ValueTracking: Remove CannotBeOrderedLessThanZero

Replace the last user of CannotBeOrderedLessThanZero with new
version. Makes assumes work in this case.

llvm/lib/Transforms/AggressiveInstCombine/AggressiveInstCombine.cpp

@@ -400,7 +400,8 @@ static bool tryToFPToSat(Instruction &I, TargetTransformInfo &TTI) {
 /// pessimistic codegen that has to account for setting errno and can enable
 /// vectorization.
 static bool foldSqrt(Instruction &I, TargetTransformInfo &TTI,
-                     TargetLibraryInfo &TLI) {
+                     TargetLibraryInfo &TLI, AssumptionCache &AC,
+                     DominatorTree &DT) {
   // Match a call to sqrt mathlib function.
   auto *Call = dyn_cast<CallInst>(&I);
   if (!Call)
@@ -424,7 +425,8 @@ static bool foldSqrt(Instruction &I, TargetTransformInfo &TTI,
   Value *Arg = Call->getArgOperand(0);
   if (TTI.haveFastSqrt(Ty) &&
       (Call->hasNoNaNs() ||
-       CannotBeOrderedLessThanZero(Arg, M->getDataLayout(), &TLI))) {
+       cannotBeOrderedLessThanZero(Arg, M->getDataLayout(), &TLI, 0, &AC, &I,
+                                   &DT))) {
     IRBuilder<> Builder(&I);
     IRBuilderBase::FastMathFlagGuard Guard(Builder);
     Builder.setFastMathFlags(Call->getFastMathFlags());
@@ -918,7 +920,8 @@ static bool foldPatternedLoads(Instruction &I, const DataLayout &DL) {
 /// occur frequently and/or have more than a constant-length pattern match.
 static bool foldUnusualPatterns(Function &F, DominatorTree &DT,
                                 TargetTransformInfo &TTI,
-                                TargetLibraryInfo &TLI, AliasAnalysis &AA) {
+                                TargetLibraryInfo &TLI, AliasAnalysis &AA,
+                                AssumptionCache &AC) {
   bool MadeChange = false;
   for (BasicBlock &BB : F) {
     // Ignore unreachable basic blocks.
@@ -943,7 +946,7 @@ static bool foldUnusualPatterns(Function &F, DominatorTree &DT,
       // NOTE: This function introduces erasing of the instruction `I`, so it
       // needs to be called at the end of this sequence, otherwise we may make
       // bugs.
-      MadeChange |= foldSqrt(I, TTI, TLI);
+      MadeChange |= foldSqrt(I, TTI, TLI, AC, DT);
     }
   }
 
@@ -964,7 +967,7 @@ static bool runImpl(Function &F, AssumptionCache &AC, TargetTransformInfo &TTI,
   const DataLayout &DL = F.getParent()->getDataLayout();
   TruncInstCombine TIC(AC, TLI, DL, DT);
   MadeChange |= TIC.run(F);
-  MadeChange |= foldUnusualPatterns(F, DT, TTI, TLI, AA);
+  MadeChange |= foldUnusualPatterns(F, DT, TTI, TLI, AA, AC);
   return MadeChange;
 }
 

llvm/test/Transforms/AggressiveInstCombine/X86/sqrt.ll

@@ -5,6 +5,7 @@ declare float @sqrtf(float)
 declare double @sqrt(double)
 declare fp128 @sqrtl(fp128)
 declare float @llvm.fabs.f32(float)
+declare void @llvm.assume(i1 noundef)
 
 ; "nnan" implies no setting of errno and the target can lower this to an
 ; instruction, so transform to an intrinsic.
@@ -46,7 +47,7 @@ define fp128 @sqrt_call_nnan_f128(fp128 %x) {
 
 define float @sqrt_call_nnan_f32_nobuiltin(float %x) {
 ; CHECK-LABEL: @sqrt_call_nnan_f32_nobuiltin(
-; CHECK-NEXT:    [[SQRT:%.*]] = call nnan float @sqrtf(float [[X:%.*]]) #[[ATTR1:[0-9]+]]
+; CHECK-NEXT:    [[SQRT:%.*]] = call nnan float @sqrtf(float [[X:%.*]]) #[[ATTR2:[0-9]+]]
 ; CHECK-NEXT:    ret float [[SQRT]]
 ;
   %sqrt = call nnan float @sqrtf(float %x) nobuiltin
@@ -74,3 +75,16 @@ define float @sqrt_call_f32_fabs(float %x) {
   %sqrt = call float @sqrtf(float %a)
   ret float %sqrt
 }
+
+define float @sqrt_call_f32_assume_oge_n0(float %x) {
+; CHECK-LABEL: @sqrt_call_f32_assume_oge_n0(
+; CHECK-NEXT:    [[IS_POS:%.*]] = fcmp oge float [[X:%.*]], -0.000000e+00
+; CHECK-NEXT:    call void @llvm.assume(i1 [[IS_POS]])
+; CHECK-NEXT:    [[SQRT1:%.*]] = call float @llvm.sqrt.f32(float [[X]])
+; CHECK-NEXT:    ret float [[SQRT1]]
+;
+  %is.pos = fcmp oge float %x, -0.0
+  call void @llvm.assume(i1 %is.pos)
+  %sqrt = call float @sqrtf(float %x)
+  ret float %sqrt
+}