nfs4proc.c 201.2 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 81
static int nfs4_proc_getattr(struct nfs_server *, struct nfs_fh *, struct nfs_fattr *, struct nfs4_label *label);
static int _nfs4_proc_getattr(struct nfs_server *server, struct nfs_fh *fhandle, struct nfs_fattr *fattr, struct nfs4_label *label);
82 83
static int nfs4_do_setattr(struct inode *inode, struct rpc_cred *cred,
			    struct nfs_fattr *fattr, struct iattr *sattr,
84 85
			    struct nfs4_state *state, struct nfs4_label *ilabel,
			    struct nfs4_label *olabel);
86
#ifdef CONFIG_NFS_V4_1
87 88 89 90
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 *);
91
#endif
92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141

#ifdef CONFIG_NFS_V4_SECURITY_LABEL
static inline struct nfs4_label *
nfs4_label_init_security(struct inode *dir, struct dentry *dentry,
	struct iattr *sattr, struct nfs4_label *label)
{
	int err;

	if (label == NULL)
		return NULL;

	if (nfs_server_capable(dir, NFS_CAP_SECURITY_LABEL) == 0)
		return NULL;

	if (NFS_SERVER(dir)->nfs_client->cl_minorversion < 2)
		return NULL;

	err = security_dentry_init_security(dentry, sattr->ia_mode,
				&dentry->d_name, (void **)&label->label, &label->len);
	if (err == 0)
		return label;

	return NULL;
}
static inline void
nfs4_label_release_security(struct nfs4_label *label)
{
	if (label)
		security_release_secctx(label->label, label->len);
}
static inline u32 *nfs4_bitmask(struct nfs_server *server, struct nfs4_label *label)
{
	if (label)
		return server->attr_bitmask;

	return server->attr_bitmask_nl;
}
#else
static inline struct nfs4_label *
nfs4_label_init_security(struct inode *dir, struct dentry *dentry,
	struct iattr *sattr, struct nfs4_label *l)
{ return NULL; }
static inline void
nfs4_label_release_security(struct nfs4_label *label)
{ return; }
static inline u32 *
nfs4_bitmask(struct nfs_server *server, struct nfs4_label *label)
{ return server->attr_bitmask; }
#endif

L
Linus Torvalds 已提交
142
/* Prevent leaks of NFSv4 errors into userland */
143
static int nfs4_map_errors(int err)
L
Linus Torvalds 已提交
144
{
145 146 147 148
	if (err >= -1000)
		return err;
	switch (err) {
	case -NFS4ERR_RESOURCE:
149 150
	case -NFS4ERR_LAYOUTTRYLATER:
	case -NFS4ERR_RECALLCONFLICT:
151
		return -EREMOTEIO;
152 153
	case -NFS4ERR_WRONGSEC:
		return -EPERM;
154 155 156
	case -NFS4ERR_BADOWNER:
	case -NFS4ERR_BADNAME:
		return -EINVAL;
157 158
	case -NFS4ERR_SHARE_DENIED:
		return -EACCES;
159 160
	case -NFS4ERR_MINOR_VERS_MISMATCH:
		return -EPROTONOSUPPORT;
161 162
	case -NFS4ERR_ACCESS:
		return -EACCES;
163 164
	case -NFS4ERR_FILE_OPEN:
		return -EBUSY;
165
	default:
L
Linus Torvalds 已提交
166
		dprintk("%s could not handle NFSv4 error %d\n",
167
				__func__, -err);
168
		break;
L
Linus Torvalds 已提交
169
	}
170
	return -EIO;
L
Linus Torvalds 已提交
171 172 173 174 175
}

/*
 * This is our standard bitmap for GETATTR requests.
 */
176
const u32 nfs4_fattr_bitmap[3] = {
L
Linus Torvalds 已提交
177 178 179 180 181 182 183 184 185 186 187 188 189
	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
190 191 192 193
	| FATTR4_WORD1_TIME_MODIFY,
#ifdef CONFIG_NFS_V4_SECURITY_LABEL
	FATTR4_WORD2_SECURITY_LABEL
#endif
L
Linus Torvalds 已提交
194 195
};

196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213
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
};

214 215 216 217 218 219
static const u32 nfs4_open_noattr_bitmap[3] = {
	FATTR4_WORD0_TYPE
	| FATTR4_WORD0_CHANGE
	| FATTR4_WORD0_FILEID,
};

220
const u32 nfs4_statfs_bitmap[3] = {
L
Linus Torvalds 已提交
221 222 223 224 225 226 227 228
	FATTR4_WORD0_FILES_AVAIL
	| FATTR4_WORD0_FILES_FREE
	| FATTR4_WORD0_FILES_TOTAL,
	FATTR4_WORD1_SPACE_AVAIL
	| FATTR4_WORD1_SPACE_FREE
	| FATTR4_WORD1_SPACE_TOTAL
};

229
const u32 nfs4_pathconf_bitmap[3] = {
L
Linus Torvalds 已提交
230 231 232 233 234
	FATTR4_WORD0_MAXLINK
	| FATTR4_WORD0_MAXNAME,
	0
};

235
const u32 nfs4_fsinfo_bitmap[3] = { FATTR4_WORD0_MAXFILESIZE
L
Linus Torvalds 已提交
236 237 238
			| FATTR4_WORD0_MAXREAD
			| FATTR4_WORD0_MAXWRITE
			| FATTR4_WORD0_LEASE_TIME,
239
			FATTR4_WORD1_TIME_DELTA
240 241
			| FATTR4_WORD1_FS_LAYOUT_TYPES,
			FATTR4_WORD2_LAYOUT_BLKSIZE
L
Linus Torvalds 已提交
242 243
};

244
const u32 nfs4_fs_locations_bitmap[3] = {
245 246 247 248 249 250 251 252 253 254 255 256 257 258 259
	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
260
	| FATTR4_WORD1_MOUNTED_ON_FILEID,
261 262
};

A
Al Viro 已提交
263
static void nfs4_setup_readdir(u64 cookie, __be32 *verifier, struct dentry *dentry,
L
Linus Torvalds 已提交
264 265
		struct nfs4_readdir_arg *readdir)
{
266
	__be32 *start, *p;
L
Linus Torvalds 已提交
267 268

	if (cookie > 2) {
269
		readdir->cookie = cookie;
L
Linus Torvalds 已提交
270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285
		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.
	 */
286
	start = p = kmap_atomic(*readdir->pages);
L
Linus Torvalds 已提交
287 288 289 290 291 292 293 294 295 296 297
	
	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
298
		p = xdr_encode_hyper(p, NFS_FILEID(dentry->d_inode));
L
Linus Torvalds 已提交
299 300 301 302 303 304 305 306 307 308 309
	}
	
	*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
310
	p = xdr_encode_hyper(p, NFS_FILEID(dentry->d_parent->d_inode));
L
Linus Torvalds 已提交
311 312 313

	readdir->pgbase = (char *)p - (char *)start;
	readdir->count -= readdir->pgbase;
314
	kunmap_atomic(start);
L
Linus Torvalds 已提交
315 316
}

317 318 319 320 321 322 323 324 325 326
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;
327
	freezable_schedule_timeout_killable_unsafe(*timeout);
328 329 330 331 332 333 334 335 336
	if (fatal_signal_pending(current))
		res = -ERESTARTSYS;
	*timeout <<= 1;
	return res;
}

/* This is the error handling routine for processes that are allowed
 * to sleep.
 */
337
static int nfs4_handle_exception(struct nfs_server *server, int errorcode, struct nfs4_exception *exception)
338 339
{
	struct nfs_client *clp = server->nfs_client;
340
	struct nfs4_state *state = exception->state;
341
	struct inode *inode = exception->inode;
342 343 344 345 346 347
	int ret = errorcode;

	exception->retry = 0;
	switch(errorcode) {
		case 0:
			return 0;
348
		case -NFS4ERR_OPENMODE:
349
			if (inode && nfs4_have_delegation(inode, FMODE_READ)) {
350
				nfs4_inode_return_delegation(inode);
351 352 353 354 355
				exception->retry = 1;
				return 0;
			}
			if (state == NULL)
				break;
356 357 358
			ret = nfs4_schedule_stateid_recovery(server, state);
			if (ret < 0)
				break;
359
			goto wait_on_recovery;
360
		case -NFS4ERR_DELEG_REVOKED:
361 362
		case -NFS4ERR_ADMIN_REVOKED:
		case -NFS4ERR_BAD_STATEID:
363 364 365 366 367
			if (inode != NULL && nfs4_have_delegation(inode, FMODE_READ)) {
				nfs_remove_bad_delegation(inode);
				exception->retry = 1;
				break;
			}
368 369
			if (state == NULL)
				break;
370 371 372
			ret = nfs4_schedule_stateid_recovery(server, state);
			if (ret < 0)
				break;
373
			goto wait_on_recovery;
374
		case -NFS4ERR_EXPIRED:
375 376 377 378 379
			if (state != NULL) {
				ret = nfs4_schedule_stateid_recovery(server, state);
				if (ret < 0)
					break;
			}
380
		case -NFS4ERR_STALE_STATEID:
381
		case -NFS4ERR_STALE_CLIENTID:
382 383
			nfs4_schedule_lease_recovery(clp);
			goto wait_on_recovery;
384
#if defined(CONFIG_NFS_V4_1)
385 386 387 388 389 390 391 392 393
		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);
394
			nfs4_schedule_session_recovery(clp->cl_session, errorcode);
395
			goto wait_on_recovery;
396
#endif /* defined(CONFIG_NFS_V4_1) */
397
		case -NFS4ERR_FILE_OPEN:
398 399 400 401 402 403 404
			if (exception->timeout > HZ) {
				/* We have retried a decent amount, time to
				 * fail
				 */
				ret = -EBUSY;
				break;
			}
405 406 407 408 409
		case -NFS4ERR_GRACE:
		case -NFS4ERR_DELAY:
			ret = nfs4_delay(server->client, &exception->timeout);
			if (ret != 0)
				break;
410
		case -NFS4ERR_RETRY_UNCACHED_REP:
411 412
		case -NFS4ERR_OLD_STATEID:
			exception->retry = 1;
413 414 415 416 417 418 419 420 421 422 423 424 425
			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);
			}
426 427 428
	}
	/* We failed to handle the error */
	return nfs4_map_errors(ret);
429
wait_on_recovery:
430 431 432 433
	ret = nfs4_wait_clnt_recover(clp);
	if (ret == 0)
		exception->retry = 1;
	return ret;
434 435 436
}


437
static void do_renew_lease(struct nfs_client *clp, unsigned long timestamp)
L
Linus Torvalds 已提交
438 439 440 441 442 443 444
{
	spin_lock(&clp->cl_lock);
	if (time_before(clp->cl_last_renewal,timestamp))
		clp->cl_last_renewal = timestamp;
	spin_unlock(&clp->cl_lock);
}

445 446 447 448 449
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
450 451
#if defined(CONFIG_NFS_V4_1)

452
static void nfs41_sequence_free_slot(struct nfs4_sequence_res *res)
A
Andy Adamson 已提交
453
{
454
	struct nfs4_session *session;
A
Andy Adamson 已提交
455
	struct nfs4_slot_table *tbl;
456
	bool send_new_highest_used_slotid = false;
A
Andy Adamson 已提交
457

458
	if (!res->sr_slot) {
A
Andy Adamson 已提交
459 460
		/* just wake up the next guy waiting since
		 * we may have not consumed a slot after all */
A
Andy Adamson 已提交
461
		dprintk("%s: No slot\n", __func__);
462
		return;
A
Andy Adamson 已提交
463
	}
464 465
	tbl = res->sr_slot->table;
	session = tbl->session;
466

467
	spin_lock(&tbl->slot_tbl_lock);
468 469 470 471 472 473
	/* 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;

474 475 476 477
	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
478
	nfs4_free_slot(tbl, res->sr_slot);
479 480 481

	if (tbl->highest_used_slotid != NFS4_NO_SLOT)
		send_new_highest_used_slotid = false;
482
out_unlock:
483
	spin_unlock(&tbl->slot_tbl_lock);
484
	res->sr_slot = NULL;
485 486
	if (send_new_highest_used_slotid)
		nfs41_server_notify_highest_slotid_update(session->clp);
A
Andy Adamson 已提交
487 488
}

489
static int nfs41_sequence_done(struct rpc_task *task, struct nfs4_sequence_res *res)
A
Andy Adamson 已提交
490
{
491
	struct nfs4_session *session;
492
	struct nfs4_slot *slot;
493
	struct nfs_client *clp;
494
	bool interrupted = false;
495
	int ret = 1;
A
Andy Adamson 已提交
496

497 498
	/* don't increment the sequence number if the task wasn't sent */
	if (!RPC_WAS_SENT(task))
A
Andy Adamson 已提交
499 500
		goto out;

501
	slot = res->sr_slot;
502
	session = slot->table->session;
503

504 505 506 507 508
	if (slot->interrupted) {
		slot->interrupted = 0;
		interrupted = true;
	}

A
Andy Adamson 已提交
509
	/* Check the SEQUENCE operation status */
510 511
	switch (res->sr_status) {
	case 0:
A
Andy Adamson 已提交
512
		/* Update the slot's sequence and clientid lease timer */
513
		++slot->seq_nr;
514
		clp = session->clp;
515
		do_renew_lease(clp, res->sr_timestamp);
516
		/* Check sequence flags */
517 518
		if (res->sr_status_flags != 0)
			nfs4_schedule_lease_recovery(clp);
519
		nfs41_update_target_slotid(slot->table, slot, res);
520
		break;
521 522 523 524 525 526 527 528 529
	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;
530 531 532 533 534
	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.
		 */
535
		dprintk("%s: slot=%u seq=%u: Operation in progress\n",
536
			__func__,
537
			slot->slot_nr,
538
			slot->seq_nr);
539
		goto out_retry;
540 541 542 543 544
	case -NFS4ERR_BADSLOT:
		/*
		 * The slot id we used was probably retired. Try again
		 * using a different slot id.
		 */
545 546
		goto retry_nowait;
	case -NFS4ERR_SEQ_MISORDERED:
547 548 549 550 551 552 553 554
		/*
		 * Was the last operation on this sequence interrupted?
		 * If so, retry after bumping the sequence number.
		 */
		if (interrupted) {
			++slot->seq_nr;
			goto retry_nowait;
		}
555 556 557 558
		/*
		 * Could this slot have been previously retired?
		 * If so, then the server may be expecting seq_nr = 1!
		 */
559 560 561 562 563
		if (slot->seq_nr != 1) {
			slot->seq_nr = 1;
			goto retry_nowait;
		}
		break;
564 565 566
	case -NFS4ERR_SEQ_FALSE_RETRY:
		++slot->seq_nr;
		goto retry_nowait;
567 568
	default:
		/* Just update the slot sequence no. */
569
		++slot->seq_nr;
A
Andy Adamson 已提交
570 571 572 573
	}
out:
	/* The session may be reset by one of the error handlers. */
	dprintk("%s: Error %d free the slot \n", __func__, res->sr_status);
574
	nfs41_sequence_free_slot(res);
575
	return ret;
576 577 578 579 580 581
retry_nowait:
	if (rpc_restart_call_prepare(task)) {
		task->tk_status = 0;
		ret = 0;
	}
	goto out;
582
out_retry:
583
	if (!rpc_restart_call(task))
584 585 586
		goto out;
	rpc_delay(task, NFS4_POLL_RETRY_MAX);
	return 0;
A
Andy Adamson 已提交
587 588
}

589 590
static int nfs4_sequence_done(struct rpc_task *task,
			       struct nfs4_sequence_res *res)
591
{
592
	if (res->sr_slot == NULL)
593 594
		return 1;
	return nfs41_sequence_done(task, res);
595 596
}

597 598 599
static void nfs41_init_sequence(struct nfs4_sequence_args *args,
		struct nfs4_sequence_res *res, int cache_reply)
{
600
	args->sa_slot = NULL;
601
	args->sa_cache_this = 0;
602
	args->sa_privileged = 0;
603 604 605 606 607
	if (cache_reply)
		args->sa_cache_this = 1;
	res->sr_slot = NULL;
}

608 609 610 611 612
static void nfs4_set_sequence_privileged(struct nfs4_sequence_args *args)
{
	args->sa_privileged = 1;
}

A
Andy Adamson 已提交
613
int nfs41_setup_sequence(struct nfs4_session *session,
A
Andy Adamson 已提交
614 615 616 617
				struct nfs4_sequence_args *args,
				struct nfs4_sequence_res *res,
				struct rpc_task *task)
{
A
Andy Adamson 已提交
618 619 620 621
	struct nfs4_slot *slot;
	struct nfs4_slot_table *tbl;

	dprintk("--> %s\n", __func__);
A
Andy Adamson 已提交
622
	/* slot already allocated? */
623
	if (res->sr_slot != NULL)
624
		goto out_success;
A
Andy Adamson 已提交
625

A
Andy Adamson 已提交
626 627
	tbl = &session->fc_slot_table;

628 629
	task->tk_timeout = 0;

A
Andy Adamson 已提交
630
	spin_lock(&tbl->slot_tbl_lock);
631
	if (test_bit(NFS4_SLOT_TBL_DRAINING, &tbl->slot_tbl_state) &&
632
	    !args->sa_privileged) {
633 634
		/* The state manager will wait until the slot table is empty */
		dprintk("%s session is draining\n", __func__);
635
		goto out_sleep;
636 637
	}

638
	slot = nfs4_alloc_slot(tbl);
639 640 641 642
	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
643
		dprintk("<-- %s: no free slots\n", __func__);
644
		goto out_sleep;
A
Andy Adamson 已提交
645 646 647
	}
	spin_unlock(&tbl->slot_tbl_lock);

648
	args->sa_slot = slot;
A
Andy Adamson 已提交
649

650 651
	dprintk("<-- %s slotid=%d seqid=%d\n", __func__,
			slot->slot_nr, slot->seq_nr);
A
Andy Adamson 已提交
652

653
	res->sr_slot = slot;
654
	res->sr_timestamp = jiffies;
655
	res->sr_status_flags = 0;
A
Andy Adamson 已提交
656 657 658 659 660
	/*
	 * 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;
661 662
out_success:
	rpc_call_start(task);
A
Andy Adamson 已提交
663
	return 0;
664
out_sleep:
665 666
	/* Privileged tasks are queued with top priority */
	if (args->sa_privileged)
667 668 669 670
		rpc_sleep_on_priority(&tbl->slot_tbl_waitq, task,
				NULL, RPC_PRIORITY_PRIVILEGED);
	else
		rpc_sleep_on(&tbl->slot_tbl_waitq, task, NULL);
671 672
	spin_unlock(&tbl->slot_tbl_lock);
	return -EAGAIN;
A
Andy Adamson 已提交
673
}
A
Andy Adamson 已提交
674
EXPORT_SYMBOL_GPL(nfs41_setup_sequence);
A
Andy Adamson 已提交
675

676
int nfs4_setup_sequence(const struct nfs_server *server,
A
Andy Adamson 已提交
677 678 679 680
			struct nfs4_sequence_args *args,
			struct nfs4_sequence_res *res,
			struct rpc_task *task)
{
681
	struct nfs4_session *session = nfs4_get_session(server);
A
Andy Adamson 已提交
682 683
	int ret = 0;

684 685
	if (session == NULL) {
		rpc_call_start(task);
686
		goto out;
687
	}
688

689
	dprintk("--> %s clp %p session %p sr_slot %d\n",
690
		__func__, session->clp, session, res->sr_slot ?
691
			res->sr_slot->slot_nr : -1);
A
Andy Adamson 已提交
692

693
	ret = nfs41_setup_sequence(session, args, res, task);
A
Andy Adamson 已提交
694 695 696 697 698 699
out:
	dprintk("<-- %s status=%d\n", __func__, ret);
	return ret;
}

struct nfs41_call_sync_data {
700
	const struct nfs_server *seq_server;
A
Andy Adamson 已提交
701 702 703 704 705 706 707
	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;
708
	struct nfs4_session *session = nfs4_get_session(data->seq_server);
A
Andy Adamson 已提交
709

710 711
	dprintk("--> %s data->seq_server %p\n", __func__, data->seq_server);

712
	nfs41_setup_sequence(session, data->seq_args, data->seq_res, task);
A
Andy Adamson 已提交
713 714
}

A
Andy Adamson 已提交
715 716 717 718
static void nfs41_call_sync_done(struct rpc_task *task, void *calldata)
{
	struct nfs41_call_sync_data *data = calldata;

719
	nfs41_sequence_done(task, data->seq_res);
A
Andy Adamson 已提交
720 721
}

722
static const struct rpc_call_ops nfs41_call_sync_ops = {
A
Andy Adamson 已提交
723
	.rpc_call_prepare = nfs41_call_sync_prepare,
A
Andy Adamson 已提交
724
	.rpc_call_done = nfs41_call_sync_done,
A
Andy Adamson 已提交
725 726
};

727 728
static int nfs4_call_sync_sequence(struct rpc_clnt *clnt,
				   struct nfs_server *server,
A
Andy Adamson 已提交
729 730
				   struct rpc_message *msg,
				   struct nfs4_sequence_args *args,
731
				   struct nfs4_sequence_res *res)
A
Andy Adamson 已提交
732 733 734 735
{
	int ret;
	struct rpc_task *task;
	struct nfs41_call_sync_data data = {
736
		.seq_server = server,
A
Andy Adamson 已提交
737 738 739 740
		.seq_args = args,
		.seq_res = res,
	};
	struct rpc_task_setup task_setup = {
741
		.rpc_client = clnt,
A
Andy Adamson 已提交
742 743 744 745 746 747 748 749 750 751 752 753 754 755 756
		.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;
}

757
#else
758
static
759 760 761 762 763
void nfs41_init_sequence(struct nfs4_sequence_args *args,
		struct nfs4_sequence_res *res, int cache_reply)
{
}

764 765 766 767 768
static void nfs4_set_sequence_privileged(struct nfs4_sequence_args *args)
{
}


769 770
static int nfs4_sequence_done(struct rpc_task *task,
			       struct nfs4_sequence_res *res)
771
{
772
	return 1;
773
}
A
Andy Adamson 已提交
774 775
#endif /* CONFIG_NFS_V4_1 */

776
static
777 778
int _nfs4_call_sync(struct rpc_clnt *clnt,
		    struct nfs_server *server,
A
Andy Adamson 已提交
779 780
		    struct rpc_message *msg,
		    struct nfs4_sequence_args *args,
781
		    struct nfs4_sequence_res *res)
A
Andy Adamson 已提交
782
{
783
	return rpc_call_sync(clnt, msg, 0);
A
Andy Adamson 已提交
784 785
}

786
static
787 788
int nfs4_call_sync(struct rpc_clnt *clnt,
		   struct nfs_server *server,
789 790 791 792 793
		   struct rpc_message *msg,
		   struct nfs4_sequence_args *args,
		   struct nfs4_sequence_res *res,
		   int cache_reply)
{
794
	nfs41_init_sequence(args, res, cache_reply);
795
	return server->nfs_client->cl_mvops->call_sync(clnt, server, msg,
796
						args, res);
797
}
A
Andy Adamson 已提交
798

799
static void update_changeattr(struct inode *dir, struct nfs4_change_info *cinfo)
L
Linus Torvalds 已提交
800
{
801
	struct nfs_inode *nfsi = NFS_I(dir);
L
Linus Torvalds 已提交
802

803
	spin_lock(&dir->i_lock);
804
	nfsi->cache_validity |= NFS_INO_INVALID_ATTR|NFS_INO_INVALID_DATA;
805
	if (!cinfo->atomic || cinfo->before != dir->i_version)
806
		nfs_force_lookup_revalidate(dir);
807
	dir->i_version = cinfo->after;
D
David Howells 已提交
808
	nfs_fscache_invalidate(dir);
809
	spin_unlock(&dir->i_lock);
L
Linus Torvalds 已提交
810 811
}

812
struct nfs4_opendata {
813
	struct kref kref;
814 815
	struct nfs_openargs o_arg;
	struct nfs_openres o_res;
816 817
	struct nfs_open_confirmargs c_arg;
	struct nfs_open_confirmres c_res;
818 819
	struct nfs4_string owner_name;
	struct nfs4_string group_name;
820
	struct nfs_fattr f_attr;
821
	struct nfs4_label *f_label;
822
	struct dentry *dir;
823
	struct dentry *dentry;
824
	struct nfs4_state_owner *owner;
825
	struct nfs4_state *state;
826
	struct iattr attrs;
827
	unsigned long timestamp;
828
	unsigned int rpc_done : 1;
829
	unsigned int is_recover : 1;
830 831
	int rpc_status;
	int cancelled;
832 833
};

834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862
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;
	}
}
863 864 865 866

static void nfs4_init_opendata_res(struct nfs4_opendata *p)
{
	p->o_res.f_attr = &p->f_attr;
867
	p->o_res.f_label = p->f_label;
868 869
	p->o_res.seqid = p->o_arg.seqid;
	p->c_res.seqid = p->c_arg.seqid;
870
	p->o_res.server = p->o_arg.server;
871
	p->o_res.access_request = p->o_arg.access;
872
	nfs_fattr_init(&p->f_attr);
873
	nfs_fattr_init_names(&p->f_attr, &p->owner_name, &p->group_name);
874 875
}

876
static struct nfs4_opendata *nfs4_opendata_alloc(struct dentry *dentry,
877
		struct nfs4_state_owner *sp, fmode_t fmode, int flags,
878
		const struct iattr *attrs,
879
		struct nfs4_label *label,
880
		enum open_claim_type4 claim,
881
		gfp_t gfp_mask)
882
{
883
	struct dentry *parent = dget_parent(dentry);
884 885 886 887
	struct inode *dir = parent->d_inode;
	struct nfs_server *server = NFS_SERVER(dir);
	struct nfs4_opendata *p;

888
	p = kzalloc(sizeof(*p), gfp_mask);
889 890
	if (p == NULL)
		goto err;
891 892 893 894 895

	p->f_label = nfs4_label_alloc(server, gfp_mask);
	if (IS_ERR(p->f_label))
		goto err_free_p;

896
	p->o_arg.seqid = nfs_alloc_seqid(&sp->so_seqid, gfp_mask);
897
	if (p->o_arg.seqid == NULL)
898
		goto err_free_label;
899 900
	nfs_sb_active(dentry->d_sb);
	p->dentry = dget(dentry);
901 902 903
	p->dir = parent;
	p->owner = sp;
	atomic_inc(&sp->so_count);
904 905
	p->o_arg.open_flags = flags;
	p->o_arg.fmode = fmode & (FMODE_READ|FMODE_WRITE);
906 907 908 909 910 911 912 913
	/* 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;
	}
914
	p->o_arg.clientid = server->nfs_client->cl_clientid;
915 916
	p->o_arg.id.create_time = ktime_to_ns(sp->so_seqid.create_time);
	p->o_arg.id.uniquifier = sp->so_seqid.owner_id;
917
	p->o_arg.name = &dentry->d_name;
918
	p->o_arg.server = server;
919
	p->o_arg.bitmask = nfs4_bitmask(server, label);
920
	p->o_arg.open_bitmap = &nfs4_fattr_bitmap[0];
921
	p->o_arg.label = label;
922 923
	p->o_arg.claim = nfs4_map_atomic_open_claim(server, claim);
	switch (p->o_arg.claim) {
924 925 926 927 928 929 930 931 932 933 934
	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);
	}
935
	if (attrs != NULL && attrs->ia_valid != 0) {
936
		__be32 verf[2];
937

938 939
		p->o_arg.u.attrs = &p->attrs;
		memcpy(&p->attrs, attrs, sizeof(p->attrs));
940 941 942 943 944

		verf[0] = jiffies;
		verf[1] = current->pid;
		memcpy(p->o_arg.u.verifier.data, verf,
				sizeof(p->o_arg.u.verifier.data));
945
	}
946 947 948
	p->c_arg.fh = &p->o_res.fh;
	p->c_arg.stateid = &p->o_res.stateid;
	p->c_arg.seqid = p->o_arg.seqid;
949
	nfs4_init_opendata_res(p);
950
	kref_init(&p->kref);
951
	return p;
952 953 954 955

err_free_label:
	nfs4_label_free(p->f_label);
err_free_p:
956 957 958 959 960 961
	kfree(p);
err:
	dput(parent);
	return NULL;
}

962
static void nfs4_opendata_free(struct kref *kref)
963
{
964 965
	struct nfs4_opendata *p = container_of(kref,
			struct nfs4_opendata, kref);
966
	struct super_block *sb = p->dentry->d_sb;
967 968

	nfs_free_seqid(p->o_arg.seqid);
969 970
	if (p->state != NULL)
		nfs4_put_open_state(p->state);
971
	nfs4_put_state_owner(p->owner);
972 973 974

	nfs4_label_free(p->f_label);

975
	dput(p->dir);
976 977
	dput(p->dentry);
	nfs_sb_deactive(sb);
978
	nfs_fattr_free_names(&p->f_attr);
979 980 981 982 983 984 985
	kfree(p);
}

static void nfs4_opendata_put(struct nfs4_opendata *p)
{
	if (p != NULL)
		kref_put(&p->kref, nfs4_opendata_free);
986 987
}

988 989 990 991 992 993 994 995
static int nfs4_wait_for_completion_rpc_task(struct rpc_task *task)
{
	int ret;

	ret = rpc_wait_for_completion_task(task);
	return ret;
}

996
static int can_open_cached(struct nfs4_state *state, fmode_t mode, int open_mode)
997 998
{
	int ret = 0;
999

1000
	if (open_mode & (O_EXCL|O_TRUNC))
1001 1002
		goto out;
	switch (mode & (FMODE_READ|FMODE_WRITE)) {
1003
		case FMODE_READ:
1004 1005
			ret |= test_bit(NFS_O_RDONLY_STATE, &state->flags) != 0
				&& state->n_rdonly != 0;
1006 1007
			break;
		case FMODE_WRITE:
1008 1009
			ret |= test_bit(NFS_O_WRONLY_STATE, &state->flags) != 0
				&& state->n_wronly != 0;
1010 1011
			break;
		case FMODE_READ|FMODE_WRITE:
1012 1013
			ret |= test_bit(NFS_O_RDWR_STATE, &state->flags) != 0
				&& state->n_rdwr != 0;
1014
	}
1015
out:
1016 1017 1018
	return ret;
}

1019
static int can_open_delegated(struct nfs_delegation *delegation, fmode_t fmode)
1020
{
1021 1022
	if (delegation == NULL)
		return 0;
1023
	if ((delegation->type & fmode) != fmode)
1024
		return 0;
1025
	if (test_bit(NFS_DELEGATION_NEED_RECLAIM, &delegation->flags))
1026
		return 0;
1027 1028
	if (test_bit(NFS_DELEGATION_RETURNING, &delegation->flags))
		return 0;
1029
	nfs_mark_delegation_referenced(delegation);
1030 1031 1032
	return 1;
}

1033
static void update_open_stateflags(struct nfs4_state *state, fmode_t fmode)
1034
{
1035
	switch (fmode) {
1036 1037 1038 1039 1040 1041 1042 1043 1044
		case FMODE_WRITE:
			state->n_wronly++;
			break;
		case FMODE_READ:
			state->n_rdonly++;
			break;
		case FMODE_READ|FMODE_WRITE:
			state->n_rdwr++;
	}
1045
	nfs4_state_set_mode_locked(state, state->state | fmode);
1046 1047
}

1048
static void nfs_set_open_stateid_locked(struct nfs4_state *state, nfs4_stateid *stateid, fmode_t fmode)
1049 1050
{
	if (test_bit(NFS_DELEGATED_STATE, &state->flags) == 0)
1051 1052
		nfs4_stateid_copy(&state->stateid, stateid);
	nfs4_stateid_copy(&state->open_stateid, stateid);
1053
	set_bit(NFS_OPEN_STATE, &state->flags);
1054
	switch (fmode) {
1055 1056 1057 1058 1059 1060 1061 1062 1063
		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);
	}
1064 1065
}

1066
static void nfs_set_open_stateid(struct nfs4_state *state, nfs4_stateid *stateid, fmode_t fmode)
1067
{
1068
	write_seqlock(&state->seqlock);
1069
	nfs_set_open_stateid_locked(state, stateid, fmode);
1070
	write_sequnlock(&state->seqlock);
1071 1072
}

1073
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
1074
{
1075 1076 1077 1078 1079
	/*
	 * Protect the call to nfs4_state_set_mode_locked and
	 * serialise the stateid update
	 */
	write_seqlock(&state->seqlock);
1080
	if (deleg_stateid != NULL) {
1081
		nfs4_stateid_copy(&state->stateid, deleg_stateid);
1082 1083 1084
		set_bit(NFS_DELEGATED_STATE, &state->flags);
	}
	if (open_stateid != NULL)
1085
		nfs_set_open_stateid_locked(state, open_stateid, fmode);
1086 1087
	write_sequnlock(&state->seqlock);
	spin_lock(&state->owner->so_lock);
1088
	update_open_stateflags(state, fmode);
1089
	spin_unlock(&state->owner->so_lock);
L
Linus Torvalds 已提交
1090 1091
}

1092
static int update_open_stateid(struct nfs4_state *state, nfs4_stateid *open_stateid, nfs4_stateid *delegation, fmode_t fmode)
1093 1094 1095 1096 1097
{
	struct nfs_inode *nfsi = NFS_I(state->inode);
	struct nfs_delegation *deleg_cur;
	int ret = 0;

1098
	fmode &= (FMODE_READ|FMODE_WRITE);
1099 1100 1101 1102 1103 1104 1105 1106

	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 ||
1107
	   test_bit(NFS_DELEGATION_RETURNING, &deleg_cur->flags) ||
1108
	    (deleg_cur->type & fmode) != fmode)
1109 1110 1111 1112
		goto no_delegation_unlock;

	if (delegation == NULL)
		delegation = &deleg_cur->stateid;
1113
	else if (!nfs4_stateid_match(&deleg_cur->stateid, delegation))
1114 1115
		goto no_delegation_unlock;

1116
	nfs_mark_delegation_referenced(deleg_cur);
1117
	__update_open_stateid(state, open_stateid, &deleg_cur->stateid, fmode);
1118 1119 1120 1121 1122 1123 1124
	ret = 1;
no_delegation_unlock:
	spin_unlock(&deleg_cur->lock);
no_delegation:
	rcu_read_unlock();

	if (!ret && open_stateid != NULL) {
1125
		__update_open_stateid(state, open_stateid, NULL, fmode);
1126 1127 1128 1129 1130 1131 1132
		ret = 1;
	}

	return ret;
}


1133
static void nfs4_return_incompatible_delegation(struct inode *inode, fmode_t fmode)
1134 1135 1136 1137 1138
{
	struct nfs_delegation *delegation;

	rcu_read_lock();
	delegation = rcu_dereference(NFS_I(inode)->delegation);
1139
	if (delegation == NULL || (delegation->type & fmode) == fmode) {
1140 1141 1142 1143
		rcu_read_unlock();
		return;
	}
	rcu_read_unlock();
1144
	nfs4_inode_return_delegation(inode);
1145 1146
}

1147
static struct nfs4_state *nfs4_try_open_cached(struct nfs4_opendata *opendata)
1148 1149 1150 1151
{
	struct nfs4_state *state = opendata->state;
	struct nfs_inode *nfsi = NFS_I(state->inode);
	struct nfs_delegation *delegation;
1152
	int open_mode = opendata->o_arg.open_flags;
1153
	fmode_t fmode = opendata->o_arg.fmode;
1154 1155 1156 1157
	nfs4_stateid stateid;
	int ret = -EAGAIN;

	for (;;) {
1158
		if (can_open_cached(state, fmode, open_mode)) {
1159
			spin_lock(&state->owner->so_lock);
1160 1161
			if (can_open_cached(state, fmode, open_mode)) {
				update_open_stateflags(state, fmode);
1162 1163 1164 1165 1166
				spin_unlock(&state->owner->so_lock);
				goto out_return_state;
			}
			spin_unlock(&state->owner->so_lock);
		}
1167 1168
		rcu_read_lock();
		delegation = rcu_dereference(nfsi->delegation);
1169
		if (!can_open_delegated(delegation, fmode)) {
1170
			rcu_read_unlock();
1171
			break;
1172
		}
1173
		/* Save the delegation */
1174
		nfs4_stateid_copy(&stateid, &delegation->stateid);
1175
		rcu_read_unlock();
1176
		nfs_release_seqid(opendata->o_arg.seqid);
1177 1178 1179 1180 1181
		if (!opendata->is_recover) {
			ret = nfs_may_open(state->inode, state->owner->so_cred, open_mode);
			if (ret != 0)
				goto out;
		}
1182
		ret = -EAGAIN;
1183 1184

		/* Try to update the stateid using the delegation */
1185
		if (update_open_stateid(state, NULL, &stateid, fmode))
1186
			goto out_return_state;
1187 1188 1189 1190 1191 1192 1193 1194
	}
out:
	return ERR_PTR(ret);
out_return_state:
	atomic_inc(&state->count);
	return state;
}

1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230 1231 1232 1233 1234 1235 1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252
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;

1253 1254
	nfs_setsecurity(inode, &data->f_attr, data->f_label);

1255 1256 1257 1258 1259 1260 1261 1262 1263 1264 1265 1266 1267
	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)
1268 1269 1270
{
	struct inode *inode;
	struct nfs4_state *state = NULL;
1271
	int ret;
1272

1273
	if (!data->rpc_done) {
1274
		state = nfs4_try_open_cached(data);
1275 1276 1277
		goto out;
	}

1278
	ret = -EAGAIN;
1279
	if (!(data->f_attr.valid & NFS_ATTR_FATTR))
1280
		goto err;
1281
	inode = nfs_fhget(data->dir->d_sb, &data->o_res.fh, &data->f_attr, data->f_label);
1282
	ret = PTR_ERR(inode);
1283
	if (IS_ERR(inode))
1284 1285
		goto err;
	ret = -ENOMEM;
1286 1287
	state = nfs4_get_open_state(inode, data->owner);
	if (state == NULL)
1288
		goto err_put_inode;
1289 1290
	if (data->o_res.delegation_type != 0)
		nfs4_opendata_check_deleg(data, state);
1291
	update_open_stateid(state, &data->o_res.stateid, NULL,
1292
			data->o_arg.fmode);
1293
	iput(inode);
1294
out:
1295
	nfs_release_seqid(data->o_arg.seqid);
1296
	return state;
1297 1298 1299 1300
err_put_inode:
	iput(inode);
err:
	return ERR_PTR(ret);
1301 1302
}

1303 1304 1305 1306 1307 1308 1309 1310
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);
}

1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326 1327
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);
}

1328 1329
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
1330 1331 1332
{
	struct nfs4_opendata *opendata;

1333
	opendata = nfs4_opendata_alloc(ctx->dentry, state->owner, 0, 0,
1334
			NULL, NULL, claim, GFP_NOFS);
T
Trond Myklebust 已提交
1335 1336 1337 1338 1339 1340 1341
	if (opendata == NULL)
		return ERR_PTR(-ENOMEM);
	opendata->state = state;
	atomic_inc(&state->count);
	return opendata;
}

1342
static int nfs4_open_recover_helper(struct nfs4_opendata *opendata, fmode_t fmode, struct nfs4_state **res)
1343
{
1344
	struct nfs4_state *newstate;
1345 1346
	int ret;

1347 1348
	opendata->o_arg.open_flags = 0;
	opendata->o_arg.fmode = fmode;
1349 1350 1351
	memset(&opendata->o_res, 0, sizeof(opendata->o_res));
	memset(&opendata->c_res, 0, sizeof(opendata->c_res));
	nfs4_init_opendata_res(opendata);
1352
	ret = _nfs4_recover_proc_open(opendata);
1353 1354
	if (ret != 0)
		return ret; 
1355
	newstate = nfs4_opendata_to_nfs4_state(opendata);
1356 1357
	if (IS_ERR(newstate))
		return PTR_ERR(newstate);
1358
	nfs4_close_state(newstate, fmode);
1359
	*res = newstate;
1360 1361 1362 1363 1364 1365 1366 1367 1368
	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_* */
1369
	clear_bit(NFS_DELEGATED_STATE, &state->flags);
1370
	clear_bit(NFS_OPEN_STATE, &state->flags);
1371 1372
	smp_rmb();
	if (state->n_rdwr != 0) {
T
Trond Myklebust 已提交
1373
		clear_bit(NFS_O_RDWR_STATE, &state->flags);
1374
		ret = nfs4_open_recover_helper(opendata, FMODE_READ|FMODE_WRITE, &newstate);
1375 1376
		if (ret != 0)
			return ret;
1377 1378
		if (newstate != state)
			return -ESTALE;
1379 1380
	}
	if (state->n_wronly != 0) {
T
Trond Myklebust 已提交
1381
		clear_bit(NFS_O_WRONLY_STATE, &state->flags);
1382
		ret = nfs4_open_recover_helper(opendata, FMODE_WRITE, &newstate);
1383 1384
		if (ret != 0)
			return ret;
1385 1386
		if (newstate != state)
			return -ESTALE;
1387 1388
	}
	if (state->n_rdonly != 0) {
T
Trond Myklebust 已提交
1389
		clear_bit(NFS_O_RDONLY_STATE, &state->flags);
1390
		ret = nfs4_open_recover_helper(opendata, FMODE_READ, &newstate);
1391 1392
		if (ret != 0)
			return ret;
1393 1394
		if (newstate != state)
			return -ESTALE;
1395
	}
1396 1397 1398 1399 1400
	/*
	 * 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 &&
1401
	    !nfs4_stateid_match(&state->stateid, &state->open_stateid)) {
1402
		write_seqlock(&state->seqlock);
1403
		if (test_bit(NFS_DELEGATED_STATE, &state->flags) == 0)
1404
			nfs4_stateid_copy(&state->stateid, &state->open_stateid);
1405
		write_sequnlock(&state->seqlock);
1406
	}
1407 1408 1409
	return 0;
}

L
Linus Torvalds 已提交
1410 1411 1412 1413
/*
 * OPEN_RECLAIM:
 * 	reclaim state on the server after a reboot.
 */
1414
static int _nfs4_do_open_reclaim(struct nfs_open_context *ctx, struct nfs4_state *state)
L
Linus Torvalds 已提交
1415
{
1416
	struct nfs_delegation *delegation;
1417
	struct nfs4_opendata *opendata;
1418
	fmode_t delegation_type = 0;
L
Linus Torvalds 已提交
1419 1420
	int status;

1421 1422
	opendata = nfs4_open_recoverdata_alloc(ctx, state,
			NFS4_OPEN_CLAIM_PREVIOUS);
T
Trond Myklebust 已提交
1423 1424
	if (IS_ERR(opendata))
		return PTR_ERR(opendata);
1425 1426
	rcu_read_lock();
	delegation = rcu_dereference(NFS_I(state->inode)->delegation);
1427
	if (delegation != NULL && test_bit(NFS_DELEGATION_NEED_RECLAIM, &delegation->flags) != 0)
1428
		delegation_type = delegation->type;
1429
	rcu_read_unlock();
1430 1431
	opendata->o_arg.u.delegation_type = delegation_type;
	status = nfs4_open_recover(opendata, state);
1432
	nfs4_opendata_put(opendata);
L
Linus Torvalds 已提交
1433 1434 1435
	return status;
}

1436
static int nfs4_do_open_reclaim(struct nfs_open_context *ctx, struct nfs4_state *state)
L
Linus Torvalds 已提交
1437 1438 1439 1440 1441
{
	struct nfs_server *server = NFS_SERVER(state->inode);
	struct nfs4_exception exception = { };
	int err;
	do {
1442
		err = _nfs4_do_open_reclaim(ctx, state);
1443 1444
		if (nfs4_clear_cap_atomic_open_v1(server, err, &exception))
			continue;
1445
		if (err != -NFS4ERR_DELAY)
1446 1447
			break;
		nfs4_handle_exception(server, err, &exception);
L
Linus Torvalds 已提交
1448 1449 1450 1451
	} while (exception.retry);
	return err;
}

1452 1453 1454 1455 1456 1457 1458
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))
1459
		return -EAGAIN;
1460
	ret = nfs4_do_open_reclaim(ctx, state);
1461 1462 1463 1464
	put_nfs_open_context(ctx);
	return ret;
}

1465
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
1466
{
1467 1468 1469 1470 1471 1472 1473 1474 1475 1476 1477 1478 1479 1480 1481 1482 1483 1484 1485 1486 1487 1488 1489 1490 1491 1492
	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:
1493
		case -NFS4ERR_OPENMODE:
1494 1495 1496 1497 1498 1499 1500 1501 1502
			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;
1503 1504 1505 1506
		case -ENOMEM:
		case -NFS4ERR_DENIED:
			/* kill_proc(fl->fl_pid, SIGLOST, 1); */
			return 0;
1507
	}
L
Linus Torvalds 已提交
1508 1509 1510
	return err;
}

1511 1512 1513 1514 1515 1516 1517 1518 1519 1520 1521 1522 1523 1524 1525 1526
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);
}

1527 1528 1529 1530 1531
static void nfs4_open_confirm_done(struct rpc_task *task, void *calldata)
{
	struct nfs4_opendata *data = calldata;

	data->rpc_status = task->tk_status;
1532
	if (data->rpc_status == 0) {
1533
		nfs4_stateid_copy(&data->o_res.stateid, &data->c_res.stateid);
1534
		nfs_confirm_seqid(&data->owner->so_seqid, 0);
1535
		renew_lease(data->o_res.server, data->timestamp);
1536
		data->rpc_done = 1;
1537
	}
1538 1539 1540 1541 1542 1543 1544 1545 1546 1547 1548
}

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! */
1549
	if (!data->rpc_done)
1550 1551
		goto out_free;
	state = nfs4_opendata_to_nfs4_state(data);
1552
	if (!IS_ERR(state))
1553
		nfs4_close_state(state, data->o_arg.fmode);
1554
out_free:
1555
	nfs4_opendata_put(data);
1556 1557 1558 1559 1560 1561 1562 1563 1564 1565 1566 1567 1568 1569
}

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;
1570 1571 1572 1573 1574 1575
	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
1576 1577
	struct rpc_task_setup task_setup_data = {
		.rpc_client = server->client,
1578
		.rpc_message = &msg,
T
Trond Myklebust 已提交
1579 1580
		.callback_ops = &nfs4_open_confirm_ops,
		.callback_data = data,
1581
		.workqueue = nfsiod_workqueue,
T
Trond Myklebust 已提交
1582 1583
		.flags = RPC_TASK_ASYNC,
	};
L
Linus Torvalds 已提交
1584 1585
	int status;

1586
	kref_get(&data->kref);
1587 1588
	data->rpc_done = 0;
	data->rpc_status = 0;
1589
	data->timestamp = jiffies;
T
Trond Myklebust 已提交
1590
	task = rpc_run_task(&task_setup_data);
1591
	if (IS_ERR(task))
1592 1593 1594 1595 1596 1597 1598
		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;
1599
	rpc_put_task(task);
L
Linus Torvalds 已提交
1600 1601 1602
	return status;
}

1603
static void nfs4_open_prepare(struct rpc_task *task, void *calldata)
L
Linus Torvalds 已提交
1604
{
1605 1606
	struct nfs4_opendata *data = calldata;
	struct nfs4_state_owner *sp = data->owner;
1607
	struct nfs_client *clp = sp->so_server->nfs_client;
1608

1609
	if (nfs_wait_on_sequence(data->o_arg.seqid, task) != 0)
1610
		goto out_wait;
1611 1612 1613 1614 1615 1616 1617
	/*
	 * 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;

1618
		if (can_open_cached(data->state, data->o_arg.fmode, data->o_arg.open_flags))
1619
			goto out_no_action;
1620 1621
		rcu_read_lock();
		delegation = rcu_dereference(NFS_I(data->state->inode)->delegation);
1622
		if (data->o_arg.claim != NFS4_OPEN_CLAIM_DELEGATE_CUR &&
1623
		    data->o_arg.claim != NFS4_OPEN_CLAIM_DELEG_CUR_FH &&
1624 1625
		    can_open_delegated(delegation, data->o_arg.fmode))
			goto unlock_no_action;
1626 1627
		rcu_read_unlock();
	}
1628
	/* Update client id. */
1629
	data->o_arg.clientid = clp->cl_clientid;
1630 1631 1632 1633
	switch (data->o_arg.claim) {
	case NFS4_OPEN_CLAIM_PREVIOUS:
	case NFS4_OPEN_CLAIM_DELEG_CUR_FH:
	case NFS4_OPEN_CLAIM_DELEG_PREV_FH:
1634
		data->o_arg.open_bitmap = &nfs4_open_noattr_bitmap[0];
1635 1636
	case NFS4_OPEN_CLAIM_FH:
		task->tk_msg.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_OPEN_NOATTR];
T
Trond Myklebust 已提交
1637 1638
		nfs_copy_fh(&data->o_res.fh, data->o_arg.fh);
	}
1639
	data->timestamp = jiffies;
1640
	if (nfs4_setup_sequence(data->o_arg.server,
1641
				&data->o_arg.seq_args,
1642 1643 1644
				&data->o_res.seq_res,
				task) != 0)
		nfs_release_seqid(data->o_arg.seqid);
1645 1646 1647 1648 1649 1650 1651 1652 1653 1654

	/* 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;
	}
1655
	return;
1656 1657
unlock_no_action:
	rcu_read_unlock();
1658 1659
out_no_action:
	task->tk_action = NULL;
1660
out_wait:
1661
	nfs4_sequence_done(task, &data->o_res.seq_res);
1662 1663
}

1664 1665 1666
static void nfs4_open_done(struct rpc_task *task, void *calldata)
{
	struct nfs4_opendata *data = calldata;
L
Linus Torvalds 已提交
1667

1668
	data->rpc_status = task->tk_status;
1669

1670 1671
	if (!nfs4_sequence_done(task, &data->o_res.seq_res))
		return;
1672

1673
	if (task->tk_status == 0) {
1674 1675
		if (data->o_res.f_attr->valid & NFS_ATTR_FATTR_TYPE) {
			switch (data->o_res.f_attr->mode & S_IFMT) {
1676 1677 1678
			case S_IFREG:
				break;
			case S_IFLNK:
1679
				data->rpc_status = -ELOOP;
1680 1681
				break;
			case S_IFDIR:
1682
				data->rpc_status = -EISDIR;
1683 1684
				break;
			default:
1685
				data->rpc_status = -ENOTDIR;
1686
			}
1687
		}
1688
		renew_lease(data->o_res.server, data->timestamp);
1689 1690
		if (!(data->o_res.rflags & NFS4_OPEN_RESULT_CONFIRM))
			nfs_confirm_seqid(&data->owner->so_seqid, 0);
1691
	}
1692
	data->rpc_done = 1;
1693
}
1694

1695 1696 1697 1698 1699 1700 1701 1702 1703
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! */
1704
	if (data->rpc_status != 0 || !data->rpc_done)
1705 1706 1707 1708 1709
		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);
1710
	if (!IS_ERR(state))
1711
		nfs4_close_state(state, data->o_arg.fmode);
1712
out_free:
1713
	nfs4_opendata_put(data);
1714 1715 1716 1717 1718 1719 1720 1721
}

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,
};

1722
static int nfs4_run_open_task(struct nfs4_opendata *data, int isrecover)
1723 1724 1725 1726 1727 1728
{
	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;
1729 1730 1731 1732 1733 1734
	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
1735 1736
	struct rpc_task_setup task_setup_data = {
		.rpc_client = server->client,
1737
		.rpc_message = &msg,
T
Trond Myklebust 已提交
1738 1739
		.callback_ops = &nfs4_open_ops,
		.callback_data = data,
1740
		.workqueue = nfsiod_workqueue,
T
Trond Myklebust 已提交
1741 1742
		.flags = RPC_TASK_ASYNC,
	};
1743 1744
	int status;

1745
	nfs41_init_sequence(&o_arg->seq_args, &o_res->seq_res, 1);
1746
	kref_get(&data->kref);
1747 1748
	data->rpc_done = 0;
	data->rpc_status = 0;
1749
	data->cancelled = 0;
1750 1751
	data->is_recover = 0;
	if (isrecover) {
1752
		nfs4_set_sequence_privileged(&o_arg->seq_args);
1753 1754
		data->is_recover = 1;
	}
T
Trond Myklebust 已提交
1755
	task = rpc_run_task(&task_setup_data);
1756 1757 1758 1759 1760 1761 1762 1763 1764 1765 1766 1767 1768 1769 1770 1771 1772 1773 1774 1775 1776 1777 1778
        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;

1779 1780
	nfs_fattr_map_and_free_names(NFS_SERVER(dir), &data->f_attr);

1781 1782 1783 1784 1785 1786 1787 1788 1789
	if (o_res->rflags & NFS4_OPEN_RESULT_CONFIRM) {
		status = _nfs4_proc_open_confirm(data);
		if (status != 0)
			return status;
	}

	return status;
}

1790 1791
static int nfs4_opendata_access(struct rpc_cred *cred,
				struct nfs4_opendata *opendata,
1792 1793
				struct nfs4_state *state, fmode_t fmode,
				int openflags)
1794 1795 1796 1797 1798 1799 1800 1801 1802 1803
{
	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;
1804
	/* don't check MAY_WRITE - a newly created file may not have
1805 1806 1807 1808 1809 1810 1811 1812
	 * 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;
1813 1814 1815 1816 1817 1818

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

1819
	if ((mask & ~cache.mask & (MAY_READ | MAY_EXEC)) == 0)
1820 1821 1822 1823
		return 0;

	/* even though OPEN succeeded, access is denied. Close the file */
	nfs4_close_state(state, fmode);
1824
	return -EACCES;
1825 1826
}

1827 1828 1829 1830 1831 1832 1833 1834 1835 1836 1837 1838
/*
 * 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);
1839 1840 1841 1842 1843 1844
	if (!data->rpc_done)
		return status;
	if (status != 0) {
		if (status == -NFS4ERR_BADNAME &&
				!(o_arg->open_flags & O_CREAT))
			return -ENOENT;
1845
		return status;
1846
	}
1847

1848 1849
	nfs_fattr_map_and_free_names(server, &data->f_attr);

1850
	if (o_arg->open_flags & O_CREAT)
1851
		update_changeattr(dir, &o_res->cinfo);
T
Trond Myklebust 已提交
1852 1853
	if ((o_res->rflags & NFS4_OPEN_RESULT_LOCKTYPE_POSIX) == 0)
		server->caps &= ~NFS_CAP_POSIX_LOCK;
L
Linus Torvalds 已提交
1854
	if(o_res->rflags & NFS4_OPEN_RESULT_CONFIRM) {
1855
		status = _nfs4_proc_open_confirm(data);
L
Linus Torvalds 已提交
1856
		if (status != 0)
1857
			return status;
L
Linus Torvalds 已提交
1858 1859
	}
	if (!(o_res->f_attr->valid & NFS_ATTR_FATTR))
1860
		_nfs4_proc_getattr(server, &o_res->fh, o_res->f_attr, o_res->f_label);
1861
	return 0;
L
Linus Torvalds 已提交
1862 1863
}

A
Andy Adamson 已提交
1864 1865 1866 1867 1868
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
1869 1870 1871 1872 1873
/*
 * OPEN_EXPIRED:
 * 	reclaim state on the server after a network partition.
 * 	Assumes caller holds the appropriate lock
 */
1874
static int _nfs4_open_expired(struct nfs_open_context *ctx, struct nfs4_state *state)
L
Linus Torvalds 已提交
1875
{
1876
	struct nfs4_opendata *opendata;
1877
	int ret;
L
Linus Torvalds 已提交
1878

1879
	opendata = nfs4_open_recoverdata_alloc(ctx, state,
1880
			NFS4_OPEN_CLAIM_FH);
T
Trond Myklebust 已提交
1881 1882
	if (IS_ERR(opendata))
		return PTR_ERR(opendata);
1883
	ret = nfs4_open_recover(opendata, state);
1884
	if (ret == -ESTALE)
1885
		d_drop(ctx->dentry);
1886
	nfs4_opendata_put(opendata);
1887
	return ret;
L
Linus Torvalds 已提交
1888 1889
}

1890
static int nfs4_do_open_expired(struct nfs_open_context *ctx, struct nfs4_state *state)
1891
{
1892
	struct nfs_server *server = NFS_SERVER(state->inode);
1893 1894 1895 1896
	struct nfs4_exception exception = { };
	int err;

	do {
1897
		err = _nfs4_open_expired(ctx, state);
1898 1899
		if (nfs4_clear_cap_atomic_open_v1(server, err, &exception))
			continue;
1900 1901 1902 1903 1904 1905 1906 1907
		switch (err) {
		default:
			goto out;
		case -NFS4ERR_GRACE:
		case -NFS4ERR_DELAY:
			nfs4_handle_exception(server, err, &exception);
			err = 0;
		}
1908
	} while (exception.retry);
1909
out:
1910 1911 1912
	return err;
}

L
Linus Torvalds 已提交
1913 1914 1915
static int nfs4_open_expired(struct nfs4_state_owner *sp, struct nfs4_state *state)
{
	struct nfs_open_context *ctx;
1916
	int ret;
L
Linus Torvalds 已提交
1917

1918 1919
	ctx = nfs4_state_find_open_context(state);
	if (IS_ERR(ctx))
1920
		return -EAGAIN;
1921
	ret = nfs4_do_open_expired(ctx, state);
1922 1923
	put_nfs_open_context(ctx);
	return ret;
L
Linus Torvalds 已提交
1924 1925
}

1926
#if defined(CONFIG_NFS_V4_1)
1927
static void nfs41_clear_delegation_stateid(struct nfs4_state *state)
1928 1929
{
	struct nfs_server *server = NFS_SERVER(state->inode);
1930
	nfs4_stateid *stateid = &state->stateid;
1931 1932 1933
	struct nfs_delegation *delegation;
	struct rpc_cred *cred = NULL;
	int status = -NFS4ERR_BAD_STATEID;
1934 1935 1936 1937 1938

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

1939 1940 1941 1942 1943 1944 1945 1946 1947 1948 1949
	/* 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();

1950 1951 1952 1953
	if (status != NFS_OK) {
		/* Free the stateid unless the server explicitly
		 * informs us the stateid is unrecognized. */
		if (status != -NFS4ERR_BAD_STATEID)
1954
			nfs41_free_stateid(server, stateid, cred);
1955
		nfs_remove_bad_delegation(state->inode);
1956

1957 1958 1959
		write_seqlock(&state->seqlock);
		nfs4_stateid_copy(&state->stateid, &state->open_stateid);
		write_sequnlock(&state->seqlock);
1960 1961
		clear_bit(NFS_DELEGATED_STATE, &state->flags);
	}
1962 1963 1964

	if (cred != NULL)
		put_rpccred(cred);
1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977
}

/**
 * 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);
1978
	nfs4_stateid *stateid = &state->open_stateid;
1979
	struct rpc_cred *cred = state->owner->so_cred;
1980 1981 1982 1983 1984 1985 1986 1987
	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;

1988
	status = nfs41_test_stateid(server, stateid, cred);
1989 1990 1991 1992
	if (status != NFS_OK) {
		/* Free the stateid unless the server explicitly
		 * informs us the stateid is unrecognized. */
		if (status != -NFS4ERR_BAD_STATEID)
1993
			nfs41_free_stateid(server, stateid, cred);
1994 1995 1996 1997

		clear_bit(NFS_O_RDONLY_STATE, &state->flags);
		clear_bit(NFS_O_WRONLY_STATE, &state->flags);
		clear_bit(NFS_O_RDWR_STATE, &state->flags);
1998
		clear_bit(NFS_OPEN_STATE, &state->flags);
1999 2000 2001 2002 2003 2004
	}
	return status;
}

static int nfs41_open_expired(struct nfs4_state_owner *sp, struct nfs4_state *state)
{
2005
	int status;
2006

2007 2008
	nfs41_clear_delegation_stateid(state);
	status = nfs41_check_open_stateid(state);
2009 2010 2011
	if (status != NFS_OK)
		status = nfs4_open_expired(sp, state);
	return status;
2012 2013 2014
}
#endif

2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 2030
/*
 * 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;
}

2031 2032 2033
static int _nfs4_open_and_get_state(struct nfs4_opendata *opendata,
		fmode_t fmode,
		int flags,
2034
		struct nfs_open_context *ctx)
2035 2036 2037
{
	struct nfs4_state_owner *sp = opendata->owner;
	struct nfs_server *server = sp->so_server;
2038
	struct dentry *dentry;
2039 2040 2041 2042 2043 2044 2045 2046 2047 2048 2049 2050 2051 2052 2053 2054 2055
	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);

2056 2057 2058 2059 2060 2061 2062 2063 2064 2065 2066 2067 2068 2069 2070
	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));
	}

2071 2072 2073 2074
	ret = nfs4_opendata_access(sp->so_cred, opendata, state, fmode, flags);
	if (ret != 0)
		goto out;

2075
	ctx->state = state;
2076 2077 2078 2079 2080
	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);
	}
2081 2082 2083 2084
out:
	return ret;
}

L
Linus Torvalds 已提交
2085
/*
2086
 * Returns a referenced nfs4_state
L
Linus Torvalds 已提交
2087
 */
2088
static int _nfs4_do_open(struct inode *dir,
2089
			struct nfs_open_context *ctx,
2090 2091
			int flags,
			struct iattr *sattr,
2092
			struct nfs4_label *label)
L
Linus Torvalds 已提交
2093 2094 2095 2096
{
	struct nfs4_state_owner  *sp;
	struct nfs4_state     *state = NULL;
	struct nfs_server       *server = NFS_SERVER(dir);
2097
	struct nfs4_opendata *opendata;
2098 2099 2100 2101
	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);
2102
	enum open_claim_type4 claim = NFS4_OPEN_CLAIM_NULL;
2103
	struct nfs4_label *olabel = NULL;
2104
	int status;
L
Linus Torvalds 已提交
2105 2106 2107

	/* Protect against reboot recovery conflicts */
	status = -ENOMEM;
2108 2109
	sp = nfs4_get_state_owner(server, cred, GFP_KERNEL);
	if (sp == NULL) {
L
Linus Torvalds 已提交
2110 2111 2112
		dprintk("nfs4_do_open: nfs4_get_state_owner failed!\n");
		goto out_err;
	}
2113 2114
	status = nfs4_recover_expired_lease(server);
	if (status != 0)
2115
		goto err_put_state_owner;
2116 2117
	if (dentry->d_inode != NULL)
		nfs4_return_incompatible_delegation(dentry->d_inode, fmode);
2118
	status = -ENOMEM;
2119 2120
	if (dentry->d_inode)
		claim = NFS4_OPEN_CLAIM_FH;
2121
	opendata = nfs4_opendata_alloc(dentry, sp, fmode, flags, sattr,
2122
			label, claim, GFP_KERNEL);
2123
	if (opendata == NULL)
T
Trond Myklebust 已提交
2124
		goto err_put_state_owner;
L
Linus Torvalds 已提交
2125

2126 2127 2128 2129 2130 2131 2132 2133
	if (label) {
		olabel = nfs4_label_alloc(server, GFP_KERNEL);
		if (IS_ERR(olabel)) {
			status = PTR_ERR(olabel);
			goto err_opendata_put;
		}
	}

2134 2135 2136
	if (ctx_th && server->attr_bitmask[2] & FATTR4_WORD2_MDSTHRESHOLD) {
		opendata->f_attr.mdsthreshold = pnfs_mdsthreshold_alloc();
		if (!opendata->f_attr.mdsthreshold)
2137
			goto err_free_label;
2138
		opendata->o_arg.open_bitmap = &nfs4_pnfs_open_bitmap[0];
2139
	}
2140 2141
	if (dentry->d_inode != NULL)
		opendata->state = nfs4_get_open_state(dentry->d_inode, sp);
2142

2143
	status = _nfs4_open_and_get_state(opendata, fmode, flags, ctx);
2144
	if (status != 0)
2145
		goto err_free_label;
2146
	state = ctx->state;
2147

2148 2149
	if ((opendata->o_arg.open_flags & O_EXCL) &&
	    (opendata->o_arg.createmode != NFS4_CREATE_GUARDED)) {
2150 2151 2152 2153 2154
		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,
2155 2156
				state, label, olabel);
		if (status == 0) {
2157
			nfs_setattr_update_inode(state->inode, sattr);
2158
			nfs_post_op_update_inode(state->inode, opendata->o_res.f_attr);
2159
			nfs_setsecurity(state->inode, opendata->o_res.f_attr, olabel);
2160
		}
2161
	}
2162 2163 2164 2165 2166 2167 2168

	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;

2169 2170
	nfs4_label_free(olabel);

2171
	nfs4_opendata_put(opendata);
L
Linus Torvalds 已提交
2172 2173
	nfs4_put_state_owner(sp);
	return 0;
2174 2175
err_free_label:
	nfs4_label_free(olabel);
2176
err_opendata_put:
2177
	kfree(opendata->f_attr.mdsthreshold);
2178
	nfs4_opendata_put(opendata);
2179 2180
err_put_state_owner:
	nfs4_put_state_owner(sp);
L
Linus Torvalds 已提交
2181 2182 2183 2184 2185
out_err:
	return status;
}


2186
static struct nfs4_state *nfs4_do_open(struct inode *dir,
2187
					struct nfs_open_context *ctx,
2188 2189
					int flags,
					struct iattr *sattr,
2190
					struct nfs4_label *label)
L
Linus Torvalds 已提交
2191
{
2192
	struct nfs_server *server = NFS_SERVER(dir);
L
Linus Torvalds 已提交
2193 2194 2195 2196 2197
	struct nfs4_exception exception = { };
	struct nfs4_state *res;
	int status;

	do {
2198
		status = _nfs4_do_open(dir, ctx, flags, sattr, label);
2199
		res = ctx->state;
L
Linus Torvalds 已提交
2200 2201 2202 2203 2204 2205 2206 2207
		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
2208
		 * doing an OPEN, we assume that nfs_increment_open_seqid() will
L
Linus Torvalds 已提交
2209 2210 2211 2212 2213
		 * 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) {
2214
			pr_warn_ratelimited("NFS: v4 server %s "
2215 2216
					" returned a bad sequence-id error!\n",
					NFS_SERVER(dir)->nfs_client->cl_hostname);
L
Linus Torvalds 已提交
2217 2218 2219
			exception.retry = 1;
			continue;
		}
2220 2221 2222 2223 2224 2225 2226 2227 2228 2229
		/*
		 * 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;
		}
2230 2231 2232 2233 2234
		if (status == -EAGAIN) {
			/* We must have found a delegation */
			exception.retry = 1;
			continue;
		}
2235 2236 2237
		if (nfs4_clear_cap_atomic_open_v1(server, status, &exception))
			continue;
		res = ERR_PTR(nfs4_handle_exception(server,
L
Linus Torvalds 已提交
2238 2239 2240 2241 2242
					status, &exception));
	} while (exception.retry);
	return res;
}

2243 2244
static int _nfs4_do_setattr(struct inode *inode, struct rpc_cred *cred,
			    struct nfs_fattr *fattr, struct iattr *sattr,
2245 2246
			    struct nfs4_state *state, struct nfs4_label *ilabel,
			    struct nfs4_label *olabel)
L
Linus Torvalds 已提交
2247
{
2248
	struct nfs_server *server = NFS_SERVER(inode);
L
Linus Torvalds 已提交
2249
        struct nfs_setattrargs  arg = {
2250
                .fh             = NFS_FH(inode),
L
Linus Torvalds 已提交
2251 2252 2253
                .iap            = sattr,
		.server		= server,
		.bitmask = server->attr_bitmask,
2254
		.label		= ilabel,
L
Linus Torvalds 已提交
2255 2256 2257
        };
        struct nfs_setattrres  res = {
		.fattr		= fattr,
2258
		.label		= olabel,
L
Linus Torvalds 已提交
2259 2260 2261
		.server		= server,
        };
        struct rpc_message msg = {
2262 2263 2264 2265
		.rpc_proc	= &nfs4_procedures[NFSPROC4_CLNT_SETATTR],
		.rpc_argp	= &arg,
		.rpc_resp	= &res,
		.rpc_cred	= cred,
L
Linus Torvalds 已提交
2266
        };
2267
	unsigned long timestamp = jiffies;
2268 2269
	fmode_t fmode;
	bool truncate;
2270
	int status;
L
Linus Torvalds 已提交
2271

2272 2273 2274 2275
	arg.bitmask = nfs4_bitmask(server, ilabel);
	if (ilabel)
		arg.bitmask = nfs4_bitmask(server, olabel);

2276
	nfs_fattr_init(fattr);
L
Linus Torvalds 已提交
2277

2278 2279 2280 2281 2282 2283 2284
	/* 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)) {
2285 2286 2287 2288
		struct nfs_lockowner lockowner = {
			.l_owner = current->files,
			.l_pid = current->tgid,
		};
2289
		nfs4_select_rw_stateid(&arg.stateid, state, FMODE_WRITE,
2290
				&lockowner);
2291
	} else
2292
		nfs4_stateid_copy(&arg.stateid, &zero_stateid);
L
Linus Torvalds 已提交
2293

2294
	status = nfs4_call_sync(server->client, server, &msg, &arg.seq_args, &res.seq_res, 1);
2295 2296
	if (status == 0 && state != NULL)
		renew_lease(server, timestamp);
2297
	return status;
L
Linus Torvalds 已提交
2298 2299
}

2300 2301
static int nfs4_do_setattr(struct inode *inode, struct rpc_cred *cred,
			   struct nfs_fattr *fattr, struct iattr *sattr,
2302 2303
			   struct nfs4_state *state, struct nfs4_label *ilabel,
			   struct nfs4_label *olabel)
L
Linus Torvalds 已提交
2304
{
2305
	struct nfs_server *server = NFS_SERVER(inode);
2306 2307
	struct nfs4_exception exception = {
		.state = state,
2308
		.inode = inode,
2309
	};
L
Linus Torvalds 已提交
2310 2311
	int err;
	do {
2312
		err = _nfs4_do_setattr(inode, cred, fattr, sattr, state, ilabel, olabel);
2313 2314
		switch (err) {
		case -NFS4ERR_OPENMODE:
2315 2316 2317 2318 2319 2320 2321
			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);
			}
2322 2323 2324 2325 2326 2327 2328 2329
			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
2330
	} while (exception.retry);
2331
out:
L
Linus Torvalds 已提交
2332 2333 2334 2335 2336 2337 2338 2339
	return err;
}

struct nfs4_closedata {
	struct inode *inode;
	struct nfs4_state *state;
	struct nfs_closeargs arg;
	struct nfs_closeres res;
2340
	struct nfs_fattr fattr;
2341
	unsigned long timestamp;
F
Fred Isaman 已提交
2342 2343
	bool roc;
	u32 roc_barrier;
L
Linus Torvalds 已提交
2344 2345
};

2346
static void nfs4_free_closedata(void *data)
2347
{
2348 2349
	struct nfs4_closedata *calldata = data;
	struct nfs4_state_owner *sp = calldata->state->owner;
2350
	struct super_block *sb = calldata->state->inode->i_sb;
2351

F
Fred Isaman 已提交
2352 2353
	if (calldata->roc)
		pnfs_roc_release(calldata->state->inode);
2354 2355 2356
	nfs4_put_open_state(calldata->state);
	nfs_free_seqid(calldata->arg.seqid);
	nfs4_put_state_owner(sp);
2357
	nfs_sb_deactive(sb);
2358 2359 2360
	kfree(calldata);
}

2361 2362 2363 2364
static void nfs4_close_clear_stateid_flags(struct nfs4_state *state,
		fmode_t fmode)
{
	spin_lock(&state->owner->so_lock);
2365 2366 2367
	clear_bit(NFS_O_RDWR_STATE, &state->flags);
	switch (fmode & (FMODE_READ|FMODE_WRITE)) {
	case FMODE_WRITE:
2368
		clear_bit(NFS_O_RDONLY_STATE, &state->flags);
2369 2370
		break;
	case FMODE_READ:
2371
		clear_bit(NFS_O_WRONLY_STATE, &state->flags);
2372 2373 2374 2375 2376 2377
		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);
	}
2378 2379 2380
	spin_unlock(&state->owner->so_lock);
}

2381
static void nfs4_close_done(struct rpc_task *task, void *data)
L
Linus Torvalds 已提交
2382
{
2383
	struct nfs4_closedata *calldata = data;
L
Linus Torvalds 已提交
2384 2385 2386
	struct nfs4_state *state = calldata->state;
	struct nfs_server *server = NFS_SERVER(calldata->inode);

2387
	dprintk("%s: begin!\n", __func__);
2388 2389
	if (!nfs4_sequence_done(task, &calldata->res.seq_res))
		return;
L
Linus Torvalds 已提交
2390 2391 2392 2393 2394
        /* 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
2395 2396 2397
			if (calldata->roc)
				pnfs_roc_set_barrier(state->inode,
						     calldata->roc_barrier);
2398
			nfs_set_open_stateid(state, &calldata->res.stateid, 0);
2399
			renew_lease(server, calldata->timestamp);
2400 2401
			nfs4_close_clear_stateid_flags(state,
					calldata->arg.fmode);
L
Linus Torvalds 已提交
2402 2403
			break;
		case -NFS4ERR_STALE_STATEID:
2404 2405
		case -NFS4ERR_OLD_STATEID:
		case -NFS4ERR_BAD_STATEID:
L
Linus Torvalds 已提交
2406
		case -NFS4ERR_EXPIRED:
2407
			if (calldata->arg.fmode == 0)
2408
				break;
L
Linus Torvalds 已提交
2409
		default:
2410 2411
			if (nfs4_async_handle_error(task, server, state) == -EAGAIN)
				rpc_restart_call_prepare(task);
L
Linus Torvalds 已提交
2412
	}
2413
	nfs_release_seqid(calldata->arg.seqid);
2414
	nfs_refresh_inode(calldata->inode, calldata->res.fattr);
2415
	dprintk("%s: done, ret = %d!\n", __func__, task->tk_status);
L
Linus Torvalds 已提交
2416 2417
}

T
Trond Myklebust 已提交
2418
static void nfs4_close_prepare(struct rpc_task *task, void *data)
L
Linus Torvalds 已提交
2419
{
T
Trond Myklebust 已提交
2420
	struct nfs4_closedata *calldata = data;
2421
	struct nfs4_state *state = calldata->state;
2422
	struct inode *inode = calldata->inode;
2423
	int call_close = 0;
2424

2425
	dprintk("%s: begin!\n", __func__);
2426
	if (nfs_wait_on_sequence(calldata->arg.seqid, task) != 0)
2427
		goto out_wait;
2428

2429 2430
	task->tk_msg.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_OPEN_DOWNGRADE];
	calldata->arg.fmode = FMODE_READ|FMODE_WRITE;
2431
	spin_lock(&state->owner->so_lock);
2432
	/* Calculate the change in open mode */
2433
	if (state->n_rdwr == 0) {
2434
		if (state->n_rdonly == 0) {
2435 2436 2437
			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;
2438 2439
		}
		if (state->n_wronly == 0) {
2440 2441 2442
			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;
2443
		}
2444
	}
2445 2446
	if (!nfs4_valid_open_stateid(state))
		call_close = 0;
2447
	spin_unlock(&state->owner->so_lock);
2448 2449

	if (!call_close) {
2450
		/* Note: exit _without_ calling nfs4_close_done */
2451
		goto out_no_action;
2452
	}
2453

F
Fred Isaman 已提交
2454
	if (calldata->arg.fmode == 0) {
2455
		task->tk_msg.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_CLOSE];
F
Fred Isaman 已提交
2456
		if (calldata->roc &&
2457 2458
		    pnfs_roc_drain(inode, &calldata->roc_barrier, task)) {
			nfs_release_seqid(calldata->arg.seqid);
2459
			goto out_wait;
2460
		    }
F
Fred Isaman 已提交
2461
	}
2462

2463
	nfs_fattr_init(calldata->res.fattr);
2464
	calldata->timestamp = jiffies;
2465
	if (nfs4_setup_sequence(NFS_SERVER(inode),
2466 2467
				&calldata->arg.seq_args,
				&calldata->res.seq_res,
2468 2469
				task) != 0)
		nfs_release_seqid(calldata->arg.seqid);
2470
	dprintk("%s: done!\n", __func__);
2471 2472 2473 2474 2475
	return;
out_no_action:
	task->tk_action = NULL;
out_wait:
	nfs4_sequence_done(task, &calldata->res.seq_res);
L
Linus Torvalds 已提交
2476 2477
}

2478
static const struct rpc_call_ops nfs4_close_ops = {
T
Trond Myklebust 已提交
2479
	.rpc_call_prepare = nfs4_close_prepare,
2480 2481 2482 2483
	.rpc_call_done = nfs4_close_done,
	.rpc_release = nfs4_free_closedata,
};

L
Linus Torvalds 已提交
2484 2485 2486 2487 2488 2489 2490 2491 2492 2493 2494
/* 
 * 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!
 */
2495
int nfs4_do_close(struct nfs4_state *state, gfp_t gfp_mask, int wait)
L
Linus Torvalds 已提交
2496
{
2497
	struct nfs_server *server = NFS_SERVER(state->inode);
L
Linus Torvalds 已提交
2498
	struct nfs4_closedata *calldata;
2499 2500
	struct nfs4_state_owner *sp = state->owner;
	struct rpc_task *task;
2501 2502 2503 2504
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_CLOSE],
		.rpc_cred = state->owner->so_cred,
	};
T
Trond Myklebust 已提交
2505 2506
	struct rpc_task_setup task_setup_data = {
		.rpc_client = server->client,
2507
		.rpc_message = &msg,
T
Trond Myklebust 已提交
2508
		.callback_ops = &nfs4_close_ops,
2509
		.workqueue = nfsiod_workqueue,
T
Trond Myklebust 已提交
2510 2511
		.flags = RPC_TASK_ASYNC,
	};
2512
	int status = -ENOMEM;
L
Linus Torvalds 已提交
2513

2514
	calldata = kzalloc(sizeof(*calldata), gfp_mask);
L
Linus Torvalds 已提交
2515
	if (calldata == NULL)
2516
		goto out;
2517
	nfs41_init_sequence(&calldata->arg.seq_args, &calldata->res.seq_res, 1);
2518
	calldata->inode = state->inode;
L
Linus Torvalds 已提交
2519
	calldata->state = state;
2520
	calldata->arg.fh = NFS_FH(state->inode);
2521
	calldata->arg.stateid = &state->open_stateid;
L
Linus Torvalds 已提交
2522
	/* Serialization for the sequence id */
2523
	calldata->arg.seqid = nfs_alloc_seqid(&state->owner->so_seqid, gfp_mask);
2524 2525
	if (calldata->arg.seqid == NULL)
		goto out_free_calldata;
2526
	calldata->arg.fmode = 0;
2527
	calldata->arg.bitmask = server->cache_consistency_bitmask;
2528
	calldata->res.fattr = &calldata->fattr;
2529
	calldata->res.seqid = calldata->arg.seqid;
2530
	calldata->res.server = server;
2531
	calldata->roc = pnfs_roc(state->inode);
2532
	nfs_sb_active(calldata->inode->i_sb);
2533

2534 2535
	msg.rpc_argp = &calldata->arg;
	msg.rpc_resp = &calldata->res;
T
Trond Myklebust 已提交
2536 2537
	task_setup_data.callback_data = calldata;
	task = rpc_run_task(&task_setup_data);
2538 2539
	if (IS_ERR(task))
		return PTR_ERR(task);
2540 2541 2542
	status = 0;
	if (wait)
		status = rpc_wait_for_completion_task(task);
2543
	rpc_put_task(task);
2544
	return status;
2545 2546 2547
out_free_calldata:
	kfree(calldata);
out:
2548 2549
	nfs4_put_open_state(state);
	nfs4_put_state_owner(sp);
2550
	return status;
L
Linus Torvalds 已提交
2551 2552
}

2553
static struct inode *
2554
nfs4_atomic_open(struct inode *dir, struct nfs_open_context *ctx, int open_flags, struct iattr *attr)
L
Linus Torvalds 已提交
2555 2556
{
	struct nfs4_state *state;
2557 2558 2559
	struct nfs4_label l = {0, 0, 0, NULL}, *label = NULL;

	label = nfs4_label_init_security(dir, ctx->dentry, attr, &l);
L
Linus Torvalds 已提交
2560

T
Trond Myklebust 已提交
2561
	/* Protect against concurrent sillydeletes */
2562
	state = nfs4_do_open(dir, ctx, open_flags, attr, label);
2563 2564 2565

	nfs4_label_release_security(label);

2566 2567
	if (IS_ERR(state))
		return ERR_CAST(state);
2568
	return state->inode;
L
Linus Torvalds 已提交
2569 2570
}

2571
static void nfs4_close_context(struct nfs_open_context *ctx, int is_sync)
T
Trond Myklebust 已提交
2572 2573 2574 2575
{
	if (ctx->state == NULL)
		return;
	if (is_sync)
2576
		nfs4_close_sync(ctx->state, ctx->mode);
T
Trond Myklebust 已提交
2577
	else
2578
		nfs4_close_state(ctx->state, ctx->mode);
T
Trond Myklebust 已提交
2579
}
L
Linus Torvalds 已提交
2580 2581 2582

static int _nfs4_server_capabilities(struct nfs_server *server, struct nfs_fh *fhandle)
{
B
Benny Halevy 已提交
2583 2584 2585
	struct nfs4_server_caps_arg args = {
		.fhandle = fhandle,
	};
L
Linus Torvalds 已提交
2586 2587 2588
	struct nfs4_server_caps_res res = {};
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_SERVER_CAPS],
B
Benny Halevy 已提交
2589
		.rpc_argp = &args,
L
Linus Torvalds 已提交
2590 2591 2592 2593
		.rpc_resp = &res,
	};
	int status;

2594
	status = nfs4_call_sync(server->client, server, &msg, &args.seq_args, &res.seq_res, 0);
L
Linus Torvalds 已提交
2595 2596
	if (status == 0) {
		memcpy(server->attr_bitmask, res.attr_bitmask, sizeof(server->attr_bitmask));
2597 2598 2599 2600 2601
		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
2602 2603 2604 2605 2606 2607
		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;
2608 2609 2610 2611 2612 2613 2614 2615 2616 2617 2618 2619 2620 2621 2622 2623
		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;
2624 2625 2626 2627 2628 2629
#ifdef CONFIG_NFS_V4_SECURITY_LABEL
		if (res.attr_bitmask[2] & FATTR4_WORD2_SECURITY_LABEL)
			server->caps |= NFS_CAP_SECURITY_LABEL;
#endif
		memcpy(server->attr_bitmask_nl, res.attr_bitmask,
				sizeof(server->attr_bitmask));
2630

2631 2632 2633 2634
		if (server->caps & NFS_CAP_SECURITY_LABEL) {
			server->attr_bitmask_nl[2] &= ~FATTR4_WORD2_SECURITY_LABEL;
			res.attr_bitmask[2] &= ~FATTR4_WORD2_SECURITY_LABEL;
		}
2635 2636 2637
		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
2638
		server->acl_bitmask = res.acl_bitmask;
2639
		server->fh_expire_type = res.fh_expire_type;
L
Linus Torvalds 已提交
2640
	}
A
Andy Adamson 已提交
2641

L
Linus Torvalds 已提交
2642 2643 2644
	return status;
}

2645
int nfs4_server_capabilities(struct nfs_server *server, struct nfs_fh *fhandle)
L
Linus Torvalds 已提交
2646 2647 2648 2649 2650 2651 2652 2653 2654 2655 2656 2657 2658 2659
{
	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)
{
2660
	u32 bitmask[3];
L
Linus Torvalds 已提交
2661
	struct nfs4_lookup_root_arg args = {
2662
		.bitmask = bitmask,
L
Linus Torvalds 已提交
2663 2664 2665
	};
	struct nfs4_lookup_res res = {
		.server = server,
2666
		.fattr = info->fattr,
L
Linus Torvalds 已提交
2667 2668 2669 2670 2671 2672 2673
		.fh = fhandle,
	};
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_LOOKUP_ROOT],
		.rpc_argp = &args,
		.rpc_resp = &res,
	};
2674

2675 2676 2677 2678 2679 2680 2681
	bitmask[0] = nfs4_fattr_bitmap[0];
	bitmask[1] = nfs4_fattr_bitmap[1];
	/*
	 * Process the label in the upcoming getfattr
	 */
	bitmask[2] = nfs4_fattr_bitmap[2] & ~FATTR4_WORD2_SECURITY_LABEL;

2682
	nfs_fattr_init(info->fattr);
2683
	return nfs4_call_sync(server->client, server, &msg, &args.seq_args, &res.seq_res, 0);
L
Linus Torvalds 已提交
2684 2685 2686 2687 2688 2689 2690 2691
}

static int nfs4_lookup_root(struct nfs_server *server, struct nfs_fh *fhandle,
		struct nfs_fsinfo *info)
{
	struct nfs4_exception exception = { };
	int err;
	do {
2692 2693 2694 2695
		err = _nfs4_lookup_root(server, fhandle, info);
		switch (err) {
		case 0:
		case -NFS4ERR_WRONGSEC:
2696
			goto out;
2697 2698 2699
		default:
			err = nfs4_handle_exception(server, err, &exception);
		}
L
Linus Torvalds 已提交
2700
	} while (exception.retry);
2701
out:
L
Linus Torvalds 已提交
2702 2703 2704
	return err;
}

2705 2706 2707 2708 2709 2710 2711
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);
2712
	if (IS_ERR(auth)) {
2713
		ret = -EACCES;
2714 2715 2716 2717 2718 2719 2720
		goto out;
	}
	ret = nfs4_lookup_root(server, fhandle, info);
out:
	return ret;
}

2721 2722 2723 2724 2725 2726 2727 2728 2729
/*
 * 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.
 */
2730
static int nfs4_find_root_sec(struct nfs_server *server, struct nfs_fh *fhandle,
2731
			      struct nfs_fsinfo *info)
L
Linus Torvalds 已提交
2732
{
2733 2734 2735 2736 2737
	/* Per 3530bis 15.33.5 */
	static const rpc_authflavor_t flav_array[] = {
		RPC_AUTH_GSS_KRB5P,
		RPC_AUTH_GSS_KRB5I,
		RPC_AUTH_GSS_KRB5,
2738
		RPC_AUTH_UNIX,			/* courtesy */
2739 2740 2741 2742
		RPC_AUTH_NULL,
	};
	int status = -EPERM;
	size_t i;
C
Chuck Lever 已提交
2743

2744
	for (i = 0; i < ARRAY_SIZE(flav_array); i++) {
2745
		status = nfs4_lookup_root_sec(server, fhandle, info, flav_array[i]);
2746
		if (status == -NFS4ERR_WRONGSEC || status == -EACCES)
2747 2748
			continue;
		break;
2749
	}
2750

2751 2752 2753 2754 2755 2756 2757 2758 2759
	/*
	 * -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;
2760 2761 2762
	return status;
}

C
Chuck Lever 已提交
2763 2764 2765 2766 2767 2768 2769 2770 2771 2772 2773 2774 2775 2776
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.
2777
 */
2778 2779
int nfs4_proc_get_rootfh(struct nfs_server *server, struct nfs_fh *fhandle,
			 struct nfs_fsinfo *info)
2780
{
C
Chuck Lever 已提交
2781 2782 2783 2784 2785 2786 2787
	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
2788 2789 2790 2791
	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
2792

2793
	return nfs4_map_errors(status);
L
Linus Torvalds 已提交
2794 2795
}

2796 2797 2798 2799 2800
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;
2801
	struct nfs4_label *label = NULL;
2802 2803 2804 2805 2806 2807 2808

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

2809 2810 2811 2812
	label = nfs4_label_alloc(server, GFP_KERNEL);
	if (IS_ERR(label))
		return PTR_ERR(label);

2813
	error = nfs4_proc_getattr(server, mntfh, fattr, label);
2814 2815
	if (error < 0) {
		dprintk("nfs4_get_root: getattr error = %d\n", -error);
2816
		goto err_free_label;
2817 2818 2819 2820 2821 2822
	}

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

2823 2824 2825
err_free_label:
	nfs4_label_free(label);

2826 2827 2828
	return error;
}

M
Manoj Naik 已提交
2829 2830 2831 2832 2833
/*
 * 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
 */
2834 2835 2836
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
2837 2838 2839 2840 2841 2842 2843 2844 2845 2846 2847 2848
{
	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;

2849
	status = nfs4_proc_fs_locations(client, dir, name, locations, page);
M
Manoj Naik 已提交
2850 2851 2852 2853
	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)) {
2854 2855
		dprintk("%s: server did not return a different fsid for"
			" a referral at %s\n", __func__, name->name);
M
Manoj Naik 已提交
2856 2857 2858
		status = -EIO;
		goto out;
	}
2859 2860
	/* Fixup attributes for the nfs_lookup() call to nfs_fhget() */
	nfs_fixup_referral_attributes(&locations->fattr);
M
Manoj Naik 已提交
2861

2862
	/* replace the lookup nfs_fattr with the locations nfs_fattr */
M
Manoj Naik 已提交
2863 2864 2865 2866 2867
	memcpy(fattr, &locations->fattr, sizeof(struct nfs_fattr));
	memset(fhandle, 0, sizeof(struct nfs_fh));
out:
	if (page)
		__free_page(page);
2868
	kfree(locations);
M
Manoj Naik 已提交
2869 2870 2871
	return status;
}

2872 2873
static int _nfs4_proc_getattr(struct nfs_server *server, struct nfs_fh *fhandle,
				struct nfs_fattr *fattr, struct nfs4_label *label)
L
Linus Torvalds 已提交
2874 2875 2876 2877 2878 2879 2880
{
	struct nfs4_getattr_arg args = {
		.fh = fhandle,
		.bitmask = server->attr_bitmask,
	};
	struct nfs4_getattr_res res = {
		.fattr = fattr,
2881
		.label = label,
L
Linus Torvalds 已提交
2882 2883 2884 2885 2886 2887 2888
		.server = server,
	};
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_GETATTR],
		.rpc_argp = &args,
		.rpc_resp = &res,
	};
2889 2890 2891

	args.bitmask = nfs4_bitmask(server, label);

2892
	nfs_fattr_init(fattr);
2893
	return nfs4_call_sync(server->client, server, &msg, &args.seq_args, &res.seq_res, 0);
L
Linus Torvalds 已提交
2894 2895
}

2896 2897
static int nfs4_proc_getattr(struct nfs_server *server, struct nfs_fh *fhandle,
				struct nfs_fattr *fattr, struct nfs4_label *label)
L
Linus Torvalds 已提交
2898 2899 2900 2901 2902
{
	struct nfs4_exception exception = { };
	int err;
	do {
		err = nfs4_handle_exception(server,
2903
				_nfs4_proc_getattr(server, fhandle, fattr, label),
L
Linus Torvalds 已提交
2904 2905 2906 2907 2908 2909 2910 2911 2912 2913 2914 2915 2916 2917 2918 2919 2920 2921 2922 2923 2924 2925 2926 2927 2928 2929
				&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)
{
2930
	struct inode *inode = dentry->d_inode;
2931
	struct rpc_cred *cred = NULL;
2932
	struct nfs4_state *state = NULL;
2933
	struct nfs4_label *label = NULL;
L
Linus Torvalds 已提交
2934 2935
	int status;

B
Benny Halevy 已提交
2936
	if (pnfs_ld_layoutret_on_setattr(inode))
2937
		pnfs_commit_and_return_layout(inode);
B
Benny Halevy 已提交
2938

2939
	nfs_fattr_init(fattr);
L
Linus Torvalds 已提交
2940
	
2941 2942 2943 2944 2945 2946 2947 2948
	/* 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;

2949
	/* Search for an existing open(O_WRITE) file */
2950 2951 2952 2953
	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
2954 2955 2956 2957
		if (ctx) {
			cred = ctx->cred;
			state = ctx->state;
		}
2958
	}
2959

2960 2961 2962 2963 2964
	label = nfs4_label_alloc(NFS_SERVER(inode), GFP_KERNEL);
	if (IS_ERR(label))
		return PTR_ERR(label);

	status = nfs4_do_setattr(inode, cred, fattr, sattr, state, NULL, label);
2965
	if (status == 0) {
2966
		nfs_setattr_update_inode(inode, sattr);
2967 2968
		nfs_setsecurity(inode, fattr, label);
	}
2969
	nfs4_label_free(label);
L
Linus Torvalds 已提交
2970 2971 2972
	return status;
}

2973 2974
static int _nfs4_proc_lookup(struct rpc_clnt *clnt, struct inode *dir,
		const struct qstr *name, struct nfs_fh *fhandle,
2975
		struct nfs_fattr *fattr, struct nfs4_label *label)
D
David Howells 已提交
2976
{
2977
	struct nfs_server *server = NFS_SERVER(dir);
D
David Howells 已提交
2978 2979 2980
	int		       status;
	struct nfs4_lookup_arg args = {
		.bitmask = server->attr_bitmask,
2981
		.dir_fh = NFS_FH(dir),
D
David Howells 已提交
2982 2983 2984 2985 2986
		.name = name,
	};
	struct nfs4_lookup_res res = {
		.server = server,
		.fattr = fattr,
2987
		.label = label,
D
David Howells 已提交
2988 2989 2990 2991 2992 2993 2994 2995
		.fh = fhandle,
	};
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_LOOKUP],
		.rpc_argp = &args,
		.rpc_resp = &res,
	};

2996 2997
	args.bitmask = nfs4_bitmask(server, label);

D
David Howells 已提交
2998 2999
	nfs_fattr_init(fattr);

L
Linus Torvalds 已提交
3000
	dprintk("NFS call  lookup %s\n", name->name);
3001
	status = nfs4_call_sync(clnt, server, &msg, &args.seq_args, &res.seq_res, 0);
L
Linus Torvalds 已提交
3002 3003 3004 3005
	dprintk("NFS reply lookup: %d\n", status);
	return status;
}

3006
static void nfs_fixup_secinfo_attributes(struct nfs_fattr *fattr)
3007 3008
{
	fattr->valid |= NFS_ATTR_FATTR_TYPE | NFS_ATTR_FATTR_MODE |
3009
		NFS_ATTR_FATTR_NLINK | NFS_ATTR_FATTR_MOUNTPOINT;
3010 3011 3012 3013
	fattr->mode = S_IFDIR | S_IRUGO | S_IXUGO;
	fattr->nlink = 2;
}

3014 3015
static int nfs4_proc_lookup_common(struct rpc_clnt **clnt, struct inode *dir,
				   struct qstr *name, struct nfs_fh *fhandle,
3016
				   struct nfs_fattr *fattr, struct nfs4_label *label)
L
Linus Torvalds 已提交
3017 3018
{
	struct nfs4_exception exception = { };
3019
	struct rpc_clnt *client = *clnt;
L
Linus Torvalds 已提交
3020 3021
	int err;
	do {
3022
		err = _nfs4_proc_lookup(client, dir, name, fhandle, fattr, label);
3023
		switch (err) {
3024
		case -NFS4ERR_BADNAME:
3025 3026
			err = -ENOENT;
			goto out;
3027
		case -NFS4ERR_MOVED:
3028
			err = nfs4_get_referral(client, dir, name, fattr, fhandle);
3029
			goto out;
3030
		case -NFS4ERR_WRONGSEC:
3031 3032 3033 3034 3035 3036 3037 3038 3039 3040 3041 3042
			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);
3043
		}
L
Linus Torvalds 已提交
3044
	} while (exception.retry);
3045 3046 3047 3048 3049 3050 3051

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

L
Linus Torvalds 已提交
3052 3053 3054
	return err;
}

3055
static int nfs4_proc_lookup(struct inode *dir, struct qstr *name,
3056 3057
			    struct nfs_fh *fhandle, struct nfs_fattr *fattr,
			    struct nfs4_label *label)
3058 3059 3060 3061
{
	int status;
	struct rpc_clnt *client = NFS_CLIENT(dir);

3062
	status = nfs4_proc_lookup_common(&client, dir, name, fhandle, fattr, label);
3063 3064 3065 3066 3067 3068 3069
	if (client != NFS_CLIENT(dir)) {
		rpc_shutdown_client(client);
		nfs_fixup_secinfo_attributes(fattr);
	}
	return status;
}

3070 3071 3072 3073 3074 3075 3076
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));

3077
	status = nfs4_proc_lookup_common(&client, dir, name, fhandle, fattr, NULL);
3078 3079 3080 3081 3082 3083 3084
	if (status < 0) {
		rpc_shutdown_client(client);
		return ERR_PTR(status);
	}
	return client;
}

L
Linus Torvalds 已提交
3085 3086
static int _nfs4_proc_access(struct inode *inode, struct nfs_access_entry *entry)
{
3087
	struct nfs_server *server = NFS_SERVER(inode);
L
Linus Torvalds 已提交
3088 3089
	struct nfs4_accessargs args = {
		.fh = NFS_FH(inode),
3090
		.bitmask = server->cache_consistency_bitmask,
3091 3092 3093
	};
	struct nfs4_accessres res = {
		.server = server,
L
Linus Torvalds 已提交
3094 3095 3096 3097 3098 3099 3100 3101
	};
	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;
3102
	int status = 0;
L
Linus Torvalds 已提交
3103 3104 3105 3106 3107 3108 3109 3110 3111 3112 3113 3114 3115 3116 3117 3118 3119

	/*
	 * 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;
	}
3120 3121 3122 3123 3124

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

3125
	status = nfs4_call_sync(server->client, server, &msg, &args.seq_args, &res.seq_res, 0);
L
Linus Torvalds 已提交
3126
	if (!status) {
3127
		nfs_access_set_mask(entry, res.access);
3128
		nfs_refresh_inode(inode, res.fattr);
L
Linus Torvalds 已提交
3129
	}
3130
	nfs_free_fattr(res.fattr);
L
Linus Torvalds 已提交
3131 3132 3133 3134 3135 3136 3137 3138 3139 3140 3141 3142 3143 3144 3145 3146 3147 3148 3149 3150 3151 3152 3153 3154 3155 3156 3157 3158 3159 3160 3161 3162
	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
3163
 * we get the post-operation mtime and size.
L
Linus Torvalds 已提交
3164 3165 3166 3167 3168 3169 3170 3171 3172 3173 3174 3175 3176
 *
 * 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
3177
	struct nfs4_readlink_res res;
L
Linus Torvalds 已提交
3178 3179 3180
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_READLINK],
		.rpc_argp = &args,
B
Benny Halevy 已提交
3181
		.rpc_resp = &res,
L
Linus Torvalds 已提交
3182 3183
	};

3184
	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
3185 3186 3187 3188 3189 3190 3191 3192 3193 3194 3195 3196 3197 3198 3199 3200
}

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;
}

/*
3201
 * This is just for mknod.  open(O_CREAT) will always do ->open_context().
L
Linus Torvalds 已提交
3202 3203 3204
 */
static int
nfs4_proc_create(struct inode *dir, struct dentry *dentry, struct iattr *sattr,
3205
		 int flags)
L
Linus Torvalds 已提交
3206
{
3207
	struct nfs4_label l, *ilabel = NULL;
3208
	struct nfs_open_context *ctx;
L
Linus Torvalds 已提交
3209 3210 3211
	struct nfs4_state *state;
	int status = 0;

3212 3213 3214 3215
	ctx = alloc_nfs_open_context(dentry, FMODE_READ);
	if (IS_ERR(ctx))
		return PTR_ERR(ctx);

3216 3217
	ilabel = nfs4_label_init_security(dir, dentry, sattr, &l);

A
Aneesh Kumar K.V 已提交
3218
	sattr->ia_mode &= ~current_umask();
3219
	state = nfs4_do_open(dir, ctx, flags, sattr, ilabel);
L
Linus Torvalds 已提交
3220 3221
	if (IS_ERR(state)) {
		status = PTR_ERR(state);
3222
		goto out;
L
Linus Torvalds 已提交
3223 3224
	}
out:
3225
	nfs4_label_release_security(ilabel);
3226
	put_nfs_open_context(ctx);
L
Linus Torvalds 已提交
3227 3228 3229 3230 3231
	return status;
}

static int _nfs4_proc_remove(struct inode *dir, struct qstr *name)
{
3232
	struct nfs_server *server = NFS_SERVER(dir);
3233
	struct nfs_removeargs args = {
L
Linus Torvalds 已提交
3234
		.fh = NFS_FH(dir),
3235
		.name = *name,
3236
	};
3237
	struct nfs_removeres res = {
3238
		.server = server,
L
Linus Torvalds 已提交
3239 3240
	};
	struct rpc_message msg = {
3241 3242 3243
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_REMOVE],
		.rpc_argp = &args,
		.rpc_resp = &res,
L
Linus Torvalds 已提交
3244
	};
3245
	int status;
L
Linus Torvalds 已提交
3246

3247
	status = nfs4_call_sync(server->client, server, &msg, &args.seq_args, &res.seq_res, 1);
3248
	if (status == 0)
3249
		update_changeattr(dir, &res.cinfo);
L
Linus Torvalds 已提交
3250 3251 3252 3253 3254 3255 3256 3257 3258 3259 3260 3261 3262 3263 3264
	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;
}

3265
static void nfs4_proc_unlink_setup(struct rpc_message *msg, struct inode *dir)
L
Linus Torvalds 已提交
3266
{
3267 3268 3269
	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
3270

3271
	res->server = server;
L
Linus Torvalds 已提交
3272
	msg->rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_REMOVE];
3273
	nfs41_init_sequence(&args->seq_args, &res->seq_res, 1);
3274 3275

	nfs_fattr_init(res->dir_attr);
L
Linus Torvalds 已提交
3276 3277
}

3278 3279
static void nfs4_proc_unlink_rpc_prepare(struct rpc_task *task, struct nfs_unlinkdata *data)
{
3280 3281 3282 3283
	nfs4_setup_sequence(NFS_SERVER(data->dir),
			&data->args.seq_args,
			&data->res.seq_res,
			task);
L
Linus Torvalds 已提交
3284 3285
}

3286
static int nfs4_proc_unlink_done(struct rpc_task *task, struct inode *dir)
L
Linus Torvalds 已提交
3287
{
3288 3289
	struct nfs_removeres *res = task->tk_msg.rpc_resp;

3290 3291
	if (!nfs4_sequence_done(task, &res->seq_res))
		return 0;
3292
	if (nfs4_async_handle_error(task, res->server, NULL) == -EAGAIN)
3293 3294 3295
		return 0;
	update_changeattr(dir, &res->cinfo);
	return 1;
L
Linus Torvalds 已提交
3296 3297
}

3298 3299 3300 3301 3302 3303 3304 3305
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;
3306
	nfs41_init_sequence(&arg->seq_args, &res->seq_res, 1);
3307 3308
}

3309 3310
static void nfs4_proc_rename_rpc_prepare(struct rpc_task *task, struct nfs_renamedata *data)
{
3311 3312 3313 3314
	nfs4_setup_sequence(NFS_SERVER(data->old_dir),
			&data->args.seq_args,
			&data->res.seq_res,
			task);
3315 3316 3317 3318 3319 3320 3321 3322 3323 3324 3325 3326 3327 3328 3329 3330 3331
}

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
3332 3333 3334
static int _nfs4_proc_rename(struct inode *old_dir, struct qstr *old_name,
		struct inode *new_dir, struct qstr *new_name)
{
3335
	struct nfs_server *server = NFS_SERVER(old_dir);
3336
	struct nfs_renameargs arg = {
L
Linus Torvalds 已提交
3337 3338 3339 3340
		.old_dir = NFS_FH(old_dir),
		.new_dir = NFS_FH(new_dir),
		.old_name = old_name,
		.new_name = new_name,
3341
	};
3342
	struct nfs_renameres res = {
3343
		.server = server,
L
Linus Torvalds 已提交
3344 3345 3346 3347 3348 3349
	};
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_RENAME],
		.rpc_argp = &arg,
		.rpc_resp = &res,
	};
3350
	int status = -ENOMEM;
3351

3352
	status = nfs4_call_sync(server->client, server, &msg, &arg.seq_args, &res.seq_res, 1);
L
Linus Torvalds 已提交
3353 3354 3355 3356 3357 3358 3359 3360 3361 3362 3363 3364 3365 3366 3367 3368 3369 3370 3371 3372 3373 3374 3375
	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)
{
3376
	struct nfs_server *server = NFS_SERVER(inode);
L
Linus Torvalds 已提交
3377 3378 3379 3380
	struct nfs4_link_arg arg = {
		.fh     = NFS_FH(inode),
		.dir_fh = NFS_FH(dir),
		.name   = name,
3381 3382 3383 3384
		.bitmask = server->attr_bitmask,
	};
	struct nfs4_link_res res = {
		.server = server,
3385
		.label = NULL,
L
Linus Torvalds 已提交
3386 3387 3388 3389
	};
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_LINK],
		.rpc_argp = &arg,
3390
		.rpc_resp = &res,
L
Linus Torvalds 已提交
3391
	};
3392 3393 3394
	int status = -ENOMEM;

	res.fattr = nfs_alloc_fattr();
3395
	if (res.fattr == NULL)
3396
		goto out;
L
Linus Torvalds 已提交
3397

3398 3399 3400 3401 3402
	res.label = nfs4_label_alloc(server, GFP_KERNEL);
	if (IS_ERR(res.label)) {
		status = PTR_ERR(res.label);
		goto out;
	}
3403
	arg.bitmask = nfs4_bitmask(server, res.label);
3404

3405
	status = nfs4_call_sync(server->client, server, &msg, &arg.seq_args, &res.seq_res, 1);
3406 3407
	if (!status) {
		update_changeattr(dir, &res.cinfo);
3408 3409 3410
		status = nfs_post_op_update_inode(inode, res.fattr);
		if (!status)
			nfs_setsecurity(inode, res.fattr, res.label);
3411
	}
3412 3413 3414 3415


	nfs4_label_free(res.label);

3416 3417
out:
	nfs_free_fattr(res.fattr);
L
Linus Torvalds 已提交
3418 3419 3420 3421 3422 3423 3424 3425 3426 3427 3428 3429 3430 3431 3432
	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;
}

3433 3434 3435 3436 3437 3438
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;
3439
	struct nfs4_label *label;
3440 3441 3442 3443 3444 3445 3446 3447 3448 3449 3450
};

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);

3451 3452 3453 3454
		data->label = nfs4_label_alloc(server, GFP_KERNEL);
		if (IS_ERR(data->label))
			goto out_free;

3455 3456 3457 3458 3459 3460 3461 3462
		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;
3463
		data->arg.bitmask = nfs4_bitmask(server, data->label);
3464 3465 3466
		data->res.server = server;
		data->res.fh = &data->fh;
		data->res.fattr = &data->fattr;
3467
		data->res.label = data->label;
3468 3469 3470
		nfs_fattr_init(data->res.fattr);
	}
	return data;
3471 3472 3473
out_free:
	kfree(data);
	return NULL;
3474 3475 3476 3477
}

static int nfs4_do_create(struct inode *dir, struct dentry *dentry, struct nfs4_createdata *data)
{
3478
	int status = nfs4_call_sync(NFS_SERVER(dir)->client, NFS_SERVER(dir), &data->msg,
3479
				    &data->arg.seq_args, &data->res.seq_res, 1);
3480 3481
	if (status == 0) {
		update_changeattr(dir, &data->res.dir_cinfo);
3482
		status = nfs_instantiate(dentry, data->res.fh, data->res.fattr, data->res.label);
3483 3484 3485 3486 3487 3488
	}
	return status;
}

static void nfs4_free_createdata(struct nfs4_createdata *data)
{
3489
	nfs4_label_free(data->label);
3490 3491 3492
	kfree(data);
}

3493
static int _nfs4_proc_symlink(struct inode *dir, struct dentry *dentry,
3494 3495
		struct page *page, unsigned int len, struct iattr *sattr,
		struct nfs4_label *label)
L
Linus Torvalds 已提交
3496
{
3497 3498
	struct nfs4_createdata *data;
	int status = -ENAMETOOLONG;
L
Linus Torvalds 已提交
3499

3500
	if (len > NFS4_MAXPATHLEN)
3501
		goto out;
3502

3503 3504 3505 3506 3507 3508 3509 3510
	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;
3511
	data->arg.label = label;
L
Linus Torvalds 已提交
3512
	
3513 3514 3515 3516
	status = nfs4_do_create(dir, dentry, data);

	nfs4_free_createdata(data);
out:
L
Linus Torvalds 已提交
3517 3518 3519
	return status;
}

3520
static int nfs4_proc_symlink(struct inode *dir, struct dentry *dentry,
3521
		struct page *page, unsigned int len, struct iattr *sattr)
L
Linus Torvalds 已提交
3522 3523
{
	struct nfs4_exception exception = { };
3524
	struct nfs4_label l, *label = NULL;
L
Linus Torvalds 已提交
3525
	int err;
3526 3527 3528

	label = nfs4_label_init_security(dir, dentry, sattr, &l);

L
Linus Torvalds 已提交
3529 3530
	do {
		err = nfs4_handle_exception(NFS_SERVER(dir),
3531
				_nfs4_proc_symlink(dir, dentry, page,
3532
							len, sattr, label),
L
Linus Torvalds 已提交
3533 3534
				&exception);
	} while (exception.retry);
3535 3536

	nfs4_label_release_security(label);
L
Linus Torvalds 已提交
3537 3538 3539 3540
	return err;
}

static int _nfs4_proc_mkdir(struct inode *dir, struct dentry *dentry,
3541
		struct iattr *sattr, struct nfs4_label *label)
L
Linus Torvalds 已提交
3542
{
3543 3544
	struct nfs4_createdata *data;
	int status = -ENOMEM;
L
Linus Torvalds 已提交
3545

3546 3547 3548 3549
	data = nfs4_alloc_createdata(dir, &dentry->d_name, sattr, NF4DIR);
	if (data == NULL)
		goto out;

3550
	data->arg.label = label;
3551 3552 3553 3554
	status = nfs4_do_create(dir, dentry, data);

	nfs4_free_createdata(data);
out:
L
Linus Torvalds 已提交
3555 3556 3557 3558 3559 3560 3561
	return status;
}

static int nfs4_proc_mkdir(struct inode *dir, struct dentry *dentry,
		struct iattr *sattr)
{
	struct nfs4_exception exception = { };
3562
	struct nfs4_label l, *label = NULL;
L
Linus Torvalds 已提交
3563
	int err;
A
Aneesh Kumar K.V 已提交
3564

3565 3566
	label = nfs4_label_init_security(dir, dentry, sattr, &l);

A
Aneesh Kumar K.V 已提交
3567
	sattr->ia_mode &= ~current_umask();
L
Linus Torvalds 已提交
3568 3569
	do {
		err = nfs4_handle_exception(NFS_SERVER(dir),
3570
				_nfs4_proc_mkdir(dir, dentry, sattr, label),
L
Linus Torvalds 已提交
3571 3572
				&exception);
	} while (exception.retry);
3573 3574
	nfs4_label_release_security(label);

L
Linus Torvalds 已提交
3575 3576 3577 3578
	return err;
}

static int _nfs4_proc_readdir(struct dentry *dentry, struct rpc_cred *cred,
B
Bryan Schumaker 已提交
3579
		u64 cookie, struct page **pages, unsigned int count, int plus)
L
Linus Torvalds 已提交
3580 3581 3582 3583
{
	struct inode		*dir = dentry->d_inode;
	struct nfs4_readdir_arg args = {
		.fh = NFS_FH(dir),
B
Bryan Schumaker 已提交
3584
		.pages = pages,
L
Linus Torvalds 已提交
3585 3586
		.pgbase = 0,
		.count = count,
3587
		.bitmask = NFS_SERVER(dentry->d_inode)->attr_bitmask,
B
Bryan Schumaker 已提交
3588
		.plus = plus,
L
Linus Torvalds 已提交
3589 3590 3591 3592 3593 3594 3595 3596 3597 3598
	};
	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;

3599
	dprintk("%s: dentry = %s/%s, cookie = %Lu\n", __func__,
3600 3601 3602
			dentry->d_parent->d_name.name,
			dentry->d_name.name,
			(unsigned long long)cookie);
3603
	nfs4_setup_readdir(cookie, NFS_I(dir)->cookieverf, dentry, &args);
L
Linus Torvalds 已提交
3604
	res.pgbase = args.pgbase;
3605
	status = nfs4_call_sync(NFS_SERVER(dir)->client, NFS_SERVER(dir), &msg, &args.seq_args, &res.seq_res, 0);
3606
	if (status >= 0) {
3607
		memcpy(NFS_I(dir)->cookieverf, res.verifier.data, NFS4_VERIFIER_SIZE);
3608 3609
		status += args.pgbase;
	}
3610 3611 3612

	nfs_invalidate_atime(dir);

3613
	dprintk("%s: returns %d\n", __func__, status);
L
Linus Torvalds 已提交
3614 3615 3616 3617
	return status;
}

static int nfs4_proc_readdir(struct dentry *dentry, struct rpc_cred *cred,
B
Bryan Schumaker 已提交
3618
		u64 cookie, struct page **pages, unsigned int count, int plus)
L
Linus Torvalds 已提交
3619 3620 3621 3622 3623 3624
{
	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
3625
					pages, count, plus),
L
Linus Torvalds 已提交
3626 3627 3628 3629 3630 3631
				&exception);
	} while (exception.retry);
	return err;
}

static int _nfs4_proc_mknod(struct inode *dir, struct dentry *dentry,
3632
		struct iattr *sattr, struct nfs4_label *label, dev_t rdev)
L
Linus Torvalds 已提交
3633
{
3634 3635 3636
	struct nfs4_createdata *data;
	int mode = sattr->ia_mode;
	int status = -ENOMEM;
L
Linus Torvalds 已提交
3637

3638 3639 3640 3641
	data = nfs4_alloc_createdata(dir, &dentry->d_name, sattr, NF4SOCK);
	if (data == NULL)
		goto out;

L
Linus Torvalds 已提交
3642
	if (S_ISFIFO(mode))
3643
		data->arg.ftype = NF4FIFO;
L
Linus Torvalds 已提交
3644
	else if (S_ISBLK(mode)) {
3645 3646 3647
		data->arg.ftype = NF4BLK;
		data->arg.u.device.specdata1 = MAJOR(rdev);
		data->arg.u.device.specdata2 = MINOR(rdev);
L
Linus Torvalds 已提交
3648 3649
	}
	else if (S_ISCHR(mode)) {
3650 3651 3652
		data->arg.ftype = NF4CHR;
		data->arg.u.device.specdata1 = MAJOR(rdev);
		data->arg.u.device.specdata2 = MINOR(rdev);
3653 3654 3655
	} else if (!S_ISSOCK(mode)) {
		status = -EINVAL;
		goto out_free;
L
Linus Torvalds 已提交
3656
	}
3657

3658
	data->arg.label = label;
3659
	status = nfs4_do_create(dir, dentry, data);
3660
out_free:
3661 3662
	nfs4_free_createdata(data);
out:
L
Linus Torvalds 已提交
3663 3664 3665 3666 3667 3668 3669
	return status;
}

static int nfs4_proc_mknod(struct inode *dir, struct dentry *dentry,
		struct iattr *sattr, dev_t rdev)
{
	struct nfs4_exception exception = { };
3670
	struct nfs4_label l, *label = NULL;
L
Linus Torvalds 已提交
3671
	int err;
A
Aneesh Kumar K.V 已提交
3672

3673 3674
	label = nfs4_label_init_security(dir, dentry, sattr, &l);

A
Aneesh Kumar K.V 已提交
3675
	sattr->ia_mode &= ~current_umask();
L
Linus Torvalds 已提交
3676 3677
	do {
		err = nfs4_handle_exception(NFS_SERVER(dir),
3678
				_nfs4_proc_mknod(dir, dentry, sattr, label, rdev),
L
Linus Torvalds 已提交
3679 3680
				&exception);
	} while (exception.retry);
3681 3682 3683

	nfs4_label_release_security(label);

L
Linus Torvalds 已提交
3684 3685 3686 3687 3688 3689 3690 3691 3692 3693
	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
3694 3695 3696
	struct nfs4_statfs_res res = {
		.fsstat = fsstat,
	};
L
Linus Torvalds 已提交
3697 3698 3699
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_STATFS],
		.rpc_argp = &args,
B
Benny Halevy 已提交
3700
		.rpc_resp = &res,
L
Linus Torvalds 已提交
3701 3702
	};

3703
	nfs_fattr_init(fsstat->fattr);
3704
	return  nfs4_call_sync(server->client, server, &msg, &args.seq_args, &res.seq_res, 0);
L
Linus Torvalds 已提交
3705 3706 3707 3708 3709 3710 3711 3712 3713 3714 3715 3716 3717 3718 3719 3720 3721 3722 3723 3724 3725
}

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
3726 3727 3728
	struct nfs4_fsinfo_res res = {
		.fsinfo = fsinfo,
	};
L
Linus Torvalds 已提交
3729 3730 3731
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_FSINFO],
		.rpc_argp = &args,
B
Benny Halevy 已提交
3732
		.rpc_resp = &res,
L
Linus Torvalds 已提交
3733 3734
	};

3735
	return nfs4_call_sync(server->client, server, &msg, &args.seq_args, &res.seq_res, 0);
L
Linus Torvalds 已提交
3736 3737 3738 3739 3740
}

static int nfs4_do_fsinfo(struct nfs_server *server, struct nfs_fh *fhandle, struct nfs_fsinfo *fsinfo)
{
	struct nfs4_exception exception = { };
3741
	unsigned long now = jiffies;
L
Linus Torvalds 已提交
3742 3743 3744
	int err;

	do {
3745 3746 3747 3748 3749 3750 3751 3752 3753 3754 3755
		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
3756 3757 3758 3759 3760 3761
	} while (exception.retry);
	return err;
}

static int nfs4_proc_fsinfo(struct nfs_server *server, struct nfs_fh *fhandle, struct nfs_fsinfo *fsinfo)
{
3762 3763
	int error;

3764
	nfs_fattr_init(fsinfo->fattr);
3765
	error = nfs4_do_fsinfo(server, fhandle, fsinfo);
3766 3767 3768
	if (error == 0) {
		/* block layout checks this! */
		server->pnfs_blksize = fsinfo->blksize;
3769
		set_pnfs_layoutdriver(server, fhandle, fsinfo->layouttype);
3770
	}
3771 3772

	return error;
L
Linus Torvalds 已提交
3773 3774 3775 3776 3777 3778 3779 3780 3781
}

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
3782 3783 3784
	struct nfs4_pathconf_res res = {
		.pathconf = pathconf,
	};
L
Linus Torvalds 已提交
3785 3786 3787
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_PATHCONF],
		.rpc_argp = &args,
B
Benny Halevy 已提交
3788
		.rpc_resp = &res,
L
Linus Torvalds 已提交
3789 3790 3791 3792 3793 3794 3795 3796
	};

	/* 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;
	}

3797
	nfs_fattr_init(pathconf->fattr);
3798
	return nfs4_call_sync(server->client, server, &msg, &args.seq_args, &res.seq_res, 0);
L
Linus Torvalds 已提交
3799 3800 3801 3802 3803 3804 3805 3806 3807 3808 3809 3810 3811 3812 3813 3814
}

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;
}

3815
int nfs4_set_rw_stateid(nfs4_stateid *stateid,
3816 3817 3818 3819 3820 3821 3822 3823
		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;
3824
	return nfs4_select_rw_stateid(stateid, ctx->state, fmode, lockowner);
3825 3826 3827
}
EXPORT_SYMBOL_GPL(nfs4_set_rw_stateid);

3828 3829 3830 3831 3832 3833 3834 3835 3836 3837 3838 3839 3840 3841 3842 3843 3844 3845 3846 3847 3848 3849 3850 3851 3852 3853 3854
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;
}

3855 3856
void __nfs4_read_done_cb(struct nfs_read_data *data)
{
3857
	nfs_invalidate_atime(data->header->inode);
3858 3859
}

3860
static int nfs4_read_done_cb(struct rpc_task *task, struct nfs_read_data *data)
L
Linus Torvalds 已提交
3861
{
3862
	struct nfs_server *server = NFS_SERVER(data->header->inode);
L
Linus Torvalds 已提交
3863

3864
	if (nfs4_async_handle_error(task, server, data->args.context->state) == -EAGAIN) {
3865
		rpc_restart_call_prepare(task);
T
Trond Myklebust 已提交
3866
		return -EAGAIN;
L
Linus Torvalds 已提交
3867
	}
3868

3869
	__nfs4_read_done_cb(data);
L
Linus Torvalds 已提交
3870
	if (task->tk_status > 0)
T
Trond Myklebust 已提交
3871 3872
		renew_lease(server, data->timestamp);
	return 0;
L
Linus Torvalds 已提交
3873 3874
}

3875 3876 3877 3878 3879 3880 3881 3882 3883 3884 3885 3886 3887 3888
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;
}

3889 3890 3891 3892 3893 3894 3895
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;
3896 3897
	if (nfs4_read_stateid_changed(task, &data->args))
		return -EAGAIN;
3898 3899
	return data->read_done_cb ? data->read_done_cb(task, data) :
				    nfs4_read_done_cb(task, data);
3900 3901
}

3902
static void nfs4_proc_read_setup(struct nfs_read_data *data, struct rpc_message *msg)
L
Linus Torvalds 已提交
3903 3904
{
	data->timestamp   = jiffies;
3905
	data->read_done_cb = nfs4_read_done_cb;
3906
	msg->rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_READ];
3907
	nfs41_init_sequence(&data->args.seq_args, &data->res.seq_res, 0);
L
Linus Torvalds 已提交
3908 3909
}

3910 3911
static void nfs4_proc_read_rpc_prepare(struct rpc_task *task, struct nfs_read_data *data)
{
3912
	if (nfs4_setup_sequence(NFS_SERVER(data->header->inode),
3913 3914
			&data->args.seq_args,
			&data->res.seq_res,
3915 3916 3917 3918
			task))
		return;
	nfs4_set_rw_stateid(&data->args.stateid, data->args.context,
			data->args.lock_context, FMODE_READ);
L
Linus Torvalds 已提交
3919 3920
}

3921
static int nfs4_write_done_cb(struct rpc_task *task, struct nfs_write_data *data)
L
Linus Torvalds 已提交
3922
{
3923
	struct inode *inode = data->header->inode;
L
Linus Torvalds 已提交
3924
	
3925
	if (nfs4_async_handle_error(task, NFS_SERVER(inode), data->args.context->state) == -EAGAIN) {
3926
		rpc_restart_call_prepare(task);
3927
		return -EAGAIN;
L
Linus Torvalds 已提交
3928
	}
3929
	if (task->tk_status >= 0) {
L
Linus Torvalds 已提交
3930
		renew_lease(NFS_SERVER(inode), data->timestamp);
3931
		nfs_post_op_update_inode_force_wcc(inode, &data->fattr);
3932
	}
3933
	return 0;
L
Linus Torvalds 已提交
3934 3935
}

3936 3937 3938 3939 3940 3941 3942 3943 3944 3945 3946 3947 3948 3949
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;
}

3950 3951 3952 3953
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;
3954 3955
	if (nfs4_write_stateid_changed(task, &data->args))
		return -EAGAIN;
3956 3957
	return data->write_done_cb ? data->write_done_cb(task, data) :
		nfs4_write_done_cb(task, data);
3958 3959
}

3960 3961
static
bool nfs4_write_need_cache_consistency_data(const struct nfs_write_data *data)
3962
{
3963 3964 3965 3966 3967 3968 3969 3970
	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
	 */
3971
	return nfs4_have_delegation(hdr->inode, FMODE_READ) == 0;
3972 3973
}

3974
static void nfs4_proc_write_setup(struct nfs_write_data *data, struct rpc_message *msg)
L
Linus Torvalds 已提交
3975
{
3976
	struct nfs_server *server = NFS_SERVER(data->header->inode);
3977

3978
	if (!nfs4_write_need_cache_consistency_data(data)) {
3979 3980 3981 3982
		data->args.bitmask = NULL;
		data->res.fattr = NULL;
	} else
		data->args.bitmask = server->cache_consistency_bitmask;
3983

3984 3985
	if (!data->write_done_cb)
		data->write_done_cb = nfs4_write_done_cb;
3986
	data->res.server = server;
L
Linus Torvalds 已提交
3987 3988
	data->timestamp   = jiffies;

3989
	msg->rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_WRITE];
3990
	nfs41_init_sequence(&data->args.seq_args, &data->res.seq_res, 1);
L
Linus Torvalds 已提交
3991 3992
}

3993 3994
static void nfs4_proc_write_rpc_prepare(struct rpc_task *task, struct nfs_write_data *data)
{
3995
	if (nfs4_setup_sequence(NFS_SERVER(data->header->inode),
3996 3997
			&data->args.seq_args,
			&data->res.seq_res,
3998 3999 4000 4001
			task))
		return;
	nfs4_set_rw_stateid(&data->args.stateid, data->args.context,
			data->args.lock_context, FMODE_WRITE);
L
Linus Torvalds 已提交
4002 4003
}

4004
static void nfs4_proc_commit_rpc_prepare(struct rpc_task *task, struct nfs_commit_data *data)
4005
{
4006 4007 4008 4009
	nfs4_setup_sequence(NFS_SERVER(data->inode),
			&data->args.seq_args,
			&data->res.seq_res,
			task);
L
Linus Torvalds 已提交
4010 4011
}

4012
static int nfs4_commit_done_cb(struct rpc_task *task, struct nfs_commit_data *data)
L
Linus Torvalds 已提交
4013 4014
{
	struct inode *inode = data->inode;
4015

4016
	if (nfs4_async_handle_error(task, NFS_SERVER(inode), NULL) == -EAGAIN) {
4017
		rpc_restart_call_prepare(task);
4018
		return -EAGAIN;
L
Linus Torvalds 已提交
4019
	}
4020
	return 0;
L
Linus Torvalds 已提交
4021 4022
}

4023
static int nfs4_commit_done(struct rpc_task *task, struct nfs_commit_data *data)
4024 4025 4026
{
	if (!nfs4_sequence_done(task, &data->res.seq_res))
		return -EAGAIN;
4027
	return data->commit_done_cb(task, data);
4028 4029
}

4030
static void nfs4_proc_commit_setup(struct nfs_commit_data *data, struct rpc_message *msg)
L
Linus Torvalds 已提交
4031
{
4032
	struct nfs_server *server = NFS_SERVER(data->inode);
4033

4034 4035
	if (data->commit_done_cb == NULL)
		data->commit_done_cb = nfs4_commit_done_cb;
4036
	data->res.server = server;
4037
	msg->rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_COMMIT];
4038
	nfs41_init_sequence(&data->args.seq_args, &data->res.seq_res, 1);
L
Linus Torvalds 已提交
4039 4040
}

4041 4042 4043 4044 4045
struct nfs4_renewdata {
	struct nfs_client	*client;
	unsigned long		timestamp;
};

L
Linus Torvalds 已提交
4046 4047 4048 4049
/*
 * nfs4_proc_async_renew(): This is not one of the nfs_rpc_ops; it is a special
 * standalone procedure for queueing an asynchronous RENEW.
 */
4050
static void nfs4_renew_release(void *calldata)
4051
{
4052 4053
	struct nfs4_renewdata *data = calldata;
	struct nfs_client *clp = data->client;
4054

4055 4056 4057
	if (atomic_read(&clp->cl_count) > 1)
		nfs4_schedule_state_renewal(clp);
	nfs_put_client(clp);
4058
	kfree(data);
4059 4060
}

4061
static void nfs4_renew_done(struct rpc_task *task, void *calldata)
L
Linus Torvalds 已提交
4062
{
4063 4064 4065
	struct nfs4_renewdata *data = calldata;
	struct nfs_client *clp = data->client;
	unsigned long timestamp = data->timestamp;
L
Linus Torvalds 已提交
4066 4067

	if (task->tk_status < 0) {
4068
		/* Unless we're shutting down, schedule state recovery! */
4069 4070 4071
		if (test_bit(NFS_CS_RENEWD, &clp->cl_res_state) == 0)
			return;
		if (task->tk_status != NFS4ERR_CB_PATH_DOWN) {
4072
			nfs4_schedule_lease_recovery(clp);
4073 4074 4075
			return;
		}
		nfs4_schedule_path_down_recovery(clp);
L
Linus Torvalds 已提交
4076
	}
4077
	do_renew_lease(clp, timestamp);
L
Linus Torvalds 已提交
4078 4079
}

4080 4081
static const struct rpc_call_ops nfs4_renew_ops = {
	.rpc_call_done = nfs4_renew_done,
4082
	.rpc_release = nfs4_renew_release,
4083 4084
};

4085
static int nfs4_proc_async_renew(struct nfs_client *clp, struct rpc_cred *cred, unsigned renew_flags)
L
Linus Torvalds 已提交
4086 4087 4088 4089
{
	struct rpc_message msg = {
		.rpc_proc	= &nfs4_procedures[NFSPROC4_CLNT_RENEW],
		.rpc_argp	= clp,
4090
		.rpc_cred	= cred,
L
Linus Torvalds 已提交
4091
	};
4092
	struct nfs4_renewdata *data;
L
Linus Torvalds 已提交
4093

4094 4095
	if (renew_flags == 0)
		return 0;
4096 4097
	if (!atomic_inc_not_zero(&clp->cl_count))
		return -EIO;
4098
	data = kmalloc(sizeof(*data), GFP_NOFS);
4099 4100 4101 4102
	if (data == NULL)
		return -ENOMEM;
	data->client = clp;
	data->timestamp = jiffies;
4103
	return rpc_call_async(clp->cl_rpcclient, &msg, RPC_TASK_TIMEOUT,
4104
			&nfs4_renew_ops, data);
L
Linus Torvalds 已提交
4105 4106
}

4107
static int nfs4_proc_renew(struct nfs_client *clp, struct rpc_cred *cred)
L
Linus Torvalds 已提交
4108 4109 4110 4111
{
	struct rpc_message msg = {
		.rpc_proc	= &nfs4_procedures[NFSPROC4_CLNT_RENEW],
		.rpc_argp	= clp,
4112
		.rpc_cred	= cred,
L
Linus Torvalds 已提交
4113 4114 4115 4116
	};
	unsigned long now = jiffies;
	int status;

4117
	status = rpc_call_sync(clp->cl_rpcclient, &msg, RPC_TASK_TIMEOUT);
L
Linus Torvalds 已提交
4118 4119
	if (status < 0)
		return status;
4120
	do_renew_lease(clp, now);
L
Linus Torvalds 已提交
4121 4122 4123
	return 0;
}

4124 4125 4126 4127 4128 4129 4130
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);
}

4131 4132
/* 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
4133 4134
 * the stack.
 */
4135
#define NFS4ACL_MAXPAGES DIV_ROUND_UP(XATTR_SIZE_MAX, PAGE_SIZE)
4136

4137 4138 4139 4140 4141 4142 4143 4144 4145
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 {
4146
		len = min_t(size_t, PAGE_SIZE, buflen);
4147 4148 4149 4150 4151 4152 4153 4154 4155 4156 4157 4158 4159 4160 4161 4162 4163 4164 4165
		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;
}

4166 4167 4168
struct nfs4_cached_acl {
	int cached;
	size_t len;
A
Andrew Morton 已提交
4169
	char data[0];
4170 4171 4172 4173 4174 4175 4176 4177 4178 4179 4180 4181 4182 4183 4184 4185 4186 4187 4188 4189 4190 4191 4192 4193 4194 4195 4196 4197 4198 4199 4200 4201 4202 4203 4204 4205 4206 4207 4208 4209 4210 4211
};

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;
}

4212
static void nfs4_write_cached_acl(struct inode *inode, struct page **pages, size_t pgbase, size_t acl_len)
4213 4214
{
	struct nfs4_cached_acl *acl;
4215
	size_t buflen = sizeof(*acl) + acl_len;
4216

4217
	if (buflen <= PAGE_SIZE) {
4218
		acl = kmalloc(buflen, GFP_KERNEL);
4219 4220 4221
		if (acl == NULL)
			goto out;
		acl->cached = 1;
4222
		_copy_from_pages(acl->data, pages, pgbase, acl_len);
4223 4224 4225 4226 4227 4228 4229 4230 4231 4232 4233
	} 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);
}

4234 4235 4236 4237 4238 4239 4240 4241 4242 4243
/*
 * 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.
 */
4244
static ssize_t __nfs4_get_acl_uncached(struct inode *inode, void *buf, size_t buflen)
4245
{
4246
	struct page *pages[NFS4ACL_MAXPAGES] = {NULL, };
4247 4248 4249 4250 4251
	struct nfs_getaclargs args = {
		.fh = NFS_FH(inode),
		.acl_pages = pages,
		.acl_len = buflen,
	};
B
Benny Halevy 已提交
4252 4253 4254
	struct nfs_getaclres res = {
		.acl_len = buflen,
	};
4255 4256 4257
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_GETACL],
		.rpc_argp = &args,
B
Benny Halevy 已提交
4258
		.rpc_resp = &res,
4259
	};
4260 4261
	unsigned int npages = DIV_ROUND_UP(buflen, PAGE_SIZE);
	int ret = -ENOMEM, i;
4262

4263 4264 4265 4266
	/* 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;
4267 4268
	if (npages > ARRAY_SIZE(pages))
		return -ERANGE;
4269

4270 4271 4272 4273
	for (i = 0; i < npages; i++) {
		pages[i] = alloc_page(GFP_KERNEL);
		if (!pages[i])
			goto out_free;
4274
	}
4275 4276 4277 4278 4279 4280

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

4281 4282
	args.acl_len = npages * PAGE_SIZE;
	args.acl_pgbase = 0;
4283

4284
	dprintk("%s  buf %p buflen %zu npages %d args.acl_len %zu\n",
4285 4286 4287
		__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);
4288 4289
	if (ret)
		goto out_free;
4290

4291 4292 4293 4294 4295
	/* 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;
4296
		ret = -ERANGE;
4297
		goto out_free;
4298
	}
4299
	nfs4_write_cached_acl(inode, pages, res.acl_data_offset, res.acl_len);
4300 4301 4302 4303 4304
	if (buf) {
		if (res.acl_len > buflen) {
			ret = -ERANGE;
			goto out_free;
		}
4305
		_copy_from_pages(buf, pages, res.acl_data_offset, res.acl_len);
4306
	}
4307 4308
out_ok:
	ret = res.acl_len;
4309
out_free:
4310 4311 4312
	for (i = 0; i < npages; i++)
		if (pages[i])
			__free_page(pages[i]);
4313 4314
	if (res.acl_scratch)
		__free_page(res.acl_scratch);
4315 4316 4317
	return ret;
}

4318 4319 4320 4321 4322 4323 4324 4325 4326 4327 4328 4329 4330
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;
}

4331 4332 4333 4334 4335 4336 4337 4338 4339 4340
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;
4341 4342
	if (NFS_I(inode)->cache_validity & NFS_INO_INVALID_ACL)
		nfs_zap_acl_cache(inode);
4343 4344
	ret = nfs4_read_cached_acl(inode, buf, buflen);
	if (ret != -ENOENT)
4345 4346
		/* -ENOENT is returned if there is no ACL or if there is an ACL
		 * but no cached acl data, just the acl length */
4347 4348 4349 4350
		return ret;
	return nfs4_get_acl_uncached(inode, buf, buflen);
}

4351
static int __nfs4_proc_set_acl(struct inode *inode, const void *buf, size_t buflen)
4352 4353 4354 4355 4356 4357 4358 4359
{
	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
4360
	struct nfs_setaclres res;
4361 4362 4363
	struct rpc_message msg = {
		.rpc_proc	= &nfs4_procedures[NFSPROC4_CLNT_SETACL],
		.rpc_argp	= &arg,
B
Benny Halevy 已提交
4364
		.rpc_resp	= &res,
4365
	};
4366
	unsigned int npages = DIV_ROUND_UP(buflen, PAGE_SIZE);
4367
	int ret, i;
4368 4369 4370

	if (!nfs4_server_supports_acls(server))
		return -EOPNOTSUPP;
4371 4372
	if (npages > ARRAY_SIZE(pages))
		return -ERANGE;
4373 4374 4375
	i = buf_to_pages_noslab(buf, buflen, arg.acl_pages, &arg.acl_pgbase);
	if (i < 0)
		return i;
4376
	nfs4_inode_return_delegation(inode);
4377
	ret = nfs4_call_sync(server->client, server, &msg, &arg.seq_args, &res.seq_res, 1);
4378 4379 4380 4381 4382 4383 4384 4385

	/*
	 * 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]);

4386 4387 4388 4389 4390 4391 4392
	/*
	 * 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);
4393 4394
	nfs_access_zap_cache(inode);
	nfs_zap_acl_cache(inode);
4395 4396 4397
	return ret;
}

4398 4399 4400 4401 4402 4403 4404 4405 4406 4407 4408 4409
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;
}

4410 4411 4412 4413 4414 4415 4416 4417 4418 4419 4420 4421 4422 4423 4424 4425 4426 4427 4428 4429 4430 4431 4432 4433 4434 4435 4436 4437 4438 4439 4440 4441 4442 4443 4444 4445 4446 4447 4448 4449 4450 4451 4452 4453 4454 4455 4456 4457 4458 4459 4460 4461 4462 4463 4464 4465 4466 4467 4468 4469 4470 4471 4472 4473 4474 4475 4476 4477 4478 4479 4480 4481 4482 4483 4484 4485 4486 4487 4488 4489 4490 4491 4492 4493 4494 4495 4496 4497 4498 4499 4500 4501 4502 4503 4504 4505 4506 4507 4508 4509 4510 4511 4512 4513 4514 4515 4516 4517 4518 4519 4520 4521 4522 4523 4524 4525 4526 4527 4528 4529 4530 4531 4532 4533 4534 4535 4536 4537 4538 4539 4540 4541 4542 4543 4544 4545 4546 4547 4548 4549 4550 4551 4552 4553 4554 4555 4556 4557 4558
#ifdef CONFIG_NFS_V4_SECURITY_LABEL
static int _nfs4_get_security_label(struct inode *inode, void *buf,
					size_t buflen)
{
	struct nfs_server *server = NFS_SERVER(inode);
	struct nfs_fattr fattr;
	struct nfs4_label label = {0, 0, buflen, buf};

	u32 bitmask[3] = { 0, 0, FATTR4_WORD2_SECURITY_LABEL };
	struct nfs4_getattr_arg args = {
		.fh		= NFS_FH(inode),
		.bitmask	= bitmask,
	};
	struct nfs4_getattr_res res = {
		.fattr		= &fattr,
		.label		= &label,
		.server		= server,
	};
	struct rpc_message msg = {
		.rpc_proc	= &nfs4_procedures[NFSPROC4_CLNT_GETATTR],
		.rpc_argp	= &args,
		.rpc_resp	= &res,
	};
	int ret;

	nfs_fattr_init(&fattr);

	ret = rpc_call_sync(server->client, &msg, 0);
	if (ret)
		return ret;
	if (!(fattr.valid & NFS_ATTR_FATTR_V4_SECURITY_LABEL))
		return -ENOENT;
	if (buflen < label.len)
		return -ERANGE;
	return 0;
}

static int nfs4_get_security_label(struct inode *inode, void *buf,
					size_t buflen)
{
	struct nfs4_exception exception = { };
	int err;

	if (!nfs_server_capable(inode, NFS_CAP_SECURITY_LABEL))
		return -EOPNOTSUPP;

	do {
		err = nfs4_handle_exception(NFS_SERVER(inode),
				_nfs4_get_security_label(inode, buf, buflen),
				&exception);
	} while (exception.retry);
	return err;
}

static int _nfs4_do_set_security_label(struct inode *inode,
		struct nfs4_label *ilabel,
		struct nfs_fattr *fattr,
		struct nfs4_label *olabel)
{

	struct iattr sattr = {0};
	struct nfs_server *server = NFS_SERVER(inode);
	const u32 bitmask[3] = { 0, 0, FATTR4_WORD2_SECURITY_LABEL };
	struct nfs_setattrargs args = {
		.fh             = NFS_FH(inode),
		.iap            = &sattr,
		.server		= server,
		.bitmask	= bitmask,
		.label		= ilabel,
	};
	struct nfs_setattrres res = {
		.fattr		= fattr,
		.label		= olabel,
		.server		= server,
	};
	struct rpc_message msg = {
		.rpc_proc       = &nfs4_procedures[NFSPROC4_CLNT_SETATTR],
		.rpc_argp       = &args,
		.rpc_resp       = &res,
	};
	int status;

	nfs4_stateid_copy(&args.stateid, &zero_stateid);

	status = rpc_call_sync(server->client, &msg, 0);
	if (status)
		dprintk("%s failed: %d\n", __func__, status);

	return status;
}

static int nfs4_do_set_security_label(struct inode *inode,
		struct nfs4_label *ilabel,
		struct nfs_fattr *fattr,
		struct nfs4_label *olabel)
{
	struct nfs4_exception exception = { };
	int err;

	do {
		err = nfs4_handle_exception(NFS_SERVER(inode),
				_nfs4_do_set_security_label(inode, ilabel,
				fattr, olabel),
				&exception);
	} while (exception.retry);
	return err;
}

static int
nfs4_set_security_label(struct dentry *dentry, const void *buf, size_t buflen)
{
	struct nfs4_label ilabel, *olabel = NULL;
	struct nfs_fattr fattr;
	struct rpc_cred *cred;
	struct inode *inode = dentry->d_inode;
	int status;

	if (!nfs_server_capable(inode, NFS_CAP_SECURITY_LABEL))
		return -EOPNOTSUPP;

	nfs_fattr_init(&fattr);

	ilabel.pi = 0;
	ilabel.lfs = 0;
	ilabel.label = (char *)buf;
	ilabel.len = buflen;

	cred = rpc_lookup_cred();
	if (IS_ERR(cred))
		return PTR_ERR(cred);

	olabel = nfs4_label_alloc(NFS_SERVER(inode), GFP_KERNEL);
	if (IS_ERR(olabel)) {
		status = -PTR_ERR(olabel);
		goto out;
	}

	status = nfs4_do_set_security_label(inode, &ilabel, &fattr, olabel);
	if (status == 0)
		nfs_setsecurity(inode, &fattr, olabel);

	nfs4_label_free(olabel);
out:
	put_rpccred(cred);
	return status;
}
#endif	/* CONFIG_NFS_V4_SECURITY_LABEL */


L
Linus Torvalds 已提交
4559
static int
4560
nfs4_async_handle_error(struct rpc_task *task, const struct nfs_server *server, struct nfs4_state *state)
L
Linus Torvalds 已提交
4561
{
4562 4563 4564
	struct nfs_client *clp = server->nfs_client;

	if (task->tk_status >= 0)
L
Linus Torvalds 已提交
4565 4566
		return 0;
	switch(task->tk_status) {
4567
		case -NFS4ERR_DELEG_REVOKED:
4568 4569
		case -NFS4ERR_ADMIN_REVOKED:
		case -NFS4ERR_BAD_STATEID:
4570 4571 4572
			if (state == NULL)
				break;
			nfs_remove_bad_delegation(state->inode);
4573 4574 4575
		case -NFS4ERR_OPENMODE:
			if (state == NULL)
				break;
4576 4577
			if (nfs4_schedule_stateid_recovery(server, state) < 0)
				goto stateid_invalid;
4578
			goto wait_on_recovery;
4579
		case -NFS4ERR_EXPIRED:
4580 4581 4582 4583
			if (state != NULL) {
				if (nfs4_schedule_stateid_recovery(server, state) < 0)
					goto stateid_invalid;
			}
L
Linus Torvalds 已提交
4584
		case -NFS4ERR_STALE_STATEID:
4585
		case -NFS4ERR_STALE_CLIENTID:
4586 4587
			nfs4_schedule_lease_recovery(clp);
			goto wait_on_recovery;
4588 4589 4590 4591 4592 4593 4594 4595 4596 4597
#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);
4598
			nfs4_schedule_session_recovery(clp->cl_session, task->tk_status);
4599 4600 4601
			task->tk_status = 0;
			return -EAGAIN;
#endif /* CONFIG_NFS_V4_1 */
L
Linus Torvalds 已提交
4602
		case -NFS4ERR_DELAY:
4603
			nfs_inc_server_stats(server, NFSIOS_DELAY);
4604
		case -NFS4ERR_GRACE:
L
Linus Torvalds 已提交
4605 4606 4607
			rpc_delay(task, NFS4_POLL_RETRY_MAX);
			task->tk_status = 0;
			return -EAGAIN;
4608
		case -NFS4ERR_RETRY_UNCACHED_REP:
L
Linus Torvalds 已提交
4609 4610 4611 4612 4613 4614
		case -NFS4ERR_OLD_STATEID:
			task->tk_status = 0;
			return -EAGAIN;
	}
	task->tk_status = nfs4_map_errors(task->tk_status);
	return 0;
4615 4616 4617
stateid_invalid:
	task->tk_status = -EIO;
	return 0;
4618
wait_on_recovery:
4619 4620 4621 4622 4623
	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
4624 4625
}

4626 4627
static void nfs4_init_boot_verifier(const struct nfs_client *clp,
				    nfs4_verifier *bootverf)
4628 4629 4630
{
	__be32 verf[2];

4631 4632 4633 4634 4635 4636
	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 {
4637 4638 4639
		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;
4640
	}
4641 4642 4643
	memcpy(bootverf->data, verf, sizeof(bootverf->data));
}

4644 4645 4646 4647 4648 4649 4650 4651 4652 4653 4654 4655 4656 4657 4658 4659 4660 4661 4662 4663 4664
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)
{
4665 4666 4667 4668
	char *nodename = clp->cl_rpcclient->cl_nodename;

	if (nfs4_client_id_uniquifier[0] != '\0')
		nodename = nfs4_client_id_uniquifier;
4669 4670
	return scnprintf(buf, len, "Linux NFSv%u.%u %s",
				clp->rpc_ops->version, clp->cl_minorversion,
4671
				nodename);
4672 4673
}

4674 4675 4676 4677 4678 4679 4680 4681 4682 4683
/**
 * 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.
 */
4684 4685 4686
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
4687 4688 4689 4690 4691
{
	nfs4_verifier sc_verifier;
	struct nfs4_setclientid setclientid = {
		.sc_verifier = &sc_verifier,
		.sc_prog = program,
A
Andy Adamson 已提交
4692
		.sc_cb_ident = clp->cl_cb_ident,
L
Linus Torvalds 已提交
4693 4694 4695 4696
	};
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_SETCLIENTID],
		.rpc_argp = &setclientid,
4697
		.rpc_resp = res,
4698
		.rpc_cred = cred,
L
Linus Torvalds 已提交
4699
	};
4700
	int status;
L
Linus Torvalds 已提交
4701

4702
	/* nfs_client_id4 */
4703
	nfs4_init_boot_verifier(clp, &sc_verifier);
4704 4705 4706 4707 4708 4709 4710 4711 4712 4713
	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));
4714
	/* cb_client4 */
4715
	rcu_read_lock();
4716
	setclientid.sc_netid_len = scnprintf(setclientid.sc_netid,
4717
				sizeof(setclientid.sc_netid), "%s",
4718 4719
				rpc_peeraddr2str(clp->cl_rpcclient,
							RPC_DISPLAY_NETID));
4720 4721
	rcu_read_unlock();
	setclientid.sc_uaddr_len = scnprintf(setclientid.sc_uaddr,
4722
				sizeof(setclientid.sc_uaddr), "%s.%u.%u",
L
Linus Torvalds 已提交
4723 4724
				clp->cl_ipaddr, port >> 8, port & 255);

4725 4726 4727 4728 4729 4730
	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
4731 4732
}

4733 4734 4735 4736 4737 4738 4739 4740
/**
 * 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.
 */
4741
int nfs4_proc_setclientid_confirm(struct nfs_client *clp,
4742 4743
		struct nfs4_setclientid_res *arg,
		struct rpc_cred *cred)
L
Linus Torvalds 已提交
4744 4745 4746
{
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_SETCLIENTID_CONFIRM],
4747
		.rpc_argp = arg,
4748
		.rpc_cred = cred,
L
Linus Torvalds 已提交
4749 4750 4751
	};
	int status;

4752 4753 4754
	dprintk("NFS call  setclientid_confirm auth=%s, (client ID %llx)\n",
		clp->cl_rpcclient->cl_auth->au_ops->au_name,
		clp->cl_clientid);
4755
	status = rpc_call_sync(clp->cl_rpcclient, &msg, RPC_TASK_TIMEOUT);
4756
	dprintk("NFS reply setclientid_confirm: %d\n", status);
L
Linus Torvalds 已提交
4757 4758 4759
	return status;
}

4760 4761
struct nfs4_delegreturndata {
	struct nfs4_delegreturnargs args;
4762
	struct nfs4_delegreturnres res;
4763 4764
	struct nfs_fh fh;
	nfs4_stateid stateid;
4765
	unsigned long timestamp;
4766
	struct nfs_fattr fattr;
4767 4768 4769 4770 4771 4772
	int rpc_status;
};

static void nfs4_delegreturn_done(struct rpc_task *task, void *calldata)
{
	struct nfs4_delegreturndata *data = calldata;
4773

4774 4775
	if (!nfs4_sequence_done(task, &data->res.seq_res))
		return;
4776

4777 4778 4779 4780
	switch (task->tk_status) {
	case -NFS4ERR_STALE_STATEID:
	case -NFS4ERR_EXPIRED:
	case 0:
4781
		renew_lease(data->res.server, data->timestamp);
4782 4783 4784 4785
		break;
	default:
		if (nfs4_async_handle_error(task, data->res.server, NULL) ==
				-EAGAIN) {
4786
			rpc_restart_call_prepare(task);
4787 4788 4789 4790
			return;
		}
	}
	data->rpc_status = task->tk_status;
4791 4792 4793 4794 4795 4796 4797
}

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

4798 4799 4800 4801 4802 4803 4804
#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;

4805 4806 4807 4808
	nfs4_setup_sequence(d_data->res.server,
			&d_data->args.seq_args,
			&d_data->res.seq_res,
			task);
4809 4810 4811
}
#endif /* CONFIG_NFS_V4_1 */

4812
static const struct rpc_call_ops nfs4_delegreturn_ops = {
4813 4814 4815
#if defined(CONFIG_NFS_V4_1)
	.rpc_call_prepare = nfs4_delegreturn_prepare,
#endif /* CONFIG_NFS_V4_1 */
4816 4817 4818 4819
	.rpc_call_done = nfs4_delegreturn_done,
	.rpc_release = nfs4_delegreturn_release,
};

4820
static int _nfs4_proc_delegreturn(struct inode *inode, struct rpc_cred *cred, const nfs4_stateid *stateid, int issync)
4821 4822
{
	struct nfs4_delegreturndata *data;
4823
	struct nfs_server *server = NFS_SERVER(inode);
4824
	struct rpc_task *task;
4825 4826 4827 4828
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_DELEGRETURN],
		.rpc_cred = cred,
	};
T
Trond Myklebust 已提交
4829 4830
	struct rpc_task_setup task_setup_data = {
		.rpc_client = server->client,
4831
		.rpc_message = &msg,
T
Trond Myklebust 已提交
4832 4833 4834
		.callback_ops = &nfs4_delegreturn_ops,
		.flags = RPC_TASK_ASYNC,
	};
4835
	int status = 0;
4836

4837
	data = kzalloc(sizeof(*data), GFP_NOFS);
4838 4839
	if (data == NULL)
		return -ENOMEM;
4840
	nfs41_init_sequence(&data->args.seq_args, &data->res.seq_res, 1);
4841 4842
	data->args.fhandle = &data->fh;
	data->args.stateid = &data->stateid;
4843
	data->args.bitmask = server->cache_consistency_bitmask;
4844
	nfs_copy_fh(&data->fh, NFS_FH(inode));
4845
	nfs4_stateid_copy(&data->stateid, stateid);
4846 4847
	data->res.fattr = &data->fattr;
	data->res.server = server;
4848
	nfs_fattr_init(data->res.fattr);
4849
	data->timestamp = jiffies;
4850 4851
	data->rpc_status = 0;

T
Trond Myklebust 已提交
4852
	task_setup_data.callback_data = data;
4853 4854
	msg.rpc_argp = &data->args;
	msg.rpc_resp = &data->res;
T
Trond Myklebust 已提交
4855
	task = rpc_run_task(&task_setup_data);
4856
	if (IS_ERR(task))
4857
		return PTR_ERR(task);
4858 4859
	if (!issync)
		goto out;
4860
	status = nfs4_wait_for_completion_rpc_task(task);
4861 4862 4863
	if (status != 0)
		goto out;
	status = data->rpc_status;
4864 4865 4866 4867
	if (status == 0)
		nfs_post_op_update_inode_force_wcc(inode, &data->fattr);
	else
		nfs_refresh_inode(inode, &data->fattr);
4868
out:
4869
	rpc_put_task(task);
4870
	return status;
L
Linus Torvalds 已提交
4871 4872
}

4873
int nfs4_proc_delegreturn(struct inode *inode, struct rpc_cred *cred, const nfs4_stateid *stateid, int issync)
L
Linus Torvalds 已提交
4874 4875 4876 4877 4878
{
	struct nfs_server *server = NFS_SERVER(inode);
	struct nfs4_exception exception = { };
	int err;
	do {
4879
		err = _nfs4_proc_delegreturn(inode, cred, stateid, issync);
L
Linus Torvalds 已提交
4880 4881 4882 4883 4884 4885 4886 4887 4888 4889 4890 4891 4892 4893 4894 4895 4896 4897 4898 4899
		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)
{
4900
	freezable_schedule_timeout_killable_unsafe(timeout);
L
Linus Torvalds 已提交
4901 4902 4903 4904 4905 4906 4907 4908 4909 4910
	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);
4911
	struct nfs_client *clp = server->nfs_client;
T
Trond Myklebust 已提交
4912
	struct nfs_lockt_args arg = {
L
Linus Torvalds 已提交
4913
		.fh = NFS_FH(inode),
T
Trond Myklebust 已提交
4914
		.fl = request,
L
Linus Torvalds 已提交
4915
	};
T
Trond Myklebust 已提交
4916 4917
	struct nfs_lockt_res res = {
		.denied = request,
L
Linus Torvalds 已提交
4918 4919 4920 4921 4922 4923 4924 4925 4926 4927
	};
	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
4928
	arg.lock_owner.clientid = clp->cl_clientid;
4929 4930 4931 4932
	status = nfs4_set_lock_state(state, request);
	if (status != 0)
		goto out;
	lsp = request->fl_u.nfs4_fl.owner;
4933
	arg.lock_owner.id = lsp->ls_seqid.owner_id;
4934
	arg.lock_owner.s_dev = server->s_dev;
4935
	status = nfs4_call_sync(server->client, server, &msg, &arg.seq_args, &res.seq_res, 1);
T
Trond Myklebust 已提交
4936 4937 4938 4939 4940 4941
	switch (status) {
		case 0:
			request->fl_type = F_UNLCK;
			break;
		case -NFS4ERR_DENIED:
			status = 0;
L
Linus Torvalds 已提交
4942
	}
4943
	request->fl_ops->fl_release_private(request);
4944
out:
L
Linus Torvalds 已提交
4945 4946 4947 4948 4949 4950 4951 4952 4953 4954 4955 4956 4957 4958 4959 4960 4961 4962 4963 4964 4965 4966 4967 4968 4969 4970 4971 4972 4973 4974 4975 4976
	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;
}

4977
struct nfs4_unlockdata {
T
Trond Myklebust 已提交
4978 4979
	struct nfs_locku_args arg;
	struct nfs_locku_res res;
4980 4981
	struct nfs4_lock_state *lsp;
	struct nfs_open_context *ctx;
T
Trond Myklebust 已提交
4982 4983
	struct file_lock fl;
	const struct nfs_server *server;
4984
	unsigned long timestamp;
4985 4986
};

T
Trond Myklebust 已提交
4987 4988 4989 4990 4991 4992 4993 4994
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;

4995
	p = kzalloc(sizeof(*p), GFP_NOFS);
T
Trond Myklebust 已提交
4996 4997 4998 4999 5000
	if (p == NULL)
		return NULL;
	p->arg.fh = NFS_FH(inode);
	p->arg.fl = &p->fl;
	p->arg.seqid = seqid;
5001
	p->res.seqid = seqid;
T
Trond Myklebust 已提交
5002 5003 5004 5005 5006 5007 5008 5009 5010 5011
	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;
}

5012
static void nfs4_locku_release_calldata(void *data)
5013
{
5014
	struct nfs4_unlockdata *calldata = data;
T
Trond Myklebust 已提交
5015
	nfs_free_seqid(calldata->arg.seqid);
5016 5017 5018
	nfs4_put_lock_state(calldata->lsp);
	put_nfs_open_context(calldata->ctx);
	kfree(calldata);
5019 5020
}

5021
static void nfs4_locku_done(struct rpc_task *task, void *data)
5022
{
5023
	struct nfs4_unlockdata *calldata = data;
5024

5025 5026
	if (!nfs4_sequence_done(task, &calldata->res.seq_res))
		return;
5027 5028
	switch (task->tk_status) {
		case 0:
5029 5030
			nfs4_stateid_copy(&calldata->lsp->ls_stateid,
					&calldata->res.stateid);
5031
			renew_lease(calldata->server, calldata->timestamp);
5032
			break;
5033 5034
		case -NFS4ERR_BAD_STATEID:
		case -NFS4ERR_OLD_STATEID:
5035 5036 5037 5038
		case -NFS4ERR_STALE_STATEID:
		case -NFS4ERR_EXPIRED:
			break;
		default:
5039
			if (nfs4_async_handle_error(task, calldata->server, NULL) == -EAGAIN)
5040
				rpc_restart_call_prepare(task);
5041
	}
5042
	nfs_release_seqid(calldata->arg.seqid);
5043 5044
}

T
Trond Myklebust 已提交
5045
static void nfs4_locku_prepare(struct rpc_task *task, void *data)
5046
{
T
Trond Myklebust 已提交
5047
	struct nfs4_unlockdata *calldata = data;
5048

T
Trond Myklebust 已提交
5049
	if (nfs_wait_on_sequence(calldata->arg.seqid, task) != 0)
5050
		goto out_wait;
5051
	if (test_bit(NFS_LOCK_INITIALIZED, &calldata->lsp->ls_flags) == 0) {
5052
		/* Note: exit _without_ running nfs4_locku_done */
5053
		goto out_no_action;
5054
	}
5055
	calldata->timestamp = jiffies;
5056
	if (nfs4_setup_sequence(calldata->server,
5057
				&calldata->arg.seq_args,
5058 5059 5060
				&calldata->res.seq_res,
				task) != 0)
		nfs_release_seqid(calldata->arg.seqid);
5061 5062 5063 5064 5065
	return;
out_no_action:
	task->tk_action = NULL;
out_wait:
	nfs4_sequence_done(task, &calldata->res.seq_res);
5066 5067
}

5068
static const struct rpc_call_ops nfs4_locku_ops = {
T
Trond Myklebust 已提交
5069
	.rpc_call_prepare = nfs4_locku_prepare,
5070
	.rpc_call_done = nfs4_locku_done,
5071
	.rpc_release = nfs4_locku_release_calldata,
5072 5073
};

5074 5075 5076 5077 5078 5079
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;
5080 5081 5082 5083
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_LOCKU],
		.rpc_cred = ctx->cred,
	};
T
Trond Myklebust 已提交
5084 5085
	struct rpc_task_setup task_setup_data = {
		.rpc_client = NFS_CLIENT(lsp->ls_state->inode),
5086
		.rpc_message = &msg,
T
Trond Myklebust 已提交
5087
		.callback_ops = &nfs4_locku_ops,
5088
		.workqueue = nfsiod_workqueue,
T
Trond Myklebust 已提交
5089 5090
		.flags = RPC_TASK_ASYNC,
	};
5091

5092 5093 5094 5095 5096
	/* 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;

5097 5098 5099 5100 5101 5102
	data = nfs4_alloc_unlockdata(fl, ctx, lsp, seqid);
	if (data == NULL) {
		nfs_free_seqid(seqid);
		return ERR_PTR(-ENOMEM);
	}

5103
	nfs41_init_sequence(&data->arg.seq_args, &data->res.seq_res, 1);
5104 5105
	msg.rpc_argp = &data->arg;
	msg.rpc_resp = &data->res;
T
Trond Myklebust 已提交
5106 5107
	task_setup_data.callback_data = data;
	return rpc_run_task(&task_setup_data);
5108 5109
}

5110 5111
static int nfs4_proc_unlck(struct nfs4_state *state, int cmd, struct file_lock *request)
{
5112 5113 5114
	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
5115
	struct nfs_seqid *seqid;
L
Linus Torvalds 已提交
5116
	struct nfs4_lock_state *lsp;
5117 5118
	struct rpc_task *task;
	int status = 0;
5119
	unsigned char fl_flags = request->fl_flags;
5120

5121
	status = nfs4_set_lock_state(state, request);
5122 5123
	/* Unlock _before_ we do the RPC call */
	request->fl_flags |= FL_EXISTS;
5124 5125 5126
	/* Exclude nfs_delegation_claim_locks() */
	mutex_lock(&sp->so_delegreturn_mutex);
	/* Exclude nfs4_reclaim_open_stateid() - note nesting! */
5127 5128 5129
	down_read(&nfsi->rwsem);
	if (do_vfs_lock(request->fl_file, request) == -ENOENT) {
		up_read(&nfsi->rwsem);
5130
		mutex_unlock(&sp->so_delegreturn_mutex);
5131
		goto out;
5132 5133
	}
	up_read(&nfsi->rwsem);
5134
	mutex_unlock(&sp->so_delegreturn_mutex);
5135
	if (status != 0)
5136 5137
		goto out;
	/* Is this a delegated lock? */
5138
	lsp = request->fl_u.nfs4_fl.owner;
5139 5140
	if (test_bit(NFS_LOCK_INITIALIZED, &lsp->ls_flags) == 0)
		goto out;
5141
	seqid = nfs_alloc_seqid(&lsp->ls_seqid, GFP_KERNEL);
5142
	status = -ENOMEM;
T
Trond Myklebust 已提交
5143
	if (seqid == NULL)
5144
		goto out;
5145
	task = nfs4_do_unlck(request, nfs_file_open_context(request->fl_file), lsp, seqid);
5146 5147
	status = PTR_ERR(task);
	if (IS_ERR(task))
5148
		goto out;
5149
	status = nfs4_wait_for_completion_rpc_task(task);
5150
	rpc_put_task(task);
5151
out:
5152
	request->fl_flags = fl_flags;
L
Linus Torvalds 已提交
5153 5154 5155
	return status;
}

5156 5157 5158 5159 5160 5161
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;
5162
	unsigned long timestamp;
5163 5164
	int rpc_status;
	int cancelled;
5165
	struct nfs_server *server;
5166 5167 5168
};

static struct nfs4_lockdata *nfs4_alloc_lockdata(struct file_lock *fl,
5169 5170
		struct nfs_open_context *ctx, struct nfs4_lock_state *lsp,
		gfp_t gfp_mask)
L
Linus Torvalds 已提交
5171
{
5172 5173
	struct nfs4_lockdata *p;
	struct inode *inode = lsp->ls_state->inode;
L
Linus Torvalds 已提交
5174
	struct nfs_server *server = NFS_SERVER(inode);
5175

5176
	p = kzalloc(sizeof(*p), gfp_mask);
5177 5178 5179 5180 5181
	if (p == NULL)
		return NULL;

	p->arg.fh = NFS_FH(inode);
	p->arg.fl = &p->fl;
5182
	p->arg.open_seqid = nfs_alloc_seqid(&lsp->ls_state->owner->so_seqid, gfp_mask);
5183 5184
	if (p->arg.open_seqid == NULL)
		goto out_free;
5185
	p->arg.lock_seqid = nfs_alloc_seqid(&lsp->ls_seqid, gfp_mask);
5186
	if (p->arg.lock_seqid == NULL)
5187
		goto out_free_seqid;
5188
	p->arg.lock_stateid = &lsp->ls_stateid;
5189
	p->arg.lock_owner.clientid = server->nfs_client->cl_clientid;
5190
	p->arg.lock_owner.id = lsp->ls_seqid.owner_id;
5191
	p->arg.lock_owner.s_dev = server->s_dev;
5192
	p->res.lock_seqid = p->arg.lock_seqid;
5193
	p->lsp = lsp;
5194
	p->server = server;
5195 5196 5197 5198
	atomic_inc(&lsp->ls_count);
	p->ctx = get_nfs_open_context(ctx);
	memcpy(&p->fl, fl, sizeof(p->fl));
	return p;
5199 5200
out_free_seqid:
	nfs_free_seqid(p->arg.open_seqid);
5201 5202 5203 5204 5205 5206 5207 5208 5209
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;
5210

5211
	dprintk("%s: begin!\n", __func__);
5212
	if (nfs_wait_on_sequence(data->arg.lock_seqid, task) != 0)
5213
		goto out_wait;
5214 5215
	/* Do we need to do an open_to_lock_owner? */
	if (!(data->arg.lock_seqid->sequence->flags & NFS_SEQID_CONFIRMED)) {
5216
		if (nfs_wait_on_sequence(data->arg.open_seqid, task) != 0) {
5217
			goto out_release_lock_seqid;
5218
		}
5219
		data->arg.open_stateid = &state->open_stateid;
5220
		data->arg.new_lock_owner = 1;
5221
		data->res.open_seqid = data->arg.open_seqid;
5222 5223
	} else
		data->arg.new_lock_owner = 0;
5224 5225 5226 5227 5228
	if (!nfs4_valid_open_stateid(state)) {
		data->rpc_status = -EBADF;
		task->tk_action = NULL;
		goto out_release_open_seqid;
	}
5229
	data->timestamp = jiffies;
5230 5231
	if (nfs4_setup_sequence(data->server,
				&data->arg.seq_args,
5232
				&data->res.seq_res,
5233
				task) == 0)
5234
		return;
5235
out_release_open_seqid:
5236 5237 5238
	nfs_release_seqid(data->arg.open_seqid);
out_release_lock_seqid:
	nfs_release_seqid(data->arg.lock_seqid);
5239 5240
out_wait:
	nfs4_sequence_done(task, &data->res.seq_res);
5241
	dprintk("%s: done!, ret = %d\n", __func__, data->rpc_status);
5242 5243
}

5244 5245 5246 5247
static void nfs4_lock_done(struct rpc_task *task, void *calldata)
{
	struct nfs4_lockdata *data = calldata;

5248
	dprintk("%s: begin!\n", __func__);
5249

5250 5251
	if (!nfs4_sequence_done(task, &data->res.seq_res))
		return;
5252

5253 5254 5255 5256 5257 5258 5259 5260
	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) {
5261
		nfs4_stateid_copy(&data->lsp->ls_stateid, &data->res.stateid);
5262
		set_bit(NFS_LOCK_INITIALIZED, &data->lsp->ls_flags);
5263
		renew_lease(NFS_SERVER(data->ctx->dentry->d_inode), data->timestamp);
5264 5265
	}
out:
5266
	dprintk("%s: done, ret = %d!\n", __func__, data->rpc_status);
5267 5268 5269 5270 5271 5272
}

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

5273
	dprintk("%s: begin!\n", __func__);
5274
	nfs_free_seqid(data->arg.open_seqid);
5275 5276 5277 5278 5279
	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))
5280
			rpc_put_task_async(task);
5281
		dprintk("%s: cancelling lock!\n", __func__);
5282 5283 5284 5285 5286
	} else
		nfs_free_seqid(data->arg.lock_seqid);
	nfs4_put_lock_state(data->lsp);
	put_nfs_open_context(data->ctx);
	kfree(data);
5287
	dprintk("%s: done!\n", __func__);
5288 5289 5290 5291 5292 5293 5294 5295
}

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,
};

5296 5297 5298 5299 5300
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:
5301
		lsp->ls_seqid.flags &= ~NFS_SEQID_CONFIRMED;
5302
		if (new_lock_owner != 0 ||
5303
		   test_bit(NFS_LOCK_INITIALIZED, &lsp->ls_flags) != 0)
5304
			nfs4_schedule_stateid_recovery(server, lsp->ls_state);
5305 5306 5307
		break;
	case -NFS4ERR_STALE_STATEID:
		lsp->ls_seqid.flags &= ~NFS_SEQID_CONFIRMED;
5308 5309
	case -NFS4ERR_EXPIRED:
		nfs4_schedule_lease_recovery(server->nfs_client);
5310 5311 5312
	};
}

5313
static int _nfs4_do_setlk(struct nfs4_state *state, int cmd, struct file_lock *fl, int recovery_type)
5314 5315 5316
{
	struct nfs4_lockdata *data;
	struct rpc_task *task;
5317 5318 5319 5320
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_LOCK],
		.rpc_cred = state->owner->so_cred,
	};
T
Trond Myklebust 已提交
5321 5322
	struct rpc_task_setup task_setup_data = {
		.rpc_client = NFS_CLIENT(state->inode),
5323
		.rpc_message = &msg,
T
Trond Myklebust 已提交
5324
		.callback_ops = &nfs4_lock_ops,
5325
		.workqueue = nfsiod_workqueue,
T
Trond Myklebust 已提交
5326 5327
		.flags = RPC_TASK_ASYNC,
	};
5328 5329
	int ret;

5330
	dprintk("%s: begin!\n", __func__);
5331
	data = nfs4_alloc_lockdata(fl, nfs_file_open_context(fl->fl_file),
5332 5333
			fl->fl_u.nfs4_fl.owner,
			recovery_type == NFS_LOCK_NEW ? GFP_KERNEL : GFP_NOFS);
5334 5335 5336 5337
	if (data == NULL)
		return -ENOMEM;
	if (IS_SETLKW(cmd))
		data->arg.block = 1;
5338
	nfs41_init_sequence(&data->arg.seq_args, &data->res.seq_res, 1);
5339 5340
	msg.rpc_argp = &data->arg;
	msg.rpc_resp = &data->res;
T
Trond Myklebust 已提交
5341
	task_setup_data.callback_data = data;
5342 5343 5344 5345 5346
	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
5347
	task = rpc_run_task(&task_setup_data);
5348
	if (IS_ERR(task))
5349 5350 5351 5352
		return PTR_ERR(task);
	ret = nfs4_wait_for_completion_rpc_task(task);
	if (ret == 0) {
		ret = data->rpc_status;
5353 5354 5355
		if (ret)
			nfs4_handle_setlk_error(data->server, data->lsp,
					data->arg.new_lock_owner, ret);
5356 5357
	} else
		data->cancelled = 1;
5358
	rpc_put_task(task);
5359
	dprintk("%s: done, ret = %d!\n", __func__, ret);
5360
	return ret;
L
Linus Torvalds 已提交
5361 5362 5363 5364
}

static int nfs4_lock_reclaim(struct nfs4_state *state, struct file_lock *request)
{
5365
	struct nfs_server *server = NFS_SERVER(state->inode);
5366 5367 5368
	struct nfs4_exception exception = {
		.inode = state->inode,
	};
5369 5370 5371
	int err;

	do {
5372 5373 5374
		/* Cache the lock if possible... */
		if (test_bit(NFS_DELEGATED_STATE, &state->flags) != 0)
			return 0;
5375
		err = _nfs4_do_setlk(state, F_SETLK, request, NFS_LOCK_RECLAIM);
5376
		if (err != -NFS4ERR_DELAY)
5377 5378 5379 5380
			break;
		nfs4_handle_exception(server, err, &exception);
	} while (exception.retry);
	return err;
L
Linus Torvalds 已提交
5381 5382 5383 5384
}

static int nfs4_lock_expired(struct nfs4_state *state, struct file_lock *request)
{
5385
	struct nfs_server *server = NFS_SERVER(state->inode);
5386 5387 5388
	struct nfs4_exception exception = {
		.inode = state->inode,
	};
5389 5390
	int err;

5391 5392 5393
	err = nfs4_set_lock_state(state, request);
	if (err != 0)
		return err;
5394
	do {
5395 5396
		if (test_bit(NFS_DELEGATED_STATE, &state->flags) != 0)
			return 0;
5397
		err = _nfs4_do_setlk(state, F_SETLK, request, NFS_LOCK_EXPIRED);
5398 5399 5400 5401 5402 5403 5404 5405
		switch (err) {
		default:
			goto out;
		case -NFS4ERR_GRACE:
		case -NFS4ERR_DELAY:
			nfs4_handle_exception(server, err, &exception);
			err = 0;
		}
5406
	} while (exception.retry);
5407
out:
5408
	return err;
L
Linus Torvalds 已提交
5409 5410
}

5411
#if defined(CONFIG_NFS_V4_1)
5412 5413 5414 5415 5416 5417 5418 5419
/**
 * 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.
 */
5420
static int nfs41_check_expired_locks(struct nfs4_state *state)
5421
{
5422
	int status, ret = -NFS4ERR_BAD_STATEID;
5423
	struct nfs4_lock_state *lsp;
5424 5425
	struct nfs_server *server = NFS_SERVER(state->inode);

5426
	list_for_each_entry(lsp, &state->lock_states, ls_locks) {
5427
		if (test_bit(NFS_LOCK_INITIALIZED, &lsp->ls_flags)) {
5428 5429 5430 5431 5432
			struct rpc_cred *cred = lsp->ls_state->owner->so_cred;

			status = nfs41_test_stateid(server,
					&lsp->ls_stateid,
					cred);
5433
			if (status != NFS_OK) {
5434 5435
				/* Free the stateid unless the server
				 * informs us the stateid is unrecognized. */
5436 5437
				if (status != -NFS4ERR_BAD_STATEID)
					nfs41_free_stateid(server,
5438 5439
							&lsp->ls_stateid,
							cred);
5440
				clear_bit(NFS_LOCK_INITIALIZED, &lsp->ls_flags);
5441 5442 5443 5444 5445 5446 5447 5448 5449 5450 5451 5452 5453 5454
				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);
5455 5456 5457
	if (status != NFS_OK)
		status = nfs4_lock_expired(state, request);
	return status;
5458 5459 5460
}
#endif

L
Linus Torvalds 已提交
5461 5462
static int _nfs4_proc_setlk(struct nfs4_state *state, int cmd, struct file_lock *request)
{
5463
	struct nfs4_state_owner *sp = state->owner;
5464
	struct nfs_inode *nfsi = NFS_I(state->inode);
5465
	unsigned char fl_flags = request->fl_flags;
5466
	unsigned int seq;
5467
	int status = -ENOLCK;
L
Linus Torvalds 已提交
5468

5469 5470 5471
	if ((fl_flags & FL_POSIX) &&
			!test_bit(NFS_STATE_POSIX_LOCKS, &state->flags))
		goto out;
5472 5473 5474 5475
	/* Is this a delegated open? */
	status = nfs4_set_lock_state(state, request);
	if (status != 0)
		goto out;
5476 5477 5478 5479
	request->fl_flags |= FL_ACCESS;
	status = do_vfs_lock(request->fl_file, request);
	if (status < 0)
		goto out;
5480
	down_read(&nfsi->rwsem);
5481 5482 5483
	if (test_bit(NFS_DELEGATED_STATE, &state->flags)) {
		/* Yes: cache locks! */
		/* ...but avoid races with delegation recall... */
5484 5485 5486
		request->fl_flags = fl_flags & ~FL_SLEEP;
		status = do_vfs_lock(request->fl_file, request);
		goto out_unlock;
5487
	}
5488 5489
	seq = raw_seqcount_begin(&sp->so_reclaim_seqcount);
	up_read(&nfsi->rwsem);
5490
	status = _nfs4_do_setlk(state, cmd, request, NFS_LOCK_NEW);
5491
	if (status != 0)
5492 5493 5494 5495
		goto out;
	down_read(&nfsi->rwsem);
	if (read_seqcount_retry(&sp->so_reclaim_seqcount, seq)) {
		status = -NFS4ERR_DELAY;
5496
		goto out_unlock;
5497
	}
5498
	/* Note: we always want to sleep here! */
5499
	request->fl_flags = fl_flags | FL_SLEEP;
5500
	if (do_vfs_lock(request->fl_file, request) < 0)
5501 5502
		printk(KERN_WARNING "NFS: %s: VFS is out of sync with lock "
			"manager!\n", __func__);
5503
out_unlock:
5504
	up_read(&nfsi->rwsem);
5505 5506
out:
	request->fl_flags = fl_flags;
L
Linus Torvalds 已提交
5507 5508 5509 5510 5511
	return status;
}

static int nfs4_proc_setlk(struct nfs4_state *state, int cmd, struct file_lock *request)
{
5512 5513
	struct nfs4_exception exception = {
		.state = state,
5514
		.inode = state->inode,
5515
	};
L
Linus Torvalds 已提交
5516 5517 5518
	int err;

	do {
5519 5520 5521
		err = _nfs4_proc_setlk(state, cmd, request);
		if (err == -NFS4ERR_DENIED)
			err = -EAGAIN;
L
Linus Torvalds 已提交
5522
		err = nfs4_handle_exception(NFS_SERVER(state->inode),
5523
				err, &exception);
L
Linus Torvalds 已提交
5524 5525 5526 5527 5528 5529 5530 5531 5532 5533 5534 5535 5536
	} 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 */
5537
	ctx = nfs_file_open_context(filp);
L
Linus Torvalds 已提交
5538 5539 5540 5541 5542
	state = ctx->state;

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

5543 5544 5545 5546 5547
	if (IS_GETLK(cmd)) {
		if (state != NULL)
			return nfs4_proc_getlk(state, F_GETLK, request);
		return 0;
	}
L
Linus Torvalds 已提交
5548 5549 5550 5551

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

5552 5553 5554 5555 5556
	if (request->fl_type == F_UNLCK) {
		if (state != NULL)
			return nfs4_proc_unlck(state, cmd, request);
		return 0;
	}
L
Linus Torvalds 已提交
5557

5558 5559
	if (state == NULL)
		return -ENOLCK;
5560 5561 5562 5563
	/*
	 * Don't rely on the VFS having checked the file open mode,
	 * since it won't do this for flock() locks.
	 */
5564
	switch (request->fl_type) {
5565 5566 5567 5568 5569 5570 5571 5572 5573
	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
5574 5575 5576 5577 5578 5579 5580 5581 5582 5583 5584 5585
	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;
}

5586
int nfs4_lock_delegation_recall(struct file_lock *fl, struct nfs4_state *state, const nfs4_stateid *stateid)
5587 5588 5589 5590 5591 5592
{
	struct nfs_server *server = NFS_SERVER(state->inode);
	int err;

	err = nfs4_set_lock_state(state, fl);
	if (err != 0)
5593
		return err;
5594
	err = _nfs4_do_setlk(state, F_SETLK, fl, NFS_LOCK_NEW);
5595
	return nfs4_handle_delegation_recall_error(server, state, stateid, err);
5596
}
5597

5598 5599
struct nfs_release_lockowner_data {
	struct nfs4_lock_state *lsp;
5600
	struct nfs_server *server;
5601 5602 5603
	struct nfs_release_lockowner_args args;
};

5604 5605
static void nfs4_release_lockowner_release(void *calldata)
{
5606
	struct nfs_release_lockowner_data *data = calldata;
5607
	nfs4_free_lock_state(data->server, data->lsp);
5608 5609 5610
	kfree(calldata);
}

5611
static const struct rpc_call_ops nfs4_release_lockowner_ops = {
5612 5613 5614
	.rpc_release = nfs4_release_lockowner_release,
};

5615
static int nfs4_release_lockowner(struct nfs_server *server, struct nfs4_lock_state *lsp)
5616
{
5617
	struct nfs_release_lockowner_data *data;
5618 5619 5620 5621 5622
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_RELEASE_LOCKOWNER],
	};

	if (server->nfs_client->cl_mvops->minor_version != 0)
5623 5624 5625 5626 5627
		return -EINVAL;
	data = kmalloc(sizeof(*data), GFP_NOFS);
	if (!data)
		return -ENOMEM;
	data->lsp = lsp;
5628
	data->server = server;
5629 5630 5631 5632 5633 5634
	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;
5635 5636
}

5637 5638
#define XATTR_NAME_NFSV4_ACL "system.nfs4_acl"

5639 5640 5641
static int nfs4_xattr_set_nfs4_acl(struct dentry *dentry, const char *key,
				   const void *buf, size_t buflen,
				   int flags, int type)
5642
{
5643 5644
	if (strcmp(key, "") != 0)
		return -EINVAL;
5645

5646
	return nfs4_proc_set_acl(dentry->d_inode, buf, buflen);
5647 5648
}

5649 5650
static int nfs4_xattr_get_nfs4_acl(struct dentry *dentry, const char *key,
				   void *buf, size_t buflen, int type)
5651
{
5652 5653
	if (strcmp(key, "") != 0)
		return -EINVAL;
5654

5655
	return nfs4_proc_get_acl(dentry->d_inode, buf, buflen);
5656 5657
}

5658 5659 5660
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)
5661
{
5662
	size_t len = sizeof(XATTR_NAME_NFSV4_ACL);
5663

5664 5665
	if (!nfs4_server_supports_acls(NFS_SERVER(dentry->d_inode)))
		return 0;
5666 5667 5668

	if (list && len <= list_len)
		memcpy(list, XATTR_NAME_NFSV4_ACL, len);
5669
	return len;
5670 5671
}

5672 5673 5674 5675 5676 5677 5678 5679 5680 5681 5682 5683 5684 5685 5686 5687 5688 5689 5690 5691 5692 5693 5694 5695 5696 5697 5698 5699 5700 5701 5702 5703 5704 5705 5706 5707 5708 5709 5710 5711 5712 5713 5714 5715 5716 5717 5718
#ifdef CONFIG_NFS_V4_SECURITY_LABEL
static inline int nfs4_server_supports_labels(struct nfs_server *server)
{
	return server->caps & NFS_CAP_SECURITY_LABEL;
}

static int nfs4_xattr_set_nfs4_label(struct dentry *dentry, const char *key,
				   const void *buf, size_t buflen,
				   int flags, int type)
{
	if (security_ismaclabel(key))
		return nfs4_set_security_label(dentry, buf, buflen);

	return -EOPNOTSUPP;
}

static int nfs4_xattr_get_nfs4_label(struct dentry *dentry, const char *key,
				   void *buf, size_t buflen, int type)
{
	if (security_ismaclabel(key))
		return nfs4_get_security_label(dentry->d_inode, buf, buflen);
	return -EOPNOTSUPP;
}

static size_t nfs4_xattr_list_nfs4_label(struct dentry *dentry, char *list,
				       size_t list_len, const char *name,
				       size_t name_len, int type)
{
	size_t len = 0;

	if (nfs_server_capable(dentry->d_inode, NFS_CAP_SECURITY_LABEL)) {
		len = security_inode_listsecurity(dentry->d_inode, NULL, 0);
		if (list && len <= list_len)
			security_inode_listsecurity(dentry->d_inode, list, len);
	}
	return len;
}

static const struct xattr_handler nfs4_xattr_nfs4_label_handler = {
	.prefix = XATTR_SECURITY_PREFIX,
	.list	= nfs4_xattr_list_nfs4_label,
	.get	= nfs4_xattr_get_nfs4_label,
	.set	= nfs4_xattr_set_nfs4_label,
};
#endif


5719 5720 5721
/*
 * nfs_fhget will use either the mounted_on_fileid or the fileid
 */
5722 5723
static void nfs_fixup_referral_attributes(struct nfs_fattr *fattr)
{
5724 5725 5726
	if (!(((fattr->valid & NFS_ATTR_FATTR_MOUNTED_ON_FILEID) ||
	       (fattr->valid & NFS_ATTR_FATTR_FILEID)) &&
	      (fattr->valid & NFS_ATTR_FATTR_FSID) &&
5727
	      (fattr->valid & NFS_ATTR_FATTR_V4_LOCATIONS)))
5728 5729 5730
		return;

	fattr->valid |= NFS_ATTR_FATTR_TYPE | NFS_ATTR_FATTR_MODE |
5731
		NFS_ATTR_FATTR_NLINK | NFS_ATTR_FATTR_V4_REFERRAL;
5732 5733 5734 5735
	fattr->mode = S_IFDIR | S_IRUGO | S_IXUGO;
	fattr->nlink = 2;
}

5736 5737 5738 5739
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)
5740 5741
{
	struct nfs_server *server = NFS_SERVER(dir);
5742
	u32 bitmask[3] = {
5743
		[0] = FATTR4_WORD0_FSID | FATTR4_WORD0_FS_LOCATIONS,
5744 5745 5746
	};
	struct nfs4_fs_locations_arg args = {
		.dir_fh = NFS_FH(dir),
5747
		.name = name,
5748 5749 5750
		.page = page,
		.bitmask = bitmask,
	};
B
Benny Halevy 已提交
5751 5752 5753
	struct nfs4_fs_locations_res res = {
		.fs_locations = fs_locations,
	};
5754 5755 5756
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_FS_LOCATIONS],
		.rpc_argp = &args,
B
Benny Halevy 已提交
5757
		.rpc_resp = &res,
5758 5759 5760
	};
	int status;

5761
	dprintk("%s: start\n", __func__);
5762 5763 5764 5765 5766 5767 5768 5769

	/* 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;

5770
	nfs_fattr_init(&fs_locations->fattr);
5771
	fs_locations->server = server;
5772
	fs_locations->nlocations = 0;
5773
	status = nfs4_call_sync(client, server, &msg, &args.seq_args, &res.seq_res, 0);
5774
	dprintk("%s: returned status = %d\n", __func__, status);
5775 5776 5777
	return status;
}

5778 5779 5780 5781
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)
5782 5783 5784 5785 5786
{
	struct nfs4_exception exception = { };
	int err;
	do {
		err = nfs4_handle_exception(NFS_SERVER(dir),
5787
				_nfs4_proc_fs_locations(client, dir, name, fs_locations, page),
5788 5789 5790 5791 5792
				&exception);
	} while (exception.retry);
	return err;
}

B
Bryan Schumaker 已提交
5793 5794 5795 5796 5797 5798 5799 5800 5801 5802 5803 5804 5805 5806 5807 5808 5809 5810 5811 5812 5813 5814
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;
}

5815 5816
int nfs4_proc_secinfo(struct inode *dir, const struct qstr *name,
		      struct nfs4_secinfo_flavors *flavors)
B
Bryan Schumaker 已提交
5817 5818 5819 5820 5821 5822 5823 5824 5825 5826 5827
{
	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;
}

5828
#ifdef CONFIG_NFS_V4_1
5829 5830 5831 5832 5833 5834 5835 5836 5837 5838 5839 5840 5841 5842 5843 5844 5845 5846 5847
/*
 * 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;
}

5848
static bool
5849 5850
nfs41_same_server_scope(struct nfs41_server_scope *a,
			struct nfs41_server_scope *b)
5851 5852 5853 5854 5855 5856 5857 5858
{
	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;
}

5859 5860 5861 5862 5863 5864
/*
 * nfs4_proc_bind_conn_to_session()
 *
 * The 4.1 client currently uses the same TCP connection for the
 * fore and backchannel.
 */
5865
int nfs4_proc_bind_conn_to_session(struct nfs_client *clp, struct rpc_cred *cred)
5866 5867 5868 5869 5870 5871 5872 5873
{
	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,
5874
		.rpc_cred = cred,
5875 5876 5877 5878 5879 5880 5881 5882 5883 5884 5885 5886 5887 5888 5889 5890 5891 5892 5893 5894 5895 5896 5897 5898 5899 5900 5901 5902 5903 5904 5905 5906 5907 5908 5909 5910 5911 5912
	};

	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
5913 5914 5915
/*
 * nfs4_proc_exchange_id()
 *
5916 5917
 * Returns zero, a negative errno, or a negative NFS4ERR status code.
 *
B
Benny Halevy 已提交
5918 5919 5920 5921 5922
 * 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.
 */
5923
int nfs4_proc_exchange_id(struct nfs_client *clp, struct rpc_cred *cred)
B
Benny Halevy 已提交
5924 5925 5926
{
	nfs4_verifier verifier;
	struct nfs41_exchange_id_args args = {
5927
		.verifier = &verifier,
B
Benny Halevy 已提交
5928
		.client = clp,
5929 5930
		.flags = EXCHGID4_FLAG_SUPP_MOVED_REFER |
			EXCHGID4_FLAG_BIND_PRINC_STATEID,
B
Benny Halevy 已提交
5931 5932
	};
	struct nfs41_exchange_id_res res = {
5933
		0
B
Benny Halevy 已提交
5934 5935 5936 5937 5938 5939 5940 5941 5942
	};
	int status;
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_EXCHANGE_ID],
		.rpc_argp = &args,
		.rpc_resp = &res,
		.rpc_cred = cred,
	};

5943
	nfs4_init_boot_verifier(clp, &verifier);
5944 5945
	args.id_len = nfs4_init_uniform_client_string(clp, args.id,
							sizeof(args.id));
5946 5947 5948
	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
5949

5950
	res.server_owner = kzalloc(sizeof(struct nfs41_server_owner),
5951
					GFP_NOFS);
5952
	if (unlikely(res.server_owner == NULL)) {
5953 5954 5955
		status = -ENOMEM;
		goto out;
	}
5956

5957
	res.server_scope = kzalloc(sizeof(struct nfs41_server_scope),
5958
					GFP_NOFS);
5959
	if (unlikely(res.server_scope == NULL)) {
5960
		status = -ENOMEM;
5961
		goto out_server_owner;
5962
	}
5963

5964
	res.impl_id = kzalloc(sizeof(struct nfs41_impl_id), GFP_NOFS);
5965
	if (unlikely(res.impl_id == NULL)) {
5966 5967 5968 5969
		status = -ENOMEM;
		goto out_server_scope;
	}

5970
	status = rpc_call_sync(clp->cl_rpcclient, &msg, RPC_TASK_TIMEOUT);
5971
	if (status == 0)
5972
		status = nfs4_check_cl_exchange_flags(res.flags);
5973

5974
	if (status == 0) {
5975 5976 5977 5978 5979
		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;

5980 5981 5982
		kfree(clp->cl_serverowner);
		clp->cl_serverowner = res.server_owner;
		res.server_owner = NULL;
5983

5984
		/* use the most recent implementation id */
5985 5986
		kfree(clp->cl_implid);
		clp->cl_implid = res.impl_id;
5987

5988
		if (clp->cl_serverscope != NULL &&
5989
		    !nfs41_same_server_scope(clp->cl_serverscope,
5990 5991 5992 5993
					     res.server_scope)) {
			dprintk("%s: server_scope mismatch detected\n",
				__func__);
			set_bit(NFS4CLNT_SERVER_SCOPE_MISMATCH, &clp->cl_state);
5994 5995
			kfree(clp->cl_serverscope);
			clp->cl_serverscope = NULL;
5996 5997
		}

5998
		if (clp->cl_serverscope == NULL) {
5999
			clp->cl_serverscope = res.server_scope;
6000 6001
			goto out;
		}
6002 6003
	} else
		kfree(res.impl_id);
6004

6005 6006
out_server_owner:
	kfree(res.server_owner);
6007
out_server_scope:
6008 6009
	kfree(res.server_scope);
out:
6010
	if (clp->cl_implid != NULL)
6011
		dprintk("NFS reply exchange_id: Server Implementation ID: "
6012
			"domain: %s, name: %s, date: %llu,%u\n",
6013
			clp->cl_implid->domain, clp->cl_implid->name,
6014 6015
			clp->cl_implid->date.seconds,
			clp->cl_implid->date.nseconds);
6016
	dprintk("NFS reply exchange_id: %d\n", status);
B
Benny Halevy 已提交
6017 6018 6019
	return status;
}

T
Trond Myklebust 已提交
6020 6021 6022 6023 6024 6025 6026 6027 6028 6029 6030 6031
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)
6032
		dprintk("NFS: Got error %d from the server %s on "
T
Trond Myklebust 已提交
6033 6034 6035 6036 6037 6038 6039 6040 6041 6042 6043 6044 6045 6046 6047 6048 6049 6050 6051 6052 6053 6054 6055 6056 6057 6058 6059 6060 6061 6062 6063 6064 6065
			"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;
6066 6067
	if (clp->cl_preserve_clid)
		goto out;
T
Trond Myklebust 已提交
6068 6069 6070 6071 6072 6073 6074 6075 6076 6077 6078 6079 6080
	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
6081 6082 6083 6084 6085 6086 6087 6088 6089 6090 6091 6092 6093 6094 6095
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 */
6096 6097 6098 6099
	nfs41_setup_sequence(data->clp->cl_session,
			&data->args->la_seq_args,
			&data->res->lr_seq_res,
			task);
A
Andy Adamson 已提交
6100 6101 6102 6103 6104 6105 6106 6107 6108 6109 6110 6111 6112
	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__);
6113 6114
	if (!nfs41_sequence_done(task, &data->res->lr_seq_res))
		return;
A
Andy Adamson 已提交
6115 6116 6117 6118 6119 6120
	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;
6121 6122
		/* fall through */
	case -NFS4ERR_RETRY_UNCACHED_REP:
6123
		rpc_restart_call_prepare(task);
A
Andy Adamson 已提交
6124 6125 6126 6127 6128
		return;
	}
	dprintk("<-- %s\n", __func__);
}

6129
static const struct rpc_call_ops nfs4_get_lease_time_ops = {
A
Andy Adamson 已提交
6130 6131 6132 6133 6134 6135 6136 6137 6138 6139 6140 6141 6142 6143 6144 6145 6146 6147 6148 6149 6150 6151 6152 6153 6154
	.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,
6155 6156
		.callback_data = &data,
		.flags = RPC_TASK_TIMEOUT,
A
Andy Adamson 已提交
6157 6158 6159
	};
	int status;

6160
	nfs41_init_sequence(&args.la_seq_args, &res.lr_seq_res, 0);
6161
	nfs4_set_sequence_privileged(&args.la_seq_args);
A
Andy Adamson 已提交
6162 6163 6164 6165 6166 6167 6168 6169 6170 6171 6172 6173 6174 6175
	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
6176 6177 6178 6179 6180 6181 6182 6183 6184 6185 6186
/*
 * 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)
{
6187 6188 6189 6190
	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
6191 6192

	/* Fore channel attributes */
6193 6194
	args->fc_attrs.max_rqst_sz = max_rqst_sz;
	args->fc_attrs.max_resp_sz = max_resp_sz;
A
Andy Adamson 已提交
6195
	args->fc_attrs.max_ops = NFS4_MAX_OPS;
6196
	args->fc_attrs.max_reqs = max_session_slots;
A
Andy Adamson 已提交
6197 6198

	dprintk("%s: Fore Channel : max_rqst_sz=%u max_resp_sz=%u "
6199
		"max_ops=%u max_reqs=%u\n",
A
Andy Adamson 已提交
6200 6201
		__func__,
		args->fc_attrs.max_rqst_sz, args->fc_attrs.max_resp_sz,
6202
		args->fc_attrs.max_ops, args->fc_attrs.max_reqs);
A
Andy Adamson 已提交
6203 6204 6205 6206 6207 6208 6209 6210 6211 6212 6213 6214 6215 6216 6217 6218

	/* 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);
}

6219
static int nfs4_verify_fore_channel_attrs(struct nfs41_create_session_args *args, struct nfs4_session *session)
6220
{
6221 6222 6223 6224 6225 6226 6227 6228 6229 6230 6231 6232 6233 6234 6235
	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;
6236 6237
	if (rcvd->max_reqs > NFS4_MAX_SLOT_TABLE)
		rcvd->max_reqs = NFS4_MAX_SLOT_TABLE;
6238
	return 0;
6239 6240
}

6241 6242 6243 6244
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;
6245

6246 6247 6248 6249 6250 6251 6252
	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: */
6253
	if (rcvd->max_ops != sent->max_ops)
6254
		return -EINVAL;
6255
	if (rcvd->max_reqs != sent->max_reqs)
6256 6257 6258
		return -EINVAL;
	return 0;
}
6259 6260 6261 6262

static int nfs4_verify_channel_attrs(struct nfs41_create_session_args *args,
				     struct nfs4_session *session)
{
6263
	int ret;
6264

6265 6266 6267 6268
	ret = nfs4_verify_fore_channel_attrs(args, session);
	if (ret)
		return ret;
	return nfs4_verify_back_channel_attrs(args, session);
6269 6270
}

6271 6272
static int _nfs4_proc_create_session(struct nfs_client *clp,
		struct rpc_cred *cred)
A
Andy Adamson 已提交
6273 6274 6275 6276 6277 6278 6279 6280 6281 6282 6283 6284 6285
{
	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,
6286
		.rpc_cred = cred,
A
Andy Adamson 已提交
6287 6288 6289 6290
	};
	int status;

	nfs4_init_channel_attrs(&args);
6291
	args.flags = (SESSION4_PERSIST | SESSION4_BACK_CHAN);
A
Andy Adamson 已提交
6292

6293
	status = rpc_call_sync(session->clp->cl_rpcclient, &msg, RPC_TASK_TIMEOUT);
A
Andy Adamson 已提交
6294

6295
	if (!status) {
6296 6297
		/* Verify the session's negotiated channel_attrs values */
		status = nfs4_verify_channel_attrs(&args, session);
A
Andy Adamson 已提交
6298 6299 6300 6301 6302 6303 6304 6305 6306 6307 6308 6309
		/* 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.
 */
6310
int nfs4_proc_create_session(struct nfs_client *clp, struct rpc_cred *cred)
A
Andy Adamson 已提交
6311 6312 6313 6314 6315 6316 6317
{
	int status;
	unsigned *ptr;
	struct nfs4_session *session = clp->cl_session;

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

6318
	status = _nfs4_proc_create_session(clp, cred);
A
Andy Adamson 已提交
6319 6320 6321
	if (status)
		goto out;

6322 6323 6324
	/* 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
6325 6326 6327 6328 6329 6330 6331 6332 6333 6334 6335
	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
6336 6337 6338 6339
/*
 * Issue the over-the-wire RPC DESTROY_SESSION.
 * The caller must serialize access to this routine.
 */
6340 6341
int nfs4_proc_destroy_session(struct nfs4_session *session,
		struct rpc_cred *cred)
A
Andy Adamson 已提交
6342
{
6343 6344 6345 6346 6347
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_DESTROY_SESSION],
		.rpc_argp = session,
		.rpc_cred = cred,
	};
A
Andy Adamson 已提交
6348 6349 6350 6351 6352 6353 6354 6355
	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;

6356
	status = rpc_call_sync(session->clp->cl_rpcclient, &msg, RPC_TASK_TIMEOUT);
A
Andy Adamson 已提交
6357 6358

	if (status)
6359
		dprintk("NFS: Got error %d from the server on DESTROY_SESSION. "
A
Andy Adamson 已提交
6360 6361 6362 6363 6364 6365
			"Session has been destroyed regardless...\n", status);

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

A
Andy Adamson 已提交
6366 6367 6368
/*
 * Renew the cl_session lease.
 */
6369 6370 6371 6372 6373 6374
struct nfs4_sequence_data {
	struct nfs_client *clp;
	struct nfs4_sequence_args args;
	struct nfs4_sequence_res res;
};

6375 6376
static void nfs41_sequence_release(void *data)
{
6377 6378
	struct nfs4_sequence_data *calldata = data;
	struct nfs_client *clp = calldata->clp;
6379

6380 6381 6382
	if (atomic_read(&clp->cl_count) > 1)
		nfs4_schedule_state_renewal(clp);
	nfs_put_client(clp);
6383
	kfree(calldata);
6384 6385
}

6386 6387 6388 6389 6390 6391 6392
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:
6393
		nfs4_schedule_lease_recovery(clp);
6394 6395 6396 6397
	}
	return 0;
}

6398
static void nfs41_sequence_call_done(struct rpc_task *task, void *data)
A
Andy Adamson 已提交
6399
{
6400 6401
	struct nfs4_sequence_data *calldata = data;
	struct nfs_client *clp = calldata->clp;
A
Andy Adamson 已提交
6402

6403 6404
	if (!nfs41_sequence_done(task, task->tk_msg.rpc_resp))
		return;
A
Andy Adamson 已提交
6405 6406 6407

	if (task->tk_status < 0) {
		dprintk("%s ERROR %d\n", __func__, task->tk_status);
6408 6409
		if (atomic_read(&clp->cl_count) == 1)
			goto out;
A
Andy Adamson 已提交
6410

6411 6412
		if (nfs41_sequence_handle_errors(task, clp) == -EAGAIN) {
			rpc_restart_call_prepare(task);
A
Andy Adamson 已提交
6413 6414 6415 6416
			return;
		}
	}
	dprintk("%s rpc_cred %p\n", __func__, task->tk_msg.rpc_cred);
6417
out:
A
Andy Adamson 已提交
6418 6419 6420 6421 6422
	dprintk("<-- %s\n", __func__);
}

static void nfs41_sequence_prepare(struct rpc_task *task, void *data)
{
6423 6424
	struct nfs4_sequence_data *calldata = data;
	struct nfs_client *clp = calldata->clp;
A
Andy Adamson 已提交
6425 6426 6427 6428 6429 6430
	struct nfs4_sequence_args *args;
	struct nfs4_sequence_res *res;

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

6431
	nfs41_setup_sequence(clp->cl_session, args, res, task);
A
Andy Adamson 已提交
6432 6433 6434 6435 6436
}

static const struct rpc_call_ops nfs41_sequence_ops = {
	.rpc_call_done = nfs41_sequence_call_done,
	.rpc_call_prepare = nfs41_sequence_prepare,
6437
	.rpc_release = nfs41_sequence_release,
A
Andy Adamson 已提交
6438 6439
};

6440 6441 6442
static struct rpc_task *_nfs41_proc_sequence(struct nfs_client *clp,
		struct rpc_cred *cred,
		bool is_privileged)
A
Andy Adamson 已提交
6443
{
6444
	struct nfs4_sequence_data *calldata;
A
Andy Adamson 已提交
6445 6446 6447 6448
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_SEQUENCE],
		.rpc_cred = cred,
	};
6449 6450 6451
	struct rpc_task_setup task_setup_data = {
		.rpc_client = clp->cl_rpcclient,
		.rpc_message = &msg,
6452
		.callback_ops = &nfs41_sequence_ops,
6453
		.flags = RPC_TASK_ASYNC | RPC_TASK_TIMEOUT,
6454
	};
A
Andy Adamson 已提交
6455

6456
	if (!atomic_inc_not_zero(&clp->cl_count))
6457
		return ERR_PTR(-EIO);
6458
	calldata = kzalloc(sizeof(*calldata), GFP_NOFS);
6459
	if (calldata == NULL) {
6460
		nfs_put_client(clp);
6461
		return ERR_PTR(-ENOMEM);
A
Andy Adamson 已提交
6462
	}
6463
	nfs41_init_sequence(&calldata->args, &calldata->res, 0);
6464 6465
	if (is_privileged)
		nfs4_set_sequence_privileged(&calldata->args);
6466 6467 6468
	msg.rpc_argp = &calldata->args;
	msg.rpc_resp = &calldata->res;
	calldata->clp = clp;
6469
	task_setup_data.callback_data = calldata;
A
Andy Adamson 已提交
6470

6471 6472 6473
	return rpc_run_task(&task_setup_data);
}

6474
static int nfs41_proc_async_sequence(struct nfs_client *clp, struct rpc_cred *cred, unsigned renew_flags)
6475 6476 6477 6478
{
	struct rpc_task *task;
	int ret = 0;

6479 6480
	if ((renew_flags & NFS4_RENEW_TIMEOUT) == 0)
		return 0;
6481
	task = _nfs41_proc_sequence(clp, cred, false);
6482 6483 6484
	if (IS_ERR(task))
		ret = PTR_ERR(task);
	else
6485
		rpc_put_task_async(task);
6486 6487 6488 6489 6490 6491 6492 6493 6494
	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;

6495
	task = _nfs41_proc_sequence(clp, cred, true);
6496 6497 6498 6499 6500
	if (IS_ERR(task)) {
		ret = PTR_ERR(task);
		goto out;
	}
	ret = rpc_wait_for_completion_task(task);
6501 6502 6503 6504 6505
	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);
6506
		ret = task->tk_status;
6507
	}
6508 6509 6510 6511
	rpc_put_task(task);
out:
	dprintk("<-- %s status=%d\n", __func__, ret);
	return ret;
A
Andy Adamson 已提交
6512 6513
}

6514 6515 6516 6517 6518 6519 6520 6521 6522 6523
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;

6524 6525 6526 6527
	nfs41_setup_sequence(calldata->clp->cl_session,
			&calldata->arg.seq_args,
			&calldata->res.seq_res,
			task);
6528 6529
}

6530 6531 6532 6533 6534 6535 6536 6537 6538
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);
6539 6540
		/* fall through */
	case -NFS4ERR_RETRY_UNCACHED_REP:
6541 6542
		return -EAGAIN;
	default:
6543
		nfs4_schedule_lease_recovery(clp);
6544 6545 6546 6547
	}
	return 0;
}

6548 6549 6550 6551 6552 6553 6554
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__);
6555 6556
	if (!nfs41_sequence_done(task, res))
		return;
6557

6558 6559 6560 6561
	if (nfs41_reclaim_complete_handle_errors(task, clp) == -EAGAIN) {
		rpc_restart_call_prepare(task);
		return;
	}
6562 6563 6564 6565 6566 6567 6568 6569 6570 6571 6572 6573 6574 6575 6576 6577 6578 6579 6580
	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.
 */
6581 6582
static int nfs41_proc_reclaim_complete(struct nfs_client *clp,
		struct rpc_cred *cred)
6583 6584 6585 6586 6587
{
	struct nfs4_reclaim_complete_data *calldata;
	struct rpc_task *task;
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_RECLAIM_COMPLETE],
6588
		.rpc_cred = cred,
6589 6590 6591 6592 6593 6594 6595 6596 6597 6598
	};
	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__);
6599
	calldata = kzalloc(sizeof(*calldata), GFP_NOFS);
6600 6601 6602 6603 6604
	if (calldata == NULL)
		goto out;
	calldata->clp = clp;
	calldata->arg.one_fs = 0;

6605
	nfs41_init_sequence(&calldata->arg.seq_args, &calldata->res.seq_res, 0);
6606
	nfs4_set_sequence_privileged(&calldata->arg.seq_args);
6607 6608 6609 6610
	msg.rpc_argp = &calldata->arg;
	msg.rpc_resp = &calldata->res;
	task_setup_data.callback_data = calldata;
	task = rpc_run_task(&task_setup_data);
6611
	if (IS_ERR(task)) {
6612
		status = PTR_ERR(task);
6613 6614
		goto out;
	}
6615 6616 6617
	status = nfs4_wait_for_completion_rpc_task(task);
	if (status == 0)
		status = task->tk_status;
6618
	rpc_put_task(task);
6619
	return 0;
6620 6621 6622 6623
out:
	dprintk("<-- %s status=%d\n", __func__, status);
	return status;
}
6624 6625 6626 6627 6628

static void
nfs4_layoutget_prepare(struct rpc_task *task, void *calldata)
{
	struct nfs4_layoutget *lgp = calldata;
F
Fred Isaman 已提交
6629
	struct nfs_server *server = NFS_SERVER(lgp->args.inode);
6630
	struct nfs4_session *session = nfs4_get_session(server);
6631 6632

	dprintk("--> %s\n", __func__);
F
Fred Isaman 已提交
6633 6634 6635 6636 6637
	/* 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.
	 */
6638
	if (nfs41_setup_sequence(session, &lgp->args.seq_args,
6639
				&lgp->res.seq_res, task))
6640
		return;
6641 6642 6643 6644 6645
	if (pnfs_choose_layoutget_stateid(&lgp->args.stateid,
					  NFS_I(lgp->args.inode)->layout,
					  lgp->args.ctx->state)) {
		rpc_exit(task, NFS4_OK);
	}
6646 6647 6648 6649 6650
}

static void nfs4_layoutget_done(struct rpc_task *task, void *calldata)
{
	struct nfs4_layoutget *lgp = calldata;
6651 6652 6653 6654
	struct inode *inode = lgp->args.inode;
	struct nfs_server *server = NFS_SERVER(inode);
	struct pnfs_layout_hdr *lo;
	struct nfs4_state *state = NULL;
6655
	unsigned long timeo, giveup;
6656 6657 6658

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

6659
	if (!nfs41_sequence_done(task, &lgp->res.seq_res))
6660
		goto out;
6661 6662 6663

	switch (task->tk_status) {
	case 0:
6664
		goto out;
6665 6666
	case -NFS4ERR_LAYOUTTRYLATER:
	case -NFS4ERR_RECALLCONFLICT:
6667 6668 6669 6670
		timeo = rpc_get_timeout(task->tk_client);
		giveup = lgp->args.timestamp + timeo;
		if (time_after(giveup, jiffies))
			task->tk_status = -NFS4ERR_DELAY;
6671 6672 6673 6674 6675 6676 6677 6678 6679 6680 6681 6682 6683 6684 6685 6686 6687
		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);
6688 6689
		}
	}
6690 6691 6692
	if (nfs4_async_handle_error(task, server, state) == -EAGAIN)
		rpc_restart_call_prepare(task);
out:
6693 6694 6695
	dprintk("<-- %s\n", __func__);
}

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 6722 6723 6724 6725 6726 6727 6728 6729 6730 6731 6732 6733 6734 6735 6736 6737 6738 6739
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;
}

6740 6741 6742
static void nfs4_layoutget_release(void *calldata)
{
	struct nfs4_layoutget *lgp = calldata;
6743 6744
	struct inode *inode = lgp->args.inode;
	struct nfs_server *server = NFS_SERVER(inode);
6745
	size_t max_pages = max_response_pages(server);
6746 6747

	dprintk("--> %s\n", __func__);
6748
	nfs4_free_pages(lgp->args.layout.pages, max_pages);
6749
	pnfs_put_layout_hdr(NFS_I(inode)->layout);
6750 6751 6752 6753 6754 6755 6756 6757 6758 6759 6760
	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,
};

6761 6762
struct pnfs_layout_segment *
nfs4_proc_layoutget(struct nfs4_layoutget *lgp, gfp_t gfp_flags)
6763
{
6764 6765
	struct inode *inode = lgp->args.inode;
	struct nfs_server *server = NFS_SERVER(inode);
6766
	size_t max_pages = max_response_pages(server);
6767 6768 6769 6770 6771
	struct rpc_task *task;
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_LAYOUTGET],
		.rpc_argp = &lgp->args,
		.rpc_resp = &lgp->res,
6772
		.rpc_cred = lgp->cred,
6773 6774 6775 6776 6777 6778 6779 6780
	};
	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,
	};
6781
	struct pnfs_layout_segment *lseg = NULL;
6782 6783 6784 6785
	int status = 0;

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

6786 6787 6788
	lgp->args.layout.pages = nfs4_alloc_pages(max_pages, gfp_flags);
	if (!lgp->args.layout.pages) {
		nfs4_layoutget_release(lgp);
6789
		return ERR_PTR(-ENOMEM);
6790 6791
	}
	lgp->args.layout.pglen = max_pages * PAGE_SIZE;
6792
	lgp->args.timestamp = jiffies;
6793

6794
	lgp->res.layoutp = &lgp->args.layout;
6795
	lgp->res.seq_res.sr_slot = NULL;
6796
	nfs41_init_sequence(&lgp->args.seq_args, &lgp->res.seq_res, 0);
6797 6798 6799 6800

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

6801 6802
	task = rpc_run_task(&task_setup_data);
	if (IS_ERR(task))
6803
		return ERR_CAST(task);
6804
	status = nfs4_wait_for_completion_rpc_task(task);
F
Fred Isaman 已提交
6805 6806
	if (status == 0)
		status = task->tk_status;
6807 6808
	/* if layoutp->len is 0, nfs4_layoutget_prepare called rpc_exit */
	if (status == 0 && lgp->res.layoutp->len)
6809
		lseg = pnfs_layout_process(lgp);
6810 6811
	rpc_put_task(task);
	dprintk("<-- %s status=%d\n", __func__, status);
6812 6813 6814
	if (status)
		return ERR_PTR(status);
	return lseg;
6815 6816
}

B
Benny Halevy 已提交
6817 6818 6819 6820 6821 6822
static void
nfs4_layoutreturn_prepare(struct rpc_task *task, void *calldata)
{
	struct nfs4_layoutreturn *lrp = calldata;

	dprintk("--> %s\n", __func__);
6823 6824 6825 6826
	nfs41_setup_sequence(lrp->clp->cl_session,
			&lrp->args.seq_args,
			&lrp->res.seq_res,
			task);
B
Benny Halevy 已提交
6827 6828 6829 6830 6831 6832 6833 6834 6835
}

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

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

6836
	if (!nfs41_sequence_done(task, &lrp->res.seq_res))
B
Benny Halevy 已提交
6837 6838 6839 6840
		return;

	server = NFS_SERVER(lrp->args.inode);
	if (nfs4_async_handle_error(task, server, NULL) == -EAGAIN) {
6841
		rpc_restart_call_prepare(task);
B
Benny Halevy 已提交
6842 6843 6844 6845 6846 6847 6848 6849
		return;
	}
	dprintk("<-- %s\n", __func__);
}

static void nfs4_layoutreturn_release(void *calldata)
{
	struct nfs4_layoutreturn *lrp = calldata;
6850
	struct pnfs_layout_hdr *lo = lrp->args.layout;
B
Benny Halevy 已提交
6851 6852

	dprintk("--> %s\n", __func__);
6853 6854 6855 6856 6857
	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);
6858
	pnfs_put_layout_hdr(lrp->args.layout);
B
Benny Halevy 已提交
6859 6860 6861 6862 6863 6864 6865 6866 6867 6868 6869 6870 6871 6872 6873 6874 6875
	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,
6876
		.rpc_cred = lrp->cred,
B
Benny Halevy 已提交
6877 6878 6879 6880 6881 6882 6883 6884 6885 6886
	};
	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__);
6887
	nfs41_init_sequence(&lrp->args.seq_args, &lrp->res.seq_res, 1);
B
Benny Halevy 已提交
6888 6889 6890 6891 6892 6893 6894 6895 6896
	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
6897 6898 6899 6900 6901 6902 6903 6904 6905 6906 6907 6908 6909 6910 6911 6912 6913 6914 6915 6916 6917 6918 6919 6920 6921 6922 6923 6924 6925 6926 6927 6928 6929 6930 6931 6932 6933 6934 6935 6936 6937 6938 6939 6940 6941 6942 6943 6944
/*
 * 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);

6945
static int
6946 6947 6948
_nfs4_proc_getdeviceinfo(struct nfs_server *server,
		struct pnfs_device *pdev,
		struct rpc_cred *cred)
6949 6950 6951 6952 6953 6954 6955 6956 6957 6958 6959
{
	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,
6960
		.rpc_cred = cred,
6961 6962 6963 6964
	};
	int status;

	dprintk("--> %s\n", __func__);
6965
	status = nfs4_call_sync(server->client, server, &msg, &args.seq_args, &res.seq_res, 0);
6966 6967 6968 6969 6970
	dprintk("<-- %s status=%d\n", __func__, status);

	return status;
}

6971 6972 6973
int nfs4_proc_getdeviceinfo(struct nfs_server *server,
		struct pnfs_device *pdev,
		struct rpc_cred *cred)
6974 6975 6976 6977 6978 6979
{
	struct nfs4_exception exception = { };
	int err;

	do {
		err = nfs4_handle_exception(server,
6980
					_nfs4_proc_getdeviceinfo(server, pdev, cred),
6981 6982 6983 6984 6985 6986
					&exception);
	} while (exception.retry);
	return err;
}
EXPORT_SYMBOL_GPL(nfs4_proc_getdeviceinfo);

A
Andy Adamson 已提交
6987 6988 6989 6990
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);
6991
	struct nfs4_session *session = nfs4_get_session(server);
A
Andy Adamson 已提交
6992

6993 6994 6995 6996
	nfs41_setup_sequence(session,
			&data->args.seq_args,
			&data->res.seq_res,
			task);
A
Andy Adamson 已提交
6997 6998 6999 7000 7001 7002 7003 7004
}

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);

7005
	if (!nfs41_sequence_done(task, &data->res.seq_res))
A
Andy Adamson 已提交
7006 7007 7008
		return;

	switch (task->tk_status) { /* Just ignore these failures */
7009 7010 7011 7012
	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
7013
		task->tk_status = 0;
7014 7015
		break;
	case 0:
A
Andy Adamson 已提交
7016 7017
		nfs_post_op_update_inode_force_wcc(data->args.inode,
						   data->res.fattr);
7018 7019 7020 7021 7022 7023 7024
		break;
	default:
		if (nfs4_async_handle_error(task, server, NULL) == -EAGAIN) {
			rpc_restart_call_prepare(task);
			return;
		}
	}
A
Andy Adamson 已提交
7025 7026 7027 7028 7029 7030
}

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

A
Andy Adamson 已提交
7031
	pnfs_cleanup_layoutcommit(data);
A
Andy Adamson 已提交
7032 7033 7034 7035 7036 7037 7038 7039 7040 7041 7042
	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
7043
nfs4_proc_layoutcommit(struct nfs4_layoutcommit_data *data, bool sync)
A
Andy Adamson 已提交
7044 7045 7046 7047 7048 7049 7050 7051 7052 7053 7054 7055 7056 7057 7058 7059 7060 7061 7062 7063 7064 7065 7066 7067
{
	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);

7068
	nfs41_init_sequence(&data->args.seq_args, &data->res.seq_res, 1);
A
Andy Adamson 已提交
7069 7070 7071
	task = rpc_run_task(&task_setup_data);
	if (IS_ERR(task))
		return PTR_ERR(task);
7072
	if (sync == false)
A
Andy Adamson 已提交
7073 7074 7075 7076 7077 7078 7079 7080 7081 7082
		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;
}
7083 7084 7085 7086 7087 7088 7089 7090 7091 7092 7093 7094 7095 7096 7097 7098 7099 7100 7101 7102 7103 7104 7105 7106 7107 7108 7109 7110 7111 7112 7113

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:
7114
			goto out;
7115 7116 7117 7118
		default:
			err = nfs4_handle_exception(server, err, &exception);
		}
	} while (exception.retry);
7119
out:
7120 7121 7122 7123 7124 7125 7126 7127 7128 7129 7130 7131 7132 7133 7134 7135 7136 7137 7138 7139 7140 7141 7142 7143 7144 7145 7146 7147 7148 7149 7150 7151 7152 7153 7154 7155 7156 7157 7158 7159 7160 7161 7162
	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;
}
7163

7164 7165 7166
static int _nfs41_test_stateid(struct nfs_server *server,
		nfs4_stateid *stateid,
		struct rpc_cred *cred)
B
Bryan Schumaker 已提交
7167 7168 7169
{
	int status;
	struct nfs41_test_stateid_args args = {
7170
		.stateid = stateid,
B
Bryan Schumaker 已提交
7171 7172 7173 7174 7175 7176
	};
	struct nfs41_test_stateid_res res;
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_TEST_STATEID],
		.rpc_argp = &args,
		.rpc_resp = &res,
7177
		.rpc_cred = cred,
B
Bryan Schumaker 已提交
7178
	};
7179

7180
	dprintk("NFS call  test_stateid %p\n", stateid);
7181
	nfs41_init_sequence(&args.seq_args, &res.seq_res, 0);
7182 7183 7184
	nfs4_set_sequence_privileged(&args.seq_args);
	status = nfs4_call_sync_sequence(server->client, server, &msg,
			&args.seq_args, &res.seq_res);
7185 7186
	if (status != NFS_OK) {
		dprintk("NFS reply test_stateid: failed, %d\n", status);
7187
		return status;
7188 7189
	}
	dprintk("NFS reply test_stateid: succeeded, %d\n", -res.status);
7190
	return -res.status;
B
Bryan Schumaker 已提交
7191 7192
}

7193 7194 7195 7196 7197
/**
 * nfs41_test_stateid - perform a TEST_STATEID operation
 *
 * @server: server / transport on which to perform the operation
 * @stateid: state ID to test
7198
 * @cred: credential
7199 7200 7201 7202 7203
 *
 * 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.
 */
7204 7205 7206
static int nfs41_test_stateid(struct nfs_server *server,
		nfs4_stateid *stateid,
		struct rpc_cred *cred)
B
Bryan Schumaker 已提交
7207 7208 7209 7210
{
	struct nfs4_exception exception = { };
	int err;
	do {
7211
		err = _nfs41_test_stateid(server, stateid, cred);
7212 7213 7214
		if (err != -NFS4ERR_DELAY)
			break;
		nfs4_handle_exception(server, err, &exception);
B
Bryan Schumaker 已提交
7215 7216 7217
	} while (exception.retry);
	return err;
}
B
Bryan Schumaker 已提交
7218

7219 7220 7221
struct nfs_free_stateid_data {
	struct nfs_server *server;
	struct nfs41_free_stateid_args args;
B
Bryan Schumaker 已提交
7222
	struct nfs41_free_stateid_res res;
7223 7224 7225 7226 7227 7228 7229 7230 7231 7232 7233 7234 7235 7236 7237 7238 7239 7240 7241 7242 7243 7244 7245 7246 7247 7248 7249 7250 7251 7252 7253 7254 7255 7256 7257 7258 7259
};

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,
7260
		struct rpc_cred *cred,
7261 7262
		bool privileged)
{
B
Bryan Schumaker 已提交
7263 7264
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_FREE_STATEID],
7265
		.rpc_cred = cred,
B
Bryan Schumaker 已提交
7266
	};
7267 7268 7269 7270 7271 7272 7273
	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
7274

7275
	dprintk("NFS call  free_stateid %p\n", stateid);
7276 7277 7278 7279 7280 7281 7282 7283 7284 7285 7286 7287 7288 7289 7290
	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
7291 7292
}

7293 7294 7295 7296 7297
/**
 * nfs41_free_stateid - perform a FREE_STATEID operation
 *
 * @server: server / transport on which to perform the operation
 * @stateid: state ID to release
7298
 * @cred: credential
7299 7300 7301 7302
 *
 * Returns NFS_OK if the server freed "stateid".  Otherwise a
 * negative NFS4ERR value is returned.
 */
7303 7304 7305
static int nfs41_free_stateid(struct nfs_server *server,
		nfs4_stateid *stateid,
		struct rpc_cred *cred)
B
Bryan Schumaker 已提交
7306
{
7307 7308 7309
	struct rpc_task *task;
	int ret;

7310
	task = _nfs41_free_stateid(server, stateid, cred, true);
7311 7312 7313 7314 7315 7316 7317
	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
7318
}
7319

7320 7321 7322
static int nfs41_free_lock_state(struct nfs_server *server, struct nfs4_lock_state *lsp)
{
	struct rpc_task *task;
7323
	struct rpc_cred *cred = lsp->ls_state->owner->so_cred;
7324

7325
	task = _nfs41_free_stateid(server, &lsp->ls_stateid, cred, false);
7326 7327 7328 7329 7330 7331 7332
	nfs4_free_lock_state(server, lsp);
	if (IS_ERR(task))
		return PTR_ERR(task);
	rpc_put_task(task);
	return 0;
}

7333 7334 7335
static bool nfs41_match_stateid(const nfs4_stateid *s1,
		const nfs4_stateid *s2)
{
7336
	if (memcmp(s1->other, s2->other, sizeof(s1->other)) != 0)
7337 7338
		return false;

7339
	if (s1->seqid == s2->seqid)
7340
		return true;
7341
	if (s1->seqid == 0 || s2->seqid == 0)
7342 7343 7344 7345 7346
		return true;

	return false;
}

7347 7348
#endif /* CONFIG_NFS_V4_1 */

7349 7350 7351
static bool nfs4_match_stateid(const nfs4_stateid *s1,
		const nfs4_stateid *s2)
{
7352
	return nfs4_stateid_match(s1, s2);
7353 7354 7355
}


7356
static const struct nfs4_state_recovery_ops nfs40_reboot_recovery_ops = {
7357
	.owner_flag_bit = NFS_OWNER_RECLAIM_REBOOT,
7358
	.state_flag_bit	= NFS_STATE_RECLAIM_REBOOT,
L
Linus Torvalds 已提交
7359 7360
	.recover_open	= nfs4_open_reclaim,
	.recover_lock	= nfs4_lock_reclaim,
7361
	.establish_clid = nfs4_init_clientid,
7362
	.get_clid_cred	= nfs4_get_setclientid_cred,
7363
	.detect_trunking = nfs40_discover_server_trunking,
L
Linus Torvalds 已提交
7364 7365
};

7366
#if defined(CONFIG_NFS_V4_1)
7367
static const struct nfs4_state_recovery_ops nfs41_reboot_recovery_ops = {
7368 7369 7370 7371
	.owner_flag_bit = NFS_OWNER_RECLAIM_REBOOT,
	.state_flag_bit	= NFS_STATE_RECLAIM_REBOOT,
	.recover_open	= nfs4_open_reclaim,
	.recover_lock	= nfs4_lock_reclaim,
7372
	.establish_clid = nfs41_init_clientid,
7373
	.get_clid_cred	= nfs4_get_exchange_id_cred,
7374
	.reclaim_complete = nfs41_proc_reclaim_complete,
7375
	.detect_trunking = nfs41_discover_server_trunking,
7376 7377 7378
};
#endif /* CONFIG_NFS_V4_1 */

7379
static const struct nfs4_state_recovery_ops nfs40_nograce_recovery_ops = {
7380 7381 7382 7383 7384
	.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,
7385
	.get_clid_cred	= nfs4_get_setclientid_cred,
7386 7387 7388
};

#if defined(CONFIG_NFS_V4_1)
7389
static const struct nfs4_state_recovery_ops nfs41_nograce_recovery_ops = {
7390
	.owner_flag_bit = NFS_OWNER_RECLAIM_NOGRACE,
7391
	.state_flag_bit	= NFS_STATE_RECLAIM_NOGRACE,
7392 7393
	.recover_open	= nfs41_open_expired,
	.recover_lock	= nfs41_lock_expired,
7394
	.establish_clid = nfs41_init_clientid,
7395
	.get_clid_cred	= nfs4_get_exchange_id_cred,
L
Linus Torvalds 已提交
7396
};
7397
#endif /* CONFIG_NFS_V4_1 */
L
Linus Torvalds 已提交
7398

7399
static const struct nfs4_state_maintenance_ops nfs40_state_renewal_ops = {
B
Benny Halevy 已提交
7400
	.sched_state_renewal = nfs4_proc_async_renew,
7401
	.get_state_renewal_cred_locked = nfs4_get_renew_cred_locked,
7402
	.renew_lease = nfs4_proc_renew,
B
Benny Halevy 已提交
7403 7404 7405
};

#if defined(CONFIG_NFS_V4_1)
7406
static const struct nfs4_state_maintenance_ops nfs41_state_renewal_ops = {
B
Benny Halevy 已提交
7407
	.sched_state_renewal = nfs41_proc_async_sequence,
7408
	.get_state_renewal_cred_locked = nfs4_get_machine_cred_locked,
7409
	.renew_lease = nfs4_proc_sequence,
B
Benny Halevy 已提交
7410 7411 7412
};
#endif

7413 7414
static const struct nfs4_minor_version_ops nfs_v4_0_minor_ops = {
	.minor_version = 0,
7415 7416 7417 7418
	.init_caps = NFS_CAP_READDIRPLUS
		| NFS_CAP_ATOMIC_OPEN
		| NFS_CAP_CHANGE_ATTR
		| NFS_CAP_POSIX_LOCK,
7419
	.call_sync = _nfs4_call_sync,
7420
	.match_stateid = nfs4_match_stateid,
7421
	.find_root_sec = nfs4_find_root_sec,
7422
	.free_lock_state = nfs4_release_lockowner,
7423 7424 7425
	.reboot_recovery_ops = &nfs40_reboot_recovery_ops,
	.nograce_recovery_ops = &nfs40_nograce_recovery_ops,
	.state_renewal_ops = &nfs40_state_renewal_ops,
7426 7427 7428 7429 7430
};

#if defined(CONFIG_NFS_V4_1)
static const struct nfs4_minor_version_ops nfs_v4_1_minor_ops = {
	.minor_version = 1,
7431 7432 7433
	.init_caps = NFS_CAP_READDIRPLUS
		| NFS_CAP_ATOMIC_OPEN
		| NFS_CAP_CHANGE_ATTR
7434
		| NFS_CAP_POSIX_LOCK
7435 7436
		| NFS_CAP_STATEID_NFSV41
		| NFS_CAP_ATOMIC_OPEN_V1,
7437
	.call_sync = nfs4_call_sync_sequence,
7438
	.match_stateid = nfs41_match_stateid,
7439
	.find_root_sec = nfs41_find_root_sec,
7440
	.free_lock_state = nfs41_free_lock_state,
7441 7442 7443
	.reboot_recovery_ops = &nfs41_reboot_recovery_ops,
	.nograce_recovery_ops = &nfs41_nograce_recovery_ops,
	.state_renewal_ops = &nfs41_state_renewal_ops,
7444 7445 7446
};
#endif

7447 7448 7449
#if defined(CONFIG_NFS_V4_2)
static const struct nfs4_minor_version_ops nfs_v4_2_minor_ops = {
	.minor_version = 2,
7450 7451 7452 7453 7454 7455
	.init_caps = NFS_CAP_READDIRPLUS
		| NFS_CAP_ATOMIC_OPEN
		| NFS_CAP_CHANGE_ATTR
		| NFS_CAP_POSIX_LOCK
		| NFS_CAP_STATEID_NFSV41
		| NFS_CAP_ATOMIC_OPEN_V1,
7456 7457 7458
	.call_sync = nfs4_call_sync_sequence,
	.match_stateid = nfs41_match_stateid,
	.find_root_sec = nfs41_find_root_sec,
7459
	.free_lock_state = nfs41_free_lock_state,
7460 7461 7462 7463 7464 7465
	.reboot_recovery_ops = &nfs41_reboot_recovery_ops,
	.nograce_recovery_ops = &nfs41_nograce_recovery_ops,
	.state_renewal_ops = &nfs41_state_renewal_ops,
};
#endif

7466 7467 7468 7469 7470
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
7471 7472 7473
#if defined(CONFIG_NFS_V4_2)
	[2] = &nfs_v4_2_minor_ops,
#endif
7474 7475
};

7476 7477 7478 7479 7480 7481 7482 7483 7484 7485 7486 7487 7488 7489 7490 7491 7492 7493 7494 7495
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,
};

7496
static const struct inode_operations nfs4_file_inode_operations = {
7497 7498 7499
	.permission	= nfs_permission,
	.getattr	= nfs_getattr,
	.setattr	= nfs_setattr,
7500 7501 7502 7503
	.getxattr	= generic_getxattr,
	.setxattr	= generic_setxattr,
	.listxattr	= generic_listxattr,
	.removexattr	= generic_removexattr,
7504 7505
};

D
David Howells 已提交
7506
const struct nfs_rpc_ops nfs_v4_clientops = {
L
Linus Torvalds 已提交
7507 7508 7509
	.version	= 4,			/* protocol version */
	.dentry_ops	= &nfs4_dentry_operations,
	.dir_inode_ops	= &nfs4_dir_inode_operations,
7510
	.file_inode_ops	= &nfs4_file_inode_operations,
7511
	.file_ops	= &nfs4_file_operations,
L
Linus Torvalds 已提交
7512
	.getroot	= nfs4_proc_get_root,
B
Bryan Schumaker 已提交
7513
	.submount	= nfs4_submount,
B
Bryan Schumaker 已提交
7514
	.try_mount	= nfs4_try_mount,
L
Linus Torvalds 已提交
7515 7516 7517 7518 7519 7520 7521 7522
	.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,
7523
	.unlink_rpc_prepare = nfs4_proc_unlink_rpc_prepare,
L
Linus Torvalds 已提交
7524 7525
	.unlink_done	= nfs4_proc_unlink_done,
	.rename		= nfs4_proc_rename,
7526
	.rename_setup	= nfs4_proc_rename_setup,
7527
	.rename_rpc_prepare = nfs4_proc_rename_rpc_prepare,
7528
	.rename_done	= nfs4_proc_rename_done,
L
Linus Torvalds 已提交
7529 7530 7531 7532 7533 7534 7535 7536 7537
	.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,
7538
	.set_capabilities = nfs4_server_capabilities,
L
Linus Torvalds 已提交
7539 7540
	.decode_dirent	= nfs4_decode_dirent,
	.read_setup	= nfs4_proc_read_setup,
7541
	.read_pageio_init = pnfs_pageio_init_read,
7542
	.read_rpc_prepare = nfs4_proc_read_rpc_prepare,
T
Trond Myklebust 已提交
7543
	.read_done	= nfs4_read_done,
L
Linus Torvalds 已提交
7544
	.write_setup	= nfs4_proc_write_setup,
7545
	.write_pageio_init = pnfs_pageio_init_write,
7546
	.write_rpc_prepare = nfs4_proc_write_rpc_prepare,
7547
	.write_done	= nfs4_write_done,
L
Linus Torvalds 已提交
7548
	.commit_setup	= nfs4_proc_commit_setup,
7549
	.commit_rpc_prepare = nfs4_proc_commit_rpc_prepare,
7550
	.commit_done	= nfs4_commit_done,
L
Linus Torvalds 已提交
7551
	.lock		= nfs4_proc_lock,
7552
	.clear_acl_cache = nfs4_zap_acl_attr,
T
Trond Myklebust 已提交
7553
	.close_context  = nfs4_close_context,
7554
	.open_context	= nfs4_atomic_open,
7555
	.have_delegation = nfs4_have_delegation,
7556
	.return_delegation = nfs4_inode_return_delegation,
7557
	.alloc_client	= nfs4_alloc_client,
7558
	.init_client	= nfs4_init_client,
7559
	.free_client	= nfs4_free_client,
7560 7561
	.create_server	= nfs4_create_server,
	.clone_server	= nfs_clone_server,
L
Linus Torvalds 已提交
7562 7563
};

7564 7565 7566 7567 7568 7569 7570 7571 7572
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,
7573 7574 7575
#ifdef CONFIG_NFS_V4_SECURITY_LABEL
	&nfs4_xattr_nfs4_label_handler,
#endif
7576 7577 7578
	NULL
};

L
Linus Torvalds 已提交
7579 7580 7581 7582 7583
/*
 * Local variables:
 *  c-basic-offset: 8
 * End:
 */