Check alignment of loads when deciding whether it is safe to execute them

unconditionally.  Besides checking the offset, also check that the underlying
object is aligned as much as the load itself.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@94875 91177308-0d34-0410-b5e6-96231b3b80d8

6 files changed, 70 insertions, 16 deletions

diff --git a/include/llvm/Transforms/Utils/Local.h b/include/llvm/Transforms/Utils/Local.h
index 3211518..bb91d80 100644
--- a/include/llvm/Transforms/Utils/Local.h
+++ b/include/llvm/Transforms/Utils/Local.h
@@ -39,7 +39,7 @@ template<typename T> class SmallVectorImpl;
 /// specified pointer, we do a quick local scan of the basic block containing
 /// ScanFrom, to determine if the address is already accessed.
 bool isSafeToLoadUnconditionally(Value *V, Instruction *ScanFrom,
-                                 const TargetData *TD = 0);
+                                 unsigned Align, const TargetData *TD = 0);
 
 //===----------------------------------------------------------------------===//
 //  Local constant propagation.
diff --git a/lib/Transforms/InstCombine/InstCombineLoadStoreAlloca.cpp b/lib/Transforms/InstCombine/InstCombineLoadStoreAlloca.cpp
index 422ffd0..2d13298 100644
--- a/lib/Transforms/InstCombine/InstCombineLoadStoreAlloca.cpp
+++ b/lib/Transforms/InstCombine/InstCombineLoadStoreAlloca.cpp
@@ -200,14 +200,15 @@ Instruction *InstCombiner::visitLoadInst(LoadInst &LI) {
     //
     if (SelectInst *SI = dyn_cast<SelectInst>(Op)) {
       // load (select (Cond, &V1, &V2))  --> select(Cond, load &V1, load &V2).
-      if (isSafeToLoadUnconditionally(SI->getOperand(1), SI, TD) &&
-          isSafeToLoadUnconditionally(SI->getOperand(2), SI, TD)) {
+      unsigned Align = LI.getAlignment();
+      if (isSafeToLoadUnconditionally(SI->getOperand(1), SI, Align, TD) &&
+          isSafeToLoadUnconditionally(SI->getOperand(2), SI, Align, TD)) {
         LoadInst *V1 = Builder->CreateLoad(SI->getOperand(1),
-                                        SI->getOperand(1)->getName()+".val");
+                                           SI->getOperand(1)->getName()+".val");
         LoadInst *V2 = Builder->CreateLoad(SI->getOperand(2),
-                                        SI->getOperand(2)->getName()+".val");
-        V1->setAlignment(LI.getAlignment());
-        V2->setAlignment(LI.getAlignment());
+                                           SI->getOperand(2)->getName()+".val");
+        V1->setAlignment(Align);
+        V2->setAlignment(Align);
         return SelectInst::Create(SI->getCondition(), V1, V2);
       }
 
diff --git a/lib/Transforms/Scalar/GVN.cpp b/lib/Transforms/Scalar/GVN.cpp
index 9c18452..6e70952 100644
--- a/lib/Transforms/Scalar/GVN.cpp
+++ b/lib/Transforms/Scalar/GVN.cpp
@@ -1651,7 +1651,8 @@ bool GVN::processNonLocalLoad(LoadInst *LI,
   if (!allSingleSucc &&
       // FIXME: REEVALUTE THIS.
       !isSafeToLoadUnconditionally(LoadPtr,
-                                   UnavailablePred->getTerminator(), TD)) {
+                                   UnavailablePred->getTerminator(),
+                                   LI->getAlignment(), TD)) {
     assert(NewInsts.empty() && "Should not have inserted instructions");
     return false;
   }
diff --git a/lib/Transforms/Scalar/TailRecursionElimination.cpp b/lib/Transforms/Scalar/TailRecursionElimination.cpp
index 4119cb9..162d902 100644
--- a/lib/Transforms/Scalar/TailRecursionElimination.cpp
+++ b/lib/Transforms/Scalar/TailRecursionElimination.cpp
@@ -211,7 +211,8 @@ bool TailCallElim::CanMoveAboveCall(Instruction *I, CallInst *CI) {
       // FIXME: Writes to memory only matter if they may alias the pointer
       // being loaded from.
       if (CI->mayWriteToMemory() ||
-          !isSafeToLoadUnconditionally(L->getPointerOperand(), L))
+          !isSafeToLoadUnconditionally(L->getPointerOperand(), L,
+                                       L->getAlignment()))
         return false;
     }
   }
diff --git a/lib/Transforms/Utils/Local.cpp b/lib/Transforms/Utils/Local.cpp
index 425f557..62c0ae0 100644
--- a/lib/Transforms/Utils/Local.cpp
+++ b/lib/Transforms/Utils/Local.cpp
@@ -75,31 +75,38 @@ static Value *getUnderlyingObjectWithOffset(Value *V, const TargetData *TD,
 /// specified pointer, we do a quick local scan of the basic block containing
 /// ScanFrom, to determine if the address is already accessed.
 bool llvm::isSafeToLoadUnconditionally(Value *V, Instruction *ScanFrom,
-                                       const TargetData *TD) {
+                                       unsigned Align, const TargetData *TD) {
   uint64_t ByteOffset = 0;
   Value *Base = V;
   if (TD)
     Base = getUnderlyingObjectWithOffset(V, TD, ByteOffset);
 
   const Type *BaseType = 0;
-  if (const AllocaInst *AI = dyn_cast<AllocaInst>(Base))
-    // If it is an alloca it is always safe to load from.
+  unsigned BaseAlign = 0;
+  if (const AllocaInst *AI = dyn_cast<AllocaInst>(Base)) {
+    // An alloca is safe to load from as load as it is suitably aligned.
     BaseType = AI->getAllocatedType();
-  else if (const GlobalValue *GV = dyn_cast<GlobalValue>(Base)) {
+    BaseAlign = AI->getAlignment();
+  } else if (const GlobalValue *GV = dyn_cast<GlobalValue>(Base)) {
     // Global variables are safe to load from but their size cannot be
     // guaranteed if they are overridden.
-    if (!isa<GlobalAlias>(GV) && !GV->mayBeOverridden())
+    if (!isa<GlobalAlias>(GV) && !GV->mayBeOverridden()) {
       BaseType = GV->getType()->getElementType();
+      BaseAlign = GV->getAlignment();
+    }
   }
+  if (TD && BaseType && BaseAlign == 0)
+    BaseAlign = TD->getPrefTypeAlignment(BaseType);
 
-  if (BaseType) {
+  if (BaseType && Align <= BaseAlign) {
     if (!TD)
       return true; // Loading directly from an alloca or global is OK.
     if (BaseType->isSized()) {
       // Check if the load is within the bounds of the underlying object.
       const PointerType *AddrTy = cast<PointerType>(V->getType());
       uint64_t LoadSize = TD->getTypeStoreSize(AddrTy->getElementType());
-      if (ByteOffset + LoadSize <= TD->getTypeAllocSize(BaseType))
+      if (ByteOffset + LoadSize <= TD->getTypeAllocSize(BaseType) &&
+          (Align == 0 || (ByteOffset % Align) == 0))
         return true;
     }
   }
diff --git a/test/Transforms/GVN/load-pre-align.ll b/test/Transforms/GVN/load-pre-align.ll
new file mode 100644
index 0000000..3a66c0b
--- /dev/null
+++ b/test/Transforms/GVN/load-pre-align.ll
@@ -0,0 +1,44 @@
+; RUN: opt < %s -gvn -S | FileCheck %s
+
+target datalayout = "e-p:32:32:32-i1:8:32-i8:8:32-i16:16:32-i32:32:32-i64:32:32-f32:32:32-f64:32:32-v64:64:64-v128:128:128-a0:0:32-n32"
+
+@p = external global i32
+
+define arm_apcscc i32 @test(i32 %n) nounwind {
+; CHECK: @test
+entry:
+  br label %for.cond
+
+; loads aligned greater than the memory should not be moved past conditionals
+; CHECK-NOT: load
+; CHECK: br i1
+
+for.cond:
+  %i.0 = phi i32 [ 0, %entry ], [ %indvar.next, %for.inc ]
+  %cmp = icmp slt i32 %i.0, %n
+  br i1 %cmp, label %for.body, label %for.cond.for.end_crit_edge
+
+for.cond.for.end_crit_edge:
+; ...but PRE can still move the load out of for.end to here.
+; CHECK: for.cond.for.end_crit_edge:
+; CHECK-NEXT: load
+  br label %for.end
+
+for.body:
+  %tmp3 = load i32* @p, align 8
+  %dec = add i32 %tmp3, -1
+  store i32 %dec, i32* @p
+  %cmp6 = icmp slt i32 %dec, 0
+  br i1 %cmp6, label %for.body.for.end_crit_edge, label %for.inc
+
+for.body.for.end_crit_edge:
+  br label %for.end
+
+for.inc:
+  %indvar.next = add i32 %i.0, 1
+  br label %for.cond
+
+for.end:
+  %tmp9 = load i32* @p, align 8
+  ret i32 %tmp9
+}