maxjhandsome
/
llvm-project
forked from OSchip/llvm-project
1
0
Fork 0
Code Issues Pull Requests Packages Releases Wiki Activity

[llvm-profgen] Refactor PerfReader to allow different types of perf scripts 

In order to support different types of perf scripts, this change tried to refactor `PerfReader` by adding the base class `PerfReaderBase` and current HybridPerfReader is derived from it for CS profile generation. Common functions like, passMM2PEvents, extract_lbrs, extract_callstack, etc. can be reused.

Next step is to add LBR only reader(for non-CS profile) and aggregated perf scripts reader(do a pre-aggregation of scripts).

Reviewed By: hoy, wenlei

Differential Revision: https://reviews.llvm.org/D107014
This commit is contained in:
wlei 2021-07-28 15:07:40 -07:00
parent 9205143f07
commit 6da9241aab
4 changed files with 96 additions and 56 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

79
llvm/tools/llvm-profgen/PerfReader.cpp
View File

@ -241,7 +241,7 @@ bool VirtualUnwinder::unwind(const HybridSample *Sample, uint64_t Repeat) {
return true;
}
void PerfReader::validateCommandLine(
void PerfReaderBase::validateCommandLine(
cl::list<std::string> &BinaryFilenames,
cl::list<std::string> &PerfTraceFilenames) {
// Allow the invalid perfscript if we only use to show binary disassembly
@ -276,16 +276,33 @@ void PerfReader::validateCommandLine(
}
}
PerfReader::PerfReader(cl::list<std::string> &BinaryFilenames,
std::unique_ptr<PerfReaderBase>
PerfReaderBase::create(cl::list<std::string> &BinaryFilenames,
cl::list<std::string> &PerfTraceFilenames) {
validateCommandLine(BinaryFilenames, PerfTraceFilenames);
PerfScriptType PerfType = extractPerfType(PerfTraceFilenames);
std::unique_ptr<PerfReaderBase> PerfReader;
if (PerfType == PERF_LBR_STACK) {
PerfReader.reset(new HybridPerfReader(BinaryFilenames));
} else if (PerfType == PERF_LBR) {
// TODO:
exitWithError("Unsupported perfscript!");
} else {
exitWithError("Unsupported perfscript!");
}
return PerfReader;
}
PerfReaderBase::PerfReaderBase(cl::list<std::string> &BinaryFilenames) {
// Load the binaries.
for (auto Filename : BinaryFilenames)
loadBinary(Filename, /*AllowNameConflict*/ false);
}
ProfiledBinary &PerfReader::loadBinary(const StringRef BinaryPath,
bool AllowNameConflict) {
ProfiledBinary &PerfReaderBase::loadBinary(const StringRef BinaryPath,
bool AllowNameConflict) {
// The binary table is currently indexed by the binary name not the full
// binary path. This is because the user-given path may not match the one
// that was actually executed.
@ -303,7 +320,7 @@ ProfiledBinary &PerfReader::loadBinary(const StringRef BinaryPath,
return Ret.first->second;
}
void PerfReader::updateBinaryAddress(const MMapEvent &Event) {
void PerfReaderBase::updateBinaryAddress(const MMapEvent &Event) {
// Load the binary.
StringRef BinaryPath = Event.BinaryPath;
StringRef BinaryName = llvm::sys::path::filename(BinaryPath);
@ -353,7 +370,7 @@ void PerfReader::updateBinaryAddress(const MMapEvent &Event) {
}
}
ProfiledBinary *PerfReader::getBinary(uint64_t Address) {
ProfiledBinary *PerfReaderBase::getBinary(uint64_t Address) {
auto Iter = AddrToBinaryMap.lower_bound(Address);
if (Iter == AddrToBinaryMap.end() || Iter->first != Address) {
if (Iter == AddrToBinaryMap.begin())
@ -415,7 +432,7 @@ static void printBranchCounter(ContextSampleCounterMap &Counter,
printSampleCounter(OrderedCounter);
}
void PerfReader::printUnwinderOutput() {
void HybridPerfReader::printUnwinderOutput() {
for (auto I : BinarySampleCounters) {
const ProfiledBinary *Binary = I.first;
outs() << "Binary(" << Binary->getName().str() << ")'s Range Counter:\n";
@ -425,7 +442,7 @@ void PerfReader::printUnwinderOutput() {
}
}
void PerfReader::unwindSamples() {
void HybridPerfReader::unwindSamples() {
for (const auto &Item : AggregatedSamples) {
const HybridSample *Sample = dyn_cast<HybridSample>(Item.first.getPtr());
VirtualUnwinder Unwinder(&BinarySampleCounters[Sample->Binary],
@ -437,9 +454,9 @@ void PerfReader::unwindSamples() {
printUnwinderOutput();
}
bool PerfReader::extractLBRStack(TraceStream &TraceIt,
SmallVectorImpl<LBREntry> &LBRStack,
ProfiledBinary *Binary) {
bool PerfReaderBase::extractLBRStack(TraceStream &TraceIt,
SmallVectorImpl<LBREntry> &LBRStack,
ProfiledBinary *Binary) {
// The raw format of LBR stack is like:
// 0x4005c8/0x4005dc/P/-/-/0 0x40062f/0x4005b0/P/-/-/0 ...
// ... 0x4005c8/0x4005dc/P/-/-/0
@ -531,8 +548,8 @@ bool PerfReader::extractLBRStack(TraceStream &TraceIt,
return !LBRStack.empty();
}
bool PerfReader::extractCallstack(TraceStream &TraceIt,
SmallVectorImpl<uint64_t> &CallStack) {
bool PerfReaderBase::extractCallstack(TraceStream &TraceIt,
SmallVectorImpl<uint64_t> &CallStack) {
// The raw format of call stack is like:
// 4005dc # leaf frame
// 400634
@ -593,7 +610,7 @@ bool PerfReader::extractCallstack(TraceStream &TraceIt,
!Binary->addressInPrologEpilog(CallStack.front());
}
void PerfReader::parseHybridSample(TraceStream &TraceIt) {
void HybridPerfReader::parseSample(TraceStream &TraceIt) {
// The raw hybird sample started with call stack in FILO order and followed
// intermediately by LBR sample
// e.g.
@ -631,7 +648,7 @@ void PerfReader::parseHybridSample(TraceStream &TraceIt) {
}
}
void PerfReader::parseMMap2Event(TraceStream &TraceIt) {
void PerfReaderBase::parseMMap2Event(TraceStream &TraceIt) {
// Parse a line like:
// PERF_RECORD_MMAP2 2113428/2113428: [0x7fd4efb57000(0x204000) @ 0
// 08:04 19532229 3585508847]: r-xp /usr/lib64/libdl-2.17.so
@ -677,26 +694,23 @@ void PerfReader::parseMMap2Event(TraceStream &TraceIt) {
TraceIt.advance();
}
void PerfReader::parseEventOrSample(TraceStream &TraceIt) {
void PerfReaderBase::parseEventOrSample(TraceStream &TraceIt) {
if (TraceIt.getCurrentLine().startswith("PERF_RECORD_MMAP2"))
parseMMap2Event(TraceIt);
else if (getPerfScriptType() == PERF_LBR_STACK)
parseHybridSample(TraceIt);
else {
// TODO: parse other type sample
TraceIt.advance();
}
else
parseSample(TraceIt);
}
void PerfReader::parseAndAggregateTrace(StringRef Filename) {
void PerfReaderBase::parseAndAggregateTrace(StringRef Filename) {
// Trace line iterator
TraceStream TraceIt(Filename);
while (!TraceIt.isAtEoF())
parseEventOrSample(TraceIt);
}
void PerfReader::checkAndSetPerfType(
cl::list<std::string> &PerfTraceFilenames) {
PerfScriptType
PerfReaderBase::extractPerfType(cl::list<std::string> &PerfTraceFilenames) {
PerfScriptType PerfType = PERF_UNKNOWN;
for (auto FileName : PerfTraceFilenames) {
PerfScriptType Type = checkPerfScriptType(FileName);
if (Type == PERF_INVALID)
@ -705,20 +719,13 @@ void PerfReader::checkAndSetPerfType(
exitWithError("Inconsistent sample among different perf scripts");
PerfType = Type;
}
return PerfType;
}
void PerfReader::generateRawProfile() {
if (getPerfScriptType() == PERF_LBR_STACK) {
// Unwind samples if it's hybird sample
unwindSamples();
} else if (getPerfScriptType() == PERF_LBR) {
// TODO: range overlap computation for regular AutoFDO
}
}
void HybridPerfReader::generateRawProfile() { unwindSamples(); }
void PerfReader::parsePerfTraces(cl::list<std::string> &PerfTraceFilenames) {
// Check and set current perfscript type
checkAndSetPerfType(PerfTraceFilenames);
void PerfReaderBase::parsePerfTraces(
cl::list<std::string> &PerfTraceFilenames) {
// Parse perf traces and do aggregation.
for (auto Filename : PerfTraceFilenames)
parseAndAggregateTrace(Filename);

54
llvm/tools/llvm-profgen/PerfReader.h
View File

@ -541,11 +541,13 @@ using BinarySampleCounterMap =
std::unordered_map<ProfiledBinary *, ContextSampleCounterMap>;
// Load binaries and read perf trace to parse the events and samples
class PerfReader {
class PerfReaderBase {
public:
PerfReader(cl::list<std::string> &BinaryFilenames,
cl::list<std::string> &PerfTraceFilenames);
PerfReaderBase(cl::list<std::string> &BinaryFilenames);
virtual ~PerfReaderBase() = default;
static std::unique_ptr<PerfReaderBase>
create(cl::list<std::string> &BinaryFilenames,
cl::list<std::string> &PerfTraceFilenames);
// A LBR sample is like:
// 0x5c6313f/0x5c63170/P/-/-/0 0x5c630e7/0x5c63130/P/-/-/0 ...
@ -614,10 +616,12 @@ public:
return BinarySampleCounters;
}
private:
protected:
/// Validate the command line input
void validateCommandLine(cl::list<std::string> &BinaryFilenames,
cl::list<std::string> &PerfTraceFilenames);
static void validateCommandLine(cl::list<std::string> &BinaryFilenames,
cl::list<std::string> &PerfTraceFilenames);
static PerfScriptType
extractPerfType(cl::list<std::string> &PerfTraceFilenames);
/// Parse a single line of a PERF_RECORD_MMAP2 event looking for a
/// mapping between the binary name and its memory layout.
///
@ -626,8 +630,6 @@ private:
void parseAndAggregateTrace(StringRef Filename);
// Parse either an MMAP event or a perf sample
void parseEventOrSample(TraceStream &TraceIt);
// Parse the hybrid sample including the call and LBR line
void parseHybridSample(TraceStream &TraceIt);
// Extract call stack from the perf trace lines
bool extractCallstack(TraceStream &TraceIt,
SmallVectorImpl<uint64_t> &CallStack);
@ -635,13 +637,12 @@ private:
bool extractLBRStack(TraceStream &TraceIt,
SmallVectorImpl<LBREntry> &LBRStack,
ProfiledBinary *Binary);
void checkAndSetPerfType(cl::list<std::string> &PerfTraceFilenames);
// Parse one sample from multiple perf lines, override this for different
// sample type
virtual void parseSample(TraceStream &TraceIt) = 0;
// Post process the profile after trace aggregation, we will do simple range
// overlap computation for AutoFDO, or unwind for CSSPGO(hybrid sample).
void generateRawProfile();
// Unwind the hybrid samples after aggregration
void unwindSamples();
void printUnwinderOutput();
virtual void generateRawProfile() = 0;
// Helper function for looking up binary in AddressBinaryMap
ProfiledBinary *getBinary(uint64_t Address);
@ -654,6 +655,31 @@ private:
PerfScriptType PerfType = PERF_UNKNOWN;
};
/*
Hybrid perf script includes a group of hybrid samples(LBRs + call stack),
which is used to generate CS profile. An example of hybrid sample:
4005dc # call stack leaf
400634
400684 # call stack root
0x4005c8/0x4005dc/P/-/-/0 0x40062f/0x4005b0/P/-/-/0 ...
... 0x4005c8/0x4005dc/P/-/-/0 # LBR Entries
*/
class HybridPerfReader : public PerfReaderBase {
public:
HybridPerfReader(cl::list<std::string> &BinaryFilenames)
: PerfReaderBase(BinaryFilenames) {
PerfType = PERF_LBR_STACK;
};
// Parse the hybrid sample including the call and LBR line
void parseSample(TraceStream &TraceIt) override;
void generateRawProfile() override;
private:
// Unwind the hybrid samples after aggregration
void unwindSamples();
void printUnwinderOutput();
};
} // end namespace sampleprof
} // end namespace llvm

3
llvm/tools/llvm-profgen/ProfiledBinary.h
View File

@ -27,6 +27,7 @@
#include "llvm/MC/MCTargetOptions.h"
#include "llvm/Object/ELFObjectFile.h"
#include "llvm/ProfileData/SampleProf.h"
#include "llvm/Support/CommandLine.h"
#include "llvm/Support/Path.h"
#include <list>
#include <set>
@ -190,7 +191,7 @@ public:
StringRef getName() const { return llvm::sys::path::filename(Path); }
uint64_t getBaseAddress() const { return BaseAddress; }
void setBaseAddress(uint64_t Address) { BaseAddress = Address; }
// Return the preferred load address for the first executable segment.
uint64_t getPreferredBaseAddress() const { return PreferredTextSegmentAddresses[0]; }
// Return the file offset for the first executable segment.

16
llvm/tools/llvm-profgen/llvm-profgen.cpp
View File

@ -49,14 +49,20 @@ int main(int argc, const char *argv[]) {
cl::HideUnrelatedOptions({&ProfGenCategory, &getColorCategory()});
cl::ParseCommandLineOptions(argc, argv, "llvm SPGO profile generator\n");
// Load binaries and parse perf events and samples
PerfReader Reader(BinaryFilenames, PerfTraceFilenames);
if (ShowDisassemblyOnly)
if (ShowDisassemblyOnly) {
for (auto BinaryPath : BinaryFilenames) {
(void)ProfiledBinary(BinaryPath);
}
return EXIT_SUCCESS;
Reader.parsePerfTraces(PerfTraceFilenames);
}
// Load binaries and parse perf events and samples
std::unique_ptr<PerfReaderBase> Reader =
PerfReaderBase::create(BinaryFilenames, PerfTraceFilenames);
Reader->parsePerfTraces(PerfTraceFilenames);
std::unique_ptr<ProfileGenerator> Generator = ProfileGenerator::create(
Reader.getBinarySampleCounters(), Reader.getPerfScriptType());
Reader->getBinarySampleCounters(), Reader->getPerfScriptType());
Generator->generateProfile();
Generator->write();