[llvm-profgen] Deduplicate PID when processing perf input

[wenlei]

Summary

When parsing mmap to retrieve PID, deduplicate them before passing PID list to perf script. Perf script would error out when there's duplicated PID in the input, however raw perf data may main duplicated PID for large binary where more than one mmap is needed to load executable segment.

[wlei]

LGTM, thanks!

[hoy]

lgtm.

llvm/tools/llvm-profgen/PerfReader.cpp

std::string PerfReaderBase::convertPerfDataToTrace(ProfiledBinary *Binary,
   Optional<StringRef> Redirects[] = {llvm::None,                // Stdin
                                      StringRef(PerfTraceFile),  // Stdout
                                      StringRef(PerfTraceFile)}; // Stderr
   sys::ExecuteAndWait(PerfPath, ScriptMMapArgs, llvm::None, Redirects);
 
   // Collect the PIDs
   TraceStream TraceIt(PerfTraceFile);
   std::string PIDs;
+  std::unordered_set<uint32_t> PIDSet;
   while (!TraceIt.isAtEoF()) {
     MMapEvent MMap;
     if (isMMap2Event(TraceIt.getCurrentLine()) &&
         extractMMap2EventForBinary(Binary, TraceIt.getCurrentLine(), MMap)) {
+      auto It = PIDSet.emplace(MMap.PID);
+      if (It.second) {
         if (!PIDs.empty()) {
           PIDs.append(",");
         }
         PIDs.append(utostr(MMap.PID));
       }
+    }
     TraceIt.advance();
   }
 
+  if (PIDs.empty()) {
+    exitWithError("No relevant mmap event is found in perf data.");
+  }
+
   // Run perf script again to retrieve events for PIDs collected above
   StringRef ScriptSampleArgs[] = {PerfPath, "script",     "--show-mmap-events",
                                   "-F",     "ip,brstack", "--pid",
                                   PIDs,     "-i",         PerfData};
   sys::ExecuteAndWait(PerfPath, ScriptSampleArgs, llvm::None, Redirects);
 
   return PerfTraceFile;
 }