svcsock.c

Linux内核源代码 为压缩文件 是<<Linux内核>>一书中的源代码

/*
 * linux/net/sunrpc/svcsock.c
 *
 * These are the RPC server socket internals.
 *
 * The server scheduling algorithm does not always distribute the load
 * evenly when servicing a single client. May need to modify the
 * svc_sock_enqueue procedure...
 *
 * TCP support is largely untested and may be a little slow. The problem
 * is that we currently do two separate recvfrom's, one for the 4-byte
 * record length, and the second for the actual record. This could possibly
 * be improved by always reading a minimum size of around 100 bytes and
 * tucking any superfluous bytes away in a temporary store. Still, that
 * leaves write requests out in the rain. An alternative may be to peek at
 * the first skb in the queue, and if it matches the next TCP sequence
 * number, to extract the record marker. Yuck.
 *
 * Copyright (C) 1995, 1996 Olaf Kirch <[e-mail]>
 */

#include <linux/sched.h>
#include <linux/errno.h>
#include <linux/fcntl.h>
#include <linux/net.h>
#include <linux/in.h>
#include <linux/inet.h>
#include <linux/udp.h>
#include <linux/version.h>
#include <linux/unistd.h>
#include <linux/malloc.h>
#include <linux/netdevice.h>
#include <linux/skbuff.h>
#include <net/sock.h>
#include <net/checksum.h>
#include <net/ip.h>
#include <asm/uaccess.h>

#include <linux/sunrpc/types.h>
#include <linux/sunrpc/xdr.h>
#include <linux/sunrpc/svcsock.h>
#include <linux/sunrpc/stats.h>

/* SMP locking strategy:
 *
 * 	svc_sock->sk_lock and svc_serv->sv_lock protect their
 *	respective structures.
 *
 *	Antideadlock ordering is sk_lock --> sv_lock.
 */

#define RPCDBG_FACILITY	RPCDBG_SVCSOCK


static struct svc_sock *svc_setup_socket(struct svc_serv *, struct socket *,
					 int *errp, int pmap_reg);
static void		svc_udp_data_ready(struct sock *, int);
static int		svc_udp_recvfrom(struct svc_rqst *);
static int		svc_udp_sendto(struct svc_rqst *);


/*
 * Queue up an idle server thread.  Must have serv->sv_lock held.
 */
static inline void
svc_serv_enqueue(struct svc_serv *serv, struct svc_rqst *rqstp)
{
	rpc_append_list(&serv->sv_threads, rqstp);
}

/*
 * Dequeue an nfsd thread.  Must have serv->sv_lock held.
 */
static inline void
svc_serv_dequeue(struct svc_serv *serv, struct svc_rqst *rqstp)
{
	rpc_remove_list(&serv->sv_threads, rqstp);
}

/*
 * Release an skbuff after use
 */
static inline void
svc_release_skb(struct svc_rqst *rqstp)
{
	struct sk_buff *skb = rqstp->rq_skbuff;

	if (!skb)
		return;
	rqstp->rq_skbuff = NULL;

	dprintk("svc: service %p, releasing skb %p\n", rqstp, skb);
	skb_free_datagram(rqstp->rq_sock->sk_sk, skb);
}

/*
 * Queue up a socket with data pending. If there are idle nfsd
 * processes, wake 'em up.
 *
 * This must be called with svsk->sk_lock held.
 */
static void
svc_sock_enqueue(struct svc_sock *svsk)
{
	struct svc_serv	*serv = svsk->sk_server;
	struct svc_rqst	*rqstp;

	/* NOTE: Local BH is already disabled by our caller. */
	spin_lock(&serv->sv_lock);

	if (serv->sv_threads && serv->sv_sockets)
		printk(KERN_ERR
			"svc_sock_enqueue: threads and sockets both waiting??\n");

	if (svsk->sk_busy) {
		/* Don't enqueue socket while daemon is receiving */
		dprintk("svc: socket %p busy, not enqueued\n", svsk->sk_sk);
		goto out_unlock;
	}

	/* Mark socket as busy. It will remain in this state until the
	 * server has processed all pending data and put the socket back
	 * on the idle list.
	 */
	svsk->sk_busy = 1;

	if ((rqstp = serv->sv_threads) != NULL) {
		dprintk("svc: socket %p served by daemon %p\n",
			svsk->sk_sk, rqstp);
		svc_serv_dequeue(serv, rqstp);
		if (rqstp->rq_sock)
			printk(KERN_ERR 
				"svc_sock_enqueue: server %p, rq_sock=%p!\n",
				rqstp, rqstp->rq_sock);
		rqstp->rq_sock = svsk;
		svsk->sk_inuse++;
		wake_up(&rqstp->rq_wait);
	} else {
		dprintk("svc: socket %p put into queue\n", svsk->sk_sk);
		rpc_append_list(&serv->sv_sockets, svsk);
		svsk->sk_qued = 1;
	}

out_unlock:
	spin_unlock(&serv->sv_lock);
}

/*
 * Dequeue the first socket.  Must be called with the serv->sv_lock held.
 */
static inline struct svc_sock *
svc_sock_dequeue(struct svc_serv *serv)
{
	struct svc_sock	*svsk;

	if ((svsk = serv->sv_sockets) != NULL)
		rpc_remove_list(&serv->sv_sockets, svsk);

	if (svsk) {
		dprintk("svc: socket %p dequeued, inuse=%d\n",
			svsk->sk_sk, svsk->sk_inuse);
		svsk->sk_qued = 0;
	}

	return svsk;
}

/*
 * Having read count bytes from a socket, check whether it
 * needs to be re-enqueued.
 */
static inline void
svc_sock_received(struct svc_sock *svsk, int count)
{
	spin_lock_bh(&svsk->sk_lock);
	if ((svsk->sk_data -= count) < 0) {
		printk(KERN_NOTICE "svc: sk_data negative!\n");
		svsk->sk_data = 0;
	}
	svsk->sk_rqstp = NULL; /* XXX */
	svsk->sk_busy = 0;
	if (svsk->sk_conn || svsk->sk_data || svsk->sk_close) {
		dprintk("svc: socket %p re-enqueued after receive\n",
						svsk->sk_sk);
		svc_sock_enqueue(svsk);
	}
	spin_unlock_bh(&svsk->sk_lock);
}

/*
 * Dequeue a new connection.
 */
static inline void
svc_sock_accepted(struct svc_sock *svsk)
{
	spin_lock_bh(&svsk->sk_lock);
        svsk->sk_busy = 0;
        svsk->sk_conn--;
        if (svsk->sk_conn || svsk->sk_data || svsk->sk_close) {
                dprintk("svc: socket %p re-enqueued after accept\n",
						svsk->sk_sk);
                svc_sock_enqueue(svsk);
        }
	spin_unlock_bh(&svsk->sk_lock);
}

/*
 * Release a socket after use.
 */
static inline void
svc_sock_release(struct svc_rqst *rqstp)
{
	struct svc_sock	*svsk = rqstp->rq_sock;

	if (!svsk)
		return;
	svc_release_skb(rqstp);
	rqstp->rq_sock = NULL;
	if (!--(svsk->sk_inuse) && svsk->sk_dead) {
		dprintk("svc: releasing dead socket\n");
		sock_release(svsk->sk_sock);
		kfree(svsk);
	}
}

/*
 * External function to wake up a server waiting for data
 */
void
svc_wake_up(struct svc_serv *serv)
{
	struct svc_rqst	*rqstp;

	spin_lock_bh(&serv->sv_lock);
	if ((rqstp = serv->sv_threads) != NULL) {
		dprintk("svc: daemon %p woken up.\n", rqstp);
		/*
		svc_serv_dequeue(serv, rqstp);
		rqstp->rq_sock = NULL;
		 */
		wake_up(&rqstp->rq_wait);
	}
	spin_unlock_bh(&serv->sv_lock);
}

/*
 * Generic sendto routine
 */
static int
svc_sendto(struct svc_rqst *rqstp, struct iovec *iov, int nr)
{
	mm_segment_t	oldfs;
	struct svc_sock	*svsk = rqstp->rq_sock;
	struct socket	*sock = svsk->sk_sock;
	struct msghdr	msg;
	int		i, buflen, len;

	for (i = buflen = 0; i < nr; i++)
		buflen += iov[i].iov_len;

	msg.msg_name    = &rqstp->rq_addr;
	msg.msg_namelen = sizeof(rqstp->rq_addr);
	msg.msg_iov     = iov;
	msg.msg_iovlen  = nr;
	msg.msg_control = NULL;
	msg.msg_controllen = 0;

	msg.msg_flags	= MSG_DONTWAIT;

	oldfs = get_fs(); set_fs(KERNEL_DS);
	len = sock_sendmsg(sock, &msg, buflen);
	set_fs(oldfs);

	dprintk("svc: socket %p sendto([%p %Zu... ], %d, %d) = %d\n",
			rqstp->rq_sock, iov[0].iov_base, iov[0].iov_len, nr, buflen, len);

	return len;
}

/*
 * Check input queue length
 */
static int
svc_recv_available(struct svc_sock *svsk)
{
	mm_segment_t	oldfs;
	struct socket	*sock = svsk->sk_sock;
	int		avail, err;

	oldfs = get_fs(); set_fs(KERNEL_DS);
	err = sock->ops->ioctl(sock, TIOCINQ, (unsigned long) &avail);
	set_fs(oldfs);

	return (err >= 0)? avail : err;
}

/*
 * Generic recvfrom routine.
 */
static int
svc_recvfrom(struct svc_rqst *rqstp, struct iovec *iov, int nr, int buflen)
{
	mm_segment_t	oldfs;
	struct msghdr	msg;
	struct socket	*sock;
	int		len, alen;

	rqstp->rq_addrlen = sizeof(rqstp->rq_addr);
	sock = rqstp->rq_sock->sk_sock;

	msg.msg_name    = &rqstp->rq_addr;
	msg.msg_namelen = sizeof(rqstp->rq_addr);
	msg.msg_iov     = iov;
	msg.msg_iovlen  = nr;
	msg.msg_control = NULL;
	msg.msg_controllen = 0;

	msg.msg_flags	= MSG_DONTWAIT;

	oldfs = get_fs(); set_fs(KERNEL_DS);
	len = sock_recvmsg(sock, &msg, buflen, MSG_DONTWAIT);
	set_fs(oldfs);

	/* sock_recvmsg doesn't fill in the name/namelen, so we must..
	 * possibly we should cache this in the svc_sock structure
	 * at accept time. FIXME
	 */
	alen = sizeof(rqstp->rq_addr);
	sock->ops->getname(sock, (struct sockaddr *)&rqstp->rq_addr, &alen, 1);

	dprintk("svc: socket %p recvfrom(%p, %Zu) = %d\n",
		rqstp->rq_sock, iov[0].iov_base, iov[0].iov_len, len);

	return len;
}

/*
 * INET callback when data has been received on the socket.
 */
static void
svc_udp_data_ready(struct sock *sk, int count)
{
	struct svc_sock	*svsk = (struct svc_sock *)(sk->user_data);

	if (!svsk)
		goto out;
	dprintk("svc: socket %p(inet %p), count=%d, busy=%d\n",
		svsk, sk, count, svsk->sk_busy);
	spin_lock_bh(&svsk->sk_lock);
	svsk->sk_data = 1;
	svc_sock_enqueue(svsk);
	spin_unlock_bh(&svsk->sk_lock);
 out:
	if (sk->sleep && waitqueue_active(sk->sleep))
		wake_up_interruptible(sk->sleep);
}

/*
 * Receive a datagram from a UDP socket.
 */
static int
svc_udp_recvfrom(struct svc_rqst *rqstp)
{
	struct svc_sock	*svsk = rqstp->rq_sock;
	struct svc_serv	*serv = svsk->sk_server;
	struct sk_buff	*skb;
	u32		*data;
	int		err, len;

	svsk->sk_data = 0;
	while ((skb = skb_recv_datagram(svsk->sk_sk, 0, 1, &err)) == NULL) {
		svc_sock_received(svsk, 0);
		if (err == -EAGAIN)
			return err;
		/* possibly an icmp error */
		dprintk("svc: recvfrom returned error %d\n", -err);
	}

	if (skb->ip_summed != CHECKSUM_UNNECESSARY) {
		unsigned int csum = skb->csum;
		csum = csum_partial(skb->h.raw, skb->len, csum);
		if ((unsigned short)csum_fold(csum)) {
			skb_free_datagram(svsk->sk_sk, skb);
			svc_sock_received(svsk, 0);
			return 0;
		}
	}

	/* There may be more data */
	svsk->sk_data = 1;

	len  = skb->len - sizeof(struct udphdr);
	data = (u32 *) (skb->h.raw + sizeof(struct udphdr));

	rqstp->rq_skbuff      = skb;
	rqstp->rq_argbuf.base = data;
	rqstp->rq_argbuf.buf  = data;
	rqstp->rq_argbuf.len  = (len >> 2);
	/* rqstp->rq_resbuf      = rqstp->rq_defbuf; */
	rqstp->rq_prot        = IPPROTO_UDP;

	/* Get sender address */
	rqstp->rq_addr.sin_family = AF_INET;
	rqstp->rq_addr.sin_port = skb->h.uh->source;
	rqstp->rq_addr.sin_addr.s_addr = skb->nh.iph->saddr;

	if (serv->sv_stats)
		serv->sv_stats->netudpcnt++;

	/* One down, maybe more to go... */
	svsk->sk_sk->stamp = skb->stamp;
	svc_sock_received(svsk, 0);

	return len;
}

static int
svc_udp_sendto(struct svc_rqst *rqstp)
{
	struct svc_buf	*bufp = &rqstp->rq_resbuf;
	int		error;

	/* Set up the first element of the reply iovec.
	 * Any other iovecs that may be in use have been taken
	 * care of by the server implementation itself.
	 */
	/* bufp->base = bufp->area; */
	bufp->iov[0].iov_base = bufp->base;
	bufp->iov[0].iov_len  = bufp->len << 2;

	error = svc_sendto(rqstp, bufp->iov, bufp->nriov);
	if (error == -ECONNREFUSED)
		/* ICMP error on earlier request. */
		error = svc_sendto(rqstp, bufp->iov, bufp->nriov);
	else if (error == -EAGAIN)
		/* Ignore and wait for re-xmit */
		error = 0;

	return error;
}

static int
svc_udp_init(struct svc_sock *svsk)
{
	svsk->sk_sk->data_ready = svc_udp_data_ready;
	svsk->sk_recvfrom = svc_udp_recvfrom;
	svsk->sk_sendto = svc_udp_sendto;

	return 0;
}

/*
 * A state change on a listening socket means there's a connection
 * pending.
 */
static void
svc_tcp_state_change1(struct sock *sk)
{
	struct svc_sock	*svsk;

	dprintk("svc: socket %p TCP (listen) state change %d\n",
			sk, sk->state);

	if  (sk->state != TCP_ESTABLISHED) {
		/* Aborted connection, SYN_RECV or whatever... */
		goto out;
	}
	if (!(svsk = (struct svc_sock *) sk->user_data)) {
		printk("svc: socket %p: no user data\n", sk);
		goto out;
	}
	spin_lock_bh(&svsk->sk_lock);
	svsk->sk_conn++;
	svc_sock_enqueue(svsk);
	spin_unlock_bh(&svsk->sk_lock);
 out:
	if (sk->sleep && waitqueue_active(sk->sleep))
		wake_up_interruptible_all(sk->sleep);
}

/*
 * A state change on a connected socket means it's dying or dead.
 */
static void
svc_tcp_state_change2(struct sock *sk)
{
	struct svc_sock	*svsk;

	dprintk("svc: socket %p TCP (connected) state change %d (svsk %p)\n",
			sk, sk->state, sk->user_data);

	if (!(svsk = (struct svc_sock *) sk->user_data)) {
		printk("svc: socket %p: no user data\n", sk);
		goto out;
	}
	spin_lock_bh(&svsk->sk_lock);
	svsk->sk_close = 1;
	svc_sock_enqueue(svsk);
	spin_unlock_bh(&svsk->sk_lock);
 out:
	if (sk->sleep && waitqueue_active(sk->sleep))
		wake_up_interruptible_all(sk->sleep);
}

static void
svc_tcp_data_ready(struct sock *sk, int count)
{
	struct svc_sock *	svsk;

	dprintk("svc: socket %p TCP data ready (svsk %p)\n",
			sk, sk->user_data);
	if (!(svsk = (struct svc_sock *)(sk->user_data)))
		goto out;
	spin_lock_bh(&svsk->sk_lock);
	svsk->sk_data++;
	svc_sock_enqueue(svsk);
	spin_unlock_bh(&svsk->sk_lock);
 out:
	if (sk->sleep && waitqueue_active(sk->sleep))
		wake_up_interruptible(sk->sleep);
}

/*
 * Accept a TCP connection
 */
static void
svc_tcp_accept(struct svc_sock *svsk)
{
	struct sockaddr_in sin;
	struct svc_serv	*serv = svsk->sk_server;
	struct socket	*sock = svsk->sk_sock;
	struct socket	*newsock;