[InstCombine] fmul nnan X, 0.0 --> copysign(0.0, X)

[spatel]

Summary

I've had this Alive tab open in my browser for weeks, and the similar D136098 reminded me of it:
https://alive2.llvm.org/ce/z/ybgM5F

llvm/lib/Transforms/InstCombine/InstCombineMulDivRem.cpp

Instruction *InstCombinerImpl::visitFMul(BinaryOperator &I) {
   if (Instruction *R = foldFPSignBitOps(I))
     return R;
 
   // X * -1.0 --> -X
   Value *Op0 = I.getOperand(0), *Op1 = I.getOperand(1);
   if (match(Op1, m_SpecificFP(-1.0)))
     return UnaryOperator::CreateFNegFMF(Op0, &I);
 
+  // With no-nans: X * 0.0 --> copysign(0.0, X)
+  if (I.hasNoNaNs() && match(Op1, m_PosZeroFP())) {
+    CallInst *CopySign = Builder.CreateIntrinsic(Intrinsic::copysign,
+                                                 {I.getType()}, {Op1, Op0}, &I);
+    return replaceInstUsesWith(I, CopySign);
+  }
+
   // -X * C --> X * -C
   Value *X, *Y;
   Constant *C;
   if (match(Op0, m_FNeg(m_Value(X))) && match(Op1, m_Constant(C)))
     if (Constant *NegC = ConstantFoldUnaryOpOperand(Instruction::FNeg, C, DL))
       return BinaryOperator::CreateFMulFMF(X, NegC, &I);
 
   // (select A, B, C) * (select A, D, E) --> select A, (B*D), (C*E)

llvm/test/Transforms/InstCombine/fmul.ll

 ;
   %shuf = shufflevector <vscale x 2 x float> insertelement (<vscale x 2 x float> undef, float 0.0, i32 0), <vscale x 2 x float> undef, <vscale x 2 x i32> zeroinitializer
   %t3 = fmul fast <vscale x 2 x float> %shuf, %z
   ret <vscale x 2 x float> %t3
 }
 
 define half @mul_zero_nnan(half %x) {
 ; CHECK-LABEL: @mul_zero_nnan(
-; CHECK-NEXT:    [[R:%.*]] = fmul nnan half [[X:%.*]], 0xH0000
-; CHECK-NEXT:    ret half [[R]]
+; CHECK-NEXT:    [[TMP1:%.*]] = call nnan half @llvm.copysign.f16(half 0xH0000, half [[X:%.*]])
+; CHECK-NEXT:    ret half [[TMP1]]
 ;
   %r = fmul nnan half %x, 0.0
   ret half %r
 }
 
+; poison propagates through vector elements
+
 define <2 x float> @mul_zero_nnan_vec_poison(<2 x float> %x) {
 ; CHECK-LABEL: @mul_zero_nnan_vec_poison(
-; CHECK-NEXT:    [[R:%.*]] = fmul nnan <2 x float> [[X:%.*]], <float 0.000000e+00, float poison>
-; CHECK-NEXT:    ret <2 x float> [[R]]
+; CHECK-NEXT:    [[TMP1:%.*]] = call nnan <2 x float> @llvm.copysign.v2f32(<2 x float> <float 0.000000e+00, float poison>, <2 x float> [[X:%.*]])
+; CHECK-NEXT:    ret <2 x float> [[TMP1]]
 ;
   %r = fmul nnan <2 x float> %x, <float 0.0, float poison>
   ret <2 x float> %r
 }
 
+; negative test - must have nnan
+
 define half @mul_zero(half %x) {
 ; CHECK-LABEL: @mul_zero(
 ; CHECK-NEXT:    [[R:%.*]] = fmul ninf nsz half [[X:%.*]], 0xH0000
 ; CHECK-NEXT:    ret half [[R]]
 ;
   %r = fmul ninf nsz half %x, 0.0
   ret half %r
 }
 
+; TODO: This could be fneg+copysign.
+
 define half @mul_negzero_nnan(half %x) {
 ; CHECK-LABEL: @mul_negzero_nnan(
 ; CHECK-NEXT:    [[R:%.*]] = fmul nnan half [[X:%.*]], 0xH8000
 ; CHECK-NEXT:    ret half [[R]]
 ;
   %r = fmul nnan half %x, -0.0
   ret half %r
 }