Pointer analysis

From Wikipedia, the free encyclopedia
Jump to navigation Jump to search

In computer science, pointer analysis, or points-to analysis, is a static code analysis technique that establishes which pointers, or heap references, can point to which variables, or storage locations. It is often a component of more complex analyses such as escape analysis. A closely related technique is shape analysis.

(This is the most common colloquial use of the term. A secondary use has pointer analysis be the collective name for both points-to analysis, defined as above, and alias analysis. Points-to and alias analysis are closely related but not always equivalent problems.)

Example[edit]

For the following example program, a points-to analysis would compute that the points-to set of p is {x, y}.

int x;
int y;
int* p = unknown() ? &x : &y;

Introduction[edit]

As a form of static analysis, fully precise pointer analysis can be shown to be undecidable.[1] Most approaches are sound, but range widely in performance and precision. For large programs, some tradeoffs may be necessary to make the analysis finish in reasonable time and space. Two examples of these tradeoffs are:[2]

  • Treating all references from a structured object as being from the object as a whole. This is known as field insensitivity or structure insensitivity.
  • Ignoring flow of control when analysing which objects are assigned to pointers. This is known as context-insensitive pointer analysis (when ignoring the context in which function calls are made) or flow-insensitive pointer analysis (when ignoring the control flow within a procedure).

The disadvantage of these simplifications is that the calculated set of objects pointed to may become less precise.

Context-Insensitive, Flow-Insensitive Algorithms[edit]

Pointer analysis algorithms are used to convert collected raw pointer usages (assignments of one pointer to another or assigning a pointer to point to another one) to a useful graph of what each pointer can point to.[3]

Steensgaard's algorithm and Andersen's algorithm are common context-insensitive, flow-insensitive algorithms for pointer analysis. They are often used in compilers, and have implementations in the LLVM codebase.

Flow-Insensitive Approaches[edit]

Many approaches to flow-insensitive pointer analysis can be understood as forms of abstract interpretation, where heap allocations are abstracted by their allocation site (i.e., a program location).[4]

Many flow-insensitive algorithms are specified in Datalog, including those in the Soot analysis framework for Java.[5]

Context-sensitive, flow-insensitive algorithms achieve higher precision, generally at the cost of some performance, by analyzing each procedure several times, once per context.[6] Most analyses use a "context-string" approach, where contexts consist of a list of entries (common choices of context entry include call sites, allocation sites, and types).[7] To ensure termination (and more generally, scalability), such analyses generally use a k-limiting approach, where the context has a fixed maximum size, and the least recently added elements are removed as needed.[8] Three common variants of context-sensitive, flow-insensitive analysis are:[9]

  • Call-site sensitivity
  • Object sensitivity
  • Type sensitivity

Call-site sensitivity[edit]

In call-site sensitivity, the points-to set of each variable (the set of abstract heap allocations each variable could point to) is further qualified by a context consisting of a list of callsites in the program. These contexts abstract the control-flow of the program.

The following program demonstrates how call-site sensitivity can achieve higher precision than a flow-insensitive, context-insensitive analysis.

int *id(int* x) {
  return x;
}
int main() {
  int y;
  int z;
  int *y2 = id(&y); // call-site 1
  int *z2 = id(&z); // call-site 2
  return 0;
}

For this program, a context-insensitive analysis would (soundly but imprecisely) conclude that x can point to either the allocation holding y or that of z, so y2 and z2 may alias, and both could point to either allocation. A callsite-sensitive analysis would analyze id twice, once for call-site 1 and once for call-site 2, and the points-to facts for x would be qualified by the call-site, enabling the analysis to deduce that when main returns, y2 can only point to the allocation holding y and z2 can only point to the allocation holding z.

Object sensitivity[edit]

In an object sensitive analysis, the points-to set of each variable is qualified by the abstract heap allocation of the receiver object of the method call. Unlike call-site sensitivity, object-sensitivity is non-syntactic or non-local: the context entries are derived during the points-to analysis itself.[10]

Type sensitivity[edit]

Type sensitivity is a variant of object sensitivity where the allocation site of the receiver object is replaced by the class/type containing the method containing the allocation site of the receiver object.[11] This results in strictly fewer contexts than would be used in an object-sensitive analysis, which generally means better performance.

References[edit]

  1. ^ Reps, Thomas (2000-01-01). "Undecidability of context-sensitive data-dependence analysis". ACM Transactions on Programming Languages and Systems. 22 (1): 162–186. doi:10.1145/345099.345137. ISSN 0164-0925. S2CID 2956433.
  2. ^ Barbara G. Ryder (2003). "Dimensions of Precision in Reference Analysis of Object-Oriented Programming Languages". Compiler Construction, 12th International Conference, CC 2003 Held as Part of the Joint European Conferences on Theory and Practice of Software, ETAPS 2003 Warsaw, Poland, April 7–11, 2003 Proceedings. pp. 126–137. doi:10.1007/3-540-36579-6_10.
  3. ^ Zyrianov, Vlas; Newman, Christian D.; Guarnera, Drew T.; Collard, Michael L.; Maletic, Jonathan I. (2019). "srcPtr: A Framework for Implementing Static Pointer Analysis Approaches" (PDF). ICPC '19: Proceedings of the 27th IEEE International Conference on Program Comprehension. Montreal, Canada: IEEE.
  4. ^ Smaragdakis, Yannis; Bravenboer, Martin; Lhoták, Ondrej (2011-01-26). "Pick your contexts well: understanding object-sensitivity". Proceedings of the 38th Annual ACM SIGPLAN-SIGACT Symposium on Principles of Programming Languages. POPL '11. Austin, Texas, USA: Association for Computing Machinery: 17–30. doi:10.1145/1926385.1926390. ISBN 978-1-4503-0490-0. S2CID 6451826.
  5. ^ Antoniadis, Tony; Triantafyllou, Konstantinos; Smaragdakis, Yannis (2017-06-18). "Porting doop to Soufflé: a tale of inter-engine portability for Datalog-based analyses". Proceedings of the 6th ACM SIGPLAN International Workshop on State of the Art in Program Analysis. SOAP 2017. Barcelona, Spain: Association for Computing Machinery: 25–30. doi:10.1145/3088515.3088522. ISBN 978-1-4503-5072-3. S2CID 3074689.
  6. ^ (Smaragdakis & Balatsouras, p. 29)
  7. ^ Thiessen, Rei; Lhoták, Ondřej (2017-06-14). "Context transformations for pointer analysis". ACM SIGPLAN Notices. 52 (6): 263–277. doi:10.1145/3140587.3062359. ISSN 0362-1340.
  8. ^ (Li et al., pp. 1:4)
  9. ^ (Smaragdakis & Balatsouras)
  10. ^ (Smaragdakis & Balatsouras, p. 37)
  11. ^ (Smaragdakis & Balatsouras, p. 39)

Bibliography[edit]