Add options for MaxLoadsPerMemcmp(OptSize).

[davidxl]

lgtm

llvm/trunk/lib/CodeGen/ExpandMemCmp.cpp

STATISTIC(NumMemCmpGreaterThanMax,
           "Number of memcmp calls with size greater than max size");
 STATISTIC(NumMemCmpInlined, "Number of inlined memcmp calls");
 
 static cl::opt<unsigned> MemCmpEqZeroNumLoadsPerBlock(
     "memcmp-num-loads-per-block", cl::Hidden, cl::init(1),
     cl::desc("The number of loads per basic block for inline expansion of "
              "memcmp that is only being compared against zero."));
 
+static cl::opt<unsigned> MaxLoadsPerMemcmp(
+    "max-loads-per-memcmp", cl::Hidden,
+    cl::desc("Set maximum number of loads used in expanded memcmp"));
+
+static cl::opt<unsigned> MaxLoadsPerMemcmpOptSize(
+    "max-loads-per-memcmp-opt-size", cl::Hidden,
+    cl::desc("Set maximum number of loads used in expanded memcmp for -Os/Oz"));
+
 namespace {
 
 
 // This class provides helper functions to expand a memcmp library call into an
 // inline expansion.
 class MemCmpExpansion {
   struct ResultBlock {
     BasicBlock *BB = nullptr;
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 auto *const Options = TTI->enableMemCmpExpansion(IsUsedForZeroCmp);
   if (!Options) return false;
 
-  const unsigned MaxNumLoads =
-      TLI->getMaxExpandSizeMemcmp(CI->getFunction()->hasOptSize());
+  const unsigned MaxNumLoads = CI->getFunction()->hasOptSize()
+      ? (MaxLoadsPerMemcmpOptSize.getNumOccurrences()
+         ? MaxLoadsPerMemcmpOptSize
+         : TLI->getMaxExpandSizeMemcmp(true))
+      : (MaxLoadsPerMemcmp.getNumOccurrences()
+         ? MaxLoadsPerMemcmp
+         : TLI->getMaxExpandSizeMemcmp(false));
 
   unsigned NumLoadsPerBlock = MemCmpEqZeroNumLoadsPerBlock.getNumOccurrences()
                                   ? MemCmpEqZeroNumLoadsPerBlock
                                   : TLI->getMemcmpEqZeroLoadsPerBlock();
 
   MemCmpExpansion Expansion(CI, SizeVal, *Options, MaxNumLoads,
                             IsUsedForZeroCmp, NumLoadsPerBlock, *DL);