nfs4proc.c 99.8 KB
Newer Older
L
Linus Torvalds 已提交
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49
/*
 *  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/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>
50
#include <linux/mount.h>
L
Linus Torvalds 已提交
51

52
#include "nfs4_fs.h"
L
Linus Torvalds 已提交
53
#include "delegation.h"
54
#include "iostat.h"
L
Linus Torvalds 已提交
55 56 57

#define NFSDBG_FACILITY		NFSDBG_PROC

58
#define NFS4_POLL_RETRY_MIN	(HZ/10)
L
Linus Torvalds 已提交
59 60
#define NFS4_POLL_RETRY_MAX	(15*HZ)

61
struct nfs4_opendata;
62
static int _nfs4_proc_open(struct nfs4_opendata *data);
L
Linus Torvalds 已提交
63
static int nfs4_do_fsinfo(struct nfs_server *, struct nfs_fh *, struct nfs_fsinfo *);
64 65
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);
66
static int nfs4_wait_clnt_recover(struct rpc_clnt *clnt, struct nfs_client *clp);
67 68
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);
L
Linus Torvalds 已提交
69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109

/* 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
};

110
const u32 nfs4_pathconf_bitmap[2] = {
L
Linus Torvalds 已提交
111 112 113 114 115 116 117 118 119 120 121 122
	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
};

123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141
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
};

A
Al Viro 已提交
142
static void nfs4_setup_readdir(u64 cookie, __be32 *verifier, struct dentry *dentry,
L
Linus Torvalds 已提交
143 144
		struct nfs4_readdir_arg *readdir)
{
145
	__be32 *start, *p;
L
Linus Torvalds 已提交
146 147 148

	BUG_ON(readdir->count < 80);
	if (cookie > 2) {
149
		readdir->cookie = cookie;
L
Linus Torvalds 已提交
150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165
		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.
	 */
166
	start = p = kmap_atomic(*readdir->pages, KM_USER0);
L
Linus Torvalds 已提交
167 168 169 170 171 172 173 174 175 176 177
	
	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
Peter Staubach 已提交
178
		p = xdr_encode_hyper(p, NFS_FILEID(dentry->d_inode));
L
Linus Torvalds 已提交
179 180 181 182 183 184 185 186 187 188 189
	}
	
	*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
Peter Staubach 已提交
190
	p = xdr_encode_hyper(p, NFS_FILEID(dentry->d_parent->d_inode));
L
Linus Torvalds 已提交
191 192 193 194 195 196

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

197
static void renew_lease(const struct nfs_server *server, unsigned long timestamp)
L
Linus Torvalds 已提交
198
{
199
	struct nfs_client *clp = server->nfs_client;
L
Linus Torvalds 已提交
200 201 202 203 204 205
	spin_lock(&clp->cl_lock);
	if (time_before(clp->cl_last_renewal,timestamp))
		clp->cl_last_renewal = timestamp;
	spin_unlock(&clp->cl_lock);
}

206
static void update_changeattr(struct inode *dir, struct nfs4_change_info *cinfo)
L
Linus Torvalds 已提交
207
{
208
	struct nfs_inode *nfsi = NFS_I(dir);
L
Linus Torvalds 已提交
209

210
	spin_lock(&dir->i_lock);
211 212
	nfsi->cache_validity |= NFS_INO_INVALID_ATTR|NFS_INO_REVAL_PAGECACHE|NFS_INO_INVALID_DATA;
	if (!cinfo->atomic || cinfo->before != nfsi->change_attr)
213
		nfs_force_lookup_revalidate(dir);
214
	nfsi->change_attr = cinfo->after;
215
	spin_unlock(&dir->i_lock);
L
Linus Torvalds 已提交
216 217
}

218
struct nfs4_opendata {
219
	struct kref kref;
220 221
	struct nfs_openargs o_arg;
	struct nfs_openres o_res;
222 223
	struct nfs_open_confirmargs c_arg;
	struct nfs_open_confirmres c_res;
224 225
	struct nfs_fattr f_attr;
	struct nfs_fattr dir_attr;
226
	struct path path;
227 228
	struct dentry *dir;
	struct nfs4_state_owner *owner;
229
	struct nfs4_state *state;
230
	struct iattr attrs;
231
	unsigned long timestamp;
232
	unsigned int rpc_done : 1;
233 234
	int rpc_status;
	int cancelled;
235 236
};

237 238 239 240 241 242 243 244 245 246

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);
}

247
static struct nfs4_opendata *nfs4_opendata_alloc(struct path *path,
248 249 250
		struct nfs4_state_owner *sp, int flags,
		const struct iattr *attrs)
{
251
	struct dentry *parent = dget_parent(path->dentry);
252 253 254 255 256 257 258 259 260 261
	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;
262 263
	p->path.mnt = mntget(path->mnt);
	p->path.dentry = dget(path->dentry);
264 265 266 267 268
	p->dir = parent;
	p->owner = sp;
	atomic_inc(&sp->so_count);
	p->o_arg.fh = NFS_FH(dir);
	p->o_arg.open_flags = flags,
269
	p->o_arg.clientid = server->nfs_client->cl_clientid;
270
	p->o_arg.id = sp->so_owner_id.id;
271
	p->o_arg.name = &p->path.dentry->d_name;
272 273 274 275 276 277 278 279 280 281 282
	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));
	}
283 284 285
	p->c_arg.fh = &p->o_res.fh;
	p->c_arg.stateid = &p->o_res.stateid;
	p->c_arg.seqid = p->o_arg.seqid;
286
	nfs4_init_opendata_res(p);
287
	kref_init(&p->kref);
288 289 290 291 292 293 294 295
	return p;
err_free:
	kfree(p);
err:
	dput(parent);
	return NULL;
}

296
static void nfs4_opendata_free(struct kref *kref)
297
{
298 299 300 301
	struct nfs4_opendata *p = container_of(kref,
			struct nfs4_opendata, kref);

	nfs_free_seqid(p->o_arg.seqid);
302 303
	if (p->state != NULL)
		nfs4_put_open_state(p->state);
304 305 306 307 308 309 310 311 312 313 314
	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);
315 316
}

317 318 319 320 321 322 323 324 325 326 327
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;
}

328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343
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;
}

344 345 346 347 348 349 350 351 352
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;
}

353
static void update_open_stateflags(struct nfs4_state *state, mode_t open_flags)
354 355 356 357 358 359 360 361 362 363 364
{
	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++;
	}
365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382
	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);
	}
383 384
}

385 386
static void nfs_set_open_stateid(struct nfs4_state *state, nfs4_stateid *stateid, int open_flags)
{
387
	write_seqlock(&state->seqlock);
388
	nfs_set_open_stateid_locked(state, stateid, open_flags);
389
	write_sequnlock(&state->seqlock);
390 391 392
}

static void update_open_stateid(struct nfs4_state *state, nfs4_stateid *open_stateid, nfs4_stateid *deleg_stateid, int open_flags)
L
Linus Torvalds 已提交
393 394
{
	open_flags &= (FMODE_READ|FMODE_WRITE);
395 396 397 398 399
	/*
	 * Protect the call to nfs4_state_set_mode_locked and
	 * serialise the stateid update
	 */
	write_seqlock(&state->seqlock);
400 401 402 403 404 405
	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);
406 407
	write_sequnlock(&state->seqlock);
	spin_lock(&state->owner->so_lock);
408
	update_open_stateflags(state, open_flags);
409
	spin_unlock(&state->owner->so_lock);
L
Linus Torvalds 已提交
410 411
}

412 413 414 415 416 417 418 419 420 421 422 423 424 425
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);
}

426
static struct nfs4_state *nfs4_try_open_cached(struct nfs4_opendata *opendata)
427 428 429 430 431 432 433 434 435 436 437
{
	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 (;;) {
438 439 440 441 442 443 444 445 446 447 448 449
		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;
450 451 452 453 454 455
		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();
456
		ret = nfs_may_open(state->inode, state->owner->so_cred, open_mode);
457 458 459 460 461 462
		unlock_kernel();
		if (ret != 0)
			goto out;
		ret = -EAGAIN;
		rcu_read_lock();
		delegation = rcu_dereference(nfsi->delegation);
463
		/* If no delegation, try a cached open */
464
		if (delegation == NULL)
465
			continue;
466 467 468 469 470 471 472 473 474 475 476 477 478 479 480
		/* Is the delegation still valid? */
		if (memcmp(stateid.data, delegation->stateid.data, sizeof(stateid.data)) != 0)
			continue;
		rcu_read_unlock();
		update_open_stateid(state, NULL, &stateid, open_mode);
		goto out_return_state;
	}
	rcu_read_unlock();
out:
	return ERR_PTR(ret);
out_return_state:
	atomic_inc(&state->count);
	return state;
}

481 482 483 484
static struct nfs4_state *nfs4_opendata_to_nfs4_state(struct nfs4_opendata *data)
{
	struct inode *inode;
	struct nfs4_state *state = NULL;
485 486
	struct nfs_delegation *delegation;
	nfs4_stateid *deleg_stateid = NULL;
487
	int ret;
488

489
	if (!data->rpc_done) {
490
		state = nfs4_try_open_cached(data);
491 492 493
		goto out;
	}

494
	ret = -EAGAIN;
495
	if (!(data->f_attr.valid & NFS_ATTR_FATTR))
496
		goto err;
497
	inode = nfs_fhget(data->dir->d_sb, &data->o_res.fh, &data->f_attr);
498
	ret = PTR_ERR(inode);
499
	if (IS_ERR(inode))
500 501
		goto err;
	ret = -ENOMEM;
502 503
	state = nfs4_get_open_state(inode, data->owner);
	if (state == NULL)
504
		goto err_put_inode;
505 506 507
	if (data->o_res.delegation_type != 0) {
		int delegation_flags = 0;

508 509 510 511 512
		rcu_read_lock();
		delegation = rcu_dereference(NFS_I(inode)->delegation);
		if (delegation)
			delegation_flags = delegation->flags;
		rcu_read_unlock();
513 514 515 516 517 518 519 520 521
		if (!(delegation_flags & NFS_DELEGATION_NEED_RECLAIM))
			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);
	}
522 523 524 525 526 527
	rcu_read_lock();
	delegation = rcu_dereference(NFS_I(inode)->delegation);
	if (delegation != NULL)
		deleg_stateid = &delegation->stateid;
	update_open_stateid(state, &data->o_res.stateid, deleg_stateid, data->o_arg.open_flags);
	rcu_read_unlock();
528
	iput(inode);
529
out:
530
	return state;
531 532 533 534
err_put_inode:
	iput(inode);
err:
	return ERR_PTR(ret);
535 536
}

537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553
static struct nfs_open_context *nfs4_state_find_open_context(struct nfs4_state *state)
{
	struct nfs_inode *nfsi = NFS_I(state->inode);
	struct nfs_open_context *ctx;

	spin_lock(&state->inode->i_lock);
	list_for_each_entry(ctx, &nfsi->open_files, list) {
		if (ctx->state != state)
			continue;
		get_nfs_open_context(ctx);
		spin_unlock(&state->inode->i_lock);
		return ctx;
	}
	spin_unlock(&state->inode->i_lock);
	return ERR_PTR(-ENOENT);
}

T
Trond Myklebust 已提交
554 555 556 557 558 559 560 561 562 563 564 565
static struct nfs4_opendata *nfs4_open_recoverdata_alloc(struct nfs_open_context *ctx, struct nfs4_state *state)
{
	struct nfs4_opendata *opendata;

	opendata = nfs4_opendata_alloc(&ctx->path, state->owner, 0, NULL);
	if (opendata == NULL)
		return ERR_PTR(-ENOMEM);
	opendata->state = state;
	atomic_inc(&state->count);
	return opendata;
}

566
static int nfs4_open_recover_helper(struct nfs4_opendata *opendata, mode_t openflags, struct nfs4_state **res)
567
{
568
	struct nfs4_state *newstate;
569 570 571
	int ret;

	opendata->o_arg.open_flags = openflags;
572 573 574
	memset(&opendata->o_res, 0, sizeof(opendata->o_res));
	memset(&opendata->c_res, 0, sizeof(opendata->c_res));
	nfs4_init_opendata_res(opendata);
575 576 577
	ret = _nfs4_proc_open(opendata);
	if (ret != 0)
		return ret; 
578
	newstate = nfs4_opendata_to_nfs4_state(opendata);
579 580 581
	if (IS_ERR(newstate))
		return PTR_ERR(newstate);
	nfs4_close_state(&opendata->path, newstate, openflags);
582
	*res = newstate;
583 584 585 586 587 588 589 590 591
	return 0;
}

static int nfs4_open_recover(struct nfs4_opendata *opendata, struct nfs4_state *state)
{
	struct nfs4_state *newstate;
	int ret;

	/* memory barrier prior to reading state->n_* */
592
	clear_bit(NFS_DELEGATED_STATE, &state->flags);
593 594
	smp_rmb();
	if (state->n_rdwr != 0) {
595
		ret = nfs4_open_recover_helper(opendata, FMODE_READ|FMODE_WRITE, &newstate);
596 597
		if (ret != 0)
			return ret;
598 599
		if (newstate != state)
			return -ESTALE;
600 601
	}
	if (state->n_wronly != 0) {
602
		ret = nfs4_open_recover_helper(opendata, FMODE_WRITE, &newstate);
603 604
		if (ret != 0)
			return ret;
605 606
		if (newstate != state)
			return -ESTALE;
607 608
	}
	if (state->n_rdonly != 0) {
609
		ret = nfs4_open_recover_helper(opendata, FMODE_READ, &newstate);
610 611
		if (ret != 0)
			return ret;
612 613
		if (newstate != state)
			return -ESTALE;
614
	}
615 616 617 618 619 620
	/*
	 * We may have performed cached opens for all three recoveries.
	 * Check if we need to update the current stateid.
	 */
	if (test_bit(NFS_DELEGATED_STATE, &state->flags) == 0 &&
	    memcmp(state->stateid.data, state->open_stateid.data, sizeof(state->stateid.data)) != 0) {
621
		write_seqlock(&state->seqlock);
622 623
		if (test_bit(NFS_DELEGATED_STATE, &state->flags) == 0)
			memcpy(state->stateid.data, state->open_stateid.data, sizeof(state->stateid.data));
624
		write_sequnlock(&state->seqlock);
625
	}
626 627 628
	return 0;
}

L
Linus Torvalds 已提交
629 630 631 632
/*
 * OPEN_RECLAIM:
 * 	reclaim state on the server after a reboot.
 */
633
static int _nfs4_do_open_reclaim(struct nfs_open_context *ctx, struct nfs4_state *state)
L
Linus Torvalds 已提交
634
{
635
	struct nfs_delegation *delegation;
636 637
	struct nfs4_opendata *opendata;
	int delegation_type = 0;
L
Linus Torvalds 已提交
638 639
	int status;

T
Trond Myklebust 已提交
640 641 642
	opendata = nfs4_open_recoverdata_alloc(ctx, state);
	if (IS_ERR(opendata))
		return PTR_ERR(opendata);
643 644
	opendata->o_arg.claim = NFS4_OPEN_CLAIM_PREVIOUS;
	opendata->o_arg.fh = NFS_FH(state->inode);
645 646 647
	rcu_read_lock();
	delegation = rcu_dereference(NFS_I(state->inode)->delegation);
	if (delegation != NULL && (delegation->flags & NFS_DELEGATION_NEED_RECLAIM) != 0)
648
		delegation_type = delegation->type;
649
	rcu_read_unlock();
650 651
	opendata->o_arg.u.delegation_type = delegation_type;
	status = nfs4_open_recover(opendata, state);
652
	nfs4_opendata_put(opendata);
L
Linus Torvalds 已提交
653 654 655
	return status;
}

656
static int nfs4_do_open_reclaim(struct nfs_open_context *ctx, struct nfs4_state *state)
L
Linus Torvalds 已提交
657 658 659 660 661
{
	struct nfs_server *server = NFS_SERVER(state->inode);
	struct nfs4_exception exception = { };
	int err;
	do {
662
		err = _nfs4_do_open_reclaim(ctx, state);
663 664 665
		if (err != -NFS4ERR_DELAY)
			break;
		nfs4_handle_exception(server, err, &exception);
L
Linus Torvalds 已提交
666 667 668 669
	} while (exception.retry);
	return err;
}

670 671 672 673 674 675 676 677
static int nfs4_open_reclaim(struct nfs4_state_owner *sp, struct nfs4_state *state)
{
	struct nfs_open_context *ctx;
	int ret;

	ctx = nfs4_state_find_open_context(state);
	if (IS_ERR(ctx))
		return PTR_ERR(ctx);
678
	ret = nfs4_do_open_reclaim(ctx, state);
679 680 681 682
	put_nfs_open_context(ctx);
	return ret;
}

683
static int _nfs4_open_delegation_recall(struct nfs_open_context *ctx, struct nfs4_state *state, const nfs4_stateid *stateid)
L
Linus Torvalds 已提交
684
{
685
	struct nfs4_opendata *opendata;
686
	int ret;
L
Linus Torvalds 已提交
687

T
Trond Myklebust 已提交
688 689 690
	opendata = nfs4_open_recoverdata_alloc(ctx, state);
	if (IS_ERR(opendata))
		return PTR_ERR(opendata);
691
	opendata->o_arg.claim = NFS4_OPEN_CLAIM_DELEGATE_CUR;
692
	memcpy(opendata->o_arg.u.delegation.data, stateid->data,
693
			sizeof(opendata->o_arg.u.delegation.data));
694
	ret = nfs4_open_recover(opendata, state);
695
	nfs4_opendata_put(opendata);
696
	return ret;
L
Linus Torvalds 已提交
697 698
}

699
int nfs4_open_delegation_recall(struct nfs_open_context *ctx, struct nfs4_state *state, const nfs4_stateid *stateid)
L
Linus Torvalds 已提交
700 701
{
	struct nfs4_exception exception = { };
702
	struct nfs_server *server = NFS_SERVER(state->inode);
L
Linus Torvalds 已提交
703 704
	int err;
	do {
705
		err = _nfs4_open_delegation_recall(ctx, state, stateid);
L
Linus Torvalds 已提交
706 707 708 709 710 711 712
		switch (err) {
			case 0:
				return err;
			case -NFS4ERR_STALE_CLIENTID:
			case -NFS4ERR_STALE_STATEID:
			case -NFS4ERR_EXPIRED:
				/* Don't recall a delegation if it was lost */
713
				nfs4_schedule_state_recovery(server->nfs_client);
L
Linus Torvalds 已提交
714 715 716 717 718 719 720
				return err;
		}
		err = nfs4_handle_exception(server, err, &exception);
	} while (exception.retry);
	return err;
}

721 722 723 724 725 726 727
static void nfs4_open_confirm_done(struct rpc_task *task, void *calldata)
{
	struct nfs4_opendata *data = calldata;

	data->rpc_status = task->tk_status;
	if (RPC_ASSASSINATED(task))
		return;
728
	if (data->rpc_status == 0) {
729 730
		memcpy(data->o_res.stateid.data, data->c_res.stateid.data,
				sizeof(data->o_res.stateid.data));
731
		nfs_confirm_seqid(&data->owner->so_seqid, 0);
732
		renew_lease(data->o_res.server, data->timestamp);
733
		data->rpc_done = 1;
734
	}
735
	nfs_increment_open_seqid(data->rpc_status, data->c_arg.seqid);
736 737 738 739 740 741 742 743 744 745 746
}

static void nfs4_open_confirm_release(void *calldata)
{
	struct nfs4_opendata *data = calldata;
	struct nfs4_state *state = NULL;

	/* If this request hasn't been cancelled, do nothing */
	if (data->cancelled == 0)
		goto out_free;
	/* In case of error, no cleanup! */
747
	if (!data->rpc_done)
748 749
		goto out_free;
	state = nfs4_opendata_to_nfs4_state(data);
750
	if (!IS_ERR(state))
751
		nfs4_close_state(&data->path, state, data->o_arg.open_flags);
752
out_free:
753
	nfs4_opendata_put(data);
754 755 756 757 758 759 760 761 762 763 764 765 766 767
}

static const struct rpc_call_ops nfs4_open_confirm_ops = {
	.rpc_call_done = nfs4_open_confirm_done,
	.rpc_release = nfs4_open_confirm_release,
};

/*
 * Note: On error, nfs4_proc_open_confirm will free the struct nfs4_opendata
 */
static int _nfs4_proc_open_confirm(struct nfs4_opendata *data)
{
	struct nfs_server *server = NFS_SERVER(data->dir->d_inode);
	struct rpc_task *task;
768 769 770 771 772 773
	struct  rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_OPEN_CONFIRM],
		.rpc_argp = &data->c_arg,
		.rpc_resp = &data->c_res,
		.rpc_cred = data->owner->so_cred,
	};
T
Trond Myklebust 已提交
774 775
	struct rpc_task_setup task_setup_data = {
		.rpc_client = server->client,
776
		.rpc_message = &msg,
T
Trond Myklebust 已提交
777 778 779 780
		.callback_ops = &nfs4_open_confirm_ops,
		.callback_data = data,
		.flags = RPC_TASK_ASYNC,
	};
L
Linus Torvalds 已提交
781 782
	int status;

783
	kref_get(&data->kref);
784 785
	data->rpc_done = 0;
	data->rpc_status = 0;
786
	data->timestamp = jiffies;
T
Trond Myklebust 已提交
787
	task = rpc_run_task(&task_setup_data);
788
	if (IS_ERR(task))
789 790 791 792 793 794 795
		return PTR_ERR(task);
	status = nfs4_wait_for_completion_rpc_task(task);
	if (status != 0) {
		data->cancelled = 1;
		smp_wmb();
	} else
		status = data->rpc_status;
796
	rpc_put_task(task);
L
Linus Torvalds 已提交
797 798 799
	return status;
}

800
static void nfs4_open_prepare(struct rpc_task *task, void *calldata)
L
Linus Torvalds 已提交
801
{
802 803
	struct nfs4_opendata *data = calldata;
	struct nfs4_state_owner *sp = data->owner;
804

805 806
	if (nfs_wait_on_sequence(data->o_arg.seqid, task) != 0)
		return;
807 808 809 810 811 812 813
	/*
	 * Check if we still need to send an OPEN call, or if we can use
	 * a delegation instead.
	 */
	if (data->state != NULL) {
		struct nfs_delegation *delegation;

814 815
		if (can_open_cached(data->state, data->o_arg.open_flags & (FMODE_READ|FMODE_WRITE|O_EXCL)))
			goto out_no_action;
816 817 818 819 820
		rcu_read_lock();
		delegation = rcu_dereference(NFS_I(data->state->inode)->delegation);
		if (delegation != NULL &&
		   (delegation->flags & NFS_DELEGATION_NEED_RECLAIM) == 0) {
			rcu_read_unlock();
821
			goto out_no_action;
822 823 824
		}
		rcu_read_unlock();
	}
825
	/* Update sequence id. */
826
	data->o_arg.id = sp->so_owner_id.id;
827
	data->o_arg.clientid = sp->so_client->cl_clientid;
T
Trond Myklebust 已提交
828
	if (data->o_arg.claim == NFS4_OPEN_CLAIM_PREVIOUS) {
829
		task->tk_msg.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_OPEN_NOATTR];
T
Trond Myklebust 已提交
830 831
		nfs_copy_fh(&data->o_res.fh, data->o_arg.fh);
	}
832
	data->timestamp = jiffies;
833
	rpc_call_start(task);
834 835 836 837
	return;
out_no_action:
	task->tk_action = NULL;

838
}
L
Linus Torvalds 已提交
839

840 841 842
static void nfs4_open_done(struct rpc_task *task, void *calldata)
{
	struct nfs4_opendata *data = calldata;
L
Linus Torvalds 已提交
843

844 845 846 847 848
	data->rpc_status = task->tk_status;
	if (RPC_ASSASSINATED(task))
		return;
	if (task->tk_status == 0) {
		switch (data->o_res.f_attr->mode & S_IFMT) {
849 850 851
			case S_IFREG:
				break;
			case S_IFLNK:
852
				data->rpc_status = -ELOOP;
853 854
				break;
			case S_IFDIR:
855
				data->rpc_status = -EISDIR;
856 857
				break;
			default:
858
				data->rpc_status = -ENOTDIR;
859
		}
860
		renew_lease(data->o_res.server, data->timestamp);
861 862
		if (!(data->o_res.rflags & NFS4_OPEN_RESULT_CONFIRM))
			nfs_confirm_seqid(&data->owner->so_seqid, 0);
863
	}
864
	nfs_increment_open_seqid(data->rpc_status, data->o_arg.seqid);
865
	data->rpc_done = 1;
866
}
867

868 869 870 871 872 873 874 875 876
static void nfs4_open_release(void *calldata)
{
	struct nfs4_opendata *data = calldata;
	struct nfs4_state *state = NULL;

	/* If this request hasn't been cancelled, do nothing */
	if (data->cancelled == 0)
		goto out_free;
	/* In case of error, no cleanup! */
877
	if (data->rpc_status != 0 || !data->rpc_done)
878 879 880 881 882
		goto out_free;
	/* In case we need an open_confirm, no cleanup! */
	if (data->o_res.rflags & NFS4_OPEN_RESULT_CONFIRM)
		goto out_free;
	state = nfs4_opendata_to_nfs4_state(data);
883
	if (!IS_ERR(state))
884
		nfs4_close_state(&data->path, state, data->o_arg.open_flags);
885
out_free:
886
	nfs4_opendata_put(data);
887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904
}

static const struct rpc_call_ops nfs4_open_ops = {
	.rpc_call_prepare = nfs4_open_prepare,
	.rpc_call_done = nfs4_open_done,
	.rpc_release = nfs4_open_release,
};

/*
 * Note: On error, nfs4_proc_open will free the struct nfs4_opendata
 */
static int _nfs4_proc_open(struct nfs4_opendata *data)
{
	struct inode *dir = data->dir->d_inode;
	struct nfs_server *server = NFS_SERVER(dir);
	struct nfs_openargs *o_arg = &data->o_arg;
	struct nfs_openres *o_res = &data->o_res;
	struct rpc_task *task;
905 906 907 908 909 910
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_OPEN],
		.rpc_argp = o_arg,
		.rpc_resp = o_res,
		.rpc_cred = data->owner->so_cred,
	};
T
Trond Myklebust 已提交
911 912
	struct rpc_task_setup task_setup_data = {
		.rpc_client = server->client,
913
		.rpc_message = &msg,
T
Trond Myklebust 已提交
914 915 916 917
		.callback_ops = &nfs4_open_ops,
		.callback_data = data,
		.flags = RPC_TASK_ASYNC,
	};
918 919
	int status;

920
	kref_get(&data->kref);
921 922
	data->rpc_done = 0;
	data->rpc_status = 0;
923
	data->cancelled = 0;
T
Trond Myklebust 已提交
924
	task = rpc_run_task(&task_setup_data);
925
	if (IS_ERR(task))
926 927 928 929 930 931 932
		return PTR_ERR(task);
	status = nfs4_wait_for_completion_rpc_task(task);
	if (status != 0) {
		data->cancelled = 1;
		smp_wmb();
	} else
		status = data->rpc_status;
933
	rpc_put_task(task);
934
	if (status != 0 || !data->rpc_done)
935 936
		return status;

937 938 939
	if (o_res->fh.size == 0)
		_nfs4_proc_lookup(dir, o_arg->name, &o_res->fh, o_res->f_attr);

940 941 942 943 944
	if (o_arg->open_flags & O_CREAT) {
		update_changeattr(dir, &o_res->cinfo);
		nfs_post_op_update_inode(dir, o_res->dir_attr);
	} else
		nfs_refresh_inode(dir, o_res->dir_attr);
L
Linus Torvalds 已提交
945
	if(o_res->rflags & NFS4_OPEN_RESULT_CONFIRM) {
946
		status = _nfs4_proc_open_confirm(data);
L
Linus Torvalds 已提交
947
		if (status != 0)
948
			return status;
L
Linus Torvalds 已提交
949 950
	}
	if (!(o_res->f_attr->valid & NFS_ATTR_FATTR))
951
		_nfs4_proc_getattr(server, &o_res->fh, o_res->f_attr);
952
	return 0;
L
Linus Torvalds 已提交
953 954
}

955
static int nfs4_recover_expired_lease(struct nfs_server *server)
956
{
957
	struct nfs_client *clp = server->nfs_client;
958
	int ret;
959

960 961 962 963 964 965
	for (;;) {
		ret = nfs4_wait_clnt_recover(server->client, clp);
		if (ret != 0)
			return ret;
		if (!test_and_clear_bit(NFS4CLNT_LEASE_EXPIRED, &clp->cl_state))
			break;
966
		nfs4_schedule_state_recovery(clp);
967 968
	}
	return 0;
969 970
}

L
Linus Torvalds 已提交
971 972 973 974 975
/*
 * OPEN_EXPIRED:
 * 	reclaim state on the server after a network partition.
 * 	Assumes caller holds the appropriate lock
 */
976
static int _nfs4_open_expired(struct nfs_open_context *ctx, struct nfs4_state *state)
L
Linus Torvalds 已提交
977
{
978
	struct nfs4_opendata *opendata;
979
	int ret;
L
Linus Torvalds 已提交
980

T
Trond Myklebust 已提交
981 982 983
	opendata = nfs4_open_recoverdata_alloc(ctx, state);
	if (IS_ERR(opendata))
		return PTR_ERR(opendata);
984 985 986
	ret = nfs4_open_recover(opendata, state);
	if (ret == -ESTALE) {
		/* Invalidate the state owner so we don't ever use it again */
987 988
		nfs4_drop_state_owner(state->owner);
		d_drop(ctx->path.dentry);
L
Linus Torvalds 已提交
989
	}
990
	nfs4_opendata_put(opendata);
991
	return ret;
L
Linus Torvalds 已提交
992 993
}

994
static inline int nfs4_do_open_expired(struct nfs_open_context *ctx, struct nfs4_state *state)
995
{
996
	struct nfs_server *server = NFS_SERVER(state->inode);
997 998 999 1000
	struct nfs4_exception exception = { };
	int err;

	do {
1001
		err = _nfs4_open_expired(ctx, state);
1002 1003 1004 1005 1006 1007
		if (err == -NFS4ERR_DELAY)
			nfs4_handle_exception(server, err, &exception);
	} while (exception.retry);
	return err;
}

L
Linus Torvalds 已提交
1008 1009 1010
static int nfs4_open_expired(struct nfs4_state_owner *sp, struct nfs4_state *state)
{
	struct nfs_open_context *ctx;
1011
	int ret;
L
Linus Torvalds 已提交
1012

1013 1014 1015
	ctx = nfs4_state_find_open_context(state);
	if (IS_ERR(ctx))
		return PTR_ERR(ctx);
1016
	ret = nfs4_do_open_expired(ctx, state);
1017 1018
	put_nfs_open_context(ctx);
	return ret;
L
Linus Torvalds 已提交
1019 1020
}

1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036
/*
 * on an EXCLUSIVE create, the server should send back a bitmask with FATTR4-*
 * fields corresponding to attributes that were used to store the verifier.
 * Make sure we clobber those fields in the later setattr call
 */
static inline void nfs4_exclusive_attrset(struct nfs4_opendata *opendata, struct iattr *sattr)
{
	if ((opendata->o_res.attrset[1] & FATTR4_WORD1_TIME_ACCESS) &&
	    !(sattr->ia_valid & ATTR_ATIME_SET))
		sattr->ia_valid |= ATTR_ATIME;

	if ((opendata->o_res.attrset[1] & FATTR4_WORD1_TIME_MODIFY) &&
	    !(sattr->ia_valid & ATTR_MTIME_SET))
		sattr->ia_valid |= ATTR_MTIME;
}

L
Linus Torvalds 已提交
1037
/*
1038
 * Returns a referenced nfs4_state
L
Linus Torvalds 已提交
1039
 */
1040
static int _nfs4_do_open(struct inode *dir, struct path *path, int flags, struct iattr *sattr, struct rpc_cred *cred, struct nfs4_state **res)
L
Linus Torvalds 已提交
1041 1042 1043 1044
{
	struct nfs4_state_owner  *sp;
	struct nfs4_state     *state = NULL;
	struct nfs_server       *server = NFS_SERVER(dir);
1045
	struct nfs_client *clp = server->nfs_client;
1046
	struct nfs4_opendata *opendata;
1047
	int status;
L
Linus Torvalds 已提交
1048 1049 1050 1051 1052 1053 1054

	/* Protect against reboot recovery conflicts */
	status = -ENOMEM;
	if (!(sp = nfs4_get_state_owner(server, cred))) {
		dprintk("nfs4_do_open: nfs4_get_state_owner failed!\n");
		goto out_err;
	}
1055 1056
	status = nfs4_recover_expired_lease(server);
	if (status != 0)
1057
		goto err_put_state_owner;
1058 1059
	if (path->dentry->d_inode != NULL)
		nfs4_return_incompatible_delegation(path->dentry->d_inode, flags & (FMODE_READ|FMODE_WRITE));
1060 1061
	down_read(&clp->cl_sem);
	status = -ENOMEM;
1062
	opendata = nfs4_opendata_alloc(path, sp, flags, sattr);
1063
	if (opendata == NULL)
1064
		goto err_release_rwsem;
L
Linus Torvalds 已提交
1065

1066 1067 1068
	if (path->dentry->d_inode != NULL)
		opendata->state = nfs4_get_open_state(path->dentry->d_inode, sp);

1069
	status = _nfs4_proc_open(opendata);
L
Linus Torvalds 已提交
1070
	if (status != 0)
1071
		goto err_opendata_put;
L
Linus Torvalds 已提交
1072

1073 1074 1075
	if (opendata->o_arg.open_flags & O_EXCL)
		nfs4_exclusive_attrset(opendata, sattr);

1076
	state = nfs4_opendata_to_nfs4_state(opendata);
1077 1078
	status = PTR_ERR(state);
	if (IS_ERR(state))
1079 1080
		goto err_opendata_put;
	nfs4_opendata_put(opendata);
L
Linus Torvalds 已提交
1081 1082 1083 1084
	nfs4_put_state_owner(sp);
	up_read(&clp->cl_sem);
	*res = state;
	return 0;
1085 1086
err_opendata_put:
	nfs4_opendata_put(opendata);
1087 1088
err_release_rwsem:
	up_read(&clp->cl_sem);
1089 1090
err_put_state_owner:
	nfs4_put_state_owner(sp);
L
Linus Torvalds 已提交
1091 1092 1093 1094 1095 1096
out_err:
	*res = NULL;
	return status;
}


1097
static struct nfs4_state *nfs4_do_open(struct inode *dir, struct path *path, int flags, struct iattr *sattr, struct rpc_cred *cred)
L
Linus Torvalds 已提交
1098 1099 1100 1101 1102 1103
{
	struct nfs4_exception exception = { };
	struct nfs4_state *res;
	int status;

	do {
1104
		status = _nfs4_do_open(dir, path, flags, sattr, cred, &res);
L
Linus Torvalds 已提交
1105 1106 1107 1108 1109 1110 1111 1112
		if (status == 0)
			break;
		/* NOTE: BAD_SEQID means the server and client disagree about the
		 * book-keeping w.r.t. state-changing operations
		 * (OPEN/CLOSE/LOCK/LOCKU...)
		 * It is actually a sign of a bug on the client or on the server.
		 *
		 * If we receive a BAD_SEQID error in the particular case of
1113
		 * doing an OPEN, we assume that nfs_increment_open_seqid() will
L
Linus Torvalds 已提交
1114 1115 1116 1117 1118
		 * have unhashed the old state_owner for us, and that we can
		 * therefore safely retry using a new one. We should still warn
		 * the user though...
		 */
		if (status == -NFS4ERR_BAD_SEQID) {
1119 1120 1121
			printk(KERN_WARNING "NFS: v4 server %s "
					" returned a bad sequence-id error!\n",
					NFS_SERVER(dir)->nfs_client->cl_hostname);
L
Linus Torvalds 已提交
1122 1123 1124
			exception.retry = 1;
			continue;
		}
1125 1126 1127 1128 1129 1130 1131 1132 1133 1134
		/*
		 * BAD_STATEID on OPEN means that the server cancelled our
		 * state before it received the OPEN_CONFIRM.
		 * Recover by retrying the request as per the discussion
		 * on Page 181 of RFC3530.
		 */
		if (status == -NFS4ERR_BAD_STATEID) {
			exception.retry = 1;
			continue;
		}
1135 1136 1137 1138 1139
		if (status == -EAGAIN) {
			/* We must have found a delegation */
			exception.retry = 1;
			continue;
		}
L
Linus Torvalds 已提交
1140 1141 1142 1143 1144 1145
		res = ERR_PTR(nfs4_handle_exception(NFS_SERVER(dir),
					status, &exception));
	} while (exception.retry);
	return res;
}

1146 1147
static int _nfs4_do_setattr(struct inode *inode, struct nfs_fattr *fattr,
                struct iattr *sattr, struct nfs4_state *state)
L
Linus Torvalds 已提交
1148
{
1149
	struct nfs_server *server = NFS_SERVER(inode);
L
Linus Torvalds 已提交
1150
        struct nfs_setattrargs  arg = {
1151
                .fh             = NFS_FH(inode),
L
Linus Torvalds 已提交
1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164
                .iap            = sattr,
		.server		= server,
		.bitmask = server->attr_bitmask,
        };
        struct nfs_setattrres  res = {
		.fattr		= fattr,
		.server		= server,
        };
        struct rpc_message msg = {
                .rpc_proc       = &nfs4_procedures[NFSPROC4_CLNT_SETATTR],
                .rpc_argp       = &arg,
                .rpc_resp       = &res,
        };
1165
	unsigned long timestamp = jiffies;
1166
	int status;
L
Linus Torvalds 已提交
1167

1168
	nfs_fattr_init(fattr);
L
Linus Torvalds 已提交
1169

1170 1171 1172
	if (nfs4_copy_delegation_stateid(&arg.stateid, inode)) {
		/* Use that stateid */
	} else if (state != NULL) {
L
Linus Torvalds 已提交
1173
		msg.rpc_cred = state->owner->so_cred;
1174 1175
		nfs4_copy_stateid(&arg.stateid, state, current->files);
	} else
L
Linus Torvalds 已提交
1176 1177
		memcpy(&arg.stateid, &zero_stateid, sizeof(arg.stateid));

1178
	status = rpc_call_sync(server->client, &msg, 0);
1179 1180
	if (status == 0 && state != NULL)
		renew_lease(server, timestamp);
1181
	return status;
L
Linus Torvalds 已提交
1182 1183
}

1184 1185
static int nfs4_do_setattr(struct inode *inode, struct nfs_fattr *fattr,
                struct iattr *sattr, struct nfs4_state *state)
L
Linus Torvalds 已提交
1186
{
1187
	struct nfs_server *server = NFS_SERVER(inode);
L
Linus Torvalds 已提交
1188 1189 1190 1191
	struct nfs4_exception exception = { };
	int err;
	do {
		err = nfs4_handle_exception(server,
1192
				_nfs4_do_setattr(inode, fattr, sattr, state),
L
Linus Torvalds 已提交
1193 1194 1195 1196 1197 1198
				&exception);
	} while (exception.retry);
	return err;
}

struct nfs4_closedata {
1199
	struct path path;
L
Linus Torvalds 已提交
1200 1201 1202 1203
	struct inode *inode;
	struct nfs4_state *state;
	struct nfs_closeargs arg;
	struct nfs_closeres res;
1204
	struct nfs_fattr fattr;
1205
	unsigned long timestamp;
L
Linus Torvalds 已提交
1206 1207
};

1208
static void nfs4_free_closedata(void *data)
1209
{
1210 1211
	struct nfs4_closedata *calldata = data;
	struct nfs4_state_owner *sp = calldata->state->owner;
1212 1213 1214 1215

	nfs4_put_open_state(calldata->state);
	nfs_free_seqid(calldata->arg.seqid);
	nfs4_put_state_owner(sp);
1216 1217
	dput(calldata->path.dentry);
	mntput(calldata->path.mnt);
1218 1219 1220
	kfree(calldata);
}

1221
static void nfs4_close_done(struct rpc_task *task, void *data)
L
Linus Torvalds 已提交
1222
{
1223
	struct nfs4_closedata *calldata = data;
L
Linus Torvalds 已提交
1224 1225 1226
	struct nfs4_state *state = calldata->state;
	struct nfs_server *server = NFS_SERVER(calldata->inode);

T
Trond Myklebust 已提交
1227 1228
	if (RPC_ASSASSINATED(task))
		return;
L
Linus Torvalds 已提交
1229 1230 1231
        /* hmm. we are done with the inode, and in the process of freeing
	 * the state_owner. we keep this around to process errors
	 */
1232
	nfs_increment_open_seqid(task->tk_status, calldata->arg.seqid);
L
Linus Torvalds 已提交
1233 1234
	switch (task->tk_status) {
		case 0:
1235
			nfs_set_open_stateid(state, &calldata->res.stateid, 0);
1236
			renew_lease(server, calldata->timestamp);
L
Linus Torvalds 已提交
1237 1238 1239 1240 1241 1242 1243 1244 1245 1246
			break;
		case -NFS4ERR_STALE_STATEID:
		case -NFS4ERR_EXPIRED:
			break;
		default:
			if (nfs4_async_handle_error(task, server) == -EAGAIN) {
				rpc_restart_call(task);
				return;
			}
	}
1247
	nfs_refresh_inode(calldata->inode, calldata->res.fattr);
L
Linus Torvalds 已提交
1248 1249
}

T
Trond Myklebust 已提交
1250
static void nfs4_close_prepare(struct rpc_task *task, void *data)
L
Linus Torvalds 已提交
1251
{
T
Trond Myklebust 已提交
1252
	struct nfs4_closedata *calldata = data;
1253
	struct nfs4_state *state = calldata->state;
1254
	int clear_rd, clear_wr, clear_rdwr;
1255

1256
	if (nfs_wait_on_sequence(calldata->arg.seqid, task) != 0)
1257
		return;
1258 1259

	clear_rd = clear_wr = clear_rdwr = 0;
1260
	spin_lock(&state->owner->so_lock);
1261
	/* Calculate the change in open mode */
1262
	if (state->n_rdwr == 0) {
1263 1264 1265 1266 1267 1268 1269 1270
		if (state->n_rdonly == 0) {
			clear_rd |= test_and_clear_bit(NFS_O_RDONLY_STATE, &state->flags);
			clear_rdwr |= test_and_clear_bit(NFS_O_RDWR_STATE, &state->flags);
		}
		if (state->n_wronly == 0) {
			clear_wr |= test_and_clear_bit(NFS_O_WRONLY_STATE, &state->flags);
			clear_rdwr |= test_and_clear_bit(NFS_O_RDWR_STATE, &state->flags);
		}
1271
	}
1272
	spin_unlock(&state->owner->so_lock);
1273
	if (!clear_rd && !clear_wr && !clear_rdwr) {
1274 1275
		/* Note: exit _without_ calling nfs4_close_done */
		task->tk_action = NULL;
1276 1277
		return;
	}
1278
	nfs_fattr_init(calldata->res.fattr);
1279
	if (test_bit(NFS_O_RDONLY_STATE, &state->flags) != 0) {
1280
		task->tk_msg.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_OPEN_DOWNGRADE];
1281 1282
		calldata->arg.open_flags = FMODE_READ;
	} else if (test_bit(NFS_O_WRONLY_STATE, &state->flags) != 0) {
1283
		task->tk_msg.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_OPEN_DOWNGRADE];
1284 1285
		calldata->arg.open_flags = FMODE_WRITE;
	}
1286
	calldata->timestamp = jiffies;
1287
	rpc_call_start(task);
L
Linus Torvalds 已提交
1288 1289
}

1290
static const struct rpc_call_ops nfs4_close_ops = {
T
Trond Myklebust 已提交
1291
	.rpc_call_prepare = nfs4_close_prepare,
1292 1293 1294 1295
	.rpc_call_done = nfs4_close_done,
	.rpc_release = nfs4_free_closedata,
};

L
Linus Torvalds 已提交
1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306
/* 
 * It is possible for data to be read/written from a mem-mapped file 
 * after the sys_close call (which hits the vfs layer as a flush).
 * This means that we can't safely call nfsv4 close on a file until 
 * the inode is cleared. This in turn means that we are not good
 * NFSv4 citizens - we do not indicate to the server to update the file's 
 * share state even when we are done with one of the three share 
 * stateid's in the inode.
 *
 * NOTE: Caller must be holding the sp->so_owner semaphore!
 */
1307
int nfs4_do_close(struct path *path, struct nfs4_state *state, int wait)
L
Linus Torvalds 已提交
1308
{
1309
	struct nfs_server *server = NFS_SERVER(state->inode);
L
Linus Torvalds 已提交
1310
	struct nfs4_closedata *calldata;
1311 1312
	struct nfs4_state_owner *sp = state->owner;
	struct rpc_task *task;
1313 1314 1315 1316
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_CLOSE],
		.rpc_cred = state->owner->so_cred,
	};
T
Trond Myklebust 已提交
1317 1318
	struct rpc_task_setup task_setup_data = {
		.rpc_client = server->client,
1319
		.rpc_message = &msg,
T
Trond Myklebust 已提交
1320 1321 1322
		.callback_ops = &nfs4_close_ops,
		.flags = RPC_TASK_ASYNC,
	};
1323
	int status = -ENOMEM;
L
Linus Torvalds 已提交
1324

1325
	calldata = kmalloc(sizeof(*calldata), GFP_KERNEL);
L
Linus Torvalds 已提交
1326
	if (calldata == NULL)
1327
		goto out;
1328
	calldata->inode = state->inode;
L
Linus Torvalds 已提交
1329
	calldata->state = state;
1330
	calldata->arg.fh = NFS_FH(state->inode);
1331
	calldata->arg.stateid = &state->open_stateid;
L
Linus Torvalds 已提交
1332
	/* Serialization for the sequence id */
1333
	calldata->arg.seqid = nfs_alloc_seqid(&state->owner->so_seqid);
1334 1335
	if (calldata->arg.seqid == NULL)
		goto out_free_calldata;
1336 1337 1338
	calldata->arg.bitmask = server->attr_bitmask;
	calldata->res.fattr = &calldata->fattr;
	calldata->res.server = server;
1339 1340
	calldata->path.mnt = mntget(path->mnt);
	calldata->path.dentry = dget(path->dentry);
1341

1342 1343
	msg.rpc_argp = &calldata->arg,
	msg.rpc_resp = &calldata->res,
T
Trond Myklebust 已提交
1344 1345
	task_setup_data.callback_data = calldata;
	task = rpc_run_task(&task_setup_data);
1346 1347
	if (IS_ERR(task))
		return PTR_ERR(task);
1348 1349 1350
	status = 0;
	if (wait)
		status = rpc_wait_for_completion_task(task);
1351
	rpc_put_task(task);
1352
	return status;
1353 1354 1355
out_free_calldata:
	kfree(calldata);
out:
1356 1357
	nfs4_put_open_state(state);
	nfs4_put_state_owner(sp);
1358
	return status;
L
Linus Torvalds 已提交
1359 1360
}

1361
static int nfs4_intent_set_file(struct nameidata *nd, struct path *path, struct nfs4_state *state)
1362 1363
{
	struct file *filp;
1364
	int ret;
1365

1366 1367
	/* If the open_intent is for execute, we have an extra check to make */
	if (nd->intent.open.flags & FMODE_EXEC) {
1368
		ret = nfs_may_open(state->inode,
1369 1370 1371 1372 1373
				state->owner->so_cred,
				nd->intent.open.flags);
		if (ret < 0)
			goto out_close;
	}
1374
	filp = lookup_instantiate_filp(nd, path->dentry, NULL);
1375 1376
	if (!IS_ERR(filp)) {
		struct nfs_open_context *ctx;
1377
		ctx = nfs_file_open_context(filp);
1378
		ctx->state = state;
1379 1380
		return 0;
	}
1381 1382
	ret = PTR_ERR(filp);
out_close:
1383
	nfs4_close_sync(path, state, nd->intent.open.flags);
1384
	return ret;
1385 1386 1387
}

struct dentry *
L
Linus Torvalds 已提交
1388 1389
nfs4_atomic_open(struct inode *dir, struct dentry *dentry, struct nameidata *nd)
{
T
Trond Myklebust 已提交
1390
	struct dentry *parent;
1391 1392 1393 1394
	struct path path = {
		.mnt = nd->mnt,
		.dentry = dentry,
	};
L
Linus Torvalds 已提交
1395 1396 1397
	struct iattr attr;
	struct rpc_cred *cred;
	struct nfs4_state *state;
1398
	struct dentry *res;
L
Linus Torvalds 已提交
1399 1400 1401 1402 1403 1404 1405 1406 1407 1408 1409

	if (nd->flags & LOOKUP_CREATE) {
		attr.ia_mode = nd->intent.open.create_mode;
		attr.ia_valid = ATTR_MODE;
		if (!IS_POSIXACL(dir))
			attr.ia_mode &= ~current->fs->umask;
	} else {
		attr.ia_valid = 0;
		BUG_ON(nd->intent.open.flags & O_CREAT);
	}

1410
	cred = rpcauth_lookupcred(NFS_CLIENT(dir)->cl_auth, 0);
L
Linus Torvalds 已提交
1411
	if (IS_ERR(cred))
1412
		return (struct dentry *)cred;
T
Trond Myklebust 已提交
1413 1414 1415
	parent = dentry->d_parent;
	/* Protect against concurrent sillydeletes */
	nfs_block_sillyrename(parent);
1416
	state = nfs4_do_open(dir, &path, nd->intent.open.flags, &attr, cred);
L
Linus Torvalds 已提交
1417
	put_rpccred(cred);
1418
	if (IS_ERR(state)) {
1419
		if (PTR_ERR(state) == -ENOENT) {
1420
			d_add(dentry, NULL);
1421 1422
			nfs_set_verifier(dentry, nfs_save_change_attribute(dir));
		}
T
Trond Myklebust 已提交
1423
		nfs_unblock_sillyrename(parent);
1424 1425
		return (struct dentry *)state;
	}
1426
	res = d_add_unique(dentry, igrab(state->inode));
1427
	if (res != NULL)
1428
		path.dentry = res;
1429
	nfs_set_verifier(path.dentry, nfs_save_change_attribute(dir));
T
Trond Myklebust 已提交
1430
	nfs_unblock_sillyrename(parent);
1431
	nfs4_intent_set_file(nd, &path, state);
1432
	return res;
L
Linus Torvalds 已提交
1433 1434 1435
}

int
1436
nfs4_open_revalidate(struct inode *dir, struct dentry *dentry, int openflags, struct nameidata *nd)
L
Linus Torvalds 已提交
1437
{
1438 1439 1440 1441
	struct path path = {
		.mnt = nd->mnt,
		.dentry = dentry,
	};
L
Linus Torvalds 已提交
1442 1443 1444
	struct rpc_cred *cred;
	struct nfs4_state *state;

1445
	cred = rpcauth_lookupcred(NFS_CLIENT(dir)->cl_auth, 0);
L
Linus Torvalds 已提交
1446 1447
	if (IS_ERR(cred))
		return PTR_ERR(cred);
1448
	state = nfs4_do_open(dir, &path, openflags, NULL, cred);
L
Linus Torvalds 已提交
1449
	put_rpccred(cred);
1450 1451 1452 1453 1454 1455 1456 1457 1458
	if (IS_ERR(state)) {
		switch (PTR_ERR(state)) {
			case -EPERM:
			case -EACCES:
			case -EDQUOT:
			case -ENOSPC:
			case -EROFS:
				lookup_instantiate_filp(nd, (struct dentry *)state, NULL);
				return 1;
1459 1460
			default:
				goto out_drop;
1461 1462
		}
	}
1463
	if (state->inode == dentry->d_inode) {
1464
		nfs_set_verifier(dentry, nfs_save_change_attribute(dir));
1465
		nfs4_intent_set_file(nd, &path, state);
L
Linus Torvalds 已提交
1466 1467
		return 1;
	}
1468
	nfs4_close_sync(&path, state, openflags);
1469 1470
out_drop:
	d_drop(dentry);
L
Linus Torvalds 已提交
1471 1472 1473 1474 1475 1476 1477 1478 1479 1480 1481 1482 1483 1484 1485 1486 1487 1488 1489 1490 1491 1492 1493 1494 1495 1496 1497 1498
	return 0;
}


static int _nfs4_server_capabilities(struct nfs_server *server, struct nfs_fh *fhandle)
{
	struct nfs4_server_caps_res res = {};
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_SERVER_CAPS],
		.rpc_argp = fhandle,
		.rpc_resp = &res,
	};
	int status;

	status = rpc_call_sync(server->client, &msg, 0);
	if (status == 0) {
		memcpy(server->attr_bitmask, res.attr_bitmask, sizeof(server->attr_bitmask));
		if (res.attr_bitmask[0] & FATTR4_WORD0_ACL)
			server->caps |= NFS_CAP_ACLS;
		if (res.has_links != 0)
			server->caps |= NFS_CAP_HARDLINKS;
		if (res.has_symlinks != 0)
			server->caps |= NFS_CAP_SYMLINKS;
		server->acl_bitmask = res.acl_bitmask;
	}
	return status;
}

1499
int nfs4_server_capabilities(struct nfs_server *server, struct nfs_fh *fhandle)
L
Linus Torvalds 已提交
1500 1501 1502 1503 1504 1505 1506 1507 1508 1509 1510 1511 1512 1513 1514 1515 1516 1517 1518
{
	struct nfs4_exception exception = { };
	int err;
	do {
		err = nfs4_handle_exception(server,
				_nfs4_server_capabilities(server, fhandle),
				&exception);
	} while (exception.retry);
	return err;
}

static int _nfs4_lookup_root(struct nfs_server *server, struct nfs_fh *fhandle,
		struct nfs_fsinfo *info)
{
	struct nfs4_lookup_root_arg args = {
		.bitmask = nfs4_fattr_bitmap,
	};
	struct nfs4_lookup_res res = {
		.server = server,
1519
		.fattr = info->fattr,
L
Linus Torvalds 已提交
1520 1521 1522 1523 1524 1525 1526
		.fh = fhandle,
	};
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_LOOKUP_ROOT],
		.rpc_argp = &args,
		.rpc_resp = &res,
	};
1527
	nfs_fattr_init(info->fattr);
L
Linus Torvalds 已提交
1528 1529 1530 1531 1532 1533 1534 1535 1536 1537 1538 1539 1540 1541 1542 1543
	return rpc_call_sync(server->client, &msg, 0);
}

static int nfs4_lookup_root(struct nfs_server *server, struct nfs_fh *fhandle,
		struct nfs_fsinfo *info)
{
	struct nfs4_exception exception = { };
	int err;
	do {
		err = nfs4_handle_exception(server,
				_nfs4_lookup_root(server, fhandle, info),
				&exception);
	} while (exception.retry);
	return err;
}

1544 1545 1546
/*
 * get the file handle for the "/" directory on the server
 */
L
Linus Torvalds 已提交
1547
static int nfs4_proc_get_root(struct nfs_server *server, struct nfs_fh *fhandle,
1548
			      struct nfs_fsinfo *info)
L
Linus Torvalds 已提交
1549 1550 1551 1552 1553 1554 1555 1556
{
	int status;

	status = nfs4_lookup_root(server, fhandle, info);
	if (status == 0)
		status = nfs4_server_capabilities(server, fhandle);
	if (status == 0)
		status = nfs4_do_fsinfo(server, fhandle, info);
1557
	return nfs4_map_errors(status);
L
Linus Torvalds 已提交
1558 1559
}

M
Manoj Naik 已提交
1560 1561 1562 1563 1564
/*
 * Get locations and (maybe) other attributes of a referral.
 * Note that we'll actually follow the referral later when
 * we detect fsid mismatch in inode revalidation
 */
1565
static int nfs4_get_referral(struct inode *dir, const struct qstr *name, struct nfs_fattr *fattr, struct nfs_fh *fhandle)
M
Manoj Naik 已提交
1566 1567 1568 1569 1570 1571 1572 1573 1574 1575 1576 1577
{
	int status = -ENOMEM;
	struct page *page = NULL;
	struct nfs4_fs_locations *locations = NULL;

	page = alloc_page(GFP_KERNEL);
	if (page == NULL)
		goto out;
	locations = kmalloc(sizeof(struct nfs4_fs_locations), GFP_KERNEL);
	if (locations == NULL)
		goto out;

1578
	status = nfs4_proc_fs_locations(dir, name, locations, page);
M
Manoj Naik 已提交
1579 1580 1581 1582 1583 1584 1585 1586 1587 1588 1589 1590 1591 1592 1593 1594 1595 1596 1597 1598 1599 1600
	if (status != 0)
		goto out;
	/* Make sure server returned a different fsid for the referral */
	if (nfs_fsid_equal(&NFS_SERVER(dir)->fsid, &locations->fattr.fsid)) {
		dprintk("%s: server did not return a different fsid for a referral at %s\n", __FUNCTION__, name->name);
		status = -EIO;
		goto out;
	}

	memcpy(fattr, &locations->fattr, sizeof(struct nfs_fattr));
	fattr->valid |= NFS_ATTR_FATTR_V4_REFERRAL;
	if (!fattr->mode)
		fattr->mode = S_IFDIR;
	memset(fhandle, 0, sizeof(struct nfs_fh));
out:
	if (page)
		__free_page(page);
	if (locations)
		kfree(locations);
	return status;
}

L
Linus Torvalds 已提交
1601 1602 1603 1604 1605 1606 1607 1608 1609 1610 1611 1612 1613 1614 1615 1616
static int _nfs4_proc_getattr(struct nfs_server *server, struct nfs_fh *fhandle, struct nfs_fattr *fattr)
{
	struct nfs4_getattr_arg args = {
		.fh = fhandle,
		.bitmask = server->attr_bitmask,
	};
	struct nfs4_getattr_res res = {
		.fattr = fattr,
		.server = server,
	};
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_GETATTR],
		.rpc_argp = &args,
		.rpc_resp = &res,
	};
	
1617
	nfs_fattr_init(fattr);
L
Linus Torvalds 已提交
1618 1619 1620 1621 1622 1623 1624 1625 1626 1627 1628 1629 1630 1631 1632 1633 1634 1635 1636 1637 1638 1639 1640 1641 1642 1643 1644 1645 1646 1647 1648 1649 1650 1651 1652 1653
	return rpc_call_sync(server->client, &msg, 0);
}

static int nfs4_proc_getattr(struct nfs_server *server, struct nfs_fh *fhandle, struct nfs_fattr *fattr)
{
	struct nfs4_exception exception = { };
	int err;
	do {
		err = nfs4_handle_exception(server,
				_nfs4_proc_getattr(server, fhandle, fattr),
				&exception);
	} while (exception.retry);
	return err;
}

/* 
 * The file is not closed if it is opened due to the a request to change
 * the size of the file. The open call will not be needed once the
 * VFS layer lookup-intents are implemented.
 *
 * Close is called when the inode is destroyed.
 * If we haven't opened the file for O_WRONLY, we
 * need to in the size_change case to obtain a stateid.
 *
 * Got race?
 * Because OPEN is always done by name in nfsv4, it is
 * possible that we opened a different file by the same
 * name.  We can recognize this race condition, but we
 * can't do anything about it besides returning an error.
 *
 * This will be fixed with VFS changes (lookup-intent).
 */
static int
nfs4_proc_setattr(struct dentry *dentry, struct nfs_fattr *fattr,
		  struct iattr *sattr)
{
1654 1655
	struct rpc_cred *cred;
	struct inode *inode = dentry->d_inode;
1656 1657
	struct nfs_open_context *ctx;
	struct nfs4_state *state = NULL;
L
Linus Torvalds 已提交
1658 1659
	int status;

1660
	nfs_fattr_init(fattr);
L
Linus Torvalds 已提交
1661
	
1662
	cred = rpcauth_lookupcred(NFS_CLIENT(inode)->cl_auth, 0);
1663 1664
	if (IS_ERR(cred))
		return PTR_ERR(cred);
1665 1666 1667 1668 1669

	/* Search for an existing open(O_WRITE) file */
	ctx = nfs_find_open_context(inode, cred, FMODE_WRITE);
	if (ctx != NULL)
		state = ctx->state;
1670

1671
	status = nfs4_do_setattr(inode, fattr, sattr, state);
1672 1673
	if (status == 0)
		nfs_setattr_update_inode(inode, sattr);
1674 1675
	if (ctx != NULL)
		put_nfs_open_context(ctx);
1676
	put_rpccred(cred);
L
Linus Torvalds 已提交
1677 1678 1679
	return status;
}

1680 1681
static int _nfs4_proc_lookupfh(struct nfs_server *server, const struct nfs_fh *dirfh,
		const struct qstr *name, struct nfs_fh *fhandle,
D
David Howells 已提交
1682 1683 1684 1685 1686 1687 1688 1689 1690 1691 1692 1693 1694 1695 1696 1697 1698 1699 1700 1701 1702 1703 1704 1705 1706 1707 1708 1709 1710 1711 1712 1713 1714 1715
		struct nfs_fattr *fattr)
{
	int		       status;
	struct nfs4_lookup_arg args = {
		.bitmask = server->attr_bitmask,
		.dir_fh = dirfh,
		.name = name,
	};
	struct nfs4_lookup_res res = {
		.server = server,
		.fattr = fattr,
		.fh = fhandle,
	};
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_LOOKUP],
		.rpc_argp = &args,
		.rpc_resp = &res,
	};

	nfs_fattr_init(fattr);

	dprintk("NFS call  lookupfh %s\n", name->name);
	status = rpc_call_sync(server->client, &msg, 0);
	dprintk("NFS reply lookupfh: %d\n", status);
	return status;
}

static int nfs4_proc_lookupfh(struct nfs_server *server, struct nfs_fh *dirfh,
			      struct qstr *name, struct nfs_fh *fhandle,
			      struct nfs_fattr *fattr)
{
	struct nfs4_exception exception = { };
	int err;
	do {
1716 1717 1718 1719 1720 1721 1722
		err = _nfs4_proc_lookupfh(server, dirfh, name, fhandle, fattr);
		/* FIXME: !!!! */
		if (err == -NFS4ERR_MOVED) {
			err = -EREMOTE;
			break;
		}
		err = nfs4_handle_exception(server, err, &exception);
D
David Howells 已提交
1723 1724 1725 1726
	} while (exception.retry);
	return err;
}

1727
static int _nfs4_proc_lookup(struct inode *dir, const struct qstr *name,
L
Linus Torvalds 已提交
1728 1729
		struct nfs_fh *fhandle, struct nfs_fattr *fattr)
{
1730
	int status;
L
Linus Torvalds 已提交
1731 1732
	
	dprintk("NFS call  lookup %s\n", name->name);
1733
	status = _nfs4_proc_lookupfh(NFS_SERVER(dir), NFS_FH(dir), name, fhandle, fattr);
M
Manoj Naik 已提交
1734 1735
	if (status == -NFS4ERR_MOVED)
		status = nfs4_get_referral(dir, name, fattr, fhandle);
L
Linus Torvalds 已提交
1736 1737 1738 1739 1740 1741 1742 1743 1744 1745 1746 1747 1748 1749 1750 1751 1752 1753
	dprintk("NFS reply lookup: %d\n", status);
	return status;
}

static int nfs4_proc_lookup(struct inode *dir, struct qstr *name, struct nfs_fh *fhandle, struct nfs_fattr *fattr)
{
	struct nfs4_exception exception = { };
	int err;
	do {
		err = nfs4_handle_exception(NFS_SERVER(dir),
				_nfs4_proc_lookup(dir, name, fhandle, fattr),
				&exception);
	} while (exception.retry);
	return err;
}

static int _nfs4_proc_access(struct inode *inode, struct nfs_access_entry *entry)
{
1754 1755
	struct nfs_server *server = NFS_SERVER(inode);
	struct nfs_fattr fattr;
L
Linus Torvalds 已提交
1756 1757
	struct nfs4_accessargs args = {
		.fh = NFS_FH(inode),
1758 1759 1760 1761 1762
		.bitmask = server->attr_bitmask,
	};
	struct nfs4_accessres res = {
		.server = server,
		.fattr = &fattr,
L
Linus Torvalds 已提交
1763 1764 1765 1766 1767 1768 1769 1770 1771 1772 1773 1774 1775 1776 1777 1778 1779 1780 1781 1782 1783 1784 1785 1786 1787 1788
	};
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_ACCESS],
		.rpc_argp = &args,
		.rpc_resp = &res,
		.rpc_cred = entry->cred,
	};
	int mode = entry->mask;
	int status;

	/*
	 * Determine which access bits we want to ask for...
	 */
	if (mode & MAY_READ)
		args.access |= NFS4_ACCESS_READ;
	if (S_ISDIR(inode->i_mode)) {
		if (mode & MAY_WRITE)
			args.access |= NFS4_ACCESS_MODIFY | NFS4_ACCESS_EXTEND | NFS4_ACCESS_DELETE;
		if (mode & MAY_EXEC)
			args.access |= NFS4_ACCESS_LOOKUP;
	} else {
		if (mode & MAY_WRITE)
			args.access |= NFS4_ACCESS_MODIFY | NFS4_ACCESS_EXTEND;
		if (mode & MAY_EXEC)
			args.access |= NFS4_ACCESS_EXECUTE;
	}
1789
	nfs_fattr_init(&fattr);
L
Linus Torvalds 已提交
1790 1791 1792 1793 1794 1795 1796 1797 1798
	status = rpc_call_sync(NFS_CLIENT(inode), &msg, 0);
	if (!status) {
		entry->mask = 0;
		if (res.access & NFS4_ACCESS_READ)
			entry->mask |= MAY_READ;
		if (res.access & (NFS4_ACCESS_MODIFY | NFS4_ACCESS_EXTEND | NFS4_ACCESS_DELETE))
			entry->mask |= MAY_WRITE;
		if (res.access & (NFS4_ACCESS_LOOKUP|NFS4_ACCESS_EXECUTE))
			entry->mask |= MAY_EXEC;
1799
		nfs_refresh_inode(inode, &fattr);
L
Linus Torvalds 已提交
1800 1801 1802 1803 1804 1805 1806 1807 1808 1809 1810 1811 1812 1813 1814 1815 1816 1817 1818 1819 1820 1821 1822 1823 1824 1825 1826 1827 1828 1829 1830 1831 1832 1833 1834 1835 1836 1837 1838 1839 1840 1841 1842 1843 1844 1845 1846 1847 1848 1849 1850 1851 1852 1853 1854 1855 1856 1857 1858 1859 1860 1861 1862 1863 1864 1865 1866 1867 1868 1869 1870 1871 1872 1873 1874 1875 1876 1877 1878 1879 1880 1881 1882 1883 1884 1885 1886 1887
	}
	return status;
}

static int nfs4_proc_access(struct inode *inode, struct nfs_access_entry *entry)
{
	struct nfs4_exception exception = { };
	int err;
	do {
		err = nfs4_handle_exception(NFS_SERVER(inode),
				_nfs4_proc_access(inode, entry),
				&exception);
	} while (exception.retry);
	return err;
}

/*
 * TODO: For the time being, we don't try to get any attributes
 * along with any of the zero-copy operations READ, READDIR,
 * READLINK, WRITE.
 *
 * In the case of the first three, we want to put the GETATTR
 * after the read-type operation -- this is because it is hard
 * to predict the length of a GETATTR response in v4, and thus
 * align the READ data correctly.  This means that the GETATTR
 * may end up partially falling into the page cache, and we should
 * shift it into the 'tail' of the xdr_buf before processing.
 * To do this efficiently, we need to know the total length
 * of data received, which doesn't seem to be available outside
 * of the RPC layer.
 *
 * In the case of WRITE, we also want to put the GETATTR after
 * the operation -- in this case because we want to make sure
 * we get the post-operation mtime and size.  This means that
 * we can't use xdr_encode_pages() as written: we need a variant
 * of it which would leave room in the 'tail' iovec.
 *
 * Both of these changes to the XDR layer would in fact be quite
 * minor, but I decided to leave them for a subsequent patch.
 */
static int _nfs4_proc_readlink(struct inode *inode, struct page *page,
		unsigned int pgbase, unsigned int pglen)
{
	struct nfs4_readlink args = {
		.fh       = NFS_FH(inode),
		.pgbase	  = pgbase,
		.pglen    = pglen,
		.pages    = &page,
	};
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_READLINK],
		.rpc_argp = &args,
		.rpc_resp = NULL,
	};

	return rpc_call_sync(NFS_CLIENT(inode), &msg, 0);
}

static int nfs4_proc_readlink(struct inode *inode, struct page *page,
		unsigned int pgbase, unsigned int pglen)
{
	struct nfs4_exception exception = { };
	int err;
	do {
		err = nfs4_handle_exception(NFS_SERVER(inode),
				_nfs4_proc_readlink(inode, page, pgbase, pglen),
				&exception);
	} while (exception.retry);
	return err;
}

/*
 * Got race?
 * We will need to arrange for the VFS layer to provide an atomic open.
 * Until then, this create/open method is prone to inefficiency and race
 * conditions due to the lookup, create, and open VFS calls from sys_open()
 * placed on the wire.
 *
 * Given the above sorry state of affairs, I'm simply sending an OPEN.
 * The file will be opened again in the subsequent VFS open call
 * (nfs4_proc_file_open).
 *
 * The open for read will just hang around to be used by any process that
 * opens the file O_RDONLY. This will all be resolved with the VFS changes.
 */

static int
nfs4_proc_create(struct inode *dir, struct dentry *dentry, struct iattr *sattr,
1888
                 int flags, struct nameidata *nd)
L
Linus Torvalds 已提交
1889
{
1890 1891 1892 1893
	struct path path = {
		.mnt = nd->mnt,
		.dentry = dentry,
	};
L
Linus Torvalds 已提交
1894 1895 1896 1897
	struct nfs4_state *state;
	struct rpc_cred *cred;
	int status = 0;

1898
	cred = rpcauth_lookupcred(NFS_CLIENT(dir)->cl_auth, 0);
L
Linus Torvalds 已提交
1899 1900 1901 1902
	if (IS_ERR(cred)) {
		status = PTR_ERR(cred);
		goto out;
	}
1903
	state = nfs4_do_open(dir, &path, flags, sattr, cred);
L
Linus Torvalds 已提交
1904
	put_rpccred(cred);
1905
	d_drop(dentry);
L
Linus Torvalds 已提交
1906 1907 1908 1909
	if (IS_ERR(state)) {
		status = PTR_ERR(state);
		goto out;
	}
1910
	d_add(dentry, igrab(state->inode));
1911
	nfs_set_verifier(dentry, nfs_save_change_attribute(dir));
L
Linus Torvalds 已提交
1912 1913
	if (flags & O_EXCL) {
		struct nfs_fattr fattr;
1914
		status = nfs4_do_setattr(state->inode, &fattr, sattr, state);
1915
		if (status == 0)
1916
			nfs_setattr_update_inode(state->inode, sattr);
1917
		nfs_post_op_update_inode(state->inode, &fattr);
1918
	}
1919
	if (status == 0 && (nd->flags & LOOKUP_OPEN) != 0)
1920
		status = nfs4_intent_set_file(nd, &path, state);
1921
	else
1922
		nfs4_close_sync(&path, state, flags);
L
Linus Torvalds 已提交
1923 1924 1925 1926 1927 1928
out:
	return status;
}

static int _nfs4_proc_remove(struct inode *dir, struct qstr *name)
{
1929
	struct nfs_server *server = NFS_SERVER(dir);
1930
	struct nfs_removeargs args = {
L
Linus Torvalds 已提交
1931
		.fh = NFS_FH(dir),
1932 1933
		.name.len = name->len,
		.name.name = name->name,
1934 1935
		.bitmask = server->attr_bitmask,
	};
1936
	struct nfs_removeres res = {
1937
		.server = server,
L
Linus Torvalds 已提交
1938 1939
	};
	struct rpc_message msg = {
1940 1941 1942
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_REMOVE],
		.rpc_argp = &args,
		.rpc_resp = &res,
L
Linus Torvalds 已提交
1943 1944 1945
	};
	int			status;

1946
	nfs_fattr_init(&res.dir_attr);
1947 1948 1949
	status = rpc_call_sync(server->client, &msg, 0);
	if (status == 0) {
		update_changeattr(dir, &res.cinfo);
1950
		nfs_post_op_update_inode(dir, &res.dir_attr);
1951
	}
L
Linus Torvalds 已提交
1952 1953 1954 1955 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1966
	return status;
}

static int nfs4_proc_remove(struct inode *dir, struct qstr *name)
{
	struct nfs4_exception exception = { };
	int err;
	do {
		err = nfs4_handle_exception(NFS_SERVER(dir),
				_nfs4_proc_remove(dir, name),
				&exception);
	} while (exception.retry);
	return err;
}

1967
static void nfs4_proc_unlink_setup(struct rpc_message *msg, struct inode *dir)
L
Linus Torvalds 已提交
1968
{
1969 1970 1971
	struct nfs_server *server = NFS_SERVER(dir);
	struct nfs_removeargs *args = msg->rpc_argp;
	struct nfs_removeres *res = msg->rpc_resp;
L
Linus Torvalds 已提交
1972

1973 1974
	args->bitmask = server->attr_bitmask;
	res->server = server;
L
Linus Torvalds 已提交
1975 1976 1977
	msg->rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_REMOVE];
}

1978
static int nfs4_proc_unlink_done(struct rpc_task *task, struct inode *dir)
L
Linus Torvalds 已提交
1979
{
1980 1981 1982 1983 1984 1985 1986
	struct nfs_removeres *res = task->tk_msg.rpc_resp;

	if (nfs4_async_handle_error(task, res->server) == -EAGAIN)
		return 0;
	update_changeattr(dir, &res->cinfo);
	nfs_post_op_update_inode(dir, &res->dir_attr);
	return 1;
L
Linus Torvalds 已提交
1987 1988 1989 1990 1991
}

static int _nfs4_proc_rename(struct inode *old_dir, struct qstr *old_name,
		struct inode *new_dir, struct qstr *new_name)
{
1992
	struct nfs_server *server = NFS_SERVER(old_dir);
L
Linus Torvalds 已提交
1993 1994 1995 1996 1997
	struct nfs4_rename_arg arg = {
		.old_dir = NFS_FH(old_dir),
		.new_dir = NFS_FH(new_dir),
		.old_name = old_name,
		.new_name = new_name,
1998 1999 2000 2001 2002 2003 2004
		.bitmask = server->attr_bitmask,
	};
	struct nfs_fattr old_fattr, new_fattr;
	struct nfs4_rename_res res = {
		.server = server,
		.old_fattr = &old_fattr,
		.new_fattr = &new_fattr,
L
Linus Torvalds 已提交
2005 2006 2007 2008 2009 2010 2011 2012
	};
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_RENAME],
		.rpc_argp = &arg,
		.rpc_resp = &res,
	};
	int			status;
	
2013 2014 2015
	nfs_fattr_init(res.old_fattr);
	nfs_fattr_init(res.new_fattr);
	status = rpc_call_sync(server->client, &msg, 0);
L
Linus Torvalds 已提交
2016 2017 2018

	if (!status) {
		update_changeattr(old_dir, &res.old_cinfo);
2019
		nfs_post_op_update_inode(old_dir, res.old_fattr);
L
Linus Torvalds 已提交
2020
		update_changeattr(new_dir, &res.new_cinfo);
2021
		nfs_post_op_update_inode(new_dir, res.new_fattr);
L
Linus Torvalds 已提交
2022 2023 2024 2025 2026 2027 2028 2029 2030 2031 2032 2033 2034 2035 2036 2037 2038 2039 2040 2041
	}
	return status;
}

static int nfs4_proc_rename(struct inode *old_dir, struct qstr *old_name,
		struct inode *new_dir, struct qstr *new_name)
{
	struct nfs4_exception exception = { };
	int err;
	do {
		err = nfs4_handle_exception(NFS_SERVER(old_dir),
				_nfs4_proc_rename(old_dir, old_name,
					new_dir, new_name),
				&exception);
	} while (exception.retry);
	return err;
}

static int _nfs4_proc_link(struct inode *inode, struct inode *dir, struct qstr *name)
{
2042
	struct nfs_server *server = NFS_SERVER(inode);
L
Linus Torvalds 已提交
2043 2044 2045 2046
	struct nfs4_link_arg arg = {
		.fh     = NFS_FH(inode),
		.dir_fh = NFS_FH(dir),
		.name   = name,
2047 2048 2049 2050 2051 2052 2053
		.bitmask = server->attr_bitmask,
	};
	struct nfs_fattr fattr, dir_attr;
	struct nfs4_link_res res = {
		.server = server,
		.fattr = &fattr,
		.dir_attr = &dir_attr,
L
Linus Torvalds 已提交
2054 2055 2056 2057
	};
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_LINK],
		.rpc_argp = &arg,
2058
		.rpc_resp = &res,
L
Linus Torvalds 已提交
2059 2060 2061
	};
	int			status;

2062 2063 2064 2065 2066 2067
	nfs_fattr_init(res.fattr);
	nfs_fattr_init(res.dir_attr);
	status = rpc_call_sync(server->client, &msg, 0);
	if (!status) {
		update_changeattr(dir, &res.cinfo);
		nfs_post_op_update_inode(dir, res.dir_attr);
2068
		nfs_post_op_update_inode(inode, res.fattr);
2069
	}
L
Linus Torvalds 已提交
2070 2071 2072 2073 2074 2075 2076 2077 2078 2079 2080 2081 2082 2083 2084 2085

	return status;
}

static int nfs4_proc_link(struct inode *inode, struct inode *dir, struct qstr *name)
{
	struct nfs4_exception exception = { };
	int err;
	do {
		err = nfs4_handle_exception(NFS_SERVER(inode),
				_nfs4_proc_link(inode, dir, name),
				&exception);
	} while (exception.retry);
	return err;
}

2086
static int _nfs4_proc_symlink(struct inode *dir, struct dentry *dentry,
2087
		struct page *page, unsigned int len, struct iattr *sattr)
L
Linus Torvalds 已提交
2088 2089
{
	struct nfs_server *server = NFS_SERVER(dir);
2090 2091
	struct nfs_fh fhandle;
	struct nfs_fattr fattr, dir_fattr;
L
Linus Torvalds 已提交
2092 2093 2094
	struct nfs4_create_arg arg = {
		.dir_fh = NFS_FH(dir),
		.server = server,
2095
		.name = &dentry->d_name,
L
Linus Torvalds 已提交
2096 2097 2098 2099 2100 2101
		.attrs = sattr,
		.ftype = NF4LNK,
		.bitmask = server->attr_bitmask,
	};
	struct nfs4_create_res res = {
		.server = server,
2102 2103
		.fh = &fhandle,
		.fattr = &fattr,
2104
		.dir_fattr = &dir_fattr,
L
Linus Torvalds 已提交
2105 2106 2107 2108 2109 2110 2111 2112
	};
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_SYMLINK],
		.rpc_argp = &arg,
		.rpc_resp = &res,
	};
	int			status;

2113
	if (len > NFS4_MAXPATHLEN)
L
Linus Torvalds 已提交
2114
		return -ENAMETOOLONG;
2115

2116 2117
	arg.u.symlink.pages = &page;
	arg.u.symlink.len = len;
2118
	nfs_fattr_init(&fattr);
2119
	nfs_fattr_init(&dir_fattr);
L
Linus Torvalds 已提交
2120 2121
	
	status = rpc_call_sync(NFS_CLIENT(dir), &msg, 0);
2122
	if (!status) {
L
Linus Torvalds 已提交
2123
		update_changeattr(dir, &res.dir_cinfo);
2124 2125 2126
		nfs_post_op_update_inode(dir, res.dir_fattr);
		status = nfs_instantiate(dentry, &fhandle, &fattr);
	}
L
Linus Torvalds 已提交
2127 2128 2129
	return status;
}

2130
static int nfs4_proc_symlink(struct inode *dir, struct dentry *dentry,
2131
		struct page *page, unsigned int len, struct iattr *sattr)
L
Linus Torvalds 已提交
2132 2133 2134 2135 2136
{
	struct nfs4_exception exception = { };
	int err;
	do {
		err = nfs4_handle_exception(NFS_SERVER(dir),
2137 2138
				_nfs4_proc_symlink(dir, dentry, page,
							len, sattr),
L
Linus Torvalds 已提交
2139 2140 2141 2142 2143 2144 2145 2146 2147 2148
				&exception);
	} while (exception.retry);
	return err;
}

static int _nfs4_proc_mkdir(struct inode *dir, struct dentry *dentry,
		struct iattr *sattr)
{
	struct nfs_server *server = NFS_SERVER(dir);
	struct nfs_fh fhandle;
2149
	struct nfs_fattr fattr, dir_fattr;
L
Linus Torvalds 已提交
2150 2151 2152 2153 2154 2155 2156 2157 2158 2159 2160 2161
	struct nfs4_create_arg arg = {
		.dir_fh = NFS_FH(dir),
		.server = server,
		.name = &dentry->d_name,
		.attrs = sattr,
		.ftype = NF4DIR,
		.bitmask = server->attr_bitmask,
	};
	struct nfs4_create_res res = {
		.server = server,
		.fh = &fhandle,
		.fattr = &fattr,
2162
		.dir_fattr = &dir_fattr,
L
Linus Torvalds 已提交
2163 2164 2165 2166 2167 2168 2169 2170
	};
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_CREATE],
		.rpc_argp = &arg,
		.rpc_resp = &res,
	};
	int			status;

2171
	nfs_fattr_init(&fattr);
2172
	nfs_fattr_init(&dir_fattr);
L
Linus Torvalds 已提交
2173 2174 2175 2176
	
	status = rpc_call_sync(NFS_CLIENT(dir), &msg, 0);
	if (!status) {
		update_changeattr(dir, &res.dir_cinfo);
2177
		nfs_post_op_update_inode(dir, res.dir_fattr);
L
Linus Torvalds 已提交
2178 2179 2180 2181 2182 2183 2184 2185 2186 2187 2188 2189 2190 2191 2192 2193 2194 2195 2196 2197 2198 2199 2200 2201 2202 2203 2204 2205 2206 2207 2208 2209 2210 2211 2212 2213 2214 2215
		status = nfs_instantiate(dentry, &fhandle, &fattr);
	}
	return status;
}

static int nfs4_proc_mkdir(struct inode *dir, struct dentry *dentry,
		struct iattr *sattr)
{
	struct nfs4_exception exception = { };
	int err;
	do {
		err = nfs4_handle_exception(NFS_SERVER(dir),
				_nfs4_proc_mkdir(dir, dentry, sattr),
				&exception);
	} while (exception.retry);
	return err;
}

static int _nfs4_proc_readdir(struct dentry *dentry, struct rpc_cred *cred,
                  u64 cookie, struct page *page, unsigned int count, int plus)
{
	struct inode		*dir = dentry->d_inode;
	struct nfs4_readdir_arg args = {
		.fh = NFS_FH(dir),
		.pages = &page,
		.pgbase = 0,
		.count = count,
		.bitmask = NFS_SERVER(dentry->d_inode)->attr_bitmask,
	};
	struct nfs4_readdir_res res;
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_READDIR],
		.rpc_argp = &args,
		.rpc_resp = &res,
		.rpc_cred = cred,
	};
	int			status;

2216 2217 2218 2219
	dprintk("%s: dentry = %s/%s, cookie = %Lu\n", __FUNCTION__,
			dentry->d_parent->d_name.name,
			dentry->d_name.name,
			(unsigned long long)cookie);
L
Linus Torvalds 已提交
2220 2221 2222 2223 2224
	nfs4_setup_readdir(cookie, NFS_COOKIEVERF(dir), dentry, &args);
	res.pgbase = args.pgbase;
	status = rpc_call_sync(NFS_CLIENT(dir), &msg, 0);
	if (status == 0)
		memcpy(NFS_COOKIEVERF(dir), res.verifier.data, NFS4_VERIFIER_SIZE);
2225 2226 2227

	nfs_invalidate_atime(dir);

2228
	dprintk("%s: returns %d\n", __FUNCTION__, status);
L
Linus Torvalds 已提交
2229 2230 2231 2232 2233 2234 2235 2236 2237 2238 2239 2240 2241 2242 2243 2244 2245 2246 2247 2248 2249 2250
	return status;
}

static int nfs4_proc_readdir(struct dentry *dentry, struct rpc_cred *cred,
                  u64 cookie, struct page *page, unsigned int count, int plus)
{
	struct nfs4_exception exception = { };
	int err;
	do {
		err = nfs4_handle_exception(NFS_SERVER(dentry->d_inode),
				_nfs4_proc_readdir(dentry, cred, cookie,
					page, count, plus),
				&exception);
	} while (exception.retry);
	return err;
}

static int _nfs4_proc_mknod(struct inode *dir, struct dentry *dentry,
		struct iattr *sattr, dev_t rdev)
{
	struct nfs_server *server = NFS_SERVER(dir);
	struct nfs_fh fh;
2251
	struct nfs_fattr fattr, dir_fattr;
L
Linus Torvalds 已提交
2252 2253 2254 2255 2256 2257 2258 2259 2260 2261 2262
	struct nfs4_create_arg arg = {
		.dir_fh = NFS_FH(dir),
		.server = server,
		.name = &dentry->d_name,
		.attrs = sattr,
		.bitmask = server->attr_bitmask,
	};
	struct nfs4_create_res res = {
		.server = server,
		.fh = &fh,
		.fattr = &fattr,
2263
		.dir_fattr = &dir_fattr,
L
Linus Torvalds 已提交
2264 2265 2266 2267 2268 2269 2270 2271 2272
	};
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_CREATE],
		.rpc_argp = &arg,
		.rpc_resp = &res,
	};
	int			status;
	int                     mode = sattr->ia_mode;

2273
	nfs_fattr_init(&fattr);
2274
	nfs_fattr_init(&dir_fattr);
L
Linus Torvalds 已提交
2275 2276 2277 2278 2279 2280 2281 2282 2283 2284 2285 2286 2287 2288 2289 2290 2291 2292 2293 2294 2295

	BUG_ON(!(sattr->ia_valid & ATTR_MODE));
	BUG_ON(!S_ISFIFO(mode) && !S_ISBLK(mode) && !S_ISCHR(mode) && !S_ISSOCK(mode));
	if (S_ISFIFO(mode))
		arg.ftype = NF4FIFO;
	else if (S_ISBLK(mode)) {
		arg.ftype = NF4BLK;
		arg.u.device.specdata1 = MAJOR(rdev);
		arg.u.device.specdata2 = MINOR(rdev);
	}
	else if (S_ISCHR(mode)) {
		arg.ftype = NF4CHR;
		arg.u.device.specdata1 = MAJOR(rdev);
		arg.u.device.specdata2 = MINOR(rdev);
	}
	else
		arg.ftype = NF4SOCK;
	
	status = rpc_call_sync(NFS_CLIENT(dir), &msg, 0);
	if (status == 0) {
		update_changeattr(dir, &res.dir_cinfo);
2296
		nfs_post_op_update_inode(dir, res.dir_fattr);
L
Linus Torvalds 已提交
2297 2298 2299 2300 2301 2302 2303 2304 2305 2306 2307 2308 2309 2310 2311 2312 2313 2314 2315 2316 2317 2318 2319 2320 2321 2322 2323 2324 2325 2326 2327
		status = nfs_instantiate(dentry, &fh, &fattr);
	}
	return status;
}

static int nfs4_proc_mknod(struct inode *dir, struct dentry *dentry,
		struct iattr *sattr, dev_t rdev)
{
	struct nfs4_exception exception = { };
	int err;
	do {
		err = nfs4_handle_exception(NFS_SERVER(dir),
				_nfs4_proc_mknod(dir, dentry, sattr, rdev),
				&exception);
	} while (exception.retry);
	return err;
}

static int _nfs4_proc_statfs(struct nfs_server *server, struct nfs_fh *fhandle,
		 struct nfs_fsstat *fsstat)
{
	struct nfs4_statfs_arg args = {
		.fh = fhandle,
		.bitmask = server->attr_bitmask,
	};
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_STATFS],
		.rpc_argp = &args,
		.rpc_resp = fsstat,
	};

2328
	nfs_fattr_init(fsstat->fattr);
L
Linus Torvalds 已提交
2329 2330 2331 2332 2333 2334 2335 2336 2337 2338 2339 2340 2341 2342 2343 2344 2345 2346 2347 2348 2349 2350 2351 2352 2353 2354 2355 2356 2357 2358 2359 2360 2361 2362 2363 2364 2365 2366 2367 2368 2369 2370 2371 2372 2373 2374
	return rpc_call_sync(server->client, &msg, 0);
}

static int nfs4_proc_statfs(struct nfs_server *server, struct nfs_fh *fhandle, struct nfs_fsstat *fsstat)
{
	struct nfs4_exception exception = { };
	int err;
	do {
		err = nfs4_handle_exception(server,
				_nfs4_proc_statfs(server, fhandle, fsstat),
				&exception);
	} while (exception.retry);
	return err;
}

static int _nfs4_do_fsinfo(struct nfs_server *server, struct nfs_fh *fhandle,
		struct nfs_fsinfo *fsinfo)
{
	struct nfs4_fsinfo_arg args = {
		.fh = fhandle,
		.bitmask = server->attr_bitmask,
	};
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_FSINFO],
		.rpc_argp = &args,
		.rpc_resp = fsinfo,
	};

	return rpc_call_sync(server->client, &msg, 0);
}

static int nfs4_do_fsinfo(struct nfs_server *server, struct nfs_fh *fhandle, struct nfs_fsinfo *fsinfo)
{
	struct nfs4_exception exception = { };
	int err;

	do {
		err = nfs4_handle_exception(server,
				_nfs4_do_fsinfo(server, fhandle, fsinfo),
				&exception);
	} while (exception.retry);
	return err;
}

static int nfs4_proc_fsinfo(struct nfs_server *server, struct nfs_fh *fhandle, struct nfs_fsinfo *fsinfo)
{
2375
	nfs_fattr_init(fsinfo->fattr);
L
Linus Torvalds 已提交
2376 2377 2378 2379 2380 2381 2382 2383 2384 2385 2386 2387 2388 2389 2390 2391 2392 2393 2394 2395 2396 2397
	return nfs4_do_fsinfo(server, fhandle, fsinfo);
}

static int _nfs4_proc_pathconf(struct nfs_server *server, struct nfs_fh *fhandle,
		struct nfs_pathconf *pathconf)
{
	struct nfs4_pathconf_arg args = {
		.fh = fhandle,
		.bitmask = server->attr_bitmask,
	};
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_PATHCONF],
		.rpc_argp = &args,
		.rpc_resp = pathconf,
	};

	/* None of the pathconf attributes are mandatory to implement */
	if ((args.bitmask[0] & nfs4_pathconf_bitmap[0]) == 0) {
		memset(pathconf, 0, sizeof(*pathconf));
		return 0;
	}

2398
	nfs_fattr_init(pathconf->fattr);
L
Linus Torvalds 已提交
2399 2400 2401 2402 2403 2404 2405 2406 2407 2408 2409 2410 2411 2412 2413 2414 2415
	return rpc_call_sync(server->client, &msg, 0);
}

static int nfs4_proc_pathconf(struct nfs_server *server, struct nfs_fh *fhandle,
		struct nfs_pathconf *pathconf)
{
	struct nfs4_exception exception = { };
	int err;

	do {
		err = nfs4_handle_exception(server,
				_nfs4_proc_pathconf(server, fhandle, pathconf),
				&exception);
	} while (exception.retry);
	return err;
}

T
Trond Myklebust 已提交
2416
static int nfs4_read_done(struct rpc_task *task, struct nfs_read_data *data)
L
Linus Torvalds 已提交
2417
{
T
Trond Myklebust 已提交
2418
	struct nfs_server *server = NFS_SERVER(data->inode);
L
Linus Torvalds 已提交
2419

T
Trond Myklebust 已提交
2420
	if (nfs4_async_handle_error(task, server) == -EAGAIN) {
L
Linus Torvalds 已提交
2421
		rpc_restart_call(task);
T
Trond Myklebust 已提交
2422
		return -EAGAIN;
L
Linus Torvalds 已提交
2423
	}
2424 2425

	nfs_invalidate_atime(data->inode);
L
Linus Torvalds 已提交
2426
	if (task->tk_status > 0)
T
Trond Myklebust 已提交
2427 2428
		renew_lease(server, data->timestamp);
	return 0;
L
Linus Torvalds 已提交
2429 2430
}

2431
static void nfs4_proc_read_setup(struct nfs_read_data *data, struct rpc_message *msg)
L
Linus Torvalds 已提交
2432 2433
{
	data->timestamp   = jiffies;
2434
	msg->rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_READ];
L
Linus Torvalds 已提交
2435 2436
}

2437
static int nfs4_write_done(struct rpc_task *task, struct nfs_write_data *data)
L
Linus Torvalds 已提交
2438 2439 2440 2441 2442
{
	struct inode *inode = data->inode;
	
	if (nfs4_async_handle_error(task, NFS_SERVER(inode)) == -EAGAIN) {
		rpc_restart_call(task);
2443
		return -EAGAIN;
L
Linus Torvalds 已提交
2444
	}
2445
	if (task->tk_status >= 0) {
L
Linus Torvalds 已提交
2446
		renew_lease(NFS_SERVER(inode), data->timestamp);
2447
		nfs_post_op_update_inode_force_wcc(inode, data->res.fattr);
2448
	}
2449
	return 0;
L
Linus Torvalds 已提交
2450 2451
}

2452
static void nfs4_proc_write_setup(struct nfs_write_data *data, struct rpc_message *msg)
L
Linus Torvalds 已提交
2453
{
2454 2455
	struct nfs_server *server = NFS_SERVER(data->inode);

2456 2457
	data->args.bitmask = server->attr_bitmask;
	data->res.server = server;
L
Linus Torvalds 已提交
2458 2459
	data->timestamp   = jiffies;

2460
	msg->rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_WRITE];
L
Linus Torvalds 已提交
2461 2462
}

2463
static int nfs4_commit_done(struct rpc_task *task, struct nfs_write_data *data)
L
Linus Torvalds 已提交
2464 2465 2466 2467 2468
{
	struct inode *inode = data->inode;
	
	if (nfs4_async_handle_error(task, NFS_SERVER(inode)) == -EAGAIN) {
		rpc_restart_call(task);
2469
		return -EAGAIN;
L
Linus Torvalds 已提交
2470
	}
2471
	nfs_refresh_inode(inode, data->res.fattr);
2472
	return 0;
L
Linus Torvalds 已提交
2473 2474
}

2475
static void nfs4_proc_commit_setup(struct nfs_write_data *data, struct rpc_message *msg)
L
Linus Torvalds 已提交
2476
{
2477
	struct nfs_server *server = NFS_SERVER(data->inode);
L
Linus Torvalds 已提交
2478
	
2479 2480
	data->args.bitmask = server->attr_bitmask;
	data->res.server = server;
2481
	msg->rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_COMMIT];
L
Linus Torvalds 已提交
2482 2483 2484 2485 2486 2487
}

/*
 * nfs4_proc_async_renew(): This is not one of the nfs_rpc_ops; it is a special
 * standalone procedure for queueing an asynchronous RENEW.
 */
2488
static void nfs4_renew_done(struct rpc_task *task, void *data)
L
Linus Torvalds 已提交
2489
{
2490
	struct nfs_client *clp = (struct nfs_client *)task->tk_msg.rpc_argp;
2491
	unsigned long timestamp = (unsigned long)data;
L
Linus Torvalds 已提交
2492 2493 2494 2495 2496 2497 2498 2499 2500 2501 2502 2503 2504 2505 2506 2507

	if (task->tk_status < 0) {
		switch (task->tk_status) {
			case -NFS4ERR_STALE_CLIENTID:
			case -NFS4ERR_EXPIRED:
			case -NFS4ERR_CB_PATH_DOWN:
				nfs4_schedule_state_recovery(clp);
		}
		return;
	}
	spin_lock(&clp->cl_lock);
	if (time_before(clp->cl_last_renewal,timestamp))
		clp->cl_last_renewal = timestamp;
	spin_unlock(&clp->cl_lock);
}

2508 2509 2510 2511
static const struct rpc_call_ops nfs4_renew_ops = {
	.rpc_call_done = nfs4_renew_done,
};

2512
int nfs4_proc_async_renew(struct nfs_client *clp, struct rpc_cred *cred)
L
Linus Torvalds 已提交
2513 2514 2515 2516
{
	struct rpc_message msg = {
		.rpc_proc	= &nfs4_procedures[NFSPROC4_CLNT_RENEW],
		.rpc_argp	= clp,
2517
		.rpc_cred	= cred,
L
Linus Torvalds 已提交
2518 2519 2520
	};

	return rpc_call_async(clp->cl_rpcclient, &msg, RPC_TASK_SOFT,
2521
			&nfs4_renew_ops, (void *)jiffies);
L
Linus Torvalds 已提交
2522 2523
}

2524
int nfs4_proc_renew(struct nfs_client *clp, struct rpc_cred *cred)
L
Linus Torvalds 已提交
2525 2526 2527 2528
{
	struct rpc_message msg = {
		.rpc_proc	= &nfs4_procedures[NFSPROC4_CLNT_RENEW],
		.rpc_argp	= clp,
2529
		.rpc_cred	= cred,
L
Linus Torvalds 已提交
2530 2531 2532 2533 2534 2535 2536 2537 2538 2539 2540 2541 2542 2543
	};
	unsigned long now = jiffies;
	int status;

	status = rpc_call_sync(clp->cl_rpcclient, &msg, 0);
	if (status < 0)
		return status;
	spin_lock(&clp->cl_lock);
	if (time_before(clp->cl_last_renewal,now))
		clp->cl_last_renewal = now;
	spin_unlock(&clp->cl_lock);
	return 0;
}

2544 2545 2546 2547 2548 2549 2550 2551 2552 2553 2554 2555 2556 2557 2558 2559 2560 2561 2562 2563 2564 2565 2566 2567 2568 2569
static inline int nfs4_server_supports_acls(struct nfs_server *server)
{
	return (server->caps & NFS_CAP_ACLS)
		&& (server->acl_bitmask & ACL4_SUPPORT_ALLOW_ACL)
		&& (server->acl_bitmask & ACL4_SUPPORT_DENY_ACL);
}

/* Assuming that XATTR_SIZE_MAX is a multiple of PAGE_CACHE_SIZE, and that
 * it's OK to put sizeof(void) * (XATTR_SIZE_MAX/PAGE_CACHE_SIZE) bytes on
 * the stack.
 */
#define NFS4ACL_MAXPAGES (XATTR_SIZE_MAX >> PAGE_CACHE_SHIFT)

static void buf_to_pages(const void *buf, size_t buflen,
		struct page **pages, unsigned int *pgbase)
{
	const void *p = buf;

	*pgbase = offset_in_page(buf);
	p -= *pgbase;
	while (p < buf + buflen) {
		*(pages++) = virt_to_page(p);
		p += PAGE_CACHE_SIZE;
	}
}

2570 2571 2572
struct nfs4_cached_acl {
	int cached;
	size_t len;
A
Andrew Morton 已提交
2573
	char data[0];
2574 2575 2576 2577 2578 2579 2580 2581 2582 2583 2584 2585 2586 2587 2588 2589 2590 2591 2592 2593 2594 2595 2596 2597 2598 2599 2600 2601 2602 2603 2604 2605 2606 2607 2608 2609 2610 2611 2612 2613 2614 2615 2616 2617 2618 2619 2620 2621 2622 2623 2624 2625 2626 2627 2628 2629 2630 2631 2632 2633 2634 2635 2636
};

static void nfs4_set_cached_acl(struct inode *inode, struct nfs4_cached_acl *acl)
{
	struct nfs_inode *nfsi = NFS_I(inode);

	spin_lock(&inode->i_lock);
	kfree(nfsi->nfs4_acl);
	nfsi->nfs4_acl = acl;
	spin_unlock(&inode->i_lock);
}

static void nfs4_zap_acl_attr(struct inode *inode)
{
	nfs4_set_cached_acl(inode, NULL);
}

static inline ssize_t nfs4_read_cached_acl(struct inode *inode, char *buf, size_t buflen)
{
	struct nfs_inode *nfsi = NFS_I(inode);
	struct nfs4_cached_acl *acl;
	int ret = -ENOENT;

	spin_lock(&inode->i_lock);
	acl = nfsi->nfs4_acl;
	if (acl == NULL)
		goto out;
	if (buf == NULL) /* user is just asking for length */
		goto out_len;
	if (acl->cached == 0)
		goto out;
	ret = -ERANGE; /* see getxattr(2) man page */
	if (acl->len > buflen)
		goto out;
	memcpy(buf, acl->data, acl->len);
out_len:
	ret = acl->len;
out:
	spin_unlock(&inode->i_lock);
	return ret;
}

static void nfs4_write_cached_acl(struct inode *inode, const char *buf, size_t acl_len)
{
	struct nfs4_cached_acl *acl;

	if (buf && acl_len <= PAGE_SIZE) {
		acl = kmalloc(sizeof(*acl) + acl_len, GFP_KERNEL);
		if (acl == NULL)
			goto out;
		acl->cached = 1;
		memcpy(acl->data, buf, acl_len);
	} else {
		acl = kmalloc(sizeof(*acl), GFP_KERNEL);
		if (acl == NULL)
			goto out;
		acl->cached = 0;
	}
	acl->len = acl_len;
out:
	nfs4_set_cached_acl(inode, acl);
}

2637
static ssize_t __nfs4_get_acl_uncached(struct inode *inode, void *buf, size_t buflen)
2638 2639 2640 2641 2642 2643 2644 2645
{
	struct page *pages[NFS4ACL_MAXPAGES];
	struct nfs_getaclargs args = {
		.fh = NFS_FH(inode),
		.acl_pages = pages,
		.acl_len = buflen,
	};
	size_t resp_len = buflen;
2646
	void *resp_buf;
2647 2648 2649 2650 2651
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_GETACL],
		.rpc_argp = &args,
		.rpc_resp = &resp_len,
	};
2652
	struct page *localpage = NULL;
2653 2654
	int ret;

2655 2656 2657 2658 2659 2660 2661 2662 2663
	if (buflen < PAGE_SIZE) {
		/* As long as we're doing a round trip to the server anyway,
		 * let's be prepared for a page of acl data. */
		localpage = alloc_page(GFP_KERNEL);
		resp_buf = page_address(localpage);
		if (localpage == NULL)
			return -ENOMEM;
		args.acl_pages[0] = localpage;
		args.acl_pgbase = 0;
J
J. Bruce Fields 已提交
2664
		resp_len = args.acl_len = PAGE_SIZE;
2665 2666 2667 2668
	} else {
		resp_buf = buf;
		buf_to_pages(buf, buflen, args.acl_pages, &args.acl_pgbase);
	}
2669
	ret = rpc_call_sync(NFS_CLIENT(inode), &msg, 0);
2670 2671 2672 2673 2674 2675 2676 2677 2678 2679 2680 2681 2682 2683 2684 2685 2686
	if (ret)
		goto out_free;
	if (resp_len > args.acl_len)
		nfs4_write_cached_acl(inode, NULL, resp_len);
	else
		nfs4_write_cached_acl(inode, resp_buf, resp_len);
	if (buf) {
		ret = -ERANGE;
		if (resp_len > buflen)
			goto out_free;
		if (localpage)
			memcpy(buf, resp_buf, resp_len);
	}
	ret = resp_len;
out_free:
	if (localpage)
		__free_page(localpage);
2687 2688 2689
	return ret;
}

2690 2691 2692 2693 2694 2695 2696 2697 2698 2699 2700 2701 2702
static ssize_t nfs4_get_acl_uncached(struct inode *inode, void *buf, size_t buflen)
{
	struct nfs4_exception exception = { };
	ssize_t ret;
	do {
		ret = __nfs4_get_acl_uncached(inode, buf, buflen);
		if (ret >= 0)
			break;
		ret = nfs4_handle_exception(NFS_SERVER(inode), ret, &exception);
	} while (exception.retry);
	return ret;
}

2703 2704 2705 2706 2707 2708 2709 2710 2711 2712 2713 2714 2715 2716 2717 2718
static ssize_t nfs4_proc_get_acl(struct inode *inode, void *buf, size_t buflen)
{
	struct nfs_server *server = NFS_SERVER(inode);
	int ret;

	if (!nfs4_server_supports_acls(server))
		return -EOPNOTSUPP;
	ret = nfs_revalidate_inode(server, inode);
	if (ret < 0)
		return ret;
	ret = nfs4_read_cached_acl(inode, buf, buflen);
	if (ret != -ENOENT)
		return ret;
	return nfs4_get_acl_uncached(inode, buf, buflen);
}

2719
static int __nfs4_proc_set_acl(struct inode *inode, const void *buf, size_t buflen)
2720 2721 2722 2723 2724 2725 2726 2727 2728 2729 2730 2731 2732 2733 2734 2735 2736
{
	struct nfs_server *server = NFS_SERVER(inode);
	struct page *pages[NFS4ACL_MAXPAGES];
	struct nfs_setaclargs arg = {
		.fh		= NFS_FH(inode),
		.acl_pages	= pages,
		.acl_len	= buflen,
	};
	struct rpc_message msg = {
		.rpc_proc	= &nfs4_procedures[NFSPROC4_CLNT_SETACL],
		.rpc_argp	= &arg,
		.rpc_resp	= NULL,
	};
	int ret;

	if (!nfs4_server_supports_acls(server))
		return -EOPNOTSUPP;
2737
	nfs_inode_return_delegation(inode);
2738
	buf_to_pages(buf, buflen, arg.acl_pages, &arg.acl_pgbase);
2739
	ret = rpc_call_sync(NFS_CLIENT(inode), &msg, 0);
2740
	nfs_zap_caches(inode);
2741 2742 2743
	return ret;
}

2744 2745 2746 2747 2748 2749 2750 2751 2752 2753 2754 2755
static int nfs4_proc_set_acl(struct inode *inode, const void *buf, size_t buflen)
{
	struct nfs4_exception exception = { };
	int err;
	do {
		err = nfs4_handle_exception(NFS_SERVER(inode),
				__nfs4_proc_set_acl(inode, buf, buflen),
				&exception);
	} while (exception.retry);
	return err;
}

L
Linus Torvalds 已提交
2756
static int
2757
nfs4_async_handle_error(struct rpc_task *task, const struct nfs_server *server)
L
Linus Torvalds 已提交
2758
{
2759
	struct nfs_client *clp = server->nfs_client;
L
Linus Torvalds 已提交
2760 2761 2762 2763 2764 2765 2766 2767 2768

	if (!clp || task->tk_status >= 0)
		return 0;
	switch(task->tk_status) {
		case -NFS4ERR_STALE_CLIENTID:
		case -NFS4ERR_STALE_STATEID:
		case -NFS4ERR_EXPIRED:
			rpc_sleep_on(&clp->cl_rpcwaitq, task, NULL, NULL);
			nfs4_schedule_state_recovery(clp);
T
Trond Myklebust 已提交
2769
			if (test_bit(NFS4CLNT_STATE_RECOVER, &clp->cl_state) == 0)
L
Linus Torvalds 已提交
2770 2771 2772 2773
				rpc_wake_up_task(task);
			task->tk_status = 0;
			return -EAGAIN;
		case -NFS4ERR_DELAY:
2774 2775 2776
			nfs_inc_server_stats((struct nfs_server *) server,
						NFSIOS_DELAY);
		case -NFS4ERR_GRACE:
L
Linus Torvalds 已提交
2777 2778 2779 2780 2781 2782 2783 2784 2785 2786 2787
			rpc_delay(task, NFS4_POLL_RETRY_MAX);
			task->tk_status = 0;
			return -EAGAIN;
		case -NFS4ERR_OLD_STATEID:
			task->tk_status = 0;
			return -EAGAIN;
	}
	task->tk_status = nfs4_map_errors(task->tk_status);
	return 0;
}

T
Trond Myklebust 已提交
2788 2789 2790 2791 2792 2793 2794 2795
static int nfs4_wait_bit_interruptible(void *word)
{
	if (signal_pending(current))
		return -ERESTARTSYS;
	schedule();
	return 0;
}

2796
static int nfs4_wait_clnt_recover(struct rpc_clnt *clnt, struct nfs_client *clp)
L
Linus Torvalds 已提交
2797 2798
{
	sigset_t oldset;
T
Trond Myklebust 已提交
2799
	int res;
L
Linus Torvalds 已提交
2800 2801 2802

	might_sleep();

2803 2804
	rwsem_acquire(&clp->cl_sem.dep_map, 0, 0, _RET_IP_);

L
Linus Torvalds 已提交
2805
	rpc_clnt_sigmask(clnt, &oldset);
T
Trond Myklebust 已提交
2806 2807 2808
	res = wait_on_bit(&clp->cl_state, NFS4CLNT_STATE_RECOVER,
			nfs4_wait_bit_interruptible,
			TASK_INTERRUPTIBLE);
L
Linus Torvalds 已提交
2809
	rpc_clnt_sigunmask(clnt, &oldset);
2810 2811

	rwsem_release(&clp->cl_sem.dep_map, 1, _RET_IP_);
L
Linus Torvalds 已提交
2812 2813 2814 2815 2816 2817 2818 2819 2820 2821 2822 2823 2824 2825 2826 2827
	return res;
}

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

	might_sleep();

	if (*timeout <= 0)
		*timeout = NFS4_POLL_RETRY_MIN;
	if (*timeout > NFS4_POLL_RETRY_MAX)
		*timeout = NFS4_POLL_RETRY_MAX;
	rpc_clnt_sigmask(clnt, &oldset);
	if (clnt->cl_intr) {
2828
		schedule_timeout_interruptible(*timeout);
L
Linus Torvalds 已提交
2829 2830
		if (signalled())
			res = -ERESTARTSYS;
2831 2832
	} else
		schedule_timeout_uninterruptible(*timeout);
L
Linus Torvalds 已提交
2833 2834 2835 2836 2837 2838 2839 2840
	rpc_clnt_sigunmask(clnt, &oldset);
	*timeout <<= 1;
	return res;
}

/* This is the error handling routine for processes that are allowed
 * to sleep.
 */
2841
static int nfs4_handle_exception(const struct nfs_server *server, int errorcode, struct nfs4_exception *exception)
L
Linus Torvalds 已提交
2842
{
2843
	struct nfs_client *clp = server->nfs_client;
L
Linus Torvalds 已提交
2844 2845 2846 2847 2848 2849 2850 2851 2852
	int ret = errorcode;

	exception->retry = 0;
	switch(errorcode) {
		case 0:
			return 0;
		case -NFS4ERR_STALE_CLIENTID:
		case -NFS4ERR_STALE_STATEID:
		case -NFS4ERR_EXPIRED:
T
Trond Myklebust 已提交
2853
			nfs4_schedule_state_recovery(clp);
L
Linus Torvalds 已提交
2854 2855 2856 2857
			ret = nfs4_wait_clnt_recover(server->client, clp);
			if (ret == 0)
				exception->retry = 1;
			break;
2858
		case -NFS4ERR_FILE_OPEN:
L
Linus Torvalds 已提交
2859 2860 2861
		case -NFS4ERR_GRACE:
		case -NFS4ERR_DELAY:
			ret = nfs4_delay(server->client, &exception->timeout);
2862 2863
			if (ret != 0)
				break;
L
Linus Torvalds 已提交
2864
		case -NFS4ERR_OLD_STATEID:
2865
			exception->retry = 1;
L
Linus Torvalds 已提交
2866 2867 2868 2869 2870
	}
	/* We failed to handle the error */
	return nfs4_map_errors(ret);
}

2871
int nfs4_proc_setclientid(struct nfs_client *clp, u32 program, unsigned short port, struct rpc_cred *cred)
L
Linus Torvalds 已提交
2872 2873 2874 2875 2876 2877 2878 2879 2880 2881
{
	nfs4_verifier sc_verifier;
	struct nfs4_setclientid setclientid = {
		.sc_verifier = &sc_verifier,
		.sc_prog = program,
	};
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_SETCLIENTID],
		.rpc_argp = &setclientid,
		.rpc_resp = clp,
2882
		.rpc_cred = cred,
L
Linus Torvalds 已提交
2883
	};
A
Al Viro 已提交
2884
	__be32 *p;
L
Linus Torvalds 已提交
2885 2886 2887
	int loop = 0;
	int status;

A
Al Viro 已提交
2888
	p = (__be32*)sc_verifier.data;
L
Linus Torvalds 已提交
2889 2890 2891 2892 2893 2894
	*p++ = htonl((u32)clp->cl_boot_time.tv_sec);
	*p = htonl((u32)clp->cl_boot_time.tv_nsec);

	for(;;) {
		setclientid.sc_name_len = scnprintf(setclientid.sc_name,
				sizeof(setclientid.sc_name), "%s/%u.%u.%u.%u %s %u",
2895
				clp->cl_ipaddr, NIPQUAD(clp->cl_addr.sin_addr),
2896
				cred->cr_ops->cr_name,
L
Linus Torvalds 已提交
2897 2898 2899 2900 2901 2902 2903 2904 2905 2906 2907 2908 2909 2910 2911 2912 2913 2914 2915 2916 2917
				clp->cl_id_uniquifier);
		setclientid.sc_netid_len = scnprintf(setclientid.sc_netid,
				sizeof(setclientid.sc_netid), "tcp");
		setclientid.sc_uaddr_len = scnprintf(setclientid.sc_uaddr,
				sizeof(setclientid.sc_uaddr), "%s.%d.%d",
				clp->cl_ipaddr, port >> 8, port & 255);

		status = rpc_call_sync(clp->cl_rpcclient, &msg, 0);
		if (status != -NFS4ERR_CLID_INUSE)
			break;
		if (signalled())
			break;
		if (loop++ & 1)
			ssleep(clp->cl_lease_time + 1);
		else
			if (++clp->cl_id_uniquifier == 0)
				break;
	}
	return status;
}

2918
static int _nfs4_proc_setclientid_confirm(struct nfs_client *clp, struct rpc_cred *cred)
L
Linus Torvalds 已提交
2919 2920 2921 2922 2923 2924
{
	struct nfs_fsinfo fsinfo;
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_SETCLIENTID_CONFIRM],
		.rpc_argp = clp,
		.rpc_resp = &fsinfo,
2925
		.rpc_cred = cred,
L
Linus Torvalds 已提交
2926 2927 2928 2929 2930 2931 2932 2933 2934 2935
	};
	unsigned long now;
	int status;

	now = jiffies;
	status = rpc_call_sync(clp->cl_rpcclient, &msg, 0);
	if (status == 0) {
		spin_lock(&clp->cl_lock);
		clp->cl_lease_time = fsinfo.lease_time * HZ;
		clp->cl_last_renewal = now;
2936
		clear_bit(NFS4CLNT_LEASE_EXPIRED, &clp->cl_state);
L
Linus Torvalds 已提交
2937 2938 2939 2940 2941
		spin_unlock(&clp->cl_lock);
	}
	return status;
}

2942
int nfs4_proc_setclientid_confirm(struct nfs_client *clp, struct rpc_cred *cred)
2943 2944 2945 2946 2947 2948 2949 2950 2951 2952 2953 2954 2955 2956 2957 2958 2959
{
	long timeout;
	int err;
	do {
		err = _nfs4_proc_setclientid_confirm(clp, cred);
		switch (err) {
			case 0:
				return err;
			case -NFS4ERR_RESOURCE:
				/* The IBM lawyers misread another document! */
			case -NFS4ERR_DELAY:
				err = nfs4_delay(clp->cl_rpcclient, &timeout);
		}
	} while (err == 0);
	return err;
}

2960 2961
struct nfs4_delegreturndata {
	struct nfs4_delegreturnargs args;
2962
	struct nfs4_delegreturnres res;
2963 2964
	struct nfs_fh fh;
	nfs4_stateid stateid;
2965
	unsigned long timestamp;
2966
	struct nfs_fattr fattr;
2967 2968 2969 2970 2971 2972 2973
	int rpc_status;
};

static void nfs4_delegreturn_done(struct rpc_task *task, void *calldata)
{
	struct nfs4_delegreturndata *data = calldata;
	data->rpc_status = task->tk_status;
2974
	if (data->rpc_status == 0)
2975
		renew_lease(data->res.server, data->timestamp);
2976 2977 2978 2979 2980 2981 2982
}

static void nfs4_delegreturn_release(void *calldata)
{
	kfree(calldata);
}

2983
static const struct rpc_call_ops nfs4_delegreturn_ops = {
2984 2985 2986 2987 2988 2989 2990
	.rpc_call_done = nfs4_delegreturn_done,
	.rpc_release = nfs4_delegreturn_release,
};

static int _nfs4_proc_delegreturn(struct inode *inode, struct rpc_cred *cred, const nfs4_stateid *stateid)
{
	struct nfs4_delegreturndata *data;
2991
	struct nfs_server *server = NFS_SERVER(inode);
2992
	struct rpc_task *task;
2993 2994 2995 2996
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_DELEGRETURN],
		.rpc_cred = cred,
	};
T
Trond Myklebust 已提交
2997 2998
	struct rpc_task_setup task_setup_data = {
		.rpc_client = server->client,
2999
		.rpc_message = &msg,
T
Trond Myklebust 已提交
3000 3001 3002
		.callback_ops = &nfs4_delegreturn_ops,
		.flags = RPC_TASK_ASYNC,
	};
3003 3004 3005 3006 3007 3008 3009
	int status;

	data = kmalloc(sizeof(*data), GFP_KERNEL);
	if (data == NULL)
		return -ENOMEM;
	data->args.fhandle = &data->fh;
	data->args.stateid = &data->stateid;
3010
	data->args.bitmask = server->attr_bitmask;
3011 3012
	nfs_copy_fh(&data->fh, NFS_FH(inode));
	memcpy(&data->stateid, stateid, sizeof(data->stateid));
3013 3014
	data->res.fattr = &data->fattr;
	data->res.server = server;
3015
	nfs_fattr_init(data->res.fattr);
3016
	data->timestamp = jiffies;
3017 3018
	data->rpc_status = 0;

T
Trond Myklebust 已提交
3019
	task_setup_data.callback_data = data;
3020 3021
	msg.rpc_argp = &data->args,
	msg.rpc_resp = &data->res,
T
Trond Myklebust 已提交
3022
	task = rpc_run_task(&task_setup_data);
3023
	if (IS_ERR(task))
3024 3025
		return PTR_ERR(task);
	status = nfs4_wait_for_completion_rpc_task(task);
3026
	if (status == 0) {
3027
		status = data->rpc_status;
3028
		if (status == 0)
3029
			nfs_refresh_inode(inode, &data->fattr);
3030
	}
3031
	rpc_put_task(task);
3032
	return status;
L
Linus Torvalds 已提交
3033 3034 3035 3036 3037 3038 3039 3040 3041 3042 3043 3044 3045 3046 3047 3048 3049 3050 3051 3052 3053 3054 3055 3056 3057 3058 3059 3060 3061
}

int nfs4_proc_delegreturn(struct inode *inode, struct rpc_cred *cred, const nfs4_stateid *stateid)
{
	struct nfs_server *server = NFS_SERVER(inode);
	struct nfs4_exception exception = { };
	int err;
	do {
		err = _nfs4_proc_delegreturn(inode, cred, stateid);
		switch (err) {
			case -NFS4ERR_STALE_STATEID:
			case -NFS4ERR_EXPIRED:
			case 0:
				return 0;
		}
		err = nfs4_handle_exception(server, err, &exception);
	} while (exception.retry);
	return err;
}

#define NFS4_LOCK_MINTIMEOUT (1 * HZ)
#define NFS4_LOCK_MAXTIMEOUT (30 * HZ)

/* 
 * sleep, with exponential backoff, and retry the LOCK operation. 
 */
static unsigned long
nfs4_set_lock_task_retry(unsigned long timeout)
{
3062
	schedule_timeout_interruptible(timeout);
L
Linus Torvalds 已提交
3063 3064 3065 3066 3067 3068 3069 3070 3071 3072
	timeout <<= 1;
	if (timeout > NFS4_LOCK_MAXTIMEOUT)
		return NFS4_LOCK_MAXTIMEOUT;
	return timeout;
}

static int _nfs4_proc_getlk(struct nfs4_state *state, int cmd, struct file_lock *request)
{
	struct inode *inode = state->inode;
	struct nfs_server *server = NFS_SERVER(inode);
3073
	struct nfs_client *clp = server->nfs_client;
T
Trond Myklebust 已提交
3074
	struct nfs_lockt_args arg = {
L
Linus Torvalds 已提交
3075
		.fh = NFS_FH(inode),
T
Trond Myklebust 已提交
3076
		.fl = request,
L
Linus Torvalds 已提交
3077
	};
T
Trond Myklebust 已提交
3078 3079
	struct nfs_lockt_res res = {
		.denied = request,
L
Linus Torvalds 已提交
3080 3081 3082 3083 3084 3085 3086 3087 3088 3089 3090
	};
	struct rpc_message msg = {
		.rpc_proc	= &nfs4_procedures[NFSPROC4_CLNT_LOCKT],
		.rpc_argp       = &arg,
		.rpc_resp       = &res,
		.rpc_cred	= state->owner->so_cred,
	};
	struct nfs4_lock_state *lsp;
	int status;

	down_read(&clp->cl_sem);
T
Trond Myklebust 已提交
3091
	arg.lock_owner.clientid = clp->cl_clientid;
3092 3093 3094 3095
	status = nfs4_set_lock_state(state, request);
	if (status != 0)
		goto out;
	lsp = request->fl_u.nfs4_fl.owner;
3096
	arg.lock_owner.id = lsp->ls_id.id;
L
Linus Torvalds 已提交
3097
	status = rpc_call_sync(server->client, &msg, 0);
T
Trond Myklebust 已提交
3098 3099 3100 3101 3102 3103
	switch (status) {
		case 0:
			request->fl_type = F_UNLCK;
			break;
		case -NFS4ERR_DENIED:
			status = 0;
L
Linus Torvalds 已提交
3104
	}
3105
	request->fl_ops->fl_release_private(request);
3106
out:
L
Linus Torvalds 已提交
3107 3108 3109 3110 3111 3112 3113 3114 3115 3116 3117 3118 3119 3120 3121 3122 3123 3124 3125 3126 3127 3128 3129 3130 3131 3132 3133 3134 3135 3136 3137 3138 3139
	up_read(&clp->cl_sem);
	return status;
}

static int nfs4_proc_getlk(struct nfs4_state *state, int cmd, struct file_lock *request)
{
	struct nfs4_exception exception = { };
	int err;

	do {
		err = nfs4_handle_exception(NFS_SERVER(state->inode),
				_nfs4_proc_getlk(state, cmd, request),
				&exception);
	} while (exception.retry);
	return err;
}

static int do_vfs_lock(struct file *file, struct file_lock *fl)
{
	int res = 0;
	switch (fl->fl_flags & (FL_POSIX|FL_FLOCK)) {
		case FL_POSIX:
			res = posix_lock_file_wait(file, fl);
			break;
		case FL_FLOCK:
			res = flock_lock_file_wait(file, fl);
			break;
		default:
			BUG();
	}
	return res;
}

3140
struct nfs4_unlockdata {
T
Trond Myklebust 已提交
3141 3142
	struct nfs_locku_args arg;
	struct nfs_locku_res res;
3143 3144
	struct nfs4_lock_state *lsp;
	struct nfs_open_context *ctx;
T
Trond Myklebust 已提交
3145 3146
	struct file_lock fl;
	const struct nfs_server *server;
3147
	unsigned long timestamp;
3148 3149
};

T
Trond Myklebust 已提交
3150 3151 3152 3153 3154 3155 3156 3157 3158 3159 3160 3161 3162 3163 3164 3165 3166 3167 3168 3169 3170 3171 3172 3173
static struct nfs4_unlockdata *nfs4_alloc_unlockdata(struct file_lock *fl,
		struct nfs_open_context *ctx,
		struct nfs4_lock_state *lsp,
		struct nfs_seqid *seqid)
{
	struct nfs4_unlockdata *p;
	struct inode *inode = lsp->ls_state->inode;

	p = kmalloc(sizeof(*p), GFP_KERNEL);
	if (p == NULL)
		return NULL;
	p->arg.fh = NFS_FH(inode);
	p->arg.fl = &p->fl;
	p->arg.seqid = seqid;
	p->arg.stateid = &lsp->ls_stateid;
	p->lsp = lsp;
	atomic_inc(&lsp->ls_count);
	/* Ensure we don't close file until we're done freeing locks! */
	p->ctx = get_nfs_open_context(ctx);
	memcpy(&p->fl, fl, sizeof(p->fl));
	p->server = NFS_SERVER(inode);
	return p;
}

3174
static void nfs4_locku_release_calldata(void *data)
3175
{
3176
	struct nfs4_unlockdata *calldata = data;
T
Trond Myklebust 已提交
3177
	nfs_free_seqid(calldata->arg.seqid);
3178 3179 3180
	nfs4_put_lock_state(calldata->lsp);
	put_nfs_open_context(calldata->ctx);
	kfree(calldata);
3181 3182
}

3183
static void nfs4_locku_done(struct rpc_task *task, void *data)
3184
{
3185
	struct nfs4_unlockdata *calldata = data;
3186

3187 3188
	if (RPC_ASSASSINATED(task))
		return;
T
Trond Myklebust 已提交
3189
	nfs_increment_lock_seqid(task->tk_status, calldata->arg.seqid);
3190 3191 3192
	switch (task->tk_status) {
		case 0:
			memcpy(calldata->lsp->ls_stateid.data,
T
Trond Myklebust 已提交
3193
					calldata->res.stateid.data,
3194
					sizeof(calldata->lsp->ls_stateid.data));
3195
			renew_lease(calldata->server, calldata->timestamp);
3196 3197 3198 3199 3200
			break;
		case -NFS4ERR_STALE_STATEID:
		case -NFS4ERR_EXPIRED:
			break;
		default:
3201
			if (nfs4_async_handle_error(task, calldata->server) == -EAGAIN)
3202 3203 3204 3205
				rpc_restart_call(task);
	}
}

T
Trond Myklebust 已提交
3206
static void nfs4_locku_prepare(struct rpc_task *task, void *data)
3207
{
T
Trond Myklebust 已提交
3208
	struct nfs4_unlockdata *calldata = data;
3209

T
Trond Myklebust 已提交
3210
	if (nfs_wait_on_sequence(calldata->arg.seqid, task) != 0)
3211 3212
		return;
	if ((calldata->lsp->ls_flags & NFS_LOCK_INITIALIZED) == 0) {
3213 3214
		/* Note: exit _without_ running nfs4_locku_done */
		task->tk_action = NULL;
3215 3216
		return;
	}
3217
	calldata->timestamp = jiffies;
3218
	rpc_call_start(task);
3219 3220
}

3221
static const struct rpc_call_ops nfs4_locku_ops = {
T
Trond Myklebust 已提交
3222
	.rpc_call_prepare = nfs4_locku_prepare,
3223
	.rpc_call_done = nfs4_locku_done,
3224
	.rpc_release = nfs4_locku_release_calldata,
3225 3226
};

3227 3228 3229 3230 3231 3232
static struct rpc_task *nfs4_do_unlck(struct file_lock *fl,
		struct nfs_open_context *ctx,
		struct nfs4_lock_state *lsp,
		struct nfs_seqid *seqid)
{
	struct nfs4_unlockdata *data;
3233 3234 3235 3236
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_LOCKU],
		.rpc_cred = ctx->cred,
	};
T
Trond Myklebust 已提交
3237 3238
	struct rpc_task_setup task_setup_data = {
		.rpc_client = NFS_CLIENT(lsp->ls_state->inode),
3239
		.rpc_message = &msg,
T
Trond Myklebust 已提交
3240 3241 3242
		.callback_ops = &nfs4_locku_ops,
		.flags = RPC_TASK_ASYNC,
	};
3243

3244 3245 3246 3247 3248
	/* Ensure this is an unlock - when canceling a lock, the
	 * canceled lock is passed in, and it won't be an unlock.
	 */
	fl->fl_type = F_UNLCK;

3249 3250 3251 3252 3253 3254
	data = nfs4_alloc_unlockdata(fl, ctx, lsp, seqid);
	if (data == NULL) {
		nfs_free_seqid(seqid);
		return ERR_PTR(-ENOMEM);
	}

3255 3256
	msg.rpc_argp = &data->arg,
	msg.rpc_resp = &data->res,
T
Trond Myklebust 已提交
3257 3258
	task_setup_data.callback_data = data;
	return rpc_run_task(&task_setup_data);
3259 3260
}

3261 3262
static int nfs4_proc_unlck(struct nfs4_state *state, int cmd, struct file_lock *request)
{
T
Trond Myklebust 已提交
3263
	struct nfs_seqid *seqid;
L
Linus Torvalds 已提交
3264
	struct nfs4_lock_state *lsp;
3265 3266
	struct rpc_task *task;
	int status = 0;
3267

3268
	status = nfs4_set_lock_state(state, request);
3269 3270 3271 3272
	/* Unlock _before_ we do the RPC call */
	request->fl_flags |= FL_EXISTS;
	if (do_vfs_lock(request->fl_file, request) == -ENOENT)
		goto out;
3273
	if (status != 0)
3274 3275 3276 3277
		goto out;
	/* Is this a delegated lock? */
	if (test_bit(NFS_DELEGATED_STATE, &state->flags))
		goto out;
3278
	lsp = request->fl_u.nfs4_fl.owner;
T
Trond Myklebust 已提交
3279
	seqid = nfs_alloc_seqid(&lsp->ls_seqid);
3280
	status = -ENOMEM;
T
Trond Myklebust 已提交
3281
	if (seqid == NULL)
3282
		goto out;
3283
	task = nfs4_do_unlck(request, nfs_file_open_context(request->fl_file), lsp, seqid);
3284 3285
	status = PTR_ERR(task);
	if (IS_ERR(task))
3286
		goto out;
3287
	status = nfs4_wait_for_completion_rpc_task(task);
3288
	rpc_put_task(task);
3289
out:
L
Linus Torvalds 已提交
3290 3291 3292
	return status;
}

3293 3294 3295 3296 3297 3298
struct nfs4_lockdata {
	struct nfs_lock_args arg;
	struct nfs_lock_res res;
	struct nfs4_lock_state *lsp;
	struct nfs_open_context *ctx;
	struct file_lock fl;
3299
	unsigned long timestamp;
3300 3301 3302 3303 3304 3305
	int rpc_status;
	int cancelled;
};

static struct nfs4_lockdata *nfs4_alloc_lockdata(struct file_lock *fl,
		struct nfs_open_context *ctx, struct nfs4_lock_state *lsp)
L
Linus Torvalds 已提交
3306
{
3307 3308
	struct nfs4_lockdata *p;
	struct inode *inode = lsp->ls_state->inode;
L
Linus Torvalds 已提交
3309
	struct nfs_server *server = NFS_SERVER(inode);
3310 3311 3312 3313 3314 3315 3316

	p = kzalloc(sizeof(*p), GFP_KERNEL);
	if (p == NULL)
		return NULL;

	p->arg.fh = NFS_FH(inode);
	p->arg.fl = &p->fl;
3317 3318 3319
	p->arg.open_seqid = nfs_alloc_seqid(&lsp->ls_state->owner->so_seqid);
	if (p->arg.open_seqid == NULL)
		goto out_free;
3320 3321
	p->arg.lock_seqid = nfs_alloc_seqid(&lsp->ls_seqid);
	if (p->arg.lock_seqid == NULL)
3322
		goto out_free_seqid;
3323
	p->arg.lock_stateid = &lsp->ls_stateid;
3324
	p->arg.lock_owner.clientid = server->nfs_client->cl_clientid;
3325
	p->arg.lock_owner.id = lsp->ls_id.id;
3326 3327 3328 3329 3330
	p->lsp = lsp;
	atomic_inc(&lsp->ls_count);
	p->ctx = get_nfs_open_context(ctx);
	memcpy(&p->fl, fl, sizeof(p->fl));
	return p;
3331 3332
out_free_seqid:
	nfs_free_seqid(p->arg.open_seqid);
3333 3334 3335 3336 3337 3338 3339 3340 3341
out_free:
	kfree(p);
	return NULL;
}

static void nfs4_lock_prepare(struct rpc_task *task, void *calldata)
{
	struct nfs4_lockdata *data = calldata;
	struct nfs4_state *state = data->lsp->ls_state;
3342

3343
	dprintk("%s: begin!\n", __FUNCTION__);
3344 3345
	if (nfs_wait_on_sequence(data->arg.lock_seqid, task) != 0)
		return;
3346 3347
	/* Do we need to do an open_to_lock_owner? */
	if (!(data->arg.lock_seqid->sequence->flags & NFS_SEQID_CONFIRMED)) {
3348 3349
		if (nfs_wait_on_sequence(data->arg.open_seqid, task) != 0)
			return;
3350 3351
		data->arg.open_stateid = &state->stateid;
		data->arg.new_lock_owner = 1;
3352 3353
	} else
		data->arg.new_lock_owner = 0;
3354
	data->timestamp = jiffies;
3355
	rpc_call_start(task);
3356 3357 3358 3359 3360 3361 3362 3363 3364 3365 3366 3367 3368 3369 3370 3371 3372 3373 3374 3375 3376 3377 3378
	dprintk("%s: done!, ret = %d\n", __FUNCTION__, data->rpc_status);
}

static void nfs4_lock_done(struct rpc_task *task, void *calldata)
{
	struct nfs4_lockdata *data = calldata;

	dprintk("%s: begin!\n", __FUNCTION__);

	data->rpc_status = task->tk_status;
	if (RPC_ASSASSINATED(task))
		goto out;
	if (data->arg.new_lock_owner != 0) {
		nfs_increment_open_seqid(data->rpc_status, data->arg.open_seqid);
		if (data->rpc_status == 0)
			nfs_confirm_seqid(&data->lsp->ls_seqid, 0);
		else
			goto out;
	}
	if (data->rpc_status == 0) {
		memcpy(data->lsp->ls_stateid.data, data->res.stateid.data,
					sizeof(data->lsp->ls_stateid.data));
		data->lsp->ls_flags |= NFS_LOCK_INITIALIZED;
3379
		renew_lease(NFS_SERVER(data->ctx->path.dentry->d_inode), data->timestamp);
3380 3381 3382 3383 3384 3385 3386 3387 3388 3389 3390
	}
	nfs_increment_lock_seqid(data->rpc_status, data->arg.lock_seqid);
out:
	dprintk("%s: done, ret = %d!\n", __FUNCTION__, data->rpc_status);
}

static void nfs4_lock_release(void *calldata)
{
	struct nfs4_lockdata *data = calldata;

	dprintk("%s: begin!\n", __FUNCTION__);
3391
	nfs_free_seqid(data->arg.open_seqid);
3392 3393 3394 3395 3396
	if (data->cancelled != 0) {
		struct rpc_task *task;
		task = nfs4_do_unlck(&data->fl, data->ctx, data->lsp,
				data->arg.lock_seqid);
		if (!IS_ERR(task))
3397
			rpc_put_task(task);
3398 3399 3400 3401 3402 3403 3404 3405 3406 3407 3408 3409 3410 3411 3412 3413 3414 3415 3416
		dprintk("%s: cancelling lock!\n", __FUNCTION__);
	} else
		nfs_free_seqid(data->arg.lock_seqid);
	nfs4_put_lock_state(data->lsp);
	put_nfs_open_context(data->ctx);
	kfree(data);
	dprintk("%s: done!\n", __FUNCTION__);
}

static const struct rpc_call_ops nfs4_lock_ops = {
	.rpc_call_prepare = nfs4_lock_prepare,
	.rpc_call_done = nfs4_lock_done,
	.rpc_release = nfs4_lock_release,
};

static int _nfs4_do_setlk(struct nfs4_state *state, int cmd, struct file_lock *fl, int reclaim)
{
	struct nfs4_lockdata *data;
	struct rpc_task *task;
3417 3418 3419 3420
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_LOCK],
		.rpc_cred = state->owner->so_cred,
	};
T
Trond Myklebust 已提交
3421 3422
	struct rpc_task_setup task_setup_data = {
		.rpc_client = NFS_CLIENT(state->inode),
3423
		.rpc_message = &msg,
T
Trond Myklebust 已提交
3424 3425 3426
		.callback_ops = &nfs4_lock_ops,
		.flags = RPC_TASK_ASYNC,
	};
3427 3428 3429
	int ret;

	dprintk("%s: begin!\n", __FUNCTION__);
3430
	data = nfs4_alloc_lockdata(fl, nfs_file_open_context(fl->fl_file),
3431 3432 3433 3434 3435 3436 3437
			fl->fl_u.nfs4_fl.owner);
	if (data == NULL)
		return -ENOMEM;
	if (IS_SETLKW(cmd))
		data->arg.block = 1;
	if (reclaim != 0)
		data->arg.reclaim = 1;
3438 3439
	msg.rpc_argp = &data->arg,
	msg.rpc_resp = &data->res,
T
Trond Myklebust 已提交
3440 3441
	task_setup_data.callback_data = data;
	task = rpc_run_task(&task_setup_data);
3442
	if (IS_ERR(task))
3443 3444 3445 3446 3447 3448 3449 3450
		return PTR_ERR(task);
	ret = nfs4_wait_for_completion_rpc_task(task);
	if (ret == 0) {
		ret = data->rpc_status;
		if (ret == -NFS4ERR_DENIED)
			ret = -EAGAIN;
	} else
		data->cancelled = 1;
3451
	rpc_put_task(task);
3452 3453
	dprintk("%s: done, ret = %d!\n", __FUNCTION__, ret);
	return ret;
L
Linus Torvalds 已提交
3454 3455 3456 3457
}

static int nfs4_lock_reclaim(struct nfs4_state *state, struct file_lock *request)
{
3458 3459 3460 3461 3462
	struct nfs_server *server = NFS_SERVER(state->inode);
	struct nfs4_exception exception = { };
	int err;

	do {
3463 3464 3465
		/* Cache the lock if possible... */
		if (test_bit(NFS_DELEGATED_STATE, &state->flags) != 0)
			return 0;
3466 3467 3468 3469 3470 3471
		err = _nfs4_do_setlk(state, F_SETLK, request, 1);
		if (err != -NFS4ERR_DELAY)
			break;
		nfs4_handle_exception(server, err, &exception);
	} while (exception.retry);
	return err;
L
Linus Torvalds 已提交
3472 3473 3474 3475
}

static int nfs4_lock_expired(struct nfs4_state *state, struct file_lock *request)
{
3476 3477 3478 3479
	struct nfs_server *server = NFS_SERVER(state->inode);
	struct nfs4_exception exception = { };
	int err;

3480 3481 3482
	err = nfs4_set_lock_state(state, request);
	if (err != 0)
		return err;
3483
	do {
3484 3485
		if (test_bit(NFS_DELEGATED_STATE, &state->flags) != 0)
			return 0;
3486 3487 3488 3489 3490 3491
		err = _nfs4_do_setlk(state, F_SETLK, request, 0);
		if (err != -NFS4ERR_DELAY)
			break;
		nfs4_handle_exception(server, err, &exception);
	} while (exception.retry);
	return err;
L
Linus Torvalds 已提交
3492 3493 3494 3495
}

static int _nfs4_proc_setlk(struct nfs4_state *state, int cmd, struct file_lock *request)
{
3496
	struct nfs_client *clp = state->owner->so_client;
3497
	unsigned char fl_flags = request->fl_flags;
L
Linus Torvalds 已提交
3498 3499
	int status;

3500 3501 3502 3503
	/* Is this a delegated open? */
	status = nfs4_set_lock_state(state, request);
	if (status != 0)
		goto out;
3504 3505 3506 3507 3508 3509 3510 3511 3512 3513 3514 3515 3516 3517 3518 3519 3520 3521
	request->fl_flags |= FL_ACCESS;
	status = do_vfs_lock(request->fl_file, request);
	if (status < 0)
		goto out;
	down_read(&clp->cl_sem);
	if (test_bit(NFS_DELEGATED_STATE, &state->flags)) {
		struct nfs_inode *nfsi = NFS_I(state->inode);
		/* Yes: cache locks! */
		down_read(&nfsi->rwsem);
		/* ...but avoid races with delegation recall... */
		if (test_bit(NFS_DELEGATED_STATE, &state->flags)) {
			request->fl_flags = fl_flags & ~FL_SLEEP;
			status = do_vfs_lock(request->fl_file, request);
			up_read(&nfsi->rwsem);
			goto out_unlock;
		}
		up_read(&nfsi->rwsem);
	}
3522 3523
	status = _nfs4_do_setlk(state, cmd, request, 0);
	if (status != 0)
3524
		goto out_unlock;
3525
	/* Note: we always want to sleep here! */
3526
	request->fl_flags = fl_flags | FL_SLEEP;
3527 3528
	if (do_vfs_lock(request->fl_file, request) < 0)
		printk(KERN_WARNING "%s: VFS is out of sync with lock manager!\n", __FUNCTION__);
3529
out_unlock:
L
Linus Torvalds 已提交
3530
	up_read(&clp->cl_sem);
3531 3532
out:
	request->fl_flags = fl_flags;
L
Linus Torvalds 已提交
3533 3534 3535 3536 3537 3538 3539 3540 3541 3542 3543 3544 3545 3546 3547 3548 3549 3550 3551 3552 3553 3554 3555 3556 3557
	return status;
}

static int nfs4_proc_setlk(struct nfs4_state *state, int cmd, struct file_lock *request)
{
	struct nfs4_exception exception = { };
	int err;

	do {
		err = nfs4_handle_exception(NFS_SERVER(state->inode),
				_nfs4_proc_setlk(state, cmd, request),
				&exception);
	} while (exception.retry);
	return err;
}

static int
nfs4_proc_lock(struct file *filp, int cmd, struct file_lock *request)
{
	struct nfs_open_context *ctx;
	struct nfs4_state *state;
	unsigned long timeout = NFS4_LOCK_MINTIMEOUT;
	int status;

	/* verify open state */
3558
	ctx = nfs_file_open_context(filp);
L
Linus Torvalds 已提交
3559 3560 3561 3562 3563 3564 3565 3566 3567 3568 3569 3570 3571 3572 3573 3574 3575 3576 3577 3578 3579 3580 3581 3582 3583 3584
	state = ctx->state;

	if (request->fl_start < 0 || request->fl_end < 0)
		return -EINVAL;

	if (IS_GETLK(cmd))
		return nfs4_proc_getlk(state, F_GETLK, request);

	if (!(IS_SETLK(cmd) || IS_SETLKW(cmd)))
		return -EINVAL;

	if (request->fl_type == F_UNLCK)
		return nfs4_proc_unlck(state, cmd, request);

	do {
		status = nfs4_proc_setlk(state, cmd, request);
		if ((status != -EAGAIN) || IS_SETLK(cmd))
			break;
		timeout = nfs4_set_lock_task_retry(timeout);
		status = -ERESTARTSYS;
		if (signalled())
			break;
	} while(status < 0);
	return status;
}

3585 3586 3587 3588 3589 3590 3591 3592 3593 3594 3595 3596 3597 3598 3599 3600 3601 3602
int nfs4_lock_delegation_recall(struct nfs4_state *state, struct file_lock *fl)
{
	struct nfs_server *server = NFS_SERVER(state->inode);
	struct nfs4_exception exception = { };
	int err;

	err = nfs4_set_lock_state(state, fl);
	if (err != 0)
		goto out;
	do {
		err = _nfs4_do_setlk(state, F_SETLK, fl, 0);
		if (err != -NFS4ERR_DELAY)
			break;
		err = nfs4_handle_exception(server, err, &exception);
	} while (exception.retry);
out:
	return err;
}
3603

3604 3605
#define XATTR_NAME_NFSV4_ACL "system.nfs4_acl"

3606 3607 3608
int nfs4_setxattr(struct dentry *dentry, const char *key, const void *buf,
		size_t buflen, int flags)
{
3609 3610 3611 3612 3613 3614 3615 3616 3617 3618
	struct inode *inode = dentry->d_inode;

	if (strcmp(key, XATTR_NAME_NFSV4_ACL) != 0)
		return -EOPNOTSUPP;

	if (!S_ISREG(inode->i_mode) &&
	    (!S_ISDIR(inode->i_mode) || inode->i_mode & S_ISVTX))
		return -EPERM;

	return nfs4_proc_set_acl(inode, buf, buflen);
3619 3620 3621 3622 3623 3624 3625 3626 3627
}

/* The getxattr man page suggests returning -ENODATA for unknown attributes,
 * and that's what we'll do for e.g. user attributes that haven't been set.
 * But we'll follow ext2/ext3's lead by returning -EOPNOTSUPP for unsupported
 * attributes in kernel-managed attribute namespaces. */
ssize_t nfs4_getxattr(struct dentry *dentry, const char *key, void *buf,
		size_t buflen)
{
3628 3629 3630 3631 3632 3633
	struct inode *inode = dentry->d_inode;

	if (strcmp(key, XATTR_NAME_NFSV4_ACL) != 0)
		return -EOPNOTSUPP;

	return nfs4_proc_get_acl(inode, buf, buflen);
3634 3635 3636 3637
}

ssize_t nfs4_listxattr(struct dentry *dentry, char *buf, size_t buflen)
{
3638
	size_t len = strlen(XATTR_NAME_NFSV4_ACL) + 1;
3639

3640 3641
	if (!nfs4_server_supports_acls(NFS_SERVER(dentry->d_inode)))
		return 0;
3642 3643 3644
	if (buf && buflen < len)
		return -ERANGE;
	if (buf)
3645 3646
		memcpy(buf, XATTR_NAME_NFSV4_ACL, len);
	return len;
3647 3648
}

3649
int nfs4_proc_fs_locations(struct inode *dir, const struct qstr *name,
3650
		struct nfs4_fs_locations *fs_locations, struct page *page)
3651 3652 3653
{
	struct nfs_server *server = NFS_SERVER(dir);
	u32 bitmask[2] = {
3654 3655
		[0] = FATTR4_WORD0_FSID | FATTR4_WORD0_FS_LOCATIONS,
		[1] = FATTR4_WORD1_MOUNTED_ON_FILEID,
3656 3657 3658
	};
	struct nfs4_fs_locations_arg args = {
		.dir_fh = NFS_FH(dir),
3659
		.name = name,
3660 3661 3662 3663 3664 3665
		.page = page,
		.bitmask = bitmask,
	};
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_FS_LOCATIONS],
		.rpc_argp = &args,
3666
		.rpc_resp = fs_locations,
3667 3668 3669 3670
	};
	int status;

	dprintk("%s: start\n", __FUNCTION__);
3671
	nfs_fattr_init(&fs_locations->fattr);
3672
	fs_locations->server = server;
3673
	fs_locations->nlocations = 0;
3674 3675 3676 3677 3678
	status = rpc_call_sync(server->client, &msg, 0);
	dprintk("%s: returned status = %d\n", __FUNCTION__, status);
	return status;
}

L
Linus Torvalds 已提交
3679 3680 3681 3682 3683 3684 3685 3686 3687 3688
struct nfs4_state_recovery_ops nfs4_reboot_recovery_ops = {
	.recover_open	= nfs4_open_reclaim,
	.recover_lock	= nfs4_lock_reclaim,
};

struct nfs4_state_recovery_ops nfs4_network_partition_recovery_ops = {
	.recover_open	= nfs4_open_expired,
	.recover_lock	= nfs4_lock_expired,
};

3689
static const struct inode_operations nfs4_file_inode_operations = {
3690 3691 3692 3693 3694 3695 3696 3697
	.permission	= nfs_permission,
	.getattr	= nfs_getattr,
	.setattr	= nfs_setattr,
	.getxattr	= nfs4_getxattr,
	.setxattr	= nfs4_setxattr,
	.listxattr	= nfs4_listxattr,
};

D
David Howells 已提交
3698
const struct nfs_rpc_ops nfs_v4_clientops = {
L
Linus Torvalds 已提交
3699 3700 3701
	.version	= 4,			/* protocol version */
	.dentry_ops	= &nfs4_dentry_operations,
	.dir_inode_ops	= &nfs4_dir_inode_operations,
3702
	.file_inode_ops	= &nfs4_file_inode_operations,
L
Linus Torvalds 已提交
3703 3704 3705
	.getroot	= nfs4_proc_get_root,
	.getattr	= nfs4_proc_getattr,
	.setattr	= nfs4_proc_setattr,
D
David Howells 已提交
3706
	.lookupfh	= nfs4_proc_lookupfh,
L
Linus Torvalds 已提交
3707 3708 3709 3710 3711 3712 3713 3714 3715 3716 3717 3718 3719 3720 3721 3722 3723
	.lookup		= nfs4_proc_lookup,
	.access		= nfs4_proc_access,
	.readlink	= nfs4_proc_readlink,
	.create		= nfs4_proc_create,
	.remove		= nfs4_proc_remove,
	.unlink_setup	= nfs4_proc_unlink_setup,
	.unlink_done	= nfs4_proc_unlink_done,
	.rename		= nfs4_proc_rename,
	.link		= nfs4_proc_link,
	.symlink	= nfs4_proc_symlink,
	.mkdir		= nfs4_proc_mkdir,
	.rmdir		= nfs4_proc_remove,
	.readdir	= nfs4_proc_readdir,
	.mknod		= nfs4_proc_mknod,
	.statfs		= nfs4_proc_statfs,
	.fsinfo		= nfs4_proc_fsinfo,
	.pathconf	= nfs4_proc_pathconf,
3724
	.set_capabilities = nfs4_server_capabilities,
L
Linus Torvalds 已提交
3725 3726
	.decode_dirent	= nfs4_decode_dirent,
	.read_setup	= nfs4_proc_read_setup,
T
Trond Myklebust 已提交
3727
	.read_done	= nfs4_read_done,
L
Linus Torvalds 已提交
3728
	.write_setup	= nfs4_proc_write_setup,
3729
	.write_done	= nfs4_write_done,
L
Linus Torvalds 已提交
3730
	.commit_setup	= nfs4_proc_commit_setup,
3731
	.commit_done	= nfs4_commit_done,
3732 3733
	.file_open      = nfs_open,
	.file_release   = nfs_release,
L
Linus Torvalds 已提交
3734
	.lock		= nfs4_proc_lock,
3735
	.clear_acl_cache = nfs4_zap_acl_attr,
L
Linus Torvalds 已提交
3736 3737 3738 3739 3740 3741 3742
};

/*
 * Local variables:
 *  c-basic-offset: 8
 * End:
 */