HAMMER 60I/Many: Mirroring

[dragonfly.git] / sys / kern / vfs_default.c

/*
 * Copyright (c) 1989, 1993
 *      The Regents of the University of California.  All rights reserved.
 *
 * This code is derived from software contributed
 * to Berkeley by John Heidemann of the UCLA Ficus project.
 *
 * The statvfs->statfs conversion code was contributed to the DragonFly
 * Project by Joerg Sonnenberger <joerg@bec.de>.
 *
 * 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.
 *
 * Source: * @(#)i405_init.c 2.10 92/04/27 UCLA Ficus project
 * $FreeBSD: src/sys/kern/vfs_default.c,v 1.28.2.7 2003/01/10 18:23:26 bde Exp $
 * $DragonFly: src/sys/kern/vfs_default.c,v 1.53 2008/06/01 19:27:35 dillon Exp $
 */

#include <sys/param.h>
#include <sys/systm.h>
#include <sys/buf.h>
#include <sys/conf.h>
#include <sys/fcntl.h>
#include <sys/file.h>
#include <sys/kernel.h>
#include <sys/lock.h>
#include <sys/malloc.h>
#include <sys/mount.h>
#include <sys/unistd.h>
#include <sys/vnode.h>
#include <sys/namei.h>
#include <sys/nlookup.h>
#include <sys/poll.h>
#include <sys/mountctl.h>

#include <machine/limits.h>

#include <vm/vm.h>
#include <vm/vm_object.h>
#include <vm/vm_page.h>
#include <vm/vm_pager.h>
#include <vm/vnode_pager.h>

static int      vop_nolookup (struct vop_old_lookup_args *);
static int      vop_nostrategy (struct vop_strategy_args *);

/*
 * This vnode table stores what we want to do if the filesystem doesn't
 * implement a particular VOP.
 *
 * If there is no specific entry here, we will return EOPNOTSUPP.
 */
struct vop_ops default_vnode_vops = {
        .vop_default            = vop_eopnotsupp,
        .vop_advlock            = (void *)vop_einval,
        .vop_fsync              = (void *)vop_null,
        .vop_ioctl              = (void *)vop_enotty,
        .vop_mmap               = (void *)vop_einval,
        .vop_old_lookup         = vop_nolookup,
        .vop_open               = vop_stdopen,
        .vop_close              = vop_stdclose,
        .vop_pathconf           = (void *)vop_einval,
        .vop_poll               = vop_nopoll,
        .vop_readlink           = (void *)vop_einval,
        .vop_reallocblks        = (void *)vop_eopnotsupp,
        .vop_revoke             = vop_stdrevoke,
        .vop_strategy           = vop_nostrategy,
        .vop_getacl             = (void *)vop_eopnotsupp,
        .vop_setacl             = (void *)vop_eopnotsupp,
        .vop_aclcheck           = (void *)vop_eopnotsupp,
        .vop_getextattr         = (void *)vop_eopnotsupp,
        .vop_setextattr         = (void *)vop_eopnotsupp,
        .vop_nresolve           = vop_compat_nresolve,
        .vop_nlookupdotdot      = vop_compat_nlookupdotdot,
        .vop_ncreate            = vop_compat_ncreate,
        .vop_nmkdir             = vop_compat_nmkdir,
        .vop_nmknod             = vop_compat_nmknod,
        .vop_nlink              = vop_compat_nlink,
        .vop_nsymlink           = vop_compat_nsymlink,
        .vop_nwhiteout          = vop_compat_nwhiteout,
        .vop_nremove            = vop_compat_nremove,
        .vop_nrmdir             = vop_compat_nrmdir,
        .vop_nrename            = vop_compat_nrename,
        .vop_mountctl           = journal_mountctl
};

VNODEOP_SET(default_vnode_vops);

int
vop_eopnotsupp(struct vop_generic_args *ap)
{
        return (EOPNOTSUPP);
}

int
vop_ebadf(struct vop_generic_args *ap)
{
        return (EBADF);
}

int
vop_enotty(struct vop_generic_args *ap)
{
        return (ENOTTY);
}

int
vop_einval(struct vop_generic_args *ap)
{
        return (EINVAL);
}

int
vop_null(struct vop_generic_args *ap)
{
        return (0);
}

int
vop_defaultop(struct vop_generic_args *ap)
{
        return (VOCALL(&default_vnode_vops, ap));
}

int
vop_panic(struct vop_generic_args *ap)
{
        panic("filesystem goof: vop_panic[%s]", ap->a_desc->sd_name);
}

/*
 * vop_compat_resolve { struct nchandle *a_nch, struct vnode *dvp }
 * XXX STOPGAP FUNCTION
 *
 * XXX OLD API ROUTINE!  WHEN ALL VFSs HAVE BEEN CLEANED UP THIS PROCEDURE
 * WILL BE REMOVED.  This procedure exists for all VFSs which have not
 * yet implemented VOP_NRESOLVE().  It converts VOP_NRESOLVE() into a 
 * vop_old_lookup() and does appropriate translations.
 *
 * Resolve a ncp for VFSs which do not support the VOP.  Eventually all
 * VFSs will support this VOP and this routine can be removed, since
 * VOP_NRESOLVE() is far less complex then the older LOOKUP/CACHEDLOOKUP
 * API.
 *
 * A locked ncp is passed in to be resolved.  The NCP is resolved by
 * figuring out the vnode (if any) and calling cache_setvp() to attach the
 * vnode to the entry.  If the entry represents a non-existant node then
 * cache_setvp() is called with a NULL vnode to resolve the entry into a
 * negative cache entry.  No vnode locks are retained and the
 * ncp is left locked on return.
 *
 * The ncp will NEVER represent "", "." or "..", or contain any slashes.
 *
 * There is a potential directory and vnode interlock.   The lock order
 * requirement is: namecache, governing directory, resolved vnode.
 */
int
vop_compat_nresolve(struct vop_nresolve_args *ap)
{
        int error;
        struct vnode *dvp;
        struct vnode *vp;
        struct nchandle *nch;
        struct namecache *ncp;
        struct componentname cnp;

        nch = ap->a_nch;        /* locked namecache node */
        ncp = nch->ncp;
        dvp = ap->a_dvp;

        /*
         * UFS currently stores all sorts of side effects, including a loop
         * variable, in the directory inode.  That needs to be fixed and the
         * other VFS's audited before we can switch to LK_SHARED.
         */
        if ((error = vget(dvp, LK_EXCLUSIVE)) != 0) {
                kprintf("[diagnostic] vop_compat_resolve: EAGAIN on ncp %p %s\n",
                        ncp, ncp->nc_name);
                return(EAGAIN);
        }

        bzero(&cnp, sizeof(cnp));
        cnp.cn_nameiop = NAMEI_LOOKUP;
        cnp.cn_flags = 0;
        cnp.cn_nameptr = ncp->nc_name;
        cnp.cn_namelen = ncp->nc_nlen;
        cnp.cn_cred = ap->a_cred;
        cnp.cn_td = curthread; /* XXX */

        /*
         * vop_old_lookup() always returns vp locked.  dvp may or may not be
         * left locked depending on CNP_PDIRUNLOCK.
         */
        error = vop_old_lookup(ap->a_head.a_ops, dvp, &vp, &cnp);
        if (error == 0)
                vn_unlock(vp);
        if ((cnp.cn_flags & CNP_PDIRUNLOCK) == 0)
                vn_unlock(dvp);
        if ((ncp->nc_flag & NCF_UNRESOLVED) == 0) {
                /* was resolved by another process while we were unlocked */
                if (error == 0)
                        vrele(vp);
        } else if (error == 0) {
                KKASSERT(vp != NULL);
                cache_setvp(nch, vp);
                vrele(vp);
        } else if (error == ENOENT) {
                KKASSERT(vp == NULL);
                if (cnp.cn_flags & CNP_ISWHITEOUT)
                        ncp->nc_flag |= NCF_WHITEOUT;
                cache_setvp(nch, NULL);
        }
        vrele(dvp);
        return (error);
}

/*
 * vop_compat_nlookupdotdot { struct vnode *a_dvp,
 *                      struct vnode **a_vpp,
 *                      struct ucred *a_cred }
 *
 * Lookup the vnode representing the parent directory of the specified
 * directory vnode.  a_dvp should not be locked.  If no error occurs *a_vpp
 * will contained the parent vnode, locked and refd, else *a_vpp will be NULL.
 *
 * This function is designed to aid NFS server-side operations and is
 * used by cache_fromdvp() to create a consistent, connected namecache
 * topology.
 *
 * As part of the NEW API work, VFSs will first split their CNP_ISDOTDOT
 * code out from their *_lookup() and create *_nlookupdotdot().  Then as time
 * permits VFSs will implement the remaining *_n*() calls and finally get
 * rid of their *_lookup() call.
 */
int
vop_compat_nlookupdotdot(struct vop_nlookupdotdot_args *ap)
{
        struct componentname cnp;
        int error;

        /*
         * UFS currently stores all sorts of side effects, including a loop
         * variable, in the directory inode.  That needs to be fixed and the
         * other VFS's audited before we can switch to LK_SHARED.
         */
        *ap->a_vpp = NULL;
        if ((error = vget(ap->a_dvp, LK_EXCLUSIVE)) != 0)
                return (error);
        if (ap->a_dvp->v_type != VDIR) {
                vput(ap->a_dvp);
                return (ENOTDIR);
        }

        bzero(&cnp, sizeof(cnp));
        cnp.cn_nameiop = NAMEI_LOOKUP;
        cnp.cn_flags = CNP_ISDOTDOT;
        cnp.cn_nameptr = "..";
        cnp.cn_namelen = 2;
        cnp.cn_cred = ap->a_cred;
        cnp.cn_td = curthread; /* XXX */

        /*
         * vop_old_lookup() always returns vp locked.  dvp may or may not be
         * left locked depending on CNP_PDIRUNLOCK.
         */
        error = vop_old_lookup(ap->a_head.a_ops, ap->a_dvp, ap->a_vpp, &cnp);
        if (error == 0)
                vn_unlock(*ap->a_vpp);
        if (cnp.cn_flags & CNP_PDIRUNLOCK)
                vrele(ap->a_dvp);
        else
                vput(ap->a_dvp);
        return (error);
}

/*
 * vop_compat_ncreate { struct nchandle *a_nch,         XXX STOPGAP FUNCTION
 *                      struct vnode *a_dvp,
 *                      struct vnode **a_vpp,
 *                      struct ucred *a_cred,
 *                      struct vattr *a_vap }
 *
 * Create a file as specified by a_vap.  Compatibility requires us to issue
 * the appropriate VOP_OLD_LOOKUP before we issue VOP_OLD_CREATE in order
 * to setup the directory inode's i_offset and i_count (e.g. in UFS).
 */
int
vop_compat_ncreate(struct vop_ncreate_args *ap)
{
        struct thread *td = curthread;
        struct componentname cnp;
        struct nchandle *nch;
        struct namecache *ncp;
        struct vnode *dvp;
        int error;

        /*
         * Sanity checks, get a locked directory vnode.
         */
        nch = ap->a_nch;                /* locked namecache node */
        dvp = ap->a_dvp;
        ncp = nch->ncp;

        if ((error = vget(dvp, LK_EXCLUSIVE)) != 0) {
                kprintf("[diagnostic] vop_compat_resolve: EAGAIN on ncp %p %s\n",
                        ncp, ncp->nc_name);
                return(EAGAIN);
        }

        /*
         * Setup the cnp for a traditional vop_old_lookup() call.  The lookup
         * caches all information required to create the entry in the
         * directory inode.  We expect a return code of EJUSTRETURN for
         * the CREATE case.  The cnp must simulated a saved-name situation.
         */
        bzero(&cnp, sizeof(cnp));
        cnp.cn_nameiop = NAMEI_CREATE;
        cnp.cn_flags = CNP_LOCKPARENT;
        cnp.cn_nameptr = ncp->nc_name;
        cnp.cn_namelen = ncp->nc_nlen;
        cnp.cn_cred = ap->a_cred;
        cnp.cn_td = td;
        *ap->a_vpp = NULL;

        error = vop_old_lookup(ap->a_head.a_ops, dvp, ap->a_vpp, &cnp);

        /*
         * EJUSTRETURN should be returned for this case, which means that
         * the VFS has setup the directory inode for the create.  The dvp we
         * passed in is expected to remain in a locked state.
         *
         * If the VOP_OLD_CREATE is successful we are responsible for updating
         * the cache state of the locked ncp that was passed to us.
         */
        if (error == EJUSTRETURN) {
                KKASSERT((cnp.cn_flags & CNP_PDIRUNLOCK) == 0);
                error = VOP_OLD_CREATE(dvp, ap->a_vpp, &cnp, ap->a_vap);
                if (error == 0) {
                        cache_setunresolved(nch);
                        cache_setvp(nch, *ap->a_vpp);
                }
        } else {
                if (error == 0) {
                        vput(*ap->a_vpp);
                        *ap->a_vpp = NULL;
                        error = EEXIST;
                }
                KKASSERT(*ap->a_vpp == NULL);
        }
        if ((cnp.cn_flags & CNP_PDIRUNLOCK) == 0)
                vn_unlock(dvp);
        vrele(dvp);
        return (error);
}

/*
 * vop_compat_nmkdir { struct nchandle *a_nch,  XXX STOPGAP FUNCTION
 *                      struct vnode *a_dvp,
 *                      struct vnode **a_vpp,
 *                      struct ucred *a_cred,
 *                      struct vattr *a_vap }
 *
 * Create a directory as specified by a_vap.  Compatibility requires us to
 * issue the appropriate VOP_OLD_LOOKUP before we issue VOP_OLD_MKDIR in
 * order to setup the directory inode's i_offset and i_count (e.g. in UFS).
 */
int
vop_compat_nmkdir(struct vop_nmkdir_args *ap)
{
        struct thread *td = curthread;
        struct componentname cnp;
        struct nchandle *nch;
        struct namecache *ncp;
        struct vnode *dvp;
        int error;

        /*
         * Sanity checks, get a locked directory vnode.
         */
        nch = ap->a_nch;                /* locked namecache node */
        ncp = nch->ncp;
        dvp = ap->a_dvp;
        if ((error = vget(dvp, LK_EXCLUSIVE)) != 0) {
                kprintf("[diagnostic] vop_compat_resolve: EAGAIN on ncp %p %s\n",
                        ncp, ncp->nc_name);
                return(EAGAIN);
        }

        /*
         * Setup the cnp for a traditional vop_old_lookup() call.  The lookup
         * caches all information required to create the entry in the
         * directory inode.  We expect a return code of EJUSTRETURN for
         * the CREATE case.  The cnp must simulated a saved-name situation.
         */
        bzero(&cnp, sizeof(cnp));
        cnp.cn_nameiop = NAMEI_CREATE;
        cnp.cn_flags = CNP_LOCKPARENT;
        cnp.cn_nameptr = ncp->nc_name;
        cnp.cn_namelen = ncp->nc_nlen;
        cnp.cn_cred = ap->a_cred;
        cnp.cn_td = td;
        *ap->a_vpp = NULL;

        error = vop_old_lookup(ap->a_head.a_ops, dvp, ap->a_vpp, &cnp);

        /*
         * EJUSTRETURN should be returned for this case, which means that
         * the VFS has setup the directory inode for the create.  The dvp we
         * passed in is expected to remain in a locked state.
         *
         * If the VOP_OLD_MKDIR is successful we are responsible for updating
         * the cache state of the locked ncp that was passed to us.
         */
        if (error == EJUSTRETURN) {
                KKASSERT((cnp.cn_flags & CNP_PDIRUNLOCK) == 0);
                error = VOP_OLD_MKDIR(dvp, ap->a_vpp, &cnp, ap->a_vap);
                if (error == 0) {
                        cache_setunresolved(nch);
                        cache_setvp(nch, *ap->a_vpp);
                }
        } else {
                if (error == 0) {
                        vput(*ap->a_vpp);
                        *ap->a_vpp = NULL;
                        error = EEXIST;
                }
                KKASSERT(*ap->a_vpp == NULL);
        }
        if ((cnp.cn_flags & CNP_PDIRUNLOCK) == 0)
                vn_unlock(dvp);
        vrele(dvp);
        return (error);
}

/*
 * vop_compat_nmknod { struct nchandle *a_nch,  XXX STOPGAP FUNCTION
 *                      struct vnode *a_dvp,
 *                      struct vnode **a_vpp,
 *                      struct ucred *a_cred,
 *                      struct vattr *a_vap }
 *
 * Create a device or fifo node as specified by a_vap.  Compatibility requires
 * us to issue the appropriate VOP_OLD_LOOKUP before we issue VOP_OLD_MKNOD
 * in order to setup the directory inode's i_offset and i_count (e.g. in UFS).
 */
int
vop_compat_nmknod(struct vop_nmknod_args *ap)
{
        struct thread *td = curthread;
        struct componentname cnp;
        struct nchandle *nch;
        struct namecache *ncp;
        struct vnode *dvp;
        int error;

        /*
         * Sanity checks, get a locked directory vnode.
         */
        nch = ap->a_nch;                /* locked namecache node */
        ncp = nch->ncp;
        dvp = ap->a_dvp;

        if ((error = vget(dvp, LK_EXCLUSIVE)) != 0) {
                kprintf("[diagnostic] vop_compat_resolve: EAGAIN on ncp %p %s\n",
                        ncp, ncp->nc_name);
                return(EAGAIN);
        }

        /*
         * Setup the cnp for a traditional vop_old_lookup() call.  The lookup
         * caches all information required to create the entry in the
         * directory inode.  We expect a return code of EJUSTRETURN for
         * the CREATE case.  The cnp must simulated a saved-name situation.
         */
        bzero(&cnp, sizeof(cnp));
        cnp.cn_nameiop = NAMEI_CREATE;
        cnp.cn_flags = CNP_LOCKPARENT;
        cnp.cn_nameptr = ncp->nc_name;
        cnp.cn_namelen = ncp->nc_nlen;
        cnp.cn_cred = ap->a_cred;
        cnp.cn_td = td;
        *ap->a_vpp = NULL;

        error = vop_old_lookup(ap->a_head.a_ops, dvp, ap->a_vpp, &cnp);

        /*
         * EJUSTRETURN should be returned for this case, which means that
         * the VFS has setup the directory inode for the create.  The dvp we
         * passed in is expected to remain in a locked state.
         *
         * If the VOP_OLD_MKNOD is successful we are responsible for updating
         * the cache state of the locked ncp that was passed to us.
         */
        if (error == EJUSTRETURN) {
                KKASSERT((cnp.cn_flags & CNP_PDIRUNLOCK) == 0);
                error = VOP_OLD_MKNOD(dvp, ap->a_vpp, &cnp, ap->a_vap);
                if (error == 0) {
                        cache_setunresolved(nch);
                        cache_setvp(nch, *ap->a_vpp);
                }
        } else {
                if (error == 0) {
                        vput(*ap->a_vpp);
                        *ap->a_vpp = NULL;
                        error = EEXIST;
                }
                KKASSERT(*ap->a_vpp == NULL);
        }
        if ((cnp.cn_flags & CNP_PDIRUNLOCK) == 0)
                vn_unlock(dvp);
        vrele(dvp);
        return (error);
}

/*
 * vop_compat_nlink { struct nchandle *a_nch,   XXX STOPGAP FUNCTION
 *                      struct vnode *a_dvp,
 *                      struct vnode *a_vp,
 *                      struct ucred *a_cred }
 *
 * The passed vp is locked and represents the source.  The passed ncp is
 * locked and represents the target to create.
 */
int
vop_compat_nlink(struct vop_nlink_args *ap)
{
        struct thread *td = curthread;
        struct componentname cnp;
        struct nchandle *nch;
        struct namecache *ncp;
        struct vnode *dvp;
        struct vnode *tvp;
        int error;

        /*
         * Sanity checks, get a locked directory vnode.
         */
        nch = ap->a_nch;                /* locked namecache node */
        ncp = nch->ncp;
        dvp = ap->a_dvp;

        if ((error = vget(dvp, LK_EXCLUSIVE)) != 0) {
                kprintf("[diagnostic] vop_compat_resolve: EAGAIN on ncp %p %s\n",
                        ncp, ncp->nc_name);
                return(EAGAIN);
        }

        /*
         * Setup the cnp for a traditional vop_old_lookup() call.  The lookup
         * caches all information required to create the entry in the
         * directory inode.  We expect a return code of EJUSTRETURN for
         * the CREATE case.  The cnp must simulated a saved-name situation.
         */
        bzero(&cnp, sizeof(cnp));
        cnp.cn_nameiop = NAMEI_CREATE;
        cnp.cn_flags = CNP_LOCKPARENT;
        cnp.cn_nameptr = ncp->nc_name;
        cnp.cn_namelen = ncp->nc_nlen;
        cnp.cn_cred = ap->a_cred;
        cnp.cn_td = td;

        tvp = NULL;
        error = vop_old_lookup(ap->a_head.a_ops, dvp, &tvp, &cnp);

        /*
         * EJUSTRETURN should be returned for this case, which means that
         * the VFS has setup the directory inode for the create.  The dvp we
         * passed in is expected to remain in a locked state.
         *
         * If the VOP_OLD_LINK is successful we are responsible for updating
         * the cache state of the locked ncp that was passed to us.
         */
        if (error == EJUSTRETURN) {
                KKASSERT((cnp.cn_flags & CNP_PDIRUNLOCK) == 0);
                error = VOP_OLD_LINK(dvp, ap->a_vp, &cnp);
                if (error == 0) {
                        cache_setunresolved(nch);
                        cache_setvp(nch, ap->a_vp);
                }
        } else {
                if (error == 0) {
                        vput(tvp);
                        error = EEXIST;
                }
        }
        if ((cnp.cn_flags & CNP_PDIRUNLOCK) == 0)
                vn_unlock(dvp);
        vrele(dvp);
        return (error);
}

int
vop_compat_nsymlink(struct vop_nsymlink_args *ap)
{
        struct thread *td = curthread;
        struct componentname cnp;
        struct nchandle *nch;
        struct namecache *ncp;
        struct vnode *dvp;
        struct vnode *vp;
        int error;

        /*
         * Sanity checks, get a locked directory vnode.
         */
        *ap->a_vpp = NULL;
        nch = ap->a_nch;                /* locked namecache node */
        ncp = nch->ncp;
        dvp = ap->a_dvp;

        if ((error = vget(dvp, LK_EXCLUSIVE)) != 0) {
                kprintf("[diagnostic] vop_compat_resolve: EAGAIN on ncp %p %s\n",
                        ncp, ncp->nc_name);
                return(EAGAIN);
        }

        /*
         * Setup the cnp for a traditional vop_old_lookup() call.  The lookup
         * caches all information required to create the entry in the
         * directory inode.  We expect a return code of EJUSTRETURN for
         * the CREATE case.  The cnp must simulated a saved-name situation.
         */
        bzero(&cnp, sizeof(cnp));
        cnp.cn_nameiop = NAMEI_CREATE;
        cnp.cn_flags = CNP_LOCKPARENT;
        cnp.cn_nameptr = ncp->nc_name;
        cnp.cn_namelen = ncp->nc_nlen;
        cnp.cn_cred = ap->a_cred;
        cnp.cn_td = td;

        vp = NULL;
        error = vop_old_lookup(ap->a_head.a_ops, dvp, &vp, &cnp);

        /*
         * EJUSTRETURN should be returned for this case, which means that
         * the VFS has setup the directory inode for the create.  The dvp we
         * passed in is expected to remain in a locked state.
         *
         * If the VOP_OLD_SYMLINK is successful we are responsible for updating
         * the cache state of the locked ncp that was passed to us.
         */
        if (error == EJUSTRETURN) {
                KKASSERT((cnp.cn_flags & CNP_PDIRUNLOCK) == 0);
                error = VOP_OLD_SYMLINK(dvp, &vp, &cnp, ap->a_vap, ap->a_target);
                if (error == 0) {
                        cache_setunresolved(nch);
                        cache_setvp(nch, vp);
                        *ap->a_vpp = vp;
                }
        } else {
                if (error == 0) {
                        vput(vp);
                        vp = NULL;
                        error = EEXIST;
                }
                KKASSERT(vp == NULL);
        }
        if ((cnp.cn_flags & CNP_PDIRUNLOCK) == 0)
                vn_unlock(dvp);
        vrele(dvp);
        return (error);
}

/*
 * vop_compat_nwhiteout { struct nchandle *a_nch,       XXX STOPGAP FUNCTION
 *                        struct vnode *a_dvp,
 *                        struct ucred *a_cred,
 *                        int a_flags }
 *
 * Issie a whiteout operation (create, lookup, or delete).  Compatibility 
 * requires us to issue the appropriate VOP_OLD_LOOKUP before we issue 
 * VOP_OLD_WHITEOUT in order to setup the directory inode's i_offset and i_count
 * (e.g. in UFS) for the NAMEI_CREATE and NAMEI_DELETE ops.  For NAMEI_LOOKUP
 * no lookup is necessary.
 */
int
vop_compat_nwhiteout(struct vop_nwhiteout_args *ap)
{
        struct thread *td = curthread;
        struct componentname cnp;
        struct nchandle *nch;
        struct namecache *ncp;
        struct vnode *dvp;
        struct vnode *vp;
        int error;

        /*
         * Sanity checks, get a locked directory vnode.
         */
        nch = ap->a_nch;                /* locked namecache node */
        ncp = nch->ncp;
        dvp = ap->a_dvp;

        if ((error = vget(dvp, LK_EXCLUSIVE)) != 0) {
                kprintf("[diagnostic] vop_compat_resolve: EAGAIN on ncp %p %s\n",
                        ncp, ncp->nc_name);
                return(EAGAIN);
        }

        /*
         * Setup the cnp for a traditional vop_old_lookup() call.  The lookup
         * caches all information required to create the entry in the
         * directory inode.  We expect a return code of EJUSTRETURN for
         * the CREATE case.  The cnp must simulated a saved-name situation.
         */
        bzero(&cnp, sizeof(cnp));
        cnp.cn_nameiop = ap->a_flags;
        cnp.cn_flags = CNP_LOCKPARENT;
        cnp.cn_nameptr = ncp->nc_name;
        cnp.cn_namelen = ncp->nc_nlen;
        cnp.cn_cred = ap->a_cred;
        cnp.cn_td = td;

        vp = NULL;

        /*
         * EJUSTRETURN should be returned for the CREATE or DELETE cases.
         * The VFS has setup the directory inode for the create.  The dvp we
         * passed in is expected to remain in a locked state.
         *
         * If the VOP_OLD_WHITEOUT is successful we are responsible for updating
         * the cache state of the locked ncp that was passed to us.
         */
        switch(ap->a_flags) {
        case NAMEI_DELETE:
                cnp.cn_flags |= CNP_DOWHITEOUT;
                /* fall through */
        case NAMEI_CREATE:
                error = vop_old_lookup(ap->a_head.a_ops, dvp, &vp, &cnp);
                if (error == EJUSTRETURN) {
                        KKASSERT((cnp.cn_flags & CNP_PDIRUNLOCK) == 0);
                        error = VOP_OLD_WHITEOUT(dvp, &cnp, ap->a_flags);
                        if (error == 0)
                                cache_setunresolved(nch);
                } else {
                        if (error == 0) {
                                vput(vp);
                                vp = NULL;
                                error = EEXIST;
                        }
                        KKASSERT(vp == NULL);
                }
                break;
        case NAMEI_LOOKUP:
                error = VOP_OLD_WHITEOUT(dvp, NULL, ap->a_flags);
                break;
        default:
                error = EINVAL;
                break;
        }
        if ((cnp.cn_flags & CNP_PDIRUNLOCK) == 0)
                vn_unlock(dvp);
        vrele(dvp);
        return (error);
}


/*
 * vop_compat_nremove { struct nchandle *a_nch,         XXX STOPGAP FUNCTION
 *                      struct vnode *a_dvp,
 *                      struct ucred *a_cred }
 */
int
vop_compat_nremove(struct vop_nremove_args *ap)
{
        struct thread *td = curthread;
        struct componentname cnp;
        struct nchandle *nch;
        struct namecache *ncp;
        struct vnode *dvp;
        struct vnode *vp;
        int error;

        /*
         * Sanity checks, get a locked directory vnode.
         */
        nch = ap->a_nch;                /* locked namecache node */
        ncp = nch->ncp;
        dvp = ap->a_dvp;

        if ((error = vget(dvp, LK_EXCLUSIVE)) != 0) {
                kprintf("[diagnostic] vop_compat_resolve: EAGAIN on ncp %p %s\n",
                        ncp, ncp->nc_name);
                return(EAGAIN);
        }

        /*
         * Setup the cnp for a traditional vop_old_lookup() call.  The lookup
         * caches all information required to delete the entry in the
         * directory inode.  We expect a return code of 0 for the DELETE
         * case (meaning that a vp has been found).  The cnp must simulated
         * a saved-name situation.
         */
        bzero(&cnp, sizeof(cnp));
        cnp.cn_nameiop = NAMEI_DELETE;
        cnp.cn_flags = CNP_LOCKPARENT;
        cnp.cn_nameptr = ncp->nc_name;
        cnp.cn_namelen = ncp->nc_nlen;
        cnp.cn_cred = ap->a_cred;
        cnp.cn_td = td;

        /*
         * The vnode must be a directory and must not represent the
         * current directory.
         */
        vp = NULL;
        error = vop_old_lookup(ap->a_head.a_ops, dvp, &vp, &cnp);
        if (error == 0 && vp->v_type == VDIR)
                error = EPERM;
        if (error == 0) {
                KKASSERT((cnp.cn_flags & CNP_PDIRUNLOCK) == 0);
                error = VOP_OLD_REMOVE(dvp, vp, &cnp);
                if (error == 0) {
                        cache_setunresolved(nch);
                        cache_setvp(nch, NULL);
                        cache_inval_vp(vp, CINV_DESTROY);
                }
        }
        if (vp) {
                if (dvp == vp)
                        vrele(vp);
                else    
                        vput(vp);
        }
        if ((cnp.cn_flags & CNP_PDIRUNLOCK) == 0)
                vn_unlock(dvp);
        vrele(dvp);
        return (error);
}

/*
 * vop_compat_nrmdir { struct nchandle *a_nch,  XXX STOPGAP FUNCTION
 *                     struct vnode *dvp,
 *                     struct ucred *a_cred }
 */
int
vop_compat_nrmdir(struct vop_nrmdir_args *ap)
{
        struct thread *td = curthread;
        struct componentname cnp;
        struct nchandle *nch;
        struct namecache *ncp;
        struct vnode *dvp;
        struct vnode *vp;
        int error;

        /*
         * Sanity checks, get a locked directory vnode.
         */
        nch = ap->a_nch;                /* locked namecache node */
        ncp = nch->ncp;
        dvp = ap->a_dvp;

        if ((error = vget(dvp, LK_EXCLUSIVE)) != 0) {
                kprintf("[diagnostic] vop_compat_resolve: EAGAIN on ncp %p %s\n",
                        ncp, ncp->nc_name);
                return(EAGAIN);
        }

        /*
         * Setup the cnp for a traditional vop_old_lookup() call.  The lookup
         * caches all information required to delete the entry in the
         * directory inode.  We expect a return code of 0 for the DELETE
         * case (meaning that a vp has been found).  The cnp must simulated
         * a saved-name situation.
         */
        bzero(&cnp, sizeof(cnp));
        cnp.cn_nameiop = NAMEI_DELETE;
        cnp.cn_flags = CNP_LOCKPARENT;
        cnp.cn_nameptr = ncp->nc_name;
        cnp.cn_namelen = ncp->nc_nlen;
        cnp.cn_cred = ap->a_cred;
        cnp.cn_td = td;

        /*
         * The vnode must be a directory and must not represent the
         * current directory.
         */
        vp = NULL;
        error = vop_old_lookup(ap->a_head.a_ops, dvp, &vp, &cnp);
        if (error == 0 && vp->v_type != VDIR)
                error = ENOTDIR;
        if (error == 0 && vp == dvp)
                error = EINVAL;
        if (error == 0 && (vp->v_flag & VROOT))
                error = EBUSY;
        if (error == 0) {
                KKASSERT((cnp.cn_flags & CNP_PDIRUNLOCK) == 0);
                error = VOP_OLD_RMDIR(dvp, vp, &cnp);

                /*
                 * Note that this invalidation will cause any process
                 * currently CD'd into the directory being removed to be
                 * disconnected from the topology and not be able to ".."
                 * back out.
                 */
                if (error == 0) {
                        cache_inval(nch, CINV_DESTROY);
                        cache_inval_vp(vp, CINV_DESTROY);
                }
        }
        if (vp) {
                if (dvp == vp)
                        vrele(vp);
                else    
                        vput(vp);
        }
        if ((cnp.cn_flags & CNP_PDIRUNLOCK) == 0)
                vn_unlock(dvp);
        vrele(dvp);
        return (error);
}

/*
 * vop_compat_nrename { struct nchandle *a_fnch,        XXX STOPGAP FUNCTION
 *                      struct nchandle *a_tnch,
 *                      struct ucred *a_cred }
 *
 * This is a fairly difficult procedure.  The old VOP_OLD_RENAME requires that
 * the source directory and vnode be unlocked and the target directory and
 * vnode (if it exists) be locked.  All arguments will be vrele'd and 
 * the targets will also be unlocked regardless of the return code.
 */
int
vop_compat_nrename(struct vop_nrename_args *ap)
{
        struct thread *td = curthread;
        struct componentname fcnp;
        struct componentname tcnp;
        struct nchandle *fnch;
        struct nchandle *tnch;
        struct namecache *fncp;
        struct namecache *tncp;
        struct vnode *fdvp, *fvp;
        struct vnode *tdvp, *tvp;
        int error;

        /*
         * Sanity checks, get referenced vnodes representing the source.
         */
        fnch = ap->a_fnch;              /* locked namecache node */
        fncp = fnch->ncp;
        fdvp = ap->a_fdvp;

        /*
         * Temporarily lock the source directory and lookup in DELETE mode to
         * check permissions.  XXX delete permissions should have been
         * checked by nlookup(), we need to add NLC_DELETE for delete
         * checking.  It is unclear whether VFS's require the directory setup
         * info NAMEI_DELETE causes to be stored in the fdvp's inode, but
         * since it isn't locked and since UFS always does a relookup of
         * the source, it is believed that the only side effect that matters
         * is the permissions check.
         */
        if ((error = vget(fdvp, LK_EXCLUSIVE)) != 0) {
                kprintf("[diagnostic] vop_compat_resolve: EAGAIN on ncp %p %s\n",
                        fncp, fncp->nc_name);
                return(EAGAIN);
        }

        bzero(&fcnp, sizeof(fcnp));
        fcnp.cn_nameiop = NAMEI_DELETE;
        fcnp.cn_flags = CNP_LOCKPARENT;
        fcnp.cn_nameptr = fncp->nc_name;
        fcnp.cn_namelen = fncp->nc_nlen;
        fcnp.cn_cred = ap->a_cred;
        fcnp.cn_td = td;

        /*
         * note: vop_old_lookup (i.e. VOP_OLD_LOOKUP) always returns a locked
         * fvp.
         */
        fvp = NULL;
        error = vop_old_lookup(ap->a_head.a_ops, fdvp, &fvp, &fcnp);
        if (error == 0 && (fvp->v_flag & VROOT)) {
                vput(fvp);      /* as if vop_old_lookup had failed */
                error = EBUSY;
        }
        if ((fcnp.cn_flags & CNP_PDIRUNLOCK) == 0) {
                fcnp.cn_flags |= CNP_PDIRUNLOCK;
                vn_unlock(fdvp);
        }
        if (error) {
                vrele(fdvp);
                return (error);
        }
        vn_unlock(fvp);

        /*
         * fdvp and fvp are now referenced and unlocked.
         *
         * Get a locked directory vnode for the target and lookup the target
         * in CREATE mode so it places the required information in the
         * directory inode.
         */
        tnch = ap->a_tnch;              /* locked namecache node */
        tncp = tnch->ncp;
        tdvp = ap->a_tdvp;
        if (error) {
                vrele(fdvp);
                vrele(fvp);
                return (error);
        }
        if ((error = vget(tdvp, LK_EXCLUSIVE)) != 0) {
                kprintf("[diagnostic] vop_compat_resolve: EAGAIN on ncp %p %s\n",
                        tncp, tncp->nc_name);
                vrele(fdvp);
                vrele(fvp);
                return(EAGAIN);
        }

        /*
         * Setup the cnp for a traditional vop_old_lookup() call.  The lookup
         * caches all information required to create the entry in the
         * target directory inode.
         */
        bzero(&tcnp, sizeof(tcnp));
        tcnp.cn_nameiop = NAMEI_RENAME;
        tcnp.cn_flags = CNP_LOCKPARENT;
        tcnp.cn_nameptr = tncp->nc_name;
        tcnp.cn_namelen = tncp->nc_nlen;
        tcnp.cn_cred = ap->a_cred;
        tcnp.cn_td = td;

        tvp = NULL;
        error = vop_old_lookup(ap->a_head.a_ops, tdvp, &tvp, &tcnp);

        if (error == EJUSTRETURN) {
                /*
                 * Target does not exist.  tvp should be NULL.
                 */
                KKASSERT(tvp == NULL);
                KKASSERT((tcnp.cn_flags & CNP_PDIRUNLOCK) == 0);
                error = VOP_OLD_RENAME(fdvp, fvp, &fcnp, tdvp, tvp, &tcnp);
                if (error == 0)
                        cache_rename(fnch, tnch);
        } else if (error == 0) {
                /*
                 * Target exists.  VOP_OLD_RENAME should correctly delete the
                 * target.
                 */
                KKASSERT((tcnp.cn_flags & CNP_PDIRUNLOCK) == 0);
                error = VOP_OLD_RENAME(fdvp, fvp, &fcnp, tdvp, tvp, &tcnp);
                if (error == 0)
                        cache_rename(fnch, tnch);
        } else {
                vrele(fdvp);
                vrele(fvp);
                if (tcnp.cn_flags & CNP_PDIRUNLOCK)
                        vrele(tdvp);
                else
                        vput(tdvp);
        }
        return (error);
}

static int
vop_nolookup(struct vop_old_lookup_args *ap)
{

        *ap->a_vpp = NULL;
        return (ENOTDIR);
}

/*
 *      vop_nostrategy:
 *
 *      Strategy routine for VFS devices that have none.
 *
 *      B_ERROR and B_INVAL must be cleared prior to calling any strategy
 *      routine.  Typically this is done for a BUF_CMD_READ strategy call.
 *      Typically B_INVAL is assumed to already be clear prior to a write
 *      and should not be cleared manually unless you just made the buffer
 *      invalid.  B_ERROR should be cleared either way.
 */

static int
vop_nostrategy (struct vop_strategy_args *ap)
{
        kprintf("No strategy for buffer at %p\n", ap->a_bio->bio_buf);
        vprint("", ap->a_vp);
        ap->a_bio->bio_buf->b_flags |= B_ERROR;
        ap->a_bio->bio_buf->b_error = EOPNOTSUPP;
        biodone(ap->a_bio);
        return (EOPNOTSUPP);
}

int
vop_stdpathconf(struct vop_pathconf_args *ap)
{

        switch (ap->a_name) {
                case _PC_LINK_MAX:
                        *ap->a_retval = LINK_MAX;
                        return (0);
                case _PC_MAX_CANON:
                        *ap->a_retval = MAX_CANON;
                        return (0);
                case _PC_MAX_INPUT:
                        *ap->a_retval = MAX_INPUT;
                        return (0);
                case _PC_PIPE_BUF:
                        *ap->a_retval = PIPE_BUF;
                        return (0);
                case _PC_CHOWN_RESTRICTED:
                        *ap->a_retval = 1;
                        return (0);
                case _PC_VDISABLE:
                        *ap->a_retval = _POSIX_VDISABLE;
                        return (0);
                default:
                        return (EINVAL);
        }
        /* NOTREACHED */
}

/*
 * Standard open.
 *
 * (struct vnode *a_vp, int a_mode, struct ucred *a_ucred, struct file *a_fp)
 *
 * a_mode: note, 'F' modes, e.g. FREAD, FWRITE
 */
int
vop_stdopen(struct vop_open_args *ap)
{
        struct vnode *vp = ap->a_vp;
        struct file *fp;

        if ((fp = ap->a_fp) != NULL) {
                switch(vp->v_type) {
                case VFIFO:
                        fp->f_type = DTYPE_FIFO;
                        break;
                default:
                        fp->f_type = DTYPE_VNODE;
                        break;
                }
                fp->f_flag = ap->a_mode & FMASK;
                fp->f_ops = &vnode_fileops;
                fp->f_data = vp;
                vref(vp);
        }
        if (ap->a_mode & FWRITE)
                ++vp->v_writecount;
        KKASSERT(vp->v_opencount >= 0 && vp->v_opencount != INT_MAX);
        ++vp->v_opencount;
        return (0);
}

/*
 * Standard close.
 *
 * (struct vnode *a_vp, int a_fflag)
 *
 * a_fflag: note, 'F' modes, e.g. FREAD, FWRITE.  same as a_mode in stdopen?
 */
int
vop_stdclose(struct vop_close_args *ap)
{
        struct vnode *vp = ap->a_vp;

        KASSERT(vp->v_opencount > 0,
                ("VOP_STDCLOSE: BAD OPENCOUNT %p %d type=%d ops=%p flgs=%08x\n",
                vp, vp->v_opencount, vp->v_type, *vp->v_ops, vp->v_flag));
        if (ap->a_fflag & FWRITE) {
                KASSERT(vp->v_writecount > 0,
                        ("VOP_STDCLOSE: BAD WRITECOUNT %p %d\n", 
                        vp, vp->v_writecount));
                --vp->v_writecount;
        }
        --vp->v_opencount;
        return (0);
}

/*
 * Return true for select/poll.
 */
int
vop_nopoll(struct vop_poll_args *ap)
{
        /*
         * Return true for read/write.  If the user asked for something
         * special, return POLLNVAL, so that clients have a way of
         * determining reliably whether or not the extended
         * functionality is present without hard-coding knowledge
         * of specific filesystem implementations.
         */
        if (ap->a_events & ~POLLSTANDARD)
                return (POLLNVAL);

        return (ap->a_events & (POLLIN | POLLOUT | POLLRDNORM | POLLWRNORM));
}

/*
 * Implement poll for local filesystems that support it.
 */
int
vop_stdpoll(struct vop_poll_args *ap)
{
        if (ap->a_events & ~POLLSTANDARD)
                return (vn_pollrecord(ap->a_vp, ap->a_events));
        return (ap->a_events & (POLLIN | POLLOUT | POLLRDNORM | POLLWRNORM));
}

/*
 * Implement standard getpages and putpages.  All filesystems must use
 * the buffer cache to back regular files.
 */
int
vop_stdgetpages(struct vop_getpages_args *ap)
{
        struct mount *mp;
        int error;

        if ((mp = ap->a_vp->v_mount) != NULL) {
                error = vnode_pager_generic_getpages(
                                ap->a_vp, ap->a_m, ap->a_count,
                                ap->a_reqpage);
        } else {
                error = VM_PAGER_BAD;
        }
        return (error);
}

int
vop_stdputpages(struct vop_putpages_args *ap)
{
        struct mount *mp;
        int error;

        if ((mp = ap->a_vp->v_mount) != NULL) {
                error = vnode_pager_generic_putpages(
                                ap->a_vp, ap->a_m, ap->a_count,
                                ap->a_sync, ap->a_rtvals);
        } else {
                error = VM_PAGER_BAD;
        }
        return (error);
}

/* 
 * vfs default ops
 * used to fill the vfs fucntion table to get reasonable default return values.
 */
int 
vfs_stdmount(struct mount *mp, char *path, caddr_t data, struct ucred *cred)
{
        return (0);
}

int     
vfs_stdunmount(struct mount *mp, int mntflags)
{
        return (0);
}

int     
vfs_stdroot(struct mount *mp, struct vnode **vpp)
{
        return (EOPNOTSUPP);
}

int     
vfs_stdstatfs(struct mount *mp, struct statfs *sbp, struct ucred *cred)
{
        return (EOPNOTSUPP);
}

/*
 * If the VFS does not implement statvfs, then call statfs and convert
 * the values.  This code was taken from libc's __cvtstatvfs() function,
 * contributed by Joerg Sonnenberger.
 */
int     
vfs_stdstatvfs(struct mount *mp, struct statvfs *sbp, struct ucred *cred)
{
        struct statfs *in;
        int error;

        in = &mp->mnt_stat;
        error = VFS_STATFS(mp, in, cred);
        if (error == 0) {
                bzero(sbp, sizeof(*sbp));

                sbp->f_bsize = in->f_bsize;
                sbp->f_frsize = in->f_bsize;
                sbp->f_blocks = in->f_blocks;
                sbp->f_bfree = in->f_bfree;
                sbp->f_bavail = in->f_bavail;
                sbp->f_files = in->f_files;
                sbp->f_ffree = in->f_ffree;

                /*
                 * XXX
                 * This field counts the number of available inodes to non-root
                 * users, but this information is not available via statfs.
                 * Just ignore this issue by returning the total number 
                 * instead.
                 */
                sbp->f_favail = in->f_ffree;

                /*
                 * XXX
                 * This field has a different meaning for statfs and statvfs.
                 * For the former it is the cookie exported for NFS and not
                 * intended for normal userland use.
                 */
                sbp->f_fsid = 0;

                sbp->f_flag = 0;
                if (in->f_flags & MNT_RDONLY)
                        sbp->f_flag |= ST_RDONLY;
                if (in->f_flags & MNT_NOSUID)
                        sbp->f_flag |= ST_NOSUID;
                sbp->f_namemax = 0;
                sbp->f_owner = in->f_owner;
                /*
                 * XXX
                 * statfs contains the type as string, statvfs expects it as
                 * enumeration.
                 */
                sbp->f_type = 0;

                sbp->f_syncreads = in->f_syncreads;
                sbp->f_syncwrites = in->f_syncwrites;
                sbp->f_asyncreads = in->f_asyncreads;
                sbp->f_asyncwrites = in->f_asyncwrites;
        }
        return (error);
}

int
vfs_stdvptofh(struct vnode *vp, struct fid *fhp)
{
        return (EOPNOTSUPP);
}

int     
vfs_stdstart(struct mount *mp, int flags)
{
        return (0);
}

int     
vfs_stdquotactl(struct mount *mp, int cmds, uid_t uid,
        caddr_t arg, struct ucred *cred)
{
        return (EOPNOTSUPP);
}

int     
vfs_stdsync(struct mount *mp, int waitfor)
{
        return (0);
}

int
vfs_stdnosync(struct mount *mp, int waitfor)
{
        return (EOPNOTSUPP);
}

int     
vfs_stdvget(struct mount *mp, ino_t ino, struct vnode **vpp)
{
        return (EOPNOTSUPP);
}

int     
vfs_stdfhtovp(struct mount *mp, struct fid *fhp, struct vnode **vpp)
{
        return (EOPNOTSUPP);
}

int 
vfs_stdcheckexp(struct mount *mp, struct sockaddr *nam, int *extflagsp,
        struct ucred **credanonp)
{
        return (EOPNOTSUPP);
}

int
vfs_stdinit(struct vfsconf *vfsp)
{
        return (0);
}

int
vfs_stduninit(struct vfsconf *vfsp)
{
        return(0);
}

int
vfs_stdextattrctl(struct mount *mp, int cmd, const char *attrname,
        caddr_t arg, struct ucred *cred)
{
        return(EOPNOTSUPP);
}

/* end of vfs default ops */