# Process management

process control block (PCB)


# process descriptor & Task_Struct

kernel把process稱作task，用doubly linked list存放，type name **task_struct**就是linux的PCB，裡面紀錄的process的一切資訊，size有1.7KB，滿大的


linux 2.6之前task_struct是存在stack最底端(stack往下長)，linux 2.6之後的版本用thread_info把task_struct包起來，也放在stack最底端


# 存取當下在cpu執行的process

> it is useful to be able to quickly look up the process descriptor of the currently executing task, which is done via the **current** macro.

> current_thread_info()
>
Finally, current dereferences the task member of thread_info to return the task_struct:
> current_thread_info()->task;
>

# process state




Kernel code often needs to change a process’s state.The preferred mechanism is using
> set_task_state(task, state); /* set task ‘task’ to state ‘state’ */

# process termination

* process自己call exit() system call
* main() return時 c compiler就會call exit()
* 中途process自己達成明確目的call exit()
* 別人send signal kill him
* 發生exception 被kernel kill掉

exit()被wrapper成do_exit()
> `code flow細節定義在kernel/exit.c`

做完之後process的所有資源都被free掉，也沒有address space了，然後被丟到EXIT_ZOMBIE state，唯一留下thread_info structure & task_struct structure，目的是寫遺書給他的parent，等parent確認過child沒有利用價值後，就叫kernel把他幹掉。

parent的作法是call wait() system call，等他的child寫遺書說他想死，wait會收到child的pid，然後可以去call release_task()才把child的資源完全釋放掉

> 1. It calls __exit_signal(), which calls __unhash_process(), which in turns calls detach_pid() to remove the process from the pidhash and remove the process from the task list.
> 2. __exit_signal() releases any remaining resources used by the now dead process and finalizes statistics and bookkeeping.
> 3. If the task was the last member of a thread group, and the leader is a zombie, then release_task() notifies the zombie leader’s parent.
> 4. release_task() calls put_task_struct() to free the pages containing the process’s kernel stack and thread_info structure and deallocate the slab cache containing the task_struct
>

# Process and thread(in linux)

Linux實作thread的方式跟大多數OS不一樣
其他os的process and thread是用完全不一樣的struct

process call fork()最後會call到 clone()
thread call clone()參數帶一些自訂的flag
process就是無帶參數的thread

> clone(SIGCHLD, 0);

thread就是有帶參數條件的process

> `clone(CLONE_VM | CLONE_FS | CLONE_FILES | CLONE_SIGHAND, 0);`
>

vfork()
vfork()跟fork()的差別只有child不複製page table，直接使用parent的address space，parent直接被block住直到child call exec() or exit()，child只能讀不能寫parent的address space，這個技術是在COW還沒發展前的技術(COW更好)，所以新版的linux漸漸淘汰掉這個機制。

> `clone(CLONE_VFORK | CLONE_VM | SIGCHLD, 0);`

<linux/sched.h> 定義clone flags如下

**fork()**+**copy on wright(COW)**

疑問: linux fork 子行程的時候，子行程直接複製parent大多數的內容，大多數子行程被生出來後又執行exec()跑去做別的事，這樣的設計不是很浪費記憶體而且複製資料也花時間不是嗎?

A:加上copy on wright技術，完全沒問題，只有在child有寫到shared的東西的時候，才會真正複製給child，如果child一出生就執行exec，那整個life cycle更不會有複製大量資料的時間

> In Linux, fork() utilizes copy-on-write pages to delay or altogether avoid copying the data that can be shared between the parent and child processes. Thus, the only overhead that is incurred during a normal fork() is the **copying of the parent's page tables** and **the assignment of a unique process descriptor struct, task_struct, for the child**.

來看這個
https://unix.stackexchange.com/questions/87551/which-file-in-kernel-specifies-fork-vfork-to-use-sys-clone-system-call

**Kernel thread & normal process的差別**

> kernel threads do not have an address space. (Their mm pointer, which points at their address space, is NULL.) They operate only in kernel-space and do not context switch into user-space. Kernel threads, however, are schedulable and preemptable, the same as normal processes.
>

Q: Can a thread of one process A share data with the thread of another process B?

A: yes, through IPC, A's thread 跟 B's thread沒有共用address space，但可透過IPC溝通