Fix refactoring mistake in "Teach InstCombine to work with smaller legal types..."

[oota-llvm.git] / lib / Transforms / InstCombine / InstCombineCompares.cpp

diff --git a/lib/Transforms/InstCombine/InstCombineCompares.cpp b/lib/Transforms/InstCombine/InstCombineCompares.cpp
index e3e5ddae80b437604fb18347431e7a63202ec964..bad46b4dab3accc67a70329f229675f931a17e48 100644 (file)
--- a/lib/Transforms/InstCombine/InstCombineCompares.cpp
+++ b/lib/Transforms/InstCombine/InstCombineCompares.cpp
@@ -12,15 +12,15 @@
 //===----------------------------------------------------------------------===//
 
 #include "InstCombine.h"
-#include "llvm/IntrinsicInst.h"
 #include "llvm/Analysis/ConstantFolding.h"
 #include "llvm/Analysis/InstructionSimplify.h"
 #include "llvm/Analysis/MemoryBuiltins.h"
-#include "llvm/DataLayout.h"
-#include "llvm/Target/TargetLibraryInfo.h"
+#include "llvm/IR/DataLayout.h"
+#include "llvm/IR/IntrinsicInst.h"
 #include "llvm/Support/ConstantRange.h"
 #include "llvm/Support/GetElementPtrTypeIterator.h"
 #include "llvm/Support/PatternMatch.h"
+#include "llvm/Target/TargetLibraryInfo.h"
 using namespace llvm;
 using namespace PatternMatch;
 
@@ -365,12 +365,11 @@ FoldCmpLoadFromIndexedGlobal(GetElementPtrInst *GEP, GlobalVariable *GV,
   // order the state machines in complexity of the generated code.
   Value *Idx = GEP->getOperand(2);
 
-  unsigned AS = GEP->getPointerAddressSpace();
   // If the index is larger than the pointer size of the target, truncate the
   // index down like the GEP would do implicitly.  We don't have to do this for
   // an inbounds GEP because the index can't be out of range.
   if (!GEP->isInBounds() &&
-      Idx->getType()->getPrimitiveSizeInBits() > TD->getPointerSizeInBits(AS))
+      Idx->getType()->getPrimitiveSizeInBits() > TD->getPointerSizeInBits())
     Idx = Builder->CreateTrunc(Idx, TD->getIntPtrType(Idx->getContext()));
 
   // If the comparison is only true for one or two elements, emit direct
@@ -529,11 +528,10 @@ static Value *EvaluateGEPOffsetExpression(User *GEP, InstCombiner &IC) {
     }
   }
 
-  unsigned AS = cast<GetElementPtrInst>(GEP)->getPointerAddressSpace();
   // Okay, we know we have a single variable index, which must be a
   // pointer/array/vector index.  If there is no offset, life is simple, return
   // the index.
-  unsigned IntPtrWidth = TD.getPointerSizeInBits(AS);
+  unsigned IntPtrWidth = TD.getPointerSizeInBits();
   if (Offset == 0) {
     // Cast to intptrty in case a truncation occurs.  If an extension is needed,
     // we don't need to bother extending: the extension won't affect where the
@@ -1228,6 +1226,16 @@ Instruction *InstCombiner::visitICmpInstWithInstAndIntCst(ICmpInst &ICI,
         ICI.setOperand(0, NewAnd);
         return &ICI;
       }
+
+      // Replace ((X & AndCST) > RHSV) with ((X & AndCST) != 0), if any
+      // bit set in (X & AndCST) will produce a result greater than RHSV.
+      if (ICI.getPredicate() == ICmpInst::ICMP_UGT) {
+        unsigned NTZ = AndCST->getValue().countTrailingZeros();
+        if ((NTZ < AndCST->getBitWidth()) &&
+            APInt::getOneBitSet(AndCST->getBitWidth(), NTZ).ugt(RHSV))
+          return new ICmpInst(ICmpInst::ICMP_NE, LHSI,
+                              Constant::getNullValue(RHS->getType()));
+      }
     }
 
     // Try to optimize things like "A[i]&42 == 0" to index computations.
@@ -1323,6 +1331,25 @@ Instruction *InstCombiner::visitICmpInstWithInstAndIntCst(ICmpInst &ICI,
       return new ICmpInst(TrueIfSigned ? ICmpInst::ICMP_NE : ICmpInst::ICMP_EQ,
                           And, Constant::getNullValue(And->getType()));
     }
+
+    // Transform (icmp pred iM (shl iM %v, N), CI)
+    // -> (icmp pred i(M-N) (trunc %v iM to i(N-N)), (trunc (CI>>N))
+    // Transform the shl to a trunc if (trunc (CI>>N)) has no loss.
+    // This enables to get rid of the shift in favor of a trunc which can be
+    // free on the target. It has the additional benefit of comparing to a
+    // smaller constant, which will be target friendly.
+    unsigned Amt = ShAmt->getLimitedValue(TypeBits-1);
+    if (Amt != 0 && RHSV.countTrailingZeros() >= Amt) {
+      Type *NTy = IntegerType::get(ICI.getContext(), TypeBits - Amt);
+      Constant *NCI = ConstantExpr::getTrunc(
+                        ConstantExpr::getAShr(RHS,
+                          ConstantInt::get(RHS->getType(), Amt)),
+                        NTy);
+      return new ICmpInst(ICI.getPredicate(),
+                          Builder->CreateTrunc(LHSI->getOperand(0), NTy),
+                          NCI);
+    }
+
     break;
   }
 
@@ -1554,8 +1581,7 @@ Instruction *InstCombiner::visitICmpInstWithCastAndCast(ICmpInst &ICI) {
   // Turn icmp (ptrtoint x), (ptrtoint/c) into a compare of the input if the
   // integer type is the same size as the pointer type.
   if (TD && LHSCI->getOpcode() == Instruction::PtrToInt &&
-      TD->getPointerSizeInBits(
-        cast<PtrToIntInst>(LHSCI)->getPointerAddressSpace()) ==
+      TD->getPointerSizeInBits() ==
          cast<IntegerType>(DestTy)->getBitWidth()) {
     Value *RHSOp = 0;
     if (Constant *RHSC = dyn_cast<Constant>(ICI.getOperand(1))) {
@@ -2359,8 +2385,25 @@ Instruction *InstCombiner::visitICmpInst(ICmpInst &I) {
         // Try not to increase register pressure.
         BO0->hasOneUse() && BO1->hasOneUse()) {
       // Determine Y and Z in the form icmp (X+Y), (X+Z).
-      Value *Y = (A == C || A == D) ? B : A;
-      Value *Z = (C == A || C == B) ? D : C;
+      Value *Y, *Z;
+      if (A == C) {
+        // C + B == C + D  ->  B == D
+        Y = B;
+        Z = D;
+      } else if (A == D) {
+        // D + B == C + D  ->  B == C
+        Y = B;
+        Z = C;
+      } else if (B == C) {
+        // A + C == C + D  ->  A == D
+        Y = A;
+        Z = D;
+      } else {
+        assert(B == D);
+        // A + D == C + D  ->  A == C
+        Y = A;
+        Z = C;
+      }
       return new ICmpInst(Pred, Y, Z);
     }
 
@@ -2898,10 +2941,6 @@ Instruction *InstCombiner::visitFCmpInst(FCmpInst &I) {
         if (!RHSF)
           break;
 
-        // We can't convert a PPC double double.
-        if (RHSF->getType()->isPPC_FP128Ty())
-          break;
-
         const fltSemantics *Sem;
         // FIXME: This shouldn't be here.
         if (LHSExt->getSrcTy()->isHalfTy())
@@ -2914,6 +2953,8 @@ Instruction *InstCombiner::visitFCmpInst(FCmpInst &I) {
           Sem = &APFloat::IEEEquad;
         else if (LHSExt->getSrcTy()->isX86_FP80Ty())
           Sem = &APFloat::x87DoubleExtended;
+        else if (LHSExt->getSrcTy()->isPPC_FP128Ty())
+          Sem = &APFloat::PPCDoubleDouble;
         else
           break;