1 //===- FuzzerFork.cpp - run fuzzing in separate subprocesses --------------===//
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 // Spawn and orchestrate separate fuzzing processes.
9 //===----------------------------------------------------------------------===//
10 
11 #include "FuzzerCommand.h"
12 #include "FuzzerFork.h"
13 #include "FuzzerIO.h"
14 #include "FuzzerInternal.h"
15 #include "FuzzerMerge.h"
16 #include "FuzzerSHA1.h"
17 #include "FuzzerTracePC.h"
18 #include "FuzzerUtil.h"
19 
20 #include <atomic>
21 #include <chrono>
22 #include <fstream>
23 #include <memory>
24 #include <mutex>
25 #include <queue>
26 #include <sstream>
27 #include <thread>
28 
29 namespace fuzzer {
30 
31 struct Stats {
32   size_t number_of_executed_units = 0;
33   size_t peak_rss_mb = 0;
34   size_t average_exec_per_sec = 0;
35 };
36 
37 static Stats ParseFinalStatsFromLog(const std::string &LogPath) {
38   std::ifstream In(LogPath);
39   std::string Line;
40   Stats Res;
41   struct {
42     const char *Name;
43     size_t *Var;
44   } NameVarPairs[] = {
45       {"stat::number_of_executed_units:", &Res.number_of_executed_units},
46       {"stat::peak_rss_mb:", &Res.peak_rss_mb},
47       {"stat::average_exec_per_sec:", &Res.average_exec_per_sec},
48       {nullptr, nullptr},
49   };
50   while (std::getline(In, Line, '\n')) {
51     if (Line.find("stat::") != 0) continue;
52     std::istringstream ISS(Line);
53     std::string Name;
54     size_t Val;
55     ISS >> Name >> Val;
56     for (size_t i = 0; NameVarPairs[i].Name; i++)
57       if (Name == NameVarPairs[i].Name)
58         *NameVarPairs[i].Var = Val;
59   }
60   return Res;
61 }
62 
63 struct FuzzJob {
64   // Inputs.
65   Command Cmd;
66   std::string CorpusDir;
67   std::string FeaturesDir;
68   std::string LogPath;
69   std::string SeedListPath;
70   std::string CFPath;
71 
72   // Fuzzing Outputs.
73   int ExitCode;
74 
75   ~FuzzJob() {
76     RemoveFile(CFPath);
77     RemoveFile(LogPath);
78     RemoveFile(SeedListPath);
79     RmDirRecursive(CorpusDir);
80     RmDirRecursive(FeaturesDir);
81   }
82 };
83 
84 struct GlobalEnv {
85   Vector<std::string> Args;
86   Vector<std::string> CorpusDirs;
87   std::string MainCorpusDir;
88   std::string TempDir;
89   Set<uint32_t> Features, Cov;
90   Vector<std::string> Files;
91   Random *Rand;
92   std::chrono::system_clock::time_point ProcessStartTime;
93   int Verbosity = 0;
94 
95   size_t NumTimeouts = 0;
96   size_t NumOOMs = 0;
97   size_t NumCrashes = 0;
98 
99 
100   size_t NumRuns = 0;
101 
102   size_t secondsSinceProcessStartUp() const {
103     return std::chrono::duration_cast<std::chrono::seconds>(
104                std::chrono::system_clock::now() - ProcessStartTime)
105         .count();
106   }
107 
108   FuzzJob *CreateNewJob(size_t JobId) {
109     Command Cmd(Args);
110     Cmd.removeFlag("fork");
111     Cmd.removeFlag("runs");
112     for (auto &C : CorpusDirs) // Remove all corpora from the args.
113       Cmd.removeArgument(C);
114     Cmd.addFlag("reload", "0");  // working in an isolated dir, no reload.
115     Cmd.addFlag("print_final_stats", "1");
116     Cmd.addFlag("print_funcs", "0");  // no need to spend time symbolizing.
117     Cmd.addFlag("max_total_time", std::to_string(std::min((size_t)300, JobId)));
118 
119     auto Job = new FuzzJob;
120     std::string Seeds;
121     if (size_t CorpusSubsetSize =
122             std::min(Files.size(), (size_t)sqrt(Files.size() + 2)))
123       for (size_t i = 0; i < CorpusSubsetSize; i++)
124         Seeds += (Seeds.empty() ? "" : ",") +
125                  Files[Rand->SkewTowardsLast(Files.size())];
126     if (!Seeds.empty()) {
127       Job->SeedListPath = std::to_string(JobId) + ".seeds";
128       WriteToFile(Seeds, Job->SeedListPath);
129       Cmd.addFlag("seed_inputs", "@" + Job->SeedListPath);
130     }
131     Job->LogPath = DirPlusFile(TempDir, std::to_string(JobId) + ".log");
132     Job->CorpusDir = DirPlusFile(TempDir, "C" + std::to_string(JobId));
133     Job->FeaturesDir = DirPlusFile(TempDir, "F" + std::to_string(JobId));
134     Job->CFPath = DirPlusFile(TempDir, std::to_string(JobId) + ".merge");
135 
136 
137     Cmd.addArgument(Job->CorpusDir);
138     Cmd.addFlag("features_dir", Job->FeaturesDir);
139 
140     for (auto &D : {Job->CorpusDir, Job->FeaturesDir}) {
141       RmDirRecursive(D);
142       MkDir(D);
143     }
144 
145     Cmd.setOutputFile(Job->LogPath);
146     Cmd.combineOutAndErr();
147 
148     Job->Cmd = Cmd;
149 
150     if (Verbosity >= 2)
151       Printf("Job %zd/%p Created: %s\n", JobId, Job,
152              Job->Cmd.toString().c_str());
153     // Start from very short runs and gradually increase them.
154     return Job;
155   }
156 
157   void RunOneMergeJob(FuzzJob *Job) {
158     auto Stats = ParseFinalStatsFromLog(Job->LogPath);
159     NumRuns += Stats.number_of_executed_units;
160 
161     Vector<SizedFile> TempFiles, MergeCandidates;
162     // Read all newly created inputs and their feature sets.
163     // Choose only those inputs that have new features.
164     GetSizedFilesFromDir(Job->CorpusDir, &TempFiles);
165     std::sort(TempFiles.begin(), TempFiles.end());
166     for (auto &F : TempFiles) {
167       auto FeatureFile = F.File;
168       FeatureFile.replace(0, Job->CorpusDir.size(), Job->FeaturesDir);
169       auto FeatureBytes = FileToVector(FeatureFile, 0, false);
170       assert((FeatureBytes.size() % sizeof(uint32_t)) == 0);
171       Vector<uint32_t> NewFeatures(FeatureBytes.size() / sizeof(uint32_t));
172       memcpy(NewFeatures.data(), FeatureBytes.data(), FeatureBytes.size());
173       for (auto Ft : NewFeatures) {
174         if (!Features.count(Ft)) {
175           MergeCandidates.push_back(F);
176           break;
177         }
178       }
179     }
180     if (MergeCandidates.empty()) return;
181 
182     Vector<std::string> FilesToAdd;
183     Set<uint32_t> NewFeatures, NewCov;
184     CrashResistantMerge(Args, {}, MergeCandidates, &FilesToAdd, Features,
185                         &NewFeatures, Cov, &NewCov, Job->CFPath, false);
186     for (auto &Path : FilesToAdd) {
187       auto U = FileToVector(Path);
188       auto NewPath = DirPlusFile(MainCorpusDir, Hash(U));
189       WriteToFile(U, NewPath);
190       Files.push_back(NewPath);
191     }
192     Features.insert(NewFeatures.begin(), NewFeatures.end());
193     Cov.insert(NewCov.begin(), NewCov.end());
194     for (auto Idx : NewCov)
195       if (auto *TE = TPC.PCTableEntryByIdx(Idx))
196         if (TPC.PcIsFuncEntry(TE))
197           PrintPC("  NEW_FUNC: %p %F %L\n", "",
198                   TPC.GetNextInstructionPc(TE->PC));
199 
200     if (!FilesToAdd.empty() || Job->ExitCode != 0)
201       Printf("#%zd: cov: %zd ft: %zd corp: %zd exec/s %zd "
202              "oom/timeout/crash: %zd/%zd/%zd time: %zds\n", NumRuns,
203              Cov.size(), Features.size(), Files.size(),
204              Stats.average_exec_per_sec,
205              NumOOMs, NumTimeouts, NumCrashes, secondsSinceProcessStartUp());
206   }
207 };
208 
209 struct JobQueue {
210   std::queue<FuzzJob *> Qu;
211   std::mutex Mu;
212 
213   void Push(FuzzJob *Job) {
214     std::lock_guard<std::mutex> Lock(Mu);
215     Qu.push(Job);
216   }
217   FuzzJob *Pop() {
218     std::lock_guard<std::mutex> Lock(Mu);
219     if (Qu.empty()) return nullptr;
220     auto Job = Qu.front();
221     Qu.pop();
222     return Job;
223   }
224 };
225 
226 void WorkerThread(std::atomic<bool> *Stop, JobQueue *FuzzQ, JobQueue *MergeQ) {
227   while (!Stop->load()) {
228     auto Job = FuzzQ->Pop();
229     // Printf("WorkerThread: job %p\n", Job);
230     if (!Job) {
231       SleepSeconds(1);
232       continue;
233     }
234     Job->ExitCode = ExecuteCommand(Job->Cmd);
235     MergeQ->Push(Job);
236   }
237 }
238 
239 // This is just a skeleton of an experimental -fork=1 feature.
240 void FuzzWithFork(Random &Rand, const FuzzingOptions &Options,
241                   const Vector<std::string> &Args,
242                   const Vector<std::string> &CorpusDirs, int NumJobs) {
243   Printf("INFO: -fork=%d: fuzzing in separate process(s)\n", NumJobs);
244 
245   GlobalEnv Env;
246   Env.Args = Args;
247   Env.CorpusDirs = CorpusDirs;
248   Env.Rand = &Rand;
249   Env.Verbosity = Options.Verbosity;
250   Env.ProcessStartTime = std::chrono::system_clock::now();
251 
252   Vector<SizedFile> SeedFiles;
253   for (auto &Dir : CorpusDirs)
254     GetSizedFilesFromDir(Dir, &SeedFiles);
255   std::sort(SeedFiles.begin(), SeedFiles.end());
256   Env.TempDir = TempPath(".dir");
257   RmDirRecursive(Env.TempDir);  // in case there is a leftover from old runs.
258   MkDir(Env.TempDir);
259 
260 
261   if (CorpusDirs.empty())
262     MkDir(Env.MainCorpusDir = DirPlusFile(Env.TempDir, "C"));
263   else
264     Env.MainCorpusDir = CorpusDirs[0];
265 
266   auto CFPath = DirPlusFile(Env.TempDir, "merge.txt");
267   CrashResistantMerge(Env.Args, {}, SeedFiles, &Env.Files, {}, &Env.Features,
268                       {}, &Env.Cov,
269                       CFPath, false);
270   RemoveFile(CFPath);
271   Printf("INFO: -fork=%d: %zd seed inputs, starting to fuzz in %s\n", NumJobs,
272          Env.Files.size(), Env.TempDir.c_str());
273 
274   int ExitCode = 0;
275 
276   JobQueue FuzzQ, MergeQ;
277   std::atomic<bool> Stop(false);
278 
279   size_t JobId = 1;
280   Vector<std::thread> Threads;
281   for (int t = 0; t < NumJobs; t++) {
282     Threads.push_back(std::thread(WorkerThread, &Stop, &FuzzQ, &MergeQ));
283     FuzzQ.Push(Env.CreateNewJob(JobId++));
284   }
285 
286   while (true) {
287     std::unique_ptr<FuzzJob> Job(MergeQ.Pop());
288     if (!Job) {
289       if (Stop)
290         break;
291       SleepSeconds(1);
292       continue;
293     }
294     ExitCode = Job->ExitCode;
295     if (ExitCode == Options.InterruptExitCode) {
296       Printf("==%lu== libFuzzer: a child was interrupted; exiting\n", GetPid());
297       Stop = true;
298       break;
299     }
300     Fuzzer::MaybeExitGracefully();
301 
302     Env.RunOneMergeJob(Job.get());
303 
304     // Continue if our crash is one of the ignorred ones.
305     if (Options.IgnoreTimeouts && ExitCode == Options.TimeoutExitCode)
306       Env.NumTimeouts++;
307     else if (Options.IgnoreOOMs && ExitCode == Options.OOMExitCode)
308       Env.NumOOMs++;
309     else if (ExitCode != 0) {
310       Env.NumCrashes++;
311       if (Options.IgnoreCrashes) {
312         std::ifstream In(Job->LogPath);
313         std::string Line;
314         while (std::getline(In, Line, '\n'))
315           if (Line.find("ERROR:") != Line.npos ||
316               Line.find("runtime error:") != Line.npos)
317             Printf("%s\n", Line.c_str());
318       } else {
319         // And exit if we don't ignore this crash.
320         Printf("INFO: log from the inner process:\n%s",
321                FileToString(Job->LogPath).c_str());
322         Stop = true;
323       }
324     }
325 
326     // Stop if we are over the time budget.
327     // This is not precise, since other threads are still running
328     // and we will wait while joining them.
329     // We also don't stop instantly: other jobs need to finish.
330     if (Options.MaxTotalTimeSec > 0 && !Stop &&
331         Env.secondsSinceProcessStartUp() >= (size_t)Options.MaxTotalTimeSec) {
332       Printf("INFO: fuzzed for %zd seconds, wrapping up soon\n",
333              Env.secondsSinceProcessStartUp());
334       Stop = true;
335     }
336     if (!Stop && Env.NumRuns >= Options.MaxNumberOfRuns) {
337       Printf("INFO: fuzzed for %zd iterations, wrapping up soon\n",
338              Env.NumRuns);
339       Stop = true;
340     }
341 
342     if (!Stop)
343       FuzzQ.Push(Env.CreateNewJob(JobId++));
344   }
345   Stop = true;
346 
347   for (auto &T : Threads)
348     T.join();
349 
350   // The workers have terminated. Don't try to remove the directory before they
351   // terminate to avoid a race condition preventing cleanup on Windows.
352   RmDirRecursive(Env.TempDir);
353 
354   // Use the exit code from the last child process.
355   Printf("INFO: exiting: %d time: %zds\n", ExitCode,
356          Env.secondsSinceProcessStartUp());
357   exit(ExitCode);
358 }
359 
360 } // namespace fuzzer
361