Scala e xchange 2013 haoyi li on metascala a tiny diy jvm
1 like2,122 views
Metascala is a minimalistic JVM implemented in approximately 3,000 lines of Scala, capable of loading and interpreting Java programs as well as itself. It is a fun experimental platform that provides insights into JVM internals, though it operates with limitations such as single-threaded execution and slow performance compared to larger JVMs like OpenJDK. Future improvements may include implementing threads, enhancing security, and optimizing the standard library interface.
Scala e xchange 2013 haoyi li on metascala a tiny diy jvm
- 1. Metascala A tiny DIY JVM https://github.com/lihaoyi/Metascala Li Haoyi [email protected] Scala Exchange 2nd Dec 2013
- 2. Who am I? Li Haoyi Write Python during the day Write Scala at night
- 3. What is Metascala? ● A JVM ● in 3000 lines of Scala ● Which can load & interpret java programs ● And can interpret itself!
- 4. Size comparison ● Metascala: ~3,000 lines ● Avian JVM: ~80,000 lines ● OpenJDK: ~1,000,000 lines
- 5. Basic Usage Create a new metascala VM Plain Old Java Object Captured variables are serialized into VM’s environment Closure’s class file is given to VM to load/parse/execute Result is extracted from VM into host environment No global state Any other classes necessary to evaluate the closure are loaded from the current Classpath
- 6. It’s Metacircular! Need to give the outer VM more than the 1mb default heap VM inside a VM! Simpler program avoids initializing the scala/java std libraries, which takes forever under double-interpretation. Takes a while (~10s) to produce result
- 7. Limitations ● Single-threaded ● Limited IO ● Slowww
- 8. Performance Comparison ● OpenJDK: 1.0x ● Metascala: ~100x ● Meta-Metascala: ~10000x
- 9. Why Metascala? ● Fun to explore the innards of the JVM ● An almost-fully secure Java runtime! ● Small size makes fiddling fun
- 10. Why Metascala? ● Fun to explore the innards of the JVM ● An almost-fully secure Java runtime! ● Small size makes fiddling fun
- 11. Quick Tour Immutable Defs: ~380 loc Native Bindings: ~650 loc Bytecode SSA transform: ~650 loc Runtime data structures: 820 loc Binary heap & Copying GC: 132 loc DIY scala-pickling: 132 loc “This is a VM”: 243 loc
- 12. Quick Tour: Tests Tests for basic Java features GC Fuzz-tests Test Metacircularity! Scala std lib usage
- 13. What’s a Heap? Fig 1. A Heap
- 14. What’s a Garbage Collector? Blit (copy) all roots to new heap Stop when you’ve scanned everything Not pseudocode Scan the already copied things for more things and copy them too
- 15. Why Metascala? ● Fun to explore the innards of the JVM ● An almost-fully secure Java runtime! ● Small size makes fiddling fun
- 17. And Limited Memory! Not an OOM Error! We throw this ourselves
- 18. Explicitly defined capabilities Every external call has to be explicitly defined and enabled
- 19. Security Characteristics ● ● ● ● Finite instruction count Finite memory Well-defined interface to outside world Doesn’t rely on Java Security Model at all! ● Still some holes…
- 20. Security Holes ● ● ● ● ● Classloader can read from anywhere Time spent classloading not accounted Memory spent on classes not accounted GC time not accounted “native” methods’ time/memory not accounted
- 21. Basic Problem Outside World User code resource consumption is bounded VM’s runtime resource usage can be made to grow arbitrarily large Classes Native method calls User Code Runtime Data Structures Garbage Collector
- 22. Possible Solution Put a VM Inside a VM! Works, ... but 10000x slowdown Outside World Fixed Unaccounted Costs Outside World Classes Native method calls User Code Runtime Data Structures Garbage Collector
- 23. Another Possible Solution Move more components into virtual runtime Difficult to bootstrap correctly WIP Outside World Native method calls Classes Garbage Collector User Code Runtime Data Structures
- 24. Why Metascala? ● Fun to explore the innards of the JVM ● An almost-fully secure Java runtime! ● Small size makes fiddling fun
- 25. Live Demo
- 26. Ugliness ● Long compile times ● Nasty JVM Interface ● Impossible Debugging
- 27. Long compile times ● ● 100 lines/s ● Twice as slow (50 lines/s) on my older machine!
- 28. Nasty JVM Interface Ideal World Initialized User Code Real World Linear Initialization Std Library Initialized User Code Std Library Clean Interfaces VM VM LazyInitialization means repeated dives back into lib/native code Nasty Language VM Interface
- 29. Java’s dirty little secret The Verbosity of Java with the Safety of C WTF! I’d never use these things!
- 30. You probably do What happens if you don’t have them Almost every Java program ever uses these things.
- 31. Next Steps ● Maximize correctness ○ Implement Threads & IO ○ Fix bugs (GC, native calls, etc.) ● Solidify security characteristics ○ Still plenty of unaccounted-for memory/processing ○ Some can be hosted “within” VM itself ● Simplify Std-Lib/VM interface ○ Try using Android Std Lib?
- 32. Possible Experiments ● Native codegen instead of an interpreter? ○ Generate/exec native code through JNI ○ Heap is already a binary blob that can be easily passed to native code ● Bytecode transforms and optimizations? ○ Already in SSA form ● Continuations, Isolates, Value Classes? ● Port the whole thing to Scala.Js?
- 33. Metascala: a tiny DIY JVM Ask me about: ● Single Static Assignment form ● Copying Garbage Collection ● sun.misc.Unsafe ● Warts of the .class file format ● Capabilities-based security ● Abandoned approaches