[OutputBackend] Make adding outputs to HashingOutputBackend thread-safe

Make adding outputs to `HashingOutputBackend` thread-safe. This is needed
when compiler can write multiple outputs at the same time to the output
backend and StringMap can lose output or crash during rehashing.
Iterating and checking the hashing outputs still requires
synchronization and is not thread-safe.

Author: cachemeifyoucan

llvm/include/llvm/Support/HashingOutputBackend.h

@@ -11,11 +11,11 @@
 
 #include "llvm/ADT/StringExtras.h"
 #include "llvm/ADT/StringMap.h"
-#include "llvm/Support/Endian.h"
 #include "llvm/Support/HashBuilder.h"
 #include "llvm/Support/VirtualOutputBackend.h"
 #include "llvm/Support/VirtualOutputConfig.h"
 #include "llvm/Support/raw_ostream.h"
+#include <mutex>
 
 namespace llvm {
 namespace vfs {
@@ -56,6 +56,7 @@ template <typename HasherT> class HashingOutputBackend : public OutputBackend {
 private:
   friend class HashingOutputFile<HasherT>;
   void addOutputFile(StringRef Path, StringRef Hash) {
+    std::lock_guard<std::mutex> Lock(OutputHashLock);
     OutputHashes[Path] = std::string(Hash);
   }
 
@@ -83,6 +84,7 @@ template <typename HasherT> class HashingOutputBackend : public OutputBackend {
   }
 
 private:
+  std::mutex OutputHashLock;
   StringMap<std::string> OutputHashes;
 };
 

llvm/unittests/Support/VirtualOutputBackendsTest.cpp

@@ -11,6 +11,8 @@
 #include "llvm/Support/BLAKE3.h"
 #include "llvm/Support/HashingOutputBackend.h"
 #include "llvm/Support/MemoryBuffer.h"
+#include "llvm/Support/ThreadPool.h"
+#include "llvm/Support/raw_ostream.h"
 #include "llvm/Testing/Support/Error.h"
 #include "gtest/gtest.h"
 
@@ -883,4 +885,33 @@ TEST(HashingBackendTest, HashOutput) {
             Backend.getHashValueForFile("file5"));
 }
 
+TEST(HashingBackendTest, ParallelHashOutput) {
+  const unsigned NumFile = 20;
+  DefaultThreadPool Pool;
+  HashingOutputBackend<BLAKE3> Backend;
+  auto getFileName = [](unsigned Idx) -> std::string {
+    std::string Name;
+    raw_string_ostream OS(Name);
+    OS << "file" << Idx;
+    return Name;
+  };
+  for (unsigned I = 0; I < NumFile; ++I) {
+    Pool.async(
+        [&](unsigned Idx) {
+          OutputFile O;
+          auto Name = getFileName(Idx);
+          EXPECT_THAT_ERROR(Backend.createFile(Name).moveInto(O), Succeeded());
+          O << "some data" << Idx;
+          EXPECT_THAT_ERROR(O.keep(), Succeeded());
+        },
+        I);
+  }
+  Pool.wait();
+
+  for (unsigned I = 0; I < NumFile; ++I) {
+    auto Name = getFileName(I);
+    auto Hash = Backend.getHashValueForFile(Name);
+    EXPECT_TRUE(Hash);
+  }
+}
 } // end namespace