nfs4proc.c 227.9 KB
Newer Older
L
Linus Torvalds 已提交
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41
/*
 *  fs/nfs/nfs4proc.c
 *
 *  Client-side procedure declarations for NFSv4.
 *
 *  Copyright (c) 2002 The Regents of the University of Michigan.
 *  All rights reserved.
 *
 *  Kendrick Smith <kmsmith@umich.edu>
 *  Andy Adamson   <andros@umich.edu>
 *
 *  Redistribution and use in source and binary forms, with or without
 *  modification, are permitted provided that the following conditions
 *  are met:
 *
 *  1. Redistributions of source code must retain the above copyright
 *     notice, this list of conditions and the following disclaimer.
 *  2. Redistributions in binary form must reproduce the above copyright
 *     notice, this list of conditions and the following disclaimer in the
 *     documentation and/or other materials provided with the distribution.
 *  3. Neither the name of the University nor the names of its
 *     contributors may be used to endorse or promote products derived
 *     from this software without specific prior written permission.
 *
 *  THIS SOFTWARE IS PROVIDED ``AS IS'' AND ANY EXPRESS OR IMPLIED
 *  WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF
 *  MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
 *  DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
 *  FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
 *  CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
 *  SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR
 *  BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
 *  LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
 *  NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
 *  SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 */

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

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

69 70
#include "nfs4trace.h"

L
Linus Torvalds 已提交
71 72
#define NFSDBG_FACILITY		NFSDBG_PROC

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

76
struct nfs4_opendata;
77
static int _nfs4_proc_open(struct nfs4_opendata *data);
78
static int _nfs4_recover_proc_open(struct nfs4_opendata *data);
L
Linus Torvalds 已提交
79
static int nfs4_do_fsinfo(struct nfs_server *, struct nfs_fh *, struct nfs_fsinfo *);
80
static int nfs4_async_handle_error(struct rpc_task *, const struct nfs_server *, struct nfs4_state *);
81
static void nfs_fixup_referral_attributes(struct nfs_fattr *fattr);
82 83
static int nfs4_proc_getattr(struct nfs_server *, struct nfs_fh *, struct nfs_fattr *, struct nfs4_label *label);
static int _nfs4_proc_getattr(struct nfs_server *server, struct nfs_fh *fhandle, struct nfs_fattr *fattr, struct nfs4_label *label);
84 85
static int nfs4_do_setattr(struct inode *inode, struct rpc_cred *cred,
			    struct nfs_fattr *fattr, struct iattr *sattr,
86 87
			    struct nfs4_state *state, struct nfs4_label *ilabel,
			    struct nfs4_label *olabel);
88
#ifdef CONFIG_NFS_V4_1
89 90 91 92
static int nfs41_test_stateid(struct nfs_server *, nfs4_stateid *,
		struct rpc_cred *);
static int nfs41_free_stateid(struct nfs_server *, nfs4_stateid *,
		struct rpc_cred *);
93
#endif
94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140

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

	if (label == NULL)
		return NULL;

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

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

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

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

L
Linus Torvalds 已提交
141
/* Prevent leaks of NFSv4 errors into userland */
142
static int nfs4_map_errors(int err)
L
Linus Torvalds 已提交
143
{
144 145 146 147
	if (err >= -1000)
		return err;
	switch (err) {
	case -NFS4ERR_RESOURCE:
148 149
	case -NFS4ERR_LAYOUTTRYLATER:
	case -NFS4ERR_RECALLCONFLICT:
150
		return -EREMOTEIO;
151
	case -NFS4ERR_WRONGSEC:
152
	case -NFS4ERR_WRONG_CRED:
153
		return -EPERM;
154 155 156
	case -NFS4ERR_BADOWNER:
	case -NFS4ERR_BADNAME:
		return -EINVAL;
157 158
	case -NFS4ERR_SHARE_DENIED:
		return -EACCES;
159 160
	case -NFS4ERR_MINOR_VERS_MISMATCH:
		return -EPROTONOSUPPORT;
161 162
	case -NFS4ERR_ACCESS:
		return -EACCES;
163 164
	case -NFS4ERR_FILE_OPEN:
		return -EBUSY;
165
	default:
L
Linus Torvalds 已提交
166
		dprintk("%s could not handle NFSv4 error %d\n",
167
				__func__, -err);
168
		break;
L
Linus Torvalds 已提交
169
	}
170
	return -EIO;
L
Linus Torvalds 已提交
171 172 173 174 175
}

/*
 * This is our standard bitmap for GETATTR requests.
 */
176
const u32 nfs4_fattr_bitmap[3] = {
L
Linus Torvalds 已提交
177 178 179 180 181 182 183 184 185 186 187 188 189
	FATTR4_WORD0_TYPE
	| FATTR4_WORD0_CHANGE
	| FATTR4_WORD0_SIZE
	| FATTR4_WORD0_FSID
	| FATTR4_WORD0_FILEID,
	FATTR4_WORD1_MODE
	| FATTR4_WORD1_NUMLINKS
	| FATTR4_WORD1_OWNER
	| FATTR4_WORD1_OWNER_GROUP
	| FATTR4_WORD1_RAWDEV
	| FATTR4_WORD1_SPACE_USED
	| FATTR4_WORD1_TIME_ACCESS
	| FATTR4_WORD1_TIME_METADATA
190 191 192 193
	| FATTR4_WORD1_TIME_MODIFY,
#ifdef CONFIG_NFS_V4_SECURITY_LABEL
	FATTR4_WORD2_SECURITY_LABEL
#endif
L
Linus Torvalds 已提交
194 195
};

196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213
static const u32 nfs4_pnfs_open_bitmap[3] = {
	FATTR4_WORD0_TYPE
	| FATTR4_WORD0_CHANGE
	| FATTR4_WORD0_SIZE
	| FATTR4_WORD0_FSID
	| FATTR4_WORD0_FILEID,
	FATTR4_WORD1_MODE
	| FATTR4_WORD1_NUMLINKS
	| FATTR4_WORD1_OWNER
	| FATTR4_WORD1_OWNER_GROUP
	| FATTR4_WORD1_RAWDEV
	| FATTR4_WORD1_SPACE_USED
	| FATTR4_WORD1_TIME_ACCESS
	| FATTR4_WORD1_TIME_METADATA
	| FATTR4_WORD1_TIME_MODIFY,
	FATTR4_WORD2_MDSTHRESHOLD
};

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

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

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

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

244
const u32 nfs4_fs_locations_bitmap[3] = {
245 246 247 248 249 250 251 252 253 254 255 256 257 258 259
	FATTR4_WORD0_TYPE
	| FATTR4_WORD0_CHANGE
	| FATTR4_WORD0_SIZE
	| FATTR4_WORD0_FSID
	| FATTR4_WORD0_FILEID
	| FATTR4_WORD0_FS_LOCATIONS,
	FATTR4_WORD1_MODE
	| FATTR4_WORD1_NUMLINKS
	| FATTR4_WORD1_OWNER
	| FATTR4_WORD1_OWNER_GROUP
	| FATTR4_WORD1_RAWDEV
	| FATTR4_WORD1_SPACE_USED
	| FATTR4_WORD1_TIME_ACCESS
	| FATTR4_WORD1_TIME_METADATA
	| FATTR4_WORD1_TIME_MODIFY
260
	| FATTR4_WORD1_MOUNTED_ON_FILEID,
261 262
};

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

	if (cookie > 2) {
269
		readdir->cookie = cookie;
L
Linus Torvalds 已提交
270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285
		memcpy(&readdir->verifier, verifier, sizeof(readdir->verifier));
		return;
	}

	readdir->cookie = 0;
	memset(&readdir->verifier, 0, sizeof(readdir->verifier));
	if (cookie == 2)
		return;
	
	/*
	 * NFSv4 servers do not return entries for '.' and '..'
	 * Therefore, we fake these entries here.  We let '.'
	 * have cookie 0 and '..' have cookie 1.  Note that
	 * when talking to the server, we always send cookie 0
	 * instead of 1 or 2.
	 */
286
	start = p = kmap_atomic(*readdir->pages);
L
Linus Torvalds 已提交
287 288 289 290 291 292 293 294 295 296 297
	
	if (cookie == 0) {
		*p++ = xdr_one;                                  /* next */
		*p++ = xdr_zero;                   /* cookie, first word */
		*p++ = xdr_one;                   /* cookie, second word */
		*p++ = xdr_one;                             /* entry len */
		memcpy(p, ".\0\0\0", 4);                        /* entry */
		p++;
		*p++ = xdr_one;                         /* bitmap length */
		*p++ = htonl(FATTR4_WORD0_FILEID);             /* bitmap */
		*p++ = htonl(8);              /* attribute buffer length */
P
Peter Staubach 已提交
298
		p = xdr_encode_hyper(p, NFS_FILEID(dentry->d_inode));
L
Linus Torvalds 已提交
299 300 301 302 303 304 305 306 307 308 309
	}
	
	*p++ = xdr_one;                                  /* next */
	*p++ = xdr_zero;                   /* cookie, first word */
	*p++ = xdr_two;                   /* cookie, second word */
	*p++ = xdr_two;                             /* entry len */
	memcpy(p, "..\0\0", 4);                         /* entry */
	p++;
	*p++ = xdr_one;                         /* bitmap length */
	*p++ = htonl(FATTR4_WORD0_FILEID);             /* bitmap */
	*p++ = htonl(8);              /* attribute buffer length */
P
Peter Staubach 已提交
310
	p = xdr_encode_hyper(p, NFS_FILEID(dentry->d_parent->d_inode));
L
Linus Torvalds 已提交
311 312 313

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

317 318 319 320 321 322 323 324 325 326
static int nfs4_delay(struct rpc_clnt *clnt, long *timeout)
{
	int res = 0;

	might_sleep();

	if (*timeout <= 0)
		*timeout = NFS4_POLL_RETRY_MIN;
	if (*timeout > NFS4_POLL_RETRY_MAX)
		*timeout = NFS4_POLL_RETRY_MAX;
327
	freezable_schedule_timeout_killable_unsafe(*timeout);
328 329 330 331 332 333 334 335 336
	if (fatal_signal_pending(current))
		res = -ERESTARTSYS;
	*timeout <<= 1;
	return res;
}

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

	exception->retry = 0;
	switch(errorcode) {
		case 0:
			return 0;
348
		case -NFS4ERR_OPENMODE:
349
			if (inode && nfs4_have_delegation(inode, FMODE_READ)) {
350
				nfs4_inode_return_delegation(inode);
351 352 353 354 355
				exception->retry = 1;
				return 0;
			}
			if (state == NULL)
				break;
356 357 358
			ret = nfs4_schedule_stateid_recovery(server, state);
			if (ret < 0)
				break;
359
			goto wait_on_recovery;
360
		case -NFS4ERR_DELEG_REVOKED:
361 362
		case -NFS4ERR_ADMIN_REVOKED:
		case -NFS4ERR_BAD_STATEID:
363 364 365 366 367
			if (inode != NULL && nfs4_have_delegation(inode, FMODE_READ)) {
				nfs_remove_bad_delegation(inode);
				exception->retry = 1;
				break;
			}
368 369
			if (state == NULL)
				break;
370 371 372
			ret = nfs4_schedule_stateid_recovery(server, state);
			if (ret < 0)
				break;
373
			goto wait_on_recovery;
374
		case -NFS4ERR_EXPIRED:
375 376 377 378 379
			if (state != NULL) {
				ret = nfs4_schedule_stateid_recovery(server, state);
				if (ret < 0)
					break;
			}
380
		case -NFS4ERR_STALE_STATEID:
381
		case -NFS4ERR_STALE_CLIENTID:
382 383
			nfs4_schedule_lease_recovery(clp);
			goto wait_on_recovery;
384 385 386 387 388
		case -NFS4ERR_MOVED:
			ret = nfs4_schedule_migration_recovery(server);
			if (ret < 0)
				break;
			goto wait_on_recovery;
389 390 391
		case -NFS4ERR_LEASE_MOVED:
			nfs4_schedule_lease_moved_recovery(clp);
			goto wait_on_recovery;
392
#if defined(CONFIG_NFS_V4_1)
393 394 395 396 397 398 399 400 401
		case -NFS4ERR_BADSESSION:
		case -NFS4ERR_BADSLOT:
		case -NFS4ERR_BAD_HIGH_SLOT:
		case -NFS4ERR_CONN_NOT_BOUND_TO_SESSION:
		case -NFS4ERR_DEADSESSION:
		case -NFS4ERR_SEQ_FALSE_RETRY:
		case -NFS4ERR_SEQ_MISORDERED:
			dprintk("%s ERROR: %d Reset session\n", __func__,
				errorcode);
402
			nfs4_schedule_session_recovery(clp->cl_session, errorcode);
403
			goto wait_on_recovery;
404
#endif /* defined(CONFIG_NFS_V4_1) */
405
		case -NFS4ERR_FILE_OPEN:
406 407 408 409 410 411 412
			if (exception->timeout > HZ) {
				/* We have retried a decent amount, time to
				 * fail
				 */
				ret = -EBUSY;
				break;
			}
413 414 415 416 417
		case -NFS4ERR_GRACE:
		case -NFS4ERR_DELAY:
			ret = nfs4_delay(server->client, &exception->timeout);
			if (ret != 0)
				break;
418
		case -NFS4ERR_RETRY_UNCACHED_REP:
419 420
		case -NFS4ERR_OLD_STATEID:
			exception->retry = 1;
421 422 423 424 425 426 427 428 429 430 431 432 433
			break;
		case -NFS4ERR_BADOWNER:
			/* The following works around a Linux server bug! */
		case -NFS4ERR_BADNAME:
			if (server->caps & NFS_CAP_UIDGID_NOMAP) {
				server->caps &= ~NFS_CAP_UIDGID_NOMAP;
				exception->retry = 1;
				printk(KERN_WARNING "NFS: v4 server %s "
						"does not accept raw "
						"uid/gids. "
						"Reenabling the idmapper.\n",
						server->nfs_client->cl_hostname);
			}
434 435 436
	}
	/* We failed to handle the error */
	return nfs4_map_errors(ret);
437
wait_on_recovery:
438
	ret = nfs4_wait_clnt_recover(clp);
439 440
	if (test_bit(NFS_MIG_FAILED, &server->mig_status))
		return -EIO;
441 442 443
	if (ret == 0)
		exception->retry = 1;
	return ret;
444 445
}

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

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

	return false;
}
460

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

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

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

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

	res->sr_slot = NULL;
}

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

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

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

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

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

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

out_start:
	rpc_call_start(task);
	return 0;

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

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

542
	if (slot == NULL)
C
Chuck Lever 已提交
543 544 545 546 547 548 549 550 551 552 553 554 555
		goto out;

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

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

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

558
static void nfs41_sequence_free_slot(struct nfs4_sequence_res *res)
A
Andy Adamson 已提交
559
{
560
	struct nfs4_session *session;
A
Andy Adamson 已提交
561
	struct nfs4_slot_table *tbl;
562
	struct nfs4_slot *slot = res->sr_slot;
563
	bool send_new_highest_used_slotid = false;
A
Andy Adamson 已提交
564

565
	tbl = slot->table;
566
	session = tbl->session;
567

568
	spin_lock(&tbl->slot_tbl_lock);
569 570 571 572 573 574
	/* 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;

575
	if (nfs41_wake_and_assign_slot(tbl, slot)) {
576 577 578
		send_new_highest_used_slotid = false;
		goto out_unlock;
	}
579
	nfs4_free_slot(tbl, slot);
580 581 582

	if (tbl->highest_used_slotid != NFS4_NO_SLOT)
		send_new_highest_used_slotid = false;
583
out_unlock:
584
	spin_unlock(&tbl->slot_tbl_lock);
585
	res->sr_slot = NULL;
586 587
	if (send_new_highest_used_slotid)
		nfs41_server_notify_highest_slotid_update(session->clp);
A
Andy Adamson 已提交
588 589
}

590
int nfs41_sequence_done(struct rpc_task *task, struct nfs4_sequence_res *res)
A
Andy Adamson 已提交
591
{
592
	struct nfs4_session *session;
593
	struct nfs4_slot *slot = res->sr_slot;
594
	struct nfs_client *clp;
595
	bool interrupted = false;
596
	int ret = 1;
A
Andy Adamson 已提交
597

598 599
	if (slot == NULL)
		goto out_noaction;
600 601
	/* don't increment the sequence number if the task wasn't sent */
	if (!RPC_WAS_SENT(task))
A
Andy Adamson 已提交
602 603
		goto out;

604
	session = slot->table->session;
605

606 607 608 609 610
	if (slot->interrupted) {
		slot->interrupted = 0;
		interrupted = true;
	}

611
	trace_nfs4_sequence_done(session, res);
A
Andy Adamson 已提交
612
	/* Check the SEQUENCE operation status */
613 614
	switch (res->sr_status) {
	case 0:
A
Andy Adamson 已提交
615
		/* Update the slot's sequence and clientid lease timer */
616
		++slot->seq_nr;
617
		clp = session->clp;
618
		do_renew_lease(clp, res->sr_timestamp);
619
		/* Check sequence flags */
620 621
		if (res->sr_status_flags != 0)
			nfs4_schedule_lease_recovery(clp);
622
		nfs41_update_target_slotid(slot->table, slot, res);
623
		break;
624 625 626 627 628 629 630 631 632
	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;
633 634 635 636 637
	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.
		 */
638
		dprintk("%s: slot=%u seq=%u: Operation in progress\n",
639
			__func__,
640
			slot->slot_nr,
641
			slot->seq_nr);
642
		goto out_retry;
643 644 645 646 647
	case -NFS4ERR_BADSLOT:
		/*
		 * The slot id we used was probably retired. Try again
		 * using a different slot id.
		 */
648 649
		goto retry_nowait;
	case -NFS4ERR_SEQ_MISORDERED:
650 651 652 653 654 655 656 657
		/*
		 * Was the last operation on this sequence interrupted?
		 * If so, retry after bumping the sequence number.
		 */
		if (interrupted) {
			++slot->seq_nr;
			goto retry_nowait;
		}
658 659 660 661
		/*
		 * Could this slot have been previously retired?
		 * If so, then the server may be expecting seq_nr = 1!
		 */
662 663 664 665 666
		if (slot->seq_nr != 1) {
			slot->seq_nr = 1;
			goto retry_nowait;
		}
		break;
667 668 669
	case -NFS4ERR_SEQ_FALSE_RETRY:
		++slot->seq_nr;
		goto retry_nowait;
670 671
	default:
		/* Just update the slot sequence no. */
672
		++slot->seq_nr;
A
Andy Adamson 已提交
673 674 675 676
	}
out:
	/* The session may be reset by one of the error handlers. */
	dprintk("%s: Error %d free the slot \n", __func__, res->sr_status);
677
	nfs41_sequence_free_slot(res);
678
out_noaction:
679
	return ret;
680 681 682 683 684 685
retry_nowait:
	if (rpc_restart_call_prepare(task)) {
		task->tk_status = 0;
		ret = 0;
	}
	goto out;
686
out_retry:
687
	if (!rpc_restart_call(task))
688 689 690
		goto out;
	rpc_delay(task, NFS4_POLL_RETRY_MAX);
	return 0;
A
Andy Adamson 已提交
691
}
692
EXPORT_SYMBOL_GPL(nfs41_sequence_done);
A
Andy Adamson 已提交
693

694 695
static int nfs4_sequence_done(struct rpc_task *task,
			       struct nfs4_sequence_res *res)
696
{
697
	if (res->sr_slot == NULL)
698
		return 1;
C
Chuck Lever 已提交
699 700
	if (!res->sr_slot->table->session)
		return nfs40_sequence_done(task, res);
701
	return nfs41_sequence_done(task, res);
702 703
}

A
Andy Adamson 已提交
704
int nfs41_setup_sequence(struct nfs4_session *session,
A
Andy Adamson 已提交
705 706 707 708
				struct nfs4_sequence_args *args,
				struct nfs4_sequence_res *res,
				struct rpc_task *task)
{
A
Andy Adamson 已提交
709 710 711 712
	struct nfs4_slot *slot;
	struct nfs4_slot_table *tbl;

	dprintk("--> %s\n", __func__);
A
Andy Adamson 已提交
713
	/* slot already allocated? */
714
	if (res->sr_slot != NULL)
715
		goto out_success;
A
Andy Adamson 已提交
716

A
Andy Adamson 已提交
717 718
	tbl = &session->fc_slot_table;

719 720
	task->tk_timeout = 0;

A
Andy Adamson 已提交
721
	spin_lock(&tbl->slot_tbl_lock);
722
	if (test_bit(NFS4_SLOT_TBL_DRAINING, &tbl->slot_tbl_state) &&
723
	    !args->sa_privileged) {
724 725
		/* The state manager will wait until the slot table is empty */
		dprintk("%s session is draining\n", __func__);
726
		goto out_sleep;
727 728
	}

729
	slot = nfs4_alloc_slot(tbl);
730 731 732 733
	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
734
		dprintk("<-- %s: no free slots\n", __func__);
735
		goto out_sleep;
A
Andy Adamson 已提交
736 737 738
	}
	spin_unlock(&tbl->slot_tbl_lock);

739
	args->sa_slot = slot;
A
Andy Adamson 已提交
740

741
	dprintk("<-- %s slotid=%u seqid=%u\n", __func__,
742
			slot->slot_nr, slot->seq_nr);
A
Andy Adamson 已提交
743

744
	res->sr_slot = slot;
745
	res->sr_timestamp = jiffies;
746
	res->sr_status_flags = 0;
A
Andy Adamson 已提交
747 748 749 750 751
	/*
	 * 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;
752
	trace_nfs4_setup_sequence(session, args);
753 754
out_success:
	rpc_call_start(task);
A
Andy Adamson 已提交
755
	return 0;
756
out_sleep:
757 758
	/* Privileged tasks are queued with top priority */
	if (args->sa_privileged)
759 760 761 762
		rpc_sleep_on_priority(&tbl->slot_tbl_waitq, task,
				NULL, RPC_PRIORITY_PRIVILEGED);
	else
		rpc_sleep_on(&tbl->slot_tbl_waitq, task, NULL);
763 764
	spin_unlock(&tbl->slot_tbl_lock);
	return -EAGAIN;
A
Andy Adamson 已提交
765
}
A
Andy Adamson 已提交
766
EXPORT_SYMBOL_GPL(nfs41_setup_sequence);
A
Andy Adamson 已提交
767

768 769 770 771
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
772
{
773
	struct nfs4_session *session = nfs4_get_session(server);
A
Andy Adamson 已提交
774 775
	int ret = 0;

C
Chuck Lever 已提交
776 777
	if (!session)
		return nfs40_setup_sequence(server, args, res, task);
778

779
	dprintk("--> %s clp %p session %p sr_slot %u\n",
780
		__func__, session->clp, session, res->sr_slot ?
781
			res->sr_slot->slot_nr : NFS4_NO_SLOT);
A
Andy Adamson 已提交
782

783
	ret = nfs41_setup_sequence(session, args, res, task);
C
Chuck Lever 已提交
784

A
Andy Adamson 已提交
785 786 787 788 789 790
	dprintk("<-- %s status=%d\n", __func__, ret);
	return ret;
}

static void nfs41_call_sync_prepare(struct rpc_task *task, void *calldata)
{
791
	struct nfs4_call_sync_data *data = calldata;
792
	struct nfs4_session *session = nfs4_get_session(data->seq_server);
A
Andy Adamson 已提交
793

794 795
	dprintk("--> %s data->seq_server %p\n", __func__, data->seq_server);

796
	nfs41_setup_sequence(session, data->seq_args, data->seq_res, task);
A
Andy Adamson 已提交
797 798
}

A
Andy Adamson 已提交
799 800
static void nfs41_call_sync_done(struct rpc_task *task, void *calldata)
{
801
	struct nfs4_call_sync_data *data = calldata;
A
Andy Adamson 已提交
802

803
	nfs41_sequence_done(task, data->seq_res);
A
Andy Adamson 已提交
804 805
}

806
static const struct rpc_call_ops nfs41_call_sync_ops = {
A
Andy Adamson 已提交
807
	.rpc_call_prepare = nfs41_call_sync_prepare,
A
Andy Adamson 已提交
808
	.rpc_call_done = nfs41_call_sync_done,
A
Andy Adamson 已提交
809 810
};

C
Chuck Lever 已提交
811 812
#else	/* !CONFIG_NFS_V4_1 */

813 814 815 816 817
static int nfs4_setup_sequence(const struct nfs_server *server,
			       struct nfs4_sequence_args *args,
			       struct nfs4_sequence_res *res,
			       struct rpc_task *task)
{
C
Chuck Lever 已提交
818
	return nfs40_setup_sequence(server, args, res, task);
819 820 821 822 823
}

static int nfs4_sequence_done(struct rpc_task *task,
			       struct nfs4_sequence_res *res)
{
C
Chuck Lever 已提交
824
	return nfs40_sequence_done(task, res);
825
}
C
Chuck Lever 已提交
826 827

#endif	/* !CONFIG_NFS_V4_1 */
828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846

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,
};

847 848
static int nfs4_call_sync_sequence(struct rpc_clnt *clnt,
				   struct nfs_server *server,
A
Andy Adamson 已提交
849 850
				   struct rpc_message *msg,
				   struct nfs4_sequence_args *args,
851
				   struct nfs4_sequence_res *res)
A
Andy Adamson 已提交
852 853 854
{
	int ret;
	struct rpc_task *task;
855
	struct nfs_client *clp = server->nfs_client;
856
	struct nfs4_call_sync_data data = {
857
		.seq_server = server,
A
Andy Adamson 已提交
858 859 860 861
		.seq_args = args,
		.seq_res = res,
	};
	struct rpc_task_setup task_setup = {
862
		.rpc_client = clnt,
A
Andy Adamson 已提交
863
		.rpc_message = msg,
864
		.callback_ops = clp->cl_mvops->call_sync_ops,
A
Andy Adamson 已提交
865 866 867 868 869 870 871 872 873 874 875 876 877
		.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;
}

878
static
879 880
int nfs4_call_sync(struct rpc_clnt *clnt,
		   struct nfs_server *server,
881 882 883 884 885
		   struct rpc_message *msg,
		   struct nfs4_sequence_args *args,
		   struct nfs4_sequence_res *res,
		   int cache_reply)
{
886
	nfs4_init_sequence(args, res, cache_reply);
887
	return nfs4_call_sync_sequence(clnt, server, msg, args, res);
888
}
A
Andy Adamson 已提交
889

890
static void update_changeattr(struct inode *dir, struct nfs4_change_info *cinfo)
L
Linus Torvalds 已提交
891
{
892
	struct nfs_inode *nfsi = NFS_I(dir);
L
Linus Torvalds 已提交
893

894
	spin_lock(&dir->i_lock);
895
	nfsi->cache_validity |= NFS_INO_INVALID_ATTR|NFS_INO_INVALID_DATA;
896
	if (!cinfo->atomic || cinfo->before != dir->i_version)
897
		nfs_force_lookup_revalidate(dir);
898
	dir->i_version = cinfo->after;
D
David Howells 已提交
899
	nfs_fscache_invalidate(dir);
900
	spin_unlock(&dir->i_lock);
L
Linus Torvalds 已提交
901 902
}

903
struct nfs4_opendata {
904
	struct kref kref;
905 906
	struct nfs_openargs o_arg;
	struct nfs_openres o_res;
907 908
	struct nfs_open_confirmargs c_arg;
	struct nfs_open_confirmres c_res;
909 910
	struct nfs4_string owner_name;
	struct nfs4_string group_name;
911
	struct nfs_fattr f_attr;
912
	struct nfs4_label *f_label;
913
	struct dentry *dir;
914
	struct dentry *dentry;
915
	struct nfs4_state_owner *owner;
916
	struct nfs4_state *state;
917
	struct iattr attrs;
918
	unsigned long timestamp;
919
	unsigned int rpc_done : 1;
920
	unsigned int file_created : 1;
921
	unsigned int is_recover : 1;
922 923
	int rpc_status;
	int cancelled;
924 925
};

926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954
static bool nfs4_clear_cap_atomic_open_v1(struct nfs_server *server,
		int err, struct nfs4_exception *exception)
{
	if (err != -EINVAL)
		return false;
	if (!(server->caps & NFS_CAP_ATOMIC_OPEN_V1))
		return false;
	server->caps &= ~NFS_CAP_ATOMIC_OPEN_V1;
	exception->retry = 1;
	return true;
}

static enum open_claim_type4
nfs4_map_atomic_open_claim(struct nfs_server *server,
		enum open_claim_type4 claim)
{
	if (server->caps & NFS_CAP_ATOMIC_OPEN_V1)
		return claim;
	switch (claim) {
	default:
		return claim;
	case NFS4_OPEN_CLAIM_FH:
		return NFS4_OPEN_CLAIM_NULL;
	case NFS4_OPEN_CLAIM_DELEG_CUR_FH:
		return NFS4_OPEN_CLAIM_DELEGATE_CUR;
	case NFS4_OPEN_CLAIM_DELEG_PREV_FH:
		return NFS4_OPEN_CLAIM_DELEGATE_PREV;
	}
}
955 956 957 958

static void nfs4_init_opendata_res(struct nfs4_opendata *p)
{
	p->o_res.f_attr = &p->f_attr;
959
	p->o_res.f_label = p->f_label;
960 961
	p->o_res.seqid = p->o_arg.seqid;
	p->c_res.seqid = p->c_arg.seqid;
962
	p->o_res.server = p->o_arg.server;
963
	p->o_res.access_request = p->o_arg.access;
964
	nfs_fattr_init(&p->f_attr);
965
	nfs_fattr_init_names(&p->f_attr, &p->owner_name, &p->group_name);
966 967
}

968
static struct nfs4_opendata *nfs4_opendata_alloc(struct dentry *dentry,
969
		struct nfs4_state_owner *sp, fmode_t fmode, int flags,
970
		const struct iattr *attrs,
971
		struct nfs4_label *label,
972
		enum open_claim_type4 claim,
973
		gfp_t gfp_mask)
974
{
975
	struct dentry *parent = dget_parent(dentry);
976 977 978 979
	struct inode *dir = parent->d_inode;
	struct nfs_server *server = NFS_SERVER(dir);
	struct nfs4_opendata *p;

980
	p = kzalloc(sizeof(*p), gfp_mask);
981 982
	if (p == NULL)
		goto err;
983 984 985 986 987

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

988
	p->o_arg.seqid = nfs_alloc_seqid(&sp->so_seqid, gfp_mask);
989
	if (p->o_arg.seqid == NULL)
990
		goto err_free_label;
991 992
	nfs_sb_active(dentry->d_sb);
	p->dentry = dget(dentry);
993 994 995
	p->dir = parent;
	p->owner = sp;
	atomic_inc(&sp->so_count);
996 997
	p->o_arg.open_flags = flags;
	p->o_arg.fmode = fmode & (FMODE_READ|FMODE_WRITE);
998 999 1000 1001 1002 1003 1004 1005
	/* 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;
	}
1006
	p->o_arg.clientid = server->nfs_client->cl_clientid;
1007 1008
	p->o_arg.id.create_time = ktime_to_ns(sp->so_seqid.create_time);
	p->o_arg.id.uniquifier = sp->so_seqid.owner_id;
1009
	p->o_arg.name = &dentry->d_name;
1010
	p->o_arg.server = server;
1011
	p->o_arg.bitmask = nfs4_bitmask(server, label);
1012
	p->o_arg.open_bitmap = &nfs4_fattr_bitmap[0];
1013
	p->o_arg.label = label;
1014 1015
	p->o_arg.claim = nfs4_map_atomic_open_claim(server, claim);
	switch (p->o_arg.claim) {
1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026
	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);
	}
1027
	if (attrs != NULL && attrs->ia_valid != 0) {
1028
		__u32 verf[2];
1029

1030 1031
		p->o_arg.u.attrs = &p->attrs;
		memcpy(&p->attrs, attrs, sizeof(p->attrs));
1032 1033 1034 1035 1036

		verf[0] = jiffies;
		verf[1] = current->pid;
		memcpy(p->o_arg.u.verifier.data, verf,
				sizeof(p->o_arg.u.verifier.data));
1037
	}
1038 1039 1040
	p->c_arg.fh = &p->o_res.fh;
	p->c_arg.stateid = &p->o_res.stateid;
	p->c_arg.seqid = p->o_arg.seqid;
1041
	nfs4_init_opendata_res(p);
1042
	kref_init(&p->kref);
1043
	return p;
1044 1045 1046 1047

err_free_label:
	nfs4_label_free(p->f_label);
err_free_p:
1048 1049 1050 1051 1052 1053
	kfree(p);
err:
	dput(parent);
	return NULL;
}

1054
static void nfs4_opendata_free(struct kref *kref)
1055
{
1056 1057
	struct nfs4_opendata *p = container_of(kref,
			struct nfs4_opendata, kref);
1058
	struct super_block *sb = p->dentry->d_sb;
1059 1060

	nfs_free_seqid(p->o_arg.seqid);
1061 1062
	if (p->state != NULL)
		nfs4_put_open_state(p->state);
1063
	nfs4_put_state_owner(p->owner);
1064 1065 1066

	nfs4_label_free(p->f_label);

1067
	dput(p->dir);
1068 1069
	dput(p->dentry);
	nfs_sb_deactive(sb);
1070
	nfs_fattr_free_names(&p->f_attr);
1071
	kfree(p->f_attr.mdsthreshold);
1072 1073 1074 1075 1076 1077 1078
	kfree(p);
}

static void nfs4_opendata_put(struct nfs4_opendata *p)
{
	if (p != NULL)
		kref_put(&p->kref, nfs4_opendata_free);
1079 1080
}

1081 1082 1083 1084 1085 1086 1087 1088
static int nfs4_wait_for_completion_rpc_task(struct rpc_task *task)
{
	int ret;

	ret = rpc_wait_for_completion_task(task);
	return ret;
}

1089
static int can_open_cached(struct nfs4_state *state, fmode_t mode, int open_mode)
1090 1091
{
	int ret = 0;
1092

1093
	if (open_mode & (O_EXCL|O_TRUNC))
1094 1095
		goto out;
	switch (mode & (FMODE_READ|FMODE_WRITE)) {
1096
		case FMODE_READ:
1097 1098
			ret |= test_bit(NFS_O_RDONLY_STATE, &state->flags) != 0
				&& state->n_rdonly != 0;
1099 1100
			break;
		case FMODE_WRITE:
1101 1102
			ret |= test_bit(NFS_O_WRONLY_STATE, &state->flags) != 0
				&& state->n_wronly != 0;
1103 1104
			break;
		case FMODE_READ|FMODE_WRITE:
1105 1106
			ret |= test_bit(NFS_O_RDWR_STATE, &state->flags) != 0
				&& state->n_rdwr != 0;
1107
	}
1108
out:
1109 1110 1111
	return ret;
}

1112
static int can_open_delegated(struct nfs_delegation *delegation, fmode_t fmode)
1113
{
1114 1115
	if (delegation == NULL)
		return 0;
1116
	if ((delegation->type & fmode) != fmode)
1117
		return 0;
1118
	if (test_bit(NFS_DELEGATION_NEED_RECLAIM, &delegation->flags))
1119
		return 0;
1120 1121
	if (test_bit(NFS_DELEGATION_RETURNING, &delegation->flags))
		return 0;
1122
	nfs_mark_delegation_referenced(delegation);
1123 1124 1125
	return 1;
}

1126
static void update_open_stateflags(struct nfs4_state *state, fmode_t fmode)
1127
{
1128
	switch (fmode) {
1129 1130 1131 1132 1133 1134 1135 1136 1137
		case FMODE_WRITE:
			state->n_wronly++;
			break;
		case FMODE_READ:
			state->n_rdonly++;
			break;
		case FMODE_READ|FMODE_WRITE:
			state->n_rdwr++;
	}
1138
	nfs4_state_set_mode_locked(state, state->state | fmode);
1139 1140
}

1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155
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);
}

1156 1157 1158 1159 1160
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;
1161 1162
	if (!nfs4_stateid_match_other(stateid, &state->open_stateid)) {
		nfs_test_and_clear_all_open_stateid(state);
1163
		return true;
1164
	}
1165 1166 1167 1168 1169
	if (nfs4_stateid_is_newer(stateid, &state->open_stateid))
		return true;
	return false;
}

1170 1171
static void nfs_clear_open_stateid_locked(struct nfs4_state *state,
		nfs4_stateid *stateid, fmode_t fmode)
1172
{
1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189
	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;
	if (!nfs_need_update_open_stateid(state, stateid))
		return;
1190
	if (test_bit(NFS_DELEGATED_STATE, &state->flags) == 0)
1191 1192
		nfs4_stateid_copy(&state->stateid, stateid);
	nfs4_stateid_copy(&state->open_stateid, stateid);
1193 1194 1195 1196 1197 1198 1199
}

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);
1200 1201
	if (test_bit(NFS_STATE_RECLAIM_NOGRACE, &state->flags))
		nfs4_schedule_state_manager(state->owner->so_server->nfs_client);
1202 1203
}

1204
static void nfs_set_open_stateid_locked(struct nfs4_state *state, nfs4_stateid *stateid, fmode_t fmode)
1205
{
1206
	switch (fmode) {
1207 1208 1209 1210 1211 1212 1213 1214 1215
		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);
	}
1216 1217 1218 1219 1220
	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);
1221 1222
}

1223
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
1224
{
1225 1226 1227 1228 1229
	/*
	 * Protect the call to nfs4_state_set_mode_locked and
	 * serialise the stateid update
	 */
	write_seqlock(&state->seqlock);
1230
	if (deleg_stateid != NULL) {
1231
		nfs4_stateid_copy(&state->stateid, deleg_stateid);
1232 1233 1234
		set_bit(NFS_DELEGATED_STATE, &state->flags);
	}
	if (open_stateid != NULL)
1235
		nfs_set_open_stateid_locked(state, open_stateid, fmode);
1236 1237
	write_sequnlock(&state->seqlock);
	spin_lock(&state->owner->so_lock);
1238
	update_open_stateflags(state, fmode);
1239
	spin_unlock(&state->owner->so_lock);
L
Linus Torvalds 已提交
1240 1241
}

1242
static int update_open_stateid(struct nfs4_state *state, nfs4_stateid *open_stateid, nfs4_stateid *delegation, fmode_t fmode)
1243 1244 1245 1246 1247
{
	struct nfs_inode *nfsi = NFS_I(state->inode);
	struct nfs_delegation *deleg_cur;
	int ret = 0;

1248
	fmode &= (FMODE_READ|FMODE_WRITE);
1249 1250 1251 1252 1253 1254 1255

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

	spin_lock(&deleg_cur->lock);
1256
	if (rcu_dereference(nfsi->delegation) != deleg_cur ||
1257
	   test_bit(NFS_DELEGATION_RETURNING, &deleg_cur->flags) ||
1258
	    (deleg_cur->type & fmode) != fmode)
1259 1260 1261 1262
		goto no_delegation_unlock;

	if (delegation == NULL)
		delegation = &deleg_cur->stateid;
1263
	else if (!nfs4_stateid_match(&deleg_cur->stateid, delegation))
1264 1265
		goto no_delegation_unlock;

1266
	nfs_mark_delegation_referenced(deleg_cur);
1267
	__update_open_stateid(state, open_stateid, &deleg_cur->stateid, fmode);
1268 1269 1270 1271 1272 1273 1274
	ret = 1;
no_delegation_unlock:
	spin_unlock(&deleg_cur->lock);
no_delegation:
	rcu_read_unlock();

	if (!ret && open_stateid != NULL) {
1275
		__update_open_stateid(state, open_stateid, NULL, fmode);
1276 1277
		ret = 1;
	}
1278 1279
	if (test_bit(NFS_STATE_RECLAIM_NOGRACE, &state->flags))
		nfs4_schedule_state_manager(state->owner->so_server->nfs_client);
1280 1281 1282 1283 1284

	return ret;
}


1285
static void nfs4_return_incompatible_delegation(struct inode *inode, fmode_t fmode)
1286 1287 1288 1289 1290
{
	struct nfs_delegation *delegation;

	rcu_read_lock();
	delegation = rcu_dereference(NFS_I(inode)->delegation);
1291
	if (delegation == NULL || (delegation->type & fmode) == fmode) {
1292 1293 1294 1295
		rcu_read_unlock();
		return;
	}
	rcu_read_unlock();
1296
	nfs4_inode_return_delegation(inode);
1297 1298
}

1299
static struct nfs4_state *nfs4_try_open_cached(struct nfs4_opendata *opendata)
1300 1301 1302 1303
{
	struct nfs4_state *state = opendata->state;
	struct nfs_inode *nfsi = NFS_I(state->inode);
	struct nfs_delegation *delegation;
1304
	int open_mode = opendata->o_arg.open_flags;
1305
	fmode_t fmode = opendata->o_arg.fmode;
1306 1307 1308 1309
	nfs4_stateid stateid;
	int ret = -EAGAIN;

	for (;;) {
1310
		if (can_open_cached(state, fmode, open_mode)) {
1311
			spin_lock(&state->owner->so_lock);
1312 1313
			if (can_open_cached(state, fmode, open_mode)) {
				update_open_stateflags(state, fmode);
1314 1315 1316 1317 1318
				spin_unlock(&state->owner->so_lock);
				goto out_return_state;
			}
			spin_unlock(&state->owner->so_lock);
		}
1319 1320
		rcu_read_lock();
		delegation = rcu_dereference(nfsi->delegation);
1321
		if (!can_open_delegated(delegation, fmode)) {
1322
			rcu_read_unlock();
1323
			break;
1324
		}
1325
		/* Save the delegation */
1326
		nfs4_stateid_copy(&stateid, &delegation->stateid);
1327
		rcu_read_unlock();
1328
		nfs_release_seqid(opendata->o_arg.seqid);
1329 1330 1331 1332 1333
		if (!opendata->is_recover) {
			ret = nfs_may_open(state->inode, state->owner->so_cred, open_mode);
			if (ret != 0)
				goto out;
		}
1334
		ret = -EAGAIN;
1335 1336

		/* Try to update the stateid using the delegation */
1337
		if (update_open_stateid(state, NULL, &stateid, fmode))
1338
			goto out_return_state;
1339 1340 1341 1342 1343 1344 1345 1346
	}
out:
	return ERR_PTR(ret);
out_return_state:
	atomic_inc(&state->count);
	return state;
}

1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365 1366 1367 1368 1369 1370 1371 1372 1373 1374 1375 1376 1377 1378 1379 1380 1381 1382 1383 1384
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;

1385 1386 1387 1388 1389 1390 1391
	if (!data->rpc_done) {
		if (data->rpc_status) {
			ret = data->rpc_status;
			goto err;
		}
		/* cached opens have already been processed */
		goto update;
1392 1393 1394 1395 1396 1397 1398 1399
	}

	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);
1400
update:
1401 1402
	update_open_stateid(state, &data->o_res.stateid, NULL,
			    data->o_arg.fmode);
1403
	atomic_inc(&state->count);
1404 1405 1406 1407 1408 1409 1410 1411 1412

	return state;
err:
	return ERR_PTR(ret);

}

static struct nfs4_state *
_nfs4_opendata_to_nfs4_state(struct nfs4_opendata *data)
1413 1414 1415
{
	struct inode *inode;
	struct nfs4_state *state = NULL;
1416
	int ret;
1417

1418
	if (!data->rpc_done) {
1419
		state = nfs4_try_open_cached(data);
1420 1421 1422
		goto out;
	}

1423
	ret = -EAGAIN;
1424
	if (!(data->f_attr.valid & NFS_ATTR_FATTR))
1425
		goto err;
1426
	inode = nfs_fhget(data->dir->d_sb, &data->o_res.fh, &data->f_attr, data->f_label);
1427
	ret = PTR_ERR(inode);
1428
	if (IS_ERR(inode))
1429 1430
		goto err;
	ret = -ENOMEM;
1431 1432
	state = nfs4_get_open_state(inode, data->owner);
	if (state == NULL)
1433
		goto err_put_inode;
1434 1435
	if (data->o_res.delegation_type != 0)
		nfs4_opendata_check_deleg(data, state);
1436
	update_open_stateid(state, &data->o_res.stateid, NULL,
1437
			data->o_arg.fmode);
1438
	iput(inode);
1439
out:
1440
	nfs_release_seqid(data->o_arg.seqid);
1441
	return state;
1442 1443 1444 1445
err_put_inode:
	iput(inode);
err:
	return ERR_PTR(ret);
1446 1447
}

1448 1449 1450 1451 1452 1453 1454 1455
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);
}

1456 1457 1458 1459 1460 1461 1462 1463 1464 1465 1466 1467 1468 1469 1470 1471 1472
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);
}

1473 1474
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
1475 1476 1477
{
	struct nfs4_opendata *opendata;

1478
	opendata = nfs4_opendata_alloc(ctx->dentry, state->owner, 0, 0,
1479
			NULL, NULL, claim, GFP_NOFS);
T
Trond Myklebust 已提交
1480 1481 1482 1483 1484 1485 1486
	if (opendata == NULL)
		return ERR_PTR(-ENOMEM);
	opendata->state = state;
	atomic_inc(&state->count);
	return opendata;
}

1487
static int nfs4_open_recover_helper(struct nfs4_opendata *opendata, fmode_t fmode, struct nfs4_state **res)
1488
{
1489
	struct nfs4_state *newstate;
1490 1491
	int ret;

1492 1493
	opendata->o_arg.open_flags = 0;
	opendata->o_arg.fmode = fmode;
1494 1495 1496
	memset(&opendata->o_res, 0, sizeof(opendata->o_res));
	memset(&opendata->c_res, 0, sizeof(opendata->c_res));
	nfs4_init_opendata_res(opendata);
1497
	ret = _nfs4_recover_proc_open(opendata);
1498 1499
	if (ret != 0)
		return ret; 
1500
	newstate = nfs4_opendata_to_nfs4_state(opendata);
1501 1502
	if (IS_ERR(newstate))
		return PTR_ERR(newstate);
1503
	nfs4_close_state(newstate, fmode);
1504
	*res = newstate;
1505 1506 1507 1508 1509 1510 1511 1512
	return 0;
}

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

1513 1514 1515 1516
	/* 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);
1517
	/* memory barrier prior to reading state->n_* */
1518
	clear_bit(NFS_DELEGATED_STATE, &state->flags);
1519
	clear_bit(NFS_OPEN_STATE, &state->flags);
1520 1521
	smp_rmb();
	if (state->n_rdwr != 0) {
1522
		ret = nfs4_open_recover_helper(opendata, FMODE_READ|FMODE_WRITE, &newstate);
1523 1524
		if (ret != 0)
			return ret;
1525 1526
		if (newstate != state)
			return -ESTALE;
1527 1528
	}
	if (state->n_wronly != 0) {
1529
		ret = nfs4_open_recover_helper(opendata, FMODE_WRITE, &newstate);
1530 1531
		if (ret != 0)
			return ret;
1532 1533
		if (newstate != state)
			return -ESTALE;
1534 1535
	}
	if (state->n_rdonly != 0) {
1536
		ret = nfs4_open_recover_helper(opendata, FMODE_READ, &newstate);
1537 1538
		if (ret != 0)
			return ret;
1539 1540
		if (newstate != state)
			return -ESTALE;
1541
	}
1542 1543 1544 1545 1546
	/*
	 * 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 &&
1547
	    !nfs4_stateid_match(&state->stateid, &state->open_stateid)) {
1548
		write_seqlock(&state->seqlock);
1549
		if (test_bit(NFS_DELEGATED_STATE, &state->flags) == 0)
1550
			nfs4_stateid_copy(&state->stateid, &state->open_stateid);
1551
		write_sequnlock(&state->seqlock);
1552
	}
1553 1554 1555
	return 0;
}

L
Linus Torvalds 已提交
1556 1557 1558 1559
/*
 * OPEN_RECLAIM:
 * 	reclaim state on the server after a reboot.
 */
1560
static int _nfs4_do_open_reclaim(struct nfs_open_context *ctx, struct nfs4_state *state)
L
Linus Torvalds 已提交
1561
{
1562
	struct nfs_delegation *delegation;
1563
	struct nfs4_opendata *opendata;
1564
	fmode_t delegation_type = 0;
L
Linus Torvalds 已提交
1565 1566
	int status;

1567 1568
	opendata = nfs4_open_recoverdata_alloc(ctx, state,
			NFS4_OPEN_CLAIM_PREVIOUS);
T
Trond Myklebust 已提交
1569 1570
	if (IS_ERR(opendata))
		return PTR_ERR(opendata);
1571 1572
	rcu_read_lock();
	delegation = rcu_dereference(NFS_I(state->inode)->delegation);
1573
	if (delegation != NULL && test_bit(NFS_DELEGATION_NEED_RECLAIM, &delegation->flags) != 0)
1574
		delegation_type = delegation->type;
1575
	rcu_read_unlock();
1576 1577
	opendata->o_arg.u.delegation_type = delegation_type;
	status = nfs4_open_recover(opendata, state);
1578
	nfs4_opendata_put(opendata);
L
Linus Torvalds 已提交
1579 1580 1581
	return status;
}

1582
static int nfs4_do_open_reclaim(struct nfs_open_context *ctx, struct nfs4_state *state)
L
Linus Torvalds 已提交
1583 1584 1585 1586 1587
{
	struct nfs_server *server = NFS_SERVER(state->inode);
	struct nfs4_exception exception = { };
	int err;
	do {
1588
		err = _nfs4_do_open_reclaim(ctx, state);
1589
		trace_nfs4_open_reclaim(ctx, 0, err);
1590 1591
		if (nfs4_clear_cap_atomic_open_v1(server, err, &exception))
			continue;
1592
		if (err != -NFS4ERR_DELAY)
1593 1594
			break;
		nfs4_handle_exception(server, err, &exception);
L
Linus Torvalds 已提交
1595 1596 1597 1598
	} while (exception.retry);
	return err;
}

1599 1600 1601 1602 1603 1604 1605
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))
1606
		return -EAGAIN;
1607
	ret = nfs4_do_open_reclaim(ctx, state);
1608 1609 1610 1611
	put_nfs_open_context(ctx);
	return ret;
}

1612
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
1613
{
1614 1615 1616 1617 1618 1619 1620 1621 1622 1623 1624 1625 1626 1627 1628 1629 1630 1631 1632 1633 1634 1635 1636
	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;
1637 1638 1639
		case -NFS4ERR_MOVED:
			nfs4_schedule_migration_recovery(server);
			return -EAGAIN;
1640 1641 1642
		case -NFS4ERR_LEASE_MOVED:
			nfs4_schedule_lease_moved_recovery(server->nfs_client);
			return -EAGAIN;
1643 1644 1645
		case -NFS4ERR_DELEG_REVOKED:
		case -NFS4ERR_ADMIN_REVOKED:
		case -NFS4ERR_BAD_STATEID:
1646
		case -NFS4ERR_OPENMODE:
1647 1648 1649 1650 1651 1652 1653 1654 1655
			nfs_inode_find_state_and_recover(state->inode,
					stateid);
			nfs4_schedule_stateid_recovery(server, state);
			return 0;
		case -NFS4ERR_DELAY:
		case -NFS4ERR_GRACE:
			set_bit(NFS_DELEGATED_STATE, &state->flags);
			ssleep(1);
			return -EAGAIN;
1656 1657 1658 1659
		case -ENOMEM:
		case -NFS4ERR_DENIED:
			/* kill_proc(fl->fl_pid, SIGLOST, 1); */
			return 0;
1660
	}
L
Linus Torvalds 已提交
1661 1662 1663
	return err;
}

1664 1665 1666 1667 1668 1669 1670 1671 1672 1673 1674 1675 1676 1677 1678 1679
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);
}

1680 1681 1682 1683
static void nfs4_open_confirm_prepare(struct rpc_task *task, void *calldata)
{
	struct nfs4_opendata *data = calldata;

1684 1685
	nfs40_setup_sequence(data->o_arg.server, &data->c_arg.seq_args,
				&data->c_res.seq_res, task);
1686 1687
}

1688 1689 1690 1691
static void nfs4_open_confirm_done(struct rpc_task *task, void *calldata)
{
	struct nfs4_opendata *data = calldata;

1692
	nfs40_sequence_done(task, &data->c_res.seq_res);
1693

1694
	data->rpc_status = task->tk_status;
1695
	if (data->rpc_status == 0) {
1696
		nfs4_stateid_copy(&data->o_res.stateid, &data->c_res.stateid);
1697
		nfs_confirm_seqid(&data->owner->so_seqid, 0);
1698
		renew_lease(data->o_res.server, data->timestamp);
1699
		data->rpc_done = 1;
1700
	}
1701 1702 1703 1704 1705 1706 1707 1708 1709 1710 1711
}

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! */
1712
	if (!data->rpc_done)
1713 1714
		goto out_free;
	state = nfs4_opendata_to_nfs4_state(data);
1715
	if (!IS_ERR(state))
1716
		nfs4_close_state(state, data->o_arg.fmode);
1717
out_free:
1718
	nfs4_opendata_put(data);
1719 1720 1721
}

static const struct rpc_call_ops nfs4_open_confirm_ops = {
1722
	.rpc_call_prepare = nfs4_open_confirm_prepare,
1723 1724 1725 1726 1727 1728 1729 1730 1731 1732 1733
	.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;
1734 1735 1736 1737 1738 1739
	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
1740 1741
	struct rpc_task_setup task_setup_data = {
		.rpc_client = server->client,
1742
		.rpc_message = &msg,
T
Trond Myklebust 已提交
1743 1744
		.callback_ops = &nfs4_open_confirm_ops,
		.callback_data = data,
1745
		.workqueue = nfsiod_workqueue,
T
Trond Myklebust 已提交
1746 1747
		.flags = RPC_TASK_ASYNC,
	};
L
Linus Torvalds 已提交
1748 1749
	int status;

1750
	nfs4_init_sequence(&data->c_arg.seq_args, &data->c_res.seq_res, 1);
1751
	kref_get(&data->kref);
1752 1753
	data->rpc_done = 0;
	data->rpc_status = 0;
1754
	data->timestamp = jiffies;
T
Trond Myklebust 已提交
1755
	task = rpc_run_task(&task_setup_data);
1756
	if (IS_ERR(task))
1757 1758 1759 1760 1761 1762 1763
		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;
1764
	rpc_put_task(task);
L
Linus Torvalds 已提交
1765 1766 1767
	return status;
}

1768
static void nfs4_open_prepare(struct rpc_task *task, void *calldata)
L
Linus Torvalds 已提交
1769
{
1770 1771
	struct nfs4_opendata *data = calldata;
	struct nfs4_state_owner *sp = data->owner;
1772
	struct nfs_client *clp = sp->so_server->nfs_client;
1773

1774
	if (nfs_wait_on_sequence(data->o_arg.seqid, task) != 0)
1775
		goto out_wait;
1776 1777 1778 1779 1780 1781 1782
	/*
	 * 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;

1783
		if (can_open_cached(data->state, data->o_arg.fmode, data->o_arg.open_flags))
1784
			goto out_no_action;
1785 1786
		rcu_read_lock();
		delegation = rcu_dereference(NFS_I(data->state->inode)->delegation);
1787
		if (data->o_arg.claim != NFS4_OPEN_CLAIM_DELEGATE_CUR &&
1788
		    data->o_arg.claim != NFS4_OPEN_CLAIM_DELEG_CUR_FH &&
1789 1790
		    can_open_delegated(delegation, data->o_arg.fmode))
			goto unlock_no_action;
1791 1792
		rcu_read_unlock();
	}
1793
	/* Update client id. */
1794
	data->o_arg.clientid = clp->cl_clientid;
1795 1796 1797 1798
	switch (data->o_arg.claim) {
	case NFS4_OPEN_CLAIM_PREVIOUS:
	case NFS4_OPEN_CLAIM_DELEG_CUR_FH:
	case NFS4_OPEN_CLAIM_DELEG_PREV_FH:
1799
		data->o_arg.open_bitmap = &nfs4_open_noattr_bitmap[0];
1800 1801
	case NFS4_OPEN_CLAIM_FH:
		task->tk_msg.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_OPEN_NOATTR];
T
Trond Myklebust 已提交
1802 1803
		nfs_copy_fh(&data->o_res.fh, data->o_arg.fh);
	}
1804
	data->timestamp = jiffies;
1805
	if (nfs4_setup_sequence(data->o_arg.server,
1806
				&data->o_arg.seq_args,
1807 1808 1809
				&data->o_res.seq_res,
				task) != 0)
		nfs_release_seqid(data->o_arg.seqid);
1810 1811 1812 1813 1814 1815 1816 1817 1818 1819

	/* 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;
	}
1820
	return;
1821 1822
unlock_no_action:
	rcu_read_unlock();
1823 1824
out_no_action:
	task->tk_action = NULL;
1825
out_wait:
1826
	nfs4_sequence_done(task, &data->o_res.seq_res);
1827 1828
}

1829 1830 1831
static void nfs4_open_done(struct rpc_task *task, void *calldata)
{
	struct nfs4_opendata *data = calldata;
L
Linus Torvalds 已提交
1832

1833
	data->rpc_status = task->tk_status;
1834

1835 1836
	if (!nfs4_sequence_done(task, &data->o_res.seq_res))
		return;
1837

1838
	if (task->tk_status == 0) {
1839 1840
		if (data->o_res.f_attr->valid & NFS_ATTR_FATTR_TYPE) {
			switch (data->o_res.f_attr->mode & S_IFMT) {
1841 1842 1843
			case S_IFREG:
				break;
			case S_IFLNK:
1844
				data->rpc_status = -ELOOP;
1845 1846
				break;
			case S_IFDIR:
1847
				data->rpc_status = -EISDIR;
1848 1849
				break;
			default:
1850
				data->rpc_status = -ENOTDIR;
1851
			}
1852
		}
1853
		renew_lease(data->o_res.server, data->timestamp);
1854 1855
		if (!(data->o_res.rflags & NFS4_OPEN_RESULT_CONFIRM))
			nfs_confirm_seqid(&data->owner->so_seqid, 0);
1856
	}
1857
	data->rpc_done = 1;
1858
}
1859

1860 1861 1862 1863 1864 1865 1866 1867 1868
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! */
1869
	if (data->rpc_status != 0 || !data->rpc_done)
1870 1871 1872 1873 1874
		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);
1875
	if (!IS_ERR(state))
1876
		nfs4_close_state(state, data->o_arg.fmode);
1877
out_free:
1878
	nfs4_opendata_put(data);
1879 1880 1881 1882 1883 1884 1885 1886
}

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,
};

1887
static int nfs4_run_open_task(struct nfs4_opendata *data, int isrecover)
1888 1889 1890 1891 1892 1893
{
	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;
1894 1895 1896 1897 1898 1899
	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
1900 1901
	struct rpc_task_setup task_setup_data = {
		.rpc_client = server->client,
1902
		.rpc_message = &msg,
T
Trond Myklebust 已提交
1903 1904
		.callback_ops = &nfs4_open_ops,
		.callback_data = data,
1905
		.workqueue = nfsiod_workqueue,
T
Trond Myklebust 已提交
1906 1907
		.flags = RPC_TASK_ASYNC,
	};
1908 1909
	int status;

1910
	nfs4_init_sequence(&o_arg->seq_args, &o_res->seq_res, 1);
1911
	kref_get(&data->kref);
1912 1913
	data->rpc_done = 0;
	data->rpc_status = 0;
1914
	data->cancelled = 0;
1915 1916
	data->is_recover = 0;
	if (isrecover) {
1917
		nfs4_set_sequence_privileged(&o_arg->seq_args);
1918 1919
		data->is_recover = 1;
	}
T
Trond Myklebust 已提交
1920
	task = rpc_run_task(&task_setup_data);
1921 1922 1923 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 1934 1935 1936 1937 1938 1939 1940 1941 1942 1943
        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;

1944 1945
	nfs_fattr_map_and_free_names(NFS_SERVER(dir), &data->f_attr);

1946 1947 1948 1949 1950 1951 1952 1953 1954
	if (o_res->rflags & NFS4_OPEN_RESULT_CONFIRM) {
		status = _nfs4_proc_open_confirm(data);
		if (status != 0)
			return status;
	}

	return status;
}

1955 1956
static int nfs4_opendata_access(struct rpc_cred *cred,
				struct nfs4_opendata *opendata,
1957 1958
				struct nfs4_state *state, fmode_t fmode,
				int openflags)
1959 1960 1961 1962 1963 1964 1965 1966 1967 1968
{
	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;
1969
	/* don't check MAY_WRITE - a newly created file may not have
1970 1971 1972 1973 1974 1975 1976 1977
	 * write mode bits, but POSIX allows the creating process to write.
	 * use openflags to check for exec, because fmode won't
	 * always have FMODE_EXEC set when file open for exec. */
	if (openflags & __FMODE_EXEC) {
		/* ONLY check for exec rights */
		mask = MAY_EXEC;
	} else if (fmode & FMODE_READ)
		mask = MAY_READ;
1978 1979 1980 1981 1982 1983

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

1984
	if ((mask & ~cache.mask & (MAY_READ | MAY_EXEC)) == 0)
1985 1986 1987 1988
		return 0;

	/* even though OPEN succeeded, access is denied. Close the file */
	nfs4_close_state(state, fmode);
1989
	return -EACCES;
1990 1991
}

1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003
/*
 * 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);
2004 2005 2006 2007 2008 2009
	if (!data->rpc_done)
		return status;
	if (status != 0) {
		if (status == -NFS4ERR_BADNAME &&
				!(o_arg->open_flags & O_CREAT))
			return -ENOENT;
2010
		return status;
2011
	}
2012

2013 2014
	nfs_fattr_map_and_free_names(server, &data->f_attr);

2015
	if (o_arg->open_flags & O_CREAT) {
2016
		update_changeattr(dir, &o_res->cinfo);
2017 2018 2019 2020 2021
		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
2022 2023
	if ((o_res->rflags & NFS4_OPEN_RESULT_LOCKTYPE_POSIX) == 0)
		server->caps &= ~NFS_CAP_POSIX_LOCK;
L
Linus Torvalds 已提交
2024
	if(o_res->rflags & NFS4_OPEN_RESULT_CONFIRM) {
2025
		status = _nfs4_proc_open_confirm(data);
L
Linus Torvalds 已提交
2026
		if (status != 0)
2027
			return status;
L
Linus Torvalds 已提交
2028 2029
	}
	if (!(o_res->f_attr->valid & NFS_ATTR_FATTR))
2030
		_nfs4_proc_getattr(server, &o_res->fh, o_res->f_attr, o_res->f_label);
2031
	return 0;
L
Linus Torvalds 已提交
2032 2033
}

A
Andy Adamson 已提交
2034 2035 2036 2037 2038
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
2039 2040 2041 2042 2043
/*
 * OPEN_EXPIRED:
 * 	reclaim state on the server after a network partition.
 * 	Assumes caller holds the appropriate lock
 */
2044
static int _nfs4_open_expired(struct nfs_open_context *ctx, struct nfs4_state *state)
L
Linus Torvalds 已提交
2045
{
2046
	struct nfs4_opendata *opendata;
2047
	int ret;
L
Linus Torvalds 已提交
2048

2049
	opendata = nfs4_open_recoverdata_alloc(ctx, state,
2050
			NFS4_OPEN_CLAIM_FH);
T
Trond Myklebust 已提交
2051 2052
	if (IS_ERR(opendata))
		return PTR_ERR(opendata);
2053
	ret = nfs4_open_recover(opendata, state);
2054
	if (ret == -ESTALE)
2055
		d_drop(ctx->dentry);
2056
	nfs4_opendata_put(opendata);
2057
	return ret;
L
Linus Torvalds 已提交
2058 2059
}

2060
static int nfs4_do_open_expired(struct nfs_open_context *ctx, struct nfs4_state *state)
2061
{
2062
	struct nfs_server *server = NFS_SERVER(state->inode);
2063 2064 2065 2066
	struct nfs4_exception exception = { };
	int err;

	do {
2067
		err = _nfs4_open_expired(ctx, state);
2068
		trace_nfs4_open_expired(ctx, 0, err);
2069 2070
		if (nfs4_clear_cap_atomic_open_v1(server, err, &exception))
			continue;
2071 2072 2073 2074 2075 2076 2077 2078
		switch (err) {
		default:
			goto out;
		case -NFS4ERR_GRACE:
		case -NFS4ERR_DELAY:
			nfs4_handle_exception(server, err, &exception);
			err = 0;
		}
2079
	} while (exception.retry);
2080
out:
2081 2082 2083
	return err;
}

L
Linus Torvalds 已提交
2084 2085 2086
static int nfs4_open_expired(struct nfs4_state_owner *sp, struct nfs4_state *state)
{
	struct nfs_open_context *ctx;
2087
	int ret;
L
Linus Torvalds 已提交
2088

2089 2090
	ctx = nfs4_state_find_open_context(state);
	if (IS_ERR(ctx))
2091
		return -EAGAIN;
2092
	ret = nfs4_do_open_expired(ctx, state);
2093 2094
	put_nfs_open_context(ctx);
	return ret;
L
Linus Torvalds 已提交
2095 2096
}

2097
#if defined(CONFIG_NFS_V4_1)
2098
static void nfs41_clear_delegation_stateid(struct nfs4_state *state)
2099 2100
{
	struct nfs_server *server = NFS_SERVER(state->inode);
2101
	nfs4_stateid *stateid = &state->stateid;
2102 2103 2104
	struct nfs_delegation *delegation;
	struct rpc_cred *cred = NULL;
	int status = -NFS4ERR_BAD_STATEID;
2105 2106 2107 2108 2109

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

2110 2111 2112 2113 2114 2115 2116 2117
	/* Get the delegation credential for use by test/free_stateid */
	rcu_read_lock();
	delegation = rcu_dereference(NFS_I(state->inode)->delegation);
	if (delegation != NULL &&
	    nfs4_stateid_match(&delegation->stateid, stateid)) {
		cred = get_rpccred(delegation->cred);
		rcu_read_unlock();
		status = nfs41_test_stateid(server, stateid, cred);
2118
		trace_nfs4_test_delegation_stateid(state, NULL, status);
2119 2120 2121
	} else
		rcu_read_unlock();

2122 2123 2124 2125
	if (status != NFS_OK) {
		/* Free the stateid unless the server explicitly
		 * informs us the stateid is unrecognized. */
		if (status != -NFS4ERR_BAD_STATEID)
2126
			nfs41_free_stateid(server, stateid, cred);
2127
		nfs_remove_bad_delegation(state->inode);
2128

2129 2130 2131
		write_seqlock(&state->seqlock);
		nfs4_stateid_copy(&state->stateid, &state->open_stateid);
		write_sequnlock(&state->seqlock);
2132 2133
		clear_bit(NFS_DELEGATED_STATE, &state->flags);
	}
2134 2135 2136

	if (cred != NULL)
		put_rpccred(cred);
2137 2138 2139 2140 2141 2142 2143 2144 2145 2146 2147 2148 2149
}

/**
 * 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);
2150
	nfs4_stateid *stateid = &state->open_stateid;
2151
	struct rpc_cred *cred = state->owner->so_cred;
2152 2153 2154 2155 2156 2157 2158 2159
	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;

2160
	status = nfs41_test_stateid(server, stateid, cred);
2161
	trace_nfs4_test_open_stateid(state, NULL, status);
2162 2163 2164 2165
	if (status != NFS_OK) {
		/* Free the stateid unless the server explicitly
		 * informs us the stateid is unrecognized. */
		if (status != -NFS4ERR_BAD_STATEID)
2166
			nfs41_free_stateid(server, stateid, cred);
2167 2168 2169 2170

		clear_bit(NFS_O_RDONLY_STATE, &state->flags);
		clear_bit(NFS_O_WRONLY_STATE, &state->flags);
		clear_bit(NFS_O_RDWR_STATE, &state->flags);
2171
		clear_bit(NFS_OPEN_STATE, &state->flags);
2172 2173 2174 2175 2176 2177
	}
	return status;
}

static int nfs41_open_expired(struct nfs4_state_owner *sp, struct nfs4_state *state)
{
2178
	int status;
2179

2180 2181
	nfs41_clear_delegation_stateid(state);
	status = nfs41_check_open_stateid(state);
2182 2183 2184
	if (status != NFS_OK)
		status = nfs4_open_expired(sp, state);
	return status;
2185 2186 2187
}
#endif

2188 2189 2190 2191 2192 2193 2194 2195 2196 2197 2198 2199 2200 2201 2202 2203
/*
 * 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;
}

2204 2205 2206
static int _nfs4_open_and_get_state(struct nfs4_opendata *opendata,
		fmode_t fmode,
		int flags,
2207
		struct nfs_open_context *ctx)
2208 2209 2210
{
	struct nfs4_state_owner *sp = opendata->owner;
	struct nfs_server *server = sp->so_server;
2211
	struct dentry *dentry;
2212 2213 2214 2215 2216 2217 2218 2219 2220 2221 2222 2223 2224 2225 2226 2227 2228
	struct nfs4_state *state;
	unsigned int seq;
	int ret;

	seq = raw_seqcount_begin(&sp->so_reclaim_seqcount);

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

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

2229 2230 2231 2232 2233 2234 2235 2236 2237 2238 2239 2240 2241 2242 2243
	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));
	}

2244 2245 2246 2247
	ret = nfs4_opendata_access(sp->so_cred, opendata, state, fmode, flags);
	if (ret != 0)
		goto out;

2248
	ctx->state = state;
2249 2250 2251 2252 2253
	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);
	}
2254 2255 2256 2257
out:
	return ret;
}

L
Linus Torvalds 已提交
2258
/*
2259
 * Returns a referenced nfs4_state
L
Linus Torvalds 已提交
2260
 */
2261
static int _nfs4_do_open(struct inode *dir,
2262
			struct nfs_open_context *ctx,
2263 2264
			int flags,
			struct iattr *sattr,
2265 2266
			struct nfs4_label *label,
			int *opened)
L
Linus Torvalds 已提交
2267 2268 2269 2270
{
	struct nfs4_state_owner  *sp;
	struct nfs4_state     *state = NULL;
	struct nfs_server       *server = NFS_SERVER(dir);
2271
	struct nfs4_opendata *opendata;
2272 2273 2274 2275
	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);
2276
	enum open_claim_type4 claim = NFS4_OPEN_CLAIM_NULL;
2277
	struct nfs4_label *olabel = NULL;
2278
	int status;
L
Linus Torvalds 已提交
2279 2280 2281

	/* Protect against reboot recovery conflicts */
	status = -ENOMEM;
2282 2283
	sp = nfs4_get_state_owner(server, cred, GFP_KERNEL);
	if (sp == NULL) {
L
Linus Torvalds 已提交
2284 2285 2286
		dprintk("nfs4_do_open: nfs4_get_state_owner failed!\n");
		goto out_err;
	}
2287 2288
	status = nfs4_recover_expired_lease(server);
	if (status != 0)
2289
		goto err_put_state_owner;
2290 2291
	if (dentry->d_inode != NULL)
		nfs4_return_incompatible_delegation(dentry->d_inode, fmode);
2292
	status = -ENOMEM;
2293 2294
	if (dentry->d_inode)
		claim = NFS4_OPEN_CLAIM_FH;
2295
	opendata = nfs4_opendata_alloc(dentry, sp, fmode, flags, sattr,
2296
			label, claim, GFP_KERNEL);
2297
	if (opendata == NULL)
T
Trond Myklebust 已提交
2298
		goto err_put_state_owner;
L
Linus Torvalds 已提交
2299

2300 2301 2302 2303 2304 2305 2306 2307
	if (label) {
		olabel = nfs4_label_alloc(server, GFP_KERNEL);
		if (IS_ERR(olabel)) {
			status = PTR_ERR(olabel);
			goto err_opendata_put;
		}
	}

2308 2309 2310 2311 2312 2313
	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;
		}
2314
		opendata->o_arg.open_bitmap = &nfs4_pnfs_open_bitmap[0];
2315
	}
2316 2317
	if (dentry->d_inode != NULL)
		opendata->state = nfs4_get_open_state(dentry->d_inode, sp);
2318

2319
	status = _nfs4_open_and_get_state(opendata, fmode, flags, ctx);
2320
	if (status != 0)
2321
		goto err_free_label;
2322
	state = ctx->state;
2323

2324 2325
	if ((opendata->o_arg.open_flags & O_EXCL) &&
	    (opendata->o_arg.createmode != NFS4_CREATE_GUARDED)) {
2326 2327 2328 2329 2330
		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,
2331 2332
				state, label, olabel);
		if (status == 0) {
2333
			nfs_setattr_update_inode(state->inode, sattr);
2334
			nfs_post_op_update_inode(state->inode, opendata->o_res.f_attr);
2335
			nfs_setsecurity(state->inode, opendata->o_res.f_attr, olabel);
2336
		}
2337
	}
2338 2339
	if (opendata->file_created)
		*opened |= FILE_CREATED;
2340

2341
	if (pnfs_use_threshold(ctx_th, opendata->f_attr.mdsthreshold, server)) {
2342
		*ctx_th = opendata->f_attr.mdsthreshold;
2343 2344
		opendata->f_attr.mdsthreshold = NULL;
	}
2345

2346 2347
	nfs4_label_free(olabel);

2348
	nfs4_opendata_put(opendata);
L
Linus Torvalds 已提交
2349 2350
	nfs4_put_state_owner(sp);
	return 0;
2351 2352
err_free_label:
	nfs4_label_free(olabel);
2353 2354
err_opendata_put:
	nfs4_opendata_put(opendata);
2355 2356
err_put_state_owner:
	nfs4_put_state_owner(sp);
L
Linus Torvalds 已提交
2357 2358 2359 2360 2361
out_err:
	return status;
}


2362
static struct nfs4_state *nfs4_do_open(struct inode *dir,
2363
					struct nfs_open_context *ctx,
2364 2365
					int flags,
					struct iattr *sattr,
2366 2367
					struct nfs4_label *label,
					int *opened)
L
Linus Torvalds 已提交
2368
{
2369
	struct nfs_server *server = NFS_SERVER(dir);
L
Linus Torvalds 已提交
2370 2371 2372 2373 2374
	struct nfs4_exception exception = { };
	struct nfs4_state *res;
	int status;

	do {
2375
		status = _nfs4_do_open(dir, ctx, flags, sattr, label, opened);
2376
		res = ctx->state;
2377
		trace_nfs4_open_file(ctx, flags, status);
L
Linus Torvalds 已提交
2378 2379 2380 2381 2382 2383 2384 2385
		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
2386
		 * doing an OPEN, we assume that nfs_increment_open_seqid() will
L
Linus Torvalds 已提交
2387 2388 2389 2390 2391
		 * 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) {
2392
			pr_warn_ratelimited("NFS: v4 server %s "
2393 2394
					" returned a bad sequence-id error!\n",
					NFS_SERVER(dir)->nfs_client->cl_hostname);
L
Linus Torvalds 已提交
2395 2396 2397
			exception.retry = 1;
			continue;
		}
2398 2399 2400 2401 2402 2403 2404 2405 2406 2407
		/*
		 * 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;
		}
2408 2409 2410 2411 2412
		if (status == -EAGAIN) {
			/* We must have found a delegation */
			exception.retry = 1;
			continue;
		}
2413 2414 2415
		if (nfs4_clear_cap_atomic_open_v1(server, status, &exception))
			continue;
		res = ERR_PTR(nfs4_handle_exception(server,
L
Linus Torvalds 已提交
2416 2417 2418 2419 2420
					status, &exception));
	} while (exception.retry);
	return res;
}

2421 2422
static int _nfs4_do_setattr(struct inode *inode, struct rpc_cred *cred,
			    struct nfs_fattr *fattr, struct iattr *sattr,
2423 2424
			    struct nfs4_state *state, struct nfs4_label *ilabel,
			    struct nfs4_label *olabel)
L
Linus Torvalds 已提交
2425
{
2426
	struct nfs_server *server = NFS_SERVER(inode);
L
Linus Torvalds 已提交
2427
        struct nfs_setattrargs  arg = {
2428
                .fh             = NFS_FH(inode),
L
Linus Torvalds 已提交
2429 2430 2431
                .iap            = sattr,
		.server		= server,
		.bitmask = server->attr_bitmask,
2432
		.label		= ilabel,
L
Linus Torvalds 已提交
2433 2434 2435
        };
        struct nfs_setattrres  res = {
		.fattr		= fattr,
2436
		.label		= olabel,
L
Linus Torvalds 已提交
2437 2438 2439
		.server		= server,
        };
        struct rpc_message msg = {
2440 2441 2442 2443
		.rpc_proc	= &nfs4_procedures[NFSPROC4_CLNT_SETATTR],
		.rpc_argp	= &arg,
		.rpc_resp	= &res,
		.rpc_cred	= cred,
L
Linus Torvalds 已提交
2444
        };
2445
	unsigned long timestamp = jiffies;
2446 2447
	fmode_t fmode;
	bool truncate;
2448
	int status;
L
Linus Torvalds 已提交
2449

2450 2451 2452 2453
	arg.bitmask = nfs4_bitmask(server, ilabel);
	if (ilabel)
		arg.bitmask = nfs4_bitmask(server, olabel);

2454
	nfs_fattr_init(fattr);
L
Linus Torvalds 已提交
2455

2456 2457 2458 2459 2460 2461
	/* 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 */
2462
	} else if (truncate && state != NULL) {
2463 2464 2465 2466
		struct nfs_lockowner lockowner = {
			.l_owner = current->files,
			.l_pid = current->tgid,
		};
2467 2468 2469 2470 2471
		if (!nfs4_valid_open_stateid(state))
			return -EBADF;
		if (nfs4_select_rw_stateid(&arg.stateid, state, FMODE_WRITE,
				&lockowner) == -EIO)
			return -EBADF;
2472
	} else
2473
		nfs4_stateid_copy(&arg.stateid, &zero_stateid);
L
Linus Torvalds 已提交
2474

2475
	status = nfs4_call_sync(server->client, server, &msg, &arg.seq_args, &res.seq_res, 1);
2476 2477
	if (status == 0 && state != NULL)
		renew_lease(server, timestamp);
2478
	return status;
L
Linus Torvalds 已提交
2479 2480
}

2481 2482
static int nfs4_do_setattr(struct inode *inode, struct rpc_cred *cred,
			   struct nfs_fattr *fattr, struct iattr *sattr,
2483 2484
			   struct nfs4_state *state, struct nfs4_label *ilabel,
			   struct nfs4_label *olabel)
L
Linus Torvalds 已提交
2485
{
2486
	struct nfs_server *server = NFS_SERVER(inode);
2487 2488
	struct nfs4_exception exception = {
		.state = state,
2489
		.inode = inode,
2490
	};
L
Linus Torvalds 已提交
2491 2492
	int err;
	do {
2493
		err = _nfs4_do_setattr(inode, cred, fattr, sattr, state, ilabel, olabel);
2494
		trace_nfs4_setattr(inode, err);
2495 2496
		switch (err) {
		case -NFS4ERR_OPENMODE:
2497 2498 2499 2500 2501 2502 2503
			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);
			}
2504 2505 2506 2507 2508 2509 2510 2511
			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
2512
	} while (exception.retry);
2513
out:
L
Linus Torvalds 已提交
2514 2515 2516 2517 2518 2519 2520 2521
	return err;
}

struct nfs4_closedata {
	struct inode *inode;
	struct nfs4_state *state;
	struct nfs_closeargs arg;
	struct nfs_closeres res;
2522
	struct nfs_fattr fattr;
2523
	unsigned long timestamp;
F
Fred Isaman 已提交
2524 2525
	bool roc;
	u32 roc_barrier;
L
Linus Torvalds 已提交
2526 2527
};

2528
static void nfs4_free_closedata(void *data)
2529
{
2530 2531
	struct nfs4_closedata *calldata = data;
	struct nfs4_state_owner *sp = calldata->state->owner;
2532
	struct super_block *sb = calldata->state->inode->i_sb;
2533

F
Fred Isaman 已提交
2534 2535
	if (calldata->roc)
		pnfs_roc_release(calldata->state->inode);
2536 2537 2538
	nfs4_put_open_state(calldata->state);
	nfs_free_seqid(calldata->arg.seqid);
	nfs4_put_state_owner(sp);
2539
	nfs_sb_deactive(sb);
2540 2541 2542
	kfree(calldata);
}

2543
static void nfs4_close_done(struct rpc_task *task, void *data)
L
Linus Torvalds 已提交
2544
{
2545
	struct nfs4_closedata *calldata = data;
L
Linus Torvalds 已提交
2546 2547 2548
	struct nfs4_state *state = calldata->state;
	struct nfs_server *server = NFS_SERVER(calldata->inode);

2549
	dprintk("%s: begin!\n", __func__);
2550 2551
	if (!nfs4_sequence_done(task, &calldata->res.seq_res))
		return;
2552
	trace_nfs4_close(state, &calldata->arg, &calldata->res, task->tk_status);
L
Linus Torvalds 已提交
2553 2554 2555 2556 2557
        /* hmm. we are done with the inode, and in the process of freeing
	 * the state_owner. we keep this around to process errors
	 */
	switch (task->tk_status) {
		case 0:
F
Fred Isaman 已提交
2558 2559 2560
			if (calldata->roc)
				pnfs_roc_set_barrier(state->inode,
						     calldata->roc_barrier);
2561
			nfs_clear_open_stateid(state, &calldata->res.stateid, 0);
2562
			renew_lease(server, calldata->timestamp);
2563
			goto out_release;
2564
		case -NFS4ERR_ADMIN_REVOKED:
L
Linus Torvalds 已提交
2565
		case -NFS4ERR_STALE_STATEID:
2566 2567
		case -NFS4ERR_OLD_STATEID:
		case -NFS4ERR_BAD_STATEID:
L
Linus Torvalds 已提交
2568
		case -NFS4ERR_EXPIRED:
2569
			if (calldata->arg.fmode == 0)
2570
				break;
L
Linus Torvalds 已提交
2571
		default:
2572
			if (nfs4_async_handle_error(task, server, state) == -EAGAIN) {
2573
				rpc_restart_call_prepare(task);
2574 2575
				goto out_release;
			}
L
Linus Torvalds 已提交
2576
	}
2577
	nfs_clear_open_stateid(state, NULL, calldata->arg.fmode);
2578
out_release:
2579
	nfs_release_seqid(calldata->arg.seqid);
2580
	nfs_refresh_inode(calldata->inode, calldata->res.fattr);
2581
	dprintk("%s: done, ret = %d!\n", __func__, task->tk_status);
L
Linus Torvalds 已提交
2582 2583
}

T
Trond Myklebust 已提交
2584
static void nfs4_close_prepare(struct rpc_task *task, void *data)
L
Linus Torvalds 已提交
2585
{
T
Trond Myklebust 已提交
2586
	struct nfs4_closedata *calldata = data;
2587
	struct nfs4_state *state = calldata->state;
2588
	struct inode *inode = calldata->inode;
2589
	int call_close = 0;
2590

2591
	dprintk("%s: begin!\n", __func__);
2592
	if (nfs_wait_on_sequence(calldata->arg.seqid, task) != 0)
2593
		goto out_wait;
2594

2595 2596
	task->tk_msg.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_OPEN_DOWNGRADE];
	calldata->arg.fmode = FMODE_READ|FMODE_WRITE;
2597
	spin_lock(&state->owner->so_lock);
2598
	/* Calculate the change in open mode */
2599
	if (state->n_rdwr == 0) {
2600
		if (state->n_rdonly == 0) {
2601 2602 2603
			call_close |= test_bit(NFS_O_RDONLY_STATE, &state->flags);
			call_close |= test_bit(NFS_O_RDWR_STATE, &state->flags);
			calldata->arg.fmode &= ~FMODE_READ;
2604 2605
		}
		if (state->n_wronly == 0) {
2606 2607 2608
			call_close |= test_bit(NFS_O_WRONLY_STATE, &state->flags);
			call_close |= test_bit(NFS_O_RDWR_STATE, &state->flags);
			calldata->arg.fmode &= ~FMODE_WRITE;
2609
		}
2610
	}
2611 2612
	if (!nfs4_valid_open_stateid(state))
		call_close = 0;
2613
	spin_unlock(&state->owner->so_lock);
2614 2615

	if (!call_close) {
2616
		/* Note: exit _without_ calling nfs4_close_done */
2617
		goto out_no_action;
2618
	}
2619

F
Fred Isaman 已提交
2620
	if (calldata->arg.fmode == 0) {
2621
		task->tk_msg.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_CLOSE];
F
Fred Isaman 已提交
2622
		if (calldata->roc &&
2623 2624
		    pnfs_roc_drain(inode, &calldata->roc_barrier, task)) {
			nfs_release_seqid(calldata->arg.seqid);
2625
			goto out_wait;
2626
		    }
F
Fred Isaman 已提交
2627
	}
2628

2629
	nfs_fattr_init(calldata->res.fattr);
2630
	calldata->timestamp = jiffies;
2631
	if (nfs4_setup_sequence(NFS_SERVER(inode),
2632 2633
				&calldata->arg.seq_args,
				&calldata->res.seq_res,
2634 2635
				task) != 0)
		nfs_release_seqid(calldata->arg.seqid);
2636
	dprintk("%s: done!\n", __func__);
2637 2638 2639 2640 2641
	return;
out_no_action:
	task->tk_action = NULL;
out_wait:
	nfs4_sequence_done(task, &calldata->res.seq_res);
L
Linus Torvalds 已提交
2642 2643
}

2644
static const struct rpc_call_ops nfs4_close_ops = {
T
Trond Myklebust 已提交
2645
	.rpc_call_prepare = nfs4_close_prepare,
2646 2647 2648 2649
	.rpc_call_done = nfs4_close_done,
	.rpc_release = nfs4_free_closedata,
};

L
Linus Torvalds 已提交
2650 2651 2652 2653 2654 2655 2656 2657 2658 2659 2660
/* 
 * 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!
 */
2661
int nfs4_do_close(struct nfs4_state *state, gfp_t gfp_mask, int wait)
L
Linus Torvalds 已提交
2662
{
2663
	struct nfs_server *server = NFS_SERVER(state->inode);
L
Linus Torvalds 已提交
2664
	struct nfs4_closedata *calldata;
2665 2666
	struct nfs4_state_owner *sp = state->owner;
	struct rpc_task *task;
2667 2668 2669 2670
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_CLOSE],
		.rpc_cred = state->owner->so_cred,
	};
T
Trond Myklebust 已提交
2671 2672
	struct rpc_task_setup task_setup_data = {
		.rpc_client = server->client,
2673
		.rpc_message = &msg,
T
Trond Myklebust 已提交
2674
		.callback_ops = &nfs4_close_ops,
2675
		.workqueue = nfsiod_workqueue,
T
Trond Myklebust 已提交
2676 2677
		.flags = RPC_TASK_ASYNC,
	};
2678
	int status = -ENOMEM;
L
Linus Torvalds 已提交
2679

2680 2681 2682
	nfs4_state_protect(server->nfs_client, NFS_SP4_MACH_CRED_CLEANUP,
		&task_setup_data.rpc_client, &msg);

2683
	calldata = kzalloc(sizeof(*calldata), gfp_mask);
L
Linus Torvalds 已提交
2684
	if (calldata == NULL)
2685
		goto out;
2686
	nfs4_init_sequence(&calldata->arg.seq_args, &calldata->res.seq_res, 1);
2687
	calldata->inode = state->inode;
L
Linus Torvalds 已提交
2688
	calldata->state = state;
2689
	calldata->arg.fh = NFS_FH(state->inode);
2690
	calldata->arg.stateid = &state->open_stateid;
L
Linus Torvalds 已提交
2691
	/* Serialization for the sequence id */
2692
	calldata->arg.seqid = nfs_alloc_seqid(&state->owner->so_seqid, gfp_mask);
2693 2694
	if (calldata->arg.seqid == NULL)
		goto out_free_calldata;
2695
	calldata->arg.fmode = 0;
2696
	calldata->arg.bitmask = server->cache_consistency_bitmask;
2697
	calldata->res.fattr = &calldata->fattr;
2698
	calldata->res.seqid = calldata->arg.seqid;
2699
	calldata->res.server = server;
2700
	calldata->roc = pnfs_roc(state->inode);
2701
	nfs_sb_active(calldata->inode->i_sb);
2702

2703 2704
	msg.rpc_argp = &calldata->arg;
	msg.rpc_resp = &calldata->res;
T
Trond Myklebust 已提交
2705 2706
	task_setup_data.callback_data = calldata;
	task = rpc_run_task(&task_setup_data);
2707 2708
	if (IS_ERR(task))
		return PTR_ERR(task);
2709 2710 2711
	status = 0;
	if (wait)
		status = rpc_wait_for_completion_task(task);
2712
	rpc_put_task(task);
2713
	return status;
2714 2715 2716
out_free_calldata:
	kfree(calldata);
out:
2717 2718
	nfs4_put_open_state(state);
	nfs4_put_state_owner(sp);
2719
	return status;
L
Linus Torvalds 已提交
2720 2721
}

2722
static struct inode *
2723 2724
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
2725 2726
{
	struct nfs4_state *state;
2727 2728 2729
	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
2730

T
Trond Myklebust 已提交
2731
	/* Protect against concurrent sillydeletes */
2732
	state = nfs4_do_open(dir, ctx, open_flags, attr, label, opened);
2733 2734 2735

	nfs4_label_release_security(label);

2736 2737
	if (IS_ERR(state))
		return ERR_CAST(state);
2738
	return state->inode;
L
Linus Torvalds 已提交
2739 2740
}

2741
static void nfs4_close_context(struct nfs_open_context *ctx, int is_sync)
T
Trond Myklebust 已提交
2742 2743 2744 2745
{
	if (ctx->state == NULL)
		return;
	if (is_sync)
2746
		nfs4_close_sync(ctx->state, ctx->mode);
T
Trond Myklebust 已提交
2747
	else
2748
		nfs4_close_state(ctx->state, ctx->mode);
T
Trond Myklebust 已提交
2749
}
L
Linus Torvalds 已提交
2750

2751 2752 2753 2754
#define FATTR4_WORD1_NFS40_MASK (2*FATTR4_WORD1_MOUNTED_ON_FILEID - 1UL)
#define FATTR4_WORD2_NFS41_MASK (2*FATTR4_WORD2_SUPPATTR_EXCLCREAT - 1UL)
#define FATTR4_WORD2_NFS42_MASK (2*FATTR4_WORD2_CHANGE_SECURITY_LABEL - 1UL)

L
Linus Torvalds 已提交
2755 2756
static int _nfs4_server_capabilities(struct nfs_server *server, struct nfs_fh *fhandle)
{
B
Benny Halevy 已提交
2757 2758 2759
	struct nfs4_server_caps_arg args = {
		.fhandle = fhandle,
	};
L
Linus Torvalds 已提交
2760 2761 2762
	struct nfs4_server_caps_res res = {};
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_SERVER_CAPS],
B
Benny Halevy 已提交
2763
		.rpc_argp = &args,
L
Linus Torvalds 已提交
2764 2765 2766 2767
		.rpc_resp = &res,
	};
	int status;

2768
	status = nfs4_call_sync(server->client, server, &msg, &args.seq_args, &res.seq_res, 0);
L
Linus Torvalds 已提交
2769
	if (status == 0) {
2770 2771 2772 2773 2774 2775 2776 2777 2778 2779 2780 2781
		/* 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
2782
		memcpy(server->attr_bitmask, res.attr_bitmask, sizeof(server->attr_bitmask));
2783 2784 2785 2786
		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|
2787 2788
				NFS_CAP_CTIME|NFS_CAP_MTIME|
				NFS_CAP_SECURITY_LABEL);
2789 2790
		if (res.attr_bitmask[0] & FATTR4_WORD0_ACL &&
				res.acl_bitmask & ACL4_SUPPORT_ALLOW_ACL)
L
Linus Torvalds 已提交
2791 2792 2793 2794 2795
			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;
2796 2797 2798 2799 2800 2801 2802 2803 2804 2805 2806 2807 2808 2809 2810 2811
		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;
2812 2813 2814 2815 2816 2817
#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));
2818
		server->attr_bitmask_nl[2] &= ~FATTR4_WORD2_SECURITY_LABEL;
2819

2820 2821 2822
		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;
2823
		server->cache_consistency_bitmask[2] = 0;
L
Linus Torvalds 已提交
2824
		server->acl_bitmask = res.acl_bitmask;
2825
		server->fh_expire_type = res.fh_expire_type;
L
Linus Torvalds 已提交
2826
	}
A
Andy Adamson 已提交
2827

L
Linus Torvalds 已提交
2828 2829 2830
	return status;
}

2831
int nfs4_server_capabilities(struct nfs_server *server, struct nfs_fh *fhandle)
L
Linus Torvalds 已提交
2832 2833 2834 2835 2836 2837 2838 2839 2840 2841 2842 2843 2844 2845
{
	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)
{
2846
	u32 bitmask[3];
L
Linus Torvalds 已提交
2847
	struct nfs4_lookup_root_arg args = {
2848
		.bitmask = bitmask,
L
Linus Torvalds 已提交
2849 2850 2851
	};
	struct nfs4_lookup_res res = {
		.server = server,
2852
		.fattr = info->fattr,
L
Linus Torvalds 已提交
2853 2854 2855 2856 2857 2858 2859
		.fh = fhandle,
	};
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_LOOKUP_ROOT],
		.rpc_argp = &args,
		.rpc_resp = &res,
	};
2860

2861 2862 2863 2864 2865 2866 2867
	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;

2868
	nfs_fattr_init(info->fattr);
2869
	return nfs4_call_sync(server->client, server, &msg, &args.seq_args, &res.seq_res, 0);
L
Linus Torvalds 已提交
2870 2871 2872 2873 2874 2875 2876 2877
}

static int nfs4_lookup_root(struct nfs_server *server, struct nfs_fh *fhandle,
		struct nfs_fsinfo *info)
{
	struct nfs4_exception exception = { };
	int err;
	do {
2878
		err = _nfs4_lookup_root(server, fhandle, info);
2879
		trace_nfs4_lookup_root(server, fhandle, info->fattr, err);
2880 2881 2882
		switch (err) {
		case 0:
		case -NFS4ERR_WRONGSEC:
2883
			goto out;
2884 2885 2886
		default:
			err = nfs4_handle_exception(server, err, &exception);
		}
L
Linus Torvalds 已提交
2887
	} while (exception.retry);
2888
out:
L
Linus Torvalds 已提交
2889 2890 2891
	return err;
}

2892 2893 2894
static int nfs4_lookup_root_sec(struct nfs_server *server, struct nfs_fh *fhandle,
				struct nfs_fsinfo *info, rpc_authflavor_t flavor)
{
2895 2896 2897
	struct rpc_auth_create_args auth_args = {
		.pseudoflavor = flavor,
	};
2898 2899 2900
	struct rpc_auth *auth;
	int ret;

2901
	auth = rpcauth_create(&auth_args, server->client);
2902
	if (IS_ERR(auth)) {
2903
		ret = -EACCES;
2904 2905 2906 2907 2908 2909 2910
		goto out;
	}
	ret = nfs4_lookup_root(server, fhandle, info);
out:
	return ret;
}

2911 2912 2913 2914 2915 2916 2917 2918 2919
/*
 * 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.
 */
2920
static int nfs4_find_root_sec(struct nfs_server *server, struct nfs_fh *fhandle,
2921
			      struct nfs_fsinfo *info)
L
Linus Torvalds 已提交
2922
{
2923 2924 2925 2926 2927
	/* Per 3530bis 15.33.5 */
	static const rpc_authflavor_t flav_array[] = {
		RPC_AUTH_GSS_KRB5P,
		RPC_AUTH_GSS_KRB5I,
		RPC_AUTH_GSS_KRB5,
2928
		RPC_AUTH_UNIX,			/* courtesy */
2929 2930 2931 2932
		RPC_AUTH_NULL,
	};
	int status = -EPERM;
	size_t i;
C
Chuck Lever 已提交
2933

2934 2935 2936 2937 2938 2939 2940 2941 2942 2943 2944 2945 2946 2947 2948 2949 2950 2951
	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;
		}
2952
	}
2953

2954 2955 2956 2957 2958 2959 2960 2961 2962
	/*
	 * -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;
2963 2964 2965
	return status;
}

C
Chuck Lever 已提交
2966 2967 2968 2969 2970 2971 2972 2973 2974 2975 2976 2977
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
2978
 * @auth_probe: probe the auth flavours
C
Chuck Lever 已提交
2979 2980
 *
 * Returns zero on success, or a negative errno.
2981
 */
2982
int nfs4_proc_get_rootfh(struct nfs_server *server, struct nfs_fh *fhandle,
2983 2984
			 struct nfs_fsinfo *info,
			 bool auth_probe)
2985
{
C
Chuck Lever 已提交
2986 2987
	int status;

2988 2989 2990 2991 2992 2993
	switch (auth_probe) {
	case false:
		status = nfs4_lookup_root(server, fhandle, info);
		if (status != -NFS4ERR_WRONGSEC)
			break;
	default:
C
Chuck Lever 已提交
2994
		status = nfs4_do_find_root_sec(server, fhandle, info);
2995
	}
C
Chuck Lever 已提交
2996

L
Linus Torvalds 已提交
2997 2998 2999 3000
	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
3001

3002
	return nfs4_map_errors(status);
L
Linus Torvalds 已提交
3003 3004
}

3005 3006 3007 3008 3009
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;
3010
	struct nfs4_label *label = NULL;
3011 3012 3013 3014 3015 3016 3017

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

3018 3019 3020 3021
	label = nfs4_label_alloc(server, GFP_KERNEL);
	if (IS_ERR(label))
		return PTR_ERR(label);

3022
	error = nfs4_proc_getattr(server, mntfh, fattr, label);
3023 3024
	if (error < 0) {
		dprintk("nfs4_get_root: getattr error = %d\n", -error);
3025
		goto err_free_label;
3026 3027 3028 3029 3030 3031
	}

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

3032 3033 3034
err_free_label:
	nfs4_label_free(label);

3035 3036 3037
	return error;
}

M
Manoj Naik 已提交
3038 3039 3040 3041 3042
/*
 * 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
 */
3043 3044 3045
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
3046 3047 3048 3049 3050 3051 3052 3053 3054 3055 3056 3057
{
	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;

3058
	status = nfs4_proc_fs_locations(client, dir, name, locations, page);
M
Manoj Naik 已提交
3059 3060
	if (status != 0)
		goto out;
3061 3062 3063 3064 3065 3066

	/*
	 * 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
3067
	if (nfs_fsid_equal(&NFS_SERVER(dir)->fsid, &locations->fattr.fsid)) {
3068 3069
		dprintk("%s: server did not return a different fsid for"
			" a referral at %s\n", __func__, name->name);
3070
		status = -NFS4ERR_MOVED;
M
Manoj Naik 已提交
3071 3072
		goto out;
	}
3073 3074
	/* Fixup attributes for the nfs_lookup() call to nfs_fhget() */
	nfs_fixup_referral_attributes(&locations->fattr);
M
Manoj Naik 已提交
3075

3076
	/* replace the lookup nfs_fattr with the locations nfs_fattr */
M
Manoj Naik 已提交
3077 3078 3079 3080 3081
	memcpy(fattr, &locations->fattr, sizeof(struct nfs_fattr));
	memset(fhandle, 0, sizeof(struct nfs_fh));
out:
	if (page)
		__free_page(page);
3082
	kfree(locations);
M
Manoj Naik 已提交
3083 3084 3085
	return status;
}

3086 3087
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
3088 3089 3090 3091 3092 3093 3094
{
	struct nfs4_getattr_arg args = {
		.fh = fhandle,
		.bitmask = server->attr_bitmask,
	};
	struct nfs4_getattr_res res = {
		.fattr = fattr,
3095
		.label = label,
L
Linus Torvalds 已提交
3096 3097 3098 3099 3100 3101 3102
		.server = server,
	};
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_GETATTR],
		.rpc_argp = &args,
		.rpc_resp = &res,
	};
3103 3104 3105

	args.bitmask = nfs4_bitmask(server, label);

3106
	nfs_fattr_init(fattr);
3107
	return nfs4_call_sync(server->client, server, &msg, &args.seq_args, &res.seq_res, 0);
L
Linus Torvalds 已提交
3108 3109
}

3110 3111
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
3112 3113 3114 3115
{
	struct nfs4_exception exception = { };
	int err;
	do {
3116 3117 3118
		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
3119 3120 3121 3122 3123 3124 3125 3126 3127 3128 3129 3130 3131 3132 3133 3134 3135 3136 3137 3138 3139 3140 3141 3142 3143 3144
				&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)
{
3145
	struct inode *inode = dentry->d_inode;
3146
	struct rpc_cred *cred = NULL;
3147
	struct nfs4_state *state = NULL;
3148
	struct nfs4_label *label = NULL;
L
Linus Torvalds 已提交
3149 3150
	int status;

B
Benny Halevy 已提交
3151
	if (pnfs_ld_layoutret_on_setattr(inode))
3152
		pnfs_commit_and_return_layout(inode);
B
Benny Halevy 已提交
3153

3154
	nfs_fattr_init(fattr);
L
Linus Torvalds 已提交
3155
	
3156 3157
	/* Deal with open(O_TRUNC) */
	if (sattr->ia_valid & ATTR_OPEN)
3158
		sattr->ia_valid &= ~(ATTR_MTIME|ATTR_CTIME);
3159 3160

	/* Optimization: if the end result is no change, don't RPC */
3161
	if ((sattr->ia_valid & ~(ATTR_FILE|ATTR_OPEN)) == 0)
3162 3163
		return 0;

3164
	/* Search for an existing open(O_WRITE) file */
3165 3166 3167 3168
	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
3169 3170 3171 3172
		if (ctx) {
			cred = ctx->cred;
			state = ctx->state;
		}
3173
	}
3174

3175 3176 3177 3178 3179
	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);
3180
	if (status == 0) {
3181
		nfs_setattr_update_inode(inode, sattr);
3182 3183
		nfs_setsecurity(inode, fattr, label);
	}
3184
	nfs4_label_free(label);
L
Linus Torvalds 已提交
3185 3186 3187
	return status;
}

3188 3189
static int _nfs4_proc_lookup(struct rpc_clnt *clnt, struct inode *dir,
		const struct qstr *name, struct nfs_fh *fhandle,
3190
		struct nfs_fattr *fattr, struct nfs4_label *label)
D
David Howells 已提交
3191
{
3192
	struct nfs_server *server = NFS_SERVER(dir);
D
David Howells 已提交
3193 3194 3195
	int		       status;
	struct nfs4_lookup_arg args = {
		.bitmask = server->attr_bitmask,
3196
		.dir_fh = NFS_FH(dir),
D
David Howells 已提交
3197 3198 3199 3200 3201
		.name = name,
	};
	struct nfs4_lookup_res res = {
		.server = server,
		.fattr = fattr,
3202
		.label = label,
D
David Howells 已提交
3203 3204 3205 3206 3207 3208 3209 3210
		.fh = fhandle,
	};
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_LOOKUP],
		.rpc_argp = &args,
		.rpc_resp = &res,
	};

3211 3212
	args.bitmask = nfs4_bitmask(server, label);

D
David Howells 已提交
3213 3214
	nfs_fattr_init(fattr);

L
Linus Torvalds 已提交
3215
	dprintk("NFS call  lookup %s\n", name->name);
3216
	status = nfs4_call_sync(clnt, server, &msg, &args.seq_args, &res.seq_res, 0);
L
Linus Torvalds 已提交
3217 3218 3219 3220
	dprintk("NFS reply lookup: %d\n", status);
	return status;
}

3221
static void nfs_fixup_secinfo_attributes(struct nfs_fattr *fattr)
3222 3223
{
	fattr->valid |= NFS_ATTR_FATTR_TYPE | NFS_ATTR_FATTR_MODE |
3224
		NFS_ATTR_FATTR_NLINK | NFS_ATTR_FATTR_MOUNTPOINT;
3225 3226 3227 3228
	fattr->mode = S_IFDIR | S_IRUGO | S_IXUGO;
	fattr->nlink = 2;
}

3229 3230
static int nfs4_proc_lookup_common(struct rpc_clnt **clnt, struct inode *dir,
				   struct qstr *name, struct nfs_fh *fhandle,
3231
				   struct nfs_fattr *fattr, struct nfs4_label *label)
L
Linus Torvalds 已提交
3232 3233
{
	struct nfs4_exception exception = { };
3234
	struct rpc_clnt *client = *clnt;
L
Linus Torvalds 已提交
3235 3236
	int err;
	do {
3237
		err = _nfs4_proc_lookup(client, dir, name, fhandle, fattr, label);
3238
		trace_nfs4_lookup(dir, name, err);
3239
		switch (err) {
3240
		case -NFS4ERR_BADNAME:
3241 3242
			err = -ENOENT;
			goto out;
3243
		case -NFS4ERR_MOVED:
3244
			err = nfs4_get_referral(client, dir, name, fattr, fhandle);
3245
			goto out;
3246
		case -NFS4ERR_WRONGSEC:
3247 3248 3249 3250 3251 3252 3253 3254 3255 3256 3257
			err = -EPERM;
			if (client != *clnt)
				goto out;
			client = nfs4_create_sec_client(client, dir, name);
			if (IS_ERR(client))
				return PTR_ERR(client);

			exception.retry = 1;
			break;
		default:
			err = nfs4_handle_exception(NFS_SERVER(dir), err, &exception);
3258
		}
L
Linus Torvalds 已提交
3259
	} while (exception.retry);
3260 3261 3262 3263 3264 3265 3266

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

L
Linus Torvalds 已提交
3267 3268 3269
	return err;
}

3270
static int nfs4_proc_lookup(struct inode *dir, struct qstr *name,
3271 3272
			    struct nfs_fh *fhandle, struct nfs_fattr *fattr,
			    struct nfs4_label *label)
3273 3274 3275 3276
{
	int status;
	struct rpc_clnt *client = NFS_CLIENT(dir);

3277
	status = nfs4_proc_lookup_common(&client, dir, name, fhandle, fattr, label);
3278 3279 3280 3281 3282 3283 3284
	if (client != NFS_CLIENT(dir)) {
		rpc_shutdown_client(client);
		nfs_fixup_secinfo_attributes(fattr);
	}
	return status;
}

3285 3286 3287 3288
struct rpc_clnt *
nfs4_proc_lookup_mountpoint(struct inode *dir, struct qstr *name,
			    struct nfs_fh *fhandle, struct nfs_fattr *fattr)
{
3289
	struct rpc_clnt *client = NFS_CLIENT(dir);
3290 3291
	int status;

3292
	status = nfs4_proc_lookup_common(&client, dir, name, fhandle, fattr, NULL);
3293
	if (status < 0)
3294
		return ERR_PTR(status);
3295
	return (client == NFS_CLIENT(dir)) ? rpc_clone_client(client) : client;
3296 3297
}

L
Linus Torvalds 已提交
3298 3299
static int _nfs4_proc_access(struct inode *inode, struct nfs_access_entry *entry)
{
3300
	struct nfs_server *server = NFS_SERVER(inode);
L
Linus Torvalds 已提交
3301 3302
	struct nfs4_accessargs args = {
		.fh = NFS_FH(inode),
3303
		.bitmask = server->cache_consistency_bitmask,
3304 3305 3306
	};
	struct nfs4_accessres res = {
		.server = server,
L
Linus Torvalds 已提交
3307 3308 3309 3310 3311 3312 3313 3314
	};
	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;
3315
	int status = 0;
L
Linus Torvalds 已提交
3316 3317 3318 3319 3320 3321 3322 3323 3324 3325 3326 3327 3328 3329 3330 3331 3332

	/*
	 * 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;
	}
3333 3334 3335 3336 3337

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

3338
	status = nfs4_call_sync(server->client, server, &msg, &args.seq_args, &res.seq_res, 0);
L
Linus Torvalds 已提交
3339
	if (!status) {
3340
		nfs_access_set_mask(entry, res.access);
3341
		nfs_refresh_inode(inode, res.fattr);
L
Linus Torvalds 已提交
3342
	}
3343
	nfs_free_fattr(res.fattr);
L
Linus Torvalds 已提交
3344 3345 3346 3347 3348 3349 3350 3351
	return status;
}

static int nfs4_proc_access(struct inode *inode, struct nfs_access_entry *entry)
{
	struct nfs4_exception exception = { };
	int err;
	do {
3352 3353 3354
		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
3355 3356 3357 3358 3359 3360 3361 3362 3363 3364 3365 3366 3367 3368 3369 3370 3371 3372 3373 3374 3375 3376
				&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
3377
 * we get the post-operation mtime and size.
L
Linus Torvalds 已提交
3378 3379 3380 3381 3382 3383 3384 3385 3386 3387 3388 3389 3390
 *
 * 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
3391
	struct nfs4_readlink_res res;
L
Linus Torvalds 已提交
3392 3393 3394
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_READLINK],
		.rpc_argp = &args,
B
Benny Halevy 已提交
3395
		.rpc_resp = &res,
L
Linus Torvalds 已提交
3396 3397
	};

3398
	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
3399 3400 3401 3402 3403 3404 3405 3406
}

static int nfs4_proc_readlink(struct inode *inode, struct page *page,
		unsigned int pgbase, unsigned int pglen)
{
	struct nfs4_exception exception = { };
	int err;
	do {
3407 3408 3409
		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
3410 3411 3412 3413 3414 3415
				&exception);
	} while (exception.retry);
	return err;
}

/*
3416
 * This is just for mknod.  open(O_CREAT) will always do ->open_context().
L
Linus Torvalds 已提交
3417 3418 3419
 */
static int
nfs4_proc_create(struct inode *dir, struct dentry *dentry, struct iattr *sattr,
3420
		 int flags)
L
Linus Torvalds 已提交
3421
{
3422
	struct nfs4_label l, *ilabel = NULL;
3423
	struct nfs_open_context *ctx;
L
Linus Torvalds 已提交
3424
	struct nfs4_state *state;
3425
	int opened = 0;
L
Linus Torvalds 已提交
3426 3427
	int status = 0;

3428 3429 3430 3431
	ctx = alloc_nfs_open_context(dentry, FMODE_READ);
	if (IS_ERR(ctx))
		return PTR_ERR(ctx);

3432 3433
	ilabel = nfs4_label_init_security(dir, dentry, sattr, &l);

A
Aneesh Kumar K.V 已提交
3434
	sattr->ia_mode &= ~current_umask();
3435
	state = nfs4_do_open(dir, ctx, flags, sattr, ilabel, &opened);
L
Linus Torvalds 已提交
3436 3437
	if (IS_ERR(state)) {
		status = PTR_ERR(state);
3438
		goto out;
L
Linus Torvalds 已提交
3439 3440
	}
out:
3441
	nfs4_label_release_security(ilabel);
3442
	put_nfs_open_context(ctx);
L
Linus Torvalds 已提交
3443 3444 3445 3446 3447
	return status;
}

static int _nfs4_proc_remove(struct inode *dir, struct qstr *name)
{
3448
	struct nfs_server *server = NFS_SERVER(dir);
3449
	struct nfs_removeargs args = {
L
Linus Torvalds 已提交
3450
		.fh = NFS_FH(dir),
3451
		.name = *name,
3452
	};
3453
	struct nfs_removeres res = {
3454
		.server = server,
L
Linus Torvalds 已提交
3455 3456
	};
	struct rpc_message msg = {
3457 3458 3459
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_REMOVE],
		.rpc_argp = &args,
		.rpc_resp = &res,
L
Linus Torvalds 已提交
3460
	};
3461
	int status;
L
Linus Torvalds 已提交
3462

3463
	status = nfs4_call_sync(server->client, server, &msg, &args.seq_args, &res.seq_res, 1);
3464
	if (status == 0)
3465
		update_changeattr(dir, &res.cinfo);
L
Linus Torvalds 已提交
3466 3467 3468 3469 3470 3471 3472 3473
	return status;
}

static int nfs4_proc_remove(struct inode *dir, struct qstr *name)
{
	struct nfs4_exception exception = { };
	int err;
	do {
3474 3475 3476
		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
3477 3478 3479 3480 3481
				&exception);
	} while (exception.retry);
	return err;
}

3482
static void nfs4_proc_unlink_setup(struct rpc_message *msg, struct inode *dir)
L
Linus Torvalds 已提交
3483
{
3484 3485 3486
	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
3487

3488
	res->server = server;
L
Linus Torvalds 已提交
3489
	msg->rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_REMOVE];
3490
	nfs4_init_sequence(&args->seq_args, &res->seq_res, 1);
3491 3492

	nfs_fattr_init(res->dir_attr);
L
Linus Torvalds 已提交
3493 3494
}

3495 3496
static void nfs4_proc_unlink_rpc_prepare(struct rpc_task *task, struct nfs_unlinkdata *data)
{
3497 3498 3499 3500
	nfs4_setup_sequence(NFS_SERVER(data->dir),
			&data->args.seq_args,
			&data->res.seq_res,
			task);
L
Linus Torvalds 已提交
3501 3502
}

3503
static int nfs4_proc_unlink_done(struct rpc_task *task, struct inode *dir)
L
Linus Torvalds 已提交
3504
{
3505 3506
	struct nfs_unlinkdata *data = task->tk_calldata;
	struct nfs_removeres *res = &data->res;
3507

3508 3509
	if (!nfs4_sequence_done(task, &res->seq_res))
		return 0;
3510
	if (nfs4_async_handle_error(task, res->server, NULL) == -EAGAIN)
3511 3512 3513
		return 0;
	update_changeattr(dir, &res->cinfo);
	return 1;
L
Linus Torvalds 已提交
3514 3515
}

3516 3517 3518 3519 3520 3521 3522 3523
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;
3524
	nfs4_init_sequence(&arg->seq_args, &res->seq_res, 1);
3525 3526
}

3527 3528
static void nfs4_proc_rename_rpc_prepare(struct rpc_task *task, struct nfs_renamedata *data)
{
3529 3530 3531 3532
	nfs4_setup_sequence(NFS_SERVER(data->old_dir),
			&data->args.seq_args,
			&data->res.seq_res,
			task);
3533 3534 3535 3536 3537
}

static int nfs4_proc_rename_done(struct rpc_task *task, struct inode *old_dir,
				 struct inode *new_dir)
{
3538 3539
	struct nfs_renamedata *data = task->tk_calldata;
	struct nfs_renameres *res = &data->res;
3540 3541 3542 3543 3544 3545 3546 3547 3548 3549 3550

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

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

L
Linus Torvalds 已提交
3551 3552
static int _nfs4_proc_link(struct inode *inode, struct inode *dir, struct qstr *name)
{
3553
	struct nfs_server *server = NFS_SERVER(inode);
L
Linus Torvalds 已提交
3554 3555 3556 3557
	struct nfs4_link_arg arg = {
		.fh     = NFS_FH(inode),
		.dir_fh = NFS_FH(dir),
		.name   = name,
3558 3559 3560 3561
		.bitmask = server->attr_bitmask,
	};
	struct nfs4_link_res res = {
		.server = server,
3562
		.label = NULL,
L
Linus Torvalds 已提交
3563 3564 3565 3566
	};
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_LINK],
		.rpc_argp = &arg,
3567
		.rpc_resp = &res,
L
Linus Torvalds 已提交
3568
	};
3569 3570 3571
	int status = -ENOMEM;

	res.fattr = nfs_alloc_fattr();
3572
	if (res.fattr == NULL)
3573
		goto out;
L
Linus Torvalds 已提交
3574

3575 3576 3577 3578 3579
	res.label = nfs4_label_alloc(server, GFP_KERNEL);
	if (IS_ERR(res.label)) {
		status = PTR_ERR(res.label);
		goto out;
	}
3580
	arg.bitmask = nfs4_bitmask(server, res.label);
3581

3582
	status = nfs4_call_sync(server->client, server, &msg, &arg.seq_args, &res.seq_res, 1);
3583 3584
	if (!status) {
		update_changeattr(dir, &res.cinfo);
3585 3586 3587
		status = nfs_post_op_update_inode(inode, res.fattr);
		if (!status)
			nfs_setsecurity(inode, res.fattr, res.label);
3588
	}
3589 3590 3591 3592


	nfs4_label_free(res.label);

3593 3594
out:
	nfs_free_fattr(res.fattr);
L
Linus Torvalds 已提交
3595 3596 3597 3598 3599 3600 3601 3602 3603 3604 3605 3606 3607 3608 3609
	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;
}

3610 3611 3612 3613 3614 3615
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;
3616
	struct nfs4_label *label;
3617 3618 3619 3620 3621 3622 3623 3624 3625 3626 3627
};

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);

3628 3629 3630 3631
		data->label = nfs4_label_alloc(server, GFP_KERNEL);
		if (IS_ERR(data->label))
			goto out_free;

3632 3633 3634 3635 3636 3637 3638 3639
		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;
3640
		data->arg.bitmask = nfs4_bitmask(server, data->label);
3641 3642 3643
		data->res.server = server;
		data->res.fh = &data->fh;
		data->res.fattr = &data->fattr;
3644
		data->res.label = data->label;
3645 3646 3647
		nfs_fattr_init(data->res.fattr);
	}
	return data;
3648 3649 3650
out_free:
	kfree(data);
	return NULL;
3651 3652 3653 3654
}

static int nfs4_do_create(struct inode *dir, struct dentry *dentry, struct nfs4_createdata *data)
{
3655
	int status = nfs4_call_sync(NFS_SERVER(dir)->client, NFS_SERVER(dir), &data->msg,
3656
				    &data->arg.seq_args, &data->res.seq_res, 1);
3657 3658
	if (status == 0) {
		update_changeattr(dir, &data->res.dir_cinfo);
3659
		status = nfs_instantiate(dentry, data->res.fh, data->res.fattr, data->res.label);
3660 3661 3662 3663 3664 3665
	}
	return status;
}

static void nfs4_free_createdata(struct nfs4_createdata *data)
{
3666
	nfs4_label_free(data->label);
3667 3668 3669
	kfree(data);
}

3670
static int _nfs4_proc_symlink(struct inode *dir, struct dentry *dentry,
3671 3672
		struct page *page, unsigned int len, struct iattr *sattr,
		struct nfs4_label *label)
L
Linus Torvalds 已提交
3673
{
3674 3675
	struct nfs4_createdata *data;
	int status = -ENAMETOOLONG;
L
Linus Torvalds 已提交
3676

3677
	if (len > NFS4_MAXPATHLEN)
3678
		goto out;
3679

3680 3681 3682 3683 3684 3685 3686 3687
	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;
3688
	data->arg.label = label;
L
Linus Torvalds 已提交
3689
	
3690 3691 3692 3693
	status = nfs4_do_create(dir, dentry, data);

	nfs4_free_createdata(data);
out:
L
Linus Torvalds 已提交
3694 3695 3696
	return status;
}

3697
static int nfs4_proc_symlink(struct inode *dir, struct dentry *dentry,
3698
		struct page *page, unsigned int len, struct iattr *sattr)
L
Linus Torvalds 已提交
3699 3700
{
	struct nfs4_exception exception = { };
3701
	struct nfs4_label l, *label = NULL;
L
Linus Torvalds 已提交
3702
	int err;
3703 3704 3705

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

L
Linus Torvalds 已提交
3706
	do {
3707 3708 3709
		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
3710 3711
				&exception);
	} while (exception.retry);
3712 3713

	nfs4_label_release_security(label);
L
Linus Torvalds 已提交
3714 3715 3716 3717
	return err;
}

static int _nfs4_proc_mkdir(struct inode *dir, struct dentry *dentry,
3718
		struct iattr *sattr, struct nfs4_label *label)
L
Linus Torvalds 已提交
3719
{
3720 3721
	struct nfs4_createdata *data;
	int status = -ENOMEM;
L
Linus Torvalds 已提交
3722

3723 3724 3725 3726
	data = nfs4_alloc_createdata(dir, &dentry->d_name, sattr, NF4DIR);
	if (data == NULL)
		goto out;

3727
	data->arg.label = label;
3728 3729 3730 3731
	status = nfs4_do_create(dir, dentry, data);

	nfs4_free_createdata(data);
out:
L
Linus Torvalds 已提交
3732 3733 3734 3735 3736 3737 3738
	return status;
}

static int nfs4_proc_mkdir(struct inode *dir, struct dentry *dentry,
		struct iattr *sattr)
{
	struct nfs4_exception exception = { };
3739
	struct nfs4_label l, *label = NULL;
L
Linus Torvalds 已提交
3740
	int err;
A
Aneesh Kumar K.V 已提交
3741

3742 3743
	label = nfs4_label_init_security(dir, dentry, sattr, &l);

A
Aneesh Kumar K.V 已提交
3744
	sattr->ia_mode &= ~current_umask();
L
Linus Torvalds 已提交
3745
	do {
3746 3747 3748
		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
3749 3750
				&exception);
	} while (exception.retry);
3751 3752
	nfs4_label_release_security(label);

L
Linus Torvalds 已提交
3753 3754 3755 3756
	return err;
}

static int _nfs4_proc_readdir(struct dentry *dentry, struct rpc_cred *cred,
B
Bryan Schumaker 已提交
3757
		u64 cookie, struct page **pages, unsigned int count, int plus)
L
Linus Torvalds 已提交
3758 3759 3760 3761
{
	struct inode		*dir = dentry->d_inode;
	struct nfs4_readdir_arg args = {
		.fh = NFS_FH(dir),
B
Bryan Schumaker 已提交
3762
		.pages = pages,
L
Linus Torvalds 已提交
3763 3764
		.pgbase = 0,
		.count = count,
3765
		.bitmask = NFS_SERVER(dentry->d_inode)->attr_bitmask,
B
Bryan Schumaker 已提交
3766
		.plus = plus,
L
Linus Torvalds 已提交
3767 3768 3769 3770 3771 3772 3773 3774 3775 3776
	};
	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;

3777 3778
	dprintk("%s: dentry = %pd2, cookie = %Lu\n", __func__,
			dentry,
3779
			(unsigned long long)cookie);
3780
	nfs4_setup_readdir(cookie, NFS_I(dir)->cookieverf, dentry, &args);
L
Linus Torvalds 已提交
3781
	res.pgbase = args.pgbase;
3782
	status = nfs4_call_sync(NFS_SERVER(dir)->client, NFS_SERVER(dir), &msg, &args.seq_args, &res.seq_res, 0);
3783
	if (status >= 0) {
3784
		memcpy(NFS_I(dir)->cookieverf, res.verifier.data, NFS4_VERIFIER_SIZE);
3785 3786
		status += args.pgbase;
	}
3787 3788 3789

	nfs_invalidate_atime(dir);

3790
	dprintk("%s: returns %d\n", __func__, status);
L
Linus Torvalds 已提交
3791 3792 3793 3794
	return status;
}

static int nfs4_proc_readdir(struct dentry *dentry, struct rpc_cred *cred,
B
Bryan Schumaker 已提交
3795
		u64 cookie, struct page **pages, unsigned int count, int plus)
L
Linus Torvalds 已提交
3796 3797 3798 3799
{
	struct nfs4_exception exception = { };
	int err;
	do {
3800 3801 3802 3803
		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
3804 3805 3806 3807 3808 3809
				&exception);
	} while (exception.retry);
	return err;
}

static int _nfs4_proc_mknod(struct inode *dir, struct dentry *dentry,
3810
		struct iattr *sattr, struct nfs4_label *label, dev_t rdev)
L
Linus Torvalds 已提交
3811
{
3812 3813 3814
	struct nfs4_createdata *data;
	int mode = sattr->ia_mode;
	int status = -ENOMEM;
L
Linus Torvalds 已提交
3815

3816 3817 3818 3819
	data = nfs4_alloc_createdata(dir, &dentry->d_name, sattr, NF4SOCK);
	if (data == NULL)
		goto out;

L
Linus Torvalds 已提交
3820
	if (S_ISFIFO(mode))
3821
		data->arg.ftype = NF4FIFO;
L
Linus Torvalds 已提交
3822
	else if (S_ISBLK(mode)) {
3823 3824 3825
		data->arg.ftype = NF4BLK;
		data->arg.u.device.specdata1 = MAJOR(rdev);
		data->arg.u.device.specdata2 = MINOR(rdev);
L
Linus Torvalds 已提交
3826 3827
	}
	else if (S_ISCHR(mode)) {
3828 3829 3830
		data->arg.ftype = NF4CHR;
		data->arg.u.device.specdata1 = MAJOR(rdev);
		data->arg.u.device.specdata2 = MINOR(rdev);
3831 3832 3833
	} else if (!S_ISSOCK(mode)) {
		status = -EINVAL;
		goto out_free;
L
Linus Torvalds 已提交
3834
	}
3835

3836
	data->arg.label = label;
3837
	status = nfs4_do_create(dir, dentry, data);
3838
out_free:
3839 3840
	nfs4_free_createdata(data);
out:
L
Linus Torvalds 已提交
3841 3842 3843 3844 3845 3846 3847
	return status;
}

static int nfs4_proc_mknod(struct inode *dir, struct dentry *dentry,
		struct iattr *sattr, dev_t rdev)
{
	struct nfs4_exception exception = { };
3848
	struct nfs4_label l, *label = NULL;
L
Linus Torvalds 已提交
3849
	int err;
A
Aneesh Kumar K.V 已提交
3850

3851 3852
	label = nfs4_label_init_security(dir, dentry, sattr, &l);

A
Aneesh Kumar K.V 已提交
3853
	sattr->ia_mode &= ~current_umask();
L
Linus Torvalds 已提交
3854
	do {
3855 3856 3857
		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
3858 3859
				&exception);
	} while (exception.retry);
3860 3861 3862

	nfs4_label_release_security(label);

L
Linus Torvalds 已提交
3863 3864 3865 3866 3867 3868 3869 3870 3871 3872
	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
3873 3874 3875
	struct nfs4_statfs_res res = {
		.fsstat = fsstat,
	};
L
Linus Torvalds 已提交
3876 3877 3878
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_STATFS],
		.rpc_argp = &args,
B
Benny Halevy 已提交
3879
		.rpc_resp = &res,
L
Linus Torvalds 已提交
3880 3881
	};

3882
	nfs_fattr_init(fsstat->fattr);
3883
	return  nfs4_call_sync(server->client, server, &msg, &args.seq_args, &res.seq_res, 0);
L
Linus Torvalds 已提交
3884 3885 3886 3887 3888 3889 3890 3891 3892 3893 3894 3895 3896 3897 3898 3899 3900 3901 3902 3903 3904
}

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
3905 3906 3907
	struct nfs4_fsinfo_res res = {
		.fsinfo = fsinfo,
	};
L
Linus Torvalds 已提交
3908 3909 3910
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_FSINFO],
		.rpc_argp = &args,
B
Benny Halevy 已提交
3911
		.rpc_resp = &res,
L
Linus Torvalds 已提交
3912 3913
	};

3914
	return nfs4_call_sync(server->client, server, &msg, &args.seq_args, &res.seq_res, 0);
L
Linus Torvalds 已提交
3915 3916 3917 3918 3919
}

static int nfs4_do_fsinfo(struct nfs_server *server, struct nfs_fh *fhandle, struct nfs_fsinfo *fsinfo)
{
	struct nfs4_exception exception = { };
3920
	unsigned long now = jiffies;
L
Linus Torvalds 已提交
3921 3922 3923
	int err;

	do {
3924
		err = _nfs4_do_fsinfo(server, fhandle, fsinfo);
3925
		trace_nfs4_fsinfo(server, fhandle, fsinfo->fattr, err);
3926 3927 3928 3929 3930 3931 3932 3933 3934 3935
		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
3936 3937 3938 3939 3940 3941
	} while (exception.retry);
	return err;
}

static int nfs4_proc_fsinfo(struct nfs_server *server, struct nfs_fh *fhandle, struct nfs_fsinfo *fsinfo)
{
3942 3943
	int error;

3944
	nfs_fattr_init(fsinfo->fattr);
3945
	error = nfs4_do_fsinfo(server, fhandle, fsinfo);
3946 3947 3948
	if (error == 0) {
		/* block layout checks this! */
		server->pnfs_blksize = fsinfo->blksize;
3949
		set_pnfs_layoutdriver(server, fhandle, fsinfo->layouttype);
3950
	}
3951 3952

	return error;
L
Linus Torvalds 已提交
3953 3954 3955 3956 3957 3958 3959 3960 3961
}

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
3962 3963 3964
	struct nfs4_pathconf_res res = {
		.pathconf = pathconf,
	};
L
Linus Torvalds 已提交
3965 3966 3967
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_PATHCONF],
		.rpc_argp = &args,
B
Benny Halevy 已提交
3968
		.rpc_resp = &res,
L
Linus Torvalds 已提交
3969 3970 3971 3972 3973 3974 3975 3976
	};

	/* 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;
	}

3977
	nfs_fattr_init(pathconf->fattr);
3978
	return nfs4_call_sync(server->client, server, &msg, &args.seq_args, &res.seq_res, 0);
L
Linus Torvalds 已提交
3979 3980 3981 3982 3983 3984 3985 3986 3987 3988 3989 3990 3991 3992 3993 3994
}

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;
}

3995
int nfs4_set_rw_stateid(nfs4_stateid *stateid,
3996 3997 3998 3999 4000 4001 4002 4003
		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;
4004
	return nfs4_select_rw_stateid(stateid, ctx->state, fmode, lockowner);
4005 4006 4007
}
EXPORT_SYMBOL_GPL(nfs4_set_rw_stateid);

4008 4009 4010 4011 4012 4013 4014
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;

4015 4016 4017
	/* If the current stateid represents a lost lock, then exit */
	if (nfs4_set_rw_stateid(&current_stateid, ctx, l_ctx, fmode) == -EIO)
		return true;
4018 4019 4020 4021 4022 4023 4024 4025 4026 4027 4028 4029 4030 4031 4032 4033 4034 4035
	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;
}

4036 4037
void __nfs4_read_done_cb(struct nfs_read_data *data)
{
4038
	nfs_invalidate_atime(data->header->inode);
4039 4040
}

4041
static int nfs4_read_done_cb(struct rpc_task *task, struct nfs_read_data *data)
L
Linus Torvalds 已提交
4042
{
4043
	struct nfs_server *server = NFS_SERVER(data->header->inode);
L
Linus Torvalds 已提交
4044

4045
	trace_nfs4_read(data, task->tk_status);
4046
	if (nfs4_async_handle_error(task, server, data->args.context->state) == -EAGAIN) {
4047
		rpc_restart_call_prepare(task);
T
Trond Myklebust 已提交
4048
		return -EAGAIN;
L
Linus Torvalds 已提交
4049
	}
4050

4051
	__nfs4_read_done_cb(data);
L
Linus Torvalds 已提交
4052
	if (task->tk_status > 0)
T
Trond Myklebust 已提交
4053 4054
		renew_lease(server, data->timestamp);
	return 0;
L
Linus Torvalds 已提交
4055 4056
}

4057 4058 4059 4060 4061 4062 4063 4064 4065 4066 4067 4068 4069 4070
static bool nfs4_read_stateid_changed(struct rpc_task *task,
		struct nfs_readargs *args)
{

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

4071 4072 4073 4074 4075 4076 4077
static int nfs4_read_done(struct rpc_task *task, struct nfs_read_data *data)
{

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

	if (!nfs4_sequence_done(task, &data->res.seq_res))
		return -EAGAIN;
4078 4079
	if (nfs4_read_stateid_changed(task, &data->args))
		return -EAGAIN;
4080 4081
	return data->read_done_cb ? data->read_done_cb(task, data) :
				    nfs4_read_done_cb(task, data);
4082 4083
}

4084
static void nfs4_proc_read_setup(struct nfs_read_data *data, struct rpc_message *msg)
L
Linus Torvalds 已提交
4085 4086
{
	data->timestamp   = jiffies;
4087
	data->read_done_cb = nfs4_read_done_cb;
4088
	msg->rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_READ];
4089
	nfs4_init_sequence(&data->args.seq_args, &data->res.seq_res, 0);
L
Linus Torvalds 已提交
4090 4091
}

4092
static int nfs4_proc_read_rpc_prepare(struct rpc_task *task, struct nfs_read_data *data)
4093
{
4094
	if (nfs4_setup_sequence(NFS_SERVER(data->header->inode),
4095 4096
			&data->args.seq_args,
			&data->res.seq_res,
4097
			task))
4098 4099 4100 4101 4102 4103 4104
		return 0;
	if (nfs4_set_rw_stateid(&data->args.stateid, data->args.context,
				data->args.lock_context, FMODE_READ) == -EIO)
		return -EIO;
	if (unlikely(test_bit(NFS_CONTEXT_BAD, &data->args.context->flags)))
		return -EIO;
	return 0;
L
Linus Torvalds 已提交
4105 4106
}

4107
static int nfs4_write_done_cb(struct rpc_task *task, struct nfs_write_data *data)
L
Linus Torvalds 已提交
4108
{
4109
	struct inode *inode = data->header->inode;
L
Linus Torvalds 已提交
4110
	
4111
	trace_nfs4_write(data, task->tk_status);
4112
	if (nfs4_async_handle_error(task, NFS_SERVER(inode), data->args.context->state) == -EAGAIN) {
4113
		rpc_restart_call_prepare(task);
4114
		return -EAGAIN;
L
Linus Torvalds 已提交
4115
	}
4116
	if (task->tk_status >= 0) {
L
Linus Torvalds 已提交
4117
		renew_lease(NFS_SERVER(inode), data->timestamp);
4118
		nfs_post_op_update_inode_force_wcc(inode, &data->fattr);
4119
	}
4120
	return 0;
L
Linus Torvalds 已提交
4121 4122
}

4123 4124 4125 4126 4127 4128 4129 4130 4131 4132 4133 4134 4135 4136
static bool nfs4_write_stateid_changed(struct rpc_task *task,
		struct nfs_writeargs *args)
{

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

4137 4138 4139 4140
static int nfs4_write_done(struct rpc_task *task, struct nfs_write_data *data)
{
	if (!nfs4_sequence_done(task, &data->res.seq_res))
		return -EAGAIN;
4141 4142
	if (nfs4_write_stateid_changed(task, &data->args))
		return -EAGAIN;
4143 4144
	return data->write_done_cb ? data->write_done_cb(task, data) :
		nfs4_write_done_cb(task, data);
4145 4146
}

4147 4148
static
bool nfs4_write_need_cache_consistency_data(const struct nfs_write_data *data)
4149
{
4150 4151 4152 4153 4154 4155 4156 4157
	const struct nfs_pgio_header *hdr = data->header;

	/* Don't request attributes for pNFS or O_DIRECT writes */
	if (data->ds_clp != NULL || hdr->dreq != NULL)
		return false;
	/* Otherwise, request attributes if and only if we don't hold
	 * a delegation
	 */
4158
	return nfs4_have_delegation(hdr->inode, FMODE_READ) == 0;
4159 4160
}

4161
static void nfs4_proc_write_setup(struct nfs_write_data *data, struct rpc_message *msg)
L
Linus Torvalds 已提交
4162
{
4163
	struct nfs_server *server = NFS_SERVER(data->header->inode);
4164

4165
	if (!nfs4_write_need_cache_consistency_data(data)) {
4166 4167 4168 4169
		data->args.bitmask = NULL;
		data->res.fattr = NULL;
	} else
		data->args.bitmask = server->cache_consistency_bitmask;
4170

4171 4172
	if (!data->write_done_cb)
		data->write_done_cb = nfs4_write_done_cb;
4173
	data->res.server = server;
L
Linus Torvalds 已提交
4174 4175
	data->timestamp   = jiffies;

4176
	msg->rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_WRITE];
4177
	nfs4_init_sequence(&data->args.seq_args, &data->res.seq_res, 1);
L
Linus Torvalds 已提交
4178 4179
}

4180
static int nfs4_proc_write_rpc_prepare(struct rpc_task *task, struct nfs_write_data *data)
4181
{
4182
	if (nfs4_setup_sequence(NFS_SERVER(data->header->inode),
4183 4184
			&data->args.seq_args,
			&data->res.seq_res,
4185
			task))
4186 4187 4188 4189 4190 4191 4192
		return 0;
	if (nfs4_set_rw_stateid(&data->args.stateid, data->args.context,
				data->args.lock_context, FMODE_WRITE) == -EIO)
		return -EIO;
	if (unlikely(test_bit(NFS_CONTEXT_BAD, &data->args.context->flags)))
		return -EIO;
	return 0;
L
Linus Torvalds 已提交
4193 4194
}

4195
static void nfs4_proc_commit_rpc_prepare(struct rpc_task *task, struct nfs_commit_data *data)
4196
{
4197 4198 4199 4200
	nfs4_setup_sequence(NFS_SERVER(data->inode),
			&data->args.seq_args,
			&data->res.seq_res,
			task);
L
Linus Torvalds 已提交
4201 4202
}

4203
static int nfs4_commit_done_cb(struct rpc_task *task, struct nfs_commit_data *data)
L
Linus Torvalds 已提交
4204 4205
{
	struct inode *inode = data->inode;
4206

4207
	trace_nfs4_commit(data, task->tk_status);
4208
	if (nfs4_async_handle_error(task, NFS_SERVER(inode), NULL) == -EAGAIN) {
4209
		rpc_restart_call_prepare(task);
4210
		return -EAGAIN;
L
Linus Torvalds 已提交
4211
	}
4212
	return 0;
L
Linus Torvalds 已提交
4213 4214
}

4215
static int nfs4_commit_done(struct rpc_task *task, struct nfs_commit_data *data)
4216 4217 4218
{
	if (!nfs4_sequence_done(task, &data->res.seq_res))
		return -EAGAIN;
4219
	return data->commit_done_cb(task, data);
4220 4221
}

4222
static void nfs4_proc_commit_setup(struct nfs_commit_data *data, struct rpc_message *msg)
L
Linus Torvalds 已提交
4223
{
4224
	struct nfs_server *server = NFS_SERVER(data->inode);
4225

4226 4227
	if (data->commit_done_cb == NULL)
		data->commit_done_cb = nfs4_commit_done_cb;
4228
	data->res.server = server;
4229
	msg->rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_COMMIT];
4230
	nfs4_init_sequence(&data->args.seq_args, &data->res.seq_res, 1);
L
Linus Torvalds 已提交
4231 4232
}

4233 4234 4235 4236 4237
struct nfs4_renewdata {
	struct nfs_client	*client;
	unsigned long		timestamp;
};

L
Linus Torvalds 已提交
4238 4239 4240 4241
/*
 * nfs4_proc_async_renew(): This is not one of the nfs_rpc_ops; it is a special
 * standalone procedure for queueing an asynchronous RENEW.
 */
4242
static void nfs4_renew_release(void *calldata)
4243
{
4244 4245
	struct nfs4_renewdata *data = calldata;
	struct nfs_client *clp = data->client;
4246

4247 4248 4249
	if (atomic_read(&clp->cl_count) > 1)
		nfs4_schedule_state_renewal(clp);
	nfs_put_client(clp);
4250
	kfree(data);
4251 4252
}

4253
static void nfs4_renew_done(struct rpc_task *task, void *calldata)
L
Linus Torvalds 已提交
4254
{
4255 4256 4257
	struct nfs4_renewdata *data = calldata;
	struct nfs_client *clp = data->client;
	unsigned long timestamp = data->timestamp;
L
Linus Torvalds 已提交
4258

4259
	trace_nfs4_renew_async(clp, task->tk_status);
4260 4261 4262 4263 4264 4265 4266
	switch (task->tk_status) {
	case 0:
		break;
	case -NFS4ERR_LEASE_MOVED:
		nfs4_schedule_lease_moved_recovery(clp);
		break;
	default:
4267
		/* Unless we're shutting down, schedule state recovery! */
4268 4269 4270
		if (test_bit(NFS_CS_RENEWD, &clp->cl_res_state) == 0)
			return;
		if (task->tk_status != NFS4ERR_CB_PATH_DOWN) {
4271
			nfs4_schedule_lease_recovery(clp);
4272 4273 4274
			return;
		}
		nfs4_schedule_path_down_recovery(clp);
L
Linus Torvalds 已提交
4275
	}
4276
	do_renew_lease(clp, timestamp);
L
Linus Torvalds 已提交
4277 4278
}

4279 4280
static const struct rpc_call_ops nfs4_renew_ops = {
	.rpc_call_done = nfs4_renew_done,
4281
	.rpc_release = nfs4_renew_release,
4282 4283
};

4284
static int nfs4_proc_async_renew(struct nfs_client *clp, struct rpc_cred *cred, unsigned renew_flags)
L
Linus Torvalds 已提交
4285 4286 4287 4288
{
	struct rpc_message msg = {
		.rpc_proc	= &nfs4_procedures[NFSPROC4_CLNT_RENEW],
		.rpc_argp	= clp,
4289
		.rpc_cred	= cred,
L
Linus Torvalds 已提交
4290
	};
4291
	struct nfs4_renewdata *data;
L
Linus Torvalds 已提交
4292

4293 4294
	if (renew_flags == 0)
		return 0;
4295 4296
	if (!atomic_inc_not_zero(&clp->cl_count))
		return -EIO;
4297
	data = kmalloc(sizeof(*data), GFP_NOFS);
4298 4299 4300 4301
	if (data == NULL)
		return -ENOMEM;
	data->client = clp;
	data->timestamp = jiffies;
4302
	return rpc_call_async(clp->cl_rpcclient, &msg, RPC_TASK_TIMEOUT,
4303
			&nfs4_renew_ops, data);
L
Linus Torvalds 已提交
4304 4305
}

4306
static int nfs4_proc_renew(struct nfs_client *clp, struct rpc_cred *cred)
L
Linus Torvalds 已提交
4307 4308 4309 4310
{
	struct rpc_message msg = {
		.rpc_proc	= &nfs4_procedures[NFSPROC4_CLNT_RENEW],
		.rpc_argp	= clp,
4311
		.rpc_cred	= cred,
L
Linus Torvalds 已提交
4312 4313 4314 4315
	};
	unsigned long now = jiffies;
	int status;

4316
	status = rpc_call_sync(clp->cl_rpcclient, &msg, RPC_TASK_TIMEOUT);
L
Linus Torvalds 已提交
4317 4318
	if (status < 0)
		return status;
4319
	do_renew_lease(clp, now);
L
Linus Torvalds 已提交
4320 4321 4322
	return 0;
}

4323 4324
static inline int nfs4_server_supports_acls(struct nfs_server *server)
{
4325
	return server->caps & NFS_CAP_ACLS;
4326 4327
}

4328 4329
/* 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
4330 4331
 * the stack.
 */
4332
#define NFS4ACL_MAXPAGES DIV_ROUND_UP(XATTR_SIZE_MAX, PAGE_SIZE)
4333

4334 4335 4336 4337 4338 4339 4340 4341 4342
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 {
4343
		len = min_t(size_t, PAGE_SIZE, buflen);
4344 4345 4346 4347 4348 4349 4350 4351 4352 4353 4354 4355 4356 4357 4358 4359 4360 4361 4362
		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;
}

4363 4364 4365
struct nfs4_cached_acl {
	int cached;
	size_t len;
A
Andrew Morton 已提交
4366
	char data[0];
4367 4368 4369 4370 4371 4372 4373 4374 4375 4376 4377 4378 4379 4380 4381 4382 4383 4384 4385 4386 4387 4388 4389 4390 4391 4392 4393 4394 4395 4396 4397 4398 4399 4400 4401 4402 4403 4404 4405 4406 4407 4408
};

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;
}

4409
static void nfs4_write_cached_acl(struct inode *inode, struct page **pages, size_t pgbase, size_t acl_len)
4410 4411
{
	struct nfs4_cached_acl *acl;
4412
	size_t buflen = sizeof(*acl) + acl_len;
4413

4414
	if (buflen <= PAGE_SIZE) {
4415
		acl = kmalloc(buflen, GFP_KERNEL);
4416 4417 4418
		if (acl == NULL)
			goto out;
		acl->cached = 1;
4419
		_copy_from_pages(acl->data, pages, pgbase, acl_len);
4420 4421 4422 4423 4424 4425 4426 4427 4428 4429 4430
	} 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);
}

4431 4432 4433 4434 4435 4436 4437 4438 4439 4440
/*
 * 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.
 */
4441
static ssize_t __nfs4_get_acl_uncached(struct inode *inode, void *buf, size_t buflen)
4442
{
4443
	struct page *pages[NFS4ACL_MAXPAGES] = {NULL, };
4444 4445 4446 4447 4448
	struct nfs_getaclargs args = {
		.fh = NFS_FH(inode),
		.acl_pages = pages,
		.acl_len = buflen,
	};
B
Benny Halevy 已提交
4449 4450 4451
	struct nfs_getaclres res = {
		.acl_len = buflen,
	};
4452 4453 4454
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_GETACL],
		.rpc_argp = &args,
B
Benny Halevy 已提交
4455
		.rpc_resp = &res,
4456
	};
4457 4458
	unsigned int npages = DIV_ROUND_UP(buflen, PAGE_SIZE);
	int ret = -ENOMEM, i;
4459

4460 4461 4462 4463
	/* 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;
4464 4465
	if (npages > ARRAY_SIZE(pages))
		return -ERANGE;
4466

4467 4468 4469 4470
	for (i = 0; i < npages; i++) {
		pages[i] = alloc_page(GFP_KERNEL);
		if (!pages[i])
			goto out_free;
4471
	}
4472 4473 4474 4475 4476 4477

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

4478 4479
	args.acl_len = npages * PAGE_SIZE;
	args.acl_pgbase = 0;
4480

4481
	dprintk("%s  buf %p buflen %zu npages %d args.acl_len %zu\n",
4482 4483 4484
		__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);
4485 4486
	if (ret)
		goto out_free;
4487

4488 4489 4490 4491 4492
	/* 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;
4493
		ret = -ERANGE;
4494
		goto out_free;
4495
	}
4496
	nfs4_write_cached_acl(inode, pages, res.acl_data_offset, res.acl_len);
4497 4498 4499 4500 4501
	if (buf) {
		if (res.acl_len > buflen) {
			ret = -ERANGE;
			goto out_free;
		}
4502
		_copy_from_pages(buf, pages, res.acl_data_offset, res.acl_len);
4503
	}
4504 4505
out_ok:
	ret = res.acl_len;
4506
out_free:
4507 4508 4509
	for (i = 0; i < npages; i++)
		if (pages[i])
			__free_page(pages[i]);
4510 4511
	if (res.acl_scratch)
		__free_page(res.acl_scratch);
4512 4513 4514
	return ret;
}

4515 4516 4517 4518 4519 4520
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);
4521
		trace_nfs4_get_acl(inode, ret);
4522 4523 4524 4525 4526 4527 4528
		if (ret >= 0)
			break;
		ret = nfs4_handle_exception(NFS_SERVER(inode), ret, &exception);
	} while (exception.retry);
	return ret;
}

4529 4530 4531 4532 4533 4534 4535 4536 4537 4538
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;
4539 4540
	if (NFS_I(inode)->cache_validity & NFS_INO_INVALID_ACL)
		nfs_zap_acl_cache(inode);
4541 4542
	ret = nfs4_read_cached_acl(inode, buf, buflen);
	if (ret != -ENOENT)
4543 4544
		/* -ENOENT is returned if there is no ACL or if there is an ACL
		 * but no cached acl data, just the acl length */
4545 4546 4547 4548
		return ret;
	return nfs4_get_acl_uncached(inode, buf, buflen);
}

4549
static int __nfs4_proc_set_acl(struct inode *inode, const void *buf, size_t buflen)
4550 4551 4552 4553 4554 4555 4556 4557
{
	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
4558
	struct nfs_setaclres res;
4559 4560 4561
	struct rpc_message msg = {
		.rpc_proc	= &nfs4_procedures[NFSPROC4_CLNT_SETACL],
		.rpc_argp	= &arg,
B
Benny Halevy 已提交
4562
		.rpc_resp	= &res,
4563
	};
4564
	unsigned int npages = DIV_ROUND_UP(buflen, PAGE_SIZE);
4565
	int ret, i;
4566 4567 4568

	if (!nfs4_server_supports_acls(server))
		return -EOPNOTSUPP;
4569 4570
	if (npages > ARRAY_SIZE(pages))
		return -ERANGE;
4571 4572 4573
	i = buf_to_pages_noslab(buf, buflen, arg.acl_pages, &arg.acl_pgbase);
	if (i < 0)
		return i;
4574
	nfs4_inode_return_delegation(inode);
4575
	ret = nfs4_call_sync(server->client, server, &msg, &arg.seq_args, &res.seq_res, 1);
4576 4577 4578 4579 4580 4581 4582 4583

	/*
	 * 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]);

4584 4585 4586 4587 4588 4589 4590
	/*
	 * 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);
4591 4592
	nfs_access_zap_cache(inode);
	nfs_zap_acl_cache(inode);
4593 4594 4595
	return ret;
}

4596 4597 4598 4599 4600
static int nfs4_proc_set_acl(struct inode *inode, const void *buf, size_t buflen)
{
	struct nfs4_exception exception = { };
	int err;
	do {
4601 4602 4603
		err = __nfs4_proc_set_acl(inode, buf, buflen);
		trace_nfs4_set_acl(inode, err);
		err = nfs4_handle_exception(NFS_SERVER(inode), err,
4604 4605 4606 4607 4608
				&exception);
	} while (exception.retry);
	return err;
}

4609 4610 4611 4612 4613 4614 4615 4616 4617
#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 };
4618
	struct nfs4_getattr_arg arg = {
4619 4620 4621 4622 4623 4624 4625 4626 4627 4628
		.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],
4629
		.rpc_argp	= &arg,
4630 4631 4632 4633 4634 4635
		.rpc_resp	= &res,
	};
	int ret;

	nfs_fattr_init(&fattr);

4636
	ret = nfs4_call_sync(server->client, server, &msg, &arg.seq_args, &res.seq_res, 0);
4637 4638 4639 4640 4641 4642 4643 4644 4645 4646 4647 4648 4649 4650 4651 4652 4653 4654 4655
	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 {
4656 4657 4658
		err = _nfs4_get_security_label(inode, buf, buflen);
		trace_nfs4_get_security_label(inode, err);
		err = nfs4_handle_exception(NFS_SERVER(inode), err,
4659 4660 4661 4662 4663 4664 4665 4666 4667 4668 4669 4670 4671 4672
				&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 };
4673
	struct nfs_setattrargs arg = {
4674 4675 4676 4677 4678 4679 4680 4681 4682 4683 4684 4685 4686
		.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],
4687
		.rpc_argp       = &arg,
4688 4689 4690 4691
		.rpc_resp       = &res,
	};
	int status;

4692
	nfs4_stateid_copy(&arg.stateid, &zero_stateid);
4693

4694
	status = nfs4_call_sync(server->client, server, &msg, &arg.seq_args, &res.seq_res, 1);
4695 4696 4697 4698 4699 4700 4701 4702 4703 4704 4705 4706 4707 4708 4709
	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 {
4710 4711 4712 4713
		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,
4714 4715 4716 4717 4718 4719 4720 4721 4722 4723 4724 4725 4726 4727 4728 4729 4730 4731 4732 4733 4734 4735 4736 4737 4738 4739 4740 4741 4742 4743 4744 4745 4746 4747 4748 4749 4750 4751 4752 4753 4754 4755 4756 4757 4758 4759
				&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
4760
static int
4761
nfs4_async_handle_error(struct rpc_task *task, const struct nfs_server *server, struct nfs4_state *state)
L
Linus Torvalds 已提交
4762
{
4763 4764 4765
	struct nfs_client *clp = server->nfs_client;

	if (task->tk_status >= 0)
L
Linus Torvalds 已提交
4766 4767
		return 0;
	switch(task->tk_status) {
4768
		case -NFS4ERR_DELEG_REVOKED:
4769 4770
		case -NFS4ERR_ADMIN_REVOKED:
		case -NFS4ERR_BAD_STATEID:
4771 4772 4773
			if (state == NULL)
				break;
			nfs_remove_bad_delegation(state->inode);
4774 4775 4776
		case -NFS4ERR_OPENMODE:
			if (state == NULL)
				break;
4777
			if (nfs4_schedule_stateid_recovery(server, state) < 0)
C
Chuck Lever 已提交
4778
				goto recovery_failed;
4779
			goto wait_on_recovery;
4780
		case -NFS4ERR_EXPIRED:
4781 4782
			if (state != NULL) {
				if (nfs4_schedule_stateid_recovery(server, state) < 0)
C
Chuck Lever 已提交
4783
					goto recovery_failed;
4784
			}
L
Linus Torvalds 已提交
4785
		case -NFS4ERR_STALE_STATEID:
4786
		case -NFS4ERR_STALE_CLIENTID:
4787 4788
			nfs4_schedule_lease_recovery(clp);
			goto wait_on_recovery;
4789 4790 4791 4792
		case -NFS4ERR_MOVED:
			if (nfs4_schedule_migration_recovery(server) < 0)
				goto recovery_failed;
			goto wait_on_recovery;
4793 4794 4795
		case -NFS4ERR_LEASE_MOVED:
			nfs4_schedule_lease_moved_recovery(clp);
			goto wait_on_recovery;
4796 4797 4798 4799 4800 4801 4802 4803 4804 4805
#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);
4806
			nfs4_schedule_session_recovery(clp->cl_session, task->tk_status);
4807
			goto wait_on_recovery;
4808
#endif /* CONFIG_NFS_V4_1 */
L
Linus Torvalds 已提交
4809
		case -NFS4ERR_DELAY:
4810
			nfs_inc_server_stats(server, NFSIOS_DELAY);
4811
		case -NFS4ERR_GRACE:
L
Linus Torvalds 已提交
4812
			rpc_delay(task, NFS4_POLL_RETRY_MAX);
4813
		case -NFS4ERR_RETRY_UNCACHED_REP:
L
Linus Torvalds 已提交
4814
		case -NFS4ERR_OLD_STATEID:
4815
			goto restart_call;
L
Linus Torvalds 已提交
4816 4817 4818
	}
	task->tk_status = nfs4_map_errors(task->tk_status);
	return 0;
C
Chuck Lever 已提交
4819
recovery_failed:
4820 4821
	task->tk_status = -EIO;
	return 0;
4822
wait_on_recovery:
4823 4824 4825
	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);
4826 4827
	if (test_bit(NFS_MIG_FAILED, &server->mig_status))
		goto recovery_failed;
4828
restart_call:
4829 4830
	task->tk_status = 0;
	return -EAGAIN;
L
Linus Torvalds 已提交
4831 4832
}

4833 4834
static void nfs4_init_boot_verifier(const struct nfs_client *clp,
				    nfs4_verifier *bootverf)
4835 4836 4837
{
	__be32 verf[2];

4838 4839 4840 4841
	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;
4842
		verf[1] = cpu_to_be32(NSEC_PER_SEC + 1);
4843
	} else {
4844
		struct nfs_net *nn = net_generic(clp->cl_net, nfs_net_id);
4845 4846
		verf[0] = cpu_to_be32(nn->boot_time.tv_sec);
		verf[1] = cpu_to_be32(nn->boot_time.tv_nsec);
4847
	}
4848 4849 4850
	memcpy(bootverf->data, verf, sizeof(bootverf->data));
}

4851 4852 4853 4854 4855 4856 4857 4858 4859 4860 4861 4862 4863 4864 4865 4866 4867 4868 4869 4870 4871
static unsigned int
nfs4_init_nonuniform_client_string(const struct nfs_client *clp,
				   char *buf, size_t len)
{
	unsigned int result;

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

static unsigned int
nfs4_init_uniform_client_string(const struct nfs_client *clp,
				char *buf, size_t len)
{
4872
	const char *nodename = clp->cl_rpcclient->cl_nodename;
4873 4874

	if (nfs4_client_id_uniquifier[0] != '\0')
4875 4876 4877 4878 4879
		return scnprintf(buf, len, "Linux NFSv%u.%u %s/%s",
				clp->rpc_ops->version,
				clp->cl_minorversion,
				nfs4_client_id_uniquifier,
				nodename);
4880 4881
	return scnprintf(buf, len, "Linux NFSv%u.%u %s",
				clp->rpc_ops->version, clp->cl_minorversion,
4882
				nodename);
4883 4884
}

4885 4886 4887 4888 4889 4890 4891 4892 4893 4894 4895 4896 4897 4898
/*
 * 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");
}

4899 4900 4901 4902 4903 4904 4905 4906 4907 4908
/**
 * 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.
 */
4909 4910 4911
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
4912 4913 4914 4915 4916
{
	nfs4_verifier sc_verifier;
	struct nfs4_setclientid setclientid = {
		.sc_verifier = &sc_verifier,
		.sc_prog = program,
A
Andy Adamson 已提交
4917
		.sc_cb_ident = clp->cl_cb_ident,
L
Linus Torvalds 已提交
4918 4919 4920 4921
	};
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_SETCLIENTID],
		.rpc_argp = &setclientid,
4922
		.rpc_resp = res,
4923
		.rpc_cred = cred,
L
Linus Torvalds 已提交
4924
	};
4925
	int status;
L
Linus Torvalds 已提交
4926

4927
	/* nfs_client_id4 */
4928
	nfs4_init_boot_verifier(clp, &sc_verifier);
4929 4930 4931 4932 4933 4934 4935 4936 4937 4938
	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));
4939
	/* cb_client4 */
4940 4941 4942 4943
	setclientid.sc_netid_len =
				nfs4_init_callback_netid(clp,
						setclientid.sc_netid,
						sizeof(setclientid.sc_netid));
4944
	setclientid.sc_uaddr_len = scnprintf(setclientid.sc_uaddr,
4945
				sizeof(setclientid.sc_uaddr), "%s.%u.%u",
L
Linus Torvalds 已提交
4946 4947
				clp->cl_ipaddr, port >> 8, port & 255);

4948 4949 4950 4951
	dprintk("NFS call  setclientid auth=%s, '%.*s'\n",
		clp->cl_rpcclient->cl_auth->au_ops->au_name,
		setclientid.sc_name_len, setclientid.sc_name);
	status = rpc_call_sync(clp->cl_rpcclient, &msg, RPC_TASK_TIMEOUT);
4952
	trace_nfs4_setclientid(clp, status);
4953 4954
	dprintk("NFS reply setclientid: %d\n", status);
	return status;
L
Linus Torvalds 已提交
4955 4956
}

4957 4958 4959 4960 4961 4962 4963 4964
/**
 * 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.
 */
4965
int nfs4_proc_setclientid_confirm(struct nfs_client *clp,
4966 4967
		struct nfs4_setclientid_res *arg,
		struct rpc_cred *cred)
L
Linus Torvalds 已提交
4968 4969 4970
{
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_SETCLIENTID_CONFIRM],
4971
		.rpc_argp = arg,
4972
		.rpc_cred = cred,
L
Linus Torvalds 已提交
4973 4974 4975
	};
	int status;

4976 4977 4978
	dprintk("NFS call  setclientid_confirm auth=%s, (client ID %llx)\n",
		clp->cl_rpcclient->cl_auth->au_ops->au_name,
		clp->cl_clientid);
4979
	status = rpc_call_sync(clp->cl_rpcclient, &msg, RPC_TASK_TIMEOUT);
4980
	trace_nfs4_setclientid_confirm(clp, status);
4981
	dprintk("NFS reply setclientid_confirm: %d\n", status);
L
Linus Torvalds 已提交
4982 4983 4984
	return status;
}

4985 4986
struct nfs4_delegreturndata {
	struct nfs4_delegreturnargs args;
4987
	struct nfs4_delegreturnres res;
4988 4989
	struct nfs_fh fh;
	nfs4_stateid stateid;
4990
	unsigned long timestamp;
4991
	struct nfs_fattr fattr;
4992 4993 4994 4995 4996 4997
	int rpc_status;
};

static void nfs4_delegreturn_done(struct rpc_task *task, void *calldata)
{
	struct nfs4_delegreturndata *data = calldata;
4998

4999 5000
	if (!nfs4_sequence_done(task, &data->res.seq_res))
		return;
5001

5002
	trace_nfs4_delegreturn_exit(&data->args, &data->res, task->tk_status);
5003 5004
	switch (task->tk_status) {
	case 0:
5005
		renew_lease(data->res.server, data->timestamp);
5006
		break;
5007 5008 5009 5010 5011 5012 5013 5014
	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;
		break;
5015 5016 5017
	default:
		if (nfs4_async_handle_error(task, data->res.server, NULL) ==
				-EAGAIN) {
5018
			rpc_restart_call_prepare(task);
5019 5020 5021 5022
			return;
		}
	}
	data->rpc_status = task->tk_status;
5023 5024 5025 5026 5027 5028 5029
}

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

5030 5031 5032 5033 5034 5035
static void nfs4_delegreturn_prepare(struct rpc_task *task, void *data)
{
	struct nfs4_delegreturndata *d_data;

	d_data = (struct nfs4_delegreturndata *)data;

5036 5037 5038 5039
	nfs4_setup_sequence(d_data->res.server,
			&d_data->args.seq_args,
			&d_data->res.seq_res,
			task);
5040 5041
}

5042
static const struct rpc_call_ops nfs4_delegreturn_ops = {
5043
	.rpc_call_prepare = nfs4_delegreturn_prepare,
5044 5045 5046 5047
	.rpc_call_done = nfs4_delegreturn_done,
	.rpc_release = nfs4_delegreturn_release,
};

5048
static int _nfs4_proc_delegreturn(struct inode *inode, struct rpc_cred *cred, const nfs4_stateid *stateid, int issync)
5049 5050
{
	struct nfs4_delegreturndata *data;
5051
	struct nfs_server *server = NFS_SERVER(inode);
5052
	struct rpc_task *task;
5053 5054 5055 5056
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_DELEGRETURN],
		.rpc_cred = cred,
	};
T
Trond Myklebust 已提交
5057 5058
	struct rpc_task_setup task_setup_data = {
		.rpc_client = server->client,
5059
		.rpc_message = &msg,
T
Trond Myklebust 已提交
5060 5061 5062
		.callback_ops = &nfs4_delegreturn_ops,
		.flags = RPC_TASK_ASYNC,
	};
5063
	int status = 0;
5064

5065
	data = kzalloc(sizeof(*data), GFP_NOFS);
5066 5067
	if (data == NULL)
		return -ENOMEM;
5068
	nfs4_init_sequence(&data->args.seq_args, &data->res.seq_res, 1);
5069 5070
	data->args.fhandle = &data->fh;
	data->args.stateid = &data->stateid;
5071
	data->args.bitmask = server->cache_consistency_bitmask;
5072
	nfs_copy_fh(&data->fh, NFS_FH(inode));
5073
	nfs4_stateid_copy(&data->stateid, stateid);
5074 5075
	data->res.fattr = &data->fattr;
	data->res.server = server;
5076
	nfs_fattr_init(data->res.fattr);
5077
	data->timestamp = jiffies;
5078 5079
	data->rpc_status = 0;

T
Trond Myklebust 已提交
5080
	task_setup_data.callback_data = data;
5081 5082
	msg.rpc_argp = &data->args;
	msg.rpc_resp = &data->res;
T
Trond Myklebust 已提交
5083
	task = rpc_run_task(&task_setup_data);
5084
	if (IS_ERR(task))
5085
		return PTR_ERR(task);
5086 5087
	if (!issync)
		goto out;
5088
	status = nfs4_wait_for_completion_rpc_task(task);
5089 5090 5091
	if (status != 0)
		goto out;
	status = data->rpc_status;
5092 5093 5094 5095
	if (status == 0)
		nfs_post_op_update_inode_force_wcc(inode, &data->fattr);
	else
		nfs_refresh_inode(inode, &data->fattr);
5096
out:
5097
	rpc_put_task(task);
5098
	return status;
L
Linus Torvalds 已提交
5099 5100
}

5101
int nfs4_proc_delegreturn(struct inode *inode, struct rpc_cred *cred, const nfs4_stateid *stateid, int issync)
L
Linus Torvalds 已提交
5102 5103 5104 5105 5106
{
	struct nfs_server *server = NFS_SERVER(inode);
	struct nfs4_exception exception = { };
	int err;
	do {
5107
		err = _nfs4_proc_delegreturn(inode, cred, stateid, issync);
5108
		trace_nfs4_delegreturn(inode, err);
L
Linus Torvalds 已提交
5109 5110 5111 5112 5113 5114 5115 5116 5117 5118 5119 5120 5121 5122 5123 5124 5125 5126 5127 5128
		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)
{
5129
	freezable_schedule_timeout_killable_unsafe(timeout);
L
Linus Torvalds 已提交
5130 5131 5132 5133 5134 5135 5136 5137 5138 5139
	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);
5140
	struct nfs_client *clp = server->nfs_client;
T
Trond Myklebust 已提交
5141
	struct nfs_lockt_args arg = {
L
Linus Torvalds 已提交
5142
		.fh = NFS_FH(inode),
T
Trond Myklebust 已提交
5143
		.fl = request,
L
Linus Torvalds 已提交
5144
	};
T
Trond Myklebust 已提交
5145 5146
	struct nfs_lockt_res res = {
		.denied = request,
L
Linus Torvalds 已提交
5147 5148 5149 5150 5151 5152 5153 5154 5155 5156
	};
	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
5157
	arg.lock_owner.clientid = clp->cl_clientid;
5158 5159 5160 5161
	status = nfs4_set_lock_state(state, request);
	if (status != 0)
		goto out;
	lsp = request->fl_u.nfs4_fl.owner;
5162
	arg.lock_owner.id = lsp->ls_seqid.owner_id;
5163
	arg.lock_owner.s_dev = server->s_dev;
5164
	status = nfs4_call_sync(server->client, server, &msg, &arg.seq_args, &res.seq_res, 1);
T
Trond Myklebust 已提交
5165 5166 5167 5168 5169 5170
	switch (status) {
		case 0:
			request->fl_type = F_UNLCK;
			break;
		case -NFS4ERR_DENIED:
			status = 0;
L
Linus Torvalds 已提交
5171
	}
5172
	request->fl_ops->fl_release_private(request);
5173
	request->fl_ops = NULL;
5174
out:
L
Linus Torvalds 已提交
5175 5176 5177 5178 5179 5180 5181 5182 5183
	return status;
}

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

	do {
5184 5185 5186
		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
5187 5188 5189 5190 5191 5192 5193 5194 5195 5196 5197 5198 5199 5200 5201 5202 5203 5204 5205 5206 5207
				&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;
}

5208
struct nfs4_unlockdata {
T
Trond Myklebust 已提交
5209 5210
	struct nfs_locku_args arg;
	struct nfs_locku_res res;
5211 5212
	struct nfs4_lock_state *lsp;
	struct nfs_open_context *ctx;
T
Trond Myklebust 已提交
5213 5214
	struct file_lock fl;
	const struct nfs_server *server;
5215
	unsigned long timestamp;
5216 5217
};

T
Trond Myklebust 已提交
5218 5219 5220 5221 5222 5223 5224 5225
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;

5226
	p = kzalloc(sizeof(*p), GFP_NOFS);
T
Trond Myklebust 已提交
5227 5228 5229 5230 5231
	if (p == NULL)
		return NULL;
	p->arg.fh = NFS_FH(inode);
	p->arg.fl = &p->fl;
	p->arg.seqid = seqid;
5232
	p->res.seqid = seqid;
T
Trond Myklebust 已提交
5233 5234 5235 5236 5237 5238 5239 5240 5241 5242
	p->arg.stateid = &lsp->ls_stateid;
	p->lsp = lsp;
	atomic_inc(&lsp->ls_count);
	/* Ensure we don't close file until we're done freeing locks! */
	p->ctx = get_nfs_open_context(ctx);
	memcpy(&p->fl, fl, sizeof(p->fl));
	p->server = NFS_SERVER(inode);
	return p;
}

5243
static void nfs4_locku_release_calldata(void *data)
5244
{
5245
	struct nfs4_unlockdata *calldata = data;
T
Trond Myklebust 已提交
5246
	nfs_free_seqid(calldata->arg.seqid);
5247 5248 5249
	nfs4_put_lock_state(calldata->lsp);
	put_nfs_open_context(calldata->ctx);
	kfree(calldata);
5250 5251
}

5252
static void nfs4_locku_done(struct rpc_task *task, void *data)
5253
{
5254
	struct nfs4_unlockdata *calldata = data;
5255

5256 5257
	if (!nfs4_sequence_done(task, &calldata->res.seq_res))
		return;
5258 5259
	switch (task->tk_status) {
		case 0:
5260 5261
			nfs4_stateid_copy(&calldata->lsp->ls_stateid,
					&calldata->res.stateid);
5262
			renew_lease(calldata->server, calldata->timestamp);
5263
			break;
5264 5265
		case -NFS4ERR_BAD_STATEID:
		case -NFS4ERR_OLD_STATEID:
5266 5267 5268 5269
		case -NFS4ERR_STALE_STATEID:
		case -NFS4ERR_EXPIRED:
			break;
		default:
5270
			if (nfs4_async_handle_error(task, calldata->server, NULL) == -EAGAIN)
5271
				rpc_restart_call_prepare(task);
5272
	}
5273
	nfs_release_seqid(calldata->arg.seqid);
5274 5275
}

T
Trond Myklebust 已提交
5276
static void nfs4_locku_prepare(struct rpc_task *task, void *data)
5277
{
T
Trond Myklebust 已提交
5278
	struct nfs4_unlockdata *calldata = data;
5279

T
Trond Myklebust 已提交
5280
	if (nfs_wait_on_sequence(calldata->arg.seqid, task) != 0)
5281
		goto out_wait;
5282
	if (test_bit(NFS_LOCK_INITIALIZED, &calldata->lsp->ls_flags) == 0) {
5283
		/* Note: exit _without_ running nfs4_locku_done */
5284
		goto out_no_action;
5285
	}
5286
	calldata->timestamp = jiffies;
5287
	if (nfs4_setup_sequence(calldata->server,
5288
				&calldata->arg.seq_args,
5289 5290 5291
				&calldata->res.seq_res,
				task) != 0)
		nfs_release_seqid(calldata->arg.seqid);
5292 5293 5294 5295 5296
	return;
out_no_action:
	task->tk_action = NULL;
out_wait:
	nfs4_sequence_done(task, &calldata->res.seq_res);
5297 5298
}

5299
static const struct rpc_call_ops nfs4_locku_ops = {
T
Trond Myklebust 已提交
5300
	.rpc_call_prepare = nfs4_locku_prepare,
5301
	.rpc_call_done = nfs4_locku_done,
5302
	.rpc_release = nfs4_locku_release_calldata,
5303 5304
};

5305 5306 5307 5308 5309 5310
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;
5311 5312 5313 5314
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_LOCKU],
		.rpc_cred = ctx->cred,
	};
T
Trond Myklebust 已提交
5315 5316
	struct rpc_task_setup task_setup_data = {
		.rpc_client = NFS_CLIENT(lsp->ls_state->inode),
5317
		.rpc_message = &msg,
T
Trond Myklebust 已提交
5318
		.callback_ops = &nfs4_locku_ops,
5319
		.workqueue = nfsiod_workqueue,
T
Trond Myklebust 已提交
5320 5321
		.flags = RPC_TASK_ASYNC,
	};
5322

5323 5324 5325
	nfs4_state_protect(NFS_SERVER(lsp->ls_state->inode)->nfs_client,
		NFS_SP4_MACH_CRED_CLEANUP, &task_setup_data.rpc_client, &msg);

5326 5327 5328 5329 5330
	/* 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;

5331 5332 5333 5334 5335 5336
	data = nfs4_alloc_unlockdata(fl, ctx, lsp, seqid);
	if (data == NULL) {
		nfs_free_seqid(seqid);
		return ERR_PTR(-ENOMEM);
	}

5337
	nfs4_init_sequence(&data->arg.seq_args, &data->res.seq_res, 1);
5338 5339
	msg.rpc_argp = &data->arg;
	msg.rpc_resp = &data->res;
T
Trond Myklebust 已提交
5340 5341
	task_setup_data.callback_data = data;
	return rpc_run_task(&task_setup_data);
5342 5343
}

5344 5345
static int nfs4_proc_unlck(struct nfs4_state *state, int cmd, struct file_lock *request)
{
5346 5347 5348
	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
5349
	struct nfs_seqid *seqid;
L
Linus Torvalds 已提交
5350
	struct nfs4_lock_state *lsp;
5351 5352
	struct rpc_task *task;
	int status = 0;
5353
	unsigned char fl_flags = request->fl_flags;
5354

5355
	status = nfs4_set_lock_state(state, request);
5356 5357
	/* Unlock _before_ we do the RPC call */
	request->fl_flags |= FL_EXISTS;
5358 5359 5360
	/* Exclude nfs_delegation_claim_locks() */
	mutex_lock(&sp->so_delegreturn_mutex);
	/* Exclude nfs4_reclaim_open_stateid() - note nesting! */
5361 5362 5363
	down_read(&nfsi->rwsem);
	if (do_vfs_lock(request->fl_file, request) == -ENOENT) {
		up_read(&nfsi->rwsem);
5364
		mutex_unlock(&sp->so_delegreturn_mutex);
5365
		goto out;
5366 5367
	}
	up_read(&nfsi->rwsem);
5368
	mutex_unlock(&sp->so_delegreturn_mutex);
5369
	if (status != 0)
5370 5371
		goto out;
	/* Is this a delegated lock? */
5372
	lsp = request->fl_u.nfs4_fl.owner;
5373 5374
	if (test_bit(NFS_LOCK_INITIALIZED, &lsp->ls_flags) == 0)
		goto out;
5375
	seqid = nfs_alloc_seqid(&lsp->ls_seqid, GFP_KERNEL);
5376
	status = -ENOMEM;
T
Trond Myklebust 已提交
5377
	if (seqid == NULL)
5378
		goto out;
5379
	task = nfs4_do_unlck(request, nfs_file_open_context(request->fl_file), lsp, seqid);
5380 5381
	status = PTR_ERR(task);
	if (IS_ERR(task))
5382
		goto out;
5383
	status = nfs4_wait_for_completion_rpc_task(task);
5384
	rpc_put_task(task);
5385
out:
5386
	request->fl_flags = fl_flags;
5387
	trace_nfs4_unlock(request, state, F_SETLK, status);
L
Linus Torvalds 已提交
5388 5389 5390
	return status;
}

5391 5392 5393 5394 5395 5396
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;
5397
	unsigned long timestamp;
5398 5399
	int rpc_status;
	int cancelled;
5400
	struct nfs_server *server;
5401 5402 5403
};

static struct nfs4_lockdata *nfs4_alloc_lockdata(struct file_lock *fl,
5404 5405
		struct nfs_open_context *ctx, struct nfs4_lock_state *lsp,
		gfp_t gfp_mask)
L
Linus Torvalds 已提交
5406
{
5407 5408
	struct nfs4_lockdata *p;
	struct inode *inode = lsp->ls_state->inode;
L
Linus Torvalds 已提交
5409
	struct nfs_server *server = NFS_SERVER(inode);
5410

5411
	p = kzalloc(sizeof(*p), gfp_mask);
5412 5413 5414 5415 5416
	if (p == NULL)
		return NULL;

	p->arg.fh = NFS_FH(inode);
	p->arg.fl = &p->fl;
5417
	p->arg.open_seqid = nfs_alloc_seqid(&lsp->ls_state->owner->so_seqid, gfp_mask);
5418 5419
	if (p->arg.open_seqid == NULL)
		goto out_free;
5420
	p->arg.lock_seqid = nfs_alloc_seqid(&lsp->ls_seqid, gfp_mask);
5421
	if (p->arg.lock_seqid == NULL)
5422
		goto out_free_seqid;
5423
	p->arg.lock_stateid = &lsp->ls_stateid;
5424
	p->arg.lock_owner.clientid = server->nfs_client->cl_clientid;
5425
	p->arg.lock_owner.id = lsp->ls_seqid.owner_id;
5426
	p->arg.lock_owner.s_dev = server->s_dev;
5427
	p->res.lock_seqid = p->arg.lock_seqid;
5428
	p->lsp = lsp;
5429
	p->server = server;
5430 5431 5432 5433
	atomic_inc(&lsp->ls_count);
	p->ctx = get_nfs_open_context(ctx);
	memcpy(&p->fl, fl, sizeof(p->fl));
	return p;
5434 5435
out_free_seqid:
	nfs_free_seqid(p->arg.open_seqid);
5436 5437 5438 5439 5440 5441 5442 5443 5444
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;
5445

5446
	dprintk("%s: begin!\n", __func__);
5447
	if (nfs_wait_on_sequence(data->arg.lock_seqid, task) != 0)
5448
		goto out_wait;
5449 5450
	/* Do we need to do an open_to_lock_owner? */
	if (!(data->arg.lock_seqid->sequence->flags & NFS_SEQID_CONFIRMED)) {
5451
		if (nfs_wait_on_sequence(data->arg.open_seqid, task) != 0) {
5452
			goto out_release_lock_seqid;
5453
		}
5454
		data->arg.open_stateid = &state->open_stateid;
5455
		data->arg.new_lock_owner = 1;
5456
		data->res.open_seqid = data->arg.open_seqid;
5457 5458
	} else
		data->arg.new_lock_owner = 0;
5459 5460 5461 5462 5463
	if (!nfs4_valid_open_stateid(state)) {
		data->rpc_status = -EBADF;
		task->tk_action = NULL;
		goto out_release_open_seqid;
	}
5464
	data->timestamp = jiffies;
5465 5466
	if (nfs4_setup_sequence(data->server,
				&data->arg.seq_args,
5467
				&data->res.seq_res,
5468
				task) == 0)
5469
		return;
5470
out_release_open_seqid:
5471 5472 5473
	nfs_release_seqid(data->arg.open_seqid);
out_release_lock_seqid:
	nfs_release_seqid(data->arg.lock_seqid);
5474 5475
out_wait:
	nfs4_sequence_done(task, &data->res.seq_res);
5476
	dprintk("%s: done!, ret = %d\n", __func__, data->rpc_status);
5477 5478
}

5479 5480 5481 5482
static void nfs4_lock_done(struct rpc_task *task, void *calldata)
{
	struct nfs4_lockdata *data = calldata;

5483
	dprintk("%s: begin!\n", __func__);
5484

5485 5486
	if (!nfs4_sequence_done(task, &data->res.seq_res))
		return;
5487

5488 5489 5490 5491 5492 5493 5494 5495
	data->rpc_status = task->tk_status;
	if (data->arg.new_lock_owner != 0) {
		if (data->rpc_status == 0)
			nfs_confirm_seqid(&data->lsp->ls_seqid, 0);
		else
			goto out;
	}
	if (data->rpc_status == 0) {
5496
		nfs4_stateid_copy(&data->lsp->ls_stateid, &data->res.stateid);
5497
		set_bit(NFS_LOCK_INITIALIZED, &data->lsp->ls_flags);
5498
		renew_lease(NFS_SERVER(data->ctx->dentry->d_inode), data->timestamp);
5499 5500
	}
out:
5501
	dprintk("%s: done, ret = %d!\n", __func__, data->rpc_status);
5502 5503 5504 5505 5506 5507
}

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

5508
	dprintk("%s: begin!\n", __func__);
5509
	nfs_free_seqid(data->arg.open_seqid);
5510 5511 5512 5513 5514
	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))
5515
			rpc_put_task_async(task);
5516
		dprintk("%s: cancelling lock!\n", __func__);
5517 5518 5519 5520 5521
	} else
		nfs_free_seqid(data->arg.lock_seqid);
	nfs4_put_lock_state(data->lsp);
	put_nfs_open_context(data->ctx);
	kfree(data);
5522
	dprintk("%s: done!\n", __func__);
5523 5524 5525 5526 5527 5528 5529 5530
}

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,
};

5531 5532 5533 5534 5535
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:
5536
		lsp->ls_seqid.flags &= ~NFS_SEQID_CONFIRMED;
5537
		if (new_lock_owner != 0 ||
5538
		   test_bit(NFS_LOCK_INITIALIZED, &lsp->ls_flags) != 0)
5539
			nfs4_schedule_stateid_recovery(server, lsp->ls_state);
5540 5541 5542
		break;
	case -NFS4ERR_STALE_STATEID:
		lsp->ls_seqid.flags &= ~NFS_SEQID_CONFIRMED;
5543 5544
	case -NFS4ERR_EXPIRED:
		nfs4_schedule_lease_recovery(server->nfs_client);
5545 5546 5547
	};
}

5548
static int _nfs4_do_setlk(struct nfs4_state *state, int cmd, struct file_lock *fl, int recovery_type)
5549 5550 5551
{
	struct nfs4_lockdata *data;
	struct rpc_task *task;
5552 5553 5554 5555
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_LOCK],
		.rpc_cred = state->owner->so_cred,
	};
T
Trond Myklebust 已提交
5556 5557
	struct rpc_task_setup task_setup_data = {
		.rpc_client = NFS_CLIENT(state->inode),
5558
		.rpc_message = &msg,
T
Trond Myklebust 已提交
5559
		.callback_ops = &nfs4_lock_ops,
5560
		.workqueue = nfsiod_workqueue,
T
Trond Myklebust 已提交
5561 5562
		.flags = RPC_TASK_ASYNC,
	};
5563 5564
	int ret;

5565
	dprintk("%s: begin!\n", __func__);
5566
	data = nfs4_alloc_lockdata(fl, nfs_file_open_context(fl->fl_file),
5567 5568
			fl->fl_u.nfs4_fl.owner,
			recovery_type == NFS_LOCK_NEW ? GFP_KERNEL : GFP_NOFS);
5569 5570 5571 5572
	if (data == NULL)
		return -ENOMEM;
	if (IS_SETLKW(cmd))
		data->arg.block = 1;
5573
	nfs4_init_sequence(&data->arg.seq_args, &data->res.seq_res, 1);
5574 5575
	msg.rpc_argp = &data->arg;
	msg.rpc_resp = &data->res;
T
Trond Myklebust 已提交
5576
	task_setup_data.callback_data = data;
5577 5578 5579 5580 5581
	if (recovery_type > NFS_LOCK_NEW) {
		if (recovery_type == NFS_LOCK_RECLAIM)
			data->arg.reclaim = NFS_LOCK_RECLAIM;
		nfs4_set_sequence_privileged(&data->arg.seq_args);
	}
T
Trond Myklebust 已提交
5582
	task = rpc_run_task(&task_setup_data);
5583
	if (IS_ERR(task))
5584 5585 5586 5587
		return PTR_ERR(task);
	ret = nfs4_wait_for_completion_rpc_task(task);
	if (ret == 0) {
		ret = data->rpc_status;
5588 5589 5590
		if (ret)
			nfs4_handle_setlk_error(data->server, data->lsp,
					data->arg.new_lock_owner, ret);
5591 5592
	} else
		data->cancelled = 1;
5593
	rpc_put_task(task);
5594
	dprintk("%s: done, ret = %d!\n", __func__, ret);
5595
	return ret;
L
Linus Torvalds 已提交
5596 5597 5598 5599
}

static int nfs4_lock_reclaim(struct nfs4_state *state, struct file_lock *request)
{
5600
	struct nfs_server *server = NFS_SERVER(state->inode);
5601 5602 5603
	struct nfs4_exception exception = {
		.inode = state->inode,
	};
5604 5605 5606
	int err;

	do {
5607 5608 5609
		/* Cache the lock if possible... */
		if (test_bit(NFS_DELEGATED_STATE, &state->flags) != 0)
			return 0;
5610
		err = _nfs4_do_setlk(state, F_SETLK, request, NFS_LOCK_RECLAIM);
5611
		trace_nfs4_lock_reclaim(request, state, F_SETLK, err);
5612
		if (err != -NFS4ERR_DELAY)
5613 5614 5615 5616
			break;
		nfs4_handle_exception(server, err, &exception);
	} while (exception.retry);
	return err;
L
Linus Torvalds 已提交
5617 5618 5619 5620
}

static int nfs4_lock_expired(struct nfs4_state *state, struct file_lock *request)
{
5621
	struct nfs_server *server = NFS_SERVER(state->inode);
5622 5623 5624
	struct nfs4_exception exception = {
		.inode = state->inode,
	};
5625 5626
	int err;

5627 5628 5629
	err = nfs4_set_lock_state(state, request);
	if (err != 0)
		return err;
5630
	if (!recover_lost_locks) {
5631 5632 5633
		set_bit(NFS_LOCK_LOST, &request->fl_u.nfs4_fl.owner->ls_flags);
		return 0;
	}
5634
	do {
5635 5636
		if (test_bit(NFS_DELEGATED_STATE, &state->flags) != 0)
			return 0;
5637
		err = _nfs4_do_setlk(state, F_SETLK, request, NFS_LOCK_EXPIRED);
5638
		trace_nfs4_lock_expired(request, state, F_SETLK, err);
5639 5640 5641 5642 5643 5644 5645 5646
		switch (err) {
		default:
			goto out;
		case -NFS4ERR_GRACE:
		case -NFS4ERR_DELAY:
			nfs4_handle_exception(server, err, &exception);
			err = 0;
		}
5647
	} while (exception.retry);
5648
out:
5649
	return err;
L
Linus Torvalds 已提交
5650 5651
}

5652
#if defined(CONFIG_NFS_V4_1)
5653 5654 5655 5656 5657 5658 5659 5660
/**
 * 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.
 */
5661
static int nfs41_check_expired_locks(struct nfs4_state *state)
5662
{
5663
	int status, ret = -NFS4ERR_BAD_STATEID;
5664
	struct nfs4_lock_state *lsp;
5665 5666
	struct nfs_server *server = NFS_SERVER(state->inode);

5667
	list_for_each_entry(lsp, &state->lock_states, ls_locks) {
5668
		if (test_bit(NFS_LOCK_INITIALIZED, &lsp->ls_flags)) {
5669 5670 5671 5672 5673
			struct rpc_cred *cred = lsp->ls_state->owner->so_cred;

			status = nfs41_test_stateid(server,
					&lsp->ls_stateid,
					cred);
5674
			trace_nfs4_test_lock_stateid(state, lsp, status);
5675
			if (status != NFS_OK) {
5676 5677
				/* Free the stateid unless the server
				 * informs us the stateid is unrecognized. */
5678 5679
				if (status != -NFS4ERR_BAD_STATEID)
					nfs41_free_stateid(server,
5680 5681
							&lsp->ls_stateid,
							cred);
5682
				clear_bit(NFS_LOCK_INITIALIZED, &lsp->ls_flags);
5683 5684 5685 5686 5687 5688 5689 5690 5691 5692 5693 5694 5695 5696
				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);
5697 5698 5699
	if (status != NFS_OK)
		status = nfs4_lock_expired(state, request);
	return status;
5700 5701 5702
}
#endif

L
Linus Torvalds 已提交
5703 5704
static int _nfs4_proc_setlk(struct nfs4_state *state, int cmd, struct file_lock *request)
{
5705
	struct nfs4_state_owner *sp = state->owner;
5706
	struct nfs_inode *nfsi = NFS_I(state->inode);
5707
	unsigned char fl_flags = request->fl_flags;
5708
	unsigned int seq;
5709
	int status = -ENOLCK;
L
Linus Torvalds 已提交
5710

5711 5712 5713
	if ((fl_flags & FL_POSIX) &&
			!test_bit(NFS_STATE_POSIX_LOCKS, &state->flags))
		goto out;
5714 5715 5716 5717
	/* Is this a delegated open? */
	status = nfs4_set_lock_state(state, request);
	if (status != 0)
		goto out;
5718 5719 5720 5721
	request->fl_flags |= FL_ACCESS;
	status = do_vfs_lock(request->fl_file, request);
	if (status < 0)
		goto out;
5722
	down_read(&nfsi->rwsem);
5723 5724 5725
	if (test_bit(NFS_DELEGATED_STATE, &state->flags)) {
		/* Yes: cache locks! */
		/* ...but avoid races with delegation recall... */
5726 5727 5728
		request->fl_flags = fl_flags & ~FL_SLEEP;
		status = do_vfs_lock(request->fl_file, request);
		goto out_unlock;
5729
	}
5730 5731
	seq = raw_seqcount_begin(&sp->so_reclaim_seqcount);
	up_read(&nfsi->rwsem);
5732
	status = _nfs4_do_setlk(state, cmd, request, NFS_LOCK_NEW);
5733
	if (status != 0)
5734 5735 5736 5737
		goto out;
	down_read(&nfsi->rwsem);
	if (read_seqcount_retry(&sp->so_reclaim_seqcount, seq)) {
		status = -NFS4ERR_DELAY;
5738
		goto out_unlock;
5739
	}
5740
	/* Note: we always want to sleep here! */
5741
	request->fl_flags = fl_flags | FL_SLEEP;
5742
	if (do_vfs_lock(request->fl_file, request) < 0)
5743 5744
		printk(KERN_WARNING "NFS: %s: VFS is out of sync with lock "
			"manager!\n", __func__);
5745
out_unlock:
5746
	up_read(&nfsi->rwsem);
5747 5748
out:
	request->fl_flags = fl_flags;
L
Linus Torvalds 已提交
5749 5750 5751 5752 5753
	return status;
}

static int nfs4_proc_setlk(struct nfs4_state *state, int cmd, struct file_lock *request)
{
5754 5755
	struct nfs4_exception exception = {
		.state = state,
5756
		.inode = state->inode,
5757
	};
L
Linus Torvalds 已提交
5758 5759 5760
	int err;

	do {
5761
		err = _nfs4_proc_setlk(state, cmd, request);
5762
		trace_nfs4_set_lock(request, state, cmd, err);
5763 5764
		if (err == -NFS4ERR_DENIED)
			err = -EAGAIN;
L
Linus Torvalds 已提交
5765
		err = nfs4_handle_exception(NFS_SERVER(state->inode),
5766
				err, &exception);
L
Linus Torvalds 已提交
5767 5768 5769 5770 5771 5772 5773 5774 5775 5776 5777 5778 5779
	} 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 */
5780
	ctx = nfs_file_open_context(filp);
L
Linus Torvalds 已提交
5781 5782 5783 5784 5785
	state = ctx->state;

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

5786 5787 5788 5789 5790
	if (IS_GETLK(cmd)) {
		if (state != NULL)
			return nfs4_proc_getlk(state, F_GETLK, request);
		return 0;
	}
L
Linus Torvalds 已提交
5791 5792 5793 5794

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

5795 5796 5797 5798 5799
	if (request->fl_type == F_UNLCK) {
		if (state != NULL)
			return nfs4_proc_unlck(state, cmd, request);
		return 0;
	}
L
Linus Torvalds 已提交
5800

5801 5802
	if (state == NULL)
		return -ENOLCK;
5803 5804 5805 5806
	/*
	 * Don't rely on the VFS having checked the file open mode,
	 * since it won't do this for flock() locks.
	 */
5807
	switch (request->fl_type) {
5808 5809 5810 5811 5812 5813 5814 5815 5816
	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
5817 5818 5819 5820 5821 5822 5823 5824 5825 5826 5827 5828
	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;
}

5829
int nfs4_lock_delegation_recall(struct file_lock *fl, struct nfs4_state *state, const nfs4_stateid *stateid)
5830 5831 5832 5833 5834 5835
{
	struct nfs_server *server = NFS_SERVER(state->inode);
	int err;

	err = nfs4_set_lock_state(state, fl);
	if (err != 0)
5836
		return err;
5837
	err = _nfs4_do_setlk(state, F_SETLK, fl, NFS_LOCK_NEW);
5838
	return nfs4_handle_delegation_recall_error(server, state, stateid, err);
5839
}
5840

5841 5842
struct nfs_release_lockowner_data {
	struct nfs4_lock_state *lsp;
5843
	struct nfs_server *server;
5844
	struct nfs_release_lockowner_args args;
5845
	struct nfs_release_lockowner_res res;
5846
	unsigned long timestamp;
5847 5848
};

5849 5850 5851 5852
static void nfs4_release_lockowner_prepare(struct rpc_task *task, void *calldata)
{
	struct nfs_release_lockowner_data *data = calldata;
	nfs40_setup_sequence(data->server,
5853
				&data->args.seq_args, &data->res.seq_res, task);
5854
	data->timestamp = jiffies;
5855 5856 5857 5858 5859
}

static void nfs4_release_lockowner_done(struct rpc_task *task, void *calldata)
{
	struct nfs_release_lockowner_data *data = calldata;
5860 5861
	struct nfs_server *server = data->server;

5862
	nfs40_sequence_done(task, &data->res.seq_res);
5863 5864 5865 5866 5867 5868 5869 5870 5871 5872 5873 5874

	switch (task->tk_status) {
	case 0:
		renew_lease(server, data->timestamp);
		break;
	case -NFS4ERR_STALE_CLIENTID:
	case -NFS4ERR_EXPIRED:
	case -NFS4ERR_LEASE_MOVED:
	case -NFS4ERR_DELAY:
		if (nfs4_async_handle_error(task, server, NULL) == -EAGAIN)
			rpc_restart_call_prepare(task);
	}
5875 5876
}

5877 5878
static void nfs4_release_lockowner_release(void *calldata)
{
5879
	struct nfs_release_lockowner_data *data = calldata;
5880
	nfs4_free_lock_state(data->server, data->lsp);
5881 5882 5883
	kfree(calldata);
}

5884
static const struct rpc_call_ops nfs4_release_lockowner_ops = {
5885 5886
	.rpc_call_prepare = nfs4_release_lockowner_prepare,
	.rpc_call_done = nfs4_release_lockowner_done,
5887 5888 5889
	.rpc_release = nfs4_release_lockowner_release,
};

5890
static int nfs4_release_lockowner(struct nfs_server *server, struct nfs4_lock_state *lsp)
5891
{
5892
	struct nfs_release_lockowner_data *data;
5893 5894 5895 5896 5897
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_RELEASE_LOCKOWNER],
	};

	if (server->nfs_client->cl_mvops->minor_version != 0)
5898
		return -EINVAL;
5899

5900 5901 5902 5903
	data = kmalloc(sizeof(*data), GFP_NOFS);
	if (!data)
		return -ENOMEM;
	data->lsp = lsp;
5904
	data->server = server;
5905 5906 5907
	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;
5908

5909
	msg.rpc_argp = &data->args;
5910 5911
	msg.rpc_resp = &data->res;
	nfs4_init_sequence(&data->args.seq_args, &data->res.seq_res, 0);
5912 5913
	rpc_call_async(server->client, &msg, 0, &nfs4_release_lockowner_ops, data);
	return 0;
5914 5915
}

5916 5917
#define XATTR_NAME_NFSV4_ACL "system.nfs4_acl"

5918 5919 5920
static int nfs4_xattr_set_nfs4_acl(struct dentry *dentry, const char *key,
				   const void *buf, size_t buflen,
				   int flags, int type)
5921
{
5922 5923
	if (strcmp(key, "") != 0)
		return -EINVAL;
5924

5925
	return nfs4_proc_set_acl(dentry->d_inode, buf, buflen);
5926 5927
}

5928 5929
static int nfs4_xattr_get_nfs4_acl(struct dentry *dentry, const char *key,
				   void *buf, size_t buflen, int type)
5930
{
5931 5932
	if (strcmp(key, "") != 0)
		return -EINVAL;
5933

5934
	return nfs4_proc_get_acl(dentry->d_inode, buf, buflen);
5935 5936
}

5937 5938 5939
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)
5940
{
5941
	size_t len = sizeof(XATTR_NAME_NFSV4_ACL);
5942

5943 5944
	if (!nfs4_server_supports_acls(NFS_SERVER(dentry->d_inode)))
		return 0;
5945 5946 5947

	if (list && len <= list_len)
		memcpy(list, XATTR_NAME_NFSV4_ACL, len);
5948
	return len;
5949 5950
}

5951 5952 5953 5954 5955 5956 5957 5958 5959 5960 5961 5962 5963 5964 5965 5966 5967 5968 5969 5970 5971 5972 5973 5974 5975 5976 5977 5978 5979 5980 5981 5982 5983 5984 5985 5986 5987 5988 5989 5990 5991 5992 5993 5994 5995 5996 5997
#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


5998 5999 6000
/*
 * nfs_fhget will use either the mounted_on_fileid or the fileid
 */
6001 6002
static void nfs_fixup_referral_attributes(struct nfs_fattr *fattr)
{
6003 6004 6005
	if (!(((fattr->valid & NFS_ATTR_FATTR_MOUNTED_ON_FILEID) ||
	       (fattr->valid & NFS_ATTR_FATTR_FILEID)) &&
	      (fattr->valid & NFS_ATTR_FATTR_FSID) &&
6006
	      (fattr->valid & NFS_ATTR_FATTR_V4_LOCATIONS)))
6007 6008 6009
		return;

	fattr->valid |= NFS_ATTR_FATTR_TYPE | NFS_ATTR_FATTR_MODE |
6010
		NFS_ATTR_FATTR_NLINK | NFS_ATTR_FATTR_V4_REFERRAL;
6011 6012 6013 6014
	fattr->mode = S_IFDIR | S_IRUGO | S_IXUGO;
	fattr->nlink = 2;
}

6015 6016 6017 6018
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)
6019 6020
{
	struct nfs_server *server = NFS_SERVER(dir);
6021
	u32 bitmask[3] = {
6022
		[0] = FATTR4_WORD0_FSID | FATTR4_WORD0_FS_LOCATIONS,
6023 6024 6025
	};
	struct nfs4_fs_locations_arg args = {
		.dir_fh = NFS_FH(dir),
6026
		.name = name,
6027 6028 6029
		.page = page,
		.bitmask = bitmask,
	};
B
Benny Halevy 已提交
6030 6031 6032
	struct nfs4_fs_locations_res res = {
		.fs_locations = fs_locations,
	};
6033 6034 6035
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_FS_LOCATIONS],
		.rpc_argp = &args,
B
Benny Halevy 已提交
6036
		.rpc_resp = &res,
6037 6038 6039
	};
	int status;

6040
	dprintk("%s: start\n", __func__);
6041 6042 6043 6044 6045 6046 6047 6048

	/* 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;

6049
	nfs_fattr_init(&fs_locations->fattr);
6050
	fs_locations->server = server;
6051
	fs_locations->nlocations = 0;
6052
	status = nfs4_call_sync(client, server, &msg, &args.seq_args, &res.seq_res, 0);
6053
	dprintk("%s: returned status = %d\n", __func__, status);
6054 6055 6056
	return status;
}

6057 6058 6059 6060
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)
6061 6062 6063 6064
{
	struct nfs4_exception exception = { };
	int err;
	do {
6065 6066 6067 6068
		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,
6069 6070 6071 6072 6073
				&exception);
	} while (exception.retry);
	return err;
}

6074 6075 6076 6077 6078 6079 6080 6081 6082 6083 6084 6085 6086 6087 6088 6089 6090 6091 6092 6093 6094 6095 6096 6097 6098 6099 6100 6101 6102 6103 6104 6105 6106 6107 6108 6109 6110 6111 6112 6113 6114 6115 6116 6117 6118 6119 6120 6121 6122 6123 6124 6125 6126 6127 6128 6129 6130 6131 6132 6133 6134 6135 6136 6137 6138 6139 6140 6141 6142 6143 6144 6145 6146 6147 6148 6149 6150 6151 6152 6153 6154 6155 6156 6157 6158 6159 6160 6161 6162 6163 6164 6165 6166 6167 6168 6169 6170 6171 6172 6173 6174 6175 6176 6177 6178 6179 6180 6181 6182 6183 6184 6185 6186 6187 6188 6189 6190 6191 6192 6193 6194 6195 6196 6197 6198 6199 6200 6201 6202 6203 6204 6205 6206 6207 6208 6209 6210 6211 6212 6213 6214 6215 6216 6217 6218 6219 6220 6221 6222 6223 6224
/*
 * 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;
}

6225 6226 6227 6228 6229 6230 6231 6232 6233 6234 6235 6236 6237 6238 6239 6240 6241 6242 6243 6244 6245 6246 6247 6248 6249 6250 6251 6252 6253 6254 6255 6256 6257 6258 6259 6260 6261 6262 6263 6264 6265 6266 6267 6268 6269 6270 6271 6272 6273 6274 6275 6276 6277 6278 6279 6280 6281 6282 6283 6284 6285 6286 6287 6288 6289 6290 6291 6292 6293 6294 6295 6296 6297 6298 6299 6300 6301 6302 6303 6304 6305 6306 6307 6308 6309 6310 6311 6312 6313 6314 6315 6316 6317 6318 6319 6320 6321 6322 6323 6324 6325 6326 6327 6328 6329 6330 6331 6332 6333 6334 6335 6336 6337 6338 6339 6340 6341 6342 6343 6344 6345 6346 6347 6348 6349 6350
/*
 * 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;
}

6351
/**
6352 6353 6354 6355 6356
 * 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.
6357
 */
6358
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
6359 6360 6361 6362 6363 6364 6365 6366 6367 6368 6369 6370 6371 6372
{
	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,
	};
6373
	struct rpc_clnt *clnt = NFS_SERVER(dir)->client;
6374
	struct rpc_cred *cred = NULL;
6375 6376 6377

	if (use_integrity) {
		clnt = NFS_SERVER(dir)->nfs_client->cl_rpcclient;
6378 6379
		cred = nfs4_get_clid_cred(NFS_SERVER(dir)->nfs_client);
		msg.rpc_cred = cred;
6380
	}
B
Bryan Schumaker 已提交
6381 6382

	dprintk("NFS call  secinfo %s\n", name->name);
6383 6384 6385 6386

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

6387 6388
	status = nfs4_call_sync(clnt, NFS_SERVER(dir), &msg, &args.seq_args,
				&res.seq_res, 0);
B
Bryan Schumaker 已提交
6389
	dprintk("NFS reply  secinfo: %d\n", status);
6390

6391 6392
	if (cred)
		put_rpccred(cred);
6393

B
Bryan Schumaker 已提交
6394 6395 6396
	return status;
}

6397 6398
int nfs4_proc_secinfo(struct inode *dir, const struct qstr *name,
		      struct nfs4_secinfo_flavors *flavors)
B
Bryan Schumaker 已提交
6399 6400 6401 6402
{
	struct nfs4_exception exception = { };
	int err;
	do {
6403 6404 6405 6406 6407 6408 6409 6410 6411 6412 6413 6414 6415 6416 6417
		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);

6418 6419
		trace_nfs4_secinfo(dir, name, err);
		err = nfs4_handle_exception(NFS_SERVER(dir), err,
B
Bryan Schumaker 已提交
6420 6421 6422 6423 6424
				&exception);
	} while (exception.retry);
	return err;
}

6425
#ifdef CONFIG_NFS_V4_1
6426 6427 6428 6429 6430 6431 6432 6433 6434 6435 6436 6437 6438 6439 6440 6441 6442 6443 6444
/*
 * 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;
}

6445
static bool
6446 6447
nfs41_same_server_scope(struct nfs41_server_scope *a,
			struct nfs41_server_scope *b)
6448 6449 6450 6451 6452 6453 6454 6455
{
	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;
}

6456 6457 6458 6459 6460 6461
/*
 * nfs4_proc_bind_conn_to_session()
 *
 * The 4.1 client currently uses the same TCP connection for the
 * fore and backchannel.
 */
6462
int nfs4_proc_bind_conn_to_session(struct nfs_client *clp, struct rpc_cred *cred)
6463 6464 6465 6466 6467 6468 6469 6470
{
	int status;
	struct nfs41_bind_conn_to_session_res res;
	struct rpc_message msg = {
		.rpc_proc =
			&nfs4_procedures[NFSPROC4_CLNT_BIND_CONN_TO_SESSION],
		.rpc_argp = clp,
		.rpc_resp = &res,
6471
		.rpc_cred = cred,
6472 6473 6474 6475 6476 6477 6478 6479 6480 6481 6482
	};

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

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

	status = rpc_call_sync(clp->cl_rpcclient, &msg, RPC_TASK_TIMEOUT);
6483
	trace_nfs4_bind_conn_to_session(clp, status);
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
	if (status == 0) {
		if (memcmp(res.session->sess_id.data,
		    clp->cl_session->sess_id.data, NFS4_MAX_SESSIONID_LEN)) {
			dprintk("NFS: %s: Session ID mismatch\n", __func__);
			status = -EIO;
			goto out_session;
		}
		if (res.dir != NFS4_CDFS4_BOTH) {
			dprintk("NFS: %s: Unexpected direction from server\n",
				__func__);
			status = -EIO;
			goto out_session;
		}
		if (res.use_conn_in_rdma_mode) {
			dprintk("NFS: %s: Server returned RDMA mode = true\n",
				__func__);
			status = -EIO;
			goto out_session;
		}
	}
out_session:
	kfree(res.session);
out:
	dprintk("<-- %s status= %d\n", __func__, status);
	return status;
}

B
Benny Halevy 已提交
6511
/*
6512 6513
 * 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
6514 6515 6516 6517 6518 6519 6520 6521 6522
 */
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)
6523 6524 6525
	},
	.allow.u.words = {
		[0] = 1 << (OP_CLOSE) |
6526 6527
		      1 << (OP_LOCKU) |
		      1 << (OP_COMMIT),
6528
		[1] = 1 << (OP_SECINFO - 32) |
6529 6530
		      1 << (OP_SECINFO_NO_NAME - 32) |
		      1 << (OP_TEST_STATEID - 32) |
6531 6532
		      1 << (OP_FREE_STATEID - 32) |
		      1 << (OP_WRITE - 32)
6533 6534 6535 6536 6537 6538
	}
};

/*
 * Select the state protection mode for client `clp' given the server results
 * from exchange_id in `sp'.
B
Benny Halevy 已提交
6539
 *
6540 6541 6542 6543 6544 6545 6546 6547 6548 6549 6550 6551 6552 6553 6554 6555 6556 6557 6558 6559 6560 6561 6562 6563 6564 6565 6566 6567 6568 6569 6570 6571 6572 6573 6574 6575 6576 6577 6578 6579 6580 6581 6582 6583 6584 6585 6586 6587 6588 6589 6590 6591
 * 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;
		}
6592 6593 6594 6595 6596 6597

		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);
		}
6598 6599 6600 6601 6602 6603

		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);
		}
6604 6605 6606 6607 6608 6609

		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);
		}
6610 6611 6612 6613 6614 6615 6616 6617 6618 6619

		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);
		}
6620 6621 6622 6623 6624 6625 6626
	}

	return 0;
}

/*
 * _nfs4_proc_exchange_id()
6627
 *
6628
 * Wrapper for EXCHANGE_ID operation.
B
Benny Halevy 已提交
6629
 */
6630 6631
static int _nfs4_proc_exchange_id(struct nfs_client *clp, struct rpc_cred *cred,
	u32 sp4_how)
B
Benny Halevy 已提交
6632 6633 6634
{
	nfs4_verifier verifier;
	struct nfs41_exchange_id_args args = {
6635
		.verifier = &verifier,
B
Benny Halevy 已提交
6636
		.client = clp,
6637
#ifdef CONFIG_NFS_V4_1_MIGRATION
6638
		.flags = EXCHGID4_FLAG_SUPP_MOVED_REFER |
6639 6640 6641 6642 6643 6644
			 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
6645 6646
	};
	struct nfs41_exchange_id_res res = {
6647
		0
B
Benny Halevy 已提交
6648 6649 6650 6651 6652 6653 6654 6655 6656
	};
	int status;
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_EXCHANGE_ID],
		.rpc_argp = &args,
		.rpc_resp = &res,
		.rpc_cred = cred,
	};

6657
	nfs4_init_boot_verifier(clp, &verifier);
6658 6659
	args.id_len = nfs4_init_uniform_client_string(clp, args.id,
							sizeof(args.id));
6660 6661 6662
	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
6663

6664
	res.server_owner = kzalloc(sizeof(struct nfs41_server_owner),
6665
					GFP_NOFS);
6666
	if (unlikely(res.server_owner == NULL)) {
6667 6668 6669
		status = -ENOMEM;
		goto out;
	}
6670

6671
	res.server_scope = kzalloc(sizeof(struct nfs41_server_scope),
6672
					GFP_NOFS);
6673
	if (unlikely(res.server_scope == NULL)) {
6674
		status = -ENOMEM;
6675
		goto out_server_owner;
6676
	}
6677

6678
	res.impl_id = kzalloc(sizeof(struct nfs41_impl_id), GFP_NOFS);
6679
	if (unlikely(res.impl_id == NULL)) {
6680 6681 6682 6683
		status = -ENOMEM;
		goto out_server_scope;
	}

6684 6685 6686 6687 6688 6689 6690 6691 6692 6693 6694 6695 6696 6697 6698 6699
	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;
	}

6700
	status = rpc_call_sync(clp->cl_rpcclient, &msg, RPC_TASK_TIMEOUT);
6701
	trace_nfs4_exchange_id(clp, status);
6702
	if (status == 0)
6703
		status = nfs4_check_cl_exchange_flags(res.flags);
6704

6705 6706 6707
	if (status == 0)
		status = nfs4_sp4_select_mode(clp, &res.state_protect);

6708
	if (status == 0) {
6709 6710 6711 6712 6713
		clp->cl_clientid = res.clientid;
		clp->cl_exchange_flags = (res.flags & ~EXCHGID4_FLAG_CONFIRMED_R);
		if (!(res.flags & EXCHGID4_FLAG_CONFIRMED_R))
			clp->cl_seqid = res.seqid;

6714 6715 6716
		kfree(clp->cl_serverowner);
		clp->cl_serverowner = res.server_owner;
		res.server_owner = NULL;
6717

6718
		/* use the most recent implementation id */
6719 6720
		kfree(clp->cl_implid);
		clp->cl_implid = res.impl_id;
6721

6722
		if (clp->cl_serverscope != NULL &&
6723
		    !nfs41_same_server_scope(clp->cl_serverscope,
6724 6725 6726 6727
					     res.server_scope)) {
			dprintk("%s: server_scope mismatch detected\n",
				__func__);
			set_bit(NFS4CLNT_SERVER_SCOPE_MISMATCH, &clp->cl_state);
6728 6729
			kfree(clp->cl_serverscope);
			clp->cl_serverscope = NULL;
6730 6731
		}

6732
		if (clp->cl_serverscope == NULL) {
6733
			clp->cl_serverscope = res.server_scope;
6734 6735
			goto out;
		}
6736 6737
	} else
		kfree(res.impl_id);
6738

6739 6740
out_server_owner:
	kfree(res.server_owner);
6741
out_server_scope:
6742 6743
	kfree(res.server_scope);
out:
6744
	if (clp->cl_implid != NULL)
6745
		dprintk("NFS reply exchange_id: Server Implementation ID: "
6746
			"domain: %s, name: %s, date: %llu,%u\n",
6747
			clp->cl_implid->domain, clp->cl_implid->name,
6748 6749
			clp->cl_implid->date.seconds,
			clp->cl_implid->date.nseconds);
6750
	dprintk("NFS reply exchange_id: %d\n", status);
B
Benny Halevy 已提交
6751 6752 6753
	return status;
}

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
/*
 * 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
6783 6784 6785 6786 6787 6788 6789 6790 6791 6792 6793
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);
6794
	trace_nfs4_destroy_clientid(clp, status);
T
Trond Myklebust 已提交
6795
	if (status)
6796
		dprintk("NFS: Got error %d from the server %s on "
T
Trond Myklebust 已提交
6797 6798 6799 6800 6801 6802 6803 6804 6805 6806 6807 6808 6809 6810 6811 6812 6813 6814 6815 6816 6817 6818 6819 6820 6821 6822 6823 6824 6825 6826 6827 6828 6829
			"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;
6830 6831
	if (clp->cl_preserve_clid)
		goto out;
6832
	cred = nfs4_get_clid_cred(clp);
T
Trond Myklebust 已提交
6833 6834 6835 6836 6837 6838 6839 6840 6841 6842 6843 6844
	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
6845 6846 6847 6848 6849 6850 6851 6852 6853 6854 6855 6856 6857 6858 6859
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 */
6860 6861 6862 6863
	nfs41_setup_sequence(data->clp->cl_session,
			&data->args->la_seq_args,
			&data->res->lr_seq_res,
			task);
A
Andy Adamson 已提交
6864 6865 6866 6867 6868 6869 6870 6871 6872 6873 6874 6875 6876
	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__);
6877 6878
	if (!nfs41_sequence_done(task, &data->res->lr_seq_res))
		return;
A
Andy Adamson 已提交
6879 6880 6881 6882 6883 6884
	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;
6885 6886
		/* fall through */
	case -NFS4ERR_RETRY_UNCACHED_REP:
6887
		rpc_restart_call_prepare(task);
A
Andy Adamson 已提交
6888 6889 6890 6891 6892
		return;
	}
	dprintk("<-- %s\n", __func__);
}

6893
static const struct rpc_call_ops nfs4_get_lease_time_ops = {
A
Andy Adamson 已提交
6894 6895 6896 6897 6898 6899 6900 6901 6902 6903 6904 6905 6906 6907 6908 6909 6910 6911 6912 6913 6914 6915 6916 6917 6918
	.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,
6919 6920
		.callback_data = &data,
		.flags = RPC_TASK_TIMEOUT,
A
Andy Adamson 已提交
6921 6922 6923
	};
	int status;

6924
	nfs4_init_sequence(&args.la_seq_args, &res.lr_seq_res, 0);
6925
	nfs4_set_sequence_privileged(&args.la_seq_args);
A
Andy Adamson 已提交
6926 6927 6928 6929 6930 6931 6932 6933 6934 6935 6936 6937 6938 6939
	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
6940 6941 6942 6943 6944 6945 6946 6947 6948 6949 6950
/*
 * 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)
{
6951 6952 6953 6954
	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
6955 6956

	/* Fore channel attributes */
6957 6958
	args->fc_attrs.max_rqst_sz = max_rqst_sz;
	args->fc_attrs.max_resp_sz = max_resp_sz;
A
Andy Adamson 已提交
6959
	args->fc_attrs.max_ops = NFS4_MAX_OPS;
6960
	args->fc_attrs.max_reqs = max_session_slots;
A
Andy Adamson 已提交
6961 6962

	dprintk("%s: Fore Channel : max_rqst_sz=%u max_resp_sz=%u "
6963
		"max_ops=%u max_reqs=%u\n",
A
Andy Adamson 已提交
6964 6965
		__func__,
		args->fc_attrs.max_rqst_sz, args->fc_attrs.max_resp_sz,
6966
		args->fc_attrs.max_ops, args->fc_attrs.max_reqs);
A
Andy Adamson 已提交
6967 6968 6969 6970 6971 6972 6973 6974 6975 6976 6977 6978 6979 6980 6981 6982

	/* 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);
}

6983
static int nfs4_verify_fore_channel_attrs(struct nfs41_create_session_args *args, struct nfs4_session *session)
6984
{
6985 6986 6987 6988 6989 6990 6991 6992 6993 6994 6995 6996 6997 6998 6999
	struct nfs4_channel_attrs *sent = &args->fc_attrs;
	struct nfs4_channel_attrs *rcvd = &session->fc_attrs;

	if (rcvd->max_resp_sz > sent->max_resp_sz)
		return -EINVAL;
	/*
	 * Our requested max_ops is the minimum we need; we're not
	 * prepared to break up compounds into smaller pieces than that.
	 * So, no point even trying to continue if the server won't
	 * cooperate:
	 */
	if (rcvd->max_ops < sent->max_ops)
		return -EINVAL;
	if (rcvd->max_reqs == 0)
		return -EINVAL;
7000 7001
	if (rcvd->max_reqs > NFS4_MAX_SLOT_TABLE)
		rcvd->max_reqs = NFS4_MAX_SLOT_TABLE;
7002
	return 0;
7003 7004
}

7005 7006 7007 7008
static int nfs4_verify_back_channel_attrs(struct nfs41_create_session_args *args, struct nfs4_session *session)
{
	struct nfs4_channel_attrs *sent = &args->bc_attrs;
	struct nfs4_channel_attrs *rcvd = &session->bc_attrs;
7009

7010 7011 7012 7013 7014 7015 7016
	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: */
7017
	if (rcvd->max_ops != sent->max_ops)
7018
		return -EINVAL;
7019
	if (rcvd->max_reqs != sent->max_reqs)
7020 7021 7022
		return -EINVAL;
	return 0;
}
7023 7024 7025 7026

static int nfs4_verify_channel_attrs(struct nfs41_create_session_args *args,
				     struct nfs4_session *session)
{
7027
	int ret;
7028

7029 7030 7031 7032
	ret = nfs4_verify_fore_channel_attrs(args, session);
	if (ret)
		return ret;
	return nfs4_verify_back_channel_attrs(args, session);
7033 7034
}

7035 7036
static int _nfs4_proc_create_session(struct nfs_client *clp,
		struct rpc_cred *cred)
A
Andy Adamson 已提交
7037 7038 7039 7040 7041 7042 7043 7044 7045 7046 7047 7048 7049
{
	struct nfs4_session *session = clp->cl_session;
	struct nfs41_create_session_args args = {
		.client = clp,
		.cb_program = NFS4_CALLBACK,
	};
	struct nfs41_create_session_res res = {
		.client = clp,
	};
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_CREATE_SESSION],
		.rpc_argp = &args,
		.rpc_resp = &res,
7050
		.rpc_cred = cred,
A
Andy Adamson 已提交
7051 7052 7053 7054
	};
	int status;

	nfs4_init_channel_attrs(&args);
7055
	args.flags = (SESSION4_PERSIST | SESSION4_BACK_CHAN);
A
Andy Adamson 已提交
7056

7057
	status = rpc_call_sync(session->clp->cl_rpcclient, &msg, RPC_TASK_TIMEOUT);
7058
	trace_nfs4_create_session(clp, status);
A
Andy Adamson 已提交
7059

7060
	if (!status) {
7061 7062
		/* Verify the session's negotiated channel_attrs values */
		status = nfs4_verify_channel_attrs(&args, session);
A
Andy Adamson 已提交
7063 7064 7065 7066 7067 7068 7069 7070 7071 7072 7073 7074
		/* Increment the clientid slot sequence id */
		clp->cl_seqid++;
	}

	return status;
}

/*
 * Issues a CREATE_SESSION operation to the server.
 * It is the responsibility of the caller to verify the session is
 * expired before calling this routine.
 */
7075
int nfs4_proc_create_session(struct nfs_client *clp, struct rpc_cred *cred)
A
Andy Adamson 已提交
7076 7077 7078 7079 7080 7081 7082
{
	int status;
	unsigned *ptr;
	struct nfs4_session *session = clp->cl_session;

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

7083
	status = _nfs4_proc_create_session(clp, cred);
A
Andy Adamson 已提交
7084 7085 7086
	if (status)
		goto out;

7087 7088 7089
	/* 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
7090 7091 7092 7093 7094 7095 7096 7097 7098 7099 7100
	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
7101 7102 7103 7104
/*
 * Issue the over-the-wire RPC DESTROY_SESSION.
 * The caller must serialize access to this routine.
 */
7105 7106
int nfs4_proc_destroy_session(struct nfs4_session *session,
		struct rpc_cred *cred)
A
Andy Adamson 已提交
7107
{
7108 7109 7110 7111 7112
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_DESTROY_SESSION],
		.rpc_argp = session,
		.rpc_cred = cred,
	};
A
Andy Adamson 已提交
7113 7114 7115 7116 7117 7118 7119 7120
	int status = 0;

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

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

7121
	status = rpc_call_sync(session->clp->cl_rpcclient, &msg, RPC_TASK_TIMEOUT);
7122
	trace_nfs4_destroy_session(session->clp, status);
A
Andy Adamson 已提交
7123 7124

	if (status)
7125
		dprintk("NFS: Got error %d from the server on DESTROY_SESSION. "
A
Andy Adamson 已提交
7126 7127 7128 7129 7130 7131
			"Session has been destroyed regardless...\n", status);

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

A
Andy Adamson 已提交
7132 7133 7134
/*
 * Renew the cl_session lease.
 */
7135 7136 7137 7138 7139 7140
struct nfs4_sequence_data {
	struct nfs_client *clp;
	struct nfs4_sequence_args args;
	struct nfs4_sequence_res res;
};

7141 7142
static void nfs41_sequence_release(void *data)
{
7143 7144
	struct nfs4_sequence_data *calldata = data;
	struct nfs_client *clp = calldata->clp;
7145

7146 7147 7148
	if (atomic_read(&clp->cl_count) > 1)
		nfs4_schedule_state_renewal(clp);
	nfs_put_client(clp);
7149
	kfree(calldata);
7150 7151
}

7152 7153 7154 7155 7156 7157 7158
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:
7159
		nfs4_schedule_lease_recovery(clp);
7160 7161 7162 7163
	}
	return 0;
}

7164
static void nfs41_sequence_call_done(struct rpc_task *task, void *data)
A
Andy Adamson 已提交
7165
{
7166 7167
	struct nfs4_sequence_data *calldata = data;
	struct nfs_client *clp = calldata->clp;
A
Andy Adamson 已提交
7168

7169 7170
	if (!nfs41_sequence_done(task, task->tk_msg.rpc_resp))
		return;
A
Andy Adamson 已提交
7171

7172
	trace_nfs4_sequence(clp, task->tk_status);
A
Andy Adamson 已提交
7173 7174
	if (task->tk_status < 0) {
		dprintk("%s ERROR %d\n", __func__, task->tk_status);
7175 7176
		if (atomic_read(&clp->cl_count) == 1)
			goto out;
A
Andy Adamson 已提交
7177

7178 7179
		if (nfs41_sequence_handle_errors(task, clp) == -EAGAIN) {
			rpc_restart_call_prepare(task);
A
Andy Adamson 已提交
7180 7181 7182 7183
			return;
		}
	}
	dprintk("%s rpc_cred %p\n", __func__, task->tk_msg.rpc_cred);
7184
out:
A
Andy Adamson 已提交
7185 7186 7187 7188 7189
	dprintk("<-- %s\n", __func__);
}

static void nfs41_sequence_prepare(struct rpc_task *task, void *data)
{
7190 7191
	struct nfs4_sequence_data *calldata = data;
	struct nfs_client *clp = calldata->clp;
A
Andy Adamson 已提交
7192 7193 7194 7195 7196 7197
	struct nfs4_sequence_args *args;
	struct nfs4_sequence_res *res;

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

7198
	nfs41_setup_sequence(clp->cl_session, args, res, task);
A
Andy Adamson 已提交
7199 7200 7201 7202 7203
}

static const struct rpc_call_ops nfs41_sequence_ops = {
	.rpc_call_done = nfs41_sequence_call_done,
	.rpc_call_prepare = nfs41_sequence_prepare,
7204
	.rpc_release = nfs41_sequence_release,
A
Andy Adamson 已提交
7205 7206
};

7207 7208 7209
static struct rpc_task *_nfs41_proc_sequence(struct nfs_client *clp,
		struct rpc_cred *cred,
		bool is_privileged)
A
Andy Adamson 已提交
7210
{
7211
	struct nfs4_sequence_data *calldata;
A
Andy Adamson 已提交
7212 7213 7214 7215
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_SEQUENCE],
		.rpc_cred = cred,
	};
7216 7217 7218
	struct rpc_task_setup task_setup_data = {
		.rpc_client = clp->cl_rpcclient,
		.rpc_message = &msg,
7219
		.callback_ops = &nfs41_sequence_ops,
7220
		.flags = RPC_TASK_ASYNC | RPC_TASK_TIMEOUT,
7221
	};
A
Andy Adamson 已提交
7222

7223
	if (!atomic_inc_not_zero(&clp->cl_count))
7224
		return ERR_PTR(-EIO);
7225
	calldata = kzalloc(sizeof(*calldata), GFP_NOFS);
7226
	if (calldata == NULL) {
7227
		nfs_put_client(clp);
7228
		return ERR_PTR(-ENOMEM);
A
Andy Adamson 已提交
7229
	}
7230
	nfs4_init_sequence(&calldata->args, &calldata->res, 0);
7231 7232
	if (is_privileged)
		nfs4_set_sequence_privileged(&calldata->args);
7233 7234 7235
	msg.rpc_argp = &calldata->args;
	msg.rpc_resp = &calldata->res;
	calldata->clp = clp;
7236
	task_setup_data.callback_data = calldata;
A
Andy Adamson 已提交
7237

7238 7239 7240
	return rpc_run_task(&task_setup_data);
}

7241
static int nfs41_proc_async_sequence(struct nfs_client *clp, struct rpc_cred *cred, unsigned renew_flags)
7242 7243 7244 7245
{
	struct rpc_task *task;
	int ret = 0;

7246 7247
	if ((renew_flags & NFS4_RENEW_TIMEOUT) == 0)
		return 0;
7248
	task = _nfs41_proc_sequence(clp, cred, false);
7249 7250 7251
	if (IS_ERR(task))
		ret = PTR_ERR(task);
	else
7252
		rpc_put_task_async(task);
7253 7254 7255 7256 7257 7258 7259 7260 7261
	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;

7262
	task = _nfs41_proc_sequence(clp, cred, true);
7263 7264 7265 7266 7267
	if (IS_ERR(task)) {
		ret = PTR_ERR(task);
		goto out;
	}
	ret = rpc_wait_for_completion_task(task);
7268 7269 7270 7271 7272
	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);
7273
		ret = task->tk_status;
7274
	}
7275 7276 7277 7278
	rpc_put_task(task);
out:
	dprintk("<-- %s status=%d\n", __func__, ret);
	return ret;
A
Andy Adamson 已提交
7279 7280
}

7281 7282 7283 7284 7285 7286 7287 7288 7289 7290
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;

7291 7292 7293 7294
	nfs41_setup_sequence(calldata->clp->cl_session,
			&calldata->arg.seq_args,
			&calldata->res.seq_res,
			task);
7295 7296
}

7297 7298 7299 7300 7301 7302 7303 7304 7305
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);
7306 7307
		/* fall through */
	case -NFS4ERR_RETRY_UNCACHED_REP:
7308 7309
		return -EAGAIN;
	default:
7310
		nfs4_schedule_lease_recovery(clp);
7311 7312 7313 7314
	}
	return 0;
}

7315 7316 7317 7318 7319 7320 7321
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__);
7322 7323
	if (!nfs41_sequence_done(task, res))
		return;
7324

7325
	trace_nfs4_reclaim_complete(clp, task->tk_status);
7326 7327 7328 7329
	if (nfs41_reclaim_complete_handle_errors(task, clp) == -EAGAIN) {
		rpc_restart_call_prepare(task);
		return;
	}
7330 7331 7332 7333 7334 7335 7336 7337 7338 7339 7340 7341 7342 7343 7344 7345 7346 7347 7348
	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.
 */
7349 7350
static int nfs41_proc_reclaim_complete(struct nfs_client *clp,
		struct rpc_cred *cred)
7351 7352 7353 7354 7355
{
	struct nfs4_reclaim_complete_data *calldata;
	struct rpc_task *task;
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_RECLAIM_COMPLETE],
7356
		.rpc_cred = cred,
7357 7358 7359 7360 7361 7362 7363 7364 7365 7366
	};
	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__);
7367
	calldata = kzalloc(sizeof(*calldata), GFP_NOFS);
7368 7369 7370 7371 7372
	if (calldata == NULL)
		goto out;
	calldata->clp = clp;
	calldata->arg.one_fs = 0;

7373
	nfs4_init_sequence(&calldata->arg.seq_args, &calldata->res.seq_res, 0);
7374
	nfs4_set_sequence_privileged(&calldata->arg.seq_args);
7375 7376 7377 7378
	msg.rpc_argp = &calldata->arg;
	msg.rpc_resp = &calldata->res;
	task_setup_data.callback_data = calldata;
	task = rpc_run_task(&task_setup_data);
7379
	if (IS_ERR(task)) {
7380
		status = PTR_ERR(task);
7381 7382
		goto out;
	}
7383 7384 7385
	status = nfs4_wait_for_completion_rpc_task(task);
	if (status == 0)
		status = task->tk_status;
7386
	rpc_put_task(task);
7387
	return 0;
7388 7389 7390 7391
out:
	dprintk("<-- %s status=%d\n", __func__, status);
	return status;
}
7392 7393 7394 7395 7396

static void
nfs4_layoutget_prepare(struct rpc_task *task, void *calldata)
{
	struct nfs4_layoutget *lgp = calldata;
F
Fred Isaman 已提交
7397
	struct nfs_server *server = NFS_SERVER(lgp->args.inode);
7398
	struct nfs4_session *session = nfs4_get_session(server);
7399 7400

	dprintk("--> %s\n", __func__);
F
Fred Isaman 已提交
7401 7402 7403 7404 7405
	/* 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.
	 */
7406
	if (nfs41_setup_sequence(session, &lgp->args.seq_args,
7407
				&lgp->res.seq_res, task))
7408
		return;
7409 7410 7411 7412 7413
	if (pnfs_choose_layoutget_stateid(&lgp->args.stateid,
					  NFS_I(lgp->args.inode)->layout,
					  lgp->args.ctx->state)) {
		rpc_exit(task, NFS4_OK);
	}
7414 7415 7416 7417 7418
}

static void nfs4_layoutget_done(struct rpc_task *task, void *calldata)
{
	struct nfs4_layoutget *lgp = calldata;
7419 7420 7421 7422
	struct inode *inode = lgp->args.inode;
	struct nfs_server *server = NFS_SERVER(inode);
	struct pnfs_layout_hdr *lo;
	struct nfs4_state *state = NULL;
7423
	unsigned long timeo, now, giveup;
7424

7425
	dprintk("--> %s tk_status => %d\n", __func__, -task->tk_status);
7426

7427
	if (!nfs41_sequence_done(task, &lgp->res.seq_res))
7428
		goto out;
7429 7430 7431

	switch (task->tk_status) {
	case 0:
7432
		goto out;
7433 7434 7435 7436
	/*
	 * NFS4ERR_LAYOUTTRYLATER is a conflict with another client
	 * (or clients) writing to the same RAID stripe
	 */
7437
	case -NFS4ERR_LAYOUTTRYLATER:
7438 7439 7440 7441
	/*
	 * NFS4ERR_RECALLCONFLICT is when conflict with self (must recall
	 * existing layout before getting a new one).
	 */
7442
	case -NFS4ERR_RECALLCONFLICT:
7443 7444
		timeo = rpc_get_timeout(task->tk_client);
		giveup = lgp->args.timestamp + timeo;
7445 7446 7447 7448 7449 7450 7451 7452 7453 7454 7455 7456 7457 7458 7459 7460 7461 7462 7463 7464
		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() */
		}
7465 7466 7467 7468 7469 7470 7471 7472 7473 7474 7475 7476 7477 7478 7479 7480 7481
		break;
	case -NFS4ERR_EXPIRED:
	case -NFS4ERR_BAD_STATEID:
		spin_lock(&inode->i_lock);
		lo = NFS_I(inode)->layout;
		if (!lo || list_empty(&lo->plh_segs)) {
			spin_unlock(&inode->i_lock);
			/* If the open stateid was bad, then recover it. */
			state = lgp->args.ctx->state;
		} else {
			LIST_HEAD(head);

			pnfs_mark_matching_lsegs_invalid(lo, &head, NULL);
			spin_unlock(&inode->i_lock);
			/* Mark the bad layout state as invalid, then
			 * retry using the open stateid. */
			pnfs_free_lseg_list(&head);
7482 7483
		}
	}
7484 7485 7486
	if (nfs4_async_handle_error(task, server, state) == -EAGAIN)
		rpc_restart_call_prepare(task);
out:
7487 7488 7489
	dprintk("<-- %s\n", __func__);
}

7490 7491 7492 7493 7494 7495 7496 7497 7498 7499 7500 7501 7502 7503 7504 7505 7506 7507 7508 7509 7510 7511 7512 7513 7514 7515 7516 7517 7518 7519 7520 7521 7522 7523 7524 7525 7526 7527 7528 7529 7530 7531 7532 7533
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;
}

7534 7535 7536
static void nfs4_layoutget_release(void *calldata)
{
	struct nfs4_layoutget *lgp = calldata;
7537 7538
	struct inode *inode = lgp->args.inode;
	struct nfs_server *server = NFS_SERVER(inode);
7539
	size_t max_pages = max_response_pages(server);
7540 7541

	dprintk("--> %s\n", __func__);
7542
	nfs4_free_pages(lgp->args.layout.pages, max_pages);
7543
	pnfs_put_layout_hdr(NFS_I(inode)->layout);
7544 7545 7546 7547 7548 7549 7550 7551 7552 7553 7554
	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,
};

7555 7556
struct pnfs_layout_segment *
nfs4_proc_layoutget(struct nfs4_layoutget *lgp, gfp_t gfp_flags)
7557
{
7558 7559
	struct inode *inode = lgp->args.inode;
	struct nfs_server *server = NFS_SERVER(inode);
7560
	size_t max_pages = max_response_pages(server);
7561 7562 7563 7564 7565
	struct rpc_task *task;
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_LAYOUTGET],
		.rpc_argp = &lgp->args,
		.rpc_resp = &lgp->res,
7566
		.rpc_cred = lgp->cred,
7567 7568 7569 7570 7571 7572 7573 7574
	};
	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,
	};
7575
	struct pnfs_layout_segment *lseg = NULL;
7576 7577 7578 7579
	int status = 0;

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

7580 7581 7582
	lgp->args.layout.pages = nfs4_alloc_pages(max_pages, gfp_flags);
	if (!lgp->args.layout.pages) {
		nfs4_layoutget_release(lgp);
7583
		return ERR_PTR(-ENOMEM);
7584 7585
	}
	lgp->args.layout.pglen = max_pages * PAGE_SIZE;
7586
	lgp->args.timestamp = jiffies;
7587

7588
	lgp->res.layoutp = &lgp->args.layout;
7589
	lgp->res.seq_res.sr_slot = NULL;
7590
	nfs4_init_sequence(&lgp->args.seq_args, &lgp->res.seq_res, 0);
7591 7592 7593 7594

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

7595 7596
	task = rpc_run_task(&task_setup_data);
	if (IS_ERR(task))
7597
		return ERR_CAST(task);
7598
	status = nfs4_wait_for_completion_rpc_task(task);
F
Fred Isaman 已提交
7599 7600
	if (status == 0)
		status = task->tk_status;
7601 7602 7603 7604
	trace_nfs4_layoutget(lgp->args.ctx,
			&lgp->args.range,
			&lgp->res.range,
			status);
7605 7606
	/* if layoutp->len is 0, nfs4_layoutget_prepare called rpc_exit */
	if (status == 0 && lgp->res.layoutp->len)
7607
		lseg = pnfs_layout_process(lgp);
7608 7609
	rpc_put_task(task);
	dprintk("<-- %s status=%d\n", __func__, status);
7610 7611 7612
	if (status)
		return ERR_PTR(status);
	return lseg;
7613 7614
}

B
Benny Halevy 已提交
7615 7616 7617 7618 7619 7620
static void
nfs4_layoutreturn_prepare(struct rpc_task *task, void *calldata)
{
	struct nfs4_layoutreturn *lrp = calldata;

	dprintk("--> %s\n", __func__);
7621 7622 7623 7624
	nfs41_setup_sequence(lrp->clp->cl_session,
			&lrp->args.seq_args,
			&lrp->res.seq_res,
			task);
B
Benny Halevy 已提交
7625 7626 7627 7628 7629 7630 7631 7632 7633
}

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

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

7634
	if (!nfs41_sequence_done(task, &lrp->res.seq_res))
B
Benny Halevy 已提交
7635 7636 7637
		return;

	server = NFS_SERVER(lrp->args.inode);
7638 7639 7640 7641 7642 7643 7644 7645
	switch (task->tk_status) {
	default:
		task->tk_status = 0;
	case 0:
		break;
	case -NFS4ERR_DELAY:
		if (nfs4_async_handle_error(task, server, NULL) != -EAGAIN)
			break;
7646
		rpc_restart_call_prepare(task);
B
Benny Halevy 已提交
7647 7648 7649 7650 7651 7652 7653 7654
		return;
	}
	dprintk("<-- %s\n", __func__);
}

static void nfs4_layoutreturn_release(void *calldata)
{
	struct nfs4_layoutreturn *lrp = calldata;
7655
	struct pnfs_layout_hdr *lo = lrp->args.layout;
B
Benny Halevy 已提交
7656 7657

	dprintk("--> %s\n", __func__);
7658 7659 7660 7661 7662
	spin_lock(&lo->plh_inode->i_lock);
	if (lrp->res.lrs_present)
		pnfs_set_layout_stateid(lo, &lrp->res.stateid, true);
	lo->plh_block_lgets--;
	spin_unlock(&lo->plh_inode->i_lock);
7663
	pnfs_put_layout_hdr(lrp->args.layout);
B
Benny Halevy 已提交
7664 7665 7666 7667 7668 7669 7670 7671 7672 7673 7674 7675 7676 7677 7678 7679 7680
	kfree(calldata);
	dprintk("<-- %s\n", __func__);
}

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

int nfs4_proc_layoutreturn(struct nfs4_layoutreturn *lrp)
{
	struct rpc_task *task;
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_LAYOUTRETURN],
		.rpc_argp = &lrp->args,
		.rpc_resp = &lrp->res,
7681
		.rpc_cred = lrp->cred,
B
Benny Halevy 已提交
7682 7683
	};
	struct rpc_task_setup task_setup_data = {
7684
		.rpc_client = NFS_SERVER(lrp->args.inode)->client,
B
Benny Halevy 已提交
7685 7686 7687 7688 7689 7690 7691
		.rpc_message = &msg,
		.callback_ops = &nfs4_layoutreturn_call_ops,
		.callback_data = lrp,
	};
	int status;

	dprintk("--> %s\n", __func__);
7692
	nfs4_init_sequence(&lrp->args.seq_args, &lrp->res.seq_res, 1);
B
Benny Halevy 已提交
7693 7694 7695 7696
	task = rpc_run_task(&task_setup_data);
	if (IS_ERR(task))
		return PTR_ERR(task);
	status = task->tk_status;
7697
	trace_nfs4_layoutreturn(lrp->args.inode, status);
B
Benny Halevy 已提交
7698 7699 7700 7701 7702
	dprintk("<-- %s status=%d\n", __func__, status);
	rpc_put_task(task);
	return status;
}

A
Andy Adamson 已提交
7703 7704 7705 7706 7707 7708 7709 7710 7711 7712 7713 7714 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
/*
 * Retrieve the list of Data Server devices from the MDS.
 */
static int _nfs4_getdevicelist(struct nfs_server *server,
				    const struct nfs_fh *fh,
				    struct pnfs_devicelist *devlist)
{
	struct nfs4_getdevicelist_args args = {
		.fh = fh,
		.layoutclass = server->pnfs_curr_ld->id,
	};
	struct nfs4_getdevicelist_res res = {
		.devlist = devlist,
	};
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_GETDEVICELIST],
		.rpc_argp = &args,
		.rpc_resp = &res,
	};
	int status;

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

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

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

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

	return err;
}
EXPORT_SYMBOL_GPL(nfs4_proc_getdevicelist);

7751
static int
7752 7753 7754
_nfs4_proc_getdeviceinfo(struct nfs_server *server,
		struct pnfs_device *pdev,
		struct rpc_cred *cred)
7755 7756 7757 7758 7759 7760 7761 7762 7763 7764 7765
{
	struct nfs4_getdeviceinfo_args args = {
		.pdev = pdev,
	};
	struct nfs4_getdeviceinfo_res res = {
		.pdev = pdev,
	};
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_GETDEVICEINFO],
		.rpc_argp = &args,
		.rpc_resp = &res,
7766
		.rpc_cred = cred,
7767 7768 7769 7770
	};
	int status;

	dprintk("--> %s\n", __func__);
7771
	status = nfs4_call_sync(server->client, server, &msg, &args.seq_args, &res.seq_res, 0);
7772 7773 7774 7775 7776
	dprintk("<-- %s status=%d\n", __func__, status);

	return status;
}

7777 7778 7779
int nfs4_proc_getdeviceinfo(struct nfs_server *server,
		struct pnfs_device *pdev,
		struct rpc_cred *cred)
7780 7781 7782 7783 7784 7785
{
	struct nfs4_exception exception = { };
	int err;

	do {
		err = nfs4_handle_exception(server,
7786
					_nfs4_proc_getdeviceinfo(server, pdev, cred),
7787 7788 7789 7790 7791 7792
					&exception);
	} while (exception.retry);
	return err;
}
EXPORT_SYMBOL_GPL(nfs4_proc_getdeviceinfo);

A
Andy Adamson 已提交
7793 7794 7795 7796
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);
7797
	struct nfs4_session *session = nfs4_get_session(server);
A
Andy Adamson 已提交
7798

7799 7800 7801 7802
	nfs41_setup_sequence(session,
			&data->args.seq_args,
			&data->res.seq_res,
			task);
A
Andy Adamson 已提交
7803 7804 7805 7806 7807 7808 7809 7810
}

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);

7811
	if (!nfs41_sequence_done(task, &data->res.seq_res))
A
Andy Adamson 已提交
7812 7813 7814
		return;

	switch (task->tk_status) { /* Just ignore these failures */
7815 7816 7817 7818
	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
7819
		task->tk_status = 0;
7820 7821 7822 7823 7824 7825 7826 7827
	case 0:
		break;
	default:
		if (nfs4_async_handle_error(task, server, NULL) == -EAGAIN) {
			rpc_restart_call_prepare(task);
			return;
		}
	}
A
Andy Adamson 已提交
7828 7829 7830 7831 7832 7833
}

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

A
Andy Adamson 已提交
7834
	pnfs_cleanup_layoutcommit(data);
7835 7836
	nfs_post_op_update_inode_force_wcc(data->args.inode,
					   data->res.fattr);
A
Andy Adamson 已提交
7837 7838 7839 7840 7841 7842 7843 7844 7845 7846 7847
	put_rpccred(data->cred);
	kfree(data);
}

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

int
7848
nfs4_proc_layoutcommit(struct nfs4_layoutcommit_data *data, bool sync)
A
Andy Adamson 已提交
7849 7850 7851 7852 7853 7854 7855 7856 7857 7858 7859 7860 7861 7862 7863 7864 7865 7866 7867 7868 7869 7870 7871 7872
{
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_LAYOUTCOMMIT],
		.rpc_argp = &data->args,
		.rpc_resp = &data->res,
		.rpc_cred = data->cred,
	};
	struct rpc_task_setup task_setup_data = {
		.task = &data->task,
		.rpc_client = NFS_CLIENT(data->args.inode),
		.rpc_message = &msg,
		.callback_ops = &nfs4_layoutcommit_ops,
		.callback_data = data,
		.flags = RPC_TASK_ASYNC,
	};
	struct rpc_task *task;
	int status = 0;

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

7873
	nfs4_init_sequence(&data->args.seq_args, &data->res.seq_res, 1);
A
Andy Adamson 已提交
7874 7875 7876
	task = rpc_run_task(&task_setup_data);
	if (IS_ERR(task))
		return PTR_ERR(task);
7877
	if (sync == false)
A
Andy Adamson 已提交
7878 7879 7880 7881 7882
		goto out;
	status = nfs4_wait_for_completion_rpc_task(task);
	if (status != 0)
		goto out;
	status = task->tk_status;
7883
	trace_nfs4_layoutcommit(data->args.inode, status);
A
Andy Adamson 已提交
7884 7885 7886 7887 7888
out:
	dprintk("%s: status %d\n", __func__, status);
	rpc_put_task(task);
	return status;
}
7889

7890 7891 7892 7893
/**
 * Use the state managment nfs_client cl_rpcclient, which uses krb5i (if
 * possible) as per RFC3530bis and RFC5661 Security Considerations sections
 */
7894 7895
static int
_nfs41_proc_secinfo_no_name(struct nfs_server *server, struct nfs_fh *fhandle,
7896 7897
		    struct nfs_fsinfo *info,
		    struct nfs4_secinfo_flavors *flavors, bool use_integrity)
7898 7899 7900 7901 7902 7903 7904 7905 7906 7907 7908 7909
{
	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,
	};
7910
	struct rpc_clnt *clnt = server->client;
7911
	struct rpc_cred *cred = NULL;
7912 7913 7914 7915
	int status;

	if (use_integrity) {
		clnt = server->nfs_client->cl_rpcclient;
7916 7917
		cred = nfs4_get_clid_cred(server->nfs_client);
		msg.rpc_cred = cred;
7918 7919 7920 7921 7922 7923 7924
	}

	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);

7925 7926
	if (cred)
		put_rpccred(cred);
7927 7928

	return status;
7929 7930 7931 7932 7933 7934 7935 7936 7937
}

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 {
7938 7939 7940 7941 7942 7943 7944 7945 7946 7947 7948 7949 7950 7951 7952 7953 7954 7955
		/* 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);

7956 7957 7958
		switch (err) {
		case 0:
		case -NFS4ERR_WRONGSEC:
7959
		case -ENOTSUPP:
7960
			goto out;
7961 7962 7963 7964
		default:
			err = nfs4_handle_exception(server, err, &exception);
		}
	} while (exception.retry);
7965
out:
7966 7967 7968 7969 7970 7971 7972 7973 7974
	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;
7975
	rpc_authflavor_t flavor = RPC_AUTH_MAXFLAVOR;
7976
	struct nfs4_secinfo_flavors *flavors;
7977 7978
	struct nfs4_secinfo4 *secinfo;
	int i;
7979 7980 7981 7982 7983 7984 7985 7986 7987 7988 7989 7990 7991 7992

	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
	 */
7993
	if (err == -NFS4ERR_WRONGSEC || err == -ENOTSUPP) {
7994 7995 7996 7997 7998 7999
		err = nfs4_find_root_sec(server, fhandle, info);
		goto out_freepage;
	}
	if (err)
		goto out_freepage;

8000 8001 8002 8003 8004 8005 8006 8007 8008 8009 8010 8011 8012 8013 8014
	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;
		}

8015 8016 8017
		if (!nfs_auth_info_match(&server->auth_info, flavor))
			flavor = RPC_AUTH_MAXFLAVOR;

8018 8019 8020 8021 8022 8023 8024 8025 8026 8027
		if (flavor != RPC_AUTH_MAXFLAVOR) {
			err = nfs4_lookup_root_sec(server, fhandle,
						   info, flavor);
			if (!err)
				break;
		}
	}

	if (flavor == RPC_AUTH_MAXFLAVOR)
		err = -EPERM;
8028 8029 8030 8031 8032 8033 8034 8035

out_freepage:
	put_page(page);
	if (err == -EACCES)
		return -EPERM;
out:
	return err;
}
8036

8037 8038 8039
static int _nfs41_test_stateid(struct nfs_server *server,
		nfs4_stateid *stateid,
		struct rpc_cred *cred)
B
Bryan Schumaker 已提交
8040 8041 8042
{
	int status;
	struct nfs41_test_stateid_args args = {
8043
		.stateid = stateid,
B
Bryan Schumaker 已提交
8044 8045 8046 8047 8048 8049
	};
	struct nfs41_test_stateid_res res;
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_TEST_STATEID],
		.rpc_argp = &args,
		.rpc_resp = &res,
8050
		.rpc_cred = cred,
B
Bryan Schumaker 已提交
8051
	};
8052 8053 8054 8055
	struct rpc_clnt *rpc_client = server->client;

	nfs4_state_protect(server->nfs_client, NFS_SP4_MACH_CRED_STATEID,
		&rpc_client, &msg);
8056

8057
	dprintk("NFS call  test_stateid %p\n", stateid);
8058
	nfs4_init_sequence(&args.seq_args, &res.seq_res, 0);
8059
	nfs4_set_sequence_privileged(&args.seq_args);
8060
	status = nfs4_call_sync_sequence(rpc_client, server, &msg,
8061
			&args.seq_args, &res.seq_res);
8062 8063
	if (status != NFS_OK) {
		dprintk("NFS reply test_stateid: failed, %d\n", status);
8064
		return status;
8065 8066
	}
	dprintk("NFS reply test_stateid: succeeded, %d\n", -res.status);
8067
	return -res.status;
B
Bryan Schumaker 已提交
8068 8069
}

8070 8071 8072 8073 8074
/**
 * nfs41_test_stateid - perform a TEST_STATEID operation
 *
 * @server: server / transport on which to perform the operation
 * @stateid: state ID to test
8075
 * @cred: credential
8076 8077 8078 8079 8080
 *
 * 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.
 */
8081 8082 8083
static int nfs41_test_stateid(struct nfs_server *server,
		nfs4_stateid *stateid,
		struct rpc_cred *cred)
B
Bryan Schumaker 已提交
8084 8085 8086 8087
{
	struct nfs4_exception exception = { };
	int err;
	do {
8088
		err = _nfs41_test_stateid(server, stateid, cred);
8089 8090 8091
		if (err != -NFS4ERR_DELAY)
			break;
		nfs4_handle_exception(server, err, &exception);
B
Bryan Schumaker 已提交
8092 8093 8094
	} while (exception.retry);
	return err;
}
B
Bryan Schumaker 已提交
8095

8096 8097 8098
struct nfs_free_stateid_data {
	struct nfs_server *server;
	struct nfs41_free_stateid_args args;
B
Bryan Schumaker 已提交
8099
	struct nfs41_free_stateid_res res;
8100 8101 8102 8103 8104 8105 8106 8107 8108 8109 8110 8111 8112 8113 8114 8115 8116 8117 8118 8119 8120 8121 8122 8123 8124 8125 8126 8127 8128
};

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

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

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

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

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

8129
static const struct rpc_call_ops nfs41_free_stateid_ops = {
8130 8131 8132 8133 8134 8135 8136
	.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,
8137
		struct rpc_cred *cred,
8138 8139
		bool privileged)
{
B
Bryan Schumaker 已提交
8140 8141
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_FREE_STATEID],
8142
		.rpc_cred = cred,
B
Bryan Schumaker 已提交
8143
	};
8144 8145 8146 8147 8148 8149 8150
	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
8151

8152 8153 8154
	nfs4_state_protect(server->nfs_client, NFS_SP4_MACH_CRED_STATEID,
		&task_setup.rpc_client, &msg);

8155
	dprintk("NFS call  free_stateid %p\n", stateid);
8156 8157 8158 8159 8160 8161 8162 8163 8164 8165
	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;
8166
	nfs4_init_sequence(&data->args.seq_args, &data->res.seq_res, 0);
8167 8168 8169 8170
	if (privileged)
		nfs4_set_sequence_privileged(&data->args.seq_args);

	return rpc_run_task(&task_setup);
B
Bryan Schumaker 已提交
8171 8172
}

8173 8174 8175 8176 8177
/**
 * nfs41_free_stateid - perform a FREE_STATEID operation
 *
 * @server: server / transport on which to perform the operation
 * @stateid: state ID to release
8178
 * @cred: credential
8179 8180 8181 8182
 *
 * Returns NFS_OK if the server freed "stateid".  Otherwise a
 * negative NFS4ERR value is returned.
 */
8183 8184 8185
static int nfs41_free_stateid(struct nfs_server *server,
		nfs4_stateid *stateid,
		struct rpc_cred *cred)
B
Bryan Schumaker 已提交
8186
{
8187 8188 8189
	struct rpc_task *task;
	int ret;

8190
	task = _nfs41_free_stateid(server, stateid, cred, true);
8191 8192 8193 8194 8195 8196 8197
	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
8198
}
8199

8200 8201 8202
static int nfs41_free_lock_state(struct nfs_server *server, struct nfs4_lock_state *lsp)
{
	struct rpc_task *task;
8203
	struct rpc_cred *cred = lsp->ls_state->owner->so_cred;
8204

8205
	task = _nfs41_free_stateid(server, &lsp->ls_stateid, cred, false);
8206 8207 8208 8209 8210 8211 8212
	nfs4_free_lock_state(server, lsp);
	if (IS_ERR(task))
		return PTR_ERR(task);
	rpc_put_task(task);
	return 0;
}

8213 8214 8215
static bool nfs41_match_stateid(const nfs4_stateid *s1,
		const nfs4_stateid *s2)
{
8216
	if (memcmp(s1->other, s2->other, sizeof(s1->other)) != 0)
8217 8218
		return false;

8219
	if (s1->seqid == s2->seqid)
8220
		return true;
8221
	if (s1->seqid == 0 || s2->seqid == 0)
8222 8223 8224 8225 8226
		return true;

	return false;
}

8227 8228
#endif /* CONFIG_NFS_V4_1 */

8229 8230 8231
static bool nfs4_match_stateid(const nfs4_stateid *s1,
		const nfs4_stateid *s2)
{
8232
	return nfs4_stateid_match(s1, s2);
8233 8234 8235
}


8236
static const struct nfs4_state_recovery_ops nfs40_reboot_recovery_ops = {
8237
	.owner_flag_bit = NFS_OWNER_RECLAIM_REBOOT,
8238
	.state_flag_bit	= NFS_STATE_RECLAIM_REBOOT,
L
Linus Torvalds 已提交
8239 8240
	.recover_open	= nfs4_open_reclaim,
	.recover_lock	= nfs4_lock_reclaim,
8241
	.establish_clid = nfs4_init_clientid,
8242
	.detect_trunking = nfs40_discover_server_trunking,
L
Linus Torvalds 已提交
8243 8244
};

8245
#if defined(CONFIG_NFS_V4_1)
8246
static const struct nfs4_state_recovery_ops nfs41_reboot_recovery_ops = {
8247 8248 8249 8250
	.owner_flag_bit = NFS_OWNER_RECLAIM_REBOOT,
	.state_flag_bit	= NFS_STATE_RECLAIM_REBOOT,
	.recover_open	= nfs4_open_reclaim,
	.recover_lock	= nfs4_lock_reclaim,
8251
	.establish_clid = nfs41_init_clientid,
8252
	.reclaim_complete = nfs41_proc_reclaim_complete,
8253
	.detect_trunking = nfs41_discover_server_trunking,
8254 8255 8256
};
#endif /* CONFIG_NFS_V4_1 */

8257
static const struct nfs4_state_recovery_ops nfs40_nograce_recovery_ops = {
8258 8259 8260 8261 8262 8263 8264 8265
	.owner_flag_bit = NFS_OWNER_RECLAIM_NOGRACE,
	.state_flag_bit	= NFS_STATE_RECLAIM_NOGRACE,
	.recover_open	= nfs4_open_expired,
	.recover_lock	= nfs4_lock_expired,
	.establish_clid = nfs4_init_clientid,
};

#if defined(CONFIG_NFS_V4_1)
8266
static const struct nfs4_state_recovery_ops nfs41_nograce_recovery_ops = {
8267
	.owner_flag_bit = NFS_OWNER_RECLAIM_NOGRACE,
8268
	.state_flag_bit	= NFS_STATE_RECLAIM_NOGRACE,
8269 8270
	.recover_open	= nfs41_open_expired,
	.recover_lock	= nfs41_lock_expired,
8271
	.establish_clid = nfs41_init_clientid,
L
Linus Torvalds 已提交
8272
};
8273
#endif /* CONFIG_NFS_V4_1 */
L
Linus Torvalds 已提交
8274

8275
static const struct nfs4_state_maintenance_ops nfs40_state_renewal_ops = {
B
Benny Halevy 已提交
8276
	.sched_state_renewal = nfs4_proc_async_renew,
8277
	.get_state_renewal_cred_locked = nfs4_get_renew_cred_locked,
8278
	.renew_lease = nfs4_proc_renew,
B
Benny Halevy 已提交
8279 8280 8281
};

#if defined(CONFIG_NFS_V4_1)
8282
static const struct nfs4_state_maintenance_ops nfs41_state_renewal_ops = {
B
Benny Halevy 已提交
8283
	.sched_state_renewal = nfs41_proc_async_sequence,
8284
	.get_state_renewal_cred_locked = nfs4_get_machine_cred_locked,
8285
	.renew_lease = nfs4_proc_sequence,
B
Benny Halevy 已提交
8286 8287 8288
};
#endif

8289
static const struct nfs4_mig_recovery_ops nfs40_mig_recovery_ops = {
8290
	.get_locations = _nfs40_proc_get_locations,
8291
	.fsid_present = _nfs40_proc_fsid_present,
8292 8293 8294 8295
};

#if defined(CONFIG_NFS_V4_1)
static const struct nfs4_mig_recovery_ops nfs41_mig_recovery_ops = {
8296
	.get_locations = _nfs41_proc_get_locations,
8297
	.fsid_present = _nfs41_proc_fsid_present,
8298 8299 8300
};
#endif	/* CONFIG_NFS_V4_1 */

8301 8302
static const struct nfs4_minor_version_ops nfs_v4_0_minor_ops = {
	.minor_version = 0,
8303 8304 8305 8306
	.init_caps = NFS_CAP_READDIRPLUS
		| NFS_CAP_ATOMIC_OPEN
		| NFS_CAP_CHANGE_ATTR
		| NFS_CAP_POSIX_LOCK,
8307 8308
	.init_client = nfs40_init_client,
	.shutdown_client = nfs40_shutdown_client,
8309
	.match_stateid = nfs4_match_stateid,
8310
	.find_root_sec = nfs4_find_root_sec,
8311
	.free_lock_state = nfs4_release_lockowner,
8312
	.call_sync_ops = &nfs40_call_sync_ops,
8313 8314 8315
	.reboot_recovery_ops = &nfs40_reboot_recovery_ops,
	.nograce_recovery_ops = &nfs40_nograce_recovery_ops,
	.state_renewal_ops = &nfs40_state_renewal_ops,
8316
	.mig_recovery_ops = &nfs40_mig_recovery_ops,
8317 8318 8319 8320 8321
};

#if defined(CONFIG_NFS_V4_1)
static const struct nfs4_minor_version_ops nfs_v4_1_minor_ops = {
	.minor_version = 1,
8322 8323 8324
	.init_caps = NFS_CAP_READDIRPLUS
		| NFS_CAP_ATOMIC_OPEN
		| NFS_CAP_CHANGE_ATTR
8325
		| NFS_CAP_POSIX_LOCK
8326 8327
		| NFS_CAP_STATEID_NFSV41
		| NFS_CAP_ATOMIC_OPEN_V1,
8328 8329
	.init_client = nfs41_init_client,
	.shutdown_client = nfs41_shutdown_client,
8330
	.match_stateid = nfs41_match_stateid,
8331
	.find_root_sec = nfs41_find_root_sec,
8332
	.free_lock_state = nfs41_free_lock_state,
8333
	.call_sync_ops = &nfs41_call_sync_ops,
8334 8335 8336
	.reboot_recovery_ops = &nfs41_reboot_recovery_ops,
	.nograce_recovery_ops = &nfs41_nograce_recovery_ops,
	.state_renewal_ops = &nfs41_state_renewal_ops,
8337
	.mig_recovery_ops = &nfs41_mig_recovery_ops,
8338 8339 8340
};
#endif

8341 8342 8343
#if defined(CONFIG_NFS_V4_2)
static const struct nfs4_minor_version_ops nfs_v4_2_minor_ops = {
	.minor_version = 2,
8344 8345 8346 8347 8348 8349
	.init_caps = NFS_CAP_READDIRPLUS
		| NFS_CAP_ATOMIC_OPEN
		| NFS_CAP_CHANGE_ATTR
		| NFS_CAP_POSIX_LOCK
		| NFS_CAP_STATEID_NFSV41
		| NFS_CAP_ATOMIC_OPEN_V1,
8350 8351
	.init_client = nfs41_init_client,
	.shutdown_client = nfs41_shutdown_client,
8352 8353
	.match_stateid = nfs41_match_stateid,
	.find_root_sec = nfs41_find_root_sec,
8354
	.free_lock_state = nfs41_free_lock_state,
8355
	.call_sync_ops = &nfs41_call_sync_ops,
8356 8357 8358 8359 8360 8361
	.reboot_recovery_ops = &nfs41_reboot_recovery_ops,
	.nograce_recovery_ops = &nfs41_nograce_recovery_ops,
	.state_renewal_ops = &nfs41_state_renewal_ops,
};
#endif

8362 8363 8364 8365 8366
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
8367 8368 8369
#if defined(CONFIG_NFS_V4_2)
	[2] = &nfs_v4_2_minor_ops,
#endif
8370 8371
};

8372
static const struct inode_operations nfs4_dir_inode_operations = {
8373 8374 8375 8376 8377 8378 8379 8380 8381 8382 8383 8384 8385 8386 8387 8388 8389 8390 8391
	.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,
};

8392
static const struct inode_operations nfs4_file_inode_operations = {
8393 8394 8395
	.permission	= nfs_permission,
	.getattr	= nfs_getattr,
	.setattr	= nfs_setattr,
8396 8397 8398 8399
	.getxattr	= generic_getxattr,
	.setxattr	= generic_setxattr,
	.listxattr	= generic_listxattr,
	.removexattr	= generic_removexattr,
8400 8401
};

D
David Howells 已提交
8402
const struct nfs_rpc_ops nfs_v4_clientops = {
L
Linus Torvalds 已提交
8403 8404 8405
	.version	= 4,			/* protocol version */
	.dentry_ops	= &nfs4_dentry_operations,
	.dir_inode_ops	= &nfs4_dir_inode_operations,
8406
	.file_inode_ops	= &nfs4_file_inode_operations,
8407
	.file_ops	= &nfs4_file_operations,
L
Linus Torvalds 已提交
8408
	.getroot	= nfs4_proc_get_root,
B
Bryan Schumaker 已提交
8409
	.submount	= nfs4_submount,
B
Bryan Schumaker 已提交
8410
	.try_mount	= nfs4_try_mount,
L
Linus Torvalds 已提交
8411 8412 8413 8414 8415 8416 8417 8418
	.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,
8419
	.unlink_rpc_prepare = nfs4_proc_unlink_rpc_prepare,
L
Linus Torvalds 已提交
8420
	.unlink_done	= nfs4_proc_unlink_done,
8421
	.rename_setup	= nfs4_proc_rename_setup,
8422
	.rename_rpc_prepare = nfs4_proc_rename_rpc_prepare,
8423
	.rename_done	= nfs4_proc_rename_done,
L
Linus Torvalds 已提交
8424 8425 8426 8427 8428 8429 8430 8431 8432
	.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,
8433
	.set_capabilities = nfs4_server_capabilities,
L
Linus Torvalds 已提交
8434 8435
	.decode_dirent	= nfs4_decode_dirent,
	.read_setup	= nfs4_proc_read_setup,
8436
	.read_rpc_prepare = nfs4_proc_read_rpc_prepare,
T
Trond Myklebust 已提交
8437
	.read_done	= nfs4_read_done,
L
Linus Torvalds 已提交
8438
	.write_setup	= nfs4_proc_write_setup,
8439
	.write_rpc_prepare = nfs4_proc_write_rpc_prepare,
8440
	.write_done	= nfs4_write_done,
L
Linus Torvalds 已提交
8441
	.commit_setup	= nfs4_proc_commit_setup,
8442
	.commit_rpc_prepare = nfs4_proc_commit_rpc_prepare,
8443
	.commit_done	= nfs4_commit_done,
L
Linus Torvalds 已提交
8444
	.lock		= nfs4_proc_lock,
8445
	.clear_acl_cache = nfs4_zap_acl_attr,
T
Trond Myklebust 已提交
8446
	.close_context  = nfs4_close_context,
8447
	.open_context	= nfs4_atomic_open,
8448
	.have_delegation = nfs4_have_delegation,
8449
	.return_delegation = nfs4_inode_return_delegation,
8450
	.alloc_client	= nfs4_alloc_client,
8451
	.init_client	= nfs4_init_client,
8452
	.free_client	= nfs4_free_client,
8453 8454
	.create_server	= nfs4_create_server,
	.clone_server	= nfs_clone_server,
L
Linus Torvalds 已提交
8455 8456
};

8457 8458 8459 8460 8461 8462 8463 8464 8465
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,
8466 8467 8468
#ifdef CONFIG_NFS_V4_SECURITY_LABEL
	&nfs4_xattr_nfs4_label_handler,
#endif
8469 8470 8471
	NULL
};

L
Linus Torvalds 已提交
8472 8473 8474 8475 8476
/*
 * Local variables:
 *  c-basic-offset: 8
 * End:
 */