《Linux内核分析》 week6作业-Linux内核fork()系统调用的创建过程

一.进程控制块PCB-stack_struct

进程在操作系统中都有一个结构，用于表示这个进程。这就是进程控制块(PCB)，在Linux中具体实现是task_struct数据结构，它主要记录了以下信息:

• 状态信息，例如可执行状态、就绪状态、阻塞状态等。
• 性质，由于unix有很多变种，进行有自己独特的性质。
• 资源，资源的链接比如内存，还有资源的限制和权限等。
• 组织，例如按照家族关系建立起来的树(父进程、子进程等)。

task_struct结构体内容非常庞大,暂时没有去分析源代码，以后有时间再去研究。

二.Linux fork执行的过程

在menu中添加一个fork的系统调用，然后用gdb开始调试.执行以下命令

qemu -kernel linux-3.18./arch/x86/boot/bzImage -initrd rootfs.img -s -s
gdb
file linux-3.18./vmlinux
target remote:

然后在sys_fork、sys_clone处设置断点，再逐步调试，观察fork系统调用的执行过程。

具体分析fork系统调用执行过程.

1.fork、vfork和clone三个系统调用都可以创建一个新进程，而且它们都是通过调用do_fork来实现进程的创建，do_fork通过传递不同的clone_flags来实现fork、clone、vfork。

long do_fork(unsigned long clone_flags,
          unsigned long stack_start,
          unsigned long stack_size,
          int __user *parent_tidptr,
          int __user *child_tidptr)
{
    struct task_struct *p;
    int trace = ;
    long nr;

    /*
1634     * Determine whether and which event to report to ptracer.  When
1635     * called from kernel_thread or CLONE_UNTRACED is explicitly
1636     * requested, no event is reported; otherwise, report if the event
1637     * for the type of forking is enabled.
1638     */
    if (!(clone_flags & CLONE_UNTRACED)) {
        if (clone_flags & CLONE_VFORK)
            trace = PTRACE_EVENT_VFORK;
        else if ((clone_flags & CSIGNAL) != SIGCHLD)
            trace = PTRACE_EVENT_CLONE;
        else
            trace = PTRACE_EVENT_FORK;

        if (likely(!ptrace_event_enabled(current, trace)))
            trace = ;
    }
1650
    p = copy_process(clone_flags, stack_start, stack_size,
             child_tidptr, NULL, trace);    #进程复制，核心函数
    /*
1654     * Do this prior waking up the new thread - the thread pointer
1655     * might get invalid after that point, if the thread exits quickly.
1656     */
    if (!IS_ERR(p)) {
        struct completion vfork;
        struct pid *pid;

        trace_sched_process_fork(current, p);

        pid = get_task_pid(p, PIDTYPE_PID);
        nr = pid_vnr(pid);

        if (clone_flags & CLONE_PARENT_SETTID)
            put_user(nr, parent_tidptr);

        if (clone_flags & CLONE_VFORK) {
            p->vfork_done = &vfork;
            init_completion(&vfork);
            get_task_struct(p);
        }

        wake_up_new_task(p);

        /* forking complete and child started to run, tell ptracer */
        if (unlikely(trace))
            ptrace_event_pid(trace, pid);

        if (clone_flags & CLONE_VFORK) {
            if (!wait_for_vfork_done(p, &vfork))
                ptrace_event_pid(PTRACE_EVENT_VFORK_DONE, pid);
        }

        put_pid(pid);
    } else {
        nr = PTR_ERR(p);
    }
    return nr;
}

do_fork()函数的核心是copy_process()，该函数完成了进程创建的绝大部分。

/*
1175 * This creates a new process as a copy of the old one,
1176 * but does not actually start it yet.
1177 *
1178 * It copies the registers, and all the appropriate
1179 * parts of the process environment (as per the clone
1180 * flags). The actual kick-off is left to the caller.
1181 */static struct task_struct *copy_process(unsigned long clone_flags,
                    unsigned long stack_start,
                    unsigned long stack_size,
                    int __user *child_tidptr,
                    struct pid *pid,
                    int trace)
{
    int retval;
    struct task_struct *p;

    if ((clone_flags & (CLONE_NEWNS|CLONE_FS)) == (CLONE_NEWNS|CLONE_FS))
        return ERR_PTR(-EINVAL);

    if ((clone_flags & (CLONE_NEWUSER|CLONE_FS)) == (CLONE_NEWUSER|CLONE_FS))
        return ERR_PTR(-EINVAL);

    /*
1199     * Thread groups must share signals as well, and detached threads
1200     * can only be started up within the thread group.
1201     */
    if ((clone_flags & CLONE_THREAD) && !(clone_flags & CLONE_SIGHAND))
        return ERR_PTR(-EINVAL);

    /*
1206     * Shared signal handlers imply shared VM. By way of the above,
1207     * thread groups also imply shared VM. Blocking this case allows
1208     * for various simplifications in other code.
1209     */
    if ((clone_flags & CLONE_SIGHAND) && !(clone_flags & CLONE_VM))
        return ERR_PTR(-EINVAL);

    /*
1214     * Siblings of global init remain as zombies on exit since they are
1215     * not reaped by their parent (swapper). To solve this and to avoid
1216     * multi-rooted process trees, prevent global and container-inits
1217     * from creating siblings.
1218     */
    if ((clone_flags & CLONE_PARENT) &&
                current->signal->flags & SIGNAL_UNKILLABLE)
        return ERR_PTR(-EINVAL);

    /*
1224     * If the new process will be in a different pid or user namespace
1225     * do not allow it to share a thread group or signal handlers or
1226     * parent with the forking task.
1227     */
    if (clone_flags & CLONE_SIGHAND) {
        if ((clone_flags & (CLONE_NEWUSER | CLONE_NEWPID)) ||
            (task_active_pid_ns(current) !=
                current->nsproxy->pid_ns_for_children))
            return ERR_PTR(-EINVAL);
    }

    retval = security_task_create(clone_flags);
    if (retval)
        goto fork_out;

    retval = -ENOMEM;
    p = dup_task_struct(current);  #为子进程创建一个新的内核栈,复制task_struct和thread_info结构，此时子进程的进程控制块和父进程完全一致。
    if (!p)
        goto fork_out;

    ftrace_graph_init_task(p);

    rt_mutex_init_task(p);

#ifdef CONFIG_PROVE_LOCKING
    DEBUG_LOCKS_WARN_ON(!p->hardirqs_enabled);
    DEBUG_LOCKS_WARN_ON(!p->softirqs_enabled);
#endif
    retval = -EAGAIN;
    if (atomic_read(&p->real_cred->user->processes) >=
            task_rlimit(p, RLIMIT_NPROC)) {
        if (p->real_cred->user != INIT_USER &&
            !capable(CAP_SYS_RESOURCE) && !capable(CAP_SYS_ADMIN))
            goto bad_fork_free;
    }
    current->flags &= ~PF_NPROC_EXCEEDED;

    retval = copy_creds(p, clone_flags);
    if (retval < )
        goto bad_fork_free;

    /*
1266     * If multiple threads are within copy_process(), then this check
1267     * triggers too late. This doesn't hurt, the check is only there
1268     * to stop root fork bombs.
1269     */
    retval = -EAGAIN;
    if (nr_threads >= max_threads)
        goto bad_fork_cleanup_count;

    if (!try_module_get(task_thread_info(p)->exec_domain->module))
        goto bad_fork_cleanup_count;

    delayacct_tsk_init(p);    /* Must remain after dup_task_struct() */
    p->flags &= ~(PF_SUPERPRIV | PF_WQ_WORKER);
    p->flags |= PF_FORKNOEXEC;
    INIT_LIST_HEAD(&p->children);
    INIT_LIST_HEAD(&p->sibling);
    rcu_copy_process(p);
    p->vfork_done = NULL;
    spin_lock_init(&p->alloc_lock);

    init_sigpending(&p->pending);

    p->utime = p->stime = p->gtime = ;
....

通过dup_task_struct()函数，为子进程创建一个新的内核栈,复制task_struct和thread_info结构。

ti=alloc_thread_info_node(task,node);
tsk->stack=ti;
setup_thread_stack(tsk,orig); //这里只是复制了thread_info

重点关注下,fork()创建子进程后，父进程从系统调用中返回，而子进程从哪开始返回.

这主要是在copy_process()中copy_thread()代码.

int copy_thread(unsigned long clone_flags, unsigned long sp,
    unsigned long arg, struct task_struct *p)
{
    struct pt_regs *childregs = task_pt_regs(p);
    struct task_struct *tsk;
    int err;

    p->thread.sp = (unsigned long) childregs; #记录进程切换时的堆栈指针
    p->thread.sp0 = (unsigned long) (childregs+);
    memset(p->thread.ptrace_bps, , sizeof(p->thread.ptrace_bps));

    if (unlikely(p->flags & PF_KTHREAD)) {
        /* kernel thread */
        memset(childregs, , sizeof(struct pt_regs));
        p->thread.ip = (unsigned long) ret_from_kernel_thread;
        task_user_gs(p) = __KERNEL_STACK_CANARY;
        childregs->ds = __USER_DS;
        childregs->es = __USER_DS;
        childregs->fs = __KERNEL_PERCPU;
        childregs->bx = sp;    /* function */
        childregs->bp = arg;
        childregs->orig_ax = -;
        childregs->cs = __KERNEL_CS | get_kernel_rpl();
        childregs->flags = X86_EFLAGS_IF | X86_EFLAGS_FIXED;
        p->thread.io_bitmap_ptr = NULL;
        return ;
    }
    *childregs = *current_pt_regs();#复制内核堆栈
    childregs->ax = ; #这也是为什么子进程的fork返回0
    if (sp)
        childregs->sp = sp; 

    p->thread.ip = (unsigned long) ret_from_fork; #子进程开始执行处
    task_user_gs(p) = get_user_gs(current_pt_regs());

    p->thread.io_bitmap_ptr = NULL;
    tsk = current;
    err = -ENOMEM;

    if (unlikely(test_tsk_thread_flag(tsk, TIF_IO_BITMAP))) {
        p->thread.io_bitmap_ptr = kmemdup(tsk->thread.io_bitmap_ptr,
                        IO_BITMAP_BYTES, GFP_KERNEL);
        if (!p->thread.io_bitmap_ptr) {
            p->thread.io_bitmap_max = ;
            return -ENOMEM;
        }
        set_tsk_thread_flag(p, TIF_IO_BITMAP);
    }

    err = ;

    /*
184     * Set a new TLS for the child thread?
185     */
    if (clone_flags & CLONE_SETTLS)
        err = do_set_thread_area(p, -,
            (struct user_desc __user *)childregs->si, );

    if (err && p->thread.io_bitmap_ptr) {
        kfree(p->thread.io_bitmap_ptr);
        p->thread.io_bitmap_max = ;
    }
    return err;
   } 

然后回到do_fork()函数中,唤醒子进程并开始运行。至此，一个进程创建就完成了。

三.实验总结

中间虽然的很多细节还不是很清楚，但是对linux 创建子进程的大体流程有了一个宏观的认识，更加深刻地理解了底层Linux 内核进程运行的机制。