af_netlink.c

《嵌入式系统设计与实例开发实验教材二源码》Linux内核移植与编译实验

/*
 * NETLINK      Kernel-user communication protocol.
 *
 * 		Authors:	Alan Cox <[e-mail]>
 * 				Alexey Kuznetsov <[e-mail]>
 *
 *		This program is free software; you can redistribute it and/or
 *		modify it under the terms of the GNU General Public License
 *		as published by the Free Software Foundation; either version
 *		2 of the License, or (at your option) any later version.
 * 
 * Tue Jun 26 14:36:48 MEST 2001 Herbert "herp" Rosmanith
 *                               added netlink_proto_exit
 *
 */

#include <linux/config.h>
#include <linux/module.h>

#include <linux/kernel.h>
#include <linux/init.h>
#include <linux/major.h>
#include <linux/signal.h>
#include <linux/sched.h>
#include <linux/errno.h>
#include <linux/string.h>
#include <linux/stat.h>
#include <linux/socket.h>
#include <linux/un.h>
#include <linux/fcntl.h>
#include <linux/termios.h>
#include <linux/sockios.h>
#include <linux/net.h>
#include <linux/fs.h>
#include <linux/slab.h>
#include <asm/uaccess.h>
#include <linux/skbuff.h>
#include <linux/netdevice.h>
#include <linux/rtnetlink.h>
#include <linux/proc_fs.h>
#include <linux/smp_lock.h>
#include <net/sock.h>
#include <net/scm.h>

#define Nprintk(a...)

#if defined(CONFIG_NETLINK_DEV) || defined(CONFIG_NETLINK_DEV_MODULE)
#define NL_EMULATE_DEV
#endif

struct netlink_opt
{
	u32			pid;
	unsigned		groups;
	u32			dst_pid;
	unsigned		dst_groups;
	unsigned long		state;
	int			(*handler)(int unit, struct sk_buff *skb);
	wait_queue_head_t	wait;
	struct netlink_callback	*cb;
	spinlock_t		cb_lock;
	void			(*data_ready)(struct sock *sk, int bytes);
};

static struct sock *nl_table[MAX_LINKS];
static DECLARE_WAIT_QUEUE_HEAD(nl_table_wait);

#ifdef NL_EMULATE_DEV
static struct socket *netlink_kernel[MAX_LINKS];
#endif

static int netlink_dump(struct sock *sk);
static void netlink_destroy_callback(struct netlink_callback *cb);

atomic_t netlink_sock_nr;

static rwlock_t nl_table_lock = RW_LOCK_UNLOCKED;
static atomic_t nl_table_users = ATOMIC_INIT(0);

static void netlink_sock_destruct(struct sock *sk)
{
	skb_queue_purge(&sk->receive_queue);

	if (!sk->dead) {
		printk("Freeing alive netlink socket %p\n", sk);
		return;
	}
	BUG_TRAP(atomic_read(&sk->rmem_alloc)==0);
	BUG_TRAP(atomic_read(&sk->wmem_alloc)==0);
	BUG_TRAP(sk->protinfo.af_netlink->cb==NULL);

	kfree(sk->protinfo.af_netlink);

	atomic_dec(&netlink_sock_nr);
#ifdef NETLINK_REFCNT_DEBUG
	printk(KERN_DEBUG "NETLINK %p released, %d are still alive\n", sk, atomic_read(&netlink_sock_nr));
#endif
}

/* This lock without WQ_FLAG_EXCLUSIVE is good on UP and it is _very_ bad on SMP.
 * Look, when several writers sleep and reader wakes them up, all but one
 * immediately hit write lock and grab all the cpus. Exclusive sleep solves
 * this, _but_ remember, it adds useless work on UP machines.
 */

static void netlink_table_grab(void)
{
	write_lock_bh(&nl_table_lock);

	if (atomic_read(&nl_table_users)) {
		DECLARE_WAITQUEUE(wait, current);

		add_wait_queue_exclusive(&nl_table_wait, &wait);
		for(;;) {
			set_current_state(TASK_UNINTERRUPTIBLE);
			if (atomic_read(&nl_table_users) == 0)
				break;
			write_unlock_bh(&nl_table_lock);
			schedule();
			write_lock_bh(&nl_table_lock);
		}

		__set_current_state(TASK_RUNNING);
		remove_wait_queue(&nl_table_wait, &wait);
	}
}

static __inline__ void netlink_table_ungrab(void)
{
	write_unlock_bh(&nl_table_lock);
	wake_up(&nl_table_wait);
}

static __inline__ void
netlink_lock_table(void)
{
	/* read_lock() synchronizes us to netlink_table_grab */

	read_lock(&nl_table_lock);
	atomic_inc(&nl_table_users);
	read_unlock(&nl_table_lock);
}

static __inline__ void
netlink_unlock_table(void)
{
	if (atomic_dec_and_test(&nl_table_users))
		wake_up(&nl_table_wait);
}

static __inline__ struct sock *netlink_lookup(int protocol, u32 pid)
{
	struct sock *sk;

	read_lock(&nl_table_lock);
	for (sk=nl_table[protocol]; sk; sk=sk->next) {
		if (sk->protinfo.af_netlink->pid == pid) {
			sock_hold(sk);
			read_unlock(&nl_table_lock);
			return sk;
		}
	}

	read_unlock(&nl_table_lock);
	return NULL;
}

extern struct proto_ops netlink_ops;

static int netlink_insert(struct sock *sk, u32 pid)
{
	int err = -EADDRINUSE;
	struct sock *osk;

	netlink_table_grab();
	for (osk=nl_table[sk->protocol]; osk; osk=osk->next) {
		if (osk->protinfo.af_netlink->pid == pid)
			break;
	}
	if (osk == NULL) {
		err = -EBUSY;
		if (sk->protinfo.af_netlink->pid == 0) {
			sk->protinfo.af_netlink->pid = pid;
			sk->next = nl_table[sk->protocol];
			nl_table[sk->protocol] = sk;
			sock_hold(sk);
			err = 0;
		}
	}
	netlink_table_ungrab();
	return err;
}

static void netlink_remove(struct sock *sk)
{
	struct sock **skp;

	netlink_table_grab();
	for (skp = &nl_table[sk->protocol]; *skp; skp = &((*skp)->next)) {
		if (*skp == sk) {
			*skp = sk->next;
			__sock_put(sk);
			break;
		}
	}
	netlink_table_ungrab();
}

static int netlink_create(struct socket *sock, int protocol)
{
	struct sock *sk;

	sock->state = SS_UNCONNECTED;

	if (sock->type != SOCK_RAW && sock->type != SOCK_DGRAM)
		return -ESOCKTNOSUPPORT;

	if (protocol<0 || protocol >= MAX_LINKS)
		return -EPROTONOSUPPORT;

	sock->ops = &netlink_ops;

	sk = sk_alloc(PF_NETLINK, GFP_KERNEL, 1);
	if (!sk)
		return -ENOMEM;

	sock_init_data(sock,sk);

	sk->protinfo.af_netlink = kmalloc(sizeof(struct netlink_opt), GFP_KERNEL);
	if (sk->protinfo.af_netlink == NULL) {
		sk_free(sk);
		return -ENOMEM;
	}
	memset(sk->protinfo.af_netlink, 0, sizeof(struct netlink_opt));

	spin_lock_init(&sk->protinfo.af_netlink->cb_lock);
	init_waitqueue_head(&sk->protinfo.af_netlink->wait);
	sk->destruct = netlink_sock_destruct;
	atomic_inc(&netlink_sock_nr);

	sk->protocol=protocol;
	return 0;
}

static int netlink_release(struct socket *sock)
{
	struct sock *sk = sock->sk;

	if (!sk)
		return 0;

	netlink_remove(sk);

	spin_lock(&sk->protinfo.af_netlink->cb_lock);
	if (sk->protinfo.af_netlink->cb) {
		sk->protinfo.af_netlink->cb->done(sk->protinfo.af_netlink->cb);
		netlink_destroy_callback(sk->protinfo.af_netlink->cb);
		sk->protinfo.af_netlink->cb = NULL;
		__sock_put(sk);
	}
	spin_unlock(&sk->protinfo.af_netlink->cb_lock);

	/* OK. Socket is unlinked, and, therefore,
	   no new packets will arrive */

	sock_orphan(sk);
	sock->sk = NULL;
	wake_up_interruptible_all(&sk->protinfo.af_netlink->wait);

	skb_queue_purge(&sk->write_queue);

	sock_put(sk);
	return 0;
}

static int netlink_autobind(struct socket *sock)
{
	struct sock *sk = sock->sk;
	struct sock *osk;
	s32 pid = current->pid;
	int err;

retry:
	netlink_table_grab();
	for (osk=nl_table[sk->protocol]; osk; osk=osk->next) {
		if (osk->protinfo.af_netlink->pid == pid) {
			/* Bind collision, search negative pid values. */
			if (pid > 0)
				pid = -4096;
			pid--;
			netlink_table_ungrab();
			goto retry;
		}
	}
	netlink_table_ungrab();

	err = netlink_insert(sk, pid);
	if (err == -EADDRINUSE)
		goto retry;
	sk->protinfo.af_netlink->groups = 0;
	return 0;
}

static int netlink_bind(struct socket *sock, struct sockaddr *addr, int addr_len)
{
	struct sock *sk = sock->sk;
	int err;
	struct sockaddr_nl *nladdr=(struct sockaddr_nl *)addr;
	
	if (nladdr->nl_family != AF_NETLINK)
		return -EINVAL;

	/* Only superuser is allowed to listen multicasts */
	if (nladdr->nl_groups && !capable(CAP_NET_ADMIN))
		return -EPERM;

	if (sk->protinfo.af_netlink->pid) {
		if (nladdr->nl_pid != sk->protinfo.af_netlink->pid)
			return -EINVAL;
		sk->protinfo.af_netlink->groups = nladdr->nl_groups;
		return 0;
	}

	if (nladdr->nl_pid == 0) {
		err = netlink_autobind(sock);
		if (err == 0)
			sk->protinfo.af_netlink->groups = nladdr->nl_groups;
		return err;
	}

	err = netlink_insert(sk, nladdr->nl_pid);
	if (err == 0)
		sk->protinfo.af_netlink->groups = nladdr->nl_groups;
	return err;
}

static int netlink_connect(struct socket *sock, struct sockaddr *addr,
			   int alen, int flags)
{
	int err = 0;
	struct sock *sk = sock->sk;
	struct sockaddr_nl *nladdr=(struct sockaddr_nl*)addr;

	if (addr->sa_family == AF_UNSPEC) {
		sk->protinfo.af_netlink->dst_pid = 0;
		sk->protinfo.af_netlink->dst_groups = 0;
		return 0;
	}
	if (addr->sa_family != AF_NETLINK)
		return -EINVAL;

	/* Only superuser is allowed to send multicasts */
	if (nladdr->nl_groups && !capable(CAP_NET_ADMIN))
		return -EPERM;

	if (!sk->protinfo.af_netlink->pid)
		err = netlink_autobind(sock);

	if (err == 0) {
		sk->protinfo.af_netlink->dst_pid = nladdr->nl_pid;
		sk->protinfo.af_netlink->dst_groups = nladdr->nl_groups;
	}

	return 0;
}

static int netlink_getname(struct socket *sock, struct sockaddr *addr, int *addr_len, int peer)
{
	struct sock *sk = sock->sk;
	struct sockaddr_nl *nladdr=(struct sockaddr_nl *)addr;
	
	nladdr->nl_family = AF_NETLINK;
	*addr_len = sizeof(*nladdr);

	if (peer) {
		nladdr->nl_pid = sk->protinfo.af_netlink->dst_pid;
		nladdr->nl_groups = sk->protinfo.af_netlink->dst_groups;
	} else {
		nladdr->nl_pid = sk->protinfo.af_netlink->pid;
		nladdr->nl_groups = sk->protinfo.af_netlink->groups;
	}
	return 0;
}

static void netlink_overrun(struct sock *sk)
{
	if (!test_and_set_bit(0, &sk->protinfo.af_netlink->state)) {
		sk->err = ENOBUFS;
		sk->error_report(sk);
	}
}

int netlink_unicast(struct sock *ssk, struct sk_buff *skb, u32 pid, int nonblock)
{
	struct sock *sk;
	int len = skb->len;
	int protocol = ssk->protocol;
	long timeo;
        DECLARE_WAITQUEUE(wait, current);

	timeo = sock_sndtimeo(ssk, nonblock);

retry:
	sk = netlink_lookup(protocol, pid);
	if (sk == NULL)
		goto no_dst;

#ifdef NL_EMULATE_DEV
	if (sk->protinfo.af_netlink->handler) {
		skb_orphan(skb);
		len = sk->protinfo.af_netlink->handler(protocol, skb);
		sock_put(sk);
		return len;
	}
#endif

	if (atomic_read(&sk->rmem_alloc) > sk->rcvbuf ||
	    test_bit(0, &sk->protinfo.af_netlink->state)) {
		if (!timeo) {
			if (ssk->protinfo.af_netlink->pid == 0)
				netlink_overrun(sk);
			sock_put(sk);
			kfree_skb(skb);
			return -EAGAIN;
		}

		__set_current_state(TASK_INTERRUPTIBLE);
		add_wait_queue(&sk->protinfo.af_netlink->wait, &wait);

		if ((atomic_read(&sk->rmem_alloc) > sk->rcvbuf ||
		    test_bit(0, &sk->protinfo.af_netlink->state)) &&
		    !sk->dead)
			timeo = schedule_timeout(timeo);

		__set_current_state(TASK_RUNNING);
		remove_wait_queue(&sk->protinfo.af_netlink->wait, &wait);
		sock_put(sk);

		if (signal_pending(current)) {
			kfree_skb(skb);
			return sock_intr_errno(timeo);
		}
		goto retry;
	}

	skb_orphan(skb);
	skb_set_owner_r(skb, sk);
	skb_queue_tail(&sk->receive_queue, skb);
	sk->data_ready(sk, len);
	sock_put(sk);
	return len;

no_dst:
	kfree_skb(skb);
	return -ECONNREFUSED;
}

static __inline__ int netlink_broadcast_deliver(struct sock *sk, struct sk_buff *skb)
{
#ifdef NL_EMULATE_DEV
	if (sk->protinfo.af_netlink->handler) {
		skb_orphan(skb);
		sk->protinfo.af_netlink->handler(sk->protocol, skb);
		return 0;
	} else
#endif
	if (atomic_read(&sk->rmem_alloc) <= sk->rcvbuf &&
	    !test_bit(0, &sk->protinfo.af_netlink->state)) {
                skb_orphan(skb);
		skb_set_owner_r(skb, sk);
		skb_queue_tail(&sk->receive_queue, skb);
		sk->data_ready(sk, skb->len);
		return 0;
	}
	return -1;
}

void netlink_broadcast(struct sock *ssk, struct sk_buff *skb, u32 pid,
		       u32 group, int allocation)
{
	struct sock *sk;
	struct sk_buff *skb2 = NULL;
	int protocol = ssk->protocol;
	int failure = 0;

	/* While we sleep in clone, do not allow to change socket list */

	netlink_lock_table();

	for (sk = nl_table[protocol]; sk; sk = sk->next) {
		if (ssk == sk)
			continue;

		if (sk->protinfo.af_netlink->pid == pid ||
		    !(sk->protinfo.af_netlink->groups&group))
			continue;

		if (failure) {
			netlink_overrun(sk);