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

			continue;
		}

		sock_hold(sk);
		if (skb2 == NULL) {
			if (atomic_read(&skb->users) != 1) {
				skb2 = skb_clone(skb, allocation);
			} else {
				skb2 = skb;
				atomic_inc(&skb->users);
			}
		}
		if (skb2 == NULL) {
			netlink_overrun(sk);
			/* Clone failed. Notify ALL listeners. */
			failure = 1;
		} else if (netlink_broadcast_deliver(sk, skb2)) {
			netlink_overrun(sk);
		} else
			skb2 = NULL;
		sock_put(sk);
	}

	netlink_unlock_table();

	if (skb2)
		kfree_skb(skb2);
	kfree_skb(skb);
}

void netlink_set_err(struct sock *ssk, u32 pid, u32 group, int code)
{
	struct sock *sk;
	int protocol = ssk->protocol;

	read_lock(&nl_table_lock);
	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;

		sk->err = code;
		sk->error_report(sk);
	}
	read_unlock(&nl_table_lock);
}

static int netlink_sendmsg(struct socket *sock, struct msghdr *msg, int len,
			   struct scm_cookie *scm)
{
	struct sock *sk = sock->sk;
	struct sockaddr_nl *addr=msg->msg_name;
	u32 dst_pid;
	u32 dst_groups;
	struct sk_buff *skb;
	int err;

	if (msg->msg_flags&MSG_OOB)
		return -EOPNOTSUPP;

	if (msg->msg_namelen) {
		if (addr->nl_family != AF_NETLINK)
			return -EINVAL;
		dst_pid = addr->nl_pid;
		dst_groups = addr->nl_groups;
		if (dst_groups && !capable(CAP_NET_ADMIN))
			return -EPERM;
	} else {
		dst_pid = sk->protinfo.af_netlink->dst_pid;
		dst_groups = sk->protinfo.af_netlink->dst_groups;
	}

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

	err = -EMSGSIZE;
	if ((unsigned)len > sk->sndbuf-32)
		goto out;
	err = -ENOBUFS;
	skb = alloc_skb(len, GFP_KERNEL);
	if (skb==NULL)
		goto out;

	NETLINK_CB(skb).pid = sk->protinfo.af_netlink->pid;
	NETLINK_CB(skb).groups = sk->protinfo.af_netlink->groups;
	NETLINK_CB(skb).dst_pid = dst_pid;
	NETLINK_CB(skb).dst_groups = dst_groups;
	memcpy(NETLINK_CREDS(skb), &scm->creds, sizeof(struct ucred));

	/* What can I do? Netlink is asynchronous, so that
	   we will have to save current capabilities to
	   check them, when this message will be delivered
	   to corresponding kernel module.   --ANK (980802)
	 */
	NETLINK_CB(skb).eff_cap = current->cap_effective;

	err = -EFAULT;
	if (memcpy_fromiovec(skb_put(skb,len), msg->msg_iov, len)) {
		kfree_skb(skb);
		goto out;
	}

	if (dst_groups) {
		atomic_inc(&skb->users);
		netlink_broadcast(sk, skb, dst_pid, dst_groups, GFP_KERNEL);
	}
	err = netlink_unicast(sk, skb, dst_pid, msg->msg_flags&MSG_DONTWAIT);

out:
	return err;
}

static int netlink_recvmsg(struct socket *sock, struct msghdr *msg, int len,
			   int flags, struct scm_cookie *scm)
{
	struct sock *sk = sock->sk;
	int noblock = flags&MSG_DONTWAIT;
	int copied;
	struct sk_buff *skb;
	int err;

	if (flags&MSG_OOB)
		return -EOPNOTSUPP;

	copied = 0;

	skb = skb_recv_datagram(sk,flags,noblock,&err);
	if (skb==NULL)
		goto out;

	msg->msg_namelen = 0;

	copied = skb->len;
	if (len < copied) {
		msg->msg_flags |= MSG_TRUNC;
		copied = len;
	}

	skb->h.raw = skb->data;
	err = skb_copy_datagram_iovec(skb, 0, msg->msg_iov, copied);

	if (msg->msg_name) {
		struct sockaddr_nl *addr = (struct sockaddr_nl*)msg->msg_name;
		addr->nl_family = AF_NETLINK;
		addr->nl_pid	= NETLINK_CB(skb).pid;
		addr->nl_groups	= NETLINK_CB(skb).dst_groups;
		msg->msg_namelen = sizeof(*addr);
	}

	scm->creds = *NETLINK_CREDS(skb);
	skb_free_datagram(sk, skb);

	if (sk->protinfo.af_netlink->cb
	    && atomic_read(&sk->rmem_alloc) <= sk->rcvbuf/2)
		netlink_dump(sk);

out:
	if (skb_queue_len(&sk->receive_queue) <= sk->rcvbuf/2) {
		if (skb_queue_len(&sk->receive_queue) == 0)
			clear_bit(0, &sk->protinfo.af_netlink->state);
		if (!test_bit(0, &sk->protinfo.af_netlink->state))
			wake_up_interruptible(&sk->protinfo.af_netlink->wait);
	}
	return err ? : copied;
}

void netlink_data_ready(struct sock *sk, int len)
{
	if (sk->protinfo.af_netlink->data_ready)
		sk->protinfo.af_netlink->data_ready(sk, len);

	if (skb_queue_len(&sk->receive_queue) <= sk->rcvbuf/2) {
		if (skb_queue_len(&sk->receive_queue) == 0)
			clear_bit(0, &sk->protinfo.af_netlink->state);
		if (!test_bit(0, &sk->protinfo.af_netlink->state))
			wake_up_interruptible(&sk->protinfo.af_netlink->wait);
	}
}

/*
 *	We export these functions to other modules. They provide a 
 *	complete set of kernel non-blocking support for message
 *	queueing.
 */

struct sock *
netlink_kernel_create(int unit, void (*input)(struct sock *sk, int len))
{
	struct socket *sock;
	struct sock *sk;

	if (unit<0 || unit>=MAX_LINKS)
		return NULL;

	if (!(sock = sock_alloc())) 
		return NULL;

	sock->type = SOCK_RAW;

	if (netlink_create(sock, unit) < 0) {
		sock_release(sock);
		return NULL;
	}
	sk = sock->sk;
	sk->data_ready = netlink_data_ready;
	if (input)
		sk->protinfo.af_netlink->data_ready = input;

	netlink_insert(sk, 0);
	return sk;
}

static void netlink_destroy_callback(struct netlink_callback *cb)
{
	if (cb->skb)
		kfree_skb(cb->skb);
	kfree(cb);
}

/*
 * It looks a bit ugly.
 * It would be better to create kernel thread.
 */

static int netlink_dump(struct sock *sk)
{
	struct netlink_callback *cb;
	struct sk_buff *skb;
	struct nlmsghdr *nlh;
	int len;
	
	skb = sock_rmalloc(sk, NLMSG_GOODSIZE, 0, GFP_KERNEL);
	if (!skb)
		return -ENOBUFS;

	spin_lock(&sk->protinfo.af_netlink->cb_lock);

	cb = sk->protinfo.af_netlink->cb;
	if (cb == NULL) {
		spin_unlock(&sk->protinfo.af_netlink->cb_lock);
		kfree_skb(skb);
		return -EINVAL;
	}

	len = cb->dump(skb, cb);

	if (len > 0) {
		spin_unlock(&sk->protinfo.af_netlink->cb_lock);
		skb_queue_tail(&sk->receive_queue, skb);
		sk->data_ready(sk, len);
		return 0;
	}

	nlh = __nlmsg_put(skb, NETLINK_CB(cb->skb).pid, cb->nlh->nlmsg_seq, NLMSG_DONE, sizeof(int));
	nlh->nlmsg_flags |= NLM_F_MULTI;
	memcpy(NLMSG_DATA(nlh), &len, sizeof(len));
	skb_queue_tail(&sk->receive_queue, skb);
	sk->data_ready(sk, skb->len);

	cb->done(cb);
	sk->protinfo.af_netlink->cb = NULL;
	spin_unlock(&sk->protinfo.af_netlink->cb_lock);

	netlink_destroy_callback(cb);
	sock_put(sk);
	return 0;
}

int netlink_dump_start(struct sock *ssk, struct sk_buff *skb,
		       struct nlmsghdr *nlh,
		       int (*dump)(struct sk_buff *skb, struct netlink_callback*),
		       int (*done)(struct netlink_callback*))
{
	struct netlink_callback *cb;
	struct sock *sk;

	cb = kmalloc(sizeof(*cb), GFP_KERNEL);
	if (cb == NULL)
		return -ENOBUFS;

	memset(cb, 0, sizeof(*cb));
	cb->dump = dump;
	cb->done = done;
	cb->nlh = nlh;
	atomic_inc(&skb->users);
	cb->skb = skb;

	sk = netlink_lookup(ssk->protocol, NETLINK_CB(skb).pid);
	if (sk == NULL) {
		netlink_destroy_callback(cb);
		return -ECONNREFUSED;
	}
	/* A dump is in progress... */
	spin_lock(&sk->protinfo.af_netlink->cb_lock);
	if (sk->protinfo.af_netlink->cb) {
		spin_unlock(&sk->protinfo.af_netlink->cb_lock);
		netlink_destroy_callback(cb);
		sock_put(sk);
		return -EBUSY;
	}
	sk->protinfo.af_netlink->cb = cb;
	spin_unlock(&sk->protinfo.af_netlink->cb_lock);

	netlink_dump(sk);
	return 0;
}

void netlink_ack(struct sk_buff *in_skb, struct nlmsghdr *nlh, int err)
{
	struct sk_buff *skb;
	struct nlmsghdr *rep;
	struct nlmsgerr *errmsg;
	int size;

	if (err == 0)
		size = NLMSG_SPACE(sizeof(struct nlmsgerr));
	else
		size = NLMSG_SPACE(4 + NLMSG_ALIGN(nlh->nlmsg_len));

	skb = alloc_skb(size, GFP_KERNEL);
	if (!skb)
		return;

	rep = __nlmsg_put(skb, NETLINK_CB(in_skb).pid, nlh->nlmsg_seq,
			  NLMSG_ERROR, sizeof(struct nlmsgerr));
	errmsg = NLMSG_DATA(rep);
	errmsg->error = err;
	memcpy(&errmsg->msg, nlh, err ? nlh->nlmsg_len : sizeof(struct nlmsghdr));
	netlink_unicast(in_skb->sk, skb, NETLINK_CB(in_skb).pid, MSG_DONTWAIT);
}


#ifdef NL_EMULATE_DEV

static rwlock_t nl_emu_lock = RW_LOCK_UNLOCKED;

/*
 *	Backward compatibility.
 */	
 
int netlink_attach(int unit, int (*function)(int, struct sk_buff *skb))
{
	struct sock *sk = netlink_kernel_create(unit, NULL);
	if (sk == NULL)
		return -ENOBUFS;
	sk->protinfo.af_netlink->handler = function;
	write_lock_bh(&nl_emu_lock);
	netlink_kernel[unit] = sk->socket;
	write_unlock_bh(&nl_emu_lock);
	return 0;
}

void netlink_detach(int unit)
{
	struct socket *sock;

	write_lock_bh(&nl_emu_lock);
	sock = netlink_kernel[unit];
	netlink_kernel[unit] = NULL;
	write_unlock_bh(&nl_emu_lock);

	sock_release(sock);
}

int netlink_post(int unit, struct sk_buff *skb)
{
	struct socket *sock;

	read_lock(&nl_emu_lock);
	sock = netlink_kernel[unit];
	if (sock) {
		struct sock *sk = sock->sk;
		memset(skb->cb, 0, sizeof(skb->cb));
		sock_hold(sk);
		read_unlock(&nl_emu_lock);

		netlink_broadcast(sk, skb, 0, ~0, GFP_ATOMIC);

		sock_put(sk);
		return 0;
	}
	read_unlock(&nl_emu_lock);
	return -EUNATCH;
}

#endif


#ifdef CONFIG_PROC_FS
static int netlink_read_proc(char *buffer, char **start, off_t offset,
			     int length, int *eof, void *data)
{
	off_t pos=0;
	off_t begin=0;
	int len=0;
	int i;
	struct sock *s;
	
	len+= sprintf(buffer,"sk       Eth Pid    Groups   "
		      "Rmem     Wmem     Dump     Locks\n");
	
	for (i=0; i<MAX_LINKS; i++) {
		read_lock(&nl_table_lock);
		for (s = nl_table[i]; s; s = s->next) {
			len+=sprintf(buffer+len,"%p %-3d %-6d %08x %-8d %-8d %p %d",
				     s,
				     s->protocol,
				     s->protinfo.af_netlink->pid,
				     s->protinfo.af_netlink->groups,
				     atomic_read(&s->rmem_alloc),
				     atomic_read(&s->wmem_alloc),
				     s->protinfo.af_netlink->cb,
				     atomic_read(&s->refcnt)
				     );

			buffer[len++]='\n';
		
			pos=begin+len;
			if(pos<offset) {
				len=0;
				begin=pos;
			}
			if(pos>offset+length) {
				read_unlock(&nl_table_lock);
				goto done;
			}
		}
		read_unlock(&nl_table_lock);
	}
	*eof = 1;

done:
	*start=buffer+(offset-begin);
	len-=(offset-begin);
	if(len>length)
		len=length;
	if(len<0)
		len=0;
	return len;
}
#endif

struct proto_ops netlink_ops = {
	family:		PF_NETLINK,

	release:	netlink_release,
	bind:		netlink_bind,
	connect:	netlink_connect,
	socketpair:	sock_no_socketpair,
	accept:		sock_no_accept,
	getname:	netlink_getname,
	poll:		datagram_poll,
	ioctl:		sock_no_ioctl,
	listen:		sock_no_listen,
	shutdown:	sock_no_shutdown,
	setsockopt:	sock_no_setsockopt,
	getsockopt:	sock_no_getsockopt,
	sendmsg:	netlink_sendmsg,
	recvmsg:	netlink_recvmsg,
	mmap:		sock_no_mmap,
	sendpage:	sock_no_sendpage,
};

struct net_proto_family netlink_family_ops = {
	PF_NETLINK,
	netlink_create
};

static int __init netlink_proto_init(void)
{
	struct sk_buff *dummy_skb;

	if (sizeof(struct netlink_skb_parms) > sizeof(dummy_skb->cb)) {
		printk(KERN_CRIT "netlink_init: panic\n");
		return -1;
	}
	sock_register(&netlink_family_ops);
#ifdef CONFIG_PROC_FS
	create_proc_read_entry("net/netlink", 0, 0, netlink_read_proc, NULL);
#endif
	return 0;
}

static void __exit netlink_proto_exit(void)
{
       sock_unregister(PF_NETLINK);
       remove_proc_entry("net/netlink", NULL);
}

module_init(netlink_proto_init);
module_exit(netlink_proto_exit);