esp path path sensitive sensitive esp program
play

ESP - - Path Path- -Sensitive Sensitive ESP Program - PowerPoint PPT Presentation

ESP - - Path Path- -Sensitive Sensitive ESP Program Verification in Program Verification in Polynomial Time Polynomial Time M. Das, S. Lerner, M. Seigle M. Das, S. Lerner, M. Seigle PLDI '02 PLDI '02 Partial program verification


  1. ESP - - Path Path- -Sensitive Sensitive ESP Program Verification in Program Verification in Polynomial Time Polynomial Time M. Das, S. Lerner, M. Seigle M. Das, S. Lerner, M. Seigle PLDI '02 PLDI '02

  2. Partial program verification Partial program verification � Verify that a program obeys a temporal safety Verify that a program obeys a temporal safety � property property � e.g. correct file opening/closing behavior e.g. correct file opening/closing behavior � � Property representable as DFA (FSM) Property representable as DFA (FSM) � Print, $uninit Close Open, Print, $error Open Close Close Open Opened Print

  3. Why it’ ’s hard: s hard: Why it � In a program, FSM may transition differently In a program, FSM may transition differently � along different execution paths along different execution paths � Path Path- -insensitive dataflow analysis will merge and insensitive dataflow analysis will merge and � lose relevant information lose relevant information � The program may satisfy the property, but we The program may satisfy the property, but we � won't be able to determine this. won't be able to determine this.

  4. Example Example void main(){ void main(){ if (dump) if (dump) f = fopen(dumpFil, "w"); f = fopen(dumpFil, "w"); if (p) if (p) x = 0; x = 0; else else x = 1; x = 1; if (dump) if (dump) fclose(f); fclose(f); } }

  5. Path- -insensitive dataflow analysis insensitive dataflow analysis Path void main(){ void main(){ [ $uninit ] if (dump) if (dump) f = fopen(dumpFil, "w"); f = fopen(dumpFil, "w"); [ $uninit, Opened ] if (p) if (p) x = 0; x = 0; else else x = 1; x = 1; [ $uninit, Opened ] if (dump) if (dump) fclose(f); fclose(f); [ $uninit, $error] } }

  6. Path- -sensitive analysis sensitive analysis Path void main(){ void main(){ [ $uninit ] if (dump) if (dump) f = fopen(dumpFil, "w"); f = fopen(dumpFil, "w"); [ $uninit , ¬ d] if (p) if (p) [ Opened , d] x = 0; x = 0; [ $uninit , ¬ d, ¬ p, x =1] else else [ $uninit , ¬ d, p, x = 0] x = 1; x = 1; [ Opened , d, ¬ p, x =1] if (dump) if (dump) [ Opened , d, p, x =0] fclose(f); fclose(f); Only one of the two } paths possible from } each state

  7. Moral of the story: Moral of the story: � Path Path- -insensitive dataflow analysis is too insensitive dataflow analysis is too � imprecise imprecise � But path But path- -sensitive analysis is overkill and too sensitive analysis is overkill and too � expensive. expensive. � The obvious solution: keep as much information The obvious solution: keep as much information � as needed, no more, no less as needed, no more, no less � the paper presents a heuristic for this the paper presents a heuristic for this �

  8. Main contributions of this paper Main contributions of this paper � An analysis framework that is An analysis framework that is only as path only as path- - � sensitive as needed to verify a property to verify a property sensitive as needed � Including an inter Including an inter- -procedural version procedural version � � Insights into developing a verification system Insights into developing a verification system � using property simulation that will scale to large using property simulation that will scale to large programs (such as gcc ) programs (such as gcc ) � This is ESP This is ESP - - Error detection via Scalable Program Error detection via Scalable Program � analysis analysis

  9. Property analysis Property analysis � An analysis framework that parametrizes how path An analysis framework that parametrizes how path- - � sensitive we choose to be. sensitive we choose to be. � Includes path Includes path- -insensitive and fully path insensitive and fully path- -sensitive sensitive � analyses as extremes. analyses as extremes. � Essentially a normal dataflow analysis, with interesting Essentially a normal dataflow analysis, with interesting � things happening at the merge points. things happening at the merge points. � path path- -insensitive insensitive - - merge everything merge everything � � path path- -sensitive sensitive - - no merges no merges � � property simulation property simulation - - merge only info "irrelevant" for the merge only info "irrelevant" for the � property being verified property being verified

  10. A few details A few details � State carried in analysis is State carried in analysis is symbolic state symbolic state � � Two components: Two components: � � abstract state abstract state ⊆ ⊆ D, where D = set of states in the D, where D = set of states in the � property FSM property FSM � execution state (as normal) execution state (as normal) � � S = domain of all symbolic states S = domain of all symbolic states � � Analysis computes dataflow facts from the Analysis computes dataflow facts from the � S domain 2 S domain 2

  11. A few details (2) A few details (2) � Key is filtering function used at merge points: Key is filtering function used at merge points: � S → � α α : 2 S : 2 S → 2 2 S � � α α cs (ss) = ss cs (ss) = ss � � gives path gives path- -sensitive analysis sensitive analysis � � α α df (ss) = { ∪ as(s), t es(s)]} df (ss) = { ss as(s), ss es(s)]} ∪ s t s � s ∈ ∈ ss s ∈ ∈ ss � gives path gives path- -insensitive dataflow analysis insensitive dataflow analysis �

  12. A few details (3) A few details (3) � Property simulation merges all those symbolic Property simulation merges all those symbolic � states that have the same property state states that have the same property state � α α as ≠ ∅ D & ss[d] ≠ = {[{d}, t es (s)] | d ∈ ∈ D & ss[d] } as = {[{d}, ss[d] es (s)] | d ∅ } t s � s ∈ ∈ ss[d] � Notation: Notation: � � ss[d] = { s | s ss[d] = { s | s ∈ ∈ ss & d ss & d ∈ ∈ as(s) } as(s) } � � “ “set of all s in ss containing d set of all s in ss containing d” ” � � Example Example � � Will see limitations of this heuristic soon Will see limitations of this heuristic soon �

  13. Path- -sensitive analysis sensitive analysis Path void main(){ void main(){ [ $uninit ] if (dump) if (dump) f = fopen(dumpFil, "w"); f = fopen(dumpFil, "w"); [ $uninit , ¬ d] if (p) if (p) [ Opened , d] x = 0; x = 0; else else [ $uninit , ¬ d, ¬ p, x =1] x = 1; x = 1; [ $uninit , ¬ d, p, x = 0] if (dump) if (dump) [ Opened , d, ¬ p, x =1] fclose(f); fclose(f); [ Opened , d, p, x =0] } }

  14. Property simulation Property simulation void main(){ void main(){ [ $uninit ] if (dump) if (dump) f = fopen(dumpFil, "w"); f = fopen(dumpFil, "w"); [ $uninit , ¬ d] if (p) if (p) [Opened, d] x = 0; x = 0; No changes to property state else else x = 1; x = 1; [ $uninit , ¬ d] [ Opened , d] if (dump) if (dump) fclose(f); fclose(f); Only one of the two } paths possible from } each state

  15. A few details (4) A few details (4) � Not all branches are possible from a particular symbolic Not all branches are possible from a particular symbolic � state state � Analysis exploits this by using a theorem prover to attempt to Analysis exploits this by using a theorem prover to attempt to � determine whether path is feasible from a given symbolic determine whether path is feasible from a given symbolic state state � Complexity O(H |E||D| (T + J + Q)) where Complexity O(H |E||D| (T + J + Q)) where � � H is the lattice height H is the lattice height � � E is the number of edges in CFG E is the number of edges in CFG � � D is the number of property states D is the number of property states � � T is the cost of one call to the flow function (includes T is the cost of one call to the flow function (includes � deciding branch feasibility), J is join, Q is deciding equality on on deciding branch feasibility), J is join, Q is deciding equality execution states. execution states.

  16. Property Analysis Property Analysis � Instantiation to constant propagation with Instantiation to constant propagation with � 2 |E||D|) property simulation – – O(V O(V 2 |E||D|) property simulation � V = number of variables V = number of variables � � Can obtain an inter Can obtain an inter- -procedural analysis using the procedural analysis using the � framework by Reps, Horwitz and Sagiv framework by Reps, Horwitz and Sagiv � the algorithm is context the algorithm is context- -sensitive for property states sensitive for property states � only (insensitive for execution states). only (insensitive for execution states).

  17. But property simulation is no But property simulation is no magic bullet magic bullet if (dump) if (dump) flag = 1; flag = 1; else else flag = 0; flag = 0; if (dump) if (dump) f = fopen(...); f = fopen(...); if (flag) if (flag) fclose(f); fclose(f);

  18. We lose information We lose information if (dump) if (dump) Property state stays same flag = 1; flag = 1; here, so analysis won’t save correlation between flag else else and dump flag = 0; flag = 0; if (dump) if (dump) Property states will be f = fopen(...); f = fopen(...); $uninit and Opened if (flag) if (flag) Potential error fclose(f); fclose(f); here!

Download Presentation
Download Policy: The content available on the website is offered to you 'AS IS' for your personal information and use only. It cannot be commercialized, licensed, or distributed on other websites without prior consent from the author. To download a presentation, simply click this link. If you encounter any difficulties during the download process, it's possible that the publisher has removed the file from their server.

Recommend


More recommend