CONCATENATION

328

June

1972

MACRO

-203-

APPENDIX H

OPERATING

331

INSTRUCTIONS

H.1

REQUIREMENTS

331

H.2

INITIALIZATION

331

H.3

COMMANDS

332

H.3.1

General Command Format

332

H.3.2

Disk

H.4

SWITCHES

Version 47

File

Command Format

332

334

June

vii

1972

-205-

MACRO

Chapter 1
Introduction

MACRO-10
ates

the

symbolic

a minimum

PDP-10 systems.

for

input

assembled
output

This

the

program

plus

and

output

symbolic

program

files.

source

program,

program,
and

easier,

processes

the

translating

gram

assigning

instructions

program which

DECtape

line

and oper-

all

‘

It is completely
a

terminal

for

printer

functiqns,

faster,
PDP-10

mnemonic

and more

can

output

can

be used

used

the
to

relocatable

and

data,

includes

and

operation

efficient.

preparing
of

source

codes

Basically,

program
the

binary

relating symbols to numeric

absolute

notification

making machine

programmer's

codes needed in machine ihstructions,
values,

PDP-10,

listing.

programming

statements

the

core memory

’For_example,

assembler performs many useful

assembler

for

impure

allowing the user to select standard peripheral

binary. object

the

language
the

input

assembly
pure

MACRO-10 is é two-pass assembler.

device independent,
devices

core

any

addresses

output

errors

for

listing

detected

pro-

the

during

the assembly process.
"MACRO-10

also contains powerful macro capabilities which allow
the

programmer

create

new

language

elements,

VERSION 47

thus

expanding

and

JUNE 1972
1-1

MACRO

-206-

adapting

the

programming

1.1

MACRO-10

riage

LANGUAGE

programs

text

ment

format

editing

perform

STATEMENTS

usually

program,

normally written

free;

feed

that

columns

are

four

separated

the

delimiting,

order

Statements

rigidly

assembler

and

translates
the

oriented

single

functions

terminal,

line

for

with

statements.

and

terminated

statements

statement

controlled

each

are

spacing

state-

car-

elements,

the

MACRO-10

These

statement,

follows

general

statement

elements
and

are
by

precedes

which

statements

are

identified

the

separating,

elements.

form:

operand,operand;comments (carriage

converts

virtually

placed

between

aid

Each

are

not

the

assemblers.

character which

written

MACRO-1l0

elements

appearance

sequence

characters.

operator

The

on a

elements

types

are

label:

specialized

prepared

sequence.

is,

specific

with

punched-card

There

used

are

return-line

numbered
as

assembler

job.

written

the

foregoing

them

into machine

instruction words.

machine

instructions

are

described

return-line

The

the

feed)

form

formats

following

paragraphs.

1.2

INSTRUCTION

There

are

two

WORD

types

and

input/output.

The

PDP-10 machine

FORMATS

machine

instructions

System Reference Manual

this manual.
the

same

fully described

listed alphabetically

primary

the PDP-10

in Appendix A of

(See monitor manuals.)

instruction statements may have

accumulator address and
may be modified by

VERSION U7

are

formats:

Monitor I/O commands, or programmed operators have

formats.

The primary

and

instruction word

(2)

two operands:

a memory address.

(1)

A memory address

indexing and indirect addressing.

JUNE 1972

-207-

MACRO

l.fi.l»_Primaty Instruction Format
Aftef brdceséing primary instruction statements, the assembler

produces machine instructions in the geheral 36-bit word format
shown below:
0

[b 1t 0o+

1 0 0 O |1

1718

l1 IO‘ o 1t ]o 6 00 00O OC OO0 OO 1 00 00 O oJ

.
INSTRUCTION

_/
INDIRECT

PART

ADDRESS

ACCUMULATOR

PART

INDEX

10-0062

Inh general, the mnemofiic,bperation code, or operator, in the sympolic statement is translated'£0 its binary equivalent and placed
in‘bits‘o—é of the chhineiinStfuctiénu The addréss operand is

evaluated and,placed‘in the address pért (bits 18-35 of the machine
instruction. The assemBlex-asSigns séqugntial binary addresses to
each statement as it,is prbcesSed by meéns of the loCation counter.
Labels are given the currénfi-value of the location counter and are

stored in the d ssembler's stbol tabi¢7 where the corresponding
binary addresses can be found if another instruction uses the same
symbol as an address

reference.

Any one of 16 pOSsible acCfimdlators may be specified in an instruction by identifyifig them 9ymbolically dr numerically as operands

in the statement folloWéd by a-comma. The indirect address bit is
set to l.when the character @.bfefixés~a @embry reference. Indexing is-specified by wrifiing thefiindéx reqister used in parentheses
immediately following the memoty reference.

(All PDP-10 accumula-

tors, except accumulator O, may be used as index registers.)

Actu-

ally, expression§ encloéed in parentheses.
(in the index register
position) are eva1uated as 36-bit quantities; their halves are ex-

changed, and then each half is added inté'the COrrespondihg half
of the binaty,Word being assembled. For example, the statements

MOVSI AC,(1.9)

MOVSI AC, (SIXBIT /DSK/)

are

equivalent

MOVSI AC,2¢01L4g@
MOVSI AC,4U46353

VERSION 47

;MOVE 1.4 TO AC)

;MOVE 1.4 TO AC)
‘

June 1972

-208-

MACRO

To illustrate this general view of assembler processing,
typical

symbolic

instruction.

defined

symbols,

with values of

This

[

)

—

178.

TEMP,

then

INDE X

10-0061

ADD,

has been translated to its octal

and stored in bits

0-8.

The first operand specifies

The effective memory address will be

taking

points

cution time by adding the contents of

address

a binary machine

ADDRESS
PART

instruction code,

accumulator

stored as

ACCUMULATOR

270,

is a

respectively.

INDIRECT
BIT

“equivalent,

here

and XR are

; STATEMENT EXAMPLE
)

the assembler and

INSTRUCTION
PART

The mnemonic

and 3,

TEMP

this:

100,

ADD AC17,@TEMP(XR)

is processed by

instruction

17,

that ACl1l7,

LABEL:

Assume

this value

the word

to be

index register

the

address

added

to ACl7.

found

exe-

to the value

the word whose

A comment following a semicolon does not affect the program in any
but

1.2.2

printed

Input/Output

listing.

Instruction Format

There are eight PDP-10
operand

in the ocutput

I/O statements;

either a peripheral device

in each statement the
number or

a device mnemonic

(see PDP-10 System Reference Manual for complete list).
operand is

a memory address.

READ:
requests
at the

The

that data be

for

170
INSTRUCTION

VErs1onN 47

read in

indexed,

example,

from a paper-tape reader,

address

DEVICE
SELECTION

INDIRECT
8T
INSTRUCTION
PART

to be

stored

given.

I/0 instruction words

The second

DATAI PTR,@NUM(4))

indirect,

format

For

first

shown below:

ADDRESS
PART
INDEX
REGISTER

way,

)
10-0063

JuNe 1972

-2091.3

Programs

trol
I/0

assembled

with

a monitor must

services.

tion

codes

and

CLOSE.

This

040

MACRO-10

exercise

memory

allocatlon,

operate

facilities

077)

over
and

such

are

CALL.

processor

services,

trapping,

which

Monitor

listed
'in Appendix A.

1.4

LOOKUP,

to allow the

manuals.

are

executive

for device

INIT,

available

central

other

monitor programmer's

commands

under

are

con-

independent

done by meanof
s programmed operators

through

control

which

use monitor

Additional monitor commands
to

MACRO

COMMUNICATION WITH MONITORS

IN,

(opera-~
OUT,

user program

modify

described

its
the

OPERATING PROCEDURES

Commands fbr'lbading and executing MACRO-10 are
contained in Appen-

dix H.

1.5

MACRO

previously

label,

The

STATEMENTS

stated

(paragraph

1.1)

operator
an
, operand

assembler

interprets

and

macro

statements

and‘optional

processes

these

consist

comments,

statements,

generating

one or more binary

instructions or data woxrds,

assembly

or performaing an

least

process.

statement

must

contain

elements and may contain all four types.

one

ten with only one operand; but others may have
many
statement

the

echoed

(«

or )).

Figures

7-1

follow1ng

used

Examples
and

and

replaces

1.5.1

before

the

control

carriage

(CTRL)

To continue a
left

return-line

of program statements are

arrow

feed

given

(«),

sequence

in Chapter

operators and operands may be represented
elther

symbolically.

them with

The

numeric

assembler
(binary)

interprets

all

symbols

wvalue.

Symbols

The programmer may create
as

the

these

7-3.

Statement labels,
numerlcally

line,

Some statements are writ-

operators

VERsI1ON 47

and

symbols to use as

operands.

symbol

-5

may

statement
consist

labels,
any

June 1972

-210-

MACRO

combination of from one to six characters of the following
set:

The 26 letters,
Ten digits, 0-9

A-Z

Three special characters:

§$ (Dollar Sign)

%
.

(Percent)
(Period)

The foregoing character set is the Radix-50 character set.

Any statement character which is not in the Radix-50 set is treated
as a symbol delimiter when encountered by the assembler.

If the first characters of a symbol are numeric, the symbol is
treated as through the numeric characters were not present. If the
first character is a period, it must not be followed by a digit.
Spaces must not be embedded in symbols. A symbol may actually have
more than six characters, but only the first six are meaningful to
MACRO-10.

MACRO-10 accepts programs written using both upper and lower case

letters and symbols (e.g., programs written using thé Teletype

Model 37).

Lower case letters are treated as upper case in symbols;

additional special characters, and lower case letters elsewhere,
are taken without change.

1.5.2

Labels

A label is the symbolic name created by the source programmer to

identify a statement.

If present, the label is written as the first

item in a statement and is terminated by a colon (:).

(Refer to

paragraph 1.5.1 for a description of how symbolic names are formed.)
1.5.3

Symbolic Addresses

A symbol used as a label to specify a symbolic address must appear

first in the statement and must be immediately followed by a colon
(:). When used in this way, a symbol is said to be defined. A

defined-symbol can reference an instruction or data word at any
point

the

program.

A label can be defined with only one value; if a programmer attempts

! to redefine a label with a different value, the second value is
VERsION 4/

June 1972

1-6

_211-

ignored and

error

Isages). The following are legal labels:
is

indicated

(see

Chapter

for

MACRO

error mes-

$SUM:
ABC:

DEF:

(Both

FOO

The

following

are

too many

characters

given

recognized

Labels

used

1.5.4

are

the

instructions,

are

used

used.

For

assembler
&

label,

example

only

the

label

first

six

ABCDEFGH:

being ABCDEF:.

for programmer

reference

addresses

for

jump

for loops and for debugging.

Operators

operator

may

one

(see

DECsystem-10

pseudo- operation

legal)

(Flrst character must not be a digit.)
(Colon must immediately follow label.)

characters

are

illegal:

TABC:
LAB

labels

assembly

the

mnemonic

machine

System Reference Manual),

pseudo-op

code

codes

which

are

instruction

a command

dlrects assembly

descrlbed

codes

to Monitor,

processing..

this manual,

and

llsted

| with all other assembler defined operators in Appendix A.
Programmers

own

may

pseudo-operators

operator

may

instruction.

but,

the

(see

power

OPDEF

macro

name,

Like pseudo-ops,

the

assembler

creating

their

pseudo-op).

which

macros

calls
a user-defined

direct

macro

assembly processing;

because of their unique power to handle repetitions and
to

extend

and

separately
or

extend

These

adapt
(see

the

assembly

Chapter

3).

language,
Operators

macros
are

are

considered

terminated

with

space

tab.

1.5.5

sSymbolic

Symbols

used

Operators

operators

must

predefined

by the'programmer.

If a statement has no label,

appear

statement,

tab,

first

carriage

the

return.

The

and must be
following

"VERSI1ON 47

are

the

assembler

the operator may

terminated
examples

by
of

space,

legal

operators:

June 1972
*

1-7

-212-

MACRO
MOV
LOC

ZIP

1.5.6

(A mnemonic machine instruction operator.)
(An assembler pseudo-op.)
(Legal only if defined by the user.)

Operands

Operands are usually the symbolic addresses of the data to be accessed when an instruction is executed, or the input data or argu-

ments or a pseudo-oOp Or mMacro instruction.

In each case, the in-

terpretation of operands in a statement depends on the statement
Operands are separated by commas, and terminated by a
operator.
semicolon

(;) or by a carriage return-line feed.

In the mnemonic machine instruction and UUO call set, if an operand is followed by a comma (spaces in the line are ignored) then
the operand is identified as an accumulator (see instruction format
If an operand is not followed by
description in paragraph 1.2.1).

a comma, then it is viewed as an address (either indexed or indirect
if negative).

1.5.7

Symbolic Operands

symbols used as operands must have a value defined by the user.
These may be symbolic references to previously defined labels where

the argument to be used by this instruction are to be found, or

+he values of symbolic operands may be constants or character
If the first operand references an accumulator, it must
strings.
be

followed by a comma.

TOTAL:

ADD AC1,TAG)

The first operand, ACl, specifies an accumulator register, determined

by the value given to the symbol ACl by the user.

The second oper-

and references a memory location, whose name or symbolic address is
TAG. If the user has equated ACl to 17, and the assembler has assigned TAG to the binary address, 000537, then the assembler inserts

17 in the accumulator field (bits 9-12) and 000537 in the address
field (bits 18-35) of this instruction.l If an accumulator is not

specified, but the operator requires one, accumulator 0 is assumed
by default. If an accumulator is specifies by the value >l78, the
four least significant bits are used.

VERsION H7

June 1972

-913-

1.5.8

The

programmer may

Such

MACRO

Comments

comments

tion,

but

are

add

not

useful

or debugging.

notes

statement

affect

assembly

program

the

following

processing

listing

The use of angle'brackets

for

(<>)

semicolon.

program

later

execu-

analysis

should be avoided in

comments because they may affect the,assembly.
Each

line

program may

contain

comment which

explains

purpose of the line and any special action it causes.
also

consist

only .a

comment;‘this

usually

done

the

A line may
at

the

begin-

ning of each routine or major program section to explain.the major
flow

of control,

entry

and

exit

points

and any

other

pertinent

information.

1.6

The

STATEMENT PROCESSING

assembler

routines.
are

The

searched

has

several

symbol

tables

arranged

corresponding

the order

in which they

are:

Macro Table

This

symbol

user—-defined

table

contains macros,

operator definitions (op-defs) 'and
(refer to the description of the SYN

synonyms

pseudo-op,

initially
items.
2.

and

paragraph

empty;
'

2.8.3).

grows

The macro

the

table

user defines

Op-Code Table - This symbol table contains all of
the operator-codes (op-codes), the UUO calls and

the assembler pseudo-operators (pseudo-ops).
Lists
of the foregoing items are given in Appendices A and
B.
and

by
3.

The op-code
is of fixed

table

is generated by the assembler
length; it cannot be changed except

reassembling MACRO.

User

Symbol

Table

user~defined

symbols

Placed

This symbol
other

in the Macro Table.
empty;
it grows
as the user
4.

Mnemonic

for

the

table
Any

than

table

contains

those

which

all

are

This table is initially
defines items. -

Table

- This table contains the mnemonics
MTAPE and TTCALL UUO's.
The mnemonic
searched only if all other measures fail.

CALLI,

symbol

found

this

table

put

into

the

this

table

macro

table as an op-def as though the user had defined it.

Examples-of

a)
b)
c)

~ Verston 47

the mnemonics

contains

are

RESET as defined by the CALLI #,g

EX
as IT
defined in CALLI #,12
OUTSRT_as‘defined in TTCALL 3,8

June 1972

MACRO

-214the macro table and the user symbol table occupy the same

Internally,

space; however,

the entries of each table are easily distinguishable

so no confusion takes place.

1.6.1

Order of Statement Evaluation

The following table shows the order in which the assembler searches
each statement

field:

Label Field
1.

Operator Field

Symbol suffixed by
colon.
If colon is
not found, no label
is present.

1.
2.
3.
4,
5.

Number
Macro/OPDEF
Machine operator
Assembler operator
Symbol

CALLl mnemonic

A single symbol could be used as a label,
depending on where

Operand Field
1.
2.
3.
4,
5.

Number
Symbol
Macro/OPDEF
Machine operator
Assembler operator

an operator, or an operand,

it is used.

The assembler first checks the operator field for a number, and if found,

assumes that no.operator is present.

Likewise, if a symbol is not a

macro, OPDEF, machine operator or assembler operator, the assembler will
If a defined symbol is found, no operator is.
search the symbol table.
present.

If a defined operator appears in an operand field,

least one word of data.

it must generate at

Statements that do not generate data may not

be used as part of operand expressions. If a statemeht'used in an
operand expressions generates more than one word pf'data,-dnly the

first word generated is meaningful.
1.6.2

Order of Expression Evaluation

Expressions are evaluated in the following order:
~

(Unary operator)

*p,

to,

B Shift,

*B, *F,

<« Shift

*tL

Logical operators

.
Multiply/Divide
Add/Subtract

At each level, operations are performed left to right.

VERsSION 47

JUNE 1972

-2151.7

MACRO

USER-DEFINED SYMBOLS

User~defined
are

symbols

generated

(e.g., TAG:).
ferent

value

are

entering

two

typesi

symbol

labels

followed

and

assignments.

immediately

Labels

colon

Symbols used as labels cannot be rede
fined with a dif-

once

they

have

been

defined.

The

value

label

the

value of the location cofinter
at the time that the label is
defined.
Assignments

are

used

represent,

symbolicaliy,

numbers

bit

patterns.

Assignments ease the doding task
in that only one line has to
be changed

(that éontaining the assignment) in orde
r to change é number or bit pat-

tern

which

changed

given

any

to the

1.7.1

used

throughout

time,

symbol

the

current

progrgm:'

value

Assignment

statements

assignment

the

may

last

value

used.

Direct Assignment

Statements

The macro inserts new symbols with
their assigned values directly
into

the

symbol

table

using

direct

assignment

statement

the

form,

symbol=value)

where
must

the

A direct
same

value

BETA=

17)

assignment

value

GAMMA=

The

new

symbol,

value

VERSION 47

number
the

expression.

symbol.

statement

may

For

also

defined

Note

that

the

equal

sign

example,

used

give

symbol

table

with

new

symbol

the

symbol:

17)
BETA)

GAMMA,

assigned

the

previously

ALPHA=

changes

ALPHA=

BETA=

The

may

immediately

value

entered

symbol

into

may

the

value

17.

changed:

assigned

the

first

example

from

June

1972

-216-

MACRO

Direct assignment statements do not generate instructions or data in the
object program. These statements are used to assign values so that symbols
can be conveniently used in other statements.
1.7.2

Local and Global Symbols

User-defined symbols may be used as local and global symbols in addition
to beging used as label and assignment symbols.

Local symbols are symbols which are known only to the program in which
| they are defined. Two separately assembled macro programs may contain
local symbols which have the same mnemonic but different definitions;
these programs, however, may be loaded and executed without conflict
since the symbols are defined as local to each program.

Global symbols are symbols which can be recognized by programs other
than the one in which it is defined. The manner in which a global
symbol is written or defined depends on where it is located: in the program in which it is defined or the program in which it is a reference to
a symbol defined elsewhere.

Global symbols located in the program in which they are defined must be
declared as available to other programs by the use of the pseudo-ops
INTERN or ENTRY (see paragraphs 2.5.14.1 and 2.5.14.3) or by the use of
the delimiter =: in their definition statement. For example, the symbol

FLAG may be declared a global symbols by:

b.
c.

INTERN FLAG (the symbol FLAG is declared internal),

ENTRY FLAG (identifies the entry point of a library subroutine) ,
FLAG=: 2¢¢ (FLAG is given the value 200 and is declared internal).
NOTE

The statement in item ¢ of the foregoing examples
(i.e., FLAG=: 2@@) is equivalent to the series
INTERN

FLAG=

FLAG

208

Global symbols located in a program in which they are references to symbols
defined in other programs must be declared as external symbols by the use
of the EXTERN pseudo-op (see paragraph 2.5.14.1) or a ## suffix. For
example,

the statement

VERSTON 47

EXTERN FLAG

June

1972

-217declares

the

symbol

MOVE

also

EXTERN

Manual).

reference.

an external

such

may

out
a

want

DDT

case,

define

(refer

the

This

available

Available
were

then

the

programmer

for

value

200

statement

symbol

MACRO

DDT Programmer's

should

define

that

but

not

have

Reference
symbol

with

type-in with

for

the

Unavailable

for

type-out

equivalent

and

will

MACRO.

loaded

FLAG=

and

this

2¢p
)

assigned
the
.

Fully

reference;

Symbols

FLAG==

statement

series:

a double equal Sign:

FLAG will

The

FLAG

typed

exterfial

FLAG as

the

programmer

symbol

@ ,FLAG

Deleted

Sometimes

that

symbol

equivalent of

MOVE

1.7.3

@ ,FLAG##

declares the

the

FLAG

MACRO

program

DDT - (assuming
containing

that

symbols

FLAG).

DDT.

to defining FLAG by:

204)

typing

FLAG$K

(the. symbol

$ represents ALT MODE)

with

declared

to DbT .
A

symbol

may

manner

defined

‘

and

internal

the

following

FLAG==:2QE).
VERs1ON 47

June 1972

-218-

MACRO
is

equivalent

INTERN FLAG)
FLAG==20¢J

The programmer may also want to define a label in MACRO but have the outThe following constructions may be
put of the label suppressed in DDT.
used:

LABEL:!
LABEL: :!

1.8

LABEL is a suppressed local symbol.
LABEL is a suppressed internal symbol.

NUMBERS

Numbers used in source program statements may be signed or unsigned, and
are interpreted by the assembler according to the radix specified by the
programmer,

where

2<radix<10

The programmer may use an assembler pseudo-op, RADIX, to set the radix
for the numbers which follow.

If the programmer does not use a RADIX

statement, the assembler assumes a radix of 8 (octal) except in the case

of the POINT pseudo-op

(see paragraph 2.5.2).

The radix may be changed for a single numeric term, by using the gualifier followed by a letter, D (for decimal), O (for octal), B (for binary),
Note that these are not control
or F (for fixed-point decimal fractions).
characters.

Thus,

1219
1848
2314

is stored as
is stored as
is stored as

+D1Q2
+01d
+B1lg

The qualifier +L is used for bit position determination of a numeric
value. +Ln generates an octal value equal to the number of 0 bits to
the left of the leftmost 1, if the numeric value n were stored in a
computer word.

ExXpression

Resultant Value
448

+Lg

VErRs1ON U4/

zZzexro

bits

0000000000. . . .0000000000

JuNe 1972

-219Expression

Resultant Value
418

The

bits

0000000000.

.0000000101

+L-1

1111111111,

.1111111111

suffixes

the

meaning

each

and

G may

number

suffix

add six zeros

¢)

add nine

add three

operators.

defined

added

which

are

numbers
to

shorthand method

follow the

given

number.

of
The

zeros

(e.g.,

5K =

5000),

(e.g.,
5M = 5000000),

zeros

Arithmetic and

and

zeros

is:

Numbers

Zzero

+L5

specifying

1.8.1

MACRO

(e.g.,

Logical

symbols

5000000000).

Operations

may

combined using

arithmetic

and

logical

The following arithmetic and logical operators may be used.
Operator

Add

Meaning

Multiply

—_e N\

Subtract

The

assembler

computes

the

Integer

Divide

AND

Inclusive

36-bit

value

series

numbers

and

‘defined symbols connected by arithmetic and logical operators, truncating from the left, if nécessary. The following examples show how
these arithmetic and logical operators arelwritten in statements.
B=

AC1+7,RHO/31)

MOVE

A+3,BETA-5)

Combinations
operators

1.8.2

65+X11-3)

MULI

are

numbers

called

Evaluating

and

defined

symbols

operations

VERSTON 47

are

arithmetic and

logical

Expressions

When combining elements of an expression,
unary operations

using

expressions.

(leading +

then done

from

-),
left

the assembler first performs

then binary
to right,

1-15

shifts.

The

logical

followed by multiplications

JUNE 1972

MACRO

~220-

and divisions,
tional

part.

to right.
For

from left to right.
Finally,

example,

first

operand
field,
and

subtractions

the

statement:

TRO

3,1+A&C)

operand field is evaluated first;

indicates
the

the

Division always

and

truncates the frac-

are performed,

left

All arithmetic operations are performed modulo 235.

TAG:

the

additions

that

this

logical AND is

sum is

placed

the

an accumulator.

performed

first,

the

the memory address

comma

terminating

the

result

this

second operand
is

added to one,

field of the machine

instruc-

tion.

To change

delimit

must

the

normal

expressions

order

and

operations,

indicate

the

angle

order of

brackets may

computation.

used

Angle brackets

always be used in pairs.

Expressions may be nested to
a pair

angle brackets.

the outermost

any

The

level,

with each expression enclosed in

innermost expression

evaluated last.

The

following are

evaluated

legal

first,

expressions:

A+B/5
<<(C=-D+B-29>%<p-U41>>4+]
|

A=<B=<C=10>>

1.8.3

Numeric Terms

A numeric term may be a digit,

a string of digits,

closed

assembler reduces numeric

in angle brackets.

36-bit value.

or an expression enterms to a

single

This is useful when specifying operations such as local

radix changes and binary

For example,

The

the

shifts,

which require

4D operator changes the

numeric term that

follows

it.

The number

single values.

local radix to decimal for the
23lo may be represented by

+D23

+D<5%¥2+13>
+D<TEN#¥*2+THREE>

but

may not be written,

+D18B-T7

VERSTON 47

June 1972

-221because
and

this

radix,

The

andis
The

the

+D operator

affects

example

second

which

shift

the

following

the

numeric
is

another

are

legal:

term

term which

taken

under

follows

the

it,

prevailing

preceded

MACRO

(77)

normally octal.

operator

followed

only
term

‘

(the

numeric

bit

term

position

(the

number

the

assumed

shifted)

point).

+F167B17
+B14g¢11B8

566B5

<MARK

A bracketed

1.8.4

may

letter

which

the

may

left

shifted

preceded by

right

following

logically
by

shifted

a numeric

left

term,

or
n,

representing

sign.

the

position,

starting with

zero

used,

(decimal)

left

three

left

the

bit

is not
number

+D1l0

number

B35

binary positions

positions,

should be

assumed;

stored
and

that

stored as

its

000000

rightmost

shifting

The
and

taken

000012;

000000

bit

placed.

stored as 240000 000000.
Binary

any

Thus,

shifted

right-hand bit of

right.

decimal.

term may

Shifting

followed

SIGN>B<PT-XXV>

numeric

Binary

A number

with

numeric term,
numbering

from

as modulo

256

but

4D10B32

000120;

the

position

in bit

and

D10B4

and

logical

operation,

rather

than

arithmetic

one.

" The following are legal binary shifts:
1BJ

400000

000000

1B17

000001

000000

1B35

000000

000001

-1B35

777777

-1B53

000000

777777

-1B792

Note

that

the

Version 47

explanation

below)

000000 000001

following

'10B32

(see

4010B32

expressions

10B

are

<42-10>=

equivalent:

10B<

<U42-10>>=

10B<4

DL2-

4D10>

JUuNe 1972

-222-

MACRO

The unary operators preceding a value are interpreted first by the asA leading plus sign has no effect,
sembler before the binary shift.
assembler to shift and then to
the
causes
but a leading minus sign
store the

2's complement.

in Section 1.3.2.
Binary shifting may operate on numeric terms, as defined
1.8.5

Left Arrow Shifting

If two expressions are combined with the operator "<«", i.e., <m>+<n>, the 36bit value of expression m is shifted V bits (where V is the value of expression n) in the direction of the arrow (left) if V is positive or against
the arrow if V is negative.

The effective magnitude of V is that of the

address of an LSH instruction.

1.8.6

Floating-Point Decimal Numbers

If a string of digits contains a decimal point, it is evaluated as a floating point decimal number, and the digits are taken radix 10. For example,
the

statement,

17.4

is stored as 205420 000000.

Floating-point decimal numbers may also be written, as in FORTRAN, with

the number followed by the letter E, followed by a signed exponent repreThe following examples are valid:
senting a power of 10.

NUM1:

NUM2:
NUM3:

1.8.7

17.2E-4)

3.85E2)
=567.825E33)

Fixed-Point Decimal Numbers

As shown in Section 1.8, 4D followed by a numeric term, is used to enter
decimal

integers.

Fixed-point decimal numbers (mixed numbers) are preceded by *F followed by

a number (not a numeric term, defined below) which normally contains a decimal point. The assembler forms these fixed-point numbers in two 36-bit
registers, the integer part in the first and the fractional part in the
second. The value is then stored in one storage word in the object program, the integer part to the left of the assumed binary point, the fractional part to the right.

VERs1ON 47

1-18

JUuNE

1972

-9293The

binary

number

shift

(B)

operator

formed

+F123.45B8

000000
346314

000173
631462

used

two

(the
(the

MACRO

to position

the

assumed point.

The

registers:

integer part)
fraction part,

left-justified)

The B operator sets the assumed poiht:after bit 8,:so fhe integer part is

placed in bits 0—8,.and'the fraction part in bits 9-35 of the storage word.
In this case, the ihteger part is trun¢ated from the left to fit the 9-bit
integer field.
the

assumed

The fraction part is moved into the 27-bit field following

point

173346

and

truncated

the

right.

The

result

is,

314631

(assumed point)

shift

assumed

Fixed-point
the

operator

does

bit

numbers

35,

are

not
and

the

assumed

a fixed-point

fractional

part

positive

000000

000173

AF123.45

000173
346314

346314
631462

+F123.45B17
4F123.45B-1

777777
777604

777604
431463

-+F123.45B17

number,

the

point

lost.

unless

a minus

sign precedes

431463

146316

-4F123.45B-1

fixed-point

-4F123.45

numbers,

if they were positive numbers,

qualifier:

Negative

1.9

appear

such

the

example above,

complemehted,

are

assembled

and then logically shifted.

ADDRESS ASSIGNMENTS

source

memory
This

statements

addresses

location

done
is

are processed,
the

instruction

incrementing
the

assigned.

the assembler assigns
and'data words

location

counter

of
each

statement which generates

storage word increments the location'cdunter by one.
may

generate

six

The

mnemonic

instruction

storage word.
pseudo-ops

storage

However,
not

WOrds,'incréménting.the

and monitor

direct

generate

1The terms monitor command
synonymous.

VERsION 47

command!

assighment

storage words,

the

consecutive
object

time

single

location

statements

program

counter

six.

generate

single

and

affect

not

object

Another statement

Statements
do

program.

a memory

some

assembler

the

loecation

(as used here) and programmed operator are
1-19

June 1972

-224-

MACRO
counter.

Other pseudo-ops and macros may generate many words in the

object program.

1.9.1

Setting and Referencing the Location Counter

The MACRO-10 programmer may set the location counter by using the
_
pseudo-ops, LOC and RELOC, which are described in Chapter 2. He may
reference the location counter directly by using the symbol, point (.).
For example, he can transfer to the second previously assigned storage

word by writing:
.-2)

JRST

1.9.2

Indirect Addressing

The character @ prefixing an operand causes the assembler to set bit 13
in the instruction word, indicating an indirect address.

For an ex-

planation of indirect addressing and effective address calculation,
see the PDP-10 System Reference Manual.
1.9.3

1Indexing

s
T1f indexing is used to increment the address field, the addresof

the index register used is entered in parentheses, as the last part
of the memory reference operand. This is normally a symbolic name
defined by a direct assignment statement, or an octal number in the
range 1-17, specifying 1 of the 15 index registers. However, the
address of the index register may be any legal expression or an expression element.

This is a symbolic, indirect, indexed, memory reference:
A:

ADD

U4,@NUM(17))
NOTE

The parentheses cause the value of the enclosed expression to be interpreted as a 36-bit word with its two
halves interchanged, e.g., (17) is effectively

000017000000_..

The 36-bit value is added to the in-

struction ang may modify it.

This is often used to

generate right half values from left half expressions;
for example, the statement

TLO AC, (1Bd)
which sets the sign bit.

VERs1ON 47

1-20

June 1972

~2251.10

MACRO

LITERALS

In a MACRO statement, a symbolic data référence may be replaced by
a

direct

([1).
sembler

signs

literal

data

an address
in

than

the data

direct representation

stores

address

representation

This

the

found

the

machine

within

called a

brackets

first word of

the

only

consist

of more

one

word

data.

direct

pseudo-ops

Literals may
bracketed

The

(such

assignments

data)

following

are

than

one

literal

statement
must,

(POINT
is a

literal

RADIX)

(i.e.,
18

which

simple

bracketed

example

LDB

4,BP,)

POINT

1@,A+3,14)

used

generate

Literals which consist
do

not

generate

data

data within

other

sets

levels.

pseudo-op which

can be

as-

that

and may generate

however,

the

user

instructions must reference by a byte pointer

BP:

table,

inserts

illegal.

nested
up

The as-

Literal

data and

least one word but no more than 18 wofds.
of

literal.
a

instruction..

may

enclose
in
d sgquare brackets

sets

insert

example

byte

the POINT

literals.

Byte

in this manner:

pointter word.)

stafement directly.

For

LDB 4,[POINT 10,A+3,14])
Literals

are

often

used as

PUSH 17,([#)

MOVE

constants

as,

for example:

(note that @ generates 6ne word of zero).

[3,14]

The following is an example of a multi-line literal:
GETCHR:

SOSG

IBUF+2

PUSHJ

P,[IN

VERSION 47

3ANY

CHARS

3;NO,

READ

3sNO

UNUSUAL

LEFT?

SOME

FOPJ

STATZ
JRST

N,T74gg20
sCHECK FOR ERRORS
[MOVEI E, [SIXBIT /INPUT ERROR
/ ]
JRST ERRPNT] ;PUBLISH ERROR MESSAG
E

JRST
'

ILDB

AC,IBUF+1

POPJ

ENDFIL]

CONDITIONS

;END OF FILE HANDLER

sPICKUP

CHAR

~June 1972

-226-

MACRO

NOTE

The

closing right square bracket does

literal

The

excessive

use

if placed after

the

especially

literals,

not

terminate

the

semicolon.

for

small

subroutines,

not recommended since they use up assembler space at the rate of four
locations
more

per

data word

difficult and may

virtual memory

The

PDP-6

tain

one

Two

version

the

of macro

Use

page

faults

(MACRO-6)

permitted

right bracket

pseudo~-ops

grams.

cause

Literals
in

also make
the

KI-1l0

debugging
processor

allocation.

statement but

Dropping

generated.

MLON and

these

MLOFF

not

only

the

right

permitted

provide

pseudo-ops

permitted

literals

bracket

dropped.

by MACRO-10.

compatibility with old

required

con-

pro-

since

MOVE AC,[SIXBIT/TEXT/)

the

legal

in MACRO-6,

literal

has

even

though

been omitted.

such an unterminated

the

literal.

closing

right

normal mode,

bracket

(])

MACRO does not

allow

The pseudo=-op

MLON

set

at the

start of

consider all code

until

such time

Thus,

carriage-return,

rather

each

assembly

following a

as the

cause

the

assembler

left bracket as part of

assembler processes a matching
line-feed does not end

the programmer must

insert a

right bracket.

literal,

right bracket.

literal,

but

The pseudo-op,

MLOFF

set by the switch /0,
which

The

places MACRO into the compatibility mode in

literals may occupy only a single line.

symbol

literals.

(current

location)

is not changed by the use of

It retains the value it had before the literal was

entered.

VERs1ON 47

June

1972

=227~

MACRO

Chapter 2
MACRO-10 Assembler

Statements—Pseudo-Ops

Assembler
certain

statements

assembler

pseudo-ops

processing

direct

operations,

the

assembler

such

perform

converting

data

binary under a selected radix, or listing selected parts of the assem-~
bled

object program.

operations

are

this

chapter,

these

assembler

processing

fully described.

NOTE

The

pseudo-op

for

constructing

name must
a

follow the

symbol

(refer

rules

Para1.5.1) and must be terminated by a
character other than those listed
in Para-

graph

graph 1.5.1 as valid symbolic characters.
(Normally, a space or tab is used as a
terminator.)

2.1

ADDRESS MODE:

MACRO-10
so

that

normally
the

function

assembles

program

what

RELOCATABLE

can

has

been

OR ABSOLUTE

programs
located

with

relocatable

anywhere

previously

binary

in memory

loaded.

When

for

addresses,

execution

desired,

the

assem-

bler will also assign absolute location addreéses, either for the entire
program or for
RELOC

and

VERSION 47

selected parts.

Two pseudo-ops

control

the

address mode:

LOC,

June 1972

~228-

MACRO
RELOC N&

This statement sets the location counter to n, which may be a number
or an expression, and causes the assembler to assign relocatable adSince most redresses to the instructions and data which follow.
locatable programs start with the location counter set to 0; the
implicit statement,

RELOC @0

begins all programs, and need not be written by the programmer who
wants his program assembled with relocatable addresses,

LOC NJ
This statement sets the location counter to n, a number or an expres-

sion, and causes the assembler to assign absolute addresses, begin ing
If the entire prowith n, to the instructions and data which follow.
must precede
statement
LOC
a
locations,
gram is to be assigned absolute
all instructions and data.
If n is not specified

(LOCJ)
zero is assumed initially.

If only a part of the program is to be assembled in absolute locations,
the LOC statement is inserted at the point where the assembler begins
assigning absolute locations.

For example, the statement,

LOC 2¢¢.0

causes the assembler to begin assigning absolute addresses, and the
next machine instruction or data word is stored at location 2008.

VErs1on 47

June 1972

MACRO

-229-

change the

statement
tinue

address mode

required.

assigning

back

the

relocatable

relocatable,

programmer wants

addresses

explicit

the

sequentially,

RELOC

assembler

con-

he writes,

RELOC J
To

switch back to

the next

sequential

absolute

assignment,

he writes,

LOC
J
Thus,

the

programmer

value

in either

not

a LOC or

required

RELOC

insert

statement,

location

and unless

counter

does,

both

the relocatable coding and the absolute coding will be assigned sequential addresses.

The

This

single quote mark

identify

relocatable

shown

geeaThH"
gel1eep

ga1@34

2F1@34

the

whether each location

skeleton coding.

listings,

ADD 1,X

;RELOC ¢ IS IMPLICIT.

LOC 190p
SUB 5,TOT

;CHANGES TO ABSOLUTE, STARTING
sWITH gg100g.
E

RELOC

3 SETS LOCATION COUNTER TO T4.

ADD 2,XAT
LOC

geagTL!
2eaaT5!

following

Pfogram

gegEge:

operating

the

and in MACRO-10

addresses.

Location Counter

When

is used here,

3 SWITCHES LOCATION COUNTER

EXP A-X+7

3BACK TO ABSOLUTE SEQUENCE.

relocatable mode,

the

assembler

object program is

determines

relocatable

absolute,

using an algorithm described in Chapter
5,

2.1l.1

Relocation

Part

This

feature

Before

program can

iterative

execution.

Execution

be‘moved into

often used

loop,

into

PHASE

other

relocate

fast memory

and

DEPHASE

Statements

locations
for

frequently used

(accumulators

0—178)

execution.

subroutine,
just prior

)

VERsION U7

June 1972

~230-

MACRO
To use this

feature,

the subroutine is assembled at sequential re-

locatable or absolute addresses along with the resfi of the program,
but the first statement before
operator,

PHASE,

the

block into which the subroutine is
All

address

subroutine contains

followed by the address of

assignments

the

the assembler

the first location of the

to be moved prior to execution.

subroutine

are

relation

the

argument of the PHASE statement.

The subroutine is terminated by a

DEPHASE statement,

no operands,

location

which requires

and which restores the

counter.

In the following example, which is the central loop in a matrix inversion,

a block transfer instruction moves the subroutine LOOP into

accumulators

11-16,

MOVE [XWD LOOPX,LOOP]
BLT LOOP+4
|

Relocatable
Address
LOOPX:

JRST LOOP
PHASE 11

L.OOP:

MOVN A
FMP

(X)

MPYR

Absolute

FADM A

Address

SOJGE X, .-3
JRST MAIN

(Y)

DEPHASE

The label LOOP represents accumulator 11,

and the point in the SOJGE

instruction represents accumulator 14.

Note that the code inside the phase to dephase program segment is
loaded into the address following the previous relocatable. code; all

labels inside the segment, however,.have the address corresponding

to the phase address.

Thus the phased code cannot, in general, be

executed until it has been moved to the address for which it was
assembled.

2.2

NAMING PROGRAMS

Normally the first statement in a program gives the name of the program using the TITLE pseudo-op.

This pseudo-op has the form

TITLE NAME
J)
in which the single operand (i.e., NAME) may contain up to 60 characters.

VERs10ON 47

JuNe 1972

_93]-

~ MACRO

'The name given will be printed at the top of each page of the program
listing; The first 6 characters of the given title will appear in the
assembled program as the program name.

assembler inserts the name .MAIN.

If no title is given,

the

The program name given in the TITLE

statement is used when debugging with DDT in order to gain access to
the program's

symbol

table.

Only one TITLE pseudo-op is~permittéd in a program; it can appear anywhere in the program but is normally the first line on the first page.
Remember that a name may be longer than 6 characters, however, only

the first 6 symbol combinations

(within the radix-50 set)

will be used

for the program name,

2.2.1

Program Subtitles

After the first page of a program listing, the first data line en-

countered on a pade may be a subtitle,
the pseudo-op SUBTTL.

Subtitles are génerated using

This pseudo-op has

the form

SUBTTL SUBTITLE)

in which the single operand (SUBTITLE)-may-contain up to 40 characters,
A subtitle is printed as the first data line.on a page and all succeeding pages until the end of the listing or until the subtitle is

changed.

If the current subtitle is changed by another SUBTTL state-

ment which is the firsf data line on a page, the new subtitle appears
on the new page and all subsequent pages.,

If the SUBTTL statement is

not the first statement on a page,
the new subtitle appears on the
next page

and all

subsequent pages.

subtitles can be changed as often as required; they do not generate
data and they do not affect the binary procedure only the listing.

They are used for informational purposes only.
2.3

PROGRAM ORIGIN

Initially all programs start with an implicit RELOC § which sets the
mode to be relocatable and the first address to be @.

wise changed,

VERSION 47

Unless other-

the code generated will be a single-segment program.

JUNE 1972

-232-

MACRO
The

programmer

using

a LOC

High~-segment
one

change

statement to

diagnostics)

spaces;

can

the

generate

to generate

(or

@§

programs)

and

2,3.1

High

HISEG

two-segment

Statements

This pseudo-op

does

(two-segment)

Segment,
op

the

have

two

the

Pseudo-Op

logical

other

for

address

starting

that directs

(Refer

Statement

assembly operations

the high segment

capability.

appear

HISEG

affect the

paragraph 6.2.1.,1,

should

(normally used

Two pseudo-oés, HISEG and TWOSEG con-

The

cept to generate information
current program into

the program by

program operation.

not

code

code.

increasing,

400000 (128K) and increasing.
trol

nature

absolute

high-segment

two-segment

starting

relocatable

the Loader

for additional
of

the

Load

reentrant

Into The

information.)

source

any way ex-

load the

the program has

to Block Type

beginning

High

This pseudo-

program.

NOTE

Whenever

possible the pseudo-op TWOSEG
should be used instead of HISEG.
This
pseudo-op provides functions which are
superior

those

HISEG,

HISEG may be followed by an optional argument which represents

the

program high-segment origin address.

must

This argument, when used,

be equal to or greater than 4Q@@@F and must be a K-bound
of 2000)

value.

(even multiple

The code produced by HISEG will execute at either

relocatable @ or relocatable 4p@@@F depending on the loading instructions

given.

HISEG must not be used if the programmer wishes

to reference data in

the low segment since locations in the low segment are referenced by
absolute addresses only.

2.3.2

TWOSEG

Statements

The TWOSEG pseudo-op generates code that directs MACRO and LOADER to

assemble and load a two-segment program in one file.

outputs a block type 3

(refer to Paragraph

LOADER to expect two segments.

VERSION U7

This pseudo-op

6.2.1.1) which signals the

An optional argument may be present

JuNe 1972

MACRO

-933which
is the first address in the high segment.
present,

400000

If no argument is

assumed.

The high segment code must be preceded by

RELOC Mflfiéfifl
or greater; the low segment code by
RELOC
@

or an argument‘indicating the low segment.,
switches

the

Each RELOC pseudo-op

relocation.

The listing prdducéd by the TWOSEG'pseudo-op shows high segment
addresses

as greater

than

400000

or the

argument of

the pseudo-op,

and low segment addresses as less than 400000 or the argument of the
pseudo-cp.

All relocatable

2.4

ENTERING DATA

2.4.1

RADIX

addresses

are

flagged with a

single quote.

Statements

When the assembler encounters a numerical value in a statement, it converts

the number to a binary representation reflecting the radix

indicated by

the programmer.

The

statement,

RADIX NJ)
where n is

a decimal number,

numerical values that follow,

2 <

n <

unless

the prevailing radix or a local radix

For example;

10,

sets the radix to n for all

another RADIX statement changes
change occurs

(see below).

if the programmer wants the assembler to interpret his

numbers as decimal quantities; then the prevailing radix must be set
to decimal before he uses decimal numbers,
RADIX 18J
The statement,

VERSION 47

RADIX 2,

sets

the prevailing radix to binary.

June 1972

-234-

MACRO

The implicit statement, RADIX 8, begins every program; if the pto—
grammer

2.4.2

wants

Entering

Data

by a

carriage

enter

Data

octal

Under

entered under

the

numbers,

the

this

Prevailing

prevailing

statement

ndt necessary.

Radix

radix

typing

the

data,

followed

return:

765432234567
)
Data may

labeled

also

contain

expressions:

U456+A+BA C+D>)

LAB:
Data may

and

entered

first

using

direct

assignment

statement

to place a symbol with an assigned value in the symbol table,
then using the symbol
gram.

For example,

to insert the

the direct

assigned value

assignment

and

in the object pro-

statements,

A=D 2
B=5J
cause
the

two

entries

sum of

address

The radix
a

the assigned

A+BJ

can also be

single value,

restored,

2.4.3

the

symbol
values

table.
in

The

following

statement enters

the object program at

symbolic

REX.

REX:

DEC

To change

changed

after which

as described

and OCT

REX contains 000000 000007
locally,

that

is,

for

the prevailing radix

in Section

a
is

single

statement

automatically

1.3,

Statements

to a local radix

for a single

statement,

the programmer

writes:

DEC N,N,N,...NJ
where all of the numbers and expressions are to be interpreted as
decimal numbers.

VERs1ON 47

The numbers or expressions

following the operator

June 1972

~235are

separated by commas,

and

each will

MACRO
generate

a word

storage.

OCT N,N,N,...N«
changes the local radix to octal for this statement only, and

generates a word of memory for each humber or expression.
The

statement,

DEC

1#,4.5,3.1416,6.93E-26,3

five

decimal words

generates

2.4.4
To

data.

Changing the Local Radix for a Single Numeric Term

change

term

the

radix

for

prefixed with

expression

4D,

used,

single
*0,

number

expression,

4B,

4F,

explained

it must

enclosed

the
in

angle

numeric

Chapter

brackets,

+D<A+B-C/222>
)

These
The

prefixes may generate

TOTAL2:MOVE

causes

When

the

contents

assembler

representation

radix.
cated

For

part

instruction word.

4D1@,ABZ
ABZ

be moved

encounters
a

36-bit

J)
to accumulator

a numeric term,

under

the

fit

example,

MOVE

required

the accumulator

assembler

the

128.

forms

the binary

prevailing or

If the quantity is a part of an instruction,
to

range 0-17

the

a word,

statement,

local

it is trun-

field.

field must have

final value

the

Ih the statement,

4D6g,ABZ
first

forming

interprets

the

integer

the

accumulator

000000000074:

address

but takes

only

36-bit

the

rightmost four bits and places them in the accumulator field of
the instruction, which results in the selection
of accumulator 148.

VErRsioN U7

JUNE 1972
2-9

MACRO

-236-

2.4.5

RADIX

Another
in

symbolic

PDP-10

0-3.

This

EXP

Several

SQUOZE

This

Loader,

expressions

mnemonic

statement

programming.

assembler,

2.4.6

Statement

radix changing

systems

bits

DDT,

into

used

available,

RADIXS50

and

bits

explained

may

other

and

place

used primarily

n,sym) which

used by

systems

add

in Appendix
in

but

a
F

programs

4-bit
of

code

this

field

manual.

the

pack

(The

RADIX50.)

Statement

numbers

and

EXP

X,4,

expressions

may

entered

using

the

EXP

state-

ment:

which

generates

generated

one word

word

generates

can

entered

expression;

five words

were

‘

forms

and other

using

the

operator,

2 )

full word

assembler

field

each

Statement

LABEL:

the

for

)

for the above example.

2.4.7

zero

+D65,HALF,B+362-4

all

zeros

primary

unknown

at LABEL.

machine

fields

If operands

instruction,

zeroed.

the

with

follow the

the

operator

statement,

Z 3,
2
the

assembler

generates
2.5

accumulator

machine

instruction:

DATA WORD

BYTE

but

000140

address

field,

and

000000.

FORMATTING

useful

Bytes of any length,

store

data

from 1 to 36 bits,

less

than

full

36-bit

may be entered by using

statement.

BYTE
The

field,

Statement

conserve memory,

words.
a

this

INPUT

2.5.1

finds

(N)

first operand
the range

X,X,X J

(n)

1-36,

is the byte

size

and must be enclosed

following are separated by commas,
an operand
from the

left

an expression,
to

the

in bits.

in parentheses.

decimal number
The operands

and are the data to be stored.

it is evaluated anhd,

specified byte

size.

VERSION 47

Bytes

if necessary, truncated
are packed

2-10

into words,

June 1972

starting at bit 0,
tions.

If,

and the words are assigned sequential storage loca-

during the packing of a word,

into the remaining bits,

the unused bits

a byte is

In the following statement,

This

too large to fit

are zeroed and the byte is

stored left-justified in the next sequential

LABEL:

MACRO

-257-

location.

three 12-bit bytes

are entered:

BYTE (12)5,177,N J

assembles at LABEL as,

0005

0177

0316,

where N=31l6.

The byte size may be altered by inserting a new byte size in parentheses
immediately

following any operand.

as delimiters,
inserted.

The following

BYTE

Notice that the parentheses

serve

so commas must not be written when a new byte size is
are

legal:

(6)5(14)NT(3)6,2,5
)

where 6 is entered in a 6-bit byte, NT in the following l4-bit byte,
6 in the following 3-bit byte,

followed by 2 and 6

in 3-bit bytes.

BYTE statement can be used to reserve null fields of any byte size.

A
If

two consecutive delimiters are found, a null field is generated.
BYTE

(18),5
J

When the assembler finds two delimiters, it assembles a null byte.
this caSe, 000000 000005.

To enter ASCII characters in a byte, the

characters are enclosed in quotation marks.
BYTE

(7)"A" )

Text handling pseudo-ops are discussed in paragraph 2.5.5.

2.5.2

POINT Statement —'Handling Bytes

Five machine instructions are available for byte manipulation.
These instructions reference a byte pointer word, which is
generated by the assembler from a POINT statement of the form,

LABEL:POINT s, address, b J

(s and b are decimal)

where the first operand s is a decimal number indicating the byte
size, the second operand is the address of the memory location
which contains the byte, and the third operand, b, is the bit
position in the word of the right-hand bit of the byte (if b
is not specified, the bit position is the nonexistent bit to the

VErRs10N 47

2-11

JuNe 1972

MACRO
left
be

-238-

bit

0).

indirect

and

MACRO-10

the

assumes

into

with

three

specified

the

size

1If

the

byte

LDB

(load-a

and

ILDB

statements.

The

example,

AB,

then

the

same.

does

load

ILDB

are

instruction

byte;

the

effect

BYTE

(615

POINT

6,AA55

NACAB2*

440620

QBOORO'

AC:

POINT

65AA

NovvV3r'

135140

Q0VPOO1*'

START:

LDB

35AB

AAQ0D
4

134140

@OA0O2"

ILDB

35AC

enters

AA which

the
is

quantity

The

goes

case,
in

located

the

byte

looks

other

byte

pointer

bit

position

The

increment

word

The

increment

string

byte

finds

size,

that

a byte

AA,

and

The

byte

then

specified

byte

(the

load

instructions

third

for

address,

byte

the

size

into

and

bit

bits,

bit

right-hand

loaded

instructions

The

deposit

accumulator

pointer

(IBP)

byte

two

accumulator

000005.

ways.

the

pointer

bytes,

and

from

6-bit

statement

its

the

word

000000

the

When the LDB is executed,

the

similar

wbrd) by the byte size.
a

this:

indicator

ing

the

manipulation mnemonic

takes

the

for

this

second

along

"increments"
of

AB:

statement

loca-

used,

V00000

byte

(IDPB)

memory

QOQ00D'

the

position

B52000

position.

struction

from

3609690

the machine

byte

PABNB1'

address

may

specified,

instructions

reference by the load byte instruction.

The

not

operand

DOPVRD

first

where

second

bits.

assembler

look

symbolic

address

accumulator)

instructions

The

following

tion

The

indexed.

and

word,

instruction

operand

the

reference

byte

(DPB)

deposits

the

in-

it,

a memory word.

increments

the

referenced

POINT

bit

This is useful when loading or depositusing

the

byte

(ILDB)

increment

same

the

and
byte

byte

pointer

increment
pointer

word.

and

word

deposit
by

the

byte

size before loading or depositing.

VERSION 47

2-12

JUNE 1972

-2392.5.3

The

IOWD Statement:

assembler

MACRO

Formatting I/0 Transfer Words

generates

I/0

transfer

words

special

format

for use in BLKI and_BLKO and all four pushdow
n instructions.
The

general

statement

IOWD

where

two

N,MJ

operands,

the

IOWD

operator.

The

left

half

the

value

first
of

is,

which

This

of the

sécond

numbers

statement

assembled

operand
n,

the

may

and

operand

expressions,

one

contains

the

word

the

generates

right
minus

half-word
one.

For

data
2's

word.
complement

contains

the

example,

IOWD 6,4D256
<
assembles

2.5.4

The

XWD

777772

Statement:

statement

written

WD
where
left
word.

bits

the

half-word,

are

enters

the

are

placed

second

the

half-words.

the

36-bit

Data

single

storage

word.

START,END)

used

Block

the

left

set

specifies

registers

The

and

specifying

the

high-order

pointer words

transfer pointer words
half

be moved,
and the right
A,,B

expression

Three examples follow:

SUM+2,DES+5)

are

operand

XWD

statements

symbol

formed

A,BJ

destination.
XWD

XWD

addresses:

half-words

and

XWD

instructions.

two

form,

operand

operand are ignored.

XWD

Entering
Two Half-Words

LHW,RHW <

first

Both

000377.

may also

the

starting

half

the

used

for

bits

two
of

18-

each

the

18-bit

the

location of

duplicate

half-

order

block transfer

contain

location

first

right

low

block

the

results

A,B.

VErsion 47

JUNE 1972

-2h0-

MACRO
2.5.5

Text Input

The assembler translates text written in full 7-bit ASCII or 6-bit
compressed ASCII. It will also format 7-bit ASCII with a null
character at the end of text, if desired. These codes are listed
in Appendix E.

are
In all three text modes, the printing symbols in the code set
translated to their binary representation.

To translate and store a single word containing as many as five

are
7-bit ASCII characters, right-justified, the input characters

enclosed in guotation marks.
"AXETM)

is stored as

0 0000000 0000000 1000001 1011000 1000101
0

null

Notice that characters are right-justified, and bit 0, which is
not used,

is set to zero.

Up to six 6-bit ASCII characters may be translated and stored,
right-justified, in a single word by enclosing the input charac
ters in single quotation marks.
'7TABLES'

is stored as

110100 100001 100010 101100 100101 110011
T

NOTE

The quotation marks (single or double) may
To
only be used to assemble a single word.
input strings of text characters, the fol-

lowing three pseudo-ops must be used.

of
2.5.5.1 ASCII, ASCIZ, and SIXBIT Statement - To enter strings
used.
are
text characters, the operators ASCII, SIXBIT, and ASCIZ
The delimiter for the string of text characters is the first nonor
blank character following the character that terminates the operat
(refer to the note on page 2.1). The binary codes are left-justified.

Unusea character positions are set to zero (null). Text is termi~nated by repeating the initial delimiter.

If the initial delimiter

is a symbol constituent, the pseudo-op must be followed by

a space

or a tab.

JunNe 1972

VErRs1oON 47

914

-24]1The

MACRO

statement

ASCII "AXE"
where

the

quotation

marks

1000001

1011000

1000101

0000000

null

The

operator

the

end of

five,

ASCIZ

text.

another

(ASCII

all

are

the

Zero)

delimiters,

guarantees

assembles

null

the number of characters

zero word

added.

For

character

a multiple of

example,

ASCIZ/"AXE"/)

assembles

followed

as,

0100010

1000001

1011000

1000101

0100010

0000000

another

0000000

word

0000000

When

thefull

6-bit

7-bit

subset may

ters

are

the

SIXBIT

ASCII

zeros.

0000000

null

code

entered,

left-justified
statement

set

using

the

sequential

the

same

Convert

lower

case

ASCII

characters.

The

not

required,

the

64-character

SIXBIT operator.
storage words.

for

ASCII

Six

charac-

Format

statements.

derive

code:

2.5.6

Add

¢.

Truncate

408

Reserving

programmer

locations

VERSION 47

the

value

result

of
to

the
the

upper

case

character.
rightmost

six

bits.

Storage

can

for use

the

characters

reserve

during

single

execution

locations,

his

blocks

many

program.

JUne 1972

-242-

MACRO
2.5.6.1

Reserving a Single Location - The number sign

fixing a symbol in an operand field,

location.

The symbol is defined,

(#),

suf-

is used to reserve a single

entered in the assembler's

symbol table,and can be referenced elsewhere in the program withcut

the

number

sign.

i \AB:

ADD

For example,

3,TEMP#

)

reserves a location called TEMP at the end of the program,
may be used to store a value entered at some other point
program.

This

feature is

useful

for storing

scalars,

which

in the

and other

quantities which may change during execution.

The pseudo-op INTEGER may be used to reserve storage locations
at the end of the program on a one-per-given name basis.
example

For

statement

the

INTEGER TEMP,FOO,BAR)

will reserve 3

locations identified as TEMP,

assignment of the

FOO and BAR.

The

locations to the names given is performed on

an alphabetical basis by the assembler rather than on the order
in which the names are given.

For example,

the order of the loca-

tions reserved by the foregoing INTEGER statement would be BAR,
FOO

then

TEMP.

Multiple word locations may be reserved by the ARRAY pseudo-op.
For

example,

the

statement

ARRAY FOO[2*3])

reserves a 2-word by 3-word array in memory which is identified by
the

name

FOO,

NOTE

the

pseudo-op TWOSEG

reserved
signed

pseudo-op
the

low

the

array
low

segment

required

used,

statement
only;

after

the

must

variables
be

thus,

RELOC

asa

VAR

back

segment.

June 1972
o

VErRSION 47
-16

-20432.5.7

VAR

MACRO

Statements

VAR )

This

statement

causes

symbols

which have

been

defined

with the # sign (array and'integer pseudo-ops)
ments

assembled

subsequent

symbol

and

the

LIT

literals

2.5.8

statements.

This

definitions

the

same

statements

not

appear

VAR

and variables

BLOCK

block

are

stored

the

suffixing

in previous statehas

effect

the

end of

program,

the

all

program.

Statements

To reserve a block of locations, the BLOCK.operator is used.
is

followed

pression
be

by a

single

the

current

reserved.

the

value

The

the

operand,
radix,

assembler
operand.

BLOCK

block

N*M

dimensions

are

BLOCK

reserve

data

words

whose

used

should

which may

indicating

increments
For

MATRIX:

reserves

type.

a number

the

number

location

an ex-

words

counter

example,

N¥M
)

words

and

starting

MATRIX

for

array

words

stored

the

specific
low

order;

segment

remember

that

two-segment

pro-

grams.

2.5.9

The

END

Statements

statement must

single

operand

first

the

given

only

the main

may

the

last

the

END

instruction to be

the

pass

operator

executed.

every

specify

Normally

the

main

program;

Since‘subprograms

program,

they

need

specify

END START)

When

statement

assembler

and

-VERSIbN 47 .

begins

start is

first
pass

not

the

encounters
The

END

label

an
also

END

starting

at the

address

this operand is
are

called

from

address.

starting address

statement,

terminates

program.

the

pass

terminates
2,

after

which

JUNE 1972

-244-

MACRO

the assembler automatically assembles all previously defined vari-

ables and literals starting at the current location.!

The following processing operations can be performed at any point
in

the

2.5.10

program.

LIT

Statements

J
LIT

This statement causes literals that have been previously defined to
If n literals have
be assembled, starting at the current location.
been defined, the next free storage location will be at location
counter plus n. Literals defined after this statement are not af-

fected.

If a LIT statement does not appear before the END statement, the
If the output
literals are XLISTed (refer to paragraph 2.6.3 ).
of literals is desired, the LIT pseudo-op should appear immediately
before

the

END

statement.

NOTE

the

also

two-segment
high

literals

2.5.11

program

segment.

given

the

will

LIT

The

high
the

must

END

segment

low

given

statement

must

the

segment.

Multi-Program Assembly

The pseudo-op PRGEND is used to compress many small files into one
This pseudo-op has the
large file to save space and disk lookups.
form PRGEND) .

PRGEND allows multiprogram assemblies, and is used

for assembling library files

short.

(LIB40)

in which all programs are very

PRGEND takes the place of all but the last END statement.

The output is a binary file which can be loaded in search mode.

The

use of PRGEND costs assembler space since the symbol tables, iiteral

tables and titles of each of the small files (i.e., programs) involved
Also, since PRGEND is functionmust be saved at the end of pass 1.
ally an END statement, macros cannot be used over it

(i.e., macros

cannot generate PRGEND as part of their expansions).

l7he END statement is also used to specify a transfer word in some
(See Section 6.2.2.4.)
output file formats.

VErRsION 47

June 1972

2-18

~245If

the

LIT

literals

2.5.12

and

VAR.statements

and variables

PASS2

are

not

MACRO

appear

the

programs,

stored at

the

end of

the

assembler

pass

pProcessing

all

Program.

Statements

PASS2
)
This

statement

switches

remaining coding;
processed

pass

only.

such as

testing macros

2.5.13

PURGE

The

PURGE

form

is:

PURGE

each

call

which

statement
age

and

used

the

delete

(refer

deletes
PURGE

the

used

the

symbols

uses

also

from
the

for

DDT.

same

2.8.2)

the

PURGE

first

SOLV

The

XPUNGE

has

the

for

debugging,

portion.

symbol

the

operator,

Its

general

for
a

symbol

both

label,

For

or macro

tables.

programs

A repeat
label.

symbols.

label,

and for

OPDEF.

purges

defined

Purged

purges

the

macro

call;

|
XPUNGE

for

The

conserve

table

space

PURGE
storis

a macro

call

statement

the

PURGE

symbol

example,

the

and/or

the

following

SOLV,SOLV
W

label.
2.5.14

assembler's

end

statement purges both:

The

primarily

pass

symbol

the

call

delete

user-created

used

Section

macro

statement

deleted

normally

programmer

OPDEF

symbol,

reused by the assembler.
If

This

defined

symbol,

operand
is

Statements

statement

where

the

Coding preceding this stat
ement will have been

the

second

SOLV

purges

the

pass

Statements

pseudo-op

form:

deletes all

local

symbols

during

XPUNGE)
VERSION 47

June 1972

246~

MACRO

1The use of this pseudo-op reduces the size of the REL file an
speeds up loading (especially of DDT). XPUNGE should be placed
just prior to the END statement.

2.5.15

Linking Subroutines

Programs usually consist of subroutines which contain references

to symbols in external programs. Since these subroutines may be_:
assembled separately, the loader must be able to identify "global"
symbols.

For a given subroutine, a global symbol is either a

symbol defined internally and available for referehce by other

subroutines, or a symbol used internally but defined in another
subroutine. Symbols defined within a subroutine, but available to
others, are considered internal. symbols which are externally
defined are considered external.

These linkages between internal and external symbols are set up

, or
declaring global symbols using the operators EXTERN, INTERN
ENTRY.

The double colon (::) may also be used.

2.5.15.1 EXTERN Statements - The EXTERN statement identifies symbols

which are defined elsewhere.

The statement,

EXTERN SQRT, CUBE,TYPE)

declares three symbols to be external. vExternal symbolsrmust not
be defined within the current subroutine. These external references
may be used only as an address or in an\éxpression»that is to bé
used‘as an address.

For example, the square root routine declared

above might be called by the statement,

P,SQRT J

PUSHJ

External symbols may be used in the same manner as any other
locatable symbol.

Examples:

EXTERN

GPDEF

WwixWD B+3sA=-5)

oUA3NGM NONAN3* MOVE
WAGRE3* NANRYB* XWD
A NR ks =
i
77 TT4% TT7TTTT3% U

VERSsION 47

re-

6sA+3
A+35A
A=

JunNe 1972

-247The

external

restrictions

symbols

are

flagged with asterisks.

for

use

the

external

The

use

Globals

ing

are

alternative
pound

than one

not permitted.

external)

There

are three

symbols:

Externals may not be used in LOC and RELOC state-

ments.

double

MACRO

external

A+B

only
illegal

additive;

-A

-EXP-A

2¥A

2% A-A

(##)

eliminates specifying

(where

expression

A and

illegal.

may

method

sign

Thus,

for

generating

following

the

EXTERN

the

are

both

therefo
the followre,

external

symbols

symbol

name.

This

statement.

For

example,

use

method

MOV @,JOBREL##

equivalent

EXTERN JOBREL
MOVE #,JOBREL

2.5.15.2

INTERN Statements - To make internal program symbols avail-

able to other programs

or ENTRY are used.
assembly of

as external

symbols,

the program,

but will make

available to other programs at load time.
INTERN

makes
symbol

the

direct

other

ment,

symbol equivalences

The statement,

MATRIX

defined

available

within

the

other

program as

programs.
label,

An internal

variable,

assignment.

2.5.15.3
it

list of

MATRIX
)

subroutine

must

the operators INTERN

These statements have no effeét on the actual

ENTRY

Statements

convenient to
programs,

place
these

Some

them

library

subroutines

have

1In order

library.

subroutines

must

common usage,
to

contain

be
the

and

called
state-

o
ENTRY NAME

VERSION 47

JUNE 1972

-243-

MACRO

where "name" is the symbolic name of the entry point of the library subroutine.

ENTRY is equivalent to INTERN with the following additional feature.
All names in a list following ENTRY are defined as internal symbols
and are placed in a list at the beginning of the library of subrou1If the loader is in library search mode, a subroutine will
tines.
be loaded if the program to be executed contains an undefined global
symbol which matches a name on the library ENTRY list.

If the MATRIX subroutine mentioned before is a library subroutine,
it must contain the

statement,

ENTRY MATRIX J)

Since library subroutines are external to programs using them, the
calling program must list them in EXTERN statements.
2.6

SUPPRESSION OF SYMBOLS

When a parameter file is used in assemblies, many symbols get
Unused defined symbols take up space
defined but are never used.

in the binary file and complicate listings of the file. Unused
and unwanted symbols may be removed from symbol tables by the use
These pseudo-ops
of a pseudo-op, either SUPPRESS or ASUPRESS.

if
control a suppress bit in each location of the symbol table;

a suppress bit is on, the symbol in that location is not output.

The suppréss bit is used in the file S.MAC so that if a bit is on
and the symbol in that location is not used later, the symbol is
not output in the CREF table.

2.6.1

SUPPRESS SYMBOL Statement

The SUPPRESS statement turns on the suppress bit for the specified
symbols,

2.6.2

ASUPPRESS Statement

The ASUPPRESS statement turns on the suppress bit for all the symbols
in the

symbol table.

VERSION 47

June 1972

_oug2.6.3

- Listing Control

MACRO

Statements

Program listings are normally printed on a line printer or a terminal
depending
printed
A

sample

the

listing

source

listing

file

program

shown

device specified.
statements

in Chapter

are

Listings

processed

are

durlng

pass

From left to right the standard columns of a-listing contain
a)

the

instruction

the

symbollc

location

counter,

data

1nstructlon

addresses

are

octal

data

form,

and

followed

comments,

Relocatable

mark

(')

Data

object-code

which may

displayed

ment format.

occur

one

either

suffixed

the

of several

left

modes

by a

single

right

half.

depending

quotation

the

state-

The possible statement formats are:

Halfword

Instruction

- two 18-bit bytes
=

9-bit

op-code,

accumulator

4-bit

code,

indirect bit,
and an 18blt

l-bit

4-bit index,
address seg-

ment

Input/Output

‘

3-bit
I/0

I/0

indicator,

device

3-bit operand,
address

an
4)

Byte

pointer

7-bit

specification,

bit,

l-bit indirect

4-bit

18-bit address

index

and

segment

6-bit byte position, 6-bit
size, 1 unused bit,
l-bit indirect address bit,
4-bit index and an 18-bit

byte

address

segment

ASCII .

- 5 seven-bit bytes

‘6)

SIXBIT

six-bit bytes.

NOTE

Refer to the DECsystem-lo System Reference
Manual

for

complete

description

word

formats.

The Iisting function is suppressed within macro expansion, therefore
only

the

macro

VERSION 47

¢all

and

any

succeeding

lines

that generate

code

are

June 1972

~250-

MACRO

listed. Line printer listings always begin at the top of a page
and up to 55 lines are printed on each page. Consecutive page
numbers are printed in the upper right-hand corner of each page.
Each page also contains a title and a subtitle.

The standard listing operations can be augmented and modified by

using the following listing control statements.

DESCRIPTION

STATEMENT

PAGE

This statement causes the assembler to skip
(A form feed
to the top of the next page.
character in the input text has the same
effect and is preferred.

XLIST J

This statement causes the assembler to stop

listing the assembled program. The listing
printout actually starts at the beginning of
pass 2 operations.

Therefore,

to suppress

all program listing, XLIST must be the first
If only a part of
statement in the program.
the program listing is to be suppressed,
XLIST is inserted at any point to stop listLiterals are XLISTed
ing from that point.
if no LIT statement is seen before the END
statement.

LIST &

Normally used following an XLIST statement

to resume listing at a particular point in
The LIST function is implicitly
the program.
contained in the END statement.

LALL J

This statement causes the assembler to list
everything that is processed including all
text, macro expansions and list control
codes suppressed in the standard listing.

XALL 2

Normally used following a LALL statement to

J
SALL

This causes suppression of all macro and repeat expansions and their text; only the in-

resume standard listing.

put file and the binary generated will be
SALL can be nullified by either XALL
listed.
or LALL and the /M switch can be used instead
of

NOSYM J

SALL.

The assembler normally prints out the symbol
table at the end of the program, but the
NOSYM statement suppresses the symbol table
printout.

VErRSION 47

June 1972

-251STATEMENT

TAPE )

MACRO
DESCRIPTION

This pseudo=~op causes the assembler to begin
assembling
the program contained in the next
gsource file in the MACRO command string.
For
example,

«R MACRO
" ¥DSK:BINAME,LPT:«TTY:,DSK:MORE

PARAM=6
‘

TAPE

sTHIS

COMMENT WILL BE

IGNORED

would set the symbol PARAM equal to 6 and then
assemble the remainder of the program from the

source file DSK:MORE.
Since MACRO is a 2-pass
the TTY: file would probably be re-

assembler,
peated

for

pass

PASS

END

PARAM=6

- TAPE
+Z

Note that all
is ignored.

text after

the

TAPE pseudo-op

PRINTX MESSAGE‘)This statement, when encountered, causes the
single operand following the PRINTX operator
to be typed out on the TTY.
This statement
is frequently used to print out conditional
information.

PRINTX

statements

are

also

used

in very long assemblies to report the progress
of the assembler through pass 1.

REMARK COMMENTS)On pass 1 the message is printed on both the
list device and TTY.
on the TTY, but only

device.

On pass
if it is

2 it is printed
not the list

The REMARK operator is used for statements
which contain only comments.
Such statements
may also be started with a semi-colon.

COMMENTW

This

pseudo-op treats

the text

first non-blank character

between the

(delimiter)

and the

next occurrence of the same character as a
comment.
If the first occurrence of the

delimiter is a right

(left)

angle bracket,

the next occurrence of the delimiter must also
be a right (left) angle bracket.
The text
may include the carriage return, line feed
sequence.

For

example,

COMMENT/THIS

IS A COMMENT

THAT

THAN

ONE

LINE

LONG

/
Internally,

the pseudo-op functions as ASCII,

but no binary is produced.

VERSION 47

June 1972

MACRO
2.7

-252-

CONDITIONAL ASSEMBLY

Parts

basis

depending

The

program may be

general

form

IFN,

where

the

first

operand,

The

assembled,

conditions

not

defined by an assembler

meets

angle brackets

the

operators

statement

and

their

Operator

Assemble

<...>

N=0,

<...>

N>0

<...>

N=0,
N<O
N=0,

IFL N, <...>
IJFLE N, <...>
IFN N, <...>
IF1l, <...>
IF2, <...>

the

following

not

brackets

symbol

IFDEF

conditional
has

been

assembled

SYMBOL,

IFNDEF

statement.

only

the

are

listed

below:

angle-bracketed

coding

blank

N>0

N<O

N#0
encountered

during

pass

encountered

during

pass

statements,

defined.

The

assembly

coding

depends

enclosed

IF:

on whether
angle

if,

<...>»

SYMBOL,

assembled

conditions

IFE

optional

requirement.

IFG

is,

within

statement

<.cuo...-ot-..>

coding

IFGE

assembled,

<...>

this

symbol

defined

this

symbol

not

pseudo-op

defined

NOTE

SYMBOL
well

The

following

Arguments

are

can

user

conditional

op-code

statements

interpreted as

the assembler makes

the character

IFIDN
IFDIF

VERSION 47

the

logical

strings

<A-Z>
<A-Z>

third

operate

7-bit ASCII

to determine if the condition

The coding within

symbol.

comparison,

character

strings.

strings,

and

character-by-character

is met.

set of

angle brackets

enclosed by the

<A=7>,<...>
<A-I>,<...>

character

first two

(1)
(2)

are
are

sets

assembled

of angle

brackets:

identical
different

June 1972

MACRO

-253-

These statements, IFIDN and IFDIF, are usually used in macro expan-

sions (see Chapter 3) where one or both arguments are dummy vari'
ables.

An alternate form is to use delimiters as in ASCII.

For example:

IFDIF/A=Z/"A=Z" y<===>

This allows the use of > inside the character string. If the first
non-blank (space or tab) character is a < character, then the < >
method is used; otherwise, the character is used as a delimiter.
N

is followed by a single
" The last pair of conditional statements
blank or not blank,
either
is
_bracketed character string, which
and which is followed by conditional coding in brackets.

The codihg enclosed in the second set of angle brackets is as'

sembled if,

the first operand is blank
the first operand is not blank

IFB <i0e>,<000e>
IFNB <.v:e23<000s.>

A blank field is either an empty field or a field containing only
the ASCII characters space (408) or tab (118).
Again, delimiters can be used as in
IFB /

Sevees?

ASSEMBLER CONTROI. STATEMENTS

2.8
2.8.1
The

weee

REPEAT Statements

statement

REPEAT N,

<...>J)

causes the assembler to repeat the coding enclosed in angle

If more than one instruction or data word is

brackets n times.

by a carriage return.
to be repeated, each is delimited

‘

example,

ADDX:
VERSION 47

For

REPEAT 3,

2-27

June 1972

MACRO

-254-

assembles

as,

ADDX:

ADD

6,(h)

ADDI 4,1
ADD €,X(L)
ADDT k4,1
ADD 6,X(4)
ADDI 4,1

Notice

line
any

that

the

level.

REPEAT

label

assembled
The

coding.

following

statement

REPEAT

statement

REPEAT

simplified

placed

statements
example

may

shows

how

the

first

nested
a

nested

interpreted.

REPEAT

3,<A
)

REPEAT

2,<BJ
C>

assembles

as,

2.8.2

NOTE

OPDEF

Brackets

indicate

repetition.

Statements

The programmer can define his own operators‘using an OPDEF statement,

which

written

OPDEF

where

the

defined

brackets.

VErRsION 47

first

SYM

second

the

form:

{[STATEMENT
]

operand

by the
The

is defined

operand,

which

operand

operator,

evaluated

enclosed
as

whose

function

square

statement,

and

the

JuNE 1972

MACRO

-255result is stored in a 36-bit word.

' OPDEF CAL1
defines CALl
may

[@39080 280284]

as an operator,

now be used

equivalent

with the value

statement

#3014g @g@1234
which is

For example,

030000

CALl

operator.

CAL1 3,1234

to,

@g3g1hg gg1234

3,1234(39029)

When MACRO-10 encounters
a user-defined operator,
single object-program storage word in the

struction word

000000.

(see Chapter 1).

fied by accumulator,

indirect,

it assembles a

format of a primary in-

The defined

36-bit value is modi-

memory address and index fields

specified by the user-defined operator.

For

example,

OPDEF CAL [MOVE 1,@SYM(2)].
CAL 1,BOL(2)))

The CAL statement is equivalent to:

MOVE 2,@SYM+BOL(4)

In this modifiication the accumulator fieids are added, the indirect
bits are logically ORed,

the memory address fields are added,

and the index register addresses are added;

>2.8.3 SYN Statements
The

statement

SYN symbol, symbol
defines the second ope;and as synonymous with the first operand,

which must have been previously defined.
symbol or a macro name.

Either operand may be a

If the first operand is a symbol, the

. second is defined as a symbol with the same value.

VERSION 47

If the first is

JuNe 1972

-256-

MACRO
a macro

name,

cally.

1If

tor,

the

first

second

becomes

a machine,

second will

interpreted

name

assembler,
in

the

SYN

statement

has

been

label

and

operator,

the

second

with

the

which

operates

identi-

or user-defined

same manner.

previously
operand

defined
is

opera-

the
as

synonymous

label.

following

are

legal

SYN

statements:

SYN K,X

JIF K=5, X=5

SYN FAD,ADD)

SYN

2.8.4

a macro

first

operand

both

The

the

END,XEND
)

Extended

Instruction

Statements

For programming convenience, some extended operation codes are provided

place

those

the MACRO-10 Assembler.
DECsystem-10

Primarily,

instructions

instruction mnemonic

and

accumulator

single

For

example:

instruction.

where

field

these
the

are

used

combination

used

denote

re-

JRST 4

equivalent

replace certain commonly used

halt

instruction.

1In

I/0

addition,

they

are

instruction-device

used

number

combinations.

The

extended

instruction

instruction
statements

they may not

have

The

field must

operator

statements

I/O

exactly

instruction

accumulator

have

are

field

one of

the

JRST

12,

the

statements,
device

primary

except

that

field.

following

Equivalent
Extended
Machine
Instructions | Instructions
JEN

extended mnemonics:

Meaning
Jump

and

enable

the

system

(priority

HALT

JRST

Halt

JRSTF

JRST

Jump

JOV

JFCL 14,

JFCL

Jump on overflow and clear

JCRY1

JFCL

Jump

JCRY

JFCL 5,

Jump on CRYZ or CRYl and clear

JEOV

JECL

Jump

RSW

DATAI 2

Read

the console switches

JCRYZ

and

restore

interrupt)

flags

Jump on CRY@ and clear
on

CRY1l

and clear

floating overflow

JUNE 1972
2-30

-2572.9
2.9.1

MACRO

MULTI-FILE ASSEMBLY
UNIVERSAL Name

UNIVERSAL files may be used to generate data, however, they are
normally used to'generate symbols, macros and opdef's (user-

defined operators).

The symbolS'géneratéd.by UNIVERSAL files need

not be declared as INTERNAL symbols since all local symbols in
files of this type are made available to all programs permitted
access to the file.

UNIVERSAL files used to generate data can save time by being set
up for a one-pass operation since symbol definition needs to be

assembled on one pass only.

This one-pass operation can be ac-

complished in either of two ways:
1)

UNIVERSAL NAME
PASS

END

UNIVERSAL NAME
IF

<END>

END

The first generates a listing; the second does not.

1f the UNIVERSAL pseudo-op is seen in a program, the NAME is stored

in a table and a flag is set. When the END statement is seen, the
‘symbol table is moved to just after the pushdown stacks and buffers;
therefore, the pushdown stacks and buffers cannot be increased during
assembly. The first assembly should use the maximum of I/0 devices
to be used later. The free core pointer is moved to after the top of

‘the moved symbol table, and pointers are stored to enable the table
to be scanned.

When assembling is done from indirect files, the universal files must
be recompiled by the /COMPIL switch. Otherwise if a REL file later
than the source exists, the universal file will not be conpiled,
and the symbol table will not be available. In addition, if the
universal routine is modified, all routines which use it must be

recompiled by either using /COMPIL or deleting all REL files.

VERSION 47

JUNE 1972

MACRO

-258-

2.9.2

The

SEARCH Name

scan

table.
with
of

statement

the

Multiple

commas;

ten

MACRO.

When

the

found,

with

the

sequence

the

order

specified.

specified

since

each

searched
be

maximum may

pseudo-op

the

seen,

table.

the

specified

the

symbol

table.

expense

Universal

symbol
them

A maximum
requires

redefined with

names

specified

MACRO

separating

name

specified

the

for

current

the

are

names

symbol

com-

cannot

all

tables

files

may

names

is cleared on

responds with

the

symbol

procedure

search
list

have

table,

table

the

symbol

been

time

but

searching

not

searched.

When

moved
future

files

not

into

the

in
the
a

symbol

current

references

The

symbol

found
until

is
on

universal

continue

assembled.

START,

universal

been

each RUN or

found,

saves

other

have

are

whenever

This

core.

names

table.

This

name

sequence

specified

auxiliary

names

VERSION 47

This

the

UNIVERSAL

specified

order

found

FIND

built.

in the

found

symbol
the

current

symbol

found

the message

output.

tables

may

UNIVERSAL

CANNOT

table

tables

This

symbol

not

core.

.UNIV

specified

may

words

are

the

symbol

tables

four

pared

symbol

they

symbol

opens

required

long

all

table

universal

cleared when MACRO

asterisk.

JUNE 1972

-259-

MACRO

Chapter 3
Macros

When writing a program,

certain coding sequences are often used

several

the

ent
To

if
do

times

entire

this,

the

~a macro
name,

with

the

only

sequence

coding

instruction.

along with

arguments

sequence

A single

list

changed.

statement

real

so,

referring

arguments,

single

defined with dummy

conveni-

statement.

arguments

generates

sequence.
3.1

can be generated by a

the macro by
the

correct

DEFINITION

OF MACROS

The first statement of a macro definition must consist of the opera-~
tor

DEFINE

followed by

the

symbolic

be constructed by the rules
name

may

entheses.

followed

name of

the macro.

for constructing symbols.

string

dummy

arguments

The

name must

The macro

enclosed

any symbols that are convenient--single letters are sufficient.
comment may

The

follow the

character

delimited by
of

group

VERSION 47

par-

The dummy arguments are separated by commas and may be

sequence,

dummy

which

angle brackets.
complete

argument

constitutes

the

The

the macro

body

list.

body

the

macro,

normally

consists

statements.

JUNE 1972

-260-

MACRO

For example, this macro computes the length of a vectors:
DEFINE VMAG (A,B)

;ROUTINE FOR THE LENGTH OF A VECTOR

FMP @

:GET THE SECOND COMPONENT

<MOVE @,A

{GET THE FIRST COMPONENT
{SQUARE IT

MOVE 1,A+1
FMP 1,1

1SQUARE IT

FMP 1,1

{SQUARE IT

.ADD THE SQUARE OF THE SECOND
.GET THE THIRD COMPONENT

FAD 1
MOVE 1,A+2

.ADD THE SQUARE OF THE THIRD

FAD 1

{USE THE FLOATING SQUARE ROOT ROUTINE

JSR FSQRT

.STORE THE LENGTH>

MOVEM B

NOTE

Storing comments in a macro takes up space.
If the comments start with a double semi-

colon (;;) the comment will not be stored;

therefore, it lists in the original definition but does not list when the macro is
expanded.

3.2

MACRO CALLS

A macro may be called by any statement containing the macro name followed by a list of arguments. The arguments are separated by commas
and may be enclosed with parentheses. If parentheses are used (indicated by an open parenthesis following the macro name), the argument string is ended by a closed parenthesis. If there are n dummy
n,
arguments in the macro definition, all arguments beyond the first
nt
if any, are ignored. If parentheses are omitted, the argume
tions
defini
string ends when all the dummy arguments of the macro
have been assigned, or when a carriage return or semicolon delimits
an

argument.

be
The arguments must be written in the order in which they are to
substituted for dummy arguements. That is, the first argument is
substituted for each appearance of the first dummy argument; the
d
second argument 1is substituted for each appearance of the secon
dummy arguemnt, etc. For example the appearance of the statement:
VMAG VEC,

LENGTH

in a program generates the instruction sequence defined above for
the macro VMAG. The character string VECT is substituted for each
ter
occurrence in the coding of the dummy argument A, and the charac
string LENGTH is substituted for the single occurrence of B in the
coding.

VERSION H7

3-2

JUNE 1972

MACRO

-261-

fields. The value of the
gtatements with a macro call may have label ion
generated.
1abel is the location of the first instruct
CAUTION

ed only by COMMA,
MACRO arguments are terminat
CLOSE PARENor
N
COLO
CARRIAGE RETURN, SEMI
string was
ment
argu
re
enti
the
n
(whe
IS
THES
These
.
sis)
nthe
pare
open
an
started with
ments
argu
in
uded
incl
characters may not be Spec
es or
spac
,
ally
ific
.
used
are
<>
unless
will
they
s;
ment
argu
e
inat
tabs do not term
itself.
d part of the argumentment
pe treateas
s, it
argu
e
inat
term
not
does
ol
The symb
to be
ols
symb
r
othe
and:
as
just permits comm
used as part.of an argument.

3.3

MACRO FORMAT

commas. However, arguments
must be separatedForby exam
a. Arguments cont
ple:
as.
comm
ain
may also

DEFINE JFQ(A,B,C)
<MOVE [A]
CAMN B
JRST C>

tion B is equal to A (a literal), the
Tf the data in toloca
C. TIf A is to be the instruction ADD 2,X%,

program jumpsmacro instruction would be written
the calling

JEQ<ADD 2,X>,B,INSTX

)

]

surrounding thed.argument are removed,
The angle brackets
and the proper coding is generate
as, semiment contains comm
is: If an argu
The general ruleothe
ment must
argu
the
rs,
mite
deli
ment
colons, or any anglre argu
nesting,
of
l
leve
y
ever
brackets. For
be enclosed in e brackets
pass
to
e,
efor
ther
ved;
remo
igs
one set of angl
ment
argu
the
os
macr
ed
nest
to
ng commas
arguments containi
each
for
kets
brac
e
by one set of angl
should be enclosed
argument,
jevel of nesting. The > does not terminate the
a comma must be used.

b. A macro need not have argfiments. The instruction:
DATAO PIP,PUNBUF(4)

ster 4,
of PUNBUF, indexed dbyby regi
which causes the contents
macro:
the
rate
gene
be
may
,
tape
r
to be punched on pape
DEFINE PUNCH

‘

The calling macro instruction could be written:
PUNCH

VERSION 47

JUNE 1672

MACRO

-262PUNCH

of
¢.

calls

the

The

macro

brackets

the

name

argument

and

without
for

three

words

first

load

can

within

are

symbols

the

arguments,

within

exactly

the

and

each

the

with

often

them

items

address
the

A
of

table

convenient
the

call,

body.

the

macro,

what

the

for

are

macro

and

labels

the

used

that

form..hijk,

user

points.

The

first

are.

labels
for

sign

(%),

called,

symbol

all

created

the

calling

assigned

overruled

Null

this

into

that

advised

created

for
not

there

address

the

last

item.

time

the

generate

list

that

the dummy

dummy

string:

macro

symbol

macro.

These

two

the

some

to
is

decimal

use

not

example,

..0001,

Created
are

there

the

the
is

considered

refer

different

sym-

called.

called,

macro

the

are

form

the

form

are

%X,

the

same

been

are

perczant

replaced

arguments

percent
normal

missing

defined

two

a macro

position

the

four

..0002,

When

sufficient

assigned

has

preceded

symbol

symbols

starting with

the

followed

symbol.

arguments

argument

suppose

generated

points

symbols

statement

that

arguments

stated

are

symbols
symbols

reason

macro

the

definition

considered

However,

For

the

where

purpose.

symbol

symbols.

arguments

macro

a -table

example,

created

not

arguments

ments.

the

necessary

is,

missing

and

appear

etc.

dummy

body

length,

Nevertheless,

each

requires

inserted

The

may

replaces

item.

follows:

stating
the

digits.

the

angle

example:

number

store

macro

used

the

outside

time

For

from

bols

generated

definition

for

used

index

called,

string

SYMBOLS

Each

list

contained

X,L(FIRST,LAST)

concerned

Created

The

string.

item,

called

labels

must

macro

are

the

explicitly

these

statement.

expression

labels

instruction

L(A,B)<3%<B-A+1>>

gives

macro

DATAO

followed

CREATED

name,

MOVEI

When

the

anywhere

DEFINE

3.4

for

macro.

with

sign

manner.

arguthe

(A,%B,%C)

VERSION 47

JUNE 1972

063If

the

macro were

called with

the

MACRO

argument

string:

(OPD,) or OPD,,

The second argument would be considered to have been declared as
null string.

This would override the % prefixed to thé second dummy

argument and would substitute the null string for each appearance of
the second dummy argument in the statement.

gument is missing.
3C.

For example:
DEFINE
<JSR

Héwever,

the third ar-

A label would be created for each occurrence of
|

TYPE(A,%B)

TYPEOUT

JRST %B
SIXBIT/A/

#B:>
This

macro

types

Teletype.

the

TYPEOUT

text

string

substituted

output routine.

for A on

Labeling

the

console

location

lowihg the text is appropriate since A may'be text of indefinite
length.

A created symbol

appropriate

for this

label

programmer would not normally reference this location.
might

fol-

since the

This macro

called by:
TYPE

HELLO

which would result in typing HELLO when the assembled macro is executed.

If the call had been:
TYPE

HELLO,BX

the

effect would

8B,

overruling the effect
of the percent sign.

3.5

the

same.

However,

catenation operator.

Rather,
or

macro

expression when
definition.

strings

<JUMP'A

bol.

for

(')

is defined as the con-

it may be a string of characters which form a complete sym-

DEFINE

When

substituted

A macro argument need not be a complete symbol

This

joined

characters

joining,

called

already

called,
If

the

VERsION 47

the
call

joined.

contained

concatenation,

by the assembler when the programmer writes
the

CONCATENATION

The apostrophe character or single quote

bol

BX would

the

performed

an apostrophe between

example,

the macro:

J(A,B,C)
B,C>

argument A

suffixed

JUMP

form

single

sym-

were:

June 1972

-264-

MACRO
J

(LE,3,X+1)

the generated code would be:

JUMPLE

3,X+1

The concatenation (') may be used in nested macros. The assembler
removes one operator when it performs concatenation if it is next
to

3.6

(before or after)

a dummy argument.

DEFAULT ARGUMENTS

Missing arguments in macros are generally replaced by nulls.
example,

For

the macro

DEFINE FOO
EXP

(A,B,C)>

A,B,C>

when called by FOO(l) would generate three words of 1, @, and #.
Default arguments may be supplied to override missing arguments.

When supplied, default arguments are written within angle brackets
(<>) after each argument. For example, the addition of default arguments 222 and 333 to arguments B and C of the foregoing example
macro would be written as

DEFINE FOO (A,B<222>, <<333>)
EXP A,B,C>

If the foregoing macro is called by FOO(1l) it would generate the
number

1,222,333.

The following example program illustrates the use of defined default
arguments.

VErRSsION 47

JUNE 1972

-265-

.MAIN

MACRO

MACRO 47(113) 1¢:1h4 28-MAR—72 PAGE 1

FOO

MAC

28-MAR-72

1¢:13

DEFINE FOOl

(A,B,C)<

DEFINE

(A<111>,B<222>,C<333

EXP

A,B,C>

FO0O2

> )<

PUOgan' E9EPE

pEgaEL"
pRamg2'

@APPPl

gAdNIE
GOURRE

BPIgID
BORUET

gEARE3 0AB0RY

POPPP1

goggel
pgepags’

pgages
2peaes

ERRORS

A,B,C>

FOO1L

(1)4

EXP1,,+

Loon (1),

EXP 1,222,3334

pppE222
@9¥333

‘

EXP

END

DETECTED

PROGRAM

BREAK

CORE

USED

3.7

g@ggge

INDEFINITE

REPEAT

It is often convenient to be able to repeat a macro one or more times
for a single call,

each repetition substituti
. successive
ng arguments

in the call statement for specifiéd arguments in the macro. This may
be done by use of the indefinite'fepeat operator, IRP.
The operator
IRP is followed by a dummy argument, which may be enclosed‘in paren-

theses.
ment's

This argument must also be contained in the DEFINE.statelist.

This

argument

broken

into

subarguments.

When

the

macro is called, the rangé of the IRP is assembled once for each
subargument,

the

successive

subarguments

being

substituted

for

appearance of the dummy argument within the range of the IRP.

each

For

example, the single argument:
<ALPHA,BETA, GAMMA>

consists

the

subarguments

ALPHA,BETA,

and

GAMMA.

The macro

de-

finition:
DEFINE
<IRP

DOEACH(A),

<A>>

and

the

call:
DOEACH<ALPHA,BETA,
GAMMA>

produce the following coding:
ALPHA

BETA

Vers1on 47

GAMMA

June 1972

MACRO
An

~266-

opening

angle

bracket must

the

argument

the

IRP

state-

ment to delimit the range of the IRP since the argument is one argument to the macro. 2 closing angle bracket must terminate the
range

ter
a

the

a time;

program

IRP.

IRPC

each character

that

uses

sometimes

IRPC

desirable

depending on conditions
operator

STOPl.

finishes

expanding

and terminates

When

A<IFE

CONV1

A>>

that

the

STOPl

only

a complete argument.
in Chapter

the
is

takes

IRP

for

An example of

Figure

7-4.

indefinite

assembler.

encountered,

the

one charac-

This

done by the

the macro

the present

repeat

processor

argument

An example:

(A)

K-A,<STOPI

value

CONVERT

stop processing

range

CONVERT

<IRP

except

given

given by

the

IRP

the repeat action.

DEFINE

Assume

then

the

call:

for

the

first

code

@,1,2,3,4,5,6,7

<IRP
IFE K-@,<STOPT
CONV1 @>
IFE

K-1,<STOP1
1>

CONV1

IFE

K-2,<STOPI

CONV1
IFE

K-3,<STOPI
3>

CONV1

The

assembly

the macro.

condition

Therefore,

fourth argument,
the

STOPI

code

arguments.

3.8

NESTING

AND

the

action

i.e.,

further scanning of
for

the

current

the macro CONV1

the

argument

(defined

REDEFINITION

that is,

The

macros may be defined within other
levels may be

outermost macros,

the macro processor,

VERSION 47

for

arguments

When the condition is met,

continues

a call

three

not encountered until the

number 3.

For ease of discussion,

nested macros.
to

the

argument

Macros may be nested;
macros.

processed which prevents

CONV1

with

not met

the STOPI

which

However,

and generates
elsewhere)

may

called

i.e.,
first

assigned to these

those

defined directly

level macros.

Macros

June 1972

-267defined within
macros

level

defined within

macros;

macros may

second

level

called

macros

may

second

level macros;

called

third

level

etc.

At the
macro

first

MACRO

beginning
processor

second

and

processing,

and

higher

may

level

first

called

macros

are

level
the

macros

normal

not yet

are

known

manner.

defined.

the

However,

When

first

level macro containing second and higher level macros is called,
all its

second

after,

the level

level macros
of

become defined to

definition

called in the normal manner.

irrelevant,

Of coufse,

macros contain third level macros,
defined

until

the

second

the processor.
and macros

may

Therebe

if these second level

the third level macros are hot

level macros

containing

them have

been

called.

When a macro of level n contains a macro of level n+l, calling the
macro results

generating

the

body

the

macro

into

the

user's

program in the normal manner until the DEFINE statement is encountered.

The

level

it does

not

appear

complete,
to

the

n+l

the macro

user's

macro

then

in the ueer's
processor

program until,

defined

program.

resumes

the macro

generating

or unless,

processor;

When the definition

the

the macro body

entire macro has

in-

been

generated.

a macro

another

ginal

The

definition

first

may be

name which

definition

example

rewritten

has

a macro

EXP

VMAG
VL:

4C:>

appears

redefined,

and

within

the

ori-

both

the

length

nesting

and

redefinition.

a vector

(A,B,%C)

VMAG

(D,E)

C,E>

(A,B)

JRST %C
HRRZ 2, (SJ)
MOVE (2)
FMP @
MOVE 1,1(2)
FMP

1,1

FAD

FMP

defined

SJ,VL

MOVE

VERSION 47

calculation

illustrate

DEFINE VMAG
<JSP

previously

the macro

eliminated.

<DEFINE

been

statement,

1,2(2)
1,1

FAD

JSR

FSQRT

MOVEM @1 (SJ)
JRST 2(3J)

3-9

JUNE 1972

-268-

MACRO

The procedure to find the length of a vector has been written as a

closed subroutine.

It need only appear once in a user's program.

From then on it can be called as a subroutine by the JSP instruction.
The first time the macro VMAG is called, the subroutine calling se-

quence is generated followed immediétely by the subroutine itself;
Before generating the subroutine,

the macro processor encounters a

This new macro is deDEFINE statement containing the name VMAG.
Henceforth,
fined and takes the place of the original macro VMAG.
However,
when VMAG is called, only the calling sequence is generated.
the original definition of VMAG is not removed until after the expansion

1is

complete.

Another example of a nested macro is given in Chapter 7, Figuré 7-4.
3.8.1

ASCII

Interpretation

If the reverse slash (\) is used as the first character of an argument in a macro call, the value of the following symbol is converted

to a 7-bit ASCII character in the current radix.
MAC

and if A=500
character

VErRs1ON 47

If the call is

(in the current radix), this generates the three ASCII

"500".

JUNE 1972

-269-

MACRO

Chapter 4
Error Detection

MACRO-10

makes

statements.

many

error

apparent

checks

error

it processes

detected,

source

the

language

assembler

prints

a Singie letter code in the left-hand margin of the progra
m listing
(and on

the

The

the

programmer

the

should

not

listing

error

however,

described

all
in

listing

examine

correction

code

are

the

each

error

required.

total

the

TTY),

the

same

line

errors

indication

the

found;

to determine

end

the

this

printed

listing,
even

requested.

indicates

caused by
the

general

illegal

preceding

three

4.1

SINGLE-LETTER ERROR CODES

Table

4-1

lists

in question.

assembler prints

Each

unless

statement

whether

TTY,

the

single-letter

class

usage

errors.

chapters

error

These

the MACRO-10

codes

this

output

errors,

language,

manual.

the

assem-

bler.

VERSION 47

June 1972

-270-

MACRO

TABLE

4-1

Error Codes

Error Code

Meaning

Argument error in

pseudo-op

Explanation
This is a broad class of errors.

which may be caused by an improper argument in a pseudo-op.

The following represent the

majority of the conditions which
would cause an A code error.

Symbol used is improperly
For example AB?CD
formed.
would result in an A code
since the character ? is
not in the Radix 50 character

set.

IFIDN comparison string is
too

large.

OPDEF of macro is SYN,.

OPDEF,

Invalid SIXBIT character
in SIXBIT/TEST Tab/

no code generated.

f. Byte size too big in byte
(>4D36)
.

Radix 50 code not absolute,
Radix 50 FOO,BAR

that is

where FOO is not @-74 absolute.

End of

line on IFx SYM

reached before an < character

seen.

i. Assignment made in an address field (e.g., MOVE

A=10).

j. Assignment of a label
(e.g., TAG: TAG=1).

k. Missing symbol in SYN SYM1,.
1.

Unknown symbol in SYN,.

m. Missing right parenthesis
())

in index

ll(zooo)p

(e.g.,

MOVE

n. Missing left parenthesis
in BYTE statement

VERSION 47

(e.g.,

No comma after repeat count

IRP not in a macro.

.JUNE 1972

-271TABLE 4-1 (Cont)
Error

MACRO

Meaning

Code

Explanation
q.

Argument

dummy

for

DEFINE

FOO

IRP(B),
r.

IRP

symbol;

for

(A)

not

example

<>>

argument

created

symbol.
S.

Multiply-defined
symbolic

error

This

reference
:

STOP1l

not in

IRP.

statement contains

which refers to
defined symbol.
bled with

the

a tag

a multiplyIt is assem-

first

value

de-

fined.

External symbol

Improper

usage

error

symbol.

The

sent

external

following

the majority

repre-

the

tions which will cause
code error.
a.

Attempting

use

both

symbol

internal

and

For

example,

EXT:

EXTERN

attempts

condi-

the

same

external

symbol.

the

statement

TXT,BRT,EXT

use

EXT

external

and

both

internal

symbol.

Using
for

e.
f.

HISEG

TWOSEG

external

the

literal

symbol

improper.

generate

The

message

symbol

CODE

typed

out

words.

GENERATED

retains

and

1lit-

defined more
symbol

M is

count.

first definition,

_VERSION_47.

symbol
of IOWD.

external

[SIXBIT //]1;NO

cnce.

pass

pseudo-op.

left half

REPEAT

eral must

symbol

an external symbol
an ARRAY size statement.

Using
a

Multiply-defined

external

IFx.

Using

EXP

symbol

index.

Using

an external symbol
a LOC, RELOC, PHASE,

error

external

Using

Literal

an AC

than
its

the error
during

If this type of error occurs
during pass 2, it is a phase
error (see below).

JuNeE 1972

-272-

MACRO

TABLE 4-1

Error Code

Meaning

(Cont)

Explanation

If a symbol is first defined as
a #-sign suffixed tag, and later

as a label, it retains the label
definition.
Examples:

ADD 3,X;

ADD 3,X#;

A: MOVE ,C; M ERROR

X: MOVE-,C; X IS ASSIGNED THE
CURRENT VALUE OF THE LOCATION
’

’

COUNTER

Multiple appearances of the TITLE
pseudo-op (which generates both
a title line and program name)
are flagged as "M" (Multiple
definition)

Number error

errors.

A number is improperly entered.

The following represent the ma-

jority of the conditions which

would cause an N-type error.

a. The number exceeds the permitted range

(e.g..

+F13.33E38).

b. A number does not follow a
B shift operator (e.g.,
.
4+D15BZ)

c. The number exceeds the cur-

rent radix (e.g., if radix
is 8 the single character
9 is acceptable but the

number 19 is not acceptable).

d. The binary shift given does
not represent an absolute
For example,
numeric.
4B<sym> is illegal if sym
is relocatable.

e. The character given after
an up arrow

(t+)

is not B, O,

L or D.

£. The expression given after
E was not a signed (+) number.

Operation code un-

Phase error

defined

The operation field of this state-

ment is undefined.

It is assem-

bled with a numeric code of #.

A symbol is assigned a value as
a label during pass 2 different

from that which it received during pass 1. In general, the assembler should generate the same
number of program locations in

pass 1 and pass 2, and any discrepancy causes a phase error.

VERSION 47

JuNe 1972

-273~

TABLE
Error

Code

4-1.

MACRO

(Cont)

Meaning

Explanation
For

example, if an assembly conditional, IF1l, generates three

instructions,
sults

such
gram

Questionable

unless

a phase error
another

re-

conditional,

IF2, generates three prolocations during pass 2.

This is a broad
errors in which
finds ambiguous

class of possible
the assembler

Q-errors

may

correct

code;

the

attempt

what

intended.

The

mer

language.

not

generate

assembler will
the

program-

following re-

present the majority of the
ditions which would cause a
Q~type
a.

error.

than

ASCII characters

are

detected

bler

before

symbol

the

is detected

"ABCDEFG"

than

ters

are

"ABC

(e.g.,
).

When

characters

the

before

are

first

SIXBIT

detected

assembler

assem-

closing

detected, only
are stored.
b.

con-

characby

the

closing

" symbol is detected.
As
in item a, only the first
6 characters are stored
when more than 6 are detected.

given

number

such

cases,

the

order
are

bits

number

E in a number is followed
by something other than a

signed

(+)

numeric (e.qg.,

—

illegal

detected

ASCII

the

big;
high-

lost.

1.9EX) .
e.

too

control

characters

not

permitted

LF,

VT,

comma

EF,

statement

character

line:
@-40

except

and

detected
after

all

are

for

HT,

ESC.

the

required fields have been
filled (e.g., MOVE 1,2,)
g.

Relocatable

ated

fore

either

HISEG

the

VERSION 47

the

code

is generassembler be-

the pseudo-op

TWOSEG

found

assembler.

JUNE 1972
4-5

MACRO

-274TABLE

Error Code

4-1

(Cont)

Explanation

Meaning
h.

Relocation error

An instruction address
pointer is detected by the
assembler which does not
have either all @#'s or all
1's in the left half of
its word location.

A LOC or RELOC pseudo-op is used
improperly.

All of the following

conditions will cause an R-type
error.

a. An expression or assignment

is made in which relocation

is not # or 1 (e.g., A+B,
A*7Z, 1/B, or X=3*B where

a and B are relocatable).

b. A BLOCK statement is written with a relocatable size
(e.g., BLOCK: A where A is
.
relocatable)

c. A relocatable variable is

used to specify an accumu-

lator

A is

(e.g.,

MOVE A,l where

relocatable).

Undefined symbol

A symbol is undefined.

Value previously
undefined

A symbol used to control the assembler is undefined prior to

‘the point at which it is first
Causes error message in
used.
pass

For example, BLOCK:A where A is
undefined.

Macro definition

error

An error occurred in defining or
calling a macro.

The page
Error messages printed during pass 1 consist of two parts.
and sequence number, if used, plus the most recently used label is
This material is then followed by +n,
printed on the first line.
where n is the (decimal) number of lines of coding between the labeled statement and the statement containing an error.

The second

line of the error message is a copy of the erroneous line of coding,

with a letter code in the 1eft—hénd margin to indicate the type of
error.

If more than one type of error occurs on the same line, more

than one letter is printed;

but if the same type of error occurs

more than once in the same line, a single letter code is printed.

VErsion 47

JUNE 1972

-275During pass
rors

are

printed

4,2

marked

the

end

listing

printed

letter

codes,

listing.

the

MACRO

and

out,

total

lines
of

containing

errors

found

er-

ERROR MESSAGES

The

following

error messages

may

typed

out

Any error meSsage preceded by a question mark

(?)

fatal

error when running under

minated by
END

PASS

This
2.

message

required
message is

This

FILE

This

ter-

paper

that manual

to start pass
issued when
tape,

cards

message

indicates that manual
required when the files
to be input are on paper tape,
cards or being input from the
terminal.

ERROR

This

‘

message

error was
mand

?INSUFFICIENT CORE
?.PDL

run

indicates

input
keyboard.

?COMMAND

(the

BATCH).

the

THE

user's. terminal.

is treated as a

the BATCH processor

LOAD

the

indicates

found

string

the

that

last

com-

input.

Not enough core is available.

OVERFLOW,TRY/P

This

message

indicates

pushdown

list

use

too

switch

size

that

the

small.

The

increases

the

of the pushdown list by 80
locations.
As many /P switches

?DEV

NOT

AVAILABLE

may

The

specified

used

desired.

device

initialized because
is using it.
?N

ERRORS

ERROR

ERRORS

DETECTED
DETECTED

These
»

the

DETECTED

three

number

END

STATEMENT

ENCOUNTERED

INPUT

the

errors

run

indicate

detected

terminated.

FILE

This message
of

VERSION 47

statements
of

user

MACRO during assembly (errors
marked by letter codes on the
listing.
Under BATCH if any error
occurs,

¢NO

cannot be

another

the

followed by one

following:

LITERAL

DEFINE

TEXT

IN
IN

CONDITIONAL
CONDITIONAL

MACRO

REPEAT

CALL

JUNE 1972

MACRO

276-

and

ON PAGE xxx AT yyy

where xxx = a page number and yyy
a sequence number Or TAG+offset.
NOTE

The foregoing type of message

usually indicates some error

other than a missing END state-

ment.

For example:

ASCIZ/TEXT

END

where TEXT has not been closed
or

JRST [statements

END

where the literal has not been

closed.

?PRGEND ERROR

that
This error message indicates the
the macro failed to restore prosymbol table for one of the

.
grams

?2T00 MANY UNIVERSALS

that
This error message indicates have
too many universal programs
The number of
been assembled.
permitted is a
rams
prog
universal

this
Macro parameter; to prevent user

error from reoccurring, the
must reassemble macro with a new

parameter which will permit the
desired assembly.

?2CANNOT FIND UNIVERSAL xxX

This message indicates that a

L
search has been made for UNIVERSA
found
program xxx but it was notd).
To
(i.e., it was not assemble

clear this error the program XXX
must be assembled.

L
xxx UNASSIGNED DEFINED AS IF EXTER NA
This message indicates that an
undefined symbol was found and

that it has been treated as if it

was an external symbol.
PROGRAM BREAK IS XXX

Wwhere xxx is the length of the low

HI-SEG BREAK IS xxX

Where xxx is the length of the

VErRSION U7

segment.

relocated high segment.

June 1972

~277ABSOLUTE

BREAK

xxx

Where

CORE

MACRO
xxx

address

USED

- Message

low

MUST HAVE

FILES

SAME

the

highest

over

indicates

segment

source

PUNIVERSAL PROGRAM(S)

seen

used

absolute

144.

the

size of the
assemble the

program.

OUTPUT SPECIFICATIONS

OTHER

This error message indicates
universal program was found

that
which

did not have either a binary or
a
listing device specified but
all of

the

following

fications.

files

For

had

example

such

the

speci-

sequence

¥ <UNIV
¥pel,List<«file
is

illegal.

would

The

legal

sequence

*¥rel, LIST<«UNIV
¥REL,LIST«FILE
?ERROR

WHILE

EXPANDING

Xxx

This
the
nal

error message indicates
that
assembler experienced an
intererror while expanding the
macro

identified
as xxx.

type

are

curs

the user

macro

4.2.1

The

LOOKUP

following

RENAME

error messages
request

filename.ext

FILE WAS
enter

then

should

involved.

can occur during a monitor

disk.

one

NOT FOUND

errors

(1)

(2)

PROTECTION

DIRECTORY

The

the

(@)

only)

FOR

form of

this

rewrite

oc-

the

error

LOOKUP,
messages

is:

following

ILLEGAL FILE NAME

PROJECT-PROGRAMMER

(used

for

NUMBER

FAILURE

(3)

FILE WAS

(4)

RENAME FILE NAME ALREADY EXISTS

(5)‘

Errors

rare;

Errors

ENTER

(9)

extremely

BEING MODIFIED

ILLEGAL SEQUENCE OF UUOS

(6)

BAD UFD OR BAD RIB

(7)

NOT A SAV FILE

(1)

NOT ENOUGH

(11)

DEVICE

(12)

NO SUCH DEVICE

CORE

NOT AVAILABLE

(13)

NOT TWO RELOC REG.

(14)

NO ROOM OR

(15)

WRITE LOCK ERRCR

VERSION 47

QUOTA

CAPABILITY

EXCEEDED

JUNE 1972

-278-

MACRO
(16)

NOT ENOUGH MONITOR TABLE SPACE

(17)

PARTIAL ALLOCATION

ONLY

(29). BLOCK NOT FREE ON ALLOCATION

CAN'T SUPERSEDE

(22)

CAN'T DELETE (RENAME)

(23)

SFD

(24)

(25)

SFD

(26)

NO-CREATE ON FOR SPECIFIED SKFD PATH

NOT

?2(V)

A NON-EMPTY DIRECTORY

FOUND
LIST

NESTED

If the error code

(ENTER) AN EXISTING DIRECTORY

(21)

(V)

EMPTY

TOO

DEEPLY

‘

is greater than 268’

LOOKUP,ENTER,

the error message:

OR RENAME ERROR

printed.

4.2.2

MACRO

I/O Error Messages

The following error messages are generated for error conditions
found during input or output operations with peripheral devices.
messages

are

The

self-explanatory.

?20UTPUT WRITE-LOCK ERROR DEVICE xxX

?0UTPUT DATA ERROR DEVICE

xxx

20UTPUT CHECKSUM OR PARITY ERROR DEVICE xxX
?0UTPUT QUOTA EXCEEDED ON DEVICE xxx
?0UTPUT BLOCK TOO LARGE DEVICE xxXx
?MONITOR DETECTED SOFTWARE INPUT ERROR DEVICE xxx
?INPUT DATA ERROR DEVICE

xxXx

?INPUT CHECKSUM OR PARITY ERROR DEVICE xxx
?INPUT BLOCK TOO LARGE DEVICE xxx

VERs1ON 47

June 1972

- -279-

MACRO

Chapter 5

Relocation

The MACRO-10
This
of

assembler will

program may

what

has

address

been

field

added to it.
PDP-10

loaded

previously
some

create
into

any

relocatable
part

loaded.

instructionsvmust

object

memory

accomplish
have

program.
function

this,

the

relocation

constant

This relocation cohstant, added at load time by the

Loader,

equals

the

difference

between

the

memory

location

‘an instruction is actually loaded into and the location it is
assembled

into.

program

loaded

into

cells

beginning

locationAl4008,-the’relocation constant k~would be 14008.
Not

all

instructions

Con51der

the

two

must

be modified

the

relocation

constant.

instructions:

MOVEI

2,.-3

MOVEI

2,1

The first is used in address manipulation and must be modified; the
second probably should not.
To accomplish the relocation, the
actual

expression

modification by
and

the

not modified.

VERSION 47

forming
an address

Linking
Point

Loader.

elements

(.)

evaluated
and marked

Integer
are

for

elements

are

absolute

relocatable

and

are

always

June 1972

-280-

MACRO
modified.!

Symbolic elements may
a

be either absolute or relo-

catable.

symbol

defined by

it may be

relocatable

absolute

depending

(=).

following the
it

equal

replaced by

sign

the

string

If
and

direct

symbol

the

Finally,

string

If it is defined as a label or a variable

assignment

the

defined

itself

(#),

statement,

expression
as

a macro,

evaluated.

it is relocatable.1

references to literals are relocatable.!

To evaluate the relocatability of an expression, consider what
happens

locatable

load

time.

element

and

constant,

the

must

added

expression evaluated.

each

Consider

re-

the

expression:

where A,B,C,
constant.

A+2%B-3%C

and D are

Adding

XR =

This

relocatable.

this

each

Assume

relocatable

the

term we

relocation

get:

(A+K)+2%¥(B+K)-3*¥(C+K)+(D+K)

expression may

rearranged

separate

the

k's,

yielding:

in = A+2%¥B-3%¥C+D+K

This

expression

addition of a
arranged

for

relocation

1In general,

result

J*K

The expression is

1¥K
N¥K

Finally,

suitable

single k.

the

the expression

the

addition

since

the

involves

the

expression can be re-

legal and fixed.

The expression is legal and relocatable,
Where n is any positive or negative integer
than 0 or 1, the expression is illegal.

expression

illegal.

involves

This

leads

any power

the

other

than

other

following conven-

tions:

Only

two values

relocatability

expression are allowed
element may

(e.g.,

not be divided

Two relocatable

Relocatable elements may

for a

complete

nK where n = @ or +1).
by a relocatable

element.

elements may not be multiplied together.
not

combined by

the

Boolean

operators.

lExcept under the LOC code or PHASE code which specifies absolute
addressing.

VERSsION U7

5-2

‘

JUNE 1972

-281If any of these rules is broken,

assembled code is flagged.
If A,

and

are

relocatable

A+B-C

A-C

A+2

storage

VERSION 47

symbols,

relocatable

fixed

illegal and the

then:

is relodatable
is

2&A-B

is illegai

word may

XWD

the.expression is

2%A-B

~the right half,.

MACRO

relocatable

the

left

half

well

For example:
A,B

JUNE 1972

~283-

MACRO

Chapter 6
Assembly Output

There

are

listing.

two

MACRO-10

The

llstlng

ops,

which were

6.1

ASSEMBLY

outputs,

binary-program

controlled by

described

Chapter

the

and

listing

program

control

pseudo-

LISTING

All MACRO-10 progréms begin with an implicit LIST statement.
:

Each

page

name,

the

begins

with

assembler

assembly

and

Form-Feed

page

PAGE

TITLE

line;

this

line

version,

the

time

number.

The

page

number

contains

assembly,

the

program's

date

incremented

pseudq-op.

If the code.listed requires more than one page, the basic page
number

given

the

title

line does not

change

but

is added and incremented for each additional page
6-3, etc.).
‘
|
The

line

second

line

contains

creation

date

printed

the

program

and

any

each page

fllename and

glven

the

(e.g.,

SUBTITLE

extensions,

subpage

number

6-1,

6-2,

11ne.

This

creation

time,

subtitle.

VERSION 47

June 1972
6-1

284~

MACRO

From left to right, the columns on a listing page contains:
a.

The 6-digit address of each storage word in the
These are normally sequential
binary program.
In the case of a
location counter assignments.
block statement, only the address of the first
An apostrophe followword allocated is listed.
ing the address indicates that the address is
relocatable.

The assembled instructions and data words shown
in one of several forms for easier reading (see
paragraph

2.6.3).

The source program statement,
programmer,

as written by the

followed by comments,

if any.

If an error is detected during assembly of a statement,

an error

code is printed on that statement's line, near the left edge of
If multiple errors of the same type occur in a partithe page.

cular statement, the error code is printed only once; but if several
errors,

each of a different type, occur in a statement,

code is printed for each error.

an error

The total number of errors is

printed at the end of the listing.

This
The program break is also printed at the end of the listing.
is
This
one.
is the highest relocatable location assembled, plus
the first location available for the next program or for patching.
6.2

BINARY

PROGRAM OUTPUT

The
The assembler produces binary program output in four formats.
choice depends on whether the program is relocatable or absolute,
and on the loading procedure to be used to load the program for
execution.

6.2.1

Relocatable Binary Programs - LINK Format

Most binary programs are output in LINK format.

Like the RELOC

statement, the LINK format output is implicit and is automatically _

produced for all relocatable MACRO-10 programs unless another format

(RIM, RIM10, RIM1OB) is explicitly requested.

The LINK format is

the only format that may be used with the Linking Loader.

The Linking Loader loads subprograms into memory, properly relocat-

ing each one and adjusting addresses to compensate for the relocation.

VERSION 47

JUNE 1972

-285It

also

links

tion between

external

and

internal

symbols

independently assembled

Linking Loader

loads

required

MACRO
to

provide

subprograms.

subroutines while

Mode.

communica-

Finally,

the

Library Search
'

Data

for

the

Linking

Loader

an identical forflat.
two

halves.

right half

The

formatted

left

half

count of

code

the number

for

data'words

are

grouped

sub-blocks

word

sub-block

preceded

word

consists

blocks.

the

relocation

2-bit‘bYtes.

block

data words

The

All

blocks

The first word of a LINK bloc
k consists of

Each

and

the

items.

word.

byte

type,

the

block.

Each

This

have

18-

relocation

corresponds

one

word

in the sub-bloék, and contains
relocation information regardin
g that

word.

is:

byte value

the

right

the

left half

both halves

relocation

occurs

half

are

relocated

relocated
relocated

These relocation words are not
included in the count;

appear

before

each

sub-block

relocation.

All

and

disks.

programs

paper

tape,

This

format

may

LINK Formats

Absolute

Programs

for
and

WORD

The

WORD N
VERSION 47

DECtape,

less

they always

ensure

proper

stored

LINK

magnetic

tape,

totally

independent

format,

including

punched

cards,

logical

divi-

It is also independent of the bloc
k

type.

words

sions in the input medium.

6.2.1.1

relocatable

programs

the

Block

Types

Block

Type

Relocatable

Data

location of

contiguous
less)

the

block

first data
of

program

word
or

data

the

block

words

(18

(N, from 1 to 18, must be less than 2000,000
octal)

JUNE 1972

-286-

MACRO
Block Type 2 Symbols

Consists of word pairs

1ST WORD

Bits 0-3 code bits

OND WORD
CODE @L4:
2ND WORD

Data (value or pointer)
Global (internal) definition
Bits 0-35 value of symbol
Local definition
Bits 0-35 value of symbol

symbol
1ST WORD Bits 4-35 radix 50 representation of
CODE 1@:

SND WORD
CODE 6g:

(see below)

Chained global requests:

Bits 0-17=0
of chain
pointer to first word
OND WORD Bits 18-3g5 defi
L.OADER
the
to
er
(ref
n
ntio

OND WORD

requirin

manual)

CODE 6¢:

Global symbol additive request: (refer to
the LOADER manual)

Block Type 3 Load Into High Segment

ry program, the
When block type 3 is present in a relocatable bina
the system has
Loader loads the program into the high segment ,if bloc
k type 3 apre—entrant (two-segment) capability. When used
pears immediately after the name block (type

6).

The first word is
XWD 3,,2

The second word is the relocation word
209900,
.8

The third word is

YWD HISEG BREAK,,TWOSEG ORIGIN

where twoseg origin is 400000 by default.

d word is negaWith the TWOSEG pseudo-op, the left half of the thir
pt to set a
tive. On a two-segment machine,this is ignored exce
erence is assumed to
LOADER flag. On a one-segment machine, the diff
assembler
be the maximum length of the high segment. A one-pass e
efor

does not know this legth at the start of pass 1, ther

VERSION 47

6-4

June 1972

MACRO

-287-

|
XWD L@@ped, ,4o089d

is used to signal two segments to a two-segment machine.

On a one-segment machine, this instruction gives the error message
TWO SEGMENTS ILLEGAL

since the LOADER does not know how much space to reserve for the
segment.

high

Block Type 4 Entry Block

This block contains a list of Radix 50 symbols, each of which may

contain a 0 or 1 in the high-order code bit.

Fach represents a

input is ignored until the néxt end block.

This block must be the

series of logical AND conditions. If all the globals in any series
are reqdested,'the following program is loaded. Otherwise, all
in a progranm.
first block
Block Type 5 End Block

a program. It contains two words, the
This is the last block in
first of Which'is the program break, that is, the location of the
first free register above the program. (Note: This word is relocatable.) It is the relocation constant for the following program
loaded. The second word is the highest absolute location seen (if
greater than 140); In a two-segment program, the two words are:
1)

the high segment break followed by

'2)

the low segment break.

Block Type 6 Name Block

The first word of this block is the program name RADIX 50). It
must appear before any type 2 blocks. The second word, if it
appears, defines the length of common. The left half of the second

word 1is used to describe the compiler type that produced the binary
file, 0 in the case of MACRO.

Version H47

June 1972

MACRO

-288-

Block

Type

Starting

The

first word

The

starting

means

Block

Type

Each

program.

this

address

the

data

Address

block

for

bits.

Internal

Request

word

The

right

one

request.

half

starting

relocatable

relocation

the

The

the

address

program may

left

value.

half

Either

the

program.

relocated

pointer

quantity

may

to
be

the
re-

locatable.,

6.2.2

Three

Absolute

output

binary

Binary

formats

programs.

Programs

are

These

available

are

for

requested

absolute

(non-relocatable)

the

assembled

RIM,

RIM10,

and

RIM10B

statements.

6.2.2.1

RIM1OB

locations

(not

statement

sembler

format

The

the

write

specified

relocatable),
beginning

out

designed

program

Format

the

program
a

RIM10B
the

object

use

with

punched

out

during

LOC

source

program

for

the

statement

the

statement.

the

causes

RIM10B

PDP-10

pass
If

program

into

the

format.

hardware

starting

first

absolute

following

two

This

read-in

LOC

as-

the

feature.

location

statements

the

are:

LOC 148¢)
RIM1gB )’

the

assembler

ing

1000,

starting

assembles

and

punches

location

duringassembly

ment

needed

RIM10B

out

the

word

(or

format

RIM10B
less)

The
his

out

program

program,

the

6-1

(Figure

6-2),

The

assembler

inserts

block,

also

I/0

inserts

each

but

entire

Figures

blocks,

absolute

programmer

Loader

data

VERS1ON 47

the

(see

data

and

punch

program with

out

1000.

counter

the

and

also

reset

the

only

one

RIM10B

6-2),

the

36-bit

start-

location
state-

program,

assembler

the

separated

transfer word
a

format,

may

followed

block

addresses

RIM10B

(IOWD)

checksum

punches

program
blank

preceding

following

17-

tape.

each

data
*

JUNE 1972

-289-

blogk as

shown

only

data

words

the

36~bit

Data

Figure 6-1.
in

the

added checksum of

blocks may contain

MACRO

The word count 'in

block,

and

the

therchecksum

IOWD

includes

the

simple

the‘IOWD and the data
words.

less

than

signs a non-consecufive loca
tion,

words.

the

assembler

as-

the current data block is
termi-

nated, and an IdWD containing the
next location is inserted,

starting

The

new data

transfer

(see

block

block is

block.

consists

two words.

The
from

Section 6.2.2.4)

read.

The

6.2.2.2
lute,

second

RIM10

follows a LOC
each

storage

address

RIM10

and

stop

the

RIM10

When

the

a program,

object

specified

the

LOC

arbitrary

read

the

format
RIM10

are

abso-

statement

assembler punches out
starting

the

absolute

the

format

statement.

"paper

in by

the

program,

statement

reader.

checksummed.

END

transfer block.
is

not

the

can be

Binary programs

statement

the

first word

executed when the

dummy word

word

writes

below,

and

Format

unblocked,

an'instruction-obtained

the

tape".

PDP-10

Read-In-Mode

hardware.

IOWD N,FIRST)
where

the

tion

at the

Pied.

The

length

end,
last

the program,

the program including

and FIRST

is the

location must

such

the

transfer

first memory location

be
a

transfer

instruction

instruc-

to be occuto

as:

begin

JRST 4,00)
For
at

example,
START

and

a program with

its

IOWD

last

location

RIM10
at

output

FINISH,

has

the

its

first

programmer

location

may

write

FINISH-START+1,START)

NOTE

VERSION 47

cases

but

where

binary

the

location

output

LOCn,

and

LIT

word

pseudo-ops),

into

the

tion

binary

skipped

by the

counter

occurs

(such
MACRO

output

location

file

increased

with

inserts
for

BLOCK,
a

each

counter.

6~7

Zero
loca-

June 1972

-290-

MACRO

6.2.2.3 RIM Format - This format, which is primarily used in PDP-6
systems, consists of a series of paired words. The first word of
each pair is a paper-tape read instruction giving the core

address of the second word.
DATAI

memory

The second word is the data word.

PTR,LOC

DATA WORD

an
The last pair of words is a transfer block. The first word is
instruction obtained from the END statement (see Section 6.2.2.4)
and .is executed when the transfer block is read. The second word
'
is a dummy word to stop the reader.
The loader that reads this format is:
LOC

CONO PTR,68
CONSO PTR,1#
JRST

.-1

DATAI PTR,B

CONSO PTR,1d

JRST
¢

JRST

.~-1

This loader is normally toggled into memory and started at location

20.

6.2.2.4 END Statements - When the programmer wants output in either
RIM or RIM10B format, he may insert an instruction or starting address as the first word in the two-word transfer block by writing
the instruction or address as an argument to the END statement.

The second word of the transfer block is zero.

1In RIM10 assemblies,

this argument is ignored.

If bits 0 through 8 of the instruction are zero, MACRO will insert

the instruction JRST 4,0, causing a halt when executed.

The END

statements

END SA)

END JRST SA)

will start automatically at address SA.

VERsTON U7

June 1972

Some

other

MACRO

-291examples:

l1st Transfer Block Word

END@XCT 1234
END ZU,SA
END

XCTel23k
JRST 4,SA
JRST 4,0

RIM 198
LOADER .

tOWD Xy, ADDRy

X1<€1710= NUMBER OF WORDS IN
15t DATA BLOCK
ADDRy=ADDRESS OF
1st DATA BLOCK

18t BLOCK

OF
PROGRAM DATA

CHECKSUM

IOWD IS INCLUDED
{N -CHECKSUM

IOWD Xq, ADDRp,

nth BLOCK
OF

PROGRAM DATA

CHECKSUM

//////////// BLANK TAPE (6 FRAMES)
JRST START

TRANSFER BLOCK

¢
10-0060

Figure

VERSTON 47

6-1

General

RIM10B Format

JunE 1972

MACRO

-292-

XWD

-1650

ST:

CONO

PTRs60

ST1

HRRI

A,RD+1

RD:

CONSO

PTR»10

JRST
DATAI

PTRs

@TBL1-RD+1(A)

XCT

TBL1-RD+1CA)

XCT
A

TBL2-RD+1
(A)D

S0JA

TBL1 :

CAME
ADD

CKSMs»ADR
CKSMs1

SKIPL

TBL2:

JRST

AOBJN
ADR:

JRST

(ADR)

CKSM,ADR

4,ST

ADRsRD
ST1

CKSM=ADR+1

Figure

VErRsION 47

6-2

RIM10B

Loader

June 1972

-293-

MACRO

Chapter 7
Programming Examples.

This

chapter

example

contains

(Figure

7-1)

four

examples

presents

MACRO-10

the logarithm of a complex argument.
ENTRY

statement

Logarithm

and

identifying

Function)

and

this

uses

programs.

routine

for

routine

CLOG

external routines,

(Figure 7-2)
used

which

contains

conditional

KAl0 or KI10 mode.
macros

which

simulate

assembly

the

KI10

(Complex

ATAN2

is the universal parameter file

produce

It defines

first

ALOG,

CABS.

DEF40.MAC

The

calculating

This routine begins with an

library

three

The second example
It

macro

the

KA-10

switches

the

version

select

accumulator

hardware

LIB40.

either

PDP-6,

conventions and

operations

the

KAl0

processor.

Example

its

(Figure

accumulators

macro

FLADD

which

generate

shown.

7-3)

and

uses

the

DEF40

macros

expanded

twice

text

but

(via

for

DMOVE,

show

binary.

the

The

pseudo-op)

DMOVEM and
effect

effect

for

FLADD.

LALL

SALL

The
lines

also

)

Example

desired

operation

VERSION 47

(Figure

7-4)

shows

take

nested

macros

ASCII

7-1

text

which

string

use

and

IRPC.

store

The

the

JuNe 1972

- MACRO

~294-

characters four per word, left-justified, with the character count
stored in the first nine bits of the first word.

The TEXT macro counts the string characters and invokes the CODE

macro to store the characters four per word.

last
The CODE macro invokes a SHIFT macro which left-justifies the
word if it is not already left-justified. The first part of the
example shows the normal listing, then SALL is set to show what
code the macros are generating.

VERSION 47

JUNE 1972

MACRO

-295-

0°Y
v
HOXH
axd

1880808

T1828¢

9TgZa8

10@8108

118889

T=4d

g1=9 9T=0

.ASOdHLHENLEHEJ

eTiyygip)

LOB30d

STA0YL

L9z

Nofuyayiygil 12 83081\COBPD PW238 NPB88 0\8 0 89

Figure

VERSION 47

7-1

RN} 19182084

MACRO Program CLOG

JUNE 1972

DDayDL¥gSon])Y0OgOo1vTIZo0YDoNY OYVHIOVH(ETTgO1h2,o9T10ToT0A128)09P1g0l.382808-L008P9Hd88a38g90a0d4v4.-#GfII:XXN€HHTEEGIE2T.-Hav-f DVTdOGW¢A ASdT1dYIS
MACRO

VERSION 47

-296-

June 1972

-297-

'MACRO

UNIVERSAL DEF4@ PARAMETER FILE FOR FORTRAN IV LIBRARY
SUBTTL

V32(343)

23=-NOV-T71

/TWE

IFNDEF PDP6.<IFNDEF KAl#,<IFNDEF KI1f,<KAlg==1>>>
IFNDEF PDP6,<PDP6==g>
; CONDITIONAL ASSEMBLY
IFNDEF KAl#,<KAlg==g>
IFNDEF

IFN

PARAMETERS

KI1J,<KIl@==g>

<PDP6!KA1¢!KI1@-PDP6-KA1@-KI1g>,
<PRINTX

MACHINE

PARAMETERS

DEFINED

WRONG>

; ACCUMULATOR ASSIGNMENTS

A=g
B=1

c=2

D=3

E=1
F=5

G=6

H=7

Q=16

;FOR JSA AND ARG ADDRESS FOR PUSHJ

P=17

;PUSH

DOWN

POINTER

IFE KAld,<
DEFINE

DOUBLE

(A,B)<

A
B>
>

IFN KAlf,<
DEFINE DOUBLE

451 ,==A%

(A,B)<

777000, ,0>

IFL %%1l.,<33l.==-3%3%1.-<1000,
,0>>
#%1 ,==%21.-<@#3300%,,9>
IFE

B,<%%1l.,==¢>

Z22,==B%1, +<<B+20@>«-B>&<@PBTT
,TTTTTT>
T,
IFL

Z%1l.,<%%2.==g>
A

BRE2
SUPPRESS

DEFINE
,

231, ,%%2.>
DMOVE (AC,M)<
IFL

M>-<@>,<

MOVE

AC,M

MOVE

AC+1,1+M>

IFGE

<& M><@>,<

MOVEL AC+1,M
MOVE AC, (AC+1)
MOVE AC+1,1(AC+1)>
>

DEFINE

DMOVN(AC,M)<
DMOVE

DFN

DEFINE

AC,M

AC,AC+1>

DMOVEM(AC,M)<

MOVEM AC,M

MOVEM AC+1,1+M

Figure

VERSION 47

7-2

Universal Parameter File

DEF40.MAC

JUNE 1972

_298-

MACRO
DEFINE FLMUL

(AC,M,%0V)<

MOVEM AC,AC+2
FMPR AC+2,1+M

JFCL

(2)

FMPR

AC+1,M

JFCL

(2)

UFA AC+1,AC+2
JECL
FMPL AC,M

JOV %0V
UFA AC+1,AC+2
FADL AC,AC+2

%OV
>

DEFINE FLDIV(AC,M,%0V)<
FDVL

AC,M

JOV %0V
MOVN
FMPR

AC+2,AC
AC+2,1+M

JFCL

(2)

UFA AC+1,AC+2
FDVR AC+2,M
JFCL

FADL

AC,AC+2

0V
>

DEFINE FLADD(AC,M,%0V)<
UFA AC+1,1+M
FADL AC,M

JOV %0V
UFA AC+1,AC+2
FADL

AC,AC+2

;END

Z0V:>
>

IFN KI1g,<

OPDEF
OFDEF
OPDEF

FLADD
FLMUL
FLDIV

DEFINE DFN

IFN

KA1lg

CONDITIONAL

[DFAD]
[ DFMP
]
[DFDV]

(A,B)<

A
DMOVN A,

<<A+1>&17-<B>>,<PRINTX "DMOVN A,A" CAN'T REPLACE "DFN A,B">

sEND OF KI1@ CONDITIONAL

END

VERSION 47

JunNe 1972

(“H®A)ON°WVAY (HBA)+OTH+V (9‘AO®WvA) v(HITA+OVHYW)

VERSION 47

Figure

OeN SHHYO0Y0¥dE sJnALOELEd

122880

7-3

Test
ASESRSR SRS

a|aunmw®

297 28

B020827 1

cepenz

IHVIS

TIVS

aNd

‘cpes”

TIVT

1WVYHOEdVV2A0NY08STg€d2z@PgBfP HAONA q°Yy

VLS

1.t,P92o3P3p8@a9388 1@@cz 1g98 2080881888yogddig]080 - IFAOW ‘T+V () STIIHHLWSISENdOTIQ9INX0HdEAJNATOW
q.SPP0B2I8F9P 2Bd@8g2zz 122388 po20e120028g WIACHWAAONWCFAOW (T+Y)
T
+
Y
A°Y
Y
9721QL190908028920083390y0 GTSP2ESE2TT 12IB18T9% a3l1eP8SPBEgy8BYeTy Aav1WOdaICnvvArdiO¢aHWT+gT#@°+8TqTgy°+°éVTy+Y HYOLS OL AYOWHW
STQ088 THT 288 ceeaed v1danvd22+Y+YY°T+Y ISITS HONTHIAYIAT
4gL9SHT1e1118gG8@pZPg88aEp97gDg TPTGBnEhEl2TTT 921Bg7981 4ycPo281clpapSPBeP8ey 1 nvTTvaadaodANvnvrvddiadJTcYq°Y2+V++gYTe°g:T2#cH+°VVY
TIVD¢ ANV ZMYNIG ATINO
8 29 ho e e

1% ZL=¥dv-S

IISSHHELL HKWO9DS2Y3S0O8YD9V8INg2cL-0¥8dV-GOMBDrV€HT(£T).Ly 84:€T ¢l-THH1¥AOTdCMLvOVIE-HILNSES°SVLDS(VHFT)¥LH%dd9d¢Td dWLO-S¥avSO-H9OVIW ¢HF
-299-

MACRO

Some Macros

JUNE 1972

-300-

dDVd ¢

MACRO

PB8880
12820

TP 0 N

€SHLOMSHWOHVLSHT2lO-9Eu¥YWTdAvL¥S-§
LSHL

"198@
cggg:°”

goese

VERSION 47

LHYLS

JUNE 1972

MACRO

-301-

J2XH

T7v71
<

&"“Z T+N= T+N=

PP088 W

Yy
180800
COP080i\

THIHTH €PT2PT

288889

100889
2pepse
384} oy

<ZZdXH

0dyT >0

1% 2L-4dv-4

Figure

VERSION 47

7-4

Store Text Character by Character

JuNe 1972

-502-

T+N= N T+N= N

118029

Lppaae

T+N==N
T+N==N

T+N= N

Yfuf ufi]

T+N==

MACRO

Z1oape

VErsION 47

JuNe 1972
7-10

~303-

MACRO

- 2% JXd

gyc1go o

47
VERSION

JuNE 1972

7-11

<
P
=
2
Z
N
A
I
>
‘
0
d
L
L
3
8
%
d
X
H
2
%
SPTHAT
dXH zZ7

NAI >°0dl).3%
MACRO
-304-

VERSION 47

June

7-12

1972

gOZNTT

Z1TPeg CITIT
6+2F==2%
<LJAIHS

VERSION 47

7-13

| pFRR==2F

.9S07P0800d9GTPITTZHTAITTLCPTITTOeZTT
T% cL-¥dv-§

-305-

MACRO

<LATHS

JuNE 1972

OHDVHL(tETIT)S:hT2l-Hd¥=-G39Vd€-T
1y

€

1] 4

0
=
fx]

I% cL-4dv-S

R Y]

¥ dH0 d sn

LWOFNVXYHIOLLSOSHMOJYIXHHHAOAVLVIdY4HELI0L4SOILVIMdgV2THLPO-2¥8dA@v9-6HALhOiVhHTVHO

LPERew. B ENT BeEPEB
MA(.RO

-306-

VERSION 47

7-14

June

1972

-307-

MACRO

Appendix A

Op Codes, Pseudo-Ops,

and Monitor 170 Commands

This

appendix

contains

complete

list

assembler defined

including machine instruction mnemo
nic codes,
monitor

programmed

operators,

operator,

and

FORTRAN

unimplemented

programmed

user

operation

operators.
code

uuo.

A.l

The

operators

assembler pseudo-ops,
A

called

ASSEMBLER PSEUDO;OPS AND MONITOR
COMMANDS
notes

Pseudo-ops
address

specify
that

operand

which

Pseudo-ops

generate

data

may

generate
be

used

data,

within

and

which

literals,

do
and

not.
in

fields.

The initial values given by MACRO~10
to I/0 instructions and FORTRAN

UUO's
and

are

ARRAY,
ARG,

for which
useful

pseudo~op,

320,

no-op

the
in

octal

checking

generates

(same

code

not

shown,

data

JUMP)

CALLI,

047,

monitor

UUO

070,

monitor

UUO

pseudo-op,

generates

data

ASCIZ,

pseudo-op,

COMMENT,

generates

data

DATA,

BLOCK,

pseudo-op,

BYTE,

pseudo-op,

CALL,

040,

data

data generated
generated

DEC,
DEC.,

generates data

monitor

VERSTION 47

given

CLOSE,

ASCII,

ASUPPRESS,

are

the

notes

listings.

033,

DEPHASE,

A-1l

data

generated

FORTRAN

pseudo-op,

DEFINE,

UUO

020,

UUO

generates

FORTRAN

pseudo-op,

data

UUO
no

data

generated

June 1972

MACRO
ENC.,

~-308034, FORTRAN UUO

END, pseudo-op , no data generated

RELOC, pseudo-op, no data generated
REMARK, pseudo-op, no data generated

077, monitor UUO

RENAME, 055, monitor UUO

EXP, pseudo-op, generates data

RERED., 030, FORTRAN UUO
RESET., 015, FORTRAN UUO

ENTER,

ENTRY, pseudo-op, no data generated

EXTERN, pseudo-op, no data generated
FIN.,

021,

GETSTS,

FORTRAN UUO

062, monitor UUO

HISEG, pseudo-op, no data generated
IFl, conditional pseudo-op
IF2, conditional pseudo-op
IFB, conditional pseudo-op
IFDEF, conditional pseudo-op
IFDIF, conditional pseudo-cp
IFE, conditional pseudo-op
IFG, conditional pseudo-op
IFGE, conditional pseudo-op

IFIDN, conditional pseudo-op
IFL, conditjonal pseudo-op
IFLE, conditional pseudo-op
IFN, conditional pseudo-op
IFNB, conditional pseudo-op
IFNDEF, conditional pseudo-op
IN, 056, monitor UUO
IN., 016, FORTRAN UUO
INBUF, 064, monitor UUO
IN., 026, FORTRAN UUO
INIT, 041, monitor UUO
INPUT, 066, monitor UUO

INTEGER, pseudo-op, generates data

INTERN, pseudo-op, no data generated
IOWD, pseudo-op, generates data
IRP, pseudo-op, no data generated
IRPC, pseudo-op, no data generated
1ALL, pseudo-op, no data generated
LIST, pseudo-op, no data generated
LIT, pseudo-op, generates data
1..0C, pseudo-op, no data generated
LOOKUP,

076, monitor UUO

MTAPE,

072, monitor UUO

MLOFF, pseudo-op, no data generated
MLON, pseudo-op, ho data generated

REPEAT, pseudo-op, no data generated
RIM, pseudo-op, no data generated
RIM10, pseudo-op, no data generated
RIM10OB, pseudo-op, no data generated
RTB., 022, FORTRAN UUO

SEARCH, pseudorop, no data generated

SETSTS, 060, monitor UUO

SIXBIT, pseudo-op, generates data

SLIST., 025, FORTRAN UUO
SQUOZE, same as RADIXS50

STATO, 061, monitor UUO
STATUS, 062, monitor UUO
STATZ, 063, monitor UUO

sTOPI, pseudo-op, no data generated
SUBTTL, pseudo-op, no data generated

SUPPRESS, pseudo-op, no data generated

SYN, pseudo-op, no data generated
TAPE, pseudo-op, no data generated
TITLE, pseudo-op, no data generated
TTCALL, 051, monitor UUO

TWOSEG, pseudo-op, no data generated
UGETF, 073, monitor UUO
UJEN, 100, monitor UUO

UNIVERSAL, pseudo-op, no data generated

USETI, 074, monitor UUO
USETO, 075, monitor UUO

VAR, pseudo-op, generates data
WIB., 023, FORTRAN UUO

XALL, pseudo-op, no data generated
XLIST, pseudo-op, no data generated
XWD, pseudo-op, generates data
7, pseudo-op, generates data

.CREF, pseudo-op, no data generated
.XCREF, pseudo-op, no data generated
.HWFRMT, pseudo-op, no data generated
.MFRMT, pseudo-op, no data generated

MTOP., 024, FORTRAN UUO
NLI., 031, FORTRAN UUO
NLO., 032, FORTRAN UUO

NOSYM, pseudo-op, no data generated
OCT, pseudo-op, generates data

OPDEF, pseudo-op, no data generated
OPEN, 050, monitor UUO
ouT, 057, monitor UUO
ouT., 017, FORTRAN UUO
OUTBUF, 065, monitor UUO
QUTF.,

QUTPUT,

027,

FORTRAN UUO

067, monitor UUO

PAGE, pseudo-op, no data generated
PASS2, pseudo-op, no data generated

PHASE, pseudo-op, no data generated

POINT, pseudo-op, generates data
PRINTX, pseudo-op, no data generated

PURGE, pseudo-op, no data generated
RADIX, pseudo-op, no data generated
RADIX50, pseudo-op, generates data
RELEAS, 071, monitor UUO

VERSION 47

JUNE 1972

~309-

A. 2

MACHINE MNEMONICS

The

following

MACRO

AND OCTAL CODES

are machine mnemon
ics

and

corresponding oct
al

codes:

ADD

270

ADDB

CAMGE

273

315

CAML

FSBRI

165

313

156

572

CAMLE

FSBRM

HRREM

271

311

HRRES

HLL

HRRM

MOovMm

400

254-4,

541

CLEAR

HALT
-

HRRI

MOVES

316

132

573

CAMN

FSC ..

500

542

HRRO

MOvMI

560

215

MOvMM

SETOB

477

216

TLN

SETOI

601

475

TLNA

605

ADDI|
ADDM

272

AND

404

MOVEM

202
214

SETMI

415

SETMM

TLCA

645

416

SETO

TLCE

643

474

TLCN

647

ANDB

407

ANDCA

CLEARB

410

403

HLLE

476

400

SETZB

607

7-20

533

211

TLNN

HLLES

MOVN!I

SETZ

603

563

210

TLNE

HRROS

MOVN

7-24

532

562

SETOM

HLLEM

HRROM

217

402

631

MOVMS

CLEARM

HLLE!

HRROI

413

401

530

ANDCAB

CLEARI

HLLI

501

ANDCAI

411

ANDCAM

CONI

412

CONO

ANDCB

440

ANDCBB

CONSO

443

7-34

CONSz

HLLM

7-30

502

HLLO

520

561

HRRS

543

MOVNM

TLO

MOVNS.

SETZI

661

550

212

403

HRRZ

401

213

TLOA

SETZM

665

402

TLOE

663

HRRZ!

551

MOovs

562

204

- MOvsi

SKIP

205

330

667

MOvsm

SKIPA

TLON

206

334

SKIPE

332

TLZA

SKIPG

224

337

625

SKIPGE

TLZE

335

623

TLZN

627

HRRzZM

ANDCB}

441

ANDCBM

DATAI

442

7-04

HLLOI

DFN

HLLOM

HRRZS

420

7-14

521

ANDCM

DATAOQ

131

522

HLLOS

18P

523

133

IDIV

MOvVSS

230

MuUL

ANDCMB

423

Div

234

HLLS

ANDCM!

421

ANDCMM

Divs

422

237

Divi

HLLZ

235

HLLZ!

+603

511

1DIvB

233

iDIVI

231

MuLs

227

MuULI

SKIPL

MULM

SKIPLE

226

333

640

232

225

TRC

IDIVM

SKIPN

TRCA

336

644

TRCE

AND/{

405

ANDM

DIVM

206

236

DPB

HLLZM

137

b12

HLLZS

1DPB

513

136

ILDB

134

AOQOBJN

253

AQBJP

EQV

252

444

HLR

AQJ

EQvVB

340

447

HLRE

AQJA

EQvV)

344

445

AQJE

EQVM

HLRE!

342

EXCH

250

HLRES

AQJG

347

FAD

AQJLE

FADB
FADL

HLRI

AOQJL

345
341

140

343

FADM

346

FADR

AOQOJGE

AOJN
AQS

350

AQSA

364

FADRI

352

AOSG

FADRM

357

AOSGE

FDV

355

FDVB

AOSL

351

AQOSLE

FDVL

353

AOSN

FDVM

366

ASH

FDVR

240

ASHC

FDVRB

244

FDVRI

BLKI

7-00

BLKO

FDVRM

7-10

FMP

BLT

* 251

CAl

FMPB

300

FMPL

304

FMPM

CAIA
CAIE

302

HLREM

HLAM
HLRO

HLRO!

144

FADRB

AQSE

CAIG

307

FMPRB

305

FMPRI

CAIL

301

FMPRM

CAILE

303

FSB

HLROM
HLROS

HLRS

HLRZ

HLR2I
HLRZM

HLRZS
HRL

HRLE
HRLE!
HRLEM
HRLES

HRLI
HRLM

HRLO
HRLOI

FMPR

CAIGE

HRLOM
HRLOS

HRLS
HRLZ
150

HRLZI

CAIN

306

CAM

FSBB

310

163

FSBL

HRLZM

CAMA

314

161

FSBM

HRLZS

1652

HRR

CAME

312

CAMG

FSBR

317

154

FSBRB

HRRE

157

HRREI

VERSION 47

644

577

621

331

642

434

ORB

S0J

437

360

SOJA

TRCN

364

646

TRN

600

IMUL

220

IMULB

ORCA

223

IMULI

221

ORCAB

457

SOJG

ORCB

470

361

606

434

SOJL

TRNN

I0R

456

365

602

ORCAM

SOJGE

TRNE

222

455

367

IMULM

ORCAI

SOJLE

363

TROA

664

SOJE

TRNA

660

IORB

A37

I0RI.

ORCBB

435

473

SOJN

TROE

436

471

366

I0ORM

ORCBI

ORCBM

S0s

472

370

TRON

SOSA

374

666

SOSE

TRZ
-

372

620

TRZA

SOSG

377

624

TRZE

622

JCRY

.255-6,

JCRYO

ORCM

256-4,

ORCMB

467

ORCMI

465

JCRY1

255-2,

JEN

264-12,

JFCL
JFFO

243

JFOV

255-1,

Jov

ORCMM
ORI

QRM
POP

255-10,

POPJ

JRA

JRST

254

PUSH.

PUSHJ

JRSTF

254-2,

JSA

ROT

266

ROTC

Jsp

265

RSwW

JSR

264

Jump

320

SOSGE

376

SOSL

TRZN

626

SOSLE

TSC

373

651

TSCA

376

TSCE

274

653

TSCN

657

SUBB

277

SUBI

TSN

611

275

TSNA

615

276

‘TDC

TSNE

650

613

TDCA

TSNN

654

617

TSO

671

TDCE

662

TSOA

656

TSOE

673

610

TSON

677

182

612

631

TSZA

635

SETA

424

SETAB

TDCN

427

TDN

JUMPA

324

JUMPE

SETAI

322

425

SETAM

TONA

426

TDNE

JUMPG

327

JUMPGE

325

JUMPL

321

675

SETCA

450

SETCAB

TDNN

453

616

TDO

TSZE

451

670

633

SETCAI

TDOA

TSZN

674

637

UFA

130

256

JUMPLE

323

JUMPN

SETCAM

326

452

SETCM

TDOE

460

672

TDON

XCT

676

XOR

LDB

655

SOSN
suB

SUBM

7-04

662

135

LSH

SETCMB

242

463

SETCMI

TDZ

433

SETCMM

TDZA

XORB

246

461

630

LSHC

462

634

TDZE

XORI

431

632

XORM

432

MOVE

200

MOVEI

201

SETM

414

SETMB

TDZN

417

636

TLC

641

JUNE 1972

-311-

MACRO

Appendix B

Summary of Pseudo-Ops

B.l1l

PSEUDO-OPS

A list of pseudo-ops and_their'functions follows:
ARRAY

Reserve

ASCIT

Seven-bit

ASCIZ

multiple

ASCII

storage.

test

Seven-bit ASCII
at

words

test,

end

with

null

character

guaranteed

ASUPPRESS

Turns

BLOCK

Reserves

BYTE

Input

COMMENT

No binéry‘produced; éame as seven-bit ASCII

DEC

Input

DEFINE

Defines

DEPHASE

Terminates

END

Last

ENTRY

Entry

point

EXP

Input

expressions

EXTERN

Idenfifies éxtérnal symbols

VERSION
47

suppress

bit

block

storage

bytes

decimal

length

for

all

symbols

cells

1-36

bits

numbers

macro
PHASE

statement

relocation mode

for

the

program

subroutine

B'fl

library

June 1972

~312-

MACRO
HISEG

Load

INTEGER

Reserve

INTERN

Define

internal

IOWD

Set

I/0

IRP

Indefinite

repeat

macro

IRPC

Indefinite

repeat

one

LALL

List

all;

LIST

List

LIT

Assemble

LOC

Assign

MLOFF

Turn

off multiline

MIL.ON

Turn

on multiline

NOSYM

Suppress

ocT

Input

OPDEF

into

high

one

segment

word of

storage

argument

symbols

transfer

expanded

normal

per

word
arguments

character

listing

macros

mode

literals

absolute

symbol

octal

addresses

literal
literal

table

feature
feature

listing

numbers

Defines user-created operator;

generates only one

word
PAGE

Start

a new

PASS?2

Terminates
cessed

listing page
pass

pass

remaining

statement are pro-

only

PHASE

Following coding relocated at execution time

POINT

Sets

PRGEND

Allows multiprogram assemblies,

PRINTX

Output on terminal or

byte

pointer word

end one

such program

listing device the rest of the

line
PURGE

Remove

RADIX

Sets

RADIX5Y

Compresses

RELOC

Implied first

statement;

REMARK

Comments

only

statement

REPEAT

Repeat

times

RIM

Prepare

RIM1g

Absolute,

unblocked,

RIM12B

Absolute,

blocked,

VERSION 47

symbol

from

prevailing

table

radix

36-bit words,

output

2-10

primarily
assigns

RIM paper-tape

output

for

system use

relocatable addresses

format

format;

checksummed output

checksums
format

JUNE 1972

MACRO

-313-

Suppress listing of macros;

SALL

binary

generated

lists only call and

Opens

symbol

SIXBIT

Input

text

SQUOZE

Same

STOPI

Stop indefiinite repeat of macro arguments

SUBTTL

Subtitle on listing

SUPPRESS

Turns on suppress bit fpr épecified symbols

SYN

Make

synonymous

TAPE

. Stop

processing

tables

Title

TWOSEG-

Assembles

UNIVERSAL

Makes

VAR

Assemble

and

6-bit

program

ASCII

above

the

listing

symbol

universal

compressed

RADIX

TITLE

current

and

loads
table

file

DDT

two

segment

available

variables

programs

other programs

suffixed with

ARRAY

INTEGER

XALL

Stop

expanded

XLIST

Stop

listing

XPURGE

Purges

XWD

Input

two

18-bit

Input

zero

word

.CREF

Resume

« XCREF

Stop

output

CREF

. HWFRMT

List

binary

half word

. MEFRMT

List binary

multi-format

B.l.1

These

Conditional

conditional

sembled

the

local

conditions

symbols

outout

Assembly

assembly

listing,

resume

pass

words

CREF

information

format

statements

the

second

the

(new)

first

IF2

Encountered

pass

IFB

Blank

IFDEF

Defined.

iFDIF

Different

during

column

are

as-

column exist.
pass

(old)

Statements

Encountered during

IFE

list mode

half

IF1

VErRsION 47

normal

Zero
|

June 1972

MACRO

-314IrG

Positive

IFGE

Zero,

IFIDN

Identical

IFL

Negative

IFLE

Zero,

IFN

Non-zero

TFNB

Not

blank

IFNDEF

Not

defined

VERSION 47

positive

negative

June 1972

-315-

MACRO

Appendix C

Summary of Character

Interpretations

The

characters

indicated.
appear

listed below have

These
text

special meaning

interpretations

strings,

Character

not apply when

Immediately

labels.

Semicolon.

all

LABEL:

comments.

Point.
Has current value of the
location counter or indicates floating point number.
Comma.

General

References
comma

Delimits
Inclusive
AND

operand

argument

Accumulator

VERSION 47

these characters

Example-

follows

Precedes

delimiter.

contexts

comments.

Meaning

Colon.

the

:
field

delimiter.

accumulator

optional.
macro

The

sTHIS

COMMENT

JRST

.+5

FIVE

LOCATIONS

JUMP

FORWARD

1.9
DEC

14,5,6

EXP

A+B,C-D

MOVEI

1,TAG

MOVEI

,TAG

MACRO

(A,B,C)

arguments.

Logical Operators

June 1972

-316-

MACRO

/
+

Example

Meaning

Chatacter

Multiplication

Arith-

Division

metic

Operators

(+A outputs the value of A)

Add

Subtract

lst character of text
string
B

In ASCII, ASCIZ and SIXBIT comment

text strings, the first non-blank
character is the delimiter.
Follows number to be shifted and
precedes binary shift count.
Exponent.
ponent in

()

ASCII/STRING/;

Precedes decimal exfloating-point numbers.

Enclose index fields.

Parentheses.

Enclose the byte
BYTE

size

7B2

F22.1E5 EXPONENT
IS

MOVEI A, (SIXBIT/ABC/)
BYTE

statements.

(7)

ADD AC1,X

(6)

DEFINE MAC(A,B,C)

Enclose the dummy argument
string in macro DEFINE
statements.

In an expression,
Angle brackets.
enclose a numeric gquantity.
In conditional assembly

state-

ments, contain a single argument, and the conditional

<A-B+5¢@/C>

IFl,

TAX

MOVE AC@,

coding.
In

REPEAT

statements,

con-

tain coding to be repeated.
enclose the macro

In macros,

Square brackets.

Delimit literals.

it
i

In OPDEF statement, contain
new operator; in ARRAY the size.
Equal sign.

Direct assignment.

Equal sign.

Direct assignment but

output

DEFINE

PUNCH

DATAQO PTP,

definitinn.

<SUB

REPEAT

DDT.

Direct asEqual sign and colon.
signment but automatically made

17,

TAG>

PUNBUF

ADD 5,[MOVEI 3,TAX]

OPDEF CAL [MOVE]
ARRAY FOO[212]
SYM=6
SYM-A+B¥D

SYM==6
FLAG=:200

internal.

Colon and exclamation point.
assignment of

DDT,

VErRs1ON 47

label,

Direct

no output

!
LABEL:

and automatically made internal.

June 1972

(4)

MACRO

-317-

il
Il

Character

Meaning
Equal

sign

ment,

and

ically made
Double

Direct
put

colon.

output

Direct

DDT,

and

assign-

colon and exclamation point.

assignment of
DDT,

and

label,

NAME:
:!

no out-

automatically
made

Quotation marks

enclose 7-bit ASCII
text, right justified, from one to
five characters.
‘

"ABCDE"

Single quotation marks enclose 6bit ASCII text, right justified,

'"TABLES'

from one

Number
as

six

sign,

tag.

characters.

Defines

symbol

used

ADD

ex-

MOVE @ ,JOBREL##

generating

symbols.

Apostrophe or single quote.

Concate-

nation character, used within macro
definitions or SIXBIT data.
Reverse

first
a

slash.

used

character of

macro

call,

the

value

the

Control

left

ation.
Left arrow.
right) M bit

arrow.

Line

DEFINE MAC

(A,B,C);

<JUMP'A

IF A=5g@g,

GENERATES

fol-

ASCII

THREE

continu-

left

(or

Indicates indirect addressing.
the indirect bit in an instruction to be set.

100+3=1004

1p@«+3=10

MOV AC,@ADDR

JuNE 1972

THIS

T7-BIT

CHARACTERS,

ASCII/528/

:
N M shift N
positions,

'SIXBIT!

MAC \ A

the

an argument

lowing symbol is converted to an
ASCII symbol in the current radix.

Causes

3,TAG#

Variable.

Alternate method
ternal

VERSION U7

LOOP==:32

automat-

internal.

Example

-319-

MACRO

Appendix D

Storage Allocation

MACRO

allocates

the

All

the

and

Most

symbols

symbol

half

sented

either

the

symbol).

The

and

the

top

etc.,

table

SIXBIT.

value

less

words

-1)

two

low

segment

grow upward

two

words

the

right

free

1dng.

(.JBREL)

The

flags

bits

symbol

and

table.

can

valfie

stored

this

value

stored

the

free

storage,

the

sixbit

name

word

left

the

repre-

symbol

first

symbol.

with

word

last referencé_in'a chain of refer
ences

second

first

Symbols

the

words

space

half.

so that additive global fixups can
be output.

VERSION 47

from

second word

than

the

have

pointer

have

are

pointer

two

symbols and macro names)

The

(e.g.,

symbols

value (i.e.,

directions:

(user

symbol

value

have

just

storage

External

the

two

from

literals,

name

36—bit va%ue

free

downward

Macros,

entries

symbol table

grows

storage

table.

the.

the

This

JUNE 1972

-320-

MACRO .

Opdefs tend to have 36-bit values and are stored like other 36bit value symbols.

Macro names are stored in the symbol table, the value is a point
to the stored text string.

The text string is stored in four (assembly parameter) word block

which have the general form

1link to next block, [¢ if last] ,, 2 characters

3)
4)

5 characters
5 characters

5 characters

However, the first such block is special

2)
3)

1link to next block

link to last block

ference
pointer to default arg; ,. <number or args expected><+9+re
count
cters
chara
5
5 characters

The number of args expected is the number of arguments in the defin
statement.

The reference count is incremented when the macro is called and
decremented when exiting from the macro.

When this count goes to

zero the macro is removed from free space.

block,
The actual arguments to a macro are stored in the same linked
are also
pbut are not in the symbol table. Repeats (2 or more times)
macro
stored the same way. The text blocks are removed when the
to
gone
has
exits or the repeat exits since the reference count
zero.

own

The addresses of the actual argument blocks are stored in a pushd
stack in order of generation.

free Default arguments are stored the same way except the list is in
left
core. The pointer to this default arg list is stored in the
.
half of the second word of the first block of the macro definition
ents are re-

The text body is stored as is, except that dummy argum

placed by special symbols.

VERsION 47

JuNe 1972

MACRO

-321-

The ASCII character RUBOUT (177) is used to signal a special character

text.

These

characters

are

291
gg2

293
a4

;end of macro
;end of dummy symbol

;end of Repeat
send of IRP or IRPC

If the character is 4<ch<77 it is illegal.

If the character is <100 then it is a dummy symbbl, the value of

the character is ANDed with 37 to get the dummy symbol number and
the corresponding pointer retrieved from the stack of pointers.
If th- symbol was not specified (i.e., no pointer) then if the 40
bit is on this is to be a created symbol and one is created, otherwise the argument is ignored.

Verbose macros can eat up a lot of storage space.

Literals are stored in four words/block per word generated (three
words if old format used).

Words

are

-3

-2:

-1:.
g

form word

relocation bits

code
pointer

to next

The pointer points to the g word of the next block.

the generated code.

The code is

Relocation is either the relocation bits f#

or 1 per half word or external pointers if externs used.

Form word is the word used for listing, this word is not checked
when comparing literals so that different forms that produce the
same code are classed as equal.

Long literals are both slow and take up extra storage, they should
be written as subroutines or inline.

Single quotes can also be used to indicate SIXBIT words, however,

one pair of single quotes is removed by the assembler if the pair
encloses a dummy argument.

VERSION 47

For example, in the macro
D-3

JuNe 1972

MACRO

B
A,

~322-

not

DEFINE

SXBT

MOVSI

1,"A"

MOVSI

2,"B"

dummy

however,

quotes

since

assembler.

the

VErRsION 47

one

argument
dummy
pair

(A)<

can

argument

and

must

quotes

(the

enclosed
be

inner

single quotes.

enclosed
pair)

will

double

removed

June 1972

~323-

MACRO

Appendix E

Text Codes

This appefidix contains a summary of MACRO=-10 text codes.
SIXBIT

Character

ASCII
7Bit*

ASCII
7 Bit*

Character

ASCII
7Bt

Space

040

041

100

140

101

042

141

102

142

044

104

144

043

SIXBIT

Character

103

143

045

105

145

046

106

146

047

107

147

(

050

110

150

)

051

151
152

052

112

053

113

153

054

114

154

055

115

155

056

116

156

057

117

)

157
160

060

120

]

061

121

161

062

122

162

23
24

3
4

063
064

63
64

S
T

123
124

s
t

163
164

065

125

165

066

126

166

067

127

167

170

070

130

071

131

171

072

132

172

;

073

[

133

173

074

134

174

075

]

135

}

175

076

136

176

077

-—

137

Delete

177

*MACRO=-10 also accepts five of the 32 control codes in 7-bit ASCII:
Horizontal Tab

Line Feed

VERSION 47

011

012

Vertical Tab

Form Feed

013

014

Carriace Return

arriag

015

JuNe 1972

-325-

MACRO

Appendix F

Radix 50 Representation

Radix

into

508

representation

bits.

scending

order

used

Each

character

from

left

condense

right;

code

for

6-character

symbols

subscripted

de-

symbols

the

symbol

i.e.,

the

’

radix

are

form

LLLLLL

45321

generated by

denotes

the

octal

the

50 8

representation

following

((C((C
#5040,
a*
) 50
50%C )
)+
#5040,
C5
) *5p+C
)
where

The

all

code

numbers

are

octal.

corresponding

Code

the

characters

(Octal)

Characters

00
01-12

Null

13-44

46
47
top

octal

four

number

VERSION 47

bltS

character
0-9

A-Z

The

are:

S
%
are

(i.e.,

taken

from

the

four

leftmost

bits

6- bit

g4-74).

F-1

June 1972

MACRO

~327-

Appendlx G

Summary of Ruies for |

Defmlng and Calling Macros

G.1

ASSEMBLER INTERPRETATTION

MACRO-lO/assembles‘macros by direct and immediate character substitu-

tions.

When a macro call is encountered, _in any field, the character

substitution is ‘made,
continues its

last

the characters are processed,

scan w1th

argument,

the

except when

character

can appear any number
of times on a

G.2
G.2.1

following

delimited by

the

and the assembler
delimiter

semicolon.

the

Macros

line.

CHARACTER HANDLING
Blanks

A macro symbol is delimited by one blank or onefitab; the character
following the delimiteis
r the start of the argument string even if
it is also a blank or tab.
Other than the first delimiter, blanks
and tabs are treated as standard characters in the argument string.
G.2.2

Angle

Brackets

brackets

are only

significant

the

argument

fields

first character of any field is a left angle bracket.

VErRs1ON 47

the

1In this case,

JUNE 1972
G-1

-328-

MACRO

no terminator or parenthesis tests are made between the left angle
bracket and its matching right bracket.

The matching brackets are

removed from the string but the scan continues until a standard
delimiter

G.2.3

found.

Parentheses

Parentheses serve only to terminate an argument scan.

They are

significant only when the first character following the blank or
1In this case, the left parentab delimiter is a left parenthesis.
thesis is removed and, if it matching right parenthesis is encoun-

tered prior to the normal termination of theargument scan, it is
removed and the

G.2.4

scan discontinued.

Commas

When a comma is encountered in an argument scan, it acts as the
delimiter of the current argument. If it delimits the last argument,
the character following it will be the first scanned after the subsitution is processed.

G.2.5

Semicolons

When a semicolon is encountered in an argument scan, the scan is
If an argument has not been satisfied, the remainder
discontinued.
It is saved, however, and will be the first
is considered null.
character scanned after the substitution is made, normally acting
as

a comment

G.2.6

flag.

Carriage

Return

A carriage return, except when pre-empted by angle brackets (see
Section G.2.2), will terminate the scan similar to the semicolon.
This can be circumvented, if desired, by the control left arrow key
described elsewhere.

G.2.7

Back-Slash

If the first character of any argument is a back-slash, it must be
directly followed by a numeric term., The value of the numeric term
is broken down into a string of ASCII digits of the current radix,

just the reverse of a fixed-point number computation.

VERSION 47

The value is

June 1972

-329-

MACRO

considered to be a 36-bit positive number having a value of 0 to

777777 777777. Leading zeros are suppressed except in the case of 0,
in which case the result is one ASCII 0. The ASCII string is substituted and the scan continued in the normal manner (no implied

.
terminator)

The default listing mode ‘is XALL, in which case the initial macro
call and all lines within its range that produce binary code are
listed. The pseudo-op LALL will cause all lines to be listed.

Substituted arguments are bracketed by *'s by the assembler.

G.3

CONCATENATION

The rule for concatenation is as follows:

For each string of apostrophes, one is removed if and only if it is
next to (either before or after) a dummy argument to that macro.

VERSION 47

June 1972

-331-

MACRO

Appendix H

Operating Instructions

H.1

The

REQUIREMENTS

following

are

MACRO-10

Minimum Core
Additional

operating

7K pure plus

Core

from

the

input device
output

and

is
Reference
DIS

additional

timesharing

two

output

The

impure

One

output

H.2

Automatically requests
ments

Equipment

requirements:

(source

devices

program

(machine

core

monitor

assign-~
needed.

input);

language

program

listing output).
If the listing
to be used as input to the Cross
(CREF) program, it must not be TTY,

or
' LPT.

INITIALIZATION

following are

VERSION 47

commands

and

corresponding

indications:

R MACRO)

Loads the MACRO-10 Assembler into core..

The

Assembler

ready

receive

command.

June 1972

-332-

MACRO
H.3

COMMANDS

H.3.1

General

MACRO-10

Command Format

general commands

are

follows:

objprog-dev:filename.ext,list~dev:filename.ext source-dev:filename.ext,...... source-n)
objprog-dev:

The device on which the object program is to be written.
MTAn:
DTAn:
PTP:
DSK:

list~-dev:

(magnetic tape)
(DECtape)
(paper-tape punch)
(disk)

The device on which the assembly listing is to
be written.

MTAn:

(magnetic

DSK:

(disk)

TTY:

(Teletype)

of these if

input to CREF!

(paper-tape punch)

PTP:

from which the source-program

The device(s)
input

Must be one

(line printer)

LPT:

source-dev:

tape)

(DECtape)

DTAn:

to assembly

MTAnN:
CDR:

to be

read.

(magnetic tape)
(card reader)
(DECtape)

DTAn:
DSK:

(disk)

(paper—-tape reader)

PTR:

TTY:

(Teletype)

If more than one file is to be assembled from a
magnetic tape, card reader, or paper tape reader,
dev: is followed by a comma for each file beyond
the

first.

Input via the Teletype is terminated by typing
CTRL Z (4%Z) to enter pass 1l; the entries must
be retyped at the beginning of pass 2.

filename.ext
(DSK:

and DTAn:

The filename and filename extension of the object

only)program file,

the listing file,

and the source

file(s).

The object program and listing devices are
separated from the source device by the left
arrow

H.3.2

symbol.

Disk File Command Format

MACRO-10

disk file commands

DSK:filename.ext

are as

follows:

[proj,prog]

11f /C switch is given, but no list-dev:

Version 47

He2

is specified, DSK:CREF.CRF is assumed.

June 1972

MACRO

-333-

the d@sk
Project-programmer number assignede to
file(s) if

[proj,prog]

area to be searched for the sourc
other than the user's project-programmer number.
The installation standard protection is assigned
to any disk file specified as output.
NOTE

If object coding output is not desired (e.g., a program is
being scanned for source language errors), objprog-dev: is
omitted. TIf an assembly listing is not desired, list~dev:
is omitted. If device is not specified, DSK is assumed.

Examples:

.R MACRO)

¥DTA3:0BJPRG,LPT: CDR )

Assemble one source program file from the card
reader; write the object code on DTA3 and call

the file OBJPRG; write the assembly listing on
the line printer.

‘

END OF PASS 1)

The source program cards must be manually re-

22 ERRORS DETECTED)
PROGRAM BREAK IS ##2537)

Number of source errors; size of object pro-

¥ C)

Return to the monitor.

.R MACRO)

Assemble the next three source files located

NO ERRORS DETECTED)
PROGRAM BREAK IS @#3552)

MTA2 for later printing.

2K CORE USED)

FMTA3: ,MTA2: MTAl:, )
—

fed for pass 2.

gram; core used by assembler.

at the present position of MTAl; write the
object program on MTA3; write the listing on

2K CORE USED)

¥ LPT: DTAl:FILE1l,FILE2,FILE5) Assemble the source files named FILEl, FILEZ2,
and FILES5 from DTAl; produce no object coding;
— NO ERRORS DETECTED)
PROGRAM BREAK IS #@1427)
2K CORE USED)

____f_,*—DSK:FILEl.MAC[lM,lZ])
NO ERRORS DETECTED)
PROGRAM BREAK IS @@@5hi

2K CORE USED)

{

write the listing on the line printer.
.

Scan the source program called FILE1l.MAC,
located in area 14, 12 on the disk, for source
language errors; produce no object coding or

assembly listing; print all error diagnostics

on the terminal.

Return to the monitor.

f_+C>
.R MACRO

¥MTAL: ,2ITTY: TTY :)
:

JMP

AOS

JFCL)

END)

VErs1oN 47

JMP

the object code program on MTAl and print the

assembly listing on the terminal.
Terminate input.

‘l‘Z)
END OF PASS

Assemble a source file from the terminal; write

Reenter terminal input.

Type first statement again.

He3

June 1972

MACRO

-334-

.MAIN

MACRO

poA808

AOS

18:14

BIBOIE

BELIBL!

JMP

208881
G:

2@-DEC-67

PAGE1)

Page heading.

First assembled.

350008

@epdp2’

A0S

Reenter second.

Second agssembled,

JFCL)

Reenter third.

geges2

255200

200099

JFCL)

Third

END))

assembled.

Reenter

END)

fourth.

Fourth assembled.

e
s
e

ERROR DETECTED)

PROGRAM

BREAK

.MAIN

MACRO

gagea2"y)

SYMBOL

Typeout of symbol

gg#gs3)

table.

14:14

2@-DEC-67

PAGE2)

TABLE)

200891')

L_2K CORE USED)

21C)
H.4

Return to the monitor.

SWITCHES

(NI © PREN 0 B

NI o)

Switches

are

used

Magnetic
Macro

tape

call

Listing
Pushdown

specify

options

as:

control

expansion

suppression
list

expansion

Cross-reference

file

All

switches

are

preceded

and

usually

occur

prior

VERSION 47

such

output.

the

slash

(/)

left

arrow

enclosed

(see

Table

parentheses,

H-1).

JUNE 1972

-335-

MACRO

Table H-1

MACRO-10 Switch Options
Switch

Meaning

Advance magnetic tape reel by one file.

Backspace magnetic tape reel by one file.

Produce listing file in -a format acceptable as input to CREF; unless the file
is named,
.CRF 1is
appear

CREF.

CRF is

assigned;
between

assigned as

the

filename;

if no list-dev:
is specified,

the

conmma

List macro expafisibns

and

the

if no extension is given,

DSK:

assumed.

/C must

left-arrow.

(same function as LALL pseudo—oé).

New format for output binary listing (.MFRMT pseudo-op)
.
0ld format for output binary listing

Print Help text

(i.e.,

(.HWFRMT pseudo-op) .

this list of switches and explanations).

Reinstate listing‘(used after list suppression by either the XLIST

pseudo-op or 5

switch).

List only call, no binary, in macro expansion (same .SALL pseudo-op).
Suppress

Sets

error printouts

on the

terminal.

the pseudo-op MLOFF which allows

literals

tb occupy on

line.

a single

This means literals may be terminated with a carriage return,
line feed instead of a right bracket.

Increase

the

size

of .the pushdown list.

many times

locations;

each /P

desired

(pushdown

increases

its

list

This

switch may appear

initially set to a size of 8010

size by 8010).

left of the left arrow.

/P must appear on thé

Suppress Q (queétionable) error indications on the listing; Q messages
indicate assumptions made during pass

of the

/Q must appear on the left

left-arrow.

Suppress

listing

(same

function as

XLIST pseudo-op).

Skip to the logiéal end of the magnetic tape.
Rewind

the magnetic tape.

Suppress all macro expansions (same function as XALL pseudo-op).
Zero the DECtape directory.

NOTE

Switches A through C and T, W, X, and 2 must immediately follow tle device or file to which the
individual switch refers.

VERSION 47

JUuNe 1972

-336-

MACRO
Examples:

.R MA(H%O)

¥MTALl:,DTA3: ,/C«PTR:)

END OF PASS 1)

bssemble one source file from the paper tape

reader; write the object code on MTAl; write

the assembly listing on DTA3 in crossreference format and call the file CREF.CRF.

The paper tape must be re-fed by the operator
for pass

—9?3 ERRORS DETECTED)
PROGRAM BREAK IS @@g4gl)

End-of-assembly messages.

__ 2K CORE USED)

¥DTA2:ASSEMB.ONE/Z,LPT:
MTAL:/W,)

NO ERRORS DETECTED)
PROGRAM BREAK IS #@5231)
__3K CORE USED)

*MTALl:/W,LPT:«MTA3:

/W, (AA),(BB))

21 ERROR DETECTED,)

Rewind MTA4 and assemble the first two source
files on it; write the object code on DTAZ2,
after zeroing the directory, and call the file

ASSEM.ONE; write the assembly listing on the
line printer.

Rewind MTAl and MTA3 and assemble files 1, 4,

and 3 (in that order) from MTA3; print the

assembly listing on the line printer; write
the object code on MTAL.

PROGRAM BREAK IS p@@655)
__2K CORE USED)

¥F00,/C FOO)
NO ERRORS DETECTED)
PROGRAM BREAK IS @@@765)

Assemble source file FOO on DSK:; write the
assembly listing on DSK in cross-reference

2K CORE USED)

ject code on DSK calling it FOO.REL.

¥4C)

Return to the monitor.

VERSTON 47

format calling the file CREF.CRF.

Write ob-

June 1972