Digital PDFs

MISC-6841C3C9

November 1983

100 pages

Original

32MB

Document:	Workstation Graphics Architecture
Order Number:	MISC-6841C3C9
Revision:
Pages:	100
Original Filename:

OCR Text

Page 1

Graphics .Architecture - PRELIMINARY

I
I
I
I
I
I

Title:

WORKSTATION GRAPHICS ARCHITECTURE

Author:

HENRY M. LEVY

Date:

7-November-198:3

Version:

l. 2

File:

Dizzy::User$l:[Aurenz.wga.mem.wga]WGA.MEM

I
I
I

ABSTRACT

This

do.cument describes, the software
be·tween
a · family
of
works.t a t ion display systems and ;a host
computer.
It
defines
commands to
generate graphics and text, and commands
to support keyboard, tablet, arid mouse
interaction.
These
commands
are
transmitted
by the host driver ' and
executed
by
a
microprocessor-based
controller in the workstation.
:

inter ~ace

Revision:
0.0-0.4

Description
Prelim. Draft
and Revisions

Author

Date

H. Levy

13-March-1982 to
10-January-1983

....

l. 0

·VSlOO Release
Base Level

H. Levy

l-March-1983

1.1

Updates based on
,
vs100 implemen. t"6.~!~:fr1

L. Samberg

20-Jline-1983

l. 2

Finalized Release for
VSlOO implementation

s. Aurenz

7-January-1984

PRELIMINARY

CONFIDEr+ltAlt

Graphics Architecture - PRELIMINARY

Page 2

COPYRIGHT (c) 1983 BY
DIGITAL EQUIPMENT CORPORATION, MAYNARD, MASS.
THIS SOFTWARE IS FURNISHED UNDER A LICENSE AND MAY BE USED AND COPIED
ONLY IN ACCORDANCE WITH THE TERMS OF SUCH LICENSE AND WITH THE
INCLUSION OF THE ABOVE COPYRIGHT NOTICE. THIS SOFTWARE OR ANY OTHER
COPIES THEREOF MAY NOT BE PROVIDED OR OTHERWISE MADE AVAILABLE TO ANY
OTHER PERSON. NO TITLE TO AND OWNERSHIP OF THE SOFTWARE IS HEREBY
TRANSFERRED.
THE INFORMATION IN THIS SOFTWARE IS SUBJECT TO CHANGE WITHOUT NOTICE
AND SHOULD NOT BE CONSTRUED AS A COMMITMENT BY DIGITAL EQUIPMENT
CORPORAT I ON.
DIGITAL ASSUMES NO RESPONSIBILITY FOR THE USE OR RELIABILITY
SOFTWARE ON EQUIPMENT WHICH IS NOT SUPPLIED BY DIGITAL.

ITS

Page 3

Graphics Architecture - PRELIMINARY

1.0

2.0
2.1
2.2

INTRODUCTION • . • . • •
. • .
DISPLAY VOCABULARY • • • • • • •

2.6
2.7
2.8
2.9

2.10
2.11

4.0
4.1

Microcode

Pixel

. . . . . .

• • 7
. • 8

. .
•
. .

Mouse

•

. . 8
. ...... 8
. . •
8
. ........... 8

• • • • • •

Two-Space
. • • •
Bitmap • • . • • •
Off set . .
. .
Extent . . . . . .
Rectangle
. . . •
Frame Buffer
•
BITBLT . • . . . •
. .
Font • •
Clipping Rectangle
Cursor .
Half tone
.

•

•
•
.
.

•
• .
• .
. .
.
•

.
.

.
.
.
.
•

•
.
.
•

.
.
.
.

•

.
.
.

. . .
• • .
• • .

•

Digitizing Tablet
• • .
. • •
ARCHITECTURAL OVERVIEW • . .
WORKSTATION OPERATION
. . . . • .
PARAMETERS TO GRAPHICS OPERATIONS

Value

4 .1.1
4 .1. 2
4 .1. 3
4 .1. 4
4. l. 5
4. l. 6
4. l. 7
4. l. 8
4. l. 9
4. l.10
4.2
4.2.1
4.2.2
4.2.2.1

4.2.2.2
4.2.2.3

4.2.2.4
4.2.2.5

. . . .

Constant .
Address
.

4•2•3•3

4.2.3.4
4.2.3.5
4.2.3.6
4.2.4
4.2.4.l

4.2.4.2
4:.2.4.2.1

4.2.4.2.1.1
4.2.4.2.l.2

MAP

•

.
.

•

. 8
. 8
. 9

.
.

.
.
.

. 9
• 9
9
9
. 9
. 9
10
10
10
10
11
13
13
13
13
13
14
14
14
14
14
15
15
16
17
21
21
22

.
.

.
.
.
.

•.
. .
• .

. . .
.
.

.
.

.
.
.
. .
. .
. .
• .

.
.
.
.
.
.
•

.
.

.
.
.
.
.

. ..
.
.
.
.
.
.
•

.
.
.
.

.
.

·
.

.
.

.
.
.
.
.
.

22
.

23
23
23
27
28

.•.
•

•

. . . .

FONT DATA STRUCTURE Header Format
. . .
The Bitmap . . . . .

.
.
.

.
.

CLIPPING RECTANGLES
Draw Curve Command . . . . . . . . . .
PATH . . . . . . . . . . . . ,
PATTERN STRING . . . . . . . . . . .
PATTERN MULTIPLIER . , . .
PATTERN STATE
. . . . . .
SECONDARY SOURCE .
. . . . . . .
SECONDARY SOURCE OFFSET
Print Text Command . . . . . . . . . .
SOURCE IMAGE
. , . . .

MASK FONT -

. . .
.
.

Extent . .
. . . .
Rectangle
. . . .
Bitmap . . . . . . . . .
Sub-Bi trnap . . . . . . . .
Halftone . .
. • . . .
List . . .
. . . .
DISPLAY GRAPHICS COMMANDS
•
THE CLIPPING MODEL
Copy Area Command
SOURCE IMAGE
SOURCE OFFSET
. . . .
SOURCE MASK
DBSTINATION IMAGE BITMAP
DESTINATION OFFSET

4.2.2.7
4.2.3
4.2.3.1
4.2.3.2

Offset.............

4.2.2.6

.;·

• .

Display . .
. . . . . . . . .
Firmware . . . . . . . .

2.3
2.4
2.5

2.12
2.13
2.14
2.15
2.16
2.17
3.0

• • •

.
.

30
31
31
32
32
33
34

34
,
.

, ,
. . . .
. ...

34
35

Page 4

kstation Graphics Architecture - PRELIMINARY

The Leftarray . . . • . • . . . . . • •
4.2.4.2.1.3
DESTINATION
IMAGE BITMAP . .
• • • •
4.2.4.3
INITIAL
DESTINATION
OFFSET
.
•
.
•
.
• •
4.2.4.4
MAP
• • • • • • • . • • .
• • • • •
4.2.4.5
CLIPPING RECTANGLES . . • . . . . . • • • •
4.2 0 4 . 6
4.2.4.7
TEXT STRING • • • • . . • • • • • .
CONTROL STRING • • . . . . • . . . . . . . •
4.2.4.8
INTER-CHARACTER PAD • . • . • • • . • • • •
4 . 2 - 4.9
SPACE PAD . • • • •
. . • • . . . . • •
4.2 . 4.10
Fill Area Command • . . . • .
• .....
4.2.5
SOURCE IMAGE • • • • . . . . . . . . . • . .
~, .2.5.1
DESTINATION IMAGE BITMAP . . . . . • . . • .
4.2.5.2
DESTINATION OFFSET • . . . • • • • • . • • •
4.2.5.3
MAP
• • • •
4.2.5.4
CLIPPING RECTANGLE . . . . . . •
4.2.5.5
4.2.5.6
PATH • • • • • • • • • • • . • • • · • •
Flood Area Command •
. . • . . • .
4.2.6
SOURCE IMAGE
. . . • • . . . •
4.2.6.l
DESTINATION IMAGE BITMAP .
• • .
4.2.6.2
SEED POINT • . . . . . . . • . • . . . • • .
4.2.6.3
CLIPPING RECTANGLE • . . .
. . .
4.2.6.4
!30Ui\!I>AR'Y' MAP • • • . . . . . . . . . . . • .
4.2.6.5
DISPLAY CURSOR COMMANDS . . . . . • . • . •
4.3
Load Cursor Command
4.3.1.
CURSOR SOURCE IMAGE
4.3.1.1
CURSOR SOURCE OFFSET . . . . . • • • . .
4.3.1.2
4.3.l.3
CURSOR SOURCE MASK . .
CURSOR MAP . . . . .
4.3.1.4
CURSOR ATTRIBUTES . . . . . . . . . . . . .
4.3.1.5
DEVICE ORIENTED COMMANDS
4.4
Attach Cursor Command
. . . . . . . . . .
4.4.1
4.4.1.1
DEVICE TYPE . . . . . . . . . . . . . . . .
Set Cursor Position Command . . . . . . .
4.4.2
4 .4 .2.1
LOCATION . . . . . . . . . . • . . . • .
4.4.3
Get Cursor Position Command • • . . . . .
4.4.4
Get Mouse Position Conunand . . . . . . . . . .
4.4.5
Set Mouse Characteristics Command . . • .
4.4.5.1
TRACKING RATIO . . . . . . . . . . . . .
4.4.5.2
THRESHOLD And SCALE FACTOR . . . •
. . .
4 . 4.6
Set Tablet Characteristics Command . . . .
4.4.6.1
TRACKING RATIO . . . . . . . . . . . . . . .
4.4.6.2
QUANTIZATION RATIO . . . . . . . . . . .
Get Tablet Position Command . . . . . .
4.4.7
4.4.8
Set Pointing Device Event Reporting . . .
4.4.8.1
ENABLE FLAG . . . . . . . . . . . . . . . .
4.5
MISCELLANEOUS COMMANDS .
. . . . . . . .
4.5.1
Move Object Command . . . . . . . . . . . . .
4.5.1.1
OBJECT TYPE . . . . . .
. . . .
4.5.1.2
OBJECT LENGTH
. . . . . . . . . . .
4.5.1.3
SOURCE
. . .
. . .
. . .
4.5.1.4
DESTINATION . . . . . .
. . .
Cl

4.5.1.5

ERRORS

4.5.1.6
4.5.1.6.1
4.5.1.6.2

NOTES . . .
DEVICE TYPES
BUFFERS

•

36
38
38
39
39
39
39
40
40

41
41
42
42
42
42
42
43
43
44
44
44

44
46
46
46
46

46
47
47
48

48
48
49
49

50
50
50
51
51
51
52
52
53

53
53
53
54
54
54
54
55

55
55
55
55

Graphics Architecture - PRELIMINARY

Page 5

DATA FORMATS
. • . • . • . • • .
WORD DATA • • • . • .
CHARACTER STRING
• • •
Report Status Command • . • . • . . . • • . .
DEVICE TYPE
. . . . . . .
. . • • .
DEVICE VERSION • • • • • • • • • • • • •
FIRMWARE VERSION • . • . . . .
. . .
VISIBLE SCREEN FRAME BUFFER BITMAP •
FREE FRAME BUFFER MEMORY • . • • • . . • . .
FREE PROGRAM MEMORY SPACE
. . •
HOST MEMORY SPACE BASE ADDRESS •
No Operation Command • . . • . • •
CONTROL AND STATUS REGISTERS • • . •
• • .
Control And Status Register (CSRO) • . • • •
Interrupt Reason Regi s ter
. . . . • . . . . • .
Device Event Register
. • . . . . .
Functi o n Parameter Register . . . . . . . •
Device Pos ition Register . . . . . • . . • .
Interrupt Vector Add r ess Register
..•
Initial izati on And I ni tial Display Requirements
Aborti ng A Request • . • •
• •.....
COMMAND PACKET FORMATS • . . • • . . . . . • .
Copy Ar ea Command Packet • •
. • . • • .
Draw Curve Command Packet
• • • • • • .
Print Text Command Packet . . . . . . . . . . .
Fill Area Command Packet .
. . . . • . .
Flood Area Command Packet
. • . . .
Load Cursor Command Packet
. . .
Attach Cursor Command Packet
... .
Set Cu rsor Position Command Packet . . . . .
Get Cursor Posit ion Command Packet . .
. . .
Get Mouse Position Command Packet
...... .
Set Mouse Cha racteristics Command Packet • . . .
Set Tablet Characte ristics Command Packet
•••
Get Ta blet Position Command Packet . . . . .
Set Poin ting Device Event Reporting
..... .
Move Object Command Packet .
Report Status Command Packet . . . • . . . .
No Operation Command Packet
......... .
CONSTANTS , OPCODES, MODIFIERS, AND ERROR CODES . .
Control And Status Register 0 Function Codes . .
Command Packet Operation Codes . . . . . . . . .
Command Packet Operation Modifiers .
. .. .
Copy Area Command Modifiers . . . . . . .
Draw Curve Command Modifiers .
. .. .
Print Text Command Modifiers
. . .
. ..
Fill Area Command Modifiers
....
Flood Area Command Modifiers . .
. ..
Load Cursor Cornmand Modifiers . . . . . . . .
Set Mouse Characteristics Command Modifiers
Set Tablet Characteristics Command Modifiers •
Interrupt Reason Values
. • . • . . . .
VAXSTATION 100 RESTRICTIONS
Number Of Planes . . . . .
Halftone Representation

4.5.1.6.3
4.5.1.6.3.1
4.5.1.6.3.2

4.5.2
4.5.2.1
4.5.2.2

59
59
60
60
60
60
60
61
61
62
62
63
63
64
64
65
65
67
68
69
71
72
74
76
78
79
79
79
79
80
80
80
80
81
81
82
83
83
83
84
84
84
85

L5.2.3
4.5.2.4
4.5.2.5
4.5.2.6

4.5.2.7
4.5.3
5.0
5.1
5.2

5.3
5.4
5.5
5.6

5.7
5.8
6.0
6.1
6.2
6.3
6.4
6.5
6.6
6.7
6.8
6.9
6.10
6.11
6.12
6.13

6 ~- 14
6.15
6.16
6.17
7.0
7.1
7.2
7.3
7.3.1
7. 3. 2

7. 3. 3
7. 3. 4
7. 3. 5
7.3.6
7. 3. 7
7. 3. 8

7.4
8.0
8.1
8.2

56
57

86
86
86

87
87

88
92
92
92

Page 6

Graphics Archite c ture - PRELIMINARY
9.0
9.1

9.2
9.3
9.4

9.5

.....
.
....... ........
. . .
....
.
. ..

GENERAL IMPLEMENTATION RESTRICTIONS
Word Access I /0
UNIBUS Window Mapping
Bitmap Sto rage Requi rements
Device Coo r dinate Management For WGA
Keyboard I nte rfac e

.....

....

92
92

92
93
93
96

.,
1 kstation Graphics Architecture - PRELIMINARY

1.0

Page 7

INTRODUCTION

This document describes the command protocol supported by workstation
display systems.
These display systems implement five basic output
commands:
1.

General bitmap copy operation,

2•

Text output,

Curve drawing,

Area fill,

Area flood,

and a collection of miscella n e ous commands.
In addition, they support
a keyboard and one or mo r e p ointing devices (mouse or tablet). The
Workstation Graphics Architecture
is intended to be general and
flexible enough to compatibly support future workstation products,
such as multi-plane (color) and higher-functionality systems.

Page 8

Graphics Architecture - PRELIMINARY
2.0

DISPLAY VOCABULARY

2.1

Display

For the purposes of this document, a physical device consisting of a
high-resolution
raster-scan
monitor,
keyboard,
pointing device,
control processor, microcode and firmware.

2.2

Firmware

A control program that is downloaded from the host into
(see
Microcode).
It
is code to be executed by
microprocessor in the display.

2.3

the
the

display
control

Microcode

A control program which is fixed in hardware (i.e.
not
loadable),
usually in ROMs or PROMs.
It is code to be executed by the microcoded
hardware in the display to perform high-speed BITBLT and graphical
operations on framebuffer memory.

2.4

Pixel

A single picture element or addressable point on a display.
Each
pixel has a value, represented by one or more bits, that describes its
state (i.e., the intensity or color of that point).

I
I
I
I
I
I
I
I
I
I
I

2.5

Two-Space

The pixels in a bitmap or on the display are organized in a Cartesian
(rectangular)
coordinate system of two dimensions.
The origin of the
coordinate grid is in the upper-left-hand corner, with the positive X
axis extending right and the positive Y axis extending down.
All
points (pixels) are thus specified by their X and Y offsets
(in
pixels) from this origin.

I
I

2.6

Bitmap

I
I
I

A data structure consisting of a rectangular array of pixel values.
The specification of a bitmap includes its starting address in memory
and its dimensions.
its Width is the X dimension in pixels,
and Its
Height
is
the Y dimension in pixels.
Its Depth indicates how many
planes, or bits/pixel, the bitmap occupies.
The starting address
points to the upper-left-hand corner of the bitmap.

Page 9

kstation Graphics Architecture - PRELIMINARY
2.7

Offset

Used to specify a point relative to some origin in two-space (See
Two-Space). The coordinates are signed 16-bit X and Y components.

2.8

Extent

An extent (or, a rectangle extent) specifies the height and width,
pixels, of a rectangle in two-space.
Both dimensions must
positive.

2.9

in
be

Rectangle

2.10

Frame Buffer

The bitmap memory used to store the current value of each
from which the physical display monitor is refreshed.

2.11

and

BITBLT

The transfer of a bit string or block from one
(bit block transfer, pronounced "bit blit").

2.12

pixel

location

another

Font

A collection of logically related images within a bitmap which may be
individually addressed by index, e.g. the symbols of a character set.

2.13

Clipping Rectangle

A rectangle used to constrain an operation on a set of pixels in a
bitmap.
The intersection of a clipping rectangle with a destination
bitmap rectangle forms a restricting boundary on the operation being
executed.

Graphics Architecture - PRELIMINARY

•

2.14

Page 10

Cursor

A small image that is displayed on the screen to indicate the current
position of a selected pointing device. The cursor is automatically
moved by the display hardware to reflect pointing device movement.
The cursor image does not interfere with the current state of the
frame buffer.

2.15

Halftone

A rectangular pattern used to tile a destination bitmap. That is, the
halftone is replicated along its height and width to fill the
destination. The point about which this replication is done is
specified as an offset relative to the origin of the destination
bitmap. Halftones are used to supply levels of shading or texture, as
in newspaper printing.

2.16

Mouse

A small box-like device capable of sensing XY movement along a
surface.
It generally has one or more buttons, the function of which
is defined by a particular application.

2.17

Digitizing Tablet

A flat rectangular surface capable of reporting the XY position of a
special "pen" or "puck" placed on it. The pen/puck also has one or
more buttons generally:
again, their function is defined by a
particular application.

Graphics Architecture - PRELIMINARY

3.0

Page 11

ARCHITECTURAL OVERVIEW

The Workstation Graphics Architecture supports a family of display
systems that connect to VAX or PDP-11 workstations.
These displays
differ in price,
performance,
size,
and physical characteristics
(e.g.,
color vs.
black and white, resolution), however they all obey
the same basic command set.
The commands are defined by this
document.
The workstation displays provide primitives to support the System
Display Architecture (SDA), a high-level software interface to display
programming. User application programs access the display via an SDA
procedural interface. This interface allows for multiple applications
to simultaneously access the display through separate windows on the
display screen.
Each application controls one or more virtual
displays, and the position of virtual display windows on the

physical

screen is controlled by a screen manager responding directly to
commands from the display operator. Thus, user application programs
will never directly generate display commands as defined in this
document;
these commands are generated only by privileged operating
system software.
The workstation display and associated input devices are
thus
multiplexed among several application processes.
Therefore,
the
display hardware must supply primitives for window
management,
movement of windows on the display, partial covering of one window by
another, and output to windows not currently visible on the screen.
Basic to all such operations is the ability to efficiently copy bitmap
rectangles from one area of the frame buff er or off-screen memory to
another.
In addition to the simple copy, the hardware also allows
construction of the destination bitmap as a function of the source or
source and destination bitmaps.
Each workstation graphics processor consists of
a
controlling
microprocessor with its private instruction and data memory, a display
monitor, a frame buffer memory from which the monitor
is refreshed,
and an interface to host memory.
It may also have some additional
memory for staging graphics operations or caching commonly used
images,
fonts,
halftones,
and so on.
The display micro-processor
typically has its private memory, the frame buffer memory,
and some
section of host memory mapped into its address space.
Operations can
be performed on data residing in any of these memories.
That is,
the
workstation processor can operate on bitmaps stored in host memory as
well as in local workstation memory.
The
instruction set of the
display is formed by the microcode and firmware installed in it.
All addresses communicated between host and display are 32-bit values.
Some hosts, displays, or interconnects may not support a full 32-bit
address space and will therefore ignore some upper bits of the
address.
All coordinates are 16-bit signed values.
The maximum size
of a bitmap dimension (height or width) is therefore 2**15-1 pixels.
In comparison,
common frame buffer dimensions for existing monitors
are on the order of 2**10 pixels on a side.

Graphics Architecture - PRELIMINARY

Page 12

Host/display communications are handled through a
shared-memory
Control and Status Register interface.
Graphics commands to the
display processor are packet-oriented. The host transmits a command
packet to the display;
the display loads the command completion
status in a control and status register and interrupts the host when
done.
Multiple command packets can be linked together to allow
processing of several commands with a single device interaction.
Other data may be transmitted spontaneously by the display to indicate
the occurrence of various events. Command packets can contain the
address of data in host memory to be read or written by the display.
Any host data passed to the display by address must be locked in host
memory for the duration of the command.

Page 13

rkstation Graphics Architecture - PRELIMINARY
4.0

WORKSTATION OPERATION

This section describes the basic graphics
operations
of
the
workstation display system.
These operations generally affect the
contents of bitmaps stored in host or display memory. The workstation
provides support for five basic graphics operations:

4.1

General bitmap copy (raster operations)

Text output

Curve drawing

Area fill

Area flood

PARAMETERS TO GRAPHICS OPERATIONS

The fundamental parameters to these basic operations are described
below. Many terms are also discussed above in the Display Vocabulary.
The components of each parameter will be described,
as will its
representation in the packet.
Only the "generic" parameters are
mentioned here;
specialized parameters associated with only one
command are mentioned in the section describing that command.
Please
refer to the section on Command Packet Formats
for
complete
specification of all packets.

4.1.1

Value -

Used in this section to denote a single word.
are built up to form the parameters.

4.1.2

Groups of

these

words

Constant -

A constant is a 16-bit binary value passed directly
in a command
packet.
In an N-plane bitmap, only the least significant N bits of
the constant are used. The format of a Constant is:
Value<l5:0>
- Constant Value

4.1.3

Address -

A 32-bit address which points to an object
(e.g.
bitmap,
font,
string)
in display-addressable memory.
Note that the low word is
given first.
The format of an Address is:
Value<l5:0>
- Low word of address

Graphics Architecture - PRELIMINARY

Value<l5:0>

Page 14

- High word of address

I
I
I

4.1.4

Offset -

An Offset specifies a point relative to some origin in two-space
(See
_ Two-Space).
It usually designates the upper-left-hand corner of a
1\!ectangle or bitmap. The coordinates are signed 16-bit X and Y
components.
- X coordinate, in pixels
Value<l5:0>
- Y coordinate, in pixels
Value<l5:0>

4.1.5

Extent -

An extent (or, a rectangle extent) specifies the height and width,
pixels, of a rectangle in two-space.
Both dimensions must
positive.
Value<l5:0>
- Rectangle Width, in pixels
Value<l5:0>
- Rectangle Height, in pixels

4.1.6

in
be

Rectangle -

Used to specify a rectangular portion of a bitmap.
A rectangle
in
two-space is specified by an Offset (of its upper-left-hand corner)
and an Extent, with the bitmap's upper-left-hand corner as the origin.
Offset<31:0>
- Coordinates of rectangle origin
Extent<31:0>
- Width and height of rectangle

4.1.7

Bitmap -

A bitmap specifies a segment of memory containing the values of all of
the pixels
in a rectangle.
The representation of the bitmap is
device-dependent.
A bitmap is specified by:
Address<31:0>
- Base address of the bitmap memory
Value<l5:0>
- Width of the bitmap, in pixels
Value<l5:0>
- Height of the bitmap, in pixels
Value<l5:0>
- Number of bits per pixel in the bitmap

4.1.8

Sub-Bitmap -

A sub-bitmap specifies a rectangular subset of the pixels in a bitmap.
It is thus specified by a bitmap, as above, plus the origin and extent
of a rectangle within that bitmap:
Bitrnap<79:0>
- Specifi c ation of a bitmap memory
Offset<31:0>
- Origin of subset rectangle within bitmap
Extent<Jl:O>
- Extent of subset rectangle

Graphics Architecture - PRELIMINARY
4.1.9

Page 15

Halftone -

A halftone is a rectangular pattern used to tile a bitmap (in the same
way tiles are used to lay a design on a floor). The Halftone
Alignment Offset specifies the origin of the halftone bitmap relative
to the origin of the destination bitmap in two-space.
e.g. if the Offset is (0,0), the Halftone will be aligned to the
upper-left-hand corner of the destination bitmap.
A halftone is
specified as a bitmap:
Bitmap<79:0>
- Specification of the bitmap
Offset<31:0>
- Alignment Offset of halftone pattern

4.1.10

List -

A list is any one-dimensional array of fixed-sized elements. Examples
of
lists are character strings, clipping rectangle lists, and
line-drawing paths. The size of the individual elements is either
known implicitly or specified by other parameters in the command. A
list is specified by its base address and its length (a COUNT of the
elements).
Address<31:0>
- Address of the start of the list
Value<lS:O>
- Length of the list (COUNT)

wo~ kstation

Graphics Architecture - PRELIMINARY

Page 16

\...

4.2

DISPLAY GRAPHICS COMMANDS

As previously stated, the workstation display executes a set of five
basic commands; each command is capable of processing some number of
different parameter types or formats.
In this section we describe the
commands and their parameters in more detail.
Before presenting the commands, it is necessary to describe a model of
the machine implemented by the workstation micro-processor. At its
most basic level, this machine inputs a source bitmap and uses it to
modify a destination bitmap in a specified way. However, in the most
general case, only certain pixels in the source may be used, and only
certain pixels in the destination may be available for modification.
Our model indicates how these sources and destinations are specified.
A picture of this machine is shown below:

v
+---------+

+---------+

SOURCE I
I
IMAGE 1-->I
I
I
+---------+

+-+

+--------+

+-----------+

SOURCE I
!Ml
IOI
!CLIPPING!
!DESTINATION!
MASK . 1-->IAl-->IFl-->IRECTANG. 1-->I
IMAGE
I
Ip I
IF I
I
I
I BITMAP

+---------+

+-+

+--------+

+-----------+

This machine is logically divided into three parts. The first part,
consisting of the source image and source mask, determines what pixels
of the source will be used to update the destination. The last part,
consisting of the clipping rectangles and destination image bitmap,
determines what pixels of the destination
are
available
for
modification.
The map box in the center specifies a function with
which source pixel values may be transformed before
replacing
destination pixels. Or, the destination pixels may be replaced with a
function of both source and destination pixels, as shown by the arrow
leading from the destination to t he map. The destination offset box
(OFF) in the center specifies where the source pixels are to be placed
relative to the origin of the destination bitmap.

'\
~-.

Page 17

wGrkstation Graphics Architecture - PRELIMINARY
4.2.1

THE CLIPPING MODEL -

Here is a graphical illustration of the "Implied Clipping" which may
take place when processing a source bitmap through the above machine.
Only a subset of these operations need take place in most cases,
depending on the user's selection of parameters.
The operation
proceeds in four steps:
[Step l]
First, the Mask parameter is processed.
It may be either a Sub-Bitmap or Rectangle mask.
Mask Bitmap

+-----------------------+
[A]

I
I
I

+---------------+---------------+
11111111111111111
11111111111111111
11111111111111111

+-------+---------------+
+-------------------------------+

Rectangle

o A Sub-Bitmap Mask specifies a Rectangle
(offset and extent) within a Mask Bitmap.
The Mask Offset specifies Point [A], and
the selected intersection (the Sub-Bitmap Mask)
is shown in "l's".
o If a Rectangle Mask is used instead of a
Sub Bitmap Mask, no clipping is needed and
[Step l] is skipped.
That is, the dimensions of the 'selected intersection'
are taken directly from the Rectangle Extent in the packet.
o The resulting Mask Rectangle (the Extent of the result)
is placed on the source as below in step 2:

Graphics Architecture - PRELIMINARY

Page 18

[Step 2]
Next, the Source is processed.
It may be either Bitmap, Constant, or Halftone.
Source Bitmap

+---------------+
[ B]

I
I

+-------!--------+

122222221111111111
122222221111111111 Resultant Mask Rectangle from [Step l]
122222221111111111

+-------1--------+
I

+---------------+
o The Source Offset, Point [B], specifies where the
resultant Mask is placed relative to the Source Bitmap's
origin. The "2's" represent the selected part
of the Source Bitmap. Note that they now also represent
the selected part of a Sub-Bitmap mask.
o If the source is a Constant or Halftone, no clipping is
necessary and [Step 2] is skipped. The Resultant Mask is
passed on unchanged to [Step 3].
The Resulting Mask is then placed on the
destination bitmap below in step 3:

~;----\.,

·J#t

Page 19

\ ~}orkstat ion Graphics Architecture - PRELIMINARY

[Step 3]
Next, the Destination bitmap and Clipping Rectangles are processed.
There may be several clipping rectangles. [Step 3] is repeated for each.
Destination Bitmap

+-----------------------------------------------+

[ D]

+-----------------+
I
I
[ cJ
I
I
I
+-----1--+
I
I

I
I

1444441331
1444441331 Resultant Mask

I
I
I

+-------------1---------------------------------+
1222221221
+-----1--+
I A Clipping Rectangle

+-----------------+

o The Resultant Mask Rectangle is placed
on the Destination Bitmap according to the Destination Offset
Point [C]. Then, one or more Clipping Rectangles may be applied
to further restrict this selected area.
The Clipping Rectangles are specified by an Offset Point [D] and
an Extent (height and width).
o In this example, the "2's" were eliminated when the Resultant Ma
was clipped to the Destination, the "J's" were eliminated when
the Clipping Rectangle was applied, and the "4's" represent the
resultant area available for modification.

'"'workstation
'..
Graphics Architecture - PRELIMINARY
;

.-~

Page 20

[Step 4]
o Now, these accumulated offsets and extents are "traced back" to
find the corresponding selected areas of the Source Bitmap
and Sub-Bitmap Mask:
Destination Bitmap

+-----------------------------------------------+
I
I
I
I
I
I

+-----+. . .
1444441
1444441

I
I
I
I
I
I

+-------+-----+---------------------------------+
Source Bitmap

+---------------+
[B]

+-----+.+
1444441 I
I 44444 I I

1-----+ I
+ ••••••• +

+---------------+
Mask Bitmap

+-----------------------+
I
I
I
I
I
I
I

[A J
+-----+ ••••••••• +

1444441
1444441

+-----+
+-------+---------------+

workstation Graphics Architecture - PRELIMINARY

Page 21

Copy Area Command -

4.2.2

The copy area command is the fundamental operation of the display
system.
Other operations are extensions of this basic copy area
function.
In its simplest form, copy area merely moves a source
bitmap to a destination bitmap. That is, the pixels in a selected
rectangular area of a destination bitmap are replaced by corresponding
pixels in an identically-sized rectangular area of a source bitmap.
Both bitmaps must be in display-addressable memory.
In addition to
simple replacement, the values moved to the destination can be either
(1) a function of the source pixel values, or (2) a function of both
source and destination pixel values. A Mask parameter may also be
used to select some subset of the source.
Finally, the user may
specify one or more Clipping Rectangles to further alter the area
affected by the operation.
The seven parameters accepted by the copy area command are:
o Source Image
o Source Off set
o Source Mask

o Destination Image Bitmap
o Destination Offset
o Map
o Clipping Rectangles
These parameters are described in detail in the following sections.
The formats of the parameters in the command packet are those shown in
Section 4.1 above. Complete packets are pictured in Section 6.

4.2.2.1

SOURCE IMAGE -

The source image parameter specifies a bitmap whose pixel values are
used to update the destination. The source image can be specified in
three ways.
1.

First, the source parameter can be a single constant value
that is used to replace all pixels in the destination. This
would be used, for example, to set the entire destination to
a single value (color or shade).

Second, the source can be specified as a
bitmap
in
display-addressable memory.
If the source and destination
bitmaps overlop, the copy operation sequences through the
pixels so that source pixels are not modified before they are
used to update the destination.

·workstation Graphics Architecture - PRELIMINARY
3.

4.2.2.2

Page 22

Third, the source can specify a halftone to be used to tile
the destination as specified by the remaining parameters. A
halftone is represeted as a bitmap.
The
bitmap
is
conceptually replicated horizontally and vertically as needed
to fill the destination.

SOURCE OFFSET -

The source offset parameter consists of a point
(two 16-bit signed
values) specifying where the source mask parameter, described below,
is to be placed relative to the source.

4.2.2.3

SOURCE MASK -

The source mask parameter defines a bounded subset of the source to be
used to modify destination pixels.
That is, the source mask can
restrict the set of source bitmap pixels used in the copy operation.
Two mask formats are available, depending on how the source subset is
to be determined.
1.

First, the mask can select a rectangular subset of the source
bitmap.
Only a rectangle extent need be specified (its
height and width),
since the source offset
parameter,
described above, determines where the origin of the rectangle
will be placed relative to the source.

Second, the mask can be specified as a bitmap of zero and one
values.
This is the most general form,
in which the
one-valued elements of the bitmap select an arbitrary set of
pixels from the source to be used in the operation. That is,
the mask bitmap is used as a template.
Once again,
the
source off set determines how the mask is to be placed
relative to the source bitmap. The mask is specified as a
sub-bitmap, as defined in the previous section .

.Either source mask format can be selected with any of the possible
source image parameters.
For example, selection of a rectangle mask
with a constant or halftone source generates a rectangle filled with
that color or halftone for use as the source. The copy area command
can thus be used as a simple rectangle fill.
More complex fill
operations can be performed with the fill area command,
to be
described later.

Page 23

workstation Graphics Architecture - PRELIMINARY
4.2.2.4

DESTINATION IMAGE BITMAP -

The destination image is always a
modified.

4.2.2.5

display-addressable

bitmap

DESTINATION OFFSET -

The destination offset parameter determines the placement of source
pixels in the destination bitmap. Refer to the Clipping Model: after
the source image and mask have been processed, the Destination Offset
is used to specify the origin of the result. The result is then
placed on the destination bitmap.

4.2.2.6

MAP - A set of source pixel

values

can

mapped

into

different set of destination pixel values by using a form of the map
parameter. A map can have three forms:
1.

Identity Map.
No map is specified.
directly replace the destination pixels.

The

source

pixels

Source Table-Lookup Map.
In this case, each source pixel
is
used as an index into a table of destination pixel values.
To map the pixels in an N-plane bitmap to pixels in an
M-plane bitmap,
the table would have 2**N entries.
Each of
these
entries
is
a
16-bit
word,
of
which
the
least-significant M bits are defined.

Page 24

~~rkstation Graphics Architecture - PRELIMINARY

consider, for example, the case below~ T~e so~rce bitmap has
two planes (Z = 2), and the destination bitmap has three
planes (Z = 3). The source pixels may therefore have values
o through 3, and the destination pixels may have values 0
through 7. In this example, then, t~e source ma~ wo~ld ~ave
2**2 = 4 locations, which contain the destination pixel
values, stored in the least-significant 3 bits of each
location.
Get pixel from source
SRC

+------------+
+------------+ I

10000001111111 I
13333333333331 I
11111100000001-+
+------------+
I
I
Table look-up
I
15
2 1 0
I
+----------+-----+
I
+->[0]
11 0 11
+----------+-----+
[l]
IO 1 OI
+----------+-----+
[2]
11 O OI
+----------+-----+
[3]
11 1 11

Replace pixel in desti
DST

+------------+
+------------+ I
+------------+ I I

15555552222221 I I
I 777777777777 I I-+
12222255555551-+

+------------+

(5) >---+

(2)
(4)
(7)

+-- - -------+-----+

If the source z parameter (the number of bits/pixel) is equal
to l, then the table may be included in the WGA command
If the source z
packet (2**1 = 2 16-bit word entries).
parameter is greater than 1, then the lookup table will not
fit in the packet, and must be referred to via a pointer to
the table's base address.
3.

Function Code.
The map may be a function code which
designates one of a number of logical operations to be
performed on the destination,
using
the
source
and
destination pixel values as operands.
This form of map assumes that the pixels in both the source
and destination planes have the same number of significant
bits.
Since only one word is needed to specify the function code
value,
it may be passed either directly (literal function
code) or indirectly ( pointer to function code), even if the z
parameters of the source and destination are > 1.
N-plane-source / N-plane-destination operations

are

performed

w~rkstation

I
I
I

I
I
I
I

I
I
I
I
I
I
I

I
I·

I
I
I
I

Page 25

Graphics Architecture - PRELIMINARY

on corresponding
values.

bits

between source and destination pixel

The user may specify any 2-operand logical function by its
"Boolean Characteristic Number". This number is generated as
below:
Consider a two-input truth table,
where the source
destination
pixel values are the
inputs,
and the
destination pixel value is the output.

and
new

Source Pixel Value
I
Destination Pixel Value
I
I
Output Pixel Value

I
I v

0
1
1

1
0
1

--------+---

I
I

c
d

+-----> abed
Since there are two
inputs,
there are
four
possible
combinations.
The result of these combinations may be
arranged, as shown above,
into a
four-bit
number
the
Characteristic
Number.
This defines a
logical mapping
function to the display:
Map Function Code Parameter:

+---------------------+
lab c di

+---------------------+
15

3 2 1 0

The Characteristic Number is a 4-bit value,
define a set of 16 functions:
Function Code I

and

Logical Function

---------------+-----------------0
1

3
4
5
6
7
8
9

src AND dst
src AND NOT dst
src
NOT src AND dst
dst
src XOR dst
src OR dst
NOT src AND NOT dst
NOT src XOR dst
NOT dst

may

Workstation Graphics Architecture - PRELIMINARY
B

c
D
E
F

src OR NOT dst
NOT src
NOT src OR dst
NOT src OR NOT dst
1

Page 26

Page 27

workstation Graphics Architecture - PRELIMINARY

Note that since these are "bitwise" operations, two N-bit
pixels may also be operands to the same 16 functions. The
selected function is applied to corresponding bits in the
operands.
Hence, the same Characteristic Number/Function
Code may be used to describe an operation between any two
operands, no matter what their size.
E~ample:

src

+---------+
I 0 1 0 1 I

+---------+
x x x x

0 0 0 0

R R R R

+---------+
dst

New dst

I 1 o o 1 I

To XOR between two 4-bit pi
use function code 6:
Source Pixel Value
I
Destination Pixel Value
I
I
Output Pixel Value
I
I
I
v
v
I v

+---------+

--------+---

+---------+

0
0
1

I 1 1 0 0 I

+---------+

0
1
0
1

I
I
I
I

0
1
1

I
+-----> 0110

4.2.2.7

= 6

CLIPPING RECTANGLES -

As shown in the Clipping Model, Copy Area operations are clipped to
the boundaries of the destination bitmap.
To clip the operation
within the destination, one or more Clipping Rectangles may be used.
A Clipping Rectangle specifies an area of the destination available
for modification. A Clipping Rectangle is defined by an Offset
(relative to the origin of the Destination Bitmap) and an Extent. The
union of all of the Clipping rectangles Provides a boundary defining
the area of the destination that can be modified. The clipping
rectangles should not overlap. A single Clipping Rectangle may be
given directly in a command packet. If more than one is required, a
List of Clipping Rectangles must be specified.

Workstation Graphics Architecture - PRELIMINARY
4.2.3

Page 28

Draw Curve Command -

The Draw Curve command is used to paint a source bitmap along a Path
defined by a given sequence of points. The path between any two
points may be either straight or curved. The Draw Curve command uses
the same parameters as the Copy Area command, and has additional
parameters to specify the path and pattern string.
The command proceeds as follows.
First, the source is determined by
the Source Bitmap, Source Mask, and Source Offset. Second, this
source image is painted through the curve by successively moving its
origin through the sequence of points within the Destination Bitmap as
determined by the Path parameter. The source is combined with the
destination as specified by the Map.
The parameters to the Draw Curve command are:
o Source Image
o Source Off set
o Source Mask
o Destination Image Bitmap
o Destination Offset
o Map.
o Clipping Rectangles
o Path
o Pattern String
o Pattern Multiplier
o Pattern State
o Secondary Source
o Secondary Source Offset
Specification of the first seven parameters is identical to the copy
area
command.
The
next sections describe the curve-specific
parameters.

Page 29

Workstation Graphics Architecture - PRELIMINARY
4.2.3.1

PATH -

A path is specified as a list of segments.
Each path segment is
described by its ending point and a Flag Word. The starting point is
the end of the previous segment. The format of a path element (point)
is:
0

+---------------+
I

X offset

Y offset

+---------------+

+---------------+
FLAGS

+---------------+
The Flag Word describes the characteristics
contains the following indicator bits:
1.

Bit 0 - the Offset-Relative I

that

Path-Relative

segment,
bit,

and

indicates

how the ending point coordinates are to be interpreted. If
bit 0 is clear (Offset-Relative mode), this point
is
interpreted as a point relative to the Destination Offset
parameter specified above. If bit 0 is set (Path-Relative
mode), this point is interpreted relative to the end of the
previous segment in the list.
2.

Bit 1 - the Draw I Move bit, indicates

whether

not

the

segment should actually be drawn.
If bit 1 is clear, the
segment will be drawn.
If bit 1 is set, the segment will not
be drawn; the position is simply advanced to the next point.
With this bit set (Move), several disconnected segments may
be drawn with a single Draw Curve command. This bit would be
set (Move), for example, to define the starting point of the
first segment in a Path. Invisible segments can be also be
used to provide additional information to the cubic spline
algorithm. An invisible segment does not advance the pattern
string, if one is specified.
3.

Bit 2 - the Straight I Curved bit, indicates how

the

source

should be moved from the starting to the ending point.
If
bit 2 is clear, a straight line is painted to the current
point.
If bit 2 is set, a curve is drawn using a cubic
spline algorithm. For a curve to be drawn through a point,
it must have a preceeding segment and succeeding segment to
define the point's tangents.
Then the curve will merge
smoothly with the preceeding and following segments without
abrupt bends.

Bit 3 - the Start Closed Figure bit, when set, indicates that
this

point is the first in a series of segments that def i nes

a closed figure (i.e.

circle, square, polygon).

Page 30

workstation Graphics Architecture - PRELIMINARY
5.

Bit 4 - the End Closed Figure bit, when set,

indicates

that

this is the last point in the closed figure.
The first and
last points (that is, the start and end points) for a closed
figure must be identical.
6.

Bit 5 - the Draw Last Point bit, when set,

indicates that the

final point in the segment will be drawn.
the endpoint will not be drawn.

If bit 5 is clear,

Therefore, by manipulating the bits in the Flag Word,
The Path may
consist of both straight-line segments and curved (cubic spline)
segments. These segments may be drawn "brush up" or "brush down".
Using this mechanism,
additional points can be supplied before the
first visible point and after the final visible point,
to define a
curve more accurately.
When drawing ~ multi-segment path using a mapping function such as
XOR,
end-point-writing can be disabled for each segment so that its
end-points are written only once.
The flag word MUST be used to specify the starting
point of closed figures.
For example, a circle would be specified by the
path:
point
point
point
point
point

point

following

and

ending

five-point

1: move to Pl, start closed figure

2:
3:
4:
5:

draw curve
draw curve
draw curve
draw curve

to
to
to
to

P2
P3
P4
Pl, end closed figure

A segment of length zero (i.e., same starting and ending point) causes
a single point to be drawn.
If the source image is specified as a halftone, the halftone pattern
is always aligned to a fixed point (the Halftone Alignment Offset).
This gives the effect of painting with a halftone.

4.2.3.2

PATTERN STRING -

The Pattern String Parameter is used to draw dashed or patterned
segments in the Path.
As mentioned above, segments may be either
"straight" or "curved". A dashed segment alternates between writing
and not writing pieces of the segment based on the sequence of set and
clear bits in the pattern string.
Patterned segments require the
specification of a secondary source parameter.
A patterned segment
alternates between the source and the secondary source based on the
sequence of set and clear bits in the pattern string.
This would be
used, for example, to draw a line that alternates between two colors.

Page 31

Workstation Graphics Architecture - PRELIMINARY

The Pattern String parameter is a 16-bit word which contains the
The pattern may be 0 to 16 bits long. The Least
drawing pattern.
Significant Bit is used first, and the pattern rotates to the right,
wrapping around as needed.
The pattern is only applied to segments that have the Draw flag bit
set as described above;
invisible (Draw flag clear} segments do not
advance the pattern. A Pattern String of length 0 draws a solid
segment (as though there were no pattern).
The Draw Curve command causes an iterated copy area of the source to
each pixel in the destination between segment endpoints. The Pattern
String specifies how or whether the destination is to be modified at
each pixel.
A command modifier bit determines the mode in which the
pattern string parameter will be used. The two modes are:
1.

Single Source Mode - In this mode, a set bit in the pattern
causes the source to be copied to the current destination
pixel along the curve path, and a cleared bit causes the
destination pixel to be skipped.
This format
lets the
existing background "show through" the spaces in the pattern.
The result is called a Dashed Line.

Alternate Source Mode - In this mode,
a set bit in the
pattern causes the source to be copied to the current
destination pixel along the curve path,
and a cleared bit
causes a different source to be copied to the destination.
This Secondary Source is passed in the
alternate source
Patterned Line.

mode

used.

The

command
result

packet

when

is called a

Thus, Single Source mode alternates between writing and not writing of
a single source. Alternate Source Mode alternates between the writing
of two different sources.

4.2.3.3

PATTERN MULTIPLIER -

The Pattern Multiplier parameter specifies the number of times each
bit
in the Pattern String should be repeated before moving on to the
next bit.
For example, if the Pattern String is '10' and the Pattern
Multiplier is 3,
the pattern '111000' will be used to generate the
patterned curves and lines in the draw command.

4.2.3~4

PATTERN STATE -

The Pattern State parameter allows a draw curve command to continue at
the
P attern String position at which a previous draw cu r ve comple~ed.
In this way, Pattern Strings can be used across multiple Draw Curve
commands.

workstation Graphics Architecture - PRELIMINARY

Page 32

The Pattern State consists of two 16-bit words, the Pattern Position
and Pattern Count. Pattern Position specifies the starting bit within
the pattern string, and must be less than the size of the string (that
is,
it runs from zero to length-1). Pattern Count specifies the
starting count to be used for that bit on its first use, and must be
less than or equal to the pattern multiplier (that is, it runs from
one to the multiplier count).
The Pattern State parameter can be updated by the display following
the completion of a Draw Curve command to indicate where in the
Pattern String the next Draw Curve command should begin. The Pattern
State can be specified as a literal in the command packet or as the
address of the two-word parameter to be updated.
If a Pattern State is specified in the Draw Curve command, scanning of
the Pattern String begins at the specified Pattern Position. The
specified starting bit is then repeated PATTERN MULTIPLIER
PATTERN
COUNT times. The scan then continues at the following bit, with each
bit used PATTERN MULTIPLIER times.

4.2.3.5

SECONDARY SOURCE -

In alternate source mode, the secondary source parameter specifies the
source that will be copied to the destination whenever a zero bit in
the pattern string is encountered.
The secondary source can be
specified as any of the formats allowed for a source in the copy area
command.

4.2.3.6

SECONDARY SOURCE OFFSET -

The secondary source offset specifies how the source mask will be
applied to the secondary source, when the secondary source is selected
in pattern string alternate source mode.

workstation Graphics Architecture - PRELIMINARY
4.2.4

Page 33

Print Text Command -

The Print Text command is the basic character string output operation
for the Workstation Graphics Architecture. The command scans along a
given text string, looking up the character indexes in a user-defined
font.
The selected character cells are written horizontally, one
after the other, in the destination bitmap.
For special formatting of output strings,
such as horizontal and
vertical adjustments (e.g.
subscripts), an optional control string
may be included.
To support simple string justification, the user may set the spacing
between character cells by specifying a fixed inter-character pad.
The width of the space character may be adjusted by specifying a fixed
additional space pad.
The Print Text command is an extension of the basic Copy Area.
The
text string is used to select bitmap cells in the font that become the
source parameters to the copy operations.
The images of the characters or symbols in a font are stored in a
bitmap.
This bitmap, plus some other indexing information, forms the
WGA font data structure. The font is specified as the address of such
a data structure in display-addressable memory, and may be used either
as a bitmap source image or as a bitmap mask.
The parameters to Print Text are:
o Source Image
o Mask Font
o Destination Image Bitmap
o Initial Destination Offset
o Map
o Clipping Rectangles
o Text String
o Control String
o Inter-character Pad
o Space Pad
The sections below describe the parameters in more detail.

workstation Graphics Architecture - PRELIMINARY
4.2.4.1

Page 34

SOURCE IMAGE -

The source image may be specified in one of three ways:
1.

First, it can be the address of a font data structure, as
defined below.
The character cells in the font bitmap are
copied directly to the destination bitmap, with no mask
involved.
The pixel values copied may, however, be modified
with a mapping function.

Second, the source image can be a constant pixel value. This
value is used as the writing color for characters whose
symbols are defined by a mask font.

Third, the source image can be a halftone. The halftone is
used as the writing color for characters whose symbols are
defined by a mask font.

4.2.4.2

MASK FONT - -

The mask font parameter is the address of a one-bit-per-pixel font
data structure that defines the symbol shapes of the characters in the
font. The mask font is used in conjunction with a constant or
halftone source to define the shape and writing color of the
characters in the text string.

4.2.4.2.1

FONT DATA STRUCTURE - -

The font parameter is specified as the address of a font data
structure in display-addressable memory.
This structure has three
parts:
o

The Header

The Image Bitmap

The Leftarray [optional]

The Header contains central information about the font, the Bitmap
contains the character images, and the Leftarray is a table of offsets
used to find the character cells in the Bitmap.
The Font parameter in the command packet is the base address of the
Header. All "pointers" in the Header, e.g. the LEFTARRAY pointer and
the pointer in the BITMAP specification, are byte offsets relative to
this address.
So to compute the base address of the Leftarray, for
example, one must add the
Leftarray pointer (offset) to the Font
parameter (Header address).
In normal usage the data structure is
contiguous.

Page 35

workstation Graphics Architecture - PRELIMINARY
4.2.4.2.1.1

Header Format bitmap

FONT BITMAP.
Specification
containing the character images

Value<l5:0>

FIRSTCHAR.
font.

Value<l5:0>

LASTCHAR. The last valid character index in the
font.
Any
character
not within the range
(FIRSTCHAR to LASTCHAR)
inclusive is an error.
The display terminates processing and reports the
error to the host.

Address<31:0>

LEFTARRAY.
A pointer to an array of 16-bit
elements.
There is one element for each character
in the font between FIRSTCHAR and LASTCHAR.
Each
element contains the X offset
(see Offset in
Dis~lay
Vocabulary)
of the left edge
of
a
character cell in the bitmap.
If the WIDTH
parameter is non-zero, then all cells have the
same width,
no Leftarray is needed,
and this
pointer need not be defined.

Value<l5:0>

BASELINE. The Y offset of the character baseline
from the top of the character cells (since origin
is upper left).

Value<l5:0>

SPACE. The index of the space character
font (32 decimal for ASCII fonts).

Value<l5:0>

WIDTH.
If all characters in a font have the same
width (a fixed-width font), then this parameter
specifies that width (in pixels), and no Leftarray
is needed.
For variable-width fonts (ones with a
Leftarray), this parameter must be zero.

4.2.4.2.1.2

the

Bitmap<79:0>

The first valid character index in the

the

The Bitmap -

The actual character images are stored in "strike" format;
that
is,
all the character images are concatenated to form a horizontal bitmap
strip. The first character is on the left and the last
is on the
right.
The characters are aligned so that they have a common
baseline. The height of the bitmap extends (at a minimum)
from the
bottom of the lowest descender to the top of the tallest character.
There is no restriction on the height of a font, or the width of the
characters within it, save for certain practical restrictions such as
memory address space. Each character in the font
is represented in
the bitmap;
absent characters have images of zero width.

workstation Graphics Architecture - PRELIMINARY
4.2.4.2.1.3

Page 36

The Leftarray -

Each character in a variable-width font is indexed through the
Leftarray.
If the WIDTH parameter is non-zero, this array is not
needed.
The following symbols are defined for the figure and explanation
below:
F = The index of the first defined character in the font.
L = The index of the last defined character in the font.
R = L - F, The range of defined characters in the font.
N = index - F, The "normalized" index of the character.
When a character index is fetched from the text string, it is
"normalized" to N, which is in the range 0 to R. Note that there are
therefore R+l characters in the font.
And the Leftarray has R+2
elements.
is
Now the off set of the left edge of character N (xl in the figure)
found in the N'th element of Leftarray, and the left edge of the next
character ( x2 in the figure)
is found in the
next
element.
Subtracting these two numbers gives the width (in pixels) of the
character N.
All of the characters have the same height - the height of the Bitmap.
Note the case for the last character in the font.
There are no
characters following it, but there is a location in Leftarray
containing the proper 'left edge' so that the last character's width
may be correctly computed as before.

Page 37

workstation Graphics Architecture - PRELIMINARY

ACCESSING CHARACTER CELLS IN A VARIABLE-WIDTH WGA FONT
Font Bitmap containing the character cells

x ------------------------------------------------->
<---------- x3 -------------------------------------------->
<---------- x2 ----------------------->
<---------- xl ------------->
<----------

+--------------------------+---------+---------------------+-----+
!Cell
I
!For
I
ICharacter I
I
!Index
I
I
l"N"
I
+--------------------------+---------+---------------------+-----+

+--> I
I
I
I
I

I
I
I
I
I
+---(FONT + BITMAP PTR)

+-----------------+

(FONT + LEFT PTR)-------+
I
I
I
I
I
Font Data Structure
I
Header
I
15
o
I

+---------------+
FONT -->I BITMAP PTR (LO) I
+---------------+

+-----------------------+

Leftarray
15

+---- - ----------+
I
I

I BITMAP PTR (HI) I

+---------------+
x
+---------------+
+---------------+

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

FIRSTCHAR
+---------------+
LASTCHAR

+---------------+
I LEFT PTR {LO) I
+---------------+
I LEFT PTR (HI) I

+---------------+
BASELINE

+---------------+
I

SPACE

WIDTH

+---------------+
I

+---------------+

I
+---------------+
+--> [ 0 ]
I

+---------------+
[N]
xl
I---+

I
I
I
I

I
+---------------+
I
[N+l]
x2
1-------+
+---------------+

I
I
I
I

I
I
I

I
I
I
I
I

+---------------+
x3
1-----------+
+---------------+
[R+l]
X
+---------------+
[R]

Page 38

workstation Graphics Architecture - PRELIMINARY
A font may be used in two ways to specify the
First, a font may be used as a source image.

character images.
In this case, each

rectangular character cell is copied to the destination.
The cell
contains both the character image and a background. A map parameter
can be used to transform the pixel values for the image and
background.
Multiplaned systems can use a source image font to
contain anti-aliased characters (characters with grey scale).
Second, the font may be used as a source mask, or, mask font.

mask

font
is always one bit per pixel, independent of the number of bits
per pixel in the destination bitmap.
The cells of the mask font
simply define the shapes of the characters. When a mask font is used,
the source image parameter specifies the writing color.
The source
image can be either a constant or a halftone. When characters are
output using a mask font, only the character shapes themselves are
written, that is, there is no background rectangle for the characters.
Mask fonts thus allow for writing of characters over other images
without obliterating the underlying rectangular area, or for writing
overstruck characters.
They also
provide
a
storage-efficient
mechanism for writing halftone or single-colored characters.

4.2.4.3

DESTINATION IMAGE BITMAP -

The destination image bitmap is identical to
bitmap specified in the copy area command.

4.2.4.4

the

destination

image

INITIAL DESTINATION OFFSET -

The Initial Destination Offset is a point in the destination image
bitmap where character writing begins.
It
is the same as the
Destination Offset mentioned in the Clipping Model, but may be used in
one of four modes. There are two modifier bits associated with the
Initial
Destination
Offset,
named
Update
and
Indirection,
respectively. They work as follows:
Initial Destinat i on Offset
Modifier Bits
U = Update, I = Indirection

U I I I mode

---+---+----------0
0
l

literal
indirect
update literal
update i ndirect

1. literal

The (x,y) offset 1s used as given in the WGA command packet.

The updated offset is not saved when the command is finished.

Page 39

workstation Graphics Architecture - PRELIMINARY

2. indirect
The (x,y) offset is stored in a 2 word block pointed to by an
address given in the WGA command packet. The offset in the
block will not be updated when the command is finished.
3. update literal
The (x,y) offset is used as given in the WGA command packet.
The updated offset is saved by storing it back in the WGA com
packet. An offset saved in this way may be referenced by re-u
the same WGA packet. This is the case in single-character ech
(see below).
4. update indirect
The (x,y) offset is stored in a 2 word block pointed to by an
address given in the WGA command packet. The updated offset w
be saved by writing it back to this same block. The address p
in the WGA packet will remain unchanged.
(Note: For efficient echoing of keyboard input characters,
a print
text command packet specifying a single character string can be cached
in display local memory. This packet will specify the address of a
single character buffer (in host memory)
and the address of the
initial x-y destination offset (also in host memory). The host need
only load the character and issue the print text command, assuming
that the x-y position is maintained by the display as indicated above.
Storing the packet in display memory saves the time required to copy
the command over the host/device interface.)

4.2.4.5 MAP - This
command.

parameter

specified

the

copy

4.2.4.6 CLIPPING RECTANGLES - The clipping rectangles
to the clipping rectangles in the copy area command.

are

area

identical

4.2.4.7 TEXT STRING - The text string is a list of 8-bit or 16-bit
character indices.
The size of the characters is determined by an
opcode modifier.

4.2.4.8 CONTROL STRING - The control string is a list of 16-bit words
that specify how the text string is to be output.
Normally,
characters are written on a single horizontal line with the spacing
determined by the width of each character as stored in the font.
However, the control string provides a series of commands that allow

for x-y position adjustment as well as text string character skipping.
~ach

control string consists of a 16-bit opcode follpwed

two 16-bit operands.

zero

The control commands are defined as follows:

Page 40

workstation Graphics Architecture - PRELIMINARY
1.

OUT(N) (opcode=O) outputs the next N characters of
string, where N is the single operand.

OUTALL (opcode=l) outputs all
text string.

SKIP(N) (opcode=2) skips the next N characters
string, where N is the single operand.

ADJUST(X,Y) (opcode=3) adjusts the current character position
by X and Y, where X and Y are signed horizontal and vertical
adjustments specified in pixels, where X and Y are the two
operands.

remaining

characters
of

the

text

the

text

If the text string is exhausted before the control string, processing
of the control string continues with the following interpretation:
(1) any ADJUST commands are obeyed,
(2)
any OUTALL commands are
ignored,
and (3) any SKIP(N) or OUT(N) commands with N > 0 cause
termination of the command with error status.
If the control string
is
exhausted
before
the text string,
the command terminates
successfully at that point.

4.2.4.9

INTER-CHARACTER PAD -

The inter-character pad parameter is a constant value specifying the
number of pixels to be added to the current destination offset
following each character printed (including the last character in the
text string).

4.2.4.10

SPACE PAD -

The space pad parameter is a constant value specifying the number of
pixels to be added to the current destination offset following each
space character printed.

workstation Graphics Architecture - PRELIMINARY
4.2.5

Page 41

Fill Area Command -

The fill area command is used to construct a source consisting of one
or more closed shapes, where each of the closed shapes is filled with
a single color or halftone value.
The generated source image is
copied to the specified location in a destination bitmap. For
example, fill can be used to place a circle filled with a single color
or shade within a destination bitmap.
The fill command is used when the boundary or boundaries of the area
to be filled are known and can be defined by a list of straight or
curved segments. The flood area command, on the other hand,
is used
when the boundary is not completely known, but the user can specify
one internal point of the closed area.
The source parameter specifies the pixel value or halftone to be used
to fill the bounded area or areas. A path list, specified as in the
draw curve command, specifies the boundaries of the areas to be
filled.
The path list defines the bounded areas within the specified
destination bitmap. Thus, this command causes one or more areas of a
destination bitmap to be filled with one or more similarly-colored
shapes.
At most one clipping rectangle can be used with fill area.
The parameters to the fill area command are:
o Source Image
o Destination Image Bitmap
o Destination Offset
o Map
o Clipping Rectangle
o Path
The parameters are described as follows:

4.2.5.1

SOURCE IMAGE -

The source image specifies the constant pixel value or halftone with
which the closed area specified by the path parameter will be filled.
The constant or halftone is specified as in the copy area command.
Bitmap source image is not allowed.

Page 42

workstation Graphics Architecture - PRELIMINARY
4.2.5.2

DESTINATION IMAGE BITMAP -

The destination image bitmap is identical to the destination image
bitmap in the copy area command.
It specifies the bitmap in which the
filled shape will be placed.

4.2.5.3

DESTINATION OFFSET -

The destination offset is identical to the destination offset
in the
copy area command, and specifies the placement of the filled image in
the destination.

4.2.5.4

MAP -

The Map parameter is identical to the map in the copy area command.

4.2.5.5

CLIPPING RECTANGLE -

The clipping rectangle parameter for the fill area command specifies a
single clipping rectangle to constrain the fill operation.
If no
clipping rectangle is supplied, the operation is
size of the destination bitmap.

4.2.5.6

constrained

the

within

the

PATH -

The path specifies one or more closed areas to be

filled

destination bitmap by the pixel values specified by the source
parameter. A path is specified as a list of segments.
The segments
are described by a path list specified exactly as in the draw curve
command. Each segment of the path can be a straight or curved line.
Multiple disjoint closed areas are generated through use of the move
flag bit.
If the path describes an open figure,
the results are
unpredictable,
but the operation will always be confined to the given
bounding box (either the clipping rectangle or the destination bitmap
border).

Workstation Graphics Architecture - PRELIMINARY
4.2.6

Page 43

Flood Area Command -

The flood area command is used to fill bounded areas of a destination
with a single color (pixel value) or halftone pattern. Moreover, the
flood command is similar to copy area with two particular differences:
(1) the source pixels are not mapped, but are copied directly, and (2)
the destination pixels that are modified are determined by a flood
algorithm, and lie within a closed area bounded by one or more pixel
values.
In more detail, the flood algorithm proceeds in several steps. First,
the algorithm determines the area of the destination to be flooded;
that is, it locates the inside and outside portion of the closed area.
Moreover, the algorithm builds a binary mask which when applied to the
destination, will select those pixels on the inside of the flooded
area.
The determination of the bounded area requires two parameters:
a seed point and a map. The seed specifies a single pixel within the
destination;
this point is known to be within the bounded area. The
boundary map is a table of zeros and ones, used to determine whether
points are internal or boundary points. The algorithm searches every
pixel adjacent to the seed pixel. The value of each pixel is mapped
through the boundary map table. If the mapped value of the pixel is
zero, then this point is an inside point, otherwise it is a boundary
point.
If an inside point is found, the algorithm continues to
examine all of its neighbors.
Processing continues until
the
neighbors of all internal points have been examined. If the seed
point is found to lie on a boundary, the algorithm terminates
immediately.
Once the boundary has been determined, the inside area is flooded with
the source. The parameters to flood area are:
o Source Image
o Destination Image Bitmap
o Seed Point
o Clipping Rectangle
o Boundary Map
The parameters are described in more detail in the following sections.

4.2.6.1

SOURCE IMAGE -

The source image can be specified in two of the three formats
available for the copy area command. For flooding, the source will be
either (1) a constant, flooding the bounded area with a single color,
or (2) a halftone, flooding the area with a halftone pattern.

Page 44

Workstation Graphics Architecture - PRELIMINARY
4.2.6.2

DESTINATION IMAGE BITMAP -

The destination image bitmap is the bitmap containing the image to
flooded.

4.2.6.3

SEED POINT -

The seed point specifies the coordinates of a single point in the
destination bitmap that lies within the internal region to be flooded.

4.2.6.4

CLIPPING RECTANGLE -

The clipping rectangle parameter for the flood area command specifies
a single clipping rectangle to constrain the flood operation. If no
clipping rectangle is supplied, the operation is
size of the destination image bitmap.

4.2.6.5

constrained

the

BOUNDARY MAP -

The Boundary Map tells the command which pixels of the closed figure
in the destination bitmap are internal (to be flooded) and which are
on the boundary (to be left alone). The Boundary Map parameter is a
table whose one-bit elements are indexed by the pixel values in the
destination. If the element is set, the corresponding pixel value is
an external point, and if the element is clear, the pixel value is an
internal point.
Note that unlike the previous commands, the Flood Area command map
used to map pixels in the destination.

For an N-bit-per-pixel destination bitmap, 2**N pixel values are
possible, and so a table with 2**N entries is needed. However, since
each table entry contains either a zero (indicating an internal point)
or a one (indicating a point on the boundary), only one bit is needed
for each entry.
If N is less than or equal to four, then the table only uses one word,
and only 2N bits of it are defined. For example, in a system with 3
bits per pixel, the 8-bit boundary map 00101110 indicates that pixel
values 1, 2, 3, and 5 are boundary values, while pixel values O, 4, 6,
and 7 are interior values.
If N is greater than four, however, more than one word i s needed. The
least-significant four bits are used to select the proper bit within a
word as before, and the remaining most-significant bits are used to
select the proper word in the boundary map.
Example: 6 bit-per-pixel destination bitmap

pixel value = 35 (hex)

workstation Graphics Architecture - PRELIMINARY

-pixel value

Page 45

boundary map
15

5 4 3 2 1 0

.....................
0
. . . . . . . . . . . . . . .. . . . . .
1

+-----------+
11 110 1 0 11

+-----------+
A ___ A _______ A

bit offset

.....................I 2

+------------------------------v

word off set +----------+-+------+

+----------------------->I

IOI

I 3

+----------+-+------+

and pixel value 35 (hex) is an internal point.

workstation Graphics Architecture - PRELIMINARY
4.3

Page 46

DISPLAY CURSOR COMMANDS

The display cursor is a small image or icon that
is automatically
displayed on the screen at a point determined by the position of the
default pointing device. The default pointing device is usually a
mouse, but may also be a graphics tablet. The cursor is automatically
moved by the display micro-processor to reflect movements of the
default pointing device. The maximum size of the cursor icon is 64x64
pixels. The icon is cached by the display and modification of the
icon bitmap in memory will not affect the im~ge on the screen.

4.3.1

Load Cursor Command -

The load cursor command initializes the cursor to a new image.
The
image is specified by source image, source mask, and map parameters,
as in the copy area command. The display device places the origin of
the cursor rectangle at the current position of the pointing device.
The command has the following parameters:
o Cursor Source Image
o Cursor Source Off set
o Cursor Source Mask
o Cursor Map
o Cursor Attributes
whose definition follows.

4.3.1.1

CURSOR SOURCE IMAGE -

This cursor source image specifies the image of the cursor icon.
parameter is specified as in the copy area command.

4.3.1.2

This

CURSOR SOURCE OFFSET -

This parameter is specified as in the copy area command.

4.3.1.3

CURSOR SOURCE MASK -

This parameter is specified as in the copy area command, and generally

defines the shape of the cursor icon.

Page 47

workstation Graphics Architecture - PRELIMINARY
4.3.1.4

CURSOR MAP -

This parameter is specified as in the copy area command.
The map
parameter defines how the cursor is added to the screen.
If no map is
specified, the pixel values specified by source and source mask are
used to replace the pixels in the frame buffer at a destination
determined by the cursor reference point. That is, the cursor image
is written to the screen. Or, some systems may use a cursor that is
XOR'ed or OR'ed onto the screen, in which case a full map would be
specified.
In any case,
the screen state is not affected by the
cursor; when the cursor is moved, the original pixel values in the
area previously occupied by the cursor are restored.

4.3.1.5

CURSOR ATTRIBUTES -

The cursor attributes parameter defines how the cursor
The bits are defined as follows:
ATTRIBUTES<O>

displayed.

This bit specifies whether the cursor should blink
or not. When set to 0, no blinking of the cursor
occurs; when set to 1, the cursor is blinked at
an implementation-defined interval.

Workstation Graphics Architecture - PRELIMINARY
4.4

Page 48

DEVICE ORIENTED COMMANDS

4.4.1

Attach Cursor Command -

At any point in time,
the cursor may "track" the position and
movements of a single pointing device connected to the display. The
Attach Cursor command tells the display which device to use.
If "No
Device" is selected, the cursor is "disconnected" from all devices,
but its position may still be changed with the "Set Cursor Position"
command.
The command has only one parameter:
o

4.4.1.1

Device Type

DEVICE TYPE -

This specifies which device will control the movements of the
The defined parameter values are:
0 = no device
1 = mouse
3 = tablet

cursor.

Page 49

Workstation Graphics Architecture - PRELIMINARY
4.4.2

Set Cursor Position Command -

This command moves the cursor to the point specified by the parameter:
o

Location

If the cursor is attached to the mouse, the mouse's position will also
be set. The mouse only reports relative movement from a given point,
and so its 'position' may be defined by the user.
The tablet, however,
reports an absolute position on the tablet
surface,
and this position may not be redefined by the user.
Therefore, an error will occur if the user attempts to set the
position of the cursor when it is attached to the graphics tablet.
While the cursor may not be moved off the
invisible by loading a zeroed mask bitmap.

4.4.2.1

screen,

can

made

LOCATION -

The Location parameter specifies an XY offset from the origin
(upper-left-hand-corner) of the visible screen. The cursor's origin
is aligned to this point.
Cursor Movement Within the Visible Screen Area
Maximum X
Visible Screen

---+

I
v

+---------------------------------------+-------+
I
I
I
I
I
I
I
I
I
I

Cursor Position •

Maximum Y ----->+ ...•..••••.•••••••••••••.•••.•.•..••••• o-------+

+---------------------------------------+-------+
I
Cursor Image must be entirely visible
If the Location is not within the 4ectangle bounded by Maximum

and

Maximum Y, the Location will be clipped to the rectangle's boundaries,
such that the entire cursor is on the screen.

Page 50

workstation Graphics Architecture - PRELIMINARY

4.4.3

Get Cursor Position Command -

This command requests the current position of the cursor.
The cursor
position is returned in a two-word field in the command packet. There
are no input parameters;
however,
the command
two-word field in addition to the header.

4.4.4

packet

contains

Get Mouse Position Command -

This command requests the current position of the mouse.
is returned in a two-word field in the command packet.

The position
There are no

input parameters;
however, the command packet contains a two-word
field in addition to the header.
Normally, the reported position will
be identical to that accumulated by the device as it moves about
its
origin point.
However,
if the mouse is attached to the cursor (via
the "Attach Cursor" command), then the position reported will conform
to the position of the cursor (and thus will be constrained to the
boundaries of the visible screen, etc.).

4.4.5

Set Mouse Characteristics Command -

This command is used to specify how the cursor tracks mouse movement
when the cursor is attached to the mouse.
The mouse has no absolute
position;
its position is reported as relative movements from a given
point.
The mouse reports CHANGE in position as it moves.
Hence,
the mouse
position is updated by adding the accumulated change on to the current
position. Of course, the change could be positive or negative.
It is frequently useful to "scale" this accumulated change in order to
alter the mouse's "action" or "responsiveness".
The WGA has two
scaling algorithms from which to choose.
This incremental movement may be scaled in one of two ways:
Linearly
or Exponentially. A command modifier is used to select which one will
be used.
The parameters to Set Mouse Characteristics are either:
o

Tracking Ratio

o
o

Threshold
Scale Factor

or:

Page 51

workstation Graphics Architecture - PRELIMINARY
4.4.5.1

TRACKING RATIO -

The tracking ratio specifies the distance relationship between device
movement and cursor movement on the screen.
It consists of two 16-bit
integers, a multiplier and a divisor. Each unit of device movement
corresponds to
cursor movement

= (device_movement*multiplier)/divisor

units of movement on the screen.
linearly.

4.4.5.2

Thus, the cursor tracks

the

device

THRESHOLD And SCALE FACTOR -

When an exponentially tracking (accelerating)
cursor is selected,
tracking is linear (with a Tracking Ratio of 1:1) up until the
Threshold is reached. When the incremental mouse movement exceeds the
threshold, the move is scaled by the Scale Factor.
For example:
Threshold = 2
Scale Factor = 3
Mouse Movement
-5
-4
-3
-2

4.4.6

Resulting Cursor Movement
-11
-8
-5
-2

-1

0
1
2
3
4

0
1
2

5
8

Set Tablet Characteristics Command -

This command is used to specify how cursor tracking
is handled when
the cursor is attached to the tablet. The tablet reports an absolute
position which is scaled/modified by
the
Tracking
Ratio
or
Quantization Ratio.
As with the Set Mouse Characteristics command,
the scaling algorithm is selected by a command modifier.
Since the
tablet
always
reports absolute positions (which correspond to
well-defined locations on the tablet surface), exponential mapping is
not defined.
The parameters to Set Tablet Characteristics are either:
o

Tracking Ratio

Page 52

Workstation Graphics Architecture - PRELIMINARY
or:
o

Quantization Ratio

4.4.6.1

TRACKING RATIO -

This parameter specifies a linear scaling ratio as in
Characteristics command.

4.4.6.2

the

Set

Mouse

QUANTIZATION RATIO -

This scaling algorithm divides, or quantizes,
the surface of the
tablet and display screen into rectangular cells. When the tablet
pointer is moved from one "input cell" to another, the output position
is moved to the corresponding "output cell". Movement within a cell
is ignored: only when the ~ointing device moves from one cell to
another is a position change reported to the host. There are two
components in this Quantization Ratio:
the Input Grid SPacing (I.GSP)
and the Output Grid SPacing (0.GSP). The I.GSP specifies the cell
size on the tablet surface, and the O.GSP specifies the cell size on
the display surface.
Each GSP parameter has X and Y components, so the user may specify
different aspect ratios for the input and output. This is useful for
"correcting" the geometrical differences between different devices.
The position reported is the upper-left-hand
cell:
it is an Offset in Two-Space.

corner

the

output

This scaling algorithm may be used to eliminate the output-coordinate
"jitter" so common in high-resolution graphics tablets.
It can also
facilitate many graphical design applications.
I.GSP = 3 X 5
O.GSP = 5 X 5

Example:

Input Grid Spacing
0,0

3,0 6,0

+---+---+---+---+---+

Output Grid Spacing
0,0

5,0

10,0

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

0,5 +---+---+---+---+---+
0,10 +---+---+---+---+---+

I
I

0,15 +---+---+---+---+---+

I
0,5 +-----+-----+-----+-----+-----+

I
I

0,10 +-----+-----+-----+-----+-----+

+-----+-----+-----+-----+-----+
I
I

; . 1
,- ~

Page 53

wo rkstation Graphics Architecture - PRELIMINARY

+---+---+---+---+---+

4.4.7

+-----+-----+-----+-----+-----+

Get Tablet Position Command -

This command requests the current position of the graphics tablet
pointing device. The position is returned in a two-word field in the
command packet.

There are no input parameters;

however, the

command

packet contains a two-word field in addition to the header. Normally,
the reported position will be identical to that transmitted by the
physical tablet device. However, if the tablet device is attached to
the cursor (via the "Attach Cursor" command), then the position
reported will conform to the position of the cursor (and thus will be
constrained to the boundaries of the visible screen, etc.).

Set Pointing Device Event Reporting -

4.4.8

If the host wants to be notified (via an interrupt) when one or more
pointing devices moves, it may enable "event-reporting" for the
desired devices. If event reporting is enabled, the display will
interrupt the host at most once every l/60'th of a second for each
device that has moved since the last interval.
There is one
parameter:
o

4.4.8.1

Enable Flag

ENABLE FLAG -

The enable flag parameter enables or disables movement event reporting
for pointing devices. The flag is bit-encoded by device. A cleared
bit disables and a set bit enables event reporting
for
the
corresponding device. This parameter is encoded as follows:
All bits 0 = no device
bit 0 set = mouse
bit 2 set = tablet

4.5

MISCELLANEOUS COMMANDS

Page 54

w6ikstation Graphics Architecture - PRELIMINARY
4.5.1

Move Object Command -

The move object command is used to move arbitrary data from
of the system to another. Some examples are:
o
o

one

part

Loading fonts into display memory
Sending control strings to the keyboard

The sections of the system involved are:
o
o

VAX and Display Memory
Peripherals (Keyboard, USART ports, etc.)

Parameters for the move object command are:
o
o
o
o

Object Type
Object Length
Source
Destination

4.5.1.l

OBJECT TYPE -

The object type parameter indicates what format the data is in and the
source and destination types.
Data Format

Source

Destination

------

-----------

reserved
data words
character strings
character strings

memory
memory
peripheral

memory
peripheral
memory

4.5.1.2

OBJECT LENGTH -

Type

-----------

1
2

The object length parameter indicates
object to be moved.

4.5.1.3

the

length

bytes

SOURCE -

Depends on object type:

Memory:

Peripheral:

Address of word-aligned and word padded buff er
Device type value

the

wo~kstation

4.5.1.4

Graphics Architecture - PRELIMINARY

Page 55

DESTINATION -

Depends on object type:
o
o

4.5.1.5

Memory: Address of word-aligned and word padded buffer
Peripheral: Device type value

ERRORS -

Invalid Device Error:

Reference to an invalid peripheral

device type or an illegal operation for a specific device.

4 • 5 • 1. 6

4.5.1.6.1

NOTES -

DEVICE TYPES -

Device type values:
0 = no device
1 = mouse
2 = keyboard

3 = tablet
4 = USART - AUX 1
5 = USART - console
6,7 =reserved

4.5.1.6.2

BUFFERS -

Buffers must be word-aligned and word padded. Padding might be needed
because all buffer must be an even number of bytes long and an extra
byte might be needed at the end of the buffer.
(The VSlOO only
supports word transfers.)

4.5.1.6.3

DATA FORMATS -

Data formats that Move Object supports:
Word Data
Character String

Workstation Graphics Architecture - PRELIMINARY
4.5.1.6.3.l

Page 56

WORD DATA Format of Word Data Buff er
Object Type 1

+== =============+
Word 0

Word 1

Word 2

+===============+
+===============+

<- -

word data buffer address
word data buffer length in bytes

+===============+
1\\\\\\\\\\\\\\\1

+===============+
Word n

+===============+

n*2

Example of a ten word buffer:

+===============+
Word 0

Word 1

Word 2

Word 3

Word 4

+===============+
+===============+
+===============+
+===============+
+===============+
Word 5

Word 6

Word 7

Word 8

Word 9

+===============+
+===============+
+===============+
+===============+
I

+===============+

<--

buff er address
buffer length = 14H (20 decimal

.· · kstation Graphics Architecture - PRELIMINARY
4.5.1.6.3.2

Page 57

CHARACTER STRING Format of Character String Buffer
Object Type 2 and 3

+===============+

I Char 11 Count I 0

<--

+===============+

I Char 31 Char 21 2

character buffer address
character buff er length in bytes

+===============+

Example of an odd number of characters:

+===============+
I Char 11

I 0

<--

+===============+

character buffer address
character buff er length = 4

I Char 31 Char 21 2

+===============+

Example of an even number of characters:

+===============+
I Char 11

+===============+

I Char 31 Char 21 2

+===============+
I f i 11

I Char 4 I 4

+===============+

<--

(padding)

character buffer address
character buff er length = 6

Wd'rkstat ion Graphics Architecture - PRELIMINARY

Page 58

Sending Control Strings to the Keyboard from the Host Processor
The WGA Move Object Command will be used to send keyboard control
strings to the LK201 keyboard. With this support the host will be
able to control the keyboard;
for example:
o
o
o
o

Enable/Disable keyclicks
Ring the bell
Set the volume of the bell
Control the LEDs
Example #1:
Ring the bell (A7H).
Move Object Parameters:
Object Type =
Object Length =
Source =
Destination =

2 (memory to peripheral)
2 (length of character buffer)
Address of character buff er
2 (Keyboard)

Character Buffer:

+===============+

+ 0

+===============+

Example #2:
Enable the bell (23H) and set the bell volume to 4 (84H).
Move Object Parameters:
Object Type =
Object Length =
Source =
Destination

2 (memory to peripheral)
4 (length of character buffer)
Address of character buff er
2 (Keyboard)

Character Buffer:

+===============+
+

+===============+
o I 84 + 2

+===============+

~ ~

~orkstation

4.5.2

Graphics Architecture - PRELIMINARY

Page 59

Report Status Command -

The Report Status command requests the return of information about the
current status and configuration of the display processor. The
information returned includes:
o Device Type
o Device Version
o Micro-code Version
o Visible Screen Frame Buffer Bitmap
o Free Frame Buff er Memory and Byte Length
o Free Program Memory Space Address and Byte Length
o Host Memory Space Base Address and Byte Length
All addresses are in the physical address space of the display
microprocessor.
The host uses the base addresses reported by this
command to map addresses in its space to addresses in the display's
space. The returned parameters are described in more detail below.

4.5.2.1 DEVICE TYPE - A 32-bit value that indicates the
device connected. This parameter is encoded as follows:
31

24123

16115

type

+-------------------------------------+
S2
Sl
Option Number
+-------------------------------------+

The low order word is a 16 bit binary option number and the high order
two bytes are the option suffix encoded in ASCII (i.e. the "AA" in
VS100-AA}. Note that the suffix is encoded in standard VAX-11 address
order (an even address points to the low-order byte).
Example:
VS100-AB

I binary

I decimal

I hexadecimal

------------+-----------+-----------+----------------------

Opt ion
ASCII "A"
ASCII "B"

0110 0100
0100 0001
0100 0010

100
65
66

Device type is: 42410064 Hex

64
41
42

Page 60

Graphics Architecture - PRELIMINARY
4.5.2.2

DEVICE VERSION -

A 16-bit binary number indicating what changes have been made
display hardware (the revision level).

4.5.2.3

the

FIRMWARE VERSION -

Two (2) 8-bit binary values representing the major and minor version
numbers of the firmware currently loaded in the display. Again, they
are stored in standard address order.
15

817

+----------------------------------+
Minor

Major

+----------------------------------+
When the Report Status command is processed
intialization, this field is returned as zero.

the

ROM

during

A bitmap specification indicating the address and size of
buffer bitmap from which the visible screen is refreshed.

the

frame

4.5.2.4

4.5.2.5

VISIBLE SCREEN FRAME BUFFER BITMAP -

FREE FRAME BUFFER MEMORY -

The base address and length of storage in the frame buff er memory not
used by the visible screen. This memory is available for use by the
host software.

4.5.2.6

FREE PROGRAM MEMORY SPACE -

The base address of additional memory in the
microprocessor's
program/data space that is available for use by the host software.
When the ROM-based report status command
is
executed
during
intialization (see below,
section on initialization and initial
display requirements), the base address of microprocessor program/data
space is returned.
This space could be used for storing frequently
used data. The size of this space will likely be much smaller than
the off-screen framebuf fer space.

Graphics Architecture - PRELIMINARY
4.5.2.7

Page 61

HOST MEMORY SPACE BASE ADDRESS -

The base address, in the display microprocessor's space, of the memory
space shared with the host. Host addresses passed to the display must
be biased by this value. The display also returns this value to the
host in CSRs during initialization.

4.5.3

No Operation Command -

The no operation command simply causes the display to read the
and respond with a command completion interrupt.
There
parameters.

packet
are no

Page 62

workstation Graphics Architecture - PRELIMINARY
5.0

CONTROL AND STATUS REGISTERS

The workstation graphics display communicates with the host via an
interface that utilizes a set of control and status registers (CSRs).
These CSRs are implemented as shared memory and not as hardware
registers.
Therefore, a strict protocol must be obeyed to avoid race
conditions when using the CSRs. The CSRs and their programming are
described in the following sections.

5.1

Control And Status Register (CSRO)

The control and status register is the main control register
transmitting commands to the display processor.
Its format is:
15
6 5
1 0

for

+---------------------------------------+
Reserved
IIEIFUNCTIONIGOI
+---------------------------------------+

CSRO

Control and Status Register

where the fields have the following uses:
GO = CSRO<O> The GO bit is set by the host to indicate that the
next command has been set up and the device should
process the command.
FUNCTION = CSR0<5:1> The function field is set by the host to
indicate the display command to execute, i.e.,
this is the command opcode.
Functions 0 through
15 are implementation-independent definitions and
16 through 31 are implementation-dependent.
IE

= CSR0<6>

The interrupt enable bit, when set, allows the
display to interrupt the host when a command is
completed, a mouse event occurs, a keyboard event
occurs, or an error occurs. There is only one
interrupt vector; the host determines the event
by scanning the interrupt reason flags.
(Note:
when the host initializes or reboots,
it should
clear IE until it is capable of receiving display
interrupts.)

Use of CSRO must follow a strict protocol. When GO is set to zero,
the device is ready for a new command. Only the host can modify
FUNCTION and GO when GO is zero.
When GO is set,
the device is
processing a command. Only the device can modify FUNCTION once GO is
set.
At initialization, the device sets GO and FUNCTION to zero to indicate
that it is ready for processing. Host can then load FUNCTION and set

GO to perform an operation.
donQ.

The device will clear

both

when

Page 63

workstation Graphics Architecture - PRELIMINARY
5.2

Interrupt Reason Register

The interrupt reason register is written by the display to indicate
the event causing an interrupt.
Interrupt reasons are bit-encoded;
each bit indicates a different condition. Bit 15 of the interrupt
reason register is an error bit.
If set, the low-order 15 bits
contain a display error code. Errors that also have bit 14 set are
diagnostic errors.
The format of the Interrupt Reason Register is:

15 14

+---------------------------------------+
IE ID I
INTERRUPT REASON
+---------------------------------------+

CSRl

Interrupt Reason Register
Like CSRO, use of CSRl must be synchronized. When interrupt reason is
zero,
the device can complete a request. Only the device can write
CSRl when it is set to zero. When interrupt reason is nonzero,
the
host has not yet examined an event and the display may not write to
CSRl. After the host has processed an interrupt,
it sets CSRl to
zero.
The device can then write the interrupt reason for the next
event.

5.3

Device Event Register

The Device Event Register indicates events on keyboard or pointing
device buttons attached to the display. The events are generally
reported as up or down transitions on numbered keys.
The format of
the Device Event Register is:
15
13 12
9 8 7
0

+-------------------------------------+
I Reserved I DEV I T I KEYCODE
+-------------------------------------+

CSR2

Keyboard Receive Register

where:
KEYCODE=CSR2<7:0> This field is written by the display to indicate
on which key a transition has occurred. The key
codes are integers from 0 to 255. The meaning of
these codes depends on the particular device
involved.
T=CSR2<8>

The

transition indicator specifies whether an
up
transition (T=O) or a down transition (T=l) has
occurred. Not all of the button-event devi c es

connected

the

display may operate in exactly

the same fashion,
though.
Some keys on the
keyboard,
for
example,
may
only
report
down-transitions, while the buttons on the mouse

Page 64

workstation Graphics Architecture - PRELIMINARY

report both up and down transitions. The user
should refer to device-specific documentation for
the particulars.
DEV=CSR2<12:9>

The device field specifies the device responsible
for generating the event. This field is encoded
as follows:

0 = no device
1 = mouse
2
keyboard
3 = tablet
4 = USART - AUX #1
5 = USART - console

5.4

Function Parameter Register

The function parameter register consists of two 16-bit CSRs, CSR3 and
CSR4.
This 32-bit register pair is loaded with the address of a
packet to be transmitted to the display.
When a command packet (or list of linked commands)
is completed (or
aborted), the number of packets successfully completed is written into
this register pair.
During initialization, the device writes in CSR3/CSR4 the location at
which host memory is mapped within its local address space. The host
must then offset all host memory addresses by this amount when
communicating with the display.
15

+---------------------------------------+
PACKET (LOW ADDRESS)
+---------------------------------------+

CSR3
CSR4

PACKET (HIGH ADDRESS)

+---------------------------------------+
Function Parameter Register

5.5

Device Position Register

The Device Position Register contains the current position of the
device whose movement generated the most recent interrupt.
If the
device is attached to the cursor, the interrupt reason will be "Cursor
Moved";
the device position will be the same as the cursor position.
If the device is not attached to the cursor, then the
interrupt will
indicate

on:

movement

that particular device.

See al3o the sections

Page 65

workstation Graphics Architecture - PRELIMINARY
Interrupt Reason Register
Set Cur~or Position
Get Cursor Position
Get Mouse Position
Get Tablet Position
Set Pointing Device Event Reporting
The format of the Device Position Register is:
15

+---------------------------------------+
I
XPOS
I
+---------------------------------------+
YPOS
I
+---------------------------------------+

CSR5
CSR6

Current Device Position
The Device Position Register should only be examined immediately
following a device event interrupt. These interrupts will occur at
most once every l/60'th of a second.
The display will update the
Device Position Register prior to interrupting. Because the update
requires two non-interlocked 16-bit writes, it would be possible for
the host to read a new XPOS with an old YPOS if the registers were
read at random times. Therefore, when servicing a device interrupt,
the host should copy the device position to a local storage area.

5.6

Interrupt Vector Address Register

The Interrupt Vector Address Register is set by the host to the
address of an entry in the UNIBUS interrupt vector block which
contains the PC and PSL of the device interrupt service routine. When
the device wants to interrupt the host, these values are used to
invoke interrupt servicing.
0

CSR7

+---------------------------------------+
Interrupt Vector Address
+---------------------------------------+

5.7

Initialization And Initial Display Requirements

The workstation graphics display must contain a ROM with code capable .
of processing a minimum of five immediate commands; that is, five
function codes that can be loaded into the function field of CSRO.
These commands are:
1.

INIT (FUNCTION

1).

The INIT

conunand

causes

the

display

micro-processor, if currently in operation, to halt execution
and re-enter the initial ROM routine. An identical function

Page 66

workstation Graphics Architecture - PRELIMINARY

is performed on power-up of the display module. When the ROM
is entered, it automatically stores the location at which
host memory is mapped within the microprocessor's address
space in the Function Parameter Registers,
CSR3 and CSR4.
When the host communicates a host memory address to the
display, it must first add this value to the host address.
This initial address is thus required by the host in order to
transmit any host-mapped command packets to the display.
2.

SEND PACKET (FUNCTION=2). The SEND PACKET command causes the
display to read and process a command packet;
the address of
the command packet is loaded by the host
into the Function
Parameter Registers, CSR3 and CSR4, prior to setting FUNCTION
CSRO .

START DISPLAY (FUNCTION=3). The START DISPLAY command starts
execution of the display micro-processor at the address
specified in the Function Parameter Register (CSR3/CSR4).
The start display command is executed during initialization
to begin execution of display microcode.
It causes the
initial display ROM to invoke the firmware at the specified
address.
Start display is an immediate function and not a message.
It
is executed by loading the start display opcode into the CSRO
function field and setting the GO bit. The firmware responds
with the "Started Executing" interrupt reason. The firmware
start address is contained in the Function Parameter Register
pair.

ABORT (FUNCTION=4).
The ABORT function code stops any
command being executed by the display. The display responds
with the "command aborted" interrupt reason, and returns the
number of successfully completed command packets in the
Function Parameter Register.

EXECUTE POWER UP SEQUENCE (FUNCTION=5). This function
causes -the display to branch to its power-up sequence.
is a more "drastic"
initialization
than
INIT.
implementations may, for example, go in to a "self-test"
here.

code
This
Some
mode

In addition, the initial display code must be capable of reading and
processing two command packet opcodes as transmitted by a send packet
request. These commands are:
1.

Report Status
returns information about the
status and addressing environment to the host,

Move Object
allows
down-line
loading
micro-processor code into display local memory.

display's
display

Workstation Graphics Architecture - PRELIMINARY
5.8

Page 67

Aborting A Request

Since some commands can take seconds or minutes to complete and
commands can be linked together,
it must be possible to abort the
current command sequence. To abort an operation, the host loads the
ABORT function code into the CSRO function field.
Note that this
violates the rules for access to CSRO.
It is possible that after the
ABORT is written and before the function field is examined by the
device, the command will complete and the device will write a zero in
the CSRO function code. Therefore, when processing an abort, the host
driver must be capable of accepting two interrupt reasons:
aborted
and command completion.
The driver must also be able to ignore an
aborted interrupt reason when no command is in progress.

Workstation Graphics Architecture - PRELIMINARY
6.0

Page 68

COMMAND PACKET FORMATS

This section describes the formats of the display command packets.
Each command packet has two parts, a header and a command message.
The header format is common among all commands and has the following
format:
8 7

+---------------+

I qual. I opcode! O

+---------------+
I

modi f i er s

I 2

modifiers

link

I 6

link

+---------------+

+- ---------- ----+
+---------------+
+---------------+

The opcode is an 8-bit value. Associated with the opcode is an 8-bit
qualifiers field that contains command-independent qualifiers. In
addition, many of the commands have command-dependent options

the

specification of one or more parameters in the packet. For each
parameter with specification options, the 32-bit opcode modifiers
field indicates which option was chosen for that command. Although
different options for a single parameter may require different size
specifications, all packets for each opcode are of fixed size. Space
is left in the packet for the largest size of each parameter offering
several options. Thus, for each opcode, each parameter always starts
at the same offset from the top of the packet, independent of the
specification options chosen for the parameters.
The single command-independent qualifier currently defined is the Wait
For

Refresh

qualifier, specified by bit zero of the 8-bit qualifiers

field.
If this bit is set to 1, the packet is processed as normal;
however, physical screen operations are syncronized with the vertical
blanking interrupt.
The link field is the 32-bit address of the next packet to be
processed, if any. By linking packets, the host can initiate several
commands with a single interaction. A zero value in the link field
terminates the list. If any error occurs, the list is aborted at the
point of error. The device returns the count of the number of packets
successfully processed in the function parameter register, CSR3. For
example, if the first packet is in error, a zero value will be

returned,

on.

the

second is an error, a one will be returned, and so

Page 69

Workstation Graphics Architecture - PRELIMINARY
6.1

Copy Area Command Packet
0

+===============+

(low) I 10

I ADDRESS

+SOURCE
IMAGE
BITMAP

(high) I 12

+---------------+
I SIZE

( x)

( y)

I 16

+( z)

SOURCE
OFFSET

I 14

I 18

+===============+
I OFFSET

(x)

I 20

( y)

I 22

+===============+
[bitmap]

+===============+
I ADDRESS

(low) I 24

+===============+
CONSTANT

+---------------+
I
I
I
I
I
I
I

( y)

SOURCE
MASK

+I OFFSET

( x)

+===============+

I 30
I 32

( y)

I EXTENT

(x)

I 34

I 36

+---------------+
+-

I 38

-+
( y)

I 40

(x)

( y)

I
-+

< z)

+===============+
I ALIGNMENT (x) I
I OFFSET

-+
<y > I

[halftone]

+---------------+
+-

I SIZE

+===============+

-+
( z)

+---------------+

[constant]

I 28

+===============+

(high) I 26
( x)

(low) I

(high) I

+---------------+
I SIZE

I ADDRESS

I
I
I
I
I
I
I

+===============+

I
I
I
I
I
I
I
I
I
I
I
I
I

+---------------+
I EXTENT
II

I
-I
( y) I
(x)

+===============+

[sub-bitmap]

[rectangle]

Page 70

Workstation Graphics Architecture - PRELIMINARY
+===============+
I ADDRESS (low) I 42
-+
+(high) I 44
I
DESTIN. +---------------+
( x) I 46
IMAGE
I SIZE
-+
BITMAP +( y) I 48
-+
+( z ) I 50
+===============+
(x) I 52
DESTIN. I OFFSET
-+
OFFSET +( y > I 54
+===============+
+===============+
I ADDRESS (low) I 56

MAP

(high) I 58
+===============+
[map address]

+===============+
I ADDRESS
(low) I 60
CLIPPING+-+
RECTANG. I
(high) I 62

+---------------+
I COUNT

+---------------+

<reserved>
66
+========= ======+
[rectangle list]

+===============+
I LITERAL MAP

+===============+
[literal map table]
+===============+
I OFFSET
( x) I
-+

( y)

+---------------+
I EXTENT

(x) I

I
+===============+
[literal rectangle]
( y)

Page 71

Workstation Graphics Architecture - PRELIMINARY
6.2

Draw Curve Command Packet

The draw curve command packet is identical to the
packet with the following additional fields:

copy

area

command

+===============+
I ADDRESS

(low) I 68

(high) I 70

+PATH

+---------------+
I COUNT

+===============+

I PATTERN LENGTH I 74

+---------------+

PATTERN
76
PATTERN
STRING +---------------+
I PATTERN MULT. I 78

+===============+
+===============+

I PATTERN POSIT. I 80

PATTERN +---------------+
STATE
I PATTERN COUNT I 82

(low) I
-+

(high) I

+===============+

[literal]

[state pointer]

+===============+

(low) I 84
-+

(high) I 86

CONSTANT

+---------------+

I SIZE

( x)

I 88

+===============+
I OFFSET

( x)

I 94

(y)

I 96

+===============+
[bitmap]

(high) I
( x)

+===============+
[constant]

-+
( z)

+===============+

-+
( y)

I 90

I 92

(low) I

-+
( z)

I ADDRESS

I SIZE

+===============+
+---------------+

+---------------+
( y)

OFFSET

I ADDRESS

+===============+
I ADDRESS

SECOND
SOURCE

+===============+

+===============+
I ALIGNMENT (x) I

I OFFSET

(y > I

+===============+
[halftone]

Page 72

Workstation Graphics Architecture - PRELIMINARY
6.3

Print Text Command Packet

The print text command is similar to the copy area command,
and has
the following format.
The differences are in specification of font
(instead of source mask), the alternative specification of initial
destination offset (instead of destination offset), and the additional
parameters.
15

+===============+
I ADDRESS

(low) I 10
-+

+===============+
I

CONSTANT

+---------------+

+===============+
I ADDRESS

(high) I

(high)! 12
SOURCE
IMAGE

+---------------+

+---------------+
I SIZE

( y)

+===============+
I

20
<reserved>

+===============+

+===============+
I ADDRESS

(low) I 24

(high)! 26

+---------------+
MASK
FONT

28
30
32
34
36
38
40

+===============+
+===============+
I ADDRESS

(low) I 42

(high) I 44

DESTIN.

+---------------+

BITMAP

IMAGE

I SIZE

( x)

I 46

( y)
+-

I 48
-+

+===============+
I

[source font]

(x) I

(low) I

I
-+

(z > I

+===============+
I ALIGNMENT (x) I

I OFFSET

( y)

+===============+

[constant]

[halftone]

Page 73

Workstation Graphics Architecture - PRELIMINARY
(z)

I 50

+====~==========+

+===============+
INITIAL I OFFSET
( x) I 52
DESTIN. +-+
OFFSET
(y) I 54
+===============+
[literal]

+===============+
I ADDRESS (low) I

+===============+
I ADDRESS (low) I 56

+===============+
I LITERAL MAP

MAP

(high)! 58
+===============+
[map address]

+===============+
!ADDRESS
(low)I 60
CLIPPING+-+
RECTANG.I
(high)! 62
+---------------+
I COUNT
64

+---------------+

<reserved>
66
+===============+
[rectangle list]

TEXT
STRING

+===============+
I ADDRESS (low) I 68
+-+
(high) I 70
+---------------+
72
I COUNT
+===============+

+===============+
I ADDRESS (low) I 74
CONTROL +-+
(high) I 76
STRING
+---------------+
78
I COUNT
+===============+
INTERCHAR
PAD

+===============+
80
I COUNT
+===============+

SPACE
PAD

+===============+
82
I COUNT
+===============+

<high) I
+===============+
[offset pointer]

+===============+
[literal map table]
+===============+
I OFFSET
( x) I
+-

-+
( y)

+---------------+
I EXTENT

(x)

+( y)

+===============+
[literal rectangle]

Page 74

Workstation Graphics Architecture - PRELIMINARY
6.4

Fill Area Command Packet

The command packet for the fill area command has the following format:
0

+===============+
I

CONSTANT

I 10

+---------------+
SOURCE
IMAGE
BITMAP

I
I
I
I
I
I
I

12
14

+===============+
I ADDRESS

(low) I

(high) I

+---------------+
I SIZE

( x)
( y)

18
20

+- <reserved>

+---------------+

( z)

+===============+
'
I ALIGNMENT

(x)

I
-+

+===============+

I OFFSET

( y)

+---------------+
[halftone]

[constant]

+===============+
I ADDRESS

+DESTIN.
IMAGE
BITMAP

(low) I 24
-+

(high) I 26

+---------------+
I SIZE

( x)

( y)

I 30

( z ) I 32

+===============+

DESTIN. I OFFSET
OFFSET +-

( x)

(y)

I 34

I 36

+===============+
+===============+

+===============+

I ADDRESS

MAP

(low) I 38
-+

I LITERAL MAP

(high) I 40

+===============+
[map address]

+===============+
I OFFSET

( x)

+===============+
[literal map table]

I 42

CLIPPING!
(y) I 44
RECT.
+---------------+
I EXTENT
(x) I 46
+-+
(y) I 48
I

+===============+

Workstation Graphics Architecture - PRELIMINARY

+===============+
I ADDRESS

PATH

(low) I 50
-+

(high)! 52

+---------------+
I COUNT

+===============+

Page 75

Page 76

Workstation Graphics Architecture - PRELIMINARY
6.5

Flood Area Command Packet

The command packet for
format:

the

15
0
+===============+
I
CONSTANT
I 10
+---------------+
12

SOURCE
IMAGE
BITMAP

flood

area

command

+===============+
I ADDRESS ( low) I

(high) I

+---------------+
I SIZE

(x)

( y)

+===============+
I
I 20
+- <reserved> -+
I
22

+---------------+
[constant]

( z) I
+---------------+
I ALIGNMENT (x) I

I OFFSET

(y >

+---------------+
[halftone]

+===============+
I ADDRESS (low) I 24
+-

(high) I 26
DESTIN. +---------------+
( x) I 28
IMAGE
I SIZE
+-+
BITMAP
( y)

I 30

-+
( z)

I 32

+===============+
SEED
POINT

+===============+
I OFFSET
(x) I 34
+-

(y) I 36
+===============+
+===============+
I OFFSET
( x) I 3 8
+-

CLIPPING!
(y) I 40
RECT.
+- - -------------+
I EXTENT
(x) I 42
+-

-+
(y)

I 44

+===============+
+===============+
i ADDRESS (low) I 46
BOUNDARY+-+
MAP
I
(high) I 48

+===============+
!BOUNDARY (low)
+-

I
-+

I BOUNDARY (high) I

has

the

following

Workstation Graphics Architecture - PRELIMINARY

+===============+

[boundary address]

+===============+

[literal boundary]

Page 77

Page 78

Workstation Graphics Architecture - PRELIMINARY

6.6

Load Cursor Command Packet

+===============+

(low) I 10

I ADDRESS

+CURSOR
SOURCE
IMAGE
BITMAP

(high) I 12

CONSTANT

+---------------+

+---------------+
I SIZE

( x)

( y)

( z)

I OFFSET

( x)

I 14

I SIZE

I 16

I 20

(y) I 22

+===============+

( z)

+===============+

( x)

CURSOR
SOURCE
MASK

I 28

-+
( y)

I 30

-+
( z)

I 32

+---------------+
I OFFSET

( x)

I 34

(y) I 36

+---------------+

I EXTENT

I EXTENT
II

( x)

I 38

-+
( y)

I 40

+===============+
[sub-bitmap]

MAP

+===============+
[rectangle]

+===============+

(low) I 42
-+

I LITERAL MAP

(high) I 44

CURSOR
ATTRIB.

I
-I
( y) I
( x)

+===============+
I ADDRESS

+===============+

[map address]

[literal map table]

+===============+
I VALUE

I ALIGNMENT

(x)

I OFFSET

( y)

+===============+
[halftone]

+---------------+
I SIZE

+- <reserved>

(high) I 26

+===============+

[constant]
(low) I 24

+===============+

[bitmap]

I ADDRESS

( x)

(y)

+===============+

(high) I

I 18

+===============+
I ADDRESS (low) I
+-+
+---------------+

+===============+
SOURCE
OFFSET

+===============+

Workstation Graphics Architecture - PRELIMINARY
6.7

Page 79

Attach Cursor Command Packet

+========:======+
DEVICE

6.8

I VALUE

+===============+

Set Cursor Position Command Packet

+===============+
I OFFSET

(x)

LOCATION+-

I 10

(y)

I 12

+===============+

6.9

Get Cursor Position Command Packet

This command has a two-word field in which

the

display

returns

the

display

returns

the

current position of the cursor.

+===============+

CURRENT I RETURN
CURSOR +POSIT I ON I OFFSET

(x)

I 10

(y) I 12

+===============+

6.10

Get Mouse Position Command Packet

This command has a two-word field in which
current position of the mouse.
+===~===========+

CURRENT I RETURN
MOUSE
+POSITION! OFFSET

(x)

I 10
-+

(y)

I 12

+===============+

Workstation Graphics Architecture - PRELIMINARY
6.11

Set Mouse Characteristics Command Packet

+===============+

TRACKING! MULTIPLIER
I 10
RATIO
+---------------+
I DIVISOR
12

+===============+
[Linear Tracking]

6.12

Page 80

+===============+
I

THRESHOLD

I 10

I SCALE FACTOR

+---------------+

+===============+

[Exponential Tracking]

Set Tablet Characteristics Command Packet

+===============+

+==============+
TRACKING! MULTIPLIER
10
RATIO
+--------------+
I DIVISOR
12

X I 10

INPUT GSP

Y I 12

+==============+

+---------------+

+- <reserved> -+

+---------------+

OUTPUT GSP X I 14
OUTPUT GSP Y I 16

+--------------+

[Linear

6.13

INPUT GSP

+---------------+

[Quantized

Tracking]

Get Tablet Position Command Packet

This command has a two-word field in which

the

display

returns

the

current position of the tablet.

+===============+

CURRENT I RETURN
TABLET +POSITION! OFFSET

(x)

I 10

(y)

I 12

+===============+

6.14

Set Pointing Device Event Reporting

This command enables or disables periodic reporting
tablet movement.
ENABLE
FLAGS

+===============+
I VALUE

+===============+

mouse

and/or

Workstation Graphics Architecture - PRELIMINARY
6.15

Page 81

Move Object Command Packet

+===============+
TYPE
LENGTH

I VALUE

I 10

+===============+
I VALUE

I 14

+===============+

OBJECT
I ADDRESS
ADDRESS +-

(low) I 16
-+

(high) I 18

DESTIN. I ADDRESS
ADDRESS +-

(low) I 20

+===============+
-+

(high)! 22

+===============+

6.16

Report Status Command Packet

For the report status command, the packet contains no input parameters
except for the opcode header. However, the packet contains space for
the following information that is filled by the display.

+===============+
DEVICE
TYPE

I VALUE

(low) I 10

(high) I 12

DEVICE +===============+
VERSION I VALUE
I 14

+===============+

MICRO- +===============+
CODE
16
I VALUE
VERSION +===============+

+===============+
I ADDRESS

(low) I 18

(high) I 20
VISIBLE
FRAME
+---------------+
( x) I 22
BUFFER I SIZE
-+
BITMAP +( y)

I 24

+( z)

I 26

+===============+
+===============+

I ADDRESS
FREE

(low) I 28

•

Workstation Graphics Architecture - PRELIMINARY
FRAME
BUFFER
MEMORY

Page 82

(high)I 30
+-----~---------+

I COUNT

(low) I 32

(high) I 34

+===============+
+===============+
I ADDRESS

(low) I 36

(high) I 38
FREE
I
PROGRAM +===============+
SPACE
(low) I 40
I COUNT
MEMORY +-+
(high) I 42
I

+===============+

+===============+
I ADDRESS

(low) I 44

(high) I 46

I COUNT

(low) I 48

+HOST
MEMORY
SPACE
BASE

+===============+
+-

(high) I 50

+===============+
6.17

No Operation Command Packet

This command has no parameters, and consists only of the packet opcode
header.

Workstation Graphics Architecture - PRELIMINARY
7.0

CONSTANTS, OPCODES, MODIFIERS, AND ERROR CODES

This section specifies the constants used as opcodes,
error codes for onyx commands and return codes.

7.1

Page 83

modifiers,

and

Control And Status Register 0 Function Codes

The following function codes are defined for CSRO functions . .
Implementation-independent functions:
INIT = 1
SEND PACKET = 2
START DISPLAY = 3
ABORT-= 4
EXECUTE POWERUP_SEQUENCE = 5
VSlOO Implementation-dependent functions:
BBA ON = 16
BBA-OFF = 17
SET-INFINITE RETRIES = 18
SET-FINITE RETRIES = 19

7.2

Command Packet Operation Codes

The following are the values of the 8-bit operation codes specified in
the low byte of the first word transmitted in a device command. These
commands are all interperted by display microcode or firmware.
NO OPERATION
COPY AREA
DRAW-CURVE
PRINT TEXT
FLOOD AREA
LOAD CURSOR
SET CURSOR POSITION
ATTACH CURSOR
GET CURSOR POSITION
MOVE OBJECT
REPORT STATUS
FILL AREA
GET MOUSE POSITION
SET-MOUSE-CHARACTERISTICS
GET-TABLET POSITION
SET-POINTING DEV REPORTING
SET-TABLET CHARACTERISTICS

= 0
= l
= 2
= 3
= 4

= 5
= 6
= 7
= 8
= 9
= 10
= 11
= 12
13

= 14
= 15
16

The following commands are interpreted only by display ROM,
and are
thus given different operation codes,
even though there is some
overlapping of the commands.
This prohibits accidental interpretation
of a command intended ~or a different machine state.

Workstation Graphics Architecture - PRELIMINARY

Page 84

MOVE OBJECT = 128
REPORT-STATUS= 129

7.3

Command Packet Operation Modifiers

The modifiers field specifies,
for each command, which optional
parameters are specified and what format is present for parameters
with multipie formats.
The following modifiers are defined below,
listed separately for each command.

7.3.1

Copy Area Command Modifiers -

PARAMETER

FIELD

SOURCE

mod<2: O>

VALUE

MEANING

0
1
2

constant
bitmap
halftone

0
1

rectangle
bitmap

SOURCE
MASK

mod<5:3>

DEST. OFF.

mod<8:6>

MAP

mod<Ll:9>

0
1
2
3
4

identity map
source map address
source map . literal
function code address
function code literal

CLIPPING
RECTANGLES

mod<l4:12>

0
1
2

none
literal rectangle
rectangle list addr.

7.3.2

(not used)

Draw Curve Command Modifiers VALUE

MEANING

mod<2:0>

0
1
2

constant
bitmap
half tone

SOURCE
MASK

mod<5:3>

rectangle
bitmap

DEST. OFF.

mod<8:6>

MAP

mod<ll:9>

PARAMETER

FIELD

SOURCE

(not used)

identity map

Workstation Graphics Architecture - PRELIMINARY

Page 85

1
2
3
4

source map address
source map literal
function code address
function code literal

CLIPPING
RECTANGLES

mod<l4:12>

0
1
2

none
literal rectangle
rectangle list addr.

DRAWING
MODE

mod<l5>

0
1

Solid Segment
Dashed/Patterned Segment

PATTERN
STATE

mod<l7:16>

0
1
2
3

Literal pattern state
Indirect pattern state
Update literal pattern sta
Update indirect pattern st

PATTERN
MODE

mod<l9:18>

0
1
2
3

No secondary source {Dashe
Constant secondary source
Bitmap secondary source
Halftone secondary source

7.3.3

Print Text Command Modifiers VALUE

MEANING

PARAMETER

FIELD

SOURCE

mod<2:0>

0
1
2

constant
source font
halftone

MASK FONT

mod<5:3>

0
1

no mask
mask font supplied

DEST. OFF.

mod<8:6>

0
1
2
3

dest offset
dest off set
update dest
update dest

MAP

mod<ll:9>

0
1
2
3
4

identity map
source map address
source map literal
function code address
function code literal

CLIPPING
RECTANGLES

mod<l4:12>

0
1

none
literal rectangle
rectangle list addr.

2
TEXT

mod<l5>

0
1

mod<l6>

0
1

STRING

CONTROL
STRING

literal
indirect
literal
indirect

8 bit characters

16 bit characters

no control string

control str i ng

Page 86

Workstation Graphics Architecture - PRELIMINARY
7.3.4

Fill Area Command Modifiers -

PARAMETER

FIELD

SOURCE

mod<2:0>

SOURCE
MASK

mod<5:3>

(not used)

DEST. OFF.

mod<8:6>

(not used)

MAP

mod<ll:9>

0
1
2
3
4

identity map
source map address
source map literal
function code address
function code literal

CLIPPING
RECTANGLE

mod<l4:12>

0
1

none
literal rectangle

7.3.5

VALUE

MEANING
constant
(not used)
half tone

0
1
2

Flood Area Command Modifiers -

PARAMETER

FIELD

SOURCE

rnod<2:0>

VALUE

MEANING

constant
(not used)
half tone

2
SOURCE
MASK

rnod<5:3>

(not used)

DEST. OFF.

mod<8:6>

(not used)

MAP

mod<ll:9>

(not used)

CLIPPING
RECTANGLE

mod<l4:12>

BOUNDARY MAP

rnod<l5>

none
literal rectangle

literal
pointer

7.3.6

Load Cursor Command Modifiers

PARAMETER

FIELD

VALUE

MEANING

SOURCE

mod <2:0>

0
1
2

constant
bitmap
halftone

___/

Page 87

workstation Graphics Architecture - PRELIMINARY

SOURCE
MASK

mod<5:3>

0
1

DEST. OFF.

mod<8:6>

MAP

mod<ll:9>

(not used)
0
1

2
3
4

7.3.7

identity map
source map address
source map literal
function code address
function code literal

Set Mouse Characteristics Command Modifiers -

PARAMETER

FIELD

TRACKING

mod<2:0>

7.3.8

rectangle
bitmap

VALUE
0
1

MEANING
Linear
Exponential

Set Tablet Characteristics Command Modifiers -

PARAMETER

FIELD

TRACKING

mod<2:0>

VALUE

MEANING

0
1

Linear
Quantized

Workstation Graphics Architecture - PRELIMINARY
7.4

Page 88

Interrupt Reason Values

The following are the values returned by the display in the
Interrupt
Reason Register (CSRl) following display interrupt.
Interrupt reasons
(bit 15 = 0) are unary encoded.
Errors
(bit 15 = 1)
are binary
encoded.

*
* WGA Completion Codes (15 reserved)
*
INT ID
INT-CD
INT-SE
INT-BE
INT-CM
INT TM
INT-MM
INT-PD

equ
equ
equ
equ
equ
equ
equ
equ

$0001
$0002
$0004
$0008
$0010
$0020
$0040
$0080

Initialisation Done
Command Done
Started Executing
Button Event
Cursor Moved
Tablet Moved
Mouse Moved
Powerup Done

* These are the error messages generated by the
* VSlOO, VS125 and VS300. Some errors are specific to
* one device. The following key indicates the error code's
* status for each device:
*
IS generated by this device
*
*
IS NOT generated by this device
*
+
New,
was not defined in previous versions
*
I
Changed,
Redefined from previous versions
*
*
* The key contains two characters. The first indicates the status

* for the VSl00/125, and the second indicates the status for the VS300.

* For example:
*
*- Error code is generated by VSl00/125 only
*
- * Error code is generated by VS300 only
*
*
** Both VSl00/125 and VS300 generate this code

*
*

VSl00/125 only, code has new definition
VS300 only, new error code added
VSl00/125 and VS300, new error code added

* WGA Hardware Error Codes (32 reserved)
*
* Error Mnemonic
NO ERROR
ERR BASE
ERR-NYI
ERR-IFC
ERR-ICC
ERR-RN
ERR RO
ERR-LD
ERR BE

Value
equ
equ
equ
equ
equ
equ
equ
equ
equ

Key

$8000
ERR BASE+O
ERR-BASE+l
ERR-BASE+2
ERR-BASE+3

**
**
**
**
**

ERR BASE+4

ERR-BASE+5
ERR-BASE+6

**-

Description
Normal Successful Completion
Error-Encountered bit
Not Yet Implemented
Invalid Function Code
Invalid Command Code
Bus Error: Non-Existant Memo
Bus Error: Retry Overtlow
Bus Error: Link Down

Bus Error: Unexplained

Workstation Graphics Architecture - PRELIMINARY
ERR AE
ERR SI
ERR II
ERR BN
ERR BNI
ERR-KBO
ERR-TBO
ERR-BBO
ERR-ITP

equ
equ
equ
equ
equ
equ
equ
equ
equ

ERR BASE+7
ERR-BASE+8
ERR-BASE+9
ERR-BASE+lO
ERR-BASE+ll
ERR-BASE+l2
ERR-BASE+l3
ERR-BASE+l4
ERR-BASE+l5

Page 89

Address Error
Spurious Interrupt
Illegal Instruction
BBA NXM (Non-Existant Memory
BBA Not Installed
Keyboard Buffer Overflow
Tablet Buff er Overflow
Button Buffer Overflow
Invalid Tablet Packet

Key

Description

**
**

*****-

* WGA Packet Error Codes (32 reserved)

* Error Mnemonic

Value

ERR ISRCMB
ERR-ISRCBW
ERR ISRCBH
ERR ISRCC
ERR-ISRCBD
ERR-ISRCD

equ
equ
equ
equ
equ
equ

ERR BASE+32
ERR-BASE+33
ERR-BASE+34
ERR-BASE+35
ERR-BASE+36
ERR-BASE+37

ERR IMSKMB
ERR-IMSKBW
ERR-IMSKBH
ERR-IMSKBD

equ
equ
equ
equ

ERR BASE+38
ERR-BASE+39
ERR-BASE+40
ERR-BASE+41

**
,• **
**
**

Invalid MSK Modifier Bits
Invalid MSK Bitmap Width
Invalid MSK Bitmap Height
Invalid MSK Bitmap Depth

ERR IDSTMB

equ

ERR BASE+44

,. **

Invalid DST-Offset Modifier

ERR IDSTBW
ERR-IDSTBH
ERR-IDSTBD

equ
equ
equ .

ERR BASE+45
ERR-BASE+46
ERR-BASE+47

**
**

Invalid DST Bitmap Width
Invalid DST Bitmap Height
Invalid DST Bitmap Depth

ERR NOAREA

equ

ERR BASE+48

No Resultant Area

ERR IMAPMB
ERR-IMAPFC
ERR ZIMAP
ERR-ZCIMAP

equ
equ
equ
equ

ERR BASE+50
ERR-BASE+51
ERR-BASE+52
ERR-BASE+53

**
-+
-+
-+

Invalid Map Modifier Bits
Invalid Map Function Code
Depth Incompatible with Map
Depth Combination Incompatib

ERR ICLPMB
ERR-ICLPRC

equ
equ

ERR BASE+54
ERR-BASE+55

**
**

Invalid ClipR Modifier Bits
Invalid ClipR Count

ERR SMC ITC
ERR-ITC-MULT
ERR-ITC-DIV

equ
equ
equ

ERR BASE+56
ERR-BASE+57
ERR-BASE+58

+- !
-+

Invalid Tracking Ratio
Invalid Tracking Multiplier
Invalid Tracking Divisor

ERR ICD
ERR-MO IBC
ERR MO IOT
ERR-MO-IDT

equ
equ
equ
equ

ERR BASE+59
ERR-BASE+60
ERR-BASE+61
ERR-BASE+62

ERR IPC

equ

**
**
**

**
**

***

ERR-BASE+63

* WGA Draw Curve Error Codes (16 reserved)

Invalid SRC Modifier Bits
Invalid SRC Bitmap Width
Invalid SRC Bitmap Height
Invalid SRC Constant
Invalid SRC Bitmap Depth
Invalid SRC Bitmap Dimension

Invalid Cursor Device
Invalid Byte Count
Invalid Object Type

Invalid Path Count

Inva l id

Devic~

Type

Workstation Graphics Architecture - PRELIMINARY

* Error Mnemonic
ERR DC IPC
ERR-DC-IPSL
ERR-DC-IPSM
ERR-DC-ICF
ERR-DC-IPSP
ERR-DC-IPSMB
ERR-DC-IPMMB
ERR-DC-IPSC
ERR-DC-ISSRCBW
ERR-DC-ISSRCBH
ERR-DC-ISSRCBD
ERR-DC-ISSRCC
ERR-DC-IDPM

equ
equ
equ
equ
equ
equ
equ
equ
equ
equ
equ
equ
equ

Page 90

Value

Key

Description

ERR BASE+64
ERR-BASE+65
ERR-BASE+66
ERR-BASE+67
ERR-BASE+68
ERR-BASE+69
ERR-BASE+70
ERR-BASE+71
ERR-BASE+72
ERR-BASE+73
ERR-BASE+74
ERR-BASE+75
ERR-BASE+76

**
. **
**
. **
**
**

**
**
**
**
*++

Invalid Path Count
Invalid Pattern Length
Invalid Pattern Multiplier
Invalid Closed Figure
Invalid Pattern Position
Invalid Pattern String Modif
Invalid Pattern Mode Modif ie
Invalid Pattern Count
Invalid Second SRC Bitmap Wi
Invalid Second SRC Bitmap He
Invalid Second SRC Bitmap De
Invalid Second SRC Constant
Incompatible Drawing/Pattern

Value

Key

Description

ERR BASE+80
ERR-BASE+81
ERR-BASE+82
ERR-BASE+83
ERR-BASE+84
ERR-BASE+85
ERR-BASE+86
ERR-BASE+87
ERR-BASE+88
ERR-BASE+89
ERR-BASE+90
ERR-BASE+91
ERR-BASE+92
ERR-BASE+93

**
**

Invalid Control String Lengt
Invalid Control String Opcod
Invalid Control String Param
Invalid Text String Length
Invalid Character Index
Text String Exhausted
No Font Present
Invalid SRC Font width
Invalid SRC Font height
Invalid SRC Font depth
Invalid MSK Font width
Invalid MSK Font height
Invalid MSK Font depth
Conflicting SRC/MSK Fonts

Value

Key

Description

ERR BASE+96
ERR-BASE+98
ERR-BASE+99
ERR-BASE+lOO

**
**

Invalid SRC Bitmap
Seed Point is on boundary
Stack Overflow
Invalid Boundary Ma~ Modif ie

* WGA Print Text Error Codes (16 reserved)
*
* Error Mnemonic
ERR PT ICSL
ERR-PT-ICSO
ERR-PT-ICSP
ERR PT ITSL
ERR-PT-IC!
ERR-PT-TSE
ERR-PT-NFP
ERR PT ISRCFW
ERR-PT-ISRCFH
ERR-PT-ISRCFD
ERR-PT-IMSKFW
ERR-PT-IMSKFH
ERR-PT-IMSKFD
ERR-PT-CSMF

equ
equ
equ
equ
equ
equ
equ
equ
equ
equ
equ
equ
equ
equ

**
**
**

* WGA Flood Area Error Codes (16 reserved)
*
* Error Mnemonic
ERR FA ISRCB
ERR-FA-SP I OB
ERR-FA-SO
ERR-FA-IBMMB

equ
equ
equ
equ

**
**

** WGA Fill_Polygon Error Codes (16 reserved)
*
* Error Mnemonic
ERR FP ISRCB

ERR-FP-SO
ERR FP IPC

Value
equ

equ
equ

ERR BASET112

ERR-BASE+ll3
ERR-BASE+ll4

Key
**
**

Description
Invalid SRC Bitmap

Stack Overflow
Invalid Point Count

Workstation Graphics Architecture - PRELIMINARY
ERR FP !CF

equ

ERR BASE+ll5

Page 91

Invalid Closed Figure

* WGA Powerup Error Codes (32 reserved)
*

* Error Mnemonic
ERR PASS
ERR-68K
ERR-RC
ERR PR
ERR-CRT
ERR-TU
ERR KU
ERR-FOE
ERR-VTO
ERR-SB
ERR-BS
ERR-BC
ERR-TTO
ERR-FOO
ERR-KTO
ERR-KST

equ
equ
equ
equ
equ
equ
equ
equ
equ
equ
equ
equ
equ
equ
equ
equ

Value

Key

Description

ERR BASE+l28
ERR-BASE+l29
ERR-BASE+l30
ERR-BASE+l31
ERR-BASE+l32
ERR-BASE+l33
ERR-BASE+l34
ERR-BASE+l35
ERR-BASE+l36
ERR-BASE+l37
ERR-BASE+l38
ERR-BASE+l39
ERR-BASE+l40
ERR-BASE+l41
ERR-BASE+l42
ERR-BASE+l43

**-

Base for test numbers
68000 CPU
ROM Checksum
Program RAM
CRTC Register
Tablet USART
Keyboard USART
FOTR Electrical Loop Back
Vsync Time Out
Screen Buff er
BBA Scratchpad RAM
BBA Copyarea Command
Tablet Time Out
FOTR Optical Loop Back
Keyboard Time Out
Keyboard Self-Test

**-

****-

***-

** WGA Load Cursor Error Codes (16 reserved)
*
* Error Mnemonic
ERR LDC IATRV
ERR-LDC-ICH
ERR-LDC-I CW
ERR-NOVALCUR

equ
equ
equ
equ

Value

Key

Description

ERR BASE+l60
ERR-BASE+l61
ERR-BASE+l62
ERR-BASE+l63

++
++
++
++

Invalid Cursor Attribute Val
Invalid Cursor Height
Invalid Cursor Width
No Valid Cursor Defined

)

Workstation Graphics Architecture - PRELIMINARY
8.0

Page 92

VAXSTATION 100 RESTRICTIONS

The following architectural restrictions apply to the VAXstation
implementation of the Workstation graphics architecture.

8.1

Number Of Planes

Since the VSlOO has only one bit plane of framebuffer
parameter may only have the value 1.

8.2

100

memory,

the

Halftone Representation

The VAXstation 100 uses only a single format for a halftone bitmap.
The halftone pattern must be specified as a square bitmap 16 pixels on
a side. Therefore, to use a "standard" four-by-four halftone pattern,
the pattern is simply replicated horizontally and vertically to form a
16xl6 pattern.

9.0

GENERAL IMPLEMENTATION RESTRICTIONS

The following restrictions apply to all VAX/UNIBUS implementations of
the Workstation graphics architecture. Hence, the term "display" is
used instead of "VSlOO".

9.1

Word Access I/O

All 16-bit word parameters must be word-aligned.
Additionally,
the
display may only access host memory by words.
Any byte strings of odd
length, then,
should be padded with an extra byte so that no
"undefined" data is accessed.

9.2

UNIBUS Window Mapping

It is often the case that the display is asked to perform an operation
between a source rectangle and a destination rectangle which overlap.
That is, both source and destination bitmaps occupy the same (or
nearly the same) area of memory.
In this case, the display firmware
must determine the proper memory-copy direction, so that no data
is
overwritten.
It does this by comparing the base addresses of the two
memory blocks in question.
For this reason, it is important that the

driver software (responsible for mapping UNIBUS memory to the display)
maintain a one-to-one correspondence between areas of memory

side

the

UBW.

Sp~cifically,

on each
an ~rea of VAX memory must neve ~) b· e

made to appear to the display as two different areas of memory.

Page 93

Workstation Graphics Architecture - PRELIMINARY
9.3

Bitmap Storage Requirements

The display represents a bitmap as a sequence of horizontal scan lines
stored in contiguous memory locations. Each scan line must begin on a
16-bit word boundary.
That is, although a bitmap can have any
horizontal width in pixels, the storage in which the bitmap is kept
must have sufficient space so that each horizontal line can be
word-aligned.
If the horizontal width in pixels is not evenly
divisible by 16, the last bits in the last word of the storage for
each horizontal line will be unused. For any bitmap of dimensions
(X,Y) on the display, the storage requirement is ((X+l5)/16)Y words.

9.4

Device Coordinate Management For WGA

The Device Position Registers, defined by the WGA, are used to hold
the current XY position of the cursor. They are also used to report
the current XY position of any pointing devices, e.g.
mouse and/or
tablet, which may be attached to the VAXstation.
The WGA also defines a phenomenon known as Event Reporting.
If
reporting is enabled for a particular device, it will interrupt the
host at a maximum rate of 60 Hz. It places its XY position in the
Cursor Position Registers and issues a "'device' moved" interrupt
reason.
The following paragraphs enumerate the
situations, and how they are implemented.
[l] State:
[la]

possible

Event~Reporting

No Device Attached to the Cursor
Mouse Event Reporting set

On the Vsync interrupt, a service routine within the VSlOO
will read the current mouse XY position.
If it has changed from the last time, it will
be placed in the Device Position Registers.

An interrupt is then

sent to the host indicating "MOUSE MOVED".
If the mouse has not moved, nothing happens.
Note that since the mouse is not connected to anything here, its
position is not confined to the visible screen.
may vary from 32767 to -32768.
[lb]

Tablet Event Reporting set

Its coordinates

Page 94

Workstation Graphics Architecture - PRELIMINARY
On the Vsync interrupt, a service routine will read the
current tablet XY position.

If it has changed from the

last time, it will be placed in the Device Position Registers.
It then issues an interrupt to the host indicating "TABLET MOVED".
If the tablet has not moved, nothing happens.
Note that the tablet position is not clipped to the visible
screen boundaries unless it's attached to the cursor.
[le]

Both Mouse and Tablet Event Reporting set

In this case, both case [la] · and [lb] happen concurrently.
However, only one device will be reported on any one Vsync.
The devices will be polled ~n a round-robin-like fashion.
On one Vsync, device (i) will be checked FIRST.

If no action,

device (i+l) is checked, and so on, until an 'active' one is found
(or until there are no more devices}.
( i +l)

On the next

Vsync,

device

will be the first one checked, and so on.
This method has the following properties:
o No more than 60 interrupts per second are sent to
no matter how many devices are reporting.

[2] State:
[2a]

No device will be 'locked out' by a higher-priorit
or a more-rapidly-interrupting device.

Each individual device may still interrupt at 60Hz
it has no 'competition'.

The worst case for device acknowledgement is 60/N
N reporting devices.

Mouse is Attached to the Cursor
Mouse Event Reporting set

All happens as in [la], except that now the interrupt reason

will be "CURSOR MOVED" instead of "MOUSE MOVED".

Of course,

since the mouse and cursor are attached, their XY positions

Workstation Graphics Architecture - PRELIMINARY

Page 95

always be the same.
This implies, of course, that the mouse XY position will now
be clipped to the visible screen.
[2b]

Tablet Event Reporting set

All happens as in case [le], except that cases [2a] and [lb] are
happening concurrently, instead of cases [la] and [lb].
[2c]

Both Mouse and Tablet Event Reporting set

This is the same as case [2b] above.
[3] State:
[3a]

Tablet is Attached to the Cursor
Mouse Event Reporting set

All happens as in case [le], except that cases [la] and [3b] are
happening concurrently, instead of cases [la] and [lb].
[3b]

Tablet Event Reporting set

All happens as in [lb], except that now the interrupt reason
will be "CURSOR MOVED" instead of "TABLET MOVED".
Since the tablet and cursor are attached, their XY positions will
always be the same as long as they are within the visible screen.
Should the tablet position stray from the visible screen,
both the tablet and the cursor XY position will be clipped to
the screen's boundaries.
[ 3c]

Both Mouse and Tablet Event Reporting set

This is the same as case [3a] above.

Workstation Graphics Architecture - PRELIMINARY
9.5

Page 96

Keyboard Interface

If the input data is a keycode, a transition bit is generated,
the
keyboard device code added, and the result is returned, where it will
become a button event to the VAXen host.
If the key is in
Autorepeating/Down-Only
mode,
no
up-transition events will be
generated. An autorepeating key will thus appear as a succession of
down-events.
If the key is in Up/Down mode, both up and down
transitions will be generated.
If the input data is an error message, or an
ignored by the VSlOO.

acknowledgement,

There are a few codes which are treated specially:
o

The Metronome character, for example, is not passed
to the host, but is used to repeat the most recently
depressed character.

The state of the control-key code is monitored. When
the 'control' key is depressed, the firmware will automaticall
issue a 'temporary autorepeat inhibit' to the keyboard to
keep control keys from autorepeating.

The 'Allups' keycode will generate up-transitions for all
up/down-mode keys which are down at the time.

Other special codes not related to actual keys will be discard

Since the user is not allowed to change the mode of the
keyboard divisions, Prefix-To-Keys-Down should never occur.
The keyboard divisions are initialized
According to the following defaults:
Division

Mode

1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.

Autorepeat down
Autorepeat down
Autorepeat down
Down only
Up down
Up down
Autorepeat down
Autorepeat down
Autorepeat down
Autorepeat down
Autorepeat down
Autorepeat down
Down only
Autorepeat down

Main array
Numeric keypad
Delete character
Return and Tab characters
Lock, Compose, and AlO
Shift keys and Control key
Horizontal arrow keys
Vertical arrow keys
Edit keypad keys
Local function keys (G99-G04)
Second function key set (G05-G09)
Third function key set (Gl0-Gl4)
HELP and MENU keys (Gl5-Gl6)
Fifth function key set (G20-G23)

In Up/Down mode, both up- and down-transitions are reported.

In Autorepeat and Down-Only mode, only down-transitions are

repor- ~ e