[RocksDB] Expose thread local perf counter for low overhead, per call level performance statistics.

Summary:
As title. No locking/atomic is needed due to thread local. There is also no need to modify the existing client interface, in order to expose related counters.

perf_context_test shows a simple example of retrieving the number of user key comparison done for each put and get call. More counters could be added later.

Sample output
./perf_context_test 1000000
==== Test PerfContextTest.KeyComparisonCount
Inserting 1000000 key/value pairs
...
total user key comparison get: 43446523
total user key comparison put: 8017877
max user key comparison get: 88939
avg user key comparison get:43

Basically, the current skiplist does well on average, but could perform poorly in extreme cases.

Test Plan: run perf_context_test <total number of entries to put/get>

Reviewers: dhruba

Differential Revision: https://reviews.facebook.net/D12225

Makefile

@@ -210,6 +210,9 @@ crc32c_test: util/crc32c_test.o $(LIBOBJECTS) $(TESTHARNESS)
 db_test: db/db_test.o $(LIBOBJECTS) $(TESTHARNESS)
 	$(CXX) db/db_test.o $(LIBOBJECTS) $(TESTHARNESS) $(EXEC_LDFLAGS) -o $@ $(LDFLAGS) $(COVERAGEFLAGS)
 
+perf_context_test: db/perf_context_test.o $(LIBOBJECTS) $(TESTHARNESS)
+	$(CXX) db/perf_context_test.o $(LIBOBJECTS) $(TESTHARNESS) $(EXEC_LDFLAGS) -o $@ $(LDFLAGS)
+
 ttl_test: utilities/ttl/ttl_test.o $(LIBOBJECTS) $(TESTHARNESS)
 	$(CXX) utilities/ttl/ttl_test.o $(LIBOBJECTS) $(TESTHARNESS) $(EXEC_LDFLAGS) -o $@  $(LDFLAGS) $(COVERAGEFLAGS)
 

db/dbformat.cc

@@ -6,6 +6,7 @@
 #include "db/dbformat.h"
 #include "port/port.h"
 #include "util/coding.h"
+#include "include/leveldb/perf_context.h"
 
 namespace leveldb {
 
@@ -53,6 +54,7 @@ int InternalKeyComparator::Compare(const Slice& akey, const Slice& bkey) const {
   //    decreasing sequence number
   //    decreasing type (though sequence# should be enough to disambiguate)
   int r = user_comparator_->Compare(ExtractUserKey(akey), ExtractUserKey(bkey));
+  perf_context.user_key_comparison_count++;
   if (r == 0) {
     const uint64_t anum = DecodeFixed64(akey.data() + akey.size() - 8);
     const uint64_t bnum = DecodeFixed64(bkey.data() + bkey.size() - 8);

db/perf_context.cc

@@ -0,0 +1,12 @@
+#include "include/leveldb/perf_context.h"
+
+
+namespace leveldb {
+
+void PerfContext::Reset() {
+  user_key_comparison_count = 0;
+}
+
+__thread PerfContext perf_context;
+
+}

db/perf_context_test.cc

@@ -0,0 +1,81 @@
+#include <iostream>
+
+#include "leveldb/db.h"
+#include "leveldb/perf_context.h"
+#include "util/testharness.h"
+
+
+namespace leveldb {
+
+// Path to the database on file system
+const std::string kDbName = test::TmpDir() + "/perf_context_test";
+
+std::shared_ptr<DB> OpenDb() {
+    DB* db;
+    Options options;
+    options.create_if_missing = true;
+    options.write_buffer_size = 1000000000;     // give it a big memtable
+    Status s = DB::Open(options, kDbName,  &db);
+    ASSERT_OK(s);
+    return std::shared_ptr<DB>(db);
+}
+
+class PerfContextTest { };
+
+int kTotalKeys = 100;
+
+TEST(PerfContextTest, KeyComparisonCount) {
+
+  DestroyDB(kDbName, Options());    // Start this test with a fresh DB
+
+  auto db = OpenDb();
+
+  WriteOptions write_options;
+  ReadOptions read_options;
+
+  uint64_t total_user_key_comparison_get = 0;
+  uint64_t total_user_key_comparison_put = 0;
+  uint64_t max_user_key_comparison_get = 0;
+
+  std::cout << "Inserting " << kTotalKeys << " key/value pairs\n...\n";
+
+  for (int i = 0; i < kTotalKeys; ++i) {
+    std::string key = "k" + std::to_string(i);
+    std::string value = "v" + std::to_string(i);
+
+    perf_context.Reset();
+    db->Put(write_options, key, value);
+    total_user_key_comparison_put += perf_context.user_key_comparison_count;
+
+    perf_context.Reset();
+    db->Get(read_options, key, &value);
+    total_user_key_comparison_get += perf_context.user_key_comparison_count;
+    max_user_key_comparison_get =
+      std::max(max_user_key_comparison_get,
+               perf_context.user_key_comparison_count);
+  }
+
+  std::cout << "total user key comparison get: "
+            << total_user_key_comparison_get << "\n"
+            << "total user key comparison put: "
+            << total_user_key_comparison_put << "\n"
+            << "max user key comparison get: "
+            << max_user_key_comparison_get << "\n"
+            << "avg user key comparison get:"
+            << total_user_key_comparison_get/kTotalKeys << "\n";
+
+}
+
+
+}
+
+
+int main(int argc, char** argv) {
+
+  if (argc > 1) {
+    leveldb::kTotalKeys = std::stoi(argv[1]);
+  }
+
+  leveldb::test::RunAllTests();
+  return 0;
+}

include/leveldb/perf_context.h

@@ -0,0 +1,24 @@
+#ifndef STORAGE_LEVELDB_INCLUDE_PERF_CONTEXT_H
+#define STORAGE_LEVELDB_INCLUDE_PERF_CONTEXT_H
+
+#include <stdint.h>
+
+namespace leveldb {
+
+// A thread local context for gathering performance counter efficiently
+// and transparently.
+
+struct PerfContext {
+
+  void Reset(); // reset all performance counters to zero
+
+
+  uint64_t user_key_comparison_count; // total number of user key comparisons
+};
+
+extern __thread PerfContext perf_context;
+
+}
+
+
+#endif