Presentation is loading. Please wait.

Presentation is loading. Please wait.

Formal Verification of fFSM Model Sachoun Park, Gihwon Kwon Department of Computer Science Kyonggi University, Korea IWFST, Shanghai, China, 2005.11.08.

Similar presentations


Presentation on theme: "Formal Verification of fFSM Model Sachoun Park, Gihwon Kwon Department of Computer Science Kyonggi University, Korea IWFST, Shanghai, China, 2005.11.08."— Presentation transcript:

1 Formal Verification of fFSM Model Sachoun Park, Gihwon Kwon Department of Computer Science Kyonggi University, Korea IWFST, Shanghai, China, 2005.11.08.

2 2 Formal Verification Minimization (e.g. bisimulation) Property specification State space: - Kripke structure - LTS (Lab.Trans.Syst.) - Marking graphs Modeling languages: - Communicating FSMs - Process algebra - Petri nets Generation init(state):= 0; next(state):= case state=0: {2,3}; state=1: 1; 1 : state; esac; Verification: - Model checking - Equivalence checking - ?? Logical results with counter-example

3 3 fFSM Model Hierarchy Concurrency Internal Event Variable State A2B C D A1 E F f1 f2 B1 B2 B3 snd rcv D1 rcv snd fFSM Dataflow wormhole (super-block with different domain inside) System-level Specification in PeaCE Event generators Display Backplane: Discrete Event

4 4 Flattened Model pFSM = (I, IT, O, M, , V), I = {coin, ready, stop, time}, O = {tilt, game_over, game_on, …} V = {randn, remain} M = {m 1, m 2, m 3, m 4 }  (m 1 ) = {m 3, m 4 },  (m 2 ) = ,  (m 3 ) = ,  (m 4 ) =   m 1 = (S 1, s 0 1, L 1, T 1 )  S 1 = {GameOff, GameOn},  L 1 = {(GameOff,  ), (GameOn,  )}  T 1 = {(GameOff, coin==1,  game_on=1, timeset=1000 , GameOn), …} m1m1 m2m2 m3m3 m4m4

5 5 fFSM Semantics {GameOff, TimeInit, Ready, Rand}, {coin},  {GameOn, TimeInit, Ready, Rand}, {timeset},  {GameOn, Wait, Ready, Rand}, ,  {game_on:=1}{remain:=1000} {GameOff, TimeInit, Ready, Rand}, {coin},  {GameOn, Wait, Ready, Rand}, ,  {game_on:=1, remain:=1000} Micro Step: Macro Step: Macro step Environmental change Micro steps Environmental change stable

6 6 Desired Properties No unused components  EF component 1  …  EF component n  Components: states, events No unreachable guards  EF (s i  EX s j ), guard(t) = g, source(t) = s i, target(t) = s j No unambiguous transitions (ambiguous transition)  AG  ((t 1  t 2 )  (t 2  t 3 )  (t 1  t 3 )), where {t 1, t 2, t 3 } is a set of outgoing transitions from the same state. No deadlocks   EFAG deadlock, where deadlock is  (t 1  …  t n ) No divergent behavior (circular transition)  AG (  stable  A[  stable U stable]) Race condition violation  AG(  stable  uadate  AX A[  uadate U stable])

7 7 Stepper: Simulation & Verification Tool F2V Model Checker SMV Translator SMV Model checke r Graph Generator *.xml fFSM simulator *.smv Model Result CTL Property Formal fFSM Verifier (F2V) Explicit model checker Properties to be checked Reachability Race condition Ambiguous transitions Circular transitions

8 8 Discussion Topics How to avoid the state explosion problem ?  State explosion problem occurs due to variables  How to abstract them away ? How to integrated heterogeneous models ?  fFSM models the control flow of a system  SDF models the data flow of a system  How to integate them ? 1. data transfer 3. result(flag) transfer Display Event source “flag ” Backplane “ a” fFSM XY a/out flag B “out” A SDF C 2. execution A SPDF C B Event source Backplane domain node_name: foo script: “stop foo ”script:“run foo ” XY a FSM reset Asynchronous Interaction Synchronous Interaction


Download ppt "Formal Verification of fFSM Model Sachoun Park, Gihwon Kwon Department of Computer Science Kyonggi University, Korea IWFST, Shanghai, China, 2005.11.08."

Similar presentations


Ads by Google