linux 内核klist分析

---

记录一下内核的的klist实现，以防以后忘记。内核目录文件所在位置：\include\linux\klist.c和\include\linux\klist.h

本文仅仅是对源代码进行分析，分析的部分参看每段代码的注释部分。

关于klist详细的文字介绍，请参考其他文章：

Linux内核中的klist分析

linux内核源代码分析----内核基础设施之klist

Linux内核链表list_head扩展---klist

linux内核部件分析（四）——更强的链表klist

---

 

 

 

---

进入正文：klist.h代码如下

---

/*
 *	klist.h - Some generic list helpers, extending struct list_head a bit.
 *
 *	Implementations are found in lib/klist.c
 *
 *
 *	Copyright (C) 2005 Patrick Mochel
 *
 *	This file is rleased under the GPL v2.
 */

#ifndef _LINUX_KLIST_H
#define _LINUX_KLIST_H

#include <linux/spinlock.h>
#include <linux/kref.h>
#include <linux/list.h>

struct klist_node;

/* 注意这里的  __attribute__ ((aligned (sizeof(void *))))对齐属性，在以后会用的，
 * 假如是4字节对齐，因此 struct klist结构体的地址的最低两位永远都是00，不可能出现
 * 1,2,3，因此内核局可以利用该地址的最低两位，一般是由 struct klist_node节点指向
 * struct klist节点的地址，因此struct klist_node节点可以利用struct klist节点对齐做一些文章。
 *
 */
struct klist {
	spinlock_t		k_lock;
	struct list_head	k_list;
	void			(*get)(struct klist_node *);
	void			(*put)(struct klist_node *);
} __attribute__ ((aligned (sizeof(void *))));


#define KLIST_INIT(_name, _get, _put)					\
	{ .k_lock	= __SPIN_LOCK_UNLOCKED(_name.k_lock),		\
	  .k_list	= LIST_HEAD_INIT(_name.k_list),			\
	  .get		= _get,						\
	  .put		= _put, }

#define DEFINE_KLIST(_name, _get, _put)					\
	struct klist _name = KLIST_INIT(_name, _get, _put)

extern void klist_init(struct klist *k, void (*get)(struct klist_node *),
		       void (*put)(struct klist_node *));

struct klist_node {
	void			*n_klist;	/* never access directly */
	struct list_head	n_node;
	struct kref		n_ref;
};

extern void klist_add_tail(struct klist_node *n, struct klist *k);
extern void klist_add_head(struct klist_node *n, struct klist *k);
extern void klist_add_behind(struct klist_node *n, struct klist_node *pos);
extern void klist_add_before(struct klist_node *n, struct klist_node *pos);

extern void klist_del(struct klist_node *n);
extern void klist_remove(struct klist_node *n);

extern int klist_node_attached(struct klist_node *n);

/* 该结构体用于klist迭代使用，也就是为了遍历klist链表中的klist_node节点 */
struct klist_iter {
	struct klist		*i_klist;
	struct klist_node	*i_cur;
};


extern void klist_iter_init(struct klist *k, struct klist_iter *i);
extern void klist_iter_init_node(struct klist *k, struct klist_iter *i,
				 struct klist_node *n);
extern void klist_iter_exit(struct klist_iter *i);
extern struct klist_node *klist_next(struct klist_iter *i);

#endif

 

klist.c

---

/*
 * klist.c - Routines for manipulating klists.
 *
 * Copyright (C) 2005 Patrick Mochel
 *
 * This file is released under the GPL v2.
 *
 * This klist interface provides a couple of structures that wrap around
 * struct list_head to provide explicit list "head" (struct klist) and list
 * "node" (struct klist_node) objects. For struct klist, a spinlock is
 * included that protects access to the actual list itself. struct
 * klist_node provides a pointer to the klist that owns it and a kref
 * reference count that indicates the number of current users of that node
 * in the list.
 *
 * The entire point is to provide an interface for iterating over a list
 * that is safe and allows for modification of the list during the
 * iteration (e.g. insertion and removal), including modification of the
 * current node on the list.
 *
 * It works using a 3rd object type - struct klist_iter - that is declared
 * and initialized before an iteration. klist_next() is used to acquire the
 * next element in the list. It returns NULL if there are no more items.
 * Internally, that routine takes the klist's lock, decrements the
 * reference count of the previous klist_node and increments the count of
 * the next klist_node. It then drops the lock and returns.
 *
 * There are primitives for adding and removing nodes to/from a klist.
 * When deleting, klist_del() will simply decrement the reference count.
 * Only when the count goes to 0 is the node removed from the list.
 * klist_remove() will try to delete the node from the list and block until
 * it is actually removed. This is useful for objects (like devices) that
 * have been removed from the system and must be freed (but must wait until
 * all accessors have finished).
 */

#include <linux/klist.h>
#include <linux/export.h>
#include <linux/sched.h>

/*
 * Use the lowest bit of n_klist to mark deleted nodes and exclude
 * dead ones from iteration.
 */
#define KNODE_DEAD		1LU
#define KNODE_KLIST_MASK	~KNODE_DEAD

/* 从klist_node节点中得到链表头 */
static struct klist *knode_klist(struct klist_node *knode)
{
	return (struct klist *)
		((unsigned long)knode->n_klist & KNODE_KLIST_MASK);
}

static bool knode_dead(struct klist_node *knode)
{
	return (unsigned long)knode->n_klist & KNODE_DEAD;
}

static void knode_set_klist(struct klist_node *knode, struct klist *klist)
{
	knode->n_klist = klist;
	/* no knode deserves to start its life dead */
	WARN_ON(knode_dead(knode));
}

/* klist_node中的n_klist指针的最低位肯定是0，因为struct klist结构体是4字节对齐的
 * 因此巧妙的利用最低位来表示该节点是否被删除
 */
static void knode_kill(struct klist_node *knode)
{
	/* and no knode should die twice ever either, see we're very humane */
	WARN_ON(knode_dead(knode));
	*(unsigned long *)&knode->n_klist |= KNODE_DEAD;
}

/**
 * klist_init - Initialize a klist structure.
 * @k: The klist we're initializing.
 * @get: The get function for the embedding object (NULL if none)
 * @put: The put function for the embedding object (NULL if none)
 *
 * Initialises the klist structure.  If the klist_node structures are
 * going to be embedded in refcounted objects (necessary for safe
 * deletion) then the get/put arguments are used to initialise
 * functions that take and release references on the embedding
 * objects.
 */
void klist_init(struct klist *k, void (*get)(struct klist_node *),
		void (*put)(struct klist_node *))
{
	INIT_LIST_HEAD(&k->k_list);
	spin_lock_init(&k->k_lock);
	k->get = get;
	k->put = put;
}
EXPORT_SYMBOL_GPL(klist_init);

static void add_head(struct klist *k, struct klist_node *n)
{
	spin_lock(&k->k_lock);
	list_add(&n->n_node, &k->k_list);
	spin_unlock(&k->k_lock);
}

static void add_tail(struct klist *k, struct klist_node *n)
{
	spin_lock(&k->k_lock);
	list_add_tail(&n->n_node, &k->k_list);
	spin_unlock(&k->k_lock);
}

static void klist_node_init(struct klist *k, struct klist_node *n)
{
	INIT_LIST_HEAD(&n->n_node);
	kref_init(&n->n_ref);
	knode_set_klist(n, k);
	if (k->get)
		k->get(n);
}

/**
 * klist_add_head - Initialize a klist_node and add it to front.
 * @n: node we're adding.
 * @k: klist it's going on.
 */
void klist_add_head(struct klist_node *n, struct klist *k)
{
	klist_node_init(k, n);
	add_head(k, n);
}
EXPORT_SYMBOL_GPL(klist_add_head);

/**
 * klist_add_tail - Initialize a klist_node and add it to back.
 * @n: node we're adding.
 * @k: klist it's going on.
 */
void klist_add_tail(struct klist_node *n, struct klist *k)
{
	klist_node_init(k, n);
	add_tail(k, n);
}
EXPORT_SYMBOL_GPL(klist_add_tail);

/**
 * klist_add_behind - Init a klist_node and add it after an existing node
 * @n: node we're adding.
 * @pos: node to put @n after
 */

/* 该函数接口是在pos后面插入一个n节点，主要原理就是list_add是在链表头加入节点，
 * 因此list_add(&n->n_node, &pos->n_node);就是以pos为链表头，在pos后面添加一个节点。
 * 
 */
void klist_add_behind(struct klist_node *n, struct klist_node *pos)
{
	struct klist *k = knode_klist(pos);

	klist_node_init(k, n);
	spin_lock(&k->k_lock);
	list_add(&n->n_node, &pos->n_node);
	spin_unlock(&k->k_lock);
}
EXPORT_SYMBOL_GPL(klist_add_behind);

/**
 * klist_add_before - Init a klist_node and add it before an existing node
 * @n: node we're adding.
 * @pos: node to put @n after
 */

/* 该函数接口是在pos前面插入一个n节点，主要原理就是list_add_tail是在链表尾部加入节点，
 * 因此list_add_tail(&n->n_node, &pos->n_node);就是以pos为链表头，在pos尾部添加一个节点。
 * 因为pos是一个双向循环链表，以pos为表头的尾部就是pos的前一个节点。
 * 
 */
void klist_add_before(struct klist_node *n, struct klist_node *pos)
{
	struct klist *k = knode_klist(pos);

	klist_node_init(k, n);
	spin_lock(&k->k_lock);
	list_add_tail(&n->n_node, &pos->n_node);
	spin_unlock(&k->k_lock);
}
EXPORT_SYMBOL_GPL(klist_add_before);

/* 该结构体用于线程阻塞 */
struct klist_waiter {
	struct list_head list;
	struct klist_node *node;
	struct task_struct *process;
	int woken;
};

static DEFINE_SPINLOCK(klist_remove_lock);
static LIST_HEAD(klist_remove_waiters);

static void klist_release(struct kref *kref)
{
	struct klist_waiter *waiter, *tmp;
	struct klist_node *n = container_of(kref, struct klist_node, n_ref);

	WARN_ON(!knode_dead(n));
	/* 从struct klist链表头中删除klist_node */
	list_del(&n->n_node);
	spin_lock(&klist_remove_lock);
	/* 遍历klist_remove_waiters链表，删除所有阻塞在删除n 节点的waiter，然后唤醒进程 */
	list_for_each_entry_safe(waiter, tmp, &klist_remove_waiters, list) {
		if (waiter->node != n)
			continue;

		list_del(&waiter->list);
		waiter->woken = 1;
		mb();
		wake_up_process(waiter->process);
	}
	spin_unlock(&klist_remove_lock);
	knode_set_klist(n, NULL);
}

static int klist_dec_and_del(struct klist_node *n)
{
	return kref_put(&n->n_ref, klist_release);
}

static void klist_put(struct klist_node *n, bool kill)
{
	struct klist *k = knode_klist(n);
	void (*put)(struct klist_node *) = k->put;

	spin_lock(&k->k_lock);
	if (kill)
		knode_kill(n);
	if (!klist_dec_and_del(n))
		put = NULL;
	spin_unlock(&k->k_lock);
	if (put)
		put(n);
}

/**
 * klist_del - Decrement the reference count of node and try to remove.
 * @n: node we're deleting.
 */
void klist_del(struct klist_node *n)
{
	klist_put(n, true);
}
EXPORT_SYMBOL_GPL(klist_del);

/**
 * klist_remove - Decrement the refcount of node and wait for it to go away.
 * @n: node we're removing.
 */
void klist_remove(struct klist_node *n)
{
	struct klist_waiter waiter;
	/* 将当前进程和要删除的klist_node放进klist_waiter结构体中
	 * 值得注意的是如果多个线程要删除同一个klist_node节点的话，这里n的地址是一样的，
	 * 这在以后唤醒进程时，将会用到，也就是说将会唤醒所有阻塞在删除n节点的进程。
	 */
	waiter.node = n;
	waiter.process = current;
	waiter.woken = 0;
	spin_lock(&klist_remove_lock);
	/* 将线程等待者放入klist_remove_waiters链表中 */
	list_add(&waiter.list, &klist_remove_waiters);
	spin_unlock(&klist_remove_lock);

	klist_del(n);

	for (;;) {
		set_current_state(TASK_UNINTERRUPTIBLE);
		if (waiter.woken)
			break;
		schedule();
	}
	__set_current_state(TASK_RUNNING);
}
EXPORT_SYMBOL_GPL(klist_remove);

/**
 * klist_node_attached - Say whether a node is bound to a list or not.
 * @n: Node that we're testing.
 */
int klist_node_attached(struct klist_node *n)
{
	return (n->n_klist != NULL);
}
EXPORT_SYMBOL_GPL(klist_node_attached);

/**
 * klist_iter_init_node - Initialize a klist_iter structure.
 * @k: klist we're iterating.
 * @i: klist_iter we're filling.
 * @n: node to start with.
 *
 * Similar to klist_iter_init(), but starts the action off with @n,
 * instead of with the list head.
 */
void klist_iter_init_node(struct klist *k, struct klist_iter *i,
			  struct klist_node *n)
{
	i->i_klist = k;
	i->i_cur = n;
	if (n)
		kref_get(&n->n_ref);
}
EXPORT_SYMBOL_GPL(klist_iter_init_node);

/**
 * klist_iter_init - Iniitalize a klist_iter structure.
 * @k: klist we're iterating.
 * @i: klist_iter structure we're filling.
 *
 * Similar to klist_iter_init_node(), but start with the list head.
 */
void klist_iter_init(struct klist *k, struct klist_iter *i)
{
	klist_iter_init_node(k, i, NULL);
}
EXPORT_SYMBOL_GPL(klist_iter_init);

/**
 * klist_iter_exit - Finish a list iteration.
 * @i: Iterator structure.
 *
 * Must be called when done iterating over list, as it decrements the
 * refcount of the current node. Necessary in case iteration exited before
 * the end of the list was reached, and always good form.
 */
void klist_iter_exit(struct klist_iter *i)
{
	if (i->i_cur) {
		klist_put(i->i_cur, false);
		i->i_cur = NULL;
	}
}
EXPORT_SYMBOL_GPL(klist_iter_exit);

static struct klist_node *to_klist_node(struct list_head *n)
{
	return container_of(n, struct klist_node, n_node);
}

/**
 * klist_next - Ante up next node in list.
 * @i: Iterator structure.
 *
 * First grab list lock. Decrement the reference count of the previous
 * node, if there was one. Grab the next node, increment its reference
 * count, drop the lock, and return that next node.
 */

/* 迭代器的主要功能 */
struct klist_node *klist_next(struct klist_iter *i)
{
	void (*put)(struct klist_node *) = i->i_klist->put;
	/* 获得迭代器的当前节点 */
	struct klist_node *last = i->i_cur;
	struct klist_node *next;

	spin_lock(&i->i_klist->k_lock);

	if (last) {
	/* 该分支表示以klist_iter中的klist_node节点为参考节点,取出参考节点中的下一个klist_node */
		next = to_klist_node(last->n_node.next);
		if (!klist_dec_and_del(last))
			put = NULL;
	} else
		/* 该分支表示klist_iter中没有指定当前klist_node节点，则从 klist中第一个节点算为参考节点，
		 * 取出参考节点中的下一个klist_node。
		 */
		next = to_klist_node(i->i_klist->k_list.next);

	i->i_cur = NULL;
	/* 将next节点与klist_iter中的头节点比较，如果链表中的某个klist_node已经被dead，那么不遍历该节点 */
	while (next != to_klist_node(&i->i_klist->k_list)) {
		if (likely(!knode_dead(next))) {
			kref_get(&next->n_ref);
			i->i_cur = next;
			break;
		}
		/* 每次next与头节点进行比较，这样会比较所有从该next节点以后到头结点之间的节点 */
		next = to_klist_node(next->n_node.next);
	}

	spin_unlock(&i->i_klist->k_lock);

	if (put && last)
		put(last);
	return i->i_cur;
}
EXPORT_SYMBOL_GPL(klist_next);