1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288
|
/*
* Copyright 2014 The WebRTC Project Authors. All rights reserved.
*
* Use of this source code is governed by a BSD-style license
* that can be found in the LICENSE file in the root of the source
* tree. An additional intellectual property rights grant can be found
* in the file PATENTS. All contributing project authors may
* be found in the AUTHORS file in the root of the source tree.
*/
#include "rtc_base/deprecated/recursive_critical_section.h"
#include <stddef.h>
#include <stdint.h>
#include <atomic>
#include <memory>
#include <set>
#include <utility>
#include <vector>
#include "api/units/time_delta.h"
#include "rtc_base/checks.h"
#include "rtc_base/event.h"
#include "rtc_base/platform_thread.h"
#include "rtc_base/thread.h"
#include "rtc_base/thread_annotations.h"
#include "test/gtest.h"
namespace webrtc {
namespace {
constexpr TimeDelta kLongTime = TimeDelta::Seconds(10);
constexpr int kNumThreads = 16;
constexpr int kOperationsToRun = 1000;
class UniqueValueVerifier {
public:
void Verify(const std::vector<int>& values) {
for (size_t i = 0; i < values.size(); ++i) {
std::pair<std::set<int>::iterator, bool> result =
all_values_.insert(values[i]);
// Each value should only be taken by one thread, so if this value
// has already been added, something went wrong.
EXPECT_TRUE(result.second)
<< " Thread=" << Thread::Current() << " value=" << values[i];
}
}
void Finalize() {}
private:
std::set<int> all_values_;
};
class CompareAndSwapVerifier {
public:
CompareAndSwapVerifier() : zero_count_(0) {}
void Verify(const std::vector<int>& values) {
for (auto v : values) {
if (v == 0) {
EXPECT_EQ(0, zero_count_) << "Thread=" << Thread::Current();
++zero_count_;
} else {
EXPECT_EQ(1, v) << " Thread=" << Thread::Current();
}
}
}
void Finalize() { EXPECT_EQ(1, zero_count_); }
private:
int zero_count_;
};
class RunnerBase {
public:
explicit RunnerBase(int value)
: threads_active_(0),
start_event_(true, false),
done_event_(true, false),
shared_value_(value) {}
bool Run() {
// Signal all threads to start.
start_event_.Set();
// Wait for all threads to finish.
return done_event_.Wait(kLongTime);
}
void SetExpectedThreadCount(int count) { threads_active_.store(count); }
int shared_value() const { return shared_value_; }
protected:
void BeforeStart() { ASSERT_TRUE(start_event_.Wait(kLongTime)); }
// Returns true if all threads have finished.
bool AfterEnd() {
if (threads_active_.fetch_sub(1) == 1) {
done_event_.Set();
return true;
}
return false;
}
std::atomic<int> threads_active_;
Event start_event_;
Event done_event_;
int shared_value_;
};
class RTC_LOCKABLE CriticalSectionLock {
public:
void Lock() RTC_EXCLUSIVE_LOCK_FUNCTION() { cs_.Enter(); }
void Unlock() RTC_UNLOCK_FUNCTION() { cs_.Leave(); }
private:
RecursiveCriticalSection cs_;
};
template <class Lock>
class LockRunner : public RunnerBase {
public:
LockRunner() : RunnerBase(0) {}
void Loop() {
BeforeStart();
lock_.Lock();
EXPECT_EQ(0, shared_value_);
int old = shared_value_;
// Use a loop to increase the chance of race.
for (int i = 0; i < kOperationsToRun; ++i) {
++shared_value_;
}
EXPECT_EQ(old + kOperationsToRun, shared_value_);
shared_value_ = 0;
lock_.Unlock();
AfterEnd();
}
private:
Lock lock_;
};
template <typename Runner>
void StartThreads(std::vector<std::unique_ptr<Thread>>* threads,
Runner* handler) {
for (int i = 0; i < kNumThreads; ++i) {
std::unique_ptr<Thread> thread(Thread::Create());
thread->Start();
thread->PostTask([handler] { handler->Loop(); });
threads->push_back(std::move(thread));
}
}
} // namespace
TEST(RecursiveCriticalSectionTest, Basic) {
// Create and start lots of threads.
LockRunner<CriticalSectionLock> runner;
std::vector<std::unique_ptr<Thread>> threads;
StartThreads(&threads, &runner);
runner.SetExpectedThreadCount(kNumThreads);
// Release the hounds!
EXPECT_TRUE(runner.Run());
EXPECT_EQ(0, runner.shared_value());
}
class PerfTestData {
public:
PerfTestData(int expected_count, Event* event)
: cache_line_barrier_1_(),
cache_line_barrier_2_(),
expected_count_(expected_count),
event_(event) {
cache_line_barrier_1_[0]++; // Avoid 'is not used'.
cache_line_barrier_2_[0]++; // Avoid 'is not used'.
}
~PerfTestData() {}
void AddToCounter(int add) {
CritScope cs(&lock_);
my_counter_ += add;
if (my_counter_ == expected_count_)
event_->Set();
}
int64_t total() const {
// Assume that only one thread is running now.
return my_counter_;
}
private:
uint8_t cache_line_barrier_1_[64];
RecursiveCriticalSection lock_;
uint8_t cache_line_barrier_2_[64];
int64_t my_counter_ = 0;
const int expected_count_;
Event* const event_;
};
class PerfTestThread {
public:
void Start(PerfTestData* data, int repeats, int id) {
RTC_DCHECK(!data_);
data_ = data;
repeats_ = repeats;
my_id_ = id;
thread_ = PlatformThread::SpawnJoinable(
[this] {
for (int i = 0; i < repeats_; ++i)
data_->AddToCounter(my_id_);
},
"CsPerf");
}
void Stop() {
RTC_DCHECK(data_);
thread_.Finalize();
repeats_ = 0;
data_ = nullptr;
my_id_ = 0;
}
private:
PlatformThread thread_;
PerfTestData* data_ = nullptr;
int repeats_ = 0;
int my_id_ = 0;
};
// Comparison of output of this test as tested on a MacBook Pro, 13-inch,
// 2017, 3,5 GHz Intel Core i7, 16 GB 2133 MHz LPDDR3,
// running macOS Mojave, 10.14.3.
//
// Native mutex implementation using fair policy (previously macOS default):
// Approximate CPU usage:
// real 4m54.612s
// user 1m20.575s
// sys 3m48.872s
// Unit test output:
// [ OK ] RecursiveCriticalSectionTest.Performance (294375 ms)
//
// Native mutex implementation using first fit policy (current macOS default):
// Approximate CPU usage:
// real 0m11.535s
// user 0m12.738s
// sys 0m31.207s
// Unit test output:
// [ OK ] RecursiveCriticalSectionTest.Performance (11444 ms)
//
// Special partially spin lock based implementation:
// Approximate CPU usage:
// real 0m2.113s
// user 0m3.014s
// sys 0m4.495s
// Unit test output:
// [ OK ] RecursiveCriticalSectionTest.Performance (1885 ms)
//
// The test is disabled by default to avoid unecessarily loading the bots.
TEST(RecursiveCriticalSectionTest, DISABLED_Performance) {
PerfTestThread threads[8];
Event event;
static const int kThreadRepeats = 10000000;
static const int kExpectedCount = kThreadRepeats * std::ssize(threads);
PerfTestData test_data(kExpectedCount, &event);
for (auto& t : threads)
t.Start(&test_data, kThreadRepeats, 1);
event.Wait(Event::kForever);
for (auto& t : threads)
t.Stop();
}
} // namespace webrtc
|