fix TLI's combineRepeatedFPDivisors interface to return the minimum user threshold

[spatel]

Summary

This fix was suggested as part of D11345.

We can avoid walking the uses of a divisor node if the target doesn't want the combineRepeatedFPDivisors transform in the first place.

This is a NFC-intended patch, but I want to make sure the new hook is named and defined as expected. I thought about returning a single 'unsigned' value and treating '0' as a special case for "don't do this transform", but that seems a bit dirty.

[samsonov]

Looks reasonable to me, although I don't often touch that code.

[spatel]

In D11531#213801, @samsonov wrote:

Looks reasonable to me, although I don't often touch that code.

Thanks, Alexey! Can I use that as an official LGTM? :)

Either way, I'll wait a bit to see if Chandler or Hal has feedback since they commented earlier.

[chandlerc]

LGTM with one tweak.

If others want to keep bike-shedding this TTI interface, fine, but let's get this in because PR24141 is hitting us.

include/llvm/Target/TargetLowering.h

2738

Please make this return the minimum number of uses that should exist for this combine. "zero" is a pretty clear way to disable this IMO, as there is never a reason to combine anything with zero uses.

include/llvm/Target/TargetLowering.h

SDValue BuildUDIV(SDNode *N, const APInt &Divisor, SelectionDAG &DAG,
                     bool IsAfterLegalization,
                     std::vector<SDNode *> *Created) const;
   virtual SDValue BuildSDIVPow2(SDNode *N, const APInt &Divisor,
                                 SelectionDAG &DAG,
                                 std::vector<SDNode *> *Created) const {
     return SDValue();
   }
 
-  /// Indicate whether this target prefers to combine the given number of FDIVs
-  /// with the same divisor.
-  virtual bool combineRepeatedFPDivisors(unsigned NumUsers) const {
+  /// Indicate whether this target prefers to combine FDIVs with the same
+  /// divisor. If the transform should be done, return the minimum number of
+  /// divisor uses that should exist before doing the transform in MinUses.
+  virtual bool combineRepeatedFPDivisors(unsigned &MinUses) const {
     return false;
   }
 
   /// Hooks for building estimates in place of slower divisions and square
   /// roots.
 
   /// Return a reciprocal square root estimate value for the input operand.
   /// The RefinementSteps output is the number of Newton-Raphson refinement

lib/CodeGen/SelectionDAG/DAGCombiner.cpp

 // Notice that this is not always beneficial. One reason is different target
 // may have different costs for FDIV and FMUL, so sometimes the cost of two
 // FDIVs may be lower than the cost of one FDIV and two FMULs. Another reason
 // is the critical path is increased from "one FDIV" to "one FDIV + one FMUL".
 SDValue DAGCombiner::combineRepeatedFPDivisors(SDNode *N) {
   if (!DAG.getTarget().Options.UnsafeFPMath)
     return SDValue();
   
+  // Skip if current node is a reciprocal.
   SDValue N0 = N->getOperand(0);
   ConstantFPSDNode *N0CFP = dyn_cast<ConstantFPSDNode>(N0);
-
-  // Skip if current node is a reciprocal.
   if (N0CFP && N0CFP->isExactlyValue(1.0))
     return SDValue();
   
+  // Exit early if the target does not want this transform or if there can't
+  // possibly be enough uses of the divisor to make the transform worthwhile.
   SDValue N1 = N->getOperand(1);
-  SmallVector<SDNode *, 4> Users;
+  unsigned MinUses;
+  if (!TLI.combineRepeatedFPDivisors(MinUses) || N1->use_size() < MinUses)
+    return SDValue();
 
   // Find all FDIV users of the same divisor.
+  SmallVector<SDNode *, 4> Users;
   for (auto *U : N1->uses()) {
     if (U->getOpcode() == ISD::FDIV && U->getOperand(1) == N1)
       Users.push_back(U);
   }
 
-  if (!TLI.combineRepeatedFPDivisors(Users.size()))
+  // Now that we have the actual number of divisor uses, make sure it meets
+  // the minimum threshold specified by the target.
+  if (Users.size() < MinUses)
     return SDValue();
 
   EVT VT = N->getValueType(0);
   SDLoc DL(N);
   SDValue FPOne = DAG.getConstantFP(1.0, DL, VT);
   // FIXME: This optimization requires some level of fast-math, so the
   // created reciprocal node should at least have the 'allowReciprocal'
   // fast-math-flag set.

lib/Target/AArch64/AArch64ISelLowering.h

private:
   SDValue LowerINT_TO_FP(SDValue Op, SelectionDAG &DAG) const;
   SDValue LowerVectorAND(SDValue Op, SelectionDAG &DAG) const;
   SDValue LowerVectorOR(SDValue Op, SelectionDAG &DAG) const;
   SDValue LowerCONCAT_VECTORS(SDValue Op, SelectionDAG &DAG) const;
   SDValue LowerFSINCOS(SDValue Op, SelectionDAG &DAG) const;
 
   SDValue BuildSDIVPow2(SDNode *N, const APInt &Divisor, SelectionDAG &DAG,
                         std::vector<SDNode *> *Created) const override;
-  bool combineRepeatedFPDivisors(unsigned NumUsers) const override;
+  bool combineRepeatedFPDivisors(unsigned &MinUses) const override;
 
   ConstraintType getConstraintType(StringRef Constraint) const override;
   unsigned getRegisterByName(const char* RegName, EVT VT,
                              SelectionDAG &DAG) const override;
 
   /// Examine constraint string and operand type and determine a weight value.
   /// The operand object must already have been set up with the operand type.
   ConstraintWeight

lib/Target/AArch64/AArch64ISelLowering.cpp

case ISD::FP_TO_SINT:
     return;
   }
 }
 
 bool AArch64TargetLowering::useLoadStackGuardNode() const {
   return true;
 }
 
-bool AArch64TargetLowering::combineRepeatedFPDivisors(unsigned NumUsers) const {
+bool AArch64TargetLowering::combineRepeatedFPDivisors(unsigned &MinUses) const {
   // Combine multiple FDIVs with the same divisor into multiple FMULs by the
   // reciprocal if there are three or more FDIVs.
-  return NumUsers > 2;
+  MinUses = 3;
+  return true;
 }
 
 TargetLoweringBase::LegalizeTypeAction
 AArch64TargetLowering::getPreferredVectorAction(EVT VT) const {
   MVT SVT = VT.getSimpleVT();
   // During type legalization, we prefer to widen v1i8, v1i16, v1i32  to v8i8,
   // v4i16, v2i32 instead of to promote.
   if (SVT == MVT::v1i8 || SVT == MVT::v1i16 || SVT == MVT::v1i32

lib/Target/PowerPC/PPCISelLowering.h

private:
     SDValue DAGCombineTruncBoolExt(SDNode *N, DAGCombinerInfo &DCI) const;
     SDValue combineFPToIntToFP(SDNode *N, DAGCombinerInfo &DCI) const;
 
     SDValue getRsqrtEstimate(SDValue Operand, DAGCombinerInfo &DCI,
                              unsigned &RefinementSteps,
                              bool &UseOneConstNR) const override;
     SDValue getRecipEstimate(SDValue Operand, DAGCombinerInfo &DCI,
                              unsigned &RefinementSteps) const override;
-    bool combineRepeatedFPDivisors(unsigned NumUsers) const override;
+    bool combineRepeatedFPDivisors(unsigned &MinUses) const override;
 
     CCAssignFn *useFastISelCCs(unsigned Flag) const;
   };
 
   namespace PPC {
     FastISel *createFastISel(FunctionLoweringInfo &FuncInfo,
                              const TargetLibraryInfo *LibInfo);
   }

lib/Target/PowerPC/PPCISelLowering.cpp

if (!Recips.isEnabled(RecipOp))
       return SDValue();
 
     RefinementSteps = Recips.getRefinementSteps(RecipOp);
     return DCI.DAG.getNode(PPCISD::FRE, SDLoc(Operand), VT, Operand);
   }
   return SDValue();
 }
 
-bool PPCTargetLowering::combineRepeatedFPDivisors(unsigned NumUsers) const {
+bool PPCTargetLowering::combineRepeatedFPDivisors(unsigned &MinUses) const {
   // Note: This functionality is used only when unsafe-fp-math is enabled, and
   // on cores with reciprocal estimates (which are used when unsafe-fp-math is
   // enabled for division), this functionality is redundant with the default
   // combiner logic (once the division -> reciprocal/multiply transformation
   // has taken place). As a result, this matters more for older cores than for
   // newer ones.
 
   // Combine multiple FDIVs with the same divisor into multiple FMULs by the
   // reciprocal if there are two or more FDIVs (for embedded cores with only
   // one FP pipeline) for three or more FDIVs (for generic OOO cores).
   switch (Subtarget.getDarwinDirective()) {
   default:
-    return NumUsers > 2;
+    MinUses = 3;
   case PPC::DIR_440:
   case PPC::DIR_A2:
   case PPC::DIR_E500mc:
   case PPC::DIR_E5500:
-    return NumUsers > 1;
+    MinUses = 2;
   }
+  return true;
 }
 
 static bool isConsecutiveLSLoc(SDValue Loc, EVT VT, LSBaseSDNode *Base,
                             unsigned Bytes, int Dist,
                             SelectionDAG &DAG) {
   if (VT.getSizeInBits() / 8 != Bytes)
     return false;
 

lib/Target/X86/X86ISelLowering.h

SDValue getRsqrtEstimate(SDValue Operand, DAGCombinerInfo &DCI,
                              unsigned &RefinementSteps,
                              bool &UseOneConstNR) const override;
 
     /// Use rcp* to speed up fdiv calculations.
     SDValue getRecipEstimate(SDValue Operand, DAGCombinerInfo &DCI,
                              unsigned &RefinementSteps) const override;
 
     /// Reassociate floating point divisions into multiply by reciprocal.
-    bool combineRepeatedFPDivisors(unsigned NumUsers) const override;
+    bool combineRepeatedFPDivisors(unsigned &MinUses) const override;
   };
 
   namespace X86 {
     FastISel *createFastISel(FunctionLoweringInfo &funcInfo,
                              const TargetLibraryInfo *libInfo);
   }
 }
 
 #endif    // X86ISELLOWERING_H

lib/Target/X86/X86ISelLowering.cpp

 }
 
 /// If we have at least two divisions that use the same divisor, convert to
 /// multplication by a reciprocal. This may need to be adjusted for a given
 /// CPU if a division's cost is not at least twice the cost of a multiplication.
 /// This is because we still need one division to calculate the reciprocal and
 /// then we need two multiplies by that reciprocal as replacements for the
 /// original divisions.
-bool X86TargetLowering::combineRepeatedFPDivisors(unsigned NumUsers) const {
-  return NumUsers > 1;
+bool X86TargetLowering::combineRepeatedFPDivisors(unsigned &MinUses) const {
+  MinUses = 2;
+  return true;
 }
 
 static bool isAllOnes(SDValue V) {
   ConstantSDNode *C = dyn_cast<ConstantSDNode>(V);
   return C && C->isAllOnesValue();
 }
 
 /// LowerToBT - Result of 'and' is compared against zero. Turn it into a BT node