[analyzer] Add support for NoRedundancy inlining mode.

We do not reanalyze a function, which has already been analyzed as an
inlined callee. As per PRELIMINARY testing, this gives over
50% run time reduction on some benchmarks without decreasing of the
number of bugs found.

Turning the mode on by default.

git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@152440 91177308-0d34-0410-b5e6-96231b3b80d8

1 files changed, 9 insertions, 2 deletions

diff --git a/lib/StaticAnalyzer/Core/CoreEngine.cpp b/lib/StaticAnalyzer/Core/CoreEngine.cpp
index e7c3d2a3b3..a350757a0f 100644
--- a/lib/StaticAnalyzer/Core/CoreEngine.cpp
+++ b/lib/StaticAnalyzer/Core/CoreEngine.cpp
@@ -214,9 +214,16 @@ bool CoreEngine::ExecuteWorkList(const LocationContext *L, unsigned Steps,
         assert (false && "BlockExit location never occur in forward analysis.");
         break;
 
-      case ProgramPoint::CallEnterKind:
-        SubEng.processCallEnter(cast<CallEnter>(Node->getLocation()), Node);
+      case ProgramPoint::CallEnterKind: {
+        CallEnter CEnter = cast<CallEnter>(Node->getLocation());
+        if (AnalyzedCallees)
+          if (const CallExpr* CE =
+              dyn_cast_or_null<CallExpr>(CEnter.getCallExpr()))
+            if (const Decl *CD = CE->getCalleeDecl())
+              AnalyzedCallees->insert(CD);
+        SubEng.processCallEnter(CEnter, Node);
         break;
+      }
 
       case ProgramPoint::CallExitKind:
         SubEng.processCallExit(Node);