[llvm] [AggressiveInstCombine] Inline strcmp/strncmp (PR #89371)

[Nikita Popov via llvm-commits]

================
@@ -922,6 +924,239 @@ static bool foldPatternedLoads(Instruction &I, const DataLayout &DL) {
   return true;
 }
 
+static cl::opt<unsigned> StrNCmpInlineThreshold(
+    "strncmp-inline-threshold", cl::init(3), cl::Hidden,
+    cl::desc("The maximum length of a constant string for a builtin string cmp "
+             "call eligible for inlining. The default value is 3."));
+
+namespace {
+class StrNCmpInliner {
+public:
+  StrNCmpInliner(CallInst *CI, LibFunc Func, Function::iterator &BBNext,
+                 DomTreeUpdater *DTU, const DataLayout &DL)
+      : CI(CI), Func(Func), BBNext(BBNext), DTU(DTU), DL(DL) {}
+
+  bool optimizeStrNCmp();
+
+private:
+  bool inlineCompare(Value *LHS, StringRef RHS, uint64_t N, bool Switched);
+
+  CallInst *CI;
+  LibFunc Func;
+  Function::iterator &BBNext;
+  DomTreeUpdater *DTU;
+  const DataLayout &DL;
+};
+
+} // namespace
+
+/// First we normalize calls to strncmp/strcmp to the form of
+/// compare(s1, s2, N), which means comparing first N bytes of s1 and s2
+/// (without considering '\0')
+///
+/// Examples:
+///
+/// \code
+///   strncmp(s, "a", 3) -> compare(s, "a", 2)
+///   strncmp(s, "abc", 3) -> compare(s, "abc", 3)
+///   strncmp(s, "a\0b", 3) -> compare(s, "a\0b", 2)
+///   strcmp(s, "a") -> compare(s, "a", 2)
+///
+///   char s2[] = {'a'}
+///   strncmp(s, s2, 3) -> compare(s, s2, 3)
+///
+///   char s2[] = {'a', 'b', 'c', 'd'}
+///   strncmp(s, s2, 3) -> compare(s, s2, 3)
+/// \endcode
+///
+/// We only handle cases that N and exactly one of s1 and s2 are constant. Cases
+/// that s1 and s2 are both constant are already handled by the instcombine
+/// pass.
+///
+/// We do not handle cases that N > StrNCmpInlineThreshold.
+///
+/// We also do not handles cases that N < 2, which are already
+/// handled by the instcombine pass.
+///
+bool StrNCmpInliner::optimizeStrNCmp() {
+  if (StrNCmpInlineThreshold < 2)
+    return false;
+
+  Value *Str1P = CI->getArgOperand(0);
+  Value *Str2P = CI->getArgOperand(1);
+  // should be handled elsewhere
+  if (Str1P == Str2P)
+    return false;
+
+  StringRef Str1, Str2;
+  bool HasStr1 = getConstantStringInfo(Str1P, Str1, false);
+  bool HasStr2 = getConstantStringInfo(Str2P, Str2, false);
+  if (!(HasStr1 ^ HasStr2))
+    return false;
+
+  // note that '\0' and characters after it are not trimmed
+  StringRef Str = HasStr1 ? Str1 : Str2;
+
+  size_t Idx = Str.find('\0');
----------------
nikic wrote:

Why do we need to manually handle \0 rather than passing true to TrimAtNul above? Is it for the case where Str does not have a terminator?

https://github.com/llvm/llvm-project/pull/89371