[AliasAnalysis] Treat invariant.start as read-memory

Summary:
We teach alias analysis that invariant.start is readonly.
This helps with GVN and memcopy optimizations that currently treat.
invariant.start as a clobber.
We need to treat this as readonly, so that DSE does not incorrectly
remove stores prior to the invariant.start

Reviewers: sanjoy, reames, majnemer, dberlin

Subscribers: llvm-commits

Differential Revision: https://reviews.llvm.org/D23214

llvm-svn: 278138

1 files changed, 26 insertions, 0 deletions

diff --git a/llvm/lib/Analysis/BasicAliasAnalysis.cpp b/llvm/lib/Analysis/BasicAliasAnalysis.cpp
index 89ae3d0..84affa1 100644
--- a/llvm/lib/Analysis/BasicAliasAnalysis.cpp
+++ b/llvm/lib/Analysis/BasicAliasAnalysis.cpp
@@ -781,6 +781,32 @@ ModRefInfo BasicAAResult::getModRefInfo(ImmutableCallSite CS,
   if (isIntrinsicCall(CS, Intrinsic::experimental_guard))
     return MRI_Ref;
 
+  // Like assumes, invariant.start intrinsics were also marked as arbitrarily
+  // writing so that proper control dependencies are maintained but they never
+  // mod any particular memory location visible to the IR.
+  // *Unlike* assumes (which are now modeled as NoModRef), invariant.start
+  // intrinsic is now modeled as reading memory. This prevents hoisting the
+  // invariant.start intrinsic over stores. Consider:
+  // *ptr = 40;
+  // *ptr = 50;
+  // invariant_start(ptr)
+  // int val = *ptr;
+  // print(val);
+  //
+  // This cannot be transformed to:
+  //
+  // *ptr = 40;
+  // invariant_start(ptr)
+  // *ptr = 50;
+  // int val = *ptr;
+  // print(val);
+  //
+  // The transformation will cause the second store to be ignored (based on
+  // rules of invariant.start)  and print 40, while the first program always
+  // prints 50.
+  if (isIntrinsicCall(CS, Intrinsic::invariant_start))
+    return MRI_Ref;
+
   // The AAResultBase base class has some smarts, lets use them.
   return AAResultBase::getModRefInfo(CS, Loc);
 }