Java開源工具在linux上的源碼分析(三)：執(zhí)行的線程
在前面的博客中（http://blog.csdn.net/raintungli/article/details/7034005）所提到的信號轉(zhuǎn)發(fā)線程，Attach Listener 線程都只是操作socket文件，并沒有去執(zhí)行比如stack 分析，或者heap的分析，真正的工作線程其實是vm thread.
（一）啟動vm thread
- jint Threads::create_vm(JavaVMInitArgs* args, bool* canTryAgain) {
- ...
- // Create the VMThread
- { TraceTime timer("Start VMThread", TraceStartupTime);
- VMThread::create();
- Thread* vmthread = VMThread::vm_thread();
- if (!os::create_thread(vmthread, os::vm_thread))
- vm_exit_during_initialization("Cannot create VM thread. Out of system resources.");
- // Wait for the VM thread to become ready, and VMThread::run to initialize
- // Monitors can have spurious returns, must always check another state flag
- {
- MutexLocker ml(Notify_lock);
- os::start_thread(vmthread);
- while (vmthread->active_handles() == NULL) {
- Notify_lock->wait();
- }
- }
- }
- ...
- }
我們可以看到，在thread.cpp里啟動了線程vm thread，在這里我們同時也稍微的略帶的講一下jvm在linux里如何啟動線程的。
通常在linux中啟動線程，是調(diào)用：
- int pthread_create((pthread_t *__thread, __const pthread_attr_t *__attr,void *(*__start_routine) (void *), void *__arg));
而在java里卻增加了os:create_thread --初始化線程 和os:start_thread--啟動線程。
我們?nèi)タ匆幌耲vm里面是如何在linux里做到的。
在os_linux.cpp中來看create_thread的方法：
- bool os::create_thread(Thread* thread, ThreadType thr_type, size_t stack_size) {
- ....
- int ret = pthread_create(&tid, &attr, (void* (*)(void*)) java_start, thread);
- ....
- }
繼續(xù)看java_start方法：
- static void *java_start(Thread *thread) {
- ....
- // handshaking with parent thread
- {
- MutexLockerEx ml(sync, Mutex::_no_safepoint_check_flag);
- // notify parent thread
- osthread->set_state(INITIALIZED);
- sync->notify_all();
- // wait until os::start_thread()
- while (osthread->get_state() == INITIALIZED) {
- sync->wait(Mutex::_no_safepoint_check_flag);
- }
- }
- // call one more level start routine
- thread->run();
- return 0;
- }
首先jvm先設(shè)置了當(dāng)前線程的狀態(tài)是Initialized, 然后notify所有的線程，
- while (osthread->get_state() == INITIALIZED) {
- sync->wait(Mutex::_no_safepoint_check_flag);
- }
不停的查看線程的當(dāng)前狀態(tài)是不是Initialized, 如果是的話，調(diào)用了sync->wait()的方法等待。
來看os:start_thread的方法 os.cpp
- void os::start_thread(Thread* thread) {
- // guard suspend/resume
- MutexLockerEx ml(thread->SR_lock(), Mutex::_no_safepoint_check_flag);
- OSThread* osthread = thread->osthread();
- osthread->set_state(RUNNABLE);
- pd_start_thread(thread);
- }
這時候設(shè)置了線程的狀態(tài)為runnable,但沒有notify線程。
在 pd_start_thread(thread)中, os_linux.cpp中：
- void os::pd_start_thread(Thread* thread) {
- OSThread * osthread = thread->osthread();
- assert(osthread->get_state() != INITIALIZED, "just checking");
- Monitor* sync_with_child = osthread->startThread_lock();
- MutexLockerEx ml(sync_with_child, Mutex::_no_safepoint_check_flag);
- sync_with_child->notify();
- }
這時候我們看到了notify 線程的操作，也就是這時候notify了線程，因為這時候的線程的狀態(tài)是RUNNABLE, 方法java_start繼續(xù)往下執(zhí)行，于是調(diào)用了thread->run()的方法。
對于線程vm Thread 也就是調(diào)用了vmthread::run方法。
vmThread.cpp
- void VMThread::run() {
- ...
- this->loop();
- ...
- }
調(diào)用了loop函數(shù)，處理了VM_Operation 的queue 關(guān)于queue的級別和優(yōu)先級處理算法：可以參考 另一篇博客：http://blog.csdn.net/raintungli/article/details/6553337
（二）Jstack 運行在vm thread里的VM_Operation
jstack 處理也就是在前面博客所提到的attach Listener 線程所做的 operation
- static jint thread_dump(AttachOperation* op, outputStream* out) {
- bool print_concurrent_locks = false;
- if (op->arg(0) != NULL && strcmp(op->arg(0), "-l") == 0) {
- print_concurrent_locks = true;
- }
- // thread stacks
- VM_PrintThreads op1(out, print_concurrent_locks);
- VMThread::execute(&op1);
- // JNI global handles
- VM_PrintJNI op2(out);
- VMThread::execute(&op2);
- // Deadlock detection
- VM_FindDeadlocks op3(out);
- VMThread::execute(&op3);
- return JNI_OK;
- }
簡單看一下類VM_PrintThreads 它 繼承了VM_Operation
- class VM_PrintThreads: public VM_Operation {
- private:
- outputStream* _out;
- bool _print_concurrent_locks;
- public:
- VM_PrintThreads() { _out = tty; _print_concurrent_locks = PrintConcurrentLocks; }
- VM_PrintThreads(outputStream* out, bool print_concurrent_locks) { _out = out; _print_concurrent_locks = print_concurrent_locks; }
- VMOp_Type type() const { return VMOp_PrintThreads; }
- void doit();
- bool doit_prologue();
- void doit_epilogue();
- };
當(dāng)調(diào)用VMThread::execute()也就是將VM_PrintThreads 放入了_vm_queue中，交給vm thread 處理，對vm thread來說取出queue里的VM_Operation,并且調(diào)用doit方法。
在jstack里，attach listener 的線程產(chǎn)生了VM_PrintThreads，VM_PrintJNI，VM_FindDeadlocks 3個operations，交給了vm thread 的線程處理。
原文鏈接：http://blog.csdn.net/raintungli/article/details/7045024
【系列文章】
