blob: 56ef8225030cf2531f1e659510b5db050ae1ead2 [file] [log] [blame]
/*
* Copyright (C) 2014 Apple Inc. All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* 1. Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
*
* THIS SOFTWARE IS PROVIDED BY APPLE INC. ``AS IS'' AND ANY
* EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
* PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL APPLE INC. OR
* CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
* EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
* PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
* PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY
* OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
* (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
* OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
#include "CPUCount.h"
#include "Interpreter.h"
#include <assert.h>
#include <cstddef>
#include <cstdlib>
#include <errno.h>
#include <fcntl.h>
#include <iostream>
#include <string>
#include <string.h>
#include <sys/mman.h>
#include <sys/stat.h>
#include <sys/types.h>
#include <sys/uio.h>
#include <unistd.h>
#include <vector>
#include "mbmalloc.h"
#define UNUSED_PARAM(variable) (void)variable
Interpreter::Interpreter(const char* fileName, bool shouldFreeAllObjects, bool useThreadId)
: m_shouldFreeAllObjects(shouldFreeAllObjects)
, m_useThreadId(useThreadId)
, m_currentThreadId(0)
, m_ops(1024)
{
m_fd = open(fileName, O_RDONLY);
if (m_fd == -1) {
fprintf(stderr, "Failed to open op file %s: ", fileName);
perror("");
exit(-1);
}
struct stat buf;
fstat(m_fd, &buf);
m_opCount = buf.st_size / sizeof(Op);
assert(m_opCount * sizeof(Op) == static_cast<size_t>(buf.st_size));
size_t maxSlot = 0;
std::vector<Op> ops(1024);
size_t remaining = m_opCount * sizeof(Op);
while (remaining) {
size_t bytes = std::min(remaining, ops.size() * sizeof(Op));
remaining -= bytes;
auto ret = read(m_fd, ops.data(), bytes);
UNUSED_PARAM(ret);
size_t opCount = bytes / sizeof(Op);
for (size_t i = 0; i < opCount; ++i) {
Op op = ops[i];
if (op.slot > maxSlot)
maxSlot = op.slot;
}
}
m_objects.resize(maxSlot + 1);
}
Interpreter::~Interpreter()
{
int result = close(m_fd);
if (result == -1) {
perror("Failed to close op file");
exit(-1);
}
}
void Interpreter::run()
{
std::vector<Op> ops(1024);
lseek(m_fd, 0, SEEK_SET);
m_remaining = m_opCount * sizeof(Op);
m_opsCursor = m_opsInBuffer = 0;
doOnSameThread(0);
for (auto thread : m_threads)
thread->stop();
for (auto thread : m_threads)
delete thread;
// A recording might not free all of its allocations.
if (!m_shouldFreeAllObjects)
return;
for (size_t i = 0; i < m_objects.size(); ++i) {
if (!m_objects[i].object)
continue;
mbfree(m_objects[i].object, m_objects[i].size);
m_objects[i] = { 0, 0 };
}
}
bool Interpreter::readOps()
{
if (!m_remaining)
return false;
size_t bytes = std::min(m_remaining, m_ops.size() * sizeof(Op));
m_remaining -= bytes;
auto ret = read(m_fd, m_ops.data(), bytes);
UNUSED_PARAM(ret);
m_opsCursor = 0;
m_opsInBuffer = bytes / sizeof(Op);
if (!m_opsInBuffer)
return false;
return true;
}
void Interpreter::doOnSameThread(ThreadId runThreadId)
{
while (true) {
if ((m_opsCursor >= m_opsInBuffer) && (!readOps())) {
if (runThreadId)
switchToThread(0);
return;
}
for (; m_opsCursor < m_opsInBuffer; ++m_opsCursor) {
Op op = m_ops[m_opsCursor];
ThreadId threadId = op.threadId;
if (m_useThreadId && (runThreadId != threadId)) {
switchToThread(threadId);
break;
}
doMallocOp(op, m_currentThreadId);
}
}
}
void Interpreter::switchToThread(ThreadId threadId)
{
if (m_currentThreadId == threadId)
return;
for (ThreadId threadIndex = static_cast<ThreadId>(m_threads.size());
threadIndex < threadId; ++threadIndex)
m_threads.push_back(new Thread(this, threadId));
ThreadId currentThreadId = m_currentThreadId;
if (threadId == 0) {
std::unique_lock<std::mutex> lock(m_threadMutex);
m_currentThreadId = threadId;
m_shouldRun.notify_one();
} else
m_threads[threadId - 1]->switchTo();
if (currentThreadId == 0) {
std::unique_lock<std::mutex> lock(m_threadMutex);
m_shouldRun.wait(lock, [this](){return m_currentThreadId == 0; });
} else
m_threads[currentThreadId - 1]->waitToRun();
}
void Interpreter::detailedReport()
{
size_t totalInUse = 0;
size_t smallInUse = 0;
size_t mediumInUse = 0;
size_t largeInUse = 0;
size_t extraLargeInUse = 0;
size_t memoryAllocated = 0;
for (size_t i = 0; i < m_objects.size(); ++i) {
if (!m_objects[i].object)
continue;
size_t objectSize = m_objects[i].size;
memoryAllocated += objectSize;
totalInUse++;
if (objectSize <= 256)
smallInUse++;
else if (objectSize <= 1024)
mediumInUse++;
else if (objectSize <= 1032192)
largeInUse++;
else
extraLargeInUse++;
}
std::cout << "0B-256B objects in use: " << smallInUse << std::endl;
std::cout << "257B-1K objects in use: " << mediumInUse << std::endl;
std::cout << " 1K-1M objects in use: " << largeInUse << std::endl;
std::cout << " 1M+ objects in use: " << extraLargeInUse << std::endl;
std::cout << " Total objects in use: " << totalInUse << std::endl;
std::cout << "Total allocated memory: " << memoryAllocated / 1024 << "kB" << std::endl;
}
static size_t compute2toPower(unsigned log2n)
{
// Check for bad alignment log2 value and return a bad alignment.
if (log2n > 64)
return 0xff00;
size_t result = 1;
while (log2n--)
result <<= 1;
return result;
}
void Interpreter::doMallocOp(Op op, ThreadId)
{
switch (op.opcode) {
case op_malloc: {
m_objects[op.slot] = { mbmalloc(op.size), op.size };
assert(m_objects[op.slot].object);
bzero(m_objects[op.slot].object, op.size);
break;
}
case op_free: {
if (!m_objects[op.slot].object)
return;
mbfree(m_objects[op.slot].object, m_objects[op.slot].size);
m_objects[op.slot] = { 0, 0 };
break;
}
case op_realloc: {
if (!m_objects[op.slot].object)
return;
m_objects[op.slot] = { mbrealloc(m_objects[op.slot].object, m_objects[op.slot].size, op.size), op.size };
break;
}
case op_align_malloc: {
size_t alignment = compute2toPower(op.alignLog2);
m_objects[op.slot] = { mbmemalign(alignment, op.size), op.size };
assert(m_objects[op.slot].object);
bzero(m_objects[op.slot].object, op.size);
break;
}
default: {
fprintf(stderr, "bad opcode: %d\n", op.opcode);
abort();
break;
}
}
}
Interpreter::Thread::Thread(Interpreter* myInterpreter, ThreadId threadId)
: m_threadId(threadId)
, m_myInterpreter(myInterpreter)
{
m_thread = std::thread(&Thread::runThread, this);
}
void Interpreter::Thread::stop()
{
m_myInterpreter->switchToThread(m_threadId);
}
Interpreter::Thread::~Thread()
{
switchTo();
m_thread.join();
}
void Interpreter::Thread::runThread()
{
waitToRun();
m_myInterpreter->doOnSameThread(m_threadId);
}
void Interpreter::Thread::waitToRun()
{
std::unique_lock<std::mutex> lock(m_myInterpreter->m_threadMutex);
m_shouldRun.wait(lock, [this](){return m_myInterpreter->m_currentThreadId == m_threadId; });
}
void Interpreter::Thread::switchTo()
{
std::unique_lock<std::mutex> lock(m_myInterpreter->m_threadMutex);
m_myInterpreter->m_currentThreadId = m_threadId;
m_shouldRun.notify_one();
}