[deliverable/linux.git] / fs / nfs / namespace.c

/*
 * linux/fs/nfs/namespace.c
 *
 * Copyright (C) 2005 Trond Myklebust <Trond.Myklebust@netapp.com>
 * - Modified by David Howells <dhowells@redhat.com>
 *
 * NFS namespace
 */

#include <linux/dcache.h>
#include <linux/mount.h>
#include <linux/namei.h>
#include <linux/nfs_fs.h>
#include <linux/string.h>
#include <linux/sunrpc/clnt.h>
#include <linux/vfs.h>
#include "internal.h"

#define NFSDBG_FACILITY		NFSDBG_VFS

static void nfs_expire_automounts(struct work_struct *work);

static LIST_HEAD(nfs_automount_list);
static DECLARE_DELAYED_WORK(nfs_automount_task, nfs_expire_automounts);
int nfs_mountpoint_expiry_timeout = 500 * HZ;

static struct vfsmount *nfs_do_submount(const struct vfsmount *mnt_parent,
					const struct dentry *dentry,
					struct nfs_fh *fh,
					struct nfs_fattr *fattr);

/*
 * nfs_path - reconstruct the path given an arbitrary dentry
 * @base - arbitrary string to prepend to the path
 * @droot - pointer to root dentry for mountpoint
 * @dentry - pointer to dentry
 * @buffer - result buffer
 * @buflen - length of buffer
 *
 * Helper function for constructing the path from the
 * root dentry to an arbitrary hashed dentry.
 *
 * This is mainly for use in figuring out the path on the
 * server side when automounting on top of an existing partition.
 */
char *nfs_path(const char *base,
	       const struct dentry *droot,
	       const struct dentry *dentry,
	       char *buffer, ssize_t buflen)
{
	char *end = buffer+buflen;
	int namelen;

	*--end = '\0';
	buflen--;
	spin_lock(&dcache_lock);
	while (!IS_ROOT(dentry) && dentry != droot) {
		namelen = dentry->d_name.len;
		buflen -= namelen + 1;
		if (buflen < 0)
			goto Elong_unlock;
		end -= namelen;
		memcpy(end, dentry->d_name.name, namelen);
		*--end = '/';
		dentry = dentry->d_parent;
	}
	spin_unlock(&dcache_lock);
	namelen = strlen(base);
	/* Strip off excess slashes in base string */
	while (namelen > 0 && base[namelen - 1] == '/')
		namelen--;
	buflen -= namelen;
	if (buflen < 0)
		goto Elong;
	end -= namelen;
	memcpy(end, base, namelen);
	return end;
Elong_unlock:
	spin_unlock(&dcache_lock);
Elong:
	return ERR_PTR(-ENAMETOOLONG);
}

/*
 * nfs_follow_mountpoint - handle crossing a mountpoint on the server
 * @dentry - dentry of mountpoint
 * @nd - nameidata info
 *
 * When we encounter a mountpoint on the server, we want to set up
 * a mountpoint on the client too, to prevent inode numbers from
 * colliding, and to allow "df" to work properly.
 * On NFSv4, we also want to allow for the fact that different
 * filesystems may be migrated to different servers in a failover
 * situation, and that different filesystems may want to use
 * different security flavours.
 */
static void * nfs_follow_mountpoint(struct dentry *dentry, struct nameidata *nd)
{
	struct vfsmount *mnt;
	struct nfs_server *server = NFS_SERVER(dentry->d_inode);
	struct dentry *parent;
	struct nfs_fh fh;
	struct nfs_fattr fattr;
	int err;

	dprintk("--> nfs_follow_mountpoint()\n");

	err = -ESTALE;
	if (IS_ROOT(dentry))
		goto out_err;

	dprintk("%s: enter\n", __func__);
	dput(nd->path.dentry);
	nd->path.dentry = dget(dentry);

	/* Look it up again */
	parent = dget_parent(nd->path.dentry);
	err = server->nfs_client->rpc_ops->lookup(parent->d_inode,
						  &nd->path.dentry->d_name,
						  &fh, &fattr);
	dput(parent);
	if (err != 0)
		goto out_err;

	if (fattr.valid & NFS_ATTR_FATTR_V4_REFERRAL)
		mnt = nfs_do_refmount(nd->path.mnt, nd->path.dentry);
	else
		mnt = nfs_do_submount(nd->path.mnt, nd->path.dentry, &fh,
				      &fattr);
	err = PTR_ERR(mnt);
	if (IS_ERR(mnt))
		goto out_err;

	mntget(mnt);
	err = do_add_mount(mnt, &nd->path, nd->path.mnt->mnt_flags|MNT_SHRINKABLE,
			   &nfs_automount_list);
	if (err < 0) {
		mntput(mnt);
		if (err == -EBUSY)
			goto out_follow;
		goto out_err;
	}
	path_put(&nd->path);
	nd->path.mnt = mnt;
	nd->path.dentry = dget(mnt->mnt_root);
	schedule_delayed_work(&nfs_automount_task, nfs_mountpoint_expiry_timeout);
out:
	dprintk("%s: done, returned %d\n", __func__, err);

	dprintk("<-- nfs_follow_mountpoint() = %d\n", err);
	return ERR_PTR(err);
out_err:
	path_put(&nd->path);
	goto out;
out_follow:
	while (d_mountpoint(nd->path.dentry) &&
	       follow_down(&nd->path.mnt, &nd->path.dentry))
		;
	err = 0;
	goto out;
}

const struct inode_operations nfs_mountpoint_inode_operations = {
	.follow_link	= nfs_follow_mountpoint,
	.getattr	= nfs_getattr,
};

const struct inode_operations nfs_referral_inode_operations = {
	.follow_link	= nfs_follow_mountpoint,
};

static void nfs_expire_automounts(struct work_struct *work)
{
	struct list_head *list = &nfs_automount_list;

	mark_mounts_for_expiry(list);
	if (!list_empty(list))
		schedule_delayed_work(&nfs_automount_task, nfs_mountpoint_expiry_timeout);
}

void nfs_release_automount_timer(void)
{
	if (list_empty(&nfs_automount_list))
		cancel_delayed_work(&nfs_automount_task);
}

/*
 * Clone a mountpoint of the appropriate type
 */
static struct vfsmount *nfs_do_clone_mount(struct nfs_server *server,
					   const char *devname,
					   struct nfs_clone_mount *mountdata)
{
#ifdef CONFIG_NFS_V4
	struct vfsmount *mnt = ERR_PTR(-EINVAL);
	switch (server->nfs_client->rpc_ops->version) {
		case 2:
		case 3:
			mnt = vfs_kern_mount(&nfs_xdev_fs_type, 0, devname, mountdata);
			break;
		case 4:
			mnt = vfs_kern_mount(&nfs4_xdev_fs_type, 0, devname, mountdata);
	}
	return mnt;
#else
	return vfs_kern_mount(&nfs_xdev_fs_type, 0, devname, mountdata);
#endif
}

/**
 * nfs_do_submount - set up mountpoint when crossing a filesystem boundary
 * @mnt_parent - mountpoint of parent directory
 * @dentry - parent directory
 * @fh - filehandle for new root dentry
 * @fattr - attributes for new root inode
 *
 */
static struct vfsmount *nfs_do_submount(const struct vfsmount *mnt_parent,
					const struct dentry *dentry,
					struct nfs_fh *fh,
					struct nfs_fattr *fattr)
{
	struct nfs_clone_mount mountdata = {
		.sb = mnt_parent->mnt_sb,
		.dentry = dentry,
		.fh = fh,
		.fattr = fattr,
	};
	struct vfsmount *mnt = ERR_PTR(-ENOMEM);
	char *page = (char *) __get_free_page(GFP_USER);
	char *devname;

	dprintk("--> nfs_do_submount()\n");

	dprintk("%s: submounting on %s/%s\n", __func__,
			dentry->d_parent->d_name.name,
			dentry->d_name.name);
	if (page == NULL)
		goto out;
	devname = nfs_devname(mnt_parent, dentry, page, PAGE_SIZE);
	mnt = (struct vfsmount *)devname;
	if (IS_ERR(devname))
		goto free_page;
	mnt = nfs_do_clone_mount(NFS_SB(mnt_parent->mnt_sb), devname, &mountdata);
free_page:
	free_page((unsigned long)page);
out:
	dprintk("%s: done\n", __func__);

	dprintk("<-- nfs_do_submount() = %p\n", mnt);
	return mnt;
}
Commit	Line	Data
55a97593 TM	1	/*
	2	* linux/fs/nfs/namespace.c
	3	*
	4	* Copyright (C) 2005 Trond Myklebust <Trond.Myklebust@netapp.com>
54ceac45	5	* - Modified by David Howells <dhowells@redhat.com>
55a97593 TM	6	*
	7	* NFS namespace
	8	*/
	9
55a97593 TM	10	#include <linux/dcache.h>
	11	#include <linux/mount.h>
	12	#include <linux/namei.h>
	13	#include <linux/nfs_fs.h>
	14	#include <linux/string.h>
	15	#include <linux/sunrpc/clnt.h>
	16	#include <linux/vfs.h>
f7b422b1	17	#include "internal.h"
55a97593 TM	18
	19	#define NFSDBG_FACILITY NFSDBG_VFS
	20
65f27f38	21	static void nfs_expire_automounts(struct work_struct *work);
f7b422b1	22
a3dab293	23	static LIST_HEAD(nfs_automount_list);
65f27f38	24	static DECLARE_DELAYED_WORK(nfs_automount_task, nfs_expire_automounts);
51d8fa6a TM	25	int nfs_mountpoint_expiry_timeout = 500 * HZ;
51d8fa6a TM	26
66f37509 AB	27	static struct vfsmount nfs_do_submount(const struct vfsmount mnt_parent,
	28	const struct dentry *dentry,
	29	struct nfs_fh *fh,
	30	struct nfs_fattr *fattr);
	31
f7b422b1 DH	32	/*
	33	* nfs_path - reconstruct the path given an arbitrary dentry
	34	* @base - arbitrary string to prepend to the path
54ceac45	35	* @droot - pointer to root dentry for mountpoint
f7b422b1 DH	36	* @dentry - pointer to dentry
	37	* @buffer - result buffer
	38	* @buflen - length of buffer
	39	*
	40	* Helper function for constructing the path from the
	41	* root dentry to an arbitrary hashed dentry.
	42	*
	43	* This is mainly for use in figuring out the path on the
	44	* server side when automounting on top of an existing partition.
	45	*/
54ceac45 DH	46	char nfs_path(const char base,
	47	const struct dentry *droot,
	48	const struct dentry *dentry,
f7b422b1 DH	49	char *buffer, ssize_t buflen)
	50	{
	51	char *end = buffer+buflen;
	52	int namelen;
	53
	54	*--end = '\0';
	55	buflen--;
	56	spin_lock(&dcache_lock);
54ceac45	57	while (!IS_ROOT(dentry) && dentry != droot) {
f7b422b1 DH	58	namelen = dentry->d_name.len;
	59	buflen -= namelen + 1;
	60	if (buflen < 0)
ce510193	61	goto Elong_unlock;
f7b422b1 DH	62	end -= namelen;
	63	memcpy(end, dentry->d_name.name, namelen);
	64	*--end = '/';
	65	dentry = dentry->d_parent;
	66	}
	67	spin_unlock(&dcache_lock);
	68	namelen = strlen(base);
	69	/* Strip off excess slashes in base string */
	70	while (namelen > 0 && base[namelen - 1] == '/')
	71	namelen--;
	72	buflen -= namelen;
	73	if (buflen < 0)
	74	goto Elong;
	75	end -= namelen;
	76	memcpy(end, base, namelen);
	77	return end;
ce510193 JT	78	Elong_unlock:
ce510193 JT	79	spin_unlock(&dcache_lock);
f7b422b1 DH	80	Elong:
	81	return ERR_PTR(-ENAMETOOLONG);
	82	}
	83
55a97593 TM	84	/*
	85	* nfs_follow_mountpoint - handle crossing a mountpoint on the server
	86	* @dentry - dentry of mountpoint
	87	* @nd - nameidata info
	88	*
	89	* When we encounter a mountpoint on the server, we want to set up
	90	* a mountpoint on the client too, to prevent inode numbers from
	91	* colliding, and to allow "df" to work properly.
	92	* On NFSv4, we also want to allow for the fact that different
	93	* filesystems may be migrated to different servers in a failover
	94	* situation, and that different filesystems may want to use
	95	* different security flavours.
	96	*/
	97	static void * nfs_follow_mountpoint(struct dentry dentry, struct nameidata nd)
	98	{
	99	struct vfsmount *mnt;
	100	struct nfs_server *server = NFS_SERVER(dentry->d_inode);
	101	struct dentry *parent;
	102	struct nfs_fh fh;
	103	struct nfs_fattr fattr;
	104	int err;
	105
54ceac45 DH	106	dprintk("--> nfs_follow_mountpoint()\n");
54ceac45 DH	107
44d5759d DL	108	err = -ESTALE;
	109	if (IS_ROOT(dentry))
	110	goto out_err;
	111
3110ff80	112	dprintk("%s: enter\n", __func__);
4ac91378 JB	113	dput(nd->path.dentry);
4ac91378 JB	114	nd->path.dentry = dget(dentry);
54ceac45	115
55a97593	116	/* Look it up again */
4ac91378	117	parent = dget_parent(nd->path.dentry);
8fa5c000	118	err = server->nfs_client->rpc_ops->lookup(parent->d_inode,
4ac91378	119	&nd->path.dentry->d_name,
8fa5c000	120	&fh, &fattr);
55a97593 TM	121	dput(parent);
	122	if (err != 0)
	123	goto out_err;
	124
6b97fd3d	125	if (fattr.valid & NFS_ATTR_FATTR_V4_REFERRAL)
4ac91378	126	mnt = nfs_do_refmount(nd->path.mnt, nd->path.dentry);
6b97fd3d	127	else
4ac91378 JB	128	mnt = nfs_do_submount(nd->path.mnt, nd->path.dentry, &fh,
4ac91378 JB	129	&fattr);
55a97593 TM	130	err = PTR_ERR(mnt);
	131	if (IS_ERR(mnt))
	132	goto out_err;
	133
	134	mntget(mnt);
8d66bf54	135	err = do_add_mount(mnt, &nd->path, nd->path.mnt->mnt_flags\|MNT_SHRINKABLE,
4ac91378	136	&nfs_automount_list);
55a97593 TM	137	if (err < 0) {
	138	mntput(mnt);
	139	if (err == -EBUSY)
	140	goto out_follow;
	141	goto out_err;
	142	}
31f31db1	143	path_put(&nd->path);
4ac91378 JB	144	nd->path.mnt = mnt;
4ac91378 JB	145	nd->path.dentry = dget(mnt->mnt_root);
51d8fa6a	146	schedule_delayed_work(&nfs_automount_task, nfs_mountpoint_expiry_timeout);
55a97593	147	out:
3110ff80	148	dprintk("%s: done, returned %d\n", __func__, err);
54ceac45 DH	149
54ceac45 DH	150	dprintk("<-- nfs_follow_mountpoint() = %d\n", err);
55a97593 TM	151	return ERR_PTR(err);
55a97593 TM	152	out_err:
1d957f9b	153	path_put(&nd->path);
55a97593 TM	154	goto out;
55a97593 TM	155	out_follow:
4ac91378 JB	156	while (d_mountpoint(nd->path.dentry) &&
4ac91378 JB	157	follow_down(&nd->path.mnt, &nd->path.dentry))
55a97593 TM	158	;
	159	err = 0;
	160	goto out;
	161	}
	162
92e1d5be	163	const struct inode_operations nfs_mountpoint_inode_operations = {
55a97593 TM	164	.follow_link = nfs_follow_mountpoint,
	165	.getattr = nfs_getattr,
	166	};
51d8fa6a	167
92e1d5be	168	const struct inode_operations nfs_referral_inode_operations = {
6b97fd3d MN	169	.follow_link = nfs_follow_mountpoint,
	170	};
	171
65f27f38	172	static void nfs_expire_automounts(struct work_struct *work)
51d8fa6a	173	{
65f27f38	174	struct list_head *list = &nfs_automount_list;
51d8fa6a TM	175
	176	mark_mounts_for_expiry(list);
	177	if (!list_empty(list))
	178	schedule_delayed_work(&nfs_automount_task, nfs_mountpoint_expiry_timeout);
	179	}
	180
	181	void nfs_release_automount_timer(void)
	182	{
3d39c691	183	if (list_empty(&nfs_automount_list))
560aef74	184	cancel_delayed_work(&nfs_automount_task);
51d8fa6a	185	}
f7b422b1 DH	186
	187	/*
	188	* Clone a mountpoint of the appropriate type
	189	*/
509de811 DH	190	static struct vfsmount nfs_do_clone_mount(struct nfs_server server,
509de811 DH	191	const char *devname,
f7b422b1 DH	192	struct nfs_clone_mount *mountdata)
	193	{
	194	#ifdef CONFIG_NFS_V4
fd08d7e9	195	struct vfsmount *mnt = ERR_PTR(-EINVAL);
40c55319	196	switch (server->nfs_client->rpc_ops->version) {
f7b422b1 DH	197	case 2:
f7b422b1 DH	198	case 3:
54ceac45	199	mnt = vfs_kern_mount(&nfs_xdev_fs_type, 0, devname, mountdata);
f7b422b1 DH	200	break;
f7b422b1 DH	201	case 4:
54ceac45	202	mnt = vfs_kern_mount(&nfs4_xdev_fs_type, 0, devname, mountdata);
f7b422b1 DH	203	}
	204	return mnt;
	205	#else
54ceac45	206	return vfs_kern_mount(&nfs_xdev_fs_type, 0, devname, mountdata);
f7b422b1 DH	207	#endif
	208	}
	209
	210	/**
	211	* nfs_do_submount - set up mountpoint when crossing a filesystem boundary
	212	* @mnt_parent - mountpoint of parent directory
	213	* @dentry - parent directory
	214	* @fh - filehandle for new root dentry
	215	* @fattr - attributes for new root inode
	216	*
	217	*/
66f37509 AB	218	static struct vfsmount nfs_do_submount(const struct vfsmount mnt_parent,
	219	const struct dentry *dentry,
	220	struct nfs_fh *fh,
	221	struct nfs_fattr *fattr)
f7b422b1 DH	222	{
	223	struct nfs_clone_mount mountdata = {
	224	.sb = mnt_parent->mnt_sb,
	225	.dentry = dentry,
	226	.fh = fh,
	227	.fattr = fattr,
	228	};
	229	struct vfsmount *mnt = ERR_PTR(-ENOMEM);
	230	char page = (char ) __get_free_page(GFP_USER);
	231	char *devname;
	232
54ceac45 DH	233	dprintk("--> nfs_do_submount()\n");
54ceac45 DH	234
3110ff80	235	dprintk("%s: submounting on %s/%s\n", __func__,
f7b422b1 DH	236	dentry->d_parent->d_name.name,
	237	dentry->d_name.name);
	238	if (page == NULL)
	239	goto out;
	240	devname = nfs_devname(mnt_parent, dentry, page, PAGE_SIZE);
	241	mnt = (struct vfsmount *)devname;
	242	if (IS_ERR(devname))
	243	goto free_page;
	244	mnt = nfs_do_clone_mount(NFS_SB(mnt_parent->mnt_sb), devname, &mountdata);
	245	free_page:
	246	free_page((unsigned long)page);
	247	out:
3110ff80	248	dprintk("%s: done\n", __func__);
54ceac45 DH	249
54ceac45 DH	250	dprintk("<-- nfs_do_submount() = %p\n", mnt);
f7b422b1 DH	251	return mnt;
f7b422b1 DH	252	}