vfs_subr.c

来自「基于组件方式开发操作系统的OSKIT源代码」· C语言 代码 · 共 2,733 行 · 第 1/5 页

/*	$NetBSD: vfs_subr.c,v 1.142 2000/11/27 08:39:44 chs Exp $	*/

/*-
 * Copyright (c) 1997, 1998 The NetBSD Foundation, Inc.
 * All rights reserved.
 *
 * This code is derived from software contributed to The NetBSD Foundation
 * by Jason R. Thorpe of the Numerical Aerospace Simulation Facility,
 * NASA Ames Research Center.
 *
 * 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 NetBSD
 *	Foundation, Inc. and its contributors.
 * 4. Neither the name of The NetBSD Foundation 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 NETBSD FOUNDATION, INC. 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 FOUNDATION 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.
 */

/*
 * Copyright (c) 1989, 1993
 *	The Regents of the University of California.  All rights reserved.
 * (c) UNIX System Laboratories, Inc.
 * All or some portions of this file are derived from material licensed
 * to the University of California by American Telephone and Telegraph
 * Co. or Unix System Laboratories, Inc. and are reproduced herein with
 * the permission of UNIX System Laboratories, Inc.
 *
 * 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.
 *
 *	@(#)vfs_subr.c	8.13 (Berkeley) 4/18/94
 */

/*
 * External virtual filesystem routines
 */

#include "opt_ddb.h"
#include "opt_compat_netbsd.h"
#include "opt_compat_43.h"

#include <sys/param.h>
#include <sys/systm.h>
#include <sys/proc.h>
#include <sys/kernel.h>
#include <sys/mount.h>
#include <sys/time.h>
#include <sys/fcntl.h>
#include <sys/vnode.h>
#include <sys/stat.h>
#include <sys/namei.h>
#include <sys/ucred.h>
#include <sys/buf.h>
#include <sys/errno.h>
#include <sys/malloc.h>
#include <sys/domain.h>
#include <sys/mbuf.h>
#include <sys/syscallargs.h>
#include <sys/device.h>
#include <sys/dirent.h>

#include <miscfs/specfs/specdev.h>
#ifndef OSKIT
#include <miscfs/genfs/genfs.h>
#include <miscfs/syncfs/syncfs.h>
#endif

#include <uvm/uvm.h>
#include <uvm/uvm_ddb.h>

#include <sys/sysctl.h>

enum vtype iftovt_tab[16] = {
	VNON, VFIFO, VCHR, VNON, VDIR, VNON, VBLK, VNON,
	VREG, VNON, VLNK, VNON, VSOCK, VNON, VNON, VBAD,
};
int	vttoif_tab[9] = {
	0, S_IFREG, S_IFDIR, S_IFBLK, S_IFCHR, S_IFLNK,
	S_IFSOCK, S_IFIFO, S_IFMT,
};

int doforce = 1;		/* 1 => permit forcible unmounting */
int prtactive = 0;		/* 1 => print out reclaim of active vnodes */

extern int dovfsusermount;	/* 1 => permit any user to mount filesystems */

/*
 * Insq/Remq for the vnode usage lists.
 */
#define	bufinsvn(bp, dp)	LIST_INSERT_HEAD(dp, bp, b_vnbufs)
#define	bufremvn(bp) {							\
	LIST_REMOVE(bp, b_vnbufs);					\
	(bp)->b_vnbufs.le_next = NOLIST;				\
}  
/* TAILQ_HEAD(freelst, vnode) vnode_free_list =	vnode free list (in vnode.h) */
struct freelst vnode_free_list = TAILQ_HEAD_INITIALIZER(vnode_free_list);
struct freelst vnode_hold_list = TAILQ_HEAD_INITIALIZER(vnode_hold_list);

struct mntlist mountlist =			/* mounted filesystem list */
    CIRCLEQ_HEAD_INITIALIZER(mountlist);
struct vfs_list_head vfs_list =			/* vfs list */
    LIST_HEAD_INITIALIZER(vfs_list);

struct nfs_public nfs_pub;			/* publicly exported FS */

#ifndef OSKIT
struct simplelock mountlist_slock = SIMPLELOCK_INITIALIZER;
static struct simplelock mntid_slock = SIMPLELOCK_INITIALIZER;
struct simplelock mntvnode_slock = SIMPLELOCK_INITIALIZER;
#endif
struct simplelock vnode_free_list_slock = SIMPLELOCK_INITIALIZER;
#ifndef OSKIT
struct simplelock spechash_slock = SIMPLELOCK_INITIALIZER;
#endif

/*
 * These define the root filesystem and device.
 */
struct mount *rootfs;
struct vnode *rootvnode;
struct device *root_device;			/* root device */

struct pool vnode_pool;				/* memory pool for vnodes */

/*
 * Local declarations.
 */
void insmntque __P((struct vnode *, struct mount *));
int getdevvp __P((dev_t, struct vnode **, enum vtype));
void vgoneall __P((struct vnode *));

#ifndef OSKIT
static int vfs_hang_addrlist __P((struct mount *, struct netexport *,
				  struct export_args *));
static int vfs_free_netcred __P((struct radix_node *, void *));
static void vfs_free_addrlist __P((struct netexport *));
#endif

#ifdef DEBUG
void printlockedvnodes __P((void));
#endif

#ifndef OSKIT
/*
 * Initialize the vnode management data structures.
 */
void
vntblinit()
{

	pool_init(&vnode_pool, sizeof(struct vnode), 0, 0, 0, "vnodepl",
	    0, pool_page_alloc_nointr, pool_page_free_nointr, M_VNODE);

	/*
	 * Initialize the filesystem syncer.
	 */
	vn_initialize_syncerd();
}

/*
 * Mark a mount point as busy. Used to synchronize access and to delay
 * unmounting. Interlock is not released on failure.
 */
int
vfs_busy(mp, flags, interlkp)
	struct mount *mp;
	int flags;
	struct simplelock *interlkp;
{
	int lkflags;

	while (mp->mnt_flag & MNT_UNMOUNT) {
		int gone;
		
		if (flags & LK_NOWAIT)
			return (ENOENT);
		if ((flags & LK_RECURSEFAIL) && mp->mnt_unmounter != NULL
		    && mp->mnt_unmounter == curproc)
			return (EDEADLK);
		if (interlkp)
			simple_unlock(interlkp);
		/*
		 * Since all busy locks are shared except the exclusive
		 * lock granted when unmounting, the only place that a
		 * wakeup needs to be done is at the release of the
		 * exclusive lock at the end of dounmount.
		 *
		 * XXX MP: add spinlock protecting mnt_wcnt here once you
		 * can atomically unlock-and-sleep.
		 */
		mp->mnt_wcnt++;
		tsleep((caddr_t)mp, PVFS, "vfs_busy", 0);
		mp->mnt_wcnt--;
		gone = mp->mnt_flag & MNT_GONE;
		
		if (mp->mnt_wcnt == 0)
			wakeup(&mp->mnt_wcnt);
		if (interlkp)
			simple_lock(interlkp);
		if (gone)
			return (ENOENT);
	}
	lkflags = LK_SHARED;
	if (interlkp)
		lkflags |= LK_INTERLOCK;
	if (lockmgr(&mp->mnt_lock, lkflags, interlkp))
		panic("vfs_busy: unexpected lock failure");
	return (0);
}

/*
 * Free a busy filesystem.
 */
void
vfs_unbusy(mp)
	struct mount *mp;
{

	lockmgr(&mp->mnt_lock, LK_RELEASE, NULL);
}

/*
 * Lookup a filesystem type, and if found allocate and initialize
 * a mount structure for it.
 *
 * Devname is usually updated by mount(8) after booting.
 */
int
vfs_rootmountalloc(fstypename, devname, mpp)
	char *fstypename;
	char *devname;
	struct mount **mpp;
{
	struct vfsops *vfsp = NULL;
	struct mount *mp;

	for (vfsp = LIST_FIRST(&vfs_list); vfsp != NULL;
	     vfsp = LIST_NEXT(vfsp, vfs_list))
		if (!strncmp(vfsp->vfs_name, fstypename, MFSNAMELEN))
			break;

	if (vfsp == NULL)
		return (ENODEV);
	mp = malloc((u_long)sizeof(struct mount), M_MOUNT, M_WAITOK);
	memset((char *)mp, 0, (u_long)sizeof(struct mount));
	lockinit(&mp->mnt_lock, PVFS, "vfslock", 0, 0);
	(void)vfs_busy(mp, LK_NOWAIT, 0);
	LIST_INIT(&mp->mnt_vnodelist);
	mp->mnt_op = vfsp;
	mp->mnt_flag = MNT_RDONLY;
	mp->mnt_vnodecovered = NULLVP;
	vfsp->vfs_refcount++;
	strncpy(mp->mnt_stat.f_fstypename, vfsp->vfs_name, MFSNAMELEN);
	mp->mnt_stat.f_mntonname[0] = '/';
	(void) copystr(devname, mp->mnt_stat.f_mntfromname, MNAMELEN - 1, 0);
	*mpp = mp;
	return (0);
}

/*
 * Lookup a mount point by filesystem identifier.
 */
struct mount *
vfs_getvfs(fsid)
	fsid_t *fsid;
{
	struct mount *mp;

	simple_lock(&mountlist_slock);
	for (mp = mountlist.cqh_first; mp != (void *)&mountlist;
	     mp = mp->mnt_list.cqe_next) {
		if (mp->mnt_stat.f_fsid.val[0] == fsid->val[0] &&
		    mp->mnt_stat.f_fsid.val[1] == fsid->val[1]) {
			simple_unlock(&mountlist_slock);
			return (mp);
		}
	}
	simple_unlock(&mountlist_slock);
	return ((struct mount *)0);
}

/*
 * Get a new unique fsid
 */
void
vfs_getnewfsid(mp)
	struct mount *mp;
{
	static u_short xxxfs_mntid;
	fsid_t tfsid;
	int mtype;

	simple_lock(&mntid_slock);
	mtype = makefstype(mp->mnt_op->vfs_name);
	mp->mnt_stat.f_fsid.val[0] = makedev(nblkdev + mtype, 0);
	mp->mnt_stat.f_fsid.val[1] = mtype;
	if (xxxfs_mntid == 0)
		++xxxfs_mntid;
	tfsid.val[0] = makedev((nblkdev + mtype) & 0xff, xxxfs_mntid);
	tfsid.val[1] = mtype;
	if (mountlist.cqh_first != (void *)&mountlist) {
		while (vfs_getvfs(&tfsid)) {
			tfsid.val[0]++;
			xxxfs_mntid++;
		}
	}
	mp->mnt_stat.f_fsid.val[0] = tfsid.val[0];
	simple_unlock(&mntid_slock);
}

/*
 * Make a 'unique' number from a mount type name.
 */
long
makefstype(type)
	const char *type;
{
	long rv;

	for (rv = 0; *type; type++) {
		rv <<= 2;
		rv ^= *type;
	}
	return rv;
}


/*
 * Set vnode attributes to VNOVAL
 */
void
vattr_null(vap)
	struct vattr *vap;
{

	vap->va_type = VNON;

	/*
	 * Assign individually so that it is safe even if size and
	 * sign of each member are varied.
	 */
	vap->va_mode = VNOVAL;
	vap->va_nlink = VNOVAL;
	vap->va_uid = VNOVAL;
	vap->va_gid = VNOVAL;
	vap->va_fsid = VNOVAL;
	vap->va_fileid = VNOVAL;
	vap->va_size = VNOVAL;
	vap->va_blocksize = VNOVAL;
	vap->va_atime.tv_sec =
	    vap->va_mtime.tv_sec =
	    vap->va_ctime.tv_sec = VNOVAL;
	vap->va_atime.tv_nsec =
	    vap->va_mtime.tv_nsec =
	    vap->va_ctime.tv_nsec = VNOVAL;
	vap->va_gen = VNOVAL;
	vap->va_flags = VNOVAL;
	vap->va_rdev = VNOVAL;
	vap->va_bytes = VNOVAL;
	vap->va_vaflags = 0;
}

/*
 * Routines having to do with the management of the vnode table.
 */
extern int (**dead_vnodeop_p) __P((void *));
long numvnodes;

/*
 * Return the next vnode from the free list.
 */
int
getnewvnode(tag, mp, vops, vpp)
	enum vtagtype tag;
	struct mount *mp;
	int (**vops) __P((void *));
	struct vnode **vpp;
{
	extern struct uvm_pagerops uvm_vnodeops;
	struct uvm_object *uobj;
	struct proc *p = curproc;	/* XXX */
	struct freelst *listhd;
	static int toggle;
	struct vnode *vp;
	int error = 0;
#ifdef DIAGNOSTIC
	int s;
#endif
	if (mp) {
		/*
		 * Mark filesystem busy while we're creating a vnode.
		 * If unmount is in progress, this will wait; if the
		 * unmount succeeds (only if umount -f), this will
		 * return an error.  If the unmount fails, we'll keep
		 * going afterwards.
		 * (This puts the per-mount vnode list logically under
		 * the protection of the vfs_busy lock).
		 */
		error = vfs_busy(mp, LK_RECURSEFAIL, 0);
		if (error && error != EDEADLK)
			return error;
	}

	/*
	 * We must choose whether to allocate a new vnode or recycle an
	 * existing one. The criterion for allocating a new one is that
	 * the total number of vnodes is less than the number desired or
	 * there are no vnodes on either free list. Generally we only
	 * want to recycle vnodes that have no buffers associated with
	 * them, so we look first on the vnode_free_list. If it is empty,
	 * we next consider vnodes with referencing buffers on the
	 * vnode_hold_list. The toggle ensures that half the time we
	 * will use a buffer from the vnode_hold_list, and half the time
	 * we will allocate a new one unless the list has grown to twice
	 * the desired size. We are reticent to recycle vnodes from the
	 * vnode_hold_list because we will lose the identity of all its
	 * referencing buffers.
	 */

	toggle ^= 1;
	if (numvnodes > 2 * desiredvnodes)
		toggle = 0;

	simple_lock(&vnode_free_list_slock);
	if (numvnodes < desiredvnodes ||
	    (TAILQ_FIRST(listhd = &vnode_free_list) == NULL &&
	    (TAILQ_FIRST(listhd = &vnode_hold_list) == NULL || toggle))) {
		simple_unlock(&vnode_free_list_slock);
		vp = pool_get(&vnode_pool, PR_WAITOK);
		memset(vp, 0, sizeof(*vp));
		simple_lock_init(&vp->v_interlock);
		numvnodes++;
	} else {
		for (vp = TAILQ_FIRST(listhd); vp != NULLVP;
		    vp = TAILQ_NEXT(vp, v_freelist)) {
			if (simple_lock_try(&vp->v_interlock)) {
				if ((vp->v_flag & VLAYER) == 0) {
					break;
				}
				if (VOP_ISLOCKED(vp) == 0)
					break;
				else
					simple_unlock(&vp->v_interlock);
			}
		}
		/*
		 * Unless this is a bad time of the month, at most
		 * the first NCPUS items on the free list are
		 * locked, so this is close enough to being empty.
		 */
		if (vp == NULLVP) {
			simple_unlock(&vnode_free_list_slock);
			if (mp && error != EDEADLK)
				vfs_unbusy(mp);
			tablefull("vnode", "increase kern.maxvnodes or NVNODE");
			*vpp = 0;
			return (ENFILE);
		}
		if (vp->v_usecount)
			panic("free vnode isn't, vp %p", vp);
		TAILQ_REMOVE(listhd, vp, v_freelist);
		/* see comment on why 0xdeadb is set at end of vgone (below) */
		vp->v_freelist.tqe_prev = (struct vnode **)0xdeadb;
		simple_unlock(&vnode_free_list_slock);
		vp->v_lease = NULL;
		if (vp->v_type != VBAD)
			vgonel(vp, p);
		else
			simple_unlock(&vp->v_interlock);
#ifdef DIAGNOSTIC
		if (vp->v_data)
			panic("cleaned vnode isn't, vp %p", vp);
		s = splbio();
		if (vp->v_numoutput)
			panic("clean vnode has pending I/O's, vp %p", vp);
		splx(s);
#endif
		vp->v_flag = 0;
		vp->v_lastr = 0;
		vp->v_ralen = 0;
		vp->v_maxra = 0;
		vp->v_lastw = 0;
		vp->v_lasta = 0;
		vp->v_cstart = 0;
		vp->v_clen = 0;
		vp->v_socket = 0;
	}
	vp->v_type = VNON;
	vp->v_vnlock = &vp->v_lock;
	lockinit(vp->v_vnlock, PVFS, "vnlock", 0, 0);
	lockinit(&vp->v_glock, PVFS, "glock", 0, 0);
	cache_purge(vp);
	vp->v_tag = tag;
	vp->v_op = vops;
	insmntque(vp, mp);
	*vpp = vp;
	vp->v_usecount = 1;
	vp->v_data = 0;
	simple_lock_init(&vp->v_uvm.u_obj.vmobjlock);

	/*