diff options
Diffstat (limited to 'tools/llvm-xray')
-rw-r--r-- | tools/llvm-xray/CMakeLists.txt | 9 | ||||
-rw-r--r-- | tools/llvm-xray/func-id-helper.cc | 60 | ||||
-rw-r--r-- | tools/llvm-xray/func-id-helper.h | 49 | ||||
-rw-r--r-- | tools/llvm-xray/xray-account.cc | 485 | ||||
-rw-r--r-- | tools/llvm-xray/xray-account.h | 109 | ||||
-rw-r--r-- | tools/llvm-xray/xray-converter.cc | 202 | ||||
-rw-r--r-- | tools/llvm-xray/xray-converter.h | 39 | ||||
-rw-r--r-- | tools/llvm-xray/xray-extract.cc | 63 | ||||
-rw-r--r-- | tools/llvm-xray/xray-record-yaml.h | 102 |
9 files changed, 1111 insertions, 7 deletions
diff --git a/tools/llvm-xray/CMakeLists.txt b/tools/llvm-xray/CMakeLists.txt index 0084e35c1b0b8..abcd7d932110a 100644 --- a/tools/llvm-xray/CMakeLists.txt +++ b/tools/llvm-xray/CMakeLists.txt @@ -1,9 +1,16 @@ set(LLVM_LINK_COMPONENTS ${LLVM_TARGETS_TO_BUILD} + DebugInfoDWARF + Object Support - Object) + Symbolize + XRay) set(LLVM_XRAY_TOOLS + func-id-helper.cc + xray-account.cc + xray-converter.cc + xray-extract.cc xray-extract.cc xray-registry.cc) diff --git a/tools/llvm-xray/func-id-helper.cc b/tools/llvm-xray/func-id-helper.cc new file mode 100644 index 0000000000000..3234010695b2b --- /dev/null +++ b/tools/llvm-xray/func-id-helper.cc @@ -0,0 +1,60 @@ +//===- xray-fc-account.cc - XRay Function Call Accounting Tool ------------===// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// +// +// Implementation of the helper tools dealing with XRay-generated function ids. +// +//===----------------------------------------------------------------------===// + +#include "func-id-helper.h" +#include "llvm/Support/Path.h" +#include <sstream> + +using namespace llvm; +using namespace xray; + +std::string FuncIdConversionHelper::SymbolOrNumber(int32_t FuncId) const { + std::ostringstream F; + auto It = FunctionAddresses.find(FuncId); + if (It == FunctionAddresses.end()) { + F << "#" << FuncId; + return F.str(); + } + + if (auto ResOrErr = Symbolizer.symbolizeCode(BinaryInstrMap, It->second)) { + auto &DI = *ResOrErr; + if (DI.FunctionName == "<invalid>") + F << "@(" << std::hex << It->second << ")"; + else + F << DI.FunctionName; + } else + handleAllErrors(ResOrErr.takeError(), [&](const ErrorInfoBase &) { + F << "@(" << std::hex << It->second << ")"; + }); + + return F.str(); +} + +std::string FuncIdConversionHelper::FileLineAndColumn(int32_t FuncId) const { + auto It = FunctionAddresses.find(FuncId); + if (It == FunctionAddresses.end()) + return "(unknown)"; + + std::ostringstream F; + auto ResOrErr = Symbolizer.symbolizeCode(BinaryInstrMap, It->second); + if (!ResOrErr) { + consumeError(ResOrErr.takeError()); + return "(unknown)"; + } + + auto &DI = *ResOrErr; + F << sys::path::filename(DI.FileName).str() << ":" << DI.Line << ":" + << DI.Column; + + return F.str(); +} diff --git a/tools/llvm-xray/func-id-helper.h b/tools/llvm-xray/func-id-helper.h new file mode 100644 index 0000000000000..7348a7100b05d --- /dev/null +++ b/tools/llvm-xray/func-id-helper.h @@ -0,0 +1,49 @@ +//===- func-id-helper.h - XRay Function ID Conversion Helpers -------------===// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// +// +// Defines helper tools dealing with XRay-generated function ids. +// +//===----------------------------------------------------------------------===// +#ifndef LLVM_TOOLS_LLVM_XRAY_FUNC_ID_HELPER_H +#define LLVM_TOOLS_LLVM_XRAY_FUNC_ID_HELPER_H + +#include "llvm/DebugInfo/Symbolize/Symbolize.h" +#include <unordered_map> + +namespace llvm { +namespace xray { + +// This class consolidates common operations related to Function IDs. +class FuncIdConversionHelper { +public: + using FunctionAddressMap = std::unordered_map<int32_t, uint64_t>; + +private: + std::string BinaryInstrMap; + symbolize::LLVMSymbolizer &Symbolizer; + const FunctionAddressMap &FunctionAddresses; + +public: + FuncIdConversionHelper(std::string BinaryInstrMap, + symbolize::LLVMSymbolizer &Symbolizer, + const FunctionAddressMap &FunctionAddresses) + : BinaryInstrMap(std::move(BinaryInstrMap)), Symbolizer(Symbolizer), + FunctionAddresses(FunctionAddresses) {} + + // Returns the symbol or a string representation of the function id. + std::string SymbolOrNumber(int32_t FuncId) const; + + // Returns the file and column from debug info for the given function id. + std::string FileLineAndColumn(int32_t FuncId) const; +}; + +} // namespace xray +} // namespace llvm + +#endif // LLVM_TOOLS_LLVM_XRAY_FUNC_ID_HELPER_H diff --git a/tools/llvm-xray/xray-account.cc b/tools/llvm-xray/xray-account.cc new file mode 100644 index 0000000000000..671a5a073eecc --- /dev/null +++ b/tools/llvm-xray/xray-account.cc @@ -0,0 +1,485 @@ +//===- xray-account.h - XRay Function Call Accounting ---------------------===// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// +// +// This file implements basic function call accounting from an XRay trace. +// +//===----------------------------------------------------------------------===// + +#include <algorithm> +#include <cassert> +#include <numeric> +#include <system_error> +#include <utility> + +#include "xray-account.h" +#include "xray-extract.h" +#include "xray-registry.h" +#include "llvm/Support/ErrorHandling.h" +#include "llvm/Support/FormatVariadic.h" +#include "llvm/XRay/Trace.h" + +using namespace llvm; +using namespace llvm::xray; + +static cl::SubCommand Account("account", "Function call accounting"); +static cl::opt<std::string> AccountInput(cl::Positional, + cl::desc("<xray log file>"), + cl::Required, cl::sub(Account)); +static cl::opt<bool> + AccountKeepGoing("keep-going", cl::desc("Keep going on errors encountered"), + cl::sub(Account), cl::init(false)); +static cl::alias AccountKeepGoing2("k", cl::aliasopt(AccountKeepGoing), + cl::desc("Alias for -keep_going"), + cl::sub(Account)); +static cl::opt<bool> AccountDeduceSiblingCalls( + "deduce-sibling-calls", + cl::desc("Deduce sibling calls when unrolling function call stacks"), + cl::sub(Account), cl::init(false)); +static cl::alias + AccountDeduceSiblingCalls2("d", cl::aliasopt(AccountDeduceSiblingCalls), + cl::desc("Alias for -deduce_sibling_calls"), + cl::sub(Account)); +static cl::opt<std::string> + AccountOutput("output", cl::value_desc("output file"), cl::init("-"), + cl::desc("output file; use '-' for stdout"), + cl::sub(Account)); +static cl::alias AccountOutput2("o", cl::aliasopt(AccountOutput), + cl::desc("Alias for -output"), + cl::sub(Account)); +enum class AccountOutputFormats { TEXT, CSV }; +static cl::opt<AccountOutputFormats> + AccountOutputFormat("format", cl::desc("output format"), + cl::values(clEnumValN(AccountOutputFormats::TEXT, + "text", "report stats in text"), + clEnumValN(AccountOutputFormats::CSV, "csv", + "report stats in csv")), + cl::sub(Account)); +static cl::alias AccountOutputFormat2("f", cl::desc("Alias of -format"), + cl::aliasopt(AccountOutputFormat), + cl::sub(Account)); + +enum class SortField { + FUNCID, + COUNT, + MIN, + MED, + PCT90, + PCT99, + MAX, + SUM, + FUNC, +}; + +static cl::opt<SortField> AccountSortOutput( + "sort", cl::desc("sort output by this field"), cl::value_desc("field"), + cl::sub(Account), cl::init(SortField::FUNCID), + cl::values(clEnumValN(SortField::FUNCID, "funcid", "function id"), + clEnumValN(SortField::COUNT, "count", "funciton call counts"), + clEnumValN(SortField::MIN, "min", "minimum function durations"), + clEnumValN(SortField::MED, "med", "median function durations"), + clEnumValN(SortField::PCT90, "90p", "90th percentile durations"), + clEnumValN(SortField::PCT99, "99p", "99th percentile durations"), + clEnumValN(SortField::MAX, "max", "maximum function durations"), + clEnumValN(SortField::SUM, "sum", "sum of call durations"), + clEnumValN(SortField::FUNC, "func", "function names"))); +static cl::alias AccountSortOutput2("s", cl::aliasopt(AccountSortOutput), + cl::desc("Alias for -sort"), + cl::sub(Account)); + +enum class SortDirection { + ASCENDING, + DESCENDING, +}; +static cl::opt<SortDirection> AccountSortOrder( + "sortorder", cl::desc("sort ordering"), cl::init(SortDirection::ASCENDING), + cl::values(clEnumValN(SortDirection::ASCENDING, "asc", "ascending"), + clEnumValN(SortDirection::DESCENDING, "dsc", "descending")), + cl::sub(Account)); +static cl::alias AccountSortOrder2("r", cl::aliasopt(AccountSortOrder), + cl::desc("Alias for -sortorder"), + cl::sub(Account)); + +static cl::opt<int> AccountTop("top", cl::desc("only show the top N results"), + cl::value_desc("N"), cl::sub(Account), + cl::init(-1)); +static cl::alias AccountTop2("p", cl::desc("Alias for -top"), + cl::aliasopt(AccountTop), cl::sub(Account)); + +static cl::opt<std::string> + AccountInstrMap("instr_map", + cl::desc("binary with the instrumentation map, or " + "a separate instrumentation map"), + cl::value_desc("binary with xray_instr_map"), + cl::sub(Account), cl::init("")); +static cl::alias AccountInstrMap2("m", cl::aliasopt(AccountInstrMap), + cl::desc("Alias for -instr_map"), + cl::sub(Account)); +static cl::opt<InstrumentationMapExtractor::InputFormats> InstrMapFormat( + "instr-map-format", cl::desc("format of instrumentation map"), + cl::values(clEnumValN(InstrumentationMapExtractor::InputFormats::ELF, "elf", + "instrumentation map in an ELF header"), + clEnumValN(InstrumentationMapExtractor::InputFormats::YAML, + "yaml", "instrumentation map in YAML")), + cl::sub(Account), cl::init(InstrumentationMapExtractor::InputFormats::ELF)); +static cl::alias InstrMapFormat2("t", cl::aliasopt(InstrMapFormat), + cl::desc("Alias for -instr-map-format"), + cl::sub(Account)); + +namespace { + +template <class T, class U> void setMinMax(std::pair<T, T> &MM, U &&V) { + if (MM.first == 0 || MM.second == 0) + MM = std::make_pair(std::forward<U>(V), std::forward<U>(V)); + else + MM = std::make_pair(std::min(MM.first, V), std::max(MM.second, V)); +} + +template <class T> T diff(T L, T R) { return std::max(L, R) - std::min(L, R); } + +} // namespace + +bool LatencyAccountant::accountRecord(const XRayRecord &Record) { + setMinMax(PerThreadMinMaxTSC[Record.TId], Record.TSC); + setMinMax(PerCPUMinMaxTSC[Record.CPU], Record.TSC); + + if (CurrentMaxTSC == 0) + CurrentMaxTSC = Record.TSC; + + if (Record.TSC < CurrentMaxTSC) + return false; + + auto &ThreadStack = PerThreadFunctionStack[Record.TId]; + switch (Record.Type) { + case RecordTypes::ENTER: { + // Function Enter + ThreadStack.emplace_back(Record.FuncId, Record.TSC); + break; + } + case RecordTypes::EXIT: { + // Function Exit + if (ThreadStack.back().first == Record.FuncId) { + const auto &Top = ThreadStack.back(); + recordLatency(Top.first, diff(Top.second, Record.TSC)); + ThreadStack.pop_back(); + break; + } + + if (!DeduceSiblingCalls) + return false; + + // Look for the parent up the stack. + auto Parent = + std::find_if(ThreadStack.rbegin(), ThreadStack.rend(), + [&](const std::pair<const int32_t, uint64_t> &E) { + return E.first == Record.FuncId; + }); + if (Parent == ThreadStack.rend()) + return false; + + // Account time for this apparently sibling call exit up the stack. + // Considering the following case: + // + // f() + // g() + // h() + // + // We might only ever see the following entries: + // + // -> f() + // -> g() + // -> h() + // <- h() + // <- f() + // + // Now we don't see the exit to g() because some older version of the XRay + // runtime wasn't instrumenting tail exits. If we don't deduce tail calls, + // we may potentially never account time for g() -- and this code would have + // already bailed out, because `<- f()` doesn't match the current "top" of + // stack where we're waiting for the exit to `g()` instead. This is not + // ideal and brittle -- so instead we provide a potentially inaccurate + // accounting of g() instead, computing it from the exit of f(). + // + // While it might be better that we account the time between `-> g()` and + // `-> h()` as the proper accounting of time for g() here, this introduces + // complexity to do correctly (need to backtrack, etc.). + // + // FIXME: Potentially implement the more complex deduction algorithm? + auto I = std::next(Parent).base(); + for (auto &E : make_range(I, ThreadStack.end())) { + recordLatency(E.first, diff(E.second, Record.TSC)); + } + ThreadStack.erase(I, ThreadStack.end()); + break; + } + } + + return true; +} + +namespace { + +// We consolidate the data into a struct which we can output in various forms. +struct ResultRow { + uint64_t Count; + double Min; + double Median; + double Pct90; + double Pct99; + double Max; + double Sum; + std::string DebugInfo; + std::string Function; +}; + +ResultRow getStats(std::vector<uint64_t> &Timings) { + assert(!Timings.empty()); + ResultRow R; + R.Sum = std::accumulate(Timings.begin(), Timings.end(), 0.0); + auto MinMax = std::minmax_element(Timings.begin(), Timings.end()); + R.Min = *MinMax.first; + R.Max = *MinMax.second; + auto MedianOff = Timings.size() / 2; + std::nth_element(Timings.begin(), Timings.begin() + MedianOff, Timings.end()); + R.Median = Timings[MedianOff]; + auto Pct90Off = std::floor(Timings.size() * 0.9); + std::nth_element(Timings.begin(), Timings.begin() + Pct90Off, Timings.end()); + R.Pct90 = Timings[Pct90Off]; + auto Pct99Off = std::floor(Timings.size() * 0.99); + std::nth_element(Timings.begin(), Timings.begin() + Pct90Off, Timings.end()); + R.Pct99 = Timings[Pct99Off]; + R.Count = Timings.size(); + return R; +} + +} // namespace + +template <class F> +void LatencyAccountant::exportStats(const XRayFileHeader &Header, F Fn) const { + using TupleType = std::tuple<int32_t, uint64_t, ResultRow>; + std::vector<TupleType> Results; + Results.reserve(FunctionLatencies.size()); + for (auto FT : FunctionLatencies) { + const auto &FuncId = FT.first; + auto &Timings = FT.second; + Results.emplace_back(FuncId, Timings.size(), getStats(Timings)); + auto &Row = std::get<2>(Results.back()); + if (Header.CycleFrequency) { + double CycleFrequency = Header.CycleFrequency; + Row.Min /= CycleFrequency; + Row.Median /= CycleFrequency; + Row.Pct90 /= CycleFrequency; + Row.Pct99 /= CycleFrequency; + Row.Max /= CycleFrequency; + Row.Sum /= CycleFrequency; + } + + Row.Function = FuncIdHelper.SymbolOrNumber(FuncId); + Row.DebugInfo = FuncIdHelper.FileLineAndColumn(FuncId); + } + + // Sort the data according to user-provided flags. + switch (AccountSortOutput) { + case SortField::FUNCID: + std::sort(Results.begin(), Results.end(), + [](const TupleType &L, const TupleType &R) { + if (AccountSortOrder == SortDirection::ASCENDING) + return std::get<0>(L) < std::get<0>(R); + if (AccountSortOrder == SortDirection::DESCENDING) + return std::get<0>(L) > std::get<0>(R); + llvm_unreachable("Unknown sort direction"); + }); + break; + case SortField::COUNT: + std::sort(Results.begin(), Results.end(), + [](const TupleType &L, const TupleType &R) { + if (AccountSortOrder == SortDirection::ASCENDING) + return std::get<1>(L) < std::get<1>(R); + if (AccountSortOrder == SortDirection::DESCENDING) + return std::get<1>(L) > std::get<1>(R); + llvm_unreachable("Unknown sort direction"); + }); + break; + default: + // Here we need to look into the ResultRow for the rest of the data that + // we want to sort by. + std::sort(Results.begin(), Results.end(), + [&](const TupleType &L, const TupleType &R) { + auto &LR = std::get<2>(L); + auto &RR = std::get<2>(R); + switch (AccountSortOutput) { + case SortField::COUNT: + if (AccountSortOrder == SortDirection::ASCENDING) + return LR.Count < RR.Count; + if (AccountSortOrder == SortDirection::DESCENDING) + return LR.Count > RR.Count; + llvm_unreachable("Unknown sort direction"); + case SortField::MIN: + if (AccountSortOrder == SortDirection::ASCENDING) + return LR.Min < RR.Min; + if (AccountSortOrder == SortDirection::DESCENDING) + return LR.Min > RR.Min; + llvm_unreachable("Unknown sort direction"); + case SortField::MED: + if (AccountSortOrder == SortDirection::ASCENDING) + return LR.Median < RR.Median; + if (AccountSortOrder == SortDirection::DESCENDING) + return LR.Median > RR.Median; + llvm_unreachable("Unknown sort direction"); + case SortField::PCT90: + if (AccountSortOrder == SortDirection::ASCENDING) + return LR.Pct90 < RR.Pct90; + if (AccountSortOrder == SortDirection::DESCENDING) + return LR.Pct90 > RR.Pct90; + llvm_unreachable("Unknown sort direction"); + case SortField::PCT99: + if (AccountSortOrder == SortDirection::ASCENDING) + return LR.Pct99 < RR.Pct99; + if (AccountSortOrder == SortDirection::DESCENDING) + return LR.Pct99 > RR.Pct99; + llvm_unreachable("Unknown sort direction"); + case SortField::MAX: + if (AccountSortOrder == SortDirection::ASCENDING) + return LR.Max < RR.Max; + if (AccountSortOrder == SortDirection::DESCENDING) + return LR.Max > RR.Max; + llvm_unreachable("Unknown sort direction"); + case SortField::SUM: + if (AccountSortOrder == SortDirection::ASCENDING) + return LR.Sum < RR.Sum; + if (AccountSortOrder == SortDirection::DESCENDING) + return LR.Sum > RR.Sum; + llvm_unreachable("Unknown sort direction"); + default: + llvm_unreachable("Unsupported sort order"); + } + }); + break; + } + + if (AccountTop > 0) + Results.erase(Results.begin() + AccountTop.getValue(), Results.end()); + + for (const auto &R : Results) + Fn(std::get<0>(R), std::get<1>(R), std::get<2>(R)); +} + +void LatencyAccountant::exportStatsAsText(raw_ostream &OS, + const XRayFileHeader &Header) const { + OS << "Functions with latencies: " << FunctionLatencies.size() << "\n"; + + // We spend some effort to make the text output more readable, so we do the + // following formatting decisions for each of the fields: + // + // - funcid: 32-bit, but we can determine the largest number and be + // between + // a minimum of 5 characters, up to 9 characters, right aligned. + // - count: 64-bit, but we can determine the largest number and be + // between + // a minimum of 5 characters, up to 9 characters, right aligned. + // - min, median, 90pct, 99pct, max: double precision, but we want to keep + // the values in seconds, with microsecond precision (0.000'001), so we + // have at most 6 significant digits, with the whole number part to be + // at + // least 1 character. For readability we'll right-align, with full 9 + // characters each. + // - debug info, function name: we format this as a concatenation of the + // debug info and the function name. + // + static constexpr char StatsHeaderFormat[] = + "{0,+9} {1,+10} [{2,+9}, {3,+9}, {4,+9}, {5,+9}, {6,+9}] {7,+9}"; + static constexpr char StatsFormat[] = + R"({0,+9} {1,+10} [{2,+9:f6}, {3,+9:f6}, {4,+9:f6}, {5,+9:f6}, {6,+9:f6}] {7,+9:f6})"; + OS << llvm::formatv(StatsHeaderFormat, "funcid", "count", "min", "med", "90p", + "99p", "max", "sum") + << llvm::formatv(" {0,-12}\n", "function"); + exportStats(Header, [&](int32_t FuncId, size_t Count, const ResultRow &Row) { + OS << llvm::formatv(StatsFormat, FuncId, Count, Row.Min, Row.Median, + Row.Pct90, Row.Pct99, Row.Max, Row.Sum) + << " " << Row.DebugInfo << ": " << Row.Function << "\n"; + }); +} + +void LatencyAccountant::exportStatsAsCSV(raw_ostream &OS, + const XRayFileHeader &Header) const { + OS << "funcid,count,min,median,90%ile,99%ile,max,sum,debug,function\n"; + exportStats(Header, [&](int32_t FuncId, size_t Count, const ResultRow &Row) { + OS << FuncId << ',' << Count << ',' << Row.Min << ',' << Row.Median << ',' + << Row.Pct90 << ',' << Row.Pct99 << ',' << Row.Max << "," << Row.Sum + << ",\"" << Row.DebugInfo << "\",\"" << Row.Function << "\"\n"; + }); +} + +using namespace llvm::xray; + +static CommandRegistration Unused(&Account, []() -> Error { + int Fd; + auto EC = sys::fs::openFileForRead(AccountInput, Fd); + if (EC) + return make_error<StringError>( + Twine("Cannot open file '") + AccountInput + "'", EC); + + Error Err = Error::success(); + xray::InstrumentationMapExtractor Extractor(AccountInstrMap, InstrMapFormat, + Err); + if (auto E = handleErrors( + std::move(Err), [&](std::unique_ptr<StringError> SE) -> Error { + if (SE->convertToErrorCode() == std::errc::no_such_file_or_directory) + return Error::success(); + return Error(std::move(SE)); + })) + return E; + + raw_fd_ostream OS(AccountOutput, EC, sys::fs::OpenFlags::F_Text); + if (EC) + return make_error<StringError>( + Twine("Cannot open file '") + AccountOutput + "' for writing.", EC); + + const auto &FunctionAddresses = Extractor.getFunctionAddresses(); + symbolize::LLVMSymbolizer::Options Opts( + symbolize::FunctionNameKind::LinkageName, true, true, false, ""); + symbolize::LLVMSymbolizer Symbolizer(Opts); + llvm::xray::FuncIdConversionHelper FuncIdHelper(AccountInstrMap, Symbolizer, + FunctionAddresses); + xray::LatencyAccountant FCA(FuncIdHelper, AccountDeduceSiblingCalls); + if (auto TraceOrErr = loadTraceFile(AccountInput)) { + auto &T = *TraceOrErr; + for (const auto &Record : T) { + if (FCA.accountRecord(Record)) + continue; + for (const auto &ThreadStack : FCA.getPerThreadFunctionStack()) { + errs() << "Thread ID: " << ThreadStack.first << "\n"; + auto Level = ThreadStack.second.size(); + for (const auto &Entry : llvm::reverse(ThreadStack.second)) + errs() << "#" << Level-- << "\t" + << FuncIdHelper.SymbolOrNumber(Entry.first) << '\n'; + } + if (!AccountKeepGoing) + return make_error<StringError>( + Twine("Failed accounting function calls in file '") + AccountInput + + "'.", + std::make_error_code(std::errc::executable_format_error)); + } + switch (AccountOutputFormat) { + case AccountOutputFormats::TEXT: + FCA.exportStatsAsText(OS, T.getFileHeader()); + break; + case AccountOutputFormats::CSV: + FCA.exportStatsAsCSV(OS, T.getFileHeader()); + break; + } + } else { + return joinErrors( + make_error<StringError>( + Twine("Failed loading input file '") + AccountInput + "'", + std::make_error_code(std::errc::executable_format_error)), + TraceOrErr.takeError()); + } + + return Error::success(); +}); diff --git a/tools/llvm-xray/xray-account.h b/tools/llvm-xray/xray-account.h new file mode 100644 index 0000000000000..cc9ba897e5373 --- /dev/null +++ b/tools/llvm-xray/xray-account.h @@ -0,0 +1,109 @@ +//===- xray-account.h - XRay Function Call Accounting ---------------------===// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// +// +// This file defines the interface for performing some basic function call +// accounting from an XRay trace. +// +//===----------------------------------------------------------------------===// +#ifndef LLVM_TOOLS_LLVM_XRAY_XRAY_ACCOUNT_H +#define LLVM_TOOLS_LLVM_XRAY_XRAY_ACCOUNT_H + +#include <map> +#include <utility> +#include <vector> + +#include "func-id-helper.h" +#include "llvm/Support/Program.h" +#include "llvm/Support/raw_ostream.h" +#include "llvm/XRay/XRayRecord.h" + +namespace llvm { +namespace xray { + +class LatencyAccountant { +public: + typedef std::map<int32_t, std::vector<uint64_t>> FunctionLatencyMap; + typedef std::map<llvm::sys::ProcessInfo::ProcessId, + std::pair<uint64_t, uint64_t>> + PerThreadMinMaxTSCMap; + typedef std::map<uint8_t, std::pair<uint64_t, uint64_t>> PerCPUMinMaxTSCMap; + typedef std::vector<std::pair<int32_t, uint64_t>> FunctionStack; + typedef std::map<llvm::sys::ProcessInfo::ProcessId, FunctionStack> + PerThreadFunctionStackMap; + +private: + PerThreadFunctionStackMap PerThreadFunctionStack; + FunctionLatencyMap FunctionLatencies; + PerThreadMinMaxTSCMap PerThreadMinMaxTSC; + PerCPUMinMaxTSCMap PerCPUMinMaxTSC; + FuncIdConversionHelper &FuncIdHelper; + + bool DeduceSiblingCalls = false; + uint64_t CurrentMaxTSC = 0; + + void recordLatency(int32_t FuncId, uint64_t Latency) { + FunctionLatencies[FuncId].push_back(Latency); + } + +public: + explicit LatencyAccountant(FuncIdConversionHelper &FuncIdHelper, + bool DeduceSiblingCalls) + : FuncIdHelper(FuncIdHelper), DeduceSiblingCalls(DeduceSiblingCalls) {} + + const FunctionLatencyMap &getFunctionLatencies() const { + return FunctionLatencies; + } + + const PerThreadMinMaxTSCMap &getPerThreadMinMaxTSC() const { + return PerThreadMinMaxTSC; + } + + const PerCPUMinMaxTSCMap &getPerCPUMinMaxTSC() const { + return PerCPUMinMaxTSC; + } + + /// Returns false in case we fail to account the provided record. This happens + /// in the following cases: + /// + /// - An exit record does not match any entry records for the same function. + /// If we've been set to deduce sibling calls, we try walking up the stack + /// and recording times for the higher level functions. + /// - A record has a TSC that's before the latest TSC that has been + /// recorded. We still record the TSC for the min-max. + /// + bool accountRecord(const XRayRecord &Record); + + const FunctionStack * + getThreadFunctionStack(llvm::sys::ProcessInfo::ProcessId TId) const { + auto I = PerThreadFunctionStack.find(TId); + if (I == PerThreadFunctionStack.end()) + return nullptr; + return &I->second; + } + + const PerThreadFunctionStackMap &getPerThreadFunctionStack() const { + return PerThreadFunctionStack; + } + + // Output Functions + // ================ + + void exportStatsAsText(raw_ostream &OS, const XRayFileHeader &Header) const; + void exportStatsAsCSV(raw_ostream &OS, const XRayFileHeader &Header) const; + +private: + // Internal helper to implement common parts of the exportStatsAs... + // functions. + template <class F> void exportStats(const XRayFileHeader &Header, F fn) const; +}; + +} // namespace xray +} // namespace llvm + +#endif // LLVM_TOOLS_LLVM_XRAY_XRAY_ACCOUNT_H diff --git a/tools/llvm-xray/xray-converter.cc b/tools/llvm-xray/xray-converter.cc new file mode 100644 index 0000000000000..31275e2902f29 --- /dev/null +++ b/tools/llvm-xray/xray-converter.cc @@ -0,0 +1,202 @@ +//===- xray-converter.cc - XRay Trace Conversion --------------------------===// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// +// +// Implements the trace conversion functions. +// +//===----------------------------------------------------------------------===// +#include "xray-converter.h" + +#include "xray-extract.h" +#include "xray-registry.h" +#include "llvm/DebugInfo/Symbolize/Symbolize.h" +#include "llvm/Support/EndianStream.h" +#include "llvm/Support/FileSystem.h" +#include "llvm/Support/YAMLTraits.h" +#include "llvm/Support/raw_ostream.h" +#include "llvm/XRay/Trace.h" +#include "llvm/XRay/YAMLXRayRecord.h" + +using namespace llvm; +using namespace xray; + +// llvm-xray convert +// ---------------------------------------------------------------------------- +static cl::SubCommand Convert("convert", "Trace Format Conversion"); +static cl::opt<std::string> ConvertInput(cl::Positional, + cl::desc("<xray log file>"), + cl::Required, cl::sub(Convert)); +enum class ConvertFormats { BINARY, YAML }; +static cl::opt<ConvertFormats> ConvertOutputFormat( + "output-format", cl::desc("output format"), + cl::values(clEnumValN(ConvertFormats::BINARY, "raw", "output in binary"), + clEnumValN(ConvertFormats::YAML, "yaml", "output in yaml")), + cl::sub(Convert)); +static cl::alias ConvertOutputFormat2("f", cl::aliasopt(ConvertOutputFormat), + cl::desc("Alias for -output-format"), + cl::sub(Convert)); +static cl::opt<std::string> + ConvertOutput("output", cl::value_desc("output file"), cl::init("-"), + cl::desc("output file; use '-' for stdout"), + cl::sub(Convert)); +static cl::alias ConvertOutput2("o", cl::aliasopt(ConvertOutput), + cl::desc("Alias for -output"), + cl::sub(Convert)); + +static cl::opt<bool> + ConvertSymbolize("symbolize", + cl::desc("symbolize function ids from the input log"), + cl::init(false), cl::sub(Convert)); +static cl::alias ConvertSymbolize2("y", cl::aliasopt(ConvertSymbolize), + cl::desc("Alias for -symbolize"), + cl::sub(Convert)); + +static cl::opt<std::string> + ConvertInstrMap("instr_map", + cl::desc("binary with the instrumentation map, or " + "a separate instrumentation map"), + cl::value_desc("binary with xray_instr_map"), + cl::sub(Convert), cl::init("")); +static cl::alias ConvertInstrMap2("m", cl::aliasopt(ConvertInstrMap), + cl::desc("Alias for -instr_map"), + cl::sub(Convert)); +static cl::opt<bool> ConvertSortInput( + "sort", + cl::desc("determines whether to sort input log records by timestamp"), + cl::sub(Convert), cl::init(true)); +static cl::alias ConvertSortInput2("s", cl::aliasopt(ConvertSortInput), + cl::desc("Alias for -sort"), + cl::sub(Convert)); +static cl::opt<InstrumentationMapExtractor::InputFormats> InstrMapFormat( + "instr-map-format", cl::desc("format of instrumentation map"), + cl::values(clEnumValN(InstrumentationMapExtractor::InputFormats::ELF, "elf", + "instrumentation map in an ELF header"), + clEnumValN(InstrumentationMapExtractor::InputFormats::YAML, + "yaml", "instrumentation map in YAML")), + cl::sub(Convert), cl::init(InstrumentationMapExtractor::InputFormats::ELF)); +static cl::alias InstrMapFormat2("t", cl::aliasopt(InstrMapFormat), + cl::desc("Alias for -instr-map-format"), + cl::sub(Convert)); + +using llvm::yaml::IO; +using llvm::yaml::Output; + +void TraceConverter::exportAsYAML(const Trace &Records, raw_ostream &OS) { + YAMLXRayTrace Trace; + const auto &FH = Records.getFileHeader(); + Trace.Header = {FH.Version, FH.Type, FH.ConstantTSC, FH.NonstopTSC, + FH.CycleFrequency}; + Trace.Records.reserve(Records.size()); + for (const auto &R : Records) { + Trace.Records.push_back({R.RecordType, R.CPU, R.Type, R.FuncId, + Symbolize ? FuncIdHelper.SymbolOrNumber(R.FuncId) + : std::to_string(R.FuncId), + R.TSC, R.TId}); + } + Output Out(OS); + Out << Trace; +} + +void TraceConverter::exportAsRAWv1(const Trace &Records, raw_ostream &OS) { + // First write out the file header, in the correct endian-appropriate format + // (XRay assumes currently little endian). + support::endian::Writer<support::endianness::little> Writer(OS); + const auto &FH = Records.getFileHeader(); + Writer.write(FH.Version); + Writer.write(FH.Type); + uint32_t Bitfield{0}; + if (FH.ConstantTSC) + Bitfield |= 1uL; + if (FH.NonstopTSC) + Bitfield |= 1uL << 1; + Writer.write(Bitfield); + Writer.write(FH.CycleFrequency); + + // There's 16 bytes of padding at the end of the file header. + static constexpr uint32_t Padding4B = 0; + Writer.write(Padding4B); + Writer.write(Padding4B); + Writer.write(Padding4B); + Writer.write(Padding4B); + + // Then write out the rest of the records, still in an endian-appropriate + // format. + for (const auto &R : Records) { + Writer.write(R.RecordType); + Writer.write(R.CPU); + switch (R.Type) { + case RecordTypes::ENTER: + Writer.write(uint8_t{0}); + break; + case RecordTypes::EXIT: + Writer.write(uint8_t{1}); + break; + } + Writer.write(R.FuncId); + Writer.write(R.TSC); + Writer.write(R.TId); + Writer.write(Padding4B); + Writer.write(Padding4B); + Writer.write(Padding4B); + } +} + +namespace llvm { +namespace xray { + +static CommandRegistration Unused(&Convert, []() -> Error { + // FIXME: Support conversion to BINARY when upgrading XRay trace versions. + int Fd; + auto EC = sys::fs::openFileForRead(ConvertInput, Fd); + if (EC) + return make_error<StringError>( + Twine("Cannot open file '") + ConvertInput + "'", EC); + + Error Err = Error::success(); + xray::InstrumentationMapExtractor Extractor(ConvertInstrMap, InstrMapFormat, + Err); + handleAllErrors(std::move(Err), + [&](const ErrorInfoBase &E) { E.log(errs()); }); + + const auto &FunctionAddresses = Extractor.getFunctionAddresses(); + symbolize::LLVMSymbolizer::Options Opts( + symbolize::FunctionNameKind::LinkageName, true, true, false, ""); + symbolize::LLVMSymbolizer Symbolizer(Opts); + llvm::xray::FuncIdConversionHelper FuncIdHelper(ConvertInstrMap, Symbolizer, + FunctionAddresses); + llvm::xray::TraceConverter TC(FuncIdHelper, ConvertSymbolize); + raw_fd_ostream OS(ConvertOutput, EC, + ConvertOutputFormat == ConvertFormats::BINARY + ? sys::fs::OpenFlags::F_None + : sys::fs::OpenFlags::F_Text); + if (EC) + return make_error<StringError>( + Twine("Cannot open file '") + ConvertOutput + "' for writing.", EC); + + if (auto TraceOrErr = loadTraceFile(ConvertInput, ConvertSortInput)) { + auto &T = *TraceOrErr; + switch (ConvertOutputFormat) { + case ConvertFormats::YAML: + TC.exportAsYAML(T, OS); + break; + case ConvertFormats::BINARY: + TC.exportAsRAWv1(T, OS); + break; + } + } else { + return joinErrors( + make_error<StringError>( + Twine("Failed loading input file '") + ConvertInput + "'.", + std::make_error_code(std::errc::executable_format_error)), + TraceOrErr.takeError()); + } + return Error::success(); +}); + +} // namespace xray +} // namespace llvm diff --git a/tools/llvm-xray/xray-converter.h b/tools/llvm-xray/xray-converter.h new file mode 100644 index 0000000000000..fa0d5e132f14d --- /dev/null +++ b/tools/llvm-xray/xray-converter.h @@ -0,0 +1,39 @@ +//===- xray-converter.h - XRay Trace Conversion ---------------------------===// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// +// +// Defines the TraceConverter class for turning binary traces into +// human-readable text and vice versa. +// +//===----------------------------------------------------------------------===// +#ifndef LLVM_TOOLS_LLVM_XRAY_XRAY_CONVERTER_H +#define LLVM_TOOLS_LLVM_XRAY_XRAY_CONVERTER_H + +#include "func-id-helper.h" +#include "llvm/XRay/XRayRecord.h" +#include "llvm/XRay/Trace.h" + +namespace llvm { +namespace xray { + +class TraceConverter { + FuncIdConversionHelper &FuncIdHelper; + bool Symbolize; + +public: + TraceConverter(FuncIdConversionHelper &FuncIdHelper, bool Symbolize = false) + : FuncIdHelper(FuncIdHelper), Symbolize(Symbolize) {} + + void exportAsYAML(const Trace &Records, raw_ostream &OS); + void exportAsRAWv1(const Trace &Records, raw_ostream &OS); +}; + +} // namespace xray +} // namespace llvm + +#endif // LLVM_TOOLS_LLVM_XRAY_XRAY_CONVERTER_H diff --git a/tools/llvm-xray/xray-extract.cc b/tools/llvm-xray/xray-extract.cc index e51b64c8ad6e8..49ecd74211374 100644 --- a/tools/llvm-xray/xray-extract.cc +++ b/tools/llvm-xray/xray-extract.cc @@ -162,8 +162,7 @@ llvm::Error LoadBinaryInstrELF( "'.", std::make_error_code(std::errc::executable_format_error)); } - auto AlwaysInstrument = Extractor.getU8(&OffsetPtr); - Entry.AlwaysInstrument = AlwaysInstrument != 0; + Entry.AlwaysInstrument = Extractor.getU8(&OffsetPtr) != 0; // We replicate the function id generation scheme implemented in the runtime // here. Ideally we should be able to break it out, or output this map from @@ -185,30 +184,82 @@ llvm::Error LoadBinaryInstrELF( return llvm::Error::success(); } +Error LoadYAMLInstrMap( + StringRef Filename, std::deque<SledEntry> &Sleds, + InstrumentationMapExtractor::FunctionAddressMap &InstrMap, + InstrumentationMapExtractor::FunctionAddressReverseMap &FunctionIds) { + int Fd; + if (auto EC = sys::fs::openFileForRead(Filename, Fd)) + return make_error<StringError>( + Twine("Failed opening file '") + Filename + "' for reading.", EC); + + uint64_t FileSize; + if (auto EC = sys::fs::file_size(Filename, FileSize)) + return make_error<StringError>( + Twine("Failed getting size of file '") + Filename + "'.", EC); + + std::error_code EC; + sys::fs::mapped_file_region MappedFile( + Fd, sys::fs::mapped_file_region::mapmode::readonly, FileSize, 0, EC); + if (EC) + return make_error<StringError>( + Twine("Failed memory-mapping file '") + Filename + "'.", EC); + + std::vector<YAMLXRaySledEntry> YAMLSleds; + Input In(StringRef(MappedFile.data(), MappedFile.size())); + In >> YAMLSleds; + if (In.error()) + return make_error<StringError>( + Twine("Failed loading YAML document from '") + Filename + "'.", + In.error()); + + for (const auto &Y : YAMLSleds) { + InstrMap[Y.FuncId] = Y.Function; + FunctionIds[Y.Function] = Y.FuncId; + Sleds.push_back( + SledEntry{Y.Address, Y.Function, Y.Kind, Y.AlwaysInstrument}); + } + return Error::success(); +} + } // namespace InstrumentationMapExtractor::InstrumentationMapExtractor(std::string Filename, InputFormats Format, Error &EC) { ErrorAsOutParameter ErrAsOutputParam(&EC); + if (Filename.empty()) { + EC = Error::success(); + return; + } switch (Format) { case InputFormats::ELF: { EC = handleErrors( LoadBinaryInstrELF(Filename, Sleds, FunctionAddresses, FunctionIds), - [](std::unique_ptr<ErrorInfoBase> E) { + [&](std::unique_ptr<ErrorInfoBase> E) { return joinErrors( make_error<StringError>( Twine("Cannot extract instrumentation map from '") + - ExtractInput + "'.", + Filename + "'.", std::make_error_code(std::errc::executable_format_error)), std::move(E)); }); break; } - default: - llvm_unreachable("Input format type not supported yet."); + case InputFormats::YAML: { + EC = handleErrors( + LoadYAMLInstrMap(Filename, Sleds, FunctionAddresses, FunctionIds), + [&](std::unique_ptr<ErrorInfoBase> E) { + return joinErrors( + make_error<StringError>( + Twine("Cannot load YAML instrumentation map from '") + + Filename + "'.", + std::make_error_code(std::errc::executable_format_error)), + std::move(E)); + }); break; } + } } void InstrumentationMapExtractor::exportAsYAML(raw_ostream &OS) { diff --git a/tools/llvm-xray/xray-record-yaml.h b/tools/llvm-xray/xray-record-yaml.h new file mode 100644 index 0000000000000..abce8ff60a94d --- /dev/null +++ b/tools/llvm-xray/xray-record-yaml.h @@ -0,0 +1,102 @@ +//===- xray-record-yaml.h - XRay Record YAML Support Definitions ----------===// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// +// +// Types and traits specialisations for YAML I/O of XRay log entries. +// +//===----------------------------------------------------------------------===// +#ifndef LLVM_TOOLS_LLVM_XRAY_XRAY_RECORD_YAML_H +#define LLVM_TOOLS_LLVM_XRAY_XRAY_RECORD_YAML_H + +#include <type_traits> + +#include "xray-record.h" +#include "llvm/Support/YAMLTraits.h" + +namespace llvm { +namespace xray { + +struct YAMLXRayFileHeader { + uint16_t Version; + uint16_t Type; + bool ConstantTSC; + bool NonstopTSC; + uint64_t CycleFrequency; +}; + +struct YAMLXRayRecord { + uint16_t RecordType; + uint8_t CPU; + RecordTypes Type; + int32_t FuncId; + std::string Function; + uint64_t TSC; + uint32_t TId; +}; + +struct YAMLXRayTrace { + YAMLXRayFileHeader Header; + std::vector<YAMLXRayRecord> Records; +}; + +using XRayRecordStorage = + std::aligned_storage<sizeof(XRayRecord), alignof(XRayRecord)>::type; + +} // namespace xray + +namespace yaml { + +// YAML Traits +// ----------- +template <> struct ScalarEnumerationTraits<xray::RecordTypes> { + static void enumeration(IO &IO, xray::RecordTypes &Type) { + IO.enumCase(Type, "function-enter", xray::RecordTypes::ENTER); + IO.enumCase(Type, "function-exit", xray::RecordTypes::EXIT); + } +}; + +template <> struct MappingTraits<xray::YAMLXRayFileHeader> { + static void mapping(IO &IO, xray::YAMLXRayFileHeader &Header) { + IO.mapRequired("version", Header.Version); + IO.mapRequired("type", Header.Type); + IO.mapRequired("constant-tsc", Header.ConstantTSC); + IO.mapRequired("nonstop-tsc", Header.NonstopTSC); + IO.mapRequired("cycle-frequency", Header.CycleFrequency); + } +}; + +template <> struct MappingTraits<xray::YAMLXRayRecord> { + static void mapping(IO &IO, xray::YAMLXRayRecord &Record) { + // FIXME: Make this type actually be descriptive + IO.mapRequired("type", Record.RecordType); + IO.mapRequired("func-id", Record.FuncId); + IO.mapOptional("function", Record.Function); + IO.mapRequired("cpu", Record.CPU); + IO.mapRequired("thread", Record.TId); + IO.mapRequired("kind", Record.Type); + IO.mapRequired("tsc", Record.TSC); + } + + static constexpr bool flow = true; +}; + +template <> struct MappingTraits<xray::YAMLXRayTrace> { + static void mapping(IO &IO, xray::YAMLXRayTrace &Trace) { + // A trace file contains two parts, the header and the list of all the + // trace records. + IO.mapRequired("header", Trace.Header); + IO.mapRequired("records", Trace.Records); + } +}; + +} // namespace yaml +} // namespace llvm + +LLVM_YAML_IS_SEQUENCE_VECTOR(xray::YAMLXRayRecord) + +#endif // LLVM_TOOLS_LLVM_XRAY_XRAY_RECORD_YAML_H |