A modular approach to on-stack replacement in LLVM

Lameed, Nurudeen; Hendren, Laurie

doi:10.1145/2451512.2451541

Cited by 14 publications

(6 citation statements)

References 16 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…D'Elia and Demetrescu [4] developed OSRKit, a library built on LLVM for 'dynamically transferring execution between different versions of a function at run time', which they define as 'OSR'. It is an improvement over its predecessor in McJIT by Lameed and Hendren [15]. In OSRKit, the old version of a function tail-calls a stub which recompiles the function and then tail-calls the new version.…”

Section: Related Workmentioning

confidence: 99%

Hop, Skip, & Jump

WangKunshan¹,

Blackburn²,

Hosking³

et al. 2018

SIGPLAN Not.

View full text Add to dashboard Cite

On-stack replacement (OSR) is a performance-critical technology for many languages, especially dynamic languages. Conventional wisdom, apparent in JavaScript engines such as V8 and SpiderMonkey, is that OSR must be implemented in a low-level (i.e., in assembly) and language-specific way. This paper presents an OSR abstraction based on Swapstack, materialized as the API for a low-level virtual machine, and shows how the abstraction of resumption protocols facilitates an elegant implementation of this API on real hardware. Using an experimental JavaScript implementation, we demonstrate that this API enables the language implementation to perform OSR without the need to deal with machine-level details. We also show that the API itself is implementable on concrete hardware. This work helps crystallize OSR abstractions and, by providing a reusable implementation, brings OSR within reach for more language implementers.

show abstract

Section: Related Workmentioning

confidence: 99%

Hop, Skip, & Jump

WangKunshan¹,

Blackburn²,

Hosking³

et al. 2018

SIGPLAN Not.

View full text Add to dashboard Cite

show abstract

“…We have discussed OSR in LLVM in Section 5.4. Also, an earlier work [29] provides support in the legacy JIT for transitions at loop headers when no state adjustment is required.…”

Section: Related Workmentioning

confidence: 99%

“…As speculative optimizations are essential for high performance, deoptimization may become necessary: it can be supported either by reconstructing a frame for the interpreter or the baseline compiled code, which both have a fixed stack layout, or by performing OSR into dynamically generated code as in Jikes RVM [15]. Following the classification proposed in [29], we use OSR to refer to both optimizing and deoptimizing transitions.…”

Section: Introductionmentioning

confidence: 99%

On-stack replacement, distilled

D’Elia

DemetrescuCamil

2018

SIGPLAN Not.

View full text Add to dashboard Cite

On-stack replacement (OSR) is essential technology for adaptive optimization, allowing changes to code actively executing in a managed runtime. The engineering aspects of OSR are well-known among VM architects, with several implementations available to date. However, OSR is yet to be explored as a general means to transfer execution between related program versions, which can pave the road to unprecedented applications that stretch beyond VMs. We aim at filling this gap with a constructive and provably correct OSR framework, allowing a class of general-purpose transformation functions to yield a special-purpose replacement. We describe and evaluate an implementation of our technique in LLVM. As a novel application of OSR, we present a feasibility study on debugging of optimized code, showing how our techniques can be used to fix variables holding incorrect values at breakpoints due to optimizations.

show abstract

“…Reference [14] proposes a modular approach to LLVM developers to facilitate the use of OSR by others through a well defined API. OSR points are restricted to beginning of loop bodies.…”

Section: A On-stack Replacementmentioning

confidence: 99%

“…On-stack replacement technology depends on extracting context from a suspended program, changing stack and code as needed, and then switching [5]. However, current dynamic on stack replacement depends on previously known, inserted in code, safe points [2] [14]. While safe points make the task easier, it limits the possible frequency of recompilations, as well as randomness, which hurts required security.…”

Section: Introductionmentioning

confidence: 99%

On-stack replacement to improve JIT-based obfuscation a preliminary study

Yusuf

El-Mahdy

Rohou

2013

2013 Second International Japan-Egypt Conference on Electronics, Communications and Computers (JEC-ECC)

View full text Add to dashboard Cite

As more devices are connecting together, more effective security techniques are needed to protect running software from hackers. One possible security technique is to continuously change the binary code running of given software by recompiling it on the fly. This switching need to be done frequently, quickly, and randomly, not constrained by specific locations in code, to make it difficult for the hacker to track the behavior of the running code or predict its functionality. In our research we are working on a technique that recompiles speculatively and concurrently with current execution, and switches to the new compiled version dynamically, at arbitrary points. This paper presents an early analytical study augmented by experimental analysis on manually applying this technique on simple kernels, to study the concept in comparison with other similar techniques.

show abstract

A modular approach to on-stack replacement in LLVM

Cited by 14 publications

References 16 publications

Hop, Skip, & Jump

Hop, Skip, & Jump

On-stack replacement, distilled

On-stack replacement to improve JIT-based obfuscation a preliminary study

Contact Info

Product

Resources

About