[SimplifyLibCalls] Add dereferenceable bytes from known callsites

Summary:
int mm(char *a, char *b) {
    return memcmp(a,b,16);
}

Currently:
define dso_local i32 @mm(i8* nocapture readonly %a, i8* nocapture readonly %b) local_unnamed_addr #1 {
entry:
  %call = tail call i32 @memcmp(i8* %a, i8* %b, i64 16)
  ret i32 %call
}

After patch:
define dso_local i32 @mm(i8* nocapture readonly %a, i8* nocapture readonly %b) local_unnamed_addr #1 {
entry:
  %call = tail call i32 @memcmp(i8* dereferenceable(16)  %a, i8* dereferenceable(16)  %b, i64 16)
  ret i32 %call
}




Reviewers: jdoerfert, efriedma

Reviewed By: jdoerfert

Subscribers: javed.absar, spatel, llvm-commits

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D66079

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@368657 91177308-0d34-0410-b5e6-96231b3b80d8

Author: davidbolvansky

include/llvm/Transforms/Utils/SimplifyLibCalls.h

@@ -167,9 +167,9 @@ class LibCallSimplifier {
   Value *optimizeMemCmp(CallInst *CI, IRBuilder<> &B);
   Value *optimizeBCmp(CallInst *CI, IRBuilder<> &B);
   Value *optimizeMemCmpBCmpCommon(CallInst *CI, IRBuilder<> &B);
-  Value *optimizeMemCpy(CallInst *CI, IRBuilder<> &B);
-  Value *optimizeMemMove(CallInst *CI, IRBuilder<> &B);
-  Value *optimizeMemSet(CallInst *CI, IRBuilder<> &B);
+  Value *optimizeMemCpy(CallInst *CI, IRBuilder<> &B, bool isIntrinsic = false);
+  Value *optimizeMemMove(CallInst *CI, IRBuilder<> &B, bool isIntrinsic = false);
+  Value *optimizeMemSet(CallInst *CI, IRBuilder<> &B, bool isIntrinsic = false);
   Value *optimizeRealloc(CallInst *CI, IRBuilder<> &B);
   Value *optimizeWcslen(CallInst *CI, IRBuilder<> &B);
   // Wrapper for all String/Memory Library Call Optimizations

lib/Transforms/Utils/SimplifyLibCalls.cpp

@@ -186,6 +186,20 @@ static bool canTransformToMemCmp(CallInst *CI, Value *Str, uint64_t Len,
   return true;
 }
 
+static void annotateDereferenceableBytes(CallInst *CI,
+                                         ArrayRef<unsigned> ArgNos,
+                                         uint64_t DerefBytes) {
+  for (unsigned ArgNo : ArgNos) {
+    if (CI->getDereferenceableBytes(ArgNo + AttributeList::FirstArgIndex) <
+        DerefBytes) {
+      CI->removeParamAttr(ArgNo, Attribute::Dereferenceable);
+      CI->removeParamAttr(ArgNo, Attribute::DereferenceableOrNull);
+      CI->addParamAttr(ArgNo, Attribute::getWithDereferenceableBytes(
+                                  CI->getContext(), DerefBytes));
+    }
+  }
+}
+
 //===----------------------------------------------------------------------===//
 // String and Memory Library Call Optimizations
 //===----------------------------------------------------------------------===//
@@ -765,9 +779,11 @@ Value *LibCallSimplifier::optimizeMemChr(CallInst *CI, IRBuilder<> &B) {
   ConstantInt *LenC = dyn_cast<ConstantInt>(CI->getArgOperand(2));
 
   // memchr(x, y, 0) -> null
-  if (LenC && LenC->isZero())
-    return Constant::getNullValue(CI->getType());
-
+  if (LenC) {
+    if (LenC->isZero())
+      return Constant::getNullValue(CI->getType());
+    annotateDereferenceableBytes(CI, {0}, LenC->getZExtValue());
+  }
   // From now on we need at least constant length and string.
   StringRef Str;
   if (!LenC || !getConstantStringInfo(SrcStr, Str, 0, /*TrimAtNul=*/false))
@@ -926,10 +942,12 @@ Value *LibCallSimplifier::optimizeMemCmpBCmpCommon(CallInst *CI,
     return Constant::getNullValue(CI->getType());
 
   // Handle constant lengths.
-  if (ConstantInt *LenC = dyn_cast<ConstantInt>(Size))
+  if (ConstantInt *LenC = dyn_cast<ConstantInt>(Size)) {
     if (Value *Res = optimizeMemCmpConstantSize(CI, LHS, RHS,
                                                 LenC->getZExtValue(), B, DL))
       return Res;
+    annotateDereferenceableBytes(CI, {0, 1}, LenC->getZExtValue());
+  }
 
   return nullptr;
 }
@@ -955,18 +973,31 @@ Value *LibCallSimplifier::optimizeBCmp(CallInst *CI, IRBuilder<> &B) {
   return optimizeMemCmpBCmpCommon(CI, B);
 }
 
-Value *LibCallSimplifier::optimizeMemCpy(CallInst *CI, IRBuilder<> &B) {
+Value *LibCallSimplifier::optimizeMemCpy(CallInst *CI, IRBuilder<> &B,
+                                         bool isIntrinsic) {
+  Value *Size = CI->getArgOperand(2);
+  if (ConstantInt *LenC = dyn_cast<ConstantInt>(Size))
+    annotateDereferenceableBytes(CI, {0, 1}, LenC->getZExtValue());
+
+  if (isIntrinsic)
+    return nullptr;
+
   // memcpy(x, y, n) -> llvm.memcpy(align 1 x, align 1 y, n)
-  B.CreateMemCpy(CI->getArgOperand(0), 1, CI->getArgOperand(1), 1,
-                 CI->getArgOperand(2));
+  B.CreateMemCpy(CI->getArgOperand(0), 1, CI->getArgOperand(1), 1, Size);
   return CI->getArgOperand(0);
 }
 
-Value *LibCallSimplifier::optimizeMemMove(CallInst *CI, IRBuilder<> &B) {
+Value *LibCallSimplifier::optimizeMemMove(CallInst *CI, IRBuilder<> &B, bool isIntrinsic) {
+  Value *Size = CI->getArgOperand(2);
+  if (ConstantInt *LenC = dyn_cast<ConstantInt>(Size))
+    annotateDereferenceableBytes(CI, {0, 1}, LenC->getZExtValue());
+
+  if (isIntrinsic)
+    return nullptr;
+
   // memmove(x, y, n) -> llvm.memmove(align 1 x, align 1 y, n)
-  B.CreateMemMove(CI->getArgOperand(0), 1, CI->getArgOperand(1), 1,
-                  CI->getArgOperand(2));
-  return CI->getArgOperand(0);
+  B.CreateMemMove( CI->getArgOperand(0), 1, CI->getArgOperand(1), 1, Size);
+  return  CI->getArgOperand(0);
 }
 
 /// Fold memset[_chk](malloc(n), 0, n) --> calloc(1, n).
@@ -1015,13 +1046,21 @@ Value *LibCallSimplifier::foldMallocMemset(CallInst *Memset, IRBuilder<> &B) {
   return Calloc;
 }
 
-Value *LibCallSimplifier::optimizeMemSet(CallInst *CI, IRBuilder<> &B) {
+Value *LibCallSimplifier::optimizeMemSet(CallInst *CI, IRBuilder<> &B,
+                                         bool isIntrinsic) {
+  Value *Size = CI->getArgOperand(2);
+  if (ConstantInt *LenC = dyn_cast<ConstantInt>(Size))
+    annotateDereferenceableBytes(CI, {0}, LenC->getZExtValue());
+
+  if (isIntrinsic)
+    return nullptr;
+
   if (auto *Calloc = foldMallocMemset(CI, B))
     return Calloc;
 
   // memset(p, v, n) -> llvm.memset(align 1 p, v, n)
   Value *Val = B.CreateIntCast(CI->getArgOperand(1), B.getInt8Ty(), false);
-  B.CreateMemSet(CI->getArgOperand(0), Val, CI->getArgOperand(2), 1);
+  B.CreateMemSet(CI->getArgOperand(0), Val, Size, 1);
   return CI->getArgOperand(0);
 }
 
@@ -2710,6 +2749,12 @@ Value *LibCallSimplifier::optimizeCall(CallInst *CI) {
     case Intrinsic::sqrt:
       return optimizeSqrt(CI, Builder);
     // TODO: Use foldMallocMemset() with memset intrinsic.
+    case Intrinsic::memset:
+      return optimizeMemSet(CI, Builder, true);
+    case Intrinsic::memcpy:
+      return optimizeMemCpy(CI, Builder, true);
+    case Intrinsic::memmove:
+      return optimizeMemMove(CI, Builder, true);
     default:
       return nullptr;
     }

test/Analysis/TypeBasedAliasAnalysis/memcpyopt.ll

@@ -6,7 +6,7 @@ target datalayout = "e-p:64:64:64"
 ; it has a TBAA tag which declares that it is unrelated.
 
 ; CHECK: @foo
-; CHECK-NEXT: tail call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 1 %p, i8* align 1 %q, i64 16, i1 false), !tbaa !0
+; CHECK-NEXT: tail call void @llvm.memcpy.p0i8.p0i8.i64(i8* align 1 dereferenceable(16) %p, i8* align 1 dereferenceable(16) %q, i64 16, i1 false), !tbaa !0
 ; CHECK-NEXT: store i8 2, i8* %s, align 1, !tbaa [[TAGA:!.*]]
 ; CHECK-NEXT: ret void
 define void @foo(i8* nocapture %p, i8* nocapture %q, i8* nocapture %s) nounwind {

test/Transforms/InstCombine/ARM/strcmp.ll

@@ -1,3 +1,4 @@
+; NOTE: Assertions have been autogenerated by utils/update_test_checks.py
 ; Test that the strcmp library call simplifier works correctly.
 ; RUN: opt < %s -instcombine -S | FileCheck %s
 
@@ -13,10 +14,11 @@ declare i32 @strcmp(i8*, i8*)
 ; strcmp("", x) -> -*x
 define arm_aapcscc i32 @test1(i8* %str2) {
 ; CHECK-LABEL: @test1(
-; CHECK: %strcmpload = load i8, i8* %str
-; CHECK: %1 = zext i8 %strcmpload to i32
-; CHECK: %2 = sub nsw i32 0, %1
-; CHECK: ret i32 %2
+; CHECK-NEXT:    [[STRCMPLOAD:%.*]] = load i8, i8* [[STR2:%.*]], align 1
+; CHECK-NEXT:    [[TMP1:%.*]] = zext i8 [[STRCMPLOAD]] to i32
+; CHECK-NEXT:    [[TMP2:%.*]] = sub nsw i32 0, [[TMP1]]
+; CHECK-NEXT:    ret i32 [[TMP2]]
+;
 
   %str1 = getelementptr inbounds [1 x i8], [1 x i8]* @null, i32 0, i32 0
   %temp1 = call arm_apcscc i32 @strcmp(i8* %str1, i8* %str2)
@@ -27,9 +29,10 @@ define arm_aapcscc i32 @test1(i8* %str2) {
 ; strcmp(x, "") -> *x
 define arm_aapcscc i32 @test2(i8* %str1) {
 ; CHECK-LABEL: @test2(
-; CHECK: %strcmpload = load i8, i8* %str
-; CHECK: %1 = zext i8 %strcmpload to i32
-; CHECK: ret i32 %1
+; CHECK-NEXT:    [[STRCMPLOAD:%.*]] = load i8, i8* [[STR1:%.*]], align 1
+; CHECK-NEXT:    [[TMP1:%.*]] = zext i8 [[STRCMPLOAD]] to i32
+; CHECK-NEXT:    ret i32 [[TMP1]]
+;
 
   %str2 = getelementptr inbounds [1 x i8], [1 x i8]* @null, i32 0, i32 0
   %temp1 = call arm_aapcscc i32 @strcmp(i8* %str1, i8* %str2)
@@ -39,7 +42,8 @@ define arm_aapcscc i32 @test2(i8* %str1) {
 ; strcmp(x, y)  -> cnst
 define arm_aapcscc i32 @test3() {
 ; CHECK-LABEL: @test3(
-; CHECK: ret i32 -1
+; CHECK-NEXT:    ret i32 -1
+;
 
   %str1 = getelementptr inbounds [5 x i8], [5 x i8]* @hell, i32 0, i32 0
   %str2 = getelementptr inbounds [6 x i8], [6 x i8]* @hello, i32 0, i32 0
@@ -49,7 +53,8 @@ define arm_aapcscc i32 @test3() {
 
 define arm_aapcscc i32 @test4() {
 ; CHECK-LABEL: @test4(
-; CHECK: ret i32 1
+; CHECK-NEXT:    ret i32 1
+;
 
   %str1 = getelementptr inbounds [5 x i8], [5 x i8]* @hell, i32 0, i32 0
   %str2 = getelementptr inbounds [1 x i8], [1 x i8]* @null, i32 0, i32 0
@@ -61,8 +66,10 @@ define arm_aapcscc i32 @test4() {
 ; (This transform is rather difficult to trigger in a useful manner)
 define arm_aapcscc i32 @test5(i1 %b) {
 ; CHECK-LABEL: @test5(
-; CHECK: %memcmp = call i32 @memcmp(i8* getelementptr inbounds ([6 x i8], [6 x i8]* @hello, i32 0, i32 0), i8* %str2, i32 5)
-; CHECK: ret i32 %memcmp
+; CHECK-NEXT:    [[STR2:%.*]] = select i1 [[B:%.*]], i8* getelementptr inbounds ([5 x i8], [5 x i8]* @hell, i32 0, i32 0), i8* getelementptr inbounds ([5 x i8], [5 x i8]* @bell, i32 0, i32 0)
+; CHECK-NEXT:    [[MEMCMP:%.*]] = call i32 @memcmp(i8* dereferenceable(5) getelementptr inbounds ([6 x i8], [6 x i8]* @hello, i32 0, i32 0), i8* dereferenceable(5) [[STR2]], i32 5)
+; CHECK-NEXT:    ret i32 [[MEMCMP]]
+;
 
   %str1 = getelementptr inbounds [6 x i8], [6 x i8]* @hello, i32 0, i32 0
   %temp1 = getelementptr inbounds [5 x i8], [5 x i8]* @hell, i32 0, i32 0
@@ -75,7 +82,8 @@ define arm_aapcscc i32 @test5(i1 %b) {
 ; strcmp(x,x)  -> 0
 define arm_aapcscc i32 @test6(i8* %str) {
 ; CHECK-LABEL: @test6(
-; CHECK: ret i32 0
+; CHECK-NEXT:    ret i32 0
+;
 
   %temp1 = call arm_aapcscc i32 @strcmp(i8* %str, i8* %str)
   ret i32 %temp1
@@ -84,10 +92,11 @@ define arm_aapcscc i32 @test6(i8* %str) {
 ; strcmp("", x) -> -*x
 define arm_aapcs_vfpcc i32 @test1_vfp(i8* %str2) {
 ; CHECK-LABEL: @test1_vfp(
-; CHECK: %strcmpload = load i8, i8* %str
-; CHECK: %1 = zext i8 %strcmpload to i32
-; CHECK: %2 = sub nsw i32 0, %1
-; CHECK: ret i32 %2
+; CHECK-NEXT:    [[STRCMPLOAD:%.*]] = load i8, i8* [[STR2:%.*]], align 1
+; CHECK-NEXT:    [[TMP1:%.*]] = zext i8 [[STRCMPLOAD]] to i32
+; CHECK-NEXT:    [[TMP2:%.*]] = sub nsw i32 0, [[TMP1]]
+; CHECK-NEXT:    ret i32 [[TMP2]]
+;
 
   %str1 = getelementptr inbounds [1 x i8], [1 x i8]* @null, i32 0, i32 0
   %temp1 = call arm_aapcs_vfpcc i32 @strcmp(i8* %str1, i8* %str2)
@@ -98,9 +107,10 @@ define arm_aapcs_vfpcc i32 @test1_vfp(i8* %str2) {
 ; strcmp(x, "") -> *x
 define arm_aapcs_vfpcc i32 @test2_vfp(i8* %str1) {
 ; CHECK-LABEL: @test2_vfp(
-; CHECK: %strcmpload = load i8, i8* %str
-; CHECK: %1 = zext i8 %strcmpload to i32
-; CHECK: ret i32 %1
+; CHECK-NEXT:    [[STRCMPLOAD:%.*]] = load i8, i8* [[STR1:%.*]], align 1
+; CHECK-NEXT:    [[TMP1:%.*]] = zext i8 [[STRCMPLOAD]] to i32
+; CHECK-NEXT:    ret i32 [[TMP1]]
+;
 
   %str2 = getelementptr inbounds [1 x i8], [1 x i8]* @null, i32 0, i32 0
   %temp1 = call arm_aapcs_vfpcc i32 @strcmp(i8* %str1, i8* %str2)
@@ -110,7 +120,8 @@ define arm_aapcs_vfpcc i32 @test2_vfp(i8* %str1) {
 ; strcmp(x, y)  -> cnst
 define arm_aapcs_vfpcc i32 @test3_vfp() {
 ; CHECK-LABEL: @test3_vfp(
-; CHECK: ret i32 -1
+; CHECK-NEXT:    ret i32 -1
+;
 
   %str1 = getelementptr inbounds [5 x i8], [5 x i8]* @hell, i32 0, i32 0
   %str2 = getelementptr inbounds [6 x i8], [6 x i8]* @hello, i32 0, i32 0
@@ -120,7 +131,8 @@ define arm_aapcs_vfpcc i32 @test3_vfp() {
 
 define arm_aapcs_vfpcc i32 @test4_vfp() {
 ; CHECK-LABEL: @test4_vfp(
-; CHECK: ret i32 1
+; CHECK-NEXT:    ret i32 1
+;
 
   %str1 = getelementptr inbounds [5 x i8], [5 x i8]* @hell, i32 0, i32 0
   %str2 = getelementptr inbounds [1 x i8], [1 x i8]* @null, i32 0, i32 0
@@ -132,8 +144,10 @@ define arm_aapcs_vfpcc i32 @test4_vfp() {
 ; (This transform is rather difficult to trigger in a useful manner)
 define arm_aapcs_vfpcc i32 @test5_vfp(i1 %b) {
 ; CHECK-LABEL: @test5_vfp(
-; CHECK: %memcmp = call i32 @memcmp(i8* getelementptr inbounds ([6 x i8], [6 x i8]* @hello, i32 0, i32 0), i8* %str2, i32 5)
-; CHECK: ret i32 %memcmp
+; CHECK-NEXT:    [[STR2:%.*]] = select i1 [[B:%.*]], i8* getelementptr inbounds ([5 x i8], [5 x i8]* @hell, i32 0, i32 0), i8* getelementptr inbounds ([5 x i8], [5 x i8]* @bell, i32 0, i32 0)
+; CHECK-NEXT:    [[MEMCMP:%.*]] = call i32 @memcmp(i8* dereferenceable(5) getelementptr inbounds ([6 x i8], [6 x i8]* @hello, i32 0, i32 0), i8* dereferenceable(5) [[STR2]], i32 5)
+; CHECK-NEXT:    ret i32 [[MEMCMP]]
+;
 
   %str1 = getelementptr inbounds [6 x i8], [6 x i8]* @hello, i32 0, i32 0
   %temp1 = getelementptr inbounds [5 x i8], [5 x i8]* @hell, i32 0, i32 0
@@ -146,7 +160,8 @@ define arm_aapcs_vfpcc i32 @test5_vfp(i1 %b) {
 ; strcmp(x,x)  -> 0
 define arm_aapcs_vfpcc i32 @test6_vfp(i8* %str) {
 ; CHECK-LABEL: @test6_vfp(
-; CHECK: ret i32 0
+; CHECK-NEXT:    ret i32 0
+;
 
   %temp1 = call arm_aapcs_vfpcc i32 @strcmp(i8* %str, i8* %str)
   ret i32 %temp1