Modeling code manipulation in JIT compilers

Heui Chan Lim; Xiyu Kang; Saumya Debray

doi:10.1145/3520313.3534656

Modeling code manipulation in JIT compilers

Heui Chan Lim, Xiyu Kang, Saumya Debray

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

1 Scopus citations

Abstract

Just-in-Time (JIT) compilers are widely used to improve the performance of interpreter-based language implementations by creating optimized code at runtime. However, bugs in the JIT compiler's code manipulation and optimization can result in the generation of incorrect code. Such bugs can be difficult to diagnose and fix, and can result in exploitable vulnerabilities. Unfortunately, existing approaches to automatic bug localization do not carry over well to such bugs. This paper discusses a different approach to analyzing JIT compiler optimization behaviors, based on using dynamic analysis to construct abstract models of the JIT compiler's optimizer and back end. By comparing the models obtained for buggy and non-buggy executions of the JIT compiler, we can pinpoint the components of the JIT compiler's internal representation that have been affected by the bug; this can then be mapped back to identify the buggy code. Our experiments with two real bugs for Google V8 JIT compiler, TurboFan, show the utility and practicality of our approach.

Original language	English (US)
Title of host publication	SOAP 2022 - Proceedings of the 11th ACM SIGPLAN International Workshop on the State Of the Art in Program Analysis, co-located with PLDI 2022
Editors	Laure Gonnord, Laura Titolo
Publisher	Association for Computing Machinery, Inc
Pages	9-15
Number of pages	7
ISBN (Electronic)	9781450392747
DOIs	https://doi.org/10.1145/3520313.3534656
State	Published - Jun 14 2022
Externally published	Yes
Event	11th ACM SIGPLAN International Workshop on the State Of the Art in Program Analysis, SOAP 2022, co-located with the 43rd ACM SIGPLAN International Conference on Programming Language Design and Implementation, PLDI 2022 - San Diego, United States Duration: Jun 14 2022 → …

Publication series

Name	SOAP 2022 - Proceedings of the 11th ACM SIGPLAN International Workshop on the State Of the Art in Program Analysis, co-located with PLDI 2022

Conference

Conference	11th ACM SIGPLAN International Workshop on the State Of the Art in Program Analysis, SOAP 2022, co-located with the 43rd ACM SIGPLAN International Conference on Programming Language Design and Implementation, PLDI 2022
Country/Territory	United States
City	San Diego
Period	6/14/22 → …

Keywords

dynamic code generation
jit compiler
optimization
program analysis

ASJC Scopus subject areas

Software
Computer Graphics and Computer-Aided Design

Access to Document

10.1145/3520313.3534656

Cite this

Lim, H. C., Kang, X., & Debray, S. (2022). Modeling code manipulation in JIT compilers. In L. Gonnord, & L. Titolo (Eds.), SOAP 2022 - Proceedings of the 11th ACM SIGPLAN International Workshop on the State Of the Art in Program Analysis, co-located with PLDI 2022 (pp. 9-15). (SOAP 2022 - Proceedings of the 11th ACM SIGPLAN International Workshop on the State Of the Art in Program Analysis, co-located with PLDI 2022). Association for Computing Machinery, Inc. https://doi.org/10.1145/3520313.3534656

Modeling code manipulation in JIT compilers. / Lim, Heui Chan; Kang, Xiyu; Debray, Saumya.
SOAP 2022 - Proceedings of the 11th ACM SIGPLAN International Workshop on the State Of the Art in Program Analysis, co-located with PLDI 2022. ed. / Laure Gonnord; Laura Titolo. Association for Computing Machinery, Inc, 2022. p. 9-15 (SOAP 2022 - Proceedings of the 11th ACM SIGPLAN International Workshop on the State Of the Art in Program Analysis, co-located with PLDI 2022).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Lim, HC, Kang, X & Debray, S 2022, Modeling code manipulation in JIT compilers. in L Gonnord & L Titolo (eds), SOAP 2022 - Proceedings of the 11th ACM SIGPLAN International Workshop on the State Of the Art in Program Analysis, co-located with PLDI 2022. SOAP 2022 - Proceedings of the 11th ACM SIGPLAN International Workshop on the State Of the Art in Program Analysis, co-located with PLDI 2022, Association for Computing Machinery, Inc, pp. 9-15, 11th ACM SIGPLAN International Workshop on the State Of the Art in Program Analysis, SOAP 2022, co-located with the 43rd ACM SIGPLAN International Conference on Programming Language Design and Implementation, PLDI 2022, San Diego, United States, 6/14/22. https://doi.org/10.1145/3520313.3534656

Lim HC, Kang X, Debray S. Modeling code manipulation in JIT compilers. In Gonnord L, Titolo L, editors, SOAP 2022 - Proceedings of the 11th ACM SIGPLAN International Workshop on the State Of the Art in Program Analysis, co-located with PLDI 2022. Association for Computing Machinery, Inc. 2022. p. 9-15. (SOAP 2022 - Proceedings of the 11th ACM SIGPLAN International Workshop on the State Of the Art in Program Analysis, co-located with PLDI 2022). doi: 10.1145/3520313.3534656

Lim, Heui Chan ; Kang, Xiyu ; Debray, Saumya. / Modeling code manipulation in JIT compilers. SOAP 2022 - Proceedings of the 11th ACM SIGPLAN International Workshop on the State Of the Art in Program Analysis, co-located with PLDI 2022. editor / Laure Gonnord ; Laura Titolo. Association for Computing Machinery, Inc, 2022. pp. 9-15 (SOAP 2022 - Proceedings of the 11th ACM SIGPLAN International Workshop on the State Of the Art in Program Analysis, co-located with PLDI 2022).

@inproceedings{8f21ea7459554067a72bc1c5c7b2529d,

title = "Modeling code manipulation in JIT compilers",

abstract = "Just-in-Time (JIT) compilers are widely used to improve the performance of interpreter-based language implementations by creating optimized code at runtime. However, bugs in the JIT compiler's code manipulation and optimization can result in the generation of incorrect code. Such bugs can be difficult to diagnose and fix, and can result in exploitable vulnerabilities. Unfortunately, existing approaches to automatic bug localization do not carry over well to such bugs. This paper discusses a different approach to analyzing JIT compiler optimization behaviors, based on using dynamic analysis to construct abstract models of the JIT compiler's optimizer and back end. By comparing the models obtained for buggy and non-buggy executions of the JIT compiler, we can pinpoint the components of the JIT compiler's internal representation that have been affected by the bug; this can then be mapped back to identify the buggy code. Our experiments with two real bugs for Google V8 JIT compiler, TurboFan, show the utility and practicality of our approach.",

keywords = "dynamic code generation, jit compiler, optimization, program analysis",

author = "Lim, {Heui Chan} and Xiyu Kang and Saumya Debray",

note = "Funding Information: This research was supported in part by the National Science Foundation under grant no. 1908313. Publisher Copyright: {\textcopyright} 2022 Owner/Author.; 11th ACM SIGPLAN International Workshop on the State Of the Art in Program Analysis, SOAP 2022, co-located with the 43rd ACM SIGPLAN International Conference on Programming Language Design and Implementation, PLDI 2022 ; Conference date: 14-06-2022",

year = "2022",

month = jun,

day = "14",

doi = "10.1145/3520313.3534656",

language = "English (US)",

series = "SOAP 2022 - Proceedings of the 11th ACM SIGPLAN International Workshop on the State Of the Art in Program Analysis, co-located with PLDI 2022",

publisher = "Association for Computing Machinery, Inc",

pages = "9--15",

editor = "Laure Gonnord and Laura Titolo",

booktitle = "SOAP 2022 - Proceedings of the 11th ACM SIGPLAN International Workshop on the State Of the Art in Program Analysis, co-located with PLDI 2022",

}

TY - GEN

T1 - Modeling code manipulation in JIT compilers

AU - Lim, Heui Chan

AU - Kang, Xiyu

AU - Debray, Saumya

PY - 2022/6/14

Y1 - 2022/6/14

N2 - Just-in-Time (JIT) compilers are widely used to improve the performance of interpreter-based language implementations by creating optimized code at runtime. However, bugs in the JIT compiler's code manipulation and optimization can result in the generation of incorrect code. Such bugs can be difficult to diagnose and fix, and can result in exploitable vulnerabilities. Unfortunately, existing approaches to automatic bug localization do not carry over well to such bugs. This paper discusses a different approach to analyzing JIT compiler optimization behaviors, based on using dynamic analysis to construct abstract models of the JIT compiler's optimizer and back end. By comparing the models obtained for buggy and non-buggy executions of the JIT compiler, we can pinpoint the components of the JIT compiler's internal representation that have been affected by the bug; this can then be mapped back to identify the buggy code. Our experiments with two real bugs for Google V8 JIT compiler, TurboFan, show the utility and practicality of our approach.

AB - Just-in-Time (JIT) compilers are widely used to improve the performance of interpreter-based language implementations by creating optimized code at runtime. However, bugs in the JIT compiler's code manipulation and optimization can result in the generation of incorrect code. Such bugs can be difficult to diagnose and fix, and can result in exploitable vulnerabilities. Unfortunately, existing approaches to automatic bug localization do not carry over well to such bugs. This paper discusses a different approach to analyzing JIT compiler optimization behaviors, based on using dynamic analysis to construct abstract models of the JIT compiler's optimizer and back end. By comparing the models obtained for buggy and non-buggy executions of the JIT compiler, we can pinpoint the components of the JIT compiler's internal representation that have been affected by the bug; this can then be mapped back to identify the buggy code. Our experiments with two real bugs for Google V8 JIT compiler, TurboFan, show the utility and practicality of our approach.

KW - dynamic code generation

KW - jit compiler

KW - optimization

KW - program analysis

UR - http://www.scopus.com/inward/record.url?scp=85146234243&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85146234243&partnerID=8YFLogxK

U2 - 10.1145/3520313.3534656

DO - 10.1145/3520313.3534656

M3 - Conference contribution

AN - SCOPUS:85146234243

T3 - SOAP 2022 - Proceedings of the 11th ACM SIGPLAN International Workshop on the State Of the Art in Program Analysis, co-located with PLDI 2022

SP - 9

EP - 15

BT - SOAP 2022 - Proceedings of the 11th ACM SIGPLAN International Workshop on the State Of the Art in Program Analysis, co-located with PLDI 2022

A2 - Gonnord, Laure

A2 - Titolo, Laura

PB - Association for Computing Machinery, Inc

T2 - 11th ACM SIGPLAN International Workshop on the State Of the Art in Program Analysis, SOAP 2022, co-located with the 43rd ACM SIGPLAN International Conference on Programming Language Design and Implementation, PLDI 2022

Y2 - 14 June 2022

ER -

Modeling code manipulation in JIT compilers

Abstract

Publication series

Conference

Keywords

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this