[llvm] r340085 - DAG: Fix isKnownNeverNaN for basic non-sNaN cases

[Matt Arsenault via llvm-commits]

Author: arsenm
Date: Fri Aug 17 14:19:22 2018
New Revision: 340085

URL: http://llvm.org/viewvc/llvm-project?rev=340085&view=rev
Log:
DAG: Fix isKnownNeverNaN for basic non-sNaN cases

fadd/fsub/fmul need to worry about infinities as well
as fdiv.

Added:
    llvm/trunk/test/CodeGen/AArch64/known-never-nan.ll
Modified:
    llvm/trunk/lib/CodeGen/SelectionDAG/SelectionDAG.cpp

Modified: llvm/trunk/lib/CodeGen/SelectionDAG/SelectionDAG.cpp
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/CodeGen/SelectionDAG/SelectionDAG.cpp?rev=340085&r1=340084&r2=340085&view=diff
==============================================================================
--- llvm/trunk/lib/CodeGen/SelectionDAG/SelectionDAG.cpp (original)
+++ llvm/trunk/lib/CodeGen/SelectionDAG/SelectionDAG.cpp Fri Aug 17 14:19:22 2018
@@ -3641,11 +3641,15 @@ bool SelectionDAG::isKnownNeverNaN(SDVal
   switch (Opcode) {
   case ISD::FADD:
   case ISD::FSUB:
-  case ISD::FMUL: {
+  case ISD::FMUL:
+  case ISD::FDIV:
+  case ISD::FREM:
+  case ISD::FSIN:
+  case ISD::FCOS: {
     if (SNaN)
       return true;
-    return isKnownNeverNaN(Op.getOperand(0), SNaN, Depth + 1) &&
-           isKnownNeverNaN(Op.getOperand(1), SNaN, Depth + 1);
+    // TODO: Need isKnownNeverInfinity
+    return false;
   }
   case ISD::FCANONICALIZE:
   case ISD::FEXP:
@@ -3668,15 +3672,6 @@ bool SelectionDAG::isKnownNeverNaN(SDVal
   case ISD::SELECT:
     return isKnownNeverNaN(Op.getOperand(1), SNaN, Depth + 1) &&
            isKnownNeverNaN(Op.getOperand(2), SNaN, Depth + 1);
-  case ISD::FDIV:
-  case ISD::FREM:
-  case ISD::FSIN:
-  case ISD::FCOS: {
-    if (SNaN)
-      return true;
-    // TODO: Need isKnownNeverInfinity
-    return false;
-  }
   case ISD::FP_EXTEND:
   case ISD::FP_ROUND: {
     if (SNaN)

Added: llvm/trunk/test/CodeGen/AArch64/known-never-nan.ll
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/test/CodeGen/AArch64/known-never-nan.ll?rev=340085&view=auto
==============================================================================
--- llvm/trunk/test/CodeGen/AArch64/known-never-nan.ll (added)
+++ llvm/trunk/test/CodeGen/AArch64/known-never-nan.ll Fri Aug 17 14:19:22 2018
@@ -0,0 +1,50 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: llc -mtriple=aarch64 < %s | FileCheck %s
+
+; This should codegen to fmaxnm with no-signed-zeros.
+define float @fmaxnm(i32 %i1, i32 %i2) #0 {
+; CHECK-LABEL: fmaxnm:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    ucvtf s0, w0
+; CHECK-NEXT:    fmov s1, #11.00000000
+; CHECK-NEXT:    ucvtf s2, w1
+; CHECK-NEXT:    fmov s3, #17.00000000
+; CHECK-NEXT:    fadd s0, s0, s1
+; CHECK-NEXT:    fadd s1, s2, s3
+; CHECK-NEXT:    fcmp s0, s1
+; CHECK-NEXT:    fcsel s0, s0, s1, pl
+; CHECK-NEXT:    ret
+  %f1 = uitofp i32 %i1 to float
+  %fadd1 = fadd float %f1, 11.0
+  %f2 = uitofp i32 %i2 to float
+  %fadd2 = fadd float %f2, 17.0
+  %cmp = fcmp uge float %fadd1, %fadd2
+  %val = select i1 %cmp, float %fadd1, float %fadd2
+  ret float %val
+}
+
+; If f1 is 0, fmul is NaN because 0.0 * -INF = NaN
+; Therefore, this is not fmaxnm.
+define float @not_fmaxnm_maybe_nan(i32 %i1, i32 %i2) #0 {
+; CHECK-LABEL: not_fmaxnm_maybe_nan:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    adrp x8, .LCPI1_0
+; CHECK-NEXT:    ldr s0, [x8, :lo12:.LCPI1_0]
+; CHECK-NEXT:    ucvtf s1, w0
+; CHECK-NEXT:    ucvtf s2, w1
+; CHECK-NEXT:    fmov s3, #17.00000000
+; CHECK-NEXT:    fmul s0, s1, s0
+; CHECK-NEXT:    fadd s1, s2, s3
+; CHECK-NEXT:    fcmp s0, s1
+; CHECK-NEXT:    fcsel s0, s0, s1, pl
+; CHECK-NEXT:    ret
+  %f1 = uitofp i32 %i1 to float
+  %fmul = fmul float %f1, 0xfff0000000000000 ; -INFINITY as 64-bit hex
+  %f2 = uitofp i32 %i2 to float
+  %fadd2 = fadd float %f2, 17.0
+  %cmp = fcmp uge float %fmul, %fadd2
+  %val = select i1 %cmp, float %fmul, float %fadd2
+  ret float %val
+}
+
+attributes #0 = { "no-signed-zeros-fp-math"="true" }