| 1 | //===- EntryPointStats.cpp --------------------------------------*- C++ -*-===// |
| 2 | // |
| 3 | // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions. |
| 4 | // See https://llvm.org/LICENSE.txt for license information. |
| 5 | // SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception |
| 6 | // |
| 7 | //===----------------------------------------------------------------------===// |
| 8 | |
| 9 | #include "clang/StaticAnalyzer/Core/PathSensitive/EntryPointStats.h" |
| 10 | #include "clang/AST/DeclBase.h" |
| 11 | #include "clang/Analysis/AnalysisDeclContext.h" |
| 12 | #include "llvm/ADT/STLExtras.h" |
| 13 | #include "llvm/ADT/StringExtras.h" |
| 14 | #include "llvm/ADT/StringRef.h" |
| 15 | #include "llvm/Support/FileSystem.h" |
| 16 | #include "llvm/Support/ManagedStatic.h" |
| 17 | #include "llvm/Support/raw_ostream.h" |
| 18 | #include <iterator> |
| 19 | |
| 20 | using namespace clang; |
| 21 | using namespace ento; |
| 22 | |
| 23 | namespace { |
| 24 | struct Registry { |
| 25 | std::vector<BoolEPStat *> BoolStats; |
| 26 | std::vector<CounterEPStat *> CounterStats; |
| 27 | std::vector<UnsignedMaxEPStat *> UnsignedMaxStats; |
| 28 | std::vector<UnsignedEPStat *> UnsignedStats; |
| 29 | |
| 30 | bool IsLocked = false; |
| 31 | |
| 32 | struct Snapshot { |
| 33 | const Decl *EntryPoint; |
| 34 | std::vector<bool> BoolStatValues; |
| 35 | std::vector<unsigned> UnsignedStatValues; |
| 36 | |
| 37 | void dumpAsCSV(llvm::raw_ostream &OS) const; |
| 38 | }; |
| 39 | |
| 40 | std::vector<Snapshot> Snapshots; |
| 41 | }; |
| 42 | } // namespace |
| 43 | |
| 44 | static llvm::ManagedStatic<Registry> StatsRegistry; |
| 45 | |
| 46 | namespace { |
| 47 | template <typename Callback> void enumerateStatVectors(const Callback &Fn) { |
| 48 | Fn(StatsRegistry->BoolStats); |
| 49 | Fn(StatsRegistry->CounterStats); |
| 50 | Fn(StatsRegistry->UnsignedMaxStats); |
| 51 | Fn(StatsRegistry->UnsignedStats); |
| 52 | } |
| 53 | } // namespace |
| 54 | |
| 55 | static void checkStatName(const EntryPointStat *M) { |
| 56 | #ifdef NDEBUG |
| 57 | return; |
| 58 | #endif // NDEBUG |
| 59 | constexpr std::array AllowedSpecialChars = { |
| 60 | '+', '-', '_', '=', ':', '(', ')', '@', '!', '~', |
| 61 | '$', '%', '^', '&', '*', '\'', ';', '<', '>', '/'}; |
| 62 | for (unsigned char C : M->name()) { |
| 63 | if (!std::isalnum(C) && !llvm::is_contained(Range: AllowedSpecialChars, Element: C)) { |
| 64 | llvm::errs() << "Stat name \"" << M->name() << "\" contains character '" |
| 65 | << C << "' (" << static_cast<int>(C) |
| 66 | << ") that is not allowed." ; |
| 67 | assert(false && "The Stat name contains unallowed character" ); |
| 68 | } |
| 69 | } |
| 70 | } |
| 71 | |
| 72 | void EntryPointStat::lockRegistry() { |
| 73 | auto CmpByNames = [](const EntryPointStat *L, const EntryPointStat *R) { |
| 74 | return L->name() < R->name(); |
| 75 | }; |
| 76 | enumerateStatVectors( |
| 77 | Fn: [CmpByNames](auto &Stats) { llvm::sort(Stats, CmpByNames); }); |
| 78 | enumerateStatVectors( |
| 79 | Fn: [](const auto &Stats) { llvm::for_each(Stats, checkStatName); }); |
| 80 | StatsRegistry->IsLocked = true; |
| 81 | } |
| 82 | |
| 83 | [[maybe_unused]] static bool isRegistered(llvm::StringLiteral Name) { |
| 84 | auto ByName = [Name](const EntryPointStat *M) { return M->name() == Name; }; |
| 85 | bool Result = false; |
| 86 | enumerateStatVectors(Fn: [ByName, &Result](const auto &Stats) { |
| 87 | Result = Result || llvm::any_of(Stats, ByName); |
| 88 | }); |
| 89 | return Result; |
| 90 | } |
| 91 | |
| 92 | BoolEPStat::BoolEPStat(llvm::StringLiteral Name) : EntryPointStat(Name) { |
| 93 | assert(!StatsRegistry->IsLocked); |
| 94 | assert(!isRegistered(Name)); |
| 95 | StatsRegistry->BoolStats.push_back(x: this); |
| 96 | } |
| 97 | |
| 98 | CounterEPStat::CounterEPStat(llvm::StringLiteral Name) : EntryPointStat(Name) { |
| 99 | assert(!StatsRegistry->IsLocked); |
| 100 | assert(!isRegistered(Name)); |
| 101 | StatsRegistry->CounterStats.push_back(x: this); |
| 102 | } |
| 103 | |
| 104 | UnsignedMaxEPStat::UnsignedMaxEPStat(llvm::StringLiteral Name) |
| 105 | : EntryPointStat(Name) { |
| 106 | assert(!StatsRegistry->IsLocked); |
| 107 | assert(!isRegistered(Name)); |
| 108 | StatsRegistry->UnsignedMaxStats.push_back(x: this); |
| 109 | } |
| 110 | |
| 111 | UnsignedEPStat::UnsignedEPStat(llvm::StringLiteral Name) |
| 112 | : EntryPointStat(Name) { |
| 113 | assert(!StatsRegistry->IsLocked); |
| 114 | assert(!isRegistered(Name)); |
| 115 | StatsRegistry->UnsignedStats.push_back(x: this); |
| 116 | } |
| 117 | |
| 118 | static std::vector<unsigned> consumeUnsignedStats() { |
| 119 | std::vector<unsigned> Result; |
| 120 | Result.reserve(n: StatsRegistry->CounterStats.size() + |
| 121 | StatsRegistry->UnsignedMaxStats.size() + |
| 122 | StatsRegistry->UnsignedStats.size()); |
| 123 | for (auto *M : StatsRegistry->CounterStats) { |
| 124 | Result.push_back(x: M->value()); |
| 125 | M->reset(); |
| 126 | } |
| 127 | for (auto *M : StatsRegistry->UnsignedMaxStats) { |
| 128 | Result.push_back(x: M->value()); |
| 129 | M->reset(); |
| 130 | } |
| 131 | for (auto *M : StatsRegistry->UnsignedStats) { |
| 132 | Result.push_back(x: M->value()); |
| 133 | M->reset(); |
| 134 | } |
| 135 | return Result; |
| 136 | } |
| 137 | |
| 138 | static std::vector<llvm::StringLiteral> getStatNames() { |
| 139 | std::vector<llvm::StringLiteral> Ret; |
| 140 | auto GetName = [](const EntryPointStat *M) { return M->name(); }; |
| 141 | enumerateStatVectors(Fn: [GetName, &Ret](const auto &Stats) { |
| 142 | transform(Stats, std::back_inserter(x&: Ret), GetName); |
| 143 | }); |
| 144 | return Ret; |
| 145 | } |
| 146 | |
| 147 | void Registry::Snapshot::dumpAsCSV(llvm::raw_ostream &OS) const { |
| 148 | OS << '"'; |
| 149 | llvm::printEscapedString( |
| 150 | Name: clang::AnalysisDeclContext::getFunctionName(D: EntryPoint), Out&: OS); |
| 151 | OS << "\", " ; |
| 152 | auto PrintAsBool = [&OS](bool B) { OS << (B ? "true" : "false" ); }; |
| 153 | llvm::interleaveComma(c: BoolStatValues, os&: OS, each_fn: PrintAsBool); |
| 154 | OS << ((BoolStatValues.empty() || UnsignedStatValues.empty()) ? "" : ", " ); |
| 155 | llvm::interleaveComma(c: UnsignedStatValues, os&: OS); |
| 156 | } |
| 157 | |
| 158 | static std::vector<bool> consumeBoolStats() { |
| 159 | std::vector<bool> Result; |
| 160 | Result.reserve(n: StatsRegistry->BoolStats.size()); |
| 161 | for (auto *M : StatsRegistry->BoolStats) { |
| 162 | Result.push_back(x: M->value()); |
| 163 | M->reset(); |
| 164 | } |
| 165 | return Result; |
| 166 | } |
| 167 | |
| 168 | void EntryPointStat::takeSnapshot(const Decl *EntryPoint) { |
| 169 | auto BoolValues = consumeBoolStats(); |
| 170 | auto UnsignedValues = consumeUnsignedStats(); |
| 171 | StatsRegistry->Snapshots.push_back( |
| 172 | x: {.EntryPoint: EntryPoint, .BoolStatValues: std::move(BoolValues), .UnsignedStatValues: std::move(UnsignedValues)}); |
| 173 | } |
| 174 | |
| 175 | void EntryPointStat::dumpStatsAsCSV(llvm::StringRef FileName) { |
| 176 | std::error_code EC; |
| 177 | llvm::raw_fd_ostream File(FileName, EC, llvm::sys::fs::OF_Text); |
| 178 | if (EC) |
| 179 | return; |
| 180 | dumpStatsAsCSV(OS&: File); |
| 181 | } |
| 182 | |
| 183 | void EntryPointStat::dumpStatsAsCSV(llvm::raw_ostream &OS) { |
| 184 | OS << "EntryPoint, " ; |
| 185 | llvm::interleaveComma(c: getStatNames(), os&: OS); |
| 186 | OS << "\n" ; |
| 187 | |
| 188 | std::vector<std::string> Rows; |
| 189 | Rows.reserve(n: StatsRegistry->Snapshots.size()); |
| 190 | for (const auto &Snapshot : StatsRegistry->Snapshots) { |
| 191 | std::string Row; |
| 192 | llvm::raw_string_ostream RowOs(Row); |
| 193 | Snapshot.dumpAsCSV(OS&: RowOs); |
| 194 | RowOs << "\n" ; |
| 195 | Rows.push_back(x: RowOs.str()); |
| 196 | } |
| 197 | llvm::sort(C&: Rows); |
| 198 | for (const auto &Row : Rows) { |
| 199 | OS << Row; |
| 200 | } |
| 201 | } |
| 202 | |