Index: llvm/lib/Analysis/ValueTracking.cpp
===================================================================
--- llvm/lib/Analysis/ValueTracking.cpp
+++ llvm/lib/Analysis/ValueTracking.cpp
@@ -4490,6 +4490,40 @@
 
     break;
   }
+  case Instruction::FDiv: {
+    const bool WantNan = (InterestedClasses & fcNan) != fcNone;
+    if (!WantNan)
+      break;
+
+    // TODO: FRem
+    KnownFPClass KnownLHS, KnownRHS;
+
+    computeKnownFPClass(Op->getOperand(1), DemandedElts,
+                        fcNan | fcInf | fcZero | fcSubnormal, KnownRHS,
+                        Depth + 1, Q, TLI);
+
+    bool KnowSomethingUseful = KnownRHS.isKnownNeverNaN() ||
+                               KnownRHS.isKnownNeverInfinity() ||
+                               KnownRHS.isKnownNeverZero();
+
+    if (KnowSomethingUseful) {
+      computeKnownFPClass(Op->getOperand(0), DemandedElts,
+                          fcNan | fcInf | fcZero, KnownLHS, Depth + 1, Q, TLI);
+    }
+
+    const Function *F = cast<Instruction>(Op)->getFunction();
+
+    // Only 0/0, Inf/Inf, Inf REM x and x REM 0 produce NaN.
+    // TODO: Track sign bit.
+    if (KnownLHS.isKnownNeverNaN() && KnownRHS.isKnownNeverNaN() &&
+        (KnownLHS.isKnownNeverInfinity() || KnownRHS.isKnownNeverInfinity()) &&
+        (KnownLHS.isKnownNeverLogicalZero(*F, Op->getType()) ||
+         KnownRHS.isKnownNeverLogicalZero(*F, Op->getType()))) {
+      Known.knownNot(fcNan);
+    }
+
+    break;
+  }
   case Instruction::FPExt: {
     // Infinity, nan and zero propagate from source.
     computeKnownFPClass(Op->getOperand(0), DemandedElts, InterestedClasses,
Index: llvm/test/Transforms/Attributor/nofpclass-fdiv.ll
===================================================================
--- llvm/test/Transforms/Attributor/nofpclass-fdiv.ll
+++ llvm/test/Transforms/Attributor/nofpclass-fdiv.ll
@@ -52,7 +52,7 @@
 }
 
 define float @ret_fdiv_ieee_nonan_noinf_nozero__nonan(float nofpclass(nan inf zero) %arg0, float nofpclass(nan) %arg1) #0 {
-; CHECK-LABEL: define float @ret_fdiv_ieee_nonan_noinf_nozero__nonan
+; CHECK-LABEL: define nofpclass(nan) float @ret_fdiv_ieee_nonan_noinf_nozero__nonan
 ; CHECK-SAME: (float nofpclass(nan inf zero) [[ARG0:%.*]], float nofpclass(nan) [[ARG1:%.*]]) #[[ATTR0]] {
 ; CHECK-NEXT:    [[FDIV:%.*]] = fdiv float [[ARG0]], [[ARG1]]
 ; CHECK-NEXT:    ret float [[FDIV]]
@@ -62,7 +62,7 @@
 }
 
 define float @ret_fdiv_ieee_nonan__nonan_noinf_nozero(float nofpclass(nan) %arg0, float nofpclass(nan inf zero) %arg1) #0 {
-; CHECK-LABEL: define float @ret_fdiv_ieee_nonan__nonan_noinf_nozero
+; CHECK-LABEL: define nofpclass(nan) float @ret_fdiv_ieee_nonan__nonan_noinf_nozero
 ; CHECK-SAME: (float nofpclass(nan) [[ARG0:%.*]], float nofpclass(nan inf zero) [[ARG1:%.*]]) #[[ATTR0]] {
 ; CHECK-NEXT:    [[FDIV:%.*]] = fdiv float [[ARG0]], [[ARG1]]
 ; CHECK-NEXT:    ret float [[FDIV]]
@@ -72,7 +72,7 @@
 }
 
 define float @ret_fdiv_ieee_nonan_nozero__nonan_noinf(float nofpclass(nan zero) %arg0, float nofpclass(nan inf) %arg1) #0 {
-; CHECK-LABEL: define float @ret_fdiv_ieee_nonan_nozero__nonan_noinf
+; CHECK-LABEL: define nofpclass(nan) float @ret_fdiv_ieee_nonan_nozero__nonan_noinf
 ; CHECK-SAME: (float nofpclass(nan zero) [[ARG0:%.*]], float nofpclass(nan inf) [[ARG1:%.*]]) #[[ATTR0]] {
 ; CHECK-NEXT:    [[FDIV:%.*]] = fdiv float [[ARG0]], [[ARG1]]
 ; CHECK-NEXT:    ret float [[FDIV]]
@@ -82,7 +82,7 @@
 }
 
 define float @ret_fdiv_ieee_nonan_noinf__nonan_nozero(float nofpclass(nan inf) %arg0, float nofpclass(nan zero) %arg1) #0 {
-; CHECK-LABEL: define float @ret_fdiv_ieee_nonan_noinf__nonan_nozero
+; CHECK-LABEL: define nofpclass(nan) float @ret_fdiv_ieee_nonan_noinf__nonan_nozero
 ; CHECK-SAME: (float nofpclass(nan inf) [[ARG0:%.*]], float nofpclass(nan zero) [[ARG1:%.*]]) #[[ATTR0]] {
 ; CHECK-NEXT:    [[FDIV:%.*]] = fdiv float [[ARG0]], [[ARG1]]
 ; CHECK-NEXT:    ret float [[FDIV]]
@@ -112,7 +112,7 @@
 }
 
 define float @ret_fdiv_daz_nonan_nozero_nosub__nonan_noinf(float nofpclass(nan zero sub) %arg0, float nofpclass(nan inf) %arg1) #1 {
-; CHECK-LABEL: define float @ret_fdiv_daz_nonan_nozero_nosub__nonan_noinf
+; CHECK-LABEL: define nofpclass(nan) float @ret_fdiv_daz_nonan_nozero_nosub__nonan_noinf
 ; CHECK-SAME: (float nofpclass(nan zero sub) [[ARG0:%.*]], float nofpclass(nan inf) [[ARG1:%.*]]) #[[ATTR1]] {
 ; CHECK-NEXT:    [[FDIV:%.*]] = fdiv float [[ARG0]], [[ARG1]]
 ; CHECK-NEXT:    ret float [[FDIV]]
@@ -122,7 +122,7 @@
 }
 
 define float @ret_fdiv_daz_nonan_noinf__nonan_nozero_nosub(float nofpclass(nan inf) %arg0, float nofpclass(nan zero sub) %arg1) #1 {
-; CHECK-LABEL: define float @ret_fdiv_daz_nonan_noinf__nonan_nozero_nosub
+; CHECK-LABEL: define nofpclass(nan) float @ret_fdiv_daz_nonan_noinf__nonan_nozero_nosub
 ; CHECK-SAME: (float nofpclass(nan inf) [[ARG0:%.*]], float nofpclass(nan zero sub) [[ARG1:%.*]]) #[[ATTR1]] {
 ; CHECK-NEXT:    [[FDIV:%.*]] = fdiv float [[ARG0]], [[ARG1]]
 ; CHECK-NEXT:    ret float [[FDIV]]
@@ -142,7 +142,7 @@
 }
 
 define float @ret_fdiv_ieee_nonan_noinf_nozero__nonan_noinf_nozero(float nofpclass(nan inf zero) %arg0, float nofpclass(nan inf zero) %arg1) #0 {
-; CHECK-LABEL: define float @ret_fdiv_ieee_nonan_noinf_nozero__nonan_noinf_nozero
+; CHECK-LABEL: define nofpclass(nan) float @ret_fdiv_ieee_nonan_noinf_nozero__nonan_noinf_nozero
 ; CHECK-SAME: (float nofpclass(nan inf zero) [[ARG0:%.*]], float nofpclass(nan inf zero) [[ARG1:%.*]]) #[[ATTR0]] {
 ; CHECK-NEXT:    [[FDIV:%.*]] = fdiv float [[ARG0]], [[ARG1]]
 ; CHECK-NEXT:    ret float [[FDIV]]
@@ -192,7 +192,7 @@
 }
 
 define float @ret_fdiv_ieee_nonan_noinf_nozero__nonan_noinf_noinf(float nofpclass(nan inf zero) %arg0, float nofpclass(nan inf) %arg1) #0 {
-; CHECK-LABEL: define float @ret_fdiv_ieee_nonan_noinf_nozero__nonan_noinf_noinf
+; CHECK-LABEL: define nofpclass(nan) float @ret_fdiv_ieee_nonan_noinf_nozero__nonan_noinf_noinf
 ; CHECK-SAME: (float nofpclass(nan inf zero) [[ARG0:%.*]], float nofpclass(nan inf) [[ARG1:%.*]]) #[[ATTR0]] {
 ; CHECK-NEXT:    [[FDIV:%.*]] = fdiv float [[ARG0]], [[ARG1]]
 ; CHECK-NEXT:    ret float [[FDIV]]
@@ -202,7 +202,7 @@
 }
 
 define float @ret_fdiv_ieee_nonan_noinf_noinf__nonan_noinf_nozero(float nofpclass(nan inf) %arg0, float nofpclass(nan inf zero) %arg1) #0 {
-; CHECK-LABEL: define float @ret_fdiv_ieee_nonan_noinf_noinf__nonan_noinf_nozero
+; CHECK-LABEL: define nofpclass(nan) float @ret_fdiv_ieee_nonan_noinf_noinf__nonan_noinf_nozero
 ; CHECK-SAME: (float nofpclass(nan inf) [[ARG0:%.*]], float nofpclass(nan inf zero) [[ARG1:%.*]]) #[[ATTR0]] {
 ; CHECK-NEXT:    [[FDIV:%.*]] = fdiv float [[ARG0]], [[ARG1]]
 ; CHECK-NEXT:    ret float [[FDIV]]
@@ -232,7 +232,7 @@
 }
 
 define float @ret_fdiv_ieee_nonan_noinf_nozero_nosub__nonan_noinf_nozero_nosub(float nofpclass(nan inf zero sub) %arg0, float nofpclass(nan inf zero sub) %arg1) #0 {
-; CHECK-LABEL: define float @ret_fdiv_ieee_nonan_noinf_nozero_nosub__nonan_noinf_nozero_nosub
+; CHECK-LABEL: define nofpclass(nan) float @ret_fdiv_ieee_nonan_noinf_nozero_nosub__nonan_noinf_nozero_nosub
 ; CHECK-SAME: (float nofpclass(nan inf zero sub) [[ARG0:%.*]], float nofpclass(nan inf zero sub) [[ARG1:%.*]]) #[[ATTR0]] {
 ; CHECK-NEXT:    [[FDIV:%.*]] = fdiv float [[ARG0]], [[ARG1]]
 ; CHECK-NEXT:    ret float [[FDIV]]
@@ -243,7 +243,7 @@
 
 ; Denormal mode doesn't matter because sources are nofpclass(sub)
 define float @ret_fdiv_daz_nonan_noinf_nozero_nosub__nonan_noinf_nozero_nosub(float nofpclass(nan inf zero sub) %arg0, float nofpclass(nan inf zero sub) %arg1) #1 {
-; CHECK-LABEL: define float @ret_fdiv_daz_nonan_noinf_nozero_nosub__nonan_noinf_nozero_nosub
+; CHECK-LABEL: define nofpclass(nan) float @ret_fdiv_daz_nonan_noinf_nozero_nosub__nonan_noinf_nozero_nosub
 ; CHECK-SAME: (float nofpclass(nan inf zero sub) [[ARG0:%.*]], float nofpclass(nan inf zero sub) [[ARG1:%.*]]) #[[ATTR1]] {
 ; CHECK-NEXT:    [[FDIV:%.*]] = fdiv float [[ARG0]], [[ARG1]]
 ; CHECK-NEXT:    ret float [[FDIV]]
@@ -253,7 +253,7 @@
 }
 
 define float @ret_fdiv_dapz_nonan_noinf_nozero_nosub__nonan_noinf_nozero_nosub(float nofpclass(nan inf zero sub) %arg0, float nofpclass(nan inf zero sub) %arg1) #2 {
-; CHECK-LABEL: define float @ret_fdiv_dapz_nonan_noinf_nozero_nosub__nonan_noinf_nozero_nosub
+; CHECK-LABEL: define nofpclass(nan) float @ret_fdiv_dapz_nonan_noinf_nozero_nosub__nonan_noinf_nozero_nosub
 ; CHECK-SAME: (float nofpclass(nan inf zero sub) [[ARG0:%.*]], float nofpclass(nan inf zero sub) [[ARG1:%.*]]) #[[ATTR2]] {
 ; CHECK-NEXT:    [[FDIV:%.*]] = fdiv float [[ARG0]], [[ARG1]]
 ; CHECK-NEXT:    ret float [[FDIV]]
@@ -263,7 +263,7 @@
 }
 
 define float @ret_fdiv_dynamic_nonan_noinf_nozero_nosub__nonan_noinf_nozero_nosub(float nofpclass(nan inf zero sub) %arg0, float nofpclass(nan inf zero sub) %arg1) #3 {
-; CHECK-LABEL: define float @ret_fdiv_dynamic_nonan_noinf_nozero_nosub__nonan_noinf_nozero_nosub
+; CHECK-LABEL: define nofpclass(nan) float @ret_fdiv_dynamic_nonan_noinf_nozero_nosub__nonan_noinf_nozero_nosub
 ; CHECK-SAME: (float nofpclass(nan inf zero sub) [[ARG0:%.*]], float nofpclass(nan inf zero sub) [[ARG1:%.*]]) #[[ATTR3]] {
 ; CHECK-NEXT:    [[FDIV:%.*]] = fdiv float [[ARG0]], [[ARG1]]
 ; CHECK-NEXT:    ret float [[FDIV]]
@@ -274,7 +274,7 @@
 
 ; Missing no-subnormal on lhs
 define float @ret_fdiv_daz_nonan_noinf_nozero__nonan_noinf_nozero_nosub(float nofpclass(nan inf zero) %arg0, float nofpclass(nan inf zero sub) %arg1) #1 {
-; CHECK-LABEL: define float @ret_fdiv_daz_nonan_noinf_nozero__nonan_noinf_nozero_nosub
+; CHECK-LABEL: define nofpclass(nan) float @ret_fdiv_daz_nonan_noinf_nozero__nonan_noinf_nozero_nosub
 ; CHECK-SAME: (float nofpclass(nan inf zero) [[ARG0:%.*]], float nofpclass(nan inf zero sub) [[ARG1:%.*]]) #[[ATTR1]] {
 ; CHECK-NEXT:    [[FDIV:%.*]] = fdiv float [[ARG0]], [[ARG1]]
 ; CHECK-NEXT:    ret float [[FDIV]]
@@ -285,7 +285,7 @@
 
 ; Missing no-subnormal on lhs
 define float @ret_fdiv_daz_nonan_noinf_nozero_nosub__nonan_noinf_nozero(float nofpclass(nan inf zero sub) %arg0, float nofpclass(nan inf zero) %arg1) #1 {
-; CHECK-LABEL: define float @ret_fdiv_daz_nonan_noinf_nozero_nosub__nonan_noinf_nozero
+; CHECK-LABEL: define nofpclass(nan) float @ret_fdiv_daz_nonan_noinf_nozero_nosub__nonan_noinf_nozero
 ; CHECK-SAME: (float nofpclass(nan inf zero sub) [[ARG0:%.*]], float nofpclass(nan inf zero) [[ARG1:%.*]]) #[[ATTR1]] {
 ; CHECK-NEXT:    [[FDIV:%.*]] = fdiv float [[ARG0]], [[ARG1]]
 ; CHECK-NEXT:    ret float [[FDIV]]