nfs4proc.c 226.4 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
#include "nfs4trace.h"

L
Linus Torvalds 已提交
71 72
#define NFSDBG_FACILITY		NFSDBG_PROC

73
#define NFS4_POLL_RETRY_MIN	(HZ/10)
L
Linus Torvalds 已提交
74 75
#define NFS4_POLL_RETRY_MAX	(15*HZ)

76
struct nfs4_opendata;
77
static int _nfs4_proc_open(struct nfs4_opendata *data);
78
static int _nfs4_recover_proc_open(struct nfs4_opendata *data);
L
Linus Torvalds 已提交
79
static int nfs4_do_fsinfo(struct nfs_server *, struct nfs_fh *, struct nfs_fsinfo *);
80
static int nfs4_async_handle_error(struct rpc_task *, const struct nfs_server *, struct nfs4_state *);
81
static void nfs_fixup_referral_attributes(struct nfs_fattr *fattr);
82 83
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);
84 85
static int nfs4_do_setattr(struct inode *inode, struct rpc_cred *cred,
			    struct nfs_fattr *fattr, struct iattr *sattr,
86 87
			    struct nfs4_state *state, struct nfs4_label *ilabel,
			    struct nfs4_label *olabel);
88
#ifdef CONFIG_NFS_V4_1
89 90 91 92
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 *);
93
#endif
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

#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;

	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
141
/* Prevent leaks of NFSv4 errors into userland */
142
static int nfs4_map_errors(int err)
L
Linus Torvalds 已提交
143
{
144 145 146 147
	if (err >= -1000)
		return err;
	switch (err) {
	case -NFS4ERR_RESOURCE:
148 149
	case -NFS4ERR_LAYOUTTRYLATER:
	case -NFS4ERR_RECALLCONFLICT:
150
		return -EREMOTEIO;
151
	case -NFS4ERR_WRONGSEC:
152
	case -NFS4ERR_WRONG_CRED:
153
		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;
384 385 386 387 388
		case -NFS4ERR_MOVED:
			ret = nfs4_schedule_migration_recovery(server);
			if (ret < 0)
				break;
			goto wait_on_recovery;
389 390 391
		case -NFS4ERR_LEASE_MOVED:
			nfs4_schedule_lease_moved_recovery(clp);
			goto wait_on_recovery;
392
#if defined(CONFIG_NFS_V4_1)
393 394 395 396 397 398 399 400 401
		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);
402
			nfs4_schedule_session_recovery(clp->cl_session, errorcode);
403
			goto wait_on_recovery;
404
#endif /* defined(CONFIG_NFS_V4_1) */
405
		case -NFS4ERR_FILE_OPEN:
406 407 408 409 410 411 412
			if (exception->timeout > HZ) {
				/* We have retried a decent amount, time to
				 * fail
				 */
				ret = -EBUSY;
				break;
			}
413 414 415 416 417
		case -NFS4ERR_GRACE:
		case -NFS4ERR_DELAY:
			ret = nfs4_delay(server->client, &exception->timeout);
			if (ret != 0)
				break;
418
		case -NFS4ERR_RETRY_UNCACHED_REP:
419 420
		case -NFS4ERR_OLD_STATEID:
			exception->retry = 1;
421 422 423 424 425 426 427 428 429 430 431 432 433
			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);
			}
434 435 436
	}
	/* We failed to handle the error */
	return nfs4_map_errors(ret);
437
wait_on_recovery:
438
	ret = nfs4_wait_clnt_recover(clp);
439 440
	if (test_bit(NFS_MIG_FAILED, &server->mig_status))
		return -EIO;
441 442 443
	if (ret == 0)
		exception->retry = 1;
	return ret;
444 445
}

446 447 448 449 450 451 452 453 454 455 456 457 458 459
/*
 * Return 'true' if 'clp' is using an rpc_client that is integrity protected
 * or 'false' otherwise.
 */
static bool _nfs4_is_integrity_protected(struct nfs_client *clp)
{
	rpc_authflavor_t flavor = clp->cl_rpcclient->cl_auth->au_flavor;

	if (flavor == RPC_AUTH_GSS_KRB5I ||
	    flavor == RPC_AUTH_GSS_KRB5P)
		return true;

	return false;
}
460

461
static void do_renew_lease(struct nfs_client *clp, unsigned long timestamp)
L
Linus Torvalds 已提交
462 463 464 465 466 467 468
{
	spin_lock(&clp->cl_lock);
	if (time_before(clp->cl_last_renewal,timestamp))
		clp->cl_last_renewal = timestamp;
	spin_unlock(&clp->cl_lock);
}

469 470 471 472 473
static void renew_lease(const struct nfs_server *server, unsigned long timestamp)
{
	do_renew_lease(server->nfs_client, timestamp);
}

474 475 476 477 478 479
struct nfs4_call_sync_data {
	const struct nfs_server *seq_server;
	struct nfs4_sequence_args *seq_args;
	struct nfs4_sequence_res *seq_res;
};

480 481 482 483 484 485 486 487 488 489 490 491 492 493 494
static void nfs4_init_sequence(struct nfs4_sequence_args *args,
			       struct nfs4_sequence_res *res, int cache_reply)
{
	args->sa_slot = NULL;
	args->sa_cache_this = cache_reply;
	args->sa_privileged = 0;

	res->sr_slot = NULL;
}

static void nfs4_set_sequence_privileged(struct nfs4_sequence_args *args)
{
	args->sa_privileged = 1;
}

C
Chuck Lever 已提交
495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555
static int nfs40_setup_sequence(const struct nfs_server *server,
				struct nfs4_sequence_args *args,
				struct nfs4_sequence_res *res,
				struct rpc_task *task)
{
	struct nfs4_slot_table *tbl = server->nfs_client->cl_slot_tbl;
	struct nfs4_slot *slot;

	/* slot already allocated? */
	if (res->sr_slot != NULL)
		goto out_start;

	spin_lock(&tbl->slot_tbl_lock);
	if (nfs4_slot_tbl_draining(tbl) && !args->sa_privileged)
		goto out_sleep;

	slot = nfs4_alloc_slot(tbl);
	if (IS_ERR(slot)) {
		if (slot == ERR_PTR(-ENOMEM))
			task->tk_timeout = HZ >> 2;
		goto out_sleep;
	}
	spin_unlock(&tbl->slot_tbl_lock);

	args->sa_slot = slot;
	res->sr_slot = slot;

out_start:
	rpc_call_start(task);
	return 0;

out_sleep:
	if (args->sa_privileged)
		rpc_sleep_on_priority(&tbl->slot_tbl_waitq, task,
				NULL, RPC_PRIORITY_PRIVILEGED);
	else
		rpc_sleep_on(&tbl->slot_tbl_waitq, task, NULL);
	spin_unlock(&tbl->slot_tbl_lock);
	return -EAGAIN;
}

static int nfs40_sequence_done(struct rpc_task *task,
			       struct nfs4_sequence_res *res)
{
	struct nfs4_slot *slot = res->sr_slot;
	struct nfs4_slot_table *tbl;

	if (!RPC_WAS_SENT(task))
		goto out;

	tbl = slot->table;
	spin_lock(&tbl->slot_tbl_lock);
	if (!nfs41_wake_and_assign_slot(tbl, slot))
		nfs4_free_slot(tbl, slot);
	spin_unlock(&tbl->slot_tbl_lock);

	res->sr_slot = NULL;
out:
	return 1;
}

A
Andy Adamson 已提交
556 557
#if defined(CONFIG_NFS_V4_1)

558
static void nfs41_sequence_free_slot(struct nfs4_sequence_res *res)
A
Andy Adamson 已提交
559
{
560
	struct nfs4_session *session;
A
Andy Adamson 已提交
561
	struct nfs4_slot_table *tbl;
562
	bool send_new_highest_used_slotid = false;
A
Andy Adamson 已提交
563

564
	if (!res->sr_slot) {
A
Andy Adamson 已提交
565 566
		/* just wake up the next guy waiting since
		 * we may have not consumed a slot after all */
A
Andy Adamson 已提交
567
		dprintk("%s: No slot\n", __func__);
568
		return;
A
Andy Adamson 已提交
569
	}
570 571
	tbl = res->sr_slot->table;
	session = tbl->session;
572

573
	spin_lock(&tbl->slot_tbl_lock);
574 575 576 577 578 579
	/* 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;

580 581 582 583
	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
584
	nfs4_free_slot(tbl, res->sr_slot);
585 586 587

	if (tbl->highest_used_slotid != NFS4_NO_SLOT)
		send_new_highest_used_slotid = false;
588
out_unlock:
589
	spin_unlock(&tbl->slot_tbl_lock);
590
	res->sr_slot = NULL;
591 592
	if (send_new_highest_used_slotid)
		nfs41_server_notify_highest_slotid_update(session->clp);
A
Andy Adamson 已提交
593 594
}

595
static int nfs41_sequence_done(struct rpc_task *task, struct nfs4_sequence_res *res)
A
Andy Adamson 已提交
596
{
597
	struct nfs4_session *session;
598
	struct nfs4_slot *slot;
599
	struct nfs_client *clp;
600
	bool interrupted = false;
601
	int ret = 1;
A
Andy Adamson 已提交
602

603 604
	/* don't increment the sequence number if the task wasn't sent */
	if (!RPC_WAS_SENT(task))
A
Andy Adamson 已提交
605 606
		goto out;

607
	slot = res->sr_slot;
608
	session = slot->table->session;
609

610 611 612 613 614
	if (slot->interrupted) {
		slot->interrupted = 0;
		interrupted = true;
	}

615
	trace_nfs4_sequence_done(session, res);
A
Andy Adamson 已提交
616
	/* Check the SEQUENCE operation status */
617 618
	switch (res->sr_status) {
	case 0:
A
Andy Adamson 已提交
619
		/* Update the slot's sequence and clientid lease timer */
620
		++slot->seq_nr;
621
		clp = session->clp;
622
		do_renew_lease(clp, res->sr_timestamp);
623
		/* Check sequence flags */
624 625
		if (res->sr_status_flags != 0)
			nfs4_schedule_lease_recovery(clp);
626
		nfs41_update_target_slotid(slot->table, slot, res);
627
		break;
628 629 630 631 632 633 634 635 636
	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;
637 638 639 640 641
	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.
		 */
642
		dprintk("%s: slot=%u seq=%u: Operation in progress\n",
643
			__func__,
644
			slot->slot_nr,
645
			slot->seq_nr);
646
		goto out_retry;
647 648 649 650 651
	case -NFS4ERR_BADSLOT:
		/*
		 * The slot id we used was probably retired. Try again
		 * using a different slot id.
		 */
652 653
		goto retry_nowait;
	case -NFS4ERR_SEQ_MISORDERED:
654 655 656 657 658 659 660 661
		/*
		 * Was the last operation on this sequence interrupted?
		 * If so, retry after bumping the sequence number.
		 */
		if (interrupted) {
			++slot->seq_nr;
			goto retry_nowait;
		}
662 663 664 665
		/*
		 * Could this slot have been previously retired?
		 * If so, then the server may be expecting seq_nr = 1!
		 */
666 667 668 669 670
		if (slot->seq_nr != 1) {
			slot->seq_nr = 1;
			goto retry_nowait;
		}
		break;
671 672 673
	case -NFS4ERR_SEQ_FALSE_RETRY:
		++slot->seq_nr;
		goto retry_nowait;
674 675
	default:
		/* Just update the slot sequence no. */
676
		++slot->seq_nr;
A
Andy Adamson 已提交
677 678 679 680
	}
out:
	/* The session may be reset by one of the error handlers. */
	dprintk("%s: Error %d free the slot \n", __func__, res->sr_status);
681
	nfs41_sequence_free_slot(res);
682
	return ret;
683 684 685 686 687 688
retry_nowait:
	if (rpc_restart_call_prepare(task)) {
		task->tk_status = 0;
		ret = 0;
	}
	goto out;
689
out_retry:
690
	if (!rpc_restart_call(task))
691 692 693
		goto out;
	rpc_delay(task, NFS4_POLL_RETRY_MAX);
	return 0;
A
Andy Adamson 已提交
694 695
}

696 697
static int nfs4_sequence_done(struct rpc_task *task,
			       struct nfs4_sequence_res *res)
698
{
699
	if (res->sr_slot == NULL)
700
		return 1;
C
Chuck Lever 已提交
701 702
	if (!res->sr_slot->table->session)
		return nfs40_sequence_done(task, res);
703
	return nfs41_sequence_done(task, res);
704 705
}

A
Andy Adamson 已提交
706
int nfs41_setup_sequence(struct nfs4_session *session,
A
Andy Adamson 已提交
707 708 709 710
				struct nfs4_sequence_args *args,
				struct nfs4_sequence_res *res,
				struct rpc_task *task)
{
A
Andy Adamson 已提交
711 712 713 714
	struct nfs4_slot *slot;
	struct nfs4_slot_table *tbl;

	dprintk("--> %s\n", __func__);
A
Andy Adamson 已提交
715
	/* slot already allocated? */
716
	if (res->sr_slot != NULL)
717
		goto out_success;
A
Andy Adamson 已提交
718

A
Andy Adamson 已提交
719 720
	tbl = &session->fc_slot_table;

721 722
	task->tk_timeout = 0;

A
Andy Adamson 已提交
723
	spin_lock(&tbl->slot_tbl_lock);
724
	if (test_bit(NFS4_SLOT_TBL_DRAINING, &tbl->slot_tbl_state) &&
725
	    !args->sa_privileged) {
726 727
		/* The state manager will wait until the slot table is empty */
		dprintk("%s session is draining\n", __func__);
728
		goto out_sleep;
729 730
	}

731
	slot = nfs4_alloc_slot(tbl);
732 733 734 735
	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
736
		dprintk("<-- %s: no free slots\n", __func__);
737
		goto out_sleep;
A
Andy Adamson 已提交
738 739 740
	}
	spin_unlock(&tbl->slot_tbl_lock);

741
	args->sa_slot = slot;
A
Andy Adamson 已提交
742

743
	dprintk("<-- %s slotid=%u seqid=%u\n", __func__,
744
			slot->slot_nr, slot->seq_nr);
A
Andy Adamson 已提交
745

746
	res->sr_slot = slot;
747
	res->sr_timestamp = jiffies;
748
	res->sr_status_flags = 0;
A
Andy Adamson 已提交
749 750 751 752 753
	/*
	 * 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;
754
	trace_nfs4_setup_sequence(session, args);
755 756
out_success:
	rpc_call_start(task);
A
Andy Adamson 已提交
757
	return 0;
758
out_sleep:
759 760
	/* Privileged tasks are queued with top priority */
	if (args->sa_privileged)
761 762 763 764
		rpc_sleep_on_priority(&tbl->slot_tbl_waitq, task,
				NULL, RPC_PRIORITY_PRIVILEGED);
	else
		rpc_sleep_on(&tbl->slot_tbl_waitq, task, NULL);
765 766
	spin_unlock(&tbl->slot_tbl_lock);
	return -EAGAIN;
A
Andy Adamson 已提交
767
}
A
Andy Adamson 已提交
768
EXPORT_SYMBOL_GPL(nfs41_setup_sequence);
A
Andy Adamson 已提交
769

770 771 772 773
static int nfs4_setup_sequence(const struct nfs_server *server,
			       struct nfs4_sequence_args *args,
			       struct nfs4_sequence_res *res,
			       struct rpc_task *task)
A
Andy Adamson 已提交
774
{
775
	struct nfs4_session *session = nfs4_get_session(server);
A
Andy Adamson 已提交
776 777
	int ret = 0;

C
Chuck Lever 已提交
778 779
	if (!session)
		return nfs40_setup_sequence(server, args, res, task);
780

781
	dprintk("--> %s clp %p session %p sr_slot %u\n",
782
		__func__, session->clp, session, res->sr_slot ?
783
			res->sr_slot->slot_nr : NFS4_NO_SLOT);
A
Andy Adamson 已提交
784

785
	ret = nfs41_setup_sequence(session, args, res, task);
C
Chuck Lever 已提交
786

A
Andy Adamson 已提交
787 788 789 790 791 792
	dprintk("<-- %s status=%d\n", __func__, ret);
	return ret;
}

static void nfs41_call_sync_prepare(struct rpc_task *task, void *calldata)
{
793
	struct nfs4_call_sync_data *data = calldata;
794
	struct nfs4_session *session = nfs4_get_session(data->seq_server);
A
Andy Adamson 已提交
795

796 797
	dprintk("--> %s data->seq_server %p\n", __func__, data->seq_server);

798
	nfs41_setup_sequence(session, data->seq_args, data->seq_res, task);
A
Andy Adamson 已提交
799 800
}

A
Andy Adamson 已提交
801 802
static void nfs41_call_sync_done(struct rpc_task *task, void *calldata)
{
803
	struct nfs4_call_sync_data *data = calldata;
A
Andy Adamson 已提交
804

805
	nfs41_sequence_done(task, data->seq_res);
A
Andy Adamson 已提交
806 807
}

808
static const struct rpc_call_ops nfs41_call_sync_ops = {
A
Andy Adamson 已提交
809
	.rpc_call_prepare = nfs41_call_sync_prepare,
A
Andy Adamson 已提交
810
	.rpc_call_done = nfs41_call_sync_done,
A
Andy Adamson 已提交
811 812
};

C
Chuck Lever 已提交
813 814
#else	/* !CONFIG_NFS_V4_1 */

815 816 817 818 819
static int nfs4_setup_sequence(const struct nfs_server *server,
			       struct nfs4_sequence_args *args,
			       struct nfs4_sequence_res *res,
			       struct rpc_task *task)
{
C
Chuck Lever 已提交
820
	return nfs40_setup_sequence(server, args, res, task);
821 822 823 824 825
}

static int nfs4_sequence_done(struct rpc_task *task,
			       struct nfs4_sequence_res *res)
{
C
Chuck Lever 已提交
826
	return nfs40_sequence_done(task, res);
827
}
C
Chuck Lever 已提交
828 829

#endif	/* !CONFIG_NFS_V4_1 */
830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848

static void nfs40_call_sync_prepare(struct rpc_task *task, void *calldata)
{
	struct nfs4_call_sync_data *data = calldata;
	nfs4_setup_sequence(data->seq_server,
				data->seq_args, data->seq_res, task);
}

static void nfs40_call_sync_done(struct rpc_task *task, void *calldata)
{
	struct nfs4_call_sync_data *data = calldata;
	nfs4_sequence_done(task, data->seq_res);
}

static const struct rpc_call_ops nfs40_call_sync_ops = {
	.rpc_call_prepare = nfs40_call_sync_prepare,
	.rpc_call_done = nfs40_call_sync_done,
};

849 850
static int nfs4_call_sync_sequence(struct rpc_clnt *clnt,
				   struct nfs_server *server,
A
Andy Adamson 已提交
851 852
				   struct rpc_message *msg,
				   struct nfs4_sequence_args *args,
853
				   struct nfs4_sequence_res *res)
A
Andy Adamson 已提交
854 855 856
{
	int ret;
	struct rpc_task *task;
857
	struct nfs_client *clp = server->nfs_client;
858
	struct nfs4_call_sync_data data = {
859
		.seq_server = server,
A
Andy Adamson 已提交
860 861 862 863
		.seq_args = args,
		.seq_res = res,
	};
	struct rpc_task_setup task_setup = {
864
		.rpc_client = clnt,
A
Andy Adamson 已提交
865
		.rpc_message = msg,
866
		.callback_ops = clp->cl_mvops->call_sync_ops,
A
Andy Adamson 已提交
867 868 869 870 871 872 873 874 875 876 877 878 879
		.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;
}

880
static
881 882
int nfs4_call_sync(struct rpc_clnt *clnt,
		   struct nfs_server *server,
883 884 885 886 887
		   struct rpc_message *msg,
		   struct nfs4_sequence_args *args,
		   struct nfs4_sequence_res *res,
		   int cache_reply)
{
888
	nfs4_init_sequence(args, res, cache_reply);
889
	return nfs4_call_sync_sequence(clnt, server, msg, args, res);
890
}
A
Andy Adamson 已提交
891

892
static void update_changeattr(struct inode *dir, struct nfs4_change_info *cinfo)
L
Linus Torvalds 已提交
893
{
894
	struct nfs_inode *nfsi = NFS_I(dir);
L
Linus Torvalds 已提交
895

896
	spin_lock(&dir->i_lock);
897
	nfsi->cache_validity |= NFS_INO_INVALID_ATTR|NFS_INO_INVALID_DATA;
898
	if (!cinfo->atomic || cinfo->before != dir->i_version)
899
		nfs_force_lookup_revalidate(dir);
900
	dir->i_version = cinfo->after;
D
David Howells 已提交
901
	nfs_fscache_invalidate(dir);
902
	spin_unlock(&dir->i_lock);
L
Linus Torvalds 已提交
903 904
}

905
struct nfs4_opendata {
906
	struct kref kref;
907 908
	struct nfs_openargs o_arg;
	struct nfs_openres o_res;
909 910
	struct nfs_open_confirmargs c_arg;
	struct nfs_open_confirmres c_res;
911 912
	struct nfs4_string owner_name;
	struct nfs4_string group_name;
913
	struct nfs_fattr f_attr;
914
	struct nfs4_label *f_label;
915
	struct dentry *dir;
916
	struct dentry *dentry;
917
	struct nfs4_state_owner *owner;
918
	struct nfs4_state *state;
919
	struct iattr attrs;
920
	unsigned long timestamp;
921
	unsigned int rpc_done : 1;
922
	unsigned int file_created : 1;
923
	unsigned int is_recover : 1;
924 925
	int rpc_status;
	int cancelled;
926 927
};

928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956
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;
	}
}
957 958 959 960

static void nfs4_init_opendata_res(struct nfs4_opendata *p)
{
	p->o_res.f_attr = &p->f_attr;
961
	p->o_res.f_label = p->f_label;
962 963
	p->o_res.seqid = p->o_arg.seqid;
	p->c_res.seqid = p->c_arg.seqid;
964
	p->o_res.server = p->o_arg.server;
965
	p->o_res.access_request = p->o_arg.access;
966
	nfs_fattr_init(&p->f_attr);
967
	nfs_fattr_init_names(&p->f_attr, &p->owner_name, &p->group_name);
968 969
}

970
static struct nfs4_opendata *nfs4_opendata_alloc(struct dentry *dentry,
971
		struct nfs4_state_owner *sp, fmode_t fmode, int flags,
972
		const struct iattr *attrs,
973
		struct nfs4_label *label,
974
		enum open_claim_type4 claim,
975
		gfp_t gfp_mask)
976
{
977
	struct dentry *parent = dget_parent(dentry);
978 979 980 981
	struct inode *dir = parent->d_inode;
	struct nfs_server *server = NFS_SERVER(dir);
	struct nfs4_opendata *p;

982
	p = kzalloc(sizeof(*p), gfp_mask);
983 984
	if (p == NULL)
		goto err;
985 986 987 988 989

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

990
	p->o_arg.seqid = nfs_alloc_seqid(&sp->so_seqid, gfp_mask);
991
	if (p->o_arg.seqid == NULL)
992
		goto err_free_label;
993 994
	nfs_sb_active(dentry->d_sb);
	p->dentry = dget(dentry);
995 996 997
	p->dir = parent;
	p->owner = sp;
	atomic_inc(&sp->so_count);
998 999
	p->o_arg.open_flags = flags;
	p->o_arg.fmode = fmode & (FMODE_READ|FMODE_WRITE);
1000 1001 1002 1003 1004 1005 1006 1007
	/* 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;
	}
1008
	p->o_arg.clientid = server->nfs_client->cl_clientid;
1009 1010
	p->o_arg.id.create_time = ktime_to_ns(sp->so_seqid.create_time);
	p->o_arg.id.uniquifier = sp->so_seqid.owner_id;
1011
	p->o_arg.name = &dentry->d_name;
1012
	p->o_arg.server = server;
1013
	p->o_arg.bitmask = nfs4_bitmask(server, label);
1014
	p->o_arg.open_bitmap = &nfs4_fattr_bitmap[0];
1015
	p->o_arg.label = label;
1016 1017
	p->o_arg.claim = nfs4_map_atomic_open_claim(server, claim);
	switch (p->o_arg.claim) {
1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028
	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);
	}
1029
	if (attrs != NULL && attrs->ia_valid != 0) {
1030
		__u32 verf[2];
1031

1032 1033
		p->o_arg.u.attrs = &p->attrs;
		memcpy(&p->attrs, attrs, sizeof(p->attrs));
1034 1035 1036 1037 1038

		verf[0] = jiffies;
		verf[1] = current->pid;
		memcpy(p->o_arg.u.verifier.data, verf,
				sizeof(p->o_arg.u.verifier.data));
1039
	}
1040 1041 1042
	p->c_arg.fh = &p->o_res.fh;
	p->c_arg.stateid = &p->o_res.stateid;
	p->c_arg.seqid = p->o_arg.seqid;
1043
	nfs4_init_opendata_res(p);
1044
	kref_init(&p->kref);
1045
	return p;
1046 1047 1048 1049

err_free_label:
	nfs4_label_free(p->f_label);
err_free_p:
1050 1051 1052 1053 1054 1055
	kfree(p);
err:
	dput(parent);
	return NULL;
}

1056
static void nfs4_opendata_free(struct kref *kref)
1057
{
1058 1059
	struct nfs4_opendata *p = container_of(kref,
			struct nfs4_opendata, kref);
1060
	struct super_block *sb = p->dentry->d_sb;
1061 1062

	nfs_free_seqid(p->o_arg.seqid);
1063 1064
	if (p->state != NULL)
		nfs4_put_open_state(p->state);
1065
	nfs4_put_state_owner(p->owner);
1066 1067 1068

	nfs4_label_free(p->f_label);

1069
	dput(p->dir);
1070 1071
	dput(p->dentry);
	nfs_sb_deactive(sb);
1072
	nfs_fattr_free_names(&p->f_attr);
1073 1074 1075 1076 1077 1078 1079
	kfree(p);
}

static void nfs4_opendata_put(struct nfs4_opendata *p)
{
	if (p != NULL)
		kref_put(&p->kref, nfs4_opendata_free);
1080 1081
}

1082 1083 1084 1085 1086 1087 1088 1089
static int nfs4_wait_for_completion_rpc_task(struct rpc_task *task)
{
	int ret;

	ret = rpc_wait_for_completion_task(task);
	return ret;
}

1090
static int can_open_cached(struct nfs4_state *state, fmode_t mode, int open_mode)
1091 1092
{
	int ret = 0;
1093

1094
	if (open_mode & (O_EXCL|O_TRUNC))
1095 1096
		goto out;
	switch (mode & (FMODE_READ|FMODE_WRITE)) {
1097
		case FMODE_READ:
1098 1099
			ret |= test_bit(NFS_O_RDONLY_STATE, &state->flags) != 0
				&& state->n_rdonly != 0;
1100 1101
			break;
		case FMODE_WRITE:
1102 1103
			ret |= test_bit(NFS_O_WRONLY_STATE, &state->flags) != 0
				&& state->n_wronly != 0;
1104 1105
			break;
		case FMODE_READ|FMODE_WRITE:
1106 1107
			ret |= test_bit(NFS_O_RDWR_STATE, &state->flags) != 0
				&& state->n_rdwr != 0;
1108
	}
1109
out:
1110 1111 1112
	return ret;
}

1113
static int can_open_delegated(struct nfs_delegation *delegation, fmode_t fmode)
1114
{
1115 1116
	if (delegation == NULL)
		return 0;
1117
	if ((delegation->type & fmode) != fmode)
1118
		return 0;
1119
	if (test_bit(NFS_DELEGATION_NEED_RECLAIM, &delegation->flags))
1120
		return 0;
1121 1122
	if (test_bit(NFS_DELEGATION_RETURNING, &delegation->flags))
		return 0;
1123
	nfs_mark_delegation_referenced(delegation);
1124 1125 1126
	return 1;
}

1127
static void update_open_stateflags(struct nfs4_state *state, fmode_t fmode)
1128
{
1129
	switch (fmode) {
1130 1131 1132 1133 1134 1135 1136 1137 1138
		case FMODE_WRITE:
			state->n_wronly++;
			break;
		case FMODE_READ:
			state->n_rdonly++;
			break;
		case FMODE_READ|FMODE_WRITE:
			state->n_rdwr++;
	}
1139
	nfs4_state_set_mode_locked(state, state->state | fmode);
1140 1141
}

1142
static void nfs_set_open_stateid_locked(struct nfs4_state *state, nfs4_stateid *stateid, fmode_t fmode)
1143 1144
{
	if (test_bit(NFS_DELEGATED_STATE, &state->flags) == 0)
1145 1146
		nfs4_stateid_copy(&state->stateid, stateid);
	nfs4_stateid_copy(&state->open_stateid, stateid);
1147
	set_bit(NFS_OPEN_STATE, &state->flags);
1148
	switch (fmode) {
1149 1150 1151 1152 1153 1154 1155 1156 1157
		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);
	}
1158 1159
}

1160
static void nfs_set_open_stateid(struct nfs4_state *state, nfs4_stateid *stateid, fmode_t fmode)
1161
{
1162
	write_seqlock(&state->seqlock);
1163
	nfs_set_open_stateid_locked(state, stateid, fmode);
1164
	write_sequnlock(&state->seqlock);
1165 1166
}

1167
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
1168
{
1169 1170 1171 1172 1173
	/*
	 * Protect the call to nfs4_state_set_mode_locked and
	 * serialise the stateid update
	 */
	write_seqlock(&state->seqlock);
1174
	if (deleg_stateid != NULL) {
1175
		nfs4_stateid_copy(&state->stateid, deleg_stateid);
1176 1177 1178
		set_bit(NFS_DELEGATED_STATE, &state->flags);
	}
	if (open_stateid != NULL)
1179
		nfs_set_open_stateid_locked(state, open_stateid, fmode);
1180 1181
	write_sequnlock(&state->seqlock);
	spin_lock(&state->owner->so_lock);
1182
	update_open_stateflags(state, fmode);
1183
	spin_unlock(&state->owner->so_lock);
L
Linus Torvalds 已提交
1184 1185
}

1186
static int update_open_stateid(struct nfs4_state *state, nfs4_stateid *open_stateid, nfs4_stateid *delegation, fmode_t fmode)
1187 1188 1189 1190 1191
{
	struct nfs_inode *nfsi = NFS_I(state->inode);
	struct nfs_delegation *deleg_cur;
	int ret = 0;

1192
	fmode &= (FMODE_READ|FMODE_WRITE);
1193 1194 1195 1196 1197 1198 1199

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

	spin_lock(&deleg_cur->lock);
1200
	if (rcu_dereference(nfsi->delegation) != deleg_cur ||
1201
	   test_bit(NFS_DELEGATION_RETURNING, &deleg_cur->flags) ||
1202
	    (deleg_cur->type & fmode) != fmode)
1203 1204 1205 1206
		goto no_delegation_unlock;

	if (delegation == NULL)
		delegation = &deleg_cur->stateid;
1207
	else if (!nfs4_stateid_match(&deleg_cur->stateid, delegation))
1208 1209
		goto no_delegation_unlock;

1210
	nfs_mark_delegation_referenced(deleg_cur);
1211
	__update_open_stateid(state, open_stateid, &deleg_cur->stateid, fmode);
1212 1213 1214 1215 1216 1217 1218
	ret = 1;
no_delegation_unlock:
	spin_unlock(&deleg_cur->lock);
no_delegation:
	rcu_read_unlock();

	if (!ret && open_stateid != NULL) {
1219
		__update_open_stateid(state, open_stateid, NULL, fmode);
1220 1221 1222 1223 1224 1225 1226
		ret = 1;
	}

	return ret;
}


1227
static void nfs4_return_incompatible_delegation(struct inode *inode, fmode_t fmode)
1228 1229 1230 1231 1232
{
	struct nfs_delegation *delegation;

	rcu_read_lock();
	delegation = rcu_dereference(NFS_I(inode)->delegation);
1233
	if (delegation == NULL || (delegation->type & fmode) == fmode) {
1234 1235 1236 1237
		rcu_read_unlock();
		return;
	}
	rcu_read_unlock();
1238
	nfs4_inode_return_delegation(inode);
1239 1240
}

1241
static struct nfs4_state *nfs4_try_open_cached(struct nfs4_opendata *opendata)
1242 1243 1244 1245
{
	struct nfs4_state *state = opendata->state;
	struct nfs_inode *nfsi = NFS_I(state->inode);
	struct nfs_delegation *delegation;
1246
	int open_mode = opendata->o_arg.open_flags;
1247
	fmode_t fmode = opendata->o_arg.fmode;
1248 1249 1250 1251
	nfs4_stateid stateid;
	int ret = -EAGAIN;

	for (;;) {
1252
		if (can_open_cached(state, fmode, open_mode)) {
1253
			spin_lock(&state->owner->so_lock);
1254 1255
			if (can_open_cached(state, fmode, open_mode)) {
				update_open_stateflags(state, fmode);
1256 1257 1258 1259 1260
				spin_unlock(&state->owner->so_lock);
				goto out_return_state;
			}
			spin_unlock(&state->owner->so_lock);
		}
1261 1262
		rcu_read_lock();
		delegation = rcu_dereference(nfsi->delegation);
1263
		if (!can_open_delegated(delegation, fmode)) {
1264
			rcu_read_unlock();
1265
			break;
1266
		}
1267
		/* Save the delegation */
1268
		nfs4_stateid_copy(&stateid, &delegation->stateid);
1269
		rcu_read_unlock();
1270
		nfs_release_seqid(opendata->o_arg.seqid);
1271 1272 1273 1274 1275
		if (!opendata->is_recover) {
			ret = nfs_may_open(state->inode, state->owner->so_cred, open_mode);
			if (ret != 0)
				goto out;
		}
1276
		ret = -EAGAIN;
1277 1278

		/* Try to update the stateid using the delegation */
1279
		if (update_open_stateid(state, NULL, &stateid, fmode))
1280
			goto out_return_state;
1281 1282 1283 1284 1285 1286 1287 1288
	}
out:
	return ERR_PTR(ret);
out_return_state:
	atomic_inc(&state->count);
	return state;
}

1289 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326
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;

1327 1328 1329 1330 1331 1332 1333
	if (!data->rpc_done) {
		if (data->rpc_status) {
			ret = data->rpc_status;
			goto err;
		}
		/* cached opens have already been processed */
		goto update;
1334 1335 1336 1337 1338 1339 1340 1341
	}

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

	if (data->o_res.delegation_type != 0)
		nfs4_opendata_check_deleg(data, state);
1342
update:
1343 1344
	update_open_stateid(state, &data->o_res.stateid, NULL,
			    data->o_arg.fmode);
1345
	atomic_inc(&state->count);
1346 1347 1348 1349 1350 1351 1352 1353 1354

	return state;
err:
	return ERR_PTR(ret);

}

static struct nfs4_state *
_nfs4_opendata_to_nfs4_state(struct nfs4_opendata *data)
1355 1356 1357
{
	struct inode *inode;
	struct nfs4_state *state = NULL;
1358
	int ret;
1359

1360
	if (!data->rpc_done) {
1361
		state = nfs4_try_open_cached(data);
1362 1363 1364
		goto out;
	}

1365
	ret = -EAGAIN;
1366
	if (!(data->f_attr.valid & NFS_ATTR_FATTR))
1367
		goto err;
1368
	inode = nfs_fhget(data->dir->d_sb, &data->o_res.fh, &data->f_attr, data->f_label);
1369
	ret = PTR_ERR(inode);
1370
	if (IS_ERR(inode))
1371 1372
		goto err;
	ret = -ENOMEM;
1373 1374
	state = nfs4_get_open_state(inode, data->owner);
	if (state == NULL)
1375
		goto err_put_inode;
1376 1377
	if (data->o_res.delegation_type != 0)
		nfs4_opendata_check_deleg(data, state);
1378
	update_open_stateid(state, &data->o_res.stateid, NULL,
1379
			data->o_arg.fmode);
1380
	iput(inode);
1381
out:
1382
	nfs_release_seqid(data->o_arg.seqid);
1383
	return state;
1384 1385 1386 1387
err_put_inode:
	iput(inode);
err:
	return ERR_PTR(ret);
1388 1389
}

1390 1391 1392 1393 1394 1395 1396 1397
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);
}

1398 1399 1400 1401 1402 1403 1404 1405 1406 1407 1408 1409 1410 1411 1412 1413 1414
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);
}

1415 1416
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
1417 1418 1419
{
	struct nfs4_opendata *opendata;

1420
	opendata = nfs4_opendata_alloc(ctx->dentry, state->owner, 0, 0,
1421
			NULL, NULL, claim, GFP_NOFS);
T
Trond Myklebust 已提交
1422 1423 1424 1425 1426 1427 1428
	if (opendata == NULL)
		return ERR_PTR(-ENOMEM);
	opendata->state = state;
	atomic_inc(&state->count);
	return opendata;
}

1429
static int nfs4_open_recover_helper(struct nfs4_opendata *opendata, fmode_t fmode, struct nfs4_state **res)
1430
{
1431
	struct nfs4_state *newstate;
1432 1433
	int ret;

1434 1435
	opendata->o_arg.open_flags = 0;
	opendata->o_arg.fmode = fmode;
1436 1437 1438
	memset(&opendata->o_res, 0, sizeof(opendata->o_res));
	memset(&opendata->c_res, 0, sizeof(opendata->c_res));
	nfs4_init_opendata_res(opendata);
1439
	ret = _nfs4_recover_proc_open(opendata);
1440 1441
	if (ret != 0)
		return ret; 
1442
	newstate = nfs4_opendata_to_nfs4_state(opendata);
1443 1444
	if (IS_ERR(newstate))
		return PTR_ERR(newstate);
1445
	nfs4_close_state(newstate, fmode);
1446
	*res = newstate;
1447 1448 1449 1450 1451 1452 1453 1454 1455
	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_* */
1456
	clear_bit(NFS_DELEGATED_STATE, &state->flags);
1457
	clear_bit(NFS_OPEN_STATE, &state->flags);
1458 1459
	smp_rmb();
	if (state->n_rdwr != 0) {
T
Trond Myklebust 已提交
1460
		clear_bit(NFS_O_RDWR_STATE, &state->flags);
1461
		ret = nfs4_open_recover_helper(opendata, FMODE_READ|FMODE_WRITE, &newstate);
1462 1463
		if (ret != 0)
			return ret;
1464 1465
		if (newstate != state)
			return -ESTALE;
1466 1467
	}
	if (state->n_wronly != 0) {
T
Trond Myklebust 已提交
1468
		clear_bit(NFS_O_WRONLY_STATE, &state->flags);
1469
		ret = nfs4_open_recover_helper(opendata, FMODE_WRITE, &newstate);
1470 1471
		if (ret != 0)
			return ret;
1472 1473
		if (newstate != state)
			return -ESTALE;
1474 1475
	}
	if (state->n_rdonly != 0) {
T
Trond Myklebust 已提交
1476
		clear_bit(NFS_O_RDONLY_STATE, &state->flags);
1477
		ret = nfs4_open_recover_helper(opendata, FMODE_READ, &newstate);
1478 1479
		if (ret != 0)
			return ret;
1480 1481
		if (newstate != state)
			return -ESTALE;
1482
	}
1483 1484 1485 1486 1487
	/*
	 * 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 &&
1488
	    !nfs4_stateid_match(&state->stateid, &state->open_stateid)) {
1489
		write_seqlock(&state->seqlock);
1490
		if (test_bit(NFS_DELEGATED_STATE, &state->flags) == 0)
1491
			nfs4_stateid_copy(&state->stateid, &state->open_stateid);
1492
		write_sequnlock(&state->seqlock);
1493
	}
1494 1495 1496
	return 0;
}

L
Linus Torvalds 已提交
1497 1498 1499 1500
/*
 * OPEN_RECLAIM:
 * 	reclaim state on the server after a reboot.
 */
1501
static int _nfs4_do_open_reclaim(struct nfs_open_context *ctx, struct nfs4_state *state)
L
Linus Torvalds 已提交
1502
{
1503
	struct nfs_delegation *delegation;
1504
	struct nfs4_opendata *opendata;
1505
	fmode_t delegation_type = 0;
L
Linus Torvalds 已提交
1506 1507
	int status;

1508 1509
	opendata = nfs4_open_recoverdata_alloc(ctx, state,
			NFS4_OPEN_CLAIM_PREVIOUS);
T
Trond Myklebust 已提交
1510 1511
	if (IS_ERR(opendata))
		return PTR_ERR(opendata);
1512 1513
	rcu_read_lock();
	delegation = rcu_dereference(NFS_I(state->inode)->delegation);
1514
	if (delegation != NULL && test_bit(NFS_DELEGATION_NEED_RECLAIM, &delegation->flags) != 0)
1515
		delegation_type = delegation->type;
1516
	rcu_read_unlock();
1517 1518
	opendata->o_arg.u.delegation_type = delegation_type;
	status = nfs4_open_recover(opendata, state);
1519
	nfs4_opendata_put(opendata);
L
Linus Torvalds 已提交
1520 1521 1522
	return status;
}

1523
static int nfs4_do_open_reclaim(struct nfs_open_context *ctx, struct nfs4_state *state)
L
Linus Torvalds 已提交
1524 1525 1526 1527 1528
{
	struct nfs_server *server = NFS_SERVER(state->inode);
	struct nfs4_exception exception = { };
	int err;
	do {
1529
		err = _nfs4_do_open_reclaim(ctx, state);
1530
		trace_nfs4_open_reclaim(ctx, 0, err);
1531 1532
		if (nfs4_clear_cap_atomic_open_v1(server, err, &exception))
			continue;
1533
		if (err != -NFS4ERR_DELAY)
1534 1535
			break;
		nfs4_handle_exception(server, err, &exception);
L
Linus Torvalds 已提交
1536 1537 1538 1539
	} while (exception.retry);
	return err;
}

1540 1541 1542 1543 1544 1545 1546
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))
1547
		return -EAGAIN;
1548
	ret = nfs4_do_open_reclaim(ctx, state);
1549 1550 1551 1552
	put_nfs_open_context(ctx);
	return ret;
}

1553
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
1554
{
1555 1556 1557 1558 1559 1560 1561 1562 1563 1564 1565 1566 1567 1568 1569 1570 1571 1572 1573 1574 1575 1576 1577
	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;
1578 1579 1580
		case -NFS4ERR_MOVED:
			nfs4_schedule_migration_recovery(server);
			return -EAGAIN;
1581 1582 1583
		case -NFS4ERR_LEASE_MOVED:
			nfs4_schedule_lease_moved_recovery(server->nfs_client);
			return -EAGAIN;
1584 1585 1586
		case -NFS4ERR_DELEG_REVOKED:
		case -NFS4ERR_ADMIN_REVOKED:
		case -NFS4ERR_BAD_STATEID:
1587
		case -NFS4ERR_OPENMODE:
1588 1589 1590 1591 1592 1593 1594 1595 1596
			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;
1597 1598 1599 1600
		case -ENOMEM:
		case -NFS4ERR_DENIED:
			/* kill_proc(fl->fl_pid, SIGLOST, 1); */
			return 0;
1601
	}
L
Linus Torvalds 已提交
1602 1603 1604
	return err;
}

1605 1606 1607 1608 1609 1610 1611 1612 1613 1614 1615 1616 1617 1618 1619 1620
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);
}

1621 1622 1623 1624 1625 1626 1627 1628
static void nfs4_open_confirm_prepare(struct rpc_task *task, void *calldata)
{
	struct nfs4_opendata *data = calldata;

	nfs40_setup_sequence(data->o_arg.server, &data->o_arg.seq_args,
				&data->o_res.seq_res, task);
}

1629 1630 1631 1632
static void nfs4_open_confirm_done(struct rpc_task *task, void *calldata)
{
	struct nfs4_opendata *data = calldata;

1633 1634
	nfs40_sequence_done(task, &data->o_res.seq_res);

1635
	data->rpc_status = task->tk_status;
1636
	if (data->rpc_status == 0) {
1637
		nfs4_stateid_copy(&data->o_res.stateid, &data->c_res.stateid);
1638
		nfs_confirm_seqid(&data->owner->so_seqid, 0);
1639
		renew_lease(data->o_res.server, data->timestamp);
1640
		data->rpc_done = 1;
1641
	}
1642 1643 1644 1645 1646 1647 1648 1649 1650 1651 1652
}

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! */
1653
	if (!data->rpc_done)
1654 1655
		goto out_free;
	state = nfs4_opendata_to_nfs4_state(data);
1656
	if (!IS_ERR(state))
1657
		nfs4_close_state(state, data->o_arg.fmode);
1658
out_free:
1659
	nfs4_opendata_put(data);
1660 1661 1662
}

static const struct rpc_call_ops nfs4_open_confirm_ops = {
1663
	.rpc_call_prepare = nfs4_open_confirm_prepare,
1664 1665 1666 1667 1668 1669 1670 1671 1672 1673 1674
	.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;
1675 1676 1677 1678 1679 1680
	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
1681 1682
	struct rpc_task_setup task_setup_data = {
		.rpc_client = server->client,
1683
		.rpc_message = &msg,
T
Trond Myklebust 已提交
1684 1685
		.callback_ops = &nfs4_open_confirm_ops,
		.callback_data = data,
1686
		.workqueue = nfsiod_workqueue,
T
Trond Myklebust 已提交
1687 1688
		.flags = RPC_TASK_ASYNC,
	};
L
Linus Torvalds 已提交
1689 1690
	int status;

1691
	nfs4_init_sequence(&data->o_arg.seq_args, &data->o_res.seq_res, 1);
1692
	kref_get(&data->kref);
1693 1694
	data->rpc_done = 0;
	data->rpc_status = 0;
1695
	data->timestamp = jiffies;
T
Trond Myklebust 已提交
1696
	task = rpc_run_task(&task_setup_data);
1697
	if (IS_ERR(task))
1698 1699 1700 1701 1702 1703 1704
		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;
1705
	rpc_put_task(task);
L
Linus Torvalds 已提交
1706 1707 1708
	return status;
}

1709
static void nfs4_open_prepare(struct rpc_task *task, void *calldata)
L
Linus Torvalds 已提交
1710
{
1711 1712
	struct nfs4_opendata *data = calldata;
	struct nfs4_state_owner *sp = data->owner;
1713
	struct nfs_client *clp = sp->so_server->nfs_client;
1714

1715
	if (nfs_wait_on_sequence(data->o_arg.seqid, task) != 0)
1716
		goto out_wait;
1717 1718 1719 1720 1721 1722 1723
	/*
	 * 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;

1724
		if (can_open_cached(data->state, data->o_arg.fmode, data->o_arg.open_flags))
1725
			goto out_no_action;
1726 1727
		rcu_read_lock();
		delegation = rcu_dereference(NFS_I(data->state->inode)->delegation);
1728
		if (data->o_arg.claim != NFS4_OPEN_CLAIM_DELEGATE_CUR &&
1729
		    data->o_arg.claim != NFS4_OPEN_CLAIM_DELEG_CUR_FH &&
1730 1731
		    can_open_delegated(delegation, data->o_arg.fmode))
			goto unlock_no_action;
1732 1733
		rcu_read_unlock();
	}
1734
	/* Update client id. */
1735
	data->o_arg.clientid = clp->cl_clientid;
1736 1737 1738 1739
	switch (data->o_arg.claim) {
	case NFS4_OPEN_CLAIM_PREVIOUS:
	case NFS4_OPEN_CLAIM_DELEG_CUR_FH:
	case NFS4_OPEN_CLAIM_DELEG_PREV_FH:
1740
		data->o_arg.open_bitmap = &nfs4_open_noattr_bitmap[0];
1741 1742
	case NFS4_OPEN_CLAIM_FH:
		task->tk_msg.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_OPEN_NOATTR];
T
Trond Myklebust 已提交
1743 1744
		nfs_copy_fh(&data->o_res.fh, data->o_arg.fh);
	}
1745
	data->timestamp = jiffies;
1746
	if (nfs4_setup_sequence(data->o_arg.server,
1747
				&data->o_arg.seq_args,
1748 1749 1750
				&data->o_res.seq_res,
				task) != 0)
		nfs_release_seqid(data->o_arg.seqid);
1751 1752 1753 1754 1755 1756 1757 1758 1759 1760

	/* 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;
	}
1761
	return;
1762 1763
unlock_no_action:
	rcu_read_unlock();
1764 1765
out_no_action:
	task->tk_action = NULL;
1766
out_wait:
1767
	nfs4_sequence_done(task, &data->o_res.seq_res);
1768 1769
}

1770 1771 1772
static void nfs4_open_done(struct rpc_task *task, void *calldata)
{
	struct nfs4_opendata *data = calldata;
L
Linus Torvalds 已提交
1773

1774
	data->rpc_status = task->tk_status;
1775

1776 1777
	if (!nfs4_sequence_done(task, &data->o_res.seq_res))
		return;
1778

1779
	if (task->tk_status == 0) {
1780 1781
		if (data->o_res.f_attr->valid & NFS_ATTR_FATTR_TYPE) {
			switch (data->o_res.f_attr->mode & S_IFMT) {
1782 1783 1784
			case S_IFREG:
				break;
			case S_IFLNK:
1785
				data->rpc_status = -ELOOP;
1786 1787
				break;
			case S_IFDIR:
1788
				data->rpc_status = -EISDIR;
1789 1790
				break;
			default:
1791
				data->rpc_status = -ENOTDIR;
1792
			}
1793
		}
1794
		renew_lease(data->o_res.server, data->timestamp);
1795 1796
		if (!(data->o_res.rflags & NFS4_OPEN_RESULT_CONFIRM))
			nfs_confirm_seqid(&data->owner->so_seqid, 0);
1797
	}
1798
	data->rpc_done = 1;
1799
}
1800

1801 1802 1803 1804 1805 1806 1807 1808 1809
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! */
1810
	if (data->rpc_status != 0 || !data->rpc_done)
1811 1812 1813 1814 1815
		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);
1816
	if (!IS_ERR(state))
1817
		nfs4_close_state(state, data->o_arg.fmode);
1818
out_free:
1819
	nfs4_opendata_put(data);
1820 1821 1822 1823 1824 1825 1826 1827
}

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,
};

1828
static int nfs4_run_open_task(struct nfs4_opendata *data, int isrecover)
1829 1830 1831 1832 1833 1834
{
	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;
1835 1836 1837 1838 1839 1840
	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
1841 1842
	struct rpc_task_setup task_setup_data = {
		.rpc_client = server->client,
1843
		.rpc_message = &msg,
T
Trond Myklebust 已提交
1844 1845
		.callback_ops = &nfs4_open_ops,
		.callback_data = data,
1846
		.workqueue = nfsiod_workqueue,
T
Trond Myklebust 已提交
1847 1848
		.flags = RPC_TASK_ASYNC,
	};
1849 1850
	int status;

1851
	nfs4_init_sequence(&o_arg->seq_args, &o_res->seq_res, 1);
1852
	kref_get(&data->kref);
1853 1854
	data->rpc_done = 0;
	data->rpc_status = 0;
1855
	data->cancelled = 0;
1856 1857
	data->is_recover = 0;
	if (isrecover) {
1858
		nfs4_set_sequence_privileged(&o_arg->seq_args);
1859 1860
		data->is_recover = 1;
	}
T
Trond Myklebust 已提交
1861
	task = rpc_run_task(&task_setup_data);
1862 1863 1864 1865 1866 1867 1868 1869 1870 1871 1872 1873 1874 1875 1876 1877 1878 1879 1880 1881 1882 1883 1884
        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;

1885 1886
	nfs_fattr_map_and_free_names(NFS_SERVER(dir), &data->f_attr);

1887 1888 1889 1890 1891 1892 1893 1894 1895
	if (o_res->rflags & NFS4_OPEN_RESULT_CONFIRM) {
		status = _nfs4_proc_open_confirm(data);
		if (status != 0)
			return status;
	}

	return status;
}

1896 1897
static int nfs4_opendata_access(struct rpc_cred *cred,
				struct nfs4_opendata *opendata,
1898 1899
				struct nfs4_state *state, fmode_t fmode,
				int openflags)
1900 1901 1902 1903 1904 1905 1906 1907 1908 1909
{
	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;
1910
	/* don't check MAY_WRITE - a newly created file may not have
1911 1912 1913 1914 1915 1916 1917 1918
	 * 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;
1919 1920 1921 1922 1923 1924

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

1925
	if ((mask & ~cache.mask & (MAY_READ | MAY_EXEC)) == 0)
1926 1927 1928 1929
		return 0;

	/* even though OPEN succeeded, access is denied. Close the file */
	nfs4_close_state(state, fmode);
1930
	return -EACCES;
1931 1932
}

1933 1934 1935 1936 1937 1938 1939 1940 1941 1942 1943 1944
/*
 * 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);
1945 1946 1947 1948 1949 1950
	if (!data->rpc_done)
		return status;
	if (status != 0) {
		if (status == -NFS4ERR_BADNAME &&
				!(o_arg->open_flags & O_CREAT))
			return -ENOENT;
1951
		return status;
1952
	}
1953

1954 1955
	nfs_fattr_map_and_free_names(server, &data->f_attr);

1956
	if (o_arg->open_flags & O_CREAT) {
1957
		update_changeattr(dir, &o_res->cinfo);
1958 1959 1960 1961 1962
		if (o_arg->open_flags & O_EXCL)
			data->file_created = 1;
		else if (o_res->cinfo.before != o_res->cinfo.after)
			data->file_created = 1;
	}
T
Trond Myklebust 已提交
1963 1964
	if ((o_res->rflags & NFS4_OPEN_RESULT_LOCKTYPE_POSIX) == 0)
		server->caps &= ~NFS_CAP_POSIX_LOCK;
L
Linus Torvalds 已提交
1965
	if(o_res->rflags & NFS4_OPEN_RESULT_CONFIRM) {
1966
		status = _nfs4_proc_open_confirm(data);
L
Linus Torvalds 已提交
1967
		if (status != 0)
1968
			return status;
L
Linus Torvalds 已提交
1969 1970
	}
	if (!(o_res->f_attr->valid & NFS_ATTR_FATTR))
1971
		_nfs4_proc_getattr(server, &o_res->fh, o_res->f_attr, o_res->f_label);
1972
	return 0;
L
Linus Torvalds 已提交
1973 1974
}

A
Andy Adamson 已提交
1975 1976 1977 1978 1979
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
1980 1981 1982 1983 1984
/*
 * OPEN_EXPIRED:
 * 	reclaim state on the server after a network partition.
 * 	Assumes caller holds the appropriate lock
 */
1985
static int _nfs4_open_expired(struct nfs_open_context *ctx, struct nfs4_state *state)
L
Linus Torvalds 已提交
1986
{
1987
	struct nfs4_opendata *opendata;
1988
	int ret;
L
Linus Torvalds 已提交
1989

1990
	opendata = nfs4_open_recoverdata_alloc(ctx, state,
1991
			NFS4_OPEN_CLAIM_FH);
T
Trond Myklebust 已提交
1992 1993
	if (IS_ERR(opendata))
		return PTR_ERR(opendata);
1994
	ret = nfs4_open_recover(opendata, state);
1995
	if (ret == -ESTALE)
1996
		d_drop(ctx->dentry);
1997
	nfs4_opendata_put(opendata);
1998
	return ret;
L
Linus Torvalds 已提交
1999 2000
}

2001
static int nfs4_do_open_expired(struct nfs_open_context *ctx, struct nfs4_state *state)
2002
{
2003
	struct nfs_server *server = NFS_SERVER(state->inode);
2004 2005 2006 2007
	struct nfs4_exception exception = { };
	int err;

	do {
2008
		err = _nfs4_open_expired(ctx, state);
2009
		trace_nfs4_open_expired(ctx, 0, err);
2010 2011
		if (nfs4_clear_cap_atomic_open_v1(server, err, &exception))
			continue;
2012 2013 2014 2015 2016 2017 2018 2019
		switch (err) {
		default:
			goto out;
		case -NFS4ERR_GRACE:
		case -NFS4ERR_DELAY:
			nfs4_handle_exception(server, err, &exception);
			err = 0;
		}
2020
	} while (exception.retry);
2021
out:
2022 2023 2024
	return err;
}

L
Linus Torvalds 已提交
2025 2026 2027
static int nfs4_open_expired(struct nfs4_state_owner *sp, struct nfs4_state *state)
{
	struct nfs_open_context *ctx;
2028
	int ret;
L
Linus Torvalds 已提交
2029

2030 2031
	ctx = nfs4_state_find_open_context(state);
	if (IS_ERR(ctx))
2032
		return -EAGAIN;
2033
	ret = nfs4_do_open_expired(ctx, state);
2034 2035
	put_nfs_open_context(ctx);
	return ret;
L
Linus Torvalds 已提交
2036 2037
}

2038
#if defined(CONFIG_NFS_V4_1)
2039
static void nfs41_clear_delegation_stateid(struct nfs4_state *state)
2040 2041
{
	struct nfs_server *server = NFS_SERVER(state->inode);
2042
	nfs4_stateid *stateid = &state->stateid;
2043 2044 2045
	struct nfs_delegation *delegation;
	struct rpc_cred *cred = NULL;
	int status = -NFS4ERR_BAD_STATEID;
2046 2047 2048 2049 2050

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

2051 2052 2053 2054 2055 2056 2057 2058
	/* 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);
2059
		trace_nfs4_test_delegation_stateid(state, NULL, status);
2060 2061 2062
	} else
		rcu_read_unlock();

2063 2064 2065 2066
	if (status != NFS_OK) {
		/* Free the stateid unless the server explicitly
		 * informs us the stateid is unrecognized. */
		if (status != -NFS4ERR_BAD_STATEID)
2067
			nfs41_free_stateid(server, stateid, cred);
2068
		nfs_remove_bad_delegation(state->inode);
2069

2070 2071 2072
		write_seqlock(&state->seqlock);
		nfs4_stateid_copy(&state->stateid, &state->open_stateid);
		write_sequnlock(&state->seqlock);
2073 2074
		clear_bit(NFS_DELEGATED_STATE, &state->flags);
	}
2075 2076 2077

	if (cred != NULL)
		put_rpccred(cred);
2078 2079 2080 2081 2082 2083 2084 2085 2086 2087 2088 2089 2090
}

/**
 * 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);
2091
	nfs4_stateid *stateid = &state->open_stateid;
2092
	struct rpc_cred *cred = state->owner->so_cred;
2093 2094 2095 2096 2097 2098 2099 2100
	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;

2101
	status = nfs41_test_stateid(server, stateid, cred);
2102
	trace_nfs4_test_open_stateid(state, NULL, status);
2103 2104 2105 2106
	if (status != NFS_OK) {
		/* Free the stateid unless the server explicitly
		 * informs us the stateid is unrecognized. */
		if (status != -NFS4ERR_BAD_STATEID)
2107
			nfs41_free_stateid(server, stateid, cred);
2108 2109 2110 2111

		clear_bit(NFS_O_RDONLY_STATE, &state->flags);
		clear_bit(NFS_O_WRONLY_STATE, &state->flags);
		clear_bit(NFS_O_RDWR_STATE, &state->flags);
2112
		clear_bit(NFS_OPEN_STATE, &state->flags);
2113 2114 2115 2116 2117 2118
	}
	return status;
}

static int nfs41_open_expired(struct nfs4_state_owner *sp, struct nfs4_state *state)
{
2119
	int status;
2120

2121 2122
	nfs41_clear_delegation_stateid(state);
	status = nfs41_check_open_stateid(state);
2123 2124 2125
	if (status != NFS_OK)
		status = nfs4_open_expired(sp, state);
	return status;
2126 2127 2128
}
#endif

2129 2130 2131 2132 2133 2134 2135 2136 2137 2138 2139 2140 2141 2142 2143 2144
/*
 * 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;
}

2145 2146 2147
static int _nfs4_open_and_get_state(struct nfs4_opendata *opendata,
		fmode_t fmode,
		int flags,
2148
		struct nfs_open_context *ctx)
2149 2150 2151
{
	struct nfs4_state_owner *sp = opendata->owner;
	struct nfs_server *server = sp->so_server;
2152
	struct dentry *dentry;
2153 2154 2155 2156 2157 2158 2159 2160 2161 2162 2163 2164 2165 2166 2167 2168 2169
	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);

2170 2171 2172 2173 2174 2175 2176 2177 2178 2179 2180 2181 2182 2183 2184
	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));
	}

2185 2186 2187 2188
	ret = nfs4_opendata_access(sp->so_cred, opendata, state, fmode, flags);
	if (ret != 0)
		goto out;

2189
	ctx->state = state;
2190 2191 2192 2193 2194
	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);
	}
2195 2196 2197 2198
out:
	return ret;
}

L
Linus Torvalds 已提交
2199
/*
2200
 * Returns a referenced nfs4_state
L
Linus Torvalds 已提交
2201
 */
2202
static int _nfs4_do_open(struct inode *dir,
2203
			struct nfs_open_context *ctx,
2204 2205
			int flags,
			struct iattr *sattr,
2206 2207
			struct nfs4_label *label,
			int *opened)
L
Linus Torvalds 已提交
2208 2209 2210 2211
{
	struct nfs4_state_owner  *sp;
	struct nfs4_state     *state = NULL;
	struct nfs_server       *server = NFS_SERVER(dir);
2212
	struct nfs4_opendata *opendata;
2213 2214 2215 2216
	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);
2217
	enum open_claim_type4 claim = NFS4_OPEN_CLAIM_NULL;
2218
	struct nfs4_label *olabel = NULL;
2219
	int status;
L
Linus Torvalds 已提交
2220 2221 2222

	/* Protect against reboot recovery conflicts */
	status = -ENOMEM;
2223 2224
	sp = nfs4_get_state_owner(server, cred, GFP_KERNEL);
	if (sp == NULL) {
L
Linus Torvalds 已提交
2225 2226 2227
		dprintk("nfs4_do_open: nfs4_get_state_owner failed!\n");
		goto out_err;
	}
2228 2229
	status = nfs4_recover_expired_lease(server);
	if (status != 0)
2230
		goto err_put_state_owner;
2231 2232
	if (dentry->d_inode != NULL)
		nfs4_return_incompatible_delegation(dentry->d_inode, fmode);
2233
	status = -ENOMEM;
2234 2235
	if (dentry->d_inode)
		claim = NFS4_OPEN_CLAIM_FH;
2236
	opendata = nfs4_opendata_alloc(dentry, sp, fmode, flags, sattr,
2237
			label, claim, GFP_KERNEL);
2238
	if (opendata == NULL)
T
Trond Myklebust 已提交
2239
		goto err_put_state_owner;
L
Linus Torvalds 已提交
2240

2241 2242 2243 2244 2245 2246 2247 2248
	if (label) {
		olabel = nfs4_label_alloc(server, GFP_KERNEL);
		if (IS_ERR(olabel)) {
			status = PTR_ERR(olabel);
			goto err_opendata_put;
		}
	}

2249 2250 2251
	if (ctx_th && server->attr_bitmask[2] & FATTR4_WORD2_MDSTHRESHOLD) {
		opendata->f_attr.mdsthreshold = pnfs_mdsthreshold_alloc();
		if (!opendata->f_attr.mdsthreshold)
2252
			goto err_free_label;
2253
		opendata->o_arg.open_bitmap = &nfs4_pnfs_open_bitmap[0];
2254
	}
2255 2256
	if (dentry->d_inode != NULL)
		opendata->state = nfs4_get_open_state(dentry->d_inode, sp);
2257

2258
	status = _nfs4_open_and_get_state(opendata, fmode, flags, ctx);
2259
	if (status != 0)
2260
		goto err_free_label;
2261
	state = ctx->state;
2262

2263 2264
	if ((opendata->o_arg.open_flags & O_EXCL) &&
	    (opendata->o_arg.createmode != NFS4_CREATE_GUARDED)) {
2265 2266 2267 2268 2269
		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,
2270 2271
				state, label, olabel);
		if (status == 0) {
2272
			nfs_setattr_update_inode(state->inode, sattr);
2273
			nfs_post_op_update_inode(state->inode, opendata->o_res.f_attr);
2274
			nfs_setsecurity(state->inode, opendata->o_res.f_attr, olabel);
2275
		}
2276
	}
2277 2278
	if (opendata->file_created)
		*opened |= FILE_CREATED;
2279 2280 2281 2282 2283 2284 2285

	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;

2286 2287
	nfs4_label_free(olabel);

2288
	nfs4_opendata_put(opendata);
L
Linus Torvalds 已提交
2289 2290
	nfs4_put_state_owner(sp);
	return 0;
2291 2292
err_free_label:
	nfs4_label_free(olabel);
2293
err_opendata_put:
2294
	kfree(opendata->f_attr.mdsthreshold);
2295
	nfs4_opendata_put(opendata);
2296 2297
err_put_state_owner:
	nfs4_put_state_owner(sp);
L
Linus Torvalds 已提交
2298 2299 2300 2301 2302
out_err:
	return status;
}


2303
static struct nfs4_state *nfs4_do_open(struct inode *dir,
2304
					struct nfs_open_context *ctx,
2305 2306
					int flags,
					struct iattr *sattr,
2307 2308
					struct nfs4_label *label,
					int *opened)
L
Linus Torvalds 已提交
2309
{
2310
	struct nfs_server *server = NFS_SERVER(dir);
L
Linus Torvalds 已提交
2311 2312 2313 2314 2315
	struct nfs4_exception exception = { };
	struct nfs4_state *res;
	int status;

	do {
2316
		status = _nfs4_do_open(dir, ctx, flags, sattr, label, opened);
2317
		res = ctx->state;
2318
		trace_nfs4_open_file(ctx, flags, status);
L
Linus Torvalds 已提交
2319 2320 2321 2322 2323 2324 2325 2326
		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
2327
		 * doing an OPEN, we assume that nfs_increment_open_seqid() will
L
Linus Torvalds 已提交
2328 2329 2330 2331 2332
		 * 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) {
2333
			pr_warn_ratelimited("NFS: v4 server %s "
2334 2335
					" returned a bad sequence-id error!\n",
					NFS_SERVER(dir)->nfs_client->cl_hostname);
L
Linus Torvalds 已提交
2336 2337 2338
			exception.retry = 1;
			continue;
		}
2339 2340 2341 2342 2343 2344 2345 2346 2347 2348
		/*
		 * 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;
		}
2349 2350 2351 2352 2353
		if (status == -EAGAIN) {
			/* We must have found a delegation */
			exception.retry = 1;
			continue;
		}
2354 2355 2356
		if (nfs4_clear_cap_atomic_open_v1(server, status, &exception))
			continue;
		res = ERR_PTR(nfs4_handle_exception(server,
L
Linus Torvalds 已提交
2357 2358 2359 2360 2361
					status, &exception));
	} while (exception.retry);
	return res;
}

2362 2363
static int _nfs4_do_setattr(struct inode *inode, struct rpc_cred *cred,
			    struct nfs_fattr *fattr, struct iattr *sattr,
2364 2365
			    struct nfs4_state *state, struct nfs4_label *ilabel,
			    struct nfs4_label *olabel)
L
Linus Torvalds 已提交
2366
{
2367
	struct nfs_server *server = NFS_SERVER(inode);
L
Linus Torvalds 已提交
2368
        struct nfs_setattrargs  arg = {
2369
                .fh             = NFS_FH(inode),
L
Linus Torvalds 已提交
2370 2371 2372
                .iap            = sattr,
		.server		= server,
		.bitmask = server->attr_bitmask,
2373
		.label		= ilabel,
L
Linus Torvalds 已提交
2374 2375 2376
        };
        struct nfs_setattrres  res = {
		.fattr		= fattr,
2377
		.label		= olabel,
L
Linus Torvalds 已提交
2378 2379 2380
		.server		= server,
        };
        struct rpc_message msg = {
2381 2382 2383 2384
		.rpc_proc	= &nfs4_procedures[NFSPROC4_CLNT_SETATTR],
		.rpc_argp	= &arg,
		.rpc_resp	= &res,
		.rpc_cred	= cred,
L
Linus Torvalds 已提交
2385
        };
2386
	unsigned long timestamp = jiffies;
2387 2388
	fmode_t fmode;
	bool truncate;
2389
	int status;
L
Linus Torvalds 已提交
2390

2391 2392 2393 2394
	arg.bitmask = nfs4_bitmask(server, ilabel);
	if (ilabel)
		arg.bitmask = nfs4_bitmask(server, olabel);

2395
	nfs_fattr_init(fattr);
L
Linus Torvalds 已提交
2396

2397 2398 2399 2400 2401 2402 2403
	/* 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)) {
2404 2405 2406 2407
		struct nfs_lockowner lockowner = {
			.l_owner = current->files,
			.l_pid = current->tgid,
		};
2408
		nfs4_select_rw_stateid(&arg.stateid, state, FMODE_WRITE,
2409
				&lockowner);
2410
	} else
2411
		nfs4_stateid_copy(&arg.stateid, &zero_stateid);
L
Linus Torvalds 已提交
2412

2413
	status = nfs4_call_sync(server->client, server, &msg, &arg.seq_args, &res.seq_res, 1);
2414 2415
	if (status == 0 && state != NULL)
		renew_lease(server, timestamp);
2416
	return status;
L
Linus Torvalds 已提交
2417 2418
}

2419 2420
static int nfs4_do_setattr(struct inode *inode, struct rpc_cred *cred,
			   struct nfs_fattr *fattr, struct iattr *sattr,
2421 2422
			   struct nfs4_state *state, struct nfs4_label *ilabel,
			   struct nfs4_label *olabel)
L
Linus Torvalds 已提交
2423
{
2424
	struct nfs_server *server = NFS_SERVER(inode);
2425 2426
	struct nfs4_exception exception = {
		.state = state,
2427
		.inode = inode,
2428
	};
L
Linus Torvalds 已提交
2429 2430
	int err;
	do {
2431
		err = _nfs4_do_setattr(inode, cred, fattr, sattr, state, ilabel, olabel);
2432
		trace_nfs4_setattr(inode, err);
2433 2434
		switch (err) {
		case -NFS4ERR_OPENMODE:
2435 2436 2437 2438 2439 2440 2441
			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);
			}
2442 2443 2444 2445 2446 2447 2448 2449
			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
2450
	} while (exception.retry);
2451
out:
L
Linus Torvalds 已提交
2452 2453 2454 2455 2456 2457 2458 2459
	return err;
}

struct nfs4_closedata {
	struct inode *inode;
	struct nfs4_state *state;
	struct nfs_closeargs arg;
	struct nfs_closeres res;
2460
	struct nfs_fattr fattr;
2461
	unsigned long timestamp;
F
Fred Isaman 已提交
2462 2463
	bool roc;
	u32 roc_barrier;
L
Linus Torvalds 已提交
2464 2465
};

2466
static void nfs4_free_closedata(void *data)
2467
{
2468 2469
	struct nfs4_closedata *calldata = data;
	struct nfs4_state_owner *sp = calldata->state->owner;
2470
	struct super_block *sb = calldata->state->inode->i_sb;
2471

F
Fred Isaman 已提交
2472 2473
	if (calldata->roc)
		pnfs_roc_release(calldata->state->inode);
2474 2475 2476
	nfs4_put_open_state(calldata->state);
	nfs_free_seqid(calldata->arg.seqid);
	nfs4_put_state_owner(sp);
2477
	nfs_sb_deactive(sb);
2478 2479 2480
	kfree(calldata);
}

2481 2482 2483 2484
static void nfs4_close_clear_stateid_flags(struct nfs4_state *state,
		fmode_t fmode)
{
	spin_lock(&state->owner->so_lock);
2485 2486 2487
	clear_bit(NFS_O_RDWR_STATE, &state->flags);
	switch (fmode & (FMODE_READ|FMODE_WRITE)) {
	case FMODE_WRITE:
2488
		clear_bit(NFS_O_RDONLY_STATE, &state->flags);
2489 2490
		break;
	case FMODE_READ:
2491
		clear_bit(NFS_O_WRONLY_STATE, &state->flags);
2492 2493 2494 2495 2496 2497
		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);
	}
2498 2499 2500
	spin_unlock(&state->owner->so_lock);
}

2501
static void nfs4_close_done(struct rpc_task *task, void *data)
L
Linus Torvalds 已提交
2502
{
2503
	struct nfs4_closedata *calldata = data;
L
Linus Torvalds 已提交
2504 2505 2506
	struct nfs4_state *state = calldata->state;
	struct nfs_server *server = NFS_SERVER(calldata->inode);

2507
	dprintk("%s: begin!\n", __func__);
2508 2509
	if (!nfs4_sequence_done(task, &calldata->res.seq_res))
		return;
2510
	trace_nfs4_close(state, &calldata->arg, &calldata->res, task->tk_status);
L
Linus Torvalds 已提交
2511 2512 2513 2514 2515
        /* 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
2516 2517 2518
			if (calldata->roc)
				pnfs_roc_set_barrier(state->inode,
						     calldata->roc_barrier);
2519
			nfs_set_open_stateid(state, &calldata->res.stateid, 0);
2520
			renew_lease(server, calldata->timestamp);
2521 2522
			nfs4_close_clear_stateid_flags(state,
					calldata->arg.fmode);
L
Linus Torvalds 已提交
2523 2524
			break;
		case -NFS4ERR_STALE_STATEID:
2525 2526
		case -NFS4ERR_OLD_STATEID:
		case -NFS4ERR_BAD_STATEID:
L
Linus Torvalds 已提交
2527
		case -NFS4ERR_EXPIRED:
2528
			if (calldata->arg.fmode == 0)
2529
				break;
L
Linus Torvalds 已提交
2530
		default:
2531 2532
			if (nfs4_async_handle_error(task, server, state) == -EAGAIN)
				rpc_restart_call_prepare(task);
L
Linus Torvalds 已提交
2533
	}
2534
	nfs_release_seqid(calldata->arg.seqid);
2535
	nfs_refresh_inode(calldata->inode, calldata->res.fattr);
2536
	dprintk("%s: done, ret = %d!\n", __func__, task->tk_status);
L
Linus Torvalds 已提交
2537 2538
}

T
Trond Myklebust 已提交
2539
static void nfs4_close_prepare(struct rpc_task *task, void *data)
L
Linus Torvalds 已提交
2540
{
T
Trond Myklebust 已提交
2541
	struct nfs4_closedata *calldata = data;
2542
	struct nfs4_state *state = calldata->state;
2543
	struct inode *inode = calldata->inode;
2544
	int call_close = 0;
2545

2546
	dprintk("%s: begin!\n", __func__);
2547
	if (nfs_wait_on_sequence(calldata->arg.seqid, task) != 0)
2548
		goto out_wait;
2549

2550 2551
	task->tk_msg.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_OPEN_DOWNGRADE];
	calldata->arg.fmode = FMODE_READ|FMODE_WRITE;
2552
	spin_lock(&state->owner->so_lock);
2553
	/* Calculate the change in open mode */
2554
	if (state->n_rdwr == 0) {
2555
		if (state->n_rdonly == 0) {
2556 2557 2558
			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;
2559 2560
		}
		if (state->n_wronly == 0) {
2561 2562 2563
			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;
2564
		}
2565
	}
2566 2567
	if (!nfs4_valid_open_stateid(state))
		call_close = 0;
2568
	spin_unlock(&state->owner->so_lock);
2569 2570

	if (!call_close) {
2571
		/* Note: exit _without_ calling nfs4_close_done */
2572
		goto out_no_action;
2573
	}
2574

F
Fred Isaman 已提交
2575
	if (calldata->arg.fmode == 0) {
2576
		task->tk_msg.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_CLOSE];
F
Fred Isaman 已提交
2577
		if (calldata->roc &&
2578 2579
		    pnfs_roc_drain(inode, &calldata->roc_barrier, task)) {
			nfs_release_seqid(calldata->arg.seqid);
2580
			goto out_wait;
2581
		    }
F
Fred Isaman 已提交
2582
	}
2583

2584
	nfs_fattr_init(calldata->res.fattr);
2585
	calldata->timestamp = jiffies;
2586
	if (nfs4_setup_sequence(NFS_SERVER(inode),
2587 2588
				&calldata->arg.seq_args,
				&calldata->res.seq_res,
2589 2590
				task) != 0)
		nfs_release_seqid(calldata->arg.seqid);
2591
	dprintk("%s: done!\n", __func__);
2592 2593 2594 2595 2596
	return;
out_no_action:
	task->tk_action = NULL;
out_wait:
	nfs4_sequence_done(task, &calldata->res.seq_res);
L
Linus Torvalds 已提交
2597 2598
}

2599
static const struct rpc_call_ops nfs4_close_ops = {
T
Trond Myklebust 已提交
2600
	.rpc_call_prepare = nfs4_close_prepare,
2601 2602 2603 2604
	.rpc_call_done = nfs4_close_done,
	.rpc_release = nfs4_free_closedata,
};

L
Linus Torvalds 已提交
2605 2606 2607 2608 2609 2610 2611 2612 2613 2614 2615
/* 
 * 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!
 */
2616
int nfs4_do_close(struct nfs4_state *state, gfp_t gfp_mask, int wait)
L
Linus Torvalds 已提交
2617
{
2618
	struct nfs_server *server = NFS_SERVER(state->inode);
L
Linus Torvalds 已提交
2619
	struct nfs4_closedata *calldata;
2620 2621
	struct nfs4_state_owner *sp = state->owner;
	struct rpc_task *task;
2622 2623 2624 2625
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_CLOSE],
		.rpc_cred = state->owner->so_cred,
	};
T
Trond Myklebust 已提交
2626 2627
	struct rpc_task_setup task_setup_data = {
		.rpc_client = server->client,
2628
		.rpc_message = &msg,
T
Trond Myklebust 已提交
2629
		.callback_ops = &nfs4_close_ops,
2630
		.workqueue = nfsiod_workqueue,
T
Trond Myklebust 已提交
2631 2632
		.flags = RPC_TASK_ASYNC,
	};
2633
	int status = -ENOMEM;
L
Linus Torvalds 已提交
2634

2635 2636 2637
	nfs4_state_protect(server->nfs_client, NFS_SP4_MACH_CRED_CLEANUP,
		&task_setup_data.rpc_client, &msg);

2638
	calldata = kzalloc(sizeof(*calldata), gfp_mask);
L
Linus Torvalds 已提交
2639
	if (calldata == NULL)
2640
		goto out;
2641
	nfs4_init_sequence(&calldata->arg.seq_args, &calldata->res.seq_res, 1);
2642
	calldata->inode = state->inode;
L
Linus Torvalds 已提交
2643
	calldata->state = state;
2644
	calldata->arg.fh = NFS_FH(state->inode);
2645
	calldata->arg.stateid = &state->open_stateid;
L
Linus Torvalds 已提交
2646
	/* Serialization for the sequence id */
2647
	calldata->arg.seqid = nfs_alloc_seqid(&state->owner->so_seqid, gfp_mask);
2648 2649
	if (calldata->arg.seqid == NULL)
		goto out_free_calldata;
2650
	calldata->arg.fmode = 0;
2651
	calldata->arg.bitmask = server->cache_consistency_bitmask;
2652
	calldata->res.fattr = &calldata->fattr;
2653
	calldata->res.seqid = calldata->arg.seqid;
2654
	calldata->res.server = server;
2655
	calldata->roc = pnfs_roc(state->inode);
2656
	nfs_sb_active(calldata->inode->i_sb);
2657

2658 2659
	msg.rpc_argp = &calldata->arg;
	msg.rpc_resp = &calldata->res;
T
Trond Myklebust 已提交
2660 2661
	task_setup_data.callback_data = calldata;
	task = rpc_run_task(&task_setup_data);
2662 2663
	if (IS_ERR(task))
		return PTR_ERR(task);
2664 2665 2666
	status = 0;
	if (wait)
		status = rpc_wait_for_completion_task(task);
2667
	rpc_put_task(task);
2668
	return status;
2669 2670 2671
out_free_calldata:
	kfree(calldata);
out:
2672 2673
	nfs4_put_open_state(state);
	nfs4_put_state_owner(sp);
2674
	return status;
L
Linus Torvalds 已提交
2675 2676
}

2677
static struct inode *
2678 2679
nfs4_atomic_open(struct inode *dir, struct nfs_open_context *ctx,
		int open_flags, struct iattr *attr, int *opened)
L
Linus Torvalds 已提交
2680 2681
{
	struct nfs4_state *state;
2682 2683 2684
	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
2685

T
Trond Myklebust 已提交
2686
	/* Protect against concurrent sillydeletes */
2687
	state = nfs4_do_open(dir, ctx, open_flags, attr, label, opened);
2688 2689 2690

	nfs4_label_release_security(label);

2691 2692
	if (IS_ERR(state))
		return ERR_CAST(state);
2693
	return state->inode;
L
Linus Torvalds 已提交
2694 2695
}

2696
static void nfs4_close_context(struct nfs_open_context *ctx, int is_sync)
T
Trond Myklebust 已提交
2697 2698 2699 2700
{
	if (ctx->state == NULL)
		return;
	if (is_sync)
2701
		nfs4_close_sync(ctx->state, ctx->mode);
T
Trond Myklebust 已提交
2702
	else
2703
		nfs4_close_state(ctx->state, ctx->mode);
T
Trond Myklebust 已提交
2704
}
L
Linus Torvalds 已提交
2705

2706 2707 2708 2709
#define FATTR4_WORD1_NFS40_MASK (2*FATTR4_WORD1_MOUNTED_ON_FILEID - 1UL)
#define FATTR4_WORD2_NFS41_MASK (2*FATTR4_WORD2_SUPPATTR_EXCLCREAT - 1UL)
#define FATTR4_WORD2_NFS42_MASK (2*FATTR4_WORD2_CHANGE_SECURITY_LABEL - 1UL)

L
Linus Torvalds 已提交
2710 2711
static int _nfs4_server_capabilities(struct nfs_server *server, struct nfs_fh *fhandle)
{
B
Benny Halevy 已提交
2712 2713 2714
	struct nfs4_server_caps_arg args = {
		.fhandle = fhandle,
	};
L
Linus Torvalds 已提交
2715 2716 2717
	struct nfs4_server_caps_res res = {};
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_SERVER_CAPS],
B
Benny Halevy 已提交
2718
		.rpc_argp = &args,
L
Linus Torvalds 已提交
2719 2720 2721 2722
		.rpc_resp = &res,
	};
	int status;

2723
	status = nfs4_call_sync(server->client, server, &msg, &args.seq_args, &res.seq_res, 0);
L
Linus Torvalds 已提交
2724
	if (status == 0) {
2725 2726 2727 2728 2729 2730 2731 2732 2733 2734 2735 2736
		/* Sanity check the server answers */
		switch (server->nfs_client->cl_minorversion) {
		case 0:
			res.attr_bitmask[1] &= FATTR4_WORD1_NFS40_MASK;
			res.attr_bitmask[2] = 0;
			break;
		case 1:
			res.attr_bitmask[2] &= FATTR4_WORD2_NFS41_MASK;
			break;
		case 2:
			res.attr_bitmask[2] &= FATTR4_WORD2_NFS42_MASK;
		}
L
Linus Torvalds 已提交
2737
		memcpy(server->attr_bitmask, res.attr_bitmask, sizeof(server->attr_bitmask));
2738 2739 2740 2741
		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|
2742 2743
				NFS_CAP_CTIME|NFS_CAP_MTIME|
				NFS_CAP_SECURITY_LABEL);
L
Linus Torvalds 已提交
2744 2745 2746 2747 2748 2749
		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;
2750 2751 2752 2753 2754 2755 2756 2757 2758 2759 2760 2761 2762 2763 2764 2765
		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;
2766 2767 2768 2769 2770 2771
#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));
2772
		server->attr_bitmask_nl[2] &= ~FATTR4_WORD2_SECURITY_LABEL;
2773

2774 2775 2776
		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;
2777
		server->cache_consistency_bitmask[2] = 0;
L
Linus Torvalds 已提交
2778
		server->acl_bitmask = res.acl_bitmask;
2779
		server->fh_expire_type = res.fh_expire_type;
L
Linus Torvalds 已提交
2780
	}
A
Andy Adamson 已提交
2781

L
Linus Torvalds 已提交
2782 2783 2784
	return status;
}

2785
int nfs4_server_capabilities(struct nfs_server *server, struct nfs_fh *fhandle)
L
Linus Torvalds 已提交
2786 2787 2788 2789 2790 2791 2792 2793 2794 2795 2796 2797 2798 2799
{
	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)
{
2800
	u32 bitmask[3];
L
Linus Torvalds 已提交
2801
	struct nfs4_lookup_root_arg args = {
2802
		.bitmask = bitmask,
L
Linus Torvalds 已提交
2803 2804 2805
	};
	struct nfs4_lookup_res res = {
		.server = server,
2806
		.fattr = info->fattr,
L
Linus Torvalds 已提交
2807 2808 2809 2810 2811 2812 2813
		.fh = fhandle,
	};
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_LOOKUP_ROOT],
		.rpc_argp = &args,
		.rpc_resp = &res,
	};
2814

2815 2816 2817 2818 2819 2820 2821
	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;

2822
	nfs_fattr_init(info->fattr);
2823
	return nfs4_call_sync(server->client, server, &msg, &args.seq_args, &res.seq_res, 0);
L
Linus Torvalds 已提交
2824 2825 2826 2827 2828 2829 2830 2831
}

static int nfs4_lookup_root(struct nfs_server *server, struct nfs_fh *fhandle,
		struct nfs_fsinfo *info)
{
	struct nfs4_exception exception = { };
	int err;
	do {
2832
		err = _nfs4_lookup_root(server, fhandle, info);
2833
		trace_nfs4_lookup_root(server, fhandle, info->fattr, err);
2834 2835 2836
		switch (err) {
		case 0:
		case -NFS4ERR_WRONGSEC:
2837
			goto out;
2838 2839 2840
		default:
			err = nfs4_handle_exception(server, err, &exception);
		}
L
Linus Torvalds 已提交
2841
	} while (exception.retry);
2842
out:
L
Linus Torvalds 已提交
2843 2844 2845
	return err;
}

2846 2847 2848
static int nfs4_lookup_root_sec(struct nfs_server *server, struct nfs_fh *fhandle,
				struct nfs_fsinfo *info, rpc_authflavor_t flavor)
{
2849 2850 2851
	struct rpc_auth_create_args auth_args = {
		.pseudoflavor = flavor,
	};
2852 2853 2854
	struct rpc_auth *auth;
	int ret;

2855
	auth = rpcauth_create(&auth_args, server->client);
2856
	if (IS_ERR(auth)) {
2857
		ret = -EACCES;
2858 2859 2860 2861 2862 2863 2864
		goto out;
	}
	ret = nfs4_lookup_root(server, fhandle, info);
out:
	return ret;
}

2865 2866 2867 2868 2869 2870 2871 2872 2873
/*
 * 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.
 */
2874
static int nfs4_find_root_sec(struct nfs_server *server, struct nfs_fh *fhandle,
2875
			      struct nfs_fsinfo *info)
L
Linus Torvalds 已提交
2876
{
2877 2878 2879 2880 2881
	/* Per 3530bis 15.33.5 */
	static const rpc_authflavor_t flav_array[] = {
		RPC_AUTH_GSS_KRB5P,
		RPC_AUTH_GSS_KRB5I,
		RPC_AUTH_GSS_KRB5,
2882
		RPC_AUTH_UNIX,			/* courtesy */
2883 2884 2885 2886
		RPC_AUTH_NULL,
	};
	int status = -EPERM;
	size_t i;
C
Chuck Lever 已提交
2887

2888 2889 2890 2891 2892 2893 2894 2895 2896 2897 2898 2899 2900 2901 2902 2903 2904 2905
	if (server->auth_info.flavor_len > 0) {
		/* try each flavor specified by user */
		for (i = 0; i < server->auth_info.flavor_len; i++) {
			status = nfs4_lookup_root_sec(server, fhandle, info,
						server->auth_info.flavors[i]);
			if (status == -NFS4ERR_WRONGSEC || status == -EACCES)
				continue;
			break;
		}
	} else {
		/* no flavors specified by user, try default list */
		for (i = 0; i < ARRAY_SIZE(flav_array); i++) {
			status = nfs4_lookup_root_sec(server, fhandle, info,
						      flav_array[i]);
			if (status == -NFS4ERR_WRONGSEC || status == -EACCES)
				continue;
			break;
		}
2906
	}
2907

2908 2909 2910 2911 2912 2913 2914 2915 2916
	/*
	 * -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;
2917 2918 2919
	return status;
}

C
Chuck Lever 已提交
2920 2921 2922 2923 2924 2925 2926 2927 2928 2929 2930 2931
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
2932
 * @auth_probe: probe the auth flavours
C
Chuck Lever 已提交
2933 2934
 *
 * Returns zero on success, or a negative errno.
2935
 */
2936
int nfs4_proc_get_rootfh(struct nfs_server *server, struct nfs_fh *fhandle,
2937 2938
			 struct nfs_fsinfo *info,
			 bool auth_probe)
2939
{
C
Chuck Lever 已提交
2940 2941
	int status;

2942 2943 2944 2945 2946 2947
	switch (auth_probe) {
	case false:
		status = nfs4_lookup_root(server, fhandle, info);
		if (status != -NFS4ERR_WRONGSEC)
			break;
	default:
C
Chuck Lever 已提交
2948
		status = nfs4_do_find_root_sec(server, fhandle, info);
2949
	}
C
Chuck Lever 已提交
2950

L
Linus Torvalds 已提交
2951 2952 2953 2954
	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
2955

2956
	return nfs4_map_errors(status);
L
Linus Torvalds 已提交
2957 2958
}

2959 2960 2961 2962 2963
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;
2964
	struct nfs4_label *label = NULL;
2965 2966 2967 2968 2969 2970 2971

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

2972 2973 2974 2975
	label = nfs4_label_alloc(server, GFP_KERNEL);
	if (IS_ERR(label))
		return PTR_ERR(label);

2976
	error = nfs4_proc_getattr(server, mntfh, fattr, label);
2977 2978
	if (error < 0) {
		dprintk("nfs4_get_root: getattr error = %d\n", -error);
2979
		goto err_free_label;
2980 2981 2982 2983 2984 2985
	}

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

2986 2987 2988
err_free_label:
	nfs4_label_free(label);

2989 2990 2991
	return error;
}

M
Manoj Naik 已提交
2992 2993 2994 2995 2996
/*
 * 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
 */
2997 2998 2999
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
3000 3001 3002 3003 3004 3005 3006 3007 3008 3009 3010 3011
{
	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;

3012
	status = nfs4_proc_fs_locations(client, dir, name, locations, page);
M
Manoj Naik 已提交
3013 3014
	if (status != 0)
		goto out;
3015 3016 3017 3018 3019 3020

	/*
	 * If the fsid didn't change, this is a migration event, not a
	 * referral.  Cause us to drop into the exception handler, which
	 * will kick off migration recovery.
	 */
M
Manoj Naik 已提交
3021
	if (nfs_fsid_equal(&NFS_SERVER(dir)->fsid, &locations->fattr.fsid)) {
3022 3023
		dprintk("%s: server did not return a different fsid for"
			" a referral at %s\n", __func__, name->name);
3024
		status = -NFS4ERR_MOVED;
M
Manoj Naik 已提交
3025 3026
		goto out;
	}
3027 3028
	/* Fixup attributes for the nfs_lookup() call to nfs_fhget() */
	nfs_fixup_referral_attributes(&locations->fattr);
M
Manoj Naik 已提交
3029

3030
	/* replace the lookup nfs_fattr with the locations nfs_fattr */
M
Manoj Naik 已提交
3031 3032 3033 3034 3035
	memcpy(fattr, &locations->fattr, sizeof(struct nfs_fattr));
	memset(fhandle, 0, sizeof(struct nfs_fh));
out:
	if (page)
		__free_page(page);
3036
	kfree(locations);
M
Manoj Naik 已提交
3037 3038 3039
	return status;
}

3040 3041
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
3042 3043 3044 3045 3046 3047 3048
{
	struct nfs4_getattr_arg args = {
		.fh = fhandle,
		.bitmask = server->attr_bitmask,
	};
	struct nfs4_getattr_res res = {
		.fattr = fattr,
3049
		.label = label,
L
Linus Torvalds 已提交
3050 3051 3052 3053 3054 3055 3056
		.server = server,
	};
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_GETATTR],
		.rpc_argp = &args,
		.rpc_resp = &res,
	};
3057 3058 3059

	args.bitmask = nfs4_bitmask(server, label);

3060
	nfs_fattr_init(fattr);
3061
	return nfs4_call_sync(server->client, server, &msg, &args.seq_args, &res.seq_res, 0);
L
Linus Torvalds 已提交
3062 3063
}

3064 3065
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
3066 3067 3068 3069
{
	struct nfs4_exception exception = { };
	int err;
	do {
3070 3071 3072
		err = _nfs4_proc_getattr(server, fhandle, fattr, label);
		trace_nfs4_getattr(server, fhandle, fattr, err);
		err = nfs4_handle_exception(server, err,
L
Linus Torvalds 已提交
3073 3074 3075 3076 3077 3078 3079 3080 3081 3082 3083 3084 3085 3086 3087 3088 3089 3090 3091 3092 3093 3094 3095 3096 3097 3098
				&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)
{
3099
	struct inode *inode = dentry->d_inode;
3100
	struct rpc_cred *cred = NULL;
3101
	struct nfs4_state *state = NULL;
3102
	struct nfs4_label *label = NULL;
L
Linus Torvalds 已提交
3103 3104
	int status;

B
Benny Halevy 已提交
3105
	if (pnfs_ld_layoutret_on_setattr(inode))
3106
		pnfs_commit_and_return_layout(inode);
B
Benny Halevy 已提交
3107

3108
	nfs_fattr_init(fattr);
L
Linus Torvalds 已提交
3109
	
3110 3111
	/* Deal with open(O_TRUNC) */
	if (sattr->ia_valid & ATTR_OPEN)
3112
		sattr->ia_valid &= ~(ATTR_MTIME|ATTR_CTIME);
3113 3114

	/* Optimization: if the end result is no change, don't RPC */
3115
	if ((sattr->ia_valid & ~(ATTR_FILE|ATTR_OPEN)) == 0)
3116 3117
		return 0;

3118
	/* Search for an existing open(O_WRITE) file */
3119 3120 3121 3122
	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
3123 3124 3125 3126
		if (ctx) {
			cred = ctx->cred;
			state = ctx->state;
		}
3127
	}
3128

3129 3130 3131 3132 3133
	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);
3134
	if (status == 0) {
3135
		nfs_setattr_update_inode(inode, sattr);
3136 3137
		nfs_setsecurity(inode, fattr, label);
	}
3138
	nfs4_label_free(label);
L
Linus Torvalds 已提交
3139 3140 3141
	return status;
}

3142 3143
static int _nfs4_proc_lookup(struct rpc_clnt *clnt, struct inode *dir,
		const struct qstr *name, struct nfs_fh *fhandle,
3144
		struct nfs_fattr *fattr, struct nfs4_label *label)
D
David Howells 已提交
3145
{
3146
	struct nfs_server *server = NFS_SERVER(dir);
D
David Howells 已提交
3147 3148 3149
	int		       status;
	struct nfs4_lookup_arg args = {
		.bitmask = server->attr_bitmask,
3150
		.dir_fh = NFS_FH(dir),
D
David Howells 已提交
3151 3152 3153 3154 3155
		.name = name,
	};
	struct nfs4_lookup_res res = {
		.server = server,
		.fattr = fattr,
3156
		.label = label,
D
David Howells 已提交
3157 3158 3159 3160 3161 3162 3163 3164
		.fh = fhandle,
	};
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_LOOKUP],
		.rpc_argp = &args,
		.rpc_resp = &res,
	};

3165 3166
	args.bitmask = nfs4_bitmask(server, label);

D
David Howells 已提交
3167 3168
	nfs_fattr_init(fattr);

L
Linus Torvalds 已提交
3169
	dprintk("NFS call  lookup %s\n", name->name);
3170
	status = nfs4_call_sync(clnt, server, &msg, &args.seq_args, &res.seq_res, 0);
L
Linus Torvalds 已提交
3171 3172 3173 3174
	dprintk("NFS reply lookup: %d\n", status);
	return status;
}

3175
static void nfs_fixup_secinfo_attributes(struct nfs_fattr *fattr)
3176 3177
{
	fattr->valid |= NFS_ATTR_FATTR_TYPE | NFS_ATTR_FATTR_MODE |
3178
		NFS_ATTR_FATTR_NLINK | NFS_ATTR_FATTR_MOUNTPOINT;
3179 3180 3181 3182
	fattr->mode = S_IFDIR | S_IRUGO | S_IXUGO;
	fattr->nlink = 2;
}

3183 3184
static int nfs4_proc_lookup_common(struct rpc_clnt **clnt, struct inode *dir,
				   struct qstr *name, struct nfs_fh *fhandle,
3185
				   struct nfs_fattr *fattr, struct nfs4_label *label)
L
Linus Torvalds 已提交
3186 3187
{
	struct nfs4_exception exception = { };
3188
	struct rpc_clnt *client = *clnt;
L
Linus Torvalds 已提交
3189 3190
	int err;
	do {
3191
		err = _nfs4_proc_lookup(client, dir, name, fhandle, fattr, label);
3192
		trace_nfs4_lookup(dir, name, err);
3193
		switch (err) {
3194
		case -NFS4ERR_BADNAME:
3195 3196
			err = -ENOENT;
			goto out;
3197
		case -NFS4ERR_MOVED:
3198
			err = nfs4_get_referral(client, dir, name, fattr, fhandle);
3199
			goto out;
3200
		case -NFS4ERR_WRONGSEC:
3201 3202 3203 3204 3205 3206 3207 3208 3209 3210 3211
			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);
3212
		}
L
Linus Torvalds 已提交
3213
	} while (exception.retry);
3214 3215 3216 3217 3218 3219 3220

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

L
Linus Torvalds 已提交
3221 3222 3223
	return err;
}

3224
static int nfs4_proc_lookup(struct inode *dir, struct qstr *name,
3225 3226
			    struct nfs_fh *fhandle, struct nfs_fattr *fattr,
			    struct nfs4_label *label)
3227 3228 3229 3230
{
	int status;
	struct rpc_clnt *client = NFS_CLIENT(dir);

3231
	status = nfs4_proc_lookup_common(&client, dir, name, fhandle, fattr, label);
3232 3233 3234 3235 3236 3237 3238
	if (client != NFS_CLIENT(dir)) {
		rpc_shutdown_client(client);
		nfs_fixup_secinfo_attributes(fattr);
	}
	return status;
}

3239 3240 3241 3242
struct rpc_clnt *
nfs4_proc_lookup_mountpoint(struct inode *dir, struct qstr *name,
			    struct nfs_fh *fhandle, struct nfs_fattr *fattr)
{
3243
	struct rpc_clnt *client = NFS_CLIENT(dir);
3244 3245
	int status;

3246
	status = nfs4_proc_lookup_common(&client, dir, name, fhandle, fattr, NULL);
3247
	if (status < 0)
3248
		return ERR_PTR(status);
3249
	return (client == NFS_CLIENT(dir)) ? rpc_clone_client(client) : client;
3250 3251
}

L
Linus Torvalds 已提交
3252 3253
static int _nfs4_proc_access(struct inode *inode, struct nfs_access_entry *entry)
{
3254
	struct nfs_server *server = NFS_SERVER(inode);
L
Linus Torvalds 已提交
3255 3256
	struct nfs4_accessargs args = {
		.fh = NFS_FH(inode),
3257
		.bitmask = server->cache_consistency_bitmask,
3258 3259 3260
	};
	struct nfs4_accessres res = {
		.server = server,
L
Linus Torvalds 已提交
3261 3262 3263 3264 3265 3266 3267 3268
	};
	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;
3269
	int status = 0;
L
Linus Torvalds 已提交
3270 3271 3272 3273 3274 3275 3276 3277 3278 3279 3280 3281 3282 3283 3284 3285 3286

	/*
	 * 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;
	}
3287 3288 3289 3290 3291

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

3292
	status = nfs4_call_sync(server->client, server, &msg, &args.seq_args, &res.seq_res, 0);
L
Linus Torvalds 已提交
3293
	if (!status) {
3294
		nfs_access_set_mask(entry, res.access);
3295
		nfs_refresh_inode(inode, res.fattr);
L
Linus Torvalds 已提交
3296
	}
3297
	nfs_free_fattr(res.fattr);
L
Linus Torvalds 已提交
3298 3299 3300 3301 3302 3303 3304 3305
	return status;
}

static int nfs4_proc_access(struct inode *inode, struct nfs_access_entry *entry)
{
	struct nfs4_exception exception = { };
	int err;
	do {
3306 3307 3308
		err = _nfs4_proc_access(inode, entry);
		trace_nfs4_access(inode, err);
		err = nfs4_handle_exception(NFS_SERVER(inode), err,
L
Linus Torvalds 已提交
3309 3310 3311 3312 3313 3314 3315 3316 3317 3318 3319 3320 3321 3322 3323 3324 3325 3326 3327 3328 3329 3330
				&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
3331
 * we get the post-operation mtime and size.
L
Linus Torvalds 已提交
3332 3333 3334 3335 3336 3337 3338 3339 3340 3341 3342 3343 3344
 *
 * 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
3345
	struct nfs4_readlink_res res;
L
Linus Torvalds 已提交
3346 3347 3348
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_READLINK],
		.rpc_argp = &args,
B
Benny Halevy 已提交
3349
		.rpc_resp = &res,
L
Linus Torvalds 已提交
3350 3351
	};

3352
	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
3353 3354 3355 3356 3357 3358 3359 3360
}

static int nfs4_proc_readlink(struct inode *inode, struct page *page,
		unsigned int pgbase, unsigned int pglen)
{
	struct nfs4_exception exception = { };
	int err;
	do {
3361 3362 3363
		err = _nfs4_proc_readlink(inode, page, pgbase, pglen);
		trace_nfs4_readlink(inode, err);
		err = nfs4_handle_exception(NFS_SERVER(inode), err,
L
Linus Torvalds 已提交
3364 3365 3366 3367 3368 3369
				&exception);
	} while (exception.retry);
	return err;
}

/*
3370
 * This is just for mknod.  open(O_CREAT) will always do ->open_context().
L
Linus Torvalds 已提交
3371 3372 3373
 */
static int
nfs4_proc_create(struct inode *dir, struct dentry *dentry, struct iattr *sattr,
3374
		 int flags)
L
Linus Torvalds 已提交
3375
{
3376
	struct nfs4_label l, *ilabel = NULL;
3377
	struct nfs_open_context *ctx;
L
Linus Torvalds 已提交
3378
	struct nfs4_state *state;
3379
	int opened = 0;
L
Linus Torvalds 已提交
3380 3381
	int status = 0;

3382 3383 3384 3385
	ctx = alloc_nfs_open_context(dentry, FMODE_READ);
	if (IS_ERR(ctx))
		return PTR_ERR(ctx);

3386 3387
	ilabel = nfs4_label_init_security(dir, dentry, sattr, &l);

A
Aneesh Kumar K.V 已提交
3388
	sattr->ia_mode &= ~current_umask();
3389
	state = nfs4_do_open(dir, ctx, flags, sattr, ilabel, &opened);
L
Linus Torvalds 已提交
3390 3391
	if (IS_ERR(state)) {
		status = PTR_ERR(state);
3392
		goto out;
L
Linus Torvalds 已提交
3393 3394
	}
out:
3395
	nfs4_label_release_security(ilabel);
3396
	put_nfs_open_context(ctx);
L
Linus Torvalds 已提交
3397 3398 3399 3400 3401
	return status;
}

static int _nfs4_proc_remove(struct inode *dir, struct qstr *name)
{
3402
	struct nfs_server *server = NFS_SERVER(dir);
3403
	struct nfs_removeargs args = {
L
Linus Torvalds 已提交
3404
		.fh = NFS_FH(dir),
3405
		.name = *name,
3406
	};
3407
	struct nfs_removeres res = {
3408
		.server = server,
L
Linus Torvalds 已提交
3409 3410
	};
	struct rpc_message msg = {
3411 3412 3413
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_REMOVE],
		.rpc_argp = &args,
		.rpc_resp = &res,
L
Linus Torvalds 已提交
3414
	};
3415
	int status;
L
Linus Torvalds 已提交
3416

3417
	status = nfs4_call_sync(server->client, server, &msg, &args.seq_args, &res.seq_res, 1);
3418
	if (status == 0)
3419
		update_changeattr(dir, &res.cinfo);
L
Linus Torvalds 已提交
3420 3421 3422 3423 3424 3425 3426 3427
	return status;
}

static int nfs4_proc_remove(struct inode *dir, struct qstr *name)
{
	struct nfs4_exception exception = { };
	int err;
	do {
3428 3429 3430
		err = _nfs4_proc_remove(dir, name);
		trace_nfs4_remove(dir, name, err);
		err = nfs4_handle_exception(NFS_SERVER(dir), err,
L
Linus Torvalds 已提交
3431 3432 3433 3434 3435
				&exception);
	} while (exception.retry);
	return err;
}

3436
static void nfs4_proc_unlink_setup(struct rpc_message *msg, struct inode *dir)
L
Linus Torvalds 已提交
3437
{
3438 3439 3440
	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
3441

3442
	res->server = server;
L
Linus Torvalds 已提交
3443
	msg->rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_REMOVE];
3444
	nfs4_init_sequence(&args->seq_args, &res->seq_res, 1);
3445 3446

	nfs_fattr_init(res->dir_attr);
L
Linus Torvalds 已提交
3447 3448
}

3449 3450
static void nfs4_proc_unlink_rpc_prepare(struct rpc_task *task, struct nfs_unlinkdata *data)
{
3451 3452 3453 3454
	nfs4_setup_sequence(NFS_SERVER(data->dir),
			&data->args.seq_args,
			&data->res.seq_res,
			task);
L
Linus Torvalds 已提交
3455 3456
}

3457
static int nfs4_proc_unlink_done(struct rpc_task *task, struct inode *dir)
L
Linus Torvalds 已提交
3458
{
3459 3460
	struct nfs_unlinkdata *data = task->tk_calldata;
	struct nfs_removeres *res = &data->res;
3461

3462 3463
	if (!nfs4_sequence_done(task, &res->seq_res))
		return 0;
3464
	if (nfs4_async_handle_error(task, res->server, NULL) == -EAGAIN)
3465 3466 3467
		return 0;
	update_changeattr(dir, &res->cinfo);
	return 1;
L
Linus Torvalds 已提交
3468 3469
}

3470 3471 3472 3473 3474 3475 3476 3477
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;
3478
	nfs4_init_sequence(&arg->seq_args, &res->seq_res, 1);
3479 3480
}

3481 3482
static void nfs4_proc_rename_rpc_prepare(struct rpc_task *task, struct nfs_renamedata *data)
{
3483 3484 3485 3486
	nfs4_setup_sequence(NFS_SERVER(data->old_dir),
			&data->args.seq_args,
			&data->res.seq_res,
			task);
3487 3488 3489 3490 3491
}

static int nfs4_proc_rename_done(struct rpc_task *task, struct inode *old_dir,
				 struct inode *new_dir)
{
3492 3493
	struct nfs_renamedata *data = task->tk_calldata;
	struct nfs_renameres *res = &data->res;
3494 3495 3496 3497 3498 3499 3500 3501 3502 3503 3504

	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
3505 3506 3507
static int _nfs4_proc_rename(struct inode *old_dir, struct qstr *old_name,
		struct inode *new_dir, struct qstr *new_name)
{
3508
	struct nfs_server *server = NFS_SERVER(old_dir);
3509
	struct nfs_renameargs arg = {
L
Linus Torvalds 已提交
3510 3511 3512 3513
		.old_dir = NFS_FH(old_dir),
		.new_dir = NFS_FH(new_dir),
		.old_name = old_name,
		.new_name = new_name,
3514
	};
3515
	struct nfs_renameres res = {
3516
		.server = server,
L
Linus Torvalds 已提交
3517 3518 3519 3520 3521 3522
	};
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_RENAME],
		.rpc_argp = &arg,
		.rpc_resp = &res,
	};
3523
	int status = -ENOMEM;
3524

3525
	status = nfs4_call_sync(server->client, server, &msg, &arg.seq_args, &res.seq_res, 1);
L
Linus Torvalds 已提交
3526 3527 3528 3529 3530 3531 3532 3533 3534 3535 3536 3537 3538
	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 {
3539 3540 3541 3542
		err = _nfs4_proc_rename(old_dir, old_name,
					new_dir, new_name);
		trace_nfs4_rename(old_dir, old_name, new_dir, new_name, err);
		err = nfs4_handle_exception(NFS_SERVER(old_dir), err,
L
Linus Torvalds 已提交
3543 3544 3545 3546 3547 3548 3549
				&exception);
	} while (exception.retry);
	return err;
}

static int _nfs4_proc_link(struct inode *inode, struct inode *dir, struct qstr *name)
{
3550
	struct nfs_server *server = NFS_SERVER(inode);
L
Linus Torvalds 已提交
3551 3552 3553 3554
	struct nfs4_link_arg arg = {
		.fh     = NFS_FH(inode),
		.dir_fh = NFS_FH(dir),
		.name   = name,
3555 3556 3557 3558
		.bitmask = server->attr_bitmask,
	};
	struct nfs4_link_res res = {
		.server = server,
3559
		.label = NULL,
L
Linus Torvalds 已提交
3560 3561 3562 3563
	};
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_LINK],
		.rpc_argp = &arg,
3564
		.rpc_resp = &res,
L
Linus Torvalds 已提交
3565
	};
3566 3567 3568
	int status = -ENOMEM;

	res.fattr = nfs_alloc_fattr();
3569
	if (res.fattr == NULL)
3570
		goto out;
L
Linus Torvalds 已提交
3571

3572 3573 3574 3575 3576
	res.label = nfs4_label_alloc(server, GFP_KERNEL);
	if (IS_ERR(res.label)) {
		status = PTR_ERR(res.label);
		goto out;
	}
3577
	arg.bitmask = nfs4_bitmask(server, res.label);
3578

3579
	status = nfs4_call_sync(server->client, server, &msg, &arg.seq_args, &res.seq_res, 1);
3580 3581
	if (!status) {
		update_changeattr(dir, &res.cinfo);
3582 3583 3584
		status = nfs_post_op_update_inode(inode, res.fattr);
		if (!status)
			nfs_setsecurity(inode, res.fattr, res.label);
3585
	}
3586 3587 3588 3589


	nfs4_label_free(res.label);

3590 3591
out:
	nfs_free_fattr(res.fattr);
L
Linus Torvalds 已提交
3592 3593 3594 3595 3596 3597 3598 3599 3600 3601 3602 3603 3604 3605 3606
	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;
}

3607 3608 3609 3610 3611 3612
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;
3613
	struct nfs4_label *label;
3614 3615 3616 3617 3618 3619 3620 3621 3622 3623 3624
};

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);

3625 3626 3627 3628
		data->label = nfs4_label_alloc(server, GFP_KERNEL);
		if (IS_ERR(data->label))
			goto out_free;

3629 3630 3631 3632 3633 3634 3635 3636
		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;
3637
		data->arg.bitmask = nfs4_bitmask(server, data->label);
3638 3639 3640
		data->res.server = server;
		data->res.fh = &data->fh;
		data->res.fattr = &data->fattr;
3641
		data->res.label = data->label;
3642 3643 3644
		nfs_fattr_init(data->res.fattr);
	}
	return data;
3645 3646 3647
out_free:
	kfree(data);
	return NULL;
3648 3649 3650 3651
}

static int nfs4_do_create(struct inode *dir, struct dentry *dentry, struct nfs4_createdata *data)
{
3652
	int status = nfs4_call_sync(NFS_SERVER(dir)->client, NFS_SERVER(dir), &data->msg,
3653
				    &data->arg.seq_args, &data->res.seq_res, 1);
3654 3655
	if (status == 0) {
		update_changeattr(dir, &data->res.dir_cinfo);
3656
		status = nfs_instantiate(dentry, data->res.fh, data->res.fattr, data->res.label);
3657 3658 3659 3660 3661 3662
	}
	return status;
}

static void nfs4_free_createdata(struct nfs4_createdata *data)
{
3663
	nfs4_label_free(data->label);
3664 3665 3666
	kfree(data);
}

3667
static int _nfs4_proc_symlink(struct inode *dir, struct dentry *dentry,
3668 3669
		struct page *page, unsigned int len, struct iattr *sattr,
		struct nfs4_label *label)
L
Linus Torvalds 已提交
3670
{
3671 3672
	struct nfs4_createdata *data;
	int status = -ENAMETOOLONG;
L
Linus Torvalds 已提交
3673

3674
	if (len > NFS4_MAXPATHLEN)
3675
		goto out;
3676

3677 3678 3679 3680 3681 3682 3683 3684
	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;
3685
	data->arg.label = label;
L
Linus Torvalds 已提交
3686
	
3687 3688 3689 3690
	status = nfs4_do_create(dir, dentry, data);

	nfs4_free_createdata(data);
out:
L
Linus Torvalds 已提交
3691 3692 3693
	return status;
}

3694
static int nfs4_proc_symlink(struct inode *dir, struct dentry *dentry,
3695
		struct page *page, unsigned int len, struct iattr *sattr)
L
Linus Torvalds 已提交
3696 3697
{
	struct nfs4_exception exception = { };
3698
	struct nfs4_label l, *label = NULL;
L
Linus Torvalds 已提交
3699
	int err;
3700 3701 3702

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

L
Linus Torvalds 已提交
3703
	do {
3704 3705 3706
		err = _nfs4_proc_symlink(dir, dentry, page, len, sattr, label);
		trace_nfs4_symlink(dir, &dentry->d_name, err);
		err = nfs4_handle_exception(NFS_SERVER(dir), err,
L
Linus Torvalds 已提交
3707 3708
				&exception);
	} while (exception.retry);
3709 3710

	nfs4_label_release_security(label);
L
Linus Torvalds 已提交
3711 3712 3713 3714
	return err;
}

static int _nfs4_proc_mkdir(struct inode *dir, struct dentry *dentry,
3715
		struct iattr *sattr, struct nfs4_label *label)
L
Linus Torvalds 已提交
3716
{
3717 3718
	struct nfs4_createdata *data;
	int status = -ENOMEM;
L
Linus Torvalds 已提交
3719

3720 3721 3722 3723
	data = nfs4_alloc_createdata(dir, &dentry->d_name, sattr, NF4DIR);
	if (data == NULL)
		goto out;

3724
	data->arg.label = label;
3725 3726 3727 3728
	status = nfs4_do_create(dir, dentry, data);

	nfs4_free_createdata(data);
out:
L
Linus Torvalds 已提交
3729 3730 3731 3732 3733 3734 3735
	return status;
}

static int nfs4_proc_mkdir(struct inode *dir, struct dentry *dentry,
		struct iattr *sattr)
{
	struct nfs4_exception exception = { };
3736
	struct nfs4_label l, *label = NULL;
L
Linus Torvalds 已提交
3737
	int err;
A
Aneesh Kumar K.V 已提交
3738

3739 3740
	label = nfs4_label_init_security(dir, dentry, sattr, &l);

A
Aneesh Kumar K.V 已提交
3741
	sattr->ia_mode &= ~current_umask();
L
Linus Torvalds 已提交
3742
	do {
3743 3744 3745
		err = _nfs4_proc_mkdir(dir, dentry, sattr, label);
		trace_nfs4_mkdir(dir, &dentry->d_name, err);
		err = nfs4_handle_exception(NFS_SERVER(dir), err,
L
Linus Torvalds 已提交
3746 3747
				&exception);
	} while (exception.retry);
3748 3749
	nfs4_label_release_security(label);

L
Linus Torvalds 已提交
3750 3751 3752 3753
	return err;
}

static int _nfs4_proc_readdir(struct dentry *dentry, struct rpc_cred *cred,
B
Bryan Schumaker 已提交
3754
		u64 cookie, struct page **pages, unsigned int count, int plus)
L
Linus Torvalds 已提交
3755 3756 3757 3758
{
	struct inode		*dir = dentry->d_inode;
	struct nfs4_readdir_arg args = {
		.fh = NFS_FH(dir),
B
Bryan Schumaker 已提交
3759
		.pages = pages,
L
Linus Torvalds 已提交
3760 3761
		.pgbase = 0,
		.count = count,
3762
		.bitmask = NFS_SERVER(dentry->d_inode)->attr_bitmask,
B
Bryan Schumaker 已提交
3763
		.plus = plus,
L
Linus Torvalds 已提交
3764 3765 3766 3767 3768 3769 3770 3771 3772 3773
	};
	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;

3774
	dprintk("%s: dentry = %s/%s, cookie = %Lu\n", __func__,
3775 3776 3777
			dentry->d_parent->d_name.name,
			dentry->d_name.name,
			(unsigned long long)cookie);
3778
	nfs4_setup_readdir(cookie, NFS_I(dir)->cookieverf, dentry, &args);
L
Linus Torvalds 已提交
3779
	res.pgbase = args.pgbase;
3780
	status = nfs4_call_sync(NFS_SERVER(dir)->client, NFS_SERVER(dir), &msg, &args.seq_args, &res.seq_res, 0);
3781
	if (status >= 0) {
3782
		memcpy(NFS_I(dir)->cookieverf, res.verifier.data, NFS4_VERIFIER_SIZE);
3783 3784
		status += args.pgbase;
	}
3785 3786 3787

	nfs_invalidate_atime(dir);

3788
	dprintk("%s: returns %d\n", __func__, status);
L
Linus Torvalds 已提交
3789 3790 3791 3792
	return status;
}

static int nfs4_proc_readdir(struct dentry *dentry, struct rpc_cred *cred,
B
Bryan Schumaker 已提交
3793
		u64 cookie, struct page **pages, unsigned int count, int plus)
L
Linus Torvalds 已提交
3794 3795 3796 3797
{
	struct nfs4_exception exception = { };
	int err;
	do {
3798 3799 3800 3801
		err = _nfs4_proc_readdir(dentry, cred, cookie,
				pages, count, plus);
		trace_nfs4_readdir(dentry->d_inode, err);
		err = nfs4_handle_exception(NFS_SERVER(dentry->d_inode), err,
L
Linus Torvalds 已提交
3802 3803 3804 3805 3806 3807
				&exception);
	} while (exception.retry);
	return err;
}

static int _nfs4_proc_mknod(struct inode *dir, struct dentry *dentry,
3808
		struct iattr *sattr, struct nfs4_label *label, dev_t rdev)
L
Linus Torvalds 已提交
3809
{
3810 3811 3812
	struct nfs4_createdata *data;
	int mode = sattr->ia_mode;
	int status = -ENOMEM;
L
Linus Torvalds 已提交
3813

3814 3815 3816 3817
	data = nfs4_alloc_createdata(dir, &dentry->d_name, sattr, NF4SOCK);
	if (data == NULL)
		goto out;

L
Linus Torvalds 已提交
3818
	if (S_ISFIFO(mode))
3819
		data->arg.ftype = NF4FIFO;
L
Linus Torvalds 已提交
3820
	else if (S_ISBLK(mode)) {
3821 3822 3823
		data->arg.ftype = NF4BLK;
		data->arg.u.device.specdata1 = MAJOR(rdev);
		data->arg.u.device.specdata2 = MINOR(rdev);
L
Linus Torvalds 已提交
3824 3825
	}
	else if (S_ISCHR(mode)) {
3826 3827 3828
		data->arg.ftype = NF4CHR;
		data->arg.u.device.specdata1 = MAJOR(rdev);
		data->arg.u.device.specdata2 = MINOR(rdev);
3829 3830 3831
	} else if (!S_ISSOCK(mode)) {
		status = -EINVAL;
		goto out_free;
L
Linus Torvalds 已提交
3832
	}
3833

3834
	data->arg.label = label;
3835
	status = nfs4_do_create(dir, dentry, data);
3836
out_free:
3837 3838
	nfs4_free_createdata(data);
out:
L
Linus Torvalds 已提交
3839 3840 3841 3842 3843 3844 3845
	return status;
}

static int nfs4_proc_mknod(struct inode *dir, struct dentry *dentry,
		struct iattr *sattr, dev_t rdev)
{
	struct nfs4_exception exception = { };
3846
	struct nfs4_label l, *label = NULL;
L
Linus Torvalds 已提交
3847
	int err;
A
Aneesh Kumar K.V 已提交
3848

3849 3850
	label = nfs4_label_init_security(dir, dentry, sattr, &l);

A
Aneesh Kumar K.V 已提交
3851
	sattr->ia_mode &= ~current_umask();
L
Linus Torvalds 已提交
3852
	do {
3853 3854 3855
		err = _nfs4_proc_mknod(dir, dentry, sattr, label, rdev);
		trace_nfs4_mknod(dir, &dentry->d_name, err);
		err = nfs4_handle_exception(NFS_SERVER(dir), err,
L
Linus Torvalds 已提交
3856 3857
				&exception);
	} while (exception.retry);
3858 3859 3860

	nfs4_label_release_security(label);

L
Linus Torvalds 已提交
3861 3862 3863 3864 3865 3866 3867 3868 3869 3870
	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
3871 3872 3873
	struct nfs4_statfs_res res = {
		.fsstat = fsstat,
	};
L
Linus Torvalds 已提交
3874 3875 3876
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_STATFS],
		.rpc_argp = &args,
B
Benny Halevy 已提交
3877
		.rpc_resp = &res,
L
Linus Torvalds 已提交
3878 3879
	};

3880
	nfs_fattr_init(fsstat->fattr);
3881
	return  nfs4_call_sync(server->client, server, &msg, &args.seq_args, &res.seq_res, 0);
L
Linus Torvalds 已提交
3882 3883 3884 3885 3886 3887 3888 3889 3890 3891 3892 3893 3894 3895 3896 3897 3898 3899 3900 3901 3902
}

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
3903 3904 3905
	struct nfs4_fsinfo_res res = {
		.fsinfo = fsinfo,
	};
L
Linus Torvalds 已提交
3906 3907 3908
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_FSINFO],
		.rpc_argp = &args,
B
Benny Halevy 已提交
3909
		.rpc_resp = &res,
L
Linus Torvalds 已提交
3910 3911
	};

3912
	return nfs4_call_sync(server->client, server, &msg, &args.seq_args, &res.seq_res, 0);
L
Linus Torvalds 已提交
3913 3914 3915 3916 3917
}

static int nfs4_do_fsinfo(struct nfs_server *server, struct nfs_fh *fhandle, struct nfs_fsinfo *fsinfo)
{
	struct nfs4_exception exception = { };
3918
	unsigned long now = jiffies;
L
Linus Torvalds 已提交
3919 3920 3921
	int err;

	do {
3922
		err = _nfs4_do_fsinfo(server, fhandle, fsinfo);
3923
		trace_nfs4_fsinfo(server, fhandle, fsinfo->fattr, err);
3924 3925 3926 3927 3928 3929 3930 3931 3932 3933
		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
3934 3935 3936 3937 3938 3939
	} while (exception.retry);
	return err;
}

static int nfs4_proc_fsinfo(struct nfs_server *server, struct nfs_fh *fhandle, struct nfs_fsinfo *fsinfo)
{
3940 3941
	int error;

3942
	nfs_fattr_init(fsinfo->fattr);
3943
	error = nfs4_do_fsinfo(server, fhandle, fsinfo);
3944 3945 3946
	if (error == 0) {
		/* block layout checks this! */
		server->pnfs_blksize = fsinfo->blksize;
3947
		set_pnfs_layoutdriver(server, fhandle, fsinfo->layouttype);
3948
	}
3949 3950

	return error;
L
Linus Torvalds 已提交
3951 3952 3953 3954 3955 3956 3957 3958 3959
}

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
3960 3961 3962
	struct nfs4_pathconf_res res = {
		.pathconf = pathconf,
	};
L
Linus Torvalds 已提交
3963 3964 3965
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_PATHCONF],
		.rpc_argp = &args,
B
Benny Halevy 已提交
3966
		.rpc_resp = &res,
L
Linus Torvalds 已提交
3967 3968 3969 3970 3971 3972 3973 3974
	};

	/* 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;
	}

3975
	nfs_fattr_init(pathconf->fattr);
3976
	return nfs4_call_sync(server->client, server, &msg, &args.seq_args, &res.seq_res, 0);
L
Linus Torvalds 已提交
3977 3978 3979 3980 3981 3982 3983 3984 3985 3986 3987 3988 3989 3990 3991 3992
}

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;
}

3993
int nfs4_set_rw_stateid(nfs4_stateid *stateid,
3994 3995 3996 3997 3998 3999 4000 4001
		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;
4002
	return nfs4_select_rw_stateid(stateid, ctx->state, fmode, lockowner);
4003 4004 4005
}
EXPORT_SYMBOL_GPL(nfs4_set_rw_stateid);

4006 4007 4008 4009 4010 4011 4012 4013 4014 4015 4016 4017 4018 4019 4020 4021 4022 4023 4024 4025 4026 4027 4028 4029 4030 4031 4032
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;
}

4033 4034
void __nfs4_read_done_cb(struct nfs_read_data *data)
{
4035
	nfs_invalidate_atime(data->header->inode);
4036 4037
}

4038
static int nfs4_read_done_cb(struct rpc_task *task, struct nfs_read_data *data)
L
Linus Torvalds 已提交
4039
{
4040
	struct nfs_server *server = NFS_SERVER(data->header->inode);
L
Linus Torvalds 已提交
4041

4042
	trace_nfs4_read(data, task->tk_status);
4043
	if (nfs4_async_handle_error(task, server, data->args.context->state) == -EAGAIN) {
4044
		rpc_restart_call_prepare(task);
T
Trond Myklebust 已提交
4045
		return -EAGAIN;
L
Linus Torvalds 已提交
4046
	}
4047

4048
	__nfs4_read_done_cb(data);
L
Linus Torvalds 已提交
4049
	if (task->tk_status > 0)
T
Trond Myklebust 已提交
4050 4051
		renew_lease(server, data->timestamp);
	return 0;
L
Linus Torvalds 已提交
4052 4053
}

4054 4055 4056 4057 4058 4059 4060 4061 4062 4063 4064 4065 4066 4067
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;
}

4068 4069 4070 4071 4072 4073 4074
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;
4075 4076
	if (nfs4_read_stateid_changed(task, &data->args))
		return -EAGAIN;
4077 4078
	return data->read_done_cb ? data->read_done_cb(task, data) :
				    nfs4_read_done_cb(task, data);
4079 4080
}

4081
static void nfs4_proc_read_setup(struct nfs_read_data *data, struct rpc_message *msg)
L
Linus Torvalds 已提交
4082 4083
{
	data->timestamp   = jiffies;
4084
	data->read_done_cb = nfs4_read_done_cb;
4085
	msg->rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_READ];
4086
	nfs4_init_sequence(&data->args.seq_args, &data->res.seq_res, 0);
L
Linus Torvalds 已提交
4087 4088
}

4089
static int nfs4_proc_read_rpc_prepare(struct rpc_task *task, struct nfs_read_data *data)
4090
{
4091
	if (nfs4_setup_sequence(NFS_SERVER(data->header->inode),
4092 4093
			&data->args.seq_args,
			&data->res.seq_res,
4094
			task))
4095 4096 4097 4098 4099 4100 4101
		return 0;
	if (nfs4_set_rw_stateid(&data->args.stateid, data->args.context,
				data->args.lock_context, FMODE_READ) == -EIO)
		return -EIO;
	if (unlikely(test_bit(NFS_CONTEXT_BAD, &data->args.context->flags)))
		return -EIO;
	return 0;
L
Linus Torvalds 已提交
4102 4103
}

4104
static int nfs4_write_done_cb(struct rpc_task *task, struct nfs_write_data *data)
L
Linus Torvalds 已提交
4105
{
4106
	struct inode *inode = data->header->inode;
L
Linus Torvalds 已提交
4107
	
4108
	trace_nfs4_write(data, task->tk_status);
4109
	if (nfs4_async_handle_error(task, NFS_SERVER(inode), data->args.context->state) == -EAGAIN) {
4110
		rpc_restart_call_prepare(task);
4111
		return -EAGAIN;
L
Linus Torvalds 已提交
4112
	}
4113
	if (task->tk_status >= 0) {
L
Linus Torvalds 已提交
4114
		renew_lease(NFS_SERVER(inode), data->timestamp);
4115
		nfs_post_op_update_inode_force_wcc(inode, &data->fattr);
4116
	}
4117
	return 0;
L
Linus Torvalds 已提交
4118 4119
}

4120 4121 4122 4123 4124 4125 4126 4127 4128 4129 4130 4131 4132 4133
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;
}

4134 4135 4136 4137
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;
4138 4139
	if (nfs4_write_stateid_changed(task, &data->args))
		return -EAGAIN;
4140 4141
	return data->write_done_cb ? data->write_done_cb(task, data) :
		nfs4_write_done_cb(task, data);
4142 4143
}

4144 4145
static
bool nfs4_write_need_cache_consistency_data(const struct nfs_write_data *data)
4146
{
4147 4148 4149 4150 4151 4152 4153 4154
	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
	 */
4155
	return nfs4_have_delegation(hdr->inode, FMODE_READ) == 0;
4156 4157
}

4158
static void nfs4_proc_write_setup(struct nfs_write_data *data, struct rpc_message *msg)
L
Linus Torvalds 已提交
4159
{
4160
	struct nfs_server *server = NFS_SERVER(data->header->inode);
4161

4162
	if (!nfs4_write_need_cache_consistency_data(data)) {
4163 4164 4165 4166
		data->args.bitmask = NULL;
		data->res.fattr = NULL;
	} else
		data->args.bitmask = server->cache_consistency_bitmask;
4167

4168 4169
	if (!data->write_done_cb)
		data->write_done_cb = nfs4_write_done_cb;
4170
	data->res.server = server;
L
Linus Torvalds 已提交
4171 4172
	data->timestamp   = jiffies;

4173
	msg->rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_WRITE];
4174
	nfs4_init_sequence(&data->args.seq_args, &data->res.seq_res, 1);
L
Linus Torvalds 已提交
4175 4176
}

4177
static int nfs4_proc_write_rpc_prepare(struct rpc_task *task, struct nfs_write_data *data)
4178
{
4179
	if (nfs4_setup_sequence(NFS_SERVER(data->header->inode),
4180 4181
			&data->args.seq_args,
			&data->res.seq_res,
4182
			task))
4183 4184 4185 4186 4187 4188 4189
		return 0;
	if (nfs4_set_rw_stateid(&data->args.stateid, data->args.context,
				data->args.lock_context, FMODE_WRITE) == -EIO)
		return -EIO;
	if (unlikely(test_bit(NFS_CONTEXT_BAD, &data->args.context->flags)))
		return -EIO;
	return 0;
L
Linus Torvalds 已提交
4190 4191
}

4192
static void nfs4_proc_commit_rpc_prepare(struct rpc_task *task, struct nfs_commit_data *data)
4193
{
4194 4195 4196 4197
	nfs4_setup_sequence(NFS_SERVER(data->inode),
			&data->args.seq_args,
			&data->res.seq_res,
			task);
L
Linus Torvalds 已提交
4198 4199
}

4200
static int nfs4_commit_done_cb(struct rpc_task *task, struct nfs_commit_data *data)
L
Linus Torvalds 已提交
4201 4202
{
	struct inode *inode = data->inode;
4203

4204
	trace_nfs4_commit(data, task->tk_status);
4205
	if (nfs4_async_handle_error(task, NFS_SERVER(inode), NULL) == -EAGAIN) {
4206
		rpc_restart_call_prepare(task);
4207
		return -EAGAIN;
L
Linus Torvalds 已提交
4208
	}
4209
	return 0;
L
Linus Torvalds 已提交
4210 4211
}

4212
static int nfs4_commit_done(struct rpc_task *task, struct nfs_commit_data *data)
4213 4214 4215
{
	if (!nfs4_sequence_done(task, &data->res.seq_res))
		return -EAGAIN;
4216
	return data->commit_done_cb(task, data);
4217 4218
}

4219
static void nfs4_proc_commit_setup(struct nfs_commit_data *data, struct rpc_message *msg)
L
Linus Torvalds 已提交
4220
{
4221
	struct nfs_server *server = NFS_SERVER(data->inode);
4222

4223 4224
	if (data->commit_done_cb == NULL)
		data->commit_done_cb = nfs4_commit_done_cb;
4225
	data->res.server = server;
4226
	msg->rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_COMMIT];
4227
	nfs4_init_sequence(&data->args.seq_args, &data->res.seq_res, 1);
L
Linus Torvalds 已提交
4228 4229
}

4230 4231 4232 4233 4234
struct nfs4_renewdata {
	struct nfs_client	*client;
	unsigned long		timestamp;
};

L
Linus Torvalds 已提交
4235 4236 4237 4238
/*
 * nfs4_proc_async_renew(): This is not one of the nfs_rpc_ops; it is a special
 * standalone procedure for queueing an asynchronous RENEW.
 */
4239
static void nfs4_renew_release(void *calldata)
4240
{
4241 4242
	struct nfs4_renewdata *data = calldata;
	struct nfs_client *clp = data->client;
4243

4244 4245 4246
	if (atomic_read(&clp->cl_count) > 1)
		nfs4_schedule_state_renewal(clp);
	nfs_put_client(clp);
4247
	kfree(data);
4248 4249
}

4250
static void nfs4_renew_done(struct rpc_task *task, void *calldata)
L
Linus Torvalds 已提交
4251
{
4252 4253 4254
	struct nfs4_renewdata *data = calldata;
	struct nfs_client *clp = data->client;
	unsigned long timestamp = data->timestamp;
L
Linus Torvalds 已提交
4255

4256
	trace_nfs4_renew_async(clp, task->tk_status);
4257 4258 4259 4260 4261 4262 4263
	switch (task->tk_status) {
	case 0:
		break;
	case -NFS4ERR_LEASE_MOVED:
		nfs4_schedule_lease_moved_recovery(clp);
		break;
	default:
4264
		/* Unless we're shutting down, schedule state recovery! */
4265 4266 4267
		if (test_bit(NFS_CS_RENEWD, &clp->cl_res_state) == 0)
			return;
		if (task->tk_status != NFS4ERR_CB_PATH_DOWN) {
4268
			nfs4_schedule_lease_recovery(clp);
4269 4270 4271
			return;
		}
		nfs4_schedule_path_down_recovery(clp);
L
Linus Torvalds 已提交
4272
	}
4273
	do_renew_lease(clp, timestamp);
L
Linus Torvalds 已提交
4274 4275
}

4276 4277
static const struct rpc_call_ops nfs4_renew_ops = {
	.rpc_call_done = nfs4_renew_done,
4278
	.rpc_release = nfs4_renew_release,
4279 4280
};

4281
static int nfs4_proc_async_renew(struct nfs_client *clp, struct rpc_cred *cred, unsigned renew_flags)
L
Linus Torvalds 已提交
4282 4283 4284 4285
{
	struct rpc_message msg = {
		.rpc_proc	= &nfs4_procedures[NFSPROC4_CLNT_RENEW],
		.rpc_argp	= clp,
4286
		.rpc_cred	= cred,
L
Linus Torvalds 已提交
4287
	};
4288
	struct nfs4_renewdata *data;
L
Linus Torvalds 已提交
4289

4290 4291
	if (renew_flags == 0)
		return 0;
4292 4293
	if (!atomic_inc_not_zero(&clp->cl_count))
		return -EIO;
4294
	data = kmalloc(sizeof(*data), GFP_NOFS);
4295 4296 4297 4298
	if (data == NULL)
		return -ENOMEM;
	data->client = clp;
	data->timestamp = jiffies;
4299
	return rpc_call_async(clp->cl_rpcclient, &msg, RPC_TASK_TIMEOUT,
4300
			&nfs4_renew_ops, data);
L
Linus Torvalds 已提交
4301 4302
}

4303
static int nfs4_proc_renew(struct nfs_client *clp, struct rpc_cred *cred)
L
Linus Torvalds 已提交
4304 4305 4306 4307
{
	struct rpc_message msg = {
		.rpc_proc	= &nfs4_procedures[NFSPROC4_CLNT_RENEW],
		.rpc_argp	= clp,
4308
		.rpc_cred	= cred,
L
Linus Torvalds 已提交
4309 4310 4311 4312
	};
	unsigned long now = jiffies;
	int status;

4313
	status = rpc_call_sync(clp->cl_rpcclient, &msg, RPC_TASK_TIMEOUT);
L
Linus Torvalds 已提交
4314 4315
	if (status < 0)
		return status;
4316
	do_renew_lease(clp, now);
L
Linus Torvalds 已提交
4317 4318 4319
	return 0;
}

4320 4321 4322 4323 4324 4325 4326
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);
}

4327 4328
/* 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
4329 4330
 * the stack.
 */
4331
#define NFS4ACL_MAXPAGES DIV_ROUND_UP(XATTR_SIZE_MAX, PAGE_SIZE)
4332

4333 4334 4335 4336 4337 4338 4339 4340 4341
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 {
4342
		len = min_t(size_t, PAGE_SIZE, buflen);
4343 4344 4345 4346 4347 4348 4349 4350 4351 4352 4353 4354 4355 4356 4357 4358 4359 4360 4361
		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;
}

4362 4363 4364
struct nfs4_cached_acl {
	int cached;
	size_t len;
A
Andrew Morton 已提交
4365
	char data[0];
4366 4367 4368 4369 4370 4371 4372 4373 4374 4375 4376 4377 4378 4379 4380 4381 4382 4383 4384 4385 4386 4387 4388 4389 4390 4391 4392 4393 4394 4395 4396 4397 4398 4399 4400 4401 4402 4403 4404 4405 4406 4407
};

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;
}

4408
static void nfs4_write_cached_acl(struct inode *inode, struct page **pages, size_t pgbase, size_t acl_len)
4409 4410
{
	struct nfs4_cached_acl *acl;
4411
	size_t buflen = sizeof(*acl) + acl_len;
4412

4413
	if (buflen <= PAGE_SIZE) {
4414
		acl = kmalloc(buflen, GFP_KERNEL);
4415 4416 4417
		if (acl == NULL)
			goto out;
		acl->cached = 1;
4418
		_copy_from_pages(acl->data, pages, pgbase, acl_len);
4419 4420 4421 4422 4423 4424 4425 4426 4427 4428 4429
	} 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);
}

4430 4431 4432 4433 4434 4435 4436 4437 4438 4439
/*
 * 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.
 */
4440
static ssize_t __nfs4_get_acl_uncached(struct inode *inode, void *buf, size_t buflen)
4441
{
4442
	struct page *pages[NFS4ACL_MAXPAGES] = {NULL, };
4443 4444 4445 4446 4447
	struct nfs_getaclargs args = {
		.fh = NFS_FH(inode),
		.acl_pages = pages,
		.acl_len = buflen,
	};
B
Benny Halevy 已提交
4448 4449 4450
	struct nfs_getaclres res = {
		.acl_len = buflen,
	};
4451 4452 4453
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_GETACL],
		.rpc_argp = &args,
B
Benny Halevy 已提交
4454
		.rpc_resp = &res,
4455
	};
4456 4457
	unsigned int npages = DIV_ROUND_UP(buflen, PAGE_SIZE);
	int ret = -ENOMEM, i;
4458

4459 4460 4461 4462
	/* 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;
4463 4464
	if (npages > ARRAY_SIZE(pages))
		return -ERANGE;
4465

4466 4467 4468 4469
	for (i = 0; i < npages; i++) {
		pages[i] = alloc_page(GFP_KERNEL);
		if (!pages[i])
			goto out_free;
4470
	}
4471 4472 4473 4474 4475 4476

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

4477 4478
	args.acl_len = npages * PAGE_SIZE;
	args.acl_pgbase = 0;
4479

4480
	dprintk("%s  buf %p buflen %zu npages %d args.acl_len %zu\n",
4481 4482 4483
		__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);
4484 4485
	if (ret)
		goto out_free;
4486

4487 4488 4489 4490 4491
	/* 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;
4492
		ret = -ERANGE;
4493
		goto out_free;
4494
	}
4495
	nfs4_write_cached_acl(inode, pages, res.acl_data_offset, res.acl_len);
4496 4497 4498 4499 4500
	if (buf) {
		if (res.acl_len > buflen) {
			ret = -ERANGE;
			goto out_free;
		}
4501
		_copy_from_pages(buf, pages, res.acl_data_offset, res.acl_len);
4502
	}
4503 4504
out_ok:
	ret = res.acl_len;
4505
out_free:
4506 4507 4508
	for (i = 0; i < npages; i++)
		if (pages[i])
			__free_page(pages[i]);
4509 4510
	if (res.acl_scratch)
		__free_page(res.acl_scratch);
4511 4512 4513
	return ret;
}

4514 4515 4516 4517 4518 4519
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);
4520
		trace_nfs4_get_acl(inode, ret);
4521 4522 4523 4524 4525 4526 4527
		if (ret >= 0)
			break;
		ret = nfs4_handle_exception(NFS_SERVER(inode), ret, &exception);
	} while (exception.retry);
	return ret;
}

4528 4529 4530 4531 4532 4533 4534 4535 4536 4537
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;
4538 4539
	if (NFS_I(inode)->cache_validity & NFS_INO_INVALID_ACL)
		nfs_zap_acl_cache(inode);
4540 4541
	ret = nfs4_read_cached_acl(inode, buf, buflen);
	if (ret != -ENOENT)
4542 4543
		/* -ENOENT is returned if there is no ACL or if there is an ACL
		 * but no cached acl data, just the acl length */
4544 4545 4546 4547
		return ret;
	return nfs4_get_acl_uncached(inode, buf, buflen);
}

4548
static int __nfs4_proc_set_acl(struct inode *inode, const void *buf, size_t buflen)
4549 4550 4551 4552 4553 4554 4555 4556
{
	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
4557
	struct nfs_setaclres res;
4558 4559 4560
	struct rpc_message msg = {
		.rpc_proc	= &nfs4_procedures[NFSPROC4_CLNT_SETACL],
		.rpc_argp	= &arg,
B
Benny Halevy 已提交
4561
		.rpc_resp	= &res,
4562
	};
4563
	unsigned int npages = DIV_ROUND_UP(buflen, PAGE_SIZE);
4564
	int ret, i;
4565 4566 4567

	if (!nfs4_server_supports_acls(server))
		return -EOPNOTSUPP;
4568 4569
	if (npages > ARRAY_SIZE(pages))
		return -ERANGE;
4570 4571 4572
	i = buf_to_pages_noslab(buf, buflen, arg.acl_pages, &arg.acl_pgbase);
	if (i < 0)
		return i;
4573
	nfs4_inode_return_delegation(inode);
4574
	ret = nfs4_call_sync(server->client, server, &msg, &arg.seq_args, &res.seq_res, 1);
4575 4576 4577 4578 4579 4580 4581 4582

	/*
	 * 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]);

4583 4584 4585 4586 4587 4588 4589
	/*
	 * 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);
4590 4591
	nfs_access_zap_cache(inode);
	nfs_zap_acl_cache(inode);
4592 4593 4594
	return ret;
}

4595 4596 4597 4598 4599
static int nfs4_proc_set_acl(struct inode *inode, const void *buf, size_t buflen)
{
	struct nfs4_exception exception = { };
	int err;
	do {
4600 4601 4602
		err = __nfs4_proc_set_acl(inode, buf, buflen);
		trace_nfs4_set_acl(inode, err);
		err = nfs4_handle_exception(NFS_SERVER(inode), err,
4603 4604 4605 4606 4607
				&exception);
	} while (exception.retry);
	return err;
}

4608 4609 4610 4611 4612 4613 4614 4615 4616
#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 };
4617
	struct nfs4_getattr_arg arg = {
4618 4619 4620 4621 4622 4623 4624 4625 4626 4627
		.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],
4628
		.rpc_argp	= &arg,
4629 4630 4631 4632 4633 4634
		.rpc_resp	= &res,
	};
	int ret;

	nfs_fattr_init(&fattr);

4635
	ret = nfs4_call_sync(server->client, server, &msg, &arg.seq_args, &res.seq_res, 0);
4636 4637 4638 4639 4640 4641 4642 4643 4644 4645 4646 4647 4648 4649 4650 4651 4652 4653 4654
	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 {
4655 4656 4657
		err = _nfs4_get_security_label(inode, buf, buflen);
		trace_nfs4_get_security_label(inode, err);
		err = nfs4_handle_exception(NFS_SERVER(inode), err,
4658 4659 4660 4661 4662 4663 4664 4665 4666 4667 4668 4669 4670 4671
				&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 };
4672
	struct nfs_setattrargs arg = {
4673 4674 4675 4676 4677 4678 4679 4680 4681 4682 4683 4684 4685
		.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],
4686
		.rpc_argp       = &arg,
4687 4688 4689 4690
		.rpc_resp       = &res,
	};
	int status;

4691
	nfs4_stateid_copy(&arg.stateid, &zero_stateid);
4692

4693
	status = nfs4_call_sync(server->client, server, &msg, &arg.seq_args, &res.seq_res, 1);
4694 4695 4696 4697 4698 4699 4700 4701 4702 4703 4704 4705 4706 4707 4708
	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 {
4709 4710 4711 4712
		err = _nfs4_do_set_security_label(inode, ilabel,
				fattr, olabel);
		trace_nfs4_set_security_label(inode, err);
		err = nfs4_handle_exception(NFS_SERVER(inode), err,
4713 4714 4715 4716 4717 4718 4719 4720 4721 4722 4723 4724 4725 4726 4727 4728 4729 4730 4731 4732 4733 4734 4735 4736 4737 4738 4739 4740 4741 4742 4743 4744 4745 4746 4747 4748 4749 4750 4751 4752 4753 4754 4755 4756 4757 4758
				&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
4759
static int
4760
nfs4_async_handle_error(struct rpc_task *task, const struct nfs_server *server, struct nfs4_state *state)
L
Linus Torvalds 已提交
4761
{
4762 4763 4764
	struct nfs_client *clp = server->nfs_client;

	if (task->tk_status >= 0)
L
Linus Torvalds 已提交
4765 4766
		return 0;
	switch(task->tk_status) {
4767
		case -NFS4ERR_DELEG_REVOKED:
4768 4769
		case -NFS4ERR_ADMIN_REVOKED:
		case -NFS4ERR_BAD_STATEID:
4770 4771 4772
			if (state == NULL)
				break;
			nfs_remove_bad_delegation(state->inode);
4773 4774 4775
		case -NFS4ERR_OPENMODE:
			if (state == NULL)
				break;
4776
			if (nfs4_schedule_stateid_recovery(server, state) < 0)
C
Chuck Lever 已提交
4777
				goto recovery_failed;
4778
			goto wait_on_recovery;
4779
		case -NFS4ERR_EXPIRED:
4780 4781
			if (state != NULL) {
				if (nfs4_schedule_stateid_recovery(server, state) < 0)
C
Chuck Lever 已提交
4782
					goto recovery_failed;
4783
			}
L
Linus Torvalds 已提交
4784
		case -NFS4ERR_STALE_STATEID:
4785
		case -NFS4ERR_STALE_CLIENTID:
4786 4787
			nfs4_schedule_lease_recovery(clp);
			goto wait_on_recovery;
4788 4789 4790 4791
		case -NFS4ERR_MOVED:
			if (nfs4_schedule_migration_recovery(server) < 0)
				goto recovery_failed;
			goto wait_on_recovery;
4792 4793 4794
		case -NFS4ERR_LEASE_MOVED:
			nfs4_schedule_lease_moved_recovery(clp);
			goto wait_on_recovery;
4795 4796 4797 4798 4799 4800 4801 4802 4803 4804
#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);
4805
			nfs4_schedule_session_recovery(clp->cl_session, task->tk_status);
4806
			goto wait_on_recovery;
4807
#endif /* CONFIG_NFS_V4_1 */
L
Linus Torvalds 已提交
4808
		case -NFS4ERR_DELAY:
4809
			nfs_inc_server_stats(server, NFSIOS_DELAY);
4810
		case -NFS4ERR_GRACE:
L
Linus Torvalds 已提交
4811
			rpc_delay(task, NFS4_POLL_RETRY_MAX);
4812
		case -NFS4ERR_RETRY_UNCACHED_REP:
L
Linus Torvalds 已提交
4813
		case -NFS4ERR_OLD_STATEID:
4814
			goto restart_call;
L
Linus Torvalds 已提交
4815 4816 4817
	}
	task->tk_status = nfs4_map_errors(task->tk_status);
	return 0;
C
Chuck Lever 已提交
4818
recovery_failed:
4819 4820
	task->tk_status = -EIO;
	return 0;
4821
wait_on_recovery:
4822 4823 4824
	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);
4825 4826
	if (test_bit(NFS_MIG_FAILED, &server->mig_status))
		goto recovery_failed;
4827
restart_call:
4828 4829
	task->tk_status = 0;
	return -EAGAIN;
L
Linus Torvalds 已提交
4830 4831
}

4832 4833
static void nfs4_init_boot_verifier(const struct nfs_client *clp,
				    nfs4_verifier *bootverf)
4834 4835 4836
{
	__be32 verf[2];

4837 4838 4839 4840
	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;
4841
		verf[1] = cpu_to_be32(NSEC_PER_SEC + 1);
4842
	} else {
4843
		struct nfs_net *nn = net_generic(clp->cl_net, nfs_net_id);
4844 4845
		verf[0] = cpu_to_be32(nn->boot_time.tv_sec);
		verf[1] = cpu_to_be32(nn->boot_time.tv_nsec);
4846
	}
4847 4848 4849
	memcpy(bootverf->data, verf, sizeof(bootverf->data));
}

4850 4851 4852 4853 4854 4855 4856 4857 4858 4859 4860 4861 4862 4863 4864 4865 4866 4867 4868 4869 4870
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)
{
4871
	const char *nodename = clp->cl_rpcclient->cl_nodename;
4872 4873

	if (nfs4_client_id_uniquifier[0] != '\0')
4874 4875 4876 4877 4878
		return scnprintf(buf, len, "Linux NFSv%u.%u %s/%s",
				clp->rpc_ops->version,
				clp->cl_minorversion,
				nfs4_client_id_uniquifier,
				nodename);
4879 4880
	return scnprintf(buf, len, "Linux NFSv%u.%u %s",
				clp->rpc_ops->version, clp->cl_minorversion,
4881
				nodename);
4882 4883
}

4884 4885 4886 4887 4888 4889 4890 4891 4892 4893
/**
 * 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.
 */
4894 4895 4896
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
4897 4898 4899 4900 4901
{
	nfs4_verifier sc_verifier;
	struct nfs4_setclientid setclientid = {
		.sc_verifier = &sc_verifier,
		.sc_prog = program,
A
Andy Adamson 已提交
4902
		.sc_cb_ident = clp->cl_cb_ident,
L
Linus Torvalds 已提交
4903 4904 4905 4906
	};
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_SETCLIENTID],
		.rpc_argp = &setclientid,
4907
		.rpc_resp = res,
4908
		.rpc_cred = cred,
L
Linus Torvalds 已提交
4909
	};
4910
	int status;
L
Linus Torvalds 已提交
4911

4912
	/* nfs_client_id4 */
4913
	nfs4_init_boot_verifier(clp, &sc_verifier);
4914 4915 4916 4917 4918 4919 4920 4921 4922 4923
	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));
4924
	/* cb_client4 */
4925
	rcu_read_lock();
4926
	setclientid.sc_netid_len = scnprintf(setclientid.sc_netid,
4927
				sizeof(setclientid.sc_netid), "%s",
4928 4929
				rpc_peeraddr2str(clp->cl_rpcclient,
							RPC_DISPLAY_NETID));
4930 4931
	rcu_read_unlock();
	setclientid.sc_uaddr_len = scnprintf(setclientid.sc_uaddr,
4932
				sizeof(setclientid.sc_uaddr), "%s.%u.%u",
L
Linus Torvalds 已提交
4933 4934
				clp->cl_ipaddr, port >> 8, port & 255);

4935 4936 4937 4938
	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);
4939
	trace_nfs4_setclientid(clp, status);
4940 4941
	dprintk("NFS reply setclientid: %d\n", status);
	return status;
L
Linus Torvalds 已提交
4942 4943
}

4944 4945 4946 4947 4948 4949 4950 4951
/**
 * 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.
 */
4952
int nfs4_proc_setclientid_confirm(struct nfs_client *clp,
4953 4954
		struct nfs4_setclientid_res *arg,
		struct rpc_cred *cred)
L
Linus Torvalds 已提交
4955 4956 4957
{
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_SETCLIENTID_CONFIRM],
4958
		.rpc_argp = arg,
4959
		.rpc_cred = cred,
L
Linus Torvalds 已提交
4960 4961 4962
	};
	int status;

4963 4964 4965
	dprintk("NFS call  setclientid_confirm auth=%s, (client ID %llx)\n",
		clp->cl_rpcclient->cl_auth->au_ops->au_name,
		clp->cl_clientid);
4966
	status = rpc_call_sync(clp->cl_rpcclient, &msg, RPC_TASK_TIMEOUT);
4967
	trace_nfs4_setclientid_confirm(clp, status);
4968
	dprintk("NFS reply setclientid_confirm: %d\n", status);
L
Linus Torvalds 已提交
4969 4970 4971
	return status;
}

4972 4973
struct nfs4_delegreturndata {
	struct nfs4_delegreturnargs args;
4974
	struct nfs4_delegreturnres res;
4975 4976
	struct nfs_fh fh;
	nfs4_stateid stateid;
4977
	unsigned long timestamp;
4978
	struct nfs_fattr fattr;
4979 4980 4981 4982 4983 4984
	int rpc_status;
};

static void nfs4_delegreturn_done(struct rpc_task *task, void *calldata)
{
	struct nfs4_delegreturndata *data = calldata;
4985

4986 4987
	if (!nfs4_sequence_done(task, &data->res.seq_res))
		return;
4988

4989
	trace_nfs4_delegreturn_exit(&data->args, &data->res, task->tk_status);
4990 4991 4992 4993
	switch (task->tk_status) {
	case -NFS4ERR_STALE_STATEID:
	case -NFS4ERR_EXPIRED:
	case 0:
4994
		renew_lease(data->res.server, data->timestamp);
4995 4996 4997 4998
		break;
	default:
		if (nfs4_async_handle_error(task, data->res.server, NULL) ==
				-EAGAIN) {
4999
			rpc_restart_call_prepare(task);
5000 5001 5002 5003
			return;
		}
	}
	data->rpc_status = task->tk_status;
5004 5005 5006 5007 5008 5009 5010
}

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

5011 5012 5013 5014 5015 5016
static void nfs4_delegreturn_prepare(struct rpc_task *task, void *data)
{
	struct nfs4_delegreturndata *d_data;

	d_data = (struct nfs4_delegreturndata *)data;

5017 5018 5019 5020
	nfs4_setup_sequence(d_data->res.server,
			&d_data->args.seq_args,
			&d_data->res.seq_res,
			task);
5021 5022
}

5023
static const struct rpc_call_ops nfs4_delegreturn_ops = {
5024
	.rpc_call_prepare = nfs4_delegreturn_prepare,
5025 5026 5027 5028
	.rpc_call_done = nfs4_delegreturn_done,
	.rpc_release = nfs4_delegreturn_release,
};

5029
static int _nfs4_proc_delegreturn(struct inode *inode, struct rpc_cred *cred, const nfs4_stateid *stateid, int issync)
5030 5031
{
	struct nfs4_delegreturndata *data;
5032
	struct nfs_server *server = NFS_SERVER(inode);
5033
	struct rpc_task *task;
5034 5035 5036 5037
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_DELEGRETURN],
		.rpc_cred = cred,
	};
T
Trond Myklebust 已提交
5038 5039
	struct rpc_task_setup task_setup_data = {
		.rpc_client = server->client,
5040
		.rpc_message = &msg,
T
Trond Myklebust 已提交
5041 5042 5043
		.callback_ops = &nfs4_delegreturn_ops,
		.flags = RPC_TASK_ASYNC,
	};
5044
	int status = 0;
5045

5046
	data = kzalloc(sizeof(*data), GFP_NOFS);
5047 5048
	if (data == NULL)
		return -ENOMEM;
5049
	nfs4_init_sequence(&data->args.seq_args, &data->res.seq_res, 1);
5050 5051
	data->args.fhandle = &data->fh;
	data->args.stateid = &data->stateid;
5052
	data->args.bitmask = server->cache_consistency_bitmask;
5053
	nfs_copy_fh(&data->fh, NFS_FH(inode));
5054
	nfs4_stateid_copy(&data->stateid, stateid);
5055 5056
	data->res.fattr = &data->fattr;
	data->res.server = server;
5057
	nfs_fattr_init(data->res.fattr);
5058
	data->timestamp = jiffies;
5059 5060
	data->rpc_status = 0;

T
Trond Myklebust 已提交
5061
	task_setup_data.callback_data = data;
5062 5063
	msg.rpc_argp = &data->args;
	msg.rpc_resp = &data->res;
T
Trond Myklebust 已提交
5064
	task = rpc_run_task(&task_setup_data);
5065
	if (IS_ERR(task))
5066
		return PTR_ERR(task);
5067 5068
	if (!issync)
		goto out;
5069
	status = nfs4_wait_for_completion_rpc_task(task);
5070 5071 5072
	if (status != 0)
		goto out;
	status = data->rpc_status;
5073 5074 5075 5076
	if (status == 0)
		nfs_post_op_update_inode_force_wcc(inode, &data->fattr);
	else
		nfs_refresh_inode(inode, &data->fattr);
5077
out:
5078
	rpc_put_task(task);
5079
	return status;
L
Linus Torvalds 已提交
5080 5081
}

5082
int nfs4_proc_delegreturn(struct inode *inode, struct rpc_cred *cred, const nfs4_stateid *stateid, int issync)
L
Linus Torvalds 已提交
5083 5084 5085 5086 5087
{
	struct nfs_server *server = NFS_SERVER(inode);
	struct nfs4_exception exception = { };
	int err;
	do {
5088
		err = _nfs4_proc_delegreturn(inode, cred, stateid, issync);
5089
		trace_nfs4_delegreturn(inode, err);
L
Linus Torvalds 已提交
5090 5091 5092 5093 5094 5095 5096 5097 5098 5099 5100 5101 5102 5103 5104 5105 5106 5107 5108 5109
		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)
{
5110
	freezable_schedule_timeout_killable_unsafe(timeout);
L
Linus Torvalds 已提交
5111 5112 5113 5114 5115 5116 5117 5118 5119 5120
	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);
5121
	struct nfs_client *clp = server->nfs_client;
T
Trond Myklebust 已提交
5122
	struct nfs_lockt_args arg = {
L
Linus Torvalds 已提交
5123
		.fh = NFS_FH(inode),
T
Trond Myklebust 已提交
5124
		.fl = request,
L
Linus Torvalds 已提交
5125
	};
T
Trond Myklebust 已提交
5126 5127
	struct nfs_lockt_res res = {
		.denied = request,
L
Linus Torvalds 已提交
5128 5129 5130 5131 5132 5133 5134 5135 5136 5137
	};
	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
5138
	arg.lock_owner.clientid = clp->cl_clientid;
5139 5140 5141 5142
	status = nfs4_set_lock_state(state, request);
	if (status != 0)
		goto out;
	lsp = request->fl_u.nfs4_fl.owner;
5143
	arg.lock_owner.id = lsp->ls_seqid.owner_id;
5144
	arg.lock_owner.s_dev = server->s_dev;
5145
	status = nfs4_call_sync(server->client, server, &msg, &arg.seq_args, &res.seq_res, 1);
T
Trond Myklebust 已提交
5146 5147 5148 5149 5150 5151
	switch (status) {
		case 0:
			request->fl_type = F_UNLCK;
			break;
		case -NFS4ERR_DENIED:
			status = 0;
L
Linus Torvalds 已提交
5152
	}
5153
	request->fl_ops->fl_release_private(request);
5154
	request->fl_ops = NULL;
5155
out:
L
Linus Torvalds 已提交
5156 5157 5158 5159 5160 5161 5162 5163 5164
	return status;
}

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

	do {
5165 5166 5167
		err = _nfs4_proc_getlk(state, cmd, request);
		trace_nfs4_get_lock(request, state, cmd, err);
		err = nfs4_handle_exception(NFS_SERVER(state->inode), err,
L
Linus Torvalds 已提交
5168 5169 5170 5171 5172 5173 5174 5175 5176 5177 5178 5179 5180 5181 5182 5183 5184 5185 5186 5187 5188
				&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;
}

5189
struct nfs4_unlockdata {
T
Trond Myklebust 已提交
5190 5191
	struct nfs_locku_args arg;
	struct nfs_locku_res res;
5192 5193
	struct nfs4_lock_state *lsp;
	struct nfs_open_context *ctx;
T
Trond Myklebust 已提交
5194 5195
	struct file_lock fl;
	const struct nfs_server *server;
5196
	unsigned long timestamp;
5197 5198
};

T
Trond Myklebust 已提交
5199 5200 5201 5202 5203 5204 5205 5206
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;

5207
	p = kzalloc(sizeof(*p), GFP_NOFS);
T
Trond Myklebust 已提交
5208 5209 5210 5211 5212
	if (p == NULL)
		return NULL;
	p->arg.fh = NFS_FH(inode);
	p->arg.fl = &p->fl;
	p->arg.seqid = seqid;
5213
	p->res.seqid = seqid;
T
Trond Myklebust 已提交
5214 5215 5216 5217 5218 5219 5220 5221 5222 5223
	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;
}

5224
static void nfs4_locku_release_calldata(void *data)
5225
{
5226
	struct nfs4_unlockdata *calldata = data;
T
Trond Myklebust 已提交
5227
	nfs_free_seqid(calldata->arg.seqid);
5228 5229 5230
	nfs4_put_lock_state(calldata->lsp);
	put_nfs_open_context(calldata->ctx);
	kfree(calldata);
5231 5232
}

5233
static void nfs4_locku_done(struct rpc_task *task, void *data)
5234
{
5235
	struct nfs4_unlockdata *calldata = data;
5236

5237 5238
	if (!nfs4_sequence_done(task, &calldata->res.seq_res))
		return;
5239 5240
	switch (task->tk_status) {
		case 0:
5241 5242
			nfs4_stateid_copy(&calldata->lsp->ls_stateid,
					&calldata->res.stateid);
5243
			renew_lease(calldata->server, calldata->timestamp);
5244
			break;
5245 5246
		case -NFS4ERR_BAD_STATEID:
		case -NFS4ERR_OLD_STATEID:
5247 5248 5249 5250
		case -NFS4ERR_STALE_STATEID:
		case -NFS4ERR_EXPIRED:
			break;
		default:
5251
			if (nfs4_async_handle_error(task, calldata->server, NULL) == -EAGAIN)
5252
				rpc_restart_call_prepare(task);
5253
	}
5254
	nfs_release_seqid(calldata->arg.seqid);
5255 5256
}

T
Trond Myklebust 已提交
5257
static void nfs4_locku_prepare(struct rpc_task *task, void *data)
5258
{
T
Trond Myklebust 已提交
5259
	struct nfs4_unlockdata *calldata = data;
5260

T
Trond Myklebust 已提交
5261
	if (nfs_wait_on_sequence(calldata->arg.seqid, task) != 0)
5262
		goto out_wait;
5263
	if (test_bit(NFS_LOCK_INITIALIZED, &calldata->lsp->ls_flags) == 0) {
5264
		/* Note: exit _without_ running nfs4_locku_done */
5265
		goto out_no_action;
5266
	}
5267
	calldata->timestamp = jiffies;
5268
	if (nfs4_setup_sequence(calldata->server,
5269
				&calldata->arg.seq_args,
5270 5271 5272
				&calldata->res.seq_res,
				task) != 0)
		nfs_release_seqid(calldata->arg.seqid);
5273 5274 5275 5276 5277
	return;
out_no_action:
	task->tk_action = NULL;
out_wait:
	nfs4_sequence_done(task, &calldata->res.seq_res);
5278 5279
}

5280
static const struct rpc_call_ops nfs4_locku_ops = {
T
Trond Myklebust 已提交
5281
	.rpc_call_prepare = nfs4_locku_prepare,
5282
	.rpc_call_done = nfs4_locku_done,
5283
	.rpc_release = nfs4_locku_release_calldata,
5284 5285
};

5286 5287 5288 5289 5290 5291
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;
5292 5293 5294 5295
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_LOCKU],
		.rpc_cred = ctx->cred,
	};
T
Trond Myklebust 已提交
5296 5297
	struct rpc_task_setup task_setup_data = {
		.rpc_client = NFS_CLIENT(lsp->ls_state->inode),
5298
		.rpc_message = &msg,
T
Trond Myklebust 已提交
5299
		.callback_ops = &nfs4_locku_ops,
5300
		.workqueue = nfsiod_workqueue,
T
Trond Myklebust 已提交
5301 5302
		.flags = RPC_TASK_ASYNC,
	};
5303

5304 5305 5306
	nfs4_state_protect(NFS_SERVER(lsp->ls_state->inode)->nfs_client,
		NFS_SP4_MACH_CRED_CLEANUP, &task_setup_data.rpc_client, &msg);

5307 5308 5309 5310 5311
	/* 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;

5312 5313 5314 5315 5316 5317
	data = nfs4_alloc_unlockdata(fl, ctx, lsp, seqid);
	if (data == NULL) {
		nfs_free_seqid(seqid);
		return ERR_PTR(-ENOMEM);
	}

5318
	nfs4_init_sequence(&data->arg.seq_args, &data->res.seq_res, 1);
5319 5320
	msg.rpc_argp = &data->arg;
	msg.rpc_resp = &data->res;
T
Trond Myklebust 已提交
5321 5322
	task_setup_data.callback_data = data;
	return rpc_run_task(&task_setup_data);
5323 5324
}

5325 5326
static int nfs4_proc_unlck(struct nfs4_state *state, int cmd, struct file_lock *request)
{
5327 5328 5329
	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
5330
	struct nfs_seqid *seqid;
L
Linus Torvalds 已提交
5331
	struct nfs4_lock_state *lsp;
5332 5333
	struct rpc_task *task;
	int status = 0;
5334
	unsigned char fl_flags = request->fl_flags;
5335

5336
	status = nfs4_set_lock_state(state, request);
5337 5338
	/* Unlock _before_ we do the RPC call */
	request->fl_flags |= FL_EXISTS;
5339 5340 5341
	/* Exclude nfs_delegation_claim_locks() */
	mutex_lock(&sp->so_delegreturn_mutex);
	/* Exclude nfs4_reclaim_open_stateid() - note nesting! */
5342 5343 5344
	down_read(&nfsi->rwsem);
	if (do_vfs_lock(request->fl_file, request) == -ENOENT) {
		up_read(&nfsi->rwsem);
5345
		mutex_unlock(&sp->so_delegreturn_mutex);
5346
		goto out;
5347 5348
	}
	up_read(&nfsi->rwsem);
5349
	mutex_unlock(&sp->so_delegreturn_mutex);
5350
	if (status != 0)
5351 5352
		goto out;
	/* Is this a delegated lock? */
5353
	lsp = request->fl_u.nfs4_fl.owner;
5354 5355
	if (test_bit(NFS_LOCK_INITIALIZED, &lsp->ls_flags) == 0)
		goto out;
5356
	seqid = nfs_alloc_seqid(&lsp->ls_seqid, GFP_KERNEL);
5357
	status = -ENOMEM;
T
Trond Myklebust 已提交
5358
	if (seqid == NULL)
5359
		goto out;
5360
	task = nfs4_do_unlck(request, nfs_file_open_context(request->fl_file), lsp, seqid);
5361 5362
	status = PTR_ERR(task);
	if (IS_ERR(task))
5363
		goto out;
5364
	status = nfs4_wait_for_completion_rpc_task(task);
5365
	rpc_put_task(task);
5366
out:
5367
	request->fl_flags = fl_flags;
5368
	trace_nfs4_unlock(request, state, F_SETLK, status);
L
Linus Torvalds 已提交
5369 5370 5371
	return status;
}

5372 5373 5374 5375 5376 5377
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;
5378
	unsigned long timestamp;
5379 5380
	int rpc_status;
	int cancelled;
5381
	struct nfs_server *server;
5382 5383 5384
};

static struct nfs4_lockdata *nfs4_alloc_lockdata(struct file_lock *fl,
5385 5386
		struct nfs_open_context *ctx, struct nfs4_lock_state *lsp,
		gfp_t gfp_mask)
L
Linus Torvalds 已提交
5387
{
5388 5389
	struct nfs4_lockdata *p;
	struct inode *inode = lsp->ls_state->inode;
L
Linus Torvalds 已提交
5390
	struct nfs_server *server = NFS_SERVER(inode);
5391

5392
	p = kzalloc(sizeof(*p), gfp_mask);
5393 5394 5395 5396 5397
	if (p == NULL)
		return NULL;

	p->arg.fh = NFS_FH(inode);
	p->arg.fl = &p->fl;
5398
	p->arg.open_seqid = nfs_alloc_seqid(&lsp->ls_state->owner->so_seqid, gfp_mask);
5399 5400
	if (p->arg.open_seqid == NULL)
		goto out_free;
5401
	p->arg.lock_seqid = nfs_alloc_seqid(&lsp->ls_seqid, gfp_mask);
5402
	if (p->arg.lock_seqid == NULL)
5403
		goto out_free_seqid;
5404
	p->arg.lock_stateid = &lsp->ls_stateid;
5405
	p->arg.lock_owner.clientid = server->nfs_client->cl_clientid;
5406
	p->arg.lock_owner.id = lsp->ls_seqid.owner_id;
5407
	p->arg.lock_owner.s_dev = server->s_dev;
5408
	p->res.lock_seqid = p->arg.lock_seqid;
5409
	p->lsp = lsp;
5410
	p->server = server;
5411 5412 5413 5414
	atomic_inc(&lsp->ls_count);
	p->ctx = get_nfs_open_context(ctx);
	memcpy(&p->fl, fl, sizeof(p->fl));
	return p;
5415 5416
out_free_seqid:
	nfs_free_seqid(p->arg.open_seqid);
5417 5418 5419 5420 5421 5422 5423 5424 5425
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;
5426

5427
	dprintk("%s: begin!\n", __func__);
5428
	if (nfs_wait_on_sequence(data->arg.lock_seqid, task) != 0)
5429
		goto out_wait;
5430 5431
	/* Do we need to do an open_to_lock_owner? */
	if (!(data->arg.lock_seqid->sequence->flags & NFS_SEQID_CONFIRMED)) {
5432
		if (nfs_wait_on_sequence(data->arg.open_seqid, task) != 0) {
5433
			goto out_release_lock_seqid;
5434
		}
5435
		data->arg.open_stateid = &state->open_stateid;
5436
		data->arg.new_lock_owner = 1;
5437
		data->res.open_seqid = data->arg.open_seqid;
5438 5439
	} else
		data->arg.new_lock_owner = 0;
5440 5441 5442 5443 5444
	if (!nfs4_valid_open_stateid(state)) {
		data->rpc_status = -EBADF;
		task->tk_action = NULL;
		goto out_release_open_seqid;
	}
5445
	data->timestamp = jiffies;
5446 5447
	if (nfs4_setup_sequence(data->server,
				&data->arg.seq_args,
5448
				&data->res.seq_res,
5449
				task) == 0)
5450
		return;
5451
out_release_open_seqid:
5452 5453 5454
	nfs_release_seqid(data->arg.open_seqid);
out_release_lock_seqid:
	nfs_release_seqid(data->arg.lock_seqid);
5455 5456
out_wait:
	nfs4_sequence_done(task, &data->res.seq_res);
5457
	dprintk("%s: done!, ret = %d\n", __func__, data->rpc_status);
5458 5459
}

5460 5461 5462 5463
static void nfs4_lock_done(struct rpc_task *task, void *calldata)
{
	struct nfs4_lockdata *data = calldata;

5464
	dprintk("%s: begin!\n", __func__);
5465

5466 5467
	if (!nfs4_sequence_done(task, &data->res.seq_res))
		return;
5468

5469 5470 5471 5472 5473 5474 5475 5476
	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) {
5477
		nfs4_stateid_copy(&data->lsp->ls_stateid, &data->res.stateid);
5478
		set_bit(NFS_LOCK_INITIALIZED, &data->lsp->ls_flags);
5479
		renew_lease(NFS_SERVER(data->ctx->dentry->d_inode), data->timestamp);
5480 5481
	}
out:
5482
	dprintk("%s: done, ret = %d!\n", __func__, data->rpc_status);
5483 5484 5485 5486 5487 5488
}

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

5489
	dprintk("%s: begin!\n", __func__);
5490
	nfs_free_seqid(data->arg.open_seqid);
5491 5492 5493 5494 5495
	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))
5496
			rpc_put_task_async(task);
5497
		dprintk("%s: cancelling lock!\n", __func__);
5498 5499 5500 5501 5502
	} else
		nfs_free_seqid(data->arg.lock_seqid);
	nfs4_put_lock_state(data->lsp);
	put_nfs_open_context(data->ctx);
	kfree(data);
5503
	dprintk("%s: done!\n", __func__);
5504 5505 5506 5507 5508 5509 5510 5511
}

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,
};

5512 5513 5514 5515 5516
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:
5517
		lsp->ls_seqid.flags &= ~NFS_SEQID_CONFIRMED;
5518
		if (new_lock_owner != 0 ||
5519
		   test_bit(NFS_LOCK_INITIALIZED, &lsp->ls_flags) != 0)
5520
			nfs4_schedule_stateid_recovery(server, lsp->ls_state);
5521 5522 5523
		break;
	case -NFS4ERR_STALE_STATEID:
		lsp->ls_seqid.flags &= ~NFS_SEQID_CONFIRMED;
5524 5525
	case -NFS4ERR_EXPIRED:
		nfs4_schedule_lease_recovery(server->nfs_client);
5526 5527 5528
	};
}

5529
static int _nfs4_do_setlk(struct nfs4_state *state, int cmd, struct file_lock *fl, int recovery_type)
5530 5531 5532
{
	struct nfs4_lockdata *data;
	struct rpc_task *task;
5533 5534 5535 5536
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_LOCK],
		.rpc_cred = state->owner->so_cred,
	};
T
Trond Myklebust 已提交
5537 5538
	struct rpc_task_setup task_setup_data = {
		.rpc_client = NFS_CLIENT(state->inode),
5539
		.rpc_message = &msg,
T
Trond Myklebust 已提交
5540
		.callback_ops = &nfs4_lock_ops,
5541
		.workqueue = nfsiod_workqueue,
T
Trond Myklebust 已提交
5542 5543
		.flags = RPC_TASK_ASYNC,
	};
5544 5545
	int ret;

5546
	dprintk("%s: begin!\n", __func__);
5547
	data = nfs4_alloc_lockdata(fl, nfs_file_open_context(fl->fl_file),
5548 5549
			fl->fl_u.nfs4_fl.owner,
			recovery_type == NFS_LOCK_NEW ? GFP_KERNEL : GFP_NOFS);
5550 5551 5552 5553
	if (data == NULL)
		return -ENOMEM;
	if (IS_SETLKW(cmd))
		data->arg.block = 1;
5554
	nfs4_init_sequence(&data->arg.seq_args, &data->res.seq_res, 1);
5555 5556
	msg.rpc_argp = &data->arg;
	msg.rpc_resp = &data->res;
T
Trond Myklebust 已提交
5557
	task_setup_data.callback_data = data;
5558 5559 5560 5561 5562
	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
5563
	task = rpc_run_task(&task_setup_data);
5564
	if (IS_ERR(task))
5565 5566 5567 5568
		return PTR_ERR(task);
	ret = nfs4_wait_for_completion_rpc_task(task);
	if (ret == 0) {
		ret = data->rpc_status;
5569 5570 5571
		if (ret)
			nfs4_handle_setlk_error(data->server, data->lsp,
					data->arg.new_lock_owner, ret);
5572 5573
	} else
		data->cancelled = 1;
5574
	rpc_put_task(task);
5575
	dprintk("%s: done, ret = %d!\n", __func__, ret);
5576
	return ret;
L
Linus Torvalds 已提交
5577 5578 5579 5580
}

static int nfs4_lock_reclaim(struct nfs4_state *state, struct file_lock *request)
{
5581
	struct nfs_server *server = NFS_SERVER(state->inode);
5582 5583 5584
	struct nfs4_exception exception = {
		.inode = state->inode,
	};
5585 5586 5587
	int err;

	do {
5588 5589 5590
		/* Cache the lock if possible... */
		if (test_bit(NFS_DELEGATED_STATE, &state->flags) != 0)
			return 0;
5591
		err = _nfs4_do_setlk(state, F_SETLK, request, NFS_LOCK_RECLAIM);
5592
		trace_nfs4_lock_reclaim(request, state, F_SETLK, err);
5593
		if (err != -NFS4ERR_DELAY)
5594 5595 5596 5597
			break;
		nfs4_handle_exception(server, err, &exception);
	} while (exception.retry);
	return err;
L
Linus Torvalds 已提交
5598 5599 5600 5601
}

static int nfs4_lock_expired(struct nfs4_state *state, struct file_lock *request)
{
5602
	struct nfs_server *server = NFS_SERVER(state->inode);
5603 5604 5605
	struct nfs4_exception exception = {
		.inode = state->inode,
	};
5606 5607
	int err;

5608 5609 5610
	err = nfs4_set_lock_state(state, request);
	if (err != 0)
		return err;
5611
	if (!recover_lost_locks) {
5612 5613 5614
		set_bit(NFS_LOCK_LOST, &request->fl_u.nfs4_fl.owner->ls_flags);
		return 0;
	}
5615
	do {
5616 5617
		if (test_bit(NFS_DELEGATED_STATE, &state->flags) != 0)
			return 0;
5618
		err = _nfs4_do_setlk(state, F_SETLK, request, NFS_LOCK_EXPIRED);
5619
		trace_nfs4_lock_expired(request, state, F_SETLK, err);
5620 5621 5622 5623 5624 5625 5626 5627
		switch (err) {
		default:
			goto out;
		case -NFS4ERR_GRACE:
		case -NFS4ERR_DELAY:
			nfs4_handle_exception(server, err, &exception);
			err = 0;
		}
5628
	} while (exception.retry);
5629
out:
5630
	return err;
L
Linus Torvalds 已提交
5631 5632
}

5633
#if defined(CONFIG_NFS_V4_1)
5634 5635 5636 5637 5638 5639 5640 5641
/**
 * 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.
 */
5642
static int nfs41_check_expired_locks(struct nfs4_state *state)
5643
{
5644
	int status, ret = -NFS4ERR_BAD_STATEID;
5645
	struct nfs4_lock_state *lsp;
5646 5647
	struct nfs_server *server = NFS_SERVER(state->inode);

5648
	list_for_each_entry(lsp, &state->lock_states, ls_locks) {
5649
		if (test_bit(NFS_LOCK_INITIALIZED, &lsp->ls_flags)) {
5650 5651 5652 5653 5654
			struct rpc_cred *cred = lsp->ls_state->owner->so_cred;

			status = nfs41_test_stateid(server,
					&lsp->ls_stateid,
					cred);
5655
			trace_nfs4_test_lock_stateid(state, lsp, status);
5656
			if (status != NFS_OK) {
5657 5658
				/* Free the stateid unless the server
				 * informs us the stateid is unrecognized. */
5659 5660
				if (status != -NFS4ERR_BAD_STATEID)
					nfs41_free_stateid(server,
5661 5662
							&lsp->ls_stateid,
							cred);
5663
				clear_bit(NFS_LOCK_INITIALIZED, &lsp->ls_flags);
5664 5665 5666 5667 5668 5669 5670 5671 5672 5673 5674 5675 5676 5677
				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);
5678 5679 5680
	if (status != NFS_OK)
		status = nfs4_lock_expired(state, request);
	return status;
5681 5682 5683
}
#endif

L
Linus Torvalds 已提交
5684 5685
static int _nfs4_proc_setlk(struct nfs4_state *state, int cmd, struct file_lock *request)
{
5686
	struct nfs4_state_owner *sp = state->owner;
5687
	struct nfs_inode *nfsi = NFS_I(state->inode);
5688
	unsigned char fl_flags = request->fl_flags;
5689
	unsigned int seq;
5690
	int status = -ENOLCK;
L
Linus Torvalds 已提交
5691

5692 5693 5694
	if ((fl_flags & FL_POSIX) &&
			!test_bit(NFS_STATE_POSIX_LOCKS, &state->flags))
		goto out;
5695 5696 5697 5698
	/* Is this a delegated open? */
	status = nfs4_set_lock_state(state, request);
	if (status != 0)
		goto out;
5699 5700 5701 5702
	request->fl_flags |= FL_ACCESS;
	status = do_vfs_lock(request->fl_file, request);
	if (status < 0)
		goto out;
5703
	down_read(&nfsi->rwsem);
5704 5705 5706
	if (test_bit(NFS_DELEGATED_STATE, &state->flags)) {
		/* Yes: cache locks! */
		/* ...but avoid races with delegation recall... */
5707 5708 5709
		request->fl_flags = fl_flags & ~FL_SLEEP;
		status = do_vfs_lock(request->fl_file, request);
		goto out_unlock;
5710
	}
5711 5712
	seq = raw_seqcount_begin(&sp->so_reclaim_seqcount);
	up_read(&nfsi->rwsem);
5713
	status = _nfs4_do_setlk(state, cmd, request, NFS_LOCK_NEW);
5714
	if (status != 0)
5715 5716 5717 5718
		goto out;
	down_read(&nfsi->rwsem);
	if (read_seqcount_retry(&sp->so_reclaim_seqcount, seq)) {
		status = -NFS4ERR_DELAY;
5719
		goto out_unlock;
5720
	}
5721
	/* Note: we always want to sleep here! */
5722
	request->fl_flags = fl_flags | FL_SLEEP;
5723
	if (do_vfs_lock(request->fl_file, request) < 0)
5724 5725
		printk(KERN_WARNING "NFS: %s: VFS is out of sync with lock "
			"manager!\n", __func__);
5726
out_unlock:
5727
	up_read(&nfsi->rwsem);
5728 5729
out:
	request->fl_flags = fl_flags;
L
Linus Torvalds 已提交
5730 5731 5732 5733 5734
	return status;
}

static int nfs4_proc_setlk(struct nfs4_state *state, int cmd, struct file_lock *request)
{
5735 5736
	struct nfs4_exception exception = {
		.state = state,
5737
		.inode = state->inode,
5738
	};
L
Linus Torvalds 已提交
5739 5740 5741
	int err;

	do {
5742
		err = _nfs4_proc_setlk(state, cmd, request);
5743
		trace_nfs4_set_lock(request, state, cmd, err);
5744 5745
		if (err == -NFS4ERR_DENIED)
			err = -EAGAIN;
L
Linus Torvalds 已提交
5746
		err = nfs4_handle_exception(NFS_SERVER(state->inode),
5747
				err, &exception);
L
Linus Torvalds 已提交
5748 5749 5750 5751 5752 5753 5754 5755 5756 5757 5758 5759 5760
	} 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 */
5761
	ctx = nfs_file_open_context(filp);
L
Linus Torvalds 已提交
5762 5763 5764 5765 5766
	state = ctx->state;

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

5767 5768 5769 5770 5771
	if (IS_GETLK(cmd)) {
		if (state != NULL)
			return nfs4_proc_getlk(state, F_GETLK, request);
		return 0;
	}
L
Linus Torvalds 已提交
5772 5773 5774 5775

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

5776 5777 5778 5779 5780
	if (request->fl_type == F_UNLCK) {
		if (state != NULL)
			return nfs4_proc_unlck(state, cmd, request);
		return 0;
	}
L
Linus Torvalds 已提交
5781

5782 5783
	if (state == NULL)
		return -ENOLCK;
5784 5785 5786 5787
	/*
	 * Don't rely on the VFS having checked the file open mode,
	 * since it won't do this for flock() locks.
	 */
5788
	switch (request->fl_type) {
5789 5790 5791 5792 5793 5794 5795 5796 5797
	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
5798 5799 5800 5801 5802 5803 5804 5805 5806 5807 5808 5809
	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;
}

5810
int nfs4_lock_delegation_recall(struct file_lock *fl, struct nfs4_state *state, const nfs4_stateid *stateid)
5811 5812 5813 5814 5815 5816
{
	struct nfs_server *server = NFS_SERVER(state->inode);
	int err;

	err = nfs4_set_lock_state(state, fl);
	if (err != 0)
5817
		return err;
5818
	err = _nfs4_do_setlk(state, F_SETLK, fl, NFS_LOCK_NEW);
5819
	return nfs4_handle_delegation_recall_error(server, state, stateid, err);
5820
}
5821

5822 5823
struct nfs_release_lockowner_data {
	struct nfs4_lock_state *lsp;
5824
	struct nfs_server *server;
5825
	struct nfs_release_lockowner_args args;
5826 5827
	struct nfs4_sequence_args seq_args;
	struct nfs4_sequence_res seq_res;
5828
	unsigned long timestamp;
5829 5830
};

5831 5832 5833 5834 5835
static void nfs4_release_lockowner_prepare(struct rpc_task *task, void *calldata)
{
	struct nfs_release_lockowner_data *data = calldata;
	nfs40_setup_sequence(data->server,
				&data->seq_args, &data->seq_res, task);
5836
	data->timestamp = jiffies;
5837 5838 5839 5840 5841
}

static void nfs4_release_lockowner_done(struct rpc_task *task, void *calldata)
{
	struct nfs_release_lockowner_data *data = calldata;
5842 5843
	struct nfs_server *server = data->server;

5844
	nfs40_sequence_done(task, &data->seq_res);
5845 5846 5847 5848 5849 5850 5851 5852 5853 5854 5855 5856

	switch (task->tk_status) {
	case 0:
		renew_lease(server, data->timestamp);
		break;
	case -NFS4ERR_STALE_CLIENTID:
	case -NFS4ERR_EXPIRED:
	case -NFS4ERR_LEASE_MOVED:
	case -NFS4ERR_DELAY:
		if (nfs4_async_handle_error(task, server, NULL) == -EAGAIN)
			rpc_restart_call_prepare(task);
	}
5857 5858
}

5859 5860
static void nfs4_release_lockowner_release(void *calldata)
{
5861
	struct nfs_release_lockowner_data *data = calldata;
5862
	nfs4_free_lock_state(data->server, data->lsp);
5863 5864 5865
	kfree(calldata);
}

5866
static const struct rpc_call_ops nfs4_release_lockowner_ops = {
5867 5868
	.rpc_call_prepare = nfs4_release_lockowner_prepare,
	.rpc_call_done = nfs4_release_lockowner_done,
5869 5870 5871
	.rpc_release = nfs4_release_lockowner_release,
};

5872
static int nfs4_release_lockowner(struct nfs_server *server, struct nfs4_lock_state *lsp)
5873
{
5874
	struct nfs_release_lockowner_data *data;
5875 5876 5877 5878 5879
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_RELEASE_LOCKOWNER],
	};

	if (server->nfs_client->cl_mvops->minor_version != 0)
5880
		return -EINVAL;
5881

5882 5883 5884
	data = kmalloc(sizeof(*data), GFP_NOFS);
	if (!data)
		return -ENOMEM;
5885
	nfs4_init_sequence(&data->seq_args, &data->seq_res, 0);
5886
	data->lsp = lsp;
5887
	data->server = server;
5888 5889 5890
	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;
5891

5892 5893 5894
	msg.rpc_argp = &data->args;
	rpc_call_async(server->client, &msg, 0, &nfs4_release_lockowner_ops, data);
	return 0;
5895 5896
}

5897 5898
#define XATTR_NAME_NFSV4_ACL "system.nfs4_acl"

5899 5900 5901
static int nfs4_xattr_set_nfs4_acl(struct dentry *dentry, const char *key,
				   const void *buf, size_t buflen,
				   int flags, int type)
5902
{
5903 5904
	if (strcmp(key, "") != 0)
		return -EINVAL;
5905

5906
	return nfs4_proc_set_acl(dentry->d_inode, buf, buflen);
5907 5908
}

5909 5910
static int nfs4_xattr_get_nfs4_acl(struct dentry *dentry, const char *key,
				   void *buf, size_t buflen, int type)
5911
{
5912 5913
	if (strcmp(key, "") != 0)
		return -EINVAL;
5914

5915
	return nfs4_proc_get_acl(dentry->d_inode, buf, buflen);
5916 5917
}

5918 5919 5920
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)
5921
{
5922
	size_t len = sizeof(XATTR_NAME_NFSV4_ACL);
5923

5924 5925
	if (!nfs4_server_supports_acls(NFS_SERVER(dentry->d_inode)))
		return 0;
5926 5927 5928

	if (list && len <= list_len)
		memcpy(list, XATTR_NAME_NFSV4_ACL, len);
5929
	return len;
5930 5931
}

5932 5933 5934 5935 5936 5937 5938 5939 5940 5941 5942 5943 5944 5945 5946 5947 5948 5949 5950 5951 5952 5953 5954 5955 5956 5957 5958 5959 5960 5961 5962 5963 5964 5965 5966 5967 5968 5969 5970 5971 5972 5973 5974 5975 5976 5977 5978
#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


5979 5980 5981
/*
 * nfs_fhget will use either the mounted_on_fileid or the fileid
 */
5982 5983
static void nfs_fixup_referral_attributes(struct nfs_fattr *fattr)
{
5984 5985 5986
	if (!(((fattr->valid & NFS_ATTR_FATTR_MOUNTED_ON_FILEID) ||
	       (fattr->valid & NFS_ATTR_FATTR_FILEID)) &&
	      (fattr->valid & NFS_ATTR_FATTR_FSID) &&
5987
	      (fattr->valid & NFS_ATTR_FATTR_V4_LOCATIONS)))
5988 5989 5990
		return;

	fattr->valid |= NFS_ATTR_FATTR_TYPE | NFS_ATTR_FATTR_MODE |
5991
		NFS_ATTR_FATTR_NLINK | NFS_ATTR_FATTR_V4_REFERRAL;
5992 5993 5994 5995
	fattr->mode = S_IFDIR | S_IRUGO | S_IXUGO;
	fattr->nlink = 2;
}

5996 5997 5998 5999
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)
6000 6001
{
	struct nfs_server *server = NFS_SERVER(dir);
6002
	u32 bitmask[3] = {
6003
		[0] = FATTR4_WORD0_FSID | FATTR4_WORD0_FS_LOCATIONS,
6004 6005 6006
	};
	struct nfs4_fs_locations_arg args = {
		.dir_fh = NFS_FH(dir),
6007
		.name = name,
6008 6009 6010
		.page = page,
		.bitmask = bitmask,
	};
B
Benny Halevy 已提交
6011 6012 6013
	struct nfs4_fs_locations_res res = {
		.fs_locations = fs_locations,
	};
6014 6015 6016
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_FS_LOCATIONS],
		.rpc_argp = &args,
B
Benny Halevy 已提交
6017
		.rpc_resp = &res,
6018 6019 6020
	};
	int status;

6021
	dprintk("%s: start\n", __func__);
6022 6023 6024 6025 6026 6027 6028 6029

	/* 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;

6030
	nfs_fattr_init(&fs_locations->fattr);
6031
	fs_locations->server = server;
6032
	fs_locations->nlocations = 0;
6033
	status = nfs4_call_sync(client, server, &msg, &args.seq_args, &res.seq_res, 0);
6034
	dprintk("%s: returned status = %d\n", __func__, status);
6035 6036 6037
	return status;
}

6038 6039 6040 6041
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)
6042 6043 6044 6045
{
	struct nfs4_exception exception = { };
	int err;
	do {
6046 6047 6048 6049
		err = _nfs4_proc_fs_locations(client, dir, name,
				fs_locations, page);
		trace_nfs4_get_fs_locations(dir, name, err);
		err = nfs4_handle_exception(NFS_SERVER(dir), err,
6050 6051 6052 6053 6054
				&exception);
	} while (exception.retry);
	return err;
}

6055 6056 6057 6058 6059 6060 6061 6062 6063 6064 6065 6066 6067 6068 6069 6070 6071 6072 6073 6074 6075 6076 6077 6078 6079 6080 6081 6082 6083 6084 6085 6086 6087 6088 6089 6090 6091 6092 6093 6094 6095 6096 6097 6098 6099 6100 6101 6102 6103 6104 6105 6106 6107 6108 6109 6110 6111 6112 6113 6114 6115 6116 6117 6118 6119 6120 6121 6122 6123 6124 6125 6126 6127 6128 6129 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 6155 6156 6157 6158 6159 6160 6161 6162 6163 6164 6165 6166 6167 6168 6169 6170 6171 6172 6173 6174 6175 6176 6177 6178 6179 6180 6181 6182 6183 6184 6185 6186 6187 6188 6189 6190 6191 6192 6193 6194 6195 6196 6197 6198 6199 6200 6201 6202 6203 6204 6205
/*
 * This operation also signals the server that this client is
 * performing migration recovery.  The server can stop returning
 * NFS4ERR_LEASE_MOVED to this client.  A RENEW operation is
 * appended to this compound to identify the client ID which is
 * performing recovery.
 */
static int _nfs40_proc_get_locations(struct inode *inode,
				     struct nfs4_fs_locations *locations,
				     struct page *page, struct rpc_cred *cred)
{
	struct nfs_server *server = NFS_SERVER(inode);
	struct rpc_clnt *clnt = server->client;
	u32 bitmask[2] = {
		[0] = FATTR4_WORD0_FSID | FATTR4_WORD0_FS_LOCATIONS,
	};
	struct nfs4_fs_locations_arg args = {
		.clientid	= server->nfs_client->cl_clientid,
		.fh		= NFS_FH(inode),
		.page		= page,
		.bitmask	= bitmask,
		.migration	= 1,		/* skip LOOKUP */
		.renew		= 1,		/* append RENEW */
	};
	struct nfs4_fs_locations_res res = {
		.fs_locations	= locations,
		.migration	= 1,
		.renew		= 1,
	};
	struct rpc_message msg = {
		.rpc_proc	= &nfs4_procedures[NFSPROC4_CLNT_FS_LOCATIONS],
		.rpc_argp	= &args,
		.rpc_resp	= &res,
		.rpc_cred	= cred,
	};
	unsigned long now = jiffies;
	int status;

	nfs_fattr_init(&locations->fattr);
	locations->server = server;
	locations->nlocations = 0;

	nfs4_init_sequence(&args.seq_args, &res.seq_res, 0);
	nfs4_set_sequence_privileged(&args.seq_args);
	status = nfs4_call_sync_sequence(clnt, server, &msg,
					&args.seq_args, &res.seq_res);
	if (status)
		return status;

	renew_lease(server, now);
	return 0;
}

#ifdef CONFIG_NFS_V4_1

/*
 * This operation also signals the server that this client is
 * performing migration recovery.  The server can stop asserting
 * SEQ4_STATUS_LEASE_MOVED for this client.  The client ID
 * performing this operation is identified in the SEQUENCE
 * operation in this compound.
 *
 * When the client supports GETATTR(fs_locations_info), it can
 * be plumbed in here.
 */
static int _nfs41_proc_get_locations(struct inode *inode,
				     struct nfs4_fs_locations *locations,
				     struct page *page, struct rpc_cred *cred)
{
	struct nfs_server *server = NFS_SERVER(inode);
	struct rpc_clnt *clnt = server->client;
	u32 bitmask[2] = {
		[0] = FATTR4_WORD0_FSID | FATTR4_WORD0_FS_LOCATIONS,
	};
	struct nfs4_fs_locations_arg args = {
		.fh		= NFS_FH(inode),
		.page		= page,
		.bitmask	= bitmask,
		.migration	= 1,		/* skip LOOKUP */
	};
	struct nfs4_fs_locations_res res = {
		.fs_locations	= locations,
		.migration	= 1,
	};
	struct rpc_message msg = {
		.rpc_proc	= &nfs4_procedures[NFSPROC4_CLNT_FS_LOCATIONS],
		.rpc_argp	= &args,
		.rpc_resp	= &res,
		.rpc_cred	= cred,
	};
	int status;

	nfs_fattr_init(&locations->fattr);
	locations->server = server;
	locations->nlocations = 0;

	nfs4_init_sequence(&args.seq_args, &res.seq_res, 0);
	nfs4_set_sequence_privileged(&args.seq_args);
	status = nfs4_call_sync_sequence(clnt, server, &msg,
					&args.seq_args, &res.seq_res);
	if (status == NFS4_OK &&
	    res.seq_res.sr_status_flags & SEQ4_STATUS_LEASE_MOVED)
		status = -NFS4ERR_LEASE_MOVED;
	return status;
}

#endif	/* CONFIG_NFS_V4_1 */

/**
 * nfs4_proc_get_locations - discover locations for a migrated FSID
 * @inode: inode on FSID that is migrating
 * @locations: result of query
 * @page: buffer
 * @cred: credential to use for this operation
 *
 * Returns NFS4_OK on success, a negative NFS4ERR status code if the
 * operation failed, or a negative errno if a local error occurred.
 *
 * On success, "locations" is filled in, but if the server has
 * no locations information, NFS_ATTR_FATTR_V4_LOCATIONS is not
 * asserted.
 *
 * -NFS4ERR_LEASE_MOVED is returned if the server still has leases
 * from this client that require migration recovery.
 */
int nfs4_proc_get_locations(struct inode *inode,
			    struct nfs4_fs_locations *locations,
			    struct page *page, struct rpc_cred *cred)
{
	struct nfs_server *server = NFS_SERVER(inode);
	struct nfs_client *clp = server->nfs_client;
	const struct nfs4_mig_recovery_ops *ops =
					clp->cl_mvops->mig_recovery_ops;
	struct nfs4_exception exception = { };
	int status;

	dprintk("%s: FSID %llx:%llx on \"%s\"\n", __func__,
		(unsigned long long)server->fsid.major,
		(unsigned long long)server->fsid.minor,
		clp->cl_hostname);
	nfs_display_fhandle(NFS_FH(inode), __func__);

	do {
		status = ops->get_locations(inode, locations, page, cred);
		if (status != -NFS4ERR_DELAY)
			break;
		nfs4_handle_exception(server, status, &exception);
	} while (exception.retry);
	return status;
}

6206 6207 6208 6209 6210 6211 6212 6213 6214 6215 6216 6217 6218 6219 6220 6221 6222 6223 6224 6225 6226 6227 6228 6229 6230 6231 6232 6233 6234 6235 6236 6237 6238 6239 6240 6241 6242 6243 6244 6245 6246 6247 6248 6249 6250 6251 6252 6253 6254 6255 6256 6257 6258 6259 6260 6261 6262 6263 6264 6265 6266 6267 6268 6269 6270 6271 6272 6273 6274 6275 6276 6277 6278 6279 6280 6281 6282 6283 6284 6285 6286 6287 6288 6289 6290 6291 6292 6293 6294 6295 6296 6297 6298 6299 6300 6301 6302 6303 6304 6305 6306 6307 6308 6309 6310 6311 6312 6313 6314 6315 6316 6317 6318 6319 6320 6321 6322 6323 6324 6325 6326 6327 6328 6329 6330 6331
/*
 * This operation also signals the server that this client is
 * performing "lease moved" recovery.  The server can stop
 * returning NFS4ERR_LEASE_MOVED to this client.  A RENEW operation
 * is appended to this compound to identify the client ID which is
 * performing recovery.
 */
static int _nfs40_proc_fsid_present(struct inode *inode, struct rpc_cred *cred)
{
	struct nfs_server *server = NFS_SERVER(inode);
	struct nfs_client *clp = NFS_SERVER(inode)->nfs_client;
	struct rpc_clnt *clnt = server->client;
	struct nfs4_fsid_present_arg args = {
		.fh		= NFS_FH(inode),
		.clientid	= clp->cl_clientid,
		.renew		= 1,		/* append RENEW */
	};
	struct nfs4_fsid_present_res res = {
		.renew		= 1,
	};
	struct rpc_message msg = {
		.rpc_proc	= &nfs4_procedures[NFSPROC4_CLNT_FSID_PRESENT],
		.rpc_argp	= &args,
		.rpc_resp	= &res,
		.rpc_cred	= cred,
	};
	unsigned long now = jiffies;
	int status;

	res.fh = nfs_alloc_fhandle();
	if (res.fh == NULL)
		return -ENOMEM;

	nfs4_init_sequence(&args.seq_args, &res.seq_res, 0);
	nfs4_set_sequence_privileged(&args.seq_args);
	status = nfs4_call_sync_sequence(clnt, server, &msg,
						&args.seq_args, &res.seq_res);
	nfs_free_fhandle(res.fh);
	if (status)
		return status;

	do_renew_lease(clp, now);
	return 0;
}

#ifdef CONFIG_NFS_V4_1

/*
 * This operation also signals the server that this client is
 * performing "lease moved" recovery.  The server can stop asserting
 * SEQ4_STATUS_LEASE_MOVED for this client.  The client ID performing
 * this operation is identified in the SEQUENCE operation in this
 * compound.
 */
static int _nfs41_proc_fsid_present(struct inode *inode, struct rpc_cred *cred)
{
	struct nfs_server *server = NFS_SERVER(inode);
	struct rpc_clnt *clnt = server->client;
	struct nfs4_fsid_present_arg args = {
		.fh		= NFS_FH(inode),
	};
	struct nfs4_fsid_present_res res = {
	};
	struct rpc_message msg = {
		.rpc_proc	= &nfs4_procedures[NFSPROC4_CLNT_FSID_PRESENT],
		.rpc_argp	= &args,
		.rpc_resp	= &res,
		.rpc_cred	= cred,
	};
	int status;

	res.fh = nfs_alloc_fhandle();
	if (res.fh == NULL)
		return -ENOMEM;

	nfs4_init_sequence(&args.seq_args, &res.seq_res, 0);
	nfs4_set_sequence_privileged(&args.seq_args);
	status = nfs4_call_sync_sequence(clnt, server, &msg,
						&args.seq_args, &res.seq_res);
	nfs_free_fhandle(res.fh);
	if (status == NFS4_OK &&
	    res.seq_res.sr_status_flags & SEQ4_STATUS_LEASE_MOVED)
		status = -NFS4ERR_LEASE_MOVED;
	return status;
}

#endif	/* CONFIG_NFS_V4_1 */

/**
 * nfs4_proc_fsid_present - Is this FSID present or absent on server?
 * @inode: inode on FSID to check
 * @cred: credential to use for this operation
 *
 * Server indicates whether the FSID is present, moved, or not
 * recognized.  This operation is necessary to clear a LEASE_MOVED
 * condition for this client ID.
 *
 * Returns NFS4_OK if the FSID is present on this server,
 * -NFS4ERR_MOVED if the FSID is no longer present, a negative
 *  NFS4ERR code if some error occurred on the server, or a
 *  negative errno if a local failure occurred.
 */
int nfs4_proc_fsid_present(struct inode *inode, struct rpc_cred *cred)
{
	struct nfs_server *server = NFS_SERVER(inode);
	struct nfs_client *clp = server->nfs_client;
	const struct nfs4_mig_recovery_ops *ops =
					clp->cl_mvops->mig_recovery_ops;
	struct nfs4_exception exception = { };
	int status;

	dprintk("%s: FSID %llx:%llx on \"%s\"\n", __func__,
		(unsigned long long)server->fsid.major,
		(unsigned long long)server->fsid.minor,
		clp->cl_hostname);
	nfs_display_fhandle(NFS_FH(inode), __func__);

	do {
		status = ops->fsid_present(inode, cred);
		if (status != -NFS4ERR_DELAY)
			break;
		nfs4_handle_exception(server, status, &exception);
	} while (exception.retry);
	return status;
}

6332
/**
6333 6334 6335 6336 6337
 * If 'use_integrity' is true and the state managment nfs_client
 * cl_rpcclient is using krb5i/p, use the integrity protected cl_rpcclient
 * and the machine credential as per RFC3530bis and RFC5661 Security
 * Considerations sections. Otherwise, just use the user cred with the
 * filesystem's rpc_client.
6338
 */
6339
static int _nfs4_proc_secinfo(struct inode *dir, const struct qstr *name, struct nfs4_secinfo_flavors *flavors, bool use_integrity)
B
Bryan Schumaker 已提交
6340 6341 6342 6343 6344 6345 6346 6347 6348 6349 6350 6351 6352 6353
{
	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,
	};
6354
	struct rpc_clnt *clnt = NFS_SERVER(dir)->client;
6355
	struct rpc_cred *cred = NULL;
6356 6357 6358

	if (use_integrity) {
		clnt = NFS_SERVER(dir)->nfs_client->cl_rpcclient;
6359 6360
		cred = nfs4_get_clid_cred(NFS_SERVER(dir)->nfs_client);
		msg.rpc_cred = cred;
6361
	}
B
Bryan Schumaker 已提交
6362 6363

	dprintk("NFS call  secinfo %s\n", name->name);
6364 6365 6366 6367

	nfs4_state_protect(NFS_SERVER(dir)->nfs_client,
		NFS_SP4_MACH_CRED_SECINFO, &clnt, &msg);

6368 6369
	status = nfs4_call_sync(clnt, NFS_SERVER(dir), &msg, &args.seq_args,
				&res.seq_res, 0);
B
Bryan Schumaker 已提交
6370
	dprintk("NFS reply  secinfo: %d\n", status);
6371

6372 6373
	if (cred)
		put_rpccred(cred);
6374

B
Bryan Schumaker 已提交
6375 6376 6377
	return status;
}

6378 6379
int nfs4_proc_secinfo(struct inode *dir, const struct qstr *name,
		      struct nfs4_secinfo_flavors *flavors)
B
Bryan Schumaker 已提交
6380 6381 6382 6383
{
	struct nfs4_exception exception = { };
	int err;
	do {
6384 6385 6386 6387 6388 6389 6390 6391 6392 6393 6394 6395 6396 6397 6398
		err = -NFS4ERR_WRONGSEC;

		/* try to use integrity protection with machine cred */
		if (_nfs4_is_integrity_protected(NFS_SERVER(dir)->nfs_client))
			err = _nfs4_proc_secinfo(dir, name, flavors, true);

		/*
		 * if unable to use integrity protection, or SECINFO with
		 * integrity protection returns NFS4ERR_WRONGSEC (which is
		 * disallowed by spec, but exists in deployed servers) use
		 * the current filesystem's rpc_client and the user cred.
		 */
		if (err == -NFS4ERR_WRONGSEC)
			err = _nfs4_proc_secinfo(dir, name, flavors, false);

6399 6400
		trace_nfs4_secinfo(dir, name, err);
		err = nfs4_handle_exception(NFS_SERVER(dir), err,
B
Bryan Schumaker 已提交
6401 6402 6403 6404 6405
				&exception);
	} while (exception.retry);
	return err;
}

6406
#ifdef CONFIG_NFS_V4_1
6407 6408 6409 6410 6411 6412 6413 6414 6415 6416 6417 6418 6419 6420 6421 6422 6423 6424 6425
/*
 * 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;
}

6426
static bool
6427 6428
nfs41_same_server_scope(struct nfs41_server_scope *a,
			struct nfs41_server_scope *b)
6429 6430 6431 6432 6433 6434 6435 6436
{
	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;
}

6437 6438 6439 6440 6441 6442
/*
 * nfs4_proc_bind_conn_to_session()
 *
 * The 4.1 client currently uses the same TCP connection for the
 * fore and backchannel.
 */
6443
int nfs4_proc_bind_conn_to_session(struct nfs_client *clp, struct rpc_cred *cred)
6444 6445 6446 6447 6448 6449 6450 6451
{
	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,
6452
		.rpc_cred = cred,
6453 6454 6455 6456 6457 6458 6459 6460 6461 6462 6463
	};

	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);
6464
	trace_nfs4_bind_conn_to_session(clp, status);
6465 6466 6467 6468 6469 6470 6471 6472 6473 6474 6475 6476 6477 6478 6479 6480 6481 6482 6483 6484 6485 6486 6487 6488 6489 6490 6491
	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
6492
/*
6493 6494
 * Minimum set of SP4_MACH_CRED operations from RFC 5661 in the enforce map
 * and operations we'd like to see to enable certain features in the allow map
6495 6496 6497 6498 6499 6500 6501 6502 6503
 */
static const struct nfs41_state_protection nfs4_sp4_mach_cred_request = {
	.how = SP4_MACH_CRED,
	.enforce.u.words = {
		[1] = 1 << (OP_BIND_CONN_TO_SESSION - 32) |
		      1 << (OP_EXCHANGE_ID - 32) |
		      1 << (OP_CREATE_SESSION - 32) |
		      1 << (OP_DESTROY_SESSION - 32) |
		      1 << (OP_DESTROY_CLIENTID - 32)
6504 6505 6506
	},
	.allow.u.words = {
		[0] = 1 << (OP_CLOSE) |
6507 6508
		      1 << (OP_LOCKU) |
		      1 << (OP_COMMIT),
6509
		[1] = 1 << (OP_SECINFO - 32) |
6510 6511
		      1 << (OP_SECINFO_NO_NAME - 32) |
		      1 << (OP_TEST_STATEID - 32) |
6512 6513
		      1 << (OP_FREE_STATEID - 32) |
		      1 << (OP_WRITE - 32)
6514 6515 6516 6517 6518 6519
	}
};

/*
 * Select the state protection mode for client `clp' given the server results
 * from exchange_id in `sp'.
B
Benny Halevy 已提交
6520
 *
6521 6522 6523 6524 6525 6526 6527 6528 6529 6530 6531 6532 6533 6534 6535 6536 6537 6538 6539 6540 6541 6542 6543 6544 6545 6546 6547 6548 6549 6550 6551 6552 6553 6554 6555 6556 6557 6558 6559 6560 6561 6562 6563 6564 6565 6566 6567 6568 6569 6570 6571 6572
 * Returns 0 on success, negative errno otherwise.
 */
static int nfs4_sp4_select_mode(struct nfs_client *clp,
				 struct nfs41_state_protection *sp)
{
	static const u32 supported_enforce[NFS4_OP_MAP_NUM_WORDS] = {
		[1] = 1 << (OP_BIND_CONN_TO_SESSION - 32) |
		      1 << (OP_EXCHANGE_ID - 32) |
		      1 << (OP_CREATE_SESSION - 32) |
		      1 << (OP_DESTROY_SESSION - 32) |
		      1 << (OP_DESTROY_CLIENTID - 32)
	};
	unsigned int i;

	if (sp->how == SP4_MACH_CRED) {
		/* Print state protect result */
		dfprintk(MOUNT, "Server SP4_MACH_CRED support:\n");
		for (i = 0; i <= LAST_NFS4_OP; i++) {
			if (test_bit(i, sp->enforce.u.longs))
				dfprintk(MOUNT, "  enforce op %d\n", i);
			if (test_bit(i, sp->allow.u.longs))
				dfprintk(MOUNT, "  allow op %d\n", i);
		}

		/* make sure nothing is on enforce list that isn't supported */
		for (i = 0; i < NFS4_OP_MAP_NUM_WORDS; i++) {
			if (sp->enforce.u.words[i] & ~supported_enforce[i]) {
				dfprintk(MOUNT, "sp4_mach_cred: disabled\n");
				return -EINVAL;
			}
		}

		/*
		 * Minimal mode - state operations are allowed to use machine
		 * credential.  Note this already happens by default, so the
		 * client doesn't have to do anything more than the negotiation.
		 *
		 * NOTE: we don't care if EXCHANGE_ID is in the list -
		 *       we're already using the machine cred for exchange_id
		 *       and will never use a different cred.
		 */
		if (test_bit(OP_BIND_CONN_TO_SESSION, sp->enforce.u.longs) &&
		    test_bit(OP_CREATE_SESSION, sp->enforce.u.longs) &&
		    test_bit(OP_DESTROY_SESSION, sp->enforce.u.longs) &&
		    test_bit(OP_DESTROY_CLIENTID, sp->enforce.u.longs)) {
			dfprintk(MOUNT, "sp4_mach_cred:\n");
			dfprintk(MOUNT, "  minimal mode enabled\n");
			set_bit(NFS_SP4_MACH_CRED_MINIMAL, &clp->cl_sp4_flags);
		} else {
			dfprintk(MOUNT, "sp4_mach_cred: disabled\n");
			return -EINVAL;
		}
6573 6574 6575 6576 6577 6578

		if (test_bit(OP_CLOSE, sp->allow.u.longs) &&
		    test_bit(OP_LOCKU, sp->allow.u.longs)) {
			dfprintk(MOUNT, "  cleanup mode enabled\n");
			set_bit(NFS_SP4_MACH_CRED_CLEANUP, &clp->cl_sp4_flags);
		}
6579 6580 6581 6582 6583 6584

		if (test_bit(OP_SECINFO, sp->allow.u.longs) &&
		    test_bit(OP_SECINFO_NO_NAME, sp->allow.u.longs)) {
			dfprintk(MOUNT, "  secinfo mode enabled\n");
			set_bit(NFS_SP4_MACH_CRED_SECINFO, &clp->cl_sp4_flags);
		}
6585 6586 6587 6588 6589 6590

		if (test_bit(OP_TEST_STATEID, sp->allow.u.longs) &&
		    test_bit(OP_FREE_STATEID, sp->allow.u.longs)) {
			dfprintk(MOUNT, "  stateid mode enabled\n");
			set_bit(NFS_SP4_MACH_CRED_STATEID, &clp->cl_sp4_flags);
		}
6591 6592 6593 6594 6595 6596 6597 6598 6599 6600

		if (test_bit(OP_WRITE, sp->allow.u.longs)) {
			dfprintk(MOUNT, "  write mode enabled\n");
			set_bit(NFS_SP4_MACH_CRED_WRITE, &clp->cl_sp4_flags);
		}

		if (test_bit(OP_COMMIT, sp->allow.u.longs)) {
			dfprintk(MOUNT, "  commit mode enabled\n");
			set_bit(NFS_SP4_MACH_CRED_COMMIT, &clp->cl_sp4_flags);
		}
6601 6602 6603 6604 6605 6606 6607
	}

	return 0;
}

/*
 * _nfs4_proc_exchange_id()
6608
 *
6609
 * Wrapper for EXCHANGE_ID operation.
B
Benny Halevy 已提交
6610
 */
6611 6612
static int _nfs4_proc_exchange_id(struct nfs_client *clp, struct rpc_cred *cred,
	u32 sp4_how)
B
Benny Halevy 已提交
6613 6614 6615
{
	nfs4_verifier verifier;
	struct nfs41_exchange_id_args args = {
6616
		.verifier = &verifier,
B
Benny Halevy 已提交
6617
		.client = clp,
6618
#ifdef CONFIG_NFS_V4_1_MIGRATION
6619
		.flags = EXCHGID4_FLAG_SUPP_MOVED_REFER |
6620 6621 6622 6623 6624 6625
			 EXCHGID4_FLAG_BIND_PRINC_STATEID |
			 EXCHGID4_FLAG_SUPP_MOVED_MIGR,
#else
		.flags = EXCHGID4_FLAG_SUPP_MOVED_REFER |
			 EXCHGID4_FLAG_BIND_PRINC_STATEID,
#endif
B
Benny Halevy 已提交
6626 6627
	};
	struct nfs41_exchange_id_res res = {
6628
		0
B
Benny Halevy 已提交
6629 6630 6631 6632 6633 6634 6635 6636 6637
	};
	int status;
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_EXCHANGE_ID],
		.rpc_argp = &args,
		.rpc_resp = &res,
		.rpc_cred = cred,
	};

6638
	nfs4_init_boot_verifier(clp, &verifier);
6639 6640
	args.id_len = nfs4_init_uniform_client_string(clp, args.id,
							sizeof(args.id));
6641 6642 6643
	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
6644

6645
	res.server_owner = kzalloc(sizeof(struct nfs41_server_owner),
6646
					GFP_NOFS);
6647
	if (unlikely(res.server_owner == NULL)) {
6648 6649 6650
		status = -ENOMEM;
		goto out;
	}
6651

6652
	res.server_scope = kzalloc(sizeof(struct nfs41_server_scope),
6653
					GFP_NOFS);
6654
	if (unlikely(res.server_scope == NULL)) {
6655
		status = -ENOMEM;
6656
		goto out_server_owner;
6657
	}
6658

6659
	res.impl_id = kzalloc(sizeof(struct nfs41_impl_id), GFP_NOFS);
6660
	if (unlikely(res.impl_id == NULL)) {
6661 6662 6663 6664
		status = -ENOMEM;
		goto out_server_scope;
	}

6665 6666 6667 6668 6669 6670 6671 6672 6673 6674 6675 6676 6677 6678 6679 6680
	switch (sp4_how) {
	case SP4_NONE:
		args.state_protect.how = SP4_NONE;
		break;

	case SP4_MACH_CRED:
		args.state_protect = nfs4_sp4_mach_cred_request;
		break;

	default:
		/* unsupported! */
		WARN_ON_ONCE(1);
		status = -EINVAL;
		goto out_server_scope;
	}

6681
	status = rpc_call_sync(clp->cl_rpcclient, &msg, RPC_TASK_TIMEOUT);
6682
	trace_nfs4_exchange_id(clp, status);
6683
	if (status == 0)
6684
		status = nfs4_check_cl_exchange_flags(res.flags);
6685

6686 6687 6688
	if (status == 0)
		status = nfs4_sp4_select_mode(clp, &res.state_protect);

6689
	if (status == 0) {
6690 6691 6692 6693 6694
		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;

6695 6696 6697
		kfree(clp->cl_serverowner);
		clp->cl_serverowner = res.server_owner;
		res.server_owner = NULL;
6698

6699
		/* use the most recent implementation id */
6700 6701
		kfree(clp->cl_implid);
		clp->cl_implid = res.impl_id;
6702

6703
		if (clp->cl_serverscope != NULL &&
6704
		    !nfs41_same_server_scope(clp->cl_serverscope,
6705 6706 6707 6708
					     res.server_scope)) {
			dprintk("%s: server_scope mismatch detected\n",
				__func__);
			set_bit(NFS4CLNT_SERVER_SCOPE_MISMATCH, &clp->cl_state);
6709 6710
			kfree(clp->cl_serverscope);
			clp->cl_serverscope = NULL;
6711 6712
		}

6713
		if (clp->cl_serverscope == NULL) {
6714
			clp->cl_serverscope = res.server_scope;
6715 6716
			goto out;
		}
6717 6718
	} else
		kfree(res.impl_id);
6719

6720 6721
out_server_owner:
	kfree(res.server_owner);
6722
out_server_scope:
6723 6724
	kfree(res.server_scope);
out:
6725
	if (clp->cl_implid != NULL)
6726
		dprintk("NFS reply exchange_id: Server Implementation ID: "
6727
			"domain: %s, name: %s, date: %llu,%u\n",
6728
			clp->cl_implid->domain, clp->cl_implid->name,
6729 6730
			clp->cl_implid->date.seconds,
			clp->cl_implid->date.nseconds);
6731
	dprintk("NFS reply exchange_id: %d\n", status);
B
Benny Halevy 已提交
6732 6733 6734
	return status;
}

6735 6736 6737 6738 6739 6740 6741 6742 6743 6744 6745 6746 6747 6748 6749 6750 6751 6752 6753 6754 6755 6756 6757 6758 6759 6760 6761 6762 6763
/*
 * nfs4_proc_exchange_id()
 *
 * Returns zero, a negative errno, or a negative NFS4ERR status code.
 *
 * 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.
 *
 * Will attempt to negotiate SP4_MACH_CRED if krb5i / krb5p auth is used.
 */
int nfs4_proc_exchange_id(struct nfs_client *clp, struct rpc_cred *cred)
{
	rpc_authflavor_t authflavor = clp->cl_rpcclient->cl_auth->au_flavor;
	int status;

	/* try SP4_MACH_CRED if krb5i/p	*/
	if (authflavor == RPC_AUTH_GSS_KRB5I ||
	    authflavor == RPC_AUTH_GSS_KRB5P) {
		status = _nfs4_proc_exchange_id(clp, cred, SP4_MACH_CRED);
		if (!status)
			return 0;
	}

	/* try SP4_NONE */
	return _nfs4_proc_exchange_id(clp, cred, SP4_NONE);
}

T
Trond Myklebust 已提交
6764 6765 6766 6767 6768 6769 6770 6771 6772 6773 6774
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);
6775
	trace_nfs4_destroy_clientid(clp, status);
T
Trond Myklebust 已提交
6776
	if (status)
6777
		dprintk("NFS: Got error %d from the server %s on "
T
Trond Myklebust 已提交
6778 6779 6780 6781 6782 6783 6784 6785 6786 6787 6788 6789 6790 6791 6792 6793 6794 6795 6796 6797 6798 6799 6800 6801 6802 6803 6804 6805 6806 6807 6808 6809 6810
			"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;
6811 6812
	if (clp->cl_preserve_clid)
		goto out;
6813
	cred = nfs4_get_clid_cred(clp);
T
Trond Myklebust 已提交
6814 6815 6816 6817 6818 6819 6820 6821 6822 6823 6824 6825
	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
6826 6827 6828 6829 6830 6831 6832 6833 6834 6835 6836 6837 6838 6839 6840
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 */
6841 6842 6843 6844
	nfs41_setup_sequence(data->clp->cl_session,
			&data->args->la_seq_args,
			&data->res->lr_seq_res,
			task);
A
Andy Adamson 已提交
6845 6846 6847 6848 6849 6850 6851 6852 6853 6854 6855 6856 6857
	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__);
6858 6859
	if (!nfs41_sequence_done(task, &data->res->lr_seq_res))
		return;
A
Andy Adamson 已提交
6860 6861 6862 6863 6864 6865
	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;
6866 6867
		/* fall through */
	case -NFS4ERR_RETRY_UNCACHED_REP:
6868
		rpc_restart_call_prepare(task);
A
Andy Adamson 已提交
6869 6870 6871 6872 6873
		return;
	}
	dprintk("<-- %s\n", __func__);
}

6874
static const struct rpc_call_ops nfs4_get_lease_time_ops = {
A
Andy Adamson 已提交
6875 6876 6877 6878 6879 6880 6881 6882 6883 6884 6885 6886 6887 6888 6889 6890 6891 6892 6893 6894 6895 6896 6897 6898 6899
	.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,
6900 6901
		.callback_data = &data,
		.flags = RPC_TASK_TIMEOUT,
A
Andy Adamson 已提交
6902 6903 6904
	};
	int status;

6905
	nfs4_init_sequence(&args.la_seq_args, &res.lr_seq_res, 0);
6906
	nfs4_set_sequence_privileged(&args.la_seq_args);
A
Andy Adamson 已提交
6907 6908 6909 6910 6911 6912 6913 6914 6915 6916 6917 6918 6919 6920
	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
6921 6922 6923 6924 6925 6926 6927 6928 6929 6930 6931
/*
 * 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)
{
6932 6933 6934 6935
	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
6936 6937

	/* Fore channel attributes */
6938 6939
	args->fc_attrs.max_rqst_sz = max_rqst_sz;
	args->fc_attrs.max_resp_sz = max_resp_sz;
A
Andy Adamson 已提交
6940
	args->fc_attrs.max_ops = NFS4_MAX_OPS;
6941
	args->fc_attrs.max_reqs = max_session_slots;
A
Andy Adamson 已提交
6942 6943

	dprintk("%s: Fore Channel : max_rqst_sz=%u max_resp_sz=%u "
6944
		"max_ops=%u max_reqs=%u\n",
A
Andy Adamson 已提交
6945 6946
		__func__,
		args->fc_attrs.max_rqst_sz, args->fc_attrs.max_resp_sz,
6947
		args->fc_attrs.max_ops, args->fc_attrs.max_reqs);
A
Andy Adamson 已提交
6948 6949 6950 6951 6952 6953 6954 6955 6956 6957 6958 6959 6960 6961 6962 6963

	/* 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);
}

6964
static int nfs4_verify_fore_channel_attrs(struct nfs41_create_session_args *args, struct nfs4_session *session)
6965
{
6966 6967 6968 6969 6970 6971 6972 6973 6974 6975 6976 6977 6978 6979 6980
	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;
6981 6982
	if (rcvd->max_reqs > NFS4_MAX_SLOT_TABLE)
		rcvd->max_reqs = NFS4_MAX_SLOT_TABLE;
6983
	return 0;
6984 6985
}

6986 6987 6988 6989
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;
6990

6991 6992 6993 6994 6995 6996 6997
	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: */
6998
	if (rcvd->max_ops != sent->max_ops)
6999
		return -EINVAL;
7000
	if (rcvd->max_reqs != sent->max_reqs)
7001 7002 7003
		return -EINVAL;
	return 0;
}
7004 7005 7006 7007

static int nfs4_verify_channel_attrs(struct nfs41_create_session_args *args,
				     struct nfs4_session *session)
{
7008
	int ret;
7009

7010 7011 7012 7013
	ret = nfs4_verify_fore_channel_attrs(args, session);
	if (ret)
		return ret;
	return nfs4_verify_back_channel_attrs(args, session);
7014 7015
}

7016 7017
static int _nfs4_proc_create_session(struct nfs_client *clp,
		struct rpc_cred *cred)
A
Andy Adamson 已提交
7018 7019 7020 7021 7022 7023 7024 7025 7026 7027 7028 7029 7030
{
	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,
7031
		.rpc_cred = cred,
A
Andy Adamson 已提交
7032 7033 7034 7035
	};
	int status;

	nfs4_init_channel_attrs(&args);
7036
	args.flags = (SESSION4_PERSIST | SESSION4_BACK_CHAN);
A
Andy Adamson 已提交
7037

7038
	status = rpc_call_sync(session->clp->cl_rpcclient, &msg, RPC_TASK_TIMEOUT);
7039
	trace_nfs4_create_session(clp, status);
A
Andy Adamson 已提交
7040

7041
	if (!status) {
7042 7043
		/* Verify the session's negotiated channel_attrs values */
		status = nfs4_verify_channel_attrs(&args, session);
A
Andy Adamson 已提交
7044 7045 7046 7047 7048 7049 7050 7051 7052 7053 7054 7055
		/* 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.
 */
7056
int nfs4_proc_create_session(struct nfs_client *clp, struct rpc_cred *cred)
A
Andy Adamson 已提交
7057 7058 7059 7060 7061 7062 7063
{
	int status;
	unsigned *ptr;
	struct nfs4_session *session = clp->cl_session;

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

7064
	status = _nfs4_proc_create_session(clp, cred);
A
Andy Adamson 已提交
7065 7066 7067
	if (status)
		goto out;

7068 7069 7070
	/* 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
7071 7072 7073 7074 7075 7076 7077 7078 7079 7080 7081
	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
7082 7083 7084 7085
/*
 * Issue the over-the-wire RPC DESTROY_SESSION.
 * The caller must serialize access to this routine.
 */
7086 7087
int nfs4_proc_destroy_session(struct nfs4_session *session,
		struct rpc_cred *cred)
A
Andy Adamson 已提交
7088
{
7089 7090 7091 7092 7093
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_DESTROY_SESSION],
		.rpc_argp = session,
		.rpc_cred = cred,
	};
A
Andy Adamson 已提交
7094 7095 7096 7097 7098 7099 7100 7101
	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;

7102
	status = rpc_call_sync(session->clp->cl_rpcclient, &msg, RPC_TASK_TIMEOUT);
7103
	trace_nfs4_destroy_session(session->clp, status);
A
Andy Adamson 已提交
7104 7105

	if (status)
7106
		dprintk("NFS: Got error %d from the server on DESTROY_SESSION. "
A
Andy Adamson 已提交
7107 7108 7109 7110 7111 7112
			"Session has been destroyed regardless...\n", status);

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

A
Andy Adamson 已提交
7113 7114 7115
/*
 * Renew the cl_session lease.
 */
7116 7117 7118 7119 7120 7121
struct nfs4_sequence_data {
	struct nfs_client *clp;
	struct nfs4_sequence_args args;
	struct nfs4_sequence_res res;
};

7122 7123
static void nfs41_sequence_release(void *data)
{
7124 7125
	struct nfs4_sequence_data *calldata = data;
	struct nfs_client *clp = calldata->clp;
7126

7127 7128 7129
	if (atomic_read(&clp->cl_count) > 1)
		nfs4_schedule_state_renewal(clp);
	nfs_put_client(clp);
7130
	kfree(calldata);
7131 7132
}

7133 7134 7135 7136 7137 7138 7139
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:
7140
		nfs4_schedule_lease_recovery(clp);
7141 7142 7143 7144
	}
	return 0;
}

7145
static void nfs41_sequence_call_done(struct rpc_task *task, void *data)
A
Andy Adamson 已提交
7146
{
7147 7148
	struct nfs4_sequence_data *calldata = data;
	struct nfs_client *clp = calldata->clp;
A
Andy Adamson 已提交
7149

7150 7151
	if (!nfs41_sequence_done(task, task->tk_msg.rpc_resp))
		return;
A
Andy Adamson 已提交
7152

7153
	trace_nfs4_sequence(clp, task->tk_status);
A
Andy Adamson 已提交
7154 7155
	if (task->tk_status < 0) {
		dprintk("%s ERROR %d\n", __func__, task->tk_status);
7156 7157
		if (atomic_read(&clp->cl_count) == 1)
			goto out;
A
Andy Adamson 已提交
7158

7159 7160
		if (nfs41_sequence_handle_errors(task, clp) == -EAGAIN) {
			rpc_restart_call_prepare(task);
A
Andy Adamson 已提交
7161 7162 7163 7164
			return;
		}
	}
	dprintk("%s rpc_cred %p\n", __func__, task->tk_msg.rpc_cred);
7165
out:
A
Andy Adamson 已提交
7166 7167 7168 7169 7170
	dprintk("<-- %s\n", __func__);
}

static void nfs41_sequence_prepare(struct rpc_task *task, void *data)
{
7171 7172
	struct nfs4_sequence_data *calldata = data;
	struct nfs_client *clp = calldata->clp;
A
Andy Adamson 已提交
7173 7174 7175 7176 7177 7178
	struct nfs4_sequence_args *args;
	struct nfs4_sequence_res *res;

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

7179
	nfs41_setup_sequence(clp->cl_session, args, res, task);
A
Andy Adamson 已提交
7180 7181 7182 7183 7184
}

static const struct rpc_call_ops nfs41_sequence_ops = {
	.rpc_call_done = nfs41_sequence_call_done,
	.rpc_call_prepare = nfs41_sequence_prepare,
7185
	.rpc_release = nfs41_sequence_release,
A
Andy Adamson 已提交
7186 7187
};

7188 7189 7190
static struct rpc_task *_nfs41_proc_sequence(struct nfs_client *clp,
		struct rpc_cred *cred,
		bool is_privileged)
A
Andy Adamson 已提交
7191
{
7192
	struct nfs4_sequence_data *calldata;
A
Andy Adamson 已提交
7193 7194 7195 7196
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_SEQUENCE],
		.rpc_cred = cred,
	};
7197 7198 7199
	struct rpc_task_setup task_setup_data = {
		.rpc_client = clp->cl_rpcclient,
		.rpc_message = &msg,
7200
		.callback_ops = &nfs41_sequence_ops,
7201
		.flags = RPC_TASK_ASYNC | RPC_TASK_TIMEOUT,
7202
	};
A
Andy Adamson 已提交
7203

7204
	if (!atomic_inc_not_zero(&clp->cl_count))
7205
		return ERR_PTR(-EIO);
7206
	calldata = kzalloc(sizeof(*calldata), GFP_NOFS);
7207
	if (calldata == NULL) {
7208
		nfs_put_client(clp);
7209
		return ERR_PTR(-ENOMEM);
A
Andy Adamson 已提交
7210
	}
7211
	nfs4_init_sequence(&calldata->args, &calldata->res, 0);
7212 7213
	if (is_privileged)
		nfs4_set_sequence_privileged(&calldata->args);
7214 7215 7216
	msg.rpc_argp = &calldata->args;
	msg.rpc_resp = &calldata->res;
	calldata->clp = clp;
7217
	task_setup_data.callback_data = calldata;
A
Andy Adamson 已提交
7218

7219 7220 7221
	return rpc_run_task(&task_setup_data);
}

7222
static int nfs41_proc_async_sequence(struct nfs_client *clp, struct rpc_cred *cred, unsigned renew_flags)
7223 7224 7225 7226
{
	struct rpc_task *task;
	int ret = 0;

7227 7228
	if ((renew_flags & NFS4_RENEW_TIMEOUT) == 0)
		return 0;
7229
	task = _nfs41_proc_sequence(clp, cred, false);
7230 7231 7232
	if (IS_ERR(task))
		ret = PTR_ERR(task);
	else
7233
		rpc_put_task_async(task);
7234 7235 7236 7237 7238 7239 7240 7241 7242
	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;

7243
	task = _nfs41_proc_sequence(clp, cred, true);
7244 7245 7246 7247 7248
	if (IS_ERR(task)) {
		ret = PTR_ERR(task);
		goto out;
	}
	ret = rpc_wait_for_completion_task(task);
7249 7250 7251 7252 7253
	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);
7254
		ret = task->tk_status;
7255
	}
7256 7257 7258 7259
	rpc_put_task(task);
out:
	dprintk("<-- %s status=%d\n", __func__, ret);
	return ret;
A
Andy Adamson 已提交
7260 7261
}

7262 7263 7264 7265 7266 7267 7268 7269 7270 7271
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;

7272 7273 7274 7275
	nfs41_setup_sequence(calldata->clp->cl_session,
			&calldata->arg.seq_args,
			&calldata->res.seq_res,
			task);
7276 7277
}

7278 7279 7280 7281 7282 7283 7284 7285 7286
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);
7287 7288
		/* fall through */
	case -NFS4ERR_RETRY_UNCACHED_REP:
7289 7290
		return -EAGAIN;
	default:
7291
		nfs4_schedule_lease_recovery(clp);
7292 7293 7294 7295
	}
	return 0;
}

7296 7297 7298 7299 7300 7301 7302
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__);
7303 7304
	if (!nfs41_sequence_done(task, res))
		return;
7305

7306
	trace_nfs4_reclaim_complete(clp, task->tk_status);
7307 7308 7309 7310
	if (nfs41_reclaim_complete_handle_errors(task, clp) == -EAGAIN) {
		rpc_restart_call_prepare(task);
		return;
	}
7311 7312 7313 7314 7315 7316 7317 7318 7319 7320 7321 7322 7323 7324 7325 7326 7327 7328 7329
	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.
 */
7330 7331
static int nfs41_proc_reclaim_complete(struct nfs_client *clp,
		struct rpc_cred *cred)
7332 7333 7334 7335 7336
{
	struct nfs4_reclaim_complete_data *calldata;
	struct rpc_task *task;
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_RECLAIM_COMPLETE],
7337
		.rpc_cred = cred,
7338 7339 7340 7341 7342 7343 7344 7345 7346 7347
	};
	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__);
7348
	calldata = kzalloc(sizeof(*calldata), GFP_NOFS);
7349 7350 7351 7352 7353
	if (calldata == NULL)
		goto out;
	calldata->clp = clp;
	calldata->arg.one_fs = 0;

7354
	nfs4_init_sequence(&calldata->arg.seq_args, &calldata->res.seq_res, 0);
7355
	nfs4_set_sequence_privileged(&calldata->arg.seq_args);
7356 7357 7358 7359
	msg.rpc_argp = &calldata->arg;
	msg.rpc_resp = &calldata->res;
	task_setup_data.callback_data = calldata;
	task = rpc_run_task(&task_setup_data);
7360
	if (IS_ERR(task)) {
7361
		status = PTR_ERR(task);
7362 7363
		goto out;
	}
7364 7365 7366
	status = nfs4_wait_for_completion_rpc_task(task);
	if (status == 0)
		status = task->tk_status;
7367
	rpc_put_task(task);
7368
	return 0;
7369 7370 7371 7372
out:
	dprintk("<-- %s status=%d\n", __func__, status);
	return status;
}
7373 7374 7375 7376 7377

static void
nfs4_layoutget_prepare(struct rpc_task *task, void *calldata)
{
	struct nfs4_layoutget *lgp = calldata;
F
Fred Isaman 已提交
7378
	struct nfs_server *server = NFS_SERVER(lgp->args.inode);
7379
	struct nfs4_session *session = nfs4_get_session(server);
7380 7381

	dprintk("--> %s\n", __func__);
F
Fred Isaman 已提交
7382 7383 7384 7385 7386
	/* 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.
	 */
7387
	if (nfs41_setup_sequence(session, &lgp->args.seq_args,
7388
				&lgp->res.seq_res, task))
7389
		return;
7390 7391 7392 7393 7394
	if (pnfs_choose_layoutget_stateid(&lgp->args.stateid,
					  NFS_I(lgp->args.inode)->layout,
					  lgp->args.ctx->state)) {
		rpc_exit(task, NFS4_OK);
	}
7395 7396 7397 7398 7399
}

static void nfs4_layoutget_done(struct rpc_task *task, void *calldata)
{
	struct nfs4_layoutget *lgp = calldata;
7400 7401 7402 7403
	struct inode *inode = lgp->args.inode;
	struct nfs_server *server = NFS_SERVER(inode);
	struct pnfs_layout_hdr *lo;
	struct nfs4_state *state = NULL;
7404
	unsigned long timeo, giveup;
7405 7406 7407

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

7408
	if (!nfs41_sequence_done(task, &lgp->res.seq_res))
7409
		goto out;
7410 7411 7412

	switch (task->tk_status) {
	case 0:
7413
		goto out;
7414 7415
	case -NFS4ERR_LAYOUTTRYLATER:
	case -NFS4ERR_RECALLCONFLICT:
7416 7417 7418 7419
		timeo = rpc_get_timeout(task->tk_client);
		giveup = lgp->args.timestamp + timeo;
		if (time_after(giveup, jiffies))
			task->tk_status = -NFS4ERR_DELAY;
7420 7421 7422 7423 7424 7425 7426 7427 7428 7429 7430 7431 7432 7433 7434 7435 7436
		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);
7437 7438
		}
	}
7439 7440 7441
	if (nfs4_async_handle_error(task, server, state) == -EAGAIN)
		rpc_restart_call_prepare(task);
out:
7442 7443 7444
	dprintk("<-- %s\n", __func__);
}

7445 7446 7447 7448 7449 7450 7451 7452 7453 7454 7455 7456 7457 7458 7459 7460 7461 7462 7463 7464 7465 7466 7467 7468 7469 7470 7471 7472 7473 7474 7475 7476 7477 7478 7479 7480 7481 7482 7483 7484 7485 7486 7487 7488
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;
}

7489 7490 7491
static void nfs4_layoutget_release(void *calldata)
{
	struct nfs4_layoutget *lgp = calldata;
7492 7493
	struct inode *inode = lgp->args.inode;
	struct nfs_server *server = NFS_SERVER(inode);
7494
	size_t max_pages = max_response_pages(server);
7495 7496

	dprintk("--> %s\n", __func__);
7497
	nfs4_free_pages(lgp->args.layout.pages, max_pages);
7498
	pnfs_put_layout_hdr(NFS_I(inode)->layout);
7499 7500 7501 7502 7503 7504 7505 7506 7507 7508 7509
	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,
};

7510 7511
struct pnfs_layout_segment *
nfs4_proc_layoutget(struct nfs4_layoutget *lgp, gfp_t gfp_flags)
7512
{
7513 7514
	struct inode *inode = lgp->args.inode;
	struct nfs_server *server = NFS_SERVER(inode);
7515
	size_t max_pages = max_response_pages(server);
7516 7517 7518 7519 7520
	struct rpc_task *task;
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_LAYOUTGET],
		.rpc_argp = &lgp->args,
		.rpc_resp = &lgp->res,
7521
		.rpc_cred = lgp->cred,
7522 7523 7524 7525 7526 7527 7528 7529
	};
	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,
	};
7530
	struct pnfs_layout_segment *lseg = NULL;
7531 7532 7533 7534
	int status = 0;

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

7535 7536 7537
	lgp->args.layout.pages = nfs4_alloc_pages(max_pages, gfp_flags);
	if (!lgp->args.layout.pages) {
		nfs4_layoutget_release(lgp);
7538
		return ERR_PTR(-ENOMEM);
7539 7540
	}
	lgp->args.layout.pglen = max_pages * PAGE_SIZE;
7541
	lgp->args.timestamp = jiffies;
7542

7543
	lgp->res.layoutp = &lgp->args.layout;
7544
	lgp->res.seq_res.sr_slot = NULL;
7545
	nfs4_init_sequence(&lgp->args.seq_args, &lgp->res.seq_res, 0);
7546 7547 7548 7549

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

7550 7551
	task = rpc_run_task(&task_setup_data);
	if (IS_ERR(task))
7552
		return ERR_CAST(task);
7553
	status = nfs4_wait_for_completion_rpc_task(task);
F
Fred Isaman 已提交
7554 7555
	if (status == 0)
		status = task->tk_status;
7556 7557 7558 7559
	trace_nfs4_layoutget(lgp->args.ctx,
			&lgp->args.range,
			&lgp->res.range,
			status);
7560 7561
	/* if layoutp->len is 0, nfs4_layoutget_prepare called rpc_exit */
	if (status == 0 && lgp->res.layoutp->len)
7562
		lseg = pnfs_layout_process(lgp);
7563 7564
	rpc_put_task(task);
	dprintk("<-- %s status=%d\n", __func__, status);
7565 7566 7567
	if (status)
		return ERR_PTR(status);
	return lseg;
7568 7569
}

B
Benny Halevy 已提交
7570 7571 7572 7573 7574 7575
static void
nfs4_layoutreturn_prepare(struct rpc_task *task, void *calldata)
{
	struct nfs4_layoutreturn *lrp = calldata;

	dprintk("--> %s\n", __func__);
7576 7577 7578 7579
	nfs41_setup_sequence(lrp->clp->cl_session,
			&lrp->args.seq_args,
			&lrp->res.seq_res,
			task);
B
Benny Halevy 已提交
7580 7581 7582 7583 7584 7585 7586 7587 7588
}

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

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

7589
	if (!nfs41_sequence_done(task, &lrp->res.seq_res))
B
Benny Halevy 已提交
7590 7591 7592 7593
		return;

	server = NFS_SERVER(lrp->args.inode);
	if (nfs4_async_handle_error(task, server, NULL) == -EAGAIN) {
7594
		rpc_restart_call_prepare(task);
B
Benny Halevy 已提交
7595 7596 7597 7598 7599 7600 7601 7602
		return;
	}
	dprintk("<-- %s\n", __func__);
}

static void nfs4_layoutreturn_release(void *calldata)
{
	struct nfs4_layoutreturn *lrp = calldata;
7603
	struct pnfs_layout_hdr *lo = lrp->args.layout;
B
Benny Halevy 已提交
7604 7605

	dprintk("--> %s\n", __func__);
7606 7607 7608 7609 7610
	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);
7611
	pnfs_put_layout_hdr(lrp->args.layout);
B
Benny Halevy 已提交
7612 7613 7614 7615 7616 7617 7618 7619 7620 7621 7622 7623 7624 7625 7626 7627 7628
	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,
7629
		.rpc_cred = lrp->cred,
B
Benny Halevy 已提交
7630 7631
	};
	struct rpc_task_setup task_setup_data = {
7632
		.rpc_client = NFS_SERVER(lrp->args.inode)->client,
B
Benny Halevy 已提交
7633 7634 7635 7636 7637 7638 7639
		.rpc_message = &msg,
		.callback_ops = &nfs4_layoutreturn_call_ops,
		.callback_data = lrp,
	};
	int status;

	dprintk("--> %s\n", __func__);
7640
	nfs4_init_sequence(&lrp->args.seq_args, &lrp->res.seq_res, 1);
B
Benny Halevy 已提交
7641 7642 7643 7644
	task = rpc_run_task(&task_setup_data);
	if (IS_ERR(task))
		return PTR_ERR(task);
	status = task->tk_status;
7645
	trace_nfs4_layoutreturn(lrp->args.inode, status);
B
Benny Halevy 已提交
7646 7647 7648 7649 7650
	dprintk("<-- %s status=%d\n", __func__, status);
	rpc_put_task(task);
	return status;
}

A
Andy Adamson 已提交
7651 7652 7653 7654 7655 7656 7657 7658 7659 7660 7661 7662 7663 7664 7665 7666 7667 7668 7669 7670 7671 7672 7673 7674 7675 7676 7677 7678 7679 7680 7681 7682 7683 7684 7685 7686 7687 7688 7689 7690 7691 7692 7693 7694 7695 7696 7697 7698
/*
 * 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);

7699
static int
7700 7701 7702
_nfs4_proc_getdeviceinfo(struct nfs_server *server,
		struct pnfs_device *pdev,
		struct rpc_cred *cred)
7703 7704 7705 7706 7707 7708 7709 7710 7711 7712 7713
{
	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,
7714
		.rpc_cred = cred,
7715 7716 7717 7718
	};
	int status;

	dprintk("--> %s\n", __func__);
7719
	status = nfs4_call_sync(server->client, server, &msg, &args.seq_args, &res.seq_res, 0);
7720 7721 7722 7723 7724
	dprintk("<-- %s status=%d\n", __func__, status);

	return status;
}

7725 7726 7727
int nfs4_proc_getdeviceinfo(struct nfs_server *server,
		struct pnfs_device *pdev,
		struct rpc_cred *cred)
7728 7729 7730 7731 7732 7733
{
	struct nfs4_exception exception = { };
	int err;

	do {
		err = nfs4_handle_exception(server,
7734
					_nfs4_proc_getdeviceinfo(server, pdev, cred),
7735 7736 7737 7738 7739 7740
					&exception);
	} while (exception.retry);
	return err;
}
EXPORT_SYMBOL_GPL(nfs4_proc_getdeviceinfo);

A
Andy Adamson 已提交
7741 7742 7743 7744
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);
7745
	struct nfs4_session *session = nfs4_get_session(server);
A
Andy Adamson 已提交
7746

7747 7748 7749 7750
	nfs41_setup_sequence(session,
			&data->args.seq_args,
			&data->res.seq_res,
			task);
A
Andy Adamson 已提交
7751 7752 7753 7754 7755 7756 7757 7758
}

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);

7759
	if (!nfs41_sequence_done(task, &data->res.seq_res))
A
Andy Adamson 已提交
7760 7761 7762
		return;

	switch (task->tk_status) { /* Just ignore these failures */
7763 7764 7765 7766
	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
7767
		task->tk_status = 0;
7768 7769
		break;
	case 0:
A
Andy Adamson 已提交
7770 7771
		nfs_post_op_update_inode_force_wcc(data->args.inode,
						   data->res.fattr);
7772 7773 7774 7775 7776 7777 7778
		break;
	default:
		if (nfs4_async_handle_error(task, server, NULL) == -EAGAIN) {
			rpc_restart_call_prepare(task);
			return;
		}
	}
A
Andy Adamson 已提交
7779 7780 7781 7782 7783 7784
}

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

A
Andy Adamson 已提交
7785
	pnfs_cleanup_layoutcommit(data);
A
Andy Adamson 已提交
7786 7787 7788 7789 7790 7791 7792 7793 7794 7795 7796
	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
7797
nfs4_proc_layoutcommit(struct nfs4_layoutcommit_data *data, bool sync)
A
Andy Adamson 已提交
7798 7799 7800 7801 7802 7803 7804 7805 7806 7807 7808 7809 7810 7811 7812 7813 7814 7815 7816 7817 7818 7819 7820 7821
{
	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);

7822
	nfs4_init_sequence(&data->args.seq_args, &data->res.seq_res, 1);
A
Andy Adamson 已提交
7823 7824 7825
	task = rpc_run_task(&task_setup_data);
	if (IS_ERR(task))
		return PTR_ERR(task);
7826
	if (sync == false)
A
Andy Adamson 已提交
7827 7828 7829 7830 7831
		goto out;
	status = nfs4_wait_for_completion_rpc_task(task);
	if (status != 0)
		goto out;
	status = task->tk_status;
7832
	trace_nfs4_layoutcommit(data->args.inode, status);
A
Andy Adamson 已提交
7833 7834 7835 7836 7837
out:
	dprintk("%s: status %d\n", __func__, status);
	rpc_put_task(task);
	return status;
}
7838

7839 7840 7841 7842
/**
 * Use the state managment nfs_client cl_rpcclient, which uses krb5i (if
 * possible) as per RFC3530bis and RFC5661 Security Considerations sections
 */
7843 7844
static int
_nfs41_proc_secinfo_no_name(struct nfs_server *server, struct nfs_fh *fhandle,
7845 7846
		    struct nfs_fsinfo *info,
		    struct nfs4_secinfo_flavors *flavors, bool use_integrity)
7847 7848 7849 7850 7851 7852 7853 7854 7855 7856 7857 7858
{
	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,
	};
7859
	struct rpc_clnt *clnt = server->client;
7860
	struct rpc_cred *cred = NULL;
7861 7862 7863 7864
	int status;

	if (use_integrity) {
		clnt = server->nfs_client->cl_rpcclient;
7865 7866
		cred = nfs4_get_clid_cred(server->nfs_client);
		msg.rpc_cred = cred;
7867 7868 7869 7870 7871 7872 7873
	}

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

7874 7875
	if (cred)
		put_rpccred(cred);
7876 7877

	return status;
7878 7879 7880 7881 7882 7883 7884 7885 7886
}

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 {
7887 7888 7889 7890 7891 7892 7893 7894 7895 7896 7897 7898 7899 7900 7901 7902 7903 7904
		/* first try using integrity protection */
		err = -NFS4ERR_WRONGSEC;

		/* try to use integrity protection with machine cred */
		if (_nfs4_is_integrity_protected(server->nfs_client))
			err = _nfs41_proc_secinfo_no_name(server, fhandle, info,
							  flavors, true);

		/*
		 * if unable to use integrity protection, or SECINFO with
		 * integrity protection returns NFS4ERR_WRONGSEC (which is
		 * disallowed by spec, but exists in deployed servers) use
		 * the current filesystem's rpc_client and the user cred.
		 */
		if (err == -NFS4ERR_WRONGSEC)
			err = _nfs41_proc_secinfo_no_name(server, fhandle, info,
							  flavors, false);

7905 7906 7907 7908
		switch (err) {
		case 0:
		case -NFS4ERR_WRONGSEC:
		case -NFS4ERR_NOTSUPP:
7909
			goto out;
7910 7911 7912 7913
		default:
			err = nfs4_handle_exception(server, err, &exception);
		}
	} while (exception.retry);
7914
out:
7915 7916 7917 7918 7919 7920 7921 7922 7923
	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;
7924
	rpc_authflavor_t flavor = RPC_AUTH_MAXFLAVOR;
7925
	struct nfs4_secinfo_flavors *flavors;
7926 7927
	struct nfs4_secinfo4 *secinfo;
	int i;
7928 7929 7930 7931 7932 7933 7934 7935 7936 7937 7938 7939 7940 7941 7942 7943 7944 7945 7946 7947 7948

	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;

7949 7950 7951 7952 7953 7954 7955 7956 7957 7958 7959 7960 7961 7962 7963
	for (i = 0; i < flavors->num_flavors; i++) {
		secinfo = &flavors->flavors[i];

		switch (secinfo->flavor) {
		case RPC_AUTH_NULL:
		case RPC_AUTH_UNIX:
		case RPC_AUTH_GSS:
			flavor = rpcauth_get_pseudoflavor(secinfo->flavor,
					&secinfo->flavor_info);
			break;
		default:
			flavor = RPC_AUTH_MAXFLAVOR;
			break;
		}

7964 7965 7966
		if (!nfs_auth_info_match(&server->auth_info, flavor))
			flavor = RPC_AUTH_MAXFLAVOR;

7967 7968 7969 7970 7971 7972 7973 7974 7975 7976
		if (flavor != RPC_AUTH_MAXFLAVOR) {
			err = nfs4_lookup_root_sec(server, fhandle,
						   info, flavor);
			if (!err)
				break;
		}
	}

	if (flavor == RPC_AUTH_MAXFLAVOR)
		err = -EPERM;
7977 7978 7979 7980 7981 7982 7983 7984

out_freepage:
	put_page(page);
	if (err == -EACCES)
		return -EPERM;
out:
	return err;
}
7985

7986 7987 7988
static int _nfs41_test_stateid(struct nfs_server *server,
		nfs4_stateid *stateid,
		struct rpc_cred *cred)
B
Bryan Schumaker 已提交
7989 7990 7991
{
	int status;
	struct nfs41_test_stateid_args args = {
7992
		.stateid = stateid,
B
Bryan Schumaker 已提交
7993 7994 7995 7996 7997 7998
	};
	struct nfs41_test_stateid_res res;
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_TEST_STATEID],
		.rpc_argp = &args,
		.rpc_resp = &res,
7999
		.rpc_cred = cred,
B
Bryan Schumaker 已提交
8000
	};
8001 8002 8003 8004
	struct rpc_clnt *rpc_client = server->client;

	nfs4_state_protect(server->nfs_client, NFS_SP4_MACH_CRED_STATEID,
		&rpc_client, &msg);
8005

8006
	dprintk("NFS call  test_stateid %p\n", stateid);
8007
	nfs4_init_sequence(&args.seq_args, &res.seq_res, 0);
8008
	nfs4_set_sequence_privileged(&args.seq_args);
8009
	status = nfs4_call_sync_sequence(rpc_client, server, &msg,
8010
			&args.seq_args, &res.seq_res);
8011 8012
	if (status != NFS_OK) {
		dprintk("NFS reply test_stateid: failed, %d\n", status);
8013
		return status;
8014 8015
	}
	dprintk("NFS reply test_stateid: succeeded, %d\n", -res.status);
8016
	return -res.status;
B
Bryan Schumaker 已提交
8017 8018
}

8019 8020 8021 8022 8023
/**
 * nfs41_test_stateid - perform a TEST_STATEID operation
 *
 * @server: server / transport on which to perform the operation
 * @stateid: state ID to test
8024
 * @cred: credential
8025 8026 8027 8028 8029
 *
 * 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.
 */
8030 8031 8032
static int nfs41_test_stateid(struct nfs_server *server,
		nfs4_stateid *stateid,
		struct rpc_cred *cred)
B
Bryan Schumaker 已提交
8033 8034 8035 8036
{
	struct nfs4_exception exception = { };
	int err;
	do {
8037
		err = _nfs41_test_stateid(server, stateid, cred);
8038 8039 8040
		if (err != -NFS4ERR_DELAY)
			break;
		nfs4_handle_exception(server, err, &exception);
B
Bryan Schumaker 已提交
8041 8042 8043
	} while (exception.retry);
	return err;
}
B
Bryan Schumaker 已提交
8044

8045 8046 8047
struct nfs_free_stateid_data {
	struct nfs_server *server;
	struct nfs41_free_stateid_args args;
B
Bryan Schumaker 已提交
8048
	struct nfs41_free_stateid_res res;
8049 8050 8051 8052 8053 8054 8055 8056 8057 8058 8059 8060 8061 8062 8063 8064 8065 8066 8067 8068 8069 8070 8071 8072 8073 8074 8075 8076 8077
};

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);
}

8078
static const struct rpc_call_ops nfs41_free_stateid_ops = {
8079 8080 8081 8082 8083 8084 8085
	.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,
8086
		struct rpc_cred *cred,
8087 8088
		bool privileged)
{
B
Bryan Schumaker 已提交
8089 8090
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_FREE_STATEID],
8091
		.rpc_cred = cred,
B
Bryan Schumaker 已提交
8092
	};
8093 8094 8095 8096 8097 8098 8099
	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
8100

8101 8102 8103
	nfs4_state_protect(server->nfs_client, NFS_SP4_MACH_CRED_STATEID,
		&task_setup.rpc_client, &msg);

8104
	dprintk("NFS call  free_stateid %p\n", stateid);
8105 8106 8107 8108 8109 8110 8111 8112 8113 8114
	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;
8115
	nfs4_init_sequence(&data->args.seq_args, &data->res.seq_res, 0);
8116 8117 8118 8119
	if (privileged)
		nfs4_set_sequence_privileged(&data->args.seq_args);

	return rpc_run_task(&task_setup);
B
Bryan Schumaker 已提交
8120 8121
}

8122 8123 8124 8125 8126
/**
 * nfs41_free_stateid - perform a FREE_STATEID operation
 *
 * @server: server / transport on which to perform the operation
 * @stateid: state ID to release
8127
 * @cred: credential
8128 8129 8130 8131
 *
 * Returns NFS_OK if the server freed "stateid".  Otherwise a
 * negative NFS4ERR value is returned.
 */
8132 8133 8134
static int nfs41_free_stateid(struct nfs_server *server,
		nfs4_stateid *stateid,
		struct rpc_cred *cred)
B
Bryan Schumaker 已提交
8135
{
8136 8137 8138
	struct rpc_task *task;
	int ret;

8139
	task = _nfs41_free_stateid(server, stateid, cred, true);
8140 8141 8142 8143 8144 8145 8146
	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
8147
}
8148

8149 8150 8151
static int nfs41_free_lock_state(struct nfs_server *server, struct nfs4_lock_state *lsp)
{
	struct rpc_task *task;
8152
	struct rpc_cred *cred = lsp->ls_state->owner->so_cred;
8153

8154
	task = _nfs41_free_stateid(server, &lsp->ls_stateid, cred, false);
8155 8156 8157 8158 8159 8160 8161
	nfs4_free_lock_state(server, lsp);
	if (IS_ERR(task))
		return PTR_ERR(task);
	rpc_put_task(task);
	return 0;
}

8162 8163 8164
static bool nfs41_match_stateid(const nfs4_stateid *s1,
		const nfs4_stateid *s2)
{
8165
	if (memcmp(s1->other, s2->other, sizeof(s1->other)) != 0)
8166 8167
		return false;

8168
	if (s1->seqid == s2->seqid)
8169
		return true;
8170
	if (s1->seqid == 0 || s2->seqid == 0)
8171 8172 8173 8174 8175
		return true;

	return false;
}

8176 8177
#endif /* CONFIG_NFS_V4_1 */

8178 8179 8180
static bool nfs4_match_stateid(const nfs4_stateid *s1,
		const nfs4_stateid *s2)
{
8181
	return nfs4_stateid_match(s1, s2);
8182 8183 8184
}


8185
static const struct nfs4_state_recovery_ops nfs40_reboot_recovery_ops = {
8186
	.owner_flag_bit = NFS_OWNER_RECLAIM_REBOOT,
8187
	.state_flag_bit	= NFS_STATE_RECLAIM_REBOOT,
L
Linus Torvalds 已提交
8188 8189
	.recover_open	= nfs4_open_reclaim,
	.recover_lock	= nfs4_lock_reclaim,
8190
	.establish_clid = nfs4_init_clientid,
8191
	.detect_trunking = nfs40_discover_server_trunking,
L
Linus Torvalds 已提交
8192 8193
};

8194
#if defined(CONFIG_NFS_V4_1)
8195
static const struct nfs4_state_recovery_ops nfs41_reboot_recovery_ops = {
8196 8197 8198 8199
	.owner_flag_bit = NFS_OWNER_RECLAIM_REBOOT,
	.state_flag_bit	= NFS_STATE_RECLAIM_REBOOT,
	.recover_open	= nfs4_open_reclaim,
	.recover_lock	= nfs4_lock_reclaim,
8200
	.establish_clid = nfs41_init_clientid,
8201
	.reclaim_complete = nfs41_proc_reclaim_complete,
8202
	.detect_trunking = nfs41_discover_server_trunking,
8203 8204 8205
};
#endif /* CONFIG_NFS_V4_1 */

8206
static const struct nfs4_state_recovery_ops nfs40_nograce_recovery_ops = {
8207 8208 8209 8210 8211 8212 8213 8214
	.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,
};

#if defined(CONFIG_NFS_V4_1)
8215
static const struct nfs4_state_recovery_ops nfs41_nograce_recovery_ops = {
8216
	.owner_flag_bit = NFS_OWNER_RECLAIM_NOGRACE,
8217
	.state_flag_bit	= NFS_STATE_RECLAIM_NOGRACE,
8218 8219
	.recover_open	= nfs41_open_expired,
	.recover_lock	= nfs41_lock_expired,
8220
	.establish_clid = nfs41_init_clientid,
L
Linus Torvalds 已提交
8221
};
8222
#endif /* CONFIG_NFS_V4_1 */
L
Linus Torvalds 已提交
8223

8224
static const struct nfs4_state_maintenance_ops nfs40_state_renewal_ops = {
B
Benny Halevy 已提交
8225
	.sched_state_renewal = nfs4_proc_async_renew,
8226
	.get_state_renewal_cred_locked = nfs4_get_renew_cred_locked,
8227
	.renew_lease = nfs4_proc_renew,
B
Benny Halevy 已提交
8228 8229 8230
};

#if defined(CONFIG_NFS_V4_1)
8231
static const struct nfs4_state_maintenance_ops nfs41_state_renewal_ops = {
B
Benny Halevy 已提交
8232
	.sched_state_renewal = nfs41_proc_async_sequence,
8233
	.get_state_renewal_cred_locked = nfs4_get_machine_cred_locked,
8234
	.renew_lease = nfs4_proc_sequence,
B
Benny Halevy 已提交
8235 8236 8237
};
#endif

8238
static const struct nfs4_mig_recovery_ops nfs40_mig_recovery_ops = {
8239
	.get_locations = _nfs40_proc_get_locations,
8240
	.fsid_present = _nfs40_proc_fsid_present,
8241 8242 8243 8244
};

#if defined(CONFIG_NFS_V4_1)
static const struct nfs4_mig_recovery_ops nfs41_mig_recovery_ops = {
8245
	.get_locations = _nfs41_proc_get_locations,
8246
	.fsid_present = _nfs41_proc_fsid_present,
8247 8248 8249
};
#endif	/* CONFIG_NFS_V4_1 */

8250 8251
static const struct nfs4_minor_version_ops nfs_v4_0_minor_ops = {
	.minor_version = 0,
8252 8253 8254 8255
	.init_caps = NFS_CAP_READDIRPLUS
		| NFS_CAP_ATOMIC_OPEN
		| NFS_CAP_CHANGE_ATTR
		| NFS_CAP_POSIX_LOCK,
8256 8257
	.init_client = nfs40_init_client,
	.shutdown_client = nfs40_shutdown_client,
8258
	.match_stateid = nfs4_match_stateid,
8259
	.find_root_sec = nfs4_find_root_sec,
8260
	.free_lock_state = nfs4_release_lockowner,
8261
	.call_sync_ops = &nfs40_call_sync_ops,
8262 8263 8264
	.reboot_recovery_ops = &nfs40_reboot_recovery_ops,
	.nograce_recovery_ops = &nfs40_nograce_recovery_ops,
	.state_renewal_ops = &nfs40_state_renewal_ops,
8265
	.mig_recovery_ops = &nfs40_mig_recovery_ops,
8266 8267 8268 8269 8270
};

#if defined(CONFIG_NFS_V4_1)
static const struct nfs4_minor_version_ops nfs_v4_1_minor_ops = {
	.minor_version = 1,
8271 8272 8273
	.init_caps = NFS_CAP_READDIRPLUS
		| NFS_CAP_ATOMIC_OPEN
		| NFS_CAP_CHANGE_ATTR
8274
		| NFS_CAP_POSIX_LOCK
8275 8276
		| NFS_CAP_STATEID_NFSV41
		| NFS_CAP_ATOMIC_OPEN_V1,
8277 8278
	.init_client = nfs41_init_client,
	.shutdown_client = nfs41_shutdown_client,
8279
	.match_stateid = nfs41_match_stateid,
8280
	.find_root_sec = nfs41_find_root_sec,
8281
	.free_lock_state = nfs41_free_lock_state,
8282
	.call_sync_ops = &nfs41_call_sync_ops,
8283 8284 8285
	.reboot_recovery_ops = &nfs41_reboot_recovery_ops,
	.nograce_recovery_ops = &nfs41_nograce_recovery_ops,
	.state_renewal_ops = &nfs41_state_renewal_ops,
8286
	.mig_recovery_ops = &nfs41_mig_recovery_ops,
8287 8288 8289
};
#endif

8290 8291 8292
#if defined(CONFIG_NFS_V4_2)
static const struct nfs4_minor_version_ops nfs_v4_2_minor_ops = {
	.minor_version = 2,
8293 8294 8295 8296 8297 8298
	.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,
8299 8300
	.init_client = nfs41_init_client,
	.shutdown_client = nfs41_shutdown_client,
8301 8302
	.match_stateid = nfs41_match_stateid,
	.find_root_sec = nfs41_find_root_sec,
8303
	.free_lock_state = nfs41_free_lock_state,
8304
	.call_sync_ops = &nfs41_call_sync_ops,
8305 8306 8307 8308 8309 8310
	.reboot_recovery_ops = &nfs41_reboot_recovery_ops,
	.nograce_recovery_ops = &nfs41_nograce_recovery_ops,
	.state_renewal_ops = &nfs41_state_renewal_ops,
};
#endif

8311 8312 8313 8314 8315
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
8316 8317 8318
#if defined(CONFIG_NFS_V4_2)
	[2] = &nfs_v4_2_minor_ops,
#endif
8319 8320
};

8321
static const struct inode_operations nfs4_dir_inode_operations = {
8322 8323 8324 8325 8326 8327 8328 8329 8330 8331 8332 8333 8334 8335 8336 8337 8338 8339 8340
	.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,
};

8341
static const struct inode_operations nfs4_file_inode_operations = {
8342 8343 8344
	.permission	= nfs_permission,
	.getattr	= nfs_getattr,
	.setattr	= nfs_setattr,
8345 8346 8347 8348
	.getxattr	= generic_getxattr,
	.setxattr	= generic_setxattr,
	.listxattr	= generic_listxattr,
	.removexattr	= generic_removexattr,
8349 8350
};

D
David Howells 已提交
8351
const struct nfs_rpc_ops nfs_v4_clientops = {
L
Linus Torvalds 已提交
8352 8353 8354
	.version	= 4,			/* protocol version */
	.dentry_ops	= &nfs4_dentry_operations,
	.dir_inode_ops	= &nfs4_dir_inode_operations,
8355
	.file_inode_ops	= &nfs4_file_inode_operations,
8356
	.file_ops	= &nfs4_file_operations,
L
Linus Torvalds 已提交
8357
	.getroot	= nfs4_proc_get_root,
B
Bryan Schumaker 已提交
8358
	.submount	= nfs4_submount,
B
Bryan Schumaker 已提交
8359
	.try_mount	= nfs4_try_mount,
L
Linus Torvalds 已提交
8360 8361 8362 8363 8364 8365 8366 8367
	.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,
8368
	.unlink_rpc_prepare = nfs4_proc_unlink_rpc_prepare,
L
Linus Torvalds 已提交
8369 8370
	.unlink_done	= nfs4_proc_unlink_done,
	.rename		= nfs4_proc_rename,
8371
	.rename_setup	= nfs4_proc_rename_setup,
8372
	.rename_rpc_prepare = nfs4_proc_rename_rpc_prepare,
8373
	.rename_done	= nfs4_proc_rename_done,
L
Linus Torvalds 已提交
8374 8375 8376 8377 8378 8379 8380 8381 8382
	.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,
8383
	.set_capabilities = nfs4_server_capabilities,
L
Linus Torvalds 已提交
8384 8385
	.decode_dirent	= nfs4_decode_dirent,
	.read_setup	= nfs4_proc_read_setup,
8386
	.read_pageio_init = pnfs_pageio_init_read,
8387
	.read_rpc_prepare = nfs4_proc_read_rpc_prepare,
T
Trond Myklebust 已提交
8388
	.read_done	= nfs4_read_done,
L
Linus Torvalds 已提交
8389
	.write_setup	= nfs4_proc_write_setup,
8390
	.write_pageio_init = pnfs_pageio_init_write,
8391
	.write_rpc_prepare = nfs4_proc_write_rpc_prepare,
8392
	.write_done	= nfs4_write_done,
L
Linus Torvalds 已提交
8393
	.commit_setup	= nfs4_proc_commit_setup,
8394
	.commit_rpc_prepare = nfs4_proc_commit_rpc_prepare,
8395
	.commit_done	= nfs4_commit_done,
L
Linus Torvalds 已提交
8396
	.lock		= nfs4_proc_lock,
8397
	.clear_acl_cache = nfs4_zap_acl_attr,
T
Trond Myklebust 已提交
8398
	.close_context  = nfs4_close_context,
8399
	.open_context	= nfs4_atomic_open,
8400
	.have_delegation = nfs4_have_delegation,
8401
	.return_delegation = nfs4_inode_return_delegation,
8402
	.alloc_client	= nfs4_alloc_client,
8403
	.init_client	= nfs4_init_client,
8404
	.free_client	= nfs4_free_client,
8405 8406
	.create_server	= nfs4_create_server,
	.clone_server	= nfs_clone_server,
L
Linus Torvalds 已提交
8407 8408
};

8409 8410 8411 8412 8413 8414 8415 8416 8417
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,
8418 8419 8420
#ifdef CONFIG_NFS_V4_SECURITY_LABEL
	&nfs4_xattr_nfs4_label_handler,
#endif
8421 8422 8423
	NULL
};

L
Linus Torvalds 已提交
8424 8425 8426 8427 8428
/*
 * Local variables:
 *  c-basic-offset: 8
 * End:
 */