shaka-packager/base/memory/shared_memory_unittest.cc

562 lines
17 KiB
C++
Raw Normal View History

// Copyright (c) 2012 The Chromium Authors. All rights reserved.
// Use of this source code is governed by a BSD-style license that can be
// found in the LICENSE file.
#include "base/basictypes.h"
#include "base/memory/scoped_ptr.h"
#include "base/memory/shared_memory.h"
#include "base/process/kill.h"
#include "base/rand_util.h"
#include "base/strings/string_number_conversions.h"
#include "base/sys_info.h"
#include "base/test/multiprocess_test.h"
#include "base/threading/platform_thread.h"
#include "base/time/time.h"
#include "testing/gtest/include/gtest/gtest.h"
#include "testing/multiprocess_func_list.h"
#if defined(OS_MACOSX)
#include "base/mac/scoped_nsautorelease_pool.h"
#endif
#if defined(OS_POSIX)
#include <sys/mman.h>
#include <sys/stat.h>
#include <sys/types.h>
#include <unistd.h>
#endif
static const int kNumThreads = 5;
static const int kNumTasks = 5;
namespace base {
namespace {
// Each thread will open the shared memory. Each thread will take a different 4
// byte int pointer, and keep changing it, with some small pauses in between.
// Verify that each thread's value in the shared memory is always correct.
class MultipleThreadMain : public PlatformThread::Delegate {
public:
explicit MultipleThreadMain(int16 id) : id_(id) {}
virtual ~MultipleThreadMain() {}
static void CleanUp() {
SharedMemory memory;
memory.Delete(s_test_name_);
}
// PlatformThread::Delegate interface.
virtual void ThreadMain() OVERRIDE {
#if defined(OS_MACOSX)
mac::ScopedNSAutoreleasePool pool;
#endif
const uint32 kDataSize = 1024;
SharedMemory memory;
bool rv = memory.CreateNamed(s_test_name_, true, kDataSize);
EXPECT_TRUE(rv);
rv = memory.Map(kDataSize);
EXPECT_TRUE(rv);
int *ptr = static_cast<int*>(memory.memory()) + id_;
EXPECT_EQ(0, *ptr);
for (int idx = 0; idx < 100; idx++) {
*ptr = idx;
PlatformThread::Sleep(base::TimeDelta::FromMilliseconds(1));
EXPECT_EQ(*ptr, idx);
}
// Reset back to 0 for the next test that uses the same name.
*ptr = 0;
memory.Close();
}
private:
int16 id_;
static const char* const s_test_name_;
DISALLOW_COPY_AND_ASSIGN(MultipleThreadMain);
};
const char* const MultipleThreadMain::s_test_name_ =
"SharedMemoryOpenThreadTest";
// TODO(port):
// This test requires the ability to pass file descriptors between processes.
// We haven't done that yet in Chrome for POSIX.
#if defined(OS_WIN)
// Each thread will open the shared memory. Each thread will take the memory,
// and keep changing it while trying to lock it, with some small pauses in
// between. Verify that each thread's value in the shared memory is always
// correct.
class MultipleLockThread : public PlatformThread::Delegate {
public:
explicit MultipleLockThread(int id) : id_(id) {}
virtual ~MultipleLockThread() {}
// PlatformThread::Delegate interface.
virtual void ThreadMain() OVERRIDE {
const uint32 kDataSize = sizeof(int);
SharedMemoryHandle handle = NULL;
{
SharedMemory memory1;
EXPECT_TRUE(memory1.CreateNamed("SharedMemoryMultipleLockThreadTest",
true, kDataSize));
EXPECT_TRUE(memory1.ShareToProcess(GetCurrentProcess(), &handle));
// TODO(paulg): Implement this once we have a posix version of
// SharedMemory::ShareToProcess.
EXPECT_TRUE(true);
}
SharedMemory memory2(handle, false);
EXPECT_TRUE(memory2.Map(kDataSize));
volatile int* const ptr = static_cast<int*>(memory2.memory());
for (int idx = 0; idx < 20; idx++) {
memory2.Lock();
int i = (id_ << 16) + idx;
*ptr = i;
PlatformThread::Sleep(TimeDelta::FromMilliseconds(1));
EXPECT_EQ(*ptr, i);
memory2.Unlock();
}
memory2.Close();
}
private:
int id_;
DISALLOW_COPY_AND_ASSIGN(MultipleLockThread);
};
#endif
} // namespace
// Android doesn't support SharedMemory::Open/Delete/
// CreateNamed(openExisting=true)
#if !defined(OS_ANDROID)
TEST(SharedMemoryTest, OpenClose) {
const uint32 kDataSize = 1024;
std::string test_name = "SharedMemoryOpenCloseTest";
// Open two handles to a memory segment, confirm that they are mapped
// separately yet point to the same space.
SharedMemory memory1;
bool rv = memory1.Delete(test_name);
EXPECT_TRUE(rv);
rv = memory1.Delete(test_name);
EXPECT_TRUE(rv);
rv = memory1.Open(test_name, false);
EXPECT_FALSE(rv);
rv = memory1.CreateNamed(test_name, false, kDataSize);
EXPECT_TRUE(rv);
rv = memory1.Map(kDataSize);
EXPECT_TRUE(rv);
SharedMemory memory2;
rv = memory2.Open(test_name, false);
EXPECT_TRUE(rv);
rv = memory2.Map(kDataSize);
EXPECT_TRUE(rv);
EXPECT_NE(memory1.memory(), memory2.memory()); // Compare the pointers.
// Make sure we don't segfault. (it actually happened!)
ASSERT_NE(memory1.memory(), static_cast<void*>(NULL));
ASSERT_NE(memory2.memory(), static_cast<void*>(NULL));
// Write data to the first memory segment, verify contents of second.
memset(memory1.memory(), '1', kDataSize);
EXPECT_EQ(memcmp(memory1.memory(), memory2.memory(), kDataSize), 0);
// Close the first memory segment, and verify the second has the right data.
memory1.Close();
char *start_ptr = static_cast<char *>(memory2.memory());
char *end_ptr = start_ptr + kDataSize;
for (char* ptr = start_ptr; ptr < end_ptr; ptr++)
EXPECT_EQ(*ptr, '1');
// Close the second memory segment.
memory2.Close();
rv = memory1.Delete(test_name);
EXPECT_TRUE(rv);
rv = memory2.Delete(test_name);
EXPECT_TRUE(rv);
}
TEST(SharedMemoryTest, OpenExclusive) {
const uint32 kDataSize = 1024;
const uint32 kDataSize2 = 2048;
std::ostringstream test_name_stream;
test_name_stream << "SharedMemoryOpenExclusiveTest."
<< Time::Now().ToDoubleT();
std::string test_name = test_name_stream.str();
// Open two handles to a memory segment and check that open_existing works
// as expected.
SharedMemory memory1;
bool rv = memory1.CreateNamed(test_name, false, kDataSize);
EXPECT_TRUE(rv);
// Memory1 knows it's size because it created it.
EXPECT_EQ(memory1.requested_size(), kDataSize);
rv = memory1.Map(kDataSize);
EXPECT_TRUE(rv);
// The mapped memory1 must be at least the size we asked for.
EXPECT_GE(memory1.mapped_size(), kDataSize);
// The mapped memory1 shouldn't exceed rounding for allocation granularity.
EXPECT_LT(memory1.mapped_size(),
kDataSize + base::SysInfo::VMAllocationGranularity());
memset(memory1.memory(), 'G', kDataSize);
SharedMemory memory2;
// Should not be able to create if openExisting is false.
rv = memory2.CreateNamed(test_name, false, kDataSize2);
EXPECT_FALSE(rv);
// Should be able to create with openExisting true.
rv = memory2.CreateNamed(test_name, true, kDataSize2);
EXPECT_TRUE(rv);
// Memory2 shouldn't know the size because we didn't create it.
EXPECT_EQ(memory2.requested_size(), 0U);
// We should be able to map the original size.
rv = memory2.Map(kDataSize);
EXPECT_TRUE(rv);
// The mapped memory2 must be at least the size of the original.
EXPECT_GE(memory2.mapped_size(), kDataSize);
// The mapped memory2 shouldn't exceed rounding for allocation granularity.
EXPECT_LT(memory2.mapped_size(),
kDataSize2 + base::SysInfo::VMAllocationGranularity());
// Verify that opening memory2 didn't truncate or delete memory 1.
char *start_ptr = static_cast<char *>(memory2.memory());
char *end_ptr = start_ptr + kDataSize;
for (char* ptr = start_ptr; ptr < end_ptr; ptr++) {
EXPECT_EQ(*ptr, 'G');
}
memory1.Close();
memory2.Close();
rv = memory1.Delete(test_name);
EXPECT_TRUE(rv);
}
#endif
// Create a set of N threads to each open a shared memory segment and write to
// it. Verify that they are always reading/writing consistent data.
TEST(SharedMemoryTest, MultipleThreads) {
MultipleThreadMain::CleanUp();
// On POSIX we have a problem when 2 threads try to create the shmem
// (a file) at exactly the same time, since create both creates the
// file and zerofills it. We solve the problem for this unit test
// (make it not flaky) by starting with 1 thread, then
// intentionally don't clean up its shmem before running with
// kNumThreads.
int threadcounts[] = { 1, kNumThreads };
for (size_t i = 0; i < arraysize(threadcounts); i++) {
int numthreads = threadcounts[i];
scoped_ptr<PlatformThreadHandle[]> thread_handles;
scoped_ptr<MultipleThreadMain*[]> thread_delegates;
thread_handles.reset(new PlatformThreadHandle[numthreads]);
thread_delegates.reset(new MultipleThreadMain*[numthreads]);
// Spawn the threads.
for (int16 index = 0; index < numthreads; index++) {
PlatformThreadHandle pth;
thread_delegates[index] = new MultipleThreadMain(index);
EXPECT_TRUE(PlatformThread::Create(0, thread_delegates[index], &pth));
thread_handles[index] = pth;
}
// Wait for the threads to finish.
for (int index = 0; index < numthreads; index++) {
PlatformThread::Join(thread_handles[index]);
delete thread_delegates[index];
}
}
MultipleThreadMain::CleanUp();
}
// TODO(port): this test requires the MultipleLockThread class
// (defined above), which requires the ability to pass file
// descriptors between processes. We haven't done that yet in Chrome
// for POSIX.
#if defined(OS_WIN)
// Create a set of threads to each open a shared memory segment and write to it
// with the lock held. Verify that they are always reading/writing consistent
// data.
TEST(SharedMemoryTest, Lock) {
PlatformThreadHandle thread_handles[kNumThreads];
MultipleLockThread* thread_delegates[kNumThreads];
// Spawn the threads.
for (int index = 0; index < kNumThreads; ++index) {
PlatformThreadHandle pth;
thread_delegates[index] = new MultipleLockThread(index);
EXPECT_TRUE(PlatformThread::Create(0, thread_delegates[index], &pth));
thread_handles[index] = pth;
}
// Wait for the threads to finish.
for (int index = 0; index < kNumThreads; ++index) {
PlatformThread::Join(thread_handles[index]);
delete thread_delegates[index];
}
}
#endif
// Allocate private (unique) shared memory with an empty string for a
// name. Make sure several of them don't point to the same thing as
// we might expect if the names are equal.
TEST(SharedMemoryTest, AnonymousPrivate) {
int i, j;
int count = 4;
bool rv;
const uint32 kDataSize = 8192;
scoped_ptr<SharedMemory[]> memories(new SharedMemory[count]);
scoped_ptr<int*[]> pointers(new int*[count]);
ASSERT_TRUE(memories.get());
ASSERT_TRUE(pointers.get());
for (i = 0; i < count; i++) {
rv = memories[i].CreateAndMapAnonymous(kDataSize);
EXPECT_TRUE(rv);
int *ptr = static_cast<int*>(memories[i].memory());
EXPECT_TRUE(ptr);
pointers[i] = ptr;
}
for (i = 0; i < count; i++) {
// zero out the first int in each except for i; for that one, make it 100.
for (j = 0; j < count; j++) {
if (i == j)
pointers[j][0] = 100;
else
pointers[j][0] = 0;
}
// make sure there is no bleeding of the 100 into the other pointers
for (j = 0; j < count; j++) {
if (i == j)
EXPECT_EQ(100, pointers[j][0]);
else
EXPECT_EQ(0, pointers[j][0]);
}
}
for (int i = 0; i < count; i++) {
memories[i].Close();
}
}
TEST(SharedMemoryTest, MapAt) {
ASSERT_TRUE(SysInfo::VMAllocationGranularity() >= sizeof(uint32));
const size_t kCount = SysInfo::VMAllocationGranularity();
const size_t kDataSize = kCount * sizeof(uint32);
SharedMemory memory;
ASSERT_TRUE(memory.CreateAndMapAnonymous(kDataSize));
ASSERT_TRUE(memory.Map(kDataSize));
uint32* ptr = static_cast<uint32*>(memory.memory());
ASSERT_NE(ptr, static_cast<void*>(NULL));
for (size_t i = 0; i < kCount; ++i) {
ptr[i] = i;
}
memory.Unmap();
off_t offset = SysInfo::VMAllocationGranularity();
ASSERT_TRUE(memory.MapAt(offset, kDataSize - offset));
offset /= sizeof(uint32);
ptr = static_cast<uint32*>(memory.memory());
ASSERT_NE(ptr, static_cast<void*>(NULL));
for (size_t i = offset; i < kCount; ++i) {
EXPECT_EQ(ptr[i - offset], i);
}
}
#if defined(OS_POSIX)
// Create a shared memory object, mmap it, and mprotect it to PROT_EXEC.
TEST(SharedMemoryTest, AnonymousExecutable) {
const uint32 kTestSize = 1 << 16;
SharedMemory shared_memory;
SharedMemoryCreateOptions options;
options.size = kTestSize;
options.executable = true;
EXPECT_TRUE(shared_memory.Create(options));
EXPECT_TRUE(shared_memory.Map(shared_memory.requested_size()));
EXPECT_EQ(0, mprotect(shared_memory.memory(), shared_memory.requested_size(),
PROT_READ | PROT_EXEC));
}
// Android supports a different permission model than POSIX for its "ashmem"
// shared memory implementation. So the tests about file permissions are not
// included on Android.
#if !defined(OS_ANDROID)
// Set a umask and restore the old mask on destruction.
class ScopedUmaskSetter {
public:
explicit ScopedUmaskSetter(mode_t target_mask) {
old_umask_ = umask(target_mask);
}
~ScopedUmaskSetter() { umask(old_umask_); }
private:
mode_t old_umask_;
DISALLOW_IMPLICIT_CONSTRUCTORS(ScopedUmaskSetter);
};
// Create a shared memory object, check its permissions.
TEST(SharedMemoryTest, FilePermissionsAnonymous) {
const uint32 kTestSize = 1 << 8;
SharedMemory shared_memory;
SharedMemoryCreateOptions options;
options.size = kTestSize;
// Set a file mode creation mask that gives all permissions.
ScopedUmaskSetter permissive_mask(S_IWGRP | S_IWOTH);
EXPECT_TRUE(shared_memory.Create(options));
int shm_fd = shared_memory.handle().fd;
struct stat shm_stat;
EXPECT_EQ(0, fstat(shm_fd, &shm_stat));
// Neither the group, nor others should be able to read the shared memory
// file.
EXPECT_FALSE(shm_stat.st_mode & S_IRWXO);
EXPECT_FALSE(shm_stat.st_mode & S_IRWXG);
}
// Create a shared memory object, check its permissions.
TEST(SharedMemoryTest, FilePermissionsNamed) {
const uint32 kTestSize = 1 << 8;
SharedMemory shared_memory;
SharedMemoryCreateOptions options;
options.size = kTestSize;
std::string shared_mem_name = "shared_perm_test-" + IntToString(getpid()) +
"-" + Uint64ToString(RandUint64());
options.name = &shared_mem_name;
// Set a file mode creation mask that gives all permissions.
ScopedUmaskSetter permissive_mask(S_IWGRP | S_IWOTH);
EXPECT_TRUE(shared_memory.Create(options));
// Clean-up the backing file name immediately, we don't need it.
EXPECT_TRUE(shared_memory.Delete(shared_mem_name));
int shm_fd = shared_memory.handle().fd;
struct stat shm_stat;
EXPECT_EQ(0, fstat(shm_fd, &shm_stat));
// Neither the group, nor others should have been able to open the shared
// memory file while its name existed.
EXPECT_FALSE(shm_stat.st_mode & S_IRWXO);
EXPECT_FALSE(shm_stat.st_mode & S_IRWXG);
}
#endif // !defined(OS_ANDROID)
#endif // defined(OS_POSIX)
// Map() will return addresses which are aligned to the platform page size, this
// varies from platform to platform though. Since we'd like to advertise a
// minimum alignment that callers can count on, test for it here.
TEST(SharedMemoryTest, MapMinimumAlignment) {
static const int kDataSize = 8192;
SharedMemory shared_memory;
ASSERT_TRUE(shared_memory.CreateAndMapAnonymous(kDataSize));
EXPECT_EQ(0U, reinterpret_cast<uintptr_t>(
shared_memory.memory()) & (SharedMemory::MAP_MINIMUM_ALIGNMENT - 1));
shared_memory.Close();
}
#if !defined(OS_IOS) // iOS does not allow multiple processes.
// On POSIX it is especially important we test shmem across processes,
// not just across threads. But the test is enabled on all platforms.
class SharedMemoryProcessTest : public MultiProcessTest {
public:
static void CleanUp() {
SharedMemory memory;
memory.Delete(s_test_name_);
}
static int TaskTestMain() {
int errors = 0;
#if defined(OS_MACOSX)
mac::ScopedNSAutoreleasePool pool;
#endif
const uint32 kDataSize = 1024;
SharedMemory memory;
bool rv = memory.CreateNamed(s_test_name_, true, kDataSize);
EXPECT_TRUE(rv);
if (rv != true)
errors++;
rv = memory.Map(kDataSize);
EXPECT_TRUE(rv);
if (rv != true)
errors++;
int *ptr = static_cast<int*>(memory.memory());
for (int idx = 0; idx < 20; idx++) {
memory.Lock();
int i = (1 << 16) + idx;
*ptr = i;
PlatformThread::Sleep(TimeDelta::FromMilliseconds(10));
if (*ptr != i)
errors++;
memory.Unlock();
}
memory.Close();
return errors;
}
private:
static const char* const s_test_name_;
};
const char* const SharedMemoryProcessTest::s_test_name_ = "MPMem";
TEST_F(SharedMemoryProcessTest, Tasks) {
SharedMemoryProcessTest::CleanUp();
ProcessHandle handles[kNumTasks];
for (int index = 0; index < kNumTasks; ++index) {
handles[index] = SpawnChild("SharedMemoryTestMain", false);
ASSERT_TRUE(handles[index]);
}
int exit_code = 0;
for (int index = 0; index < kNumTasks; ++index) {
EXPECT_TRUE(WaitForExitCode(handles[index], &exit_code));
EXPECT_EQ(0, exit_code);
}
SharedMemoryProcessTest::CleanUp();
}
MULTIPROCESS_TEST_MAIN(SharedMemoryTestMain) {
return SharedMemoryProcessTest::TaskTestMain();
}
#endif // !OS_IOS
} // namespace base