The ART of Sharing Points-to Analysis: Reusing Points-to Analysis Results Safely and(…)
summary
A scheme called ART encodes flow-sensitive, context-insensitive points-to results so a consumer can regenerate them without recomputing fixed points and verify safety by a single fixpoint check that detects tampering. Implementation uses Soot as producer and an ART regenerator in the Eclipse OpenJ9 JIT as a consumer. Regeneration takes under 1% of the original analysis time, ART adds about 4% storage overhead, and integrating regeneration costs roughly 10% extra per-method compilation while improving end-to-end runtime.
AI-written summary · about
chapters (12)
- 0:00Introduction
- 1:26Precise PTA in JIT compilers
- 3:09Producer-consumer PTA results
- 4:10Fix-point consistency checks
- 6:16Tampered PTA results detection
- 9:01ART encoding of PTA results
- 10:47Interprocedural PTA summaries
- 14:41ART encoder and regenerator
- 15:46Implementation and evaluation
- 16:50Performance impact benchmarks
- 17:10Summary and conclusions
- 17:30Q&A
related talks
29mHow (Not) To Write Java Pointer Analyses After 2020
16mHow (Not) to Write Java Pointer Analyses after 2020
37mA Static Tainting Analysis Method for Aspect-Oriented Programs
15mMaking Pointer Analysis More Precise by Unleashing the Power of Selective Context Sensitivity
42mAn Opinionated Overview on Static Analysis for Java
36mThe art of static code analysis
18mSHARP: fast incremental context-sensitive pointer analysis for Java
59mTools & Tips to Modernize Your Java code with AI