nfs_syscalls.c

早期freebsd实现

/*
 * Copyright (c) 1989, 1993
 *	The Regents of the University of California.  All rights reserved.
 *
 * This code is derived from software contributed to Berkeley by
 * Rick Macklem at The University of Guelph.
 *
 * Redistribution and use in source and binary forms, with or without
 * modification, are permitted provided that the following conditions
 * are met:
 * 1. Redistributions of source code must retain the above copyright
 *    notice, this list of conditions and the following disclaimer.
 * 2. Redistributions in binary form must reproduce the above copyright
 *    notice, this list of conditions and the following disclaimer in the
 *    documentation and/or other materials provided with the distribution.
 * 3. All advertising materials mentioning features or use of this software
 *    must display the following acknowledgement:
 *	This product includes software developed by the University of
 *	California, Berkeley and its contributors.
 * 4. Neither the name of the University nor the names of its contributors
 *    may be used to endorse or promote products derived from this software
 *    without specific prior written permission.
 *
 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
 * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
 * SUCH DAMAGE.
 *
 *	@(#)nfs_syscalls.c	8.3 (Berkeley) 1/4/94
 */

#include <sys/param.h>
#include <sys/systm.h>
#include <sys/kernel.h>
#include <sys/file.h>
#include <sys/stat.h>
#include <sys/vnode.h>
#include <sys/mount.h>
#include <sys/proc.h>
#include <sys/uio.h>
#include <sys/malloc.h>
#include <sys/buf.h>
#include <sys/mbuf.h>
#include <sys/socket.h>
#include <sys/socketvar.h>
#include <sys/domain.h>
#include <sys/protosw.h>
#include <sys/namei.h>
#include <sys/syslog.h>

#include <netinet/in.h>
#include <netinet/tcp.h>
#ifdef ISO
#include <netiso/iso.h>
#endif
#include <nfs/rpcv2.h>
#include <nfs/nfsv2.h>
#include <nfs/nfs.h>
#include <nfs/nfsrvcache.h>
#include <nfs/nfsmount.h>
#include <nfs/nfsnode.h>
#include <nfs/nqnfs.h>
#include <nfs/nfsrtt.h>

/* Global defs. */
extern u_long nfs_prog, nfs_vers;
extern int (*nfsrv_procs[NFS_NPROCS])();
extern struct proc *nfs_iodwant[NFS_MAXASYNCDAEMON];
extern int nfs_numasync;
extern time_t nqnfsstarttime;
extern struct nfsrv_req nsrvq_head;
extern struct nfsd nfsd_head;
extern int nqsrv_writeslack;
extern int nfsrtton;
struct nfssvc_sock *nfs_udpsock, *nfs_cltpsock;
int nuidhash_max = NFS_MAXUIDHASH;
static int nfs_numnfsd = 0;
int nfsd_waiting = 0;
static int notstarted = 1;
static int modify_flag = 0;
static struct nfsdrt nfsdrt;
void nfsrv_cleancache(), nfsrv_rcv(), nfsrv_wakenfsd(), nfs_sndunlock();
static void nfsd_rt();
void nfsrv_slpderef(), nfsrv_init();

#define	TRUE	1
#define	FALSE	0

static int nfs_asyncdaemon[NFS_MAXASYNCDAEMON];
/*
 * NFS server system calls
 * getfh() lives here too, but maybe should move to kern/vfs_syscalls.c
 */

/*
 * Get file handle system call
 */
struct getfh_args {
	char	*fname;
	fhandle_t *fhp;
};
getfh(p, uap, retval)
	struct proc *p;
	register struct getfh_args *uap;
	int *retval;
{
	register struct vnode *vp;
	fhandle_t fh;
	int error;
	struct nameidata nd;

	/*
	 * Must be super user
	 */
	if (error = suser(p->p_ucred, &p->p_acflag))
		return (error);
	NDINIT(&nd, LOOKUP, FOLLOW | LOCKLEAF, UIO_USERSPACE, uap->fname, p);
	if (error = namei(&nd))
		return (error);
	vp = nd.ni_vp;
	bzero((caddr_t)&fh, sizeof(fh));
	fh.fh_fsid = vp->v_mount->mnt_stat.f_fsid;
	error = VFS_VPTOFH(vp, &fh.fh_fid);
	vput(vp);
	if (error)
		return (error);
	error = copyout((caddr_t)&fh, (caddr_t)uap->fhp, sizeof (fh));
	return (error);
}

static struct nfssvc_sock nfssvc_sockhead;

/*
 * Nfs server psuedo system call for the nfsd's
 * Based on the flag value it either:
 * - adds a socket to the selection list
 * - remains in the kernel as an nfsd
 * - remains in the kernel as an nfsiod
 */
struct nfssvc_args {
	int flag;
	caddr_t argp;
};
nfssvc(p, uap, retval)
	struct proc *p;
	register struct nfssvc_args *uap;
	int *retval;
{
	struct nameidata nd;
	struct file *fp;
	struct mbuf *nam;
	struct nfsd_args nfsdarg;
	struct nfsd_srvargs nfsd_srvargs, *nsd = &nfsd_srvargs;
	struct nfsd_cargs ncd;
	struct nfsd *nfsd;
	struct nfssvc_sock *slp;
	struct nfsuid *nuidp, **nuh;
	struct nfsmount *nmp;
	int error;

	/*
	 * Must be super user
	 */
	if (error = suser(p->p_ucred, &p->p_acflag))
		return (error);
	while (nfssvc_sockhead.ns_flag & SLP_INIT) {
		nfssvc_sockhead.ns_flag |= SLP_WANTINIT;
		(void) tsleep((caddr_t)&nfssvc_sockhead, PSOCK, "nfsd init", 0);
	}
	if (uap->flag & NFSSVC_BIOD)
		error = nfssvc_iod(p);
	else if (uap->flag & NFSSVC_MNTD) {
		if (error = copyin(uap->argp, (caddr_t)&ncd, sizeof (ncd)))
			return (error);
		NDINIT(&nd, LOOKUP, FOLLOW | LOCKLEAF, UIO_USERSPACE,
			ncd.ncd_dirp, p);
		if (error = namei(&nd))
			return (error);
		if ((nd.ni_vp->v_flag & VROOT) == 0)
			error = EINVAL;
		nmp = VFSTONFS(nd.ni_vp->v_mount);
		vput(nd.ni_vp);
		if (error)
			return (error);
		if ((nmp->nm_flag & NFSMNT_MNTD) &&
			(uap->flag & NFSSVC_GOTAUTH) == 0)
			return (0);
		nmp->nm_flag |= NFSMNT_MNTD;
		error = nqnfs_clientd(nmp, p->p_ucred, &ncd, uap->flag,
			uap->argp, p);
	} else if (uap->flag & NFSSVC_ADDSOCK) {
		if (error = copyin(uap->argp, (caddr_t)&nfsdarg,
		    sizeof(nfsdarg)))
			return (error);
		if (error = getsock(p->p_fd, nfsdarg.sock, &fp))
			return (error);
		/*
		 * Get the client address for connected sockets.
		 */
		if (nfsdarg.name == NULL || nfsdarg.namelen == 0)
			nam = (struct mbuf *)0;
		else if (error = sockargs(&nam, nfsdarg.name, nfsdarg.namelen,
			MT_SONAME))
			return (error);
		error = nfssvc_addsock(fp, nam);
	} else {
		if (error = copyin(uap->argp, (caddr_t)nsd, sizeof (*nsd)))
			return (error);
		if ((uap->flag & NFSSVC_AUTHIN) && (nfsd = nsd->nsd_nfsd) &&
			(nfsd->nd_slp->ns_flag & SLP_VALID)) {
			slp = nfsd->nd_slp;

			/*
			 * First check to see if another nfsd has already
			 * added this credential.
			 */
			nuidp = slp->ns_uidh[NUIDHASH(nsd->nsd_uid)];
			while (nuidp) {
				if (nuidp->nu_uid == nsd->nsd_uid)
					break;
				nuidp = nuidp->nu_hnext;
			}
			if (!nuidp) {
			    /*
			     * Nope, so we will.
			     */
			    if (slp->ns_numuids < nuidhash_max) {
				slp->ns_numuids++;
				nuidp = (struct nfsuid *)
				   malloc(sizeof (struct nfsuid), M_NFSUID,
					M_WAITOK);
			    } else
				nuidp = (struct nfsuid *)0;
			    if ((slp->ns_flag & SLP_VALID) == 0) {
				if (nuidp)
				    free((caddr_t)nuidp, M_NFSUID);
			    } else {
				if (nuidp == (struct nfsuid *)0) {
				    nuidp = slp->ns_lruprev;
				    remque(nuidp);
				    if (nuidp->nu_hprev)
					nuidp->nu_hprev->nu_hnext =
					    nuidp->nu_hnext;
				    if (nuidp->nu_hnext)
					nuidp->nu_hnext->nu_hprev =
					    nuidp->nu_hprev;
			        }
				nuidp->nu_cr = nsd->nsd_cr;
				if (nuidp->nu_cr.cr_ngroups > NGROUPS)
					nuidp->nu_cr.cr_ngroups = NGROUPS;
				nuidp->nu_cr.cr_ref = 1;
				nuidp->nu_uid = nsd->nsd_uid;
				insque(nuidp, (struct nfsuid *)slp);
				nuh = &slp->ns_uidh[NUIDHASH(nsd->nsd_uid)];
				if (nuidp->nu_hnext = *nuh)
				    nuidp->nu_hnext->nu_hprev = nuidp;
				nuidp->nu_hprev = (struct nfsuid *)0;
				*nuh = nuidp;
			    }
			}
		}
		if ((uap->flag & NFSSVC_AUTHINFAIL) && (nfsd = nsd->nsd_nfsd))
			nfsd->nd_flag |= NFSD_AUTHFAIL;
		error = nfssvc_nfsd(nsd, uap->argp, p);
	}
	if (error == EINTR || error == ERESTART)
		error = 0;
	return (error);
}

/*
 * Adds a socket to the list for servicing by nfsds.
 */
nfssvc_addsock(fp, mynam)
	struct file *fp;
	struct mbuf *mynam;
{
	register struct mbuf *m;
	register int siz;
	register struct nfssvc_sock *slp;
	register struct socket *so;
	struct nfssvc_sock *tslp;
	int error, s;

	so = (struct socket *)fp->f_data;
	tslp = (struct nfssvc_sock *)0;
	/*
	 * Add it to the list, as required.
	 */
	if (so->so_proto->pr_protocol == IPPROTO_UDP) {
		tslp = nfs_udpsock;
		if (tslp->ns_flag & SLP_VALID) {
			m_freem(mynam);
			return (EPERM);
		}
#ifdef ISO
	} else if (so->so_proto->pr_protocol == ISOPROTO_CLTP) {
		tslp = nfs_cltpsock;
		if (tslp->ns_flag & SLP_VALID) {
			m_freem(mynam);
			return (EPERM);
		}
#endif /* ISO */
	}
	if (so->so_type == SOCK_STREAM)
		siz = NFS_MAXPACKET + sizeof (u_long);
	else
		siz = NFS_MAXPACKET;
	if (error = soreserve(so, siz, siz)) {
		m_freem(mynam);
		return (error);
	}

	/*
	 * Set protocol specific options { for now TCP only } and
	 * reserve some space. For datagram sockets, this can get called
	 * repeatedly for the same socket, but that isn't harmful.
	 */
	if (so->so_type == SOCK_STREAM) {
		MGET(m, M_WAIT, MT_SOOPTS);
		*mtod(m, int *) = 1;
		m->m_len = sizeof(int);
		sosetopt(so, SOL_SOCKET, SO_KEEPALIVE, m);
	}
	if (so->so_proto->pr_domain->dom_family == AF_INET &&
	    so->so_proto->pr_protocol == IPPROTO_TCP) {
		MGET(m, M_WAIT, MT_SOOPTS);
		*mtod(m, int *) = 1;
		m->m_len = sizeof(int);
		sosetopt(so, IPPROTO_TCP, TCP_NODELAY, m);
	}
	so->so_rcv.sb_flags &= ~SB_NOINTR;
	so->so_rcv.sb_timeo = 0;
	so->so_snd.sb_flags &= ~SB_NOINTR;
	so->so_snd.sb_timeo = 0;
	if (tslp)
		slp = tslp;
	else {
		slp = (struct nfssvc_sock *)
			malloc(sizeof (struct nfssvc_sock), M_NFSSVC, M_WAITOK);
		bzero((caddr_t)slp, sizeof (struct nfssvc_sock));
		slp->ns_prev = nfssvc_sockhead.ns_prev;
		slp->ns_prev->ns_next = slp;
		slp->ns_next = &nfssvc_sockhead;
		nfssvc_sockhead.ns_prev = slp;
		slp->ns_lrunext = slp->ns_lruprev = (struct nfsuid *)slp;
	}
	slp->ns_so = so;
	slp->ns_nam = mynam;
	fp->f_count++;
	slp->ns_fp = fp;
	s = splnet();
	so->so_upcallarg = (caddr_t)slp;
	so->so_upcall = nfsrv_rcv;
	slp->ns_flag = (SLP_VALID | SLP_NEEDQ);
	nfsrv_wakenfsd(slp);
	splx(s);
	return (0);
}

/*
 * Called by nfssvc() for nfsds. Just loops around servicing rpc requests
 * until it is killed by a signal.
 */
nfssvc_nfsd(nsd, argp, p)
	struct nfsd_srvargs *nsd;
	caddr_t argp;
	struct proc *p;
{
	register struct mbuf *m, *nam2;
	register int siz;
	register struct nfssvc_sock *slp;
	register struct socket *so;
	register int *solockp;
	struct nfsd *nd = nsd->nsd_nfsd;
	struct mbuf *mreq, *nam;
	struct timeval starttime;
	struct nfsuid *uidp;
	int error, cacherep, s;
	int sotype;

	s = splnet();
	if (nd == (struct nfsd *)0) {
		nsd->nsd_nfsd = nd = (struct nfsd *)
			malloc(sizeof (struct nfsd), M_NFSD, M_WAITOK);
		bzero((caddr_t)nd, sizeof (struct nfsd));
		nd->nd_procp = p;
		nd->nd_cr.cr_ref = 1;
		insque(nd, &nfsd_head);
		nd->nd_nqlflag = NQL_NOVAL;
		nfs_numnfsd++;
	}
	/*
	 * Loop getting rpc requests until SIGKILL.
	 */
	for (;;) {
		if ((nd->nd_flag & NFSD_REQINPROG) == 0) {
			while (nd->nd_slp == (struct nfssvc_sock *)0 &&
				 (nfsd_head.nd_flag & NFSD_CHECKSLP) == 0) {
				nd->nd_flag |= NFSD_WAITING;
				nfsd_waiting++;
				error = tsleep((caddr_t)nd, PSOCK | PCATCH, "nfsd", 0);
				nfsd_waiting--;
				if (error)
					goto done;
			}
			if (nd->nd_slp == (struct nfssvc_sock *)0 &&
				(nfsd_head.nd_flag & NFSD_CHECKSLP)) {
				slp = nfssvc_sockhead.ns_next;
				while (slp != &nfssvc_sockhead) {
				    if ((slp->ns_flag & (SLP_VALID | SLP_DOREC))
					== (SLP_VALID | SLP_DOREC)) {
					    slp->ns_flag &= ~SLP_DOREC;
					    slp->ns_sref++;
					    nd->nd_slp = slp;
					    break;
				    }
				    slp = slp->ns_next;
				}
				if (slp == &nfssvc_sockhead)
					nfsd_head.nd_flag &= ~NFSD_CHECKSLP;
			}
			if ((slp = nd->nd_slp) == (struct nfssvc_sock *)0)
				continue;
			if (slp->ns_flag & SLP_VALID) {
				if (slp->ns_flag & SLP_DISCONN)
					nfsrv_zapsock(slp);
				else if (slp->ns_flag & SLP_NEEDQ) {
					slp->ns_flag &= ~SLP_NEEDQ;
					(void) nfs_sndlock(&slp->ns_solock,
						(struct nfsreq *)0);