[profile] Static counter allocation for value profiling (part-1)

Differential Revision: http://reviews.llvm.org/D20459

llvm-svn: 270336

include/llvm/ProfileData/InstrProf.h

@@ -58,6 +58,20 @@ inline StringRef getInstrProfDataSectionName(bool AddSegment) {
                     : INSTR_PROF_DATA_SECT_NAME_STR;
 }
 
+/// Return the name of data section containing pointers to value profile
+/// counters/nodes.
+inline StringRef getInstrProfValuesSectionName(bool AddSegment) {
+  return AddSegment ? "__DATA," INSTR_PROF_VALS_SECT_NAME_STR
+                    : INSTR_PROF_VALS_SECT_NAME_STR;
+}
+
+/// Return the name of data section containing nodes holdling value
+/// profiling data.
+inline StringRef getInstrProfVNodesSectionName(bool AddSegment) {
+  return AddSegment ? "__DATA," INSTR_PROF_VNODES_SECT_NAME_STR
+                    : INSTR_PROF_VNODES_SECT_NAME_STR;
+}
+
 /// Return the name profile runtime entry point to do value profiling
 /// for a given site.
 inline StringRef getInstrProfValueProfFuncName() {
@@ -80,6 +94,12 @@ inline StringRef getInstrProfDataVarPrefix() { return "__profd_"; }
 /// Return the name prefix of profile counter variables.
 inline StringRef getInstrProfCountersVarPrefix() { return "__profc_"; }
 
+/// Return the name prefix of value profile variables.
+inline StringRef getInstrProfValuesVarPrefix() { return "__profvp_"; }
+
+/// Return the name of value profile node array variables:
+inline StringRef getInstrProfVNodesVarName() { return "__llvm_prf_vnodes"; }
+
 /// Return the name prefix of the COMDAT group for instrumentation variables
 /// associated with a COMDAT function.
 inline StringRef getInstrProfComdatPrefix() { return "__profv_"; }

include/llvm/ProfileData/InstrProfData.inc

@@ -80,7 +80,7 @@ INSTR_PROF_DATA(const IntPtrT, llvm::Type::getInt64PtrTy(Ctx), CounterPtr, \
 INSTR_PROF_DATA(const IntPtrT, llvm::Type::getInt8PtrTy(Ctx), FunctionPointer, \
                 FunctionAddr)
 INSTR_PROF_DATA(IntPtrT, llvm::Type::getInt8PtrTy(Ctx), Values, \
-                ConstantPointerNull::get(Int8PtrTy))
+                ValuesPtrExpr)
 INSTR_PROF_DATA(const uint32_t, llvm::Type::getInt32Ty(Ctx), NumCounters, \
                 ConstantInt::get(llvm::Type::getInt32Ty(Ctx), NumCounters))
 INSTR_PROF_DATA(const uint16_t, Int16ArrayTy, NumValueSites[IPVK_Last+1], \
@@ -597,6 +597,12 @@ ValueProfData *serializeValueProfDataFrom(ValueProfRecordClosure *Closure,
 #define INSTR_PROF_DATA_SECT_NAME __llvm_prf_data
 #define INSTR_PROF_NAME_SECT_NAME __llvm_prf_names
 #define INSTR_PROF_CNTS_SECT_NAME __llvm_prf_cnts
+/* Array of pointers. Each pointer points to a list
+ * of value nodes associated with one value site.
+ */
+#define INSTR_PROF_VALS_SECT_NAME __llvm_prf_values
+/* Value profile nodes section. */
+#define INSTR_PROF_VNODES_SECT_NAME __llvm_prf_vnodes
 #define INSTR_PROF_COVMAP_SECT_NAME __llvm_covmap
 
 #define INSTR_PROF_DATA_SECT_NAME_STR                                          \
@@ -607,6 +613,10 @@ ValueProfData *serializeValueProfDataFrom(ValueProfRecordClosure *Closure,
   INSTR_PROF_QUOTE(INSTR_PROF_CNTS_SECT_NAME)
 #define INSTR_PROF_COVMAP_SECT_NAME_STR                                        \
   INSTR_PROF_QUOTE(INSTR_PROF_COVMAP_SECT_NAME)
+#define INSTR_PROF_VALS_SECT_NAME_STR                                          \
+  INSTR_PROF_QUOTE(INSTR_PROF_VALS_SECT_NAME)
+#define INSTR_PROF_VNODES_SECT_NAME_STR                                        \
+  INSTR_PROF_QUOTE(INSTR_PROF_VNODES_SECT_NAME)
 
 /* Macros to define start/stop section symbol for a given
  * section on Linux. For instance

include/llvm/Transforms/InstrProfiling.h

@@ -84,6 +84,9 @@ private:
   /// Emit the section with compressed function names.
   void emitNameData();
 
+  /// Emit value nodes section for value profiling.
+  void emitVNodes();
+
   /// Emit runtime registration functions for each profile data variable.
   void emitRegistration();
 

lib/Transforms/Instrumentation/InstrProfiling.cpp

@@ -31,6 +31,20 @@ cl::opt<bool> DoNameCompression("enable-name-compression",
                                 cl::desc("Enable name string compression"),
                                 cl::init(true));
 
+cl::opt<bool> ValueProfileStaticAlloc(
+    "vp-static-alloc",
+    cl::desc("Do static counter allocation for value profiler"),
+    cl::init(true));
+cl::opt<double> NumCountersPerValueSite(
+    "vp-counters-per-site",
+    cl::desc("The average number of profile counters allocated "
+             "per value profiling site."),
+    // This is set to a very small value because in real programs, only
+    // a very small percentage of value sites have non-zero targets, e.g, 1/30.
+    // For those sites with non-zero profile, the average number of targets
+    // is usually smaller than 2.
+    cl::init(1.0));
+
 class InstrProfilingLegacyPass : public ModulePass {
   InstrProfiling InstrProf;
 
@@ -141,6 +155,7 @@ bool InstrProfiling::run(Module &M) {
   if (!MadeChange)
     return false;
 
+  emitVNodes();
   emitNameData();
   emitRegistration();
   emitRuntimeHook();
@@ -288,6 +303,20 @@ static inline Comdat *getOrCreateProfileComdat(Module &M, Function &F,
   return M.getOrInsertComdat(StringRef(getVarName(Inc, ComdatPrefix)));
 }
 
+static bool needsRuntimeRegistrationOfSectionRange(const Module &M) {
+  // Don't do this for Darwin.  compiler-rt uses linker magic.
+  if (Triple(M.getTargetTriple()).isOSDarwin())
+    return false;
+
+  // Use linker script magic to get data/cnts/name start/end.
+  if (Triple(M.getTargetTriple()).isOSLinux() ||
+      Triple(M.getTargetTriple()).isOSFreeBSD() ||
+      Triple(M.getTargetTriple()).isPS4CPU())
+    return false;
+
+  return true;
+}
+
 GlobalVariable *
 InstrProfiling::getOrCreateRegionCounters(InstrProfIncrementInst *Inc) {
   GlobalVariable *NamePtr = Inc->getName();
@@ -321,10 +350,34 @@ InstrProfiling::getOrCreateRegionCounters(InstrProfIncrementInst *Inc) {
   CounterPtr->setAlignment(8);
   CounterPtr->setComdat(ProfileVarsComdat);
 
-  // Create data variable.
   auto *Int8PtrTy = Type::getInt8PtrTy(Ctx);
+  // Allocate statically the array of pointers to value profile nodes for
+  // the current function.
+  Constant *ValuesPtrExpr = ConstantPointerNull::get(Int8PtrTy);
+  if (ValueProfileStaticAlloc && !needsRuntimeRegistrationOfSectionRange(*M)) {
+
+    uint64_t NS = 0;
+    for (uint32_t Kind = IPVK_First; Kind <= IPVK_Last; ++Kind)
+      NS += PD.NumValueSites[Kind];
+    if (NS) {
+      ArrayType *ValuesTy = ArrayType::get(Type::getInt64Ty(Ctx), NS);
+
+      auto *ValuesVar =
+          new GlobalVariable(*M, ValuesTy, false, NamePtr->getLinkage(),
+                             Constant::getNullValue(ValuesTy),
+                             getVarName(Inc, getInstrProfValuesVarPrefix()));
+      ValuesVar->setVisibility(NamePtr->getVisibility());
+      ValuesVar->setSection(getInstrProfValuesSectionName(isMachO()));
+      ValuesVar->setAlignment(8);
+      ValuesVar->setComdat(ProfileVarsComdat);
+      ValuesPtrExpr =
+          ConstantExpr::getBitCast(ValuesVar, llvm::Type::getInt8PtrTy(Ctx));
+    }
+  }
+
+  // Create data variable.
   auto *Int16Ty = Type::getInt16Ty(Ctx);
-  auto *Int16ArrayTy = ArrayType::get(Int16Ty, IPVK_Last+1);
+  auto *Int16ArrayTy = ArrayType::get(Int16Ty, IPVK_Last + 1);
   Type *DataTypes[] = {
     #define INSTR_PROF_DATA(Type, LLVMType, Name, Init) LLVMType,
     #include "llvm/ProfileData/InstrProfData.inc"
@@ -367,18 +420,49 @@ InstrProfiling::getOrCreateRegionCounters(InstrProfIncrementInst *Inc) {
   return CounterPtr;
 }
 
-static bool needsRuntimeRegistrationOfSectionRange(const Module &M) {
-  // Don't do this for Darwin.  compiler-rt uses linker magic.
-  if (Triple(M.getTargetTriple()).isOSDarwin())
-    return false;
+void InstrProfiling::emitVNodes() {
+  if (!ValueProfileStaticAlloc)
+    return;
 
-  // Use linker script magic to get data/cnts/name start/end.
-  if (Triple(M.getTargetTriple()).isOSLinux() ||
-      Triple(M.getTargetTriple()).isOSFreeBSD() ||
-      Triple(M.getTargetTriple()).isPS4CPU())
-    return false;
+  // For now only support this on platforms that do
+  // not require runtime registration to discover
+  // named section start/end.
+  if (needsRuntimeRegistrationOfSectionRange(*M))
+    return;
 
-  return true;
+  size_t TotalNS = 0;
+  for (auto &PD : ProfileDataMap) {
+    for (uint32_t Kind = IPVK_First; Kind <= IPVK_Last; ++Kind)
+      TotalNS += PD.second.NumValueSites[Kind];
+  }
+
+  if (!TotalNS)
+    return;
+
+  uint64_t NumCounters = TotalNS * NumCountersPerValueSite;
+  // Heuristic for small programs with very few total value sites.
+  // The default value of vp-counters-per-site is chosen based on
+  // the observation that large apps usually have a low percentage
+  // of value sites that actually have any profile data, and thus
+  // the average number of counters per site is low. For small
+  // apps with very few sites, this may not be true. Bump up the
+  // number of counters in this case.
+  if (NumCounters < 10)
+    NumCounters *= 2;
+
+  auto &Ctx = M->getContext();
+  Type *VNodeTypes[] = {
+#define INSTR_PROF_VALUE_NODE(Type, LLVMType, Name, Init) LLVMType,
+#include "llvm/ProfileData/InstrProfData.inc"
+  };
+  auto *VNodeTy = StructType::get(Ctx, makeArrayRef(VNodeTypes));
+
+  ArrayType *VNodesTy = ArrayType::get(VNodeTy, NumCounters);
+  auto *VNodesVar = new GlobalVariable(
+      *M, VNodesTy, false, llvm::GlobalValue::PrivateLinkage,
+      Constant::getNullValue(VNodesTy), getInstrProfVNodesVarName());
+  VNodesVar->setSection(getInstrProfVNodesSectionName(isMachO()));
+  UsedVars.push_back(VNodesVar);
 }
 
 void InstrProfiling::emitNameData() {

test/Instrumentation/InstrProfiling/icall.ll

@@ -0,0 +1,31 @@
+;; Check that static counters are allocated for value profiler
+
+; RUN: opt < %s -mtriple=x86_64-apple-macosx10.10.0 -vp-static-alloc=true -instrprof -S | FileCheck %s --check-prefix=STATIC
+; RUN: opt < %s -mtriple=x86_64-unknown-linux -instrprof -vp-static-alloc=true -S | FileCheck %s --check-prefix=STATIC
+; RUN: opt < %s -mtriple=x86_64-apple-macosx10.10.0 -vp-static-alloc=false -instrprof -S | FileCheck %s --check-prefix=DYN
+; RUN: opt < %s -mtriple=x86_64-unknown-linux -instrprof -vp-static-alloc=false -S | FileCheck %s --check-prefix=DYN
+
+
+@__profn_foo = private constant [3 x i8] c"foo"
+
+define i32 @foo(i32 ()* ) {
+  call void @llvm.instrprof.increment(i8* getelementptr inbounds ([3 x i8], [3 x i8]* @__profn_foo, i32 0, i32 0), i64 12884901887, i32 1, i32 0)
+  %2 = ptrtoint i32 ()* %0 to i64
+  call void @llvm.instrprof.value.profile(i8* getelementptr inbounds ([3 x i8], [3 x i8]* @__profn_foo, i32 0, i32 0), i64 12884901887, i64 %2, i32 0, i32 0)
+  %3 = tail call i32 %0()
+  ret i32 %3
+}
+
+; Function Attrs: nounwind
+declare void @llvm.instrprof.increment(i8*, i64, i32, i32) #0
+
+; Function Attrs: nounwind
+declare void @llvm.instrprof.value.profile(i8*, i64, i64, i32, i32) #0
+
+attributes #0 = { nounwind }
+
+; STATIC: @__profvp_foo
+; STATIC: @__llvm_prf_vnodes
+
+; DYN-NOT: @__profvp_foo
+; DYN-NOT: @__llvm_prf_vnodes