[InstCombine] Calloc-ed strings optimizations
Summary:
Example cases:
strlen(calloc(...)) -> 0
Reviewers: efriedma, bkramer
Reviewed By: bkramer
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D47059
llvm-svn: 332990
diff --git a/llvm/lib/Analysis/MemoryBuiltins.cpp b/llvm/lib/Analysis/MemoryBuiltins.cpp
index deacdb9..988eacb 100644
--- a/llvm/lib/Analysis/MemoryBuiltins.cpp
+++ b/llvm/lib/Analysis/MemoryBuiltins.cpp
@@ -589,7 +589,7 @@
// Handle strdup-like functions separately.
if (FnData->AllocTy == StrDupLike) {
- APInt Size(IntTyBits, GetStringLength(CS.getArgument(0)));
+ APInt Size(IntTyBits, GetStringLength(CS.getArgument(0), TLI));
if (!Size)
return unknown();
diff --git a/llvm/lib/Analysis/ValueTracking.cpp b/llvm/lib/Analysis/ValueTracking.cpp
index 4442df8..95e7a9b 100644
--- a/llvm/lib/Analysis/ValueTracking.cpp
+++ b/llvm/lib/Analysis/ValueTracking.cpp
@@ -3372,10 +3372,39 @@
return NullIndex + 1;
}
+static bool isStringFromCalloc(const Value *Str, const TargetLibraryInfo *TLI) {
+ const CallInst *Calloc = dyn_cast<CallInst>(Str);
+ if (!Calloc)
+ return false;
+
+ const Function *InnerCallee = Calloc->getCalledFunction();
+ if (!InnerCallee)
+ return false;
+
+ LibFunc Func;
+ if (!TLI->getLibFunc(*InnerCallee, Func) || !TLI->has(Func) ||
+ Func != LibFunc_calloc)
+ return false;
+
+ const ConstantInt *N = dyn_cast<ConstantInt>(Calloc->getOperand(0));
+ const ConstantInt *Size = dyn_cast<ConstantInt>(Calloc->getOperand(1));
+
+ if (!N || !Size)
+ return false;
+
+ if (N->isNullValue() || Size->isNullValue())
+ return false;
+
+ return true;
+}
+
/// If we can compute the length of the string pointed to by
/// the specified pointer, return 'len+1'. If we can't, return 0.
-uint64_t llvm::GetStringLength(const Value *V, unsigned CharSize) {
- if (!V->getType()->isPointerTy()) return 0;
+uint64_t llvm::GetStringLength(const Value *V, const TargetLibraryInfo *TLI, unsigned CharSize) {
+ if (!V->getType()->isPointerTy())
+ return 0;
+ if (isStringFromCalloc(V, TLI))
+ return 1;
SmallPtrSet<const PHINode*, 32> PHIs;
uint64_t Len = GetStringLengthH(V, PHIs, CharSize);
diff --git a/llvm/lib/Transforms/Utils/SimplifyLibCalls.cpp b/llvm/lib/Transforms/Utils/SimplifyLibCalls.cpp
index d97eb05..49869ac 100644
--- a/llvm/lib/Transforms/Utils/SimplifyLibCalls.cpp
+++ b/llvm/lib/Transforms/Utils/SimplifyLibCalls.cpp
@@ -160,7 +160,7 @@
Value *Src = CI->getArgOperand(1);
// See if we can get the length of the input string.
- uint64_t Len = GetStringLength(Src);
+ uint64_t Len = GetStringLength(Src, TLI);
if (Len == 0)
return nullptr;
--Len; // Unbias length.
@@ -205,7 +205,7 @@
return nullptr;
// See if we can get the length of the input string.
- uint64_t SrcLen = GetStringLength(Src);
+ uint64_t SrcLen = GetStringLength(Src, TLI);
if (SrcLen == 0)
return nullptr;
--SrcLen; // Unbias length.
@@ -234,7 +234,7 @@
// of the input string and turn this into memchr.
ConstantInt *CharC = dyn_cast<ConstantInt>(CI->getArgOperand(1));
if (!CharC) {
- uint64_t Len = GetStringLength(SrcStr);
+ uint64_t Len = GetStringLength(SrcStr, TLI);
if (Len == 0 || !FT->getParamType(1)->isIntegerTy(32)) // memchr needs i32.
return nullptr;
@@ -313,8 +313,8 @@
return B.CreateZExt(B.CreateLoad(Str1P, "strcmpload"), CI->getType());
// strcmp(P, "x") -> memcmp(P, "x", 2)
- uint64_t Len1 = GetStringLength(Str1P);
- uint64_t Len2 = GetStringLength(Str2P);
+ uint64_t Len1 = GetStringLength(Str1P, TLI);
+ uint64_t Len2 = GetStringLength(Str2P, TLI);
if (Len1 && Len2) {
return emitMemCmp(Str1P, Str2P,
ConstantInt::get(DL.getIntPtrType(CI->getContext()),
@@ -370,7 +370,7 @@
return Src;
// See if we can get the length of the input string.
- uint64_t Len = GetStringLength(Src);
+ uint64_t Len = GetStringLength(Src, TLI);
if (Len == 0)
return nullptr;
@@ -390,7 +390,7 @@
}
// See if we can get the length of the input string.
- uint64_t Len = GetStringLength(Src);
+ uint64_t Len = GetStringLength(Src, TLI);
if (Len == 0)
return nullptr;
@@ -412,7 +412,7 @@
Value *LenOp = CI->getArgOperand(2);
// See if we can get the length of the input string.
- uint64_t SrcLen = GetStringLength(Src);
+ uint64_t SrcLen = GetStringLength(Src, TLI);
if (SrcLen == 0)
return nullptr;
--SrcLen;
@@ -448,7 +448,7 @@
Value *Src = CI->getArgOperand(0);
// Constant folding: strlen("xyz") -> 3
- if (uint64_t Len = GetStringLength(Src, CharSize))
+ if (uint64_t Len = GetStringLength(Src, TLI, CharSize))
return ConstantInt::get(CI->getType(), Len - 1);
// If s is a constant pointer pointing to a string literal, we can fold
@@ -512,8 +512,8 @@
// strlen(x?"foo":"bars") --> x ? 3 : 4
if (SelectInst *SI = dyn_cast<SelectInst>(Src)) {
- uint64_t LenTrue = GetStringLength(SI->getTrueValue(), CharSize);
- uint64_t LenFalse = GetStringLength(SI->getFalseValue(), CharSize);
+ uint64_t LenTrue = GetStringLength(SI->getTrueValue(), TLI, CharSize);
+ uint64_t LenFalse = GetStringLength(SI->getFalseValue(), TLI, CharSize);
if (LenTrue && LenFalse) {
ORE.emit([&]() {
return OptimizationRemark("instcombine", "simplify-libcalls", CI)
@@ -2142,7 +2142,7 @@
}
// fputs(s,F) --> fwrite(s,1,strlen(s),F)
- uint64_t Len = GetStringLength(CI->getArgOperand(0));
+ uint64_t Len = GetStringLength(CI->getArgOperand(0), TLI);
if (!Len)
return nullptr;
@@ -2565,7 +2565,7 @@
if (OnlyLowerUnknownSize)
return false;
if (isString) {
- uint64_t Len = GetStringLength(CI->getArgOperand(SizeOp));
+ uint64_t Len = GetStringLength(CI->getArgOperand(SizeOp), TLI);
// If the length is 0 we don't know how long it is and so we can't
// remove the check.
if (Len == 0)
@@ -2637,7 +2637,7 @@
return nullptr;
// Maybe we can stil fold __st[rp]cpy_chk to __memcpy_chk.
- uint64_t Len = GetStringLength(Src);
+ uint64_t Len = GetStringLength(Src, TLI);
if (Len == 0)
return nullptr;
@@ -2716,4 +2716,4 @@
FortifiedLibCallSimplifier::FortifiedLibCallSimplifier(
const TargetLibraryInfo *TLI, bool OnlyLowerUnknownSize)
- : TLI(TLI), OnlyLowerUnknownSize(OnlyLowerUnknownSize) {}
+ : TLI(TLI), OnlyLowerUnknownSize(OnlyLowerUnknownSize) {}
\ No newline at end of file