nfs4proc.c

/*
 *  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 <kmsmith@umich.edu>
 *  Andy Adamson   <andros@umich.edu>
 *
 *  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/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/namei.h>
#include <linux/mount.h>
#include <linux/module.h>
#include <linux/sunrpc/bc_xprt.h>

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

#define NFSDBG_FACILITY		NFSDBG_PROC

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

#define NFS4_MAX_LOOP_ON_RECOVER (10)

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 *, struct nfs4_state *);
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 */
static int nfs4_map_errors(int err)
{
	if (err >= -1000)
		return err;
	switch (err) {
	case -NFS4ERR_RESOURCE:
		return -EREMOTEIO;
	default:
		dprintk("%s could not handle NFSv4 error %d\n",
				__func__, -err);
		break;
	}
	return -EIO;
}

/*
 * 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 int nfs4_wait_clnt_recover(struct nfs_client *clp)
{
	int res;

	might_sleep();

	res = wait_on_bit(&clp->cl_state, NFS4CLNT_MANAGER_RUNNING,
			nfs_wait_bit_killable, TASK_KILLABLE);
	return res;
}

static int nfs4_delay(struct rpc_clnt *clnt, long *timeout)
{
	int res = 0;

	might_sleep();

	if (*timeout <= 0)
		*timeout = NFS4_POLL_RETRY_MIN;
	if (*timeout > NFS4_POLL_RETRY_MAX)
		*timeout = NFS4_POLL_RETRY_MAX;
	schedule_timeout_killable(*timeout);
	if (fatal_signal_pending(current))
		res = -ERESTARTSYS;
	*timeout <<= 1;
	return res;
}

/* This is the error handling routine for processes that are allowed
 * to sleep.
 */
static int nfs4_handle_exception(const struct nfs_server *server, int errorcode, struct nfs4_exception *exception)
{
	struct nfs_client *clp = server->nfs_client;
	struct nfs4_state *state = exception->state;
	int ret = errorcode;

	exception->retry = 0;
	switch(errorcode) {
		case 0:
			return 0;
		case -NFS4ERR_ADMIN_REVOKED:
		case -NFS4ERR_BAD_STATEID:
		case -NFS4ERR_OPENMODE:
			if (state == NULL)
				break;
			nfs4_state_mark_reclaim_nograce(clp, state);
		case -NFS4ERR_STALE_CLIENTID:
		case -NFS4ERR_STALE_STATEID:
		case -NFS4ERR_EXPIRED:
			nfs4_schedule_state_recovery(clp);
			ret = nfs4_wait_clnt_recover(clp);
			if (ret == 0)
				exception->retry = 1;
#if !defined(CONFIG_NFS_V4_1)
			break;
#else /* !defined(CONFIG_NFS_V4_1) */
			if (!nfs4_has_session(server->nfs_client))
				break;
			/* FALLTHROUGH */
		case -NFS4ERR_BADSESSION:
		case -NFS4ERR_BADSLOT:
		case -NFS4ERR_BAD_HIGH_SLOT:
		case -NFS4ERR_CONN_NOT_BOUND_TO_SESSION:
		case -NFS4ERR_DEADSESSION:
		case -NFS4ERR_SEQ_FALSE_RETRY:
		case -NFS4ERR_SEQ_MISORDERED:
			dprintk("%s ERROR: %d Reset session\n", __func__,
				errorcode);
			nfs4_schedule_state_recovery(clp);
			exception->retry = 1;
			break;
#endif /* !defined(CONFIG_NFS_V4_1) */
		case -NFS4ERR_FILE_OPEN:
			if (exception->timeout > HZ) {
				/* We have retried a decent amount, time to
				 * fail
				 */
				ret = -EBUSY;
				break;
			}
		case -NFS4ERR_GRACE:
		case -NFS4ERR_DELAY:
			ret = nfs4_delay(server->client, &exception->timeout);
			if (ret != 0)
				break;
		case -NFS4ERR_OLD_STATEID:
			exception->retry = 1;
	}
	/* We failed to handle the error */
	return nfs4_map_errors(ret);
}


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);
}

#if defined(CONFIG_NFS_V4_1)

/*
 * nfs4_free_slot - free a slot and efficiently update slot table.
 *
 * freeing a slot is trivially done by clearing its respective bit
 * in the bitmap.
 * If the freed slotid equals highest_used_slotid we want to update it
 * so that the server would be able to size down the slot table if needed,
 * otherwise we know that the highest_used_slotid is still in use.
 * When updating highest_used_slotid there may be "holes" in the bitmap
 * so we need to scan down from highest_used_slotid to 0 looking for the now
 * highest slotid in use.
 * If none found, highest_used_slotid is set to -1.
 *
 * Must be called while holding tbl->slot_tbl_lock
 */
static void
nfs4_free_slot(struct nfs4_slot_table *tbl, u8 free_slotid)
{
	int slotid = free_slotid;

	/* clear used bit in bitmap */
	__clear_bit(slotid, tbl->used_slots);

	/* update highest_used_slotid when it is freed */
	if (slotid == tbl->highest_used_slotid) {
		slotid = find_last_bit(tbl->used_slots, tbl->max_slots);
		if (slotid < tbl->max_slots)
			tbl->highest_used_slotid = slotid;
		else
			tbl->highest_used_slotid = -1;
	}
	dprintk("%s: free_slotid %u highest_used_slotid %d\n", __func__,
		free_slotid, tbl->highest_used_slotid);
}

static void nfs41_sequence_free_slot(const struct nfs_client *clp,
			      struct nfs4_sequence_res *res)
{
	struct nfs4_slot_table *tbl;

	tbl = &clp->cl_session->fc_slot_table;
	if (res->sr_slotid == NFS4_MAX_SLOT_TABLE) {
		/* just wake up the next guy waiting since
		 * we may have not consumed a slot after all */
		dprintk("%s: No slot\n", __func__);
		return;
	}

	spin_lock(&tbl->slot_tbl_lock);
	nfs4_free_slot(tbl, res->sr_slotid);

	/* Signal state manager thread if session is drained */
	if (test_bit(NFS4CLNT_SESSION_DRAINING, &clp->cl_state)) {
		if (tbl->highest_used_slotid == -1) {
			dprintk("%s COMPLETE: Session Drained\n", __func__);
			complete(&clp->cl_session->complete);
		}
	} else
		rpc_wake_up_next(&tbl->slot_tbl_waitq);
	spin_unlock(&tbl->slot_tbl_lock);
	res->sr_slotid = NFS4_MAX_SLOT_TABLE;
}

static void nfs41_sequence_done(struct nfs_client *clp,
				struct nfs4_sequence_res *res,
				int rpc_status)
{
	unsigned long timestamp;
	struct nfs4_slot_table *tbl;
	struct nfs4_slot *slot;

	/*
	 * sr_status remains 1 if an RPC level error occurred. The server
	 * may or may not have processed the sequence operation..
	 * Proceed as if the server received and processed the sequence
	 * operation.
	 */
	if (res->sr_status == 1)
		res->sr_status = NFS_OK;

	/* -ERESTARTSYS can result in skipping nfs41_sequence_setup */
	if (res->sr_slotid == NFS4_MAX_SLOT_TABLE)
		goto out;

	/* Check the SEQUENCE operation status */
	if (res->sr_status == 0) {
		tbl = &clp->cl_session->fc_slot_table;
		slot = tbl->slots + res->sr_slotid;
		/* Update the slot's sequence and clientid lease timer */
		++slot->seq_nr;
		timestamp = res->sr_renewal_time;
		spin_lock(&clp->cl_lock);
		if (time_before(clp->cl_last_renewal, timestamp))
			clp->cl_last_renewal = timestamp;
		spin_unlock(&clp->cl_lock);
		/* Check sequence flags */
		nfs41_handle_sequence_flag_errors(clp, res->sr_status_flags);
	}
out:
	/* The session may be reset by one of the error handlers. */
	dprintk("%s: Error %d free the slot \n", __func__, res->sr_status);
	nfs41_sequence_free_slot(clp, res);
}

/*
 * nfs4_find_slot - efficiently look for a free slot
 *
 * nfs4_find_slot looks for an unset bit in the used_slots bitmap.
 * If found, we mark the slot as used, update the highest_used_slotid,
 * and respectively set up the sequence operation args.
 * The slot number is returned if found, or NFS4_MAX_SLOT_TABLE otherwise.
 *
 * Note: must be called with under the slot_tbl_lock.
 */
static u8
nfs4_find_slot(struct nfs4_slot_table *tbl, struct rpc_task *task)
{
	int slotid;
	u8 ret_id = NFS4_MAX_SLOT_TABLE;
	BUILD_BUG_ON((u8)NFS4_MAX_SLOT_TABLE != (int)NFS4_MAX_SLOT_TABLE);

	dprintk("--> %s used_slots=%04lx highest_used=%d max_slots=%d\n",
		__func__, tbl->used_slots[0], tbl->highest_used_slotid,
		tbl->max_slots);
	slotid = find_first_zero_bit(tbl->used_slots, tbl->max_slots);
	if (slotid >= tbl->max_slots)
		goto out;
	__set_bit(slotid, tbl->used_slots);
	if (slotid > tbl->highest_used_slotid)
		tbl->highest_used_slotid = slotid;
	ret_id = slotid;
out:
	dprintk("<-- %s used_slots=%04lx highest_used=%d slotid=%d \n",
		__func__, tbl->used_slots[0], tbl->highest_used_slotid, ret_id);
	return ret_id;
}

static int nfs41_setup_sequence(struct nfs4_session *session,
				struct nfs4_sequence_args *args,
				struct nfs4_sequence_res *res,
				int cache_reply,
				struct rpc_task *task)
{
	struct nfs4_slot *slot;
	struct nfs4_slot_table *tbl;
	u8 slotid;

	dprintk("--> %s\n", __func__);
	/* slot already allocated? */
	if (res->sr_slotid != NFS4_MAX_SLOT_TABLE)
		return 0;

	memset(res, 0, sizeof(*res));
	res->sr_slotid = NFS4_MAX_SLOT_TABLE;
	tbl = &session->fc_slot_table;

	spin_lock(&tbl->slot_tbl_lock);
	if (test_bit(NFS4CLNT_SESSION_DRAINING, &session->clp->cl_state)) {
		/*
		 * The state manager will wait until the slot table is empty.
		 * Schedule the reset thread
		 */
		rpc_sleep_on(&tbl->slot_tbl_waitq, task, NULL);
		spin_unlock(&tbl->slot_tbl_lock);
		dprintk("%s Schedule Session Reset\n", __func__);
		return -EAGAIN;
	}

	slotid = nfs4_find_slot(tbl, task);
	if (slotid == NFS4_MAX_SLOT_TABLE) {
		rpc_sleep_on(&tbl->slot_tbl_waitq, task, NULL);
		spin_unlock(&tbl->slot_tbl_lock);
		dprintk("<-- %s: no free slots\n", __func__);
		return -EAGAIN;
	}
	spin_unlock(&tbl->slot_tbl_lock);

	slot = tbl->slots + slotid;
	args->sa_session = session;
	args->sa_slotid = slotid;
	args->sa_cache_this = cache_reply;

	dprintk("<-- %s slotid=%d seqid=%d\n", __func__, slotid, slot->seq_nr);

	res->sr_session = session;
	res->sr_slotid = slotid;
	res->sr_renewal_time = jiffies;
	/*
	 * sr_status is only set in decode_sequence, and so will remain
	 * set to 1 if an rpc level failure occurs.
	 */
	res->sr_status = 1;
	return 0;
}

int nfs4_setup_sequence(struct nfs_client *clp,
			struct nfs4_sequence_args *args,
			struct nfs4_sequence_res *res,
			int cache_reply,
			struct rpc_task *task)
{
	int ret = 0;

	dprintk("--> %s clp %p session %p sr_slotid %d\n",
		__func__, clp, clp->cl_session, res->sr_slotid);

	if (!nfs4_has_session(clp))
		goto out;
	ret = nfs41_setup_sequence(clp->cl_session, args, res, cache_reply,
				   task);
	if (ret && ret != -EAGAIN) {
		/* terminate rpc task */
		task->tk_status = ret;
		task->tk_action = NULL;
	}
out:
	dprintk("<-- %s status=%d\n", __func__, ret);
	return ret;
}

struct nfs41_call_sync_data {
	struct nfs_client *clp;
	struct nfs4_sequence_args *seq_args;
	struct nfs4_sequence_res *seq_res;
	int cache_reply;
};

static void nfs41_call_sync_prepare(struct rpc_task *task, void *calldata)
{
	struct nfs41_call_sync_data *data = calldata;

	dprintk("--> %s data->clp->cl_session %p\n", __func__,
		data->clp->cl_session);
	if (nfs4_setup_sequence(data->clp, data->seq_args,
				data->seq_res, data->cache_reply, task))
		return;
	rpc_call_start(task);
}

static void nfs41_call_sync_done(struct rpc_task *task, void *calldata)
{
	struct nfs41_call_sync_data *data = calldata;

	nfs41_sequence_done(data->clp, data->seq_res, task->tk_status);
}

struct rpc_call_ops nfs41_call_sync_ops = {
	.rpc_call_prepare = nfs41_call_sync_prepare,
	.rpc_call_done = nfs41_call_sync_done,
};

static int nfs4_call_sync_sequence(struct nfs_client *clp,
				   struct rpc_clnt *clnt,
				   struct rpc_message *msg,
				   struct nfs4_sequence_args *args,
				   struct nfs4_sequence_res *res,
				   int cache_reply)
{
	int ret;
	struct rpc_task *task;
	struct nfs41_call_sync_data data = {
		.clp = clp,
		.seq_args = args,
		.seq_res = res,
		.cache_reply = cache_reply,
	};
	struct rpc_task_setup task_setup = {
		.rpc_client = clnt,
		.rpc_message = msg,
		.callback_ops = &nfs41_call_sync_ops,
		.callback_data = &data
	};

	res->sr_slotid = NFS4_MAX_SLOT_TABLE;
	task = rpc_run_task(&task_setup);
	if (IS_ERR(task))
		ret = PTR_ERR(task);
	else {
		ret = task->tk_status;
		rpc_put_task(task);
	}
	return ret;
}

int _nfs4_call_sync_session(struct nfs_server *server,
			    struct rpc_message *msg,
			    struct nfs4_sequence_args *args,
			    struct nfs4_sequence_res *res,
			    int cache_reply)
{
	return nfs4_call_sync_sequence(server->nfs_client, server->client,
				       msg, args, res, cache_reply);
}

#endif /* CONFIG_NFS_V4_1 */

int _nfs4_call_sync(struct nfs_server *server,
		    struct rpc_message *msg,
		    struct nfs4_sequence_args *args,
		    struct nfs4_sequence_res *res,
		    int cache_reply)
{
	args->sa_session = res->sr_session = NULL;
	return rpc_call_sync(server->client, msg, 0);
}

#define nfs4_call_sync(server, msg, args, res, cache_reply) \
	(server)->nfs_client->cl_call_sync((server), (msg), &(args)->seq_args, \
			&(res)->seq_res, (cache_reply))

static void nfs4_sequence_done(const struct nfs_server *server,
			       struct nfs4_sequence_res *res, int rpc_status)
{
#ifdef CONFIG_NFS_V4_1
	if (nfs4_has_session(server->nfs_client))
		nfs41_sequence_done(server->nfs_client, res, rpc_status);
#endif /* CONFIG_NFS_V4_1 */
}

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)
		nfs_force_lookup_revalidate(dir);
	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.seqid = p->o_arg.seqid;
	p->c_res.seqid = p->c_arg.seqid;
	p->o_res.server = p->o_arg.server;
	nfs_fattr_init(&p->f_attr);
	nfs_fattr_init(&p->dir_attr);
	p->o_res.seq_res.sr_slotid = NFS4_MAX_SLOT_TABLE;
}

static struct nfs4_opendata *nfs4_opendata_alloc(struct path *path,
		struct nfs4_state_owner *sp, fmode_t fmode, 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.fmode = fmode & (FMODE_READ|FMODE_WRITE);
	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) {
		if (nfs4_has_persistent_session(server->nfs_client)) {
			/* GUARDED */
			p->o_arg.u.attrs = &p->attrs;
			memcpy(&p->attrs, attrs, sizeof(p->attrs));
		} else { /* EXCLUSIVE4_1 */
			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);
	path_put(&p->path);
	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)
{
	int ret;

	ret = rpc_wait_for_completion_task(task);
	return ret;
}

static int can_open_cached(struct nfs4_state *state, fmode_t mode, int open_mode)
{
	int ret = 0;

	if (open_mode & O_EXCL)
		goto out;
	switch (mode & (FMODE_READ|FMODE_WRITE)) {
		case FMODE_READ:
			ret |= test_bit(NFS_O_RDONLY_STATE, &state->flags) != 0
				&& state->n_rdonly != 0;
			break;
		case FMODE_WRITE:
			ret |= test_bit(NFS_O_WRONLY_STATE, &state->flags) != 0
				&& state->n_wronly != 0;
			break;
		case FMODE_READ|FMODE_WRITE:
			ret |= test_bit(NFS_O_RDWR_STATE, &state->flags) != 0
				&& state->n_rdwr != 0;
	}
out:
	return ret;
}

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

static void update_open_stateflags(struct nfs4_state *state, fmode_t fmode)
{
	switch (fmode) {
		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 | fmode);
}

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

static void __update_open_stateid(struct nfs4_state *state, nfs4_stateid *open_stateid, const nfs4_stateid *deleg_stateid, fmode_t fmode)
{
	/*
	 * 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, fmode);
	write_sequnlock(&state->seqlock);
	spin_lock(&state->owner->so_lock);
	update_open_stateflags(state, fmode);
	spin_unlock(&state->owner->so_lock);
}

static int update_open_stateid(struct nfs4_state *state, nfs4_stateid *open_stateid, nfs4_stateid *delegation, fmode_t fmode)
{
	struct nfs_inode *nfsi = NFS_I(state->inode);
	struct nfs_delegation *deleg_cur;
	int ret = 0;

	fmode &= (FMODE_READ|FMODE_WRITE);

	rcu_read_lock();
	deleg_cur = rcu_dereference(nfsi->delegation);
	if (deleg_cur == NULL)
		goto no_delegation;

	spin_lock(&deleg_cur->lock);
	if (nfsi->delegation != deleg_cur ||
	    (deleg_cur->type & fmode) != fmode)
		goto no_delegation_unlock;

	if (delegation == NULL)
		delegation = &deleg_cur->stateid;
	else if (memcmp(deleg_cur->stateid.data, delegation->data, NFS4_STATEID_SIZE) != 0)
		goto no_delegation_unlock;

	nfs_mark_delegation_referenced(deleg_cur);
	__update_open_stateid(state, open_stateid, &deleg_cur->stateid, fmode);
	ret = 1;
no_delegation_unlock:
	spin_unlock(&deleg_cur->lock);
no_delegation:
	rcu_read_unlock();

	if (!ret && open_stateid != NULL) {
		__update_open_stateid(state, open_stateid, NULL, fmode);
		ret = 1;
	}

	return ret;
}


static void nfs4_return_incompatible_delegation(struct inode *inode, fmode_t fmode)
{
	struct nfs_delegation *delegation;

	rcu_read_lock();
	delegation = rcu_dereference(NFS_I(inode)->delegation);
	if (delegation == NULL || (delegation->type & fmode) == fmode) {
		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 & O_EXCL;
	fmode_t fmode = opendata->o_arg.fmode;
	nfs4_stateid stateid;
	int ret = -EAGAIN;

	for (;;) {
		if (can_open_cached(state, fmode, open_mode)) {
			spin_lock(&state->owner->so_lock);
			if (can_open_cached(state, fmode, open_mode)) {
				update_open_stateflags(state, fmode);
				spin_unlock(&state->owner->so_lock);
				goto out_return_state;
			}
			spin_unlock(&state->owner->so_lock);
		}
		rcu_read_lock();
		delegation = rcu_dereference(nfsi->delegation);
		if (delegation == NULL ||
		    !can_open_delegated(delegation, fmode)) {
			rcu_read_unlock();
			break;
		}
		/* Save the delegation */
		memcpy(stateid.data, delegation->stateid.data, sizeof(stateid.data));
		rcu_read_unlock();
		ret = nfs_may_open(state->inode, state->owner->so_cred, open_mode);
		if (ret != 0)
			goto out;
		ret = -EAGAIN;

		/* Try to update the stateid using the delegation */
		if (update_open_stateid(state, NULL, &stateid, fmode))
			goto out_return_state;
	}
out:
	return ERR_PTR(ret);
out_return_state:
	atomic_inc(&state->count);
	return state;
}

static struct nfs4_state *nfs4_opendata_to_nfs4_state(struct nfs4_opendata *data)
{
	struct inode *inode;
	struct nfs4_state *state = NULL;
	struct nfs_delegation *delegation;
	int ret;

	if (!data->rpc_done) {
		state = nfs4_try_open_cached(data);
		goto out;
	}

	ret = -EAGAIN;
	if (!(data->f_attr.valid & NFS_ATTR_FATTR))
		goto err;
	inode = nfs_fhget(data->dir->d_sb, &data->o_res.fh, &data->f_attr);
	ret = PTR_ERR(inode);
	if (IS_ERR(inode))
		goto err;
	ret = -ENOMEM;
	state = nfs4_get_open_state(inode, data->owner);
	if (state == NULL)
		goto err_put_inode;
	if (data->o_res.delegation_type != 0) {
		int delegation_flags = 0;

		rcu_read_lock();
		delegation = rcu_dereference(NFS_I(inode)->delegation);
		if (delegation)
			delegation_flags = delegation->flags;
		rcu_read_unlock();
		if ((delegation_flags & 1UL<<NFS_DELEGATION_NEED_RECLAIM) == 0)
			nfs_inode_set_delegation(state->inode,
					data->owner->so_cred,
					&data->o_res);
		else
			nfs_inode_reclaim_delegation(state->inode,
					data->owner->so_cred,
					&data->o_res);
	}

	update_open_stateid(state, &data->o_res.stateid, NULL,
			data->o_arg.fmode);
	iput(inode);
out:
	return state;
err_put_inode:
	iput(inode);
err:
	return ERR_PTR(ret);
}

static struct nfs_open_context *nfs4_state_find_open_context(struct nfs4_state *state)
{