Fix BasicAA incorrect assumption on GEP

[mehdi_amini]

Summary

This is fixing pr31761: BasicAA is deducing NoAlias
on the result of the GEP if the base pointer is itself NoAlias.

This is possible only if the NoAlias on the base pointer is
deduced with a non-sized query: this should guarantee that
the pointers are belonging to different memory allocation
and that the GEP can't legally jump from one to another.

[mehdi_amini]

Update after discussing with Chandler on IRC: making the query unsized should be enough here!

If we deduce NoAlias with an unsized query, it should mean that a GEP can't
reach one allocation from another.

[hfinkel]

LGTM. This should go into the branch too.

llvm/lib/Analysis/BasicAliasAnalysis.cpp

1194 ↗

(On Diff #86027)

I had to look at this for a few moments to realize why this was broken: UnknownSize might be unknown, but it is still assumed to be positive, so we only get a one-sided range. Thus, we must use UnknownSize for both sizes.

llvm/test/Transforms/SLPVectorizer/X86/tiny-tree.ll

5 ↗

(On Diff #86027)

We don't need to touch this file now.

llvm/trunk/lib/Analysis/BasicAliasAnalysis.cpp

if (const GEPOperator *GEP2 = dyn_cast<GEPOperator>(V2)) {
     // pointer, we know they cannot alias.
 
     // If both accesses are unknown size, we can't do anything useful here.
     if (V1Size == MemoryLocation::UnknownSize &&
         V2Size == MemoryLocation::UnknownSize)
       return MayAlias;
 
     AliasResult R = aliasCheck(UnderlyingV1, MemoryLocation::UnknownSize,
-                               AAMDNodes(), V2, V2Size, V2AAInfo,
-                               nullptr, UnderlyingV2);
+                               AAMDNodes(), V2, MemoryLocation::UnknownSize,
+                               V2AAInfo, nullptr, UnderlyingV2);
     if (R != MustAlias)
       // If V2 may alias GEP base pointer, conservatively returns MayAlias.
       // If V2 is known not to alias GEP base pointer, then the two values
-      // cannot alias per GEP semantics: "A pointer value formed from a
-      // getelementptr instruction is associated with the addresses associated
-      // with the first operand of the getelementptr".
+      // cannot alias per GEP semantics: "Any memory access must be done through
+      // a pointer value associated with an address range of the memory access,
+      // otherwise the behavior is undefined.".
       return R;
 
     // If the max search depth is reached the result is undefined
     if (GEP1MaxLookupReached)
       return MayAlias;
   }
 
   // In the two GEP Case, if there is no difference in the offsets of the

llvm/trunk/test/Analysis/BasicAA/pr31761.ll

+; RUN: opt < %s -basicaa -aa-eval -print-all-alias-modref-info -disable-output 2>&1 | FileCheck %s
+
+
+target datalayout = "e-m:o-i64:64-f80:128-n8:16:32:64-S128"
+target triple = "x86_64-apple-macosx10.12.0"
+
+%struct.blam = type { i32, i32 }
+
+
+; CHECK-DAG: MayAlias: i32* %tmp, i32* %tmp3
+
+define i1 @ham(%struct.blam* %arg)  {
+  %isNull = icmp eq %struct.blam* %arg, null
+  %tmp = getelementptr  %struct.blam, %struct.blam* %arg, i64 0, i32 0
+  %tmp2 = getelementptr  %struct.blam, %struct.blam* %arg, i64 0, i32 1
+  %select = select i1 %isNull, i32* null, i32* %tmp2
+  %tmp3 = getelementptr  i32, i32* %select, i32 -1
+  ret i1 true
+}