HomeHome

A new abstraction for summary-based pointer analysis

Marcio Buss; Daniel Brand; Vugranam Sreedhar; Stephen A. Edwards

Title:
A new abstraction for summary-based pointer analysis
Author(s):
Buss, Marcio
Brand, Daniel
Sreedhar, Vugranam
Edwards, Stephen A.
Date:
Type:
Reports
Department(s):
Computer Science
Persistent URL:
Series:
IBM Technical Report
Part Number:
RC24104
Publisher:
IBM
Publisher Location:
Yorktown Heights, NY
Abstract:
W e propose a new abstraction for pointer analysis based on the principle of matching pointer dereferences or "memory fetches" with pointer assignments. Our abstraction, the Assign-Fetch Graph (AFG), has several advantages over traditional points-to graphs. It leads to concise procedure summaries for pointer analysis and similar computations; each procedure' s summary information can be used effectively in arbitrary calling contexts. Different analysis variations can be formulated on the AFG we present two possible variations. One is based on a standard insensitive view of the program; the other takes some statement ordering into account and produces results that are both more precise and more quickly computed. Our abstraction also facilitates incremental summary-based pointer analysis, which is not generally possible with conventional points-to graphs.
Subject(s):
Computer science
Item views
172
Metadata:
text | xml
Suggested Citation:
Marcio Buss, Daniel Brand, Vugranam Sreedhar, Stephen A. Edwards, , A new abstraction for summary-based pointer analysis, Columbia University Academic Commons, .

Columbia University Libraries | Policies | FAQ