[ExpandMemCmp] Properly expand `bcmp` to an equality pattern.

[courbet]

Summary

Before that change, constant-size bcmp would miss an opportunity to generate
a more efficient equality pattern and would generate a -1/0-1 pattern
instead.

[courbet]

Rebase on base test.

[nikic]

LG

[courbet]

Update llc tests too.

llvm/lib/CodeGen/ExpandMemCmp.cpp

 ///  br label %endblock
 /// endblock:                                         ; preds = %res_block,
 /// %loadbb3
 ///  %phi.res = phi i32 [ %48, %loadbb3 ], [ %11, %res_block ]
 ///  ret i32 %phi.res
 static bool expandMemCmp(CallInst *CI, const TargetTransformInfo *TTI,
                          const TargetLowering *TLI, const DataLayout *DL,
                          ProfileSummaryInfo *PSI, BlockFrequencyInfo *BFI,
-                         DomTreeUpdater *DTU) {
+                         DomTreeUpdater *DTU, const bool IsBCmp) {
   NumMemCmpCalls++;
 
   // Early exit from expansion if -Oz.
   if (CI->getFunction()->hasMinSize())
     return false;
 
   // Early exit from expansion if size is not a constant.
   ConstantInt *SizeCast = dyn_cast<ConstantInt>(CI->getArgOperand(2));
   if (!SizeCast) {
     NumMemCmpNotConstant++;
     return false;
   }
   const uint64_t SizeVal = SizeCast->getZExtValue();
 
   if (SizeVal == 0) {
     return false;
   }
   // TTI call to check if target would like to expand memcmp. Also, get the
   // available load sizes.
-  const bool IsUsedForZeroCmp = isOnlyUsedInZeroEqualityComparison(CI);
+  const bool IsUsedForZeroCmp =
+      IsBCmp || isOnlyUsedInZeroEqualityComparison(CI);
   bool OptForSize = CI->getFunction()->hasOptSize() ||
                     llvm::shouldOptimizeForSize(CI->getParent(), PSI, BFI);
   auto Options = TTI->enableMemCmpExpansion(OptForSize,
                                             IsUsedForZeroCmp);
   if (!Options) return false;
 
   if (MemCmpEqZeroNumLoadsPerBlock.getNumOccurrences())
     Options.NumLoadsPerBlock = MemCmpEqZeroNumLoadsPerBlock;
bool ExpandMemCmpPass::runOnBlock(BasicBlock &BB, const TargetLibraryInfo *TLI,
   for (Instruction& I : BB) {
     CallInst *CI = dyn_cast<CallInst>(&I);
     if (!CI) {
       continue;
     }
     LibFunc Func;
     if (TLI->getLibFunc(*CI, Func) &&
         (Func == LibFunc_memcmp || Func == LibFunc_bcmp) &&
-        expandMemCmp(CI, TTI, TL, &DL, PSI, BFI, DTU)) {
+        expandMemCmp(CI, TTI, TL, &DL, PSI, BFI, DTU, Func == LibFunc_bcmp)) {
       return true;
     }
   }
   return false;
 }
 
 PreservedAnalyses
 ExpandMemCmpPass::runImpl(Function &F, const TargetLibraryInfo *TLI,

llvm/test/CodeGen/X86/memcmp-optsize-x32.ll

; X86-NEXT:    retl
   %m = tail call i32 @memcmp(i8* %X, i8* getelementptr inbounds ([65 x i8], [65 x i8]* @.str, i32 0, i32 0), i32 64) nounwind
   %c = icmp eq i32 %m, 0
   ret i1 %c
 }
 
 define i32 @bcmp_length2(i8* %X, i8* %Y) nounwind optsize {
 ; X86-LABEL: bcmp_length2:
 ; X86:       # %bb.0:
-; X86-NEXT:    movl {{[0-9]+}}(%esp), %eax
 ; X86-NEXT:    movl {{[0-9]+}}(%esp), %ecx
-; X86-NEXT:    movzwl (%ecx), %ecx
+; X86-NEXT:    movl {{[0-9]+}}(%esp), %eax
 ; X86-NEXT:    movzwl (%eax), %edx
-; X86-NEXT:    rolw $8, %cx
-; X86-NEXT:    rolw $8, %dx
-; X86-NEXT:    movzwl %cx, %eax
-; X86-NEXT:    movzwl %dx, %ecx
-; X86-NEXT:    subl %ecx, %eax
+; X86-NEXT:    xorl %eax, %eax
+; X86-NEXT:    cmpw (%ecx), %dx
+; X86-NEXT:    setne %al
 ; X86-NEXT:    retl
   %m = tail call i32 @bcmp(i8* %X, i8* %Y, i32 2) nounwind
   ret i32 %m
 }

llvm/test/CodeGen/X86/memcmp-optsize.ll

; X64-AVX2-NEXT:    retq
   %m = tail call i32 @memcmp(i8* %X, i8* getelementptr inbounds ([65 x i8], [65 x i8]* @.str, i32 0, i32 0), i64 64) nounwind
   %c = icmp eq i32 %m, 0
   ret i1 %c
 }
 
 define i32 @bcmp_length2(i8* %X, i8* %Y) nounwind optsize {
 ; X64-LABEL: bcmp_length2:
 ; X64:       # %bb.0:
-; X64-NEXT:    movzwl (%rdi), %eax
-; X64-NEXT:    movzwl (%rsi), %ecx
-; X64-NEXT:    rolw $8, %ax
-; X64-NEXT:    rolw $8, %cx
-; X64-NEXT:    movzwl %ax, %eax
-; X64-NEXT:    movzwl %cx, %ecx
-; X64-NEXT:    subl %ecx, %eax
+; X64-NEXT:    movzwl (%rdi), %ecx
+; X64-NEXT:    xorl %eax, %eax
+; X64-NEXT:    cmpw (%rsi), %cx
+; X64-NEXT:    setne %al
 ; X64-NEXT:    retq
   %m = tail call i32 @bcmp(i8* %X, i8* %Y, i64 2) nounwind
   ret i32 %m
 }

llvm/test/CodeGen/X86/memcmp-pgso-x32.ll

; X86-NEXT:    retl
   %m = tail call i32 @memcmp(i8* %X, i8* getelementptr inbounds ([65 x i8], [65 x i8]* @.str, i32 0, i32 0), i32 64) nounwind
   %c = icmp eq i32 %m, 0
   ret i1 %c
 }
 
 define i32 @bcmp_length2(i8* %X, i8* %Y) nounwind !prof !14 {
 ; X86-LABEL: bcmp_length2:
 ; X86:       # %bb.0:
-; X86-NEXT:    movl {{[0-9]+}}(%esp), %eax
 ; X86-NEXT:    movl {{[0-9]+}}(%esp), %ecx
-; X86-NEXT:    movzwl (%ecx), %ecx
+; X86-NEXT:    movl {{[0-9]+}}(%esp), %eax
 ; X86-NEXT:    movzwl (%eax), %edx
-; X86-NEXT:    rolw $8, %cx
-; X86-NEXT:    rolw $8, %dx
-; X86-NEXT:    movzwl %cx, %eax
-; X86-NEXT:    movzwl %dx, %ecx
-; X86-NEXT:    subl %ecx, %eax
+; X86-NEXT:    xorl %eax, %eax
+; X86-NEXT:    cmpw (%ecx), %dx
+; X86-NEXT:    setne %al
 ; X86-NEXT:    retl
   %m = tail call i32 @bcmp(i8* %X, i8* %Y, i32 2) nounwind
   ret i32 %m
 }
 
 !llvm.module.flags = !{!0}
 !0 = !{i32 1, !"ProfileSummary", !1}
 !1 = !{!2, !3, !4, !5, !6, !7, !8, !9}

llvm/test/CodeGen/X86/memcmp-pgso.ll

; X64-AVX2-NEXT:    retq
   %m = tail call i32 @memcmp(i8* %X, i8* getelementptr inbounds ([65 x i8], [65 x i8]* @.str, i32 0, i32 0), i64 64) nounwind
   %c = icmp eq i32 %m, 0
   ret i1 %c
 }
 
 define i32 @bcmp_length2(i8* %X, i8* %Y) nounwind !prof !14 {
 ; X64-LABEL: bcmp_length2:
 ; X64:       # %bb.0:
-; X64-NEXT:    movzwl (%rdi), %eax
-; X64-NEXT:    movzwl (%rsi), %ecx
-; X64-NEXT:    rolw $8, %ax
-; X64-NEXT:    rolw $8, %cx
-; X64-NEXT:    movzwl %ax, %eax
-; X64-NEXT:    movzwl %cx, %ecx
-; X64-NEXT:    subl %ecx, %eax
+; X64-NEXT:    movzwl (%rdi), %ecx
+; X64-NEXT:    xorl %eax, %eax
+; X64-NEXT:    cmpw (%rsi), %cx
+; X64-NEXT:    setne %al
 ; X64-NEXT:    retq
   %m = tail call i32 @bcmp(i8* %X, i8* %Y, i64 2) nounwind
   ret i32 %m
 }
 
 !llvm.module.flags = !{!0}
 !0 = !{i32 1, !"ProfileSummary", !1}
 !1 = !{!2, !3, !4, !5, !6, !7, !8, !9}

llvm/test/Transforms/ExpandMemCmp/X86/bcmp.ll

 ; NOTE: Assertions have been autogenerated by utils/update_test_checks.py
 ; RUN: opt -S -expandmemcmp -memcmp-num-loads-per-block=1 -mtriple=x86_64-unknown-unknown -data-layout=e-m:o-i64:64-f80:128-n8:16:32:64-S128         < %s | FileCheck %s --check-prefix=X64
 
 declare i32 @bcmp(i8* nocapture, i8* nocapture, i64)
 
 define i32 @bcmp8(i8* nocapture readonly %x, i8* nocapture readonly %y)  {
 ; X64-LABEL: @bcmp8(
 ; X64-NEXT:    [[TMP1:%.*]] = bitcast i8* [[X:%.*]] to i64*
 ; X64-NEXT:    [[TMP2:%.*]] = bitcast i8* [[Y:%.*]] to i64*
 ; X64-NEXT:    [[TMP3:%.*]] = load i64, i64* [[TMP1]], align 1
 ; X64-NEXT:    [[TMP4:%.*]] = load i64, i64* [[TMP2]], align 1
-; X64-NEXT:    [[TMP5:%.*]] = call i64 @llvm.bswap.i64(i64 [[TMP3]])
-; X64-NEXT:    [[TMP6:%.*]] = call i64 @llvm.bswap.i64(i64 [[TMP4]])
-; X64-NEXT:    [[TMP7:%.*]] = icmp ugt i64 [[TMP5]], [[TMP6]]
-; X64-NEXT:    [[TMP8:%.*]] = icmp ult i64 [[TMP5]], [[TMP6]]
-; X64-NEXT:    [[TMP9:%.*]] = zext i1 [[TMP7]] to i32
-; X64-NEXT:    [[TMP10:%.*]] = zext i1 [[TMP8]] to i32
-; X64-NEXT:    [[TMP11:%.*]] = sub i32 [[TMP9]], [[TMP10]]
-; X64-NEXT:    ret i32 [[TMP11]]
+; X64-NEXT:    [[TMP5:%.*]] = icmp ne i64 [[TMP3]], [[TMP4]]
+; X64-NEXT:    [[TMP6:%.*]] = zext i1 [[TMP5]] to i32
+; X64-NEXT:    ret i32 [[TMP6]]
 ;
   %call = tail call i32 @bcmp(i8* %x, i8* %y, i64 8)
   ret i32 %call
 }