Smalltalk80LanguageImplementation:Chapter 26

From 흡혈양파의 번역工房
Revision as of 07:42, 14 July 2015 by Onionmixer (talk | contribs) (collect error)
Jump to navigation Jump to search
Chapter 26 The Implementation

The Implementation

Two major components of the Smalltalk-80 system can be distinguished: the virtual image and the virtual machine.

  1. The virtual image consists of all of the objects in the system.
  2. The virtual machine consists of the hardware devices and machine language (or microcode) routines that give dynamics to the objects in the virtual image.


The system implementer's task is to create a virtual machine. A virtual image can then be loaded into the virtual machine and the Smalltalk-80 system becomes the interactive entity described in earlier chapters.


The overview of the Smalltalk-80 implementation given in this chapter is organized in a top-down fashion, starting with the source methods written by programmers. These methods are translated by a compiler into sequences of eight-bit instructions called bytecodes. The compiler and bytecodes are the subject of this chapter's first section. The bytecodes produced by the compiler are instructions for an interpreter, which is described in the second section. Below the interpreter in the implementation is an object memory that stores the objects that make up the virtual image. The object memory is described in the third section of this chapter. At the bottom of any implementation is the hardware. The fourth and final section of this chapter discusses the hardware required to implement the interpreter and object memory. Chapters 27-30 give a detailed specification of the virtual machine's interpreter and object memory.


The Compiler

Source methods written by programmers are represented in the Smalltalk-80 system as instances of String. The Strings contain sequences of characters that conform to the syntax introduced in the first part of this book. For example, the following source method might describe how instances of class Rectangle respond to the unary message center. The center message is used to find the Point equidistant from a Rectangle's four sides.

center
    orgin + corner / 2


Source methods are translated by the system's compiler into sequences of instructions for a stack-oriented interpreter. The instructions are eight-bit numbers called bytecodes. For example, the bytecodes corresponding to the source method shown above are,

0, 1, 176, 119, 185, 124


Since a bytecode's value gives us little indication of its meaning to the interpreter, this chapter will accompany lists of bytecodes with comments about their functions. Any part of a bytecode's comment that depends on the context of the method in which it appears will be parenthesized. The unparenthesized part of the comment describes its general function. For example, the bytecode 0 always instructs the interpreter to push the value of the receiver's first instance variable on its stack. The fact that the variable is named origin depends on the fact that this method is used by Rectangles, so origin is parenthesized. The commented form of the bytecodes for Rectangle center is shown below.

Rectangle center
0 push the value of the receiver's first instance variable (origin) onto the stack
1 push the value of the receiver's second instance variable (corner) onto the stack
176 send a binary message with the selector +
119 push the SmallInteger 2 onto the stack
185 send a binary message with the selector /
124 return the object on top of the stack as the value of the message (center)


The stack mentioned in some of the bytecodes is used for several purposes. In this method, it is used to hold the receiver, arguments, and results of the two messages that are sent. The stack is also used as the source of the result to be returned from the center method. The stack is maintained by the interpreter and will be described in greater detail in the next section. A description of all the types of bytecodes will appear at the end of this section.


A programmer does not interact directly with the compiler. When a new source method is added to a class (Rectangle in this example), the class asks the compiler for an instance of CompiledMethod containing the bytecode translation of the source method. The class provides the compiler with some necessary information not given in the source method, including the names of the receiver's instance variables and the dictionaries containing accessible shared variables (global, class, and pool variables). The compiler translates the source text into a CompiledMethod and the class stores the method in its message dictionary. For example, the CompiledMethod shown above is stored in Rectangle's message dictionary associated with the selector center.


Another example of the bytecodes compiled from a source method illustrates the use of a store bytecode. The message extent: to a Rectangle changes the receiver's width and height to be equal to the x and y coordinates of the argument (a Point). The receiver's upper left corner (origin) is kept the same and the lower right corner (corner) is moved.

extent: newExtent
    corner  origin + newExtent
Rectangle extent:
0 push the value of the receiver's first instance variable (origin) onto the stack
16 push the argument (newExtent) onto the stack
176 send a binary message with the selector +
97 pop the top object off of the stack and store it in the receiver's second instance variable (corner)
120 return the receiver as the value of the message (extent:)


The forms of source methods and compiled bytecodes are different in several respects. The variable names in a source method are converted into instructions to push objects on the stack, the selectors are converted into instructions to send messages, and the uparrow is converted into an instruction to return a result. The order of the corresponding components is also different in a source method and compiled bytecodes. Despite these differences in form, the source method and compiled bytecodes describe the same actions.


Compiled Methods

The compiler creates an instance of CompiledMethod to hold the bytecode translation of a source method. In addition to the bytecodes themselves, a CompiledMethod contains a set of objects called its literal frame. The literal frame contains any objects that could not be referred to directly by bytecodes. All of the objects in Rectangle center and Rectangle extent: were referred to directly by bytecodes, so the CompiledMethods for these methods do not need literal frames. As an example of a CompiledMethod with a literal frame, consider the method for Rectangle intersects:. The intersects: message inquires whether one Rectangle (the receiver) overlaps another Rectangle (the argument).

intersects: aRectangle
    (origin max: aRectangle origin) < (corner min: aRectangle corner)


The four message selectors, max:, origin, rain:, and corner are not in the set that can be directly referenced by bytecodes. These selectors are included in the CompiledMethod's literal frame and the send bytecodes refer to the selectors by their position in the literal frame. A CompiledMethod's literal frame will be shown after its bytecodes.

Rectangle intersects:
0 push the value of the receiver's first instance variable (origin) onto the stack
16 push the argument (aRectangle)
209 send a unary message with the selector in the second literal frame location (origin)
224 send a single argument message with the selector in the first literal frame location (max:)
1 push the value of the receiver's second instance variable (corner) onto the stack
16 push the argument (aRectangle) onto the stack
211 send a unary message with the selector in the fourth literal frame location (corner)
226 send a single argument message with the selector in the third literal frame location (min:)
178 send a binary message with the selector <
124 return the object on top of the stack as the value of the message (intersects:)
literal frame
    #max:
    #origin
    #min:
    #corner


The categories of objects that can be referred to directly by bytecodes are:

  • the receiver and arguments of the invoking message
  • the values of the receiver's instance variables
  • the values of any temporary variables required by the method
  • seven special constants (true, false, nil, -1, 0, 1, and 2)
  • 32 special message selectors


The 32 special message selectors are listen below.

+ - < >
< = > = = ~=
* / \ @
bitShift: \\ bitAnd: bitOr:
(at:) (at:put:) (size) (next)
(nextPut:) (atEnd) = = class
blockCopy: value value: (do:)
(new) (new:) (x) (y)


The selectors in parentheses may be replaced with other selectors by modifying the compiler and recompiling all methods in the system. The other selectors are built into the virtual machine.


Any objects referred to in a CompiledMethod's bytecodes that do not fall into one of the categories above must appear in its literal frame.

  • shared variables (global, class, and pool)
  • most literal constants (numbers, characters, strings, arrays, and symbols)
  • most message selectors (those that are not special)


Objects of these three types may be intermixed in the literal frame. If an object in the literal frame is referenced twice in the same method, it need only appear in the literal frame once. The two bytecodes that refer to the object will refer to the same location in the literal frame.


Two types of object that were referred to above, temporary variables and shared variables, have not been used in the example methods. The following example method for Rectangle merge: uses both types. The merge: message is used to find a Rectangle that includes the areas in both the receiver and the argument.

merge: aRectangle
    | minPoint maxPoint |
    minPoint  origin min: aRectangle origin.
    maxPoint  corner max: aRectangle corner.
    Rectangle origin: minPoint
        corner: maxPoint


When a CompiledMethod uses temporary variables (maxPoint and minPoint in this example), the number required is specified in the first line of its printed form. When a CompiledMethod uses a shared variable (Rectangle in this example) an instance of Association is included in its literal frame. All CompiledMethods that refer to a particular shared variable's name include the same Association in their literal frames.

Rectangle merge: requires 2 temporary variables
0 push the value of the receiver's first instance variable (origin) onto the stack
16 push the contents of the first temporary frame location (the argument aRectangle) onto the stack
209 send a unary message with the selector in the second literal frame location (origin)
224 send the single argument message with the selector in the first literal frame location (min:)
105 pop the top object off of the stack and store in the second temporary frame location (minPoint)
1 push the value of the receiver's second instance variable (corner) onto the stack
16 push the contents of the first temporary frame location (the argument aRectangle) onto the stack
211 send a unary message with the selector in the fourth literal frame location (corner)
226 send a single argument message with the selector in the third literal frame location (max:)
106 pop the top object off of the stack and store it in the third temporary frame location (maxPoint)
69 push the value of the shared variable in the sixth literal frame location(Rectangle) onto the stack
17 push the contents of the second temporary frame location (minPoint) onto the stack
18 push the contents of the third temporary frame location (maxPoint) onto the stack
244 send the two argument message with the selector in the fifth literal frame location (origin:corner:)
124 return the object on top of the stack as the value of the message (merge:)
literal frame
    #min:
    #origin
    #max:
    #corner
    #origin:corner:
    Association: #Rectangle ➛ Rectangle


❏ Temporary Variables Temporary variables are created for a particular execution of a CompiledMethod and cease to exist when the execution is complete. The CompiledMethod indicates to the interpreter how many temporary variables will be required. The arguments of the invoking message and the values of the temporary variables are stored together in the temporary frame. The arguments are stored first and the temporary variable values immediately after. They are accessed by the same type of bytecode (whose comments refer to a temporary frame location). Since merge: takes a single argument, its two temporary variables use the second and third locations in the temporary frame. The compiler enforces the fact that the values of the argument names cannot be changed by never issuing a store bytecode referring to the part of the temporary frame inhabited by the arguments.


❏ Shared Variables Shared variables are found in dictionaries.

  • global variables in a dictionary whose names can appear in any method
  • class variables in a dictionary whose names can only appear in the methods of a single class and its subclasses
  • pool variables in a dictionary whose names can appear in the methods of several classes


Shared variables are the individual associations that make up these dictionaries. The system represents associations in general, and shared variables in particular, with instances of Association. When the compiler encounters the name of a shared variable in a source method, the Association with the same name is included in the CompiledMethod's literal frame. The bytecodes that access shared variables indicate the location of an Association in the literal frame. The actual value of the variable is stored in an instance variable of the Association. In the CompiledMethod for Rectangle merge: shown above, class Rectangle is referenced by including the Association from the global dictionary whose name is the symbol Rectangle and whose value is the class Rectangle.


The Bytecodes

The interpreter understands 256 bytecode instructions that fall into five categories: pushes, stores, sends, returns, and jumps. This section gives a general description of each type of bytecode without going into detail about which bytecode represents which instruction. Chapter 28 describes the .exact meaning of each bytecode. Since more than 256 instructions for the interpreter are needed, some of the bytecodes take extensions. An extension is one or two bytes following the bytecode, that further specify the instruction. An extension is not an instruction on its own, it is only a part of an instruction.


❏ Push Bytecodes A push bytecode indicates the source of an object to be added to the top of the interpreter's stack. The sources include

  • the receiver of the message that invoked the CompiledMethod
  • the instance variables of the receiver
  • the temporary frame (the arguments of the message and the temporary variables)
  • the literal frame of the CompiledMethod
  • the top of the stack (i.e., this bytecode duplicates the top of the stack)


Examples of most of the types of push bytecode have been included in the examples. The bytecode that duplicates the top of the stack is used to implement cascaded messages.


Two different types of push bytecode use the literal frame as their source. One is used to push literal constants and the other to push the value of shared variables. Literal constants are stored directly in the literal frame, but the values of shared variables are stored in an Association that is pointed to by the literal frame. The following example method uses one shared variable and one literal constant.

incrementIndex
    Index + Index + 4
ExampleClass incrementIndex
64 push the value of the shared variable in the first literal frame location(Index) onto the stack
33 push the constant in the second literal frame location (4) onto the stack
176 send a binary message with the selector +
129,192 store the object on top of the stack in the shared variable in the first literal frame location (Index)
124 return the object on top of the stack as the value of the message (incrementIndex)
literal frame
    Association: #Index ➛ 260
    4


❏ Store Bytecodes The bytecodes compiled from an assignment expression end with a store bytecode. The bytecodes before the store bytecode compute the new value of a variable and leave it on top of the stack. A store bytecode indicates the variable whose value should be changed. The variables that can be changed are

  • the instance variables of the receiver
  • temporary variables
  • shared variables


Some of the store bytecodes remove the object to be stored from the stack, and others leave the object on top of the stack, after storing it.


❏ Send Bytecodes A send bytecode specifies the selector of a message to be sent and how many arguments it should have. The receiver and arguments of the message are taken off the interpreter's stack, the receiver from below the arguments. By the time the bytecode following the send is executed, the message's result will have replaced its receiver and arguments on the top of the stack. The details of sending messages and returning results is the subject of the next sections of this chapter. A set of 32 send bytecodes refer directly to the special selectors listed earlier. The other send bytecodes refer to their selectors in the literal frame.


❏ Return Bytecodes When a return bytecode is encountered, the CompiledMethod in which it was found has been completely executed. Therefore a value is returned for the message that invoked that CompiledMethod. The value is usually found on top of the stack. Four special return bytecodes return the message receiver (self), true, false, and nil.


❏ Jump Bytecodes Ordinarily, the interpreter executes the bytecodes sequentially in the order they appear in a CompiledMethod. The jump bytecodes indicate that the next bytecode to execute is not the one following the jump. There are two varieties of jump, unconditional and conditional. The unconditional jumps transfer control whenever they are encountered. The conditional jumps will only transfer control if the top of the stack is a specified value. Some of the conditional jumps transfer if the top object on the stack is true and others if it is false. The jump bytecodes are used to implement efficient control structures.


The control structures that are so optimized by the compiler are the conditional selection messages to Booleans (ifTrue:, ifFalse:, and ifTrue:ifFalse:), some of the logical operation messages to Booleans (and" and or:), and the conditional repetition messages to blocks (whileTrue: and whileFalse:). The jump bytecodes indicate the next bytecode to be executed relative to the position of the jump. In other words, they tell the interpreter how many bytecodes to skip. The following method for Rectangle includesPoint: uses a conditional jump.

includesPoint: aPoint
    origin < = aPoint
        ifTrue: [aPoint < corner]
        ifFalse: [false]


Rectangle includesPoint:
0 push the value of the receiver's first instance variable (origin) onto the stack
16 push the contents of the first temporary frame location (the argument aPoint) onto the stack
180 send a binary message with the selector < =
155 jump ahead 4 bytecodes if the object on top of the stack is false
16 push the contents of the first temporary frame location (the argument aPoint) onto the stack
1 push the value of the receiver's second instance variable (corner) onto the stack
178 send a binary message with the selector <
124 return the object on top of the stack as the value of the message (includesPoint:)


The Interpreter

The Smalltalk-80 interpreter executes the bytecode instructions found in CompiledMethods. The interpreter uses five pieces of information and repeatedly performs a three-step cycle.

  1. The CompiledMethod whose bytecodes are being executed.
  2. The location of the next bytecode to be executed in that CompiledMethod. This is the interpreter's instruction pointer.
  3. The receiver and arguments of the message that invoked the CompiledMethod.
  4. Any temporary variables needed by the CompiledMethod.
  5. A stack.


The execution of most bytecodes involves the interpreter's stack. Push bytecodes tell where to find objects to add to the stack. Store bytecodes tell where to put objects found on the stack. Send bytecodes remove the receiver and arguments of messages from the stack. When the result of a message is computed, it is pushed onto the stack.


The Cycle of the Interpreter

  1. Fetch the bytecode from the CompiledMethod indicated by the instruction pointer.
  2. Increment the instruction pointer.
  3. Perform the function specified by the bytecode.


As an example of the interpreter's function, we will trace its execution of the CompiledMethod for Rectangle center. The state of the interpreter will be displayed after each of its cycles. The instruction pointer will be indicated by an arrow pointing at the next bytecode in the CompiledMethod to be executed.

➧ 0 push the value of the receiver's first instance variable (origin) onto the stack


The receiver, arguments, temporary variables, and objects on the stack will be shown as normally printed (their responses to printString). For example, if a message is sent to a Rectangle, the receiver will be shown as

Receiver||100@100 corner: 200@200


At the start of execution, the stack is empty and the instruction pointer indicates the first bytecode in the CompiledMethod. This CompiledMethod does not require temporaries and the invoking message did not have arguments, so these two categories are also empty.

Method for Rectangle center
➧ 0 push the value of the receiver's first instance variable (origin) onto the stack
1 push the value of the receiver's second instance variable (corner) onto the stack
176 send a binary message with the selector +
119 push the SmallInteger 2 onto the stack
185 send a binary message with the selector /
124 return the object on top of the stack as the value of the message (center)
Receiver 100@100 corner: 200@200
Arguments
Temporary Variables
Stack


Following one cycle of the interpreter, the instruction pointer has been advanced and the value of the receiver's first instance variable has been copied onto the stack.

Method for Rectangle center
0 push the value of the receiver's first instance variable (origin) onto the stack
➧ 1 push the value of the receiver's second instance variable (corner) onto the stack
176 send a binary message with the selector +
119 push the SmallInteger 2 onto the stack
185 send a binary message with the selector /
124 return the object on top of the stack as the value of the message (center)
Receiver 100@100 corner: 200@200
Arguments
Temporary Variables
Stack 100@100


The interpreter's second cycle has an effect similar to the first. The top of the stack is shown toward the bottom of the page. This corresponds to the commonly used convention that memory locations are shown with addresses increasing toward the bottom of the page.

Method for Rectangle center
0 push the value of the receiver's first instance variable (origin) onto the stack
1 push the value of the receiver's second instance variable (corner) onto the stack
➧ 176 send a binary message with the selector +
119 push the SmallInteger 2 onto the stack
185 send a binary message with the selector /
124 return the object on top of the stack as the value of the message (center)
Receiver 100@100 corner: 200@200
Arguments
Temporary Variables
Stack 100@100
200@200


The interpreter's third cycle encounters a send bytecode. It removes two objects from the stack and uses them as the receiver and argument of a message with selector +. The procedure for sending the message will not be described in detail here. For the moment, it is only necessary to know that eventually the result of the + message will be pushed onto the stack. Sending messages will be described in later sections.

Method for Rectangle center
Method for Rectangle center
0 push the value of the receiver's first instance variable (origin) onto the stack
1 push the value of the receiver's second instance variable (corner) onto the stack
176 send a binary message with the selector +
➧ 119 push the SmallInteger 2 onto the stack
185 send a binary message with the selector /
124 return the object on top of the stack as the value of the message (center)
Receiver 100@100 corner: 200@200
Arguments
Temporary Variables
Stack 300@300


The interpreter's next cycle pushes the constant 2 onto the stack.

Method for Rectangle center
0 push the value of the receiver's first instance variable (origin) onto the stack
1 push the value of the receiver's second instance variable (corner) onto the stack
176 send a binary message with the selector +
119 push the SmallInteger 2 onto the stack
➧ 185 send a binary message with the selector /
124 return the object on top of the stack as the value of the message (center)
Receiver 100@100 corner: 200@200
Arguments
Temporary Variables
Stack 2


The interpreter's next cycle sends another message whose result replaces its receiver and arguments on the stack.

Method for Rectangle center
0 push the value of the receiver's first instance variable (origin) onto the stack
1 push the value of the receiver's second instance variable (corner) onto the stack
176 send a binary message with the selector +
119 push the SmallInteger 2 onto the stack
185 send a binary message with the selector /
➧ 124 return the object on top of the stack as the value of the message (center)
Receiver 100@100 corner: 200@200
Arguments
Temporary Variables
Stack 150@150


The final bytecode returns a result to the center message. The result is found on the stack (150@150). It is clear by this point that a return bytecode must involve pushing the result onto another stack. The details of returning a value to a message will be described after the description of sending a message.


Contexts

Push, store, and jump bytecodes require only small changes to the state of the interpreter. Objects may be moved to or from the stack, and the instruction pointer is always changed; but most of the state remains the same. Send and return bytecodes may require much larger changes to the interpreter's state. When a message is sent, all five parts of the interpreter's state may have to be changed in order to execute a different CompiledMethod in response to this new message. The interpreter's old state must be remembered because the bytecodes after the send must be executed after the value of the message is returned.


The interpreter saves its state in objects called contexts. There will be many contexts in the system at any one time. The context that represents the current state of the interpreter is called the active context. When a send bytecode in the active context's CompiledMethod requires a new CompiledMethod to be executed, the active context becomes suspended and a new context is created and made active. The suspended context retains the state associated with the original CompiledMethod until that context becomes active again. A context must remember the context that it suspended so that the suspended context can be resumed when a result is returned. The suspended context is called the new context's sender.


The form used to show the interpreter's state in the last section will be used to show contexts as well. The active context will be indicated by the word Active in its top delimiter. Suspended contexts will say Suspended. For example, consider a context representing the execution of the CompiledMethod for Rectangle rightCenter with a receiver of 100@ 100 corner: 200@200. The source method for Rectangle rightCenter is

rightCenter
     self right @ self center y


The interpreter's state following execution of the first bytecode is shown below. The sender is some other context in the system.

Active
Method for Rectangle rightCenter
112 push the receiver (self) onto the stack
➧ 208 send a unary message with the selector in the first literal (right)
112 push the receiver (self) onto the stack
209 send the unary message with the selector in the second literal (center)
207 send the unary message with the selector y
187 send the unary message with the selector @
124 return the object on top of the stack as the value of the message (rightCenter)
literal frame
    #right
    #center
Receiver 100@100 corner: 200@200
Arguments
Temporary Variables
Stack 100@100 corner: 200@200
Sender ⇩


After the next bytecode is executed, that context will be suspended. The object pushed by the first bytecode has been removed to be used as the receiver of a new context, which becomes active. The new active context is shown above the suspended context.

Active
Method for Rectangle right
➧ 1 push the value of the receiver's second instance variable (corner) onto the stack
206 send a unary message with the selector x
124 return the object on top of the stack as the value of the message (right)
Receiver 100@100 corner: 200@200
Arguments
Temporary Varialbes
Stack
Sender ⇩


Suspended
Method for Rectangle rightCenter
112 push the receiver (self) onto the stack
208 send a unary message with the selector in the first literal (right)
➧ 112 push the receiver (self) onto the stack
209 send the unary message with the selector in the second literal (center)
207 send the unary message with the selector y
187 send the unary message with the selector @
124 return the object on top of the stack as the value of the message (rightCenter)
literal frame
    #right
    #center
Receiver 100@100 corner: 200@200
Arguments
Temporary Variables
Stack
Sender ⇩


The next cycle of the interpreter advances the new context instead of the previous one.

Active
Method for Rectangle right
1 push the value of the receiver's second instance variable (corner) onto the stack
➧ 206 send a unary message with the selector x
124 return the object on top of the stack as the value of the message (right)
Receiver 100@100 corner: 200@200
Arguments
Temporary Variables
Stack 200@200
Sender ⇩


Suspended
Method for Rectangle rightCenter
112 push the receiver (self) onto the stack
208 send a unary message with the selector in the first literal (right)
➧ 112 push the receiver (self) onto the stack
209 send the unary message with the selector in the second literal (center)
207 send the unary message with the selector y
187 send the unary message with the selector @
124 return the object on top of the stack as the value of the message (rightCenter)
literal frame
    #right
    #center
Receiver 100@100 corner: 200@200
Arguments
Temporary Variables
Stack
Sender ⇩


In the next cycle, another message is sent, perhaps creating another context. Instead of following the response of this new message (x), we will skip to the point that this context returns a value (to right). When the result of x has been returned, the new context looks like this:

Active
Method for Rectangle right
1 push the value of the receiver's second instance variable (corner) onto the stack
206 send a unary message with the selector x
➧ 124 return the object on top of the stack as the value of the message (right)
Receiver 100@100 corner: 200@200
Arguments
Temporary Variables
Stack 200
Sender ⇩


Suspended
Method for Rectangle rightCenter
112 push the receiver (self) onto the stack
208 send a unary message with the selector in the first literal (right)
➧ 112 push the receiver (self) onto the stack
209 send the unary message with the selector in the second literal (center)
207 send the unary message with the selector y
187 send the unary message with the selector @
124 return the object on top of the stack as the value of the message (rightCenter)
literal frame
    #right
    #center
Receiver 100@100 corner: 200@200
Arguments
Temporary Variables
Stack
Sender ⇩


The next bytecode returns the value on the top of the active context's stack (200) as the value of the message that created the context (right). The active context's sender becomes the active context again and the returned value is pushed on its stack.

Active
Method for Rectangle rightCenter
112 push the receiver (self) onto the stack
208 send a unary message with the selector in the first literal (right)
➧ 112 push the receiver (self) onto the stack
209 send the unary message with the selector in the second literal (center)
207 send the unary message with the selector y
187 send the unary message with the selector ®
124 return the object on top of the stack as the value of the message(rightCenter)
literal frame
    #right
    #center
Receiver 100@100 corner: 200@200
Arguments
Temporary Variables
Stack 200
Sender ⇩


Block Contexts

The contexts illustrated in the last section are represented in the system by instances of MethodContext. A MethodContext represents the execution of a CompiledMethod in response to a message. There is another type of context in the system, which is represented by instances of BlockContext. A BlockContext represents a block in a source method that is not part of an optimized control structure. The compilation of the optimized control structures was described in the earlier section on jump bytecodes. The bytecodes compiled from a nonoptimized control structure are illustrated by the following hypothetical method in Collection. This method returns a collection of the classes of the receiver's elements.

classes
    self collect: [ :element | element class]


Active
Collection classes requires 1 temporary variable
112 push the receiver (self) onto the stack
137 push the active context (thisContext) onto the stack
118 push the SmallInteger 1 onto the stack
200 send a single argument message with the selector blockCopy:
164,4. jump around the next 4 bytes
104 pop the top object off of the stack and store in the first temporary frame location (element)
16 push the contents of the first temporary frame location (element) onto the stack
199 send a unary message with the selector class
125 return the object on top of the stack as the value of the block
224 send a single argument message with the selector in the first literal frame location (collect:)
124 return the object on top of the stack as the value of the message (classes)
literal frame
    #collect:


A new BlockContext is created by the blockCopy: message to the active context. The bytecode that pushes the active context was not described along with the rest of the push bytecodes since the function of contexts had not been described at that point. The argument to blockCopy: (1 in this example) indicates the number of block arguments the block requires. The BlockContext shares much of the state of the active context that creates it. The receiver, arguments, temporary variables, CompiledMethod, and sender are all the same. The BlockContext has its own instruction pointer and stack. Upon returning from the blockCopy: message, the newly created BlockContext is on the stack of the active context and the next instruction jumps around the bytecodes that describe the actions of the block. The active context gave the BlockContext an initial instruction pointer pointing to the bytecode after this jump. The compiler always uses an extended (two-byte) jump after a blockCopy: so that the BlockContext's initial instruction pointer is always two more than the active context's instruction pointer when it receives the blockCopy: message.


The method for Collection classes creates a BlockContext, but does not execute its bytecodes. When the collection receives the collect: message, it will repeatedly send value: messages to the BlockContext with the elements of the collection as arguments. A BlockContext responds to value: by becoming the active context, which causes its bytecodes to be executed by the interpreter. Before the BlockContext becomes active, the argument to value: is pushed onto the BlockContext's stack. The first bytecode executed by the BlockContext stores this value in a temporary variable used for the block argument.


A BlockContext can return a value in two ways. After the bytecodes in the block have been executed, the final value on the stack is returned as the value of the message value or value:. The block can also return a value to the message that invoked the CompiledMethod that created the BlockContext. This is done with the regular return bytecodes. The hypothetical method for Collection containsInstanceOf: uses both types of return from a BlockContext.

containsInstanceOf: aClass
    self do: [ :element | (element isKindOf: aClass) ifTrue: [true]].
    false


Collection containsInstanceOf: requires 1 temporary variable
112 push the receiver (self) onto the stack
137 push the active context (thisContext) onto the stack
118 push the SmallInteger 1 onto the stack
200 send a single argument message with the selector blockCopy:
164,8 jump around the next 8 bytes
105 pop the top object off of the stack and store in the second temporary frame location (element)
17 push the contents of the second temporary frame location (element) onto the stack
16 push the contents of the first temporary frame location (aClass) onto the stack
224 send a single argument message with the selector in the first literal frame location (isKindOf:)
152 pop the top object off of the stack and jump around 1 byte if it is false
121 return true as the value of the message (containsInstanceOf:)
115 push nil onto the stack
125 return the object on top of the stack as the value of the block
203 send the single argument message with the selector do:
135 pop the top object off the stack
122 return false as the value of the message (containsInstanceOf:)
literal frame
    #isKindOf:


Messages

When a send bytecode is encountered, the interpreter finds the CompiledMethod indicated by the message as follows.

  1. Find the message receiver. The receiver is below the arguments on the stack. The number of arguments is indicated in the send bytecode.
  2. Access a message dictionary. The original message dictionary is found in the receiver's class.
  3. Look up the message selector in the message dictionary. The selector is indicated in the send bytecode.
  4. If the selector is found, the associated CompiledMethod describes the response to the message.
  5. If the selector is not found, a new message dictionary must be searched (returning to step 3). The new message dictionary will be found in the superclass of the last class whose message dictionary was searched. This cycle may be repeated several times, traveling up the superclass chain.


If the selector is not found in the receiver's class nor in any of its superclasses, an error is reported, and execution of the bytecodes following the send is suspended.


❏ Superclass Sends A variation of the send bytecodes called supersends uses a slightly different algorithm to find the CompiledMethod associated with a message. Everything is the same except for the second step, which specifies the original message dictionary to search. When a super-send is encountered, the following second step is substituted.


2. Access a message dictionary. The original message dictionary is found in the superclass of the class in which the currently executing CompiledMethod was found.


Super-send bytecodes are used when super is used as the receiver of a message in a source method. The bytecode used to push the receiver will be the same as if self had been used, but a super-send bytecode will be used to describe the selector.


As an example of the use of a super-send, imagine a subclass of Rectangle called ShadedRectangle that adds an instance variable named shade. A Rectangle might respond to the message shade: by producing a new ShadedRectangle. ShadedRectangle provides a new method for the message intersect:, returning a ShadedRectangle instead of a Rectangle. This method must use super to access its own ability to actually compute the intersection.

intersect: aRectangle
    (super intersect: aRectangle)
        shade: shade


Active
ShadedRectangle intersect:
112 push the receiver (self) onto the stack
16 push the contents of the first temporary frame location (the argument aRectangle) onto the stack
133,33 send to super a single argument message with the selector in the second literal frame location (intersect:)
2 push the value of the receiver's third instance variable (shade) onto the stack
224 send a single argument message with the selector in the first literal frame location (shade:)
124 return the object on top of the stack as the value of the message (intersect:)
literal frame
    #shade:
    #intersect:
    Association: #ShadedRectangle ➛ ShadedRectangle


It is important to note that the initial class searched in response to a super-send will be the superclass of the receiver's class only if the CompiledMethod containing the super-send was originally found the receiver's class. If the CompiledMethod was originally found in a superclass of the receiver's class, the search will start in that class's superclass. Since the interpreter's state does not include the class in which it found each CompiledMethod, that information is included in the CompiledMethod itself. Every CompiledMethod that includes a super-send bytecode refers to the class in whose message dictionary it is found. The last entry of the literal frame of those CompiledMethods contains an association referring to the class.


Primitive Methods

The interpreter's actions after finding a CompiledMethod depend on whether or not the CompiledMethod indicates that a primitive method may be able to respond to the message. If no primitive method is indicated, a new MethodContext is created and made active as described in previous sections. If a primitive method is indicated in the CompiledMethod, the interpreter may be able to respond to the message without actually executing the bytecodes. For example, one of the primitive methods is associated with the + message to instances of SmallInteger.

+ addend
    <primitive: 1>
    super + addend


SmallInteger + associated with primitive #1
112 push the receiver (self) onto the stack
16 push the contents of the first temporary frame location (the argument addend) onto the stack
133,32 send to super a single argument message with the selector in the first literal frame location(+)
124 return the object on top of the stack as the value of the message (+)
literal frame
    #+


Even if a primitive method is indicated for a CompiledMethod, the interpreter may not be able to respond successfully. For example, the argument of the + message might not be another instance of SmallInteger or the sum might not be representable by a SmallInteger. If the interpreter cannot execute the primitive for some reason, the primitive is said to fail. When a primitive fails, the bytecodes in the CompiledMethod are executed as if the primitive method had not been indicated. The method for SmallInteger + indicates that the + method in the superclass (Integer) will be used if the primitive fails.


There are about a hundred primitive methods in the system that perform four types of operation. The exact function of all of the primitives will be described in Chapter 29.

  1. Arithmetic
  2. Storage management
  3. Control
  4. Input-output


The Object Memory

The object memory provides the interpreter with an interface to the objects that make up the Smalltalk-80 virtual image. Each object is associated with a unique identifier called its object pointer. The object memory and interpreter communicate about objects with object pointers. The size of object pointers determines the maximum number of objects a Smalltalk-80 system can contain. This number is not fixed by anything about the language, but the implementation described in this book uses 16-bit object pointers, allowing 65536 objects to be referenced. Implementation of the Smalltalk-80 system with larger object references will require changing certain parts of the virtual machine specification. It is not within the scope of this book to detail the relevant changes.


The object memory associates each object pointer with a set of other object pointers, Every object pointer is associated with the object pointer of a class. If an object has instance variables, its object pointer is also associated with the object pointers of their values. The individual instance variables are referred to by zero-relative integer indices. The value of an instance variable can be changed, but the class associated with an object cannot be changed. The object memory provides the following five fundamental functions to the interpreter.

1. Access the value of an object's instance variable. The object pointer of the instance and the index of the instance variable must be supplied. The object pointer of the instance variable's value is returned.
2. Change the value of an object's instance variable. The object pointer of the instance and the index of the instance variable must be supplied. The object pointer of the new value must also be supplied.
3. Access an object's class. The object pointer of the instance must be supplied. The object pointer of the instance's class is returned.
4. Create a new object. The object pointer of the new object's class and the number of instance variables it should have must be supplied. The object pointer of the new instance is returned.
5. Find the number of instance variables an object has. The object's pointer must be supplied. The number of instance variables is returned.


There is no explicit function of the object memory to remove an object no longer being used because these objects are reclaimed automatically. An object is reclaimed when there are no object pointers to it from other objects. This reclamation can be accomplished either by reference counting or garbage collection.


There are two additional features of the object memory that provide efficient representation of numerical information. The first of these sets aside certain object pointers for instances of class SmallInteger. The second allows objects to contain integer values instead of object pointers.


❏ Representation of Small Integers The instances of class SmallInteger represent the integers-16384 through 16383. Each of these instances is assigned a unique object pointer. These object pointers all have a 1 in the low-order bit position and the two's complement representation of their value in the high-order 15 bits. An instance of SmallInteger needs no instance storage since both its class and its value can be determined from its object pointer. Two additional functions are provided by the object memory to convert back and forth between SmallInteger object pointers and numerical values.


6. Find the numerical value represented by a SmallInteger. The object pointer of the SmallInteger must be supplied. The two's complement value is returned.
7. Find the SmallInteger representing a numerical value. The two's complement value must be supplied. A SmallInteger object pointer is returned.


This representation for SmallIntegers implies that there can be 32768 instances of the other classes in the system. It also implies that equality (=) and equivalence (= =) will be the same for instances of SmallInteger. Integers outside the range-16384 through 16383 are represented by instances of class LargePositiveInteger or LargeNegativeInteger. There may be several instances representing the same value, so equality and equivalence are different.


❏ Collections of Integer Values Another special representation is included for objects representing collections of integers. Instead of storing the object pointers of the SmallIntegers representing the contents of the collection, the actual numerical values are stored. The values in these special collections are constrained to be positive. There are two varieties of collection, one limiting its values to be less than 256 and the other limiting its values to be less than 65536. The object memory provides functions analogous to the first five listed in this section, but for objects whose contents are numerical values instead of object pointers.


The distinction between objects that contain object pointers and those that contain integer values is never visible to the Smalltalk-80 programmer. When one of these special numerical collections is accessed by sending it a message, the object pointer of an object representing the value is returned. The nature of these special collections is only evident in that they may refuse to store objects that do not represent integers within the proper range.


The Hardware

The Smalltalk-80 implementation has been described as a virtual machine to avoid unnecessary hardware dependencies. It is naturally assumed that the hardware will include a processor and more than enough memory to store the virtual image and the machine language routines simulating the interpreter and object memory. The current size of the virtual image requires at least a half megabyte of memory.


The size of the processor and the organization of the memory are not actually constrained by the virtual machine specification. Since object pointers are 16 bits, the most convenient arrangement would be a 16-bit processor and a memory of 16-bit words. As with the processor and memory of any system, the faster the better.


The other hardware requirements are imposed by the primitives that the virtual image depends on. These input-output devices and clocks are listed below.

  1. A bitmap display. It is most convenient if the bitmap being displayed can be located in the object memory, although this is not absolutely necessary.
  2. A pointing device.
  3. Three buttons associated with the pointing device. It is most convenient if these are physically located on the device.
  4. A keyboard, either decoded ASCII or undecoded ALTO.
  5. A disk. The standard Smalltalk-80 virtual image contains only a skeleton disk system that must be tailored to the actual disk used.
  6. A millisecond timer.
  7. A real time clock with one second resolution.


Notes