nfs4proc.c 99.5 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

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

217
struct nfs4_opendata {
218
	struct kref kref;
219 220
	struct nfs_openargs o_arg;
	struct nfs_openres o_res;
221 222
	struct nfs_open_confirmargs c_arg;
	struct nfs_open_confirmres c_res;
223 224
	struct nfs_fattr f_attr;
	struct nfs_fattr dir_attr;
225
	struct path path;
226 227
	struct dentry *dir;
	struct nfs4_state_owner *owner;
228
	struct nfs4_state *state;
229
	struct iattr attrs;
230
	unsigned long timestamp;
231
	unsigned int rpc_done : 1;
232 233
	int rpc_status;
	int cancelled;
234 235
};

236 237 238 239 240 241 242 243 244 245

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);
}

246
static struct nfs4_opendata *nfs4_opendata_alloc(struct path *path,
247 248 249
		struct nfs4_state_owner *sp, int flags,
		const struct iattr *attrs)
{
250
	struct dentry *parent = dget_parent(path->dentry);
251 252 253 254 255 256 257 258 259 260
	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;
261 262
	p->path.mnt = mntget(path->mnt);
	p->path.dentry = dget(path->dentry);
263 264 265 266 267
	p->dir = parent;
	p->owner = sp;
	atomic_inc(&sp->so_count);
	p->o_arg.fh = NFS_FH(dir);
	p->o_arg.open_flags = flags,
268
	p->o_arg.clientid = server->nfs_client->cl_clientid;
269
	p->o_arg.id = sp->so_owner_id.id;
270
	p->o_arg.name = &p->path.dentry->d_name;
271 272 273 274 275 276 277 278 279 280 281
	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));
	}
282 283 284
	p->c_arg.fh = &p->o_res.fh;
	p->c_arg.stateid = &p->o_res.stateid;
	p->c_arg.seqid = p->o_arg.seqid;
285
	nfs4_init_opendata_res(p);
286
	kref_init(&p->kref);
287 288 289 290 291 292 293 294
	return p;
err_free:
	kfree(p);
err:
	dput(parent);
	return NULL;
}

295
static void nfs4_opendata_free(struct kref *kref)
296
{
297 298 299 300
	struct nfs4_opendata *p = container_of(kref,
			struct nfs4_opendata, kref);

	nfs_free_seqid(p->o_arg.seqid);
301 302
	if (p->state != NULL)
		nfs4_put_open_state(p->state);
303 304 305 306 307 308 309 310 311 312 313
	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);
314 315
}

316 317 318 319 320 321 322 323 324 325 326
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;
}

327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342
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;
}

343 344 345 346 347 348 349 350 351
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;
}

352
static void update_open_stateflags(struct nfs4_state *state, mode_t open_flags)
353 354 355 356 357 358 359 360 361 362 363
{
	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++;
	}
364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381
	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);
	}
382 383
}

384 385
static void nfs_set_open_stateid(struct nfs4_state *state, nfs4_stateid *stateid, int open_flags)
{
386
	write_seqlock(&state->seqlock);
387
	nfs_set_open_stateid_locked(state, stateid, open_flags);
388
	write_sequnlock(&state->seqlock);
389 390 391
}

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
392 393
{
	open_flags &= (FMODE_READ|FMODE_WRITE);
394 395 396 397 398
	/*
	 * Protect the call to nfs4_state_set_mode_locked and
	 * serialise the stateid update
	 */
	write_seqlock(&state->seqlock);
399 400 401 402 403 404
	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);
405 406
	write_sequnlock(&state->seqlock);
	spin_lock(&state->owner->so_lock);
407
	update_open_stateflags(state, open_flags);
408
	spin_unlock(&state->owner->so_lock);
L
Linus Torvalds 已提交
409 410
}

411 412 413 414 415 416 417 418 419 420 421 422 423 424
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);
}

425
static struct nfs4_state *nfs4_try_open_cached(struct nfs4_opendata *opendata)
426 427 428 429 430 431 432 433 434 435 436
{
	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 (;;) {
437 438 439 440 441 442 443 444 445 446 447 448
		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;
449 450 451 452 453 454
		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();
455
		ret = nfs_may_open(state->inode, state->owner->so_cred, open_mode);
456 457 458 459 460 461
		unlock_kernel();
		if (ret != 0)
			goto out;
		ret = -EAGAIN;
		rcu_read_lock();
		delegation = rcu_dereference(nfsi->delegation);
462
		/* If no delegation, try a cached open */
463
		if (delegation == NULL)
464
			continue;
465 466 467 468 469 470 471 472 473 474 475 476 477 478 479
		/* 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;
}

480 481 482 483
static struct nfs4_state *nfs4_opendata_to_nfs4_state(struct nfs4_opendata *data)
{
	struct inode *inode;
	struct nfs4_state *state = NULL;
484 485
	struct nfs_delegation *delegation;
	nfs4_stateid *deleg_stateid = NULL;
486
	int ret;
487

488
	if (!data->rpc_done) {
489
		state = nfs4_try_open_cached(data);
490 491 492
		goto out;
	}

493
	ret = -EAGAIN;
494
	if (!(data->f_attr.valid & NFS_ATTR_FATTR))
495
		goto err;
496
	inode = nfs_fhget(data->dir->d_sb, &data->o_res.fh, &data->f_attr);
497
	ret = PTR_ERR(inode);
498
	if (IS_ERR(inode))
499 500
		goto err;
	ret = -ENOMEM;
501 502
	state = nfs4_get_open_state(inode, data->owner);
	if (state == NULL)
503
		goto err_put_inode;
504 505 506
	if (data->o_res.delegation_type != 0) {
		int delegation_flags = 0;

507 508 509 510 511
		rcu_read_lock();
		delegation = rcu_dereference(NFS_I(inode)->delegation);
		if (delegation)
			delegation_flags = delegation->flags;
		rcu_read_unlock();
512 513 514 515 516 517 518 519 520
		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);
	}
521 522 523 524 525 526
	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();
527
	iput(inode);
528
out:
529
	return state;
530 531 532 533
err_put_inode:
	iput(inode);
err:
	return ERR_PTR(ret);
534 535
}

536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552
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
553 554 555 556 557 558 559 560 561 562 563 564
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;
}

565
static int nfs4_open_recover_helper(struct nfs4_opendata *opendata, mode_t openflags, struct nfs4_state **res)
566
{
567
	struct nfs4_state *newstate;
568 569 570
	int ret;

	opendata->o_arg.open_flags = openflags;
571 572 573
	memset(&opendata->o_res, 0, sizeof(opendata->o_res));
	memset(&opendata->c_res, 0, sizeof(opendata->c_res));
	nfs4_init_opendata_res(opendata);
574 575 576
	ret = _nfs4_proc_open(opendata);
	if (ret != 0)
		return ret; 
577
	newstate = nfs4_opendata_to_nfs4_state(opendata);
578 579 580
	if (IS_ERR(newstate))
		return PTR_ERR(newstate);
	nfs4_close_state(&opendata->path, newstate, openflags);
581
	*res = newstate;
582 583 584 585 586 587 588 589 590
	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_* */
591
	clear_bit(NFS_DELEGATED_STATE, &state->flags);
592 593
	smp_rmb();
	if (state->n_rdwr != 0) {
594
		ret = nfs4_open_recover_helper(opendata, FMODE_READ|FMODE_WRITE, &newstate);
595 596
		if (ret != 0)
			return ret;
597 598
		if (newstate != state)
			return -ESTALE;
599 600
	}
	if (state->n_wronly != 0) {
601
		ret = nfs4_open_recover_helper(opendata, FMODE_WRITE, &newstate);
602 603
		if (ret != 0)
			return ret;
604 605
		if (newstate != state)
			return -ESTALE;
606 607
	}
	if (state->n_rdonly != 0) {
608
		ret = nfs4_open_recover_helper(opendata, FMODE_READ, &newstate);
609 610
		if (ret != 0)
			return ret;
611 612
		if (newstate != state)
			return -ESTALE;
613
	}
614 615 616 617 618 619
	/*
	 * 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) {
620
		write_seqlock(&state->seqlock);
621 622
		if (test_bit(NFS_DELEGATED_STATE, &state->flags) == 0)
			memcpy(state->stateid.data, state->open_stateid.data, sizeof(state->stateid.data));
623
		write_sequnlock(&state->seqlock);
624
	}
625 626 627
	return 0;
}

L
Linus Torvalds 已提交
628 629 630 631
/*
 * OPEN_RECLAIM:
 * 	reclaim state on the server after a reboot.
 */
632
static int _nfs4_do_open_reclaim(struct nfs_open_context *ctx, struct nfs4_state *state)
L
Linus Torvalds 已提交
633
{
634
	struct nfs_delegation *delegation;
635 636
	struct nfs4_opendata *opendata;
	int delegation_type = 0;
L
Linus Torvalds 已提交
637 638
	int status;

T
Trond Myklebust 已提交
639 640 641
	opendata = nfs4_open_recoverdata_alloc(ctx, state);
	if (IS_ERR(opendata))
		return PTR_ERR(opendata);
642 643
	opendata->o_arg.claim = NFS4_OPEN_CLAIM_PREVIOUS;
	opendata->o_arg.fh = NFS_FH(state->inode);
644 645 646
	rcu_read_lock();
	delegation = rcu_dereference(NFS_I(state->inode)->delegation);
	if (delegation != NULL && (delegation->flags & NFS_DELEGATION_NEED_RECLAIM) != 0)
647
		delegation_type = delegation->type;
648
	rcu_read_unlock();
649 650
	opendata->o_arg.u.delegation_type = delegation_type;
	status = nfs4_open_recover(opendata, state);
651
	nfs4_opendata_put(opendata);
L
Linus Torvalds 已提交
652 653 654
	return status;
}

655
static int nfs4_do_open_reclaim(struct nfs_open_context *ctx, struct nfs4_state *state)
L
Linus Torvalds 已提交
656 657 658 659 660
{
	struct nfs_server *server = NFS_SERVER(state->inode);
	struct nfs4_exception exception = { };
	int err;
	do {
661
		err = _nfs4_do_open_reclaim(ctx, state);
662 663 664
		if (err != -NFS4ERR_DELAY)
			break;
		nfs4_handle_exception(server, err, &exception);
L
Linus Torvalds 已提交
665 666 667 668
	} while (exception.retry);
	return err;
}

669 670 671 672 673 674 675 676
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);
677
	ret = nfs4_do_open_reclaim(ctx, state);
678 679 680 681
	put_nfs_open_context(ctx);
	return ret;
}

682
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
683
{
684
	struct nfs4_opendata *opendata;
685
	int ret;
L
Linus Torvalds 已提交
686

T
Trond Myklebust 已提交
687 688 689
	opendata = nfs4_open_recoverdata_alloc(ctx, state);
	if (IS_ERR(opendata))
		return PTR_ERR(opendata);
690
	opendata->o_arg.claim = NFS4_OPEN_CLAIM_DELEGATE_CUR;
691
	memcpy(opendata->o_arg.u.delegation.data, stateid->data,
692
			sizeof(opendata->o_arg.u.delegation.data));
693
	ret = nfs4_open_recover(opendata, state);
694
	nfs4_opendata_put(opendata);
695
	return ret;
L
Linus Torvalds 已提交
696 697
}

698
int nfs4_open_delegation_recall(struct nfs_open_context *ctx, struct nfs4_state *state, const nfs4_stateid *stateid)
L
Linus Torvalds 已提交
699 700
{
	struct nfs4_exception exception = { };
701
	struct nfs_server *server = NFS_SERVER(state->inode);
L
Linus Torvalds 已提交
702 703
	int err;
	do {
704
		err = _nfs4_open_delegation_recall(ctx, state, stateid);
L
Linus Torvalds 已提交
705 706 707 708 709 710 711
		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 */
712
				nfs4_schedule_state_recovery(server->nfs_client);
L
Linus Torvalds 已提交
713 714 715 716 717 718 719
				return err;
		}
		err = nfs4_handle_exception(server, err, &exception);
	} while (exception.retry);
	return err;
}

720
static void nfs4_open_confirm_prepare(struct rpc_task *task, void *calldata)
L
Linus Torvalds 已提交
721
{
722 723 724 725 726 727
	struct nfs4_opendata *data = calldata;
	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,
L
Linus Torvalds 已提交
728
	};
729
	data->timestamp = jiffies;
730 731 732 733 734 735 736 737 738 739
	rpc_call_setup(task, &msg, 0);
}

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;
740
	if (data->rpc_status == 0) {
741 742
		memcpy(data->o_res.stateid.data, data->c_res.stateid.data,
				sizeof(data->o_res.stateid.data));
743
		renew_lease(data->o_res.server, data->timestamp);
744
		data->rpc_done = 1;
745
	}
746
	nfs_confirm_seqid(&data->owner->so_seqid, data->rpc_status);
747
	nfs_increment_open_seqid(data->rpc_status, data->c_arg.seqid);
748 749 750 751 752 753 754 755 756 757 758
}

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! */
759
	if (!data->rpc_done)
760 761 762
		goto out_free;
	nfs_confirm_seqid(&data->owner->so_seqid, 0);
	state = nfs4_opendata_to_nfs4_state(data);
763
	if (!IS_ERR(state))
764
		nfs4_close_state(&data->path, state, data->o_arg.open_flags);
765
out_free:
766
	nfs4_opendata_put(data);
767 768 769 770 771 772 773 774 775 776 777 778 779 780 781
}

static const struct rpc_call_ops nfs4_open_confirm_ops = {
	.rpc_call_prepare = nfs4_open_confirm_prepare,
	.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;
L
Linus Torvalds 已提交
782 783
	int status;

784
	kref_get(&data->kref);
785 786
	data->rpc_done = 0;
	data->rpc_status = 0;
787
	task = rpc_run_task(server->client, RPC_TASK_ASYNC, &nfs4_open_confirm_ops, 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;
L
Linus Torvalds 已提交
804 805
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_OPEN],
806 807
		.rpc_argp = &data->o_arg,
		.rpc_resp = &data->o_res,
L
Linus Torvalds 已提交
808 809
		.rpc_cred = sp->so_cred,
	};
810 811 812
	
	if (nfs_wait_on_sequence(data->o_arg.seqid, task) != 0)
		return;
813 814 815 816 817 818 819
	/*
	 * 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;

820 821
		if (can_open_cached(data->state, data->o_arg.open_flags & (FMODE_READ|FMODE_WRITE|O_EXCL)))
			goto out_no_action;
822 823 824 825 826
		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();
827
			goto out_no_action;
828 829 830
		}
		rcu_read_unlock();
	}
831
	/* Update sequence id. */
832
	data->o_arg.id = sp->so_owner_id.id;
833
	data->o_arg.clientid = sp->so_client->cl_clientid;
T
Trond Myklebust 已提交
834
	if (data->o_arg.claim == NFS4_OPEN_CLAIM_PREVIOUS) {
835
		msg.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_OPEN_NOATTR];
T
Trond Myklebust 已提交
836 837
		nfs_copy_fh(&data->o_res.fh, data->o_arg.fh);
	}
838
	data->timestamp = jiffies;
839
	rpc_call_setup(task, &msg, 0);
840 841 842 843
	return;
out_no_action:
	task->tk_action = NULL;

844
}
L
Linus Torvalds 已提交
845

846 847 848
static void nfs4_open_done(struct rpc_task *task, void *calldata)
{
	struct nfs4_opendata *data = calldata;
L
Linus Torvalds 已提交
849

850 851 852 853 854
	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) {
855 856 857
			case S_IFREG:
				break;
			case S_IFLNK:
858
				data->rpc_status = -ELOOP;
859 860
				break;
			case S_IFDIR:
861
				data->rpc_status = -EISDIR;
862 863
				break;
			default:
864
				data->rpc_status = -ENOTDIR;
865
		}
866
		renew_lease(data->o_res.server, data->timestamp);
867 868
		if (!(data->o_res.rflags & NFS4_OPEN_RESULT_CONFIRM))
			nfs_confirm_seqid(&data->owner->so_seqid, 0);
869
	}
870
	nfs_increment_open_seqid(data->rpc_status, data->o_arg.seqid);
871
	data->rpc_done = 1;
872
}
873

874 875 876 877 878 879 880 881 882
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! */
883
	if (data->rpc_status != 0 || !data->rpc_done)
884 885 886 887 888 889
		goto out_free;
	/* In case we need an open_confirm, no cleanup! */
	if (data->o_res.rflags & NFS4_OPEN_RESULT_CONFIRM)
		goto out_free;
	nfs_confirm_seqid(&data->owner->so_seqid, 0);
	state = nfs4_opendata_to_nfs4_state(data);
890
	if (!IS_ERR(state))
891
		nfs4_close_state(&data->path, state, data->o_arg.open_flags);
892
out_free:
893
	nfs4_opendata_put(data);
894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913
}

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;
	int status;

914
	kref_get(&data->kref);
915 916
	data->rpc_done = 0;
	data->rpc_status = 0;
917
	data->cancelled = 0;
918
	task = rpc_run_task(server->client, RPC_TASK_ASYNC, &nfs4_open_ops, data);
919
	if (IS_ERR(task))
920 921 922 923 924 925 926
		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;
927
	rpc_put_task(task);
928
	if (status != 0 || !data->rpc_done)
929 930
		return status;

931 932 933
	if (o_res->fh.size == 0)
		_nfs4_proc_lookup(dir, o_arg->name, &o_res->fh, o_res->f_attr);

934 935 936 937 938
	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
939
	if(o_res->rflags & NFS4_OPEN_RESULT_CONFIRM) {
940
		status = _nfs4_proc_open_confirm(data);
L
Linus Torvalds 已提交
941
		if (status != 0)
942
			return status;
L
Linus Torvalds 已提交
943 944
	}
	if (!(o_res->f_attr->valid & NFS_ATTR_FATTR))
945
		_nfs4_proc_getattr(server, &o_res->fh, o_res->f_attr);
946
	return 0;
L
Linus Torvalds 已提交
947 948
}

949
static int nfs4_recover_expired_lease(struct nfs_server *server)
950
{
951
	struct nfs_client *clp = server->nfs_client;
952
	int ret;
953

954 955 956 957 958 959
	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;
960
		nfs4_schedule_state_recovery(clp);
961 962
	}
	return 0;
963 964
}

L
Linus Torvalds 已提交
965 966 967 968 969
/*
 * OPEN_EXPIRED:
 * 	reclaim state on the server after a network partition.
 * 	Assumes caller holds the appropriate lock
 */
970
static int _nfs4_open_expired(struct nfs_open_context *ctx, struct nfs4_state *state)
L
Linus Torvalds 已提交
971
{
972
	struct nfs4_opendata *opendata;
973
	int ret;
L
Linus Torvalds 已提交
974

T
Trond Myklebust 已提交
975 976 977
	opendata = nfs4_open_recoverdata_alloc(ctx, state);
	if (IS_ERR(opendata))
		return PTR_ERR(opendata);
978 979 980
	ret = nfs4_open_recover(opendata, state);
	if (ret == -ESTALE) {
		/* Invalidate the state owner so we don't ever use it again */
981 982
		nfs4_drop_state_owner(state->owner);
		d_drop(ctx->path.dentry);
L
Linus Torvalds 已提交
983
	}
984
	nfs4_opendata_put(opendata);
985
	return ret;
L
Linus Torvalds 已提交
986 987
}

988
static inline int nfs4_do_open_expired(struct nfs_open_context *ctx, struct nfs4_state *state)
989
{
990
	struct nfs_server *server = NFS_SERVER(state->inode);
991 992 993 994
	struct nfs4_exception exception = { };
	int err;

	do {
995
		err = _nfs4_open_expired(ctx, state);
996 997 998 999 1000 1001
		if (err == -NFS4ERR_DELAY)
			nfs4_handle_exception(server, err, &exception);
	} while (exception.retry);
	return err;
}

L
Linus Torvalds 已提交
1002 1003 1004
static int nfs4_open_expired(struct nfs4_state_owner *sp, struct nfs4_state *state)
{
	struct nfs_open_context *ctx;
1005
	int ret;
L
Linus Torvalds 已提交
1006

1007 1008 1009
	ctx = nfs4_state_find_open_context(state);
	if (IS_ERR(ctx))
		return PTR_ERR(ctx);
1010
	ret = nfs4_do_open_expired(ctx, state);
1011 1012
	put_nfs_open_context(ctx);
	return ret;
L
Linus Torvalds 已提交
1013 1014
}

1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030
/*
 * 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
1031
/*
1032
 * Returns a referenced nfs4_state
L
Linus Torvalds 已提交
1033
 */
1034
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
1035 1036 1037 1038
{
	struct nfs4_state_owner  *sp;
	struct nfs4_state     *state = NULL;
	struct nfs_server       *server = NFS_SERVER(dir);
1039
	struct nfs_client *clp = server->nfs_client;
1040
	struct nfs4_opendata *opendata;
1041
	int status;
L
Linus Torvalds 已提交
1042 1043 1044 1045 1046 1047 1048

	/* 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;
	}
1049 1050
	status = nfs4_recover_expired_lease(server);
	if (status != 0)
1051
		goto err_put_state_owner;
1052 1053
	if (path->dentry->d_inode != NULL)
		nfs4_return_incompatible_delegation(path->dentry->d_inode, flags & (FMODE_READ|FMODE_WRITE));
1054 1055
	down_read(&clp->cl_sem);
	status = -ENOMEM;
1056
	opendata = nfs4_opendata_alloc(path, sp, flags, sattr);
1057
	if (opendata == NULL)
1058
		goto err_release_rwsem;
L
Linus Torvalds 已提交
1059

1060 1061 1062
	if (path->dentry->d_inode != NULL)
		opendata->state = nfs4_get_open_state(path->dentry->d_inode, sp);

1063
	status = _nfs4_proc_open(opendata);
L
Linus Torvalds 已提交
1064
	if (status != 0)
1065
		goto err_opendata_put;
L
Linus Torvalds 已提交
1066

1067 1068 1069
	if (opendata->o_arg.open_flags & O_EXCL)
		nfs4_exclusive_attrset(opendata, sattr);

1070
	state = nfs4_opendata_to_nfs4_state(opendata);
1071 1072
	status = PTR_ERR(state);
	if (IS_ERR(state))
1073 1074
		goto err_opendata_put;
	nfs4_opendata_put(opendata);
L
Linus Torvalds 已提交
1075 1076 1077 1078
	nfs4_put_state_owner(sp);
	up_read(&clp->cl_sem);
	*res = state;
	return 0;
1079 1080
err_opendata_put:
	nfs4_opendata_put(opendata);
1081 1082
err_release_rwsem:
	up_read(&clp->cl_sem);
1083 1084
err_put_state_owner:
	nfs4_put_state_owner(sp);
L
Linus Torvalds 已提交
1085 1086 1087 1088 1089 1090
out_err:
	*res = NULL;
	return status;
}


1091
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
1092 1093 1094 1095 1096 1097
{
	struct nfs4_exception exception = { };
	struct nfs4_state *res;
	int status;

	do {
1098
		status = _nfs4_do_open(dir, path, flags, sattr, cred, &res);
L
Linus Torvalds 已提交
1099 1100 1101 1102 1103 1104 1105 1106
		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
1107
		 * doing an OPEN, we assume that nfs_increment_open_seqid() will
L
Linus Torvalds 已提交
1108 1109 1110 1111 1112
		 * 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) {
1113 1114 1115
			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
1116 1117 1118
			exception.retry = 1;
			continue;
		}
1119 1120 1121 1122 1123 1124 1125 1126 1127 1128
		/*
		 * 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;
		}
1129 1130 1131 1132 1133
		if (status == -EAGAIN) {
			/* We must have found a delegation */
			exception.retry = 1;
			continue;
		}
L
Linus Torvalds 已提交
1134 1135 1136 1137 1138 1139
		res = ERR_PTR(nfs4_handle_exception(NFS_SERVER(dir),
					status, &exception));
	} while (exception.retry);
	return res;
}

1140 1141
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
1142
{
1143
	struct nfs_server *server = NFS_SERVER(inode);
L
Linus Torvalds 已提交
1144
        struct nfs_setattrargs  arg = {
1145
                .fh             = NFS_FH(inode),
L
Linus Torvalds 已提交
1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158
                .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,
        };
1159
	unsigned long timestamp = jiffies;
1160
	int status;
L
Linus Torvalds 已提交
1161

1162
	nfs_fattr_init(fattr);
L
Linus Torvalds 已提交
1163

1164 1165 1166
	if (nfs4_copy_delegation_stateid(&arg.stateid, inode)) {
		/* Use that stateid */
	} else if (state != NULL) {
L
Linus Torvalds 已提交
1167
		msg.rpc_cred = state->owner->so_cred;
1168 1169
		nfs4_copy_stateid(&arg.stateid, state, current->files);
	} else
L
Linus Torvalds 已提交
1170 1171
		memcpy(&arg.stateid, &zero_stateid, sizeof(arg.stateid));

1172
	status = rpc_call_sync(server->client, &msg, 0);
1173 1174
	if (status == 0 && state != NULL)
		renew_lease(server, timestamp);
1175
	return status;
L
Linus Torvalds 已提交
1176 1177
}

1178 1179
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
1180
{
1181
	struct nfs_server *server = NFS_SERVER(inode);
L
Linus Torvalds 已提交
1182 1183 1184 1185
	struct nfs4_exception exception = { };
	int err;
	do {
		err = nfs4_handle_exception(server,
1186
				_nfs4_do_setattr(inode, fattr, sattr, state),
L
Linus Torvalds 已提交
1187 1188 1189 1190 1191 1192
				&exception);
	} while (exception.retry);
	return err;
}

struct nfs4_closedata {
1193
	struct path path;
L
Linus Torvalds 已提交
1194 1195 1196 1197
	struct inode *inode;
	struct nfs4_state *state;
	struct nfs_closeargs arg;
	struct nfs_closeres res;
1198
	struct nfs_fattr fattr;
1199
	unsigned long timestamp;
L
Linus Torvalds 已提交
1200 1201
};

1202
static void nfs4_free_closedata(void *data)
1203
{
1204 1205
	struct nfs4_closedata *calldata = data;
	struct nfs4_state_owner *sp = calldata->state->owner;
1206 1207 1208 1209

	nfs4_put_open_state(calldata->state);
	nfs_free_seqid(calldata->arg.seqid);
	nfs4_put_state_owner(sp);
1210 1211
	dput(calldata->path.dentry);
	mntput(calldata->path.mnt);
1212 1213 1214
	kfree(calldata);
}

1215
static void nfs4_close_done(struct rpc_task *task, void *data)
L
Linus Torvalds 已提交
1216
{
1217
	struct nfs4_closedata *calldata = data;
L
Linus Torvalds 已提交
1218 1219 1220
	struct nfs4_state *state = calldata->state;
	struct nfs_server *server = NFS_SERVER(calldata->inode);

T
Trond Myklebust 已提交
1221 1222
	if (RPC_ASSASSINATED(task))
		return;
L
Linus Torvalds 已提交
1223 1224 1225
        /* hmm. we are done with the inode, and in the process of freeing
	 * the state_owner. we keep this around to process errors
	 */
1226
	nfs_increment_open_seqid(task->tk_status, calldata->arg.seqid);
L
Linus Torvalds 已提交
1227 1228
	switch (task->tk_status) {
		case 0:
1229
			nfs_set_open_stateid(state, &calldata->res.stateid, 0);
1230
			renew_lease(server, calldata->timestamp);
L
Linus Torvalds 已提交
1231 1232 1233 1234 1235 1236 1237 1238 1239 1240
			break;
		case -NFS4ERR_STALE_STATEID:
		case -NFS4ERR_EXPIRED:
			break;
		default:
			if (nfs4_async_handle_error(task, server) == -EAGAIN) {
				rpc_restart_call(task);
				return;
			}
	}
1241
	nfs_refresh_inode(calldata->inode, calldata->res.fattr);
L
Linus Torvalds 已提交
1242 1243
}

T
Trond Myklebust 已提交
1244
static void nfs4_close_prepare(struct rpc_task *task, void *data)
L
Linus Torvalds 已提交
1245
{
T
Trond Myklebust 已提交
1246
	struct nfs4_closedata *calldata = data;
1247
	struct nfs4_state *state = calldata->state;
L
Linus Torvalds 已提交
1248 1249 1250 1251
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_CLOSE],
		.rpc_argp = &calldata->arg,
		.rpc_resp = &calldata->res,
1252
		.rpc_cred = state->owner->so_cred,
L
Linus Torvalds 已提交
1253
	};
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) {
L
Linus Torvalds 已提交
1280
		msg.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_OPEN_DOWNGRADE];
1281 1282 1283 1284 1285
		calldata->arg.open_flags = FMODE_READ;
	} else if (test_bit(NFS_O_WRONLY_STATE, &state->flags) != 0) {
		msg.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_OPEN_DOWNGRADE];
		calldata->arg.open_flags = FMODE_WRITE;
	}
1286
	calldata->timestamp = jiffies;
1287
	rpc_call_setup(task, &msg, 0);
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)
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
	int status = -ENOMEM;
L
Linus Torvalds 已提交
1314

1315
	calldata = kmalloc(sizeof(*calldata), GFP_KERNEL);
L
Linus Torvalds 已提交
1316
	if (calldata == NULL)
1317
		goto out;
1318
	calldata->inode = state->inode;
L
Linus Torvalds 已提交
1319
	calldata->state = state;
1320
	calldata->arg.fh = NFS_FH(state->inode);
1321
	calldata->arg.stateid = &state->open_stateid;
L
Linus Torvalds 已提交
1322
	/* Serialization for the sequence id */
1323
	calldata->arg.seqid = nfs_alloc_seqid(&state->owner->so_seqid);
1324 1325
	if (calldata->arg.seqid == NULL)
		goto out_free_calldata;
1326 1327 1328
	calldata->arg.bitmask = server->attr_bitmask;
	calldata->res.fattr = &calldata->fattr;
	calldata->res.server = server;
1329 1330
	calldata->path.mnt = mntget(path->mnt);
	calldata->path.dentry = dget(path->dentry);
1331

1332 1333 1334 1335 1336
	task = rpc_run_task(server->client, RPC_TASK_ASYNC, &nfs4_close_ops, calldata);
	if (IS_ERR(task))
		return PTR_ERR(task);
	rpc_put_task(task);
	return 0;
1337 1338 1339
out_free_calldata:
	kfree(calldata);
out:
1340 1341
	nfs4_put_open_state(state);
	nfs4_put_state_owner(sp);
1342
	return status;
L
Linus Torvalds 已提交
1343 1344
}

1345
static int nfs4_intent_set_file(struct nameidata *nd, struct path *path, struct nfs4_state *state)
1346 1347
{
	struct file *filp;
1348
	int ret;
1349

1350 1351
	/* If the open_intent is for execute, we have an extra check to make */
	if (nd->intent.open.flags & FMODE_EXEC) {
1352
		ret = nfs_may_open(state->inode,
1353 1354 1355 1356 1357
				state->owner->so_cred,
				nd->intent.open.flags);
		if (ret < 0)
			goto out_close;
	}
1358
	filp = lookup_instantiate_filp(nd, path->dentry, NULL);
1359 1360
	if (!IS_ERR(filp)) {
		struct nfs_open_context *ctx;
1361
		ctx = nfs_file_open_context(filp);
1362
		ctx->state = state;
1363 1364
		return 0;
	}
1365 1366
	ret = PTR_ERR(filp);
out_close:
1367
	nfs4_close_state(path, state, nd->intent.open.flags);
1368
	return ret;
1369 1370 1371
}

struct dentry *
L
Linus Torvalds 已提交
1372 1373
nfs4_atomic_open(struct inode *dir, struct dentry *dentry, struct nameidata *nd)
{
1374 1375 1376 1377
	struct path path = {
		.mnt = nd->mnt,
		.dentry = dentry,
	};
L
Linus Torvalds 已提交
1378 1379 1380
	struct iattr attr;
	struct rpc_cred *cred;
	struct nfs4_state *state;
1381
	struct dentry *res;
L
Linus Torvalds 已提交
1382 1383 1384 1385 1386 1387 1388 1389 1390 1391 1392

	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);
	}

1393
	cred = rpcauth_lookupcred(NFS_CLIENT(dir)->cl_auth, 0);
L
Linus Torvalds 已提交
1394
	if (IS_ERR(cred))
1395
		return (struct dentry *)cred;
1396
	state = nfs4_do_open(dir, &path, nd->intent.open.flags, &attr, cred);
L
Linus Torvalds 已提交
1397
	put_rpccred(cred);
1398 1399 1400 1401 1402
	if (IS_ERR(state)) {
		if (PTR_ERR(state) == -ENOENT)
			d_add(dentry, NULL);
		return (struct dentry *)state;
	}
1403
	res = d_add_unique(dentry, igrab(state->inode));
1404
	if (res != NULL)
1405
		path.dentry = res;
1406
	nfs4_intent_set_file(nd, &path, state);
1407
	return res;
L
Linus Torvalds 已提交
1408 1409 1410
}

int
1411
nfs4_open_revalidate(struct inode *dir, struct dentry *dentry, int openflags, struct nameidata *nd)
L
Linus Torvalds 已提交
1412
{
1413 1414 1415 1416
	struct path path = {
		.mnt = nd->mnt,
		.dentry = dentry,
	};
L
Linus Torvalds 已提交
1417 1418 1419
	struct rpc_cred *cred;
	struct nfs4_state *state;

1420
	cred = rpcauth_lookupcred(NFS_CLIENT(dir)->cl_auth, 0);
L
Linus Torvalds 已提交
1421 1422
	if (IS_ERR(cred))
		return PTR_ERR(cred);
1423
	state = nfs4_do_open(dir, &path, openflags, NULL, cred);
L
Linus Torvalds 已提交
1424
	put_rpccred(cred);
1425 1426 1427 1428 1429 1430 1431 1432 1433
	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;
1434 1435
			default:
				goto out_drop;
1436 1437
		}
	}
1438
	if (state->inode == dentry->d_inode) {
1439
		nfs4_intent_set_file(nd, &path, state);
L
Linus Torvalds 已提交
1440 1441
		return 1;
	}
1442
	nfs4_close_state(&path, state, openflags);
1443 1444
out_drop:
	d_drop(dentry);
L
Linus Torvalds 已提交
1445 1446 1447 1448 1449 1450 1451 1452 1453 1454 1455 1456 1457 1458 1459 1460 1461 1462 1463 1464 1465 1466 1467 1468 1469 1470 1471 1472
	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;
}

1473
int nfs4_server_capabilities(struct nfs_server *server, struct nfs_fh *fhandle)
L
Linus Torvalds 已提交
1474 1475 1476 1477 1478 1479 1480 1481 1482 1483 1484 1485 1486 1487 1488 1489 1490 1491 1492
{
	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,
1493
		.fattr = info->fattr,
L
Linus Torvalds 已提交
1494 1495 1496 1497 1498 1499 1500
		.fh = fhandle,
	};
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_LOOKUP_ROOT],
		.rpc_argp = &args,
		.rpc_resp = &res,
	};
1501
	nfs_fattr_init(info->fattr);
L
Linus Torvalds 已提交
1502 1503 1504 1505 1506 1507 1508 1509 1510 1511 1512 1513 1514 1515 1516 1517
	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;
}

1518 1519 1520
/*
 * get the file handle for the "/" directory on the server
 */
L
Linus Torvalds 已提交
1521
static int nfs4_proc_get_root(struct nfs_server *server, struct nfs_fh *fhandle,
1522
			      struct nfs_fsinfo *info)
L
Linus Torvalds 已提交
1523 1524 1525 1526 1527 1528 1529 1530
{
	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);
1531
	return nfs4_map_errors(status);
L
Linus Torvalds 已提交
1532 1533
}

M
Manoj Naik 已提交
1534 1535 1536 1537 1538
/*
 * 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
 */
1539
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
1540 1541 1542 1543 1544 1545 1546 1547 1548 1549 1550 1551
{
	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;

1552
	status = nfs4_proc_fs_locations(dir, name, locations, page);
M
Manoj Naik 已提交
1553 1554 1555 1556 1557 1558 1559 1560 1561 1562 1563 1564 1565 1566 1567 1568 1569 1570 1571 1572 1573 1574
	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
1575 1576 1577 1578 1579 1580 1581 1582 1583 1584 1585 1586 1587 1588 1589 1590
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,
	};
	
1591
	nfs_fattr_init(fattr);
L
Linus Torvalds 已提交
1592 1593 1594 1595 1596 1597 1598 1599 1600 1601 1602 1603 1604 1605 1606 1607 1608 1609 1610 1611 1612 1613 1614 1615 1616 1617 1618 1619 1620 1621 1622 1623 1624 1625 1626 1627
	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)
{
1628 1629
	struct rpc_cred *cred;
	struct inode *inode = dentry->d_inode;
1630 1631
	struct nfs_open_context *ctx;
	struct nfs4_state *state = NULL;
L
Linus Torvalds 已提交
1632 1633
	int status;

1634
	nfs_fattr_init(fattr);
L
Linus Torvalds 已提交
1635
	
1636
	cred = rpcauth_lookupcred(NFS_CLIENT(inode)->cl_auth, 0);
1637 1638
	if (IS_ERR(cred))
		return PTR_ERR(cred);
1639 1640 1641 1642 1643

	/* Search for an existing open(O_WRITE) file */
	ctx = nfs_find_open_context(inode, cred, FMODE_WRITE);
	if (ctx != NULL)
		state = ctx->state;
1644

1645
	status = nfs4_do_setattr(inode, fattr, sattr, state);
1646 1647
	if (status == 0)
		nfs_setattr_update_inode(inode, sattr);
1648 1649
	if (ctx != NULL)
		put_nfs_open_context(ctx);
1650
	put_rpccred(cred);
L
Linus Torvalds 已提交
1651 1652 1653
	return status;
}

1654 1655
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
1656 1657 1658 1659 1660 1661 1662 1663 1664 1665 1666 1667 1668 1669 1670 1671 1672 1673 1674 1675 1676 1677 1678 1679 1680 1681 1682 1683 1684 1685 1686 1687 1688 1689
		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 {
1690 1691 1692 1693 1694 1695 1696
		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
1697 1698 1699 1700
	} while (exception.retry);
	return err;
}

1701
static int _nfs4_proc_lookup(struct inode *dir, const struct qstr *name,
L
Linus Torvalds 已提交
1702 1703
		struct nfs_fh *fhandle, struct nfs_fattr *fattr)
{
1704
	int status;
L
Linus Torvalds 已提交
1705 1706
	
	dprintk("NFS call  lookup %s\n", name->name);
1707
	status = _nfs4_proc_lookupfh(NFS_SERVER(dir), NFS_FH(dir), name, fhandle, fattr);
M
Manoj Naik 已提交
1708 1709
	if (status == -NFS4ERR_MOVED)
		status = nfs4_get_referral(dir, name, fattr, fhandle);
L
Linus Torvalds 已提交
1710 1711 1712 1713 1714 1715 1716 1717 1718 1719 1720 1721 1722 1723 1724 1725 1726 1727
	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)
{
1728 1729
	struct nfs_server *server = NFS_SERVER(inode);
	struct nfs_fattr fattr;
L
Linus Torvalds 已提交
1730 1731
	struct nfs4_accessargs args = {
		.fh = NFS_FH(inode),
1732 1733 1734 1735 1736
		.bitmask = server->attr_bitmask,
	};
	struct nfs4_accessres res = {
		.server = server,
		.fattr = &fattr,
L
Linus Torvalds 已提交
1737 1738 1739 1740 1741 1742 1743 1744 1745 1746 1747 1748 1749 1750 1751 1752 1753 1754 1755 1756 1757 1758 1759 1760 1761 1762
	};
	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;
	}
1763
	nfs_fattr_init(&fattr);
L
Linus Torvalds 已提交
1764 1765 1766 1767 1768 1769 1770 1771 1772
	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;
1773
		nfs_refresh_inode(inode, &fattr);
L
Linus Torvalds 已提交
1774 1775 1776 1777 1778 1779 1780 1781 1782 1783 1784 1785 1786 1787 1788 1789 1790 1791 1792 1793 1794 1795 1796 1797 1798 1799 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
	}
	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,
1862
                 int flags, struct nameidata *nd)
L
Linus Torvalds 已提交
1863
{
1864 1865 1866 1867
	struct path path = {
		.mnt = nd->mnt,
		.dentry = dentry,
	};
L
Linus Torvalds 已提交
1868 1869 1870 1871
	struct nfs4_state *state;
	struct rpc_cred *cred;
	int status = 0;

1872
	cred = rpcauth_lookupcred(NFS_CLIENT(dir)->cl_auth, 0);
L
Linus Torvalds 已提交
1873 1874 1875 1876
	if (IS_ERR(cred)) {
		status = PTR_ERR(cred);
		goto out;
	}
1877
	state = nfs4_do_open(dir, &path, flags, sattr, cred);
L
Linus Torvalds 已提交
1878 1879 1880 1881 1882
	put_rpccred(cred);
	if (IS_ERR(state)) {
		status = PTR_ERR(state);
		goto out;
	}
1883
	d_instantiate(dentry, igrab(state->inode));
L
Linus Torvalds 已提交
1884 1885
	if (flags & O_EXCL) {
		struct nfs_fattr fattr;
1886
		status = nfs4_do_setattr(state->inode, &fattr, sattr, state);
1887
		if (status == 0)
1888
			nfs_setattr_update_inode(state->inode, sattr);
1889
		nfs_post_op_update_inode(state->inode, &fattr);
1890
	}
1891
	if (status == 0 && (nd->flags & LOOKUP_OPEN) != 0)
1892
		status = nfs4_intent_set_file(nd, &path, state);
1893
	else
1894
		nfs4_close_state(&path, state, flags);
L
Linus Torvalds 已提交
1895 1896 1897 1898 1899 1900
out:
	return status;
}

static int _nfs4_proc_remove(struct inode *dir, struct qstr *name)
{
1901
	struct nfs_server *server = NFS_SERVER(dir);
1902
	struct nfs_removeargs args = {
L
Linus Torvalds 已提交
1903
		.fh = NFS_FH(dir),
1904 1905
		.name.len = name->len,
		.name.name = name->name,
1906 1907
		.bitmask = server->attr_bitmask,
	};
1908
	struct nfs_removeres res = {
1909
		.server = server,
L
Linus Torvalds 已提交
1910 1911
	};
	struct rpc_message msg = {
1912 1913 1914
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_REMOVE],
		.rpc_argp = &args,
		.rpc_resp = &res,
L
Linus Torvalds 已提交
1915 1916 1917
	};
	int			status;

1918
	nfs_fattr_init(&res.dir_attr);
1919 1920 1921
	status = rpc_call_sync(server->client, &msg, 0);
	if (status == 0) {
		update_changeattr(dir, &res.cinfo);
1922
		nfs_post_op_update_inode(dir, &res.dir_attr);
1923
	}
L
Linus Torvalds 已提交
1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 1934 1935 1936 1937 1938
	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;
}

1939
static void nfs4_proc_unlink_setup(struct rpc_message *msg, struct inode *dir)
L
Linus Torvalds 已提交
1940
{
1941 1942 1943
	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
1944

1945 1946
	args->bitmask = server->attr_bitmask;
	res->server = server;
L
Linus Torvalds 已提交
1947 1948 1949
	msg->rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_REMOVE];
}

1950
static int nfs4_proc_unlink_done(struct rpc_task *task, struct inode *dir)
L
Linus Torvalds 已提交
1951
{
1952 1953 1954 1955 1956 1957 1958
	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
1959 1960 1961 1962 1963
}

static int _nfs4_proc_rename(struct inode *old_dir, struct qstr *old_name,
		struct inode *new_dir, struct qstr *new_name)
{
1964
	struct nfs_server *server = NFS_SERVER(old_dir);
L
Linus Torvalds 已提交
1965 1966 1967 1968 1969
	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,
1970 1971 1972 1973 1974 1975 1976
		.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
1977 1978 1979 1980 1981 1982 1983 1984
	};
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_RENAME],
		.rpc_argp = &arg,
		.rpc_resp = &res,
	};
	int			status;
	
1985 1986 1987
	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
1988 1989 1990

	if (!status) {
		update_changeattr(old_dir, &res.old_cinfo);
1991
		nfs_post_op_update_inode(old_dir, res.old_fattr);
L
Linus Torvalds 已提交
1992
		update_changeattr(new_dir, &res.new_cinfo);
1993
		nfs_post_op_update_inode(new_dir, res.new_fattr);
L
Linus Torvalds 已提交
1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013
	}
	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)
{
2014
	struct nfs_server *server = NFS_SERVER(inode);
L
Linus Torvalds 已提交
2015 2016 2017 2018
	struct nfs4_link_arg arg = {
		.fh     = NFS_FH(inode),
		.dir_fh = NFS_FH(dir),
		.name   = name,
2019 2020 2021 2022 2023 2024 2025
		.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
2026 2027 2028 2029
	};
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_LINK],
		.rpc_argp = &arg,
2030
		.rpc_resp = &res,
L
Linus Torvalds 已提交
2031 2032 2033
	};
	int			status;

2034 2035 2036 2037 2038 2039
	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);
2040
		nfs_post_op_update_inode(inode, res.fattr);
2041
	}
L
Linus Torvalds 已提交
2042 2043 2044 2045 2046 2047 2048 2049 2050 2051 2052 2053 2054 2055 2056 2057

	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;
}

2058
static int _nfs4_proc_symlink(struct inode *dir, struct dentry *dentry,
2059
		struct page *page, unsigned int len, struct iattr *sattr)
L
Linus Torvalds 已提交
2060 2061
{
	struct nfs_server *server = NFS_SERVER(dir);
2062 2063
	struct nfs_fh fhandle;
	struct nfs_fattr fattr, dir_fattr;
L
Linus Torvalds 已提交
2064 2065 2066
	struct nfs4_create_arg arg = {
		.dir_fh = NFS_FH(dir),
		.server = server,
2067
		.name = &dentry->d_name,
L
Linus Torvalds 已提交
2068 2069 2070 2071 2072 2073
		.attrs = sattr,
		.ftype = NF4LNK,
		.bitmask = server->attr_bitmask,
	};
	struct nfs4_create_res res = {
		.server = server,
2074 2075
		.fh = &fhandle,
		.fattr = &fattr,
2076
		.dir_fattr = &dir_fattr,
L
Linus Torvalds 已提交
2077 2078 2079 2080 2081 2082 2083 2084
	};
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_SYMLINK],
		.rpc_argp = &arg,
		.rpc_resp = &res,
	};
	int			status;

2085
	if (len > NFS4_MAXPATHLEN)
L
Linus Torvalds 已提交
2086
		return -ENAMETOOLONG;
2087

2088 2089
	arg.u.symlink.pages = &page;
	arg.u.symlink.len = len;
2090
	nfs_fattr_init(&fattr);
2091
	nfs_fattr_init(&dir_fattr);
L
Linus Torvalds 已提交
2092 2093
	
	status = rpc_call_sync(NFS_CLIENT(dir), &msg, 0);
2094
	if (!status) {
L
Linus Torvalds 已提交
2095
		update_changeattr(dir, &res.dir_cinfo);
2096 2097 2098
		nfs_post_op_update_inode(dir, res.dir_fattr);
		status = nfs_instantiate(dentry, &fhandle, &fattr);
	}
L
Linus Torvalds 已提交
2099 2100 2101
	return status;
}

2102
static int nfs4_proc_symlink(struct inode *dir, struct dentry *dentry,
2103
		struct page *page, unsigned int len, struct iattr *sattr)
L
Linus Torvalds 已提交
2104 2105 2106 2107 2108
{
	struct nfs4_exception exception = { };
	int err;
	do {
		err = nfs4_handle_exception(NFS_SERVER(dir),
2109 2110
				_nfs4_proc_symlink(dir, dentry, page,
							len, sattr),
L
Linus Torvalds 已提交
2111 2112 2113 2114 2115 2116 2117 2118 2119 2120
				&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;
2121
	struct nfs_fattr fattr, dir_fattr;
L
Linus Torvalds 已提交
2122 2123 2124 2125 2126 2127 2128 2129 2130 2131 2132 2133
	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,
2134
		.dir_fattr = &dir_fattr,
L
Linus Torvalds 已提交
2135 2136 2137 2138 2139 2140 2141 2142
	};
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_CREATE],
		.rpc_argp = &arg,
		.rpc_resp = &res,
	};
	int			status;

2143
	nfs_fattr_init(&fattr);
2144
	nfs_fattr_init(&dir_fattr);
L
Linus Torvalds 已提交
2145 2146 2147 2148
	
	status = rpc_call_sync(NFS_CLIENT(dir), &msg, 0);
	if (!status) {
		update_changeattr(dir, &res.dir_cinfo);
2149
		nfs_post_op_update_inode(dir, res.dir_fattr);
L
Linus Torvalds 已提交
2150 2151 2152 2153 2154 2155 2156 2157 2158 2159 2160 2161 2162 2163 2164 2165 2166 2167 2168 2169 2170 2171 2172 2173 2174 2175 2176 2177 2178 2179 2180 2181 2182 2183 2184 2185 2186 2187
		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;

2188 2189 2190 2191
	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
2192 2193 2194 2195 2196
	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);
2197
	dprintk("%s: returns %d\n", __FUNCTION__, status);
L
Linus Torvalds 已提交
2198 2199 2200 2201 2202 2203 2204 2205 2206 2207 2208 2209 2210 2211 2212 2213 2214 2215 2216 2217 2218 2219
	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;
2220
	struct nfs_fattr fattr, dir_fattr;
L
Linus Torvalds 已提交
2221 2222 2223 2224 2225 2226 2227 2228 2229 2230 2231
	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,
2232
		.dir_fattr = &dir_fattr,
L
Linus Torvalds 已提交
2233 2234 2235 2236 2237 2238 2239 2240 2241
	};
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_CREATE],
		.rpc_argp = &arg,
		.rpc_resp = &res,
	};
	int			status;
	int                     mode = sattr->ia_mode;

2242
	nfs_fattr_init(&fattr);
2243
	nfs_fattr_init(&dir_fattr);
L
Linus Torvalds 已提交
2244 2245 2246 2247 2248 2249 2250 2251 2252 2253 2254 2255 2256 2257 2258 2259 2260 2261 2262 2263 2264

	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);
2265
		nfs_post_op_update_inode(dir, res.dir_fattr);
L
Linus Torvalds 已提交
2266 2267 2268 2269 2270 2271 2272 2273 2274 2275 2276 2277 2278 2279 2280 2281 2282 2283 2284 2285 2286 2287 2288 2289 2290 2291 2292 2293 2294 2295 2296
		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,
	};

2297
	nfs_fattr_init(fsstat->fattr);
L
Linus Torvalds 已提交
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 2328 2329 2330 2331 2332 2333 2334 2335 2336 2337 2338 2339 2340 2341 2342 2343
	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)
{
2344
	nfs_fattr_init(fsinfo->fattr);
L
Linus Torvalds 已提交
2345 2346 2347 2348 2349 2350 2351 2352 2353 2354 2355 2356 2357 2358 2359 2360 2361 2362 2363 2364 2365 2366
	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;
	}

2367
	nfs_fattr_init(pathconf->fattr);
L
Linus Torvalds 已提交
2368 2369 2370 2371 2372 2373 2374 2375 2376 2377 2378 2379 2380 2381 2382 2383 2384
	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
2385
static int nfs4_read_done(struct rpc_task *task, struct nfs_read_data *data)
L
Linus Torvalds 已提交
2386
{
T
Trond Myklebust 已提交
2387
	struct nfs_server *server = NFS_SERVER(data->inode);
L
Linus Torvalds 已提交
2388

T
Trond Myklebust 已提交
2389
	if (nfs4_async_handle_error(task, server) == -EAGAIN) {
L
Linus Torvalds 已提交
2390
		rpc_restart_call(task);
T
Trond Myklebust 已提交
2391
		return -EAGAIN;
L
Linus Torvalds 已提交
2392 2393
	}
	if (task->tk_status > 0)
T
Trond Myklebust 已提交
2394 2395
		renew_lease(server, data->timestamp);
	return 0;
L
Linus Torvalds 已提交
2396 2397
}

T
Trond Myklebust 已提交
2398
static void nfs4_proc_read_setup(struct nfs_read_data *data)
L
Linus Torvalds 已提交
2399 2400 2401 2402 2403 2404 2405 2406 2407 2408
{
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_READ],
		.rpc_argp = &data->args,
		.rpc_resp = &data->res,
		.rpc_cred = data->cred,
	};

	data->timestamp   = jiffies;

T
Trond Myklebust 已提交
2409
	rpc_call_setup(&data->task, &msg, 0);
L
Linus Torvalds 已提交
2410 2411
}

2412
static int nfs4_write_done(struct rpc_task *task, struct nfs_write_data *data)
L
Linus Torvalds 已提交
2413 2414 2415 2416 2417
{
	struct inode *inode = data->inode;
	
	if (nfs4_async_handle_error(task, NFS_SERVER(inode)) == -EAGAIN) {
		rpc_restart_call(task);
2418
		return -EAGAIN;
L
Linus Torvalds 已提交
2419
	}
2420
	if (task->tk_status >= 0) {
L
Linus Torvalds 已提交
2421
		renew_lease(NFS_SERVER(inode), data->timestamp);
2422 2423
		nfs_post_op_update_inode(inode, data->res.fattr);
	}
2424
	return 0;
L
Linus Torvalds 已提交
2425 2426
}

2427
static void nfs4_proc_write_setup(struct nfs_write_data *data, int how)
L
Linus Torvalds 已提交
2428 2429 2430 2431 2432 2433 2434 2435
{
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_WRITE],
		.rpc_argp = &data->args,
		.rpc_resp = &data->res,
		.rpc_cred = data->cred,
	};
	struct inode *inode = data->inode;
2436
	struct nfs_server *server = NFS_SERVER(inode);
L
Linus Torvalds 已提交
2437 2438 2439 2440 2441 2442 2443 2444 2445 2446
	int stable;
	
	if (how & FLUSH_STABLE) {
		if (!NFS_I(inode)->ncommit)
			stable = NFS_FILE_SYNC;
		else
			stable = NFS_DATA_SYNC;
	} else
		stable = NFS_UNSTABLE;
	data->args.stable = stable;
2447 2448
	data->args.bitmask = server->attr_bitmask;
	data->res.server = server;
L
Linus Torvalds 已提交
2449 2450 2451 2452

	data->timestamp   = jiffies;

	/* Finalize the task. */
2453
	rpc_call_setup(&data->task, &msg, 0);
L
Linus Torvalds 已提交
2454 2455
}

2456
static int nfs4_commit_done(struct rpc_task *task, struct nfs_write_data *data)
L
Linus Torvalds 已提交
2457 2458 2459 2460 2461
{
	struct inode *inode = data->inode;
	
	if (nfs4_async_handle_error(task, NFS_SERVER(inode)) == -EAGAIN) {
		rpc_restart_call(task);
2462
		return -EAGAIN;
L
Linus Torvalds 已提交
2463
	}
2464 2465
	if (task->tk_status >= 0)
		nfs_post_op_update_inode(inode, data->res.fattr);
2466
	return 0;
L
Linus Torvalds 已提交
2467 2468
}

2469
static void nfs4_proc_commit_setup(struct nfs_write_data *data, int how)
L
Linus Torvalds 已提交
2470 2471 2472 2473 2474 2475 2476
{
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_COMMIT],
		.rpc_argp = &data->args,
		.rpc_resp = &data->res,
		.rpc_cred = data->cred,
	};	
2477
	struct nfs_server *server = NFS_SERVER(data->inode);
L
Linus Torvalds 已提交
2478
	
2479 2480 2481
	data->args.bitmask = server->attr_bitmask;
	data->res.server = server;

2482
	rpc_call_setup(&data->task, &msg, 0);
L
Linus Torvalds 已提交
2483 2484 2485 2486 2487 2488
}

/*
 * nfs4_proc_async_renew(): This is not one of the nfs_rpc_ops; it is a special
 * standalone procedure for queueing an asynchronous RENEW.
 */
2489
static void nfs4_renew_done(struct rpc_task *task, void *data)
L
Linus Torvalds 已提交
2490
{
2491
	struct nfs_client *clp = (struct nfs_client *)task->tk_msg.rpc_argp;
2492
	unsigned long timestamp = (unsigned long)data;
L
Linus Torvalds 已提交
2493 2494 2495 2496 2497 2498 2499 2500 2501 2502 2503 2504 2505 2506 2507 2508

	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);
}

2509 2510 2511 2512
static const struct rpc_call_ops nfs4_renew_ops = {
	.rpc_call_done = nfs4_renew_done,
};

2513
int nfs4_proc_async_renew(struct nfs_client *clp, struct rpc_cred *cred)
L
Linus Torvalds 已提交
2514 2515 2516 2517
{
	struct rpc_message msg = {
		.rpc_proc	= &nfs4_procedures[NFSPROC4_CLNT_RENEW],
		.rpc_argp	= clp,
2518
		.rpc_cred	= cred,
L
Linus Torvalds 已提交
2519 2520 2521
	};

	return rpc_call_async(clp->cl_rpcclient, &msg, RPC_TASK_SOFT,
2522
			&nfs4_renew_ops, (void *)jiffies);
L
Linus Torvalds 已提交
2523 2524
}

2525
int nfs4_proc_renew(struct nfs_client *clp, struct rpc_cred *cred)
L
Linus Torvalds 已提交
2526 2527 2528 2529
{
	struct rpc_message msg = {
		.rpc_proc	= &nfs4_procedures[NFSPROC4_CLNT_RENEW],
		.rpc_argp	= clp,
2530
		.rpc_cred	= cred,
L
Linus Torvalds 已提交
2531 2532 2533 2534 2535 2536 2537 2538 2539 2540 2541 2542 2543 2544
	};
	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;
}

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 2570
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;
	}
}

2571 2572 2573
struct nfs4_cached_acl {
	int cached;
	size_t len;
A
Andrew Morton 已提交
2574
	char data[0];
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 2637
};

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);
}

2638
static ssize_t __nfs4_get_acl_uncached(struct inode *inode, void *buf, size_t buflen)
2639 2640 2641 2642 2643 2644 2645 2646
{
	struct page *pages[NFS4ACL_MAXPAGES];
	struct nfs_getaclargs args = {
		.fh = NFS_FH(inode),
		.acl_pages = pages,
		.acl_len = buflen,
	};
	size_t resp_len = buflen;
2647
	void *resp_buf;
2648 2649 2650 2651 2652
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_GETACL],
		.rpc_argp = &args,
		.rpc_resp = &resp_len,
	};
2653
	struct page *localpage = NULL;
2654 2655
	int ret;

2656 2657 2658 2659 2660 2661 2662 2663 2664
	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
2665
		resp_len = args.acl_len = PAGE_SIZE;
2666 2667 2668 2669
	} else {
		resp_buf = buf;
		buf_to_pages(buf, buflen, args.acl_pages, &args.acl_pgbase);
	}
2670
	ret = rpc_call_sync(NFS_CLIENT(inode), &msg, 0);
2671 2672 2673 2674 2675 2676 2677 2678 2679 2680 2681 2682 2683 2684 2685 2686 2687
	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);
2688 2689 2690
	return ret;
}

2691 2692 2693 2694 2695 2696 2697 2698 2699 2700 2701 2702 2703
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;
}

2704 2705 2706 2707 2708 2709 2710 2711 2712 2713 2714 2715 2716 2717 2718 2719
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);
}

2720
static int __nfs4_proc_set_acl(struct inode *inode, const void *buf, size_t buflen)
2721 2722 2723 2724 2725 2726 2727 2728 2729 2730 2731 2732 2733 2734 2735 2736 2737
{
	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;
2738
	nfs_inode_return_delegation(inode);
2739
	buf_to_pages(buf, buflen, arg.acl_pages, &arg.acl_pgbase);
2740
	ret = rpc_call_sync(NFS_CLIENT(inode), &msg, 0);
2741
	nfs_zap_caches(inode);
2742 2743 2744
	return ret;
}

2745 2746 2747 2748 2749 2750 2751 2752 2753 2754 2755 2756
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
2757
static int
2758
nfs4_async_handle_error(struct rpc_task *task, const struct nfs_server *server)
L
Linus Torvalds 已提交
2759
{
2760
	struct nfs_client *clp = server->nfs_client;
L
Linus Torvalds 已提交
2761 2762 2763 2764 2765 2766 2767 2768 2769

	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
2770
			if (test_bit(NFS4CLNT_STATE_RECOVER, &clp->cl_state) == 0)
L
Linus Torvalds 已提交
2771 2772 2773 2774
				rpc_wake_up_task(task);
			task->tk_status = 0;
			return -EAGAIN;
		case -NFS4ERR_DELAY:
2775 2776 2777
			nfs_inc_server_stats((struct nfs_server *) server,
						NFSIOS_DELAY);
		case -NFS4ERR_GRACE:
L
Linus Torvalds 已提交
2778 2779 2780 2781 2782 2783 2784 2785 2786 2787 2788
			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
2789 2790 2791 2792 2793 2794 2795 2796
static int nfs4_wait_bit_interruptible(void *word)
{
	if (signal_pending(current))
		return -ERESTARTSYS;
	schedule();
	return 0;
}

2797
static int nfs4_wait_clnt_recover(struct rpc_clnt *clnt, struct nfs_client *clp)
L
Linus Torvalds 已提交
2798 2799
{
	sigset_t oldset;
T
Trond Myklebust 已提交
2800
	int res;
L
Linus Torvalds 已提交
2801 2802 2803

	might_sleep();

2804 2805
	rwsem_acquire(&clp->cl_sem.dep_map, 0, 0, _RET_IP_);

L
Linus Torvalds 已提交
2806
	rpc_clnt_sigmask(clnt, &oldset);
T
Trond Myklebust 已提交
2807 2808 2809
	res = wait_on_bit(&clp->cl_state, NFS4CLNT_STATE_RECOVER,
			nfs4_wait_bit_interruptible,
			TASK_INTERRUPTIBLE);
L
Linus Torvalds 已提交
2810
	rpc_clnt_sigunmask(clnt, &oldset);
2811 2812

	rwsem_release(&clp->cl_sem.dep_map, 1, _RET_IP_);
L
Linus Torvalds 已提交
2813 2814 2815 2816 2817 2818 2819 2820 2821 2822 2823 2824 2825 2826 2827 2828
	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) {
2829
		schedule_timeout_interruptible(*timeout);
L
Linus Torvalds 已提交
2830 2831
		if (signalled())
			res = -ERESTARTSYS;
2832 2833
	} else
		schedule_timeout_uninterruptible(*timeout);
L
Linus Torvalds 已提交
2834 2835 2836 2837 2838 2839 2840 2841
	rpc_clnt_sigunmask(clnt, &oldset);
	*timeout <<= 1;
	return res;
}

/* This is the error handling routine for processes that are allowed
 * to sleep.
 */
2842
static int nfs4_handle_exception(const struct nfs_server *server, int errorcode, struct nfs4_exception *exception)
L
Linus Torvalds 已提交
2843
{
2844
	struct nfs_client *clp = server->nfs_client;
L
Linus Torvalds 已提交
2845 2846 2847 2848 2849 2850 2851 2852 2853
	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
2854
			nfs4_schedule_state_recovery(clp);
L
Linus Torvalds 已提交
2855 2856 2857 2858
			ret = nfs4_wait_clnt_recover(server->client, clp);
			if (ret == 0)
				exception->retry = 1;
			break;
2859
		case -NFS4ERR_FILE_OPEN:
L
Linus Torvalds 已提交
2860 2861 2862
		case -NFS4ERR_GRACE:
		case -NFS4ERR_DELAY:
			ret = nfs4_delay(server->client, &exception->timeout);
2863 2864
			if (ret != 0)
				break;
L
Linus Torvalds 已提交
2865
		case -NFS4ERR_OLD_STATEID:
2866
			exception->retry = 1;
L
Linus Torvalds 已提交
2867 2868 2869 2870 2871
	}
	/* We failed to handle the error */
	return nfs4_map_errors(ret);
}

2872
int nfs4_proc_setclientid(struct nfs_client *clp, u32 program, unsigned short port, struct rpc_cred *cred)
L
Linus Torvalds 已提交
2873 2874 2875 2876 2877 2878 2879 2880 2881 2882
{
	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,
2883
		.rpc_cred = cred,
L
Linus Torvalds 已提交
2884
	};
A
Al Viro 已提交
2885
	__be32 *p;
L
Linus Torvalds 已提交
2886 2887 2888
	int loop = 0;
	int status;

A
Al Viro 已提交
2889
	p = (__be32*)sc_verifier.data;
L
Linus Torvalds 已提交
2890 2891 2892 2893 2894 2895
	*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",
2896
				clp->cl_ipaddr, NIPQUAD(clp->cl_addr.sin_addr),
2897
				cred->cr_ops->cr_name,
L
Linus Torvalds 已提交
2898 2899 2900 2901 2902 2903 2904 2905 2906 2907 2908 2909 2910 2911 2912 2913 2914 2915 2916 2917 2918
				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;
}

2919
static int _nfs4_proc_setclientid_confirm(struct nfs_client *clp, struct rpc_cred *cred)
L
Linus Torvalds 已提交
2920 2921 2922 2923 2924 2925
{
	struct nfs_fsinfo fsinfo;
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_SETCLIENTID_CONFIRM],
		.rpc_argp = clp,
		.rpc_resp = &fsinfo,
2926
		.rpc_cred = cred,
L
Linus Torvalds 已提交
2927 2928 2929 2930 2931 2932 2933 2934 2935 2936
	};
	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;
2937
		clear_bit(NFS4CLNT_LEASE_EXPIRED, &clp->cl_state);
L
Linus Torvalds 已提交
2938 2939 2940 2941 2942
		spin_unlock(&clp->cl_lock);
	}
	return status;
}

2943
int nfs4_proc_setclientid_confirm(struct nfs_client *clp, struct rpc_cred *cred)
2944 2945 2946 2947 2948 2949 2950 2951 2952 2953 2954 2955 2956 2957 2958 2959 2960
{
	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;
}

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

static void nfs4_delegreturn_prepare(struct rpc_task *task, void *calldata)
L
Linus Torvalds 已提交
2973
{
2974
	struct nfs4_delegreturndata *data = calldata;
L
Linus Torvalds 已提交
2975 2976
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_DELEGRETURN],
2977
		.rpc_argp = &data->args,
2978
		.rpc_resp = &data->res,
2979
		.rpc_cred = data->cred,
L
Linus Torvalds 已提交
2980
	};
2981
	nfs_fattr_init(data->res.fattr);
2982 2983
	rpc_call_setup(task, &msg, 0);
}
L
Linus Torvalds 已提交
2984

2985 2986 2987 2988
static void nfs4_delegreturn_done(struct rpc_task *task, void *calldata)
{
	struct nfs4_delegreturndata *data = calldata;
	data->rpc_status = task->tk_status;
2989
	if (data->rpc_status == 0)
2990
		renew_lease(data->res.server, data->timestamp);
2991 2992 2993 2994 2995 2996 2997 2998 2999 3000
}

static void nfs4_delegreturn_release(void *calldata)
{
	struct nfs4_delegreturndata *data = calldata;

	put_rpccred(data->cred);
	kfree(calldata);
}

3001
static const struct rpc_call_ops nfs4_delegreturn_ops = {
3002 3003 3004 3005 3006 3007 3008 3009
	.rpc_call_prepare = nfs4_delegreturn_prepare,
	.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;
3010
	struct nfs_server *server = NFS_SERVER(inode);
3011 3012 3013 3014 3015 3016 3017 3018
	struct rpc_task *task;
	int status;

	data = kmalloc(sizeof(*data), GFP_KERNEL);
	if (data == NULL)
		return -ENOMEM;
	data->args.fhandle = &data->fh;
	data->args.stateid = &data->stateid;
3019
	data->args.bitmask = server->attr_bitmask;
3020 3021
	nfs_copy_fh(&data->fh, NFS_FH(inode));
	memcpy(&data->stateid, stateid, sizeof(data->stateid));
3022 3023
	data->res.fattr = &data->fattr;
	data->res.server = server;
3024
	data->cred = get_rpccred(cred);
3025
	data->timestamp = jiffies;
3026 3027 3028
	data->rpc_status = 0;

	task = rpc_run_task(NFS_CLIENT(inode), RPC_TASK_ASYNC, &nfs4_delegreturn_ops, data);
3029
	if (IS_ERR(task))
3030 3031
		return PTR_ERR(task);
	status = nfs4_wait_for_completion_rpc_task(task);
3032
	if (status == 0) {
3033
		status = data->rpc_status;
3034 3035 3036
		if (status == 0)
			nfs_post_op_update_inode(inode, &data->fattr);
	}
3037
	rpc_put_task(task);
3038
	return status;
L
Linus Torvalds 已提交
3039 3040 3041 3042 3043 3044 3045 3046 3047 3048 3049 3050 3051 3052 3053 3054 3055 3056 3057 3058 3059 3060 3061 3062 3063 3064 3065 3066 3067
}

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)
{
3068
	schedule_timeout_interruptible(timeout);
L
Linus Torvalds 已提交
3069 3070 3071 3072 3073 3074 3075 3076 3077 3078
	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);
3079
	struct nfs_client *clp = server->nfs_client;
T
Trond Myklebust 已提交
3080
	struct nfs_lockt_args arg = {
L
Linus Torvalds 已提交
3081
		.fh = NFS_FH(inode),
T
Trond Myklebust 已提交
3082
		.fl = request,
L
Linus Torvalds 已提交
3083
	};
T
Trond Myklebust 已提交
3084 3085
	struct nfs_lockt_res res = {
		.denied = request,
L
Linus Torvalds 已提交
3086 3087 3088 3089 3090 3091 3092 3093 3094 3095 3096
	};
	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
3097
	arg.lock_owner.clientid = clp->cl_clientid;
3098 3099 3100 3101
	status = nfs4_set_lock_state(state, request);
	if (status != 0)
		goto out;
	lsp = request->fl_u.nfs4_fl.owner;
3102
	arg.lock_owner.id = lsp->ls_id.id;
L
Linus Torvalds 已提交
3103
	status = rpc_call_sync(server->client, &msg, 0);
T
Trond Myklebust 已提交
3104 3105 3106 3107 3108 3109
	switch (status) {
		case 0:
			request->fl_type = F_UNLCK;
			break;
		case -NFS4ERR_DENIED:
			status = 0;
L
Linus Torvalds 已提交
3110
	}
3111
	request->fl_ops->fl_release_private(request);
3112
out:
L
Linus Torvalds 已提交
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 3140 3141 3142 3143 3144 3145
	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;
}

3146
struct nfs4_unlockdata {
T
Trond Myklebust 已提交
3147 3148
	struct nfs_locku_args arg;
	struct nfs_locku_res res;
3149 3150
	struct nfs4_lock_state *lsp;
	struct nfs_open_context *ctx;
T
Trond Myklebust 已提交
3151 3152
	struct file_lock fl;
	const struct nfs_server *server;
3153
	unsigned long timestamp;
3154 3155
};

T
Trond Myklebust 已提交
3156 3157 3158 3159 3160 3161 3162 3163 3164 3165 3166 3167 3168 3169 3170 3171 3172 3173 3174 3175 3176 3177 3178 3179
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;
}

3180
static void nfs4_locku_release_calldata(void *data)
3181
{
3182
	struct nfs4_unlockdata *calldata = data;
T
Trond Myklebust 已提交
3183
	nfs_free_seqid(calldata->arg.seqid);
3184 3185 3186
	nfs4_put_lock_state(calldata->lsp);
	put_nfs_open_context(calldata->ctx);
	kfree(calldata);
3187 3188
}

3189
static void nfs4_locku_done(struct rpc_task *task, void *data)
3190
{
3191
	struct nfs4_unlockdata *calldata = data;
3192

3193 3194
	if (RPC_ASSASSINATED(task))
		return;
T
Trond Myklebust 已提交
3195
	nfs_increment_lock_seqid(task->tk_status, calldata->arg.seqid);
3196 3197 3198
	switch (task->tk_status) {
		case 0:
			memcpy(calldata->lsp->ls_stateid.data,
T
Trond Myklebust 已提交
3199
					calldata->res.stateid.data,
3200
					sizeof(calldata->lsp->ls_stateid.data));
3201
			renew_lease(calldata->server, calldata->timestamp);
3202 3203 3204 3205 3206
			break;
		case -NFS4ERR_STALE_STATEID:
		case -NFS4ERR_EXPIRED:
			break;
		default:
3207
			if (nfs4_async_handle_error(task, calldata->server) == -EAGAIN)
3208 3209 3210 3211
				rpc_restart_call(task);
	}
}

T
Trond Myklebust 已提交
3212
static void nfs4_locku_prepare(struct rpc_task *task, void *data)
3213
{
T
Trond Myklebust 已提交
3214
	struct nfs4_unlockdata *calldata = data;
L
Linus Torvalds 已提交
3215 3216
	struct rpc_message msg = {
		.rpc_proc	= &nfs4_procedures[NFSPROC4_CLNT_LOCKU],
3217 3218 3219
		.rpc_argp       = &calldata->arg,
		.rpc_resp       = &calldata->res,
		.rpc_cred	= calldata->lsp->ls_state->owner->so_cred,
L
Linus Torvalds 已提交
3220
	};
3221

T
Trond Myklebust 已提交
3222
	if (nfs_wait_on_sequence(calldata->arg.seqid, task) != 0)
3223 3224
		return;
	if ((calldata->lsp->ls_flags & NFS_LOCK_INITIALIZED) == 0) {
3225 3226
		/* Note: exit _without_ running nfs4_locku_done */
		task->tk_action = NULL;
3227 3228
		return;
	}
3229
	calldata->timestamp = jiffies;
3230 3231 3232
	rpc_call_setup(task, &msg, 0);
}

3233
static const struct rpc_call_ops nfs4_locku_ops = {
T
Trond Myklebust 已提交
3234
	.rpc_call_prepare = nfs4_locku_prepare,
3235
	.rpc_call_done = nfs4_locku_done,
3236
	.rpc_release = nfs4_locku_release_calldata,
3237 3238
};

3239 3240 3241 3242 3243 3244 3245
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;

3246 3247 3248 3249 3250
	/* 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;

3251 3252 3253 3254 3255 3256
	data = nfs4_alloc_unlockdata(fl, ctx, lsp, seqid);
	if (data == NULL) {
		nfs_free_seqid(seqid);
		return ERR_PTR(-ENOMEM);
	}

3257
	return rpc_run_task(NFS_CLIENT(lsp->ls_state->inode), RPC_TASK_ASYNC, &nfs4_locku_ops, data);
3258 3259
}

3260 3261
static int nfs4_proc_unlck(struct nfs4_state *state, int cmd, struct file_lock *request)
{
T
Trond Myklebust 已提交
3262
	struct nfs_seqid *seqid;
L
Linus Torvalds 已提交
3263
	struct nfs4_lock_state *lsp;
3264 3265
	struct rpc_task *task;
	int status = 0;
3266

3267
	status = nfs4_set_lock_state(state, request);
3268 3269 3270 3271
	/* Unlock _before_ we do the RPC call */
	request->fl_flags |= FL_EXISTS;
	if (do_vfs_lock(request->fl_file, request) == -ENOENT)
		goto out;
3272
	if (status != 0)
3273 3274 3275 3276
		goto out;
	/* Is this a delegated lock? */
	if (test_bit(NFS_DELEGATED_STATE, &state->flags))
		goto out;
3277
	lsp = request->fl_u.nfs4_fl.owner;
T
Trond Myklebust 已提交
3278
	seqid = nfs_alloc_seqid(&lsp->ls_seqid);
3279
	status = -ENOMEM;
T
Trond Myklebust 已提交
3280
	if (seqid == NULL)
3281
		goto out;
3282
	task = nfs4_do_unlck(request, nfs_file_open_context(request->fl_file), lsp, seqid);
3283 3284
	status = PTR_ERR(task);
	if (IS_ERR(task))
3285
		goto out;
3286
	status = nfs4_wait_for_completion_rpc_task(task);
3287
	rpc_put_task(task);
3288
out:
L
Linus Torvalds 已提交
3289 3290 3291
	return status;
}

3292 3293 3294 3295 3296 3297
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;
3298
	unsigned long timestamp;
3299 3300 3301 3302 3303 3304
	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
3305
{
3306 3307
	struct nfs4_lockdata *p;
	struct inode *inode = lsp->ls_state->inode;
L
Linus Torvalds 已提交
3308
	struct nfs_server *server = NFS_SERVER(inode);
3309 3310 3311 3312 3313 3314 3315 3316 3317 3318 3319

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

	p->arg.fh = NFS_FH(inode);
	p->arg.fl = &p->fl;
	p->arg.lock_seqid = nfs_alloc_seqid(&lsp->ls_seqid);
	if (p->arg.lock_seqid == NULL)
		goto out_free;
	p->arg.lock_stateid = &lsp->ls_stateid;
3320
	p->arg.lock_owner.clientid = server->nfs_client->cl_clientid;
3321
	p->arg.lock_owner.id = lsp->ls_id.id;
3322 3323 3324 3325 3326 3327 3328 3329 3330 3331 3332 3333 3334 3335 3336
	p->lsp = lsp;
	atomic_inc(&lsp->ls_count);
	p->ctx = get_nfs_open_context(ctx);
	memcpy(&p->fl, fl, sizeof(p->fl));
	return p;
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;
	struct nfs4_state_owner *sp = state->owner;
L
Linus Torvalds 已提交
3337
	struct rpc_message msg = {
3338 3339 3340 3341
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_LOCK],
		.rpc_argp = &data->arg,
		.rpc_resp = &data->res,
		.rpc_cred = sp->so_cred,
L
Linus Torvalds 已提交
3342
	};
3343

3344 3345 3346 3347 3348 3349 3350 3351 3352
	if (nfs_wait_on_sequence(data->arg.lock_seqid, task) != 0)
		return;
	dprintk("%s: begin!\n", __FUNCTION__);
	/* Do we need to do an open_to_lock_owner? */
	if (!(data->arg.lock_seqid->sequence->flags & NFS_SEQID_CONFIRMED)) {
		data->arg.open_seqid = nfs_alloc_seqid(&sp->so_seqid);
		if (data->arg.open_seqid == NULL) {
			data->rpc_status = -ENOMEM;
			task->tk_action = NULL;
3353
			goto out;
3354
		}
3355 3356 3357
		data->arg.open_stateid = &state->stateid;
		data->arg.new_lock_owner = 1;
	}
3358
	data->timestamp = jiffies;
3359
	rpc_call_setup(task, &msg, 0);
3360
out:
3361 3362 3363 3364 3365 3366 3367 3368 3369 3370 3371 3372 3373 3374 3375 3376 3377 3378 3379 3380 3381 3382 3383
	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;
3384
		renew_lease(NFS_SERVER(data->ctx->path.dentry->d_inode), data->timestamp);
3385 3386 3387 3388 3389 3390 3391 3392 3393 3394 3395 3396 3397 3398 3399 3400 3401 3402
	}
	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__);
	if (data->arg.open_seqid != NULL)
		nfs_free_seqid(data->arg.open_seqid);
	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))
3403
			rpc_put_task(task);
3404 3405 3406 3407 3408 3409 3410 3411 3412 3413 3414 3415 3416 3417 3418 3419 3420 3421 3422 3423 3424 3425
		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;
	int ret;

	dprintk("%s: begin!\n", __FUNCTION__);
3426
	data = nfs4_alloc_lockdata(fl, nfs_file_open_context(fl->fl_file),
3427 3428 3429 3430 3431 3432 3433 3434 3435
			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;
	task = rpc_run_task(NFS_CLIENT(state->inode), RPC_TASK_ASYNC,
			&nfs4_lock_ops, data);
3436
	if (IS_ERR(task))
3437 3438 3439 3440 3441 3442 3443 3444
		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;
3445
	rpc_put_task(task);
3446 3447
	dprintk("%s: done, ret = %d!\n", __FUNCTION__, ret);
	return ret;
L
Linus Torvalds 已提交
3448 3449 3450 3451
}

static int nfs4_lock_reclaim(struct nfs4_state *state, struct file_lock *request)
{
3452 3453 3454 3455 3456
	struct nfs_server *server = NFS_SERVER(state->inode);
	struct nfs4_exception exception = { };
	int err;

	do {
3457 3458 3459
		/* Cache the lock if possible... */
		if (test_bit(NFS_DELEGATED_STATE, &state->flags) != 0)
			return 0;
3460 3461 3462 3463 3464 3465
		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
3466 3467 3468 3469
}

static int nfs4_lock_expired(struct nfs4_state *state, struct file_lock *request)
{
3470 3471 3472 3473
	struct nfs_server *server = NFS_SERVER(state->inode);
	struct nfs4_exception exception = { };
	int err;

3474 3475 3476
	err = nfs4_set_lock_state(state, request);
	if (err != 0)
		return err;
3477
	do {
3478 3479
		if (test_bit(NFS_DELEGATED_STATE, &state->flags) != 0)
			return 0;
3480 3481 3482 3483 3484 3485
		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
3486 3487 3488 3489
}

static int _nfs4_proc_setlk(struct nfs4_state *state, int cmd, struct file_lock *request)
{
3490
	struct nfs_client *clp = state->owner->so_client;
3491
	unsigned char fl_flags = request->fl_flags;
L
Linus Torvalds 已提交
3492 3493
	int status;

3494 3495 3496 3497
	/* Is this a delegated open? */
	status = nfs4_set_lock_state(state, request);
	if (status != 0)
		goto out;
3498 3499 3500 3501 3502 3503 3504 3505 3506 3507 3508 3509 3510 3511 3512 3513 3514 3515
	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);
	}
3516 3517
	status = _nfs4_do_setlk(state, cmd, request, 0);
	if (status != 0)
3518
		goto out_unlock;
3519
	/* Note: we always want to sleep here! */
3520
	request->fl_flags = fl_flags | FL_SLEEP;
3521 3522
	if (do_vfs_lock(request->fl_file, request) < 0)
		printk(KERN_WARNING "%s: VFS is out of sync with lock manager!\n", __FUNCTION__);
3523
out_unlock:
L
Linus Torvalds 已提交
3524
	up_read(&clp->cl_sem);
3525 3526
out:
	request->fl_flags = fl_flags;
L
Linus Torvalds 已提交
3527 3528 3529 3530 3531 3532 3533 3534 3535 3536 3537 3538 3539 3540 3541 3542 3543 3544 3545 3546 3547 3548 3549 3550 3551
	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 */
3552
	ctx = nfs_file_open_context(filp);
L
Linus Torvalds 已提交
3553 3554 3555 3556 3557 3558 3559 3560 3561 3562 3563 3564 3565 3566 3567 3568 3569 3570 3571 3572 3573 3574 3575 3576 3577 3578
	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;
}

3579 3580 3581 3582 3583 3584 3585 3586 3587 3588 3589 3590 3591 3592 3593 3594 3595 3596
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;
}
3597

3598 3599
#define XATTR_NAME_NFSV4_ACL "system.nfs4_acl"

3600 3601 3602
int nfs4_setxattr(struct dentry *dentry, const char *key, const void *buf,
		size_t buflen, int flags)
{
3603 3604 3605 3606 3607 3608 3609 3610 3611 3612
	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);
3613 3614 3615 3616 3617 3618 3619 3620 3621
}

/* 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)
{
3622 3623 3624 3625 3626 3627
	struct inode *inode = dentry->d_inode;

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

	return nfs4_proc_get_acl(inode, buf, buflen);
3628 3629 3630 3631
}

ssize_t nfs4_listxattr(struct dentry *dentry, char *buf, size_t buflen)
{
3632
	size_t len = strlen(XATTR_NAME_NFSV4_ACL) + 1;
3633

3634 3635
	if (!nfs4_server_supports_acls(NFS_SERVER(dentry->d_inode)))
		return 0;
3636 3637 3638
	if (buf && buflen < len)
		return -ERANGE;
	if (buf)
3639 3640
		memcpy(buf, XATTR_NAME_NFSV4_ACL, len);
	return len;
3641 3642
}

3643
int nfs4_proc_fs_locations(struct inode *dir, const struct qstr *name,
3644
		struct nfs4_fs_locations *fs_locations, struct page *page)
3645 3646 3647
{
	struct nfs_server *server = NFS_SERVER(dir);
	u32 bitmask[2] = {
3648 3649
		[0] = FATTR4_WORD0_FSID | FATTR4_WORD0_FS_LOCATIONS,
		[1] = FATTR4_WORD1_MOUNTED_ON_FILEID,
3650 3651 3652
	};
	struct nfs4_fs_locations_arg args = {
		.dir_fh = NFS_FH(dir),
3653
		.name = name,
3654 3655 3656 3657 3658 3659
		.page = page,
		.bitmask = bitmask,
	};
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_FS_LOCATIONS],
		.rpc_argp = &args,
3660
		.rpc_resp = fs_locations,
3661 3662 3663 3664
	};
	int status;

	dprintk("%s: start\n", __FUNCTION__);
3665
	nfs_fattr_init(&fs_locations->fattr);
3666
	fs_locations->server = server;
3667
	fs_locations->nlocations = 0;
3668 3669 3670 3671 3672
	status = rpc_call_sync(server->client, &msg, 0);
	dprintk("%s: returned status = %d\n", __FUNCTION__, status);
	return status;
}

L
Linus Torvalds 已提交
3673 3674 3675 3676 3677 3678 3679 3680 3681 3682
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,
};

3683
static const struct inode_operations nfs4_file_inode_operations = {
3684 3685 3686 3687 3688 3689 3690 3691
	.permission	= nfs_permission,
	.getattr	= nfs_getattr,
	.setattr	= nfs_setattr,
	.getxattr	= nfs4_getxattr,
	.setxattr	= nfs4_setxattr,
	.listxattr	= nfs4_listxattr,
};

D
David Howells 已提交
3692
const struct nfs_rpc_ops nfs_v4_clientops = {
L
Linus Torvalds 已提交
3693 3694 3695
	.version	= 4,			/* protocol version */
	.dentry_ops	= &nfs4_dentry_operations,
	.dir_inode_ops	= &nfs4_dir_inode_operations,
3696
	.file_inode_ops	= &nfs4_file_inode_operations,
L
Linus Torvalds 已提交
3697 3698 3699
	.getroot	= nfs4_proc_get_root,
	.getattr	= nfs4_proc_getattr,
	.setattr	= nfs4_proc_setattr,
D
David Howells 已提交
3700
	.lookupfh	= nfs4_proc_lookupfh,
L
Linus Torvalds 已提交
3701 3702 3703 3704 3705 3706 3707 3708 3709 3710 3711 3712 3713 3714 3715 3716 3717
	.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,
3718
	.set_capabilities = nfs4_server_capabilities,
L
Linus Torvalds 已提交
3719 3720
	.decode_dirent	= nfs4_decode_dirent,
	.read_setup	= nfs4_proc_read_setup,
T
Trond Myklebust 已提交
3721
	.read_done	= nfs4_read_done,
L
Linus Torvalds 已提交
3722
	.write_setup	= nfs4_proc_write_setup,
3723
	.write_done	= nfs4_write_done,
L
Linus Torvalds 已提交
3724
	.commit_setup	= nfs4_proc_commit_setup,
3725
	.commit_done	= nfs4_commit_done,
3726 3727
	.file_open      = nfs_open,
	.file_release   = nfs_release,
L
Linus Torvalds 已提交
3728
	.lock		= nfs4_proc_lock,
3729
	.clear_acl_cache = nfs4_zap_acl_attr,
L
Linus Torvalds 已提交
3730 3731 3732 3733 3734 3735 3736
};

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