CodeGen: Include a function hash in instrumentation based profiling
The hash itself is still the number of counters, which isn't all that
useful, but this separates the API changes from the actual
implementation of the hash and will make it easier to transition to
the ProfileData library once it's implemented.
llvm-svn: 204186
diff --git a/clang/lib/CodeGen/CodeGenPGO.cpp b/clang/lib/CodeGen/CodeGenPGO.cpp
index 207cb66..1f2fe38 100644
--- a/clang/lib/CodeGen/CodeGenPGO.cpp
+++ b/clang/lib/CodeGen/CodeGenPGO.cpp
@@ -58,10 +58,15 @@
ReportBadPGOData(CGM, "pgo data file has malformed function entry");
return;
}
- while (*--CurPtr != ' ')
- ;
StringRef FuncName(FuncStart, CurPtr - FuncStart);
+ // Skip over the function hash.
+ CurPtr = strchr(++CurPtr, '\n');
+ if (!CurPtr) {
+ ReportBadPGOData(CGM, "pgo data file is missing the function hash");
+ return;
+ }
+
// Read the number of counters.
char *EndPtr;
unsigned NumCounters = strtol(++CurPtr, &EndPtr, 10);
@@ -99,7 +104,7 @@
MaxFunctionCount = MaxCount;
}
-bool PGOProfileData::getFunctionCounts(StringRef FuncName,
+bool PGOProfileData::getFunctionCounts(StringRef FuncName, uint64_t &FuncHash,
std::vector<uint64_t> &Counts) {
// Find the relevant section of the pgo-data file.
llvm::StringMap<unsigned>::const_iterator OffsetIter =
@@ -111,11 +116,15 @@
// Skip over the function name.
CurPtr = strchr(CurPtr, '\n');
assert(CurPtr && "pgo-data has corrupted function entry");
- while (*--CurPtr != ' ')
- ;
+
+ char *EndPtr;
+ // Read the function hash.
+ FuncHash = strtoll(++CurPtr, &EndPtr, 10);
+ assert(EndPtr != CurPtr && *EndPtr == '\n' &&
+ "pgo-data file has corrupted function hash");
+ CurPtr = EndPtr;
// Read the number of counters.
- char *EndPtr;
unsigned NumCounters = strtol(++CurPtr, &EndPtr, 10);
assert(EndPtr != CurPtr && *EndPtr == '\n' && NumCounters > 0 &&
"pgo-data file has corrupted number of counters");
@@ -246,15 +255,17 @@
// Create data variable.
auto *Int32Ty = llvm::Type::getInt32Ty(Ctx);
+ auto *Int64Ty = llvm::Type::getInt64Ty(Ctx);
auto *Int8PtrTy = llvm::Type::getInt8PtrTy(Ctx);
auto *Int64PtrTy = llvm::Type::getInt64PtrTy(Ctx);
llvm::Type *DataTypes[] = {
- Int32Ty, Int32Ty, Int8PtrTy, Int64PtrTy
+ Int32Ty, Int32Ty, Int64Ty, Int8PtrTy, Int64PtrTy
};
auto *DataTy = llvm::StructType::get(Ctx, makeArrayRef(DataTypes));
llvm::Constant *DataVals[] = {
llvm::ConstantInt::get(Int32Ty, getFuncName().size()),
llvm::ConstantInt::get(Int32Ty, NumRegionCounters),
+ llvm::ConstantInt::get(Int64Ty, FunctionHash),
llvm::ConstantExpr::getBitCast(Name, Int8PtrTy),
llvm::ConstantExpr::getBitCast(RegionCounters, Int64PtrTy)
};
@@ -847,6 +858,8 @@
else if (const BlockDecl *BD = dyn_cast_or_null<BlockDecl>(D))
Walker.VisitBlockDecl(BD);
NumRegionCounters = Walker.NextCounter;
+ // FIXME: The number of counters isn't sufficient for the hash
+ FunctionHash = NumRegionCounters;
}
void CodeGenPGO::computeRegionCounts(const Decl *D) {
@@ -905,8 +918,9 @@
// ignore counts when the input changes in various ways, e.g., by comparing a
// hash value based on some characteristics of the input.
RegionCounts = new std::vector<uint64_t>();
- if (PGOData->getFunctionCounts(getFuncName(), *RegionCounts) ||
- RegionCounts->size() != NumRegionCounters) {
+ uint64_t Hash;
+ if (PGOData->getFunctionCounts(getFuncName(), Hash, *RegionCounts) ||
+ Hash != FunctionHash || RegionCounts->size() != NumRegionCounters) {
delete RegionCounts;
RegionCounts = 0;
}