[PATCH] D117804: [TargetLowering][InstCombine] Simplify BSwap demanded bits code a little. NFC

[Craig Topper via Phabricator via llvm-commits]

craig.topper created this revision.
craig.topper added reviewers: spatel, RKSimon, lebedev.ri.
Herald added a subscriber: hiraditya.
craig.topper requested review of this revision.
Herald added a project: LLVM.

Use alignDown instead of &= ~7.

Replace ResultBit with NLZ. (BitWidth - NLZ - NTZ == 8) so
(BitWidth - NTZ - 8 == NLZ).


Repository:
  rG LLVM Github Monorepo

https://reviews.llvm.org/D117804

Files:
  llvm/lib/CodeGen/SelectionDAG/TargetLowering.cpp
  llvm/lib/Transforms/InstCombine/InstCombineSimplifyDemanded.cpp


Index: llvm/lib/Transforms/InstCombine/InstCombineSimplifyDemanded.cpp
===================================================================
--- llvm/lib/Transforms/InstCombine/InstCombineSimplifyDemanded.cpp
+++ llvm/lib/Transforms/InstCombine/InstCombineSimplifyDemanded.cpp
@@ -800,22 +800,19 @@
         // Round NTZ down to the next byte.  If we have 11 trailing zeros, then
         // we need all the bits down to bit 8.  Likewise, round NLZ.  If we
         // have 14 leading zeros, round to 8.
-        NLZ &= ~7;
-        NTZ &= ~7;
+        NLZ = alignDown(NLZ, 8);
+        NTZ = alignDown(NTZ, 8);
         // If we need exactly one byte, we can do this transformation.
         if (BitWidth-NLZ-NTZ == 8) {
-          unsigned ResultBit = NTZ;
-          unsigned InputBit = BitWidth-NTZ-8;
-
           // Replace this with either a left or right shift to get the byte into
           // the right place.
           Instruction *NewVal;
-          if (InputBit > ResultBit)
+          if (NLZ > NTZ)
             NewVal = BinaryOperator::CreateLShr(II->getArgOperand(0),
-                    ConstantInt::get(I->getType(), InputBit-ResultBit));
+                    ConstantInt::get(I->getType(), NLZ - NTZ));
           else
             NewVal = BinaryOperator::CreateShl(II->getArgOperand(0),
-                    ConstantInt::get(I->getType(), ResultBit-InputBit));
+                    ConstantInt::get(I->getType(), NTZ - NLZ));
           NewVal->takeName(I);
           return InsertNewInstWith(NewVal, *I);
         }
Index: llvm/lib/CodeGen/SelectionDAG/TargetLowering.cpp
===================================================================
--- llvm/lib/CodeGen/SelectionDAG/TargetLowering.cpp
+++ llvm/lib/CodeGen/SelectionDAG/TargetLowering.cpp
@@ -1815,20 +1815,16 @@
     // Round NTZ down to the next byte.  If we have 11 trailing zeros, then
     // we need all the bits down to bit 8.  Likewise, round NLZ.  If we
     // have 14 leading zeros, round to 8.
-    NLZ &= ~7;
-    NTZ &= ~7;
+    NLZ = alignDown(NLZ, 8);
+    NTZ = alignDown(NTZ, 8);
     // If we need exactly one byte, we can do this transformation.
     if (BitWidth - NLZ - NTZ == 8) {
-      unsigned ResultBit = NTZ;
-      unsigned InputBit = BitWidth - NTZ - 8;
-
       // Replace this with either a left or right shift to get the byte into
       // the right place.
-      unsigned ShiftOpcode = InputBit > ResultBit ? ISD::SRL : ISD::SHL;
+      unsigned ShiftOpcode = NLZ > NTZ ? ISD::SRL : ISD::SHL;
       if (!TLO.LegalOperations() || isOperationLegal(ShiftOpcode, VT)) {
         EVT ShiftAmtTy = getShiftAmountTy(VT, DL);
-        unsigned ShiftAmount =
-            InputBit > ResultBit ? InputBit - ResultBit : ResultBit - InputBit;
+        unsigned ShiftAmount = NLZ > NTZ ? NLZ - NTZ : NTZ - NLZ;
         SDValue ShAmt = TLO.DAG.getConstant(ShiftAmount, dl, ShiftAmtTy);
         SDValue NewOp = TLO.DAG.getNode(ShiftOpcode, dl, VT, Src, ShAmt);
         return TLO.CombineTo(Op, NewOp);


-------------- next part --------------
A non-text attachment was scrubbed...
Name: D117804.401676.patch
Type: text/x-patch
Size: 3008 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20220120/d0c03d09/attachment.bin>