[llvm] r274656 - Work around PR28400 a bit harder.

[Sean Silva via llvm-commits]

Author: silvas
Date: Wed Jul  6 14:05:41 2016
New Revision: 274656

URL: http://llvm.org/viewvc/llvm-project?rev=274656&view=rev
Log:
Work around PR28400 a bit harder.

We were still crashing in the "no change" case because LVI was not
getting invalidated.

See the thread "Should analyses be able to hold AssertingVH to IR?
(related to PR28400)" for more discussion.

Added:
    llvm/trunk/test/Transforms/JumpThreading/crash-assertingvh.ll
Modified:
    llvm/trunk/lib/Transforms/Scalar/JumpThreading.cpp

Modified: llvm/trunk/lib/Transforms/Scalar/JumpThreading.cpp
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/Transforms/Scalar/JumpThreading.cpp?rev=274656&r1=274655&r2=274656&view=diff
==============================================================================
--- llvm/trunk/lib/Transforms/Scalar/JumpThreading.cpp (original)
+++ llvm/trunk/lib/Transforms/Scalar/JumpThreading.cpp Wed Jul  6 14:05:41 2016
@@ -148,11 +148,14 @@ PreservedAnalyses JumpThreadingPass::run
   }
   bool Changed =
       runImpl(F, &TLI, &LVI, HasProfileData, std::move(BFI), std::move(BPI));
+
+  // FIXME: We need to invalidate LVI to avoid PR28400. Is there a better
+  // solution?
+  AM.invalidate<LazyValueAnalysis>(F);
+
   if (!Changed)
     return PreservedAnalyses::all();
   PreservedAnalyses PA;
-  // FIXME: Not preserving LVI! We need it to be invalidated so that we
-  // don't run into issues like PR28400. Is there a better solution?
   PA.preserve<GlobalsAA>();
   return PA;
 }

Added: llvm/trunk/test/Transforms/JumpThreading/crash-assertingvh.ll
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/test/Transforms/JumpThreading/crash-assertingvh.ll?rev=274656&view=auto
==============================================================================
--- llvm/trunk/test/Transforms/JumpThreading/crash-assertingvh.ll (added)
+++ llvm/trunk/test/Transforms/JumpThreading/crash-assertingvh.ll Wed Jul  6 14:05:41 2016
@@ -0,0 +1,19 @@
+; RUN: opt -disable-output < %s -passes='module(function(jump-threading),globaldce)'
+
+target datalayout = "e-m:e-i64:64-f80:128-n8:16:32:64-S128"
+target triple = "x86_64-unknown-linux-gnu"
+
+declare i32 @bar()
+
+define internal i32 @foo() {
+entry:
+  %call4 = call i32 @bar()
+  %cmp5 = icmp eq i32 %call4, 0
+  br i1 %cmp5, label %if.then6, label %if.end8
+
+if.then6:
+  ret i32 0
+
+if.end8:
+  ret i32 1
+}