RBuilder and ByteSurgeon
0 likes583 views
The document discusses working with bytecode through the use of two tools: IRBuilder and ByteSurgeon. IRBuilder allows for easy generation of bytecode by providing a symbolic assembler interface. ByteSurgeon allows for transformation of existing bytecode through insertion, deletion, and replacement of instructions. Advanced features of ByteSurgeon like metavariables allow access to values on the stack before and after method sends for instrumentation purposes like logging.
![Jump Bytecodes
• Control Flow inside one method
• Used to implement control-flow efficiently
• Example:
^ 1<2 ifTrue: ['true']
9 <76> pushConstant: 1
10 <77> pushConstant: 2
11 <B2> send: <
12 <99> jumpFalse: 15
13 <20> pushConstant: 'true'
14 <90> jumpTo: 16
15 <73> pushConstant: nil
16 <7C> returnTop
Marcus Denker](https://image.slidesharecdn.com/dcc02bytecodeandbytesurgeon-090827193952-phpapp02/85/RBuilder-and-ByteSurgeon-11-320.jpg)
![IRBuilder: Symbolic Jumps
• Jump targets are resolved:
• Example: false ifTrue: [’true’] ifFalse: [’false’]
iRMethod := IRBuilder new
numRargs: 1;
addTemps: #(self); "receiver"
pushLiteral: false;
jumpAheadTo: #false if: false;
pushLiteral: 'true';
"ifTrue: ['true']"
jumpAheadTo: #end;
jumpAheadTarget: #false;
pushLiteral: 'false';
"ifFalse: ['false']"
jumpAheadTarget: #end;
returnTop;
ir.
Marcus Denker](https://image.slidesharecdn.com/dcc02bytecodeandbytesurgeon-090827193952-phpapp02/85/RBuilder-and-ByteSurgeon-25-320.jpg)
![Logging with Bytesurgen
• Goal: Change the method without changing program
text
• Example:
(Example>>#example)instrumentSend: [:send |
send insertBefore:
‘Transcript show: ‘’sending #test’’ ‘.
]
Marcus Denker](https://image.slidesharecdn.com/dcc02bytecodeandbytesurgeon-090827193952-phpapp02/85/RBuilder-and-ByteSurgeon-32-320.jpg)
![Logging: Step by Step
(Example>>#example)instrumentSend: [:send |
send insertBefore:
‘Transcript show: ‘’sending #test’’ ‘.
]
Example >> #example
Class Name of Method
>>: - takes a name of a method
- returns the CompiledMethod object
Marcus Denker](https://image.slidesharecdn.com/dcc02bytecodeandbytesurgeon-090827193952-phpapp02/85/RBuilder-and-ByteSurgeon-33-320.jpg)
![Logging: Step by Step
(Example>>#example)instrumentSend: [:send |
send insertBefore:
‘Transcript show: ‘’sending #test’’ ‘.
]
• instrumentSend:
– takes a block as an argument
– evaluates it for all send bytecodes
Marcus Denker](https://image.slidesharecdn.com/dcc02bytecodeandbytesurgeon-090827193952-phpapp02/85/RBuilder-and-ByteSurgeon-34-320.jpg)
![Logging: Step by Step
(Example>>#example)instrumentSend: [:send |
send insertBefore:
‘Transcript show: ‘’sending #test’’ ‘.
]
• The block has one parameter: send
• It is executed for each send bytecode in the method
Marcus Denker](https://image.slidesharecdn.com/dcc02bytecodeandbytesurgeon-090827193952-phpapp02/85/RBuilder-and-ByteSurgeon-35-320.jpg)
![Logging: Step by Step
(Example>>#example)instrumentSend: [:send |
send insertBefore:
‘Transcript show: ‘’sending #test’’ ‘.
]
• Objects describing bytecode understand how to
insert code
– insertBefor
– insertAfter
– replace
Marcus Denker](https://image.slidesharecdn.com/dcc02bytecodeandbytesurgeon-090827193952-phpapp02/85/RBuilder-and-ByteSurgeon-36-320.jpg)
![Logging: Step by Step
(Example>>#example)instrumentSend: [:send |
send insertBefore:
‘Transcript show: ‘’sending #test’’ ‘.
]
• The code to be inserted.
• Double quoting for string inside string
–Transcript show: ’sending #test’
Marcus Denker](https://image.slidesharecdn.com/dcc02bytecodeandbytesurgeon-090827193952-phpapp02/85/RBuilder-and-ByteSurgeon-37-320.jpg)
![ByteSurgeon Usage
• On Methods or Classes:
MyClass instrument: [.... ].
(MyClass>>#myMethod) instrument: [.... ].
• Different instrument methods:
– instrument:
– instrumentSend:
– instrumentTempVarRead:
– instrumentTempVarStore:
– instrumentTempVarAccess:
– same for InstVar
Marcus Denker](https://image.slidesharecdn.com/dcc02bytecodeandbytesurgeon-090827193952-phpapp02/85/RBuilder-and-ByteSurgeon-39-320.jpg)
![Advanced ByteSurgeon
• With Bytesurgeon, something like:
(Example>>#example)instrumentSend: [:send |
send insertAfter:
‘Logger logSendTo: ?’ .
]
• How can we access the receiver of the send?
• Solution: Metavariable
Marcus Denker](https://image.slidesharecdn.com/dcc02bytecodeandbytesurgeon-090827193952-phpapp02/85/RBuilder-and-ByteSurgeon-42-320.jpg)
![Advanced ByteSurgeon
• With Bytesurgeon, something like:
(Example>>#example)instrumentSend: [:send |
send insertAfter:
‘Logger logSendTo: <meta: #receiver>’ .
]
• How can we access the receiver of the send?
• Solution: Metavariable
Marcus Denker](https://image.slidesharecdn.com/dcc02bytecodeandbytesurgeon-090827193952-phpapp02/85/RBuilder-and-ByteSurgeon-43-320.jpg)
RBuilder and ByteSurgeon
- 1. Working with Bytecodes: IRBuilder and ByteSurgeon Marcus Denker
- 2. Reasons for working with Bytecode • Generating Bytecode – Implementing compilers for other languages – Experimentation with new language features • Bytecode Transformation – Adaptation of running Systems – Tracing / Debugging – New language features Marcus Denker
- 3. Overview 1. Introduction to Squeak Bytecodes 2. Generating Bytecode with IRBuilder 3. Introduction to ByteSurgeon Marcus Denker
- 4. The Squeak Virtual Machine • From last lecture: – Virtual machine provides a virtual processor – Bytecode: The ‘machine-code’ of the virtual machine – Smalltalk (like Java): Stack machine • Today: – Closer look at Squeak bytecode Marcus Denker
- 5. Bytecode in the CompiledMethod • CompiledMethods format: (Number>>#asInteger)inspect Header Number of temps, literals... Array of all Literals Literal Objects Bytecode Pointer to Trailer Source Marcus Denker
- 6. Example: Number>>asInteger • Smalltalk code: Number>>asInteger "Answer an Integer nearest the receiver toward zero." ^self truncated • Symbolic Bytecode 9 <70> self 10 <D0> send: truncated 11 <7C> returnTop Marcus Denker
- 7. Example: Step by Step • 9 <70> self – The receiver (self) is pushed on the stack • 10 <D0> send: truncated – Bytecode 208: send literal selector 1 – Get the selector from the first literal – start message lookup in the class of the object that is top of the stack – result is pushed on the stack • 11 <7C> returnTop – return the object on top of the stack to the calling method Marcus Denker
- 8. Squeak Bytecodes • 256 Bytecodes, four groups: – Stack Bytecodes • Stack manipulation: push / pop / dup – Send Bytecodes • Invoke Methods – Return Bytecodes • Return to caller – Jump Bytecodes • Control flow inside a method Marcus Denker
- 9. Stack Bytecodes • Push values on the stack, e.g., temps, instVars, literals – e.g: 16 - 31: push instance variable • Push Constants (False/True/Nil/1/0/2/-1) • Push self, thisContext • Duplicate top of stack • Pop Marcus Denker
- 10. Sends and Returns • Sends: receiver is on top of stack – Normal send – Super Sends – Hard-coded sends for efficiency, e.g. +, - • Returns – Return top of stack to the sender – Return from a block – Special bytecodes for return self, nil, true, false (for efficiency) Marcus Denker
- 11. Jump Bytecodes • Control Flow inside one method • Used to implement control-flow efficiently • Example: ^ 1<2 ifTrue: ['true'] 9 <76> pushConstant: 1 10 <77> pushConstant: 2 11 <B2> send: < 12 <99> jumpFalse: 15 13 <20> pushConstant: 'true' 14 <90> jumpTo: 16 15 <73> pushConstant: nil 16 <7C> returnTop Marcus Denker
- 12. What you should have learned... • ... dealing with bytecodes directly is possible, but very boring. • We want reusable abstractions that hide the details (e.g. the different send bytecodes) • We would like to have frameworks for – Generating bytecode easily – Transforming bytecode
- 13. Generating Bytecodes • IRBuilder: A tool for generating bytecode • Part of the new compiler for Squeak 3.9 • Idea: a symbolic Assembler for Squeak Marcus Denker
- 14. IRBuilder: Simple Example • Number>>asInteger iRMethod := IRBuilder new numRargs: 1; addTemps: #(self); "receiver" pushTemp: #self; send: #truncated; returnTop; ir. aCompiledMethod := iRMethod compiledMethod. aCompiledMethod valueWithReceiver:3.5 arguments: #() Marcus Denker
- 15. IRBuilder: Step by Step • Number>>asInteger iRMethod := IRBuilder new – Make a instance of IRBuilder Marcus Denker
- 16. IRBuilder: Step by Step • Number>>asInteger iRMethod := IRBuilder new numRargs: 1; – Define arguments. Note: “self” is default argument Marcus Denker
- 17. IRBuilder: Step by Step • Number>>asInteger iRMethod := IRBuilder new numRargs: 1; addTemps: #(self); "receiver" – define temporary variables. Note: arguments are temps Marcus Denker
- 18. IRBuilder: Step by Step • Number>>asInteger iRMethod := IRBuilder new numRargs: 1; addTemps: #(self); "receiver" pushTemp: #self – push “self” on the stack Marcus Denker
- 19. IRBuilder: Step by Step • Number>>asInteger iRMethod := IRBuilder new numRargs: 1; addTemps: #(self); "receiver" pushTemp: #self send: #truncated; – call method truncated on “self” Marcus Denker
- 20. IRBuilder: Step by Step • Number>>asInteger iRMethod := IRBuilder new numRargs: 1; addTemps: #(self); "receiver" pushTemp: #self send: #truncated; returnTop; – return Top of Stack Marcus Denker
- 21. IRBuilder: Step by Step • Number>>asInteger iRMethod := IRBuilder new numRargs: 1; addTemps: #(self); "receiver" pushTemp: #self send: #truncated; returnTop; ir. – tell IRBuilder to generate Intermediate Representation (IR) Marcus Denker
- 22. IRBuilder: Step by Step • Number>>asInteger iRMethod := IRBuilder new numRargs: 1; addTemps: #(self); "receiver" pushTemp: #self send: #truncated; returnTop; ir. aCompiledMethod := iRMethod compiledMethod. – Generate method from IR Marcus Denker
- 23. IRBuilder: Step by Step • Number>>asInteger iRMethod := IRBuilder new numRargs: 1; addTemps: #(self); "receiver" pushTemp: #self send: #truncated; returnTop; ir. aCompiledMethod := iRMethod compiledMethod. aCompiledMethod valueWithReceiver:3.5 arguments: #() – Execute the method with reveiver 3.5 and no arguments. – “3.5 truncated” Marcus Denker
- 24. IRBuilder: Stack Manipulation • popTop - remove the top of stack • pushDup - push top of stack on the stack • pushLiteral: • pushReceiver - push self • pushThisContext Marcus Denker
- 25. IRBuilder: Symbolic Jumps • Jump targets are resolved: • Example: false ifTrue: [’true’] ifFalse: [’false’] iRMethod := IRBuilder new numRargs: 1; addTemps: #(self); "receiver" pushLiteral: false; jumpAheadTo: #false if: false; pushLiteral: 'true'; "ifTrue: ['true']" jumpAheadTo: #end; jumpAheadTarget: #false; pushLiteral: 'false'; "ifFalse: ['false']" jumpAheadTarget: #end; returnTop; ir. Marcus Denker
- 26. IRBuiler: Instance Variables • Access by offset • Read: getField: – receiver on top of stack • Write: setField: – receiver and value on stack • Example: set the first instance variable to 2 iRMethod := IRBuilder new numRargs: 1; addTemps: #(self); "receiver" pushLiteral: 2; pushTemp: #self; setField: 1; pushTemp: #self; returnTop; ir. aCompiledMethod := iRMethod compiledMethod. aCompiledMethod valueWithReceiver: 1@2 arguments: #() Marcus Denker
- 27. IRBuilder: Temporary Variables • Accessed by name • Define with addTemp: / addTemps: • Read with pushTemp: • Write with storeTemp: • Examle: set variables a and b, return value of a iRMethod := IRBuilder new numRargs: 1; addTemps: #(self); "receiver" addTemps: #(a b); pushLiteral: 1; storeTemp: #a; pushLiteral: 2; storeTemp: #b; pushTemp: #a; returnTop; ir. Marcus Denker
- 28. IRBuilder: Sends • normal send builder pushLiteral: ‘hello’ builder send: #size; • super send .... builder send: #selector toSuperOf: aClass; – The second parameter specifies the class were the lookup starts. Marcus Denker
- 29. IRBuilder: Lessons learned • IRBuilder: Easy bytecode generation – Jumps – Instance variable – Temporary variables – Sends • Next: Manipulating bytecode
- 30. ByteSurgeon • Library for bytecode transformation in Smalltalk • Full flexibility of Smalltalk Runtime • Provides high-level API • For Squeak, but portable • Runtime transformation needed for • Adaptation of running systems • Tracing / debugging • New language features (MOP, AOP) Marcus Denker
- 31. Example: Logging • Goal: logging message send. • First way: Just edit the text: example self test. example Transcript show: ‘sending #test’. self test. Marcus Denker
- 32. Logging with Bytesurgen • Goal: Change the method without changing program text • Example: (Example>>#example)instrumentSend: [:send | send insertBefore: ‘Transcript show: ‘’sending #test’’ ‘. ] Marcus Denker
- 33. Logging: Step by Step (Example>>#example)instrumentSend: [:send | send insertBefore: ‘Transcript show: ‘’sending #test’’ ‘. ] Example >> #example Class Name of Method >>: - takes a name of a method - returns the CompiledMethod object Marcus Denker
- 34. Logging: Step by Step (Example>>#example)instrumentSend: [:send | send insertBefore: ‘Transcript show: ‘’sending #test’’ ‘. ] • instrumentSend: – takes a block as an argument – evaluates it for all send bytecodes Marcus Denker
- 35. Logging: Step by Step (Example>>#example)instrumentSend: [:send | send insertBefore: ‘Transcript show: ‘’sending #test’’ ‘. ] • The block has one parameter: send • It is executed for each send bytecode in the method Marcus Denker
- 36. Logging: Step by Step (Example>>#example)instrumentSend: [:send | send insertBefore: ‘Transcript show: ‘’sending #test’’ ‘. ] • Objects describing bytecode understand how to insert code – insertBefor – insertAfter – replace Marcus Denker
- 37. Logging: Step by Step (Example>>#example)instrumentSend: [:send | send insertBefore: ‘Transcript show: ‘’sending #test’’ ‘. ] • The code to be inserted. • Double quoting for string inside string –Transcript show: ’sending #test’ Marcus Denker
- 38. Inside ByteSurgeon • Uses IRBuilder internally Decompile Compile Bytecode IR Bytecode • Transformation (Code inlining) done on IR Marcus Denker
- 39. ByteSurgeon Usage • On Methods or Classes: MyClass instrument: [.... ]. (MyClass>>#myMethod) instrument: [.... ]. • Different instrument methods: – instrument: – instrumentSend: – instrumentTempVarRead: – instrumentTempVarStore: – instrumentTempVarAccess: – same for InstVar Marcus Denker
- 40. ByteSurgeon: Lessons learned • ByteSurgeon: Tool for editing bytecode – Simple example – Based on IRBuilder • Next: Advanced ByteSurgeon
- 41. Advanced ByteSurgeon: • Goal: extend a send with after logging example self test. example self test. Logger logSendTo: self. Marcus Denker
- 42. Advanced ByteSurgeon • With Bytesurgeon, something like: (Example>>#example)instrumentSend: [:send | send insertAfter: ‘Logger logSendTo: ?’ . ] • How can we access the receiver of the send? • Solution: Metavariable Marcus Denker
- 43. Advanced ByteSurgeon • With Bytesurgeon, something like: (Example>>#example)instrumentSend: [:send | send insertAfter: ‘Logger logSendTo: <meta: #receiver>’ . ] • How can we access the receiver of the send? • Solution: Metavariable Marcus Denker
- 44. Implementation Metavariables • Stack during send: receiver arg1 arg2 result before after • Problem 1: After send, receiver is not available • Problem II: Before send, receiver is deep in the stack Marcus Denker
- 45. Metavariables: Implementation • Solution: ByteSurgeon generates preamble – Pop the arguments into temps – Pop the receiver into temps – Rebuild the stack – Do the send – Now we can acces the receiver even after the send Marcus Denker
- 46. Metavariables: Implementation 25 <70> self 26 <81 40> storeIntoTemp: 0 Preamble 28 <D0> send: test 29 <41> pushLit: Transcript 30 <10> pushTemp: 0 Inlined Code 31 <E2> send: show: 32 <87> pop 33 <87> pop 34 <78> returnSelf Marcus Denker
- 47. End • Short overview of Squeak bytecode • Introduction to bytecode generation with IRBuilder • Manipulating bytecode with ByteSurgeon • Questions?