1. Dataflow Analysis for Software Product Lines May, 2013 SSS Dataflow Analysis for Software Product Lines DFA-4-SPL Claus Brabrand IT University of Copenhagen Universidade Federal de Pernambuco [ brabrand@itu.dk ] Márcio Ribeiro Universidade Federal de Alagoas Universidade Federal de Pernambuco [ mmr3@cin.ufpe.br ] Paulo Borba Universidade Federal de Pernambuco [ phmb@cin.ufpe.br ] Társis Toledo Universidade Federal de Pernambuco [ twt@cin.ufpe.br ] AOSD 2012 and TAOSD 2013

2. Dataflow Analysis for Software Product Lines May, 2013 SSS Statically Analyzing Software Product Lines in Minutes instead of Years SPL LIFT Eric Bodden Technische Universität Darmstadt [ bodden@acm.org ] Társis Tolêdo Universidade Federal de Pernambuco [ mmr3@cin.ufpe.br ] Márcio Ribeiro Universidade Federal de Alagoas [ twt@cin.ufpe.br ] Mira Mezini Technische Universität Darmstadt [ phmb@cin.ufpe.br ] Claus Brabrand IT University of Copenhagen [ brabrand@itu.dk ] Paulo Borba Universidade Federal de Pernambuco [ phmb@cin.ufpe.br ] PLDI 2013

3. [ 3 ] Dataflow Analysis for Software Product Lines May, 2013 SSS Abstract Software product lines (SPLs) developed using annotative approaches such as conditional compilation come with an inherent risk of constructing erroneous products. For this reason, it is essential to be able to analyze such SPLs. However, as dataflow analysis techniques are not able to deal with SPLs, developers must generate and analyze all valid products individually, which is expensive for non-trivial SPLs. In this paper, we demonstrate how to take any standard intraprocedural dataflow analysis and automatically turn it into a feature-sensitive dataflow analysis in five different ways where the last is a combination of the other four. All analyses are capable of analyzing all valid products of an SPL without having to generate all of them explicitly. We have implemented all analyses using SOOT’s intraprocedural dataflow analysis framework and experimentally evaluated four of them according to their performance and memory characteristics on five qualitatively different SPLs. On our benchmarks, the combined analysis strategy is up to almost eight times faster than the brute-force approach.

4. [ 4 ] Dataflow Analysis for Software Product Lines May, 2013 SSS Introduction: Software Product Lines (SPL) Dataflow Analysis (DFA) DFA-4-SPL: A0 (brute force): (feature in-sensitive) A1 (consecutive): (feature sensitive) A2 (simultaneous): (feature sensitive) A3 (shared simultaneous): (feature sensitive) SPL LIFT (graph encoding): Evaluation and Results: A1 vs SPL LIFT

5. [ 5 ] Dataflow Analysis for Software Product Lines May, 2013 SSS Introduction Traditional Software Development: One program = One product Product Line: A ”family” of products (of N ”similar” products): customize SPL: (Family of Programs)

6. [ 6 ] Dataflow Analysis for Software Product Lines May, 2013 SSS Software Product Line SPLs based on Conditional Compilation: #ifdef (  )... #endif Logo logo;... logo.use(); #ifdef (VIDEO) logo = new Logo(); #endif Example (SPL fragment)  : f  F |  | 

7. [ 7 ] Dataflow Analysis for Software Product Lines May, 2013 SSS Software Product Line SPL: Feature Model: (e.g.: ψ FM ≡ VIDEO  COLOR ) Family of Programs: COLOR VIDEO COLOR  VIDEO VIDEO Ø { Video } { Color, Video } Configurations: Ø, {Color}, {Video}, {Color,Video} VALID { Color } customize 2F2F Set of Features: F = { COLOR, VIDEO } 2F2F

8. [ 8 ] Dataflow Analysis for Software Product Lines May, 2013 SSS Software Product Line SPLs based on Conditional Compilation: #ifdef (  )... #endif Logo logo;... logo.use(); #ifdef (VIDEO) logo = new Logo(); #endif Example (SPL fragment) *** null-pointer exception! in configurations: {Ø, {COLOR}}  : f  F |  | 

9. [ 9 ] Dataflow Analysis for Software Product Lines May, 2013 SSS result 0100101 1110110 1010011 1110111 0100101 1110110 1010011 1110111 Analysis of SPLs The Compilation Process:...and for Software Product Lines: 0100101 1110110 1010011 1110111 result compile run ERROR! generate 0100101 1110110 1010011 1110111 result run ERROR! ANALYZE! Feature-sensitive data-flow analysis ! run compile ANALYZE! ERROR! 2F2F

10. [ 10 ] Dataflow Analysis for Software Product Lines May, 2013 SSS Dataflow Analysis Dataflow Analysis: 1) Control-flow graph 2) Lattice (finite height) 3) Transfer functions (monotone) L Example: "sign-of-x analysis"

11. [ 11 ] Dataflow Analysis for Software Product Lines May, 2013 SSS Analyzing a Program 1) Program 2) Build CFG 3) Make Equations 4) Solve equations: fixed-point computation (iteration) 5) SOLUTION (least fixed point): Annotated with program points

12. [ 12 ] Dataflow Analysis for Software Product Lines May, 2013 SSS Introduction: Software Product Lines (SPL) Dataflow Analysis (DFA) DFA-4-SPL: A0 (brute force): (feature in-sensitive) A1 (consecutive): (feature sensitive) A2 (simultaneous): (feature sensitive) A3 (shared simultaneous): (feature sensitive) SPL LIFT (graph encoding): Evaluation and Results: A1 vs SPL LIFT

13. [ 13 ] Dataflow Analysis for Software Product Lines May, 2013 SSS A0 A0 (brute force): void m() { int x=0; ifdef(A) x++; ifdef(B) x--; } c = {A}:c = {B}:c = {A,B}: int x = 0; x++; x--;int x = 0; x++; x--;int x = 0; x++; x--; 0 _ | + 0 _ | - 0 _ | 0/+ + ψ FM = A ∨ B L feature in-sensitive! N = O (2 F ) compilations!

14. [ 14 ] Dataflow Analysis for Software Product Lines May, 2013 SSS int x = 0; x++; x--; A: B: int x = 0; x++; x--; A: B: int x = 0; x++; x--; A: B: A1 A1 (consecutive): void m() { int x=0; ifdef(A) x++; ifdef(B) x--; } c = {A}: 0 _ | + ✗ ✓ ✓ ψ FM = A ∨ B L c = {B}:c = {A,B}: 0 _ | - 0 _ | 0/+ + ✗ ✓✓ ✓ ✓ ✓ + 0 feature sensitive!

15. [ 15 ] Dataflow Analysis for Software Product Lines May, 2013 SSS x++; + ({A} =, {B} =, {A,B} = ) A2 A2 (simultaneous): void m() { int x=0; ifdef(A) x++; ifdef(B) x--; } ∀ c ∈ {{A},{B},{A,B}}: int x = 0;x--; 0 _ | 0 _ | - 0 _ | 0/+ + A: B: ✓ ({A} =, {B} =, {A,B} = ) ✓✓ ✓✓ ✓✓ ✗ ✗ ψ FM = A ∨ B L 0 + feature sensitive!

16. [ 16 ] Dataflow Analysis for Software Product Lines May, 2013 SSS x--; + x++; 0 ( [[ψ ∧ ¬A ]] =, [[ψ ∧ A ]] =, [[ψ ∧ ¬A ]] =, [[ψ ∧ A ]] = ) ( [[ψ ]] =, [[ψ ]] = ) A3 A3 (shared): void m() { int x=0; ifdef(A) x++; ifdef(B) x--; } ψ FM = A ∨ B: int x = 0; A: B: _ | ( [[ψ]] = ) 0 (A ∨ B) ∧ ¬A ∧ ¬B ≡ false can use BDD representation ! (compact+efficient) - 0/+ i.e., invalid given wrt. the feature model, ψ ! ψ FM = A ∨ B L 0 ∧ ¬A ∧A∧A + ∧ ¬B ∧B∧B ∧B∧B feature sensitive!

17. [ 17 ] Dataflow Analysis for Software Product Lines May, 2013 SSS true [ (A ∧ B) ∧ ¬A] ∨ [ true ∧ A] = A true A∧BA∧B true ∧ ¬A = ¬A true SPL LIFT IFDS:A0: λS. (S – {x}) ∪ {y} {x} {y} SPL LIFT (IFDS ➞ IDE):A2: ( {A} = {x}, {B} = {x}, {A,B} = {x,y} ) 0 x y 0 x y 0 x y 0 x y λS. (S – {x}) ∪ {y} A:A: #ifdef (A) ( {A} = {y}, {B} = {x}, {A,B} = {y} ) A ¬A LIFT: ■Reps ■Horwitz ■Sagiv fixed-point iteration graph reachability ➔ ¬AA

18. [ 18 ] Dataflow Analysis for Software Product Lines May, 2013 SSS Introduction: Software Product Lines (SPL) Dataflow Analysis (DFA) DFA-4-SPL: A0 (brute force): (feature in-sensitive) A1 (consecutive): (feature sensitive) A2 (simultaneous): (feature sensitive) A3 (shared simultaneous): (feature sensitive) SPL LIFT (graph encoding): Evaluation and Results: A1 vs SPL LIFT

19. [ 19 ] Dataflow Analysis for Software Product Lines May, 2013 SSS Results Results for SPL LIFT (interprocedural analysis): In fact, analyzing all valid configs is only slightly slower than analyzing one config ! "Statically Analyzing Software Product Lines... in Minutes instead of Years" :-) Reaching DefinitionsPossible TypesUninitialized Variables SPL benchmark # valid configs A1SPL LIFT A1SPL LIFT A1SPL LIFT Lampiro43m30s42s13s4s3m09s1m25s MM 082624m29s59s2m06s3s27m39s2m13s GPL1,872days8m48s9h03m39s42sdays7m09s Berkeley DB unknown years12m04syears24syears10m18s

20. Dataflow Analysis for Software Product Lines May, 2013 SSS *) Thanks

21. Dataflow Analysis for Software Product Lines May, 2013 SSS BONUS SLIDES

22. [ 22 ] Dataflow Analysis for Software Product Lines May, 2013 SSS Overview A0 (brute force) A1 (consecutive) A2 (simultaneous) A3 (shared) A* (combo) IFDS ➞ IDE (lift) FASTER (intra-procedural) PLDI 2013 IFDS (graph repr) A3+BDD (esp. inter- procedural) no re-compile! caching! sharing! combo! graph encoding! repr! AOSD 2012 TAOSD 2013

23. [ 23 ] Dataflow Analysis for Software Product Lines May, 2013 SSS Errors Logo logo; logo.use(); #ifdef (VIDEO) logo = new Logo(); #endif *** null-pointer exception! in configurations: {Ø, {COLOR}} Logo logo; print(logo); #ifdef (VIDEO) logo = new Logo(); #endif *** uninitialized variable! in configurations: {Ø, {COLOR}} Logo logo;... #ifdef (VIDEO) logo = new Logo(); #endif *** unused variable! in configurations: {Ø, {COLOR}}

24. [ 24 ] Dataflow Analysis for Software Product Lines May, 2013 SSS Related Work (DFA) Path-sensitive DFA: Idea of “conditionally executed statements” Compute different analysis info along different paths (~ A1, A2, A3) to improve precision or to optimize “hot paths” Predicated DFA: Guard lattice values by propositional logic predicates (~ A3), yielding “optimistic dataflow values” that are kept distinct during analysis (~ A2 and A3) “Constant Propagation with Conditional Branches” ( Wegman and Zadeck ) TOPLAS 1991 “Predicated Array Data-Flow Analysis for Run-time Parallelization” ( Moon, Hall, and Murphy ) ICS 1998 Our work: Automatically lift any DFA to SPLs (with ψ FM ) ⇒ feature-sensitive analysis for analyzing entire program family

25. [ 25 ] Dataflow Analysis for Software Product Lines May, 2013 SSS Related Work (Lifting for SPLs) Model Checking: Type Checking: Parsing: Testing: Model Checking Lots of Systems: Efficient Verification of Temporal Properties in Software Product Lines” ( Classen, Heymans, Schobbens, Legay, and Raskin ) ICSE 2010 Model checks all SPLs at the same time (3.5x faster) than one by one! (similar goal, diff techniques) Type checking ↔ DFA (similar goal, diff techniques) Our: auto lift any DFA (uninit vars, null pointers,...) “Type Safety for Feature-Oriented Product Lines” ( Apel, Kastner, Grösslinger, and Lengauer ) ASE 2010 “Type-Checking Software Product Lines - A Formal Approach” ( Kastner and Apel ) ASE 2008 “Variability-Aware Parsing in the Presence of Lexical Macros & C.C.” ( Kastner, Giarrusso, Rendel, Erdweg, Ostermann, and Berger ) OOPSLA 2011 “Reducing Combinatorics in Testing Product Lines” ( Hwan, Kim, Batory, and Khurshid ) AOSD 2011 Select relevant feature combinations for a given test case Uses (hardwired) DFA (w/o FM) to compute reachability (similar techniques, diff goal): Split and merging parsing (~A3) and also uses instrumentation

26. [ 26 ] Dataflow Analysis for Software Product Lines May, 2013 SSS Emerging Interfaces

27. [ 27 ] Dataflow Analysis for Software Product Lines May, 2013 SSS Emerging Interfaces "A Tool for Improving Maintainability of Preprocessor-based Product Lines" ( Márcio Ribeiro, Társis Tolêdo, Paulo Borba, Claus Brabrand ) *** Best Tool Award *** CBSoft 2011:

28. [ 28 ] Dataflow Analysis for Software Product Lines May, 2013 SSS Summary A0 A1 A2 A3 void m() { int x=0; ifdef(A) x++; ifdef(B) x--; } Analyzing program: ψ FM = A ∨ B

29. [ 29 ] Dataflow Analysis for Software Product Lines May, 2013 SSS Specification: A0, A1, A2, A3 A0 A1 A2 A3

30. [ 30 ] Dataflow Analysis for Software Product Lines May, 2013 SSS Evaluation Five (qualitatively different) SPL benchmarks:

31. [ 31 ] Dataflow Analysis for Software Product Lines May, 2013 SSS Total Time (incl. compile) In practice: 4x (Reaching Definitions) 7x 3x 1x ( no re-compile !) Feature sensitive (A1, A2, and A3) all faster than A0

32. [ 32 ] Dataflow Analysis for Software Product Lines May, 2013 SSS Analysis Time (excl. compile) In practice: ( caching !) (Reaching Definitions) A2 faster than A1 A3 faster than A2 ( sharing !)

33. [ 33 ] Dataflow Analysis for Software Product Lines May, 2013 SSS Beyond the Sum of all Methods For a method with x # valid configurations, which of analyses A1 vs A2 vs A3 is fastest? Statistically significant differences between A1, A2, and A3 for all N, except between A2 and A3 for N=4 (underlined above).

34. [ 34 ] Dataflow Analysis for Software Product Lines May, 2013 SSS Combo Analysis Strategy: A * Intraprocedurally combined analysis strategy, A*: A* consistently fastest ( combo !)

35. [ 35 ] Dataflow Analysis for Software Product Lines May, 2013 SSS Analysis Time (excl. compile) In theory: In practice: T IME ( A 3 ) : Depends on degree of sharing in SPL ! ( caching !) (Reaching Definitions) A2 faster than A1 A3 faster than A2 ( sharing !)

36. [ 36 ] Dataflow Analysis for Software Product Lines May, 2013 SSS Memory Usage In theory: In practice: (Reaching Definitions) S PACE ( A 3 ) : Depends on degree of sharing in SPL !

37. [ 37 ] Dataflow Analysis for Software Product Lines May, 2013 SSS Analysis Time (excl. compile) In practice: (Reaching Definitions) Nx1 ≠ 1xN ?! Caching! A2 faster than A1

38. [ 38 ] Dataflow Analysis for Software Product Lines May, 2013 SSS Caching (A1 vs A2) Cache misses (A1 vs A2): Cache enabled: This is the "normal condition" (for reference) Cache disabled*: As hypothesized, this indeed affects A1 more than A2 i.e., A2 has better cache properties than A1 *) we flush the L2 cache, by traversing an 8MB “bogus array” to invalidate cache!

39. [ 39 ] Dataflow Analysis for Software Product Lines May, 2013 SSS IFDEF normalization Refactor "undisciplined" (lexical) ifdefs into "disciplined" (syntactic) ifdefs: Normalize "ifdef"s (by transformation):

40. [ 40 ] Dataflow Analysis for Software Product Lines May, 2013 SSS Lexical #ifdef  Syntactic ifdef Simple transformation: We do not handle non-syntactic '#ifdef's: Fair assumption (also in CIDE) Nested ifdef's also give rise to a conj. of formulas

41. [ 41 ] Dataflow Analysis for Software Product Lines May, 2013 SSS BDD (Binary Decision Diagram) Compact and efficient representation for boolean functions (aka., set of set of names) FAST: negation, conjunction, disjunction, equality !  =  F(A,B,C) =F(A,B,C) = A  (B  C)   A C minimized BDD B    A BB CCCC BDD

42. [ 42 ] Dataflow Analysis for Software Product Lines May, 2013 SSS Formula ~ Set of Configurations Definitions (given F, set of feature names): f  F feature name c  2 F configuration ( set of feature names ) c  F X  2 2 set of config's ( set of set of feature names ) X  2 F Example ifdef s: F [[ B  A ]] [[ A  (B  C) ]] F = {A,B} F = {A,B,C} = { {A}, {B}, {A,B} } = { {A,B}, {A,C}, {A,B,C} }

43. [ 43 ] Dataflow Analysis for Software Product Lines May, 2013 SSS Feature Model (Example) Feature Model: Feature set: Formula: Set of configurations:  FM  Car  Engine  (1.0  1.4)  Air  1.4 { { Car, Engine, 1.0 }, { Car, Engine, 1.4 }, { Car, Engine, 1.4, Air } } F = {Car, Engine, 1.0, 1.4, Air} Note: | [[  FM ]] | = 3 < 32 = |2 F | [[ ]] = Engine 1.0 Air 1.4

44. [ 44 ] Dataflow Analysis for Software Product Lines May, 2013 SSS Conditional Compilation The ' ifdef ' construction: Syntactic variant of lexical #ifdef Propositional Logic: where f  F (finite set of feature names) Example: STM : 'ifdef' '('  ')' STM  : f  F |  |  status.print("you die"); ifdef (DeluxeVersion && ColorDisplay) { player.redraw(Color.red); Audio.play("crash.wav"); } lives = lives - 1; A ifdef (A) {... }

45. [ 45 ] Dataflow Analysis for Software Product Lines May, 2013 SSS CASE 1: "COPY" A3: Lazy Splitting (using BDDs) CASE 2: "APPLY"CASE 3: "SPLIT"  : S [  = l,... ] l ' = f S ( l )      : S [  = l,... ] [  = l ',... ] l ' = f S ( l )  : S [  = l,... ] [  = l,  = l',...] l ' = f S ( l )   = Ø  =  Ø    

46. [ 46 ] Dataflow Analysis for Software Product Lines May, 2013 SSS A0, A1, A2, and A3 A0 A1 A2 A3