nfs4proc.c 233.8 KB
Newer Older
L
Linus Torvalds 已提交
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41
/*
 *  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 *, long *);
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_FILE_OPEN:
		return -EBUSY;
163
	default:
L
Linus Torvalds 已提交
164
		dprintk("%s could not handle NFSv4 error %d\n",
165
				__func__, -err);
166
		break;
L
Linus Torvalds 已提交
167
	}
168
	return -EIO;
L
Linus Torvalds 已提交
169 170 171 172 173
}

/*
 * This is our standard bitmap for GETATTR requests.
 */
174
const u32 nfs4_fattr_bitmap[3] = {
L
Linus Torvalds 已提交
175 176 177 178 179 180 181 182 183 184 185 186 187
	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
188 189
	| FATTR4_WORD1_TIME_MODIFY
	| FATTR4_WORD1_MOUNTED_ON_FILEID,
190 191 192
#ifdef CONFIG_NFS_V4_SECURITY_LABEL
	FATTR4_WORD2_SECURITY_LABEL
#endif
L
Linus Torvalds 已提交
193 194
};

195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212
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
};

213 214 215 216 217 218
static const u32 nfs4_open_noattr_bitmap[3] = {
	FATTR4_WORD0_TYPE
	| FATTR4_WORD0_CHANGE
	| FATTR4_WORD0_FILEID,
};

219
const u32 nfs4_statfs_bitmap[3] = {
L
Linus Torvalds 已提交
220 221 222 223 224 225 226 227
	FATTR4_WORD0_FILES_AVAIL
	| FATTR4_WORD0_FILES_FREE
	| FATTR4_WORD0_FILES_TOTAL,
	FATTR4_WORD1_SPACE_AVAIL
	| FATTR4_WORD1_SPACE_FREE
	| FATTR4_WORD1_SPACE_TOTAL
};

228
const u32 nfs4_pathconf_bitmap[3] = {
L
Linus Torvalds 已提交
229 230 231 232 233
	FATTR4_WORD0_MAXLINK
	| FATTR4_WORD0_MAXNAME,
	0
};

234
const u32 nfs4_fsinfo_bitmap[3] = { FATTR4_WORD0_MAXFILESIZE
L
Linus Torvalds 已提交
235 236 237
			| FATTR4_WORD0_MAXREAD
			| FATTR4_WORD0_MAXWRITE
			| FATTR4_WORD0_LEASE_TIME,
238
			FATTR4_WORD1_TIME_DELTA
239 240
			| FATTR4_WORD1_FS_LAYOUT_TYPES,
			FATTR4_WORD2_LAYOUT_BLKSIZE
L
Linus Torvalds 已提交
241 242
};

243
const u32 nfs4_fs_locations_bitmap[3] = {
244 245 246 247 248 249 250 251 252 253 254 255 256 257 258
	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
259
	| FATTR4_WORD1_MOUNTED_ON_FILEID,
260 261
};

A
Al Viro 已提交
262
static void nfs4_setup_readdir(u64 cookie, __be32 *verifier, struct dentry *dentry,
L
Linus Torvalds 已提交
263 264
		struct nfs4_readdir_arg *readdir)
{
265
	__be32 *start, *p;
L
Linus Torvalds 已提交
266 267

	if (cookie > 2) {
268
		readdir->cookie = cookie;
L
Linus Torvalds 已提交
269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284
		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.
	 */
285
	start = p = kmap_atomic(*readdir->pages);
L
Linus Torvalds 已提交
286 287 288 289 290 291 292 293 294 295 296
	
	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
297
		p = xdr_encode_hyper(p, NFS_FILEID(dentry->d_inode));
L
Linus Torvalds 已提交
298 299 300 301 302 303 304 305 306 307 308
	}
	
	*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
309
	p = xdr_encode_hyper(p, NFS_FILEID(dentry->d_parent->d_inode));
L
Linus Torvalds 已提交
310 311 312

	readdir->pgbase = (char *)p - (char *)start;
	readdir->count -= readdir->pgbase;
313
	kunmap_atomic(start);
L
Linus Torvalds 已提交
314 315
}

316 317 318 319 320 321 322 323 324 325 326 327 328 329
static long nfs4_update_delay(long *timeout)
{
	long ret;
	if (!timeout)
		return NFS4_POLL_RETRY_MAX;
	if (*timeout <= 0)
		*timeout = NFS4_POLL_RETRY_MIN;
	if (*timeout > NFS4_POLL_RETRY_MAX)
		*timeout = NFS4_POLL_RETRY_MAX;
	ret = *timeout;
	*timeout <<= 1;
	return ret;
}

330 331 332 333 334 335
static int nfs4_delay(struct rpc_clnt *clnt, long *timeout)
{
	int res = 0;

	might_sleep();

336 337
	freezable_schedule_timeout_killable_unsafe(
		nfs4_update_delay(timeout));
338 339 340 341 342 343 344 345
	if (fatal_signal_pending(current))
		res = -ERESTARTSYS;
	return res;
}

/* This is the error handling routine for processes that are allowed
 * to sleep.
 */
A
Anna Schumaker 已提交
346
int nfs4_handle_exception(struct nfs_server *server, int errorcode, struct nfs4_exception *exception)
347 348
{
	struct nfs_client *clp = server->nfs_client;
349
	struct nfs4_state *state = exception->state;
350
	struct inode *inode = exception->inode;
351 352 353 354 355 356
	int ret = errorcode;

	exception->retry = 0;
	switch(errorcode) {
		case 0:
			return 0;
357
		case -NFS4ERR_OPENMODE:
358
			if (inode && nfs4_have_delegation(inode, FMODE_READ)) {
359
				nfs4_inode_return_delegation(inode);
360 361 362 363 364
				exception->retry = 1;
				return 0;
			}
			if (state == NULL)
				break;
365 366 367
			ret = nfs4_schedule_stateid_recovery(server, state);
			if (ret < 0)
				break;
368
			goto wait_on_recovery;
369
		case -NFS4ERR_DELEG_REVOKED:
370 371 372 373
		case -NFS4ERR_ADMIN_REVOKED:
		case -NFS4ERR_BAD_STATEID:
			if (state == NULL)
				break;
374 375 376
			ret = nfs4_schedule_stateid_recovery(server, state);
			if (ret < 0)
				break;
377
			goto wait_on_recovery;
378
		case -NFS4ERR_EXPIRED:
379 380 381 382 383
			if (state != NULL) {
				ret = nfs4_schedule_stateid_recovery(server, state);
				if (ret < 0)
					break;
			}
384
		case -NFS4ERR_STALE_STATEID:
385
		case -NFS4ERR_STALE_CLIENTID:
386 387
			nfs4_schedule_lease_recovery(clp);
			goto wait_on_recovery;
388 389 390 391 392
		case -NFS4ERR_MOVED:
			ret = nfs4_schedule_migration_recovery(server);
			if (ret < 0)
				break;
			goto wait_on_recovery;
393 394 395
		case -NFS4ERR_LEASE_MOVED:
			nfs4_schedule_lease_moved_recovery(clp);
			goto wait_on_recovery;
396
#if defined(CONFIG_NFS_V4_1)
397 398 399 400 401 402 403 404 405
		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);
406
			nfs4_schedule_session_recovery(clp->cl_session, errorcode);
407
			goto wait_on_recovery;
408
#endif /* defined(CONFIG_NFS_V4_1) */
409
		case -NFS4ERR_FILE_OPEN:
410 411 412 413 414 415 416
			if (exception->timeout > HZ) {
				/* We have retried a decent amount, time to
				 * fail
				 */
				ret = -EBUSY;
				break;
			}
417 418 419 420 421
		case -NFS4ERR_GRACE:
		case -NFS4ERR_DELAY:
			ret = nfs4_delay(server->client, &exception->timeout);
			if (ret != 0)
				break;
422
		case -NFS4ERR_RETRY_UNCACHED_REP:
423 424
		case -NFS4ERR_OLD_STATEID:
			exception->retry = 1;
425 426 427 428 429 430 431 432 433 434 435 436 437
			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);
			}
438 439 440
	}
	/* We failed to handle the error */
	return nfs4_map_errors(ret);
441
wait_on_recovery:
442
	ret = nfs4_wait_clnt_recover(clp);
443 444
	if (test_bit(NFS_MIG_FAILED, &server->mig_status))
		return -EIO;
445 446 447
	if (ret == 0)
		exception->retry = 1;
	return ret;
448 449
}

450 451 452 453 454 455 456 457 458 459 460 461 462 463
/*
 * 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;
}
464

465
static void do_renew_lease(struct nfs_client *clp, unsigned long timestamp)
L
Linus Torvalds 已提交
466 467 468 469 470 471 472
{
	spin_lock(&clp->cl_lock);
	if (time_before(clp->cl_last_renewal,timestamp))
		clp->cl_last_renewal = timestamp;
	spin_unlock(&clp->cl_lock);
}

473 474 475 476 477
static void renew_lease(const struct nfs_server *server, unsigned long timestamp)
{
	do_renew_lease(server->nfs_client, timestamp);
}

478 479 480 481 482 483
struct nfs4_call_sync_data {
	const struct nfs_server *seq_server;
	struct nfs4_sequence_args *seq_args;
	struct nfs4_sequence_res *seq_res;
};

484 485 486 487 488 489 490 491 492 493 494 495 496 497 498
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;
}

499 500 501 502
int nfs40_setup_sequence(struct nfs4_slot_table *tbl,
			 struct nfs4_sequence_args *args,
			 struct nfs4_sequence_res *res,
			 struct rpc_task *task)
C
Chuck Lever 已提交
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
{
	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;
}
538
EXPORT_SYMBOL_GPL(nfs40_setup_sequence);
C
Chuck Lever 已提交
539 540 541 542 543 544 545

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;

546
	if (slot == NULL)
C
Chuck Lever 已提交
547 548 549 550 551 552 553 554 555 556 557 558 559
		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
560 561
#if defined(CONFIG_NFS_V4_1)

562
static void nfs41_sequence_free_slot(struct nfs4_sequence_res *res)
A
Andy Adamson 已提交
563
{
564
	struct nfs4_session *session;
A
Andy Adamson 已提交
565
	struct nfs4_slot_table *tbl;
566
	struct nfs4_slot *slot = res->sr_slot;
567
	bool send_new_highest_used_slotid = false;
A
Andy Adamson 已提交
568

569
	tbl = slot->table;
570
	session = tbl->session;
571

572
	spin_lock(&tbl->slot_tbl_lock);
573 574 575 576 577 578
	/* 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;

579
	if (nfs41_wake_and_assign_slot(tbl, slot)) {
580 581 582
		send_new_highest_used_slotid = false;
		goto out_unlock;
	}
583
	nfs4_free_slot(tbl, slot);
584 585 586

	if (tbl->highest_used_slotid != NFS4_NO_SLOT)
		send_new_highest_used_slotid = false;
587
out_unlock:
588
	spin_unlock(&tbl->slot_tbl_lock);
589
	res->sr_slot = NULL;
590 591
	if (send_new_highest_used_slotid)
		nfs41_server_notify_highest_slotid_update(session->clp);
A
Andy Adamson 已提交
592 593
}

594
int nfs41_sequence_done(struct rpc_task *task, struct nfs4_sequence_res *res)
A
Andy Adamson 已提交
595
{
596
	struct nfs4_session *session;
597
	struct nfs4_slot *slot = res->sr_slot;
598
	struct nfs_client *clp;
599
	bool interrupted = false;
600
	int ret = 1;
A
Andy Adamson 已提交
601

602 603
	if (slot == NULL)
		goto out_noaction;
604 605
	/* don't increment the sequence number if the task wasn't sent */
	if (!RPC_WAS_SENT(task))
A
Andy Adamson 已提交
606 607
		goto out;

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
out_noaction:
683
	return ret;
684 685 686 687 688 689
retry_nowait:
	if (rpc_restart_call_prepare(task)) {
		task->tk_status = 0;
		ret = 0;
	}
	goto out;
690
out_retry:
691
	if (!rpc_restart_call(task))
692 693 694
		goto out;
	rpc_delay(task, NFS4_POLL_RETRY_MAX);
	return 0;
A
Andy Adamson 已提交
695
}
696
EXPORT_SYMBOL_GPL(nfs41_sequence_done);
A
Andy Adamson 已提交
697

P
Peng Tao 已提交
698
int nfs4_sequence_done(struct rpc_task *task, struct nfs4_sequence_res *res)
699
{
700
	if (res->sr_slot == NULL)
701
		return 1;
C
Chuck Lever 已提交
702 703
	if (!res->sr_slot->table->session)
		return nfs40_sequence_done(task, res);
704
	return nfs41_sequence_done(task, res);
705
}
P
Peng Tao 已提交
706
EXPORT_SYMBOL_GPL(nfs4_sequence_done);
707

A
Andy Adamson 已提交
708
int nfs41_setup_sequence(struct nfs4_session *session,
A
Andy Adamson 已提交
709 710 711 712
				struct nfs4_sequence_args *args,
				struct nfs4_sequence_res *res,
				struct rpc_task *task)
{
A
Andy Adamson 已提交
713 714 715 716
	struct nfs4_slot *slot;
	struct nfs4_slot_table *tbl;

	dprintk("--> %s\n", __func__);
A
Andy Adamson 已提交
717
	/* slot already allocated? */
718
	if (res->sr_slot != NULL)
719
		goto out_success;
A
Andy Adamson 已提交
720

A
Andy Adamson 已提交
721 722
	tbl = &session->fc_slot_table;

723 724
	task->tk_timeout = 0;

A
Andy Adamson 已提交
725
	spin_lock(&tbl->slot_tbl_lock);
726
	if (test_bit(NFS4_SLOT_TBL_DRAINING, &tbl->slot_tbl_state) &&
727
	    !args->sa_privileged) {
728 729
		/* The state manager will wait until the slot table is empty */
		dprintk("%s session is draining\n", __func__);
730
		goto out_sleep;
731 732
	}

733
	slot = nfs4_alloc_slot(tbl);
734 735 736 737
	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
738
		dprintk("<-- %s: no free slots\n", __func__);
739
		goto out_sleep;
A
Andy Adamson 已提交
740 741 742
	}
	spin_unlock(&tbl->slot_tbl_lock);

743
	args->sa_slot = slot;
A
Andy Adamson 已提交
744

745
	dprintk("<-- %s slotid=%u seqid=%u\n", __func__,
746
			slot->slot_nr, slot->seq_nr);
A
Andy Adamson 已提交
747

748
	res->sr_slot = slot;
749
	res->sr_timestamp = jiffies;
750
	res->sr_status_flags = 0;
A
Andy Adamson 已提交
751 752 753 754 755
	/*
	 * 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;
756
	trace_nfs4_setup_sequence(session, args);
757 758
out_success:
	rpc_call_start(task);
A
Andy Adamson 已提交
759
	return 0;
760
out_sleep:
761 762
	/* Privileged tasks are queued with top priority */
	if (args->sa_privileged)
763 764 765 766
		rpc_sleep_on_priority(&tbl->slot_tbl_waitq, task,
				NULL, RPC_PRIORITY_PRIVILEGED);
	else
		rpc_sleep_on(&tbl->slot_tbl_waitq, task, NULL);
767 768
	spin_unlock(&tbl->slot_tbl_lock);
	return -EAGAIN;
A
Andy Adamson 已提交
769
}
A
Andy Adamson 已提交
770
EXPORT_SYMBOL_GPL(nfs41_setup_sequence);
A
Andy Adamson 已提交
771

772 773 774 775
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
776
{
777
	struct nfs4_session *session = nfs4_get_session(server);
A
Andy Adamson 已提交
778 779
	int ret = 0;

C
Chuck Lever 已提交
780
	if (!session)
781 782
		return nfs40_setup_sequence(server->nfs_client->cl_slot_tbl,
					    args, res, task);
783

784
	dprintk("--> %s clp %p session %p sr_slot %u\n",
785
		__func__, session->clp, session, res->sr_slot ?
786
			res->sr_slot->slot_nr : NFS4_NO_SLOT);
A
Andy Adamson 已提交
787

788
	ret = nfs41_setup_sequence(session, args, res, task);
C
Chuck Lever 已提交
789

A
Andy Adamson 已提交
790 791 792 793 794 795
	dprintk("<-- %s status=%d\n", __func__, ret);
	return ret;
}

static void nfs41_call_sync_prepare(struct rpc_task *task, void *calldata)
{
796
	struct nfs4_call_sync_data *data = calldata;
797
	struct nfs4_session *session = nfs4_get_session(data->seq_server);
A
Andy Adamson 已提交
798

799 800
	dprintk("--> %s data->seq_server %p\n", __func__, data->seq_server);

801
	nfs41_setup_sequence(session, data->seq_args, data->seq_res, task);
A
Andy Adamson 已提交
802 803
}

A
Andy Adamson 已提交
804 805
static void nfs41_call_sync_done(struct rpc_task *task, void *calldata)
{
806
	struct nfs4_call_sync_data *data = calldata;
A
Andy Adamson 已提交
807

808
	nfs41_sequence_done(task, data->seq_res);
A
Andy Adamson 已提交
809 810
}

811
static const struct rpc_call_ops nfs41_call_sync_ops = {
A
Andy Adamson 已提交
812
	.rpc_call_prepare = nfs41_call_sync_prepare,
A
Andy Adamson 已提交
813
	.rpc_call_done = nfs41_call_sync_done,
A
Andy Adamson 已提交
814 815
};

C
Chuck Lever 已提交
816 817
#else	/* !CONFIG_NFS_V4_1 */

818 819 820 821 822
static int nfs4_setup_sequence(const struct nfs_server *server,
			       struct nfs4_sequence_args *args,
			       struct nfs4_sequence_res *res,
			       struct rpc_task *task)
{
823 824
	return nfs40_setup_sequence(server->nfs_client->cl_slot_tbl,
				    args, res, task);
825 826
}

P
Peng Tao 已提交
827 828
int nfs4_sequence_done(struct rpc_task *task,
		       struct nfs4_sequence_res *res)
829
{
C
Chuck Lever 已提交
830
	return nfs40_sequence_done(task, res);
831
}
P
Peng Tao 已提交
832
EXPORT_SYMBOL_GPL(nfs4_sequence_done);
C
Chuck Lever 已提交
833 834

#endif	/* !CONFIG_NFS_V4_1 */
835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853

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,
};

854 855
static int nfs4_call_sync_sequence(struct rpc_clnt *clnt,
				   struct nfs_server *server,
A
Andy Adamson 已提交
856 857
				   struct rpc_message *msg,
				   struct nfs4_sequence_args *args,
858
				   struct nfs4_sequence_res *res)
A
Andy Adamson 已提交
859 860 861
{
	int ret;
	struct rpc_task *task;
862
	struct nfs_client *clp = server->nfs_client;
863
	struct nfs4_call_sync_data data = {
864
		.seq_server = server,
A
Andy Adamson 已提交
865 866 867 868
		.seq_args = args,
		.seq_res = res,
	};
	struct rpc_task_setup task_setup = {
869
		.rpc_client = clnt,
A
Andy Adamson 已提交
870
		.rpc_message = msg,
871
		.callback_ops = clp->cl_mvops->call_sync_ops,
A
Andy Adamson 已提交
872 873 874 875 876 877 878 879 880 881 882 883 884
		.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;
}

885 886
int nfs4_call_sync(struct rpc_clnt *clnt,
		   struct nfs_server *server,
887 888 889 890 891
		   struct rpc_message *msg,
		   struct nfs4_sequence_args *args,
		   struct nfs4_sequence_res *res,
		   int cache_reply)
{
892
	nfs4_init_sequence(args, res, cache_reply);
893
	return nfs4_call_sync_sequence(clnt, server, msg, args, res);
894
}
A
Andy Adamson 已提交
895

896
static void update_changeattr(struct inode *dir, struct nfs4_change_info *cinfo)
L
Linus Torvalds 已提交
897
{
898
	struct nfs_inode *nfsi = NFS_I(dir);
L
Linus Torvalds 已提交
899

900
	spin_lock(&dir->i_lock);
901
	nfsi->cache_validity |= NFS_INO_INVALID_ATTR|NFS_INO_INVALID_DATA;
902
	if (!cinfo->atomic || cinfo->before != dir->i_version)
903
		nfs_force_lookup_revalidate(dir);
904
	dir->i_version = cinfo->after;
905
	nfsi->attr_gencount = nfs_inc_attr_generation_counter();
D
David Howells 已提交
906
	nfs_fscache_invalidate(dir);
907
	spin_unlock(&dir->i_lock);
L
Linus Torvalds 已提交
908 909
}

910
struct nfs4_opendata {
911
	struct kref kref;
912 913
	struct nfs_openargs o_arg;
	struct nfs_openres o_res;
914 915
	struct nfs_open_confirmargs c_arg;
	struct nfs_open_confirmres c_res;
916 917
	struct nfs4_string owner_name;
	struct nfs4_string group_name;
918
	struct nfs_fattr f_attr;
919
	struct nfs4_label *f_label;
920
	struct dentry *dir;
921
	struct dentry *dentry;
922
	struct nfs4_state_owner *owner;
923
	struct nfs4_state *state;
924
	struct iattr attrs;
925
	unsigned long timestamp;
926
	unsigned int rpc_done : 1;
927
	unsigned int file_created : 1;
928
	unsigned int is_recover : 1;
929 930
	int rpc_status;
	int cancelled;
931 932
};

933 934 935 936 937 938 939 940 941 942 943 944
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;
}

945 946 947 948 949 950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968 969
static u32
nfs4_map_atomic_open_share(struct nfs_server *server,
		fmode_t fmode, int openflags)
{
	u32 res = 0;

	switch (fmode & (FMODE_READ | FMODE_WRITE)) {
	case FMODE_READ:
		res = NFS4_SHARE_ACCESS_READ;
		break;
	case FMODE_WRITE:
		res = NFS4_SHARE_ACCESS_WRITE;
		break;
	case FMODE_READ|FMODE_WRITE:
		res = NFS4_SHARE_ACCESS_BOTH;
	}
	if (!(server->caps & NFS_CAP_ATOMIC_OPEN_V1))
		goto out;
	/* Want no delegation if we're using O_DIRECT */
	if (openflags & O_DIRECT)
		res |= NFS4_SHARE_WANT_NO_DELEG;
out:
	return res;
}

970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986
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;
	}
}
987 988 989 990

static void nfs4_init_opendata_res(struct nfs4_opendata *p)
{
	p->o_res.f_attr = &p->f_attr;
991
	p->o_res.f_label = p->f_label;
992 993
	p->o_res.seqid = p->o_arg.seqid;
	p->c_res.seqid = p->c_arg.seqid;
994
	p->o_res.server = p->o_arg.server;
995
	p->o_res.access_request = p->o_arg.access;
996
	nfs_fattr_init(&p->f_attr);
997
	nfs_fattr_init_names(&p->f_attr, &p->owner_name, &p->group_name);
998 999
}

1000
static struct nfs4_opendata *nfs4_opendata_alloc(struct dentry *dentry,
1001
		struct nfs4_state_owner *sp, fmode_t fmode, int flags,
1002
		const struct iattr *attrs,
1003
		struct nfs4_label *label,
1004
		enum open_claim_type4 claim,
1005
		gfp_t gfp_mask)
1006
{
1007
	struct dentry *parent = dget_parent(dentry);
1008 1009
	struct inode *dir = parent->d_inode;
	struct nfs_server *server = NFS_SERVER(dir);
1010
	struct nfs_seqid *(*alloc_seqid)(struct nfs_seqid_counter *, gfp_t);
1011 1012
	struct nfs4_opendata *p;

1013
	p = kzalloc(sizeof(*p), gfp_mask);
1014 1015
	if (p == NULL)
		goto err;
1016 1017 1018 1019 1020

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

1021 1022
	alloc_seqid = server->nfs_client->cl_mvops->alloc_seqid;
	p->o_arg.seqid = alloc_seqid(&sp->so_seqid, gfp_mask);
1023
	if (IS_ERR(p->o_arg.seqid))
1024
		goto err_free_label;
1025 1026
	nfs_sb_active(dentry->d_sb);
	p->dentry = dget(dentry);
1027 1028 1029
	p->dir = parent;
	p->owner = sp;
	atomic_inc(&sp->so_count);
1030 1031
	p->o_arg.open_flags = flags;
	p->o_arg.fmode = fmode & (FMODE_READ|FMODE_WRITE);
1032 1033
	p->o_arg.share_access = nfs4_map_atomic_open_share(server,
			fmode, flags);
1034 1035 1036 1037 1038 1039 1040 1041
	/* 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;
	}
1042
	p->o_arg.clientid = server->nfs_client->cl_clientid;
1043 1044
	p->o_arg.id.create_time = ktime_to_ns(sp->so_seqid.create_time);
	p->o_arg.id.uniquifier = sp->so_seqid.owner_id;
1045
	p->o_arg.name = &dentry->d_name;
1046
	p->o_arg.server = server;
1047
	p->o_arg.bitmask = nfs4_bitmask(server, label);
1048
	p->o_arg.open_bitmap = &nfs4_fattr_bitmap[0];
1049
	p->o_arg.label = label;
1050 1051
	p->o_arg.claim = nfs4_map_atomic_open_claim(server, claim);
	switch (p->o_arg.claim) {
1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062
	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);
	}
1063
	if (attrs != NULL && attrs->ia_valid != 0) {
1064
		__u32 verf[2];
1065

1066 1067
		p->o_arg.u.attrs = &p->attrs;
		memcpy(&p->attrs, attrs, sizeof(p->attrs));
1068 1069 1070 1071 1072

		verf[0] = jiffies;
		verf[1] = current->pid;
		memcpy(p->o_arg.u.verifier.data, verf,
				sizeof(p->o_arg.u.verifier.data));
1073
	}
1074 1075 1076
	p->c_arg.fh = &p->o_res.fh;
	p->c_arg.stateid = &p->o_res.stateid;
	p->c_arg.seqid = p->o_arg.seqid;
1077
	nfs4_init_opendata_res(p);
1078
	kref_init(&p->kref);
1079
	return p;
1080 1081 1082 1083

err_free_label:
	nfs4_label_free(p->f_label);
err_free_p:
1084 1085 1086 1087 1088 1089
	kfree(p);
err:
	dput(parent);
	return NULL;
}

1090
static void nfs4_opendata_free(struct kref *kref)
1091
{
1092 1093
	struct nfs4_opendata *p = container_of(kref,
			struct nfs4_opendata, kref);
1094
	struct super_block *sb = p->dentry->d_sb;
1095 1096

	nfs_free_seqid(p->o_arg.seqid);
1097 1098
	if (p->state != NULL)
		nfs4_put_open_state(p->state);
1099
	nfs4_put_state_owner(p->owner);
1100 1101 1102

	nfs4_label_free(p->f_label);

1103
	dput(p->dir);
1104 1105
	dput(p->dentry);
	nfs_sb_deactive(sb);
1106
	nfs_fattr_free_names(&p->f_attr);
1107
	kfree(p->f_attr.mdsthreshold);
1108 1109 1110 1111 1112 1113 1114
	kfree(p);
}

static void nfs4_opendata_put(struct nfs4_opendata *p)
{
	if (p != NULL)
		kref_put(&p->kref, nfs4_opendata_free);
1115 1116
}

1117 1118 1119 1120 1121 1122 1123 1124
static int nfs4_wait_for_completion_rpc_task(struct rpc_task *task)
{
	int ret;

	ret = rpc_wait_for_completion_task(task);
	return ret;
}

1125
static int can_open_cached(struct nfs4_state *state, fmode_t mode, int open_mode)
1126 1127
{
	int ret = 0;
1128

1129
	if (open_mode & (O_EXCL|O_TRUNC))
1130 1131
		goto out;
	switch (mode & (FMODE_READ|FMODE_WRITE)) {
1132
		case FMODE_READ:
1133 1134
			ret |= test_bit(NFS_O_RDONLY_STATE, &state->flags) != 0
				&& state->n_rdonly != 0;
1135 1136
			break;
		case FMODE_WRITE:
1137 1138
			ret |= test_bit(NFS_O_WRONLY_STATE, &state->flags) != 0
				&& state->n_wronly != 0;
1139 1140
			break;
		case FMODE_READ|FMODE_WRITE:
1141 1142
			ret |= test_bit(NFS_O_RDWR_STATE, &state->flags) != 0
				&& state->n_rdwr != 0;
1143
	}
1144
out:
1145 1146 1147
	return ret;
}

1148
static int can_open_delegated(struct nfs_delegation *delegation, fmode_t fmode)
1149
{
1150 1151
	if (delegation == NULL)
		return 0;
1152
	if ((delegation->type & fmode) != fmode)
1153
		return 0;
1154 1155
	if (test_bit(NFS_DELEGATION_RETURNING, &delegation->flags))
		return 0;
1156
	nfs_mark_delegation_referenced(delegation);
1157 1158 1159
	return 1;
}

1160
static void update_open_stateflags(struct nfs4_state *state, fmode_t fmode)
1161
{
1162
	switch (fmode) {
1163 1164 1165 1166 1167 1168 1169 1170 1171
		case FMODE_WRITE:
			state->n_wronly++;
			break;
		case FMODE_READ:
			state->n_rdonly++;
			break;
		case FMODE_READ|FMODE_WRITE:
			state->n_rdwr++;
	}
1172
	nfs4_state_set_mode_locked(state, state->state | fmode);
1173 1174
}

1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189
static void nfs_test_and_clear_all_open_stateid(struct nfs4_state *state)
{
	struct nfs_client *clp = state->owner->so_server->nfs_client;
	bool need_recover = false;

	if (test_and_clear_bit(NFS_O_RDONLY_STATE, &state->flags) && state->n_rdonly)
		need_recover = true;
	if (test_and_clear_bit(NFS_O_WRONLY_STATE, &state->flags) && state->n_wronly)
		need_recover = true;
	if (test_and_clear_bit(NFS_O_RDWR_STATE, &state->flags) && state->n_rdwr)
		need_recover = true;
	if (need_recover)
		nfs4_state_mark_reclaim_nograce(clp, state);
}

1190 1191 1192 1193 1194
static bool nfs_need_update_open_stateid(struct nfs4_state *state,
		nfs4_stateid *stateid)
{
	if (test_and_set_bit(NFS_OPEN_STATE, &state->flags) == 0)
		return true;
1195 1196
	if (!nfs4_stateid_match_other(stateid, &state->open_stateid)) {
		nfs_test_and_clear_all_open_stateid(state);
1197
		return true;
1198
	}
1199 1200 1201 1202 1203
	if (nfs4_stateid_is_newer(stateid, &state->open_stateid))
		return true;
	return false;
}

T
Trond Myklebust 已提交
1204 1205 1206 1207 1208 1209 1210 1211 1212 1213
static void nfs_resync_open_stateid_locked(struct nfs4_state *state)
{
	if (state->n_wronly)
		set_bit(NFS_O_WRONLY_STATE, &state->flags);
	if (state->n_rdonly)
		set_bit(NFS_O_RDONLY_STATE, &state->flags);
	if (state->n_rdwr)
		set_bit(NFS_O_RDWR_STATE, &state->flags);
}

1214 1215
static void nfs_clear_open_stateid_locked(struct nfs4_state *state,
		nfs4_stateid *stateid, fmode_t fmode)
1216
{
1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230 1231
	clear_bit(NFS_O_RDWR_STATE, &state->flags);
	switch (fmode & (FMODE_READ|FMODE_WRITE)) {
	case FMODE_WRITE:
		clear_bit(NFS_O_RDONLY_STATE, &state->flags);
		break;
	case FMODE_READ:
		clear_bit(NFS_O_WRONLY_STATE, &state->flags);
		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);
	}
	if (stateid == NULL)
		return;
T
Trond Myklebust 已提交
1232 1233 1234 1235
	/* Handle races with OPEN */
	if (!nfs4_stateid_match_other(stateid, &state->open_stateid) ||
	    !nfs4_stateid_is_newer(stateid, &state->open_stateid)) {
		nfs_resync_open_stateid_locked(state);
1236
		return;
T
Trond Myklebust 已提交
1237
	}
1238
	if (test_bit(NFS_DELEGATED_STATE, &state->flags) == 0)
1239 1240
		nfs4_stateid_copy(&state->stateid, stateid);
	nfs4_stateid_copy(&state->open_stateid, stateid);
1241 1242 1243 1244 1245 1246 1247
}

static void nfs_clear_open_stateid(struct nfs4_state *state, nfs4_stateid *stateid, fmode_t fmode)
{
	write_seqlock(&state->seqlock);
	nfs_clear_open_stateid_locked(state, stateid, fmode);
	write_sequnlock(&state->seqlock);
1248 1249
	if (test_bit(NFS_STATE_RECLAIM_NOGRACE, &state->flags))
		nfs4_schedule_state_manager(state->owner->so_server->nfs_client);
1250 1251
}

1252
static void nfs_set_open_stateid_locked(struct nfs4_state *state, nfs4_stateid *stateid, fmode_t fmode)
1253
{
1254
	switch (fmode) {
1255 1256 1257 1258 1259 1260 1261 1262 1263
		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);
	}
1264 1265 1266 1267 1268
	if (!nfs_need_update_open_stateid(state, stateid))
		return;
	if (test_bit(NFS_DELEGATED_STATE, &state->flags) == 0)
		nfs4_stateid_copy(&state->stateid, stateid);
	nfs4_stateid_copy(&state->open_stateid, stateid);
1269 1270
}

1271
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
1272
{
1273 1274 1275 1276 1277
	/*
	 * Protect the call to nfs4_state_set_mode_locked and
	 * serialise the stateid update
	 */
	write_seqlock(&state->seqlock);
1278
	if (deleg_stateid != NULL) {
1279
		nfs4_stateid_copy(&state->stateid, deleg_stateid);
1280 1281 1282
		set_bit(NFS_DELEGATED_STATE, &state->flags);
	}
	if (open_stateid != NULL)
1283
		nfs_set_open_stateid_locked(state, open_stateid, fmode);
1284 1285
	write_sequnlock(&state->seqlock);
	spin_lock(&state->owner->so_lock);
1286
	update_open_stateflags(state, fmode);
1287
	spin_unlock(&state->owner->so_lock);
L
Linus Torvalds 已提交
1288 1289
}

1290
static int update_open_stateid(struct nfs4_state *state, nfs4_stateid *open_stateid, nfs4_stateid *delegation, fmode_t fmode)
1291 1292 1293 1294 1295
{
	struct nfs_inode *nfsi = NFS_I(state->inode);
	struct nfs_delegation *deleg_cur;
	int ret = 0;

1296
	fmode &= (FMODE_READ|FMODE_WRITE);
1297 1298 1299 1300 1301 1302 1303

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

	spin_lock(&deleg_cur->lock);
1304
	if (rcu_dereference(nfsi->delegation) != deleg_cur ||
1305
	   test_bit(NFS_DELEGATION_RETURNING, &deleg_cur->flags) ||
1306
	    (deleg_cur->type & fmode) != fmode)
1307 1308 1309 1310
		goto no_delegation_unlock;

	if (delegation == NULL)
		delegation = &deleg_cur->stateid;
1311
	else if (!nfs4_stateid_match(&deleg_cur->stateid, delegation))
1312 1313
		goto no_delegation_unlock;

1314
	nfs_mark_delegation_referenced(deleg_cur);
1315
	__update_open_stateid(state, open_stateid, &deleg_cur->stateid, fmode);
1316 1317 1318 1319 1320 1321 1322
	ret = 1;
no_delegation_unlock:
	spin_unlock(&deleg_cur->lock);
no_delegation:
	rcu_read_unlock();

	if (!ret && open_stateid != NULL) {
1323
		__update_open_stateid(state, open_stateid, NULL, fmode);
1324 1325
		ret = 1;
	}
1326 1327
	if (test_bit(NFS_STATE_RECLAIM_NOGRACE, &state->flags))
		nfs4_schedule_state_manager(state->owner->so_server->nfs_client);
1328 1329 1330 1331

	return ret;
}

1332 1333 1334 1335 1336 1337 1338 1339 1340 1341 1342 1343 1344 1345 1346 1347 1348
static bool nfs4_update_lock_stateid(struct nfs4_lock_state *lsp,
		const nfs4_stateid *stateid)
{
	struct nfs4_state *state = lsp->ls_state;
	bool ret = false;

	spin_lock(&state->state_lock);
	if (!nfs4_stateid_match_other(stateid, &lsp->ls_stateid))
		goto out_noupdate;
	if (!nfs4_stateid_is_newer(stateid, &lsp->ls_stateid))
		goto out_noupdate;
	nfs4_stateid_copy(&lsp->ls_stateid, stateid);
	ret = true;
out_noupdate:
	spin_unlock(&state->state_lock);
	return ret;
}
1349

1350
static void nfs4_return_incompatible_delegation(struct inode *inode, fmode_t fmode)
1351 1352 1353 1354 1355
{
	struct nfs_delegation *delegation;

	rcu_read_lock();
	delegation = rcu_dereference(NFS_I(inode)->delegation);
1356
	if (delegation == NULL || (delegation->type & fmode) == fmode) {
1357 1358 1359 1360
		rcu_read_unlock();
		return;
	}
	rcu_read_unlock();
1361
	nfs4_inode_return_delegation(inode);
1362 1363
}

1364
static struct nfs4_state *nfs4_try_open_cached(struct nfs4_opendata *opendata)
1365 1366 1367 1368
{
	struct nfs4_state *state = opendata->state;
	struct nfs_inode *nfsi = NFS_I(state->inode);
	struct nfs_delegation *delegation;
1369
	int open_mode = opendata->o_arg.open_flags;
1370
	fmode_t fmode = opendata->o_arg.fmode;
1371 1372 1373 1374
	nfs4_stateid stateid;
	int ret = -EAGAIN;

	for (;;) {
1375
		spin_lock(&state->owner->so_lock);
1376
		if (can_open_cached(state, fmode, open_mode)) {
1377
			update_open_stateflags(state, fmode);
1378
			spin_unlock(&state->owner->so_lock);
1379
			goto out_return_state;
1380
		}
1381
		spin_unlock(&state->owner->so_lock);
1382 1383
		rcu_read_lock();
		delegation = rcu_dereference(nfsi->delegation);
1384
		if (!can_open_delegated(delegation, fmode)) {
1385
			rcu_read_unlock();
1386
			break;
1387
		}
1388
		/* Save the delegation */
1389
		nfs4_stateid_copy(&stateid, &delegation->stateid);
1390
		rcu_read_unlock();
1391
		nfs_release_seqid(opendata->o_arg.seqid);
1392 1393 1394 1395 1396
		if (!opendata->is_recover) {
			ret = nfs_may_open(state->inode, state->owner->so_cred, open_mode);
			if (ret != 0)
				goto out;
		}
1397
		ret = -EAGAIN;
1398 1399

		/* Try to update the stateid using the delegation */
1400
		if (update_open_stateid(state, NULL, &stateid, fmode))
1401
			goto out_return_state;
1402 1403 1404 1405 1406 1407 1408 1409
	}
out:
	return ERR_PTR(ret);
out_return_state:
	atomic_inc(&state->count);
	return state;
}

1410 1411 1412 1413 1414 1415 1416 1417 1418 1419 1420 1421 1422 1423 1424 1425 1426 1427 1428 1429 1430 1431 1432 1433 1434 1435 1436 1437 1438 1439 1440 1441 1442 1443 1444 1445 1446 1447
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;

1448 1449 1450 1451 1452 1453 1454
	if (!data->rpc_done) {
		if (data->rpc_status) {
			ret = data->rpc_status;
			goto err;
		}
		/* cached opens have already been processed */
		goto update;
1455 1456 1457 1458 1459 1460 1461 1462
	}

	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);
1463
update:
1464 1465
	update_open_stateid(state, &data->o_res.stateid, NULL,
			    data->o_arg.fmode);
1466
	atomic_inc(&state->count);
1467 1468 1469 1470 1471 1472 1473 1474 1475

	return state;
err:
	return ERR_PTR(ret);

}

static struct nfs4_state *
_nfs4_opendata_to_nfs4_state(struct nfs4_opendata *data)
1476 1477 1478
{
	struct inode *inode;
	struct nfs4_state *state = NULL;
1479
	int ret;
1480

1481
	if (!data->rpc_done) {
1482
		state = nfs4_try_open_cached(data);
1483 1484 1485
		goto out;
	}

1486
	ret = -EAGAIN;
1487
	if (!(data->f_attr.valid & NFS_ATTR_FATTR))
1488
		goto err;
1489
	inode = nfs_fhget(data->dir->d_sb, &data->o_res.fh, &data->f_attr, data->f_label);
1490
	ret = PTR_ERR(inode);
1491
	if (IS_ERR(inode))
1492 1493
		goto err;
	ret = -ENOMEM;
1494 1495
	state = nfs4_get_open_state(inode, data->owner);
	if (state == NULL)
1496
		goto err_put_inode;
1497 1498
	if (data->o_res.delegation_type != 0)
		nfs4_opendata_check_deleg(data, state);
1499
	update_open_stateid(state, &data->o_res.stateid, NULL,
1500
			data->o_arg.fmode);
1501
	iput(inode);
1502
out:
1503
	nfs_release_seqid(data->o_arg.seqid);
1504
	return state;
1505 1506 1507 1508
err_put_inode:
	iput(inode);
err:
	return ERR_PTR(ret);
1509 1510
}

1511 1512 1513 1514 1515 1516 1517 1518
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);
}

1519 1520 1521 1522 1523 1524 1525 1526 1527 1528 1529 1530 1531 1532 1533 1534 1535
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);
}

1536 1537
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
1538 1539 1540
{
	struct nfs4_opendata *opendata;

1541
	opendata = nfs4_opendata_alloc(ctx->dentry, state->owner, 0, 0,
1542
			NULL, NULL, claim, GFP_NOFS);
T
Trond Myklebust 已提交
1543 1544 1545 1546 1547 1548 1549
	if (opendata == NULL)
		return ERR_PTR(-ENOMEM);
	opendata->state = state;
	atomic_inc(&state->count);
	return opendata;
}

1550
static int nfs4_open_recover_helper(struct nfs4_opendata *opendata, fmode_t fmode, struct nfs4_state **res)
1551
{
1552
	struct nfs4_state *newstate;
1553 1554
	int ret;

1555 1556
	opendata->o_arg.open_flags = 0;
	opendata->o_arg.fmode = fmode;
1557 1558 1559
	opendata->o_arg.share_access = nfs4_map_atomic_open_share(
			NFS_SB(opendata->dentry->d_sb),
			fmode, 0);
1560 1561 1562
	memset(&opendata->o_res, 0, sizeof(opendata->o_res));
	memset(&opendata->c_res, 0, sizeof(opendata->c_res));
	nfs4_init_opendata_res(opendata);
1563
	ret = _nfs4_recover_proc_open(opendata);
1564 1565
	if (ret != 0)
		return ret; 
1566
	newstate = nfs4_opendata_to_nfs4_state(opendata);
1567 1568
	if (IS_ERR(newstate))
		return PTR_ERR(newstate);
1569
	nfs4_close_state(newstate, fmode);
1570
	*res = newstate;
1571 1572 1573 1574 1575 1576 1577 1578
	return 0;
}

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

1579 1580 1581 1582
	/* Don't trigger recovery in nfs_test_and_clear_all_open_stateid */
	clear_bit(NFS_O_RDWR_STATE, &state->flags);
	clear_bit(NFS_O_WRONLY_STATE, &state->flags);
	clear_bit(NFS_O_RDONLY_STATE, &state->flags);
1583
	/* memory barrier prior to reading state->n_* */
1584
	clear_bit(NFS_DELEGATED_STATE, &state->flags);
1585
	clear_bit(NFS_OPEN_STATE, &state->flags);
1586 1587
	smp_rmb();
	if (state->n_rdwr != 0) {
1588
		ret = nfs4_open_recover_helper(opendata, FMODE_READ|FMODE_WRITE, &newstate);
1589 1590
		if (ret != 0)
			return ret;
1591 1592
		if (newstate != state)
			return -ESTALE;
1593 1594
	}
	if (state->n_wronly != 0) {
1595
		ret = nfs4_open_recover_helper(opendata, FMODE_WRITE, &newstate);
1596 1597
		if (ret != 0)
			return ret;
1598 1599
		if (newstate != state)
			return -ESTALE;
1600 1601
	}
	if (state->n_rdonly != 0) {
1602
		ret = nfs4_open_recover_helper(opendata, FMODE_READ, &newstate);
1603 1604
		if (ret != 0)
			return ret;
1605 1606
		if (newstate != state)
			return -ESTALE;
1607
	}
1608 1609 1610 1611 1612
	/*
	 * 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 &&
1613
	    !nfs4_stateid_match(&state->stateid, &state->open_stateid)) {
1614
		write_seqlock(&state->seqlock);
1615
		if (test_bit(NFS_DELEGATED_STATE, &state->flags) == 0)
1616
			nfs4_stateid_copy(&state->stateid, &state->open_stateid);
1617
		write_sequnlock(&state->seqlock);
1618
	}
1619 1620 1621
	return 0;
}

L
Linus Torvalds 已提交
1622 1623 1624 1625
/*
 * OPEN_RECLAIM:
 * 	reclaim state on the server after a reboot.
 */
1626
static int _nfs4_do_open_reclaim(struct nfs_open_context *ctx, struct nfs4_state *state)
L
Linus Torvalds 已提交
1627
{
1628
	struct nfs_delegation *delegation;
1629
	struct nfs4_opendata *opendata;
1630
	fmode_t delegation_type = 0;
L
Linus Torvalds 已提交
1631 1632
	int status;

1633 1634
	opendata = nfs4_open_recoverdata_alloc(ctx, state,
			NFS4_OPEN_CLAIM_PREVIOUS);
T
Trond Myklebust 已提交
1635 1636
	if (IS_ERR(opendata))
		return PTR_ERR(opendata);
1637 1638
	rcu_read_lock();
	delegation = rcu_dereference(NFS_I(state->inode)->delegation);
1639
	if (delegation != NULL && test_bit(NFS_DELEGATION_NEED_RECLAIM, &delegation->flags) != 0)
1640
		delegation_type = delegation->type;
1641
	rcu_read_unlock();
1642 1643
	opendata->o_arg.u.delegation_type = delegation_type;
	status = nfs4_open_recover(opendata, state);
1644
	nfs4_opendata_put(opendata);
L
Linus Torvalds 已提交
1645 1646 1647
	return status;
}

1648
static int nfs4_do_open_reclaim(struct nfs_open_context *ctx, struct nfs4_state *state)
L
Linus Torvalds 已提交
1649 1650 1651 1652 1653
{
	struct nfs_server *server = NFS_SERVER(state->inode);
	struct nfs4_exception exception = { };
	int err;
	do {
1654
		err = _nfs4_do_open_reclaim(ctx, state);
1655
		trace_nfs4_open_reclaim(ctx, 0, err);
1656 1657
		if (nfs4_clear_cap_atomic_open_v1(server, err, &exception))
			continue;
1658
		if (err != -NFS4ERR_DELAY)
1659 1660
			break;
		nfs4_handle_exception(server, err, &exception);
L
Linus Torvalds 已提交
1661 1662 1663 1664
	} while (exception.retry);
	return err;
}

1665 1666 1667 1668 1669 1670 1671
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))
1672
		return -EAGAIN;
1673
	ret = nfs4_do_open_reclaim(ctx, state);
1674 1675 1676 1677
	put_nfs_open_context(ctx);
	return ret;
}

1678
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
1679
{
1680 1681 1682 1683 1684 1685 1686 1687 1688 1689 1690 1691 1692 1693 1694 1695 1696 1697 1698 1699 1700 1701 1702
	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;
1703 1704 1705
		case -NFS4ERR_MOVED:
			nfs4_schedule_migration_recovery(server);
			return -EAGAIN;
1706 1707 1708
		case -NFS4ERR_LEASE_MOVED:
			nfs4_schedule_lease_moved_recovery(server->nfs_client);
			return -EAGAIN;
1709 1710 1711
		case -NFS4ERR_DELEG_REVOKED:
		case -NFS4ERR_ADMIN_REVOKED:
		case -NFS4ERR_BAD_STATEID:
1712
		case -NFS4ERR_OPENMODE:
1713 1714 1715
			nfs_inode_find_state_and_recover(state->inode,
					stateid);
			nfs4_schedule_stateid_recovery(server, state);
1716
			return -EAGAIN;
1717 1718 1719 1720 1721
		case -NFS4ERR_DELAY:
		case -NFS4ERR_GRACE:
			set_bit(NFS_DELEGATED_STATE, &state->flags);
			ssleep(1);
			return -EAGAIN;
1722 1723 1724 1725
		case -ENOMEM:
		case -NFS4ERR_DENIED:
			/* kill_proc(fl->fl_pid, SIGLOST, 1); */
			return 0;
1726
	}
L
Linus Torvalds 已提交
1727 1728 1729
	return err;
}

1730 1731 1732 1733 1734 1735 1736 1737 1738 1739 1740 1741 1742 1743 1744 1745
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);
}

1746 1747 1748 1749
static void nfs4_open_confirm_prepare(struct rpc_task *task, void *calldata)
{
	struct nfs4_opendata *data = calldata;

1750 1751
	nfs40_setup_sequence(data->o_arg.server->nfs_client->cl_slot_tbl,
			     &data->c_arg.seq_args, &data->c_res.seq_res, task);
1752 1753
}

1754 1755 1756 1757
static void nfs4_open_confirm_done(struct rpc_task *task, void *calldata)
{
	struct nfs4_opendata *data = calldata;

1758
	nfs40_sequence_done(task, &data->c_res.seq_res);
1759

1760
	data->rpc_status = task->tk_status;
1761
	if (data->rpc_status == 0) {
1762
		nfs4_stateid_copy(&data->o_res.stateid, &data->c_res.stateid);
1763
		nfs_confirm_seqid(&data->owner->so_seqid, 0);
1764
		renew_lease(data->o_res.server, data->timestamp);
1765
		data->rpc_done = 1;
1766
	}
1767 1768 1769 1770 1771 1772 1773 1774 1775 1776 1777
}

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! */
1778
	if (!data->rpc_done)
1779 1780
		goto out_free;
	state = nfs4_opendata_to_nfs4_state(data);
1781
	if (!IS_ERR(state))
1782
		nfs4_close_state(state, data->o_arg.fmode);
1783
out_free:
1784
	nfs4_opendata_put(data);
1785 1786 1787
}

static const struct rpc_call_ops nfs4_open_confirm_ops = {
1788
	.rpc_call_prepare = nfs4_open_confirm_prepare,
1789 1790 1791 1792 1793 1794 1795 1796 1797 1798 1799
	.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;
1800 1801 1802 1803 1804 1805
	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
1806 1807
	struct rpc_task_setup task_setup_data = {
		.rpc_client = server->client,
1808
		.rpc_message = &msg,
T
Trond Myklebust 已提交
1809 1810
		.callback_ops = &nfs4_open_confirm_ops,
		.callback_data = data,
1811
		.workqueue = nfsiod_workqueue,
T
Trond Myklebust 已提交
1812 1813
		.flags = RPC_TASK_ASYNC,
	};
L
Linus Torvalds 已提交
1814 1815
	int status;

1816
	nfs4_init_sequence(&data->c_arg.seq_args, &data->c_res.seq_res, 1);
1817
	kref_get(&data->kref);
1818 1819
	data->rpc_done = 0;
	data->rpc_status = 0;
1820
	data->timestamp = jiffies;
T
Trond Myklebust 已提交
1821
	task = rpc_run_task(&task_setup_data);
1822
	if (IS_ERR(task))
1823 1824 1825 1826 1827 1828 1829
		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;
1830
	rpc_put_task(task);
L
Linus Torvalds 已提交
1831 1832 1833
	return status;
}

1834
static void nfs4_open_prepare(struct rpc_task *task, void *calldata)
L
Linus Torvalds 已提交
1835
{
1836 1837
	struct nfs4_opendata *data = calldata;
	struct nfs4_state_owner *sp = data->owner;
1838
	struct nfs_client *clp = sp->so_server->nfs_client;
1839

1840
	if (nfs_wait_on_sequence(data->o_arg.seqid, task) != 0)
1841
		goto out_wait;
1842 1843 1844 1845 1846 1847 1848
	/*
	 * 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;

1849
		if (can_open_cached(data->state, data->o_arg.fmode, data->o_arg.open_flags))
1850
			goto out_no_action;
1851 1852
		rcu_read_lock();
		delegation = rcu_dereference(NFS_I(data->state->inode)->delegation);
1853
		if (data->o_arg.claim != NFS4_OPEN_CLAIM_DELEGATE_CUR &&
1854
		    data->o_arg.claim != NFS4_OPEN_CLAIM_DELEG_CUR_FH &&
1855 1856
		    can_open_delegated(delegation, data->o_arg.fmode))
			goto unlock_no_action;
1857 1858
		rcu_read_unlock();
	}
1859
	/* Update client id. */
1860
	data->o_arg.clientid = clp->cl_clientid;
1861 1862 1863 1864
	switch (data->o_arg.claim) {
	case NFS4_OPEN_CLAIM_PREVIOUS:
	case NFS4_OPEN_CLAIM_DELEG_CUR_FH:
	case NFS4_OPEN_CLAIM_DELEG_PREV_FH:
1865
		data->o_arg.open_bitmap = &nfs4_open_noattr_bitmap[0];
1866 1867
	case NFS4_OPEN_CLAIM_FH:
		task->tk_msg.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_OPEN_NOATTR];
T
Trond Myklebust 已提交
1868 1869
		nfs_copy_fh(&data->o_res.fh, data->o_arg.fh);
	}
1870
	data->timestamp = jiffies;
1871
	if (nfs4_setup_sequence(data->o_arg.server,
1872
				&data->o_arg.seq_args,
1873 1874 1875
				&data->o_res.seq_res,
				task) != 0)
		nfs_release_seqid(data->o_arg.seqid);
1876 1877 1878 1879 1880 1881 1882 1883 1884 1885

	/* 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;
	}
1886
	return;
1887 1888
unlock_no_action:
	rcu_read_unlock();
1889 1890
out_no_action:
	task->tk_action = NULL;
1891
out_wait:
1892
	nfs4_sequence_done(task, &data->o_res.seq_res);
1893 1894
}

1895 1896 1897
static void nfs4_open_done(struct rpc_task *task, void *calldata)
{
	struct nfs4_opendata *data = calldata;
L
Linus Torvalds 已提交
1898

1899
	data->rpc_status = task->tk_status;
1900

1901 1902
	if (!nfs4_sequence_done(task, &data->o_res.seq_res))
		return;
1903

1904
	if (task->tk_status == 0) {
1905 1906
		if (data->o_res.f_attr->valid & NFS_ATTR_FATTR_TYPE) {
			switch (data->o_res.f_attr->mode & S_IFMT) {
1907 1908 1909
			case S_IFREG:
				break;
			case S_IFLNK:
1910
				data->rpc_status = -ELOOP;
1911 1912
				break;
			case S_IFDIR:
1913
				data->rpc_status = -EISDIR;
1914 1915
				break;
			default:
1916
				data->rpc_status = -ENOTDIR;
1917
			}
1918
		}
1919
		renew_lease(data->o_res.server, data->timestamp);
1920 1921
		if (!(data->o_res.rflags & NFS4_OPEN_RESULT_CONFIRM))
			nfs_confirm_seqid(&data->owner->so_seqid, 0);
1922
	}
1923
	data->rpc_done = 1;
1924
}
1925

1926 1927 1928 1929 1930 1931 1932 1933 1934
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! */
1935
	if (data->rpc_status != 0 || !data->rpc_done)
1936 1937 1938 1939 1940
		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);
1941
	if (!IS_ERR(state))
1942
		nfs4_close_state(state, data->o_arg.fmode);
1943
out_free:
1944
	nfs4_opendata_put(data);
1945 1946 1947 1948 1949 1950 1951 1952
}

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,
};

1953
static int nfs4_run_open_task(struct nfs4_opendata *data, int isrecover)
1954 1955 1956 1957 1958 1959
{
	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;
1960 1961 1962 1963 1964 1965
	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
1966 1967
	struct rpc_task_setup task_setup_data = {
		.rpc_client = server->client,
1968
		.rpc_message = &msg,
T
Trond Myklebust 已提交
1969 1970
		.callback_ops = &nfs4_open_ops,
		.callback_data = data,
1971
		.workqueue = nfsiod_workqueue,
T
Trond Myklebust 已提交
1972 1973
		.flags = RPC_TASK_ASYNC,
	};
1974 1975
	int status;

1976
	nfs4_init_sequence(&o_arg->seq_args, &o_res->seq_res, 1);
1977
	kref_get(&data->kref);
1978 1979
	data->rpc_done = 0;
	data->rpc_status = 0;
1980
	data->cancelled = 0;
1981 1982
	data->is_recover = 0;
	if (isrecover) {
1983
		nfs4_set_sequence_privileged(&o_arg->seq_args);
1984 1985
		data->is_recover = 1;
	}
T
Trond Myklebust 已提交
1986
	task = rpc_run_task(&task_setup_data);
1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009
        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;

2010 2011
	nfs_fattr_map_and_free_names(NFS_SERVER(dir), &data->f_attr);

2012 2013 2014 2015 2016 2017 2018 2019 2020
	if (o_res->rflags & NFS4_OPEN_RESULT_CONFIRM) {
		status = _nfs4_proc_open_confirm(data);
		if (status != 0)
			return status;
	}

	return status;
}

2021 2022 2023 2024 2025 2026 2027 2028
/*
 * Additional permission checks in order to distinguish between an
 * open for read, and an open for execute. This works around the
 * fact that NFSv4 OPEN treats read and execute permissions as being
 * the same.
 * Note that in the non-execute case, we want to turn off permission
 * checking if we just created a new file (POSIX open() semantics).
 */
2029 2030
static int nfs4_opendata_access(struct rpc_cred *cred,
				struct nfs4_opendata *opendata,
2031 2032
				struct nfs4_state *state, fmode_t fmode,
				int openflags)
2033 2034 2035 2036 2037 2038 2039 2040 2041 2042
{
	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;
2043 2044 2045 2046
	/*
	 * Use openflags to check for exec, because fmode won't
	 * always have FMODE_EXEC set when file open for exec.
	 */
2047 2048 2049
	if (openflags & __FMODE_EXEC) {
		/* ONLY check for exec rights */
		mask = MAY_EXEC;
2050
	} else if ((fmode & FMODE_READ) && !opendata->file_created)
2051
		mask = MAY_READ;
2052 2053 2054 2055 2056 2057

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

2058
	if ((mask & ~cache.mask & (MAY_READ | MAY_EXEC)) == 0)
2059 2060 2061 2062
		return 0;

	/* even though OPEN succeeded, access is denied. Close the file */
	nfs4_close_state(state, fmode);
2063
	return -EACCES;
2064 2065
}

2066 2067 2068 2069 2070 2071 2072 2073 2074 2075 2076 2077
/*
 * 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);
2078 2079 2080 2081 2082 2083
	if (!data->rpc_done)
		return status;
	if (status != 0) {
		if (status == -NFS4ERR_BADNAME &&
				!(o_arg->open_flags & O_CREAT))
			return -ENOENT;
2084
		return status;
2085
	}
2086

2087 2088
	nfs_fattr_map_and_free_names(server, &data->f_attr);

2089
	if (o_arg->open_flags & O_CREAT) {
2090
		update_changeattr(dir, &o_res->cinfo);
2091 2092 2093 2094 2095
		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
2096 2097
	if ((o_res->rflags & NFS4_OPEN_RESULT_LOCKTYPE_POSIX) == 0)
		server->caps &= ~NFS_CAP_POSIX_LOCK;
L
Linus Torvalds 已提交
2098
	if(o_res->rflags & NFS4_OPEN_RESULT_CONFIRM) {
2099
		status = _nfs4_proc_open_confirm(data);
L
Linus Torvalds 已提交
2100
		if (status != 0)
2101
			return status;
L
Linus Torvalds 已提交
2102 2103
	}
	if (!(o_res->f_attr->valid & NFS_ATTR_FATTR))
2104
		nfs4_proc_getattr(server, &o_res->fh, o_res->f_attr, o_res->f_label);
2105
	return 0;
L
Linus Torvalds 已提交
2106 2107
}

A
Andy Adamson 已提交
2108 2109 2110 2111 2112
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
2113 2114 2115 2116 2117
/*
 * OPEN_EXPIRED:
 * 	reclaim state on the server after a network partition.
 * 	Assumes caller holds the appropriate lock
 */
2118
static int _nfs4_open_expired(struct nfs_open_context *ctx, struct nfs4_state *state)
L
Linus Torvalds 已提交
2119
{
2120
	struct nfs4_opendata *opendata;
2121
	int ret;
L
Linus Torvalds 已提交
2122

2123
	opendata = nfs4_open_recoverdata_alloc(ctx, state,
2124
			NFS4_OPEN_CLAIM_FH);
T
Trond Myklebust 已提交
2125 2126
	if (IS_ERR(opendata))
		return PTR_ERR(opendata);
2127
	ret = nfs4_open_recover(opendata, state);
2128
	if (ret == -ESTALE)
2129
		d_drop(ctx->dentry);
2130
	nfs4_opendata_put(opendata);
2131
	return ret;
L
Linus Torvalds 已提交
2132 2133
}

2134
static int nfs4_do_open_expired(struct nfs_open_context *ctx, struct nfs4_state *state)
2135
{
2136
	struct nfs_server *server = NFS_SERVER(state->inode);
2137 2138 2139 2140
	struct nfs4_exception exception = { };
	int err;

	do {
2141
		err = _nfs4_open_expired(ctx, state);
2142
		trace_nfs4_open_expired(ctx, 0, err);
2143 2144
		if (nfs4_clear_cap_atomic_open_v1(server, err, &exception))
			continue;
2145 2146 2147 2148 2149 2150 2151 2152
		switch (err) {
		default:
			goto out;
		case -NFS4ERR_GRACE:
		case -NFS4ERR_DELAY:
			nfs4_handle_exception(server, err, &exception);
			err = 0;
		}
2153
	} while (exception.retry);
2154
out:
2155 2156 2157
	return err;
}

L
Linus Torvalds 已提交
2158 2159 2160
static int nfs4_open_expired(struct nfs4_state_owner *sp, struct nfs4_state *state)
{
	struct nfs_open_context *ctx;
2161
	int ret;
L
Linus Torvalds 已提交
2162

2163 2164
	ctx = nfs4_state_find_open_context(state);
	if (IS_ERR(ctx))
2165
		return -EAGAIN;
2166
	ret = nfs4_do_open_expired(ctx, state);
2167 2168
	put_nfs_open_context(ctx);
	return ret;
L
Linus Torvalds 已提交
2169 2170
}

2171 2172 2173 2174 2175 2176 2177 2178 2179 2180 2181 2182 2183 2184 2185 2186 2187 2188 2189 2190 2191 2192
static void nfs_finish_clear_delegation_stateid(struct nfs4_state *state)
{
	nfs_remove_bad_delegation(state->inode);
	write_seqlock(&state->seqlock);
	nfs4_stateid_copy(&state->stateid, &state->open_stateid);
	write_sequnlock(&state->seqlock);
	clear_bit(NFS_DELEGATED_STATE, &state->flags);
}

static void nfs40_clear_delegation_stateid(struct nfs4_state *state)
{
	if (rcu_access_pointer(NFS_I(state->inode)->delegation) != NULL)
		nfs_finish_clear_delegation_stateid(state);
}

static int nfs40_open_expired(struct nfs4_state_owner *sp, struct nfs4_state *state)
{
	/* NFSv4.0 doesn't allow for delegation recovery on open expire */
	nfs40_clear_delegation_stateid(state);
	return nfs4_open_expired(sp, state);
}

2193
#if defined(CONFIG_NFS_V4_1)
2194
static void nfs41_check_delegation_stateid(struct nfs4_state *state)
2195 2196
{
	struct nfs_server *server = NFS_SERVER(state->inode);
2197
	nfs4_stateid stateid;
2198
	struct nfs_delegation *delegation;
2199 2200
	struct rpc_cred *cred;
	int status;
2201

2202 2203 2204
	/* Get the delegation credential for use by test/free_stateid */
	rcu_read_lock();
	delegation = rcu_dereference(NFS_I(state->inode)->delegation);
2205
	if (delegation == NULL) {
2206
		rcu_read_unlock();
2207 2208 2209 2210 2211 2212 2213 2214
		return;
	}

	nfs4_stateid_copy(&stateid, &delegation->stateid);
	cred = get_rpccred(delegation->cred);
	rcu_read_unlock();
	status = nfs41_test_stateid(server, &stateid, cred);
	trace_nfs4_test_delegation_stateid(state, NULL, status);
2215

2216 2217 2218 2219
	if (status != NFS_OK) {
		/* Free the stateid unless the server explicitly
		 * informs us the stateid is unrecognized. */
		if (status != -NFS4ERR_BAD_STATEID)
2220 2221
			nfs41_free_stateid(server, &stateid, cred);
		nfs_finish_clear_delegation_stateid(state);
2222
	}
2223

2224
	put_rpccred(cred);
2225 2226 2227 2228 2229 2230 2231 2232 2233 2234 2235 2236 2237
}

/**
 * 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);
2238
	nfs4_stateid *stateid = &state->open_stateid;
2239
	struct rpc_cred *cred = state->owner->so_cred;
2240 2241 2242 2243 2244 2245 2246 2247
	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;

2248
	status = nfs41_test_stateid(server, stateid, cred);
2249
	trace_nfs4_test_open_stateid(state, NULL, status);
2250 2251 2252 2253
	if (status != NFS_OK) {
		/* Free the stateid unless the server explicitly
		 * informs us the stateid is unrecognized. */
		if (status != -NFS4ERR_BAD_STATEID)
2254
			nfs41_free_stateid(server, stateid, cred);
2255 2256 2257 2258

		clear_bit(NFS_O_RDONLY_STATE, &state->flags);
		clear_bit(NFS_O_WRONLY_STATE, &state->flags);
		clear_bit(NFS_O_RDWR_STATE, &state->flags);
2259
		clear_bit(NFS_OPEN_STATE, &state->flags);
2260 2261 2262 2263 2264 2265
	}
	return status;
}

static int nfs41_open_expired(struct nfs4_state_owner *sp, struct nfs4_state *state)
{
2266
	int status;
2267

2268
	nfs41_check_delegation_stateid(state);
2269
	status = nfs41_check_open_stateid(state);
2270 2271 2272
	if (status != NFS_OK)
		status = nfs4_open_expired(sp, state);
	return status;
2273 2274 2275
}
#endif

2276 2277 2278 2279 2280 2281 2282 2283 2284 2285 2286 2287 2288 2289 2290 2291
/*
 * 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;
}

2292 2293 2294
static int _nfs4_open_and_get_state(struct nfs4_opendata *opendata,
		fmode_t fmode,
		int flags,
2295
		struct nfs_open_context *ctx)
2296 2297 2298
{
	struct nfs4_state_owner *sp = opendata->owner;
	struct nfs_server *server = sp->so_server;
2299
	struct dentry *dentry;
2300 2301 2302 2303 2304 2305 2306
	struct nfs4_state *state;
	unsigned int seq;
	int ret;

	seq = raw_seqcount_begin(&sp->so_reclaim_seqcount);

	ret = _nfs4_proc_open(opendata);
2307
	if (ret != 0)
2308 2309 2310 2311 2312 2313 2314 2315 2316
		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);

2317 2318 2319 2320 2321 2322 2323 2324 2325 2326 2327 2328 2329 2330 2331
	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));
	}

2332 2333 2334 2335
	ret = nfs4_opendata_access(sp->so_cred, opendata, state, fmode, flags);
	if (ret != 0)
		goto out;

2336
	ctx->state = state;
2337 2338 2339 2340 2341
	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);
	}
2342 2343 2344 2345
out:
	return ret;
}

L
Linus Torvalds 已提交
2346
/*
2347
 * Returns a referenced nfs4_state
L
Linus Torvalds 已提交
2348
 */
2349
static int _nfs4_do_open(struct inode *dir,
2350
			struct nfs_open_context *ctx,
2351 2352
			int flags,
			struct iattr *sattr,
2353 2354
			struct nfs4_label *label,
			int *opened)
L
Linus Torvalds 已提交
2355 2356 2357 2358
{
	struct nfs4_state_owner  *sp;
	struct nfs4_state     *state = NULL;
	struct nfs_server       *server = NFS_SERVER(dir);
2359
	struct nfs4_opendata *opendata;
2360 2361 2362 2363
	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);
2364
	enum open_claim_type4 claim = NFS4_OPEN_CLAIM_NULL;
2365
	struct nfs4_label *olabel = NULL;
2366
	int status;
L
Linus Torvalds 已提交
2367 2368 2369

	/* Protect against reboot recovery conflicts */
	status = -ENOMEM;
2370 2371
	sp = nfs4_get_state_owner(server, cred, GFP_KERNEL);
	if (sp == NULL) {
L
Linus Torvalds 已提交
2372 2373 2374
		dprintk("nfs4_do_open: nfs4_get_state_owner failed!\n");
		goto out_err;
	}
2375 2376
	status = nfs4_recover_expired_lease(server);
	if (status != 0)
2377
		goto err_put_state_owner;
2378 2379
	if (dentry->d_inode != NULL)
		nfs4_return_incompatible_delegation(dentry->d_inode, fmode);
2380
	status = -ENOMEM;
2381 2382
	if (dentry->d_inode)
		claim = NFS4_OPEN_CLAIM_FH;
2383
	opendata = nfs4_opendata_alloc(dentry, sp, fmode, flags, sattr,
2384
			label, claim, GFP_KERNEL);
2385
	if (opendata == NULL)
T
Trond Myklebust 已提交
2386
		goto err_put_state_owner;
L
Linus Torvalds 已提交
2387

2388 2389 2390 2391 2392 2393 2394 2395
	if (label) {
		olabel = nfs4_label_alloc(server, GFP_KERNEL);
		if (IS_ERR(olabel)) {
			status = PTR_ERR(olabel);
			goto err_opendata_put;
		}
	}

2396 2397 2398 2399 2400 2401
	if (server->attr_bitmask[2] & FATTR4_WORD2_MDSTHRESHOLD) {
		if (!opendata->f_attr.mdsthreshold) {
			opendata->f_attr.mdsthreshold = pnfs_mdsthreshold_alloc();
			if (!opendata->f_attr.mdsthreshold)
				goto err_free_label;
		}
2402
		opendata->o_arg.open_bitmap = &nfs4_pnfs_open_bitmap[0];
2403
	}
2404 2405
	if (dentry->d_inode != NULL)
		opendata->state = nfs4_get_open_state(dentry->d_inode, sp);
2406

2407
	status = _nfs4_open_and_get_state(opendata, fmode, flags, ctx);
2408
	if (status != 0)
2409
		goto err_free_label;
2410
	state = ctx->state;
2411

2412 2413
	if ((opendata->o_arg.open_flags & O_EXCL) &&
	    (opendata->o_arg.createmode != NFS4_CREATE_GUARDED)) {
2414 2415 2416 2417 2418
		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,
2419 2420
				state, label, olabel);
		if (status == 0) {
2421 2422
			nfs_setattr_update_inode(state->inode, sattr,
					opendata->o_res.f_attr);
2423
			nfs_setsecurity(state->inode, opendata->o_res.f_attr, olabel);
2424
		}
2425
	}
2426 2427
	if (opendata->file_created)
		*opened |= FILE_CREATED;
2428

2429
	if (pnfs_use_threshold(ctx_th, opendata->f_attr.mdsthreshold, server)) {
2430
		*ctx_th = opendata->f_attr.mdsthreshold;
2431 2432
		opendata->f_attr.mdsthreshold = NULL;
	}
2433

2434 2435
	nfs4_label_free(olabel);

2436
	nfs4_opendata_put(opendata);
L
Linus Torvalds 已提交
2437 2438
	nfs4_put_state_owner(sp);
	return 0;
2439 2440
err_free_label:
	nfs4_label_free(olabel);
2441 2442
err_opendata_put:
	nfs4_opendata_put(opendata);
2443 2444
err_put_state_owner:
	nfs4_put_state_owner(sp);
L
Linus Torvalds 已提交
2445 2446 2447 2448 2449
out_err:
	return status;
}


2450
static struct nfs4_state *nfs4_do_open(struct inode *dir,
2451
					struct nfs_open_context *ctx,
2452 2453
					int flags,
					struct iattr *sattr,
2454 2455
					struct nfs4_label *label,
					int *opened)
L
Linus Torvalds 已提交
2456
{
2457
	struct nfs_server *server = NFS_SERVER(dir);
L
Linus Torvalds 已提交
2458 2459 2460 2461 2462
	struct nfs4_exception exception = { };
	struct nfs4_state *res;
	int status;

	do {
2463
		status = _nfs4_do_open(dir, ctx, flags, sattr, label, opened);
2464
		res = ctx->state;
2465
		trace_nfs4_open_file(ctx, flags, status);
L
Linus Torvalds 已提交
2466 2467 2468 2469 2470 2471 2472 2473
		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
2474
		 * doing an OPEN, we assume that nfs_increment_open_seqid() will
L
Linus Torvalds 已提交
2475 2476 2477 2478 2479
		 * 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) {
2480
			pr_warn_ratelimited("NFS: v4 server %s "
2481 2482
					" returned a bad sequence-id error!\n",
					NFS_SERVER(dir)->nfs_client->cl_hostname);
L
Linus Torvalds 已提交
2483 2484 2485
			exception.retry = 1;
			continue;
		}
2486 2487 2488 2489 2490 2491 2492 2493 2494 2495
		/*
		 * 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;
		}
2496 2497 2498 2499 2500
		if (status == -EAGAIN) {
			/* We must have found a delegation */
			exception.retry = 1;
			continue;
		}
2501 2502 2503
		if (nfs4_clear_cap_atomic_open_v1(server, status, &exception))
			continue;
		res = ERR_PTR(nfs4_handle_exception(server,
L
Linus Torvalds 已提交
2504 2505 2506 2507 2508
					status, &exception));
	} while (exception.retry);
	return res;
}

2509 2510
static int _nfs4_do_setattr(struct inode *inode, struct rpc_cred *cred,
			    struct nfs_fattr *fattr, struct iattr *sattr,
2511 2512
			    struct nfs4_state *state, struct nfs4_label *ilabel,
			    struct nfs4_label *olabel)
L
Linus Torvalds 已提交
2513
{
2514
	struct nfs_server *server = NFS_SERVER(inode);
L
Linus Torvalds 已提交
2515
        struct nfs_setattrargs  arg = {
2516
                .fh             = NFS_FH(inode),
L
Linus Torvalds 已提交
2517 2518 2519
                .iap            = sattr,
		.server		= server,
		.bitmask = server->attr_bitmask,
2520
		.label		= ilabel,
L
Linus Torvalds 已提交
2521 2522 2523
        };
        struct nfs_setattrres  res = {
		.fattr		= fattr,
2524
		.label		= olabel,
L
Linus Torvalds 已提交
2525 2526 2527
		.server		= server,
        };
        struct rpc_message msg = {
2528 2529 2530 2531
		.rpc_proc	= &nfs4_procedures[NFSPROC4_CLNT_SETATTR],
		.rpc_argp	= &arg,
		.rpc_resp	= &res,
		.rpc_cred	= cred,
L
Linus Torvalds 已提交
2532
        };
2533
	unsigned long timestamp = jiffies;
2534 2535
	fmode_t fmode;
	bool truncate;
2536
	int status;
L
Linus Torvalds 已提交
2537

2538 2539 2540 2541
	arg.bitmask = nfs4_bitmask(server, ilabel);
	if (ilabel)
		arg.bitmask = nfs4_bitmask(server, olabel);

2542
	nfs_fattr_init(fattr);
L
Linus Torvalds 已提交
2543

2544 2545 2546 2547 2548 2549
	/* 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 */
2550
	} else if (truncate && state != NULL) {
2551 2552 2553 2554
		struct nfs_lockowner lockowner = {
			.l_owner = current->files,
			.l_pid = current->tgid,
		};
2555 2556 2557 2558 2559
		if (!nfs4_valid_open_stateid(state))
			return -EBADF;
		if (nfs4_select_rw_stateid(&arg.stateid, state, FMODE_WRITE,
				&lockowner) == -EIO)
			return -EBADF;
2560
	} else
2561
		nfs4_stateid_copy(&arg.stateid, &zero_stateid);
L
Linus Torvalds 已提交
2562

2563
	status = nfs4_call_sync(server->client, server, &msg, &arg.seq_args, &res.seq_res, 1);
2564 2565
	if (status == 0 && state != NULL)
		renew_lease(server, timestamp);
2566
	return status;
L
Linus Torvalds 已提交
2567 2568
}

2569 2570
static int nfs4_do_setattr(struct inode *inode, struct rpc_cred *cred,
			   struct nfs_fattr *fattr, struct iattr *sattr,
2571 2572
			   struct nfs4_state *state, struct nfs4_label *ilabel,
			   struct nfs4_label *olabel)
L
Linus Torvalds 已提交
2573
{
2574
	struct nfs_server *server = NFS_SERVER(inode);
2575 2576
	struct nfs4_exception exception = {
		.state = state,
2577
		.inode = inode,
2578
	};
L
Linus Torvalds 已提交
2579 2580
	int err;
	do {
2581
		err = _nfs4_do_setattr(inode, cred, fattr, sattr, state, ilabel, olabel);
2582
		trace_nfs4_setattr(inode, err);
2583 2584
		switch (err) {
		case -NFS4ERR_OPENMODE:
2585 2586 2587 2588 2589 2590 2591
			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);
			}
2592 2593 2594 2595 2596 2597 2598 2599
			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
2600
	} while (exception.retry);
2601
out:
L
Linus Torvalds 已提交
2602 2603 2604 2605 2606 2607 2608 2609
	return err;
}

struct nfs4_closedata {
	struct inode *inode;
	struct nfs4_state *state;
	struct nfs_closeargs arg;
	struct nfs_closeres res;
2610
	struct nfs_fattr fattr;
2611
	unsigned long timestamp;
F
Fred Isaman 已提交
2612 2613
	bool roc;
	u32 roc_barrier;
L
Linus Torvalds 已提交
2614 2615
};

2616
static void nfs4_free_closedata(void *data)
2617
{
2618 2619
	struct nfs4_closedata *calldata = data;
	struct nfs4_state_owner *sp = calldata->state->owner;
2620
	struct super_block *sb = calldata->state->inode->i_sb;
2621

F
Fred Isaman 已提交
2622 2623
	if (calldata->roc)
		pnfs_roc_release(calldata->state->inode);
2624 2625 2626
	nfs4_put_open_state(calldata->state);
	nfs_free_seqid(calldata->arg.seqid);
	nfs4_put_state_owner(sp);
2627
	nfs_sb_deactive(sb);
2628 2629 2630
	kfree(calldata);
}

2631
static void nfs4_close_done(struct rpc_task *task, void *data)
L
Linus Torvalds 已提交
2632
{
2633
	struct nfs4_closedata *calldata = data;
L
Linus Torvalds 已提交
2634 2635
	struct nfs4_state *state = calldata->state;
	struct nfs_server *server = NFS_SERVER(calldata->inode);
2636
	nfs4_stateid *res_stateid = NULL;
L
Linus Torvalds 已提交
2637

2638
	dprintk("%s: begin!\n", __func__);
2639 2640
	if (!nfs4_sequence_done(task, &calldata->res.seq_res))
		return;
2641
	trace_nfs4_close(state, &calldata->arg, &calldata->res, task->tk_status);
L
Linus Torvalds 已提交
2642 2643 2644 2645 2646
        /* 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:
2647 2648
			res_stateid = &calldata->res.stateid;
			if (calldata->arg.fmode == 0 && calldata->roc)
F
Fred Isaman 已提交
2649 2650
				pnfs_roc_set_barrier(state->inode,
						     calldata->roc_barrier);
2651
			renew_lease(server, calldata->timestamp);
2652
			break;
2653
		case -NFS4ERR_ADMIN_REVOKED:
L
Linus Torvalds 已提交
2654
		case -NFS4ERR_STALE_STATEID:
2655 2656
		case -NFS4ERR_OLD_STATEID:
		case -NFS4ERR_BAD_STATEID:
L
Linus Torvalds 已提交
2657
		case -NFS4ERR_EXPIRED:
2658
			if (!nfs4_stateid_match(&calldata->arg.stateid,
2659
						&state->open_stateid)) {
2660 2661 2662
				rpc_restart_call_prepare(task);
				goto out_release;
			}
2663
			if (calldata->arg.fmode == 0)
2664
				break;
L
Linus Torvalds 已提交
2665
		default:
2666
			if (nfs4_async_handle_error(task, server, state, NULL) == -EAGAIN) {
2667
				rpc_restart_call_prepare(task);
2668 2669
				goto out_release;
			}
L
Linus Torvalds 已提交
2670
	}
2671
	nfs_clear_open_stateid(state, res_stateid, calldata->arg.fmode);
2672
out_release:
2673
	nfs_release_seqid(calldata->arg.seqid);
2674
	nfs_refresh_inode(calldata->inode, calldata->res.fattr);
2675
	dprintk("%s: done, ret = %d!\n", __func__, task->tk_status);
L
Linus Torvalds 已提交
2676 2677
}

T
Trond Myklebust 已提交
2678
static void nfs4_close_prepare(struct rpc_task *task, void *data)
L
Linus Torvalds 已提交
2679
{
T
Trond Myklebust 已提交
2680
	struct nfs4_closedata *calldata = data;
2681
	struct nfs4_state *state = calldata->state;
2682
	struct inode *inode = calldata->inode;
2683
	bool is_rdonly, is_wronly, is_rdwr;
2684
	int call_close = 0;
2685

2686
	dprintk("%s: begin!\n", __func__);
2687
	if (nfs_wait_on_sequence(calldata->arg.seqid, task) != 0)
2688
		goto out_wait;
2689

2690
	task->tk_msg.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_OPEN_DOWNGRADE];
2691
	spin_lock(&state->owner->so_lock);
2692 2693 2694
	is_rdwr = test_bit(NFS_O_RDWR_STATE, &state->flags);
	is_rdonly = test_bit(NFS_O_RDONLY_STATE, &state->flags);
	is_wronly = test_bit(NFS_O_WRONLY_STATE, &state->flags);
2695
	nfs4_stateid_copy(&calldata->arg.stateid, &state->open_stateid);
2696
	/* Calculate the change in open mode */
2697
	calldata->arg.fmode = 0;
2698
	if (state->n_rdwr == 0) {
2699 2700 2701 2702 2703 2704 2705 2706 2707 2708 2709 2710 2711 2712
		if (state->n_rdonly == 0)
			call_close |= is_rdonly;
		else if (is_rdonly)
			calldata->arg.fmode |= FMODE_READ;
		if (state->n_wronly == 0)
			call_close |= is_wronly;
		else if (is_wronly)
			calldata->arg.fmode |= FMODE_WRITE;
	} else if (is_rdwr)
		calldata->arg.fmode |= FMODE_READ|FMODE_WRITE;

	if (calldata->arg.fmode == 0)
		call_close |= is_rdwr;

2713 2714
	if (!nfs4_valid_open_stateid(state))
		call_close = 0;
2715
	spin_unlock(&state->owner->so_lock);
2716 2717

	if (!call_close) {
2718
		/* Note: exit _without_ calling nfs4_close_done */
2719
		goto out_no_action;
2720
	}
2721

F
Fred Isaman 已提交
2722
	if (calldata->arg.fmode == 0) {
2723
		task->tk_msg.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_CLOSE];
F
Fred Isaman 已提交
2724
		if (calldata->roc &&
2725 2726
		    pnfs_roc_drain(inode, &calldata->roc_barrier, task)) {
			nfs_release_seqid(calldata->arg.seqid);
2727
			goto out_wait;
2728
		    }
F
Fred Isaman 已提交
2729
	}
2730 2731 2732
	calldata->arg.share_access =
		nfs4_map_atomic_open_share(NFS_SERVER(inode),
				calldata->arg.fmode, 0);
2733

2734
	nfs_fattr_init(calldata->res.fattr);
2735
	calldata->timestamp = jiffies;
2736
	if (nfs4_setup_sequence(NFS_SERVER(inode),
2737 2738
				&calldata->arg.seq_args,
				&calldata->res.seq_res,
2739 2740
				task) != 0)
		nfs_release_seqid(calldata->arg.seqid);
2741
	dprintk("%s: done!\n", __func__);
2742 2743 2744 2745 2746
	return;
out_no_action:
	task->tk_action = NULL;
out_wait:
	nfs4_sequence_done(task, &calldata->res.seq_res);
L
Linus Torvalds 已提交
2747 2748
}

2749
static const struct rpc_call_ops nfs4_close_ops = {
T
Trond Myklebust 已提交
2750
	.rpc_call_prepare = nfs4_close_prepare,
2751 2752 2753 2754
	.rpc_call_done = nfs4_close_done,
	.rpc_release = nfs4_free_closedata,
};

P
Peng Tao 已提交
2755 2756
static bool nfs4_roc(struct inode *inode)
{
2757
	if (!nfs_have_layout(inode))
P
Peng Tao 已提交
2758 2759 2760 2761
		return false;
	return pnfs_roc(inode);
}

L
Linus Torvalds 已提交
2762 2763 2764 2765 2766 2767 2768 2769 2770 2771 2772
/* 
 * 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!
 */
2773
int nfs4_do_close(struct nfs4_state *state, gfp_t gfp_mask, int wait)
L
Linus Torvalds 已提交
2774
{
2775
	struct nfs_server *server = NFS_SERVER(state->inode);
2776
	struct nfs_seqid *(*alloc_seqid)(struct nfs_seqid_counter *, gfp_t);
L
Linus Torvalds 已提交
2777
	struct nfs4_closedata *calldata;
2778 2779
	struct nfs4_state_owner *sp = state->owner;
	struct rpc_task *task;
2780 2781 2782 2783
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_CLOSE],
		.rpc_cred = state->owner->so_cred,
	};
T
Trond Myklebust 已提交
2784 2785
	struct rpc_task_setup task_setup_data = {
		.rpc_client = server->client,
2786
		.rpc_message = &msg,
T
Trond Myklebust 已提交
2787
		.callback_ops = &nfs4_close_ops,
2788
		.workqueue = nfsiod_workqueue,
T
Trond Myklebust 已提交
2789 2790
		.flags = RPC_TASK_ASYNC,
	};
2791
	int status = -ENOMEM;
L
Linus Torvalds 已提交
2792

2793 2794 2795
	nfs4_state_protect(server->nfs_client, NFS_SP4_MACH_CRED_CLEANUP,
		&task_setup_data.rpc_client, &msg);

2796
	calldata = kzalloc(sizeof(*calldata), gfp_mask);
L
Linus Torvalds 已提交
2797
	if (calldata == NULL)
2798
		goto out;
2799
	nfs4_init_sequence(&calldata->arg.seq_args, &calldata->res.seq_res, 1);
2800
	calldata->inode = state->inode;
L
Linus Torvalds 已提交
2801
	calldata->state = state;
2802
	calldata->arg.fh = NFS_FH(state->inode);
L
Linus Torvalds 已提交
2803
	/* Serialization for the sequence id */
2804 2805
	alloc_seqid = server->nfs_client->cl_mvops->alloc_seqid;
	calldata->arg.seqid = alloc_seqid(&state->owner->so_seqid, gfp_mask);
2806
	if (IS_ERR(calldata->arg.seqid))
2807
		goto out_free_calldata;
2808
	calldata->arg.fmode = 0;
2809
	calldata->arg.bitmask = server->cache_consistency_bitmask;
2810
	calldata->res.fattr = &calldata->fattr;
2811
	calldata->res.seqid = calldata->arg.seqid;
2812
	calldata->res.server = server;
P
Peng Tao 已提交
2813
	calldata->roc = nfs4_roc(state->inode);
2814
	nfs_sb_active(calldata->inode->i_sb);
2815

2816 2817
	msg.rpc_argp = &calldata->arg;
	msg.rpc_resp = &calldata->res;
T
Trond Myklebust 已提交
2818 2819
	task_setup_data.callback_data = calldata;
	task = rpc_run_task(&task_setup_data);
2820 2821
	if (IS_ERR(task))
		return PTR_ERR(task);
2822 2823 2824
	status = 0;
	if (wait)
		status = rpc_wait_for_completion_task(task);
2825
	rpc_put_task(task);
2826
	return status;
2827 2828 2829
out_free_calldata:
	kfree(calldata);
out:
2830 2831
	nfs4_put_open_state(state);
	nfs4_put_state_owner(sp);
2832
	return status;
L
Linus Torvalds 已提交
2833 2834
}

2835
static struct inode *
2836 2837
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
2838 2839
{
	struct nfs4_state *state;
2840 2841 2842
	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
2843

T
Trond Myklebust 已提交
2844
	/* Protect against concurrent sillydeletes */
2845
	state = nfs4_do_open(dir, ctx, open_flags, attr, label, opened);
2846 2847 2848

	nfs4_label_release_security(label);

2849 2850
	if (IS_ERR(state))
		return ERR_CAST(state);
2851
	return state->inode;
L
Linus Torvalds 已提交
2852 2853
}

2854
static void nfs4_close_context(struct nfs_open_context *ctx, int is_sync)
T
Trond Myklebust 已提交
2855 2856 2857 2858
{
	if (ctx->state == NULL)
		return;
	if (is_sync)
2859
		nfs4_close_sync(ctx->state, ctx->mode);
T
Trond Myklebust 已提交
2860
	else
2861
		nfs4_close_state(ctx->state, ctx->mode);
T
Trond Myklebust 已提交
2862
}
L
Linus Torvalds 已提交
2863

2864 2865
#define FATTR4_WORD1_NFS40_MASK (2*FATTR4_WORD1_MOUNTED_ON_FILEID - 1UL)
#define FATTR4_WORD2_NFS41_MASK (2*FATTR4_WORD2_SUPPATTR_EXCLCREAT - 1UL)
2866
#define FATTR4_WORD2_NFS42_MASK (2*FATTR4_WORD2_SECURITY_LABEL - 1UL)
2867

L
Linus Torvalds 已提交
2868 2869
static int _nfs4_server_capabilities(struct nfs_server *server, struct nfs_fh *fhandle)
{
B
Benny Halevy 已提交
2870 2871 2872
	struct nfs4_server_caps_arg args = {
		.fhandle = fhandle,
	};
L
Linus Torvalds 已提交
2873 2874 2875
	struct nfs4_server_caps_res res = {};
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_SERVER_CAPS],
B
Benny Halevy 已提交
2876
		.rpc_argp = &args,
L
Linus Torvalds 已提交
2877 2878 2879 2880
		.rpc_resp = &res,
	};
	int status;

2881
	status = nfs4_call_sync(server->client, server, &msg, &args.seq_args, &res.seq_res, 0);
L
Linus Torvalds 已提交
2882
	if (status == 0) {
2883 2884 2885 2886 2887 2888 2889 2890 2891 2892 2893 2894
		/* 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
2895
		memcpy(server->attr_bitmask, res.attr_bitmask, sizeof(server->attr_bitmask));
2896 2897 2898 2899
		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|
2900 2901
				NFS_CAP_CTIME|NFS_CAP_MTIME|
				NFS_CAP_SECURITY_LABEL);
2902 2903
		if (res.attr_bitmask[0] & FATTR4_WORD0_ACL &&
				res.acl_bitmask & ACL4_SUPPORT_ALLOW_ACL)
L
Linus Torvalds 已提交
2904 2905 2906 2907 2908
			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;
2909 2910 2911 2912 2913 2914 2915 2916 2917 2918 2919 2920 2921 2922 2923 2924
		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;
2925 2926 2927 2928 2929 2930
#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));
2931
		server->attr_bitmask_nl[2] &= ~FATTR4_WORD2_SECURITY_LABEL;
2932

2933 2934 2935
		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;
2936
		server->cache_consistency_bitmask[2] = 0;
L
Linus Torvalds 已提交
2937
		server->acl_bitmask = res.acl_bitmask;
2938
		server->fh_expire_type = res.fh_expire_type;
L
Linus Torvalds 已提交
2939
	}
A
Andy Adamson 已提交
2940

L
Linus Torvalds 已提交
2941 2942 2943
	return status;
}

2944
int nfs4_server_capabilities(struct nfs_server *server, struct nfs_fh *fhandle)
L
Linus Torvalds 已提交
2945 2946 2947 2948 2949 2950 2951 2952 2953 2954 2955 2956 2957 2958
{
	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)
{
2959
	u32 bitmask[3];
L
Linus Torvalds 已提交
2960
	struct nfs4_lookup_root_arg args = {
2961
		.bitmask = bitmask,
L
Linus Torvalds 已提交
2962 2963 2964
	};
	struct nfs4_lookup_res res = {
		.server = server,
2965
		.fattr = info->fattr,
L
Linus Torvalds 已提交
2966 2967 2968 2969 2970 2971 2972
		.fh = fhandle,
	};
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_LOOKUP_ROOT],
		.rpc_argp = &args,
		.rpc_resp = &res,
	};
2973

2974 2975 2976 2977 2978 2979 2980
	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;

2981
	nfs_fattr_init(info->fattr);
2982
	return nfs4_call_sync(server->client, server, &msg, &args.seq_args, &res.seq_res, 0);
L
Linus Torvalds 已提交
2983 2984 2985 2986 2987 2988 2989 2990
}

static int nfs4_lookup_root(struct nfs_server *server, struct nfs_fh *fhandle,
		struct nfs_fsinfo *info)
{
	struct nfs4_exception exception = { };
	int err;
	do {
2991
		err = _nfs4_lookup_root(server, fhandle, info);
2992
		trace_nfs4_lookup_root(server, fhandle, info->fattr, err);
2993 2994 2995
		switch (err) {
		case 0:
		case -NFS4ERR_WRONGSEC:
2996
			goto out;
2997 2998 2999
		default:
			err = nfs4_handle_exception(server, err, &exception);
		}
L
Linus Torvalds 已提交
3000
	} while (exception.retry);
3001
out:
L
Linus Torvalds 已提交
3002 3003 3004
	return err;
}

3005 3006 3007
static int nfs4_lookup_root_sec(struct nfs_server *server, struct nfs_fh *fhandle,
				struct nfs_fsinfo *info, rpc_authflavor_t flavor)
{
3008 3009 3010
	struct rpc_auth_create_args auth_args = {
		.pseudoflavor = flavor,
	};
3011 3012 3013
	struct rpc_auth *auth;
	int ret;

3014
	auth = rpcauth_create(&auth_args, server->client);
3015
	if (IS_ERR(auth)) {
3016
		ret = -EACCES;
3017 3018 3019 3020 3021 3022 3023
		goto out;
	}
	ret = nfs4_lookup_root(server, fhandle, info);
out:
	return ret;
}

3024 3025 3026 3027 3028 3029 3030 3031 3032
/*
 * 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.
 */
3033
static int nfs4_find_root_sec(struct nfs_server *server, struct nfs_fh *fhandle,
3034
			      struct nfs_fsinfo *info)
L
Linus Torvalds 已提交
3035
{
3036 3037 3038 3039 3040
	/* Per 3530bis 15.33.5 */
	static const rpc_authflavor_t flav_array[] = {
		RPC_AUTH_GSS_KRB5P,
		RPC_AUTH_GSS_KRB5I,
		RPC_AUTH_GSS_KRB5,
3041
		RPC_AUTH_UNIX,			/* courtesy */
3042 3043 3044 3045
		RPC_AUTH_NULL,
	};
	int status = -EPERM;
	size_t i;
C
Chuck Lever 已提交
3046

3047 3048 3049 3050 3051 3052 3053 3054 3055 3056 3057 3058 3059 3060 3061 3062 3063 3064
	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;
		}
3065
	}
3066

3067 3068 3069 3070 3071 3072 3073 3074 3075
	/*
	 * -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;
3076 3077 3078
	return status;
}

C
Chuck Lever 已提交
3079 3080 3081 3082 3083 3084 3085 3086 3087 3088 3089 3090
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
3091
 * @auth_probe: probe the auth flavours
C
Chuck Lever 已提交
3092 3093
 *
 * Returns zero on success, or a negative errno.
3094
 */
3095
int nfs4_proc_get_rootfh(struct nfs_server *server, struct nfs_fh *fhandle,
3096 3097
			 struct nfs_fsinfo *info,
			 bool auth_probe)
3098
{
C
Chuck Lever 已提交
3099 3100
	int status;

3101 3102 3103 3104 3105 3106
	switch (auth_probe) {
	case false:
		status = nfs4_lookup_root(server, fhandle, info);
		if (status != -NFS4ERR_WRONGSEC)
			break;
	default:
C
Chuck Lever 已提交
3107
		status = nfs4_do_find_root_sec(server, fhandle, info);
3108
	}
C
Chuck Lever 已提交
3109

L
Linus Torvalds 已提交
3110 3111 3112 3113
	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
3114

3115
	return nfs4_map_errors(status);
L
Linus Torvalds 已提交
3116 3117
}

3118 3119 3120 3121 3122
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;
3123
	struct nfs4_label *label = NULL;
3124 3125 3126 3127 3128 3129 3130

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

3131 3132 3133 3134
	label = nfs4_label_alloc(server, GFP_KERNEL);
	if (IS_ERR(label))
		return PTR_ERR(label);

3135
	error = nfs4_proc_getattr(server, mntfh, fattr, label);
3136 3137
	if (error < 0) {
		dprintk("nfs4_get_root: getattr error = %d\n", -error);
3138
		goto err_free_label;
3139 3140 3141 3142 3143 3144
	}

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

3145 3146 3147
err_free_label:
	nfs4_label_free(label);

3148 3149 3150
	return error;
}

M
Manoj Naik 已提交
3151 3152 3153 3154 3155
/*
 * 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
 */
3156 3157 3158
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
3159 3160 3161 3162 3163 3164 3165 3166 3167 3168 3169 3170
{
	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;

3171
	status = nfs4_proc_fs_locations(client, dir, name, locations, page);
M
Manoj Naik 已提交
3172 3173
	if (status != 0)
		goto out;
3174 3175 3176 3177 3178 3179

	/*
	 * 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
3180
	if (nfs_fsid_equal(&NFS_SERVER(dir)->fsid, &locations->fattr.fsid)) {
3181 3182
		dprintk("%s: server did not return a different fsid for"
			" a referral at %s\n", __func__, name->name);
3183
		status = -NFS4ERR_MOVED;
M
Manoj Naik 已提交
3184 3185
		goto out;
	}
3186 3187
	/* Fixup attributes for the nfs_lookup() call to nfs_fhget() */
	nfs_fixup_referral_attributes(&locations->fattr);
M
Manoj Naik 已提交
3188

3189
	/* replace the lookup nfs_fattr with the locations nfs_fattr */
M
Manoj Naik 已提交
3190 3191 3192 3193 3194
	memcpy(fattr, &locations->fattr, sizeof(struct nfs_fattr));
	memset(fhandle, 0, sizeof(struct nfs_fh));
out:
	if (page)
		__free_page(page);
3195
	kfree(locations);
M
Manoj Naik 已提交
3196 3197 3198
	return status;
}

3199 3200
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
3201 3202 3203 3204 3205 3206 3207
{
	struct nfs4_getattr_arg args = {
		.fh = fhandle,
		.bitmask = server->attr_bitmask,
	};
	struct nfs4_getattr_res res = {
		.fattr = fattr,
3208
		.label = label,
L
Linus Torvalds 已提交
3209 3210 3211 3212 3213 3214 3215
		.server = server,
	};
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_GETATTR],
		.rpc_argp = &args,
		.rpc_resp = &res,
	};
3216 3217 3218

	args.bitmask = nfs4_bitmask(server, label);

3219
	nfs_fattr_init(fattr);
3220
	return nfs4_call_sync(server->client, server, &msg, &args.seq_args, &res.seq_res, 0);
L
Linus Torvalds 已提交
3221 3222
}

3223 3224
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
3225 3226 3227 3228
{
	struct nfs4_exception exception = { };
	int err;
	do {
3229 3230 3231
		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
3232 3233 3234 3235 3236 3237 3238 3239 3240 3241 3242 3243 3244 3245 3246 3247 3248 3249 3250 3251 3252 3253 3254 3255 3256 3257
				&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)
{
3258
	struct inode *inode = dentry->d_inode;
3259
	struct rpc_cred *cred = NULL;
3260
	struct nfs4_state *state = NULL;
3261
	struct nfs4_label *label = NULL;
L
Linus Torvalds 已提交
3262 3263
	int status;

3264 3265 3266
	if (pnfs_ld_layoutret_on_setattr(inode) &&
	    sattr->ia_valid & ATTR_SIZE &&
	    sattr->ia_size < i_size_read(inode))
3267
		pnfs_commit_and_return_layout(inode);
B
Benny Halevy 已提交
3268

3269
	nfs_fattr_init(fattr);
L
Linus Torvalds 已提交
3270
	
3271 3272
	/* Deal with open(O_TRUNC) */
	if (sattr->ia_valid & ATTR_OPEN)
3273
		sattr->ia_valid &= ~(ATTR_MTIME|ATTR_CTIME);
3274 3275

	/* Optimization: if the end result is no change, don't RPC */
3276
	if ((sattr->ia_valid & ~(ATTR_FILE|ATTR_OPEN)) == 0)
3277 3278
		return 0;

3279
	/* Search for an existing open(O_WRITE) file */
3280 3281 3282 3283
	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
3284 3285 3286 3287
		if (ctx) {
			cred = ctx->cred;
			state = ctx->state;
		}
3288
	}
3289

3290 3291 3292 3293 3294
	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);
3295
	if (status == 0) {
3296
		nfs_setattr_update_inode(inode, sattr, fattr);
3297 3298
		nfs_setsecurity(inode, fattr, label);
	}
3299
	nfs4_label_free(label);
L
Linus Torvalds 已提交
3300 3301 3302
	return status;
}

3303 3304
static int _nfs4_proc_lookup(struct rpc_clnt *clnt, struct inode *dir,
		const struct qstr *name, struct nfs_fh *fhandle,
3305
		struct nfs_fattr *fattr, struct nfs4_label *label)
D
David Howells 已提交
3306
{
3307
	struct nfs_server *server = NFS_SERVER(dir);
D
David Howells 已提交
3308 3309 3310
	int		       status;
	struct nfs4_lookup_arg args = {
		.bitmask = server->attr_bitmask,
3311
		.dir_fh = NFS_FH(dir),
D
David Howells 已提交
3312 3313 3314 3315 3316
		.name = name,
	};
	struct nfs4_lookup_res res = {
		.server = server,
		.fattr = fattr,
3317
		.label = label,
D
David Howells 已提交
3318 3319 3320 3321 3322 3323 3324 3325
		.fh = fhandle,
	};
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_LOOKUP],
		.rpc_argp = &args,
		.rpc_resp = &res,
	};

3326 3327
	args.bitmask = nfs4_bitmask(server, label);

D
David Howells 已提交
3328 3329
	nfs_fattr_init(fattr);

L
Linus Torvalds 已提交
3330
	dprintk("NFS call  lookup %s\n", name->name);
3331
	status = nfs4_call_sync(clnt, server, &msg, &args.seq_args, &res.seq_res, 0);
L
Linus Torvalds 已提交
3332 3333 3334 3335
	dprintk("NFS reply lookup: %d\n", status);
	return status;
}

3336
static void nfs_fixup_secinfo_attributes(struct nfs_fattr *fattr)
3337 3338
{
	fattr->valid |= NFS_ATTR_FATTR_TYPE | NFS_ATTR_FATTR_MODE |
3339
		NFS_ATTR_FATTR_NLINK | NFS_ATTR_FATTR_MOUNTPOINT;
3340 3341 3342 3343
	fattr->mode = S_IFDIR | S_IRUGO | S_IXUGO;
	fattr->nlink = 2;
}

3344 3345
static int nfs4_proc_lookup_common(struct rpc_clnt **clnt, struct inode *dir,
				   struct qstr *name, struct nfs_fh *fhandle,
3346
				   struct nfs_fattr *fattr, struct nfs4_label *label)
L
Linus Torvalds 已提交
3347 3348
{
	struct nfs4_exception exception = { };
3349
	struct rpc_clnt *client = *clnt;
L
Linus Torvalds 已提交
3350 3351
	int err;
	do {
3352
		err = _nfs4_proc_lookup(client, dir, name, fhandle, fattr, label);
3353
		trace_nfs4_lookup(dir, name, err);
3354
		switch (err) {
3355
		case -NFS4ERR_BADNAME:
3356 3357
			err = -ENOENT;
			goto out;
3358
		case -NFS4ERR_MOVED:
3359
			err = nfs4_get_referral(client, dir, name, fattr, fhandle);
3360
			goto out;
3361
		case -NFS4ERR_WRONGSEC:
3362 3363 3364
			err = -EPERM;
			if (client != *clnt)
				goto out;
3365
			client = nfs4_negotiate_security(client, dir, name);
3366 3367 3368 3369 3370 3371 3372
			if (IS_ERR(client))
				return PTR_ERR(client);

			exception.retry = 1;
			break;
		default:
			err = nfs4_handle_exception(NFS_SERVER(dir), err, &exception);
3373
		}
L
Linus Torvalds 已提交
3374
	} while (exception.retry);
3375 3376 3377 3378 3379 3380 3381

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

L
Linus Torvalds 已提交
3382 3383 3384
	return err;
}

3385
static int nfs4_proc_lookup(struct inode *dir, struct qstr *name,
3386 3387
			    struct nfs_fh *fhandle, struct nfs_fattr *fattr,
			    struct nfs4_label *label)
3388 3389 3390 3391
{
	int status;
	struct rpc_clnt *client = NFS_CLIENT(dir);

3392
	status = nfs4_proc_lookup_common(&client, dir, name, fhandle, fattr, label);
3393 3394 3395 3396 3397 3398 3399
	if (client != NFS_CLIENT(dir)) {
		rpc_shutdown_client(client);
		nfs_fixup_secinfo_attributes(fattr);
	}
	return status;
}

3400 3401 3402 3403
struct rpc_clnt *
nfs4_proc_lookup_mountpoint(struct inode *dir, struct qstr *name,
			    struct nfs_fh *fhandle, struct nfs_fattr *fattr)
{
3404
	struct rpc_clnt *client = NFS_CLIENT(dir);
3405 3406
	int status;

3407
	status = nfs4_proc_lookup_common(&client, dir, name, fhandle, fattr, NULL);
3408
	if (status < 0)
3409
		return ERR_PTR(status);
3410
	return (client == NFS_CLIENT(dir)) ? rpc_clone_client(client) : client;
3411 3412
}

L
Linus Torvalds 已提交
3413 3414
static int _nfs4_proc_access(struct inode *inode, struct nfs_access_entry *entry)
{
3415
	struct nfs_server *server = NFS_SERVER(inode);
L
Linus Torvalds 已提交
3416 3417
	struct nfs4_accessargs args = {
		.fh = NFS_FH(inode),
3418
		.bitmask = server->cache_consistency_bitmask,
3419 3420 3421
	};
	struct nfs4_accessres res = {
		.server = server,
L
Linus Torvalds 已提交
3422 3423 3424 3425 3426 3427 3428 3429
	};
	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;
3430
	int status = 0;
L
Linus Torvalds 已提交
3431 3432 3433 3434 3435 3436 3437 3438 3439 3440 3441 3442 3443 3444 3445 3446 3447

	/*
	 * 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;
	}
3448 3449 3450 3451 3452

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

3453
	status = nfs4_call_sync(server->client, server, &msg, &args.seq_args, &res.seq_res, 0);
L
Linus Torvalds 已提交
3454
	if (!status) {
3455
		nfs_access_set_mask(entry, res.access);
3456
		nfs_refresh_inode(inode, res.fattr);
L
Linus Torvalds 已提交
3457
	}
3458
	nfs_free_fattr(res.fattr);
L
Linus Torvalds 已提交
3459 3460 3461 3462 3463 3464 3465 3466
	return status;
}

static int nfs4_proc_access(struct inode *inode, struct nfs_access_entry *entry)
{
	struct nfs4_exception exception = { };
	int err;
	do {
3467 3468 3469
		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
3470 3471 3472 3473 3474 3475 3476 3477 3478 3479 3480 3481 3482 3483 3484 3485 3486 3487 3488 3489 3490 3491
				&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
3492
 * we get the post-operation mtime and size.
L
Linus Torvalds 已提交
3493 3494 3495 3496 3497 3498 3499 3500 3501 3502 3503 3504 3505
 *
 * 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
3506
	struct nfs4_readlink_res res;
L
Linus Torvalds 已提交
3507 3508 3509
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_READLINK],
		.rpc_argp = &args,
B
Benny Halevy 已提交
3510
		.rpc_resp = &res,
L
Linus Torvalds 已提交
3511 3512
	};

3513
	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
3514 3515 3516 3517 3518 3519 3520 3521
}

static int nfs4_proc_readlink(struct inode *inode, struct page *page,
		unsigned int pgbase, unsigned int pglen)
{
	struct nfs4_exception exception = { };
	int err;
	do {
3522 3523 3524
		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
3525 3526 3527 3528 3529 3530
				&exception);
	} while (exception.retry);
	return err;
}

/*
3531
 * This is just for mknod.  open(O_CREAT) will always do ->open_context().
L
Linus Torvalds 已提交
3532 3533 3534
 */
static int
nfs4_proc_create(struct inode *dir, struct dentry *dentry, struct iattr *sattr,
3535
		 int flags)
L
Linus Torvalds 已提交
3536
{
3537
	struct nfs4_label l, *ilabel = NULL;
3538
	struct nfs_open_context *ctx;
L
Linus Torvalds 已提交
3539
	struct nfs4_state *state;
3540
	int opened = 0;
L
Linus Torvalds 已提交
3541 3542
	int status = 0;

3543 3544 3545 3546
	ctx = alloc_nfs_open_context(dentry, FMODE_READ);
	if (IS_ERR(ctx))
		return PTR_ERR(ctx);

3547 3548
	ilabel = nfs4_label_init_security(dir, dentry, sattr, &l);

A
Aneesh Kumar K.V 已提交
3549
	sattr->ia_mode &= ~current_umask();
3550
	state = nfs4_do_open(dir, ctx, flags, sattr, ilabel, &opened);
L
Linus Torvalds 已提交
3551 3552
	if (IS_ERR(state)) {
		status = PTR_ERR(state);
3553
		goto out;
L
Linus Torvalds 已提交
3554 3555
	}
out:
3556
	nfs4_label_release_security(ilabel);
3557
	put_nfs_open_context(ctx);
L
Linus Torvalds 已提交
3558 3559 3560 3561 3562
	return status;
}

static int _nfs4_proc_remove(struct inode *dir, struct qstr *name)
{
3563
	struct nfs_server *server = NFS_SERVER(dir);
3564
	struct nfs_removeargs args = {
L
Linus Torvalds 已提交
3565
		.fh = NFS_FH(dir),
3566
		.name = *name,
3567
	};
3568
	struct nfs_removeres res = {
3569
		.server = server,
L
Linus Torvalds 已提交
3570 3571
	};
	struct rpc_message msg = {
3572 3573 3574
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_REMOVE],
		.rpc_argp = &args,
		.rpc_resp = &res,
L
Linus Torvalds 已提交
3575
	};
3576
	int status;
L
Linus Torvalds 已提交
3577

3578
	status = nfs4_call_sync(server->client, server, &msg, &args.seq_args, &res.seq_res, 1);
3579
	if (status == 0)
3580
		update_changeattr(dir, &res.cinfo);
L
Linus Torvalds 已提交
3581 3582 3583 3584 3585 3586 3587 3588
	return status;
}

static int nfs4_proc_remove(struct inode *dir, struct qstr *name)
{
	struct nfs4_exception exception = { };
	int err;
	do {
3589 3590 3591
		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
3592 3593 3594 3595 3596
				&exception);
	} while (exception.retry);
	return err;
}

3597
static void nfs4_proc_unlink_setup(struct rpc_message *msg, struct inode *dir)
L
Linus Torvalds 已提交
3598
{
3599 3600 3601
	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
3602

3603
	res->server = server;
L
Linus Torvalds 已提交
3604
	msg->rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_REMOVE];
3605
	nfs4_init_sequence(&args->seq_args, &res->seq_res, 1);
3606 3607

	nfs_fattr_init(res->dir_attr);
L
Linus Torvalds 已提交
3608 3609
}

3610 3611
static void nfs4_proc_unlink_rpc_prepare(struct rpc_task *task, struct nfs_unlinkdata *data)
{
3612 3613 3614 3615
	nfs4_setup_sequence(NFS_SERVER(data->dir),
			&data->args.seq_args,
			&data->res.seq_res,
			task);
L
Linus Torvalds 已提交
3616 3617
}

3618
static int nfs4_proc_unlink_done(struct rpc_task *task, struct inode *dir)
L
Linus Torvalds 已提交
3619
{
3620 3621
	struct nfs_unlinkdata *data = task->tk_calldata;
	struct nfs_removeres *res = &data->res;
3622

3623 3624
	if (!nfs4_sequence_done(task, &res->seq_res))
		return 0;
3625 3626
	if (nfs4_async_handle_error(task, res->server, NULL,
				    &data->timeout) == -EAGAIN)
3627 3628 3629
		return 0;
	update_changeattr(dir, &res->cinfo);
	return 1;
L
Linus Torvalds 已提交
3630 3631
}

3632 3633 3634 3635 3636 3637 3638 3639
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;
3640
	nfs4_init_sequence(&arg->seq_args, &res->seq_res, 1);
3641 3642
}

3643 3644
static void nfs4_proc_rename_rpc_prepare(struct rpc_task *task, struct nfs_renamedata *data)
{
3645 3646 3647 3648
	nfs4_setup_sequence(NFS_SERVER(data->old_dir),
			&data->args.seq_args,
			&data->res.seq_res,
			task);
3649 3650 3651 3652 3653
}

static int nfs4_proc_rename_done(struct rpc_task *task, struct inode *old_dir,
				 struct inode *new_dir)
{
3654 3655
	struct nfs_renamedata *data = task->tk_calldata;
	struct nfs_renameres *res = &data->res;
3656 3657 3658

	if (!nfs4_sequence_done(task, &res->seq_res))
		return 0;
3659
	if (nfs4_async_handle_error(task, res->server, NULL, &data->timeout) == -EAGAIN)
3660 3661 3662 3663 3664 3665 3666
		return 0;

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

L
Linus Torvalds 已提交
3667 3668
static int _nfs4_proc_link(struct inode *inode, struct inode *dir, struct qstr *name)
{
3669
	struct nfs_server *server = NFS_SERVER(inode);
L
Linus Torvalds 已提交
3670 3671 3672 3673
	struct nfs4_link_arg arg = {
		.fh     = NFS_FH(inode),
		.dir_fh = NFS_FH(dir),
		.name   = name,
3674 3675 3676 3677
		.bitmask = server->attr_bitmask,
	};
	struct nfs4_link_res res = {
		.server = server,
3678
		.label = NULL,
L
Linus Torvalds 已提交
3679 3680 3681 3682
	};
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_LINK],
		.rpc_argp = &arg,
3683
		.rpc_resp = &res,
L
Linus Torvalds 已提交
3684
	};
3685 3686 3687
	int status = -ENOMEM;

	res.fattr = nfs_alloc_fattr();
3688
	if (res.fattr == NULL)
3689
		goto out;
L
Linus Torvalds 已提交
3690

3691 3692 3693 3694 3695
	res.label = nfs4_label_alloc(server, GFP_KERNEL);
	if (IS_ERR(res.label)) {
		status = PTR_ERR(res.label);
		goto out;
	}
3696
	arg.bitmask = nfs4_bitmask(server, res.label);
3697

3698
	status = nfs4_call_sync(server->client, server, &msg, &arg.seq_args, &res.seq_res, 1);
3699 3700
	if (!status) {
		update_changeattr(dir, &res.cinfo);
3701 3702 3703
		status = nfs_post_op_update_inode(inode, res.fattr);
		if (!status)
			nfs_setsecurity(inode, res.fattr, res.label);
3704
	}
3705 3706 3707 3708


	nfs4_label_free(res.label);

3709 3710
out:
	nfs_free_fattr(res.fattr);
L
Linus Torvalds 已提交
3711 3712 3713 3714 3715 3716 3717 3718 3719 3720 3721 3722 3723 3724 3725
	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;
}

3726 3727 3728 3729 3730 3731
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;
3732
	struct nfs4_label *label;
3733 3734 3735 3736 3737 3738 3739 3740 3741 3742 3743
};

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);

3744 3745 3746 3747
		data->label = nfs4_label_alloc(server, GFP_KERNEL);
		if (IS_ERR(data->label))
			goto out_free;

3748 3749 3750 3751 3752 3753 3754 3755
		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;
3756
		data->arg.bitmask = nfs4_bitmask(server, data->label);
3757 3758 3759
		data->res.server = server;
		data->res.fh = &data->fh;
		data->res.fattr = &data->fattr;
3760
		data->res.label = data->label;
3761 3762 3763
		nfs_fattr_init(data->res.fattr);
	}
	return data;
3764 3765 3766
out_free:
	kfree(data);
	return NULL;
3767 3768 3769 3770
}

static int nfs4_do_create(struct inode *dir, struct dentry *dentry, struct nfs4_createdata *data)
{
3771
	int status = nfs4_call_sync(NFS_SERVER(dir)->client, NFS_SERVER(dir), &data->msg,
3772
				    &data->arg.seq_args, &data->res.seq_res, 1);
3773 3774
	if (status == 0) {
		update_changeattr(dir, &data->res.dir_cinfo);
3775
		status = nfs_instantiate(dentry, data->res.fh, data->res.fattr, data->res.label);
3776 3777 3778 3779 3780 3781
	}
	return status;
}

static void nfs4_free_createdata(struct nfs4_createdata *data)
{
3782
	nfs4_label_free(data->label);
3783 3784 3785
	kfree(data);
}

3786
static int _nfs4_proc_symlink(struct inode *dir, struct dentry *dentry,
3787 3788
		struct page *page, unsigned int len, struct iattr *sattr,
		struct nfs4_label *label)
L
Linus Torvalds 已提交
3789
{
3790 3791
	struct nfs4_createdata *data;
	int status = -ENAMETOOLONG;
L
Linus Torvalds 已提交
3792

3793
	if (len > NFS4_MAXPATHLEN)
3794
		goto out;
3795

3796 3797 3798 3799 3800 3801 3802 3803
	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;
3804
	data->arg.label = label;
L
Linus Torvalds 已提交
3805
	
3806 3807 3808 3809
	status = nfs4_do_create(dir, dentry, data);

	nfs4_free_createdata(data);
out:
L
Linus Torvalds 已提交
3810 3811 3812
	return status;
}

3813
static int nfs4_proc_symlink(struct inode *dir, struct dentry *dentry,
3814
		struct page *page, unsigned int len, struct iattr *sattr)
L
Linus Torvalds 已提交
3815 3816
{
	struct nfs4_exception exception = { };
3817
	struct nfs4_label l, *label = NULL;
L
Linus Torvalds 已提交
3818
	int err;
3819 3820 3821

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

L
Linus Torvalds 已提交
3822
	do {
3823 3824 3825
		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
3826 3827
				&exception);
	} while (exception.retry);
3828 3829

	nfs4_label_release_security(label);
L
Linus Torvalds 已提交
3830 3831 3832 3833
	return err;
}

static int _nfs4_proc_mkdir(struct inode *dir, struct dentry *dentry,
3834
		struct iattr *sattr, struct nfs4_label *label)
L
Linus Torvalds 已提交
3835
{
3836 3837
	struct nfs4_createdata *data;
	int status = -ENOMEM;
L
Linus Torvalds 已提交
3838

3839 3840 3841 3842
	data = nfs4_alloc_createdata(dir, &dentry->d_name, sattr, NF4DIR);
	if (data == NULL)
		goto out;

3843
	data->arg.label = label;
3844 3845 3846 3847
	status = nfs4_do_create(dir, dentry, data);

	nfs4_free_createdata(data);
out:
L
Linus Torvalds 已提交
3848 3849 3850 3851 3852 3853 3854
	return status;
}

static int nfs4_proc_mkdir(struct inode *dir, struct dentry *dentry,
		struct iattr *sattr)
{
	struct nfs4_exception exception = { };
3855
	struct nfs4_label l, *label = NULL;
L
Linus Torvalds 已提交
3856
	int err;
A
Aneesh Kumar K.V 已提交
3857

3858 3859
	label = nfs4_label_init_security(dir, dentry, sattr, &l);

A
Aneesh Kumar K.V 已提交
3860
	sattr->ia_mode &= ~current_umask();
L
Linus Torvalds 已提交
3861
	do {
3862 3863 3864
		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
3865 3866
				&exception);
	} while (exception.retry);
3867 3868
	nfs4_label_release_security(label);

L
Linus Torvalds 已提交
3869 3870 3871 3872
	return err;
}

static int _nfs4_proc_readdir(struct dentry *dentry, struct rpc_cred *cred,
B
Bryan Schumaker 已提交
3873
		u64 cookie, struct page **pages, unsigned int count, int plus)
L
Linus Torvalds 已提交
3874 3875 3876 3877
{
	struct inode		*dir = dentry->d_inode;
	struct nfs4_readdir_arg args = {
		.fh = NFS_FH(dir),
B
Bryan Schumaker 已提交
3878
		.pages = pages,
L
Linus Torvalds 已提交
3879 3880
		.pgbase = 0,
		.count = count,
3881
		.bitmask = NFS_SERVER(dentry->d_inode)->attr_bitmask,
B
Bryan Schumaker 已提交
3882
		.plus = plus,
L
Linus Torvalds 已提交
3883 3884 3885 3886 3887 3888 3889 3890 3891 3892
	};
	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;

3893 3894
	dprintk("%s: dentry = %pd2, cookie = %Lu\n", __func__,
			dentry,
3895
			(unsigned long long)cookie);
3896
	nfs4_setup_readdir(cookie, NFS_I(dir)->cookieverf, dentry, &args);
L
Linus Torvalds 已提交
3897
	res.pgbase = args.pgbase;
3898
	status = nfs4_call_sync(NFS_SERVER(dir)->client, NFS_SERVER(dir), &msg, &args.seq_args, &res.seq_res, 0);
3899
	if (status >= 0) {
3900
		memcpy(NFS_I(dir)->cookieverf, res.verifier.data, NFS4_VERIFIER_SIZE);
3901 3902
		status += args.pgbase;
	}
3903 3904 3905

	nfs_invalidate_atime(dir);

3906
	dprintk("%s: returns %d\n", __func__, status);
L
Linus Torvalds 已提交
3907 3908 3909 3910
	return status;
}

static int nfs4_proc_readdir(struct dentry *dentry, struct rpc_cred *cred,
B
Bryan Schumaker 已提交
3911
		u64 cookie, struct page **pages, unsigned int count, int plus)
L
Linus Torvalds 已提交
3912 3913 3914 3915
{
	struct nfs4_exception exception = { };
	int err;
	do {
3916 3917 3918 3919
		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
3920 3921 3922 3923 3924 3925
				&exception);
	} while (exception.retry);
	return err;
}

static int _nfs4_proc_mknod(struct inode *dir, struct dentry *dentry,
3926
		struct iattr *sattr, struct nfs4_label *label, dev_t rdev)
L
Linus Torvalds 已提交
3927
{
3928 3929 3930
	struct nfs4_createdata *data;
	int mode = sattr->ia_mode;
	int status = -ENOMEM;
L
Linus Torvalds 已提交
3931

3932 3933 3934 3935
	data = nfs4_alloc_createdata(dir, &dentry->d_name, sattr, NF4SOCK);
	if (data == NULL)
		goto out;

L
Linus Torvalds 已提交
3936
	if (S_ISFIFO(mode))
3937
		data->arg.ftype = NF4FIFO;
L
Linus Torvalds 已提交
3938
	else if (S_ISBLK(mode)) {
3939 3940 3941
		data->arg.ftype = NF4BLK;
		data->arg.u.device.specdata1 = MAJOR(rdev);
		data->arg.u.device.specdata2 = MINOR(rdev);
L
Linus Torvalds 已提交
3942 3943
	}
	else if (S_ISCHR(mode)) {
3944 3945 3946
		data->arg.ftype = NF4CHR;
		data->arg.u.device.specdata1 = MAJOR(rdev);
		data->arg.u.device.specdata2 = MINOR(rdev);
3947 3948 3949
	} else if (!S_ISSOCK(mode)) {
		status = -EINVAL;
		goto out_free;
L
Linus Torvalds 已提交
3950
	}
3951

3952
	data->arg.label = label;
3953
	status = nfs4_do_create(dir, dentry, data);
3954
out_free:
3955 3956
	nfs4_free_createdata(data);
out:
L
Linus Torvalds 已提交
3957 3958 3959 3960 3961 3962 3963
	return status;
}

static int nfs4_proc_mknod(struct inode *dir, struct dentry *dentry,
		struct iattr *sattr, dev_t rdev)
{
	struct nfs4_exception exception = { };
3964
	struct nfs4_label l, *label = NULL;
L
Linus Torvalds 已提交
3965
	int err;
A
Aneesh Kumar K.V 已提交
3966

3967 3968
	label = nfs4_label_init_security(dir, dentry, sattr, &l);

A
Aneesh Kumar K.V 已提交
3969
	sattr->ia_mode &= ~current_umask();
L
Linus Torvalds 已提交
3970
	do {
3971 3972 3973
		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
3974 3975
				&exception);
	} while (exception.retry);
3976 3977 3978

	nfs4_label_release_security(label);

L
Linus Torvalds 已提交
3979 3980 3981 3982 3983 3984 3985 3986 3987 3988
	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
3989 3990 3991
	struct nfs4_statfs_res res = {
		.fsstat = fsstat,
	};
L
Linus Torvalds 已提交
3992 3993 3994
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_STATFS],
		.rpc_argp = &args,
B
Benny Halevy 已提交
3995
		.rpc_resp = &res,
L
Linus Torvalds 已提交
3996 3997
	};

3998
	nfs_fattr_init(fsstat->fattr);
3999
	return  nfs4_call_sync(server->client, server, &msg, &args.seq_args, &res.seq_res, 0);
L
Linus Torvalds 已提交
4000 4001 4002 4003 4004 4005 4006 4007 4008 4009 4010 4011 4012 4013 4014 4015 4016 4017 4018 4019 4020
}

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
4021 4022 4023
	struct nfs4_fsinfo_res res = {
		.fsinfo = fsinfo,
	};
L
Linus Torvalds 已提交
4024 4025 4026
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_FSINFO],
		.rpc_argp = &args,
B
Benny Halevy 已提交
4027
		.rpc_resp = &res,
L
Linus Torvalds 已提交
4028 4029
	};

4030
	return nfs4_call_sync(server->client, server, &msg, &args.seq_args, &res.seq_res, 0);
L
Linus Torvalds 已提交
4031 4032 4033 4034 4035
}

static int nfs4_do_fsinfo(struct nfs_server *server, struct nfs_fh *fhandle, struct nfs_fsinfo *fsinfo)
{
	struct nfs4_exception exception = { };
4036
	unsigned long now = jiffies;
L
Linus Torvalds 已提交
4037 4038 4039
	int err;

	do {
4040
		err = _nfs4_do_fsinfo(server, fhandle, fsinfo);
4041
		trace_nfs4_fsinfo(server, fhandle, fsinfo->fattr, err);
4042 4043 4044 4045 4046 4047 4048 4049 4050 4051
		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
4052 4053 4054 4055 4056 4057
	} while (exception.retry);
	return err;
}

static int nfs4_proc_fsinfo(struct nfs_server *server, struct nfs_fh *fhandle, struct nfs_fsinfo *fsinfo)
{
4058 4059
	int error;

4060
	nfs_fattr_init(fsinfo->fattr);
4061
	error = nfs4_do_fsinfo(server, fhandle, fsinfo);
4062 4063 4064
	if (error == 0) {
		/* block layout checks this! */
		server->pnfs_blksize = fsinfo->blksize;
4065
		set_pnfs_layoutdriver(server, fhandle, fsinfo->layouttype);
4066
	}
4067 4068

	return error;
L
Linus Torvalds 已提交
4069 4070 4071 4072 4073 4074 4075 4076 4077
}

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
4078 4079 4080
	struct nfs4_pathconf_res res = {
		.pathconf = pathconf,
	};
L
Linus Torvalds 已提交
4081 4082 4083
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_PATHCONF],
		.rpc_argp = &args,
B
Benny Halevy 已提交
4084
		.rpc_resp = &res,
L
Linus Torvalds 已提交
4085 4086 4087 4088 4089 4090 4091 4092
	};

	/* 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;
	}

4093
	nfs_fattr_init(pathconf->fattr);
4094
	return nfs4_call_sync(server->client, server, &msg, &args.seq_args, &res.seq_res, 0);
L
Linus Torvalds 已提交
4095 4096 4097 4098 4099 4100 4101 4102 4103 4104 4105 4106 4107 4108 4109 4110
}

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;
}

4111
int nfs4_set_rw_stateid(nfs4_stateid *stateid,
4112 4113 4114 4115 4116 4117 4118 4119
		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;
4120
	return nfs4_select_rw_stateid(stateid, ctx->state, fmode, lockowner);
4121 4122 4123
}
EXPORT_SYMBOL_GPL(nfs4_set_rw_stateid);

4124 4125 4126 4127 4128 4129 4130
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;

4131 4132 4133
	/* If the current stateid represents a lost lock, then exit */
	if (nfs4_set_rw_stateid(&current_stateid, ctx, l_ctx, fmode) == -EIO)
		return true;
4134 4135 4136 4137 4138 4139 4140 4141 4142 4143 4144 4145 4146 4147 4148 4149 4150 4151
	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;
}

4152
void __nfs4_read_done_cb(struct nfs_pgio_header *hdr)
4153
{
4154
	nfs_invalidate_atime(hdr->inode);
4155 4156
}

4157
static int nfs4_read_done_cb(struct rpc_task *task, struct nfs_pgio_header *hdr)
L
Linus Torvalds 已提交
4158
{
4159
	struct nfs_server *server = NFS_SERVER(hdr->inode);
L
Linus Torvalds 已提交
4160

4161 4162
	trace_nfs4_read(hdr, task->tk_status);
	if (nfs4_async_handle_error(task, server,
4163 4164
				    hdr->args.context->state,
				    NULL) == -EAGAIN) {
4165
		rpc_restart_call_prepare(task);
T
Trond Myklebust 已提交
4166
		return -EAGAIN;
L
Linus Torvalds 已提交
4167
	}
4168

4169
	__nfs4_read_done_cb(hdr);
L
Linus Torvalds 已提交
4170
	if (task->tk_status > 0)
4171
		renew_lease(server, hdr->timestamp);
T
Trond Myklebust 已提交
4172
	return 0;
L
Linus Torvalds 已提交
4173 4174
}

4175
static bool nfs4_read_stateid_changed(struct rpc_task *task,
4176
		struct nfs_pgio_args *args)
4177 4178 4179 4180 4181 4182 4183 4184 4185 4186 4187 4188
{

	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;
}

4189
static int nfs4_read_done(struct rpc_task *task, struct nfs_pgio_header *hdr)
4190 4191 4192 4193
{

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

4194
	if (!nfs4_sequence_done(task, &hdr->res.seq_res))
4195
		return -EAGAIN;
4196
	if (nfs4_read_stateid_changed(task, &hdr->args))
4197
		return -EAGAIN;
4198 4199
	return hdr->pgio_done_cb ? hdr->pgio_done_cb(task, hdr) :
				    nfs4_read_done_cb(task, hdr);
4200 4201
}

4202 4203
static void nfs4_proc_read_setup(struct nfs_pgio_header *hdr,
				 struct rpc_message *msg)
L
Linus Torvalds 已提交
4204
{
4205 4206
	hdr->timestamp   = jiffies;
	hdr->pgio_done_cb = nfs4_read_done_cb;
4207
	msg->rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_READ];
4208
	nfs4_init_sequence(&hdr->args.seq_args, &hdr->res.seq_res, 0);
L
Linus Torvalds 已提交
4209 4210
}

4211 4212
static int nfs4_proc_pgio_rpc_prepare(struct rpc_task *task,
				      struct nfs_pgio_header *hdr)
4213
{
4214 4215 4216
	if (nfs4_setup_sequence(NFS_SERVER(hdr->inode),
			&hdr->args.seq_args,
			&hdr->res.seq_res,
4217
			task))
4218
		return 0;
4219 4220 4221
	if (nfs4_set_rw_stateid(&hdr->args.stateid, hdr->args.context,
				hdr->args.lock_context,
				hdr->rw_ops->rw_mode) == -EIO)
4222
		return -EIO;
4223
	if (unlikely(test_bit(NFS_CONTEXT_BAD, &hdr->args.context->flags)))
4224 4225
		return -EIO;
	return 0;
L
Linus Torvalds 已提交
4226 4227
}

4228 4229
static int nfs4_write_done_cb(struct rpc_task *task,
			      struct nfs_pgio_header *hdr)
L
Linus Torvalds 已提交
4230
{
4231
	struct inode *inode = hdr->inode;
4232

4233 4234
	trace_nfs4_write(hdr, task->tk_status);
	if (nfs4_async_handle_error(task, NFS_SERVER(inode),
4235 4236
				    hdr->args.context->state,
				    NULL) == -EAGAIN) {
4237
		rpc_restart_call_prepare(task);
4238
		return -EAGAIN;
L
Linus Torvalds 已提交
4239
	}
4240
	if (task->tk_status >= 0) {
4241
		renew_lease(NFS_SERVER(inode), hdr->timestamp);
4242
		nfs_writeback_update_inode(hdr);
4243
	}
4244
	return 0;
L
Linus Torvalds 已提交
4245 4246
}

4247
static bool nfs4_write_stateid_changed(struct rpc_task *task,
4248
		struct nfs_pgio_args *args)
4249 4250 4251 4252 4253 4254 4255 4256 4257 4258 4259 4260
{

	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;
}

4261
static int nfs4_write_done(struct rpc_task *task, struct nfs_pgio_header *hdr)
4262
{
4263
	if (!nfs4_sequence_done(task, &hdr->res.seq_res))
4264
		return -EAGAIN;
4265
	if (nfs4_write_stateid_changed(task, &hdr->args))
4266
		return -EAGAIN;
4267 4268
	return hdr->pgio_done_cb ? hdr->pgio_done_cb(task, hdr) :
		nfs4_write_done_cb(task, hdr);
4269 4270
}

4271
static
4272
bool nfs4_write_need_cache_consistency_data(struct nfs_pgio_header *hdr)
4273
{
4274
	/* Don't request attributes for pNFS or O_DIRECT writes */
4275
	if (hdr->ds_clp != NULL || hdr->dreq != NULL)
4276 4277 4278 4279
		return false;
	/* Otherwise, request attributes if and only if we don't hold
	 * a delegation
	 */
4280
	return nfs4_have_delegation(hdr->inode, FMODE_READ) == 0;
4281 4282
}

4283 4284
static void nfs4_proc_write_setup(struct nfs_pgio_header *hdr,
				  struct rpc_message *msg)
L
Linus Torvalds 已提交
4285
{
4286
	struct nfs_server *server = NFS_SERVER(hdr->inode);
4287

4288 4289 4290
	if (!nfs4_write_need_cache_consistency_data(hdr)) {
		hdr->args.bitmask = NULL;
		hdr->res.fattr = NULL;
4291
	} else
4292
		hdr->args.bitmask = server->cache_consistency_bitmask;
4293

4294 4295 4296 4297
	if (!hdr->pgio_done_cb)
		hdr->pgio_done_cb = nfs4_write_done_cb;
	hdr->res.server = server;
	hdr->timestamp   = jiffies;
L
Linus Torvalds 已提交
4298

4299
	msg->rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_WRITE];
4300
	nfs4_init_sequence(&hdr->args.seq_args, &hdr->res.seq_res, 1);
L
Linus Torvalds 已提交
4301 4302
}

4303
static void nfs4_proc_commit_rpc_prepare(struct rpc_task *task, struct nfs_commit_data *data)
4304
{
4305 4306 4307 4308
	nfs4_setup_sequence(NFS_SERVER(data->inode),
			&data->args.seq_args,
			&data->res.seq_res,
			task);
L
Linus Torvalds 已提交
4309 4310
}

4311
static int nfs4_commit_done_cb(struct rpc_task *task, struct nfs_commit_data *data)
L
Linus Torvalds 已提交
4312 4313
{
	struct inode *inode = data->inode;
4314

4315
	trace_nfs4_commit(data, task->tk_status);
4316 4317
	if (nfs4_async_handle_error(task, NFS_SERVER(inode),
				    NULL, NULL) == -EAGAIN) {
4318
		rpc_restart_call_prepare(task);
4319
		return -EAGAIN;
L
Linus Torvalds 已提交
4320
	}
4321
	return 0;
L
Linus Torvalds 已提交
4322 4323
}

4324
static int nfs4_commit_done(struct rpc_task *task, struct nfs_commit_data *data)
4325 4326 4327
{
	if (!nfs4_sequence_done(task, &data->res.seq_res))
		return -EAGAIN;
4328
	return data->commit_done_cb(task, data);
4329 4330
}

4331
static void nfs4_proc_commit_setup(struct nfs_commit_data *data, struct rpc_message *msg)
L
Linus Torvalds 已提交
4332
{
4333
	struct nfs_server *server = NFS_SERVER(data->inode);
4334

4335 4336
	if (data->commit_done_cb == NULL)
		data->commit_done_cb = nfs4_commit_done_cb;
4337
	data->res.server = server;
4338
	msg->rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_COMMIT];
4339
	nfs4_init_sequence(&data->args.seq_args, &data->res.seq_res, 1);
L
Linus Torvalds 已提交
4340 4341
}

4342 4343 4344 4345 4346
struct nfs4_renewdata {
	struct nfs_client	*client;
	unsigned long		timestamp;
};

L
Linus Torvalds 已提交
4347 4348 4349 4350
/*
 * nfs4_proc_async_renew(): This is not one of the nfs_rpc_ops; it is a special
 * standalone procedure for queueing an asynchronous RENEW.
 */
4351
static void nfs4_renew_release(void *calldata)
4352
{
4353 4354
	struct nfs4_renewdata *data = calldata;
	struct nfs_client *clp = data->client;
4355

4356 4357 4358
	if (atomic_read(&clp->cl_count) > 1)
		nfs4_schedule_state_renewal(clp);
	nfs_put_client(clp);
4359
	kfree(data);
4360 4361
}

4362
static void nfs4_renew_done(struct rpc_task *task, void *calldata)
L
Linus Torvalds 已提交
4363
{
4364 4365 4366
	struct nfs4_renewdata *data = calldata;
	struct nfs_client *clp = data->client;
	unsigned long timestamp = data->timestamp;
L
Linus Torvalds 已提交
4367

4368
	trace_nfs4_renew_async(clp, task->tk_status);
4369 4370 4371 4372 4373 4374 4375
	switch (task->tk_status) {
	case 0:
		break;
	case -NFS4ERR_LEASE_MOVED:
		nfs4_schedule_lease_moved_recovery(clp);
		break;
	default:
4376
		/* Unless we're shutting down, schedule state recovery! */
4377 4378 4379
		if (test_bit(NFS_CS_RENEWD, &clp->cl_res_state) == 0)
			return;
		if (task->tk_status != NFS4ERR_CB_PATH_DOWN) {
4380
			nfs4_schedule_lease_recovery(clp);
4381 4382 4383
			return;
		}
		nfs4_schedule_path_down_recovery(clp);
L
Linus Torvalds 已提交
4384
	}
4385
	do_renew_lease(clp, timestamp);
L
Linus Torvalds 已提交
4386 4387
}

4388 4389
static const struct rpc_call_ops nfs4_renew_ops = {
	.rpc_call_done = nfs4_renew_done,
4390
	.rpc_release = nfs4_renew_release,
4391 4392
};

4393
static int nfs4_proc_async_renew(struct nfs_client *clp, struct rpc_cred *cred, unsigned renew_flags)
L
Linus Torvalds 已提交
4394 4395 4396 4397
{
	struct rpc_message msg = {
		.rpc_proc	= &nfs4_procedures[NFSPROC4_CLNT_RENEW],
		.rpc_argp	= clp,
4398
		.rpc_cred	= cred,
L
Linus Torvalds 已提交
4399
	};
4400
	struct nfs4_renewdata *data;
L
Linus Torvalds 已提交
4401

4402 4403
	if (renew_flags == 0)
		return 0;
4404 4405
	if (!atomic_inc_not_zero(&clp->cl_count))
		return -EIO;
4406
	data = kmalloc(sizeof(*data), GFP_NOFS);
4407 4408 4409 4410
	if (data == NULL)
		return -ENOMEM;
	data->client = clp;
	data->timestamp = jiffies;
4411
	return rpc_call_async(clp->cl_rpcclient, &msg, RPC_TASK_TIMEOUT,
4412
			&nfs4_renew_ops, data);
L
Linus Torvalds 已提交
4413 4414
}

4415
static int nfs4_proc_renew(struct nfs_client *clp, struct rpc_cred *cred)
L
Linus Torvalds 已提交
4416 4417 4418 4419
{
	struct rpc_message msg = {
		.rpc_proc	= &nfs4_procedures[NFSPROC4_CLNT_RENEW],
		.rpc_argp	= clp,
4420
		.rpc_cred	= cred,
L
Linus Torvalds 已提交
4421 4422 4423 4424
	};
	unsigned long now = jiffies;
	int status;

4425
	status = rpc_call_sync(clp->cl_rpcclient, &msg, RPC_TASK_TIMEOUT);
L
Linus Torvalds 已提交
4426 4427
	if (status < 0)
		return status;
4428
	do_renew_lease(clp, now);
L
Linus Torvalds 已提交
4429 4430 4431
	return 0;
}

4432 4433
static inline int nfs4_server_supports_acls(struct nfs_server *server)
{
4434
	return server->caps & NFS_CAP_ACLS;
4435 4436
}

4437 4438
/* 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
4439 4440
 * the stack.
 */
4441
#define NFS4ACL_MAXPAGES DIV_ROUND_UP(XATTR_SIZE_MAX, PAGE_SIZE)
4442

4443 4444 4445 4446 4447 4448 4449 4450 4451
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 {
4452
		len = min_t(size_t, PAGE_SIZE, buflen);
4453 4454 4455 4456 4457 4458 4459 4460 4461 4462 4463 4464 4465 4466 4467 4468 4469 4470 4471
		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;
}

4472 4473 4474
struct nfs4_cached_acl {
	int cached;
	size_t len;
A
Andrew Morton 已提交
4475
	char data[0];
4476 4477 4478 4479 4480 4481 4482 4483 4484 4485 4486 4487 4488 4489 4490 4491 4492 4493 4494 4495 4496 4497 4498 4499 4500 4501 4502 4503 4504 4505 4506 4507 4508 4509 4510 4511 4512 4513 4514 4515 4516 4517
};

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;
}

4518
static void nfs4_write_cached_acl(struct inode *inode, struct page **pages, size_t pgbase, size_t acl_len)
4519 4520
{
	struct nfs4_cached_acl *acl;
4521
	size_t buflen = sizeof(*acl) + acl_len;
4522

4523
	if (buflen <= PAGE_SIZE) {
4524
		acl = kmalloc(buflen, GFP_KERNEL);
4525 4526 4527
		if (acl == NULL)
			goto out;
		acl->cached = 1;
4528
		_copy_from_pages(acl->data, pages, pgbase, acl_len);
4529 4530 4531 4532 4533 4534 4535 4536 4537 4538 4539
	} 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);
}

4540 4541 4542 4543 4544 4545 4546 4547 4548 4549
/*
 * 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.
 */
4550
static ssize_t __nfs4_get_acl_uncached(struct inode *inode, void *buf, size_t buflen)
4551
{
4552
	struct page *pages[NFS4ACL_MAXPAGES] = {NULL, };
4553 4554 4555 4556 4557
	struct nfs_getaclargs args = {
		.fh = NFS_FH(inode),
		.acl_pages = pages,
		.acl_len = buflen,
	};
B
Benny Halevy 已提交
4558 4559 4560
	struct nfs_getaclres res = {
		.acl_len = buflen,
	};
4561 4562 4563
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_GETACL],
		.rpc_argp = &args,
B
Benny Halevy 已提交
4564
		.rpc_resp = &res,
4565
	};
4566 4567
	unsigned int npages = DIV_ROUND_UP(buflen, PAGE_SIZE);
	int ret = -ENOMEM, i;
4568

4569 4570 4571 4572
	/* 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;
4573 4574
	if (npages > ARRAY_SIZE(pages))
		return -ERANGE;
4575

4576 4577 4578 4579
	for (i = 0; i < npages; i++) {
		pages[i] = alloc_page(GFP_KERNEL);
		if (!pages[i])
			goto out_free;
4580
	}
4581 4582 4583 4584 4585 4586

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

4587 4588
	args.acl_len = npages * PAGE_SIZE;
	args.acl_pgbase = 0;
4589

4590
	dprintk("%s  buf %p buflen %zu npages %d args.acl_len %zu\n",
4591 4592 4593
		__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);
4594 4595
	if (ret)
		goto out_free;
4596

4597 4598 4599 4600 4601
	/* 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;
4602
		ret = -ERANGE;
4603
		goto out_free;
4604
	}
4605
	nfs4_write_cached_acl(inode, pages, res.acl_data_offset, res.acl_len);
4606 4607 4608 4609 4610
	if (buf) {
		if (res.acl_len > buflen) {
			ret = -ERANGE;
			goto out_free;
		}
4611
		_copy_from_pages(buf, pages, res.acl_data_offset, res.acl_len);
4612
	}
4613 4614
out_ok:
	ret = res.acl_len;
4615
out_free:
4616 4617 4618
	for (i = 0; i < npages; i++)
		if (pages[i])
			__free_page(pages[i]);
4619 4620
	if (res.acl_scratch)
		__free_page(res.acl_scratch);
4621 4622 4623
	return ret;
}

4624 4625 4626 4627 4628 4629
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);
4630
		trace_nfs4_get_acl(inode, ret);
4631 4632 4633 4634 4635 4636 4637
		if (ret >= 0)
			break;
		ret = nfs4_handle_exception(NFS_SERVER(inode), ret, &exception);
	} while (exception.retry);
	return ret;
}

4638 4639 4640 4641 4642 4643 4644 4645 4646 4647
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;
4648 4649
	if (NFS_I(inode)->cache_validity & NFS_INO_INVALID_ACL)
		nfs_zap_acl_cache(inode);
4650 4651
	ret = nfs4_read_cached_acl(inode, buf, buflen);
	if (ret != -ENOENT)
4652 4653
		/* -ENOENT is returned if there is no ACL or if there is an ACL
		 * but no cached acl data, just the acl length */
4654 4655 4656 4657
		return ret;
	return nfs4_get_acl_uncached(inode, buf, buflen);
}

4658
static int __nfs4_proc_set_acl(struct inode *inode, const void *buf, size_t buflen)
4659 4660 4661 4662 4663 4664 4665 4666
{
	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
4667
	struct nfs_setaclres res;
4668 4669 4670
	struct rpc_message msg = {
		.rpc_proc	= &nfs4_procedures[NFSPROC4_CLNT_SETACL],
		.rpc_argp	= &arg,
B
Benny Halevy 已提交
4671
		.rpc_resp	= &res,
4672
	};
4673
	unsigned int npages = DIV_ROUND_UP(buflen, PAGE_SIZE);
4674
	int ret, i;
4675 4676 4677

	if (!nfs4_server_supports_acls(server))
		return -EOPNOTSUPP;
4678 4679
	if (npages > ARRAY_SIZE(pages))
		return -ERANGE;
4680 4681 4682
	i = buf_to_pages_noslab(buf, buflen, arg.acl_pages, &arg.acl_pgbase);
	if (i < 0)
		return i;
4683
	nfs4_inode_return_delegation(inode);
4684
	ret = nfs4_call_sync(server->client, server, &msg, &arg.seq_args, &res.seq_res, 1);
4685 4686 4687 4688 4689 4690 4691 4692

	/*
	 * 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]);

4693 4694 4695 4696 4697 4698 4699
	/*
	 * 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);
4700 4701
	nfs_access_zap_cache(inode);
	nfs_zap_acl_cache(inode);
4702 4703 4704
	return ret;
}

4705 4706 4707 4708 4709
static int nfs4_proc_set_acl(struct inode *inode, const void *buf, size_t buflen)
{
	struct nfs4_exception exception = { };
	int err;
	do {
4710 4711 4712
		err = __nfs4_proc_set_acl(inode, buf, buflen);
		trace_nfs4_set_acl(inode, err);
		err = nfs4_handle_exception(NFS_SERVER(inode), err,
4713 4714 4715 4716 4717
				&exception);
	} while (exception.retry);
	return err;
}

4718 4719 4720 4721 4722 4723 4724 4725 4726
#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 };
4727
	struct nfs4_getattr_arg arg = {
4728 4729 4730 4731 4732 4733 4734 4735 4736 4737
		.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],
4738
		.rpc_argp	= &arg,
4739 4740 4741 4742 4743 4744
		.rpc_resp	= &res,
	};
	int ret;

	nfs_fattr_init(&fattr);

4745
	ret = nfs4_call_sync(server->client, server, &msg, &arg.seq_args, &res.seq_res, 0);
4746 4747 4748 4749 4750 4751 4752 4753 4754 4755 4756 4757 4758 4759 4760 4761 4762 4763 4764
	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 {
4765 4766 4767
		err = _nfs4_get_security_label(inode, buf, buflen);
		trace_nfs4_get_security_label(inode, err);
		err = nfs4_handle_exception(NFS_SERVER(inode), err,
4768 4769 4770 4771 4772 4773 4774 4775 4776 4777 4778 4779 4780 4781
				&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 };
4782
	struct nfs_setattrargs arg = {
4783 4784 4785 4786 4787 4788 4789 4790 4791 4792 4793 4794 4795
		.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],
4796
		.rpc_argp       = &arg,
4797 4798 4799 4800
		.rpc_resp       = &res,
	};
	int status;

4801
	nfs4_stateid_copy(&arg.stateid, &zero_stateid);
4802

4803
	status = nfs4_call_sync(server->client, server, &msg, &arg.seq_args, &res.seq_res, 1);
4804 4805 4806 4807 4808 4809 4810 4811 4812 4813 4814 4815 4816 4817 4818
	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 {
4819 4820 4821 4822
		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,
4823 4824 4825 4826 4827 4828 4829 4830 4831 4832 4833 4834 4835 4836 4837 4838 4839 4840 4841 4842 4843 4844 4845 4846 4847 4848 4849 4850 4851 4852 4853 4854 4855 4856 4857 4858 4859 4860 4861 4862 4863 4864 4865 4866 4867 4868
				&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
4869
static int
4870 4871
nfs4_async_handle_error(struct rpc_task *task, const struct nfs_server *server,
			struct nfs4_state *state, long *timeout)
L
Linus Torvalds 已提交
4872
{
4873 4874 4875
	struct nfs_client *clp = server->nfs_client;

	if (task->tk_status >= 0)
L
Linus Torvalds 已提交
4876 4877
		return 0;
	switch(task->tk_status) {
4878
		case -NFS4ERR_DELEG_REVOKED:
4879 4880 4881 4882 4883
		case -NFS4ERR_ADMIN_REVOKED:
		case -NFS4ERR_BAD_STATEID:
		case -NFS4ERR_OPENMODE:
			if (state == NULL)
				break;
4884
			if (nfs4_schedule_stateid_recovery(server, state) < 0)
C
Chuck Lever 已提交
4885
				goto recovery_failed;
4886
			goto wait_on_recovery;
4887
		case -NFS4ERR_EXPIRED:
4888 4889
			if (state != NULL) {
				if (nfs4_schedule_stateid_recovery(server, state) < 0)
C
Chuck Lever 已提交
4890
					goto recovery_failed;
4891
			}
L
Linus Torvalds 已提交
4892
		case -NFS4ERR_STALE_STATEID:
4893
		case -NFS4ERR_STALE_CLIENTID:
4894 4895
			nfs4_schedule_lease_recovery(clp);
			goto wait_on_recovery;
4896 4897 4898 4899
		case -NFS4ERR_MOVED:
			if (nfs4_schedule_migration_recovery(server) < 0)
				goto recovery_failed;
			goto wait_on_recovery;
4900 4901 4902
		case -NFS4ERR_LEASE_MOVED:
			nfs4_schedule_lease_moved_recovery(clp);
			goto wait_on_recovery;
4903 4904 4905 4906 4907 4908 4909 4910 4911 4912
#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);
4913
			nfs4_schedule_session_recovery(clp->cl_session, task->tk_status);
4914
			goto wait_on_recovery;
4915
#endif /* CONFIG_NFS_V4_1 */
L
Linus Torvalds 已提交
4916
		case -NFS4ERR_DELAY:
4917
			nfs_inc_server_stats(server, NFSIOS_DELAY);
4918 4919
			rpc_delay(task, nfs4_update_delay(timeout));
			goto restart_call;
4920
		case -NFS4ERR_GRACE:
L
Linus Torvalds 已提交
4921
			rpc_delay(task, NFS4_POLL_RETRY_MAX);
4922
		case -NFS4ERR_RETRY_UNCACHED_REP:
L
Linus Torvalds 已提交
4923
		case -NFS4ERR_OLD_STATEID:
4924
			goto restart_call;
L
Linus Torvalds 已提交
4925 4926 4927
	}
	task->tk_status = nfs4_map_errors(task->tk_status);
	return 0;
C
Chuck Lever 已提交
4928
recovery_failed:
4929 4930
	task->tk_status = -EIO;
	return 0;
4931
wait_on_recovery:
4932 4933 4934
	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);
4935 4936
	if (test_bit(NFS_MIG_FAILED, &server->mig_status))
		goto recovery_failed;
4937
restart_call:
4938 4939
	task->tk_status = 0;
	return -EAGAIN;
L
Linus Torvalds 已提交
4940 4941
}

4942 4943
static void nfs4_init_boot_verifier(const struct nfs_client *clp,
				    nfs4_verifier *bootverf)
4944 4945 4946
{
	__be32 verf[2];

4947 4948 4949 4950
	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;
4951
		verf[1] = cpu_to_be32(NSEC_PER_SEC + 1);
4952
	} else {
4953
		struct nfs_net *nn = net_generic(clp->cl_net, nfs_net_id);
4954 4955
		verf[0] = cpu_to_be32(nn->boot_time.tv_sec);
		verf[1] = cpu_to_be32(nn->boot_time.tv_nsec);
4956
	}
4957 4958 4959
	memcpy(bootverf->data, verf, sizeof(bootverf->data));
}

4960
static unsigned int
4961
nfs4_init_nonuniform_client_string(struct nfs_client *clp,
4962 4963 4964 4965
				   char *buf, size_t len)
{
	unsigned int result;

4966 4967 4968
	if (clp->cl_owner_id != NULL)
		return strlcpy(buf, clp->cl_owner_id, len);

4969 4970 4971 4972 4973 4974 4975 4976
	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();
4977
	clp->cl_owner_id = kstrdup(buf, GFP_KERNEL);
4978 4979 4980 4981
	return result;
}

static unsigned int
4982
nfs4_init_uniform_client_string(struct nfs_client *clp,
4983 4984
				char *buf, size_t len)
{
4985
	const char *nodename = clp->cl_rpcclient->cl_nodename;
4986 4987 4988 4989
	unsigned int result;

	if (clp->cl_owner_id != NULL)
		return strlcpy(buf, clp->cl_owner_id, len);
4990 4991

	if (nfs4_client_id_uniquifier[0] != '\0')
4992
		result = scnprintf(buf, len, "Linux NFSv%u.%u %s/%s",
4993 4994 4995 4996
				clp->rpc_ops->version,
				clp->cl_minorversion,
				nfs4_client_id_uniquifier,
				nodename);
4997 4998
	else
		result = scnprintf(buf, len, "Linux NFSv%u.%u %s",
4999
				clp->rpc_ops->version, clp->cl_minorversion,
5000
				nodename);
5001 5002
	clp->cl_owner_id = kstrdup(buf, GFP_KERNEL);
	return result;
5003 5004
}

5005 5006 5007 5008 5009 5010 5011 5012 5013 5014 5015 5016 5017 5018
/*
 * nfs4_callback_up_net() starts only "tcp" and "tcp6" callback
 * services.  Advertise one based on the address family of the
 * clientaddr.
 */
static unsigned int
nfs4_init_callback_netid(const struct nfs_client *clp, char *buf, size_t len)
{
	if (strchr(clp->cl_ipaddr, ':') != NULL)
		return scnprintf(buf, len, "tcp6");
	else
		return scnprintf(buf, len, "tcp");
}

5019 5020 5021 5022 5023 5024 5025 5026 5027 5028 5029 5030
static void nfs4_setclientid_done(struct rpc_task *task, void *calldata)
{
	struct nfs4_setclientid *sc = calldata;

	if (task->tk_status == 0)
		sc->sc_cred = get_rpccred(task->tk_rqstp->rq_cred);
}

static const struct rpc_call_ops nfs4_setclientid_ops = {
	.rpc_call_done = nfs4_setclientid_done,
};

5031 5032 5033 5034 5035 5036 5037 5038 5039 5040
/**
 * 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.
 */
5041 5042 5043
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
5044 5045 5046 5047 5048
{
	nfs4_verifier sc_verifier;
	struct nfs4_setclientid setclientid = {
		.sc_verifier = &sc_verifier,
		.sc_prog = program,
A
Andy Adamson 已提交
5049
		.sc_cb_ident = clp->cl_cb_ident,
L
Linus Torvalds 已提交
5050 5051 5052 5053
	};
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_SETCLIENTID],
		.rpc_argp = &setclientid,
5054
		.rpc_resp = res,
5055
		.rpc_cred = cred,
L
Linus Torvalds 已提交
5056
	};
5057 5058 5059 5060 5061 5062 5063 5064
	struct rpc_task *task;
	struct rpc_task_setup task_setup_data = {
		.rpc_client = clp->cl_rpcclient,
		.rpc_message = &msg,
		.callback_ops = &nfs4_setclientid_ops,
		.callback_data = &setclientid,
		.flags = RPC_TASK_TIMEOUT,
	};
5065
	int status;
L
Linus Torvalds 已提交
5066

5067
	/* nfs_client_id4 */
5068
	nfs4_init_boot_verifier(clp, &sc_verifier);
5069 5070 5071 5072 5073 5074 5075 5076 5077 5078
	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));
5079
	/* cb_client4 */
5080 5081 5082 5083
	setclientid.sc_netid_len =
				nfs4_init_callback_netid(clp,
						setclientid.sc_netid,
						sizeof(setclientid.sc_netid));
5084
	setclientid.sc_uaddr_len = scnprintf(setclientid.sc_uaddr,
5085
				sizeof(setclientid.sc_uaddr), "%s.%u.%u",
L
Linus Torvalds 已提交
5086 5087
				clp->cl_ipaddr, port >> 8, port & 255);

5088 5089 5090
	dprintk("NFS call  setclientid auth=%s, '%.*s'\n",
		clp->cl_rpcclient->cl_auth->au_ops->au_name,
		setclientid.sc_name_len, setclientid.sc_name);
5091 5092 5093 5094 5095 5096 5097 5098 5099 5100 5101 5102
	task = rpc_run_task(&task_setup_data);
	if (IS_ERR(task)) {
		status = PTR_ERR(task);
		goto out;
	}
	status = task->tk_status;
	if (setclientid.sc_cred) {
		clp->cl_acceptor = rpcauth_stringify_acceptor(setclientid.sc_cred);
		put_rpccred(setclientid.sc_cred);
	}
	rpc_put_task(task);
out:
5103
	trace_nfs4_setclientid(clp, status);
5104 5105
	dprintk("NFS reply setclientid: %d\n", status);
	return status;
L
Linus Torvalds 已提交
5106 5107
}

5108 5109 5110 5111 5112 5113 5114 5115
/**
 * 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.
 */
5116
int nfs4_proc_setclientid_confirm(struct nfs_client *clp,
5117 5118
		struct nfs4_setclientid_res *arg,
		struct rpc_cred *cred)
L
Linus Torvalds 已提交
5119 5120 5121
{
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_SETCLIENTID_CONFIRM],
5122
		.rpc_argp = arg,
5123
		.rpc_cred = cred,
L
Linus Torvalds 已提交
5124 5125 5126
	};
	int status;

5127 5128 5129
	dprintk("NFS call  setclientid_confirm auth=%s, (client ID %llx)\n",
		clp->cl_rpcclient->cl_auth->au_ops->au_name,
		clp->cl_clientid);
5130
	status = rpc_call_sync(clp->cl_rpcclient, &msg, RPC_TASK_TIMEOUT);
5131
	trace_nfs4_setclientid_confirm(clp, status);
5132
	dprintk("NFS reply setclientid_confirm: %d\n", status);
L
Linus Torvalds 已提交
5133 5134 5135
	return status;
}

5136 5137
struct nfs4_delegreturndata {
	struct nfs4_delegreturnargs args;
5138
	struct nfs4_delegreturnres res;
5139 5140
	struct nfs_fh fh;
	nfs4_stateid stateid;
5141
	unsigned long timestamp;
5142
	struct nfs_fattr fattr;
5143
	int rpc_status;
5144 5145 5146
	struct inode *inode;
	bool roc;
	u32 roc_barrier;
5147 5148 5149 5150 5151
};

static void nfs4_delegreturn_done(struct rpc_task *task, void *calldata)
{
	struct nfs4_delegreturndata *data = calldata;
5152

5153 5154
	if (!nfs4_sequence_done(task, &data->res.seq_res))
		return;
5155

5156
	trace_nfs4_delegreturn_exit(&data->args, &data->res, task->tk_status);
5157 5158
	switch (task->tk_status) {
	case 0:
5159
		renew_lease(data->res.server, data->timestamp);
5160 5161 5162 5163 5164 5165 5166
	case -NFS4ERR_ADMIN_REVOKED:
	case -NFS4ERR_DELEG_REVOKED:
	case -NFS4ERR_BAD_STATEID:
	case -NFS4ERR_OLD_STATEID:
	case -NFS4ERR_STALE_STATEID:
	case -NFS4ERR_EXPIRED:
		task->tk_status = 0;
5167 5168
		if (data->roc)
			pnfs_roc_set_barrier(data->inode, data->roc_barrier);
5169
		break;
5170
	default:
5171 5172
		if (nfs4_async_handle_error(task, data->res.server,
					    NULL, NULL) == -EAGAIN) {
5173
			rpc_restart_call_prepare(task);
5174 5175 5176 5177
			return;
		}
	}
	data->rpc_status = task->tk_status;
5178 5179 5180 5181
}

static void nfs4_delegreturn_release(void *calldata)
{
5182
	struct nfs4_delegreturndata *data = calldata;
5183
	struct inode *inode = data->inode;
5184

5185 5186 5187 5188 5189
	if (inode) {
		if (data->roc)
			pnfs_roc_release(inode);
		nfs_iput_and_deactive(inode);
	}
5190 5191 5192
	kfree(calldata);
}

5193 5194 5195 5196 5197 5198
static void nfs4_delegreturn_prepare(struct rpc_task *task, void *data)
{
	struct nfs4_delegreturndata *d_data;

	d_data = (struct nfs4_delegreturndata *)data;

5199 5200 5201 5202
	if (d_data->roc &&
	    pnfs_roc_drain(d_data->inode, &d_data->roc_barrier, task))
		return;

5203 5204 5205 5206
	nfs4_setup_sequence(d_data->res.server,
			&d_data->args.seq_args,
			&d_data->res.seq_res,
			task);
5207 5208
}

5209
static const struct rpc_call_ops nfs4_delegreturn_ops = {
5210
	.rpc_call_prepare = nfs4_delegreturn_prepare,
5211 5212 5213 5214
	.rpc_call_done = nfs4_delegreturn_done,
	.rpc_release = nfs4_delegreturn_release,
};

5215
static int _nfs4_proc_delegreturn(struct inode *inode, struct rpc_cred *cred, const nfs4_stateid *stateid, int issync)
5216 5217
{
	struct nfs4_delegreturndata *data;
5218
	struct nfs_server *server = NFS_SERVER(inode);
5219
	struct rpc_task *task;
5220 5221 5222 5223
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_DELEGRETURN],
		.rpc_cred = cred,
	};
T
Trond Myklebust 已提交
5224 5225
	struct rpc_task_setup task_setup_data = {
		.rpc_client = server->client,
5226
		.rpc_message = &msg,
T
Trond Myklebust 已提交
5227 5228 5229
		.callback_ops = &nfs4_delegreturn_ops,
		.flags = RPC_TASK_ASYNC,
	};
5230
	int status = 0;
5231

5232
	data = kzalloc(sizeof(*data), GFP_NOFS);
5233 5234
	if (data == NULL)
		return -ENOMEM;
5235
	nfs4_init_sequence(&data->args.seq_args, &data->res.seq_res, 1);
5236 5237
	data->args.fhandle = &data->fh;
	data->args.stateid = &data->stateid;
5238
	data->args.bitmask = server->cache_consistency_bitmask;
5239
	nfs_copy_fh(&data->fh, NFS_FH(inode));
5240
	nfs4_stateid_copy(&data->stateid, stateid);
5241 5242
	data->res.fattr = &data->fattr;
	data->res.server = server;
5243
	nfs_fattr_init(data->res.fattr);
5244
	data->timestamp = jiffies;
5245
	data->rpc_status = 0;
5246 5247 5248
	data->inode = nfs_igrab_and_active(inode);
	if (data->inode)
		data->roc = nfs4_roc(inode);
5249

T
Trond Myklebust 已提交
5250
	task_setup_data.callback_data = data;
5251 5252
	msg.rpc_argp = &data->args;
	msg.rpc_resp = &data->res;
T
Trond Myklebust 已提交
5253
	task = rpc_run_task(&task_setup_data);
5254
	if (IS_ERR(task))
5255
		return PTR_ERR(task);
5256 5257
	if (!issync)
		goto out;
5258
	status = nfs4_wait_for_completion_rpc_task(task);
5259 5260 5261
	if (status != 0)
		goto out;
	status = data->rpc_status;
5262 5263 5264 5265
	if (status == 0)
		nfs_post_op_update_inode_force_wcc(inode, &data->fattr);
	else
		nfs_refresh_inode(inode, &data->fattr);
5266
out:
5267
	rpc_put_task(task);
5268
	return status;
L
Linus Torvalds 已提交
5269 5270
}

5271
int nfs4_proc_delegreturn(struct inode *inode, struct rpc_cred *cred, const nfs4_stateid *stateid, int issync)
L
Linus Torvalds 已提交
5272 5273 5274 5275 5276
{
	struct nfs_server *server = NFS_SERVER(inode);
	struct nfs4_exception exception = { };
	int err;
	do {
5277
		err = _nfs4_proc_delegreturn(inode, cred, stateid, issync);
5278
		trace_nfs4_delegreturn(inode, err);
L
Linus Torvalds 已提交
5279 5280 5281 5282 5283 5284 5285 5286 5287 5288 5289 5290 5291 5292 5293 5294 5295 5296 5297 5298
		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)
{
5299
	freezable_schedule_timeout_killable_unsafe(timeout);
L
Linus Torvalds 已提交
5300 5301 5302 5303 5304 5305 5306 5307 5308 5309
	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);
5310
	struct nfs_client *clp = server->nfs_client;
T
Trond Myklebust 已提交
5311
	struct nfs_lockt_args arg = {
L
Linus Torvalds 已提交
5312
		.fh = NFS_FH(inode),
T
Trond Myklebust 已提交
5313
		.fl = request,
L
Linus Torvalds 已提交
5314
	};
T
Trond Myklebust 已提交
5315 5316
	struct nfs_lockt_res res = {
		.denied = request,
L
Linus Torvalds 已提交
5317 5318 5319 5320 5321 5322 5323 5324 5325 5326
	};
	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
5327
	arg.lock_owner.clientid = clp->cl_clientid;
5328 5329 5330 5331
	status = nfs4_set_lock_state(state, request);
	if (status != 0)
		goto out;
	lsp = request->fl_u.nfs4_fl.owner;
5332
	arg.lock_owner.id = lsp->ls_seqid.owner_id;
5333
	arg.lock_owner.s_dev = server->s_dev;
5334
	status = nfs4_call_sync(server->client, server, &msg, &arg.seq_args, &res.seq_res, 1);
T
Trond Myklebust 已提交
5335 5336 5337 5338 5339 5340
	switch (status) {
		case 0:
			request->fl_type = F_UNLCK;
			break;
		case -NFS4ERR_DENIED:
			status = 0;
L
Linus Torvalds 已提交
5341
	}
5342
	request->fl_ops->fl_release_private(request);
5343
	request->fl_ops = NULL;
5344
out:
L
Linus Torvalds 已提交
5345 5346 5347 5348 5349 5350 5351 5352 5353
	return status;
}

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

	do {
5354 5355 5356
		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
5357 5358 5359 5360 5361 5362 5363 5364 5365 5366 5367 5368 5369 5370 5371 5372 5373 5374 5375 5376 5377
				&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;
}

5378
struct nfs4_unlockdata {
T
Trond Myklebust 已提交
5379 5380
	struct nfs_locku_args arg;
	struct nfs_locku_res res;
5381 5382
	struct nfs4_lock_state *lsp;
	struct nfs_open_context *ctx;
T
Trond Myklebust 已提交
5383 5384
	struct file_lock fl;
	const struct nfs_server *server;
5385
	unsigned long timestamp;
5386 5387
};

T
Trond Myklebust 已提交
5388 5389 5390 5391 5392 5393 5394 5395
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;

5396
	p = kzalloc(sizeof(*p), GFP_NOFS);
T
Trond Myklebust 已提交
5397 5398 5399 5400 5401
	if (p == NULL)
		return NULL;
	p->arg.fh = NFS_FH(inode);
	p->arg.fl = &p->fl;
	p->arg.seqid = seqid;
5402
	p->res.seqid = seqid;
T
Trond Myklebust 已提交
5403 5404 5405 5406 5407 5408 5409 5410 5411
	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;
}

5412
static void nfs4_locku_release_calldata(void *data)
5413
{
5414
	struct nfs4_unlockdata *calldata = data;
T
Trond Myklebust 已提交
5415
	nfs_free_seqid(calldata->arg.seqid);
5416 5417 5418
	nfs4_put_lock_state(calldata->lsp);
	put_nfs_open_context(calldata->ctx);
	kfree(calldata);
5419 5420
}

5421
static void nfs4_locku_done(struct rpc_task *task, void *data)
5422
{
5423
	struct nfs4_unlockdata *calldata = data;
5424

5425 5426
	if (!nfs4_sequence_done(task, &calldata->res.seq_res))
		return;
5427 5428
	switch (task->tk_status) {
		case 0:
5429
			renew_lease(calldata->server, calldata->timestamp);
5430 5431 5432 5433
			do_vfs_lock(calldata->fl.fl_file, &calldata->fl);
			if (nfs4_update_lock_stateid(calldata->lsp,
					&calldata->res.stateid))
				break;
5434 5435
		case -NFS4ERR_BAD_STATEID:
		case -NFS4ERR_OLD_STATEID:
5436 5437
		case -NFS4ERR_STALE_STATEID:
		case -NFS4ERR_EXPIRED:
5438 5439 5440
			if (!nfs4_stateid_match(&calldata->arg.stateid,
						&calldata->lsp->ls_stateid))
				rpc_restart_call_prepare(task);
5441 5442
			break;
		default:
5443 5444
			if (nfs4_async_handle_error(task, calldata->server,
						    NULL, NULL) == -EAGAIN)
5445
				rpc_restart_call_prepare(task);
5446
	}
5447
	nfs_release_seqid(calldata->arg.seqid);
5448 5449
}

T
Trond Myklebust 已提交
5450
static void nfs4_locku_prepare(struct rpc_task *task, void *data)
5451
{
T
Trond Myklebust 已提交
5452
	struct nfs4_unlockdata *calldata = data;
5453

T
Trond Myklebust 已提交
5454
	if (nfs_wait_on_sequence(calldata->arg.seqid, task) != 0)
5455
		goto out_wait;
5456
	nfs4_stateid_copy(&calldata->arg.stateid, &calldata->lsp->ls_stateid);
5457
	if (test_bit(NFS_LOCK_INITIALIZED, &calldata->lsp->ls_flags) == 0) {
5458
		/* Note: exit _without_ running nfs4_locku_done */
5459
		goto out_no_action;
5460
	}
5461
	calldata->timestamp = jiffies;
5462
	if (nfs4_setup_sequence(calldata->server,
5463
				&calldata->arg.seq_args,
5464 5465 5466
				&calldata->res.seq_res,
				task) != 0)
		nfs_release_seqid(calldata->arg.seqid);
5467 5468 5469 5470 5471
	return;
out_no_action:
	task->tk_action = NULL;
out_wait:
	nfs4_sequence_done(task, &calldata->res.seq_res);
5472 5473
}

5474
static const struct rpc_call_ops nfs4_locku_ops = {
T
Trond Myklebust 已提交
5475
	.rpc_call_prepare = nfs4_locku_prepare,
5476
	.rpc_call_done = nfs4_locku_done,
5477
	.rpc_release = nfs4_locku_release_calldata,
5478 5479
};

5480 5481 5482 5483 5484 5485
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;
5486 5487 5488 5489
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_LOCKU],
		.rpc_cred = ctx->cred,
	};
T
Trond Myklebust 已提交
5490 5491
	struct rpc_task_setup task_setup_data = {
		.rpc_client = NFS_CLIENT(lsp->ls_state->inode),
5492
		.rpc_message = &msg,
T
Trond Myklebust 已提交
5493
		.callback_ops = &nfs4_locku_ops,
5494
		.workqueue = nfsiod_workqueue,
T
Trond Myklebust 已提交
5495 5496
		.flags = RPC_TASK_ASYNC,
	};
5497

5498 5499 5500
	nfs4_state_protect(NFS_SERVER(lsp->ls_state->inode)->nfs_client,
		NFS_SP4_MACH_CRED_CLEANUP, &task_setup_data.rpc_client, &msg);

5501 5502 5503 5504 5505
	/* 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;

5506 5507 5508 5509 5510 5511
	data = nfs4_alloc_unlockdata(fl, ctx, lsp, seqid);
	if (data == NULL) {
		nfs_free_seqid(seqid);
		return ERR_PTR(-ENOMEM);
	}

5512
	nfs4_init_sequence(&data->arg.seq_args, &data->res.seq_res, 1);
5513 5514
	msg.rpc_argp = &data->arg;
	msg.rpc_resp = &data->res;
T
Trond Myklebust 已提交
5515 5516
	task_setup_data.callback_data = data;
	return rpc_run_task(&task_setup_data);
5517 5518
}

5519 5520
static int nfs4_proc_unlck(struct nfs4_state *state, int cmd, struct file_lock *request)
{
5521 5522 5523
	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
5524
	struct nfs_seqid *seqid;
L
Linus Torvalds 已提交
5525
	struct nfs4_lock_state *lsp;
5526
	struct rpc_task *task;
5527
	struct nfs_seqid *(*alloc_seqid)(struct nfs_seqid_counter *, gfp_t);
5528
	int status = 0;
5529
	unsigned char fl_flags = request->fl_flags;
5530

5531
	status = nfs4_set_lock_state(state, request);
5532 5533
	/* Unlock _before_ we do the RPC call */
	request->fl_flags |= FL_EXISTS;
5534 5535 5536
	/* Exclude nfs_delegation_claim_locks() */
	mutex_lock(&sp->so_delegreturn_mutex);
	/* Exclude nfs4_reclaim_open_stateid() - note nesting! */
5537 5538 5539
	down_read(&nfsi->rwsem);
	if (do_vfs_lock(request->fl_file, request) == -ENOENT) {
		up_read(&nfsi->rwsem);
5540
		mutex_unlock(&sp->so_delegreturn_mutex);
5541
		goto out;
5542 5543
	}
	up_read(&nfsi->rwsem);
5544
	mutex_unlock(&sp->so_delegreturn_mutex);
5545
	if (status != 0)
5546 5547
		goto out;
	/* Is this a delegated lock? */
5548
	lsp = request->fl_u.nfs4_fl.owner;
5549 5550
	if (test_bit(NFS_LOCK_INITIALIZED, &lsp->ls_flags) == 0)
		goto out;
5551 5552
	alloc_seqid = NFS_SERVER(inode)->nfs_client->cl_mvops->alloc_seqid;
	seqid = alloc_seqid(&lsp->ls_seqid, GFP_KERNEL);
5553
	status = -ENOMEM;
5554
	if (IS_ERR(seqid))
5555
		goto out;
5556
	task = nfs4_do_unlck(request, nfs_file_open_context(request->fl_file), lsp, seqid);
5557 5558
	status = PTR_ERR(task);
	if (IS_ERR(task))
5559
		goto out;
5560
	status = nfs4_wait_for_completion_rpc_task(task);
5561
	rpc_put_task(task);
5562
out:
5563
	request->fl_flags = fl_flags;
5564
	trace_nfs4_unlock(request, state, F_SETLK, status);
L
Linus Torvalds 已提交
5565 5566 5567
	return status;
}

5568 5569 5570 5571 5572 5573
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;
5574
	unsigned long timestamp;
5575 5576
	int rpc_status;
	int cancelled;
5577
	struct nfs_server *server;
5578 5579 5580
};

static struct nfs4_lockdata *nfs4_alloc_lockdata(struct file_lock *fl,
5581 5582
		struct nfs_open_context *ctx, struct nfs4_lock_state *lsp,
		gfp_t gfp_mask)
L
Linus Torvalds 已提交
5583
{
5584 5585
	struct nfs4_lockdata *p;
	struct inode *inode = lsp->ls_state->inode;
L
Linus Torvalds 已提交
5586
	struct nfs_server *server = NFS_SERVER(inode);
5587
	struct nfs_seqid *(*alloc_seqid)(struct nfs_seqid_counter *, gfp_t);
5588

5589
	p = kzalloc(sizeof(*p), gfp_mask);
5590 5591 5592 5593 5594
	if (p == NULL)
		return NULL;

	p->arg.fh = NFS_FH(inode);
	p->arg.fl = &p->fl;
5595
	p->arg.open_seqid = nfs_alloc_seqid(&lsp->ls_state->owner->so_seqid, gfp_mask);
5596
	if (IS_ERR(p->arg.open_seqid))
5597
		goto out_free;
5598 5599
	alloc_seqid = server->nfs_client->cl_mvops->alloc_seqid;
	p->arg.lock_seqid = alloc_seqid(&lsp->ls_seqid, gfp_mask);
5600
	if (IS_ERR(p->arg.lock_seqid))
5601
		goto out_free_seqid;
5602
	p->arg.lock_owner.clientid = server->nfs_client->cl_clientid;
5603
	p->arg.lock_owner.id = lsp->ls_seqid.owner_id;
5604
	p->arg.lock_owner.s_dev = server->s_dev;
5605
	p->res.lock_seqid = p->arg.lock_seqid;
5606
	p->lsp = lsp;
5607
	p->server = server;
5608 5609 5610 5611
	atomic_inc(&lsp->ls_count);
	p->ctx = get_nfs_open_context(ctx);
	memcpy(&p->fl, fl, sizeof(p->fl));
	return p;
5612 5613
out_free_seqid:
	nfs_free_seqid(p->arg.open_seqid);
5614 5615 5616 5617 5618 5619 5620 5621 5622
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;
5623

5624
	dprintk("%s: begin!\n", __func__);
5625
	if (nfs_wait_on_sequence(data->arg.lock_seqid, task) != 0)
5626
		goto out_wait;
5627
	/* Do we need to do an open_to_lock_owner? */
5628
	if (!test_bit(NFS_LOCK_INITIALIZED, &data->lsp->ls_flags)) {
5629
		if (nfs_wait_on_sequence(data->arg.open_seqid, task) != 0) {
5630
			goto out_release_lock_seqid;
5631
		}
5632 5633
		nfs4_stateid_copy(&data->arg.open_stateid,
				&state->open_stateid);
5634
		data->arg.new_lock_owner = 1;
5635
		data->res.open_seqid = data->arg.open_seqid;
5636
	} else {
5637
		data->arg.new_lock_owner = 0;
5638 5639 5640
		nfs4_stateid_copy(&data->arg.lock_stateid,
				&data->lsp->ls_stateid);
	}
5641 5642 5643 5644 5645
	if (!nfs4_valid_open_stateid(state)) {
		data->rpc_status = -EBADF;
		task->tk_action = NULL;
		goto out_release_open_seqid;
	}
5646
	data->timestamp = jiffies;
5647 5648
	if (nfs4_setup_sequence(data->server,
				&data->arg.seq_args,
5649
				&data->res.seq_res,
5650
				task) == 0)
5651
		return;
5652
out_release_open_seqid:
5653 5654 5655
	nfs_release_seqid(data->arg.open_seqid);
out_release_lock_seqid:
	nfs_release_seqid(data->arg.lock_seqid);
5656 5657
out_wait:
	nfs4_sequence_done(task, &data->res.seq_res);
5658
	dprintk("%s: done!, ret = %d\n", __func__, data->rpc_status);
5659 5660
}

5661 5662 5663
static void nfs4_lock_done(struct rpc_task *task, void *calldata)
{
	struct nfs4_lockdata *data = calldata;
5664
	struct nfs4_lock_state *lsp = data->lsp;
5665

5666
	dprintk("%s: begin!\n", __func__);
5667

5668 5669
	if (!nfs4_sequence_done(task, &data->res.seq_res))
		return;
5670

5671
	data->rpc_status = task->tk_status;
5672 5673
	switch (task->tk_status) {
	case 0:
5674 5675
		renew_lease(NFS_SERVER(data->ctx->dentry->d_inode),
				data->timestamp);
5676 5677 5678 5679 5680 5681 5682
		if (data->arg.new_lock) {
			data->fl.fl_flags &= ~(FL_SLEEP | FL_ACCESS);
			if (do_vfs_lock(data->fl.fl_file, &data->fl) < 0) {
				rpc_restart_call_prepare(task);
				break;
			}
		}
5683 5684 5685 5686 5687 5688
		if (data->arg.new_lock_owner != 0) {
			nfs_confirm_seqid(&lsp->ls_seqid, 0);
			nfs4_stateid_copy(&lsp->ls_stateid, &data->res.stateid);
			set_bit(NFS_LOCK_INITIALIZED, &lsp->ls_flags);
		} else if (!nfs4_update_lock_stateid(lsp, &data->res.stateid))
			rpc_restart_call_prepare(task);
5689 5690 5691 5692 5693 5694 5695 5696 5697 5698 5699 5700
		break;
	case -NFS4ERR_BAD_STATEID:
	case -NFS4ERR_OLD_STATEID:
	case -NFS4ERR_STALE_STATEID:
	case -NFS4ERR_EXPIRED:
		if (data->arg.new_lock_owner != 0) {
			if (!nfs4_stateid_match(&data->arg.open_stateid,
						&lsp->ls_state->open_stateid))
				rpc_restart_call_prepare(task);
		} else if (!nfs4_stateid_match(&data->arg.lock_stateid,
						&lsp->ls_stateid))
				rpc_restart_call_prepare(task);
5701
	}
5702
	dprintk("%s: done, ret = %d!\n", __func__, data->rpc_status);
5703 5704 5705 5706 5707 5708
}

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

5709
	dprintk("%s: begin!\n", __func__);
5710
	nfs_free_seqid(data->arg.open_seqid);
5711 5712 5713 5714 5715
	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))
5716
			rpc_put_task_async(task);
5717
		dprintk("%s: cancelling lock!\n", __func__);
5718 5719 5720 5721 5722
	} else
		nfs_free_seqid(data->arg.lock_seqid);
	nfs4_put_lock_state(data->lsp);
	put_nfs_open_context(data->ctx);
	kfree(data);
5723
	dprintk("%s: done!\n", __func__);
5724 5725 5726 5727 5728 5729 5730 5731
}

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,
};

5732 5733 5734 5735 5736
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:
5737
		lsp->ls_seqid.flags &= ~NFS_SEQID_CONFIRMED;
5738
		if (new_lock_owner != 0 ||
5739
		   test_bit(NFS_LOCK_INITIALIZED, &lsp->ls_flags) != 0)
5740
			nfs4_schedule_stateid_recovery(server, lsp->ls_state);
5741 5742 5743
		break;
	case -NFS4ERR_STALE_STATEID:
		lsp->ls_seqid.flags &= ~NFS_SEQID_CONFIRMED;
5744 5745
	case -NFS4ERR_EXPIRED:
		nfs4_schedule_lease_recovery(server->nfs_client);
5746 5747 5748
	};
}

5749
static int _nfs4_do_setlk(struct nfs4_state *state, int cmd, struct file_lock *fl, int recovery_type)
5750 5751 5752
{
	struct nfs4_lockdata *data;
	struct rpc_task *task;
5753 5754 5755 5756
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_LOCK],
		.rpc_cred = state->owner->so_cred,
	};
T
Trond Myklebust 已提交
5757 5758
	struct rpc_task_setup task_setup_data = {
		.rpc_client = NFS_CLIENT(state->inode),
5759
		.rpc_message = &msg,
T
Trond Myklebust 已提交
5760
		.callback_ops = &nfs4_lock_ops,
5761
		.workqueue = nfsiod_workqueue,
T
Trond Myklebust 已提交
5762 5763
		.flags = RPC_TASK_ASYNC,
	};
5764 5765
	int ret;

5766
	dprintk("%s: begin!\n", __func__);
5767
	data = nfs4_alloc_lockdata(fl, nfs_file_open_context(fl->fl_file),
5768 5769
			fl->fl_u.nfs4_fl.owner,
			recovery_type == NFS_LOCK_NEW ? GFP_KERNEL : GFP_NOFS);
5770 5771 5772 5773
	if (data == NULL)
		return -ENOMEM;
	if (IS_SETLKW(cmd))
		data->arg.block = 1;
5774
	nfs4_init_sequence(&data->arg.seq_args, &data->res.seq_res, 1);
5775 5776
	msg.rpc_argp = &data->arg;
	msg.rpc_resp = &data->res;
T
Trond Myklebust 已提交
5777
	task_setup_data.callback_data = data;
5778 5779 5780 5781
	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);
5782 5783
	} else
		data->arg.new_lock = 1;
T
Trond Myklebust 已提交
5784
	task = rpc_run_task(&task_setup_data);
5785
	if (IS_ERR(task))
5786 5787 5788 5789
		return PTR_ERR(task);
	ret = nfs4_wait_for_completion_rpc_task(task);
	if (ret == 0) {
		ret = data->rpc_status;
5790 5791 5792
		if (ret)
			nfs4_handle_setlk_error(data->server, data->lsp,
					data->arg.new_lock_owner, ret);
5793 5794
	} else
		data->cancelled = 1;
5795
	rpc_put_task(task);
5796
	dprintk("%s: done, ret = %d!\n", __func__, ret);
5797
	return ret;
L
Linus Torvalds 已提交
5798 5799 5800 5801
}

static int nfs4_lock_reclaim(struct nfs4_state *state, struct file_lock *request)
{
5802
	struct nfs_server *server = NFS_SERVER(state->inode);
5803 5804 5805
	struct nfs4_exception exception = {
		.inode = state->inode,
	};
5806 5807 5808
	int err;

	do {
5809 5810 5811
		/* Cache the lock if possible... */
		if (test_bit(NFS_DELEGATED_STATE, &state->flags) != 0)
			return 0;
5812
		err = _nfs4_do_setlk(state, F_SETLK, request, NFS_LOCK_RECLAIM);
5813
		trace_nfs4_lock_reclaim(request, state, F_SETLK, err);
5814
		if (err != -NFS4ERR_DELAY)
5815 5816 5817 5818
			break;
		nfs4_handle_exception(server, err, &exception);
	} while (exception.retry);
	return err;
L
Linus Torvalds 已提交
5819 5820 5821 5822
}

static int nfs4_lock_expired(struct nfs4_state *state, struct file_lock *request)
{
5823
	struct nfs_server *server = NFS_SERVER(state->inode);
5824 5825 5826
	struct nfs4_exception exception = {
		.inode = state->inode,
	};
5827 5828
	int err;

5829 5830 5831
	err = nfs4_set_lock_state(state, request);
	if (err != 0)
		return err;
5832
	if (!recover_lost_locks) {
5833 5834 5835
		set_bit(NFS_LOCK_LOST, &request->fl_u.nfs4_fl.owner->ls_flags);
		return 0;
	}
5836
	do {
5837 5838
		if (test_bit(NFS_DELEGATED_STATE, &state->flags) != 0)
			return 0;
5839
		err = _nfs4_do_setlk(state, F_SETLK, request, NFS_LOCK_EXPIRED);
5840
		trace_nfs4_lock_expired(request, state, F_SETLK, err);
5841 5842 5843 5844 5845 5846 5847 5848
		switch (err) {
		default:
			goto out;
		case -NFS4ERR_GRACE:
		case -NFS4ERR_DELAY:
			nfs4_handle_exception(server, err, &exception);
			err = 0;
		}
5849
	} while (exception.retry);
5850
out:
5851
	return err;
L
Linus Torvalds 已提交
5852 5853
}

5854
#if defined(CONFIG_NFS_V4_1)
5855 5856 5857 5858 5859 5860 5861 5862
/**
 * 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.
 */
5863
static int nfs41_check_expired_locks(struct nfs4_state *state)
5864
{
5865
	int status, ret = -NFS4ERR_BAD_STATEID;
5866
	struct nfs4_lock_state *lsp;
5867 5868
	struct nfs_server *server = NFS_SERVER(state->inode);

5869
	list_for_each_entry(lsp, &state->lock_states, ls_locks) {
5870
		if (test_bit(NFS_LOCK_INITIALIZED, &lsp->ls_flags)) {
5871 5872 5873 5874 5875
			struct rpc_cred *cred = lsp->ls_state->owner->so_cred;

			status = nfs41_test_stateid(server,
					&lsp->ls_stateid,
					cred);
5876
			trace_nfs4_test_lock_stateid(state, lsp, status);
5877
			if (status != NFS_OK) {
5878 5879
				/* Free the stateid unless the server
				 * informs us the stateid is unrecognized. */
5880 5881
				if (status != -NFS4ERR_BAD_STATEID)
					nfs41_free_stateid(server,
5882 5883
							&lsp->ls_stateid,
							cred);
5884
				clear_bit(NFS_LOCK_INITIALIZED, &lsp->ls_flags);
5885 5886 5887 5888 5889 5890 5891 5892 5893 5894 5895 5896 5897 5898
				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);
5899 5900 5901
	if (status != NFS_OK)
		status = nfs4_lock_expired(state, request);
	return status;
5902 5903 5904
}
#endif

L
Linus Torvalds 已提交
5905 5906
static int _nfs4_proc_setlk(struct nfs4_state *state, int cmd, struct file_lock *request)
{
5907
	struct nfs_inode *nfsi = NFS_I(state->inode);
5908
	unsigned char fl_flags = request->fl_flags;
5909
	int status = -ENOLCK;
L
Linus Torvalds 已提交
5910

5911 5912 5913
	if ((fl_flags & FL_POSIX) &&
			!test_bit(NFS_STATE_POSIX_LOCKS, &state->flags))
		goto out;
5914 5915 5916 5917
	/* Is this a delegated open? */
	status = nfs4_set_lock_state(state, request);
	if (status != 0)
		goto out;
5918 5919 5920 5921
	request->fl_flags |= FL_ACCESS;
	status = do_vfs_lock(request->fl_file, request);
	if (status < 0)
		goto out;
5922
	down_read(&nfsi->rwsem);
5923 5924 5925
	if (test_bit(NFS_DELEGATED_STATE, &state->flags)) {
		/* Yes: cache locks! */
		/* ...but avoid races with delegation recall... */
5926 5927
		request->fl_flags = fl_flags & ~FL_SLEEP;
		status = do_vfs_lock(request->fl_file, request);
5928
		up_read(&nfsi->rwsem);
5929 5930
		goto out;
	}
5931
	up_read(&nfsi->rwsem);
5932
	status = _nfs4_do_setlk(state, cmd, request, NFS_LOCK_NEW);
5933 5934
out:
	request->fl_flags = fl_flags;
L
Linus Torvalds 已提交
5935 5936 5937 5938 5939
	return status;
}

static int nfs4_proc_setlk(struct nfs4_state *state, int cmd, struct file_lock *request)
{
5940 5941
	struct nfs4_exception exception = {
		.state = state,
5942
		.inode = state->inode,
5943
	};
L
Linus Torvalds 已提交
5944 5945 5946
	int err;

	do {
5947
		err = _nfs4_proc_setlk(state, cmd, request);
5948
		trace_nfs4_set_lock(request, state, cmd, err);
5949 5950
		if (err == -NFS4ERR_DENIED)
			err = -EAGAIN;
L
Linus Torvalds 已提交
5951
		err = nfs4_handle_exception(NFS_SERVER(state->inode),
5952
				err, &exception);
L
Linus Torvalds 已提交
5953 5954 5955 5956 5957 5958 5959 5960 5961 5962 5963 5964 5965
	} 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 */
5966
	ctx = nfs_file_open_context(filp);
L
Linus Torvalds 已提交
5967 5968 5969 5970 5971
	state = ctx->state;

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

5972 5973 5974 5975 5976
	if (IS_GETLK(cmd)) {
		if (state != NULL)
			return nfs4_proc_getlk(state, F_GETLK, request);
		return 0;
	}
L
Linus Torvalds 已提交
5977 5978 5979 5980

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

5981 5982 5983 5984 5985
	if (request->fl_type == F_UNLCK) {
		if (state != NULL)
			return nfs4_proc_unlck(state, cmd, request);
		return 0;
	}
L
Linus Torvalds 已提交
5986

5987 5988
	if (state == NULL)
		return -ENOLCK;
5989 5990 5991 5992
	/*
	 * Don't rely on the VFS having checked the file open mode,
	 * since it won't do this for flock() locks.
	 */
5993
	switch (request->fl_type) {
5994 5995 5996 5997 5998 5999 6000 6001 6002
	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
6003 6004 6005 6006 6007 6008 6009 6010 6011 6012 6013 6014
	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;
}

6015
int nfs4_lock_delegation_recall(struct file_lock *fl, struct nfs4_state *state, const nfs4_stateid *stateid)
6016 6017 6018 6019 6020 6021
{
	struct nfs_server *server = NFS_SERVER(state->inode);
	int err;

	err = nfs4_set_lock_state(state, fl);
	if (err != 0)
6022
		return err;
6023
	err = _nfs4_do_setlk(state, F_SETLK, fl, NFS_LOCK_NEW);
6024
	return nfs4_handle_delegation_recall_error(server, state, stateid, err);
6025
}
6026

6027 6028
struct nfs_release_lockowner_data {
	struct nfs4_lock_state *lsp;
6029
	struct nfs_server *server;
6030
	struct nfs_release_lockowner_args args;
6031
	struct nfs_release_lockowner_res res;
6032
	unsigned long timestamp;
6033 6034
};

6035 6036 6037
static void nfs4_release_lockowner_prepare(struct rpc_task *task, void *calldata)
{
	struct nfs_release_lockowner_data *data = calldata;
6038
	struct nfs_server *server = data->server;
6039 6040
	nfs40_setup_sequence(server->nfs_client->cl_slot_tbl,
			     &data->args.seq_args, &data->res.seq_res, task);
6041
	data->args.lock_owner.clientid = server->nfs_client->cl_clientid;
6042
	data->timestamp = jiffies;
6043 6044 6045 6046 6047
}

static void nfs4_release_lockowner_done(struct rpc_task *task, void *calldata)
{
	struct nfs_release_lockowner_data *data = calldata;
6048 6049
	struct nfs_server *server = data->server;

6050
	nfs40_sequence_done(task, &data->res.seq_res);
6051 6052 6053 6054 6055 6056 6057

	switch (task->tk_status) {
	case 0:
		renew_lease(server, data->timestamp);
		break;
	case -NFS4ERR_STALE_CLIENTID:
	case -NFS4ERR_EXPIRED:
6058 6059
		nfs4_schedule_lease_recovery(server->nfs_client);
		break;
6060 6061
	case -NFS4ERR_LEASE_MOVED:
	case -NFS4ERR_DELAY:
6062 6063
		if (nfs4_async_handle_error(task, server,
					    NULL, NULL) == -EAGAIN)
6064 6065
			rpc_restart_call_prepare(task);
	}
6066 6067
}

6068 6069
static void nfs4_release_lockowner_release(void *calldata)
{
6070
	struct nfs_release_lockowner_data *data = calldata;
6071
	nfs4_free_lock_state(data->server, data->lsp);
6072 6073 6074
	kfree(calldata);
}

6075
static const struct rpc_call_ops nfs4_release_lockowner_ops = {
6076 6077
	.rpc_call_prepare = nfs4_release_lockowner_prepare,
	.rpc_call_done = nfs4_release_lockowner_done,
6078 6079 6080
	.rpc_release = nfs4_release_lockowner_release,
};

6081 6082
static void
nfs4_release_lockowner(struct nfs_server *server, struct nfs4_lock_state *lsp)
6083
{
6084
	struct nfs_release_lockowner_data *data;
6085 6086 6087 6088 6089
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_RELEASE_LOCKOWNER],
	};

	if (server->nfs_client->cl_mvops->minor_version != 0)
6090
		return;
6091

6092 6093
	data = kmalloc(sizeof(*data), GFP_NOFS);
	if (!data)
6094
		return;
6095
	data->lsp = lsp;
6096
	data->server = server;
6097 6098 6099
	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;
6100

6101
	msg.rpc_argp = &data->args;
6102 6103
	msg.rpc_resp = &data->res;
	nfs4_init_sequence(&data->args.seq_args, &data->res.seq_res, 0);
6104
	rpc_call_async(server->client, &msg, 0, &nfs4_release_lockowner_ops, data);
6105 6106
}

6107 6108
#define XATTR_NAME_NFSV4_ACL "system.nfs4_acl"

6109 6110 6111
static int nfs4_xattr_set_nfs4_acl(struct dentry *dentry, const char *key,
				   const void *buf, size_t buflen,
				   int flags, int type)
6112
{
6113 6114
	if (strcmp(key, "") != 0)
		return -EINVAL;
6115

6116
	return nfs4_proc_set_acl(dentry->d_inode, buf, buflen);
6117 6118
}

6119 6120
static int nfs4_xattr_get_nfs4_acl(struct dentry *dentry, const char *key,
				   void *buf, size_t buflen, int type)
6121
{
6122 6123
	if (strcmp(key, "") != 0)
		return -EINVAL;
6124

6125
	return nfs4_proc_get_acl(dentry->d_inode, buf, buflen);
6126 6127
}

6128 6129 6130
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)
6131
{
6132
	size_t len = sizeof(XATTR_NAME_NFSV4_ACL);
6133

6134 6135
	if (!nfs4_server_supports_acls(NFS_SERVER(dentry->d_inode)))
		return 0;
6136 6137 6138

	if (list && len <= list_len)
		memcpy(list, XATTR_NAME_NFSV4_ACL, len);
6139
	return len;
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
#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


6189 6190 6191
/*
 * nfs_fhget will use either the mounted_on_fileid or the fileid
 */
6192 6193
static void nfs_fixup_referral_attributes(struct nfs_fattr *fattr)
{
6194 6195 6196
	if (!(((fattr->valid & NFS_ATTR_FATTR_MOUNTED_ON_FILEID) ||
	       (fattr->valid & NFS_ATTR_FATTR_FILEID)) &&
	      (fattr->valid & NFS_ATTR_FATTR_FSID) &&
6197
	      (fattr->valid & NFS_ATTR_FATTR_V4_LOCATIONS)))
6198 6199 6200
		return;

	fattr->valid |= NFS_ATTR_FATTR_TYPE | NFS_ATTR_FATTR_MODE |
6201
		NFS_ATTR_FATTR_NLINK | NFS_ATTR_FATTR_V4_REFERRAL;
6202 6203 6204 6205
	fattr->mode = S_IFDIR | S_IRUGO | S_IXUGO;
	fattr->nlink = 2;
}

6206 6207 6208 6209
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)
6210 6211
{
	struct nfs_server *server = NFS_SERVER(dir);
6212
	u32 bitmask[3] = {
6213
		[0] = FATTR4_WORD0_FSID | FATTR4_WORD0_FS_LOCATIONS,
6214 6215 6216
	};
	struct nfs4_fs_locations_arg args = {
		.dir_fh = NFS_FH(dir),
6217
		.name = name,
6218 6219 6220
		.page = page,
		.bitmask = bitmask,
	};
B
Benny Halevy 已提交
6221 6222 6223
	struct nfs4_fs_locations_res res = {
		.fs_locations = fs_locations,
	};
6224 6225 6226
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_FS_LOCATIONS],
		.rpc_argp = &args,
B
Benny Halevy 已提交
6227
		.rpc_resp = &res,
6228 6229 6230
	};
	int status;

6231
	dprintk("%s: start\n", __func__);
6232 6233 6234 6235 6236 6237 6238 6239

	/* 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;

6240
	nfs_fattr_init(&fs_locations->fattr);
6241
	fs_locations->server = server;
6242
	fs_locations->nlocations = 0;
6243
	status = nfs4_call_sync(client, server, &msg, &args.seq_args, &res.seq_res, 0);
6244
	dprintk("%s: returned status = %d\n", __func__, status);
6245 6246 6247
	return status;
}

6248 6249 6250 6251
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)
6252 6253 6254 6255
{
	struct nfs4_exception exception = { };
	int err;
	do {
6256 6257 6258 6259
		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,
6260 6261 6262 6263 6264
				&exception);
	} while (exception.retry);
	return err;
}

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 6332 6333 6334 6335 6336 6337 6338 6339 6340 6341 6342 6343 6344 6345 6346 6347 6348 6349 6350 6351 6352 6353 6354 6355 6356 6357 6358 6359 6360 6361 6362 6363 6364 6365 6366 6367 6368 6369 6370 6371 6372 6373 6374 6375 6376 6377 6378 6379 6380 6381 6382 6383 6384 6385 6386 6387 6388 6389 6390 6391 6392 6393 6394 6395 6396 6397 6398 6399 6400 6401 6402 6403 6404 6405 6406 6407 6408 6409 6410 6411 6412 6413 6414 6415
/*
 * 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;
}

6416 6417 6418 6419 6420 6421 6422 6423 6424 6425 6426 6427 6428 6429 6430 6431 6432 6433 6434 6435 6436 6437 6438 6439 6440 6441 6442 6443 6444 6445 6446 6447 6448 6449 6450 6451 6452 6453 6454 6455 6456 6457 6458 6459 6460 6461 6462 6463 6464 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 6492 6493 6494 6495 6496 6497 6498 6499 6500 6501 6502 6503 6504 6505 6506 6507 6508 6509 6510 6511 6512 6513 6514 6515 6516 6517 6518 6519 6520 6521 6522 6523 6524 6525 6526 6527 6528 6529 6530 6531 6532 6533 6534 6535 6536 6537 6538 6539 6540 6541
/*
 * 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;
}

6542
/**
6543 6544 6545 6546 6547
 * 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.
6548
 */
6549
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
6550 6551 6552 6553 6554 6555 6556 6557 6558 6559 6560 6561 6562 6563
{
	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,
	};
6564
	struct rpc_clnt *clnt = NFS_SERVER(dir)->client;
6565
	struct rpc_cred *cred = NULL;
6566 6567 6568

	if (use_integrity) {
		clnt = NFS_SERVER(dir)->nfs_client->cl_rpcclient;
6569 6570
		cred = nfs4_get_clid_cred(NFS_SERVER(dir)->nfs_client);
		msg.rpc_cred = cred;
6571
	}
B
Bryan Schumaker 已提交
6572 6573

	dprintk("NFS call  secinfo %s\n", name->name);
6574 6575 6576 6577

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

6578 6579
	status = nfs4_call_sync(clnt, NFS_SERVER(dir), &msg, &args.seq_args,
				&res.seq_res, 0);
B
Bryan Schumaker 已提交
6580
	dprintk("NFS reply  secinfo: %d\n", status);
6581

6582 6583
	if (cred)
		put_rpccred(cred);
6584

B
Bryan Schumaker 已提交
6585 6586 6587
	return status;
}

6588 6589
int nfs4_proc_secinfo(struct inode *dir, const struct qstr *name,
		      struct nfs4_secinfo_flavors *flavors)
B
Bryan Schumaker 已提交
6590 6591 6592 6593
{
	struct nfs4_exception exception = { };
	int err;
	do {
6594 6595 6596 6597 6598 6599 6600 6601 6602 6603 6604 6605 6606 6607 6608
		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);

6609 6610
		trace_nfs4_secinfo(dir, name, err);
		err = nfs4_handle_exception(NFS_SERVER(dir), err,
B
Bryan Schumaker 已提交
6611 6612 6613 6614 6615
				&exception);
	} while (exception.retry);
	return err;
}

6616
#ifdef CONFIG_NFS_V4_1
6617 6618 6619 6620 6621 6622 6623 6624 6625 6626 6627 6628 6629 6630 6631 6632 6633 6634 6635
/*
 * 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;
}

6636
static bool
6637 6638
nfs41_same_server_scope(struct nfs41_server_scope *a,
			struct nfs41_server_scope *b)
6639 6640 6641 6642 6643 6644 6645 6646
{
	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;
}

6647 6648 6649 6650 6651 6652
/*
 * nfs4_proc_bind_conn_to_session()
 *
 * The 4.1 client currently uses the same TCP connection for the
 * fore and backchannel.
 */
6653
int nfs4_proc_bind_conn_to_session(struct nfs_client *clp, struct rpc_cred *cred)
6654 6655
{
	int status;
6656 6657 6658 6659
	struct nfs41_bind_conn_to_session_args args = {
		.client = clp,
		.dir = NFS4_CDFC4_FORE_OR_BOTH,
	};
6660 6661 6662 6663
	struct nfs41_bind_conn_to_session_res res;
	struct rpc_message msg = {
		.rpc_proc =
			&nfs4_procedures[NFSPROC4_CLNT_BIND_CONN_TO_SESSION],
6664
		.rpc_argp = &args,
6665
		.rpc_resp = &res,
6666
		.rpc_cred = cred,
6667 6668 6669 6670
	};

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

6671 6672 6673
	nfs4_copy_sessionid(&args.sessionid, &clp->cl_session->sess_id);
	if (!(clp->cl_session->flags & SESSION4_BACK_CHAN))
		args.dir = NFS4_CDFC4_FORE;
6674 6675

	status = rpc_call_sync(clp->cl_rpcclient, &msg, RPC_TASK_TIMEOUT);
6676
	trace_nfs4_bind_conn_to_session(clp, status);
6677
	if (status == 0) {
6678
		if (memcmp(res.sessionid.data,
6679 6680 6681
		    clp->cl_session->sess_id.data, NFS4_MAX_SESSIONID_LEN)) {
			dprintk("NFS: %s: Session ID mismatch\n", __func__);
			status = -EIO;
6682
			goto out;
6683
		}
6684
		if ((res.dir & args.dir) != res.dir || res.dir == 0) {
6685 6686 6687
			dprintk("NFS: %s: Unexpected direction from server\n",
				__func__);
			status = -EIO;
6688
			goto out;
6689
		}
6690
		if (res.use_conn_in_rdma_mode != args.use_conn_in_rdma_mode) {
6691 6692 6693
			dprintk("NFS: %s: Server returned RDMA mode = true\n",
				__func__);
			status = -EIO;
6694
			goto out;
6695 6696 6697 6698 6699 6700 6701
		}
	}
out:
	dprintk("<-- %s status= %d\n", __func__, status);
	return status;
}

B
Benny Halevy 已提交
6702
/*
6703 6704
 * 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
6705 6706 6707 6708 6709 6710 6711 6712 6713
 */
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)
6714 6715 6716
	},
	.allow.u.words = {
		[0] = 1 << (OP_CLOSE) |
6717 6718
		      1 << (OP_LOCKU) |
		      1 << (OP_COMMIT),
6719
		[1] = 1 << (OP_SECINFO - 32) |
6720 6721
		      1 << (OP_SECINFO_NO_NAME - 32) |
		      1 << (OP_TEST_STATEID - 32) |
6722 6723
		      1 << (OP_FREE_STATEID - 32) |
		      1 << (OP_WRITE - 32)
6724 6725 6726 6727 6728 6729
	}
};

/*
 * Select the state protection mode for client `clp' given the server results
 * from exchange_id in `sp'.
B
Benny Halevy 已提交
6730
 *
6731 6732 6733 6734 6735 6736 6737 6738 6739 6740 6741 6742 6743 6744 6745 6746 6747 6748 6749 6750 6751 6752 6753 6754 6755 6756 6757 6758 6759 6760 6761 6762 6763 6764 6765 6766 6767 6768 6769 6770 6771 6772 6773 6774 6775 6776 6777 6778 6779 6780 6781 6782
 * 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;
		}
6783 6784 6785 6786 6787 6788

		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);
		}
6789 6790 6791 6792 6793 6794

		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);
		}
6795 6796 6797 6798 6799 6800

		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);
		}
6801 6802 6803 6804 6805 6806 6807 6808 6809 6810

		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);
		}
6811 6812 6813 6814 6815 6816 6817
	}

	return 0;
}

/*
 * _nfs4_proc_exchange_id()
6818
 *
6819
 * Wrapper for EXCHANGE_ID operation.
B
Benny Halevy 已提交
6820
 */
6821 6822
static int _nfs4_proc_exchange_id(struct nfs_client *clp, struct rpc_cred *cred,
	u32 sp4_how)
B
Benny Halevy 已提交
6823 6824 6825
{
	nfs4_verifier verifier;
	struct nfs41_exchange_id_args args = {
6826
		.verifier = &verifier,
B
Benny Halevy 已提交
6827
		.client = clp,
6828
#ifdef CONFIG_NFS_V4_1_MIGRATION
6829
		.flags = EXCHGID4_FLAG_SUPP_MOVED_REFER |
6830 6831 6832 6833 6834 6835
			 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
6836 6837
	};
	struct nfs41_exchange_id_res res = {
6838
		0
B
Benny Halevy 已提交
6839 6840 6841 6842 6843 6844 6845 6846 6847
	};
	int status;
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_EXCHANGE_ID],
		.rpc_argp = &args,
		.rpc_resp = &res,
		.rpc_cred = cred,
	};

6848
	nfs4_init_boot_verifier(clp, &verifier);
6849 6850
	args.id_len = nfs4_init_uniform_client_string(clp, args.id,
							sizeof(args.id));
6851 6852 6853
	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
6854

6855
	res.server_owner = kzalloc(sizeof(struct nfs41_server_owner),
6856
					GFP_NOFS);
6857
	if (unlikely(res.server_owner == NULL)) {
6858 6859 6860
		status = -ENOMEM;
		goto out;
	}
6861

6862
	res.server_scope = kzalloc(sizeof(struct nfs41_server_scope),
6863
					GFP_NOFS);
6864
	if (unlikely(res.server_scope == NULL)) {
6865
		status = -ENOMEM;
6866
		goto out_server_owner;
6867
	}
6868

6869
	res.impl_id = kzalloc(sizeof(struct nfs41_impl_id), GFP_NOFS);
6870
	if (unlikely(res.impl_id == NULL)) {
6871 6872 6873 6874
		status = -ENOMEM;
		goto out_server_scope;
	}

6875 6876 6877 6878 6879 6880 6881 6882 6883 6884 6885 6886 6887 6888 6889 6890
	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;
	}

6891
	status = rpc_call_sync(clp->cl_rpcclient, &msg, RPC_TASK_TIMEOUT);
6892
	trace_nfs4_exchange_id(clp, status);
6893
	if (status == 0)
6894
		status = nfs4_check_cl_exchange_flags(res.flags);
6895

6896 6897 6898
	if (status == 0)
		status = nfs4_sp4_select_mode(clp, &res.state_protect);

6899
	if (status == 0) {
6900
		clp->cl_clientid = res.clientid;
6901 6902 6903 6904 6905
		clp->cl_exchange_flags = res.flags;
		/* Client ID is not confirmed */
		if (!(res.flags & EXCHGID4_FLAG_CONFIRMED_R)) {
			clear_bit(NFS4_SESSION_ESTABLISHED,
					&clp->cl_session->session_state);
6906
			clp->cl_seqid = res.seqid;
6907
		}
6908

6909 6910 6911
		kfree(clp->cl_serverowner);
		clp->cl_serverowner = res.server_owner;
		res.server_owner = NULL;
6912

6913
		/* use the most recent implementation id */
6914 6915
		kfree(clp->cl_implid);
		clp->cl_implid = res.impl_id;
6916

6917
		if (clp->cl_serverscope != NULL &&
6918
		    !nfs41_same_server_scope(clp->cl_serverscope,
6919 6920 6921 6922
					     res.server_scope)) {
			dprintk("%s: server_scope mismatch detected\n",
				__func__);
			set_bit(NFS4CLNT_SERVER_SCOPE_MISMATCH, &clp->cl_state);
6923 6924
			kfree(clp->cl_serverscope);
			clp->cl_serverscope = NULL;
6925 6926
		}

6927
		if (clp->cl_serverscope == NULL) {
6928
			clp->cl_serverscope = res.server_scope;
6929 6930
			goto out;
		}
6931 6932
	} else
		kfree(res.impl_id);
6933

6934 6935
out_server_owner:
	kfree(res.server_owner);
6936
out_server_scope:
6937 6938
	kfree(res.server_scope);
out:
6939
	if (clp->cl_implid != NULL)
6940
		dprintk("NFS reply exchange_id: Server Implementation ID: "
6941
			"domain: %s, name: %s, date: %llu,%u\n",
6942
			clp->cl_implid->domain, clp->cl_implid->name,
6943 6944
			clp->cl_implid->date.seconds,
			clp->cl_implid->date.nseconds);
6945
	dprintk("NFS reply exchange_id: %d\n", status);
B
Benny Halevy 已提交
6946 6947 6948
	return status;
}

6949 6950 6951 6952 6953 6954 6955 6956 6957 6958 6959 6960 6961 6962 6963 6964 6965 6966 6967 6968 6969 6970 6971 6972 6973 6974 6975 6976 6977
/*
 * 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
6978 6979 6980 6981 6982 6983 6984 6985 6986 6987 6988
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);
6989
	trace_nfs4_destroy_clientid(clp, status);
T
Trond Myklebust 已提交
6990
	if (status)
6991
		dprintk("NFS: Got error %d from the server %s on "
T
Trond Myklebust 已提交
6992 6993 6994 6995 6996 6997 6998 6999 7000 7001 7002 7003 7004 7005 7006 7007 7008 7009 7010 7011 7012 7013 7014 7015 7016 7017 7018 7019 7020 7021 7022 7023 7024
			"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;
7025 7026
	if (clp->cl_preserve_clid)
		goto out;
7027
	cred = nfs4_get_clid_cred(clp);
T
Trond Myklebust 已提交
7028 7029 7030 7031 7032 7033 7034 7035 7036 7037 7038 7039
	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
7040 7041 7042 7043 7044 7045 7046 7047 7048 7049 7050 7051 7052 7053 7054
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 */
7055 7056 7057 7058
	nfs41_setup_sequence(data->clp->cl_session,
			&data->args->la_seq_args,
			&data->res->lr_seq_res,
			task);
A
Andy Adamson 已提交
7059 7060 7061 7062 7063 7064 7065 7066 7067 7068 7069 7070 7071
	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__);
7072 7073
	if (!nfs41_sequence_done(task, &data->res->lr_seq_res))
		return;
A
Andy Adamson 已提交
7074 7075 7076 7077 7078 7079
	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;
7080 7081
		/* fall through */
	case -NFS4ERR_RETRY_UNCACHED_REP:
7082
		rpc_restart_call_prepare(task);
A
Andy Adamson 已提交
7083 7084 7085 7086 7087
		return;
	}
	dprintk("<-- %s\n", __func__);
}

7088
static const struct rpc_call_ops nfs4_get_lease_time_ops = {
A
Andy Adamson 已提交
7089 7090 7091 7092 7093 7094 7095 7096 7097 7098 7099 7100 7101 7102 7103 7104 7105 7106 7107 7108 7109 7110 7111 7112 7113
	.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,
7114 7115
		.callback_data = &data,
		.flags = RPC_TASK_TIMEOUT,
A
Andy Adamson 已提交
7116 7117 7118
	};
	int status;

7119
	nfs4_init_sequence(&args.la_seq_args, &res.lr_seq_res, 0);
7120
	nfs4_set_sequence_privileged(&args.la_seq_args);
A
Andy Adamson 已提交
7121 7122 7123 7124 7125 7126 7127 7128 7129 7130 7131 7132 7133 7134
	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
7135 7136 7137 7138 7139 7140 7141 7142 7143 7144 7145
/*
 * 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)
{
7146 7147 7148 7149
	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
7150 7151

	/* Fore channel attributes */
7152 7153
	args->fc_attrs.max_rqst_sz = max_rqst_sz;
	args->fc_attrs.max_resp_sz = max_resp_sz;
A
Andy Adamson 已提交
7154
	args->fc_attrs.max_ops = NFS4_MAX_OPS;
7155
	args->fc_attrs.max_reqs = max_session_slots;
A
Andy Adamson 已提交
7156 7157

	dprintk("%s: Fore Channel : max_rqst_sz=%u max_resp_sz=%u "
7158
		"max_ops=%u max_reqs=%u\n",
A
Andy Adamson 已提交
7159 7160
		__func__,
		args->fc_attrs.max_rqst_sz, args->fc_attrs.max_resp_sz,
7161
		args->fc_attrs.max_ops, args->fc_attrs.max_reqs);
A
Andy Adamson 已提交
7162 7163 7164 7165 7166 7167 7168 7169 7170 7171 7172 7173 7174 7175 7176 7177

	/* 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);
}

7178 7179
static int nfs4_verify_fore_channel_attrs(struct nfs41_create_session_args *args,
		struct nfs41_create_session_res *res)
7180
{
7181
	struct nfs4_channel_attrs *sent = &args->fc_attrs;
7182
	struct nfs4_channel_attrs *rcvd = &res->fc_attrs;
7183 7184 7185 7186 7187 7188 7189 7190 7191 7192 7193 7194 7195

	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;
7196 7197
	if (rcvd->max_reqs > NFS4_MAX_SLOT_TABLE)
		rcvd->max_reqs = NFS4_MAX_SLOT_TABLE;
7198
	return 0;
7199 7200
}

7201 7202
static int nfs4_verify_back_channel_attrs(struct nfs41_create_session_args *args,
		struct nfs41_create_session_res *res)
7203 7204
{
	struct nfs4_channel_attrs *sent = &args->bc_attrs;
7205
	struct nfs4_channel_attrs *rcvd = &res->bc_attrs;
7206

7207 7208
	if (!(res->flags & SESSION4_BACK_CHAN))
		goto out;
7209 7210 7211 7212 7213 7214 7215
	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: */
7216
	if (rcvd->max_ops != sent->max_ops)
7217
		return -EINVAL;
7218
	if (rcvd->max_reqs != sent->max_reqs)
7219
		return -EINVAL;
7220
out:
7221 7222
	return 0;
}
7223 7224

static int nfs4_verify_channel_attrs(struct nfs41_create_session_args *args,
7225
				     struct nfs41_create_session_res *res)
7226
{
7227
	int ret;
7228

7229
	ret = nfs4_verify_fore_channel_attrs(args, res);
7230 7231
	if (ret)
		return ret;
7232 7233 7234 7235 7236 7237 7238
	return nfs4_verify_back_channel_attrs(args, res);
}

static void nfs4_update_session(struct nfs4_session *session,
		struct nfs41_create_session_res *res)
{
	nfs4_copy_sessionid(&session->sess_id, &res->sessionid);
7239 7240 7241
	/* Mark client id and session as being confirmed */
	session->clp->cl_exchange_flags |= EXCHGID4_FLAG_CONFIRMED_R;
	set_bit(NFS4_SESSION_ESTABLISHED, &session->session_state);
7242 7243
	session->flags = res->flags;
	memcpy(&session->fc_attrs, &res->fc_attrs, sizeof(session->fc_attrs));
7244 7245 7246
	if (res->flags & SESSION4_BACK_CHAN)
		memcpy(&session->bc_attrs, &res->bc_attrs,
				sizeof(session->bc_attrs));
7247 7248
}

7249 7250
static int _nfs4_proc_create_session(struct nfs_client *clp,
		struct rpc_cred *cred)
A
Andy Adamson 已提交
7251 7252 7253 7254
{
	struct nfs4_session *session = clp->cl_session;
	struct nfs41_create_session_args args = {
		.client = clp,
7255 7256
		.clientid = clp->cl_clientid,
		.seqid = clp->cl_seqid,
A
Andy Adamson 已提交
7257 7258
		.cb_program = NFS4_CALLBACK,
	};
7259 7260
	struct nfs41_create_session_res res;

A
Andy Adamson 已提交
7261 7262 7263 7264
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_CREATE_SESSION],
		.rpc_argp = &args,
		.rpc_resp = &res,
7265
		.rpc_cred = cred,
A
Andy Adamson 已提交
7266 7267 7268 7269
	};
	int status;

	nfs4_init_channel_attrs(&args);
7270
	args.flags = (SESSION4_PERSIST | SESSION4_BACK_CHAN);
A
Andy Adamson 已提交
7271

7272
	status = rpc_call_sync(session->clp->cl_rpcclient, &msg, RPC_TASK_TIMEOUT);
7273
	trace_nfs4_create_session(clp, status);
A
Andy Adamson 已提交
7274

7275
	if (!status) {
7276
		/* Verify the session's negotiated channel_attrs values */
7277
		status = nfs4_verify_channel_attrs(&args, &res);
A
Andy Adamson 已提交
7278
		/* Increment the clientid slot sequence id */
7279 7280 7281 7282 7283
		if (clp->cl_seqid == res.seqid)
			clp->cl_seqid++;
		if (status)
			goto out;
		nfs4_update_session(session, &res);
A
Andy Adamson 已提交
7284
	}
7285
out:
A
Andy Adamson 已提交
7286 7287 7288 7289 7290 7291 7292 7293
	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.
 */
7294
int nfs4_proc_create_session(struct nfs_client *clp, struct rpc_cred *cred)
A
Andy Adamson 已提交
7295 7296 7297 7298 7299 7300 7301
{
	int status;
	unsigned *ptr;
	struct nfs4_session *session = clp->cl_session;

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

7302
	status = _nfs4_proc_create_session(clp, cred);
A
Andy Adamson 已提交
7303 7304 7305
	if (status)
		goto out;

7306 7307 7308
	/* 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
7309 7310 7311 7312 7313 7314 7315 7316 7317 7318 7319
	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
7320 7321 7322 7323
/*
 * Issue the over-the-wire RPC DESTROY_SESSION.
 * The caller must serialize access to this routine.
 */
7324 7325
int nfs4_proc_destroy_session(struct nfs4_session *session,
		struct rpc_cred *cred)
A
Andy Adamson 已提交
7326
{
7327 7328 7329 7330 7331
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_DESTROY_SESSION],
		.rpc_argp = session,
		.rpc_cred = cred,
	};
A
Andy Adamson 已提交
7332 7333 7334 7335 7336
	int status = 0;

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

	/* session is still being setup */
7337 7338
	if (!test_and_clear_bit(NFS4_SESSION_ESTABLISHED, &session->session_state))
		return 0;
A
Andy Adamson 已提交
7339

7340
	status = rpc_call_sync(session->clp->cl_rpcclient, &msg, RPC_TASK_TIMEOUT);
7341
	trace_nfs4_destroy_session(session->clp, status);
A
Andy Adamson 已提交
7342 7343

	if (status)
7344
		dprintk("NFS: Got error %d from the server on DESTROY_SESSION. "
A
Andy Adamson 已提交
7345 7346 7347 7348 7349 7350
			"Session has been destroyed regardless...\n", status);

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

A
Andy Adamson 已提交
7351 7352 7353
/*
 * Renew the cl_session lease.
 */
7354 7355 7356 7357 7358 7359
struct nfs4_sequence_data {
	struct nfs_client *clp;
	struct nfs4_sequence_args args;
	struct nfs4_sequence_res res;
};

7360 7361
static void nfs41_sequence_release(void *data)
{
7362 7363
	struct nfs4_sequence_data *calldata = data;
	struct nfs_client *clp = calldata->clp;
7364

7365 7366 7367
	if (atomic_read(&clp->cl_count) > 1)
		nfs4_schedule_state_renewal(clp);
	nfs_put_client(clp);
7368
	kfree(calldata);
7369 7370
}

7371 7372 7373 7374 7375 7376 7377
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:
7378
		nfs4_schedule_lease_recovery(clp);
7379 7380 7381 7382
	}
	return 0;
}

7383
static void nfs41_sequence_call_done(struct rpc_task *task, void *data)
A
Andy Adamson 已提交
7384
{
7385 7386
	struct nfs4_sequence_data *calldata = data;
	struct nfs_client *clp = calldata->clp;
A
Andy Adamson 已提交
7387

7388 7389
	if (!nfs41_sequence_done(task, task->tk_msg.rpc_resp))
		return;
A
Andy Adamson 已提交
7390

7391
	trace_nfs4_sequence(clp, task->tk_status);
A
Andy Adamson 已提交
7392 7393
	if (task->tk_status < 0) {
		dprintk("%s ERROR %d\n", __func__, task->tk_status);
7394 7395
		if (atomic_read(&clp->cl_count) == 1)
			goto out;
A
Andy Adamson 已提交
7396

7397 7398
		if (nfs41_sequence_handle_errors(task, clp) == -EAGAIN) {
			rpc_restart_call_prepare(task);
A
Andy Adamson 已提交
7399 7400 7401 7402
			return;
		}
	}
	dprintk("%s rpc_cred %p\n", __func__, task->tk_msg.rpc_cred);
7403
out:
A
Andy Adamson 已提交
7404 7405 7406 7407 7408
	dprintk("<-- %s\n", __func__);
}

static void nfs41_sequence_prepare(struct rpc_task *task, void *data)
{
7409 7410
	struct nfs4_sequence_data *calldata = data;
	struct nfs_client *clp = calldata->clp;
A
Andy Adamson 已提交
7411 7412 7413 7414 7415 7416
	struct nfs4_sequence_args *args;
	struct nfs4_sequence_res *res;

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

7417
	nfs41_setup_sequence(clp->cl_session, args, res, task);
A
Andy Adamson 已提交
7418 7419 7420 7421 7422
}

static const struct rpc_call_ops nfs41_sequence_ops = {
	.rpc_call_done = nfs41_sequence_call_done,
	.rpc_call_prepare = nfs41_sequence_prepare,
7423
	.rpc_release = nfs41_sequence_release,
A
Andy Adamson 已提交
7424 7425
};

7426 7427 7428
static struct rpc_task *_nfs41_proc_sequence(struct nfs_client *clp,
		struct rpc_cred *cred,
		bool is_privileged)
A
Andy Adamson 已提交
7429
{
7430
	struct nfs4_sequence_data *calldata;
A
Andy Adamson 已提交
7431 7432 7433 7434
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_SEQUENCE],
		.rpc_cred = cred,
	};
7435 7436 7437
	struct rpc_task_setup task_setup_data = {
		.rpc_client = clp->cl_rpcclient,
		.rpc_message = &msg,
7438
		.callback_ops = &nfs41_sequence_ops,
7439
		.flags = RPC_TASK_ASYNC | RPC_TASK_TIMEOUT,
7440
	};
A
Andy Adamson 已提交
7441

7442
	if (!atomic_inc_not_zero(&clp->cl_count))
7443
		return ERR_PTR(-EIO);
7444
	calldata = kzalloc(sizeof(*calldata), GFP_NOFS);
7445
	if (calldata == NULL) {
7446
		nfs_put_client(clp);
7447
		return ERR_PTR(-ENOMEM);
A
Andy Adamson 已提交
7448
	}
7449
	nfs4_init_sequence(&calldata->args, &calldata->res, 0);
7450 7451
	if (is_privileged)
		nfs4_set_sequence_privileged(&calldata->args);
7452 7453 7454
	msg.rpc_argp = &calldata->args;
	msg.rpc_resp = &calldata->res;
	calldata->clp = clp;
7455
	task_setup_data.callback_data = calldata;
A
Andy Adamson 已提交
7456

7457 7458 7459
	return rpc_run_task(&task_setup_data);
}

7460
static int nfs41_proc_async_sequence(struct nfs_client *clp, struct rpc_cred *cred, unsigned renew_flags)
7461 7462 7463 7464
{
	struct rpc_task *task;
	int ret = 0;

7465
	if ((renew_flags & NFS4_RENEW_TIMEOUT) == 0)
7466
		return -EAGAIN;
7467
	task = _nfs41_proc_sequence(clp, cred, false);
7468 7469 7470
	if (IS_ERR(task))
		ret = PTR_ERR(task);
	else
7471
		rpc_put_task_async(task);
7472 7473 7474 7475 7476 7477 7478 7479 7480
	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;

7481
	task = _nfs41_proc_sequence(clp, cred, true);
7482 7483 7484 7485 7486
	if (IS_ERR(task)) {
		ret = PTR_ERR(task);
		goto out;
	}
	ret = rpc_wait_for_completion_task(task);
7487 7488 7489 7490 7491
	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);
7492
		ret = task->tk_status;
7493
	}
7494 7495 7496 7497
	rpc_put_task(task);
out:
	dprintk("<-- %s status=%d\n", __func__, ret);
	return ret;
A
Andy Adamson 已提交
7498 7499
}

7500 7501 7502 7503 7504 7505 7506 7507 7508 7509
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;

7510 7511 7512 7513
	nfs41_setup_sequence(calldata->clp->cl_session,
			&calldata->arg.seq_args,
			&calldata->res.seq_res,
			task);
7514 7515
}

7516 7517 7518 7519 7520 7521 7522 7523 7524
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);
7525 7526
		/* fall through */
	case -NFS4ERR_RETRY_UNCACHED_REP:
7527 7528
		return -EAGAIN;
	default:
7529
		nfs4_schedule_lease_recovery(clp);
7530 7531 7532 7533
	}
	return 0;
}

7534 7535 7536 7537 7538 7539 7540
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__);
7541 7542
	if (!nfs41_sequence_done(task, res))
		return;
7543

7544
	trace_nfs4_reclaim_complete(clp, task->tk_status);
7545 7546 7547 7548
	if (nfs41_reclaim_complete_handle_errors(task, clp) == -EAGAIN) {
		rpc_restart_call_prepare(task);
		return;
	}
7549 7550 7551 7552 7553 7554 7555 7556 7557 7558 7559 7560 7561 7562 7563 7564 7565 7566 7567
	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.
 */
7568 7569
static int nfs41_proc_reclaim_complete(struct nfs_client *clp,
		struct rpc_cred *cred)
7570 7571 7572 7573 7574
{
	struct nfs4_reclaim_complete_data *calldata;
	struct rpc_task *task;
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_RECLAIM_COMPLETE],
7575
		.rpc_cred = cred,
7576 7577 7578 7579 7580 7581 7582 7583 7584 7585
	};
	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__);
7586
	calldata = kzalloc(sizeof(*calldata), GFP_NOFS);
7587 7588 7589 7590 7591
	if (calldata == NULL)
		goto out;
	calldata->clp = clp;
	calldata->arg.one_fs = 0;

7592
	nfs4_init_sequence(&calldata->arg.seq_args, &calldata->res.seq_res, 0);
7593
	nfs4_set_sequence_privileged(&calldata->arg.seq_args);
7594 7595 7596 7597
	msg.rpc_argp = &calldata->arg;
	msg.rpc_resp = &calldata->res;
	task_setup_data.callback_data = calldata;
	task = rpc_run_task(&task_setup_data);
7598
	if (IS_ERR(task)) {
7599
		status = PTR_ERR(task);
7600 7601
		goto out;
	}
7602 7603 7604
	status = nfs4_wait_for_completion_rpc_task(task);
	if (status == 0)
		status = task->tk_status;
7605
	rpc_put_task(task);
7606
	return 0;
7607 7608 7609 7610
out:
	dprintk("<-- %s status=%d\n", __func__, status);
	return status;
}
7611 7612 7613 7614 7615

static void
nfs4_layoutget_prepare(struct rpc_task *task, void *calldata)
{
	struct nfs4_layoutget *lgp = calldata;
F
Fred Isaman 已提交
7616
	struct nfs_server *server = NFS_SERVER(lgp->args.inode);
7617
	struct nfs4_session *session = nfs4_get_session(server);
7618 7619

	dprintk("--> %s\n", __func__);
F
Fred Isaman 已提交
7620 7621 7622 7623 7624
	/* 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.
	 */
7625
	if (nfs41_setup_sequence(session, &lgp->args.seq_args,
7626
				&lgp->res.seq_res, task))
7627
		return;
7628 7629
	if (pnfs_choose_layoutget_stateid(&lgp->args.stateid,
					  NFS_I(lgp->args.inode)->layout,
7630
					  &lgp->args.range,
7631 7632 7633
					  lgp->args.ctx->state)) {
		rpc_exit(task, NFS4_OK);
	}
7634 7635 7636 7637 7638
}

static void nfs4_layoutget_done(struct rpc_task *task, void *calldata)
{
	struct nfs4_layoutget *lgp = calldata;
7639 7640 7641 7642
	struct inode *inode = lgp->args.inode;
	struct nfs_server *server = NFS_SERVER(inode);
	struct pnfs_layout_hdr *lo;
	struct nfs4_state *state = NULL;
7643
	unsigned long timeo, now, giveup;
7644

7645
	dprintk("--> %s tk_status => %d\n", __func__, -task->tk_status);
7646

7647
	if (!nfs41_sequence_done(task, &lgp->res.seq_res))
7648
		goto out;
7649 7650 7651

	switch (task->tk_status) {
	case 0:
7652
		goto out;
7653 7654 7655 7656
	/*
	 * NFS4ERR_LAYOUTTRYLATER is a conflict with another client
	 * (or clients) writing to the same RAID stripe
	 */
7657
	case -NFS4ERR_LAYOUTTRYLATER:
7658 7659 7660 7661
	/*
	 * NFS4ERR_RECALLCONFLICT is when conflict with self (must recall
	 * existing layout before getting a new one).
	 */
7662
	case -NFS4ERR_RECALLCONFLICT:
7663 7664
		timeo = rpc_get_timeout(task->tk_client);
		giveup = lgp->args.timestamp + timeo;
7665 7666 7667 7668 7669 7670 7671 7672 7673 7674 7675 7676 7677 7678 7679 7680 7681 7682 7683 7684
		now = jiffies;
		if (time_after(giveup, now)) {
			unsigned long delay;

			/* Delay for:
			 * - Not less then NFS4_POLL_RETRY_MIN.
			 * - One last time a jiffie before we give up
			 * - exponential backoff (time_now minus start_attempt)
			 */
			delay = max_t(unsigned long, NFS4_POLL_RETRY_MIN,
				    min((giveup - now - 1),
					now - lgp->args.timestamp));

			dprintk("%s: NFS4ERR_RECALLCONFLICT waiting %lu\n",
				__func__, delay);
			rpc_delay(task, delay);
			task->tk_status = 0;
			rpc_restart_call_prepare(task);
			goto out; /* Do not call nfs4_async_handle_error() */
		}
7685 7686 7687 7688 7689 7690 7691 7692 7693 7694 7695 7696
		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);

7697 7698 7699 7700
			/*
			 * Mark the bad layout state as invalid, then retry
			 * with the current stateid.
			 */
7701 7702 7703
			pnfs_mark_matching_lsegs_invalid(lo, &head, NULL);
			spin_unlock(&inode->i_lock);
			pnfs_free_lseg_list(&head);
7704 7705 7706
	
			task->tk_status = 0;
			rpc_restart_call_prepare(task);
7707 7708
		}
	}
7709
	if (nfs4_async_handle_error(task, server, state, NULL) == -EAGAIN)
7710 7711
		rpc_restart_call_prepare(task);
out:
7712 7713 7714
	dprintk("<-- %s\n", __func__);
}

7715 7716 7717 7718 7719 7720 7721 7722 7723 7724 7725 7726 7727 7728 7729 7730 7731 7732 7733 7734 7735 7736 7737 7738 7739 7740 7741 7742 7743 7744 7745 7746 7747 7748 7749 7750 7751 7752 7753 7754 7755 7756 7757 7758
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;
}

7759 7760 7761
static void nfs4_layoutget_release(void *calldata)
{
	struct nfs4_layoutget *lgp = calldata;
7762 7763
	struct inode *inode = lgp->args.inode;
	struct nfs_server *server = NFS_SERVER(inode);
7764
	size_t max_pages = max_response_pages(server);
7765 7766

	dprintk("--> %s\n", __func__);
7767
	nfs4_free_pages(lgp->args.layout.pages, max_pages);
7768
	pnfs_put_layout_hdr(NFS_I(inode)->layout);
7769 7770 7771 7772 7773 7774 7775 7776 7777 7778 7779
	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,
};

7780 7781
struct pnfs_layout_segment *
nfs4_proc_layoutget(struct nfs4_layoutget *lgp, gfp_t gfp_flags)
7782
{
7783 7784
	struct inode *inode = lgp->args.inode;
	struct nfs_server *server = NFS_SERVER(inode);
7785
	size_t max_pages = max_response_pages(server);
7786 7787 7788 7789 7790
	struct rpc_task *task;
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_LAYOUTGET],
		.rpc_argp = &lgp->args,
		.rpc_resp = &lgp->res,
7791
		.rpc_cred = lgp->cred,
7792 7793 7794 7795 7796 7797 7798 7799
	};
	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,
	};
7800
	struct pnfs_layout_segment *lseg = NULL;
7801 7802 7803 7804
	int status = 0;

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

7805 7806 7807
	/* nfs4_layoutget_release calls pnfs_put_layout_hdr */
	pnfs_get_layout_hdr(NFS_I(inode)->layout);

7808 7809 7810
	lgp->args.layout.pages = nfs4_alloc_pages(max_pages, gfp_flags);
	if (!lgp->args.layout.pages) {
		nfs4_layoutget_release(lgp);
7811
		return ERR_PTR(-ENOMEM);
7812 7813
	}
	lgp->args.layout.pglen = max_pages * PAGE_SIZE;
7814
	lgp->args.timestamp = jiffies;
7815

7816
	lgp->res.layoutp = &lgp->args.layout;
7817
	lgp->res.seq_res.sr_slot = NULL;
7818
	nfs4_init_sequence(&lgp->args.seq_args, &lgp->res.seq_res, 0);
7819

7820 7821
	task = rpc_run_task(&task_setup_data);
	if (IS_ERR(task))
7822
		return ERR_CAST(task);
7823
	status = nfs4_wait_for_completion_rpc_task(task);
F
Fred Isaman 已提交
7824 7825
	if (status == 0)
		status = task->tk_status;
7826 7827 7828 7829
	trace_nfs4_layoutget(lgp->args.ctx,
			&lgp->args.range,
			&lgp->res.range,
			status);
7830 7831
	/* if layoutp->len is 0, nfs4_layoutget_prepare called rpc_exit */
	if (status == 0 && lgp->res.layoutp->len)
7832
		lseg = pnfs_layout_process(lgp);
7833 7834
	rpc_put_task(task);
	dprintk("<-- %s status=%d\n", __func__, status);
7835 7836 7837
	if (status)
		return ERR_PTR(status);
	return lseg;
7838 7839
}

B
Benny Halevy 已提交
7840 7841 7842 7843 7844 7845
static void
nfs4_layoutreturn_prepare(struct rpc_task *task, void *calldata)
{
	struct nfs4_layoutreturn *lrp = calldata;

	dprintk("--> %s\n", __func__);
7846 7847 7848 7849
	nfs41_setup_sequence(lrp->clp->cl_session,
			&lrp->args.seq_args,
			&lrp->res.seq_res,
			task);
B
Benny Halevy 已提交
7850 7851 7852 7853 7854 7855 7856 7857 7858
}

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

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

7859
	if (!nfs41_sequence_done(task, &lrp->res.seq_res))
B
Benny Halevy 已提交
7860 7861 7862
		return;

	server = NFS_SERVER(lrp->args.inode);
7863 7864 7865 7866 7867 7868
	switch (task->tk_status) {
	default:
		task->tk_status = 0;
	case 0:
		break;
	case -NFS4ERR_DELAY:
7869
		if (nfs4_async_handle_error(task, server, NULL, NULL) != -EAGAIN)
7870
			break;
7871
		rpc_restart_call_prepare(task);
B
Benny Halevy 已提交
7872 7873 7874 7875 7876 7877 7878 7879
		return;
	}
	dprintk("<-- %s\n", __func__);
}

static void nfs4_layoutreturn_release(void *calldata)
{
	struct nfs4_layoutreturn *lrp = calldata;
7880
	struct pnfs_layout_hdr *lo = lrp->args.layout;
B
Benny Halevy 已提交
7881 7882

	dprintk("--> %s\n", __func__);
7883 7884 7885
	spin_lock(&lo->plh_inode->i_lock);
	if (lrp->res.lrs_present)
		pnfs_set_layout_stateid(lo, &lrp->res.stateid, true);
7886
	pnfs_clear_layoutreturn_waitbit(lo);
7887 7888
	clear_bit(NFS_LAYOUT_RETURN_BEFORE_CLOSE, &lo->plh_flags);
	rpc_wake_up(&NFS_SERVER(lo->plh_inode)->roc_rpcwaitq);
7889 7890
	lo->plh_block_lgets--;
	spin_unlock(&lo->plh_inode->i_lock);
7891
	pnfs_put_layout_hdr(lrp->args.layout);
7892
	nfs_iput_and_deactive(lrp->inode);
B
Benny Halevy 已提交
7893 7894 7895 7896 7897 7898 7899 7900 7901 7902
	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,
};

7903
int nfs4_proc_layoutreturn(struct nfs4_layoutreturn *lrp, bool sync)
B
Benny Halevy 已提交
7904 7905 7906 7907 7908 7909
{
	struct rpc_task *task;
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_LAYOUTRETURN],
		.rpc_argp = &lrp->args,
		.rpc_resp = &lrp->res,
7910
		.rpc_cred = lrp->cred,
B
Benny Halevy 已提交
7911 7912
	};
	struct rpc_task_setup task_setup_data = {
7913
		.rpc_client = NFS_SERVER(lrp->args.inode)->client,
B
Benny Halevy 已提交
7914 7915 7916 7917
		.rpc_message = &msg,
		.callback_ops = &nfs4_layoutreturn_call_ops,
		.callback_data = lrp,
	};
7918
	int status = 0;
B
Benny Halevy 已提交
7919 7920

	dprintk("--> %s\n", __func__);
7921 7922 7923 7924 7925 7926 7927 7928
	if (!sync) {
		lrp->inode = nfs_igrab_and_active(lrp->args.inode);
		if (!lrp->inode) {
			nfs4_layoutreturn_release(lrp);
			return -EAGAIN;
		}
		task_setup_data.flags |= RPC_TASK_ASYNC;
	}
7929
	nfs4_init_sequence(&lrp->args.seq_args, &lrp->res.seq_res, 1);
B
Benny Halevy 已提交
7930 7931 7932
	task = rpc_run_task(&task_setup_data);
	if (IS_ERR(task))
		return PTR_ERR(task);
7933 7934
	if (sync)
		status = task->tk_status;
7935
	trace_nfs4_layoutreturn(lrp->args.inode, status);
B
Benny Halevy 已提交
7936 7937 7938 7939 7940
	dprintk("<-- %s status=%d\n", __func__, status);
	rpc_put_task(task);
	return status;
}

7941
static int
7942 7943 7944
_nfs4_proc_getdeviceinfo(struct nfs_server *server,
		struct pnfs_device *pdev,
		struct rpc_cred *cred)
7945 7946 7947
{
	struct nfs4_getdeviceinfo_args args = {
		.pdev = pdev,
7948 7949
		.notify_types = NOTIFY_DEVICEID4_CHANGE |
			NOTIFY_DEVICEID4_DELETE,
7950 7951 7952 7953 7954 7955 7956 7957
	};
	struct nfs4_getdeviceinfo_res res = {
		.pdev = pdev,
	};
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_GETDEVICEINFO],
		.rpc_argp = &args,
		.rpc_resp = &res,
7958
		.rpc_cred = cred,
7959 7960 7961 7962
	};
	int status;

	dprintk("--> %s\n", __func__);
7963
	status = nfs4_call_sync(server->client, server, &msg, &args.seq_args, &res.seq_res, 0);
7964 7965
	if (res.notification & ~args.notify_types)
		dprintk("%s: unsupported notification\n", __func__);
7966 7967
	if (res.notification != args.notify_types)
		pdev->nocache = 1;
7968

7969 7970 7971 7972 7973
	dprintk("<-- %s status=%d\n", __func__, status);

	return status;
}

7974 7975 7976
int nfs4_proc_getdeviceinfo(struct nfs_server *server,
		struct pnfs_device *pdev,
		struct rpc_cred *cred)
7977 7978 7979 7980 7981 7982
{
	struct nfs4_exception exception = { };
	int err;

	do {
		err = nfs4_handle_exception(server,
7983
					_nfs4_proc_getdeviceinfo(server, pdev, cred),
7984 7985 7986 7987 7988 7989
					&exception);
	} while (exception.retry);
	return err;
}
EXPORT_SYMBOL_GPL(nfs4_proc_getdeviceinfo);

A
Andy Adamson 已提交
7990 7991 7992 7993
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);
7994
	struct nfs4_session *session = nfs4_get_session(server);
A
Andy Adamson 已提交
7995

7996 7997 7998 7999
	nfs41_setup_sequence(session,
			&data->args.seq_args,
			&data->res.seq_res,
			task);
A
Andy Adamson 已提交
8000 8001 8002 8003 8004 8005 8006 8007
}

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);

8008
	if (!nfs41_sequence_done(task, &data->res.seq_res))
A
Andy Adamson 已提交
8009 8010 8011
		return;

	switch (task->tk_status) { /* Just ignore these failures */
8012 8013 8014 8015
	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
8016
		task->tk_status = 0;
8017 8018 8019
	case 0:
		break;
	default:
8020
		if (nfs4_async_handle_error(task, server, NULL, NULL) == -EAGAIN) {
8021 8022 8023 8024
			rpc_restart_call_prepare(task);
			return;
		}
	}
A
Andy Adamson 已提交
8025 8026 8027 8028 8029 8030
}

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

A
Andy Adamson 已提交
8031
	pnfs_cleanup_layoutcommit(data);
8032 8033
	nfs_post_op_update_inode_force_wcc(data->args.inode,
					   data->res.fattr);
A
Andy Adamson 已提交
8034
	put_rpccred(data->cred);
8035
	nfs_iput_and_deactive(data->inode);
A
Andy Adamson 已提交
8036 8037 8038 8039 8040 8041 8042 8043 8044 8045
	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
8046
nfs4_proc_layoutcommit(struct nfs4_layoutcommit_data *data, bool sync)
A
Andy Adamson 已提交
8047 8048 8049 8050 8051 8052 8053 8054 8055 8056 8057 8058 8059 8060 8061 8062 8063 8064 8065 8066 8067 8068 8069
{
	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,
	};
	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);

8070 8071 8072 8073 8074 8075 8076 8077
	if (!sync) {
		data->inode = nfs_igrab_and_active(data->args.inode);
		if (data->inode == NULL) {
			nfs4_layoutcommit_release(data);
			return -EAGAIN;
		}
		task_setup_data.flags = RPC_TASK_ASYNC;
	}
8078
	nfs4_init_sequence(&data->args.seq_args, &data->res.seq_res, 1);
A
Andy Adamson 已提交
8079 8080 8081
	task = rpc_run_task(&task_setup_data);
	if (IS_ERR(task))
		return PTR_ERR(task);
8082 8083
	if (sync)
		status = task->tk_status;
8084
	trace_nfs4_layoutcommit(data->args.inode, status);
A
Andy Adamson 已提交
8085 8086 8087 8088
	dprintk("%s: status %d\n", __func__, status);
	rpc_put_task(task);
	return status;
}
8089

8090 8091 8092 8093
/**
 * Use the state managment nfs_client cl_rpcclient, which uses krb5i (if
 * possible) as per RFC3530bis and RFC5661 Security Considerations sections
 */
8094 8095
static int
_nfs41_proc_secinfo_no_name(struct nfs_server *server, struct nfs_fh *fhandle,
8096 8097
		    struct nfs_fsinfo *info,
		    struct nfs4_secinfo_flavors *flavors, bool use_integrity)
8098 8099 8100 8101 8102 8103 8104 8105 8106 8107 8108 8109
{
	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,
	};
8110
	struct rpc_clnt *clnt = server->client;
8111
	struct rpc_cred *cred = NULL;
8112 8113 8114 8115
	int status;

	if (use_integrity) {
		clnt = server->nfs_client->cl_rpcclient;
8116 8117
		cred = nfs4_get_clid_cred(server->nfs_client);
		msg.rpc_cred = cred;
8118 8119 8120 8121 8122 8123 8124
	}

	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);

8125 8126
	if (cred)
		put_rpccred(cred);
8127 8128

	return status;
8129 8130 8131 8132 8133 8134 8135 8136 8137
}

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 {
8138 8139 8140 8141 8142 8143 8144 8145 8146 8147 8148 8149 8150 8151 8152 8153 8154 8155
		/* 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);

8156 8157 8158
		switch (err) {
		case 0:
		case -NFS4ERR_WRONGSEC:
8159
		case -ENOTSUPP:
8160
			goto out;
8161 8162 8163 8164
		default:
			err = nfs4_handle_exception(server, err, &exception);
		}
	} while (exception.retry);
8165
out:
8166 8167 8168 8169 8170 8171 8172 8173 8174
	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;
8175
	rpc_authflavor_t flavor = RPC_AUTH_MAXFLAVOR;
8176
	struct nfs4_secinfo_flavors *flavors;
8177 8178
	struct nfs4_secinfo4 *secinfo;
	int i;
8179 8180 8181 8182 8183 8184 8185 8186 8187 8188 8189 8190 8191 8192

	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
	 */
8193
	if (err == -NFS4ERR_WRONGSEC || err == -ENOTSUPP) {
8194 8195 8196 8197 8198 8199
		err = nfs4_find_root_sec(server, fhandle, info);
		goto out_freepage;
	}
	if (err)
		goto out_freepage;

8200 8201 8202 8203 8204 8205 8206 8207 8208 8209 8210 8211 8212 8213 8214
	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;
		}

8215 8216 8217
		if (!nfs_auth_info_match(&server->auth_info, flavor))
			flavor = RPC_AUTH_MAXFLAVOR;

8218 8219 8220 8221 8222 8223 8224 8225 8226 8227
		if (flavor != RPC_AUTH_MAXFLAVOR) {
			err = nfs4_lookup_root_sec(server, fhandle,
						   info, flavor);
			if (!err)
				break;
		}
	}

	if (flavor == RPC_AUTH_MAXFLAVOR)
		err = -EPERM;
8228 8229 8230 8231 8232 8233 8234 8235

out_freepage:
	put_page(page);
	if (err == -EACCES)
		return -EPERM;
out:
	return err;
}
8236

8237 8238 8239
static int _nfs41_test_stateid(struct nfs_server *server,
		nfs4_stateid *stateid,
		struct rpc_cred *cred)
B
Bryan Schumaker 已提交
8240 8241 8242
{
	int status;
	struct nfs41_test_stateid_args args = {
8243
		.stateid = stateid,
B
Bryan Schumaker 已提交
8244 8245 8246 8247 8248 8249
	};
	struct nfs41_test_stateid_res res;
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_TEST_STATEID],
		.rpc_argp = &args,
		.rpc_resp = &res,
8250
		.rpc_cred = cred,
B
Bryan Schumaker 已提交
8251
	};
8252 8253 8254 8255
	struct rpc_clnt *rpc_client = server->client;

	nfs4_state_protect(server->nfs_client, NFS_SP4_MACH_CRED_STATEID,
		&rpc_client, &msg);
8256

8257
	dprintk("NFS call  test_stateid %p\n", stateid);
8258
	nfs4_init_sequence(&args.seq_args, &res.seq_res, 0);
8259
	nfs4_set_sequence_privileged(&args.seq_args);
8260
	status = nfs4_call_sync_sequence(rpc_client, server, &msg,
8261
			&args.seq_args, &res.seq_res);
8262 8263
	if (status != NFS_OK) {
		dprintk("NFS reply test_stateid: failed, %d\n", status);
8264
		return status;
8265 8266
	}
	dprintk("NFS reply test_stateid: succeeded, %d\n", -res.status);
8267
	return -res.status;
B
Bryan Schumaker 已提交
8268 8269
}

8270 8271 8272 8273 8274
/**
 * nfs41_test_stateid - perform a TEST_STATEID operation
 *
 * @server: server / transport on which to perform the operation
 * @stateid: state ID to test
8275
 * @cred: credential
8276 8277 8278 8279 8280
 *
 * 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.
 */
8281 8282 8283
static int nfs41_test_stateid(struct nfs_server *server,
		nfs4_stateid *stateid,
		struct rpc_cred *cred)
B
Bryan Schumaker 已提交
8284 8285 8286 8287
{
	struct nfs4_exception exception = { };
	int err;
	do {
8288
		err = _nfs41_test_stateid(server, stateid, cred);
8289 8290 8291
		if (err != -NFS4ERR_DELAY)
			break;
		nfs4_handle_exception(server, err, &exception);
B
Bryan Schumaker 已提交
8292 8293 8294
	} while (exception.retry);
	return err;
}
B
Bryan Schumaker 已提交
8295

8296 8297 8298
struct nfs_free_stateid_data {
	struct nfs_server *server;
	struct nfs41_free_stateid_args args;
B
Bryan Schumaker 已提交
8299
	struct nfs41_free_stateid_res res;
8300 8301 8302 8303 8304 8305 8306 8307 8308 8309 8310 8311 8312 8313 8314 8315 8316 8317 8318
};

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:
8319
		if (nfs4_async_handle_error(task, data->server, NULL, NULL) == -EAGAIN)
8320 8321 8322 8323 8324 8325 8326 8327 8328
			rpc_restart_call_prepare(task);
	}
}

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

8329
static const struct rpc_call_ops nfs41_free_stateid_ops = {
8330 8331 8332 8333 8334 8335 8336
	.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,
8337
		struct rpc_cred *cred,
8338 8339
		bool privileged)
{
B
Bryan Schumaker 已提交
8340 8341
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_FREE_STATEID],
8342
		.rpc_cred = cred,
B
Bryan Schumaker 已提交
8343
	};
8344 8345 8346 8347 8348 8349 8350
	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
8351

8352 8353 8354
	nfs4_state_protect(server->nfs_client, NFS_SP4_MACH_CRED_STATEID,
		&task_setup.rpc_client, &msg);

8355
	dprintk("NFS call  free_stateid %p\n", stateid);
8356 8357 8358 8359 8360 8361 8362 8363 8364 8365
	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;
8366
	nfs4_init_sequence(&data->args.seq_args, &data->res.seq_res, 0);
8367 8368 8369 8370
	if (privileged)
		nfs4_set_sequence_privileged(&data->args.seq_args);

	return rpc_run_task(&task_setup);
B
Bryan Schumaker 已提交
8371 8372
}

8373 8374 8375 8376 8377
/**
 * nfs41_free_stateid - perform a FREE_STATEID operation
 *
 * @server: server / transport on which to perform the operation
 * @stateid: state ID to release
8378
 * @cred: credential
8379 8380 8381 8382
 *
 * Returns NFS_OK if the server freed "stateid".  Otherwise a
 * negative NFS4ERR value is returned.
 */
8383 8384 8385
static int nfs41_free_stateid(struct nfs_server *server,
		nfs4_stateid *stateid,
		struct rpc_cred *cred)
B
Bryan Schumaker 已提交
8386
{
8387 8388 8389
	struct rpc_task *task;
	int ret;

8390
	task = _nfs41_free_stateid(server, stateid, cred, true);
8391 8392 8393 8394 8395 8396 8397
	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
8398
}
8399

8400 8401
static void
nfs41_free_lock_state(struct nfs_server *server, struct nfs4_lock_state *lsp)
8402 8403
{
	struct rpc_task *task;
8404
	struct rpc_cred *cred = lsp->ls_state->owner->so_cred;
8405

8406
	task = _nfs41_free_stateid(server, &lsp->ls_stateid, cred, false);
8407 8408
	nfs4_free_lock_state(server, lsp);
	if (IS_ERR(task))
8409
		return;
8410 8411 8412
	rpc_put_task(task);
}

8413 8414 8415
static bool nfs41_match_stateid(const nfs4_stateid *s1,
		const nfs4_stateid *s2)
{
8416
	if (memcmp(s1->other, s2->other, sizeof(s1->other)) != 0)
8417 8418
		return false;

8419
	if (s1->seqid == s2->seqid)
8420
		return true;
8421
	if (s1->seqid == 0 || s2->seqid == 0)
8422 8423 8424 8425 8426
		return true;

	return false;
}

8427 8428
#endif /* CONFIG_NFS_V4_1 */

8429 8430 8431
static bool nfs4_match_stateid(const nfs4_stateid *s1,
		const nfs4_stateid *s2)
{
8432
	return nfs4_stateid_match(s1, s2);
8433 8434 8435
}


8436
static const struct nfs4_state_recovery_ops nfs40_reboot_recovery_ops = {
8437
	.owner_flag_bit = NFS_OWNER_RECLAIM_REBOOT,
8438
	.state_flag_bit	= NFS_STATE_RECLAIM_REBOOT,
L
Linus Torvalds 已提交
8439 8440
	.recover_open	= nfs4_open_reclaim,
	.recover_lock	= nfs4_lock_reclaim,
8441
	.establish_clid = nfs4_init_clientid,
8442
	.detect_trunking = nfs40_discover_server_trunking,
L
Linus Torvalds 已提交
8443 8444
};

8445
#if defined(CONFIG_NFS_V4_1)
8446
static const struct nfs4_state_recovery_ops nfs41_reboot_recovery_ops = {
8447 8448 8449 8450
	.owner_flag_bit = NFS_OWNER_RECLAIM_REBOOT,
	.state_flag_bit	= NFS_STATE_RECLAIM_REBOOT,
	.recover_open	= nfs4_open_reclaim,
	.recover_lock	= nfs4_lock_reclaim,
8451
	.establish_clid = nfs41_init_clientid,
8452
	.reclaim_complete = nfs41_proc_reclaim_complete,
8453
	.detect_trunking = nfs41_discover_server_trunking,
8454 8455 8456
};
#endif /* CONFIG_NFS_V4_1 */

8457
static const struct nfs4_state_recovery_ops nfs40_nograce_recovery_ops = {
8458 8459
	.owner_flag_bit = NFS_OWNER_RECLAIM_NOGRACE,
	.state_flag_bit	= NFS_STATE_RECLAIM_NOGRACE,
8460
	.recover_open	= nfs40_open_expired,
8461 8462 8463 8464 8465
	.recover_lock	= nfs4_lock_expired,
	.establish_clid = nfs4_init_clientid,
};

#if defined(CONFIG_NFS_V4_1)
8466
static const struct nfs4_state_recovery_ops nfs41_nograce_recovery_ops = {
8467
	.owner_flag_bit = NFS_OWNER_RECLAIM_NOGRACE,
8468
	.state_flag_bit	= NFS_STATE_RECLAIM_NOGRACE,
8469 8470
	.recover_open	= nfs41_open_expired,
	.recover_lock	= nfs41_lock_expired,
8471
	.establish_clid = nfs41_init_clientid,
L
Linus Torvalds 已提交
8472
};
8473
#endif /* CONFIG_NFS_V4_1 */
L
Linus Torvalds 已提交
8474

8475
static const struct nfs4_state_maintenance_ops nfs40_state_renewal_ops = {
B
Benny Halevy 已提交
8476
	.sched_state_renewal = nfs4_proc_async_renew,
8477
	.get_state_renewal_cred_locked = nfs4_get_renew_cred_locked,
8478
	.renew_lease = nfs4_proc_renew,
B
Benny Halevy 已提交
8479 8480 8481
};

#if defined(CONFIG_NFS_V4_1)
8482
static const struct nfs4_state_maintenance_ops nfs41_state_renewal_ops = {
B
Benny Halevy 已提交
8483
	.sched_state_renewal = nfs41_proc_async_sequence,
8484
	.get_state_renewal_cred_locked = nfs4_get_machine_cred_locked,
8485
	.renew_lease = nfs4_proc_sequence,
B
Benny Halevy 已提交
8486 8487 8488
};
#endif

8489
static const struct nfs4_mig_recovery_ops nfs40_mig_recovery_ops = {
8490
	.get_locations = _nfs40_proc_get_locations,
8491
	.fsid_present = _nfs40_proc_fsid_present,
8492 8493 8494 8495
};

#if defined(CONFIG_NFS_V4_1)
static const struct nfs4_mig_recovery_ops nfs41_mig_recovery_ops = {
8496
	.get_locations = _nfs41_proc_get_locations,
8497
	.fsid_present = _nfs41_proc_fsid_present,
8498 8499 8500
};
#endif	/* CONFIG_NFS_V4_1 */

8501 8502
static const struct nfs4_minor_version_ops nfs_v4_0_minor_ops = {
	.minor_version = 0,
8503 8504 8505 8506
	.init_caps = NFS_CAP_READDIRPLUS
		| NFS_CAP_ATOMIC_OPEN
		| NFS_CAP_CHANGE_ATTR
		| NFS_CAP_POSIX_LOCK,
8507 8508
	.init_client = nfs40_init_client,
	.shutdown_client = nfs40_shutdown_client,
8509
	.match_stateid = nfs4_match_stateid,
8510
	.find_root_sec = nfs4_find_root_sec,
8511
	.free_lock_state = nfs4_release_lockowner,
8512
	.alloc_seqid = nfs_alloc_seqid,
8513
	.call_sync_ops = &nfs40_call_sync_ops,
8514 8515 8516
	.reboot_recovery_ops = &nfs40_reboot_recovery_ops,
	.nograce_recovery_ops = &nfs40_nograce_recovery_ops,
	.state_renewal_ops = &nfs40_state_renewal_ops,
8517
	.mig_recovery_ops = &nfs40_mig_recovery_ops,
8518 8519 8520
};

#if defined(CONFIG_NFS_V4_1)
8521 8522 8523 8524 8525 8526
static struct nfs_seqid *
nfs_alloc_no_seqid(struct nfs_seqid_counter *arg1, gfp_t arg2)
{
	return NULL;
}

8527 8528
static const struct nfs4_minor_version_ops nfs_v4_1_minor_ops = {
	.minor_version = 1,
8529 8530 8531
	.init_caps = NFS_CAP_READDIRPLUS
		| NFS_CAP_ATOMIC_OPEN
		| NFS_CAP_CHANGE_ATTR
8532
		| NFS_CAP_POSIX_LOCK
8533
		| NFS_CAP_STATEID_NFSV41
A
Anna Schumaker 已提交
8534
		| NFS_CAP_ATOMIC_OPEN_V1,
8535 8536
	.init_client = nfs41_init_client,
	.shutdown_client = nfs41_shutdown_client,
8537
	.match_stateid = nfs41_match_stateid,
8538
	.find_root_sec = nfs41_find_root_sec,
8539
	.free_lock_state = nfs41_free_lock_state,
8540
	.alloc_seqid = nfs_alloc_no_seqid,
8541
	.call_sync_ops = &nfs41_call_sync_ops,
8542 8543 8544
	.reboot_recovery_ops = &nfs41_reboot_recovery_ops,
	.nograce_recovery_ops = &nfs41_nograce_recovery_ops,
	.state_renewal_ops = &nfs41_state_renewal_ops,
8545
	.mig_recovery_ops = &nfs41_mig_recovery_ops,
8546 8547 8548
};
#endif

8549 8550 8551
#if defined(CONFIG_NFS_V4_2)
static const struct nfs4_minor_version_ops nfs_v4_2_minor_ops = {
	.minor_version = 2,
8552 8553 8554 8555 8556
	.init_caps = NFS_CAP_READDIRPLUS
		| NFS_CAP_ATOMIC_OPEN
		| NFS_CAP_CHANGE_ATTR
		| NFS_CAP_POSIX_LOCK
		| NFS_CAP_STATEID_NFSV41
A
Anna Schumaker 已提交
8557
		| NFS_CAP_ATOMIC_OPEN_V1
A
Anna Schumaker 已提交
8558
		| NFS_CAP_ALLOCATE
A
Anna Schumaker 已提交
8559
		| NFS_CAP_DEALLOCATE
A
Anna Schumaker 已提交
8560
		| NFS_CAP_SEEK,
8561 8562
	.init_client = nfs41_init_client,
	.shutdown_client = nfs41_shutdown_client,
8563 8564
	.match_stateid = nfs41_match_stateid,
	.find_root_sec = nfs41_find_root_sec,
8565
	.free_lock_state = nfs41_free_lock_state,
8566
	.call_sync_ops = &nfs41_call_sync_ops,
8567
	.alloc_seqid = nfs_alloc_no_seqid,
8568 8569 8570 8571 8572 8573
	.reboot_recovery_ops = &nfs41_reboot_recovery_ops,
	.nograce_recovery_ops = &nfs41_nograce_recovery_ops,
	.state_renewal_ops = &nfs41_state_renewal_ops,
};
#endif

8574 8575 8576 8577 8578
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
8579 8580 8581
#if defined(CONFIG_NFS_V4_2)
	[2] = &nfs_v4_2_minor_ops,
#endif
8582 8583
};

8584
static const struct inode_operations nfs4_dir_inode_operations = {
8585 8586 8587 8588 8589 8590 8591 8592 8593 8594 8595 8596 8597 8598 8599 8600 8601 8602 8603
	.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,
};

8604
static const struct inode_operations nfs4_file_inode_operations = {
8605 8606 8607
	.permission	= nfs_permission,
	.getattr	= nfs_getattr,
	.setattr	= nfs_setattr,
8608 8609 8610 8611
	.getxattr	= generic_getxattr,
	.setxattr	= generic_setxattr,
	.listxattr	= generic_listxattr,
	.removexattr	= generic_removexattr,
8612 8613
};

D
David Howells 已提交
8614
const struct nfs_rpc_ops nfs_v4_clientops = {
L
Linus Torvalds 已提交
8615 8616 8617
	.version	= 4,			/* protocol version */
	.dentry_ops	= &nfs4_dentry_operations,
	.dir_inode_ops	= &nfs4_dir_inode_operations,
8618
	.file_inode_ops	= &nfs4_file_inode_operations,
8619
	.file_ops	= &nfs4_file_operations,
L
Linus Torvalds 已提交
8620
	.getroot	= nfs4_proc_get_root,
B
Bryan Schumaker 已提交
8621
	.submount	= nfs4_submount,
B
Bryan Schumaker 已提交
8622
	.try_mount	= nfs4_try_mount,
L
Linus Torvalds 已提交
8623 8624 8625 8626 8627 8628 8629 8630
	.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,
8631
	.unlink_rpc_prepare = nfs4_proc_unlink_rpc_prepare,
L
Linus Torvalds 已提交
8632
	.unlink_done	= nfs4_proc_unlink_done,
8633
	.rename_setup	= nfs4_proc_rename_setup,
8634
	.rename_rpc_prepare = nfs4_proc_rename_rpc_prepare,
8635
	.rename_done	= nfs4_proc_rename_done,
L
Linus Torvalds 已提交
8636 8637 8638 8639 8640 8641 8642 8643 8644
	.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,
8645
	.set_capabilities = nfs4_server_capabilities,
L
Linus Torvalds 已提交
8646
	.decode_dirent	= nfs4_decode_dirent,
8647
	.pgio_rpc_prepare = nfs4_proc_pgio_rpc_prepare,
L
Linus Torvalds 已提交
8648
	.read_setup	= nfs4_proc_read_setup,
T
Trond Myklebust 已提交
8649
	.read_done	= nfs4_read_done,
L
Linus Torvalds 已提交
8650
	.write_setup	= nfs4_proc_write_setup,
8651
	.write_done	= nfs4_write_done,
L
Linus Torvalds 已提交
8652
	.commit_setup	= nfs4_proc_commit_setup,
8653
	.commit_rpc_prepare = nfs4_proc_commit_rpc_prepare,
8654
	.commit_done	= nfs4_commit_done,
L
Linus Torvalds 已提交
8655
	.lock		= nfs4_proc_lock,
8656
	.clear_acl_cache = nfs4_zap_acl_attr,
T
Trond Myklebust 已提交
8657
	.close_context  = nfs4_close_context,
8658
	.open_context	= nfs4_atomic_open,
8659
	.have_delegation = nfs4_have_delegation,
8660
	.return_delegation = nfs4_inode_return_delegation,
8661
	.alloc_client	= nfs4_alloc_client,
8662
	.init_client	= nfs4_init_client,
8663
	.free_client	= nfs4_free_client,
8664 8665
	.create_server	= nfs4_create_server,
	.clone_server	= nfs_clone_server,
L
Linus Torvalds 已提交
8666 8667
};

8668 8669 8670 8671 8672 8673 8674 8675 8676
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,
8677 8678 8679
#ifdef CONFIG_NFS_V4_SECURITY_LABEL
	&nfs4_xattr_nfs4_label_handler,
#endif
8680 8681 8682
	NULL
};

L
Linus Torvalds 已提交
8683 8684 8685 8686 8687
/*
 * Local variables:
 *  c-basic-offset: 8
 * End:
 */