nfs4proc.c

linux 内核源代码

/*
 *  fs/nfs/nfs4proc.c
 *
 *  Client-side procedure declarations for NFSv4.
 *
 *  Copyright (c) 2002 The Regents of the University of Michigan.
 *  All rights reserved.
 *
 *  Kendrick Smith <[e-mail]>
 *  Andy Adamson   <[e-mail]>
 *
 *  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. 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 ``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.
 */

#include <linux/mm.h>
#include <linux/utsname.h>
#include <linux/delay.h>
#include <linux/errno.h>
#include <linux/string.h>
#include <linux/sunrpc/clnt.h>
#include <linux/nfs.h>
#include <linux/nfs4.h>
#include <linux/nfs_fs.h>
#include <linux/nfs_page.h>
#include <linux/smp_lock.h>
#include <linux/namei.h>
#include <linux/mount.h>

#include "nfs4_fs.h"
#include "delegation.h"
#include "iostat.h"

#define NFSDBG_FACILITY		NFSDBG_PROC

#define NFS4_POLL_RETRY_MIN	(HZ/10)
#define NFS4_POLL_RETRY_MAX	(15*HZ)

struct nfs4_opendata;
static int _nfs4_proc_open(struct nfs4_opendata *data);
static int nfs4_do_fsinfo(struct nfs_server *, struct nfs_fh *, struct nfs_fsinfo *);
static int nfs4_async_handle_error(struct rpc_task *, const struct nfs_server *);
static int nfs4_handle_exception(const struct nfs_server *server, int errorcode, struct nfs4_exception *exception);
static int nfs4_wait_clnt_recover(struct rpc_clnt *clnt, struct nfs_client *clp);
static int _nfs4_proc_lookup(struct inode *dir, const struct qstr *name, struct nfs_fh *fhandle, struct nfs_fattr *fattr);
static int _nfs4_proc_getattr(struct nfs_server *server, struct nfs_fh *fhandle, struct nfs_fattr *fattr);

/* Prevent leaks of NFSv4 errors into userland */
int nfs4_map_errors(int err)
{
	if (err < -1000) {
		dprintk("%s could not handle NFSv4 error %d\n",
				__FUNCTION__, -err);
		return -EIO;
	}
	return err;
}

/*
 * This is our standard bitmap for GETATTR requests.
 */
const u32 nfs4_fattr_bitmap[2] = {
	FATTR4_WORD0_TYPE
	| FATTR4_WORD0_CHANGE
	| FATTR4_WORD0_SIZE
	| FATTR4_WORD0_FSID
	| FATTR4_WORD0_FILEID,
	FATTR4_WORD1_MODE
	| FATTR4_WORD1_NUMLINKS
	| FATTR4_WORD1_OWNER
	| FATTR4_WORD1_OWNER_GROUP
	| FATTR4_WORD1_RAWDEV
	| FATTR4_WORD1_SPACE_USED
	| FATTR4_WORD1_TIME_ACCESS
	| FATTR4_WORD1_TIME_METADATA
	| FATTR4_WORD1_TIME_MODIFY
};

const u32 nfs4_statfs_bitmap[2] = {
	FATTR4_WORD0_FILES_AVAIL
	| FATTR4_WORD0_FILES_FREE
	| FATTR4_WORD0_FILES_TOTAL,
	FATTR4_WORD1_SPACE_AVAIL
	| FATTR4_WORD1_SPACE_FREE
	| FATTR4_WORD1_SPACE_TOTAL
};

const u32 nfs4_pathconf_bitmap[2] = {
	FATTR4_WORD0_MAXLINK
	| FATTR4_WORD0_MAXNAME,
	0
};

const u32 nfs4_fsinfo_bitmap[2] = { FATTR4_WORD0_MAXFILESIZE
			| FATTR4_WORD0_MAXREAD
			| FATTR4_WORD0_MAXWRITE
			| FATTR4_WORD0_LEASE_TIME,
			0
};

const u32 nfs4_fs_locations_bitmap[2] = {
	FATTR4_WORD0_TYPE
	| FATTR4_WORD0_CHANGE
	| FATTR4_WORD0_SIZE
	| FATTR4_WORD0_FSID
	| FATTR4_WORD0_FILEID
	| FATTR4_WORD0_FS_LOCATIONS,
	FATTR4_WORD1_MODE
	| FATTR4_WORD1_NUMLINKS
	| FATTR4_WORD1_OWNER
	| FATTR4_WORD1_OWNER_GROUP
	| FATTR4_WORD1_RAWDEV
	| FATTR4_WORD1_SPACE_USED
	| FATTR4_WORD1_TIME_ACCESS
	| FATTR4_WORD1_TIME_METADATA
	| FATTR4_WORD1_TIME_MODIFY
	| FATTR4_WORD1_MOUNTED_ON_FILEID
};

static void nfs4_setup_readdir(u64 cookie, __be32 *verifier, struct dentry *dentry,
		struct nfs4_readdir_arg *readdir)
{
	__be32 *start, *p;

	BUG_ON(readdir->count < 80);
	if (cookie > 2) {
		readdir->cookie = cookie;
		memcpy(&readdir->verifier, verifier, sizeof(readdir->verifier));
		return;
	}

	readdir->cookie = 0;
	memset(&readdir->verifier, 0, sizeof(readdir->verifier));
	if (cookie == 2)
		return;
	
	/*
	 * NFSv4 servers do not return entries for '.' and '..'
	 * Therefore, we fake these entries here.  We let '.'
	 * have cookie 0 and '..' have cookie 1.  Note that
	 * when talking to the server, we always send cookie 0
	 * instead of 1 or 2.
	 */
	start = p = kmap_atomic(*readdir->pages, KM_USER0);
	
	if (cookie == 0) {
		*p++ = xdr_one;                                  /* next */
		*p++ = xdr_zero;                   /* cookie, first word */
		*p++ = xdr_one;                   /* cookie, second word */
		*p++ = xdr_one;                             /* entry len */
		memcpy(p, ".\0\0\0", 4);                        /* entry */
		p++;
		*p++ = xdr_one;                         /* bitmap length */
		*p++ = htonl(FATTR4_WORD0_FILEID);             /* bitmap */
		*p++ = htonl(8);              /* attribute buffer length */
		p = xdr_encode_hyper(p, NFS_FILEID(dentry->d_inode));
	}
	
	*p++ = xdr_one;                                  /* next */
	*p++ = xdr_zero;                   /* cookie, first word */
	*p++ = xdr_two;                   /* cookie, second word */
	*p++ = xdr_two;                             /* entry len */
	memcpy(p, "..\0\0", 4);                         /* entry */
	p++;
	*p++ = xdr_one;                         /* bitmap length */
	*p++ = htonl(FATTR4_WORD0_FILEID);             /* bitmap */
	*p++ = htonl(8);              /* attribute buffer length */
	p = xdr_encode_hyper(p, NFS_FILEID(dentry->d_parent->d_inode));

	readdir->pgbase = (char *)p - (char *)start;
	readdir->count -= readdir->pgbase;
	kunmap_atomic(start, KM_USER0);
}

static void renew_lease(const struct nfs_server *server, unsigned long timestamp)
{
	struct nfs_client *clp = server->nfs_client;
	spin_lock(&clp->cl_lock);
	if (time_before(clp->cl_last_renewal,timestamp))
		clp->cl_last_renewal = timestamp;
	spin_unlock(&clp->cl_lock);
}

static void update_changeattr(struct inode *dir, struct nfs4_change_info *cinfo)
{
	struct nfs_inode *nfsi = NFS_I(dir);

	spin_lock(&dir->i_lock);
	nfsi->cache_validity |= NFS_INO_INVALID_ATTR|NFS_INO_REVAL_PAGECACHE|NFS_INO_INVALID_DATA;
	if (!cinfo->atomic || cinfo->before != nfsi->change_attr)
		nfsi->cache_change_attribute = jiffies;
	nfsi->change_attr = cinfo->after;
	spin_unlock(&dir->i_lock);
}

struct nfs4_opendata {
	struct kref kref;
	struct nfs_openargs o_arg;
	struct nfs_openres o_res;
	struct nfs_open_confirmargs c_arg;
	struct nfs_open_confirmres c_res;
	struct nfs_fattr f_attr;
	struct nfs_fattr dir_attr;
	struct path path;
	struct dentry *dir;
	struct nfs4_state_owner *owner;
	struct nfs4_state *state;
	struct iattr attrs;
	unsigned long timestamp;
	unsigned int rpc_done : 1;
	int rpc_status;
	int cancelled;
};


static void nfs4_init_opendata_res(struct nfs4_opendata *p)
{
	p->o_res.f_attr = &p->f_attr;
	p->o_res.dir_attr = &p->dir_attr;
	p->o_res.server = p->o_arg.server;
	nfs_fattr_init(&p->f_attr);
	nfs_fattr_init(&p->dir_attr);
}

static struct nfs4_opendata *nfs4_opendata_alloc(struct path *path,
		struct nfs4_state_owner *sp, int flags,
		const struct iattr *attrs)
{
	struct dentry *parent = dget_parent(path->dentry);
	struct inode *dir = parent->d_inode;
	struct nfs_server *server = NFS_SERVER(dir);
	struct nfs4_opendata *p;

	p = kzalloc(sizeof(*p), GFP_KERNEL);
	if (p == NULL)
		goto err;
	p->o_arg.seqid = nfs_alloc_seqid(&sp->so_seqid);
	if (p->o_arg.seqid == NULL)
		goto err_free;
	p->path.mnt = mntget(path->mnt);
	p->path.dentry = dget(path->dentry);
	p->dir = parent;
	p->owner = sp;
	atomic_inc(&sp->so_count);
	p->o_arg.fh = NFS_FH(dir);
	p->o_arg.open_flags = flags,
	p->o_arg.clientid = server->nfs_client->cl_clientid;
	p->o_arg.id = sp->so_owner_id.id;
	p->o_arg.name = &p->path.dentry->d_name;
	p->o_arg.server = server;
	p->o_arg.bitmask = server->attr_bitmask;
	p->o_arg.claim = NFS4_OPEN_CLAIM_NULL;
	if (flags & O_EXCL) {
		u32 *s = (u32 *) p->o_arg.u.verifier.data;
		s[0] = jiffies;
		s[1] = current->pid;
	} else if (flags & O_CREAT) {
		p->o_arg.u.attrs = &p->attrs;
		memcpy(&p->attrs, attrs, sizeof(p->attrs));
	}
	p->c_arg.fh = &p->o_res.fh;
	p->c_arg.stateid = &p->o_res.stateid;
	p->c_arg.seqid = p->o_arg.seqid;
	nfs4_init_opendata_res(p);
	kref_init(&p->kref);
	return p;
err_free:
	kfree(p);
err:
	dput(parent);
	return NULL;
}

static void nfs4_opendata_free(struct kref *kref)
{
	struct nfs4_opendata *p = container_of(kref,
			struct nfs4_opendata, kref);

	nfs_free_seqid(p->o_arg.seqid);
	if (p->state != NULL)
		nfs4_put_open_state(p->state);
	nfs4_put_state_owner(p->owner);
	dput(p->dir);
	dput(p->path.dentry);
	mntput(p->path.mnt);
	kfree(p);
}

static void nfs4_opendata_put(struct nfs4_opendata *p)
{
	if (p != NULL)
		kref_put(&p->kref, nfs4_opendata_free);
}

static int nfs4_wait_for_completion_rpc_task(struct rpc_task *task)
{
	sigset_t oldset;
	int ret;

	rpc_clnt_sigmask(task->tk_client, &oldset);
	ret = rpc_wait_for_completion_task(task);
	rpc_clnt_sigunmask(task->tk_client, &oldset);
	return ret;
}

static int can_open_cached(struct nfs4_state *state, int mode)
{
	int ret = 0;
	switch (mode & (FMODE_READ|FMODE_WRITE|O_EXCL)) {
		case FMODE_READ:
			ret |= test_bit(NFS_O_RDONLY_STATE, &state->flags) != 0;
			break;
		case FMODE_WRITE:
			ret |= test_bit(NFS_O_WRONLY_STATE, &state->flags) != 0;
			break;
		case FMODE_READ|FMODE_WRITE:
			ret |= test_bit(NFS_O_RDWR_STATE, &state->flags) != 0;
	}
	return ret;
}

static int can_open_delegated(struct nfs_delegation *delegation, mode_t open_flags)
{
	if ((delegation->type & open_flags) != open_flags)
		return 0;
	if (delegation->flags & NFS_DELEGATION_NEED_RECLAIM)
		return 0;
	return 1;
}

static void update_open_stateflags(struct nfs4_state *state, mode_t open_flags)
{
	switch (open_flags) {
		case FMODE_WRITE:
			state->n_wronly++;
			break;
		case FMODE_READ:
			state->n_rdonly++;
			break;
		case FMODE_READ|FMODE_WRITE:
			state->n_rdwr++;
	}
	nfs4_state_set_mode_locked(state, state->state | open_flags);
}

static void nfs_set_open_stateid_locked(struct nfs4_state *state, nfs4_stateid *stateid, int open_flags)
{
	if (test_bit(NFS_DELEGATED_STATE, &state->flags) == 0)
		memcpy(state->stateid.data, stateid->data, sizeof(state->stateid.data));
	memcpy(state->open_stateid.data, stateid->data, sizeof(state->open_stateid.data));
	switch (open_flags) {
		case FMODE_READ:
			set_bit(NFS_O_RDONLY_STATE, &state->flags);
			break;
		case FMODE_WRITE:
			set_bit(NFS_O_WRONLY_STATE, &state->flags);
			break;
		case FMODE_READ|FMODE_WRITE:
			set_bit(NFS_O_RDWR_STATE, &state->flags);
	}
}

static void nfs_set_open_stateid(struct nfs4_state *state, nfs4_stateid *stateid, int open_flags)
{
	write_seqlock(&state->seqlock);
	nfs_set_open_stateid_locked(state, stateid, open_flags);
	write_sequnlock(&state->seqlock);
}

static void update_open_stateid(struct nfs4_state *state, nfs4_stateid *open_stateid, nfs4_stateid *deleg_stateid, int open_flags)
{
	open_flags &= (FMODE_READ|FMODE_WRITE);
	/*
	 * Protect the call to nfs4_state_set_mode_locked and
	 * serialise the stateid update
	 */
	write_seqlock(&state->seqlock);
	if (deleg_stateid != NULL) {
		memcpy(state->stateid.data, deleg_stateid->data, sizeof(state->stateid.data));
		set_bit(NFS_DELEGATED_STATE, &state->flags);
	}
	if (open_stateid != NULL)
		nfs_set_open_stateid_locked(state, open_stateid, open_flags);
	write_sequnlock(&state->seqlock);
	spin_lock(&state->owner->so_lock);
	update_open_stateflags(state, open_flags);
	spin_unlock(&state->owner->so_lock);
}

static void nfs4_return_incompatible_delegation(struct inode *inode, mode_t open_flags)
{
	struct nfs_delegation *delegation;

	rcu_read_lock();
	delegation = rcu_dereference(NFS_I(inode)->delegation);
	if (delegation == NULL || (delegation->type & open_flags) == open_flags) {
		rcu_read_unlock();
		return;
	}
	rcu_read_unlock();
	nfs_inode_return_delegation(inode);
}

static struct nfs4_state *nfs4_try_open_cached(struct nfs4_opendata *opendata)
{
	struct nfs4_state *state = opendata->state;
	struct nfs_inode *nfsi = NFS_I(state->inode);
	struct nfs_delegation *delegation;
	int open_mode = opendata->o_arg.open_flags & (FMODE_READ|FMODE_WRITE|O_EXCL);
	nfs4_stateid stateid;
	int ret = -EAGAIN;

	rcu_read_lock();
	delegation = rcu_dereference(nfsi->delegation);
	for (;;) {
		if (can_open_cached(state, open_mode)) {
			spin_lock(&state->owner->so_lock);
			if (can_open_cached(state, open_mode)) {
				update_open_stateflags(state, open_mode);
				spin_unlock(&state->owner->so_lock);
				rcu_read_unlock();
				goto out_return_state;
			}
			spin_unlock(&state->owner->so_lock);
		}
		if (delegation == NULL)
			break;
		if (!can_open_delegated(delegation, open_mode))
			break;
		/* Save the delegation */
		memcpy(stateid.data, delegation->stateid.data, sizeof(stateid.data));
		rcu_read_unlock();
		lock_kernel();
		ret = nfs_may_open(state->inode, state->owner->so_cred, open_mode);
		unlock_kernel();
		if (ret != 0)
			goto out;
		ret = -EAGAIN;
		rcu_read_lock();
		delegation = rcu_dereference(nfsi->delegation);
		/* If no delegation, try a cached open */
		if (delegation == NULL)
			continue;
		/* Is the delegation still valid? */
		if (memcmp(stateid.data, delegation->stateid.data, sizeof(stateid.data)) != 0)