[XRay][compiler-rt] FDRLogWriter Abstraction

Summary:
This change introduces an `FDRLogWriter` type which is responsible for
serialising metadata and function records to character buffers. This is
the first step in a refactoring of the implementation of the FDR runtime
to allow for more granular testing of the individual components of the
implementation.

The main contribution of this change is a means of hiding the details of
how specific records are written to a buffer, and for managing the
extents of these buffers. We make use of C++ features (templates and
some metaprogramming) to reduce repetition in the act of writing out
specific kinds of records to the buffer.

In this process, we make a number of changes across both LLVM and
compiler-rt to allow us to use the `Trace` abstraction defined in the
LLVM project in the testing of the runtime implementation. This gives us
a closer end-to-end test which version-locks the runtime implementation
with the loading implementation in LLVM.

We also allow using gmock in compiler-rt unit tests, by adding the
requisite definitions in the `AddCompilerRT.cmake` module.

Finally, we've gone ahead and updated the FDR logging implementation to
use the FDRLogWriter for the lowest-level record-writing details.

Following patches will isolate the state machine transitions which
manage the set-up and tear-down of the buffers we're using in multiple
threads.

Reviewers: mboerger, eizan

Subscribers: mgorny, jfb, llvm-commits

Differential Revision: https://reviews.llvm.org/D52220

git-svn-id: https://llvm.org/svn/llvm-project/compiler-rt/trunk@342518 91177308-0d34-0410-b5e6-96231b3b80d8

6 files changed, 279 insertions, 87 deletions

diff --git a/lib/xray/tests/CMakeLists.txt b/lib/xray/tests/CMakeLists.txt
index 11f373167..6e12297f9 100644
--- a/lib/xray/tests/CMakeLists.txt
+++ b/lib/xray/tests/CMakeLists.txt
@@ -19,9 +19,13 @@ set(XRAY_UNITTEST_CFLAGS
   ${XRAY_CFLAGS}
   ${COMPILER_RT_UNITTEST_CFLAGS}
   ${COMPILER_RT_GTEST_CFLAGS}
+  ${COMPILER_RT_GMOCK_CFLAGS}
   -I${COMPILER_RT_SOURCE_DIR}/include
   -I${COMPILER_RT_SOURCE_DIR}/lib/xray
-  -I${COMPILER_RT_SOURCE_DIR}/lib)
+  -I${COMPILER_RT_SOURCE_DIR}/lib
+  -I${LLVM_MAIN_SRC_DIR}/include
+  -I${LLVM_INCLUDE_DIR}
+  )
 
 function(add_xray_lib library)
   add_library(${library} STATIC ${ARGN})
@@ -68,11 +72,14 @@ macro(add_xray_unittest testname)
         COMPILE_DEPS ${TEST_SOURCES} ${COMPILER_RT_GTEST_SOURCE}
         ${XRAY_HEADERS} ${XRAY_ALL_SOURCE_FILES_ABS_PATHS}
         RUNTIME "${XRAY_RUNTIME_LIBS}"
-        DEPS gtest xray llvm-xray
+        DEPS gtest xray llvm-xray LLVMXRay LLVMTestingSupport
         CFLAGS ${XRAY_UNITTEST_CFLAGS}
-        LINK_FLAGS ${TARGET_LINK_FLAGS} ${XRAY_UNITTEST_LINK_FLAGS})
+        LINK_FLAGS ${TARGET_LINK_FLAGS} ${XRAY_UNITTEST_LINK_FLAGS}
+        -lLLVMXRay -lLLVMSupport -lLLVMTestingSupport -ltinfo
+        )
       set_target_properties(XRayUnitTests
-        PROPERTIES RUNTIME_OUTPUT_DIRECTORY ${CMAKE_CURRENT_BINARY_DIR})
+        PROPERTIES
+        RUNTIME_OUTPUT_DIRECTORY ${CMAKE_CURRENT_BINARY_DIR})
     endforeach()
   endif()
 endmacro()
diff --git a/lib/xray/tests/unit/CMakeLists.txt b/lib/xray/tests/unit/CMakeLists.txt
index 63c46e41c..b1ead627e 100644
--- a/lib/xray/tests/unit/CMakeLists.txt
+++ b/lib/xray/tests/unit/CMakeLists.txt
@@ -8,3 +8,7 @@ add_xray_unittest(XRayFunctionCallTrieTest SOURCES
   function_call_trie_test.cc xray_unit_test_main.cc)
 add_xray_unittest(XRayProfileCollectorTest SOURCES
   profile_collector_test.cc xray_unit_test_main.cc)
+
+add_xray_unittest(XRayFDRLoggingTest SOURCES
+  fdr_log_writer_test.cc
+  xray_unit_test_main.cc)
diff --git a/lib/xray/tests/unit/fdr_log_writer_test.cc b/lib/xray/tests/unit/fdr_log_writer_test.cc
new file mode 100644
index 000000000..3a2138cd8
--- /dev/null
+++ b/lib/xray/tests/unit/fdr_log_writer_test.cc
@@ -0,0 +1,92 @@
+//===-- fdr_log_writer_test.cc --------------------------------------------===//
+//
+//                     The LLVM Compiler Infrastructure
+//
+// This file is distributed under the University of Illinois Open Source
+// License. See LICENSE.TXT for details.
+//
+//===----------------------------------------------------------------------===//
+//
+// This file is a part of XRay, a function call tracing system.
+//
+//===----------------------------------------------------------------------===//
+#include <time.h>
+
+#include "xray/xray_records.h"
+#include "xray_fdr_log_writer.h"
+#include "llvm/Support/DataExtractor.h"
+#include "llvm/Testing/Support/Error.h"
+#include "llvm/XRay/Trace.h"
+#include "gmock/gmock.h"
+#include "gtest/gtest.h"
+
+namespace __xray {
+namespace {
+
+static constexpr size_t kSize = 4096;
+
+using ::llvm::HasValue;
+using ::testing::Eq;
+using ::testing::SizeIs;
+
+// Exercise the common code path where we initialize a buffer and are able to
+// write some records successfully.
+TEST(FdrLogWriterTest, WriteSomeRecords) {
+  bool Success = false;
+  BufferQueue Buffers(kSize, 1, Success);
+  BufferQueue::Buffer B;
+  ASSERT_EQ(Buffers.getBuffer(B), BufferQueue::ErrorCode::Ok);
+
+  FDRLogWriter Writer(B);
+  MetadataRecord Preamble[] = {
+      createMetadataRecord<MetadataRecord::RecordKinds::NewBuffer>(int32_t{1}),
+      createMetadataRecord<MetadataRecord::RecordKinds::WalltimeMarker>(
+          int64_t{1}, int32_t{2}),
+      createMetadataRecord<MetadataRecord::RecordKinds::Pid>(int32_t{1}),
+  };
+  ASSERT_THAT(Writer.writeMetadataRecords(Preamble),
+              Eq(sizeof(MetadataRecord) * 3));
+  ASSERT_TRUE(Writer.writeMetadata<MetadataRecord::RecordKinds::NewCPUId>(1));
+  ASSERT_TRUE(
+      Writer.writeFunction(FDRLogWriter::FunctionRecordKind::Enter, 1, 1));
+  ASSERT_TRUE(
+      Writer.writeFunction(FDRLogWriter::FunctionRecordKind::Exit, 1, 1));
+  ASSERT_EQ(Buffers.releaseBuffer(B), BufferQueue::ErrorCode::Ok);
+  ASSERT_EQ(B.Data, nullptr);
+  ASSERT_EQ(Buffers.finalize(), BufferQueue::ErrorCode::Ok);
+
+  // We then need to go through each element of the Buffers, and re-create a
+  // flat buffer that we would see if they were laid out in a file. This also
+  // means we need to write out the header manually.
+  // TODO: Isolate the file header writing.
+  std::string Serialized;
+  std::aligned_storage<sizeof(XRayFileHeader), alignof(XRayFileHeader)>::type
+      HeaderStorage;
+  auto *Header = reinterpret_cast<XRayFileHeader *>(&HeaderStorage);
+  new (Header) XRayFileHeader();
+  Header->Version = 3;
+  Header->Type = FileTypes::FDR_LOG;
+  Header->CycleFrequency = 3e9;
+  Header->ConstantTSC = 1;
+  Header->NonstopTSC = 1;
+  Serialized.append(reinterpret_cast<const char *>(&HeaderStorage),
+                    sizeof(XRayFileHeader));
+  size_t BufferCount = 0;
+  Buffers.apply([&](const BufferQueue::Buffer &B) {
+    ++BufferCount;
+    auto Size = atomic_load_relaxed(&B.Extents);
+    auto Extents =
+        createMetadataRecord<MetadataRecord::RecordKinds::BufferExtents>(Size);
+    Serialized.append(reinterpret_cast<const char *>(&Extents),
+                      sizeof(Extents));
+    Serialized.append(reinterpret_cast<const char *>(B.Data), Size);
+  });
+  ASSERT_EQ(BufferCount, 1u);
+
+  llvm::DataExtractor DE(Serialized, true, 8);
+  auto TraceOrErr = llvm::xray::loadTrace(DE);
+  EXPECT_THAT_EXPECTED(TraceOrErr, HasValue(SizeIs(2)));
+}
+
+} // namespace
+} // namespace __xray
diff --git a/lib/xray/xray_fdr_log_records.h b/lib/xray/xray_fdr_log_records.h
index 87096d4fc..e7b1ee562 100644
--- a/lib/xray/xray_fdr_log_records.h
+++ b/lib/xray/xray_fdr_log_records.h
@@ -12,6 +12,9 @@
 //===----------------------------------------------------------------------===//
 #ifndef XRAY_XRAY_FDR_LOG_RECORDS_H
 #define XRAY_XRAY_FDR_LOG_RECORDS_H
+#include <cstdint>
+
+namespace __xray {
 
 enum class RecordType : uint8_t { Function, Metadata };
 
@@ -68,4 +71,6 @@ struct alignas(8) FunctionRecord {
 
 static_assert(sizeof(FunctionRecord) == 8, "Wrong size for FunctionRecord.");
 
+} // namespace __xray
+
 #endif // XRAY_XRAY_FDR_LOG_RECORDS_H
diff --git a/lib/xray/xray_fdr_log_writer.h b/lib/xray/xray_fdr_log_writer.h
new file mode 100644
index 000000000..28af356e7
--- /dev/null
+++ b/lib/xray/xray_fdr_log_writer.h
@@ -0,0 +1,119 @@
+//===-- xray_fdr_log_writer.h ---------------------------------------------===//
+//
+//                     The LLVM Compiler Infrastructure
+//
+// This file is distributed under the University of Illinois Open Source
+// License. See LICENSE.TXT for details.
+//
+//===----------------------------------------------------------------------===//
+//
+// This file is a part of XRay, a function call tracing system.
+//
+//===----------------------------------------------------------------------===//
+#ifndef COMPILER_RT_LIB_XRAY_XRAY_FDR_LOG_WRITER_H_
+#define COMPILER_RT_LIB_XRAY_XRAY_FDR_LOG_WRITER_H_
+
+#include "xray_buffer_queue.h"
+#include "xray_fdr_log_records.h"
+#include <functional>
+#include <tuple>
+#include <type_traits>
+#include <utility>
+
+namespace __xray {
+
+template <size_t Index> struct SerializerImpl {
+  template <class Tuple,
+            typename std::enable_if<
+                Index<std::tuple_size<
+                          typename std::remove_reference<Tuple>::type>::value,
+                      int>::type = 0> static void serializeTo(char *Buffer,
+                                                              Tuple &&T) {
+    auto P = reinterpret_cast<const char *>(&std::get<Index>(T));
+    constexpr auto Size = sizeof(std::get<Index>(T));
+    internal_memcpy(Buffer, P, Size);
+    SerializerImpl<Index + 1>::serializeTo(Buffer + Size,
+                                           std::forward<Tuple>(T));
+  }
+
+  template <class Tuple,
+            typename std::enable_if<
+                Index >= std::tuple_size<typename std::remove_reference<
+                             Tuple>::type>::value,
+                int>::type = 0>
+  static void serializeTo(char *, Tuple &&){};
+};
+
+using Serializer = SerializerImpl<0>;
+
+template <MetadataRecord::RecordKinds Kind, class... DataTypes>
+MetadataRecord createMetadataRecord(DataTypes &&... Ds) {
+  MetadataRecord R;
+  R.Type = 1;
+  R.RecordKind = static_cast<uint8_t>(Kind);
+  Serializer::serializeTo(R.Data,
+                          std::make_tuple(std::forward<DataTypes>(Ds)...));
+  return R;
+}
+
+class FDRLogWriter {
+  BufferQueue::Buffer &Buffer;
+  char *NextRecord = nullptr;
+
+  template <class T> void writeRecord(const T &R) {
+    internal_memcpy(NextRecord, reinterpret_cast<const char *>(&R), sizeof(T));
+    NextRecord += sizeof(T);
+    atomic_fetch_add(&Buffer.Extents, sizeof(T), memory_order_acq_rel);
+  }
+
+public:
+  explicit FDRLogWriter(BufferQueue::Buffer &B, char *P)
+      : Buffer(B), NextRecord(P) {
+    DCHECK_NE(Buffer.Data, nullptr);
+    DCHECK_NE(NextRecord, nullptr);
+  }
+
+  explicit FDRLogWriter(BufferQueue::Buffer &B)
+      : FDRLogWriter(B, static_cast<char *>(B.Data)) {}
+
+  template <MetadataRecord::RecordKinds Kind, class... Data>
+  bool writeMetadata(Data &&... Ds) {
+    // TODO: Check boundary conditions:
+    // 1) Buffer is full, and cannot handle one metadata record.
+    // 2) Buffer queue is finalising.
+    writeRecord(createMetadataRecord<Kind>(std::forward<Data>(Ds)...));
+    return true;
+  }
+
+  template <size_t N> size_t writeMetadataRecords(MetadataRecord (&Recs)[N]) {
+    constexpr auto Size = sizeof(MetadataRecord) * N;
+    internal_memcpy(NextRecord, reinterpret_cast<const char *>(Recs), Size);
+    NextRecord += Size;
+    atomic_fetch_add(&Buffer.Extents, Size, memory_order_acq_rel);
+    return Size;
+  }
+
+  enum class FunctionRecordKind : uint8_t {
+    Enter = 0x00,
+    Exit = 0x01,
+    TailExit = 0x02,
+    EnterArg = 0x03,
+  };
+
+  bool writeFunction(FunctionRecordKind Kind, int32_t FuncId, int32_t Delta) {
+    FunctionRecord R;
+    R.Type = 0;
+    R.RecordKind = uint8_t(Kind);
+    R.FuncId = FuncId;
+    R.TSCDelta = Delta;
+    writeRecord(R);
+    return true;
+  }
+
+  char *getNextRecord() const { return NextRecord; }
+
+}; // namespace __xray
+
+} // namespace __xray
+
+#endif // COMPILER-RT_LIB_XRAY_XRAY_FDR_LOG_WRITER_H_
diff --git a/lib/xray/xray_fdr_logging.cc b/lib/xray/xray_fdr_logging.cc
index 131c137b6..bb9492bc9 100644
--- a/lib/xray/xray_fdr_logging.cc
+++ b/lib/xray/xray_fdr_logging.cc
@@ -34,6 +34,7 @@
 #include "xray_buffer_queue.h"
 #include "xray_defs.h"
 #include "xray_fdr_flags.h"
+#include "xray_fdr_log_writer.h"
 #include "xray_flags.h"
 #include "xray_recursion_guard.h"
 #include "xray_tsc.h"
@@ -138,59 +139,35 @@ static ThreadLocalData &getThreadLocalData() {
 
 static void writeNewBufferPreamble(tid_t Tid, timespec TS,
                                    pid_t Pid) XRAY_NEVER_INSTRUMENT {
-  static constexpr int InitRecordsCount = 3;
+  static_assert(sizeof(time_t) <= 8, "time_t needs to be at most 8 bytes");
   auto &TLD = getThreadLocalData();
-  MetadataRecord Metadata[InitRecordsCount];
-  {
-    // Write out a MetadataRecord to signify that this is the start of a new
-    // buffer, associated with a particular thread, with a new CPU.  For the
-    // data, we have 15 bytes to squeeze as much information as we can.  At this
-    // point we only write down the following bytes:
-    //   - Thread ID (tid_t, cast to 4 bytes type due to Darwin being 8 bytes)
-    auto &NewBuffer = Metadata[0];
-    NewBuffer.Type = uint8_t(RecordType::Metadata);
-    NewBuffer.RecordKind = uint8_t(MetadataRecord::RecordKinds::NewBuffer);
-    int32_t tid = static_cast<int32_t>(Tid);
-    internal_memcpy(&NewBuffer.Data, &tid, sizeof(tid));
-  }
-
-  // Also write the WalltimeMarker record.
-  {
-    static_assert(sizeof(time_t) <= 8, "time_t needs to be at most 8 bytes");
-    auto &WalltimeMarker = Metadata[1];
-    WalltimeMarker.Type = uint8_t(RecordType::Metadata);
-    WalltimeMarker.RecordKind =
-        uint8_t(MetadataRecord::RecordKinds::WalltimeMarker);
-
-    // We only really need microsecond precision here, and enforce across
-    // platforms that we need 64-bit seconds and 32-bit microseconds encoded in
-    // the Metadata record.
-    int32_t Micros = TS.tv_nsec / 1000;
-    int64_t Seconds = TS.tv_sec;
-    internal_memcpy(WalltimeMarker.Data, &Seconds, sizeof(Seconds));
-    internal_memcpy(WalltimeMarker.Data + sizeof(Seconds), &Micros,
-                    sizeof(Micros));
-  }
-
-  // Also write the Pid record.
-  {
-    // Write out a MetadataRecord that contains the current pid
-    auto &PidMetadata = Metadata[2];
-    PidMetadata.Type = uint8_t(RecordType::Metadata);
-    PidMetadata.RecordKind = uint8_t(MetadataRecord::RecordKinds::Pid);
-    int32_t pid = static_cast<int32_t>(Pid);
-    internal_memcpy(&PidMetadata.Data, &pid, sizeof(pid));
-  }
+  MetadataRecord Metadata[] = {
+      // Write out a MetadataRecord to signify that this is the start of a new
+      // buffer, associated with a particular thread, with a new CPU. For the
+      // data, we have 15 bytes to squeeze as much information as we can. At
+      // this point we only write down the following bytes:
+      //   - Thread ID (tid_t, cast to 4 bytes type due to Darwin being 8 bytes)
+      createMetadataRecord<MetadataRecord::RecordKinds::NewBuffer>(
+          static_cast<int32_t>(Tid)),
+
+      // Also write the WalltimeMarker record. We only really need microsecond
+      // precision here, and enforce across platforms that we need 64-bit
+      // seconds and 32-bit microseconds encoded in the Metadata record.
+      createMetadataRecord<MetadataRecord::RecordKinds::WalltimeMarker>(
+          static_cast<int64_t>(TS.tv_sec),
+          static_cast<int32_t>(TS.tv_nsec / 1000)),
+
+      // Also write the Pid record.
+      createMetadataRecord<MetadataRecord::RecordKinds::Pid>(
+          static_cast<int32_t>(Pid)),
+  };
 
   TLD.NumConsecutiveFnEnters = 0;
   TLD.NumTailCalls = 0;
   if (TLD.BQ == nullptr || TLD.BQ->finalizing())
     return;
-  internal_memcpy(TLD.RecordPtr, Metadata,
-                  sizeof(MetadataRecord) * InitRecordsCount);
-  TLD.RecordPtr += sizeof(MetadataRecord) * InitRecordsCount;
-  atomic_store(&TLD.Buffer.Extents, sizeof(MetadataRecord) * InitRecordsCount,
-               memory_order_release);
+  FDRLogWriter Writer(TLD.Buffer);
+  TLD.RecordPtr += Writer.writeMetadataRecords(Metadata);
 }
 
 static void setupNewBuffer(int (*wall_clock_reader)(
@@ -198,6 +175,7 @@ static void setupNewBuffer(int (*wall_clock_reader)(
   auto &TLD = getThreadLocalData();
   auto &B = TLD.Buffer;
   TLD.RecordPtr = static_cast<char *>(B.Data);
+  atomic_store(&B.Extents, 0, memory_order_release);
   tid_t Tid = GetTid();
   timespec TS{0, 0};
   pid_t Pid = internal_getpid();
@@ -221,52 +199,38 @@ static void decrementExtents(size_t Subtract) {
 static void writeNewCPUIdMetadata(uint16_t CPU,
                                   uint64_t TSC) XRAY_NEVER_INSTRUMENT {
   auto &TLD = getThreadLocalData();
-  MetadataRecord NewCPUId;
-  NewCPUId.Type = uint8_t(RecordType::Metadata);
-  NewCPUId.RecordKind = uint8_t(MetadataRecord::RecordKinds::NewCPUId);
+  FDRLogWriter W(TLD.Buffer, TLD.RecordPtr);
 
   // The data for the New CPU will contain the following bytes:
   //   - CPU ID (uint16_t, 2 bytes)
   //   - Full TSC (uint64_t, 8 bytes)
   // Total = 10 bytes.
-  internal_memcpy(&NewCPUId.Data, &CPU, sizeof(CPU));
-  internal_memcpy(&NewCPUId.Data[sizeof(CPU)], &TSC, sizeof(TSC));
-  internal_memcpy(TLD.RecordPtr, &NewCPUId, sizeof(MetadataRecord));
-  TLD.RecordPtr += sizeof(MetadataRecord);
+  W.writeMetadata<MetadataRecord::RecordKinds::NewCPUId>(CPU, TSC);
+  TLD.RecordPtr = W.getNextRecord();
   TLD.NumConsecutiveFnEnters = 0;
   TLD.NumTailCalls = 0;
-  incrementExtents(sizeof(MetadataRecord));
 }
 
 static void writeTSCWrapMetadata(uint64_t TSC) XRAY_NEVER_INSTRUMENT {
   auto &TLD = getThreadLocalData();
-  MetadataRecord TSCWrap;
-  TSCWrap.Type = uint8_t(RecordType::Metadata);
-  TSCWrap.RecordKind = uint8_t(MetadataRecord::RecordKinds::TSCWrap);
+  FDRLogWriter W(TLD.Buffer, TLD.RecordPtr);
 
   // The data for the TSCWrap record contains the following bytes:
   //   - Full TSC (uint64_t, 8 bytes)
   // Total = 8 bytes.
-  internal_memcpy(&TSCWrap.Data, &TSC, sizeof(TSC));
-  internal_memcpy(TLD.RecordPtr, &TSCWrap, sizeof(MetadataRecord));
-  TLD.RecordPtr += sizeof(MetadataRecord);
+  W.writeMetadata<MetadataRecord::RecordKinds::TSCWrap>(TSC);
+  TLD.RecordPtr = W.getNextRecord();
   TLD.NumConsecutiveFnEnters = 0;
   TLD.NumTailCalls = 0;
-  incrementExtents(sizeof(MetadataRecord));
 }
 
 // Call Argument metadata records store the arguments to a function in the
 // order of their appearance; holes are not supported by the buffer format.
 static void writeCallArgumentMetadata(uint64_t A) XRAY_NEVER_INSTRUMENT {
   auto &TLD = getThreadLocalData();
-  MetadataRecord CallArg;
-  CallArg.Type = uint8_t(RecordType::Metadata);
-  CallArg.RecordKind = uint8_t(MetadataRecord::RecordKinds::CallArgument);
-
-  internal_memcpy(CallArg.Data, &A, sizeof(A));
-  internal_memcpy(TLD.RecordPtr, &CallArg, sizeof(MetadataRecord));
-  TLD.RecordPtr += sizeof(MetadataRecord);
-  incrementExtents(sizeof(MetadataRecord));
+  FDRLogWriter W(TLD.Buffer, TLD.RecordPtr);
+  W.writeMetadata<MetadataRecord::RecordKinds::CallArgument>(A);
+  TLD.RecordPtr = W.getNextRecord();
 }
 
 static void writeFunctionRecord(int32_t FuncId, uint32_t TSCDelta,
@@ -279,8 +243,8 @@ static void writeFunctionRecord(int32_t FuncId, uint32_t TSCDelta,
     return;
   }
 
-  FunctionRecord FuncRecord;
-  FuncRecord.Type = uint8_t(RecordType::Function);
+  auto &TLD = getThreadLocalData();
+  FDRLogWriter W(TLD.Buffer, TLD.RecordPtr);
 
   // Only take 28 bits of the function id.
   //
@@ -289,27 +253,25 @@ static void writeFunctionRecord(int32_t FuncId, uint32_t TSCDelta,
   // to the first 28 bits. To do this properly, this means we need to mask the
   // function id with (2 ^ 28) - 1 == 0x0fffffff.
   //
-  FuncRecord.FuncId = FuncId & MaxFuncId;
-  FuncRecord.TSCDelta = TSCDelta;
+  auto TruncatedId = FuncId & MaxFuncId;
+  auto Kind = FDRLogWriter::FunctionRecordKind::Enter;
 
-  auto &TLD = getThreadLocalData();
   switch (EntryType) {
   case XRayEntryType::ENTRY:
     ++TLD.NumConsecutiveFnEnters;
-    FuncRecord.RecordKind = uint8_t(FunctionRecord::RecordKinds::FunctionEnter);
     break;
   case XRayEntryType::LOG_ARGS_ENTRY:
     // We should not rewind functions with logged args.
     TLD.NumConsecutiveFnEnters = 0;
     TLD.NumTailCalls = 0;
-    FuncRecord.RecordKind = uint8_t(FunctionRecord::RecordKinds::FunctionEnter);
+    Kind = FDRLogWriter::FunctionRecordKind::EnterArg;
     break;
   case XRayEntryType::EXIT:
     // If we've decided to log the function exit, we will never erase the log
     // before it.
     TLD.NumConsecutiveFnEnters = 0;
     TLD.NumTailCalls = 0;
-    FuncRecord.RecordKind = uint8_t(FunctionRecord::RecordKinds::FunctionExit);
+    Kind = FDRLogWriter::FunctionRecordKind::Exit;
     break;
   case XRayEntryType::TAIL:
     // If we just entered the function we're tail exiting from or erased every
@@ -324,8 +286,7 @@ static void writeFunctionRecord(int32_t FuncId, uint32_t TSCDelta,
       TLD.NumTailCalls = 0;
       TLD.NumConsecutiveFnEnters = 0;
     }
-    FuncRecord.RecordKind =
-        uint8_t(FunctionRecord::RecordKinds::FunctionTailExit);
+    Kind = FDRLogWriter::FunctionRecordKind::TailExit;
     break;
   case XRayEntryType::CUSTOM_EVENT: {
     // This is a bug in patching, so we'll report it once and move on.
@@ -346,9 +307,8 @@ static void writeFunctionRecord(int32_t FuncId, uint32_t TSCDelta,
   }
   }
 
-  internal_memcpy(TLD.RecordPtr, &FuncRecord, sizeof(FunctionRecord));
-  TLD.RecordPtr += sizeof(FunctionRecord);
-  incrementExtents(sizeof(FunctionRecord));
+  W.writeFunction(Kind, TruncatedId, TSCDelta);
+  TLD.RecordPtr = W.getNextRecord();
 }
 
 static atomic_uint64_t TicksPerSec{0};
@@ -423,6 +383,9 @@ static void rewindRecentCall(uint64_t TSC, uint64_t &LastTSC,
 static bool releaseThreadLocalBuffer(BufferQueue &BQArg) {
   auto &TLD = getThreadLocalData();
   auto EC = BQArg.releaseBuffer(TLD.Buffer);
+  if (TLD.Buffer.Data == nullptr)
+    return true;
+
   if (EC != BufferQueue::ErrorCode::Ok) {
     Report("Failed to release buffer at %p; error=%s\n", TLD.Buffer.Data,
            BufferQueue::getErrorString(EC));
@@ -1174,6 +1137,8 @@ XRayLogInitStatus fdrLoggingInit(UNUSED size_t BufferSize,
       auto &TLD = *reinterpret_cast<ThreadLocalData *>(TLDPtr);
       if (TLD.BQ == nullptr)
         return;
+      if (TLD.Buffer.Data == nullptr)
+        return;
       auto EC = TLD.BQ->releaseBuffer(TLD.Buffer);
       if (EC != BufferQueue::ErrorCode::Ok)
         Report("At thread exit, failed to release buffer at %p; error=%s\n",