nfs4proc.c 190.7 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
/*
 *  fs/nfs/nfs4proc.c
 *
 *  Client-side procedure declarations for NFSv4.
 *
 *  Copyright (c) 2002 The Regents of the University of Michigan.
 *  All rights reserved.
 *
 *  Kendrick Smith <kmsmith@umich.edu>
 *  Andy Adamson   <andros@umich.edu>
 *
 *  Redistribution and use in source and binary forms, with or without
 *  modification, are permitted provided that the following conditions
 *  are met:
 *
 *  1. Redistributions of source code must retain the above copyright
 *     notice, this list of conditions and the following disclaimer.
 *  2. Redistributions in binary form must reproduce the above copyright
 *     notice, this list of conditions and the following disclaimer in the
 *     documentation and/or other materials provided with the distribution.
 *  3. Neither the name of the University nor the names of its
 *     contributors may be used to endorse or promote products derived
 *     from this software without specific prior written permission.
 *
 *  THIS SOFTWARE IS PROVIDED ``AS IS'' AND ANY EXPRESS OR IMPLIED
 *  WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF
 *  MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
 *  DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
 *  FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
 *  CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
 *  SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR
 *  BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
 *  LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
 *  NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
 *  SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 */

#include <linux/mm.h>
#include <linux/delay.h>
#include <linux/errno.h>
#include <linux/string.h>
42 43
#include <linux/ratelimit.h>
#include <linux/printk.h>
44
#include <linux/slab.h>
L
Linus Torvalds 已提交
45 46 47 48 49
#include <linux/sunrpc/clnt.h>
#include <linux/nfs.h>
#include <linux/nfs4.h>
#include <linux/nfs_fs.h>
#include <linux/nfs_page.h>
50
#include <linux/nfs_mount.h>
L
Linus Torvalds 已提交
51
#include <linux/namei.h>
52
#include <linux/mount.h>
B
Benny Halevy 已提交
53
#include <linux/module.h>
54
#include <linux/nfs_idmap.h>
55
#include <linux/xattr.h>
56
#include <linux/utsname.h>
57
#include <linux/freezer.h>
L
Linus Torvalds 已提交
58

59
#include "nfs4_fs.h"
L
Linus Torvalds 已提交
60
#include "delegation.h"
61
#include "internal.h"
62
#include "iostat.h"
A
Andy Adamson 已提交
63
#include "callback.h"
64
#include "pnfs.h"
65
#include "netns.h"
66
#include "nfs4session.h"
D
David Howells 已提交
67
#include "fscache.h"
L
Linus Torvalds 已提交
68 69 70

#define NFSDBG_FACILITY		NFSDBG_PROC

71
#define NFS4_POLL_RETRY_MIN	(HZ/10)
L
Linus Torvalds 已提交
72 73
#define NFS4_POLL_RETRY_MAX	(15*HZ)

74
struct nfs4_opendata;
75
static int _nfs4_proc_open(struct nfs4_opendata *data);
76
static int _nfs4_recover_proc_open(struct nfs4_opendata *data);
L
Linus Torvalds 已提交
77
static int nfs4_do_fsinfo(struct nfs_server *, struct nfs_fh *, struct nfs_fsinfo *);
78
static int nfs4_async_handle_error(struct rpc_task *, const struct nfs_server *, struct nfs4_state *);
79
static void nfs_fixup_referral_attributes(struct nfs_fattr *fattr);
80
static int nfs4_proc_getattr(struct nfs_server *, struct nfs_fh *, struct nfs_fattr *);
81
static int _nfs4_proc_getattr(struct nfs_server *server, struct nfs_fh *fhandle, struct nfs_fattr *fattr);
82 83 84
static int nfs4_do_setattr(struct inode *inode, struct rpc_cred *cred,
			    struct nfs_fattr *fattr, struct iattr *sattr,
			    struct nfs4_state *state);
85
#ifdef CONFIG_NFS_V4_1
86 87 88 89
static int nfs41_test_stateid(struct nfs_server *, nfs4_stateid *,
		struct rpc_cred *);
static int nfs41_free_stateid(struct nfs_server *, nfs4_stateid *,
		struct rpc_cred *);
90
#endif
L
Linus Torvalds 已提交
91
/* Prevent leaks of NFSv4 errors into userland */
92
static int nfs4_map_errors(int err)
L
Linus Torvalds 已提交
93
{
94 95 96 97
	if (err >= -1000)
		return err;
	switch (err) {
	case -NFS4ERR_RESOURCE:
98 99
	case -NFS4ERR_LAYOUTTRYLATER:
	case -NFS4ERR_RECALLCONFLICT:
100
		return -EREMOTEIO;
101 102
	case -NFS4ERR_WRONGSEC:
		return -EPERM;
103 104 105
	case -NFS4ERR_BADOWNER:
	case -NFS4ERR_BADNAME:
		return -EINVAL;
106 107
	case -NFS4ERR_SHARE_DENIED:
		return -EACCES;
108 109
	case -NFS4ERR_MINOR_VERS_MISMATCH:
		return -EPROTONOSUPPORT;
110 111
	case -NFS4ERR_ACCESS:
		return -EACCES;
112 113
	case -NFS4ERR_FILE_OPEN:
		return -EBUSY;
114
	default:
L
Linus Torvalds 已提交
115
		dprintk("%s could not handle NFSv4 error %d\n",
116
				__func__, -err);
117
		break;
L
Linus Torvalds 已提交
118
	}
119
	return -EIO;
L
Linus Torvalds 已提交
120 121 122 123 124
}

/*
 * This is our standard bitmap for GETATTR requests.
 */
125
const u32 nfs4_fattr_bitmap[3] = {
L
Linus Torvalds 已提交
126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141
	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
};

142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159
static const u32 nfs4_pnfs_open_bitmap[3] = {
	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,
	FATTR4_WORD2_MDSTHRESHOLD
};

160 161 162 163 164 165
static const u32 nfs4_open_noattr_bitmap[3] = {
	FATTR4_WORD0_TYPE
	| FATTR4_WORD0_CHANGE
	| FATTR4_WORD0_FILEID,
};

L
Linus Torvalds 已提交
166 167 168 169 170 171 172 173 174
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
};

175
const u32 nfs4_pathconf_bitmap[2] = {
L
Linus Torvalds 已提交
176 177 178 179 180
	FATTR4_WORD0_MAXLINK
	| FATTR4_WORD0_MAXNAME,
	0
};

181
const u32 nfs4_fsinfo_bitmap[3] = { FATTR4_WORD0_MAXFILESIZE
L
Linus Torvalds 已提交
182 183 184
			| FATTR4_WORD0_MAXREAD
			| FATTR4_WORD0_MAXWRITE
			| FATTR4_WORD0_LEASE_TIME,
185
			FATTR4_WORD1_TIME_DELTA
186 187
			| FATTR4_WORD1_FS_LAYOUT_TYPES,
			FATTR4_WORD2_LAYOUT_BLKSIZE
L
Linus Torvalds 已提交
188 189
};

190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208
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
209
static void nfs4_setup_readdir(u64 cookie, __be32 *verifier, struct dentry *dentry,
L
Linus Torvalds 已提交
210 211
		struct nfs4_readdir_arg *readdir)
{
212
	__be32 *start, *p;
L
Linus Torvalds 已提交
213 214

	if (cookie > 2) {
215
		readdir->cookie = cookie;
L
Linus Torvalds 已提交
216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231
		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.
	 */
232
	start = p = kmap_atomic(*readdir->pages);
L
Linus Torvalds 已提交
233 234 235 236 237 238 239 240 241 242 243
	
	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
244
		p = xdr_encode_hyper(p, NFS_FILEID(dentry->d_inode));
L
Linus Torvalds 已提交
245 246 247 248 249 250 251 252 253 254 255
	}
	
	*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
256
	p = xdr_encode_hyper(p, NFS_FILEID(dentry->d_parent->d_inode));
L
Linus Torvalds 已提交
257 258 259

	readdir->pgbase = (char *)p - (char *)start;
	readdir->count -= readdir->pgbase;
260
	kunmap_atomic(start);
L
Linus Torvalds 已提交
261 262
}

263 264 265 266 267 268 269 270 271 272
static int nfs4_delay(struct rpc_clnt *clnt, long *timeout)
{
	int res = 0;

	might_sleep();

	if (*timeout <= 0)
		*timeout = NFS4_POLL_RETRY_MIN;
	if (*timeout > NFS4_POLL_RETRY_MAX)
		*timeout = NFS4_POLL_RETRY_MAX;
273
	freezable_schedule_timeout_killable(*timeout);
274 275 276 277 278 279 280 281 282
	if (fatal_signal_pending(current))
		res = -ERESTARTSYS;
	*timeout <<= 1;
	return res;
}

/* This is the error handling routine for processes that are allowed
 * to sleep.
 */
283
static int nfs4_handle_exception(struct nfs_server *server, int errorcode, struct nfs4_exception *exception)
284 285
{
	struct nfs_client *clp = server->nfs_client;
286
	struct nfs4_state *state = exception->state;
287
	struct inode *inode = exception->inode;
288 289 290 291 292 293
	int ret = errorcode;

	exception->retry = 0;
	switch(errorcode) {
		case 0:
			return 0;
294
		case -NFS4ERR_OPENMODE:
295
			if (inode && nfs4_have_delegation(inode, FMODE_READ)) {
296
				nfs4_inode_return_delegation(inode);
297 298 299 300 301
				exception->retry = 1;
				return 0;
			}
			if (state == NULL)
				break;
302 303 304
			ret = nfs4_schedule_stateid_recovery(server, state);
			if (ret < 0)
				break;
305
			goto wait_on_recovery;
306
		case -NFS4ERR_DELEG_REVOKED:
307 308
		case -NFS4ERR_ADMIN_REVOKED:
		case -NFS4ERR_BAD_STATEID:
309 310 311 312 313
			if (inode != NULL && nfs4_have_delegation(inode, FMODE_READ)) {
				nfs_remove_bad_delegation(inode);
				exception->retry = 1;
				break;
			}
314 315
			if (state == NULL)
				break;
316 317 318
			ret = nfs4_schedule_stateid_recovery(server, state);
			if (ret < 0)
				break;
319
			goto wait_on_recovery;
320
		case -NFS4ERR_EXPIRED:
321 322 323 324 325
			if (state != NULL) {
				ret = nfs4_schedule_stateid_recovery(server, state);
				if (ret < 0)
					break;
			}
326
		case -NFS4ERR_STALE_STATEID:
327
		case -NFS4ERR_STALE_CLIENTID:
328 329
			nfs4_schedule_lease_recovery(clp);
			goto wait_on_recovery;
330
#if defined(CONFIG_NFS_V4_1)
331 332 333 334 335 336 337 338 339
		case -NFS4ERR_BADSESSION:
		case -NFS4ERR_BADSLOT:
		case -NFS4ERR_BAD_HIGH_SLOT:
		case -NFS4ERR_CONN_NOT_BOUND_TO_SESSION:
		case -NFS4ERR_DEADSESSION:
		case -NFS4ERR_SEQ_FALSE_RETRY:
		case -NFS4ERR_SEQ_MISORDERED:
			dprintk("%s ERROR: %d Reset session\n", __func__,
				errorcode);
340
			nfs4_schedule_session_recovery(clp->cl_session, errorcode);
341
			goto wait_on_recovery;
342
#endif /* defined(CONFIG_NFS_V4_1) */
343
		case -NFS4ERR_FILE_OPEN:
344 345 346 347 348 349 350
			if (exception->timeout > HZ) {
				/* We have retried a decent amount, time to
				 * fail
				 */
				ret = -EBUSY;
				break;
			}
351 352 353 354 355
		case -NFS4ERR_GRACE:
		case -NFS4ERR_DELAY:
			ret = nfs4_delay(server->client, &exception->timeout);
			if (ret != 0)
				break;
356
		case -NFS4ERR_RETRY_UNCACHED_REP:
357 358
		case -NFS4ERR_OLD_STATEID:
			exception->retry = 1;
359 360 361 362 363 364 365 366 367 368 369 370 371
			break;
		case -NFS4ERR_BADOWNER:
			/* The following works around a Linux server bug! */
		case -NFS4ERR_BADNAME:
			if (server->caps & NFS_CAP_UIDGID_NOMAP) {
				server->caps &= ~NFS_CAP_UIDGID_NOMAP;
				exception->retry = 1;
				printk(KERN_WARNING "NFS: v4 server %s "
						"does not accept raw "
						"uid/gids. "
						"Reenabling the idmapper.\n",
						server->nfs_client->cl_hostname);
			}
372 373 374
	}
	/* We failed to handle the error */
	return nfs4_map_errors(ret);
375
wait_on_recovery:
376 377 378 379
	ret = nfs4_wait_clnt_recover(clp);
	if (ret == 0)
		exception->retry = 1;
	return ret;
380 381 382
}


383
static void do_renew_lease(struct nfs_client *clp, unsigned long timestamp)
L
Linus Torvalds 已提交
384 385 386 387 388 389 390
{
	spin_lock(&clp->cl_lock);
	if (time_before(clp->cl_last_renewal,timestamp))
		clp->cl_last_renewal = timestamp;
	spin_unlock(&clp->cl_lock);
}

391 392 393 394 395
static void renew_lease(const struct nfs_server *server, unsigned long timestamp)
{
	do_renew_lease(server->nfs_client, timestamp);
}

A
Andy Adamson 已提交
396 397
#if defined(CONFIG_NFS_V4_1)

398
static void nfs41_sequence_free_slot(struct nfs4_sequence_res *res)
A
Andy Adamson 已提交
399
{
400
	struct nfs4_session *session;
A
Andy Adamson 已提交
401
	struct nfs4_slot_table *tbl;
402
	bool send_new_highest_used_slotid = false;
A
Andy Adamson 已提交
403

404
	if (!res->sr_slot) {
A
Andy Adamson 已提交
405 406
		/* just wake up the next guy waiting since
		 * we may have not consumed a slot after all */
A
Andy Adamson 已提交
407
		dprintk("%s: No slot\n", __func__);
408
		return;
A
Andy Adamson 已提交
409
	}
410 411
	tbl = res->sr_slot->table;
	session = tbl->session;
412

413
	spin_lock(&tbl->slot_tbl_lock);
414 415 416 417 418 419
	/* Be nice to the server: try to ensure that the last transmitted
	 * value for highest_user_slotid <= target_highest_slotid
	 */
	if (tbl->highest_used_slotid > tbl->target_highest_slotid)
		send_new_highest_used_slotid = true;

420 421 422 423
	if (nfs41_wake_and_assign_slot(tbl, res->sr_slot)) {
		send_new_highest_used_slotid = false;
		goto out_unlock;
	}
T
Trond Myklebust 已提交
424
	nfs4_free_slot(tbl, res->sr_slot);
425 426 427

	if (tbl->highest_used_slotid != NFS4_NO_SLOT)
		send_new_highest_used_slotid = false;
428
out_unlock:
429
	spin_unlock(&tbl->slot_tbl_lock);
430
	res->sr_slot = NULL;
431 432
	if (send_new_highest_used_slotid)
		nfs41_server_notify_highest_slotid_update(session->clp);
A
Andy Adamson 已提交
433 434
}

435
static int nfs41_sequence_done(struct rpc_task *task, struct nfs4_sequence_res *res)
A
Andy Adamson 已提交
436
{
437
	struct nfs4_session *session;
438
	struct nfs4_slot *slot;
439
	struct nfs_client *clp;
440
	bool interrupted = false;
441
	int ret = 1;
A
Andy Adamson 已提交
442

443 444
	/* don't increment the sequence number if the task wasn't sent */
	if (!RPC_WAS_SENT(task))
A
Andy Adamson 已提交
445 446
		goto out;

447
	slot = res->sr_slot;
448
	session = slot->table->session;
449

450 451 452 453 454
	if (slot->interrupted) {
		slot->interrupted = 0;
		interrupted = true;
	}

A
Andy Adamson 已提交
455
	/* Check the SEQUENCE operation status */
456 457
	switch (res->sr_status) {
	case 0:
A
Andy Adamson 已提交
458
		/* Update the slot's sequence and clientid lease timer */
459
		++slot->seq_nr;
460
		clp = session->clp;
461
		do_renew_lease(clp, res->sr_timestamp);
462
		/* Check sequence flags */
463 464
		if (res->sr_status_flags != 0)
			nfs4_schedule_lease_recovery(clp);
465
		nfs41_update_target_slotid(slot->table, slot, res);
466
		break;
467 468 469 470 471 472 473 474 475
	case 1:
		/*
		 * sr_status remains 1 if an RPC level error occurred.
		 * The server may or may not have processed the sequence
		 * operation..
		 * Mark the slot as having hosted an interrupted RPC call.
		 */
		slot->interrupted = 1;
		goto out;
476 477 478 479 480
	case -NFS4ERR_DELAY:
		/* The server detected a resend of the RPC call and
		 * returned NFS4ERR_DELAY as per Section 2.10.6.2
		 * of RFC5661.
		 */
481
		dprintk("%s: slot=%u seq=%u: Operation in progress\n",
482
			__func__,
483
			slot->slot_nr,
484
			slot->seq_nr);
485
		goto out_retry;
486 487 488 489 490
	case -NFS4ERR_BADSLOT:
		/*
		 * The slot id we used was probably retired. Try again
		 * using a different slot id.
		 */
491 492
		goto retry_nowait;
	case -NFS4ERR_SEQ_MISORDERED:
493 494 495 496 497 498 499 500
		/*
		 * Was the last operation on this sequence interrupted?
		 * If so, retry after bumping the sequence number.
		 */
		if (interrupted) {
			++slot->seq_nr;
			goto retry_nowait;
		}
501 502 503 504
		/*
		 * Could this slot have been previously retired?
		 * If so, then the server may be expecting seq_nr = 1!
		 */
505 506 507 508 509
		if (slot->seq_nr != 1) {
			slot->seq_nr = 1;
			goto retry_nowait;
		}
		break;
510 511 512
	case -NFS4ERR_SEQ_FALSE_RETRY:
		++slot->seq_nr;
		goto retry_nowait;
513 514
	default:
		/* Just update the slot sequence no. */
515
		++slot->seq_nr;
A
Andy Adamson 已提交
516 517 518 519
	}
out:
	/* The session may be reset by one of the error handlers. */
	dprintk("%s: Error %d free the slot \n", __func__, res->sr_status);
520
	nfs41_sequence_free_slot(res);
521
	return ret;
522 523 524 525 526 527
retry_nowait:
	if (rpc_restart_call_prepare(task)) {
		task->tk_status = 0;
		ret = 0;
	}
	goto out;
528
out_retry:
529
	if (!rpc_restart_call(task))
530 531 532
		goto out;
	rpc_delay(task, NFS4_POLL_RETRY_MAX);
	return 0;
A
Andy Adamson 已提交
533 534
}

535 536
static int nfs4_sequence_done(struct rpc_task *task,
			       struct nfs4_sequence_res *res)
537
{
538
	if (res->sr_slot == NULL)
539 540
		return 1;
	return nfs41_sequence_done(task, res);
541 542
}

543 544 545
static void nfs41_init_sequence(struct nfs4_sequence_args *args,
		struct nfs4_sequence_res *res, int cache_reply)
{
546
	args->sa_slot = NULL;
547
	args->sa_cache_this = 0;
548
	args->sa_privileged = 0;
549 550 551 552 553
	if (cache_reply)
		args->sa_cache_this = 1;
	res->sr_slot = NULL;
}

554 555 556 557 558
static void nfs4_set_sequence_privileged(struct nfs4_sequence_args *args)
{
	args->sa_privileged = 1;
}

A
Andy Adamson 已提交
559
int nfs41_setup_sequence(struct nfs4_session *session,
A
Andy Adamson 已提交
560 561 562 563
				struct nfs4_sequence_args *args,
				struct nfs4_sequence_res *res,
				struct rpc_task *task)
{
A
Andy Adamson 已提交
564 565 566 567
	struct nfs4_slot *slot;
	struct nfs4_slot_table *tbl;

	dprintk("--> %s\n", __func__);
A
Andy Adamson 已提交
568
	/* slot already allocated? */
569
	if (res->sr_slot != NULL)
570
		goto out_success;
A
Andy Adamson 已提交
571

A
Andy Adamson 已提交
572 573
	tbl = &session->fc_slot_table;

574 575
	task->tk_timeout = 0;

A
Andy Adamson 已提交
576
	spin_lock(&tbl->slot_tbl_lock);
577
	if (test_bit(NFS4_SLOT_TBL_DRAINING, &tbl->slot_tbl_state) &&
578
	    !args->sa_privileged) {
579 580
		/* The state manager will wait until the slot table is empty */
		dprintk("%s session is draining\n", __func__);
581
		goto out_sleep;
582 583
	}

584
	slot = nfs4_alloc_slot(tbl);
585 586 587 588
	if (IS_ERR(slot)) {
		/* If out of memory, try again in 1/4 second */
		if (slot == ERR_PTR(-ENOMEM))
			task->tk_timeout = HZ >> 2;
A
Andy Adamson 已提交
589
		dprintk("<-- %s: no free slots\n", __func__);
590
		goto out_sleep;
A
Andy Adamson 已提交
591 592 593
	}
	spin_unlock(&tbl->slot_tbl_lock);

594
	args->sa_slot = slot;
A
Andy Adamson 已提交
595

596 597
	dprintk("<-- %s slotid=%d seqid=%d\n", __func__,
			slot->slot_nr, slot->seq_nr);
A
Andy Adamson 已提交
598

599
	res->sr_slot = slot;
600
	res->sr_timestamp = jiffies;
601
	res->sr_status_flags = 0;
A
Andy Adamson 已提交
602 603 604 605 606
	/*
	 * sr_status is only set in decode_sequence, and so will remain
	 * set to 1 if an rpc level failure occurs.
	 */
	res->sr_status = 1;
607 608
out_success:
	rpc_call_start(task);
A
Andy Adamson 已提交
609
	return 0;
610
out_sleep:
611 612
	/* Privileged tasks are queued with top priority */
	if (args->sa_privileged)
613 614 615 616
		rpc_sleep_on_priority(&tbl->slot_tbl_waitq, task,
				NULL, RPC_PRIORITY_PRIVILEGED);
	else
		rpc_sleep_on(&tbl->slot_tbl_waitq, task, NULL);
617 618
	spin_unlock(&tbl->slot_tbl_lock);
	return -EAGAIN;
A
Andy Adamson 已提交
619
}
A
Andy Adamson 已提交
620
EXPORT_SYMBOL_GPL(nfs41_setup_sequence);
A
Andy Adamson 已提交
621

622
int nfs4_setup_sequence(const struct nfs_server *server,
A
Andy Adamson 已提交
623 624 625 626
			struct nfs4_sequence_args *args,
			struct nfs4_sequence_res *res,
			struct rpc_task *task)
{
627
	struct nfs4_session *session = nfs4_get_session(server);
A
Andy Adamson 已提交
628 629
	int ret = 0;

630 631
	if (session == NULL) {
		rpc_call_start(task);
632
		goto out;
633
	}
634

635
	dprintk("--> %s clp %p session %p sr_slot %d\n",
636
		__func__, session->clp, session, res->sr_slot ?
637
			res->sr_slot->slot_nr : -1);
A
Andy Adamson 已提交
638

639
	ret = nfs41_setup_sequence(session, args, res, task);
A
Andy Adamson 已提交
640 641 642 643 644 645
out:
	dprintk("<-- %s status=%d\n", __func__, ret);
	return ret;
}

struct nfs41_call_sync_data {
646
	const struct nfs_server *seq_server;
A
Andy Adamson 已提交
647 648 649 650 651 652 653
	struct nfs4_sequence_args *seq_args;
	struct nfs4_sequence_res *seq_res;
};

static void nfs41_call_sync_prepare(struct rpc_task *task, void *calldata)
{
	struct nfs41_call_sync_data *data = calldata;
654
	struct nfs4_session *session = nfs4_get_session(data->seq_server);
A
Andy Adamson 已提交
655

656 657
	dprintk("--> %s data->seq_server %p\n", __func__, data->seq_server);

658
	nfs41_setup_sequence(session, data->seq_args, data->seq_res, task);
A
Andy Adamson 已提交
659 660
}

A
Andy Adamson 已提交
661 662 663 664
static void nfs41_call_sync_done(struct rpc_task *task, void *calldata)
{
	struct nfs41_call_sync_data *data = calldata;

665
	nfs41_sequence_done(task, data->seq_res);
A
Andy Adamson 已提交
666 667
}

668
static const struct rpc_call_ops nfs41_call_sync_ops = {
A
Andy Adamson 已提交
669
	.rpc_call_prepare = nfs41_call_sync_prepare,
A
Andy Adamson 已提交
670
	.rpc_call_done = nfs41_call_sync_done,
A
Andy Adamson 已提交
671 672
};

673 674
static int nfs4_call_sync_sequence(struct rpc_clnt *clnt,
				   struct nfs_server *server,
A
Andy Adamson 已提交
675 676
				   struct rpc_message *msg,
				   struct nfs4_sequence_args *args,
677
				   struct nfs4_sequence_res *res)
A
Andy Adamson 已提交
678 679 680 681
{
	int ret;
	struct rpc_task *task;
	struct nfs41_call_sync_data data = {
682
		.seq_server = server,
A
Andy Adamson 已提交
683 684 685 686
		.seq_args = args,
		.seq_res = res,
	};
	struct rpc_task_setup task_setup = {
687
		.rpc_client = clnt,
A
Andy Adamson 已提交
688 689 690 691 692 693 694 695 696 697 698 699 700 701 702
		.rpc_message = msg,
		.callback_ops = &nfs41_call_sync_ops,
		.callback_data = &data
	};

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

703
#else
704
static
705 706 707 708 709
void nfs41_init_sequence(struct nfs4_sequence_args *args,
		struct nfs4_sequence_res *res, int cache_reply)
{
}

710 711 712 713 714
static void nfs4_set_sequence_privileged(struct nfs4_sequence_args *args)
{
}


715 716
static int nfs4_sequence_done(struct rpc_task *task,
			       struct nfs4_sequence_res *res)
717
{
718
	return 1;
719
}
A
Andy Adamson 已提交
720 721
#endif /* CONFIG_NFS_V4_1 */

722
static
723 724
int _nfs4_call_sync(struct rpc_clnt *clnt,
		    struct nfs_server *server,
A
Andy Adamson 已提交
725 726
		    struct rpc_message *msg,
		    struct nfs4_sequence_args *args,
727
		    struct nfs4_sequence_res *res)
A
Andy Adamson 已提交
728
{
729
	return rpc_call_sync(clnt, msg, 0);
A
Andy Adamson 已提交
730 731
}

732
static
733 734
int nfs4_call_sync(struct rpc_clnt *clnt,
		   struct nfs_server *server,
735 736 737 738 739
		   struct rpc_message *msg,
		   struct nfs4_sequence_args *args,
		   struct nfs4_sequence_res *res,
		   int cache_reply)
{
740
	nfs41_init_sequence(args, res, cache_reply);
741
	return server->nfs_client->cl_mvops->call_sync(clnt, server, msg,
742
						args, res);
743
}
A
Andy Adamson 已提交
744

745
static void update_changeattr(struct inode *dir, struct nfs4_change_info *cinfo)
L
Linus Torvalds 已提交
746
{
747
	struct nfs_inode *nfsi = NFS_I(dir);
L
Linus Torvalds 已提交
748

749
	spin_lock(&dir->i_lock);
750
	nfsi->cache_validity |= NFS_INO_INVALID_ATTR|NFS_INO_INVALID_DATA;
751
	if (!cinfo->atomic || cinfo->before != dir->i_version)
752
		nfs_force_lookup_revalidate(dir);
753
	dir->i_version = cinfo->after;
D
David Howells 已提交
754
	nfs_fscache_invalidate(dir);
755
	spin_unlock(&dir->i_lock);
L
Linus Torvalds 已提交
756 757
}

758
struct nfs4_opendata {
759
	struct kref kref;
760 761
	struct nfs_openargs o_arg;
	struct nfs_openres o_res;
762 763
	struct nfs_open_confirmargs c_arg;
	struct nfs_open_confirmres c_res;
764 765
	struct nfs4_string owner_name;
	struct nfs4_string group_name;
766 767
	struct nfs_fattr f_attr;
	struct dentry *dir;
768
	struct dentry *dentry;
769
	struct nfs4_state_owner *owner;
770
	struct nfs4_state *state;
771
	struct iattr attrs;
772
	unsigned long timestamp;
773
	unsigned int rpc_done : 1;
774
	unsigned int is_recover : 1;
775 776
	int rpc_status;
	int cancelled;
777 778
};

779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807
static bool nfs4_clear_cap_atomic_open_v1(struct nfs_server *server,
		int err, struct nfs4_exception *exception)
{
	if (err != -EINVAL)
		return false;
	if (!(server->caps & NFS_CAP_ATOMIC_OPEN_V1))
		return false;
	server->caps &= ~NFS_CAP_ATOMIC_OPEN_V1;
	exception->retry = 1;
	return true;
}

static enum open_claim_type4
nfs4_map_atomic_open_claim(struct nfs_server *server,
		enum open_claim_type4 claim)
{
	if (server->caps & NFS_CAP_ATOMIC_OPEN_V1)
		return claim;
	switch (claim) {
	default:
		return claim;
	case NFS4_OPEN_CLAIM_FH:
		return NFS4_OPEN_CLAIM_NULL;
	case NFS4_OPEN_CLAIM_DELEG_CUR_FH:
		return NFS4_OPEN_CLAIM_DELEGATE_CUR;
	case NFS4_OPEN_CLAIM_DELEG_PREV_FH:
		return NFS4_OPEN_CLAIM_DELEGATE_PREV;
	}
}
808 809 810 811

static void nfs4_init_opendata_res(struct nfs4_opendata *p)
{
	p->o_res.f_attr = &p->f_attr;
812 813
	p->o_res.seqid = p->o_arg.seqid;
	p->c_res.seqid = p->c_arg.seqid;
814
	p->o_res.server = p->o_arg.server;
815
	p->o_res.access_request = p->o_arg.access;
816
	nfs_fattr_init(&p->f_attr);
817
	nfs_fattr_init_names(&p->f_attr, &p->owner_name, &p->group_name);
818 819
}

820
static struct nfs4_opendata *nfs4_opendata_alloc(struct dentry *dentry,
821
		struct nfs4_state_owner *sp, fmode_t fmode, int flags,
822
		const struct iattr *attrs,
823
		enum open_claim_type4 claim,
824
		gfp_t gfp_mask)
825
{
826
	struct dentry *parent = dget_parent(dentry);
827 828 829 830
	struct inode *dir = parent->d_inode;
	struct nfs_server *server = NFS_SERVER(dir);
	struct nfs4_opendata *p;

831
	p = kzalloc(sizeof(*p), gfp_mask);
832 833
	if (p == NULL)
		goto err;
834
	p->o_arg.seqid = nfs_alloc_seqid(&sp->so_seqid, gfp_mask);
835 836
	if (p->o_arg.seqid == NULL)
		goto err_free;
837 838
	nfs_sb_active(dentry->d_sb);
	p->dentry = dget(dentry);
839 840 841
	p->dir = parent;
	p->owner = sp;
	atomic_inc(&sp->so_count);
842 843
	p->o_arg.open_flags = flags;
	p->o_arg.fmode = fmode & (FMODE_READ|FMODE_WRITE);
844 845 846 847 848 849 850 851
	/* don't put an ACCESS op in OPEN compound if O_EXCL, because ACCESS
	 * will return permission denied for all bits until close */
	if (!(flags & O_EXCL)) {
		/* ask server to check for all possible rights as results
		 * are cached */
		p->o_arg.access = NFS4_ACCESS_READ | NFS4_ACCESS_MODIFY |
				  NFS4_ACCESS_EXTEND | NFS4_ACCESS_EXECUTE;
	}
852
	p->o_arg.clientid = server->nfs_client->cl_clientid;
853 854
	p->o_arg.id.create_time = ktime_to_ns(sp->so_seqid.create_time);
	p->o_arg.id.uniquifier = sp->so_seqid.owner_id;
855
	p->o_arg.name = &dentry->d_name;
856 857
	p->o_arg.server = server;
	p->o_arg.bitmask = server->attr_bitmask;
858
	p->o_arg.open_bitmap = &nfs4_fattr_bitmap[0];
859 860
	p->o_arg.claim = nfs4_map_atomic_open_claim(server, claim);
	switch (p->o_arg.claim) {
861 862 863 864 865 866 867 868 869 870 871
	case NFS4_OPEN_CLAIM_NULL:
	case NFS4_OPEN_CLAIM_DELEGATE_CUR:
	case NFS4_OPEN_CLAIM_DELEGATE_PREV:
		p->o_arg.fh = NFS_FH(dir);
		break;
	case NFS4_OPEN_CLAIM_PREVIOUS:
	case NFS4_OPEN_CLAIM_FH:
	case NFS4_OPEN_CLAIM_DELEG_CUR_FH:
	case NFS4_OPEN_CLAIM_DELEG_PREV_FH:
		p->o_arg.fh = NFS_FH(dentry->d_inode);
	}
872
	if (attrs != NULL && attrs->ia_valid != 0) {
873
		__be32 verf[2];
874

875 876
		p->o_arg.u.attrs = &p->attrs;
		memcpy(&p->attrs, attrs, sizeof(p->attrs));
877 878 879 880 881

		verf[0] = jiffies;
		verf[1] = current->pid;
		memcpy(p->o_arg.u.verifier.data, verf,
				sizeof(p->o_arg.u.verifier.data));
882
	}
883 884 885
	p->c_arg.fh = &p->o_res.fh;
	p->c_arg.stateid = &p->o_res.stateid;
	p->c_arg.seqid = p->o_arg.seqid;
886
	nfs4_init_opendata_res(p);
887
	kref_init(&p->kref);
888 889 890 891 892 893 894 895
	return p;
err_free:
	kfree(p);
err:
	dput(parent);
	return NULL;
}

896
static void nfs4_opendata_free(struct kref *kref)
897
{
898 899
	struct nfs4_opendata *p = container_of(kref,
			struct nfs4_opendata, kref);
900
	struct super_block *sb = p->dentry->d_sb;
901 902

	nfs_free_seqid(p->o_arg.seqid);
903 904
	if (p->state != NULL)
		nfs4_put_open_state(p->state);
905 906
	nfs4_put_state_owner(p->owner);
	dput(p->dir);
907 908
	dput(p->dentry);
	nfs_sb_deactive(sb);
909
	nfs_fattr_free_names(&p->f_attr);
910 911 912 913 914 915 916
	kfree(p);
}

static void nfs4_opendata_put(struct nfs4_opendata *p)
{
	if (p != NULL)
		kref_put(&p->kref, nfs4_opendata_free);
917 918
}

919 920 921 922 923 924 925 926
static int nfs4_wait_for_completion_rpc_task(struct rpc_task *task)
{
	int ret;

	ret = rpc_wait_for_completion_task(task);
	return ret;
}

927
static int can_open_cached(struct nfs4_state *state, fmode_t mode, int open_mode)
928 929
{
	int ret = 0;
930

931
	if (open_mode & (O_EXCL|O_TRUNC))
932 933
		goto out;
	switch (mode & (FMODE_READ|FMODE_WRITE)) {
934
		case FMODE_READ:
935 936
			ret |= test_bit(NFS_O_RDONLY_STATE, &state->flags) != 0
				&& state->n_rdonly != 0;
937 938
			break;
		case FMODE_WRITE:
939 940
			ret |= test_bit(NFS_O_WRONLY_STATE, &state->flags) != 0
				&& state->n_wronly != 0;
941 942
			break;
		case FMODE_READ|FMODE_WRITE:
943 944
			ret |= test_bit(NFS_O_RDWR_STATE, &state->flags) != 0
				&& state->n_rdwr != 0;
945
	}
946
out:
947 948 949
	return ret;
}

950
static int can_open_delegated(struct nfs_delegation *delegation, fmode_t fmode)
951
{
952 953
	if (delegation == NULL)
		return 0;
954
	if ((delegation->type & fmode) != fmode)
955
		return 0;
956
	if (test_bit(NFS_DELEGATION_NEED_RECLAIM, &delegation->flags))
957
		return 0;
958 959
	if (test_bit(NFS_DELEGATION_RETURNING, &delegation->flags))
		return 0;
960
	nfs_mark_delegation_referenced(delegation);
961 962 963
	return 1;
}

964
static void update_open_stateflags(struct nfs4_state *state, fmode_t fmode)
965
{
966
	switch (fmode) {
967 968 969 970 971 972 973 974 975
		case FMODE_WRITE:
			state->n_wronly++;
			break;
		case FMODE_READ:
			state->n_rdonly++;
			break;
		case FMODE_READ|FMODE_WRITE:
			state->n_rdwr++;
	}
976
	nfs4_state_set_mode_locked(state, state->state | fmode);
977 978
}

979
static void nfs_set_open_stateid_locked(struct nfs4_state *state, nfs4_stateid *stateid, fmode_t fmode)
980 981
{
	if (test_bit(NFS_DELEGATED_STATE, &state->flags) == 0)
982 983
		nfs4_stateid_copy(&state->stateid, stateid);
	nfs4_stateid_copy(&state->open_stateid, stateid);
984
	set_bit(NFS_OPEN_STATE, &state->flags);
985
	switch (fmode) {
986 987 988 989 990 991 992 993 994
		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);
	}
995 996
}

997
static void nfs_set_open_stateid(struct nfs4_state *state, nfs4_stateid *stateid, fmode_t fmode)
998
{
999
	write_seqlock(&state->seqlock);
1000
	nfs_set_open_stateid_locked(state, stateid, fmode);
1001
	write_sequnlock(&state->seqlock);
1002 1003
}

1004
static void __update_open_stateid(struct nfs4_state *state, nfs4_stateid *open_stateid, const nfs4_stateid *deleg_stateid, fmode_t fmode)
L
Linus Torvalds 已提交
1005
{
1006 1007 1008 1009 1010
	/*
	 * Protect the call to nfs4_state_set_mode_locked and
	 * serialise the stateid update
	 */
	write_seqlock(&state->seqlock);
1011
	if (deleg_stateid != NULL) {
1012
		nfs4_stateid_copy(&state->stateid, deleg_stateid);
1013 1014 1015
		set_bit(NFS_DELEGATED_STATE, &state->flags);
	}
	if (open_stateid != NULL)
1016
		nfs_set_open_stateid_locked(state, open_stateid, fmode);
1017 1018
	write_sequnlock(&state->seqlock);
	spin_lock(&state->owner->so_lock);
1019
	update_open_stateflags(state, fmode);
1020
	spin_unlock(&state->owner->so_lock);
L
Linus Torvalds 已提交
1021 1022
}

1023
static int update_open_stateid(struct nfs4_state *state, nfs4_stateid *open_stateid, nfs4_stateid *delegation, fmode_t fmode)
1024 1025 1026 1027 1028
{
	struct nfs_inode *nfsi = NFS_I(state->inode);
	struct nfs_delegation *deleg_cur;
	int ret = 0;

1029
	fmode &= (FMODE_READ|FMODE_WRITE);
1030 1031 1032 1033 1034 1035 1036 1037

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

	spin_lock(&deleg_cur->lock);
	if (nfsi->delegation != deleg_cur ||
1038
	   test_bit(NFS_DELEGATION_RETURNING, &deleg_cur->flags) ||
1039
	    (deleg_cur->type & fmode) != fmode)
1040 1041 1042 1043
		goto no_delegation_unlock;

	if (delegation == NULL)
		delegation = &deleg_cur->stateid;
1044
	else if (!nfs4_stateid_match(&deleg_cur->stateid, delegation))
1045 1046
		goto no_delegation_unlock;

1047
	nfs_mark_delegation_referenced(deleg_cur);
1048
	__update_open_stateid(state, open_stateid, &deleg_cur->stateid, fmode);
1049 1050 1051 1052 1053 1054 1055
	ret = 1;
no_delegation_unlock:
	spin_unlock(&deleg_cur->lock);
no_delegation:
	rcu_read_unlock();

	if (!ret && open_stateid != NULL) {
1056
		__update_open_stateid(state, open_stateid, NULL, fmode);
1057 1058 1059 1060 1061 1062 1063
		ret = 1;
	}

	return ret;
}


1064
static void nfs4_return_incompatible_delegation(struct inode *inode, fmode_t fmode)
1065 1066 1067 1068 1069
{
	struct nfs_delegation *delegation;

	rcu_read_lock();
	delegation = rcu_dereference(NFS_I(inode)->delegation);
1070
	if (delegation == NULL || (delegation->type & fmode) == fmode) {
1071 1072 1073 1074
		rcu_read_unlock();
		return;
	}
	rcu_read_unlock();
1075
	nfs4_inode_return_delegation(inode);
1076 1077
}

1078
static struct nfs4_state *nfs4_try_open_cached(struct nfs4_opendata *opendata)
1079 1080 1081 1082
{
	struct nfs4_state *state = opendata->state;
	struct nfs_inode *nfsi = NFS_I(state->inode);
	struct nfs_delegation *delegation;
1083
	int open_mode = opendata->o_arg.open_flags;
1084
	fmode_t fmode = opendata->o_arg.fmode;
1085 1086 1087 1088
	nfs4_stateid stateid;
	int ret = -EAGAIN;

	for (;;) {
1089
		if (can_open_cached(state, fmode, open_mode)) {
1090
			spin_lock(&state->owner->so_lock);
1091 1092
			if (can_open_cached(state, fmode, open_mode)) {
				update_open_stateflags(state, fmode);
1093 1094 1095 1096 1097
				spin_unlock(&state->owner->so_lock);
				goto out_return_state;
			}
			spin_unlock(&state->owner->so_lock);
		}
1098 1099
		rcu_read_lock();
		delegation = rcu_dereference(nfsi->delegation);
1100
		if (!can_open_delegated(delegation, fmode)) {
1101
			rcu_read_unlock();
1102
			break;
1103
		}
1104
		/* Save the delegation */
1105
		nfs4_stateid_copy(&stateid, &delegation->stateid);
1106
		rcu_read_unlock();
1107
		nfs_release_seqid(opendata->o_arg.seqid);
1108 1109 1110 1111 1112
		if (!opendata->is_recover) {
			ret = nfs_may_open(state->inode, state->owner->so_cred, open_mode);
			if (ret != 0)
				goto out;
		}
1113
		ret = -EAGAIN;
1114 1115

		/* Try to update the stateid using the delegation */
1116
		if (update_open_stateid(state, NULL, &stateid, fmode))
1117
			goto out_return_state;
1118 1119 1120 1121 1122 1123 1124 1125
	}
out:
	return ERR_PTR(ret);
out_return_state:
	atomic_inc(&state->count);
	return state;
}

1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 1195 1196
static void
nfs4_opendata_check_deleg(struct nfs4_opendata *data, struct nfs4_state *state)
{
	struct nfs_client *clp = NFS_SERVER(state->inode)->nfs_client;
	struct nfs_delegation *delegation;
	int delegation_flags = 0;

	rcu_read_lock();
	delegation = rcu_dereference(NFS_I(state->inode)->delegation);
	if (delegation)
		delegation_flags = delegation->flags;
	rcu_read_unlock();
	if (data->o_arg.claim == NFS4_OPEN_CLAIM_DELEGATE_CUR) {
		pr_err_ratelimited("NFS: Broken NFSv4 server %s is "
				   "returning a delegation for "
				   "OPEN(CLAIM_DELEGATE_CUR)\n",
				   clp->cl_hostname);
	} else if ((delegation_flags & 1UL<<NFS_DELEGATION_NEED_RECLAIM) == 0)
		nfs_inode_set_delegation(state->inode,
					 data->owner->so_cred,
					 &data->o_res);
	else
		nfs_inode_reclaim_delegation(state->inode,
					     data->owner->so_cred,
					     &data->o_res);
}

/*
 * Check the inode attributes against the CLAIM_PREVIOUS returned attributes
 * and update the nfs4_state.
 */
static struct nfs4_state *
_nfs4_opendata_reclaim_to_nfs4_state(struct nfs4_opendata *data)
{
	struct inode *inode = data->state->inode;
	struct nfs4_state *state = data->state;
	int ret;

	if (!data->rpc_done) {
		ret = data->rpc_status;
		goto err;
	}

	ret = -ESTALE;
	if (!(data->f_attr.valid & NFS_ATTR_FATTR_TYPE) ||
	    !(data->f_attr.valid & NFS_ATTR_FATTR_FILEID) ||
	    !(data->f_attr.valid & NFS_ATTR_FATTR_CHANGE))
		goto err;

	ret = -ENOMEM;
	state = nfs4_get_open_state(inode, data->owner);
	if (state == NULL)
		goto err;

	ret = nfs_refresh_inode(inode, &data->f_attr);
	if (ret)
		goto err;

	if (data->o_res.delegation_type != 0)
		nfs4_opendata_check_deleg(data, state);
	update_open_stateid(state, &data->o_res.stateid, NULL,
			    data->o_arg.fmode);

	return state;
err:
	return ERR_PTR(ret);

}

static struct nfs4_state *
_nfs4_opendata_to_nfs4_state(struct nfs4_opendata *data)
1197 1198 1199
{
	struct inode *inode;
	struct nfs4_state *state = NULL;
1200
	int ret;
1201

1202
	if (!data->rpc_done) {
1203
		state = nfs4_try_open_cached(data);
1204 1205 1206
		goto out;
	}

1207
	ret = -EAGAIN;
1208
	if (!(data->f_attr.valid & NFS_ATTR_FATTR))
1209
		goto err;
1210
	inode = nfs_fhget(data->dir->d_sb, &data->o_res.fh, &data->f_attr);
1211
	ret = PTR_ERR(inode);
1212
	if (IS_ERR(inode))
1213 1214
		goto err;
	ret = -ENOMEM;
1215 1216
	state = nfs4_get_open_state(inode, data->owner);
	if (state == NULL)
1217
		goto err_put_inode;
1218 1219
	if (data->o_res.delegation_type != 0)
		nfs4_opendata_check_deleg(data, state);
1220
	update_open_stateid(state, &data->o_res.stateid, NULL,
1221
			data->o_arg.fmode);
1222
	iput(inode);
1223
out:
1224
	nfs_release_seqid(data->o_arg.seqid);
1225
	return state;
1226 1227 1228 1229
err_put_inode:
	iput(inode);
err:
	return ERR_PTR(ret);
1230 1231
}

1232 1233 1234 1235 1236 1237 1238 1239
static struct nfs4_state *
nfs4_opendata_to_nfs4_state(struct nfs4_opendata *data)
{
	if (data->o_arg.claim == NFS4_OPEN_CLAIM_PREVIOUS)
		return _nfs4_opendata_reclaim_to_nfs4_state(data);
	return _nfs4_opendata_to_nfs4_state(data);
}

1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254 1255 1256
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);
}

1257 1258
static struct nfs4_opendata *nfs4_open_recoverdata_alloc(struct nfs_open_context *ctx,
		struct nfs4_state *state, enum open_claim_type4 claim)
T
Trond Myklebust 已提交
1259 1260 1261
{
	struct nfs4_opendata *opendata;

1262 1263
	opendata = nfs4_opendata_alloc(ctx->dentry, state->owner, 0, 0,
			NULL, claim, GFP_NOFS);
T
Trond Myklebust 已提交
1264 1265 1266 1267 1268 1269 1270
	if (opendata == NULL)
		return ERR_PTR(-ENOMEM);
	opendata->state = state;
	atomic_inc(&state->count);
	return opendata;
}

1271
static int nfs4_open_recover_helper(struct nfs4_opendata *opendata, fmode_t fmode, struct nfs4_state **res)
1272
{
1273
	struct nfs4_state *newstate;
1274 1275
	int ret;

1276 1277
	opendata->o_arg.open_flags = 0;
	opendata->o_arg.fmode = fmode;
1278 1279 1280
	memset(&opendata->o_res, 0, sizeof(opendata->o_res));
	memset(&opendata->c_res, 0, sizeof(opendata->c_res));
	nfs4_init_opendata_res(opendata);
1281
	ret = _nfs4_recover_proc_open(opendata);
1282 1283
	if (ret != 0)
		return ret; 
1284
	newstate = nfs4_opendata_to_nfs4_state(opendata);
1285 1286
	if (IS_ERR(newstate))
		return PTR_ERR(newstate);
1287
	nfs4_close_state(newstate, fmode);
1288
	*res = newstate;
1289 1290 1291 1292 1293 1294 1295 1296 1297
	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_* */
1298
	clear_bit(NFS_DELEGATED_STATE, &state->flags);
1299
	clear_bit(NFS_OPEN_STATE, &state->flags);
1300 1301
	smp_rmb();
	if (state->n_rdwr != 0) {
T
Trond Myklebust 已提交
1302
		clear_bit(NFS_O_RDWR_STATE, &state->flags);
1303
		ret = nfs4_open_recover_helper(opendata, FMODE_READ|FMODE_WRITE, &newstate);
1304 1305
		if (ret != 0)
			return ret;
1306 1307
		if (newstate != state)
			return -ESTALE;
1308 1309
	}
	if (state->n_wronly != 0) {
T
Trond Myklebust 已提交
1310
		clear_bit(NFS_O_WRONLY_STATE, &state->flags);
1311
		ret = nfs4_open_recover_helper(opendata, FMODE_WRITE, &newstate);
1312 1313
		if (ret != 0)
			return ret;
1314 1315
		if (newstate != state)
			return -ESTALE;
1316 1317
	}
	if (state->n_rdonly != 0) {
T
Trond Myklebust 已提交
1318
		clear_bit(NFS_O_RDONLY_STATE, &state->flags);
1319
		ret = nfs4_open_recover_helper(opendata, FMODE_READ, &newstate);
1320 1321
		if (ret != 0)
			return ret;
1322 1323
		if (newstate != state)
			return -ESTALE;
1324
	}
1325 1326 1327 1328 1329
	/*
	 * 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 &&
1330
	    !nfs4_stateid_match(&state->stateid, &state->open_stateid)) {
1331
		write_seqlock(&state->seqlock);
1332
		if (test_bit(NFS_DELEGATED_STATE, &state->flags) == 0)
1333
			nfs4_stateid_copy(&state->stateid, &state->open_stateid);
1334
		write_sequnlock(&state->seqlock);
1335
	}
1336 1337 1338
	return 0;
}

L
Linus Torvalds 已提交
1339 1340 1341 1342
/*
 * OPEN_RECLAIM:
 * 	reclaim state on the server after a reboot.
 */
1343
static int _nfs4_do_open_reclaim(struct nfs_open_context *ctx, struct nfs4_state *state)
L
Linus Torvalds 已提交
1344
{
1345
	struct nfs_delegation *delegation;
1346
	struct nfs4_opendata *opendata;
1347
	fmode_t delegation_type = 0;
L
Linus Torvalds 已提交
1348 1349
	int status;

1350 1351
	opendata = nfs4_open_recoverdata_alloc(ctx, state,
			NFS4_OPEN_CLAIM_PREVIOUS);
T
Trond Myklebust 已提交
1352 1353
	if (IS_ERR(opendata))
		return PTR_ERR(opendata);
1354 1355
	rcu_read_lock();
	delegation = rcu_dereference(NFS_I(state->inode)->delegation);
1356
	if (delegation != NULL && test_bit(NFS_DELEGATION_NEED_RECLAIM, &delegation->flags) != 0)
1357
		delegation_type = delegation->type;
1358
	rcu_read_unlock();
1359 1360
	opendata->o_arg.u.delegation_type = delegation_type;
	status = nfs4_open_recover(opendata, state);
1361
	nfs4_opendata_put(opendata);
L
Linus Torvalds 已提交
1362 1363 1364
	return status;
}

1365
static int nfs4_do_open_reclaim(struct nfs_open_context *ctx, struct nfs4_state *state)
L
Linus Torvalds 已提交
1366 1367 1368 1369 1370
{
	struct nfs_server *server = NFS_SERVER(state->inode);
	struct nfs4_exception exception = { };
	int err;
	do {
1371
		err = _nfs4_do_open_reclaim(ctx, state);
1372 1373
		if (nfs4_clear_cap_atomic_open_v1(server, err, &exception))
			continue;
1374
		if (err != -NFS4ERR_DELAY)
1375 1376
			break;
		nfs4_handle_exception(server, err, &exception);
L
Linus Torvalds 已提交
1377 1378 1379 1380
	} while (exception.retry);
	return err;
}

1381 1382 1383 1384 1385 1386 1387
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))
1388
		return -EAGAIN;
1389
	ret = nfs4_do_open_reclaim(ctx, state);
1390 1391 1392 1393
	put_nfs_open_context(ctx);
	return ret;
}

1394
static int nfs4_handle_delegation_recall_error(struct nfs_server *server, struct nfs4_state *state, const nfs4_stateid *stateid, int err)
L
Linus Torvalds 已提交
1395
{
1396 1397 1398 1399 1400 1401 1402 1403 1404 1405 1406 1407 1408 1409 1410 1411 1412 1413 1414 1415 1416 1417 1418 1419 1420 1421
	switch (err) {
		default:
			printk(KERN_ERR "NFS: %s: unhandled error "
					"%d.\n", __func__, err);
		case 0:
		case -ENOENT:
		case -ESTALE:
			break;
		case -NFS4ERR_BADSESSION:
		case -NFS4ERR_BADSLOT:
		case -NFS4ERR_BAD_HIGH_SLOT:
		case -NFS4ERR_CONN_NOT_BOUND_TO_SESSION:
		case -NFS4ERR_DEADSESSION:
			set_bit(NFS_DELEGATED_STATE, &state->flags);
			nfs4_schedule_session_recovery(server->nfs_client->cl_session, err);
			return -EAGAIN;
		case -NFS4ERR_STALE_CLIENTID:
		case -NFS4ERR_STALE_STATEID:
			set_bit(NFS_DELEGATED_STATE, &state->flags);
		case -NFS4ERR_EXPIRED:
			/* Don't recall a delegation if it was lost */
			nfs4_schedule_lease_recovery(server->nfs_client);
			return -EAGAIN;
		case -NFS4ERR_DELEG_REVOKED:
		case -NFS4ERR_ADMIN_REVOKED:
		case -NFS4ERR_BAD_STATEID:
1422
		case -NFS4ERR_OPENMODE:
1423 1424 1425 1426 1427 1428 1429 1430 1431
			nfs_inode_find_state_and_recover(state->inode,
					stateid);
			nfs4_schedule_stateid_recovery(server, state);
			return 0;
		case -NFS4ERR_DELAY:
		case -NFS4ERR_GRACE:
			set_bit(NFS_DELEGATED_STATE, &state->flags);
			ssleep(1);
			return -EAGAIN;
1432 1433 1434 1435
		case -ENOMEM:
		case -NFS4ERR_DENIED:
			/* kill_proc(fl->fl_pid, SIGLOST, 1); */
			return 0;
1436
	}
L
Linus Torvalds 已提交
1437 1438 1439
	return err;
}

1440 1441 1442 1443 1444 1445 1446 1447 1448 1449 1450 1451 1452 1453 1454 1455
int nfs4_open_delegation_recall(struct nfs_open_context *ctx, struct nfs4_state *state, const nfs4_stateid *stateid)
{
	struct nfs_server *server = NFS_SERVER(state->inode);
	struct nfs4_opendata *opendata;
	int err;

	opendata = nfs4_open_recoverdata_alloc(ctx, state,
			NFS4_OPEN_CLAIM_DELEG_CUR_FH);
	if (IS_ERR(opendata))
		return PTR_ERR(opendata);
	nfs4_stateid_copy(&opendata->o_arg.u.delegation, stateid);
	err = nfs4_open_recover(opendata, state);
	nfs4_opendata_put(opendata);
	return nfs4_handle_delegation_recall_error(server, state, stateid, err);
}

1456 1457 1458 1459 1460
static void nfs4_open_confirm_done(struct rpc_task *task, void *calldata)
{
	struct nfs4_opendata *data = calldata;

	data->rpc_status = task->tk_status;
1461
	if (data->rpc_status == 0) {
1462
		nfs4_stateid_copy(&data->o_res.stateid, &data->c_res.stateid);
1463
		nfs_confirm_seqid(&data->owner->so_seqid, 0);
1464
		renew_lease(data->o_res.server, data->timestamp);
1465
		data->rpc_done = 1;
1466
	}
1467 1468 1469 1470 1471 1472 1473 1474 1475 1476 1477
}

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! */
1478
	if (!data->rpc_done)
1479 1480
		goto out_free;
	state = nfs4_opendata_to_nfs4_state(data);
1481
	if (!IS_ERR(state))
1482
		nfs4_close_state(state, data->o_arg.fmode);
1483
out_free:
1484
	nfs4_opendata_put(data);
1485 1486 1487 1488 1489 1490 1491 1492 1493 1494 1495 1496 1497 1498
}

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

/*
 * Note: On error, nfs4_proc_open_confirm will free the struct nfs4_opendata
 */
static int _nfs4_proc_open_confirm(struct nfs4_opendata *data)
{
	struct nfs_server *server = NFS_SERVER(data->dir->d_inode);
	struct rpc_task *task;
1499 1500 1501 1502 1503 1504
	struct  rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_OPEN_CONFIRM],
		.rpc_argp = &data->c_arg,
		.rpc_resp = &data->c_res,
		.rpc_cred = data->owner->so_cred,
	};
T
Trond Myklebust 已提交
1505 1506
	struct rpc_task_setup task_setup_data = {
		.rpc_client = server->client,
1507
		.rpc_message = &msg,
T
Trond Myklebust 已提交
1508 1509
		.callback_ops = &nfs4_open_confirm_ops,
		.callback_data = data,
1510
		.workqueue = nfsiod_workqueue,
T
Trond Myklebust 已提交
1511 1512
		.flags = RPC_TASK_ASYNC,
	};
L
Linus Torvalds 已提交
1513 1514
	int status;

1515
	kref_get(&data->kref);
1516 1517
	data->rpc_done = 0;
	data->rpc_status = 0;
1518
	data->timestamp = jiffies;
T
Trond Myklebust 已提交
1519
	task = rpc_run_task(&task_setup_data);
1520
	if (IS_ERR(task))
1521 1522 1523 1524 1525 1526 1527
		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;
1528
	rpc_put_task(task);
L
Linus Torvalds 已提交
1529 1530 1531
	return status;
}

1532
static void nfs4_open_prepare(struct rpc_task *task, void *calldata)
L
Linus Torvalds 已提交
1533
{
1534 1535
	struct nfs4_opendata *data = calldata;
	struct nfs4_state_owner *sp = data->owner;
1536
	struct nfs_client *clp = sp->so_server->nfs_client;
1537

1538
	if (nfs_wait_on_sequence(data->o_arg.seqid, task) != 0)
1539
		goto out_wait;
1540 1541 1542 1543 1544 1545 1546
	/*
	 * 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;

1547
		if (can_open_cached(data->state, data->o_arg.fmode, data->o_arg.open_flags))
1548
			goto out_no_action;
1549 1550
		rcu_read_lock();
		delegation = rcu_dereference(NFS_I(data->state->inode)->delegation);
1551
		if (data->o_arg.claim != NFS4_OPEN_CLAIM_DELEGATE_CUR &&
1552
		    data->o_arg.claim != NFS4_OPEN_CLAIM_DELEG_CUR_FH &&
1553 1554
		    can_open_delegated(delegation, data->o_arg.fmode))
			goto unlock_no_action;
1555 1556
		rcu_read_unlock();
	}
1557
	/* Update client id. */
1558
	data->o_arg.clientid = clp->cl_clientid;
1559 1560 1561 1562
	switch (data->o_arg.claim) {
	case NFS4_OPEN_CLAIM_PREVIOUS:
	case NFS4_OPEN_CLAIM_DELEG_CUR_FH:
	case NFS4_OPEN_CLAIM_DELEG_PREV_FH:
1563
		data->o_arg.open_bitmap = &nfs4_open_noattr_bitmap[0];
1564 1565
	case NFS4_OPEN_CLAIM_FH:
		task->tk_msg.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_OPEN_NOATTR];
T
Trond Myklebust 已提交
1566 1567
		nfs_copy_fh(&data->o_res.fh, data->o_arg.fh);
	}
1568
	data->timestamp = jiffies;
1569
	if (nfs4_setup_sequence(data->o_arg.server,
1570
				&data->o_arg.seq_args,
1571 1572 1573
				&data->o_res.seq_res,
				task) != 0)
		nfs_release_seqid(data->o_arg.seqid);
1574 1575 1576 1577 1578 1579 1580 1581 1582 1583

	/* Set the create mode (note dependency on the session type) */
	data->o_arg.createmode = NFS4_CREATE_UNCHECKED;
	if (data->o_arg.open_flags & O_EXCL) {
		data->o_arg.createmode = NFS4_CREATE_EXCLUSIVE;
		if (nfs4_has_persistent_session(clp))
			data->o_arg.createmode = NFS4_CREATE_GUARDED;
		else if (clp->cl_mvops->minor_version > 0)
			data->o_arg.createmode = NFS4_CREATE_EXCLUSIVE4_1;
	}
1584
	return;
1585 1586
unlock_no_action:
	rcu_read_unlock();
1587 1588
out_no_action:
	task->tk_action = NULL;
1589
out_wait:
1590
	nfs4_sequence_done(task, &data->o_res.seq_res);
1591 1592
}

1593 1594 1595
static void nfs4_open_done(struct rpc_task *task, void *calldata)
{
	struct nfs4_opendata *data = calldata;
L
Linus Torvalds 已提交
1596

1597
	data->rpc_status = task->tk_status;
1598

1599 1600
	if (!nfs4_sequence_done(task, &data->o_res.seq_res))
		return;
1601

1602
	if (task->tk_status == 0) {
1603 1604
		if (data->o_res.f_attr->valid & NFS_ATTR_FATTR_TYPE) {
			switch (data->o_res.f_attr->mode & S_IFMT) {
1605 1606 1607
			case S_IFREG:
				break;
			case S_IFLNK:
1608
				data->rpc_status = -ELOOP;
1609 1610
				break;
			case S_IFDIR:
1611
				data->rpc_status = -EISDIR;
1612 1613
				break;
			default:
1614
				data->rpc_status = -ENOTDIR;
1615
			}
1616
		}
1617
		renew_lease(data->o_res.server, data->timestamp);
1618 1619
		if (!(data->o_res.rflags & NFS4_OPEN_RESULT_CONFIRM))
			nfs_confirm_seqid(&data->owner->so_seqid, 0);
1620
	}
1621
	data->rpc_done = 1;
1622
}
1623

1624 1625 1626 1627 1628 1629 1630 1631 1632
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! */
1633
	if (data->rpc_status != 0 || !data->rpc_done)
1634 1635 1636 1637 1638
		goto out_free;
	/* In case we need an open_confirm, no cleanup! */
	if (data->o_res.rflags & NFS4_OPEN_RESULT_CONFIRM)
		goto out_free;
	state = nfs4_opendata_to_nfs4_state(data);
1639
	if (!IS_ERR(state))
1640
		nfs4_close_state(state, data->o_arg.fmode);
1641
out_free:
1642
	nfs4_opendata_put(data);
1643 1644 1645 1646 1647 1648 1649 1650
}

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,
};

1651
static int nfs4_run_open_task(struct nfs4_opendata *data, int isrecover)
1652 1653 1654 1655 1656 1657
{
	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;
1658 1659 1660 1661 1662 1663
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_OPEN],
		.rpc_argp = o_arg,
		.rpc_resp = o_res,
		.rpc_cred = data->owner->so_cred,
	};
T
Trond Myklebust 已提交
1664 1665
	struct rpc_task_setup task_setup_data = {
		.rpc_client = server->client,
1666
		.rpc_message = &msg,
T
Trond Myklebust 已提交
1667 1668
		.callback_ops = &nfs4_open_ops,
		.callback_data = data,
1669
		.workqueue = nfsiod_workqueue,
T
Trond Myklebust 已提交
1670 1671
		.flags = RPC_TASK_ASYNC,
	};
1672 1673
	int status;

1674
	nfs41_init_sequence(&o_arg->seq_args, &o_res->seq_res, 1);
1675
	kref_get(&data->kref);
1676 1677
	data->rpc_done = 0;
	data->rpc_status = 0;
1678
	data->cancelled = 0;
1679 1680
	data->is_recover = 0;
	if (isrecover) {
1681
		nfs4_set_sequence_privileged(&o_arg->seq_args);
1682 1683
		data->is_recover = 1;
	}
T
Trond Myklebust 已提交
1684
	task = rpc_run_task(&task_setup_data);
1685 1686 1687 1688 1689 1690 1691 1692 1693 1694 1695 1696 1697 1698 1699 1700 1701 1702 1703 1704 1705 1706 1707
        if (IS_ERR(task))
                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;
        rpc_put_task(task);

	return status;
}

static int _nfs4_recover_proc_open(struct nfs4_opendata *data)
{
	struct inode *dir = data->dir->d_inode;
	struct nfs_openres *o_res = &data->o_res;
        int status;

	status = nfs4_run_open_task(data, 1);
	if (status != 0 || !data->rpc_done)
		return status;

1708 1709
	nfs_fattr_map_and_free_names(NFS_SERVER(dir), &data->f_attr);

1710 1711 1712 1713 1714 1715 1716 1717 1718
	if (o_res->rflags & NFS4_OPEN_RESULT_CONFIRM) {
		status = _nfs4_proc_open_confirm(data);
		if (status != 0)
			return status;
	}

	return status;
}

1719 1720
static int nfs4_opendata_access(struct rpc_cred *cred,
				struct nfs4_opendata *opendata,
1721 1722
				struct nfs4_state *state, fmode_t fmode,
				int openflags)
1723 1724 1725 1726 1727 1728 1729 1730 1731 1732
{
	struct nfs_access_entry cache;
	u32 mask;

	/* access call failed or for some reason the server doesn't
	 * support any access modes -- defer access call until later */
	if (opendata->o_res.access_supported == 0)
		return 0;

	mask = 0;
1733
	/* don't check MAY_WRITE - a newly created file may not have
1734 1735 1736 1737 1738 1739 1740 1741
	 * write mode bits, but POSIX allows the creating process to write.
	 * use openflags to check for exec, because fmode won't
	 * always have FMODE_EXEC set when file open for exec. */
	if (openflags & __FMODE_EXEC) {
		/* ONLY check for exec rights */
		mask = MAY_EXEC;
	} else if (fmode & FMODE_READ)
		mask = MAY_READ;
1742 1743 1744 1745 1746 1747

	cache.cred = cred;
	cache.jiffies = jiffies;
	nfs_access_set_mask(&cache, opendata->o_res.access_result);
	nfs_access_add_cache(state->inode, &cache);

1748
	if ((mask & ~cache.mask & (MAY_READ | MAY_EXEC)) == 0)
1749 1750 1751 1752
		return 0;

	/* even though OPEN succeeded, access is denied. Close the file */
	nfs4_close_state(state, fmode);
1753
	return -EACCES;
1754 1755
}

1756 1757 1758 1759 1760 1761 1762 1763 1764 1765 1766 1767
/*
 * 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;
	int status;

	status = nfs4_run_open_task(data, 0);
1768 1769 1770 1771 1772 1773
	if (!data->rpc_done)
		return status;
	if (status != 0) {
		if (status == -NFS4ERR_BADNAME &&
				!(o_arg->open_flags & O_CREAT))
			return -ENOENT;
1774
		return status;
1775
	}
1776

1777 1778
	nfs_fattr_map_and_free_names(server, &data->f_attr);

1779
	if (o_arg->open_flags & O_CREAT)
1780
		update_changeattr(dir, &o_res->cinfo);
T
Trond Myklebust 已提交
1781 1782
	if ((o_res->rflags & NFS4_OPEN_RESULT_LOCKTYPE_POSIX) == 0)
		server->caps &= ~NFS_CAP_POSIX_LOCK;
L
Linus Torvalds 已提交
1783
	if(o_res->rflags & NFS4_OPEN_RESULT_CONFIRM) {
1784
		status = _nfs4_proc_open_confirm(data);
L
Linus Torvalds 已提交
1785
		if (status != 0)
1786
			return status;
L
Linus Torvalds 已提交
1787 1788
	}
	if (!(o_res->f_attr->valid & NFS_ATTR_FATTR))
1789
		_nfs4_proc_getattr(server, &o_res->fh, o_res->f_attr);
1790
	return 0;
L
Linus Torvalds 已提交
1791 1792
}

A
Andy Adamson 已提交
1793 1794 1795 1796 1797
static int nfs4_recover_expired_lease(struct nfs_server *server)
{
	return nfs4_client_recover_expired_lease(server->nfs_client);
}

L
Linus Torvalds 已提交
1798 1799 1800 1801 1802
/*
 * OPEN_EXPIRED:
 * 	reclaim state on the server after a network partition.
 * 	Assumes caller holds the appropriate lock
 */
1803
static int _nfs4_open_expired(struct nfs_open_context *ctx, struct nfs4_state *state)
L
Linus Torvalds 已提交
1804
{
1805
	struct nfs4_opendata *opendata;
1806
	int ret;
L
Linus Torvalds 已提交
1807

1808
	opendata = nfs4_open_recoverdata_alloc(ctx, state,
1809
			NFS4_OPEN_CLAIM_FH);
T
Trond Myklebust 已提交
1810 1811
	if (IS_ERR(opendata))
		return PTR_ERR(opendata);
1812
	ret = nfs4_open_recover(opendata, state);
1813
	if (ret == -ESTALE)
1814
		d_drop(ctx->dentry);
1815
	nfs4_opendata_put(opendata);
1816
	return ret;
L
Linus Torvalds 已提交
1817 1818
}

1819
static int nfs4_do_open_expired(struct nfs_open_context *ctx, struct nfs4_state *state)
1820
{
1821
	struct nfs_server *server = NFS_SERVER(state->inode);
1822 1823 1824 1825
	struct nfs4_exception exception = { };
	int err;

	do {
1826
		err = _nfs4_open_expired(ctx, state);
1827 1828
		if (nfs4_clear_cap_atomic_open_v1(server, err, &exception))
			continue;
1829 1830 1831 1832 1833 1834 1835 1836
		switch (err) {
		default:
			goto out;
		case -NFS4ERR_GRACE:
		case -NFS4ERR_DELAY:
			nfs4_handle_exception(server, err, &exception);
			err = 0;
		}
1837
	} while (exception.retry);
1838
out:
1839 1840 1841
	return err;
}

L
Linus Torvalds 已提交
1842 1843 1844
static int nfs4_open_expired(struct nfs4_state_owner *sp, struct nfs4_state *state)
{
	struct nfs_open_context *ctx;
1845
	int ret;
L
Linus Torvalds 已提交
1846

1847 1848
	ctx = nfs4_state_find_open_context(state);
	if (IS_ERR(ctx))
1849
		return -EAGAIN;
1850
	ret = nfs4_do_open_expired(ctx, state);
1851 1852
	put_nfs_open_context(ctx);
	return ret;
L
Linus Torvalds 已提交
1853 1854
}

1855
#if defined(CONFIG_NFS_V4_1)
1856
static void nfs41_clear_delegation_stateid(struct nfs4_state *state)
1857 1858
{
	struct nfs_server *server = NFS_SERVER(state->inode);
1859
	nfs4_stateid *stateid = &state->stateid;
1860 1861 1862
	struct nfs_delegation *delegation;
	struct rpc_cred *cred = NULL;
	int status = -NFS4ERR_BAD_STATEID;
1863 1864 1865 1866 1867

	/* If a state reset has been done, test_stateid is unneeded */
	if (test_bit(NFS_DELEGATED_STATE, &state->flags) == 0)
		return;

1868 1869 1870 1871 1872 1873 1874 1875 1876 1877 1878
	/* Get the delegation credential for use by test/free_stateid */
	rcu_read_lock();
	delegation = rcu_dereference(NFS_I(state->inode)->delegation);
	if (delegation != NULL &&
	    nfs4_stateid_match(&delegation->stateid, stateid)) {
		cred = get_rpccred(delegation->cred);
		rcu_read_unlock();
		status = nfs41_test_stateid(server, stateid, cred);
	} else
		rcu_read_unlock();

1879 1880 1881 1882
	if (status != NFS_OK) {
		/* Free the stateid unless the server explicitly
		 * informs us the stateid is unrecognized. */
		if (status != -NFS4ERR_BAD_STATEID)
1883
			nfs41_free_stateid(server, stateid, cred);
1884
		nfs_remove_bad_delegation(state->inode);
1885

1886 1887 1888
		write_seqlock(&state->seqlock);
		nfs4_stateid_copy(&state->stateid, &state->open_stateid);
		write_sequnlock(&state->seqlock);
1889 1890
		clear_bit(NFS_DELEGATED_STATE, &state->flags);
	}
1891 1892 1893

	if (cred != NULL)
		put_rpccred(cred);
1894 1895 1896 1897 1898 1899 1900 1901 1902 1903 1904 1905 1906
}

/**
 * nfs41_check_open_stateid - possibly free an open stateid
 *
 * @state: NFSv4 state for an inode
 *
 * Returns NFS_OK if recovery for this stateid is now finished.
 * Otherwise a negative NFS4ERR value is returned.
 */
static int nfs41_check_open_stateid(struct nfs4_state *state)
{
	struct nfs_server *server = NFS_SERVER(state->inode);
1907
	nfs4_stateid *stateid = &state->open_stateid;
1908
	struct rpc_cred *cred = state->owner->so_cred;
1909 1910 1911 1912 1913 1914 1915 1916
	int status;

	/* If a state reset has been done, test_stateid is unneeded */
	if ((test_bit(NFS_O_RDONLY_STATE, &state->flags) == 0) &&
	    (test_bit(NFS_O_WRONLY_STATE, &state->flags) == 0) &&
	    (test_bit(NFS_O_RDWR_STATE, &state->flags) == 0))
		return -NFS4ERR_BAD_STATEID;

1917
	status = nfs41_test_stateid(server, stateid, cred);
1918 1919 1920 1921
	if (status != NFS_OK) {
		/* Free the stateid unless the server explicitly
		 * informs us the stateid is unrecognized. */
		if (status != -NFS4ERR_BAD_STATEID)
1922
			nfs41_free_stateid(server, stateid, cred);
1923 1924 1925 1926

		clear_bit(NFS_O_RDONLY_STATE, &state->flags);
		clear_bit(NFS_O_WRONLY_STATE, &state->flags);
		clear_bit(NFS_O_RDWR_STATE, &state->flags);
1927
		clear_bit(NFS_OPEN_STATE, &state->flags);
1928 1929 1930 1931 1932 1933
	}
	return status;
}

static int nfs41_open_expired(struct nfs4_state_owner *sp, struct nfs4_state *state)
{
1934
	int status;
1935

1936 1937
	nfs41_clear_delegation_stateid(state);
	status = nfs41_check_open_stateid(state);
1938 1939 1940
	if (status != NFS_OK)
		status = nfs4_open_expired(sp, state);
	return status;
1941 1942 1943
}
#endif

1944 1945 1946 1947 1948 1949 1950 1951 1952 1953 1954 1955 1956 1957 1958 1959
/*
 * 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;
}

1960 1961 1962
static int _nfs4_open_and_get_state(struct nfs4_opendata *opendata,
		fmode_t fmode,
		int flags,
1963
		struct nfs_open_context *ctx)
1964 1965 1966
{
	struct nfs4_state_owner *sp = opendata->owner;
	struct nfs_server *server = sp->so_server;
1967
	struct dentry *dentry;
1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984
	struct nfs4_state *state;
	unsigned int seq;
	int ret;

	seq = raw_seqcount_begin(&sp->so_reclaim_seqcount);

	ret = _nfs4_proc_open(opendata);
	if (ret != 0)
		goto out;

	state = nfs4_opendata_to_nfs4_state(opendata);
	ret = PTR_ERR(state);
	if (IS_ERR(state))
		goto out;
	if (server->caps & NFS_CAP_POSIX_LOCK)
		set_bit(NFS_STATE_POSIX_LOCKS, &state->flags);

1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999
	dentry = opendata->dentry;
	if (dentry->d_inode == NULL) {
		/* FIXME: Is this d_drop() ever needed? */
		d_drop(dentry);
		dentry = d_add_unique(dentry, igrab(state->inode));
		if (dentry == NULL) {
			dentry = opendata->dentry;
		} else if (dentry != ctx->dentry) {
			dput(ctx->dentry);
			ctx->dentry = dget(dentry);
		}
		nfs_set_verifier(dentry,
				nfs_save_change_attribute(opendata->dir->d_inode));
	}

2000 2001 2002 2003
	ret = nfs4_opendata_access(sp->so_cred, opendata, state, fmode, flags);
	if (ret != 0)
		goto out;

2004
	ctx->state = state;
2005 2006 2007 2008 2009
	if (dentry->d_inode == state->inode) {
		nfs_inode_attach_open_context(ctx);
		if (read_seqcount_retry(&sp->so_reclaim_seqcount, seq))
			nfs4_schedule_stateid_recovery(server, state);
	}
2010 2011 2012 2013
out:
	return ret;
}

L
Linus Torvalds 已提交
2014
/*
2015
 * Returns a referenced nfs4_state
L
Linus Torvalds 已提交
2016
 */
2017
static int _nfs4_do_open(struct inode *dir,
2018
			struct nfs_open_context *ctx,
2019
			int flags,
2020
			struct iattr *sattr)
L
Linus Torvalds 已提交
2021 2022 2023 2024
{
	struct nfs4_state_owner  *sp;
	struct nfs4_state     *state = NULL;
	struct nfs_server       *server = NFS_SERVER(dir);
2025
	struct nfs4_opendata *opendata;
2026 2027 2028 2029
	struct dentry *dentry = ctx->dentry;
	struct rpc_cred *cred = ctx->cred;
	struct nfs4_threshold **ctx_th = &ctx->mdsthreshold;
	fmode_t fmode = ctx->mode & (FMODE_READ|FMODE_WRITE|FMODE_EXEC);
2030
	enum open_claim_type4 claim = NFS4_OPEN_CLAIM_NULL;
2031
	int status;
L
Linus Torvalds 已提交
2032 2033 2034

	/* Protect against reboot recovery conflicts */
	status = -ENOMEM;
2035 2036
	sp = nfs4_get_state_owner(server, cred, GFP_KERNEL);
	if (sp == NULL) {
L
Linus Torvalds 已提交
2037 2038 2039
		dprintk("nfs4_do_open: nfs4_get_state_owner failed!\n");
		goto out_err;
	}
2040 2041
	status = nfs4_recover_expired_lease(server);
	if (status != 0)
2042
		goto err_put_state_owner;
2043 2044
	if (dentry->d_inode != NULL)
		nfs4_return_incompatible_delegation(dentry->d_inode, fmode);
2045
	status = -ENOMEM;
2046 2047
	if (dentry->d_inode)
		claim = NFS4_OPEN_CLAIM_FH;
2048
	opendata = nfs4_opendata_alloc(dentry, sp, fmode, flags, sattr,
2049
			claim, GFP_KERNEL);
2050
	if (opendata == NULL)
T
Trond Myklebust 已提交
2051
		goto err_put_state_owner;
L
Linus Torvalds 已提交
2052

2053 2054 2055 2056
	if (ctx_th && server->attr_bitmask[2] & FATTR4_WORD2_MDSTHRESHOLD) {
		opendata->f_attr.mdsthreshold = pnfs_mdsthreshold_alloc();
		if (!opendata->f_attr.mdsthreshold)
			goto err_opendata_put;
2057
		opendata->o_arg.open_bitmap = &nfs4_pnfs_open_bitmap[0];
2058
	}
2059 2060
	if (dentry->d_inode != NULL)
		opendata->state = nfs4_get_open_state(dentry->d_inode, sp);
2061

2062
	status = _nfs4_open_and_get_state(opendata, fmode, flags, ctx);
2063 2064
	if (status != 0)
		goto err_opendata_put;
2065
	state = ctx->state;
2066

2067 2068
	if ((opendata->o_arg.open_flags & O_EXCL) &&
	    (opendata->o_arg.createmode != NFS4_CREATE_GUARDED)) {
2069 2070 2071 2072 2073 2074 2075 2076 2077 2078
		nfs4_exclusive_attrset(opendata, sattr);

		nfs_fattr_init(opendata->o_res.f_attr);
		status = nfs4_do_setattr(state->inode, cred,
				opendata->o_res.f_attr, sattr,
				state);
		if (status == 0)
			nfs_setattr_update_inode(state->inode, sattr);
		nfs_post_op_update_inode(state->inode, opendata->o_res.f_attr);
	}
2079 2080 2081 2082 2083 2084 2085

	if (pnfs_use_threshold(ctx_th, opendata->f_attr.mdsthreshold, server))
		*ctx_th = opendata->f_attr.mdsthreshold;
	else
		kfree(opendata->f_attr.mdsthreshold);
	opendata->f_attr.mdsthreshold = NULL;

2086
	nfs4_opendata_put(opendata);
L
Linus Torvalds 已提交
2087 2088
	nfs4_put_state_owner(sp);
	return 0;
2089
err_opendata_put:
2090
	kfree(opendata->f_attr.mdsthreshold);
2091
	nfs4_opendata_put(opendata);
2092 2093
err_put_state_owner:
	nfs4_put_state_owner(sp);
L
Linus Torvalds 已提交
2094 2095 2096 2097 2098
out_err:
	return status;
}


2099
static struct nfs4_state *nfs4_do_open(struct inode *dir,
2100
					struct nfs_open_context *ctx,
2101
					int flags,
2102
					struct iattr *sattr)
L
Linus Torvalds 已提交
2103
{
2104
	struct nfs_server *server = NFS_SERVER(dir);
L
Linus Torvalds 已提交
2105 2106 2107 2108 2109
	struct nfs4_exception exception = { };
	struct nfs4_state *res;
	int status;

	do {
2110 2111
		status = _nfs4_do_open(dir, ctx, flags, sattr);
		res = ctx->state;
L
Linus Torvalds 已提交
2112 2113 2114 2115 2116 2117 2118 2119
		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
2120
		 * doing an OPEN, we assume that nfs_increment_open_seqid() will
L
Linus Torvalds 已提交
2121 2122 2123 2124 2125
		 * 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) {
2126
			pr_warn_ratelimited("NFS: v4 server %s "
2127 2128
					" returned a bad sequence-id error!\n",
					NFS_SERVER(dir)->nfs_client->cl_hostname);
L
Linus Torvalds 已提交
2129 2130 2131
			exception.retry = 1;
			continue;
		}
2132 2133 2134 2135 2136 2137 2138 2139 2140 2141
		/*
		 * 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;
		}
2142 2143 2144 2145 2146
		if (status == -EAGAIN) {
			/* We must have found a delegation */
			exception.retry = 1;
			continue;
		}
2147 2148 2149
		if (nfs4_clear_cap_atomic_open_v1(server, status, &exception))
			continue;
		res = ERR_PTR(nfs4_handle_exception(server,
L
Linus Torvalds 已提交
2150 2151 2152 2153 2154
					status, &exception));
	} while (exception.retry);
	return res;
}

2155 2156 2157
static int _nfs4_do_setattr(struct inode *inode, struct rpc_cred *cred,
			    struct nfs_fattr *fattr, struct iattr *sattr,
			    struct nfs4_state *state)
L
Linus Torvalds 已提交
2158
{
2159
	struct nfs_server *server = NFS_SERVER(inode);
L
Linus Torvalds 已提交
2160
        struct nfs_setattrargs  arg = {
2161
                .fh             = NFS_FH(inode),
L
Linus Torvalds 已提交
2162 2163 2164 2165 2166 2167 2168 2169 2170
                .iap            = sattr,
		.server		= server,
		.bitmask = server->attr_bitmask,
        };
        struct nfs_setattrres  res = {
		.fattr		= fattr,
		.server		= server,
        };
        struct rpc_message msg = {
2171 2172 2173 2174
		.rpc_proc	= &nfs4_procedures[NFSPROC4_CLNT_SETATTR],
		.rpc_argp	= &arg,
		.rpc_resp	= &res,
		.rpc_cred	= cred,
L
Linus Torvalds 已提交
2175
        };
2176
	unsigned long timestamp = jiffies;
2177 2178
	fmode_t fmode;
	bool truncate;
2179
	int status;
L
Linus Torvalds 已提交
2180

2181
	nfs_fattr_init(fattr);
L
Linus Torvalds 已提交
2182

2183 2184 2185 2186 2187 2188 2189
	/* Servers should only apply open mode checks for file size changes */
	truncate = (sattr->ia_valid & ATTR_SIZE) ? true : false;
	fmode = truncate ? FMODE_WRITE : FMODE_READ;

	if (nfs4_copy_delegation_stateid(&arg.stateid, inode, fmode)) {
		/* Use that stateid */
	} else if (truncate && state != NULL && nfs4_valid_open_stateid(state)) {
2190 2191 2192 2193
		struct nfs_lockowner lockowner = {
			.l_owner = current->files,
			.l_pid = current->tgid,
		};
2194
		nfs4_select_rw_stateid(&arg.stateid, state, FMODE_WRITE,
2195
				&lockowner);
2196
	} else
2197
		nfs4_stateid_copy(&arg.stateid, &zero_stateid);
L
Linus Torvalds 已提交
2198

2199
	status = nfs4_call_sync(server->client, server, &msg, &arg.seq_args, &res.seq_res, 1);
2200 2201
	if (status == 0 && state != NULL)
		renew_lease(server, timestamp);
2202
	return status;
L
Linus Torvalds 已提交
2203 2204
}

2205 2206 2207
static int nfs4_do_setattr(struct inode *inode, struct rpc_cred *cred,
			   struct nfs_fattr *fattr, struct iattr *sattr,
			   struct nfs4_state *state)
L
Linus Torvalds 已提交
2208
{
2209
	struct nfs_server *server = NFS_SERVER(inode);
2210 2211
	struct nfs4_exception exception = {
		.state = state,
2212
		.inode = inode,
2213
	};
L
Linus Torvalds 已提交
2214 2215
	int err;
	do {
2216 2217 2218
		err = _nfs4_do_setattr(inode, cred, fattr, sattr, state);
		switch (err) {
		case -NFS4ERR_OPENMODE:
2219 2220 2221 2222 2223 2224 2225
			if (!(sattr->ia_valid & ATTR_SIZE)) {
				pr_warn_once("NFSv4: server %s is incorrectly "
						"applying open mode checks to "
						"a SETATTR that is not "
						"changing file size.\n",
						server->nfs_client->cl_hostname);
			}
2226 2227 2228 2229 2230 2231 2232 2233
			if (state && !(state->state & FMODE_WRITE)) {
				err = -EBADF;
				if (sattr->ia_valid & ATTR_OPEN)
					err = -EACCES;
				goto out;
			}
		}
		err = nfs4_handle_exception(server, err, &exception);
L
Linus Torvalds 已提交
2234
	} while (exception.retry);
2235
out:
L
Linus Torvalds 已提交
2236 2237 2238 2239 2240 2241 2242 2243
	return err;
}

struct nfs4_closedata {
	struct inode *inode;
	struct nfs4_state *state;
	struct nfs_closeargs arg;
	struct nfs_closeres res;
2244
	struct nfs_fattr fattr;
2245
	unsigned long timestamp;
F
Fred Isaman 已提交
2246 2247
	bool roc;
	u32 roc_barrier;
L
Linus Torvalds 已提交
2248 2249
};

2250
static void nfs4_free_closedata(void *data)
2251
{
2252 2253
	struct nfs4_closedata *calldata = data;
	struct nfs4_state_owner *sp = calldata->state->owner;
2254
	struct super_block *sb = calldata->state->inode->i_sb;
2255

F
Fred Isaman 已提交
2256 2257
	if (calldata->roc)
		pnfs_roc_release(calldata->state->inode);
2258 2259 2260
	nfs4_put_open_state(calldata->state);
	nfs_free_seqid(calldata->arg.seqid);
	nfs4_put_state_owner(sp);
2261
	nfs_sb_deactive(sb);
2262 2263 2264
	kfree(calldata);
}

2265 2266 2267 2268
static void nfs4_close_clear_stateid_flags(struct nfs4_state *state,
		fmode_t fmode)
{
	spin_lock(&state->owner->so_lock);
2269 2270 2271
	clear_bit(NFS_O_RDWR_STATE, &state->flags);
	switch (fmode & (FMODE_READ|FMODE_WRITE)) {
	case FMODE_WRITE:
2272
		clear_bit(NFS_O_RDONLY_STATE, &state->flags);
2273 2274
		break;
	case FMODE_READ:
2275
		clear_bit(NFS_O_WRONLY_STATE, &state->flags);
2276 2277 2278 2279 2280 2281
		break;
	case 0:
		clear_bit(NFS_O_RDONLY_STATE, &state->flags);
		clear_bit(NFS_O_WRONLY_STATE, &state->flags);
		clear_bit(NFS_OPEN_STATE, &state->flags);
	}
2282 2283 2284
	spin_unlock(&state->owner->so_lock);
}

2285
static void nfs4_close_done(struct rpc_task *task, void *data)
L
Linus Torvalds 已提交
2286
{
2287
	struct nfs4_closedata *calldata = data;
L
Linus Torvalds 已提交
2288 2289 2290
	struct nfs4_state *state = calldata->state;
	struct nfs_server *server = NFS_SERVER(calldata->inode);

2291
	dprintk("%s: begin!\n", __func__);
2292 2293
	if (!nfs4_sequence_done(task, &calldata->res.seq_res))
		return;
L
Linus Torvalds 已提交
2294 2295 2296 2297 2298
        /* hmm. we are done with the inode, and in the process of freeing
	 * the state_owner. we keep this around to process errors
	 */
	switch (task->tk_status) {
		case 0:
F
Fred Isaman 已提交
2299 2300 2301
			if (calldata->roc)
				pnfs_roc_set_barrier(state->inode,
						     calldata->roc_barrier);
2302
			nfs_set_open_stateid(state, &calldata->res.stateid, 0);
2303
			renew_lease(server, calldata->timestamp);
2304 2305
			nfs4_close_clear_stateid_flags(state,
					calldata->arg.fmode);
L
Linus Torvalds 已提交
2306 2307
			break;
		case -NFS4ERR_STALE_STATEID:
2308 2309
		case -NFS4ERR_OLD_STATEID:
		case -NFS4ERR_BAD_STATEID:
L
Linus Torvalds 已提交
2310
		case -NFS4ERR_EXPIRED:
2311
			if (calldata->arg.fmode == 0)
2312
				break;
L
Linus Torvalds 已提交
2313
		default:
2314 2315
			if (nfs4_async_handle_error(task, server, state) == -EAGAIN)
				rpc_restart_call_prepare(task);
L
Linus Torvalds 已提交
2316
	}
2317
	nfs_release_seqid(calldata->arg.seqid);
2318
	nfs_refresh_inode(calldata->inode, calldata->res.fattr);
2319
	dprintk("%s: done, ret = %d!\n", __func__, task->tk_status);
L
Linus Torvalds 已提交
2320 2321
}

T
Trond Myklebust 已提交
2322
static void nfs4_close_prepare(struct rpc_task *task, void *data)
L
Linus Torvalds 已提交
2323
{
T
Trond Myklebust 已提交
2324
	struct nfs4_closedata *calldata = data;
2325
	struct nfs4_state *state = calldata->state;
2326
	struct inode *inode = calldata->inode;
2327
	int call_close = 0;
2328

2329
	dprintk("%s: begin!\n", __func__);
2330
	if (nfs_wait_on_sequence(calldata->arg.seqid, task) != 0)
2331
		goto out_wait;
2332

2333 2334
	task->tk_msg.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_OPEN_DOWNGRADE];
	calldata->arg.fmode = FMODE_READ|FMODE_WRITE;
2335
	spin_lock(&state->owner->so_lock);
2336
	/* Calculate the change in open mode */
2337
	if (state->n_rdwr == 0) {
2338
		if (state->n_rdonly == 0) {
2339 2340 2341
			call_close |= test_bit(NFS_O_RDONLY_STATE, &state->flags);
			call_close |= test_bit(NFS_O_RDWR_STATE, &state->flags);
			calldata->arg.fmode &= ~FMODE_READ;
2342 2343
		}
		if (state->n_wronly == 0) {
2344 2345 2346
			call_close |= test_bit(NFS_O_WRONLY_STATE, &state->flags);
			call_close |= test_bit(NFS_O_RDWR_STATE, &state->flags);
			calldata->arg.fmode &= ~FMODE_WRITE;
2347
		}
2348
	}
2349 2350
	if (!nfs4_valid_open_stateid(state))
		call_close = 0;
2351
	spin_unlock(&state->owner->so_lock);
2352 2353

	if (!call_close) {
2354
		/* Note: exit _without_ calling nfs4_close_done */
2355
		goto out_no_action;
2356
	}
2357

F
Fred Isaman 已提交
2358
	if (calldata->arg.fmode == 0) {
2359
		task->tk_msg.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_CLOSE];
F
Fred Isaman 已提交
2360
		if (calldata->roc &&
2361 2362
		    pnfs_roc_drain(inode, &calldata->roc_barrier, task)) {
			nfs_release_seqid(calldata->arg.seqid);
2363
			goto out_wait;
2364
		    }
F
Fred Isaman 已提交
2365
	}
2366

2367
	nfs_fattr_init(calldata->res.fattr);
2368
	calldata->timestamp = jiffies;
2369
	if (nfs4_setup_sequence(NFS_SERVER(inode),
2370 2371
				&calldata->arg.seq_args,
				&calldata->res.seq_res,
2372 2373
				task) != 0)
		nfs_release_seqid(calldata->arg.seqid);
2374
	dprintk("%s: done!\n", __func__);
2375 2376 2377 2378 2379
	return;
out_no_action:
	task->tk_action = NULL;
out_wait:
	nfs4_sequence_done(task, &calldata->res.seq_res);
L
Linus Torvalds 已提交
2380 2381
}

2382
static const struct rpc_call_ops nfs4_close_ops = {
T
Trond Myklebust 已提交
2383
	.rpc_call_prepare = nfs4_close_prepare,
2384 2385 2386 2387
	.rpc_call_done = nfs4_close_done,
	.rpc_release = nfs4_free_closedata,
};

L
Linus Torvalds 已提交
2388 2389 2390 2391 2392 2393 2394 2395 2396 2397 2398
/* 
 * 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!
 */
2399
int nfs4_do_close(struct nfs4_state *state, gfp_t gfp_mask, int wait)
L
Linus Torvalds 已提交
2400
{
2401
	struct nfs_server *server = NFS_SERVER(state->inode);
L
Linus Torvalds 已提交
2402
	struct nfs4_closedata *calldata;
2403 2404
	struct nfs4_state_owner *sp = state->owner;
	struct rpc_task *task;
2405 2406 2407 2408
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_CLOSE],
		.rpc_cred = state->owner->so_cred,
	};
T
Trond Myklebust 已提交
2409 2410
	struct rpc_task_setup task_setup_data = {
		.rpc_client = server->client,
2411
		.rpc_message = &msg,
T
Trond Myklebust 已提交
2412
		.callback_ops = &nfs4_close_ops,
2413
		.workqueue = nfsiod_workqueue,
T
Trond Myklebust 已提交
2414 2415
		.flags = RPC_TASK_ASYNC,
	};
2416
	int status = -ENOMEM;
L
Linus Torvalds 已提交
2417

2418
	calldata = kzalloc(sizeof(*calldata), gfp_mask);
L
Linus Torvalds 已提交
2419
	if (calldata == NULL)
2420
		goto out;
2421
	nfs41_init_sequence(&calldata->arg.seq_args, &calldata->res.seq_res, 1);
2422
	calldata->inode = state->inode;
L
Linus Torvalds 已提交
2423
	calldata->state = state;
2424
	calldata->arg.fh = NFS_FH(state->inode);
2425
	calldata->arg.stateid = &state->open_stateid;
L
Linus Torvalds 已提交
2426
	/* Serialization for the sequence id */
2427
	calldata->arg.seqid = nfs_alloc_seqid(&state->owner->so_seqid, gfp_mask);
2428 2429
	if (calldata->arg.seqid == NULL)
		goto out_free_calldata;
2430
	calldata->arg.fmode = 0;
2431
	calldata->arg.bitmask = server->cache_consistency_bitmask;
2432
	calldata->res.fattr = &calldata->fattr;
2433
	calldata->res.seqid = calldata->arg.seqid;
2434
	calldata->res.server = server;
2435
	calldata->roc = pnfs_roc(state->inode);
2436
	nfs_sb_active(calldata->inode->i_sb);
2437

2438 2439
	msg.rpc_argp = &calldata->arg;
	msg.rpc_resp = &calldata->res;
T
Trond Myklebust 已提交
2440 2441
	task_setup_data.callback_data = calldata;
	task = rpc_run_task(&task_setup_data);
2442 2443
	if (IS_ERR(task))
		return PTR_ERR(task);
2444 2445 2446
	status = 0;
	if (wait)
		status = rpc_wait_for_completion_task(task);
2447
	rpc_put_task(task);
2448
	return status;
2449 2450 2451
out_free_calldata:
	kfree(calldata);
out:
2452 2453
	nfs4_put_open_state(state);
	nfs4_put_state_owner(sp);
2454
	return status;
L
Linus Torvalds 已提交
2455 2456
}

2457
static struct inode *
2458
nfs4_atomic_open(struct inode *dir, struct nfs_open_context *ctx, int open_flags, struct iattr *attr)
L
Linus Torvalds 已提交
2459 2460 2461
{
	struct nfs4_state *state;

T
Trond Myklebust 已提交
2462
	/* Protect against concurrent sillydeletes */
2463
	state = nfs4_do_open(dir, ctx, open_flags, attr);
2464 2465
	if (IS_ERR(state))
		return ERR_CAST(state);
2466
	return state->inode;
L
Linus Torvalds 已提交
2467 2468
}

2469
static void nfs4_close_context(struct nfs_open_context *ctx, int is_sync)
T
Trond Myklebust 已提交
2470 2471 2472 2473
{
	if (ctx->state == NULL)
		return;
	if (is_sync)
2474
		nfs4_close_sync(ctx->state, ctx->mode);
T
Trond Myklebust 已提交
2475
	else
2476
		nfs4_close_state(ctx->state, ctx->mode);
T
Trond Myklebust 已提交
2477
}
L
Linus Torvalds 已提交
2478 2479 2480

static int _nfs4_server_capabilities(struct nfs_server *server, struct nfs_fh *fhandle)
{
B
Benny Halevy 已提交
2481 2482 2483
	struct nfs4_server_caps_arg args = {
		.fhandle = fhandle,
	};
L
Linus Torvalds 已提交
2484 2485 2486
	struct nfs4_server_caps_res res = {};
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_SERVER_CAPS],
B
Benny Halevy 已提交
2487
		.rpc_argp = &args,
L
Linus Torvalds 已提交
2488 2489 2490 2491
		.rpc_resp = &res,
	};
	int status;

2492
	status = nfs4_call_sync(server->client, server, &msg, &args.seq_args, &res.seq_res, 0);
L
Linus Torvalds 已提交
2493 2494
	if (status == 0) {
		memcpy(server->attr_bitmask, res.attr_bitmask, sizeof(server->attr_bitmask));
2495 2496 2497 2498 2499
		server->caps &= ~(NFS_CAP_ACLS|NFS_CAP_HARDLINKS|
				NFS_CAP_SYMLINKS|NFS_CAP_FILEID|
				NFS_CAP_MODE|NFS_CAP_NLINK|NFS_CAP_OWNER|
				NFS_CAP_OWNER_GROUP|NFS_CAP_ATIME|
				NFS_CAP_CTIME|NFS_CAP_MTIME);
L
Linus Torvalds 已提交
2500 2501 2502 2503 2504 2505
		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;
2506 2507 2508 2509 2510 2511 2512 2513 2514 2515 2516 2517 2518 2519 2520 2521 2522
		if (res.attr_bitmask[0] & FATTR4_WORD0_FILEID)
			server->caps |= NFS_CAP_FILEID;
		if (res.attr_bitmask[1] & FATTR4_WORD1_MODE)
			server->caps |= NFS_CAP_MODE;
		if (res.attr_bitmask[1] & FATTR4_WORD1_NUMLINKS)
			server->caps |= NFS_CAP_NLINK;
		if (res.attr_bitmask[1] & FATTR4_WORD1_OWNER)
			server->caps |= NFS_CAP_OWNER;
		if (res.attr_bitmask[1] & FATTR4_WORD1_OWNER_GROUP)
			server->caps |= NFS_CAP_OWNER_GROUP;
		if (res.attr_bitmask[1] & FATTR4_WORD1_TIME_ACCESS)
			server->caps |= NFS_CAP_ATIME;
		if (res.attr_bitmask[1] & FATTR4_WORD1_TIME_METADATA)
			server->caps |= NFS_CAP_CTIME;
		if (res.attr_bitmask[1] & FATTR4_WORD1_TIME_MODIFY)
			server->caps |= NFS_CAP_MTIME;

2523 2524 2525
		memcpy(server->cache_consistency_bitmask, res.attr_bitmask, sizeof(server->cache_consistency_bitmask));
		server->cache_consistency_bitmask[0] &= FATTR4_WORD0_CHANGE|FATTR4_WORD0_SIZE;
		server->cache_consistency_bitmask[1] &= FATTR4_WORD1_TIME_METADATA|FATTR4_WORD1_TIME_MODIFY;
L
Linus Torvalds 已提交
2526
		server->acl_bitmask = res.acl_bitmask;
2527
		server->fh_expire_type = res.fh_expire_type;
L
Linus Torvalds 已提交
2528
	}
A
Andy Adamson 已提交
2529

L
Linus Torvalds 已提交
2530 2531 2532
	return status;
}

2533
int nfs4_server_capabilities(struct nfs_server *server, struct nfs_fh *fhandle)
L
Linus Torvalds 已提交
2534 2535 2536 2537 2538 2539 2540 2541 2542 2543 2544 2545 2546 2547 2548 2549 2550 2551 2552
{
	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,
2553
		.fattr = info->fattr,
L
Linus Torvalds 已提交
2554 2555 2556 2557 2558 2559 2560
		.fh = fhandle,
	};
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_LOOKUP_ROOT],
		.rpc_argp = &args,
		.rpc_resp = &res,
	};
2561

2562
	nfs_fattr_init(info->fattr);
2563
	return nfs4_call_sync(server->client, server, &msg, &args.seq_args, &res.seq_res, 0);
L
Linus Torvalds 已提交
2564 2565 2566 2567 2568 2569 2570 2571
}

static int nfs4_lookup_root(struct nfs_server *server, struct nfs_fh *fhandle,
		struct nfs_fsinfo *info)
{
	struct nfs4_exception exception = { };
	int err;
	do {
2572 2573 2574 2575
		err = _nfs4_lookup_root(server, fhandle, info);
		switch (err) {
		case 0:
		case -NFS4ERR_WRONGSEC:
2576
			goto out;
2577 2578 2579
		default:
			err = nfs4_handle_exception(server, err, &exception);
		}
L
Linus Torvalds 已提交
2580
	} while (exception.retry);
2581
out:
L
Linus Torvalds 已提交
2582 2583 2584
	return err;
}

2585 2586 2587 2588 2589 2590 2591
static int nfs4_lookup_root_sec(struct nfs_server *server, struct nfs_fh *fhandle,
				struct nfs_fsinfo *info, rpc_authflavor_t flavor)
{
	struct rpc_auth *auth;
	int ret;

	auth = rpcauth_create(flavor, server->client);
2592
	if (IS_ERR(auth)) {
2593
		ret = -EACCES;
2594 2595 2596 2597 2598 2599 2600
		goto out;
	}
	ret = nfs4_lookup_root(server, fhandle, info);
out:
	return ret;
}

2601 2602 2603 2604 2605 2606 2607 2608 2609
/*
 * Retry pseudoroot lookup with various security flavors.  We do this when:
 *
 *   NFSv4.0: the PUTROOTFH operation returns NFS4ERR_WRONGSEC
 *   NFSv4.1: the server does not support the SECINFO_NO_NAME operation
 *
 * Returns zero on success, or a negative NFS4ERR value, or a
 * negative errno value.
 */
2610
static int nfs4_find_root_sec(struct nfs_server *server, struct nfs_fh *fhandle,
2611
			      struct nfs_fsinfo *info)
L
Linus Torvalds 已提交
2612
{
2613 2614 2615 2616 2617
	/* Per 3530bis 15.33.5 */
	static const rpc_authflavor_t flav_array[] = {
		RPC_AUTH_GSS_KRB5P,
		RPC_AUTH_GSS_KRB5I,
		RPC_AUTH_GSS_KRB5,
2618
		RPC_AUTH_UNIX,			/* courtesy */
2619 2620 2621 2622
		RPC_AUTH_NULL,
	};
	int status = -EPERM;
	size_t i;
C
Chuck Lever 已提交
2623

2624
	for (i = 0; i < ARRAY_SIZE(flav_array); i++) {
2625
		status = nfs4_lookup_root_sec(server, fhandle, info, flav_array[i]);
2626
		if (status == -NFS4ERR_WRONGSEC || status == -EACCES)
2627 2628
			continue;
		break;
2629
	}
2630

2631 2632 2633 2634 2635 2636 2637 2638 2639
	/*
	 * -EACCESS could mean that the user doesn't have correct permissions
	 * to access the mount.  It could also mean that we tried to mount
	 * with a gss auth flavor, but rpc.gssd isn't running.  Either way,
	 * existing mount programs don't handle -EACCES very well so it should
	 * be mapped to -EPERM instead.
	 */
	if (status == -EACCES)
		status = -EPERM;
2640 2641 2642
	return status;
}

C
Chuck Lever 已提交
2643 2644 2645 2646 2647 2648 2649 2650 2651 2652 2653 2654 2655 2656
static int nfs4_do_find_root_sec(struct nfs_server *server,
		struct nfs_fh *fhandle, struct nfs_fsinfo *info)
{
	int mv = server->nfs_client->cl_minorversion;
	return nfs_v4_minor_ops[mv]->find_root_sec(server, fhandle, info);
}

/**
 * nfs4_proc_get_rootfh - get file handle for server's pseudoroot
 * @server: initialized nfs_server handle
 * @fhandle: we fill in the pseudo-fs root file handle
 * @info: we fill in an FSINFO struct
 *
 * Returns zero on success, or a negative errno.
2657
 */
2658 2659
int nfs4_proc_get_rootfh(struct nfs_server *server, struct nfs_fh *fhandle,
			 struct nfs_fsinfo *info)
2660
{
C
Chuck Lever 已提交
2661 2662 2663 2664 2665 2666 2667
	int status;

	status = nfs4_lookup_root(server, fhandle, info);
	if ((status == -NFS4ERR_WRONGSEC) &&
	    !(server->flags & NFS_MOUNT_SECFLAVOUR))
		status = nfs4_do_find_root_sec(server, fhandle, info);

L
Linus Torvalds 已提交
2668 2669 2670 2671
	if (status == 0)
		status = nfs4_server_capabilities(server, fhandle);
	if (status == 0)
		status = nfs4_do_fsinfo(server, fhandle, info);
C
Chuck Lever 已提交
2672

2673
	return nfs4_map_errors(status);
L
Linus Torvalds 已提交
2674 2675
}

2676 2677 2678 2679 2680 2681 2682 2683 2684 2685 2686 2687 2688 2689 2690 2691 2692 2693 2694 2695 2696 2697 2698 2699 2700
static int nfs4_proc_get_root(struct nfs_server *server, struct nfs_fh *mntfh,
			      struct nfs_fsinfo *info)
{
	int error;
	struct nfs_fattr *fattr = info->fattr;

	error = nfs4_server_capabilities(server, mntfh);
	if (error < 0) {
		dprintk("nfs4_get_root: getcaps error = %d\n", -error);
		return error;
	}

	error = nfs4_proc_getattr(server, mntfh, fattr);
	if (error < 0) {
		dprintk("nfs4_get_root: getattr error = %d\n", -error);
		return error;
	}

	if (fattr->valid & NFS_ATTR_FATTR_FSID &&
	    !nfs_fsid_equal(&server->fsid, &fattr->fsid))
		memcpy(&server->fsid, &fattr->fsid, sizeof(server->fsid));

	return error;
}

M
Manoj Naik 已提交
2701 2702 2703 2704 2705
/*
 * 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
 */
2706 2707 2708
static int nfs4_get_referral(struct rpc_clnt *client, struct inode *dir,
			     const struct qstr *name, struct nfs_fattr *fattr,
			     struct nfs_fh *fhandle)
M
Manoj Naik 已提交
2709 2710 2711 2712 2713 2714 2715 2716 2717 2718 2719 2720
{
	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;

2721
	status = nfs4_proc_fs_locations(client, dir, name, locations, page);
M
Manoj Naik 已提交
2722 2723 2724 2725
	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)) {
2726 2727
		dprintk("%s: server did not return a different fsid for"
			" a referral at %s\n", __func__, name->name);
M
Manoj Naik 已提交
2728 2729 2730
		status = -EIO;
		goto out;
	}
2731 2732
	/* Fixup attributes for the nfs_lookup() call to nfs_fhget() */
	nfs_fixup_referral_attributes(&locations->fattr);
M
Manoj Naik 已提交
2733

2734
	/* replace the lookup nfs_fattr with the locations nfs_fattr */
M
Manoj Naik 已提交
2735 2736 2737 2738 2739
	memcpy(fattr, &locations->fattr, sizeof(struct nfs_fattr));
	memset(fhandle, 0, sizeof(struct nfs_fh));
out:
	if (page)
		__free_page(page);
2740
	kfree(locations);
M
Manoj Naik 已提交
2741 2742 2743
	return status;
}

L
Linus Torvalds 已提交
2744 2745 2746 2747 2748 2749 2750 2751 2752 2753 2754 2755 2756 2757 2758 2759
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,
	};
	
2760
	nfs_fattr_init(fattr);
2761
	return nfs4_call_sync(server->client, server, &msg, &args.seq_args, &res.seq_res, 0);
L
Linus Torvalds 已提交
2762 2763 2764 2765 2766 2767 2768 2769 2770 2771 2772 2773 2774 2775 2776 2777 2778 2779 2780 2781 2782 2783 2784 2785 2786 2787 2788 2789 2790 2791 2792 2793 2794 2795 2796
}

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)
{
2797
	struct inode *inode = dentry->d_inode;
2798
	struct rpc_cred *cred = NULL;
2799
	struct nfs4_state *state = NULL;
L
Linus Torvalds 已提交
2800 2801
	int status;

B
Benny Halevy 已提交
2802
	if (pnfs_ld_layoutret_on_setattr(inode))
2803
		pnfs_commit_and_return_layout(inode);
B
Benny Halevy 已提交
2804

2805
	nfs_fattr_init(fattr);
L
Linus Torvalds 已提交
2806
	
2807 2808 2809 2810 2811 2812 2813 2814
	/* Deal with open(O_TRUNC) */
	if (sattr->ia_valid & ATTR_OPEN)
		sattr->ia_valid &= ~(ATTR_MTIME|ATTR_CTIME|ATTR_OPEN);

	/* Optimization: if the end result is no change, don't RPC */
	if ((sattr->ia_valid & ~(ATTR_FILE)) == 0)
		return 0;

2815
	/* Search for an existing open(O_WRITE) file */
2816 2817 2818 2819
	if (sattr->ia_valid & ATTR_FILE) {
		struct nfs_open_context *ctx;

		ctx = nfs_file_open_context(sattr->ia_file);
N
Neil Brown 已提交
2820 2821 2822 2823
		if (ctx) {
			cred = ctx->cred;
			state = ctx->state;
		}
2824
	}
2825

2826
	status = nfs4_do_setattr(inode, cred, fattr, sattr, state);
2827 2828
	if (status == 0)
		nfs_setattr_update_inode(inode, sattr);
L
Linus Torvalds 已提交
2829 2830 2831
	return status;
}

2832 2833 2834
static int _nfs4_proc_lookup(struct rpc_clnt *clnt, struct inode *dir,
		const struct qstr *name, struct nfs_fh *fhandle,
		struct nfs_fattr *fattr)
D
David Howells 已提交
2835
{
2836
	struct nfs_server *server = NFS_SERVER(dir);
D
David Howells 已提交
2837 2838 2839
	int		       status;
	struct nfs4_lookup_arg args = {
		.bitmask = server->attr_bitmask,
2840
		.dir_fh = NFS_FH(dir),
D
David Howells 已提交
2841 2842 2843 2844 2845 2846 2847 2848 2849 2850 2851 2852 2853 2854 2855
		.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);

L
Linus Torvalds 已提交
2856
	dprintk("NFS call  lookup %s\n", name->name);
2857
	status = nfs4_call_sync(clnt, server, &msg, &args.seq_args, &res.seq_res, 0);
L
Linus Torvalds 已提交
2858 2859 2860 2861
	dprintk("NFS reply lookup: %d\n", status);
	return status;
}

2862
static void nfs_fixup_secinfo_attributes(struct nfs_fattr *fattr)
2863 2864
{
	fattr->valid |= NFS_ATTR_FATTR_TYPE | NFS_ATTR_FATTR_MODE |
2865
		NFS_ATTR_FATTR_NLINK | NFS_ATTR_FATTR_MOUNTPOINT;
2866 2867 2868 2869
	fattr->mode = S_IFDIR | S_IRUGO | S_IXUGO;
	fattr->nlink = 2;
}

2870 2871 2872
static int nfs4_proc_lookup_common(struct rpc_clnt **clnt, struct inode *dir,
				   struct qstr *name, struct nfs_fh *fhandle,
				   struct nfs_fattr *fattr)
L
Linus Torvalds 已提交
2873 2874
{
	struct nfs4_exception exception = { };
2875
	struct rpc_clnt *client = *clnt;
L
Linus Torvalds 已提交
2876 2877
	int err;
	do {
2878 2879
		err = _nfs4_proc_lookup(client, dir, name, fhandle, fattr);
		switch (err) {
2880
		case -NFS4ERR_BADNAME:
2881 2882
			err = -ENOENT;
			goto out;
2883
		case -NFS4ERR_MOVED:
2884
			err = nfs4_get_referral(client, dir, name, fattr, fhandle);
2885
			goto out;
2886
		case -NFS4ERR_WRONGSEC:
2887 2888 2889 2890 2891 2892 2893 2894 2895 2896 2897 2898
			err = -EPERM;
			if (client != *clnt)
				goto out;

			client = nfs4_create_sec_client(client, dir, name);
			if (IS_ERR(client))
				return PTR_ERR(client);

			exception.retry = 1;
			break;
		default:
			err = nfs4_handle_exception(NFS_SERVER(dir), err, &exception);
2899
		}
L
Linus Torvalds 已提交
2900
	} while (exception.retry);
2901 2902 2903 2904 2905 2906 2907

out:
	if (err == 0)
		*clnt = client;
	else if (client != *clnt)
		rpc_shutdown_client(client);

L
Linus Torvalds 已提交
2908 2909 2910
	return err;
}

2911
static int nfs4_proc_lookup(struct inode *dir, struct qstr *name,
2912 2913 2914 2915 2916 2917 2918 2919 2920 2921 2922 2923 2924
			    struct nfs_fh *fhandle, struct nfs_fattr *fattr)
{
	int status;
	struct rpc_clnt *client = NFS_CLIENT(dir);

	status = nfs4_proc_lookup_common(&client, dir, name, fhandle, fattr);
	if (client != NFS_CLIENT(dir)) {
		rpc_shutdown_client(client);
		nfs_fixup_secinfo_attributes(fattr);
	}
	return status;
}

2925 2926 2927 2928 2929 2930 2931 2932 2933 2934 2935 2936 2937 2938 2939
struct rpc_clnt *
nfs4_proc_lookup_mountpoint(struct inode *dir, struct qstr *name,
			    struct nfs_fh *fhandle, struct nfs_fattr *fattr)
{
	int status;
	struct rpc_clnt *client = rpc_clone_client(NFS_CLIENT(dir));

	status = nfs4_proc_lookup_common(&client, dir, name, fhandle, fattr);
	if (status < 0) {
		rpc_shutdown_client(client);
		return ERR_PTR(status);
	}
	return client;
}

L
Linus Torvalds 已提交
2940 2941
static int _nfs4_proc_access(struct inode *inode, struct nfs_access_entry *entry)
{
2942
	struct nfs_server *server = NFS_SERVER(inode);
L
Linus Torvalds 已提交
2943 2944
	struct nfs4_accessargs args = {
		.fh = NFS_FH(inode),
2945
		.bitmask = server->cache_consistency_bitmask,
2946 2947 2948
	};
	struct nfs4_accessres res = {
		.server = server,
L
Linus Torvalds 已提交
2949 2950 2951 2952 2953 2954 2955 2956 2957 2958 2959 2960 2961 2962 2963 2964 2965 2966 2967 2968 2969 2970 2971 2972 2973 2974
	};
	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;
	}
2975 2976 2977 2978 2979

	res.fattr = nfs_alloc_fattr();
	if (res.fattr == NULL)
		return -ENOMEM;

2980
	status = nfs4_call_sync(server->client, server, &msg, &args.seq_args, &res.seq_res, 0);
L
Linus Torvalds 已提交
2981
	if (!status) {
2982
		nfs_access_set_mask(entry, res.access);
2983
		nfs_refresh_inode(inode, res.fattr);
L
Linus Torvalds 已提交
2984
	}
2985
	nfs_free_fattr(res.fattr);
L
Linus Torvalds 已提交
2986 2987 2988 2989 2990 2991 2992 2993 2994 2995 2996 2997 2998 2999 3000 3001 3002 3003 3004 3005 3006 3007 3008 3009 3010 3011 3012 3013 3014 3015 3016 3017
	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
3018
 * we get the post-operation mtime and size.
L
Linus Torvalds 已提交
3019 3020 3021 3022 3023 3024 3025 3026 3027 3028 3029 3030 3031
 *
 * 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,
	};
B
Benny Halevy 已提交
3032
	struct nfs4_readlink_res res;
L
Linus Torvalds 已提交
3033 3034 3035
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_READLINK],
		.rpc_argp = &args,
B
Benny Halevy 已提交
3036
		.rpc_resp = &res,
L
Linus Torvalds 已提交
3037 3038
	};

3039
	return nfs4_call_sync(NFS_SERVER(inode)->client, NFS_SERVER(inode), &msg, &args.seq_args, &res.seq_res, 0);
L
Linus Torvalds 已提交
3040 3041 3042 3043 3044 3045 3046 3047 3048 3049 3050 3051 3052 3053 3054 3055
}

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;
}

/*
3056
 * This is just for mknod.  open(O_CREAT) will always do ->open_context().
L
Linus Torvalds 已提交
3057 3058 3059
 */
static int
nfs4_proc_create(struct inode *dir, struct dentry *dentry, struct iattr *sattr,
3060
		 int flags)
L
Linus Torvalds 已提交
3061
{
3062
	struct nfs_open_context *ctx;
L
Linus Torvalds 已提交
3063 3064 3065
	struct nfs4_state *state;
	int status = 0;

3066 3067 3068 3069
	ctx = alloc_nfs_open_context(dentry, FMODE_READ);
	if (IS_ERR(ctx))
		return PTR_ERR(ctx);

A
Aneesh Kumar K.V 已提交
3070
	sattr->ia_mode &= ~current_umask();
3071
	state = nfs4_do_open(dir, ctx, flags, sattr);
L
Linus Torvalds 已提交
3072 3073
	if (IS_ERR(state)) {
		status = PTR_ERR(state);
3074
		goto out;
L
Linus Torvalds 已提交
3075 3076
	}
out:
3077
	put_nfs_open_context(ctx);
L
Linus Torvalds 已提交
3078 3079 3080 3081 3082
	return status;
}

static int _nfs4_proc_remove(struct inode *dir, struct qstr *name)
{
3083
	struct nfs_server *server = NFS_SERVER(dir);
3084
	struct nfs_removeargs args = {
L
Linus Torvalds 已提交
3085
		.fh = NFS_FH(dir),
3086
		.name = *name,
3087
	};
3088
	struct nfs_removeres res = {
3089
		.server = server,
L
Linus Torvalds 已提交
3090 3091
	};
	struct rpc_message msg = {
3092 3093 3094
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_REMOVE],
		.rpc_argp = &args,
		.rpc_resp = &res,
L
Linus Torvalds 已提交
3095
	};
3096
	int status;
L
Linus Torvalds 已提交
3097

3098
	status = nfs4_call_sync(server->client, server, &msg, &args.seq_args, &res.seq_res, 1);
3099
	if (status == 0)
3100
		update_changeattr(dir, &res.cinfo);
L
Linus Torvalds 已提交
3101 3102 3103 3104 3105 3106 3107 3108 3109 3110 3111 3112 3113 3114 3115
	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;
}

3116
static void nfs4_proc_unlink_setup(struct rpc_message *msg, struct inode *dir)
L
Linus Torvalds 已提交
3117
{
3118 3119 3120
	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
3121

3122
	res->server = server;
L
Linus Torvalds 已提交
3123
	msg->rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_REMOVE];
3124
	nfs41_init_sequence(&args->seq_args, &res->seq_res, 1);
L
Linus Torvalds 已提交
3125 3126
}

3127 3128
static void nfs4_proc_unlink_rpc_prepare(struct rpc_task *task, struct nfs_unlinkdata *data)
{
3129 3130 3131 3132
	nfs4_setup_sequence(NFS_SERVER(data->dir),
			&data->args.seq_args,
			&data->res.seq_res,
			task);
L
Linus Torvalds 已提交
3133 3134
}

3135
static int nfs4_proc_unlink_done(struct rpc_task *task, struct inode *dir)
L
Linus Torvalds 已提交
3136
{
3137 3138
	struct nfs_removeres *res = task->tk_msg.rpc_resp;

3139 3140
	if (!nfs4_sequence_done(task, &res->seq_res))
		return 0;
3141
	if (nfs4_async_handle_error(task, res->server, NULL) == -EAGAIN)
3142 3143 3144
		return 0;
	update_changeattr(dir, &res->cinfo);
	return 1;
L
Linus Torvalds 已提交
3145 3146
}

3147 3148 3149 3150 3151 3152 3153 3154
static void nfs4_proc_rename_setup(struct rpc_message *msg, struct inode *dir)
{
	struct nfs_server *server = NFS_SERVER(dir);
	struct nfs_renameargs *arg = msg->rpc_argp;
	struct nfs_renameres *res = msg->rpc_resp;

	msg->rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_RENAME];
	res->server = server;
3155
	nfs41_init_sequence(&arg->seq_args, &res->seq_res, 1);
3156 3157
}

3158 3159
static void nfs4_proc_rename_rpc_prepare(struct rpc_task *task, struct nfs_renamedata *data)
{
3160 3161 3162 3163
	nfs4_setup_sequence(NFS_SERVER(data->old_dir),
			&data->args.seq_args,
			&data->res.seq_res,
			task);
3164 3165 3166 3167 3168 3169 3170 3171 3172 3173 3174 3175 3176 3177 3178 3179 3180
}

static int nfs4_proc_rename_done(struct rpc_task *task, struct inode *old_dir,
				 struct inode *new_dir)
{
	struct nfs_renameres *res = task->tk_msg.rpc_resp;

	if (!nfs4_sequence_done(task, &res->seq_res))
		return 0;
	if (nfs4_async_handle_error(task, res->server, NULL) == -EAGAIN)
		return 0;

	update_changeattr(old_dir, &res->old_cinfo);
	update_changeattr(new_dir, &res->new_cinfo);
	return 1;
}

L
Linus Torvalds 已提交
3181 3182 3183
static int _nfs4_proc_rename(struct inode *old_dir, struct qstr *old_name,
		struct inode *new_dir, struct qstr *new_name)
{
3184
	struct nfs_server *server = NFS_SERVER(old_dir);
3185
	struct nfs_renameargs arg = {
L
Linus Torvalds 已提交
3186 3187 3188 3189
		.old_dir = NFS_FH(old_dir),
		.new_dir = NFS_FH(new_dir),
		.old_name = old_name,
		.new_name = new_name,
3190
	};
3191
	struct nfs_renameres res = {
3192
		.server = server,
L
Linus Torvalds 已提交
3193 3194 3195 3196 3197 3198
	};
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_RENAME],
		.rpc_argp = &arg,
		.rpc_resp = &res,
	};
3199
	int status = -ENOMEM;
L
Linus Torvalds 已提交
3200
	
3201
	status = nfs4_call_sync(server->client, server, &msg, &arg.seq_args, &res.seq_res, 1);
L
Linus Torvalds 已提交
3202 3203 3204 3205 3206 3207 3208 3209 3210 3211 3212 3213 3214 3215 3216 3217 3218 3219 3220 3221 3222 3223 3224
	if (!status) {
		update_changeattr(old_dir, &res.old_cinfo);
		update_changeattr(new_dir, &res.new_cinfo);
	}
	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)
{
3225
	struct nfs_server *server = NFS_SERVER(inode);
L
Linus Torvalds 已提交
3226 3227 3228 3229
	struct nfs4_link_arg arg = {
		.fh     = NFS_FH(inode),
		.dir_fh = NFS_FH(dir),
		.name   = name,
3230 3231 3232 3233
		.bitmask = server->attr_bitmask,
	};
	struct nfs4_link_res res = {
		.server = server,
L
Linus Torvalds 已提交
3234 3235 3236 3237
	};
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_LINK],
		.rpc_argp = &arg,
3238
		.rpc_resp = &res,
L
Linus Torvalds 已提交
3239
	};
3240 3241 3242
	int status = -ENOMEM;

	res.fattr = nfs_alloc_fattr();
3243
	if (res.fattr == NULL)
3244
		goto out;
L
Linus Torvalds 已提交
3245

3246
	status = nfs4_call_sync(server->client, server, &msg, &arg.seq_args, &res.seq_res, 1);
3247 3248
	if (!status) {
		update_changeattr(dir, &res.cinfo);
3249
		nfs_post_op_update_inode(inode, res.fattr);
3250
	}
3251 3252
out:
	nfs_free_fattr(res.fattr);
L
Linus Torvalds 已提交
3253 3254 3255 3256 3257 3258 3259 3260 3261 3262 3263 3264 3265 3266 3267
	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;
}

3268 3269 3270 3271 3272 3273 3274 3275 3276 3277 3278 3279 3280 3281 3282 3283 3284 3285 3286 3287 3288 3289 3290 3291 3292 3293 3294 3295 3296 3297 3298 3299 3300 3301 3302 3303
struct nfs4_createdata {
	struct rpc_message msg;
	struct nfs4_create_arg arg;
	struct nfs4_create_res res;
	struct nfs_fh fh;
	struct nfs_fattr fattr;
};

static struct nfs4_createdata *nfs4_alloc_createdata(struct inode *dir,
		struct qstr *name, struct iattr *sattr, u32 ftype)
{
	struct nfs4_createdata *data;

	data = kzalloc(sizeof(*data), GFP_KERNEL);
	if (data != NULL) {
		struct nfs_server *server = NFS_SERVER(dir);

		data->msg.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_CREATE];
		data->msg.rpc_argp = &data->arg;
		data->msg.rpc_resp = &data->res;
		data->arg.dir_fh = NFS_FH(dir);
		data->arg.server = server;
		data->arg.name = name;
		data->arg.attrs = sattr;
		data->arg.ftype = ftype;
		data->arg.bitmask = server->attr_bitmask;
		data->res.server = server;
		data->res.fh = &data->fh;
		data->res.fattr = &data->fattr;
		nfs_fattr_init(data->res.fattr);
	}
	return data;
}

static int nfs4_do_create(struct inode *dir, struct dentry *dentry, struct nfs4_createdata *data)
{
3304
	int status = nfs4_call_sync(NFS_SERVER(dir)->client, NFS_SERVER(dir), &data->msg,
3305
				    &data->arg.seq_args, &data->res.seq_res, 1);
3306 3307 3308 3309 3310 3311 3312 3313 3314 3315 3316 3317
	if (status == 0) {
		update_changeattr(dir, &data->res.dir_cinfo);
		status = nfs_instantiate(dentry, data->res.fh, data->res.fattr);
	}
	return status;
}

static void nfs4_free_createdata(struct nfs4_createdata *data)
{
	kfree(data);
}

3318
static int _nfs4_proc_symlink(struct inode *dir, struct dentry *dentry,
3319
		struct page *page, unsigned int len, struct iattr *sattr)
L
Linus Torvalds 已提交
3320
{
3321 3322
	struct nfs4_createdata *data;
	int status = -ENAMETOOLONG;
L
Linus Torvalds 已提交
3323

3324
	if (len > NFS4_MAXPATHLEN)
3325
		goto out;
3326

3327 3328 3329 3330 3331 3332 3333 3334
	status = -ENOMEM;
	data = nfs4_alloc_createdata(dir, &dentry->d_name, sattr, NF4LNK);
	if (data == NULL)
		goto out;

	data->msg.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_SYMLINK];
	data->arg.u.symlink.pages = &page;
	data->arg.u.symlink.len = len;
L
Linus Torvalds 已提交
3335
	
3336 3337 3338 3339
	status = nfs4_do_create(dir, dentry, data);

	nfs4_free_createdata(data);
out:
L
Linus Torvalds 已提交
3340 3341 3342
	return status;
}

3343
static int nfs4_proc_symlink(struct inode *dir, struct dentry *dentry,
3344
		struct page *page, unsigned int len, struct iattr *sattr)
L
Linus Torvalds 已提交
3345 3346 3347 3348 3349
{
	struct nfs4_exception exception = { };
	int err;
	do {
		err = nfs4_handle_exception(NFS_SERVER(dir),
3350 3351
				_nfs4_proc_symlink(dir, dentry, page,
							len, sattr),
L
Linus Torvalds 已提交
3352 3353 3354 3355 3356 3357 3358 3359
				&exception);
	} while (exception.retry);
	return err;
}

static int _nfs4_proc_mkdir(struct inode *dir, struct dentry *dentry,
		struct iattr *sattr)
{
3360 3361
	struct nfs4_createdata *data;
	int status = -ENOMEM;
L
Linus Torvalds 已提交
3362

3363 3364 3365 3366 3367 3368 3369 3370
	data = nfs4_alloc_createdata(dir, &dentry->d_name, sattr, NF4DIR);
	if (data == NULL)
		goto out;

	status = nfs4_do_create(dir, dentry, data);

	nfs4_free_createdata(data);
out:
L
Linus Torvalds 已提交
3371 3372 3373 3374 3375 3376 3377 3378
	return status;
}

static int nfs4_proc_mkdir(struct inode *dir, struct dentry *dentry,
		struct iattr *sattr)
{
	struct nfs4_exception exception = { };
	int err;
A
Aneesh Kumar K.V 已提交
3379 3380

	sattr->ia_mode &= ~current_umask();
L
Linus Torvalds 已提交
3381 3382 3383 3384 3385 3386 3387 3388 3389
	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,
B
Bryan Schumaker 已提交
3390
		u64 cookie, struct page **pages, unsigned int count, int plus)
L
Linus Torvalds 已提交
3391 3392 3393 3394
{
	struct inode		*dir = dentry->d_inode;
	struct nfs4_readdir_arg args = {
		.fh = NFS_FH(dir),
B
Bryan Schumaker 已提交
3395
		.pages = pages,
L
Linus Torvalds 已提交
3396 3397
		.pgbase = 0,
		.count = count,
3398
		.bitmask = NFS_SERVER(dentry->d_inode)->attr_bitmask,
B
Bryan Schumaker 已提交
3399
		.plus = plus,
L
Linus Torvalds 已提交
3400 3401 3402 3403 3404 3405 3406 3407 3408 3409
	};
	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;

3410
	dprintk("%s: dentry = %s/%s, cookie = %Lu\n", __func__,
3411 3412 3413
			dentry->d_parent->d_name.name,
			dentry->d_name.name,
			(unsigned long long)cookie);
3414
	nfs4_setup_readdir(cookie, NFS_I(dir)->cookieverf, dentry, &args);
L
Linus Torvalds 已提交
3415
	res.pgbase = args.pgbase;
3416
	status = nfs4_call_sync(NFS_SERVER(dir)->client, NFS_SERVER(dir), &msg, &args.seq_args, &res.seq_res, 0);
3417
	if (status >= 0) {
3418
		memcpy(NFS_I(dir)->cookieverf, res.verifier.data, NFS4_VERIFIER_SIZE);
3419 3420
		status += args.pgbase;
	}
3421 3422 3423

	nfs_invalidate_atime(dir);

3424
	dprintk("%s: returns %d\n", __func__, status);
L
Linus Torvalds 已提交
3425 3426 3427 3428
	return status;
}

static int nfs4_proc_readdir(struct dentry *dentry, struct rpc_cred *cred,
B
Bryan Schumaker 已提交
3429
		u64 cookie, struct page **pages, unsigned int count, int plus)
L
Linus Torvalds 已提交
3430 3431 3432 3433 3434 3435
{
	struct nfs4_exception exception = { };
	int err;
	do {
		err = nfs4_handle_exception(NFS_SERVER(dentry->d_inode),
				_nfs4_proc_readdir(dentry, cred, cookie,
B
Bryan Schumaker 已提交
3436
					pages, count, plus),
L
Linus Torvalds 已提交
3437 3438 3439 3440 3441 3442 3443 3444
				&exception);
	} while (exception.retry);
	return err;
}

static int _nfs4_proc_mknod(struct inode *dir, struct dentry *dentry,
		struct iattr *sattr, dev_t rdev)
{
3445 3446 3447
	struct nfs4_createdata *data;
	int mode = sattr->ia_mode;
	int status = -ENOMEM;
L
Linus Torvalds 已提交
3448

3449 3450 3451 3452
	data = nfs4_alloc_createdata(dir, &dentry->d_name, sattr, NF4SOCK);
	if (data == NULL)
		goto out;

L
Linus Torvalds 已提交
3453
	if (S_ISFIFO(mode))
3454
		data->arg.ftype = NF4FIFO;
L
Linus Torvalds 已提交
3455
	else if (S_ISBLK(mode)) {
3456 3457 3458
		data->arg.ftype = NF4BLK;
		data->arg.u.device.specdata1 = MAJOR(rdev);
		data->arg.u.device.specdata2 = MINOR(rdev);
L
Linus Torvalds 已提交
3459 3460
	}
	else if (S_ISCHR(mode)) {
3461 3462 3463
		data->arg.ftype = NF4CHR;
		data->arg.u.device.specdata1 = MAJOR(rdev);
		data->arg.u.device.specdata2 = MINOR(rdev);
3464 3465 3466
	} else if (!S_ISSOCK(mode)) {
		status = -EINVAL;
		goto out_free;
L
Linus Torvalds 已提交
3467 3468
	}
	
3469
	status = nfs4_do_create(dir, dentry, data);
3470
out_free:
3471 3472
	nfs4_free_createdata(data);
out:
L
Linus Torvalds 已提交
3473 3474 3475 3476 3477 3478 3479 3480
	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;
A
Aneesh Kumar K.V 已提交
3481 3482

	sattr->ia_mode &= ~current_umask();
L
Linus Torvalds 已提交
3483 3484 3485 3486 3487 3488 3489 3490 3491 3492 3493 3494 3495 3496 3497
	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,
	};
B
Benny Halevy 已提交
3498 3499 3500
	struct nfs4_statfs_res res = {
		.fsstat = fsstat,
	};
L
Linus Torvalds 已提交
3501 3502 3503
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_STATFS],
		.rpc_argp = &args,
B
Benny Halevy 已提交
3504
		.rpc_resp = &res,
L
Linus Torvalds 已提交
3505 3506
	};

3507
	nfs_fattr_init(fsstat->fattr);
3508
	return  nfs4_call_sync(server->client, server, &msg, &args.seq_args, &res.seq_res, 0);
L
Linus Torvalds 已提交
3509 3510 3511 3512 3513 3514 3515 3516 3517 3518 3519 3520 3521 3522 3523 3524 3525 3526 3527 3528 3529
}

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,
	};
B
Benny Halevy 已提交
3530 3531 3532
	struct nfs4_fsinfo_res res = {
		.fsinfo = fsinfo,
	};
L
Linus Torvalds 已提交
3533 3534 3535
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_FSINFO],
		.rpc_argp = &args,
B
Benny Halevy 已提交
3536
		.rpc_resp = &res,
L
Linus Torvalds 已提交
3537 3538
	};

3539
	return nfs4_call_sync(server->client, server, &msg, &args.seq_args, &res.seq_res, 0);
L
Linus Torvalds 已提交
3540 3541 3542 3543 3544
}

static int nfs4_do_fsinfo(struct nfs_server *server, struct nfs_fh *fhandle, struct nfs_fsinfo *fsinfo)
{
	struct nfs4_exception exception = { };
3545
	unsigned long now = jiffies;
L
Linus Torvalds 已提交
3546 3547 3548
	int err;

	do {
3549 3550 3551 3552 3553 3554 3555 3556 3557 3558 3559
		err = _nfs4_do_fsinfo(server, fhandle, fsinfo);
		if (err == 0) {
			struct nfs_client *clp = server->nfs_client;

			spin_lock(&clp->cl_lock);
			clp->cl_lease_time = fsinfo->lease_time * HZ;
			clp->cl_last_renewal = now;
			spin_unlock(&clp->cl_lock);
			break;
		}
		err = nfs4_handle_exception(server, err, &exception);
L
Linus Torvalds 已提交
3560 3561 3562 3563 3564 3565
	} while (exception.retry);
	return err;
}

static int nfs4_proc_fsinfo(struct nfs_server *server, struct nfs_fh *fhandle, struct nfs_fsinfo *fsinfo)
{
3566 3567
	int error;

3568
	nfs_fattr_init(fsinfo->fattr);
3569
	error = nfs4_do_fsinfo(server, fhandle, fsinfo);
3570 3571 3572
	if (error == 0) {
		/* block layout checks this! */
		server->pnfs_blksize = fsinfo->blksize;
3573
		set_pnfs_layoutdriver(server, fhandle, fsinfo->layouttype);
3574
	}
3575 3576

	return error;
L
Linus Torvalds 已提交
3577 3578 3579 3580 3581 3582 3583 3584 3585
}

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,
	};
B
Benny Halevy 已提交
3586 3587 3588
	struct nfs4_pathconf_res res = {
		.pathconf = pathconf,
	};
L
Linus Torvalds 已提交
3589 3590 3591
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_PATHCONF],
		.rpc_argp = &args,
B
Benny Halevy 已提交
3592
		.rpc_resp = &res,
L
Linus Torvalds 已提交
3593 3594 3595 3596 3597 3598 3599 3600
	};

	/* 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;
	}

3601
	nfs_fattr_init(pathconf->fattr);
3602
	return nfs4_call_sync(server->client, server, &msg, &args.seq_args, &res.seq_res, 0);
L
Linus Torvalds 已提交
3603 3604 3605 3606 3607 3608 3609 3610 3611 3612 3613 3614 3615 3616 3617 3618
}

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;
}

3619
int nfs4_set_rw_stateid(nfs4_stateid *stateid,
3620 3621 3622 3623 3624 3625 3626 3627
		const struct nfs_open_context *ctx,
		const struct nfs_lock_context *l_ctx,
		fmode_t fmode)
{
	const struct nfs_lockowner *lockowner = NULL;

	if (l_ctx != NULL)
		lockowner = &l_ctx->lockowner;
3628
	return nfs4_select_rw_stateid(stateid, ctx->state, fmode, lockowner);
3629 3630 3631
}
EXPORT_SYMBOL_GPL(nfs4_set_rw_stateid);

3632 3633 3634 3635 3636 3637 3638 3639 3640 3641 3642 3643 3644 3645 3646 3647 3648 3649 3650 3651 3652 3653 3654 3655 3656 3657 3658
static bool nfs4_stateid_is_current(nfs4_stateid *stateid,
		const struct nfs_open_context *ctx,
		const struct nfs_lock_context *l_ctx,
		fmode_t fmode)
{
	nfs4_stateid current_stateid;

	if (nfs4_set_rw_stateid(&current_stateid, ctx, l_ctx, fmode))
		return false;
	return nfs4_stateid_match(stateid, &current_stateid);
}

static bool nfs4_error_stateid_expired(int err)
{
	switch (err) {
	case -NFS4ERR_DELEG_REVOKED:
	case -NFS4ERR_ADMIN_REVOKED:
	case -NFS4ERR_BAD_STATEID:
	case -NFS4ERR_STALE_STATEID:
	case -NFS4ERR_OLD_STATEID:
	case -NFS4ERR_OPENMODE:
	case -NFS4ERR_EXPIRED:
		return true;
	}
	return false;
}

3659 3660
void __nfs4_read_done_cb(struct nfs_read_data *data)
{
3661
	nfs_invalidate_atime(data->header->inode);
3662 3663
}

3664
static int nfs4_read_done_cb(struct rpc_task *task, struct nfs_read_data *data)
L
Linus Torvalds 已提交
3665
{
3666
	struct nfs_server *server = NFS_SERVER(data->header->inode);
L
Linus Torvalds 已提交
3667

3668
	if (nfs4_async_handle_error(task, server, data->args.context->state) == -EAGAIN) {
3669
		rpc_restart_call_prepare(task);
T
Trond Myklebust 已提交
3670
		return -EAGAIN;
L
Linus Torvalds 已提交
3671
	}
3672

3673
	__nfs4_read_done_cb(data);
L
Linus Torvalds 已提交
3674
	if (task->tk_status > 0)
T
Trond Myklebust 已提交
3675 3676
		renew_lease(server, data->timestamp);
	return 0;
L
Linus Torvalds 已提交
3677 3678
}

3679 3680 3681 3682 3683 3684 3685 3686 3687 3688 3689 3690 3691 3692
static bool nfs4_read_stateid_changed(struct rpc_task *task,
		struct nfs_readargs *args)
{

	if (!nfs4_error_stateid_expired(task->tk_status) ||
		nfs4_stateid_is_current(&args->stateid,
				args->context,
				args->lock_context,
				FMODE_READ))
		return false;
	rpc_restart_call_prepare(task);
	return true;
}

3693 3694 3695 3696 3697 3698 3699
static int nfs4_read_done(struct rpc_task *task, struct nfs_read_data *data)
{

	dprintk("--> %s\n", __func__);

	if (!nfs4_sequence_done(task, &data->res.seq_res))
		return -EAGAIN;
3700 3701
	if (nfs4_read_stateid_changed(task, &data->args))
		return -EAGAIN;
3702 3703
	return data->read_done_cb ? data->read_done_cb(task, data) :
				    nfs4_read_done_cb(task, data);
3704 3705
}

3706
static void nfs4_proc_read_setup(struct nfs_read_data *data, struct rpc_message *msg)
L
Linus Torvalds 已提交
3707 3708
{
	data->timestamp   = jiffies;
3709
	data->read_done_cb = nfs4_read_done_cb;
3710
	msg->rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_READ];
3711
	nfs41_init_sequence(&data->args.seq_args, &data->res.seq_res, 0);
L
Linus Torvalds 已提交
3712 3713
}

3714 3715
static void nfs4_proc_read_rpc_prepare(struct rpc_task *task, struct nfs_read_data *data)
{
3716
	if (nfs4_setup_sequence(NFS_SERVER(data->header->inode),
3717 3718
			&data->args.seq_args,
			&data->res.seq_res,
3719 3720 3721 3722
			task))
		return;
	nfs4_set_rw_stateid(&data->args.stateid, data->args.context,
			data->args.lock_context, FMODE_READ);
L
Linus Torvalds 已提交
3723 3724
}

3725
static int nfs4_write_done_cb(struct rpc_task *task, struct nfs_write_data *data)
L
Linus Torvalds 已提交
3726
{
3727
	struct inode *inode = data->header->inode;
L
Linus Torvalds 已提交
3728
	
3729
	if (nfs4_async_handle_error(task, NFS_SERVER(inode), data->args.context->state) == -EAGAIN) {
3730
		rpc_restart_call_prepare(task);
3731
		return -EAGAIN;
L
Linus Torvalds 已提交
3732
	}
3733
	if (task->tk_status >= 0) {
L
Linus Torvalds 已提交
3734
		renew_lease(NFS_SERVER(inode), data->timestamp);
3735
		nfs_post_op_update_inode_force_wcc(inode, &data->fattr);
3736
	}
3737
	return 0;
L
Linus Torvalds 已提交
3738 3739
}

3740 3741 3742 3743 3744 3745 3746 3747 3748 3749 3750 3751 3752 3753
static bool nfs4_write_stateid_changed(struct rpc_task *task,
		struct nfs_writeargs *args)
{

	if (!nfs4_error_stateid_expired(task->tk_status) ||
		nfs4_stateid_is_current(&args->stateid,
				args->context,
				args->lock_context,
				FMODE_WRITE))
		return false;
	rpc_restart_call_prepare(task);
	return true;
}

3754 3755 3756 3757
static int nfs4_write_done(struct rpc_task *task, struct nfs_write_data *data)
{
	if (!nfs4_sequence_done(task, &data->res.seq_res))
		return -EAGAIN;
3758 3759
	if (nfs4_write_stateid_changed(task, &data->args))
		return -EAGAIN;
3760 3761
	return data->write_done_cb ? data->write_done_cb(task, data) :
		nfs4_write_done_cb(task, data);
3762 3763
}

3764 3765
static
bool nfs4_write_need_cache_consistency_data(const struct nfs_write_data *data)
3766
{
3767 3768 3769 3770 3771 3772 3773 3774
	const struct nfs_pgio_header *hdr = data->header;

	/* Don't request attributes for pNFS or O_DIRECT writes */
	if (data->ds_clp != NULL || hdr->dreq != NULL)
		return false;
	/* Otherwise, request attributes if and only if we don't hold
	 * a delegation
	 */
3775
	return nfs4_have_delegation(hdr->inode, FMODE_READ) == 0;
3776 3777
}

3778
static void nfs4_proc_write_setup(struct nfs_write_data *data, struct rpc_message *msg)
L
Linus Torvalds 已提交
3779
{
3780
	struct nfs_server *server = NFS_SERVER(data->header->inode);
3781

3782
	if (!nfs4_write_need_cache_consistency_data(data)) {
3783 3784 3785 3786
		data->args.bitmask = NULL;
		data->res.fattr = NULL;
	} else
		data->args.bitmask = server->cache_consistency_bitmask;
3787

3788 3789
	if (!data->write_done_cb)
		data->write_done_cb = nfs4_write_done_cb;
3790
	data->res.server = server;
L
Linus Torvalds 已提交
3791 3792
	data->timestamp   = jiffies;

3793
	msg->rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_WRITE];
3794
	nfs41_init_sequence(&data->args.seq_args, &data->res.seq_res, 1);
L
Linus Torvalds 已提交
3795 3796
}

3797 3798
static void nfs4_proc_write_rpc_prepare(struct rpc_task *task, struct nfs_write_data *data)
{
3799
	if (nfs4_setup_sequence(NFS_SERVER(data->header->inode),
3800 3801
			&data->args.seq_args,
			&data->res.seq_res,
3802 3803 3804 3805
			task))
		return;
	nfs4_set_rw_stateid(&data->args.stateid, data->args.context,
			data->args.lock_context, FMODE_WRITE);
L
Linus Torvalds 已提交
3806 3807
}

3808
static void nfs4_proc_commit_rpc_prepare(struct rpc_task *task, struct nfs_commit_data *data)
3809
{
3810 3811 3812 3813
	nfs4_setup_sequence(NFS_SERVER(data->inode),
			&data->args.seq_args,
			&data->res.seq_res,
			task);
L
Linus Torvalds 已提交
3814 3815
}

3816
static int nfs4_commit_done_cb(struct rpc_task *task, struct nfs_commit_data *data)
L
Linus Torvalds 已提交
3817 3818
{
	struct inode *inode = data->inode;
3819

3820
	if (nfs4_async_handle_error(task, NFS_SERVER(inode), NULL) == -EAGAIN) {
3821
		rpc_restart_call_prepare(task);
3822
		return -EAGAIN;
L
Linus Torvalds 已提交
3823
	}
3824
	return 0;
L
Linus Torvalds 已提交
3825 3826
}

3827
static int nfs4_commit_done(struct rpc_task *task, struct nfs_commit_data *data)
3828 3829 3830
{
	if (!nfs4_sequence_done(task, &data->res.seq_res))
		return -EAGAIN;
3831
	return data->commit_done_cb(task, data);
3832 3833
}

3834
static void nfs4_proc_commit_setup(struct nfs_commit_data *data, struct rpc_message *msg)
L
Linus Torvalds 已提交
3835
{
3836
	struct nfs_server *server = NFS_SERVER(data->inode);
3837

3838 3839
	if (data->commit_done_cb == NULL)
		data->commit_done_cb = nfs4_commit_done_cb;
3840
	data->res.server = server;
3841
	msg->rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_COMMIT];
3842
	nfs41_init_sequence(&data->args.seq_args, &data->res.seq_res, 1);
L
Linus Torvalds 已提交
3843 3844
}

3845 3846 3847 3848 3849
struct nfs4_renewdata {
	struct nfs_client	*client;
	unsigned long		timestamp;
};

L
Linus Torvalds 已提交
3850 3851 3852 3853
/*
 * nfs4_proc_async_renew(): This is not one of the nfs_rpc_ops; it is a special
 * standalone procedure for queueing an asynchronous RENEW.
 */
3854
static void nfs4_renew_release(void *calldata)
3855
{
3856 3857
	struct nfs4_renewdata *data = calldata;
	struct nfs_client *clp = data->client;
3858

3859 3860 3861
	if (atomic_read(&clp->cl_count) > 1)
		nfs4_schedule_state_renewal(clp);
	nfs_put_client(clp);
3862
	kfree(data);
3863 3864
}

3865
static void nfs4_renew_done(struct rpc_task *task, void *calldata)
L
Linus Torvalds 已提交
3866
{
3867 3868 3869
	struct nfs4_renewdata *data = calldata;
	struct nfs_client *clp = data->client;
	unsigned long timestamp = data->timestamp;
L
Linus Torvalds 已提交
3870 3871

	if (task->tk_status < 0) {
3872
		/* Unless we're shutting down, schedule state recovery! */
3873 3874 3875
		if (test_bit(NFS_CS_RENEWD, &clp->cl_res_state) == 0)
			return;
		if (task->tk_status != NFS4ERR_CB_PATH_DOWN) {
3876
			nfs4_schedule_lease_recovery(clp);
3877 3878 3879
			return;
		}
		nfs4_schedule_path_down_recovery(clp);
L
Linus Torvalds 已提交
3880
	}
3881
	do_renew_lease(clp, timestamp);
L
Linus Torvalds 已提交
3882 3883
}

3884 3885
static const struct rpc_call_ops nfs4_renew_ops = {
	.rpc_call_done = nfs4_renew_done,
3886
	.rpc_release = nfs4_renew_release,
3887 3888
};

3889
static int nfs4_proc_async_renew(struct nfs_client *clp, struct rpc_cred *cred, unsigned renew_flags)
L
Linus Torvalds 已提交
3890 3891 3892 3893
{
	struct rpc_message msg = {
		.rpc_proc	= &nfs4_procedures[NFSPROC4_CLNT_RENEW],
		.rpc_argp	= clp,
3894
		.rpc_cred	= cred,
L
Linus Torvalds 已提交
3895
	};
3896
	struct nfs4_renewdata *data;
L
Linus Torvalds 已提交
3897

3898 3899
	if (renew_flags == 0)
		return 0;
3900 3901
	if (!atomic_inc_not_zero(&clp->cl_count))
		return -EIO;
3902
	data = kmalloc(sizeof(*data), GFP_NOFS);
3903 3904 3905 3906
	if (data == NULL)
		return -ENOMEM;
	data->client = clp;
	data->timestamp = jiffies;
3907
	return rpc_call_async(clp->cl_rpcclient, &msg, RPC_TASK_TIMEOUT,
3908
			&nfs4_renew_ops, data);
L
Linus Torvalds 已提交
3909 3910
}

3911
static int nfs4_proc_renew(struct nfs_client *clp, struct rpc_cred *cred)
L
Linus Torvalds 已提交
3912 3913 3914 3915
{
	struct rpc_message msg = {
		.rpc_proc	= &nfs4_procedures[NFSPROC4_CLNT_RENEW],
		.rpc_argp	= clp,
3916
		.rpc_cred	= cred,
L
Linus Torvalds 已提交
3917 3918 3919 3920
	};
	unsigned long now = jiffies;
	int status;

3921
	status = rpc_call_sync(clp->cl_rpcclient, &msg, RPC_TASK_TIMEOUT);
L
Linus Torvalds 已提交
3922 3923
	if (status < 0)
		return status;
3924
	do_renew_lease(clp, now);
L
Linus Torvalds 已提交
3925 3926 3927
	return 0;
}

3928 3929 3930 3931 3932 3933 3934
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);
}

3935 3936
/* Assuming that XATTR_SIZE_MAX is a multiple of PAGE_SIZE, and that
 * it's OK to put sizeof(void) * (XATTR_SIZE_MAX/PAGE_SIZE) bytes on
3937 3938
 * the stack.
 */
3939
#define NFS4ACL_MAXPAGES DIV_ROUND_UP(XATTR_SIZE_MAX, PAGE_SIZE)
3940

3941 3942 3943 3944 3945 3946 3947 3948 3949
static int buf_to_pages_noslab(const void *buf, size_t buflen,
		struct page **pages, unsigned int *pgbase)
{
	struct page *newpage, **spages;
	int rc = 0;
	size_t len;
	spages = pages;

	do {
3950
		len = min_t(size_t, PAGE_SIZE, buflen);
3951 3952 3953 3954 3955 3956 3957 3958 3959 3960 3961 3962 3963 3964 3965 3966 3967 3968 3969
		newpage = alloc_page(GFP_KERNEL);

		if (newpage == NULL)
			goto unwind;
		memcpy(page_address(newpage), buf, len);
                buf += len;
                buflen -= len;
		*pages++ = newpage;
		rc++;
	} while (buflen != 0);

	return rc;

unwind:
	for(; rc > 0; rc--)
		__free_page(spages[rc-1]);
	return -ENOMEM;
}

3970 3971 3972
struct nfs4_cached_acl {
	int cached;
	size_t len;
A
Andrew Morton 已提交
3973
	char data[0];
3974 3975 3976 3977 3978 3979 3980 3981 3982 3983 3984 3985 3986 3987 3988 3989 3990 3991 3992 3993 3994 3995 3996 3997 3998 3999 4000 4001 4002 4003 4004 4005 4006 4007 4008 4009 4010 4011 4012 4013 4014 4015
};

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;
}

4016
static void nfs4_write_cached_acl(struct inode *inode, struct page **pages, size_t pgbase, size_t acl_len)
4017 4018
{
	struct nfs4_cached_acl *acl;
4019
	size_t buflen = sizeof(*acl) + acl_len;
4020

4021
	if (buflen <= PAGE_SIZE) {
4022
		acl = kmalloc(buflen, GFP_KERNEL);
4023 4024 4025
		if (acl == NULL)
			goto out;
		acl->cached = 1;
4026
		_copy_from_pages(acl->data, pages, pgbase, acl_len);
4027 4028 4029 4030 4031 4032 4033 4034 4035 4036 4037
	} 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);
}

4038 4039 4040 4041 4042 4043 4044 4045 4046 4047
/*
 * The getxattr API returns the required buffer length when called with a
 * NULL buf. The NFSv4 acl tool then calls getxattr again after allocating
 * the required buf.  On a NULL buf, we send a page of data to the server
 * guessing that the ACL request can be serviced by a page. If so, we cache
 * up to the page of ACL data, and the 2nd call to getxattr is serviced by
 * the cache. If not so, we throw away the page, and cache the required
 * length. The next getxattr call will then produce another round trip to
 * the server, this time with the input buf of the required size.
 */
4048
static ssize_t __nfs4_get_acl_uncached(struct inode *inode, void *buf, size_t buflen)
4049
{
4050
	struct page *pages[NFS4ACL_MAXPAGES] = {NULL, };
4051 4052 4053 4054 4055
	struct nfs_getaclargs args = {
		.fh = NFS_FH(inode),
		.acl_pages = pages,
		.acl_len = buflen,
	};
B
Benny Halevy 已提交
4056 4057 4058
	struct nfs_getaclres res = {
		.acl_len = buflen,
	};
4059 4060 4061
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_GETACL],
		.rpc_argp = &args,
B
Benny Halevy 已提交
4062
		.rpc_resp = &res,
4063
	};
4064 4065
	unsigned int npages = DIV_ROUND_UP(buflen, PAGE_SIZE);
	int ret = -ENOMEM, i;
4066

4067 4068 4069 4070
	/* As long as we're doing a round trip to the server anyway,
	 * let's be prepared for a page of acl data. */
	if (npages == 0)
		npages = 1;
4071 4072
	if (npages > ARRAY_SIZE(pages))
		return -ERANGE;
4073

4074 4075 4076 4077
	for (i = 0; i < npages; i++) {
		pages[i] = alloc_page(GFP_KERNEL);
		if (!pages[i])
			goto out_free;
4078
	}
4079 4080 4081 4082 4083 4084

	/* for decoding across pages */
	res.acl_scratch = alloc_page(GFP_KERNEL);
	if (!res.acl_scratch)
		goto out_free;

4085 4086
	args.acl_len = npages * PAGE_SIZE;
	args.acl_pgbase = 0;
4087

4088
	dprintk("%s  buf %p buflen %zu npages %d args.acl_len %zu\n",
4089 4090 4091
		__func__, buf, buflen, npages, args.acl_len);
	ret = nfs4_call_sync(NFS_SERVER(inode)->client, NFS_SERVER(inode),
			     &msg, &args.seq_args, &res.seq_res, 0);
4092 4093
	if (ret)
		goto out_free;
4094

4095 4096 4097 4098 4099
	/* Handle the case where the passed-in buffer is too short */
	if (res.acl_flags & NFS4_ACL_TRUNC) {
		/* Did the user only issue a request for the acl length? */
		if (buf == NULL)
			goto out_ok;
4100
		ret = -ERANGE;
4101
		goto out_free;
4102
	}
4103
	nfs4_write_cached_acl(inode, pages, res.acl_data_offset, res.acl_len);
4104 4105 4106 4107 4108
	if (buf) {
		if (res.acl_len > buflen) {
			ret = -ERANGE;
			goto out_free;
		}
4109
		_copy_from_pages(buf, pages, res.acl_data_offset, res.acl_len);
4110
	}
4111 4112
out_ok:
	ret = res.acl_len;
4113
out_free:
4114 4115 4116
	for (i = 0; i < npages; i++)
		if (pages[i])
			__free_page(pages[i]);
4117 4118
	if (res.acl_scratch)
		__free_page(res.acl_scratch);
4119 4120 4121
	return ret;
}

4122 4123 4124 4125 4126 4127 4128 4129 4130 4131 4132 4133 4134
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;
}

4135 4136 4137 4138 4139 4140 4141 4142 4143 4144
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;
4145 4146
	if (NFS_I(inode)->cache_validity & NFS_INO_INVALID_ACL)
		nfs_zap_acl_cache(inode);
4147 4148
	ret = nfs4_read_cached_acl(inode, buf, buflen);
	if (ret != -ENOENT)
4149 4150
		/* -ENOENT is returned if there is no ACL or if there is an ACL
		 * but no cached acl data, just the acl length */
4151 4152 4153 4154
		return ret;
	return nfs4_get_acl_uncached(inode, buf, buflen);
}

4155
static int __nfs4_proc_set_acl(struct inode *inode, const void *buf, size_t buflen)
4156 4157 4158 4159 4160 4161 4162 4163
{
	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,
	};
B
Benny Halevy 已提交
4164
	struct nfs_setaclres res;
4165 4166 4167
	struct rpc_message msg = {
		.rpc_proc	= &nfs4_procedures[NFSPROC4_CLNT_SETACL],
		.rpc_argp	= &arg,
B
Benny Halevy 已提交
4168
		.rpc_resp	= &res,
4169
	};
4170
	unsigned int npages = DIV_ROUND_UP(buflen, PAGE_SIZE);
4171
	int ret, i;
4172 4173 4174

	if (!nfs4_server_supports_acls(server))
		return -EOPNOTSUPP;
4175 4176
	if (npages > ARRAY_SIZE(pages))
		return -ERANGE;
4177 4178 4179
	i = buf_to_pages_noslab(buf, buflen, arg.acl_pages, &arg.acl_pgbase);
	if (i < 0)
		return i;
4180
	nfs4_inode_return_delegation(inode);
4181
	ret = nfs4_call_sync(server->client, server, &msg, &arg.seq_args, &res.seq_res, 1);
4182 4183 4184 4185 4186 4187 4188 4189

	/*
	 * Free each page after tx, so the only ref left is
	 * held by the network stack
	 */
	for (; i > 0; i--)
		put_page(pages[i-1]);

4190 4191 4192 4193 4194 4195 4196
	/*
	 * Acl update can result in inode attribute update.
	 * so mark the attribute cache invalid.
	 */
	spin_lock(&inode->i_lock);
	NFS_I(inode)->cache_validity |= NFS_INO_INVALID_ATTR;
	spin_unlock(&inode->i_lock);
4197 4198
	nfs_access_zap_cache(inode);
	nfs_zap_acl_cache(inode);
4199 4200 4201
	return ret;
}

4202 4203 4204 4205 4206 4207 4208 4209 4210 4211 4212 4213
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
4214
static int
4215
nfs4_async_handle_error(struct rpc_task *task, const struct nfs_server *server, struct nfs4_state *state)
L
Linus Torvalds 已提交
4216
{
4217 4218 4219
	struct nfs_client *clp = server->nfs_client;

	if (task->tk_status >= 0)
L
Linus Torvalds 已提交
4220 4221
		return 0;
	switch(task->tk_status) {
4222
		case -NFS4ERR_DELEG_REVOKED:
4223 4224
		case -NFS4ERR_ADMIN_REVOKED:
		case -NFS4ERR_BAD_STATEID:
4225 4226 4227
			if (state == NULL)
				break;
			nfs_remove_bad_delegation(state->inode);
4228 4229 4230
		case -NFS4ERR_OPENMODE:
			if (state == NULL)
				break;
4231 4232
			if (nfs4_schedule_stateid_recovery(server, state) < 0)
				goto stateid_invalid;
4233
			goto wait_on_recovery;
4234
		case -NFS4ERR_EXPIRED:
4235 4236 4237 4238
			if (state != NULL) {
				if (nfs4_schedule_stateid_recovery(server, state) < 0)
					goto stateid_invalid;
			}
L
Linus Torvalds 已提交
4239
		case -NFS4ERR_STALE_STATEID:
4240
		case -NFS4ERR_STALE_CLIENTID:
4241 4242
			nfs4_schedule_lease_recovery(clp);
			goto wait_on_recovery;
4243 4244 4245 4246 4247 4248 4249 4250 4251 4252
#if defined(CONFIG_NFS_V4_1)
		case -NFS4ERR_BADSESSION:
		case -NFS4ERR_BADSLOT:
		case -NFS4ERR_BAD_HIGH_SLOT:
		case -NFS4ERR_DEADSESSION:
		case -NFS4ERR_CONN_NOT_BOUND_TO_SESSION:
		case -NFS4ERR_SEQ_FALSE_RETRY:
		case -NFS4ERR_SEQ_MISORDERED:
			dprintk("%s ERROR %d, Reset session\n", __func__,
				task->tk_status);
4253
			nfs4_schedule_session_recovery(clp->cl_session, task->tk_status);
4254 4255 4256
			task->tk_status = 0;
			return -EAGAIN;
#endif /* CONFIG_NFS_V4_1 */
L
Linus Torvalds 已提交
4257
		case -NFS4ERR_DELAY:
4258
			nfs_inc_server_stats(server, NFSIOS_DELAY);
4259
		case -NFS4ERR_GRACE:
L
Linus Torvalds 已提交
4260 4261 4262
			rpc_delay(task, NFS4_POLL_RETRY_MAX);
			task->tk_status = 0;
			return -EAGAIN;
4263
		case -NFS4ERR_RETRY_UNCACHED_REP:
L
Linus Torvalds 已提交
4264 4265 4266 4267 4268 4269
		case -NFS4ERR_OLD_STATEID:
			task->tk_status = 0;
			return -EAGAIN;
	}
	task->tk_status = nfs4_map_errors(task->tk_status);
	return 0;
4270 4271 4272
stateid_invalid:
	task->tk_status = -EIO;
	return 0;
4273
wait_on_recovery:
4274 4275 4276 4277 4278
	rpc_sleep_on(&clp->cl_rpcwaitq, task, NULL);
	if (test_bit(NFS4CLNT_MANAGER_RUNNING, &clp->cl_state) == 0)
		rpc_wake_up_queued_task(&clp->cl_rpcwaitq, task);
	task->tk_status = 0;
	return -EAGAIN;
L
Linus Torvalds 已提交
4279 4280
}

4281 4282
static void nfs4_init_boot_verifier(const struct nfs_client *clp,
				    nfs4_verifier *bootverf)
4283 4284 4285
{
	__be32 verf[2];

4286 4287 4288 4289 4290 4291
	if (test_bit(NFS4CLNT_PURGE_STATE, &clp->cl_state)) {
		/* An impossible timestamp guarantees this value
		 * will never match a generated boot time. */
		verf[0] = 0;
		verf[1] = (__be32)(NSEC_PER_SEC + 1);
	} else {
4292 4293 4294
		struct nfs_net *nn = net_generic(clp->cl_net, nfs_net_id);
		verf[0] = (__be32)nn->boot_time.tv_sec;
		verf[1] = (__be32)nn->boot_time.tv_nsec;
4295
	}
4296 4297 4298
	memcpy(bootverf->data, verf, sizeof(bootverf->data));
}

4299 4300 4301 4302 4303 4304 4305 4306 4307 4308 4309 4310 4311 4312 4313 4314 4315 4316 4317 4318 4319
static unsigned int
nfs4_init_nonuniform_client_string(const struct nfs_client *clp,
				   char *buf, size_t len)
{
	unsigned int result;

	rcu_read_lock();
	result = scnprintf(buf, len, "Linux NFSv4.0 %s/%s %s",
				clp->cl_ipaddr,
				rpc_peeraddr2str(clp->cl_rpcclient,
							RPC_DISPLAY_ADDR),
				rpc_peeraddr2str(clp->cl_rpcclient,
							RPC_DISPLAY_PROTO));
	rcu_read_unlock();
	return result;
}

static unsigned int
nfs4_init_uniform_client_string(const struct nfs_client *clp,
				char *buf, size_t len)
{
4320 4321 4322 4323
	char *nodename = clp->cl_rpcclient->cl_nodename;

	if (nfs4_client_id_uniquifier[0] != '\0')
		nodename = nfs4_client_id_uniquifier;
4324 4325
	return scnprintf(buf, len, "Linux NFSv%u.%u %s",
				clp->rpc_ops->version, clp->cl_minorversion,
4326
				nodename);
4327 4328
}

4329 4330 4331 4332 4333 4334 4335 4336 4337 4338
/**
 * nfs4_proc_setclientid - Negotiate client ID
 * @clp: state data structure
 * @program: RPC program for NFSv4 callback service
 * @port: IP port number for NFS4 callback service
 * @cred: RPC credential to use for this call
 * @res: where to place the result
 *
 * Returns zero, a negative errno, or a negative NFS4ERR status code.
 */
4339 4340 4341
int nfs4_proc_setclientid(struct nfs_client *clp, u32 program,
		unsigned short port, struct rpc_cred *cred,
		struct nfs4_setclientid_res *res)
L
Linus Torvalds 已提交
4342 4343 4344 4345 4346
{
	nfs4_verifier sc_verifier;
	struct nfs4_setclientid setclientid = {
		.sc_verifier = &sc_verifier,
		.sc_prog = program,
A
Andy Adamson 已提交
4347
		.sc_cb_ident = clp->cl_cb_ident,
L
Linus Torvalds 已提交
4348 4349 4350 4351
	};
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_SETCLIENTID],
		.rpc_argp = &setclientid,
4352
		.rpc_resp = res,
4353
		.rpc_cred = cred,
L
Linus Torvalds 已提交
4354
	};
4355
	int status;
L
Linus Torvalds 已提交
4356

4357
	/* nfs_client_id4 */
4358
	nfs4_init_boot_verifier(clp, &sc_verifier);
4359 4360 4361 4362 4363 4364 4365 4366 4367 4368
	if (test_bit(NFS_CS_MIGRATION, &clp->cl_flags))
		setclientid.sc_name_len =
				nfs4_init_uniform_client_string(clp,
						setclientid.sc_name,
						sizeof(setclientid.sc_name));
	else
		setclientid.sc_name_len =
				nfs4_init_nonuniform_client_string(clp,
						setclientid.sc_name,
						sizeof(setclientid.sc_name));
4369
	/* cb_client4 */
4370
	rcu_read_lock();
4371
	setclientid.sc_netid_len = scnprintf(setclientid.sc_netid,
4372
				sizeof(setclientid.sc_netid), "%s",
4373 4374
				rpc_peeraddr2str(clp->cl_rpcclient,
							RPC_DISPLAY_NETID));
4375 4376
	rcu_read_unlock();
	setclientid.sc_uaddr_len = scnprintf(setclientid.sc_uaddr,
4377
				sizeof(setclientid.sc_uaddr), "%s.%u.%u",
L
Linus Torvalds 已提交
4378 4379
				clp->cl_ipaddr, port >> 8, port & 255);

4380 4381 4382 4383 4384 4385
	dprintk("NFS call  setclientid auth=%s, '%.*s'\n",
		clp->cl_rpcclient->cl_auth->au_ops->au_name,
		setclientid.sc_name_len, setclientid.sc_name);
	status = rpc_call_sync(clp->cl_rpcclient, &msg, RPC_TASK_TIMEOUT);
	dprintk("NFS reply setclientid: %d\n", status);
	return status;
L
Linus Torvalds 已提交
4386 4387
}

4388 4389 4390 4391 4392 4393 4394 4395
/**
 * nfs4_proc_setclientid_confirm - Confirm client ID
 * @clp: state data structure
 * @res: result of a previous SETCLIENTID
 * @cred: RPC credential to use for this call
 *
 * Returns zero, a negative errno, or a negative NFS4ERR status code.
 */
4396
int nfs4_proc_setclientid_confirm(struct nfs_client *clp,
4397 4398
		struct nfs4_setclientid_res *arg,
		struct rpc_cred *cred)
L
Linus Torvalds 已提交
4399 4400 4401
{
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_SETCLIENTID_CONFIRM],
4402
		.rpc_argp = arg,
4403
		.rpc_cred = cred,
L
Linus Torvalds 已提交
4404 4405 4406
	};
	int status;

4407 4408 4409
	dprintk("NFS call  setclientid_confirm auth=%s, (client ID %llx)\n",
		clp->cl_rpcclient->cl_auth->au_ops->au_name,
		clp->cl_clientid);
4410
	status = rpc_call_sync(clp->cl_rpcclient, &msg, RPC_TASK_TIMEOUT);
4411
	dprintk("NFS reply setclientid_confirm: %d\n", status);
L
Linus Torvalds 已提交
4412 4413 4414
	return status;
}

4415 4416
struct nfs4_delegreturndata {
	struct nfs4_delegreturnargs args;
4417
	struct nfs4_delegreturnres res;
4418 4419
	struct nfs_fh fh;
	nfs4_stateid stateid;
4420
	unsigned long timestamp;
4421
	struct nfs_fattr fattr;
4422 4423 4424 4425 4426 4427
	int rpc_status;
};

static void nfs4_delegreturn_done(struct rpc_task *task, void *calldata)
{
	struct nfs4_delegreturndata *data = calldata;
4428

4429 4430
	if (!nfs4_sequence_done(task, &data->res.seq_res))
		return;
4431

4432 4433 4434 4435
	switch (task->tk_status) {
	case -NFS4ERR_STALE_STATEID:
	case -NFS4ERR_EXPIRED:
	case 0:
4436
		renew_lease(data->res.server, data->timestamp);
4437 4438 4439 4440
		break;
	default:
		if (nfs4_async_handle_error(task, data->res.server, NULL) ==
				-EAGAIN) {
4441
			rpc_restart_call_prepare(task);
4442 4443 4444 4445
			return;
		}
	}
	data->rpc_status = task->tk_status;
4446 4447 4448 4449 4450 4451 4452
}

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

4453 4454 4455 4456 4457 4458 4459
#if defined(CONFIG_NFS_V4_1)
static void nfs4_delegreturn_prepare(struct rpc_task *task, void *data)
{
	struct nfs4_delegreturndata *d_data;

	d_data = (struct nfs4_delegreturndata *)data;

4460 4461 4462 4463
	nfs4_setup_sequence(d_data->res.server,
			&d_data->args.seq_args,
			&d_data->res.seq_res,
			task);
4464 4465 4466
}
#endif /* CONFIG_NFS_V4_1 */

4467
static const struct rpc_call_ops nfs4_delegreturn_ops = {
4468 4469 4470
#if defined(CONFIG_NFS_V4_1)
	.rpc_call_prepare = nfs4_delegreturn_prepare,
#endif /* CONFIG_NFS_V4_1 */
4471 4472 4473 4474
	.rpc_call_done = nfs4_delegreturn_done,
	.rpc_release = nfs4_delegreturn_release,
};

4475
static int _nfs4_proc_delegreturn(struct inode *inode, struct rpc_cred *cred, const nfs4_stateid *stateid, int issync)
4476 4477
{
	struct nfs4_delegreturndata *data;
4478
	struct nfs_server *server = NFS_SERVER(inode);
4479
	struct rpc_task *task;
4480 4481 4482 4483
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_DELEGRETURN],
		.rpc_cred = cred,
	};
T
Trond Myklebust 已提交
4484 4485
	struct rpc_task_setup task_setup_data = {
		.rpc_client = server->client,
4486
		.rpc_message = &msg,
T
Trond Myklebust 已提交
4487 4488 4489
		.callback_ops = &nfs4_delegreturn_ops,
		.flags = RPC_TASK_ASYNC,
	};
4490
	int status = 0;
4491

4492
	data = kzalloc(sizeof(*data), GFP_NOFS);
4493 4494
	if (data == NULL)
		return -ENOMEM;
4495
	nfs41_init_sequence(&data->args.seq_args, &data->res.seq_res, 1);
4496 4497
	data->args.fhandle = &data->fh;
	data->args.stateid = &data->stateid;
4498
	data->args.bitmask = server->cache_consistency_bitmask;
4499
	nfs_copy_fh(&data->fh, NFS_FH(inode));
4500
	nfs4_stateid_copy(&data->stateid, stateid);
4501 4502
	data->res.fattr = &data->fattr;
	data->res.server = server;
4503
	nfs_fattr_init(data->res.fattr);
4504
	data->timestamp = jiffies;
4505 4506
	data->rpc_status = 0;

T
Trond Myklebust 已提交
4507
	task_setup_data.callback_data = data;
4508 4509
	msg.rpc_argp = &data->args;
	msg.rpc_resp = &data->res;
T
Trond Myklebust 已提交
4510
	task = rpc_run_task(&task_setup_data);
4511
	if (IS_ERR(task))
4512
		return PTR_ERR(task);
4513 4514
	if (!issync)
		goto out;
4515
	status = nfs4_wait_for_completion_rpc_task(task);
4516 4517 4518
	if (status != 0)
		goto out;
	status = data->rpc_status;
4519 4520 4521 4522
	if (status == 0)
		nfs_post_op_update_inode_force_wcc(inode, &data->fattr);
	else
		nfs_refresh_inode(inode, &data->fattr);
4523
out:
4524
	rpc_put_task(task);
4525
	return status;
L
Linus Torvalds 已提交
4526 4527
}

4528
int nfs4_proc_delegreturn(struct inode *inode, struct rpc_cred *cred, const nfs4_stateid *stateid, int issync)
L
Linus Torvalds 已提交
4529 4530 4531 4532 4533
{
	struct nfs_server *server = NFS_SERVER(inode);
	struct nfs4_exception exception = { };
	int err;
	do {
4534
		err = _nfs4_proc_delegreturn(inode, cred, stateid, issync);
L
Linus Torvalds 已提交
4535 4536 4537 4538 4539 4540 4541 4542 4543 4544 4545 4546 4547 4548 4549 4550 4551 4552 4553 4554
		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)
{
4555
	freezable_schedule_timeout_killable(timeout);
L
Linus Torvalds 已提交
4556 4557 4558 4559 4560 4561 4562 4563 4564 4565
	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);
4566
	struct nfs_client *clp = server->nfs_client;
T
Trond Myklebust 已提交
4567
	struct nfs_lockt_args arg = {
L
Linus Torvalds 已提交
4568
		.fh = NFS_FH(inode),
T
Trond Myklebust 已提交
4569
		.fl = request,
L
Linus Torvalds 已提交
4570
	};
T
Trond Myklebust 已提交
4571 4572
	struct nfs_lockt_res res = {
		.denied = request,
L
Linus Torvalds 已提交
4573 4574 4575 4576 4577 4578 4579 4580 4581 4582
	};
	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;

T
Trond Myklebust 已提交
4583
	arg.lock_owner.clientid = clp->cl_clientid;
4584 4585 4586 4587
	status = nfs4_set_lock_state(state, request);
	if (status != 0)
		goto out;
	lsp = request->fl_u.nfs4_fl.owner;
4588
	arg.lock_owner.id = lsp->ls_seqid.owner_id;
4589
	arg.lock_owner.s_dev = server->s_dev;
4590
	status = nfs4_call_sync(server->client, server, &msg, &arg.seq_args, &res.seq_res, 1);
T
Trond Myklebust 已提交
4591 4592 4593 4594 4595 4596
	switch (status) {
		case 0:
			request->fl_type = F_UNLCK;
			break;
		case -NFS4ERR_DENIED:
			status = 0;
L
Linus Torvalds 已提交
4597
	}
4598
	request->fl_ops->fl_release_private(request);
4599
out:
L
Linus Torvalds 已提交
4600 4601 4602 4603 4604 4605 4606 4607 4608 4609 4610 4611 4612 4613 4614 4615 4616 4617 4618 4619 4620 4621 4622 4623 4624 4625 4626 4627 4628 4629 4630 4631
	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;
}

4632
struct nfs4_unlockdata {
T
Trond Myklebust 已提交
4633 4634
	struct nfs_locku_args arg;
	struct nfs_locku_res res;
4635 4636
	struct nfs4_lock_state *lsp;
	struct nfs_open_context *ctx;
T
Trond Myklebust 已提交
4637 4638
	struct file_lock fl;
	const struct nfs_server *server;
4639
	unsigned long timestamp;
4640 4641
};

T
Trond Myklebust 已提交
4642 4643 4644 4645 4646 4647 4648 4649
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;

4650
	p = kzalloc(sizeof(*p), GFP_NOFS);
T
Trond Myklebust 已提交
4651 4652 4653 4654 4655
	if (p == NULL)
		return NULL;
	p->arg.fh = NFS_FH(inode);
	p->arg.fl = &p->fl;
	p->arg.seqid = seqid;
4656
	p->res.seqid = seqid;
T
Trond Myklebust 已提交
4657 4658 4659 4660 4661 4662 4663 4664 4665 4666
	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;
}

4667
static void nfs4_locku_release_calldata(void *data)
4668
{
4669
	struct nfs4_unlockdata *calldata = data;
T
Trond Myklebust 已提交
4670
	nfs_free_seqid(calldata->arg.seqid);
4671 4672 4673
	nfs4_put_lock_state(calldata->lsp);
	put_nfs_open_context(calldata->ctx);
	kfree(calldata);
4674 4675
}

4676
static void nfs4_locku_done(struct rpc_task *task, void *data)
4677
{
4678
	struct nfs4_unlockdata *calldata = data;
4679

4680 4681
	if (!nfs4_sequence_done(task, &calldata->res.seq_res))
		return;
4682 4683
	switch (task->tk_status) {
		case 0:
4684 4685
			nfs4_stateid_copy(&calldata->lsp->ls_stateid,
					&calldata->res.stateid);
4686
			renew_lease(calldata->server, calldata->timestamp);
4687
			break;
4688 4689
		case -NFS4ERR_BAD_STATEID:
		case -NFS4ERR_OLD_STATEID:
4690 4691 4692 4693
		case -NFS4ERR_STALE_STATEID:
		case -NFS4ERR_EXPIRED:
			break;
		default:
4694
			if (nfs4_async_handle_error(task, calldata->server, NULL) == -EAGAIN)
4695
				rpc_restart_call_prepare(task);
4696
	}
4697
	nfs_release_seqid(calldata->arg.seqid);
4698 4699
}

T
Trond Myklebust 已提交
4700
static void nfs4_locku_prepare(struct rpc_task *task, void *data)
4701
{
T
Trond Myklebust 已提交
4702
	struct nfs4_unlockdata *calldata = data;
4703

T
Trond Myklebust 已提交
4704
	if (nfs_wait_on_sequence(calldata->arg.seqid, task) != 0)
4705
		goto out_wait;
4706
	if (test_bit(NFS_LOCK_INITIALIZED, &calldata->lsp->ls_flags) == 0) {
4707
		/* Note: exit _without_ running nfs4_locku_done */
4708
		goto out_no_action;
4709
	}
4710
	calldata->timestamp = jiffies;
4711
	if (nfs4_setup_sequence(calldata->server,
4712
				&calldata->arg.seq_args,
4713 4714 4715
				&calldata->res.seq_res,
				task) != 0)
		nfs_release_seqid(calldata->arg.seqid);
4716 4717 4718 4719 4720
	return;
out_no_action:
	task->tk_action = NULL;
out_wait:
	nfs4_sequence_done(task, &calldata->res.seq_res);
4721 4722
}

4723
static const struct rpc_call_ops nfs4_locku_ops = {
T
Trond Myklebust 已提交
4724
	.rpc_call_prepare = nfs4_locku_prepare,
4725
	.rpc_call_done = nfs4_locku_done,
4726
	.rpc_release = nfs4_locku_release_calldata,
4727 4728
};

4729 4730 4731 4732 4733 4734
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;
4735 4736 4737 4738
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_LOCKU],
		.rpc_cred = ctx->cred,
	};
T
Trond Myklebust 已提交
4739 4740
	struct rpc_task_setup task_setup_data = {
		.rpc_client = NFS_CLIENT(lsp->ls_state->inode),
4741
		.rpc_message = &msg,
T
Trond Myklebust 已提交
4742
		.callback_ops = &nfs4_locku_ops,
4743
		.workqueue = nfsiod_workqueue,
T
Trond Myklebust 已提交
4744 4745
		.flags = RPC_TASK_ASYNC,
	};
4746

4747 4748 4749 4750 4751
	/* 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;

4752 4753 4754 4755 4756 4757
	data = nfs4_alloc_unlockdata(fl, ctx, lsp, seqid);
	if (data == NULL) {
		nfs_free_seqid(seqid);
		return ERR_PTR(-ENOMEM);
	}

4758
	nfs41_init_sequence(&data->arg.seq_args, &data->res.seq_res, 1);
4759 4760
	msg.rpc_argp = &data->arg;
	msg.rpc_resp = &data->res;
T
Trond Myklebust 已提交
4761 4762
	task_setup_data.callback_data = data;
	return rpc_run_task(&task_setup_data);
4763 4764
}

4765 4766
static int nfs4_proc_unlck(struct nfs4_state *state, int cmd, struct file_lock *request)
{
4767 4768 4769
	struct inode *inode = state->inode;
	struct nfs4_state_owner *sp = state->owner;
	struct nfs_inode *nfsi = NFS_I(inode);
T
Trond Myklebust 已提交
4770
	struct nfs_seqid *seqid;
L
Linus Torvalds 已提交
4771
	struct nfs4_lock_state *lsp;
4772 4773
	struct rpc_task *task;
	int status = 0;
4774
	unsigned char fl_flags = request->fl_flags;
4775

4776
	status = nfs4_set_lock_state(state, request);
4777 4778
	/* Unlock _before_ we do the RPC call */
	request->fl_flags |= FL_EXISTS;
4779 4780 4781
	/* Exclude nfs_delegation_claim_locks() */
	mutex_lock(&sp->so_delegreturn_mutex);
	/* Exclude nfs4_reclaim_open_stateid() - note nesting! */
4782 4783 4784
	down_read(&nfsi->rwsem);
	if (do_vfs_lock(request->fl_file, request) == -ENOENT) {
		up_read(&nfsi->rwsem);
4785
		mutex_unlock(&sp->so_delegreturn_mutex);
4786
		goto out;
4787 4788
	}
	up_read(&nfsi->rwsem);
4789
	mutex_unlock(&sp->so_delegreturn_mutex);
4790
	if (status != 0)
4791 4792
		goto out;
	/* Is this a delegated lock? */
4793
	lsp = request->fl_u.nfs4_fl.owner;
4794 4795
	if (test_bit(NFS_LOCK_INITIALIZED, &lsp->ls_flags) == 0)
		goto out;
4796
	seqid = nfs_alloc_seqid(&lsp->ls_seqid, GFP_KERNEL);
4797
	status = -ENOMEM;
T
Trond Myklebust 已提交
4798
	if (seqid == NULL)
4799
		goto out;
4800
	task = nfs4_do_unlck(request, nfs_file_open_context(request->fl_file), lsp, seqid);
4801 4802
	status = PTR_ERR(task);
	if (IS_ERR(task))
4803
		goto out;
4804
	status = nfs4_wait_for_completion_rpc_task(task);
4805
	rpc_put_task(task);
4806
out:
4807
	request->fl_flags = fl_flags;
L
Linus Torvalds 已提交
4808 4809 4810
	return status;
}

4811 4812 4813 4814 4815 4816
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;
4817
	unsigned long timestamp;
4818 4819
	int rpc_status;
	int cancelled;
4820
	struct nfs_server *server;
4821 4822 4823
};

static struct nfs4_lockdata *nfs4_alloc_lockdata(struct file_lock *fl,
4824 4825
		struct nfs_open_context *ctx, struct nfs4_lock_state *lsp,
		gfp_t gfp_mask)
L
Linus Torvalds 已提交
4826
{
4827 4828
	struct nfs4_lockdata *p;
	struct inode *inode = lsp->ls_state->inode;
L
Linus Torvalds 已提交
4829
	struct nfs_server *server = NFS_SERVER(inode);
4830

4831
	p = kzalloc(sizeof(*p), gfp_mask);
4832 4833 4834 4835 4836
	if (p == NULL)
		return NULL;

	p->arg.fh = NFS_FH(inode);
	p->arg.fl = &p->fl;
4837
	p->arg.open_seqid = nfs_alloc_seqid(&lsp->ls_state->owner->so_seqid, gfp_mask);
4838 4839
	if (p->arg.open_seqid == NULL)
		goto out_free;
4840
	p->arg.lock_seqid = nfs_alloc_seqid(&lsp->ls_seqid, gfp_mask);
4841
	if (p->arg.lock_seqid == NULL)
4842
		goto out_free_seqid;
4843
	p->arg.lock_stateid = &lsp->ls_stateid;
4844
	p->arg.lock_owner.clientid = server->nfs_client->cl_clientid;
4845
	p->arg.lock_owner.id = lsp->ls_seqid.owner_id;
4846
	p->arg.lock_owner.s_dev = server->s_dev;
4847
	p->res.lock_seqid = p->arg.lock_seqid;
4848
	p->lsp = lsp;
4849
	p->server = server;
4850 4851 4852 4853
	atomic_inc(&lsp->ls_count);
	p->ctx = get_nfs_open_context(ctx);
	memcpy(&p->fl, fl, sizeof(p->fl));
	return p;
4854 4855
out_free_seqid:
	nfs_free_seqid(p->arg.open_seqid);
4856 4857 4858 4859 4860 4861 4862 4863 4864
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;
4865

4866
	dprintk("%s: begin!\n", __func__);
4867
	if (nfs_wait_on_sequence(data->arg.lock_seqid, task) != 0)
4868
		goto out_wait;
4869 4870
	/* Do we need to do an open_to_lock_owner? */
	if (!(data->arg.lock_seqid->sequence->flags & NFS_SEQID_CONFIRMED)) {
4871
		if (nfs_wait_on_sequence(data->arg.open_seqid, task) != 0) {
4872
			goto out_release_lock_seqid;
4873
		}
4874
		data->arg.open_stateid = &state->open_stateid;
4875
		data->arg.new_lock_owner = 1;
4876
		data->res.open_seqid = data->arg.open_seqid;
4877 4878
	} else
		data->arg.new_lock_owner = 0;
4879 4880 4881 4882 4883
	if (!nfs4_valid_open_stateid(state)) {
		data->rpc_status = -EBADF;
		task->tk_action = NULL;
		goto out_release_open_seqid;
	}
4884
	data->timestamp = jiffies;
4885 4886
	if (nfs4_setup_sequence(data->server,
				&data->arg.seq_args,
4887
				&data->res.seq_res,
4888
				task) == 0)
4889
		return;
4890
out_release_open_seqid:
4891 4892 4893
	nfs_release_seqid(data->arg.open_seqid);
out_release_lock_seqid:
	nfs_release_seqid(data->arg.lock_seqid);
4894 4895
out_wait:
	nfs4_sequence_done(task, &data->res.seq_res);
4896
	dprintk("%s: done!, ret = %d\n", __func__, data->rpc_status);
4897 4898
}

4899 4900 4901 4902
static void nfs4_lock_done(struct rpc_task *task, void *calldata)
{
	struct nfs4_lockdata *data = calldata;

4903
	dprintk("%s: begin!\n", __func__);
4904

4905 4906
	if (!nfs4_sequence_done(task, &data->res.seq_res))
		return;
4907

4908 4909 4910 4911 4912 4913 4914 4915
	data->rpc_status = task->tk_status;
	if (data->arg.new_lock_owner != 0) {
		if (data->rpc_status == 0)
			nfs_confirm_seqid(&data->lsp->ls_seqid, 0);
		else
			goto out;
	}
	if (data->rpc_status == 0) {
4916
		nfs4_stateid_copy(&data->lsp->ls_stateid, &data->res.stateid);
4917
		set_bit(NFS_LOCK_INITIALIZED, &data->lsp->ls_flags);
4918
		renew_lease(NFS_SERVER(data->ctx->dentry->d_inode), data->timestamp);
4919 4920
	}
out:
4921
	dprintk("%s: done, ret = %d!\n", __func__, data->rpc_status);
4922 4923 4924 4925 4926 4927
}

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

4928
	dprintk("%s: begin!\n", __func__);
4929
	nfs_free_seqid(data->arg.open_seqid);
4930 4931 4932 4933 4934
	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))
4935
			rpc_put_task_async(task);
4936
		dprintk("%s: cancelling lock!\n", __func__);
4937 4938 4939 4940 4941
	} else
		nfs_free_seqid(data->arg.lock_seqid);
	nfs4_put_lock_state(data->lsp);
	put_nfs_open_context(data->ctx);
	kfree(data);
4942
	dprintk("%s: done!\n", __func__);
4943 4944 4945 4946 4947 4948 4949 4950
}

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,
};

4951 4952 4953 4954 4955
static void nfs4_handle_setlk_error(struct nfs_server *server, struct nfs4_lock_state *lsp, int new_lock_owner, int error)
{
	switch (error) {
	case -NFS4ERR_ADMIN_REVOKED:
	case -NFS4ERR_BAD_STATEID:
4956
		lsp->ls_seqid.flags &= ~NFS_SEQID_CONFIRMED;
4957
		if (new_lock_owner != 0 ||
4958
		   test_bit(NFS_LOCK_INITIALIZED, &lsp->ls_flags) != 0)
4959
			nfs4_schedule_stateid_recovery(server, lsp->ls_state);
4960 4961 4962
		break;
	case -NFS4ERR_STALE_STATEID:
		lsp->ls_seqid.flags &= ~NFS_SEQID_CONFIRMED;
4963 4964
	case -NFS4ERR_EXPIRED:
		nfs4_schedule_lease_recovery(server->nfs_client);
4965 4966 4967
	};
}

4968
static int _nfs4_do_setlk(struct nfs4_state *state, int cmd, struct file_lock *fl, int recovery_type)
4969 4970 4971
{
	struct nfs4_lockdata *data;
	struct rpc_task *task;
4972 4973 4974 4975
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_LOCK],
		.rpc_cred = state->owner->so_cred,
	};
T
Trond Myklebust 已提交
4976 4977
	struct rpc_task_setup task_setup_data = {
		.rpc_client = NFS_CLIENT(state->inode),
4978
		.rpc_message = &msg,
T
Trond Myklebust 已提交
4979
		.callback_ops = &nfs4_lock_ops,
4980
		.workqueue = nfsiod_workqueue,
T
Trond Myklebust 已提交
4981 4982
		.flags = RPC_TASK_ASYNC,
	};
4983 4984
	int ret;

4985
	dprintk("%s: begin!\n", __func__);
4986
	data = nfs4_alloc_lockdata(fl, nfs_file_open_context(fl->fl_file),
4987 4988
			fl->fl_u.nfs4_fl.owner,
			recovery_type == NFS_LOCK_NEW ? GFP_KERNEL : GFP_NOFS);
4989 4990 4991 4992
	if (data == NULL)
		return -ENOMEM;
	if (IS_SETLKW(cmd))
		data->arg.block = 1;
4993
	nfs41_init_sequence(&data->arg.seq_args, &data->res.seq_res, 1);
4994 4995
	msg.rpc_argp = &data->arg;
	msg.rpc_resp = &data->res;
T
Trond Myklebust 已提交
4996
	task_setup_data.callback_data = data;
4997 4998 4999 5000 5001
	if (recovery_type > NFS_LOCK_NEW) {
		if (recovery_type == NFS_LOCK_RECLAIM)
			data->arg.reclaim = NFS_LOCK_RECLAIM;
		nfs4_set_sequence_privileged(&data->arg.seq_args);
	}
T
Trond Myklebust 已提交
5002
	task = rpc_run_task(&task_setup_data);
5003
	if (IS_ERR(task))
5004 5005 5006 5007
		return PTR_ERR(task);
	ret = nfs4_wait_for_completion_rpc_task(task);
	if (ret == 0) {
		ret = data->rpc_status;
5008 5009 5010
		if (ret)
			nfs4_handle_setlk_error(data->server, data->lsp,
					data->arg.new_lock_owner, ret);
5011 5012
	} else
		data->cancelled = 1;
5013
	rpc_put_task(task);
5014
	dprintk("%s: done, ret = %d!\n", __func__, ret);
5015
	return ret;
L
Linus Torvalds 已提交
5016 5017 5018 5019
}

static int nfs4_lock_reclaim(struct nfs4_state *state, struct file_lock *request)
{
5020
	struct nfs_server *server = NFS_SERVER(state->inode);
5021 5022 5023
	struct nfs4_exception exception = {
		.inode = state->inode,
	};
5024 5025 5026
	int err;

	do {
5027 5028 5029
		/* Cache the lock if possible... */
		if (test_bit(NFS_DELEGATED_STATE, &state->flags) != 0)
			return 0;
5030
		err = _nfs4_do_setlk(state, F_SETLK, request, NFS_LOCK_RECLAIM);
5031
		if (err != -NFS4ERR_DELAY)
5032 5033 5034 5035
			break;
		nfs4_handle_exception(server, err, &exception);
	} while (exception.retry);
	return err;
L
Linus Torvalds 已提交
5036 5037 5038 5039
}

static int nfs4_lock_expired(struct nfs4_state *state, struct file_lock *request)
{
5040
	struct nfs_server *server = NFS_SERVER(state->inode);
5041 5042 5043
	struct nfs4_exception exception = {
		.inode = state->inode,
	};
5044 5045
	int err;

5046 5047 5048
	err = nfs4_set_lock_state(state, request);
	if (err != 0)
		return err;
5049
	do {
5050 5051
		if (test_bit(NFS_DELEGATED_STATE, &state->flags) != 0)
			return 0;
5052
		err = _nfs4_do_setlk(state, F_SETLK, request, NFS_LOCK_EXPIRED);
5053 5054 5055 5056 5057 5058 5059 5060
		switch (err) {
		default:
			goto out;
		case -NFS4ERR_GRACE:
		case -NFS4ERR_DELAY:
			nfs4_handle_exception(server, err, &exception);
			err = 0;
		}
5061
	} while (exception.retry);
5062
out:
5063
	return err;
L
Linus Torvalds 已提交
5064 5065
}

5066
#if defined(CONFIG_NFS_V4_1)
5067 5068 5069 5070 5071 5072 5073 5074
/**
 * nfs41_check_expired_locks - possibly free a lock stateid
 *
 * @state: NFSv4 state for an inode
 *
 * Returns NFS_OK if recovery for this stateid is now finished.
 * Otherwise a negative NFS4ERR value is returned.
 */
5075
static int nfs41_check_expired_locks(struct nfs4_state *state)
5076
{
5077
	int status, ret = -NFS4ERR_BAD_STATEID;
5078
	struct nfs4_lock_state *lsp;
5079 5080
	struct nfs_server *server = NFS_SERVER(state->inode);

5081
	list_for_each_entry(lsp, &state->lock_states, ls_locks) {
5082
		if (test_bit(NFS_LOCK_INITIALIZED, &lsp->ls_flags)) {
5083 5084 5085 5086 5087
			struct rpc_cred *cred = lsp->ls_state->owner->so_cred;

			status = nfs41_test_stateid(server,
					&lsp->ls_stateid,
					cred);
5088
			if (status != NFS_OK) {
5089 5090
				/* Free the stateid unless the server
				 * informs us the stateid is unrecognized. */
5091 5092
				if (status != -NFS4ERR_BAD_STATEID)
					nfs41_free_stateid(server,
5093 5094
							&lsp->ls_stateid,
							cred);
5095
				clear_bit(NFS_LOCK_INITIALIZED, &lsp->ls_flags);
5096 5097 5098 5099 5100 5101 5102 5103 5104 5105 5106 5107 5108 5109
				ret = status;
			}
		}
	};

	return ret;
}

static int nfs41_lock_expired(struct nfs4_state *state, struct file_lock *request)
{
	int status = NFS_OK;

	if (test_bit(LK_STATE_IN_USE, &state->flags))
		status = nfs41_check_expired_locks(state);
5110 5111 5112
	if (status != NFS_OK)
		status = nfs4_lock_expired(state, request);
	return status;
5113 5114 5115
}
#endif

L
Linus Torvalds 已提交
5116 5117
static int _nfs4_proc_setlk(struct nfs4_state *state, int cmd, struct file_lock *request)
{
5118
	struct nfs4_state_owner *sp = state->owner;
5119
	struct nfs_inode *nfsi = NFS_I(state->inode);
5120
	unsigned char fl_flags = request->fl_flags;
5121
	unsigned int seq;
5122
	int status = -ENOLCK;
L
Linus Torvalds 已提交
5123

5124 5125 5126
	if ((fl_flags & FL_POSIX) &&
			!test_bit(NFS_STATE_POSIX_LOCKS, &state->flags))
		goto out;
5127 5128 5129 5130
	/* Is this a delegated open? */
	status = nfs4_set_lock_state(state, request);
	if (status != 0)
		goto out;
5131 5132 5133 5134
	request->fl_flags |= FL_ACCESS;
	status = do_vfs_lock(request->fl_file, request);
	if (status < 0)
		goto out;
5135
	down_read(&nfsi->rwsem);
5136 5137 5138
	if (test_bit(NFS_DELEGATED_STATE, &state->flags)) {
		/* Yes: cache locks! */
		/* ...but avoid races with delegation recall... */
5139 5140 5141
		request->fl_flags = fl_flags & ~FL_SLEEP;
		status = do_vfs_lock(request->fl_file, request);
		goto out_unlock;
5142
	}
5143 5144
	seq = raw_seqcount_begin(&sp->so_reclaim_seqcount);
	up_read(&nfsi->rwsem);
5145
	status = _nfs4_do_setlk(state, cmd, request, NFS_LOCK_NEW);
5146
	if (status != 0)
5147 5148 5149 5150
		goto out;
	down_read(&nfsi->rwsem);
	if (read_seqcount_retry(&sp->so_reclaim_seqcount, seq)) {
		status = -NFS4ERR_DELAY;
5151
		goto out_unlock;
5152
	}
5153
	/* Note: we always want to sleep here! */
5154
	request->fl_flags = fl_flags | FL_SLEEP;
5155
	if (do_vfs_lock(request->fl_file, request) < 0)
5156 5157
		printk(KERN_WARNING "NFS: %s: VFS is out of sync with lock "
			"manager!\n", __func__);
5158
out_unlock:
5159
	up_read(&nfsi->rwsem);
5160 5161
out:
	request->fl_flags = fl_flags;
L
Linus Torvalds 已提交
5162 5163 5164 5165 5166
	return status;
}

static int nfs4_proc_setlk(struct nfs4_state *state, int cmd, struct file_lock *request)
{
5167 5168
	struct nfs4_exception exception = {
		.state = state,
5169
		.inode = state->inode,
5170
	};
L
Linus Torvalds 已提交
5171 5172 5173
	int err;

	do {
5174 5175 5176
		err = _nfs4_proc_setlk(state, cmd, request);
		if (err == -NFS4ERR_DENIED)
			err = -EAGAIN;
L
Linus Torvalds 已提交
5177
		err = nfs4_handle_exception(NFS_SERVER(state->inode),
5178
				err, &exception);
L
Linus Torvalds 已提交
5179 5180 5181 5182 5183 5184 5185 5186 5187 5188 5189 5190 5191
	} 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 */
5192
	ctx = nfs_file_open_context(filp);
L
Linus Torvalds 已提交
5193 5194 5195 5196 5197
	state = ctx->state;

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

5198 5199 5200 5201 5202
	if (IS_GETLK(cmd)) {
		if (state != NULL)
			return nfs4_proc_getlk(state, F_GETLK, request);
		return 0;
	}
L
Linus Torvalds 已提交
5203 5204 5205 5206

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

5207 5208 5209 5210 5211
	if (request->fl_type == F_UNLCK) {
		if (state != NULL)
			return nfs4_proc_unlck(state, cmd, request);
		return 0;
	}
L
Linus Torvalds 已提交
5212

5213 5214
	if (state == NULL)
		return -ENOLCK;
5215 5216 5217 5218
	/*
	 * Don't rely on the VFS having checked the file open mode,
	 * since it won't do this for flock() locks.
	 */
5219
	switch (request->fl_type) {
5220 5221 5222 5223 5224 5225 5226 5227 5228
	case F_RDLCK:
		if (!(filp->f_mode & FMODE_READ))
			return -EBADF;
		break;
	case F_WRLCK:
		if (!(filp->f_mode & FMODE_WRITE))
			return -EBADF;
	}

L
Linus Torvalds 已提交
5229 5230 5231 5232 5233 5234 5235 5236 5237 5238 5239 5240
	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;
}

5241
int nfs4_lock_delegation_recall(struct file_lock *fl, struct nfs4_state *state, const nfs4_stateid *stateid)
5242 5243 5244 5245 5246 5247
{
	struct nfs_server *server = NFS_SERVER(state->inode);
	int err;

	err = nfs4_set_lock_state(state, fl);
	if (err != 0)
5248
		return err;
5249
	err = _nfs4_do_setlk(state, F_SETLK, fl, NFS_LOCK_NEW);
5250
	return nfs4_handle_delegation_recall_error(server, state, stateid, err);
5251
}
5252

5253 5254
struct nfs_release_lockowner_data {
	struct nfs4_lock_state *lsp;
5255
	struct nfs_server *server;
5256 5257 5258
	struct nfs_release_lockowner_args args;
};

5259 5260
static void nfs4_release_lockowner_release(void *calldata)
{
5261
	struct nfs_release_lockowner_data *data = calldata;
5262
	nfs4_free_lock_state(data->server, data->lsp);
5263 5264 5265
	kfree(calldata);
}

5266
static const struct rpc_call_ops nfs4_release_lockowner_ops = {
5267 5268 5269
	.rpc_release = nfs4_release_lockowner_release,
};

5270
static int nfs4_release_lockowner(struct nfs_server *server, struct nfs4_lock_state *lsp)
5271
{
5272
	struct nfs_release_lockowner_data *data;
5273 5274 5275 5276 5277
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_RELEASE_LOCKOWNER],
	};

	if (server->nfs_client->cl_mvops->minor_version != 0)
5278 5279 5280 5281 5282
		return -EINVAL;
	data = kmalloc(sizeof(*data), GFP_NOFS);
	if (!data)
		return -ENOMEM;
	data->lsp = lsp;
5283
	data->server = server;
5284 5285 5286 5287 5288 5289
	data->args.lock_owner.clientid = server->nfs_client->cl_clientid;
	data->args.lock_owner.id = lsp->ls_seqid.owner_id;
	data->args.lock_owner.s_dev = server->s_dev;
	msg.rpc_argp = &data->args;
	rpc_call_async(server->client, &msg, 0, &nfs4_release_lockowner_ops, data);
	return 0;
5290 5291
}

5292 5293
#define XATTR_NAME_NFSV4_ACL "system.nfs4_acl"

5294 5295 5296
static int nfs4_xattr_set_nfs4_acl(struct dentry *dentry, const char *key,
				   const void *buf, size_t buflen,
				   int flags, int type)
5297
{
5298 5299
	if (strcmp(key, "") != 0)
		return -EINVAL;
5300

5301
	return nfs4_proc_set_acl(dentry->d_inode, buf, buflen);
5302 5303
}

5304 5305
static int nfs4_xattr_get_nfs4_acl(struct dentry *dentry, const char *key,
				   void *buf, size_t buflen, int type)
5306
{
5307 5308
	if (strcmp(key, "") != 0)
		return -EINVAL;
5309

5310
	return nfs4_proc_get_acl(dentry->d_inode, buf, buflen);
5311 5312
}

5313 5314 5315
static size_t nfs4_xattr_list_nfs4_acl(struct dentry *dentry, char *list,
				       size_t list_len, const char *name,
				       size_t name_len, int type)
5316
{
5317
	size_t len = sizeof(XATTR_NAME_NFSV4_ACL);
5318

5319 5320
	if (!nfs4_server_supports_acls(NFS_SERVER(dentry->d_inode)))
		return 0;
5321 5322 5323

	if (list && len <= list_len)
		memcpy(list, XATTR_NAME_NFSV4_ACL, len);
5324
	return len;
5325 5326
}

5327 5328 5329
/*
 * nfs_fhget will use either the mounted_on_fileid or the fileid
 */
5330 5331
static void nfs_fixup_referral_attributes(struct nfs_fattr *fattr)
{
5332 5333 5334
	if (!(((fattr->valid & NFS_ATTR_FATTR_MOUNTED_ON_FILEID) ||
	       (fattr->valid & NFS_ATTR_FATTR_FILEID)) &&
	      (fattr->valid & NFS_ATTR_FATTR_FSID) &&
5335
	      (fattr->valid & NFS_ATTR_FATTR_V4_LOCATIONS)))
5336 5337 5338
		return;

	fattr->valid |= NFS_ATTR_FATTR_TYPE | NFS_ATTR_FATTR_MODE |
5339
		NFS_ATTR_FATTR_NLINK | NFS_ATTR_FATTR_V4_REFERRAL;
5340 5341 5342 5343
	fattr->mode = S_IFDIR | S_IRUGO | S_IXUGO;
	fattr->nlink = 2;
}

5344 5345 5346 5347
static int _nfs4_proc_fs_locations(struct rpc_clnt *client, struct inode *dir,
				   const struct qstr *name,
				   struct nfs4_fs_locations *fs_locations,
				   struct page *page)
5348 5349 5350
{
	struct nfs_server *server = NFS_SERVER(dir);
	u32 bitmask[2] = {
5351
		[0] = FATTR4_WORD0_FSID | FATTR4_WORD0_FS_LOCATIONS,
5352 5353 5354
	};
	struct nfs4_fs_locations_arg args = {
		.dir_fh = NFS_FH(dir),
5355
		.name = name,
5356 5357 5358
		.page = page,
		.bitmask = bitmask,
	};
B
Benny Halevy 已提交
5359 5360 5361
	struct nfs4_fs_locations_res res = {
		.fs_locations = fs_locations,
	};
5362 5363 5364
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_FS_LOCATIONS],
		.rpc_argp = &args,
B
Benny Halevy 已提交
5365
		.rpc_resp = &res,
5366 5367 5368
	};
	int status;

5369
	dprintk("%s: start\n", __func__);
5370 5371 5372 5373 5374 5375 5376 5377

	/* Ask for the fileid of the absent filesystem if mounted_on_fileid
	 * is not supported */
	if (NFS_SERVER(dir)->attr_bitmask[1] & FATTR4_WORD1_MOUNTED_ON_FILEID)
		bitmask[1] |= FATTR4_WORD1_MOUNTED_ON_FILEID;
	else
		bitmask[0] |= FATTR4_WORD0_FILEID;

5378
	nfs_fattr_init(&fs_locations->fattr);
5379
	fs_locations->server = server;
5380
	fs_locations->nlocations = 0;
5381
	status = nfs4_call_sync(client, server, &msg, &args.seq_args, &res.seq_res, 0);
5382
	dprintk("%s: returned status = %d\n", __func__, status);
5383 5384 5385
	return status;
}

5386 5387 5388 5389
int nfs4_proc_fs_locations(struct rpc_clnt *client, struct inode *dir,
			   const struct qstr *name,
			   struct nfs4_fs_locations *fs_locations,
			   struct page *page)
5390 5391 5392 5393 5394
{
	struct nfs4_exception exception = { };
	int err;
	do {
		err = nfs4_handle_exception(NFS_SERVER(dir),
5395
				_nfs4_proc_fs_locations(client, dir, name, fs_locations, page),
5396 5397 5398 5399 5400
				&exception);
	} while (exception.retry);
	return err;
}

B
Bryan Schumaker 已提交
5401 5402 5403 5404 5405 5406 5407 5408 5409 5410 5411 5412 5413 5414 5415 5416 5417 5418 5419 5420 5421 5422
static int _nfs4_proc_secinfo(struct inode *dir, const struct qstr *name, struct nfs4_secinfo_flavors *flavors)
{
	int status;
	struct nfs4_secinfo_arg args = {
		.dir_fh = NFS_FH(dir),
		.name   = name,
	};
	struct nfs4_secinfo_res res = {
		.flavors     = flavors,
	};
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_SECINFO],
		.rpc_argp = &args,
		.rpc_resp = &res,
	};

	dprintk("NFS call  secinfo %s\n", name->name);
	status = nfs4_call_sync(NFS_SERVER(dir)->client, NFS_SERVER(dir), &msg, &args.seq_args, &res.seq_res, 0);
	dprintk("NFS reply  secinfo: %d\n", status);
	return status;
}

5423 5424
int nfs4_proc_secinfo(struct inode *dir, const struct qstr *name,
		      struct nfs4_secinfo_flavors *flavors)
B
Bryan Schumaker 已提交
5425 5426 5427 5428 5429 5430 5431 5432 5433 5434 5435
{
	struct nfs4_exception exception = { };
	int err;
	do {
		err = nfs4_handle_exception(NFS_SERVER(dir),
				_nfs4_proc_secinfo(dir, name, flavors),
				&exception);
	} while (exception.retry);
	return err;
}

5436
#ifdef CONFIG_NFS_V4_1
5437 5438 5439 5440 5441 5442 5443 5444 5445 5446 5447 5448 5449 5450 5451 5452 5453 5454 5455
/*
 * Check the exchange flags returned by the server for invalid flags, having
 * both PNFS and NON_PNFS flags set, and not having one of NON_PNFS, PNFS, or
 * DS flags set.
 */
static int nfs4_check_cl_exchange_flags(u32 flags)
{
	if (flags & ~EXCHGID4_FLAG_MASK_R)
		goto out_inval;
	if ((flags & EXCHGID4_FLAG_USE_PNFS_MDS) &&
	    (flags & EXCHGID4_FLAG_USE_NON_PNFS))
		goto out_inval;
	if (!(flags & (EXCHGID4_FLAG_MASK_PNFS)))
		goto out_inval;
	return NFS_OK;
out_inval:
	return -NFS4ERR_INVAL;
}

5456
static bool
5457 5458
nfs41_same_server_scope(struct nfs41_server_scope *a,
			struct nfs41_server_scope *b)
5459 5460 5461 5462 5463 5464 5465 5466
{
	if (a->server_scope_sz == b->server_scope_sz &&
	    memcmp(a->server_scope, b->server_scope, a->server_scope_sz) == 0)
		return true;

	return false;
}

5467 5468 5469 5470 5471 5472
/*
 * nfs4_proc_bind_conn_to_session()
 *
 * The 4.1 client currently uses the same TCP connection for the
 * fore and backchannel.
 */
5473
int nfs4_proc_bind_conn_to_session(struct nfs_client *clp, struct rpc_cred *cred)
5474 5475 5476 5477 5478 5479 5480 5481
{
	int status;
	struct nfs41_bind_conn_to_session_res res;
	struct rpc_message msg = {
		.rpc_proc =
			&nfs4_procedures[NFSPROC4_CLNT_BIND_CONN_TO_SESSION],
		.rpc_argp = clp,
		.rpc_resp = &res,
5482
		.rpc_cred = cred,
5483 5484 5485 5486 5487 5488 5489 5490 5491 5492 5493 5494 5495 5496 5497 5498 5499 5500 5501 5502 5503 5504 5505 5506 5507 5508 5509 5510 5511 5512 5513 5514 5515 5516 5517 5518 5519 5520
	};

	dprintk("--> %s\n", __func__);

	res.session = kzalloc(sizeof(struct nfs4_session), GFP_NOFS);
	if (unlikely(res.session == NULL)) {
		status = -ENOMEM;
		goto out;
	}

	status = rpc_call_sync(clp->cl_rpcclient, &msg, RPC_TASK_TIMEOUT);
	if (status == 0) {
		if (memcmp(res.session->sess_id.data,
		    clp->cl_session->sess_id.data, NFS4_MAX_SESSIONID_LEN)) {
			dprintk("NFS: %s: Session ID mismatch\n", __func__);
			status = -EIO;
			goto out_session;
		}
		if (res.dir != NFS4_CDFS4_BOTH) {
			dprintk("NFS: %s: Unexpected direction from server\n",
				__func__);
			status = -EIO;
			goto out_session;
		}
		if (res.use_conn_in_rdma_mode) {
			dprintk("NFS: %s: Server returned RDMA mode = true\n",
				__func__);
			status = -EIO;
			goto out_session;
		}
	}
out_session:
	kfree(res.session);
out:
	dprintk("<-- %s status= %d\n", __func__, status);
	return status;
}

B
Benny Halevy 已提交
5521 5522 5523
/*
 * nfs4_proc_exchange_id()
 *
5524 5525
 * Returns zero, a negative errno, or a negative NFS4ERR status code.
 *
B
Benny Halevy 已提交
5526 5527 5528 5529 5530
 * Since the clientid has expired, all compounds using sessions
 * associated with the stale clientid will be returning
 * NFS4ERR_BADSESSION in the sequence operation, and will therefore
 * be in some phase of session reset.
 */
5531
int nfs4_proc_exchange_id(struct nfs_client *clp, struct rpc_cred *cred)
B
Benny Halevy 已提交
5532 5533 5534
{
	nfs4_verifier verifier;
	struct nfs41_exchange_id_args args = {
5535
		.verifier = &verifier,
B
Benny Halevy 已提交
5536
		.client = clp,
5537 5538
		.flags = EXCHGID4_FLAG_SUPP_MOVED_REFER |
			EXCHGID4_FLAG_BIND_PRINC_STATEID,
B
Benny Halevy 已提交
5539 5540
	};
	struct nfs41_exchange_id_res res = {
5541
		0
B
Benny Halevy 已提交
5542 5543 5544 5545 5546 5547 5548 5549 5550
	};
	int status;
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_EXCHANGE_ID],
		.rpc_argp = &args,
		.rpc_resp = &res,
		.rpc_cred = cred,
	};

5551
	nfs4_init_boot_verifier(clp, &verifier);
5552 5553
	args.id_len = nfs4_init_uniform_client_string(clp, args.id,
							sizeof(args.id));
5554 5555 5556
	dprintk("NFS call  exchange_id auth=%s, '%.*s'\n",
		clp->cl_rpcclient->cl_auth->au_ops->au_name,
		args.id_len, args.id);
B
Benny Halevy 已提交
5557

5558
	res.server_owner = kzalloc(sizeof(struct nfs41_server_owner),
5559
					GFP_NOFS);
5560
	if (unlikely(res.server_owner == NULL)) {
5561 5562 5563
		status = -ENOMEM;
		goto out;
	}
5564

5565
	res.server_scope = kzalloc(sizeof(struct nfs41_server_scope),
5566
					GFP_NOFS);
5567
	if (unlikely(res.server_scope == NULL)) {
5568
		status = -ENOMEM;
5569
		goto out_server_owner;
5570
	}
5571

5572
	res.impl_id = kzalloc(sizeof(struct nfs41_impl_id), GFP_NOFS);
5573
	if (unlikely(res.impl_id == NULL)) {
5574 5575 5576 5577
		status = -ENOMEM;
		goto out_server_scope;
	}

5578
	status = rpc_call_sync(clp->cl_rpcclient, &msg, RPC_TASK_TIMEOUT);
5579
	if (status == 0)
5580
		status = nfs4_check_cl_exchange_flags(res.flags);
5581

5582
	if (status == 0) {
5583 5584 5585 5586 5587
		clp->cl_clientid = res.clientid;
		clp->cl_exchange_flags = (res.flags & ~EXCHGID4_FLAG_CONFIRMED_R);
		if (!(res.flags & EXCHGID4_FLAG_CONFIRMED_R))
			clp->cl_seqid = res.seqid;

5588 5589 5590
		kfree(clp->cl_serverowner);
		clp->cl_serverowner = res.server_owner;
		res.server_owner = NULL;
5591

5592
		/* use the most recent implementation id */
5593 5594
		kfree(clp->cl_implid);
		clp->cl_implid = res.impl_id;
5595

5596
		if (clp->cl_serverscope != NULL &&
5597
		    !nfs41_same_server_scope(clp->cl_serverscope,
5598 5599 5600 5601
					     res.server_scope)) {
			dprintk("%s: server_scope mismatch detected\n",
				__func__);
			set_bit(NFS4CLNT_SERVER_SCOPE_MISMATCH, &clp->cl_state);
5602 5603
			kfree(clp->cl_serverscope);
			clp->cl_serverscope = NULL;
5604 5605
		}

5606
		if (clp->cl_serverscope == NULL) {
5607
			clp->cl_serverscope = res.server_scope;
5608 5609
			goto out;
		}
5610 5611
	} else
		kfree(res.impl_id);
5612

5613 5614
out_server_owner:
	kfree(res.server_owner);
5615
out_server_scope:
5616 5617
	kfree(res.server_scope);
out:
5618
	if (clp->cl_implid != NULL)
5619
		dprintk("NFS reply exchange_id: Server Implementation ID: "
5620
			"domain: %s, name: %s, date: %llu,%u\n",
5621
			clp->cl_implid->domain, clp->cl_implid->name,
5622 5623
			clp->cl_implid->date.seconds,
			clp->cl_implid->date.nseconds);
5624
	dprintk("NFS reply exchange_id: %d\n", status);
B
Benny Halevy 已提交
5625 5626 5627
	return status;
}

T
Trond Myklebust 已提交
5628 5629 5630 5631 5632 5633 5634 5635 5636 5637 5638 5639
static int _nfs4_proc_destroy_clientid(struct nfs_client *clp,
		struct rpc_cred *cred)
{
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_DESTROY_CLIENTID],
		.rpc_argp = clp,
		.rpc_cred = cred,
	};
	int status;

	status = rpc_call_sync(clp->cl_rpcclient, &msg, RPC_TASK_TIMEOUT);
	if (status)
5640
		dprintk("NFS: Got error %d from the server %s on "
T
Trond Myklebust 已提交
5641 5642 5643 5644 5645 5646 5647 5648 5649 5650 5651 5652 5653 5654 5655 5656 5657 5658 5659 5660 5661 5662 5663 5664 5665 5666 5667 5668 5669 5670 5671 5672 5673
			"DESTROY_CLIENTID.", status, clp->cl_hostname);
	return status;
}

static int nfs4_proc_destroy_clientid(struct nfs_client *clp,
		struct rpc_cred *cred)
{
	unsigned int loop;
	int ret;

	for (loop = NFS4_MAX_LOOP_ON_RECOVER; loop != 0; loop--) {
		ret = _nfs4_proc_destroy_clientid(clp, cred);
		switch (ret) {
		case -NFS4ERR_DELAY:
		case -NFS4ERR_CLIENTID_BUSY:
			ssleep(1);
			break;
		default:
			return ret;
		}
	}
	return 0;
}

int nfs4_destroy_clientid(struct nfs_client *clp)
{
	struct rpc_cred *cred;
	int ret = 0;

	if (clp->cl_mvops->minor_version < 1)
		goto out;
	if (clp->cl_exchange_flags == 0)
		goto out;
5674 5675
	if (clp->cl_preserve_clid)
		goto out;
T
Trond Myklebust 已提交
5676 5677 5678 5679 5680 5681 5682 5683 5684 5685 5686 5687 5688
	cred = nfs4_get_exchange_id_cred(clp);
	ret = nfs4_proc_destroy_clientid(clp, cred);
	if (cred)
		put_rpccred(cred);
	switch (ret) {
	case 0:
	case -NFS4ERR_STALE_CLIENTID:
		clp->cl_exchange_flags = 0;
	}
out:
	return ret;
}

A
Andy Adamson 已提交
5689 5690 5691 5692 5693 5694 5695 5696 5697 5698 5699 5700 5701 5702 5703
struct nfs4_get_lease_time_data {
	struct nfs4_get_lease_time_args *args;
	struct nfs4_get_lease_time_res *res;
	struct nfs_client *clp;
};

static void nfs4_get_lease_time_prepare(struct rpc_task *task,
					void *calldata)
{
	struct nfs4_get_lease_time_data *data =
			(struct nfs4_get_lease_time_data *)calldata;

	dprintk("--> %s\n", __func__);
	/* just setup sequence, do not trigger session recovery
	   since we're invoked within one */
5704 5705 5706 5707
	nfs41_setup_sequence(data->clp->cl_session,
			&data->args->la_seq_args,
			&data->res->lr_seq_res,
			task);
A
Andy Adamson 已提交
5708 5709 5710 5711 5712 5713 5714 5715 5716 5717 5718 5719 5720
	dprintk("<-- %s\n", __func__);
}

/*
 * Called from nfs4_state_manager thread for session setup, so don't recover
 * from sequence operation or clientid errors.
 */
static void nfs4_get_lease_time_done(struct rpc_task *task, void *calldata)
{
	struct nfs4_get_lease_time_data *data =
			(struct nfs4_get_lease_time_data *)calldata;

	dprintk("--> %s\n", __func__);
5721 5722
	if (!nfs41_sequence_done(task, &data->res->lr_seq_res))
		return;
A
Andy Adamson 已提交
5723 5724 5725 5726 5727 5728
	switch (task->tk_status) {
	case -NFS4ERR_DELAY:
	case -NFS4ERR_GRACE:
		dprintk("%s Retry: tk_status %d\n", __func__, task->tk_status);
		rpc_delay(task, NFS4_POLL_RETRY_MIN);
		task->tk_status = 0;
5729 5730
		/* fall through */
	case -NFS4ERR_RETRY_UNCACHED_REP:
5731
		rpc_restart_call_prepare(task);
A
Andy Adamson 已提交
5732 5733 5734 5735 5736
		return;
	}
	dprintk("<-- %s\n", __func__);
}

5737
static const struct rpc_call_ops nfs4_get_lease_time_ops = {
A
Andy Adamson 已提交
5738 5739 5740 5741 5742 5743 5744 5745 5746 5747 5748 5749 5750 5751 5752 5753 5754 5755 5756 5757 5758 5759 5760 5761 5762
	.rpc_call_prepare = nfs4_get_lease_time_prepare,
	.rpc_call_done = nfs4_get_lease_time_done,
};

int nfs4_proc_get_lease_time(struct nfs_client *clp, struct nfs_fsinfo *fsinfo)
{
	struct rpc_task *task;
	struct nfs4_get_lease_time_args args;
	struct nfs4_get_lease_time_res res = {
		.lr_fsinfo = fsinfo,
	};
	struct nfs4_get_lease_time_data data = {
		.args = &args,
		.res = &res,
		.clp = clp,
	};
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_GET_LEASE_TIME],
		.rpc_argp = &args,
		.rpc_resp = &res,
	};
	struct rpc_task_setup task_setup = {
		.rpc_client = clp->cl_rpcclient,
		.rpc_message = &msg,
		.callback_ops = &nfs4_get_lease_time_ops,
5763 5764
		.callback_data = &data,
		.flags = RPC_TASK_TIMEOUT,
A
Andy Adamson 已提交
5765 5766 5767
	};
	int status;

5768
	nfs41_init_sequence(&args.la_seq_args, &res.lr_seq_res, 0);
5769
	nfs4_set_sequence_privileged(&args.la_seq_args);
A
Andy Adamson 已提交
5770 5771 5772 5773 5774 5775 5776 5777 5778 5779 5780 5781 5782 5783
	dprintk("--> %s\n", __func__);
	task = rpc_run_task(&task_setup);

	if (IS_ERR(task))
		status = PTR_ERR(task);
	else {
		status = task->tk_status;
		rpc_put_task(task);
	}
	dprintk("<-- %s return %d\n", __func__, status);

	return status;
}

A
Andy Adamson 已提交
5784 5785 5786 5787 5788 5789 5790 5791 5792 5793 5794
/*
 * Initialize the values to be used by the client in CREATE_SESSION
 * If nfs4_init_session set the fore channel request and response sizes,
 * use them.
 *
 * Set the back channel max_resp_sz_cached to zero to force the client to
 * always set csa_cachethis to FALSE because the current implementation
 * of the back channel DRC only supports caching the CB_SEQUENCE operation.
 */
static void nfs4_init_channel_attrs(struct nfs41_create_session_args *args)
{
5795 5796 5797 5798 5799
	unsigned int max_rqst_sz, max_resp_sz;

	max_rqst_sz = NFS_MAX_FILE_IO_SIZE + nfs41_maxwrite_overhead;
	max_resp_sz = NFS_MAX_FILE_IO_SIZE + nfs41_maxread_overhead;

A
Andy Adamson 已提交
5800
	/* Fore channel attributes */
5801 5802
	args->fc_attrs.max_rqst_sz = max_rqst_sz;
	args->fc_attrs.max_resp_sz = max_resp_sz;
A
Andy Adamson 已提交
5803
	args->fc_attrs.max_ops = NFS4_MAX_OPS;
5804
	args->fc_attrs.max_reqs = max_session_slots;
A
Andy Adamson 已提交
5805 5806

	dprintk("%s: Fore Channel : max_rqst_sz=%u max_resp_sz=%u "
5807
		"max_ops=%u max_reqs=%u\n",
A
Andy Adamson 已提交
5808 5809
		__func__,
		args->fc_attrs.max_rqst_sz, args->fc_attrs.max_resp_sz,
5810
		args->fc_attrs.max_ops, args->fc_attrs.max_reqs);
A
Andy Adamson 已提交
5811 5812 5813 5814 5815 5816 5817 5818 5819 5820 5821 5822 5823 5824 5825 5826

	/* Back channel attributes */
	args->bc_attrs.max_rqst_sz = PAGE_SIZE;
	args->bc_attrs.max_resp_sz = PAGE_SIZE;
	args->bc_attrs.max_resp_sz_cached = 0;
	args->bc_attrs.max_ops = NFS4_MAX_BACK_CHANNEL_OPS;
	args->bc_attrs.max_reqs = 1;

	dprintk("%s: Back Channel : max_rqst_sz=%u max_resp_sz=%u "
		"max_resp_sz_cached=%u max_ops=%u max_reqs=%u\n",
		__func__,
		args->bc_attrs.max_rqst_sz, args->bc_attrs.max_resp_sz,
		args->bc_attrs.max_resp_sz_cached, args->bc_attrs.max_ops,
		args->bc_attrs.max_reqs);
}

5827
static int nfs4_verify_fore_channel_attrs(struct nfs41_create_session_args *args, struct nfs4_session *session)
5828
{
5829 5830 5831 5832 5833 5834 5835 5836 5837 5838 5839 5840 5841 5842 5843
	struct nfs4_channel_attrs *sent = &args->fc_attrs;
	struct nfs4_channel_attrs *rcvd = &session->fc_attrs;

	if (rcvd->max_resp_sz > sent->max_resp_sz)
		return -EINVAL;
	/*
	 * Our requested max_ops is the minimum we need; we're not
	 * prepared to break up compounds into smaller pieces than that.
	 * So, no point even trying to continue if the server won't
	 * cooperate:
	 */
	if (rcvd->max_ops < sent->max_ops)
		return -EINVAL;
	if (rcvd->max_reqs == 0)
		return -EINVAL;
5844 5845
	if (rcvd->max_reqs > NFS4_MAX_SLOT_TABLE)
		rcvd->max_reqs = NFS4_MAX_SLOT_TABLE;
5846
	return 0;
5847 5848
}

5849 5850 5851 5852
static int nfs4_verify_back_channel_attrs(struct nfs41_create_session_args *args, struct nfs4_session *session)
{
	struct nfs4_channel_attrs *sent = &args->bc_attrs;
	struct nfs4_channel_attrs *rcvd = &session->bc_attrs;
5853

5854 5855 5856 5857 5858 5859 5860
	if (rcvd->max_rqst_sz > sent->max_rqst_sz)
		return -EINVAL;
	if (rcvd->max_resp_sz < sent->max_resp_sz)
		return -EINVAL;
	if (rcvd->max_resp_sz_cached > sent->max_resp_sz_cached)
		return -EINVAL;
	/* These would render the backchannel useless: */
5861
	if (rcvd->max_ops != sent->max_ops)
5862
		return -EINVAL;
5863
	if (rcvd->max_reqs != sent->max_reqs)
5864 5865 5866
		return -EINVAL;
	return 0;
}
5867 5868 5869 5870

static int nfs4_verify_channel_attrs(struct nfs41_create_session_args *args,
				     struct nfs4_session *session)
{
5871
	int ret;
5872

5873 5874 5875 5876
	ret = nfs4_verify_fore_channel_attrs(args, session);
	if (ret)
		return ret;
	return nfs4_verify_back_channel_attrs(args, session);
5877 5878
}

5879 5880
static int _nfs4_proc_create_session(struct nfs_client *clp,
		struct rpc_cred *cred)
A
Andy Adamson 已提交
5881 5882 5883 5884 5885 5886 5887 5888 5889 5890 5891 5892 5893
{
	struct nfs4_session *session = clp->cl_session;
	struct nfs41_create_session_args args = {
		.client = clp,
		.cb_program = NFS4_CALLBACK,
	};
	struct nfs41_create_session_res res = {
		.client = clp,
	};
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_CREATE_SESSION],
		.rpc_argp = &args,
		.rpc_resp = &res,
5894
		.rpc_cred = cred,
A
Andy Adamson 已提交
5895 5896 5897 5898
	};
	int status;

	nfs4_init_channel_attrs(&args);
5899
	args.flags = (SESSION4_PERSIST | SESSION4_BACK_CHAN);
A
Andy Adamson 已提交
5900

5901
	status = rpc_call_sync(session->clp->cl_rpcclient, &msg, RPC_TASK_TIMEOUT);
A
Andy Adamson 已提交
5902

5903
	if (!status) {
5904 5905
		/* Verify the session's negotiated channel_attrs values */
		status = nfs4_verify_channel_attrs(&args, session);
A
Andy Adamson 已提交
5906 5907 5908 5909 5910 5911 5912 5913 5914 5915 5916 5917
		/* Increment the clientid slot sequence id */
		clp->cl_seqid++;
	}

	return status;
}

/*
 * Issues a CREATE_SESSION operation to the server.
 * It is the responsibility of the caller to verify the session is
 * expired before calling this routine.
 */
5918
int nfs4_proc_create_session(struct nfs_client *clp, struct rpc_cred *cred)
A
Andy Adamson 已提交
5919 5920 5921 5922 5923 5924 5925
{
	int status;
	unsigned *ptr;
	struct nfs4_session *session = clp->cl_session;

	dprintk("--> %s clp=%p session=%p\n", __func__, clp, session);

5926
	status = _nfs4_proc_create_session(clp, cred);
A
Andy Adamson 已提交
5927 5928 5929
	if (status)
		goto out;

5930 5931 5932
	/* Init or reset the session slot tables */
	status = nfs4_setup_session_slot_tables(session);
	dprintk("slot table setup returned %d\n", status);
A
Andy Adamson 已提交
5933 5934 5935 5936 5937 5938 5939 5940 5941 5942 5943
	if (status)
		goto out;

	ptr = (unsigned *)&session->sess_id.data[0];
	dprintk("%s client>seqid %d sessionid %u:%u:%u:%u\n", __func__,
		clp->cl_seqid, ptr[0], ptr[1], ptr[2], ptr[3]);
out:
	dprintk("<-- %s\n", __func__);
	return status;
}

A
Andy Adamson 已提交
5944 5945 5946 5947
/*
 * Issue the over-the-wire RPC DESTROY_SESSION.
 * The caller must serialize access to this routine.
 */
5948 5949
int nfs4_proc_destroy_session(struct nfs4_session *session,
		struct rpc_cred *cred)
A
Andy Adamson 已提交
5950
{
5951 5952 5953 5954 5955
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_DESTROY_SESSION],
		.rpc_argp = session,
		.rpc_cred = cred,
	};
A
Andy Adamson 已提交
5956 5957 5958 5959 5960 5961 5962 5963
	int status = 0;

	dprintk("--> nfs4_proc_destroy_session\n");

	/* session is still being setup */
	if (session->clp->cl_cons_state != NFS_CS_READY)
		return status;

5964
	status = rpc_call_sync(session->clp->cl_rpcclient, &msg, RPC_TASK_TIMEOUT);
A
Andy Adamson 已提交
5965 5966

	if (status)
5967
		dprintk("NFS: Got error %d from the server on DESTROY_SESSION. "
A
Andy Adamson 已提交
5968 5969 5970 5971 5972 5973
			"Session has been destroyed regardless...\n", status);

	dprintk("<-- nfs4_proc_destroy_session\n");
	return status;
}

A
Andy Adamson 已提交
5974 5975 5976
/*
 * Renew the cl_session lease.
 */
5977 5978 5979 5980 5981 5982
struct nfs4_sequence_data {
	struct nfs_client *clp;
	struct nfs4_sequence_args args;
	struct nfs4_sequence_res res;
};

5983 5984
static void nfs41_sequence_release(void *data)
{
5985 5986
	struct nfs4_sequence_data *calldata = data;
	struct nfs_client *clp = calldata->clp;
5987

5988 5989 5990
	if (atomic_read(&clp->cl_count) > 1)
		nfs4_schedule_state_renewal(clp);
	nfs_put_client(clp);
5991
	kfree(calldata);
5992 5993
}

5994 5995 5996 5997 5998 5999 6000
static int nfs41_sequence_handle_errors(struct rpc_task *task, struct nfs_client *clp)
{
	switch(task->tk_status) {
	case -NFS4ERR_DELAY:
		rpc_delay(task, NFS4_POLL_RETRY_MAX);
		return -EAGAIN;
	default:
6001
		nfs4_schedule_lease_recovery(clp);
6002 6003 6004 6005
	}
	return 0;
}

6006
static void nfs41_sequence_call_done(struct rpc_task *task, void *data)
A
Andy Adamson 已提交
6007
{
6008 6009
	struct nfs4_sequence_data *calldata = data;
	struct nfs_client *clp = calldata->clp;
A
Andy Adamson 已提交
6010

6011 6012
	if (!nfs41_sequence_done(task, task->tk_msg.rpc_resp))
		return;
A
Andy Adamson 已提交
6013 6014 6015

	if (task->tk_status < 0) {
		dprintk("%s ERROR %d\n", __func__, task->tk_status);
6016 6017
		if (atomic_read(&clp->cl_count) == 1)
			goto out;
A
Andy Adamson 已提交
6018

6019 6020
		if (nfs41_sequence_handle_errors(task, clp) == -EAGAIN) {
			rpc_restart_call_prepare(task);
A
Andy Adamson 已提交
6021 6022 6023 6024
			return;
		}
	}
	dprintk("%s rpc_cred %p\n", __func__, task->tk_msg.rpc_cred);
6025
out:
A
Andy Adamson 已提交
6026 6027 6028 6029 6030
	dprintk("<-- %s\n", __func__);
}

static void nfs41_sequence_prepare(struct rpc_task *task, void *data)
{
6031 6032
	struct nfs4_sequence_data *calldata = data;
	struct nfs_client *clp = calldata->clp;
A
Andy Adamson 已提交
6033 6034 6035 6036 6037 6038
	struct nfs4_sequence_args *args;
	struct nfs4_sequence_res *res;

	args = task->tk_msg.rpc_argp;
	res = task->tk_msg.rpc_resp;

6039
	nfs41_setup_sequence(clp->cl_session, args, res, task);
A
Andy Adamson 已提交
6040 6041 6042 6043 6044
}

static const struct rpc_call_ops nfs41_sequence_ops = {
	.rpc_call_done = nfs41_sequence_call_done,
	.rpc_call_prepare = nfs41_sequence_prepare,
6045
	.rpc_release = nfs41_sequence_release,
A
Andy Adamson 已提交
6046 6047
};

6048 6049 6050
static struct rpc_task *_nfs41_proc_sequence(struct nfs_client *clp,
		struct rpc_cred *cred,
		bool is_privileged)
A
Andy Adamson 已提交
6051
{
6052
	struct nfs4_sequence_data *calldata;
A
Andy Adamson 已提交
6053 6054 6055 6056
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_SEQUENCE],
		.rpc_cred = cred,
	};
6057 6058 6059
	struct rpc_task_setup task_setup_data = {
		.rpc_client = clp->cl_rpcclient,
		.rpc_message = &msg,
6060
		.callback_ops = &nfs41_sequence_ops,
6061
		.flags = RPC_TASK_ASYNC | RPC_TASK_TIMEOUT,
6062
	};
A
Andy Adamson 已提交
6063

6064
	if (!atomic_inc_not_zero(&clp->cl_count))
6065
		return ERR_PTR(-EIO);
6066
	calldata = kzalloc(sizeof(*calldata), GFP_NOFS);
6067
	if (calldata == NULL) {
6068
		nfs_put_client(clp);
6069
		return ERR_PTR(-ENOMEM);
A
Andy Adamson 已提交
6070
	}
6071
	nfs41_init_sequence(&calldata->args, &calldata->res, 0);
6072 6073
	if (is_privileged)
		nfs4_set_sequence_privileged(&calldata->args);
6074 6075 6076
	msg.rpc_argp = &calldata->args;
	msg.rpc_resp = &calldata->res;
	calldata->clp = clp;
6077
	task_setup_data.callback_data = calldata;
A
Andy Adamson 已提交
6078

6079 6080 6081
	return rpc_run_task(&task_setup_data);
}

6082
static int nfs41_proc_async_sequence(struct nfs_client *clp, struct rpc_cred *cred, unsigned renew_flags)
6083 6084 6085 6086
{
	struct rpc_task *task;
	int ret = 0;

6087 6088
	if ((renew_flags & NFS4_RENEW_TIMEOUT) == 0)
		return 0;
6089
	task = _nfs41_proc_sequence(clp, cred, false);
6090 6091 6092
	if (IS_ERR(task))
		ret = PTR_ERR(task);
	else
6093
		rpc_put_task_async(task);
6094 6095 6096 6097 6098 6099 6100 6101 6102
	dprintk("<-- %s status=%d\n", __func__, ret);
	return ret;
}

static int nfs4_proc_sequence(struct nfs_client *clp, struct rpc_cred *cred)
{
	struct rpc_task *task;
	int ret;

6103
	task = _nfs41_proc_sequence(clp, cred, true);
6104 6105 6106 6107 6108
	if (IS_ERR(task)) {
		ret = PTR_ERR(task);
		goto out;
	}
	ret = rpc_wait_for_completion_task(task);
6109 6110 6111 6112 6113
	if (!ret) {
		struct nfs4_sequence_res *res = task->tk_msg.rpc_resp;

		if (task->tk_status == 0)
			nfs41_handle_sequence_flag_errors(clp, res->sr_status_flags);
6114
		ret = task->tk_status;
6115
	}
6116 6117 6118 6119
	rpc_put_task(task);
out:
	dprintk("<-- %s status=%d\n", __func__, ret);
	return ret;
A
Andy Adamson 已提交
6120 6121
}

6122 6123 6124 6125 6126 6127 6128 6129 6130 6131
struct nfs4_reclaim_complete_data {
	struct nfs_client *clp;
	struct nfs41_reclaim_complete_args arg;
	struct nfs41_reclaim_complete_res res;
};

static void nfs4_reclaim_complete_prepare(struct rpc_task *task, void *data)
{
	struct nfs4_reclaim_complete_data *calldata = data;

6132 6133 6134 6135
	nfs41_setup_sequence(calldata->clp->cl_session,
			&calldata->arg.seq_args,
			&calldata->res.seq_res,
			task);
6136 6137
}

6138 6139 6140 6141 6142 6143 6144 6145 6146
static int nfs41_reclaim_complete_handle_errors(struct rpc_task *task, struct nfs_client *clp)
{
	switch(task->tk_status) {
	case 0:
	case -NFS4ERR_COMPLETE_ALREADY:
	case -NFS4ERR_WRONG_CRED: /* What to do here? */
		break;
	case -NFS4ERR_DELAY:
		rpc_delay(task, NFS4_POLL_RETRY_MAX);
6147 6148
		/* fall through */
	case -NFS4ERR_RETRY_UNCACHED_REP:
6149 6150
		return -EAGAIN;
	default:
6151
		nfs4_schedule_lease_recovery(clp);
6152 6153 6154 6155
	}
	return 0;
}

6156 6157 6158 6159 6160 6161 6162
static void nfs4_reclaim_complete_done(struct rpc_task *task, void *data)
{
	struct nfs4_reclaim_complete_data *calldata = data;
	struct nfs_client *clp = calldata->clp;
	struct nfs4_sequence_res *res = &calldata->res.seq_res;

	dprintk("--> %s\n", __func__);
6163 6164
	if (!nfs41_sequence_done(task, res))
		return;
6165

6166 6167 6168 6169
	if (nfs41_reclaim_complete_handle_errors(task, clp) == -EAGAIN) {
		rpc_restart_call_prepare(task);
		return;
	}
6170 6171 6172 6173 6174 6175 6176 6177 6178 6179 6180 6181 6182 6183 6184 6185 6186 6187 6188
	dprintk("<-- %s\n", __func__);
}

static void nfs4_free_reclaim_complete_data(void *data)
{
	struct nfs4_reclaim_complete_data *calldata = data;

	kfree(calldata);
}

static const struct rpc_call_ops nfs4_reclaim_complete_call_ops = {
	.rpc_call_prepare = nfs4_reclaim_complete_prepare,
	.rpc_call_done = nfs4_reclaim_complete_done,
	.rpc_release = nfs4_free_reclaim_complete_data,
};

/*
 * Issue a global reclaim complete.
 */
6189 6190
static int nfs41_proc_reclaim_complete(struct nfs_client *clp,
		struct rpc_cred *cred)
6191 6192 6193 6194 6195
{
	struct nfs4_reclaim_complete_data *calldata;
	struct rpc_task *task;
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_RECLAIM_COMPLETE],
6196
		.rpc_cred = cred,
6197 6198 6199 6200 6201 6202 6203 6204 6205 6206
	};
	struct rpc_task_setup task_setup_data = {
		.rpc_client = clp->cl_rpcclient,
		.rpc_message = &msg,
		.callback_ops = &nfs4_reclaim_complete_call_ops,
		.flags = RPC_TASK_ASYNC,
	};
	int status = -ENOMEM;

	dprintk("--> %s\n", __func__);
6207
	calldata = kzalloc(sizeof(*calldata), GFP_NOFS);
6208 6209 6210 6211 6212
	if (calldata == NULL)
		goto out;
	calldata->clp = clp;
	calldata->arg.one_fs = 0;

6213
	nfs41_init_sequence(&calldata->arg.seq_args, &calldata->res.seq_res, 0);
6214
	nfs4_set_sequence_privileged(&calldata->arg.seq_args);
6215 6216 6217 6218
	msg.rpc_argp = &calldata->arg;
	msg.rpc_resp = &calldata->res;
	task_setup_data.callback_data = calldata;
	task = rpc_run_task(&task_setup_data);
6219
	if (IS_ERR(task)) {
6220
		status = PTR_ERR(task);
6221 6222
		goto out;
	}
6223 6224 6225
	status = nfs4_wait_for_completion_rpc_task(task);
	if (status == 0)
		status = task->tk_status;
6226
	rpc_put_task(task);
6227
	return 0;
6228 6229 6230 6231
out:
	dprintk("<-- %s status=%d\n", __func__, status);
	return status;
}
6232 6233 6234 6235 6236

static void
nfs4_layoutget_prepare(struct rpc_task *task, void *calldata)
{
	struct nfs4_layoutget *lgp = calldata;
F
Fred Isaman 已提交
6237
	struct nfs_server *server = NFS_SERVER(lgp->args.inode);
6238
	struct nfs4_session *session = nfs4_get_session(server);
6239 6240

	dprintk("--> %s\n", __func__);
F
Fred Isaman 已提交
6241 6242 6243 6244 6245
	/* Note the is a race here, where a CB_LAYOUTRECALL can come in
	 * right now covering the LAYOUTGET we are about to send.
	 * However, that is not so catastrophic, and there seems
	 * to be no way to prevent it completely.
	 */
6246
	if (nfs41_setup_sequence(session, &lgp->args.seq_args,
6247
				&lgp->res.seq_res, task))
6248
		return;
6249 6250 6251 6252 6253
	if (pnfs_choose_layoutget_stateid(&lgp->args.stateid,
					  NFS_I(lgp->args.inode)->layout,
					  lgp->args.ctx->state)) {
		rpc_exit(task, NFS4_OK);
	}
6254 6255 6256 6257 6258
}

static void nfs4_layoutget_done(struct rpc_task *task, void *calldata)
{
	struct nfs4_layoutget *lgp = calldata;
6259 6260 6261 6262
	struct inode *inode = lgp->args.inode;
	struct nfs_server *server = NFS_SERVER(inode);
	struct pnfs_layout_hdr *lo;
	struct nfs4_state *state = NULL;
6263
	unsigned long timeo, giveup;
6264 6265 6266

	dprintk("--> %s\n", __func__);

6267
	if (!nfs41_sequence_done(task, &lgp->res.seq_res))
6268
		goto out;
6269 6270 6271

	switch (task->tk_status) {
	case 0:
6272
		goto out;
6273 6274
	case -NFS4ERR_LAYOUTTRYLATER:
	case -NFS4ERR_RECALLCONFLICT:
6275 6276 6277 6278
		timeo = rpc_get_timeout(task->tk_client);
		giveup = lgp->args.timestamp + timeo;
		if (time_after(giveup, jiffies))
			task->tk_status = -NFS4ERR_DELAY;
6279 6280 6281 6282 6283 6284 6285 6286 6287 6288 6289 6290 6291 6292 6293 6294 6295
		break;
	case -NFS4ERR_EXPIRED:
	case -NFS4ERR_BAD_STATEID:
		spin_lock(&inode->i_lock);
		lo = NFS_I(inode)->layout;
		if (!lo || list_empty(&lo->plh_segs)) {
			spin_unlock(&inode->i_lock);
			/* If the open stateid was bad, then recover it. */
			state = lgp->args.ctx->state;
		} else {
			LIST_HEAD(head);

			pnfs_mark_matching_lsegs_invalid(lo, &head, NULL);
			spin_unlock(&inode->i_lock);
			/* Mark the bad layout state as invalid, then
			 * retry using the open stateid. */
			pnfs_free_lseg_list(&head);
6296 6297
		}
	}
6298 6299 6300
	if (nfs4_async_handle_error(task, server, state) == -EAGAIN)
		rpc_restart_call_prepare(task);
out:
6301 6302 6303
	dprintk("<-- %s\n", __func__);
}

6304 6305 6306 6307 6308 6309 6310 6311 6312 6313 6314 6315 6316 6317 6318 6319 6320 6321 6322 6323 6324 6325 6326 6327 6328 6329 6330 6331 6332 6333 6334 6335 6336 6337 6338 6339 6340 6341 6342 6343 6344 6345 6346 6347
static size_t max_response_pages(struct nfs_server *server)
{
	u32 max_resp_sz = server->nfs_client->cl_session->fc_attrs.max_resp_sz;
	return nfs_page_array_len(0, max_resp_sz);
}

static void nfs4_free_pages(struct page **pages, size_t size)
{
	int i;

	if (!pages)
		return;

	for (i = 0; i < size; i++) {
		if (!pages[i])
			break;
		__free_page(pages[i]);
	}
	kfree(pages);
}

static struct page **nfs4_alloc_pages(size_t size, gfp_t gfp_flags)
{
	struct page **pages;
	int i;

	pages = kcalloc(size, sizeof(struct page *), gfp_flags);
	if (!pages) {
		dprintk("%s: can't alloc array of %zu pages\n", __func__, size);
		return NULL;
	}

	for (i = 0; i < size; i++) {
		pages[i] = alloc_page(gfp_flags);
		if (!pages[i]) {
			dprintk("%s: failed to allocate page\n", __func__);
			nfs4_free_pages(pages, size);
			return NULL;
		}
	}

	return pages;
}

6348 6349 6350
static void nfs4_layoutget_release(void *calldata)
{
	struct nfs4_layoutget *lgp = calldata;
6351 6352
	struct inode *inode = lgp->args.inode;
	struct nfs_server *server = NFS_SERVER(inode);
6353
	size_t max_pages = max_response_pages(server);
6354 6355

	dprintk("--> %s\n", __func__);
6356
	nfs4_free_pages(lgp->args.layout.pages, max_pages);
6357
	pnfs_put_layout_hdr(NFS_I(inode)->layout);
6358 6359 6360 6361 6362 6363 6364 6365 6366 6367 6368
	put_nfs_open_context(lgp->args.ctx);
	kfree(calldata);
	dprintk("<-- %s\n", __func__);
}

static const struct rpc_call_ops nfs4_layoutget_call_ops = {
	.rpc_call_prepare = nfs4_layoutget_prepare,
	.rpc_call_done = nfs4_layoutget_done,
	.rpc_release = nfs4_layoutget_release,
};

6369 6370
struct pnfs_layout_segment *
nfs4_proc_layoutget(struct nfs4_layoutget *lgp, gfp_t gfp_flags)
6371
{
6372 6373
	struct inode *inode = lgp->args.inode;
	struct nfs_server *server = NFS_SERVER(inode);
6374
	size_t max_pages = max_response_pages(server);
6375 6376 6377 6378 6379
	struct rpc_task *task;
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_LAYOUTGET],
		.rpc_argp = &lgp->args,
		.rpc_resp = &lgp->res,
6380
		.rpc_cred = lgp->cred,
6381 6382 6383 6384 6385 6386 6387 6388
	};
	struct rpc_task_setup task_setup_data = {
		.rpc_client = server->client,
		.rpc_message = &msg,
		.callback_ops = &nfs4_layoutget_call_ops,
		.callback_data = lgp,
		.flags = RPC_TASK_ASYNC,
	};
6389
	struct pnfs_layout_segment *lseg = NULL;
6390 6391 6392 6393
	int status = 0;

	dprintk("--> %s\n", __func__);

6394 6395 6396
	lgp->args.layout.pages = nfs4_alloc_pages(max_pages, gfp_flags);
	if (!lgp->args.layout.pages) {
		nfs4_layoutget_release(lgp);
6397
		return ERR_PTR(-ENOMEM);
6398 6399
	}
	lgp->args.layout.pglen = max_pages * PAGE_SIZE;
6400
	lgp->args.timestamp = jiffies;
6401

6402
	lgp->res.layoutp = &lgp->args.layout;
6403
	lgp->res.seq_res.sr_slot = NULL;
6404
	nfs41_init_sequence(&lgp->args.seq_args, &lgp->res.seq_res, 0);
6405 6406 6407 6408

	/* nfs4_layoutget_release calls pnfs_put_layout_hdr */
	pnfs_get_layout_hdr(NFS_I(inode)->layout);

6409 6410
	task = rpc_run_task(&task_setup_data);
	if (IS_ERR(task))
6411
		return ERR_CAST(task);
6412
	status = nfs4_wait_for_completion_rpc_task(task);
F
Fred Isaman 已提交
6413 6414
	if (status == 0)
		status = task->tk_status;
6415 6416
	/* if layoutp->len is 0, nfs4_layoutget_prepare called rpc_exit */
	if (status == 0 && lgp->res.layoutp->len)
6417
		lseg = pnfs_layout_process(lgp);
6418 6419
	rpc_put_task(task);
	dprintk("<-- %s status=%d\n", __func__, status);
6420 6421 6422
	if (status)
		return ERR_PTR(status);
	return lseg;
6423 6424
}

B
Benny Halevy 已提交
6425 6426 6427 6428 6429 6430
static void
nfs4_layoutreturn_prepare(struct rpc_task *task, void *calldata)
{
	struct nfs4_layoutreturn *lrp = calldata;

	dprintk("--> %s\n", __func__);
6431 6432 6433 6434
	nfs41_setup_sequence(lrp->clp->cl_session,
			&lrp->args.seq_args,
			&lrp->res.seq_res,
			task);
B
Benny Halevy 已提交
6435 6436 6437 6438 6439 6440 6441 6442 6443
}

static void nfs4_layoutreturn_done(struct rpc_task *task, void *calldata)
{
	struct nfs4_layoutreturn *lrp = calldata;
	struct nfs_server *server;

	dprintk("--> %s\n", __func__);

6444
	if (!nfs41_sequence_done(task, &lrp->res.seq_res))
B
Benny Halevy 已提交
6445 6446 6447 6448
		return;

	server = NFS_SERVER(lrp->args.inode);
	if (nfs4_async_handle_error(task, server, NULL) == -EAGAIN) {
6449
		rpc_restart_call_prepare(task);
B
Benny Halevy 已提交
6450 6451 6452 6453 6454 6455 6456 6457
		return;
	}
	dprintk("<-- %s\n", __func__);
}

static void nfs4_layoutreturn_release(void *calldata)
{
	struct nfs4_layoutreturn *lrp = calldata;
6458
	struct pnfs_layout_hdr *lo = lrp->args.layout;
B
Benny Halevy 已提交
6459 6460

	dprintk("--> %s\n", __func__);
6461 6462 6463 6464 6465
	spin_lock(&lo->plh_inode->i_lock);
	if (lrp->res.lrs_present)
		pnfs_set_layout_stateid(lo, &lrp->res.stateid, true);
	lo->plh_block_lgets--;
	spin_unlock(&lo->plh_inode->i_lock);
6466
	pnfs_put_layout_hdr(lrp->args.layout);
B
Benny Halevy 已提交
6467 6468 6469 6470 6471 6472 6473 6474 6475 6476 6477 6478 6479 6480 6481 6482 6483
	kfree(calldata);
	dprintk("<-- %s\n", __func__);
}

static const struct rpc_call_ops nfs4_layoutreturn_call_ops = {
	.rpc_call_prepare = nfs4_layoutreturn_prepare,
	.rpc_call_done = nfs4_layoutreturn_done,
	.rpc_release = nfs4_layoutreturn_release,
};

int nfs4_proc_layoutreturn(struct nfs4_layoutreturn *lrp)
{
	struct rpc_task *task;
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_LAYOUTRETURN],
		.rpc_argp = &lrp->args,
		.rpc_resp = &lrp->res,
6484
		.rpc_cred = lrp->cred,
B
Benny Halevy 已提交
6485 6486 6487 6488 6489 6490 6491 6492 6493 6494
	};
	struct rpc_task_setup task_setup_data = {
		.rpc_client = lrp->clp->cl_rpcclient,
		.rpc_message = &msg,
		.callback_ops = &nfs4_layoutreturn_call_ops,
		.callback_data = lrp,
	};
	int status;

	dprintk("--> %s\n", __func__);
6495
	nfs41_init_sequence(&lrp->args.seq_args, &lrp->res.seq_res, 1);
B
Benny Halevy 已提交
6496 6497 6498 6499 6500 6501 6502 6503 6504
	task = rpc_run_task(&task_setup_data);
	if (IS_ERR(task))
		return PTR_ERR(task);
	status = task->tk_status;
	dprintk("<-- %s status=%d\n", __func__, status);
	rpc_put_task(task);
	return status;
}

A
Andy Adamson 已提交
6505 6506 6507 6508 6509 6510 6511 6512 6513 6514 6515 6516 6517 6518 6519 6520 6521 6522 6523 6524 6525 6526 6527 6528 6529 6530 6531 6532 6533 6534 6535 6536 6537 6538 6539 6540 6541 6542 6543 6544 6545 6546 6547 6548 6549 6550 6551 6552
/*
 * Retrieve the list of Data Server devices from the MDS.
 */
static int _nfs4_getdevicelist(struct nfs_server *server,
				    const struct nfs_fh *fh,
				    struct pnfs_devicelist *devlist)
{
	struct nfs4_getdevicelist_args args = {
		.fh = fh,
		.layoutclass = server->pnfs_curr_ld->id,
	};
	struct nfs4_getdevicelist_res res = {
		.devlist = devlist,
	};
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_GETDEVICELIST],
		.rpc_argp = &args,
		.rpc_resp = &res,
	};
	int status;

	dprintk("--> %s\n", __func__);
	status = nfs4_call_sync(server->client, server, &msg, &args.seq_args,
				&res.seq_res, 0);
	dprintk("<-- %s status=%d\n", __func__, status);
	return status;
}

int nfs4_proc_getdevicelist(struct nfs_server *server,
			    const struct nfs_fh *fh,
			    struct pnfs_devicelist *devlist)
{
	struct nfs4_exception exception = { };
	int err;

	do {
		err = nfs4_handle_exception(server,
				_nfs4_getdevicelist(server, fh, devlist),
				&exception);
	} while (exception.retry);

	dprintk("%s: err=%d, num_devs=%u\n", __func__,
		err, devlist->num_devs);

	return err;
}
EXPORT_SYMBOL_GPL(nfs4_proc_getdevicelist);

6553
static int
6554 6555 6556
_nfs4_proc_getdeviceinfo(struct nfs_server *server,
		struct pnfs_device *pdev,
		struct rpc_cred *cred)
6557 6558 6559 6560 6561 6562 6563 6564 6565 6566 6567
{
	struct nfs4_getdeviceinfo_args args = {
		.pdev = pdev,
	};
	struct nfs4_getdeviceinfo_res res = {
		.pdev = pdev,
	};
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_GETDEVICEINFO],
		.rpc_argp = &args,
		.rpc_resp = &res,
6568
		.rpc_cred = cred,
6569 6570 6571 6572
	};
	int status;

	dprintk("--> %s\n", __func__);
6573
	status = nfs4_call_sync(server->client, server, &msg, &args.seq_args, &res.seq_res, 0);
6574 6575 6576 6577 6578
	dprintk("<-- %s status=%d\n", __func__, status);

	return status;
}

6579 6580 6581
int nfs4_proc_getdeviceinfo(struct nfs_server *server,
		struct pnfs_device *pdev,
		struct rpc_cred *cred)
6582 6583 6584 6585 6586 6587
{
	struct nfs4_exception exception = { };
	int err;

	do {
		err = nfs4_handle_exception(server,
6588
					_nfs4_proc_getdeviceinfo(server, pdev, cred),
6589 6590 6591 6592 6593 6594
					&exception);
	} while (exception.retry);
	return err;
}
EXPORT_SYMBOL_GPL(nfs4_proc_getdeviceinfo);

A
Andy Adamson 已提交
6595 6596 6597 6598
static void nfs4_layoutcommit_prepare(struct rpc_task *task, void *calldata)
{
	struct nfs4_layoutcommit_data *data = calldata;
	struct nfs_server *server = NFS_SERVER(data->args.inode);
6599
	struct nfs4_session *session = nfs4_get_session(server);
A
Andy Adamson 已提交
6600

6601 6602 6603 6604
	nfs41_setup_sequence(session,
			&data->args.seq_args,
			&data->res.seq_res,
			task);
A
Andy Adamson 已提交
6605 6606 6607 6608 6609 6610 6611 6612
}

static void
nfs4_layoutcommit_done(struct rpc_task *task, void *calldata)
{
	struct nfs4_layoutcommit_data *data = calldata;
	struct nfs_server *server = NFS_SERVER(data->args.inode);

6613
	if (!nfs41_sequence_done(task, &data->res.seq_res))
A
Andy Adamson 已提交
6614 6615 6616
		return;

	switch (task->tk_status) { /* Just ignore these failures */
6617 6618 6619 6620
	case -NFS4ERR_DELEG_REVOKED: /* layout was recalled */
	case -NFS4ERR_BADIOMODE:     /* no IOMODE_RW layout for range */
	case -NFS4ERR_BADLAYOUT:     /* no layout */
	case -NFS4ERR_GRACE:	    /* loca_recalim always false */
A
Andy Adamson 已提交
6621
		task->tk_status = 0;
6622 6623
		break;
	case 0:
A
Andy Adamson 已提交
6624 6625
		nfs_post_op_update_inode_force_wcc(data->args.inode,
						   data->res.fattr);
6626 6627 6628 6629 6630 6631 6632
		break;
	default:
		if (nfs4_async_handle_error(task, server, NULL) == -EAGAIN) {
			rpc_restart_call_prepare(task);
			return;
		}
	}
A
Andy Adamson 已提交
6633 6634 6635 6636 6637 6638
}

static void nfs4_layoutcommit_release(void *calldata)
{
	struct nfs4_layoutcommit_data *data = calldata;

A
Andy Adamson 已提交
6639
	pnfs_cleanup_layoutcommit(data);
A
Andy Adamson 已提交
6640 6641 6642 6643 6644 6645 6646 6647 6648 6649 6650
	put_rpccred(data->cred);
	kfree(data);
}

static const struct rpc_call_ops nfs4_layoutcommit_ops = {
	.rpc_call_prepare = nfs4_layoutcommit_prepare,
	.rpc_call_done = nfs4_layoutcommit_done,
	.rpc_release = nfs4_layoutcommit_release,
};

int
6651
nfs4_proc_layoutcommit(struct nfs4_layoutcommit_data *data, bool sync)
A
Andy Adamson 已提交
6652 6653 6654 6655 6656 6657 6658 6659 6660 6661 6662 6663 6664 6665 6666 6667 6668 6669 6670 6671 6672 6673 6674 6675
{
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_LAYOUTCOMMIT],
		.rpc_argp = &data->args,
		.rpc_resp = &data->res,
		.rpc_cred = data->cred,
	};
	struct rpc_task_setup task_setup_data = {
		.task = &data->task,
		.rpc_client = NFS_CLIENT(data->args.inode),
		.rpc_message = &msg,
		.callback_ops = &nfs4_layoutcommit_ops,
		.callback_data = data,
		.flags = RPC_TASK_ASYNC,
	};
	struct rpc_task *task;
	int status = 0;

	dprintk("NFS: %4d initiating layoutcommit call. sync %d "
		"lbw: %llu inode %lu\n",
		data->task.tk_pid, sync,
		data->args.lastbytewritten,
		data->args.inode->i_ino);

6676
	nfs41_init_sequence(&data->args.seq_args, &data->res.seq_res, 1);
A
Andy Adamson 已提交
6677 6678 6679
	task = rpc_run_task(&task_setup_data);
	if (IS_ERR(task))
		return PTR_ERR(task);
6680
	if (sync == false)
A
Andy Adamson 已提交
6681 6682 6683 6684 6685 6686 6687 6688 6689 6690
		goto out;
	status = nfs4_wait_for_completion_rpc_task(task);
	if (status != 0)
		goto out;
	status = task->tk_status;
out:
	dprintk("%s: status %d\n", __func__, status);
	rpc_put_task(task);
	return status;
}
6691 6692 6693 6694 6695 6696 6697 6698 6699 6700 6701 6702 6703 6704 6705 6706 6707 6708 6709 6710 6711 6712 6713 6714 6715 6716 6717 6718 6719 6720 6721

static int
_nfs41_proc_secinfo_no_name(struct nfs_server *server, struct nfs_fh *fhandle,
		    struct nfs_fsinfo *info, struct nfs4_secinfo_flavors *flavors)
{
	struct nfs41_secinfo_no_name_args args = {
		.style = SECINFO_STYLE_CURRENT_FH,
	};
	struct nfs4_secinfo_res res = {
		.flavors = flavors,
	};
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_SECINFO_NO_NAME],
		.rpc_argp = &args,
		.rpc_resp = &res,
	};
	return nfs4_call_sync(server->client, server, &msg, &args.seq_args, &res.seq_res, 0);
}

static int
nfs41_proc_secinfo_no_name(struct nfs_server *server, struct nfs_fh *fhandle,
			   struct nfs_fsinfo *info, struct nfs4_secinfo_flavors *flavors)
{
	struct nfs4_exception exception = { };
	int err;
	do {
		err = _nfs41_proc_secinfo_no_name(server, fhandle, info, flavors);
		switch (err) {
		case 0:
		case -NFS4ERR_WRONGSEC:
		case -NFS4ERR_NOTSUPP:
6722
			goto out;
6723 6724 6725 6726
		default:
			err = nfs4_handle_exception(server, err, &exception);
		}
	} while (exception.retry);
6727
out:
6728 6729 6730 6731 6732 6733 6734 6735 6736 6737 6738 6739 6740 6741 6742 6743 6744 6745 6746 6747 6748 6749 6750 6751 6752 6753 6754 6755 6756 6757 6758 6759 6760 6761 6762 6763 6764 6765 6766 6767 6768 6769 6770
	return err;
}

static int
nfs41_find_root_sec(struct nfs_server *server, struct nfs_fh *fhandle,
		    struct nfs_fsinfo *info)
{
	int err;
	struct page *page;
	rpc_authflavor_t flavor;
	struct nfs4_secinfo_flavors *flavors;

	page = alloc_page(GFP_KERNEL);
	if (!page) {
		err = -ENOMEM;
		goto out;
	}

	flavors = page_address(page);
	err = nfs41_proc_secinfo_no_name(server, fhandle, info, flavors);

	/*
	 * Fall back on "guess and check" method if
	 * the server doesn't support SECINFO_NO_NAME
	 */
	if (err == -NFS4ERR_WRONGSEC || err == -NFS4ERR_NOTSUPP) {
		err = nfs4_find_root_sec(server, fhandle, info);
		goto out_freepage;
	}
	if (err)
		goto out_freepage;

	flavor = nfs_find_best_sec(flavors);
	if (err == 0)
		err = nfs4_lookup_root_sec(server, fhandle, info, flavor);

out_freepage:
	put_page(page);
	if (err == -EACCES)
		return -EPERM;
out:
	return err;
}
6771

6772 6773 6774
static int _nfs41_test_stateid(struct nfs_server *server,
		nfs4_stateid *stateid,
		struct rpc_cred *cred)
B
Bryan Schumaker 已提交
6775 6776 6777
{
	int status;
	struct nfs41_test_stateid_args args = {
6778
		.stateid = stateid,
B
Bryan Schumaker 已提交
6779 6780 6781 6782 6783 6784
	};
	struct nfs41_test_stateid_res res;
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_TEST_STATEID],
		.rpc_argp = &args,
		.rpc_resp = &res,
6785
		.rpc_cred = cred,
B
Bryan Schumaker 已提交
6786
	};
6787

6788
	dprintk("NFS call  test_stateid %p\n", stateid);
6789
	nfs41_init_sequence(&args.seq_args, &res.seq_res, 0);
6790 6791 6792
	nfs4_set_sequence_privileged(&args.seq_args);
	status = nfs4_call_sync_sequence(server->client, server, &msg,
			&args.seq_args, &res.seq_res);
6793 6794
	if (status != NFS_OK) {
		dprintk("NFS reply test_stateid: failed, %d\n", status);
6795
		return status;
6796 6797
	}
	dprintk("NFS reply test_stateid: succeeded, %d\n", -res.status);
6798
	return -res.status;
B
Bryan Schumaker 已提交
6799 6800
}

6801 6802 6803 6804 6805
/**
 * nfs41_test_stateid - perform a TEST_STATEID operation
 *
 * @server: server / transport on which to perform the operation
 * @stateid: state ID to test
6806
 * @cred: credential
6807 6808 6809 6810 6811
 *
 * Returns NFS_OK if the server recognizes that "stateid" is valid.
 * Otherwise a negative NFS4ERR value is returned if the operation
 * failed or the state ID is not currently valid.
 */
6812 6813 6814
static int nfs41_test_stateid(struct nfs_server *server,
		nfs4_stateid *stateid,
		struct rpc_cred *cred)
B
Bryan Schumaker 已提交
6815 6816 6817 6818
{
	struct nfs4_exception exception = { };
	int err;
	do {
6819
		err = _nfs41_test_stateid(server, stateid, cred);
6820 6821 6822
		if (err != -NFS4ERR_DELAY)
			break;
		nfs4_handle_exception(server, err, &exception);
B
Bryan Schumaker 已提交
6823 6824 6825
	} while (exception.retry);
	return err;
}
B
Bryan Schumaker 已提交
6826

6827 6828 6829
struct nfs_free_stateid_data {
	struct nfs_server *server;
	struct nfs41_free_stateid_args args;
B
Bryan Schumaker 已提交
6830
	struct nfs41_free_stateid_res res;
6831 6832 6833 6834 6835 6836 6837 6838 6839 6840 6841 6842 6843 6844 6845 6846 6847 6848 6849 6850 6851 6852 6853 6854 6855 6856 6857 6858 6859 6860 6861 6862 6863 6864 6865 6866 6867
};

static void nfs41_free_stateid_prepare(struct rpc_task *task, void *calldata)
{
	struct nfs_free_stateid_data *data = calldata;
	nfs41_setup_sequence(nfs4_get_session(data->server),
			&data->args.seq_args,
			&data->res.seq_res,
			task);
}

static void nfs41_free_stateid_done(struct rpc_task *task, void *calldata)
{
	struct nfs_free_stateid_data *data = calldata;

	nfs41_sequence_done(task, &data->res.seq_res);

	switch (task->tk_status) {
	case -NFS4ERR_DELAY:
		if (nfs4_async_handle_error(task, data->server, NULL) == -EAGAIN)
			rpc_restart_call_prepare(task);
	}
}

static void nfs41_free_stateid_release(void *calldata)
{
	kfree(calldata);
}

const struct rpc_call_ops nfs41_free_stateid_ops = {
	.rpc_call_prepare = nfs41_free_stateid_prepare,
	.rpc_call_done = nfs41_free_stateid_done,
	.rpc_release = nfs41_free_stateid_release,
};

static struct rpc_task *_nfs41_free_stateid(struct nfs_server *server,
		nfs4_stateid *stateid,
6868
		struct rpc_cred *cred,
6869 6870
		bool privileged)
{
B
Bryan Schumaker 已提交
6871 6872
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_FREE_STATEID],
6873
		.rpc_cred = cred,
B
Bryan Schumaker 已提交
6874
	};
6875 6876 6877 6878 6879 6880 6881
	struct rpc_task_setup task_setup = {
		.rpc_client = server->client,
		.rpc_message = &msg,
		.callback_ops = &nfs41_free_stateid_ops,
		.flags = RPC_TASK_ASYNC,
	};
	struct nfs_free_stateid_data *data;
B
Bryan Schumaker 已提交
6882

6883
	dprintk("NFS call  free_stateid %p\n", stateid);
6884 6885 6886 6887 6888 6889 6890 6891 6892 6893 6894 6895 6896 6897 6898
	data = kmalloc(sizeof(*data), GFP_NOFS);
	if (!data)
		return ERR_PTR(-ENOMEM);
	data->server = server;
	nfs4_stateid_copy(&data->args.stateid, stateid);

	task_setup.callback_data = data;

	msg.rpc_argp = &data->args;
	msg.rpc_resp = &data->res;
	nfs41_init_sequence(&data->args.seq_args, &data->res.seq_res, 0);
	if (privileged)
		nfs4_set_sequence_privileged(&data->args.seq_args);

	return rpc_run_task(&task_setup);
B
Bryan Schumaker 已提交
6899 6900
}

6901 6902 6903 6904 6905
/**
 * nfs41_free_stateid - perform a FREE_STATEID operation
 *
 * @server: server / transport on which to perform the operation
 * @stateid: state ID to release
6906
 * @cred: credential
6907 6908 6909 6910
 *
 * Returns NFS_OK if the server freed "stateid".  Otherwise a
 * negative NFS4ERR value is returned.
 */
6911 6912 6913
static int nfs41_free_stateid(struct nfs_server *server,
		nfs4_stateid *stateid,
		struct rpc_cred *cred)
B
Bryan Schumaker 已提交
6914
{
6915 6916 6917
	struct rpc_task *task;
	int ret;

6918
	task = _nfs41_free_stateid(server, stateid, cred, true);
6919 6920 6921 6922 6923 6924 6925
	if (IS_ERR(task))
		return PTR_ERR(task);
	ret = rpc_wait_for_completion_task(task);
	if (!ret)
		ret = task->tk_status;
	rpc_put_task(task);
	return ret;
B
Bryan Schumaker 已提交
6926
}
6927

6928 6929 6930
static int nfs41_free_lock_state(struct nfs_server *server, struct nfs4_lock_state *lsp)
{
	struct rpc_task *task;
6931
	struct rpc_cred *cred = lsp->ls_state->owner->so_cred;
6932

6933
	task = _nfs41_free_stateid(server, &lsp->ls_stateid, cred, false);
6934 6935 6936 6937 6938 6939 6940
	nfs4_free_lock_state(server, lsp);
	if (IS_ERR(task))
		return PTR_ERR(task);
	rpc_put_task(task);
	return 0;
}

6941 6942 6943
static bool nfs41_match_stateid(const nfs4_stateid *s1,
		const nfs4_stateid *s2)
{
6944
	if (memcmp(s1->other, s2->other, sizeof(s1->other)) != 0)
6945 6946
		return false;

6947
	if (s1->seqid == s2->seqid)
6948
		return true;
6949
	if (s1->seqid == 0 || s2->seqid == 0)
6950 6951 6952 6953 6954
		return true;

	return false;
}

6955 6956
#endif /* CONFIG_NFS_V4_1 */

6957 6958 6959
static bool nfs4_match_stateid(const nfs4_stateid *s1,
		const nfs4_stateid *s2)
{
6960
	return nfs4_stateid_match(s1, s2);
6961 6962 6963
}


6964
static const struct nfs4_state_recovery_ops nfs40_reboot_recovery_ops = {
6965
	.owner_flag_bit = NFS_OWNER_RECLAIM_REBOOT,
6966
	.state_flag_bit	= NFS_STATE_RECLAIM_REBOOT,
L
Linus Torvalds 已提交
6967 6968
	.recover_open	= nfs4_open_reclaim,
	.recover_lock	= nfs4_lock_reclaim,
6969
	.establish_clid = nfs4_init_clientid,
6970
	.get_clid_cred	= nfs4_get_setclientid_cred,
6971
	.detect_trunking = nfs40_discover_server_trunking,
L
Linus Torvalds 已提交
6972 6973
};

6974
#if defined(CONFIG_NFS_V4_1)
6975
static const struct nfs4_state_recovery_ops nfs41_reboot_recovery_ops = {
6976 6977 6978 6979
	.owner_flag_bit = NFS_OWNER_RECLAIM_REBOOT,
	.state_flag_bit	= NFS_STATE_RECLAIM_REBOOT,
	.recover_open	= nfs4_open_reclaim,
	.recover_lock	= nfs4_lock_reclaim,
6980
	.establish_clid = nfs41_init_clientid,
6981
	.get_clid_cred	= nfs4_get_exchange_id_cred,
6982
	.reclaim_complete = nfs41_proc_reclaim_complete,
6983
	.detect_trunking = nfs41_discover_server_trunking,
6984 6985 6986
};
#endif /* CONFIG_NFS_V4_1 */

6987
static const struct nfs4_state_recovery_ops nfs40_nograce_recovery_ops = {
6988 6989 6990 6991 6992
	.owner_flag_bit = NFS_OWNER_RECLAIM_NOGRACE,
	.state_flag_bit	= NFS_STATE_RECLAIM_NOGRACE,
	.recover_open	= nfs4_open_expired,
	.recover_lock	= nfs4_lock_expired,
	.establish_clid = nfs4_init_clientid,
6993
	.get_clid_cred	= nfs4_get_setclientid_cred,
6994 6995 6996
};

#if defined(CONFIG_NFS_V4_1)
6997
static const struct nfs4_state_recovery_ops nfs41_nograce_recovery_ops = {
6998
	.owner_flag_bit = NFS_OWNER_RECLAIM_NOGRACE,
6999
	.state_flag_bit	= NFS_STATE_RECLAIM_NOGRACE,
7000 7001
	.recover_open	= nfs41_open_expired,
	.recover_lock	= nfs41_lock_expired,
7002
	.establish_clid = nfs41_init_clientid,
7003
	.get_clid_cred	= nfs4_get_exchange_id_cred,
L
Linus Torvalds 已提交
7004
};
7005
#endif /* CONFIG_NFS_V4_1 */
L
Linus Torvalds 已提交
7006

7007
static const struct nfs4_state_maintenance_ops nfs40_state_renewal_ops = {
B
Benny Halevy 已提交
7008
	.sched_state_renewal = nfs4_proc_async_renew,
7009
	.get_state_renewal_cred_locked = nfs4_get_renew_cred_locked,
7010
	.renew_lease = nfs4_proc_renew,
B
Benny Halevy 已提交
7011 7012 7013
};

#if defined(CONFIG_NFS_V4_1)
7014
static const struct nfs4_state_maintenance_ops nfs41_state_renewal_ops = {
B
Benny Halevy 已提交
7015
	.sched_state_renewal = nfs41_proc_async_sequence,
7016
	.get_state_renewal_cred_locked = nfs4_get_machine_cred_locked,
7017
	.renew_lease = nfs4_proc_sequence,
B
Benny Halevy 已提交
7018 7019 7020
};
#endif

7021 7022
static const struct nfs4_minor_version_ops nfs_v4_0_minor_ops = {
	.minor_version = 0,
7023 7024 7025 7026
	.init_caps = NFS_CAP_READDIRPLUS
		| NFS_CAP_ATOMIC_OPEN
		| NFS_CAP_CHANGE_ATTR
		| NFS_CAP_POSIX_LOCK,
7027
	.call_sync = _nfs4_call_sync,
7028
	.match_stateid = nfs4_match_stateid,
7029
	.find_root_sec = nfs4_find_root_sec,
7030
	.free_lock_state = nfs4_release_lockowner,
7031 7032 7033
	.reboot_recovery_ops = &nfs40_reboot_recovery_ops,
	.nograce_recovery_ops = &nfs40_nograce_recovery_ops,
	.state_renewal_ops = &nfs40_state_renewal_ops,
7034 7035 7036 7037 7038
};

#if defined(CONFIG_NFS_V4_1)
static const struct nfs4_minor_version_ops nfs_v4_1_minor_ops = {
	.minor_version = 1,
7039 7040 7041
	.init_caps = NFS_CAP_READDIRPLUS
		| NFS_CAP_ATOMIC_OPEN
		| NFS_CAP_CHANGE_ATTR
7042
		| NFS_CAP_POSIX_LOCK
7043 7044
		| NFS_CAP_STATEID_NFSV41
		| NFS_CAP_ATOMIC_OPEN_V1,
7045
	.call_sync = nfs4_call_sync_sequence,
7046
	.match_stateid = nfs41_match_stateid,
7047
	.find_root_sec = nfs41_find_root_sec,
7048
	.free_lock_state = nfs41_free_lock_state,
7049 7050 7051
	.reboot_recovery_ops = &nfs41_reboot_recovery_ops,
	.nograce_recovery_ops = &nfs41_nograce_recovery_ops,
	.state_renewal_ops = &nfs41_state_renewal_ops,
7052 7053 7054 7055 7056 7057 7058 7059 7060 7061
};
#endif

const struct nfs4_minor_version_ops *nfs_v4_minor_ops[] = {
	[0] = &nfs_v4_0_minor_ops,
#if defined(CONFIG_NFS_V4_1)
	[1] = &nfs_v4_1_minor_ops,
#endif
};

7062 7063 7064 7065 7066 7067 7068 7069 7070 7071 7072 7073 7074 7075 7076 7077 7078 7079 7080 7081
const struct inode_operations nfs4_dir_inode_operations = {
	.create		= nfs_create,
	.lookup		= nfs_lookup,
	.atomic_open	= nfs_atomic_open,
	.link		= nfs_link,
	.unlink		= nfs_unlink,
	.symlink	= nfs_symlink,
	.mkdir		= nfs_mkdir,
	.rmdir		= nfs_rmdir,
	.mknod		= nfs_mknod,
	.rename		= nfs_rename,
	.permission	= nfs_permission,
	.getattr	= nfs_getattr,
	.setattr	= nfs_setattr,
	.getxattr	= generic_getxattr,
	.setxattr	= generic_setxattr,
	.listxattr	= generic_listxattr,
	.removexattr	= generic_removexattr,
};

7082
static const struct inode_operations nfs4_file_inode_operations = {
7083 7084 7085
	.permission	= nfs_permission,
	.getattr	= nfs_getattr,
	.setattr	= nfs_setattr,
7086 7087 7088 7089
	.getxattr	= generic_getxattr,
	.setxattr	= generic_setxattr,
	.listxattr	= generic_listxattr,
	.removexattr	= generic_removexattr,
7090 7091
};

D
David Howells 已提交
7092
const struct nfs_rpc_ops nfs_v4_clientops = {
L
Linus Torvalds 已提交
7093 7094 7095
	.version	= 4,			/* protocol version */
	.dentry_ops	= &nfs4_dentry_operations,
	.dir_inode_ops	= &nfs4_dir_inode_operations,
7096
	.file_inode_ops	= &nfs4_file_inode_operations,
7097
	.file_ops	= &nfs4_file_operations,
L
Linus Torvalds 已提交
7098
	.getroot	= nfs4_proc_get_root,
B
Bryan Schumaker 已提交
7099
	.submount	= nfs4_submount,
B
Bryan Schumaker 已提交
7100
	.try_mount	= nfs4_try_mount,
L
Linus Torvalds 已提交
7101 7102 7103 7104 7105 7106 7107 7108
	.getattr	= nfs4_proc_getattr,
	.setattr	= nfs4_proc_setattr,
	.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,
7109
	.unlink_rpc_prepare = nfs4_proc_unlink_rpc_prepare,
L
Linus Torvalds 已提交
7110 7111
	.unlink_done	= nfs4_proc_unlink_done,
	.rename		= nfs4_proc_rename,
7112
	.rename_setup	= nfs4_proc_rename_setup,
7113
	.rename_rpc_prepare = nfs4_proc_rename_rpc_prepare,
7114
	.rename_done	= nfs4_proc_rename_done,
L
Linus Torvalds 已提交
7115 7116 7117 7118 7119 7120 7121 7122 7123
	.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,
7124
	.set_capabilities = nfs4_server_capabilities,
L
Linus Torvalds 已提交
7125 7126
	.decode_dirent	= nfs4_decode_dirent,
	.read_setup	= nfs4_proc_read_setup,
7127
	.read_pageio_init = pnfs_pageio_init_read,
7128
	.read_rpc_prepare = nfs4_proc_read_rpc_prepare,
T
Trond Myklebust 已提交
7129
	.read_done	= nfs4_read_done,
L
Linus Torvalds 已提交
7130
	.write_setup	= nfs4_proc_write_setup,
7131
	.write_pageio_init = pnfs_pageio_init_write,
7132
	.write_rpc_prepare = nfs4_proc_write_rpc_prepare,
7133
	.write_done	= nfs4_write_done,
L
Linus Torvalds 已提交
7134
	.commit_setup	= nfs4_proc_commit_setup,
7135
	.commit_rpc_prepare = nfs4_proc_commit_rpc_prepare,
7136
	.commit_done	= nfs4_commit_done,
L
Linus Torvalds 已提交
7137
	.lock		= nfs4_proc_lock,
7138
	.clear_acl_cache = nfs4_zap_acl_attr,
T
Trond Myklebust 已提交
7139
	.close_context  = nfs4_close_context,
7140
	.open_context	= nfs4_atomic_open,
7141
	.have_delegation = nfs4_have_delegation,
7142
	.return_delegation = nfs4_inode_return_delegation,
7143
	.alloc_client	= nfs4_alloc_client,
7144
	.init_client	= nfs4_init_client,
7145
	.free_client	= nfs4_free_client,
7146 7147
	.create_server	= nfs4_create_server,
	.clone_server	= nfs_clone_server,
L
Linus Torvalds 已提交
7148 7149
};

7150 7151 7152 7153 7154 7155 7156 7157 7158 7159 7160 7161
static const struct xattr_handler nfs4_xattr_nfs4_acl_handler = {
	.prefix	= XATTR_NAME_NFSV4_ACL,
	.list	= nfs4_xattr_list_nfs4_acl,
	.get	= nfs4_xattr_get_nfs4_acl,
	.set	= nfs4_xattr_set_nfs4_acl,
};

const struct xattr_handler *nfs4_xattr_handlers[] = {
	&nfs4_xattr_nfs4_acl_handler,
	NULL
};

L
Linus Torvalds 已提交
7162 7163 7164 7165 7166
/*
 * Local variables:
 *  c-basic-offset: 8
 * End:
 */