nfs4proc.c 172.6 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
#include <linux/sunrpc/clnt.h>
46
#include <linux/sunrpc/gss_api.h>
L
Linus Torvalds 已提交
47 48 49 50
#include <linux/nfs.h>
#include <linux/nfs4.h>
#include <linux/nfs_fs.h>
#include <linux/nfs_page.h>
51
#include <linux/nfs_mount.h>
L
Linus Torvalds 已提交
52
#include <linux/namei.h>
53
#include <linux/mount.h>
B
Benny Halevy 已提交
54
#include <linux/module.h>
55
#include <linux/nfs_idmap.h>
56
#include <linux/sunrpc/bc_xprt.h>
57
#include <linux/xattr.h>
58
#include <linux/utsname.h>
59
#include <linux/freezer.h>
L
Linus Torvalds 已提交
60

61
#include "nfs4_fs.h"
L
Linus Torvalds 已提交
62
#include "delegation.h"
63
#include "internal.h"
64
#include "iostat.h"
A
Andy Adamson 已提交
65
#include "callback.h"
66
#include "pnfs.h"
L
Linus Torvalds 已提交
67 68 69

#define NFSDBG_FACILITY		NFSDBG_PROC

70
#define NFS4_POLL_RETRY_MIN	(HZ/10)
L
Linus Torvalds 已提交
71 72
#define NFS4_POLL_RETRY_MAX	(15*HZ)

73 74
#define NFS4_MAX_LOOP_ON_RECOVER (10)

75 76
static unsigned short max_session_slots = NFS4_DEF_SLOT_TABLE_SIZE;

77
struct nfs4_opendata;
78
static int _nfs4_proc_open(struct nfs4_opendata *data);
79
static int _nfs4_recover_proc_open(struct nfs4_opendata *data);
L
Linus Torvalds 已提交
80
static int nfs4_do_fsinfo(struct nfs_server *, struct nfs_fh *, struct nfs_fsinfo *);
81
static int nfs4_async_handle_error(struct rpc_task *, const struct nfs_server *, struct nfs4_state *);
82
static int _nfs4_proc_getattr(struct nfs_server *server, struct nfs_fh *fhandle, struct nfs_fattr *fattr);
83 84 85
static int nfs4_do_setattr(struct inode *inode, struct rpc_cred *cred,
			    struct nfs_fattr *fattr, struct iattr *sattr,
			    struct nfs4_state *state);
86
#ifdef CONFIG_NFS_V4_1
87 88
static int nfs41_test_stateid(struct nfs_server *, nfs4_stateid *);
static int nfs41_free_stateid(struct nfs_server *, nfs4_stateid *);
89
#endif
L
Linus Torvalds 已提交
90
/* Prevent leaks of NFSv4 errors into userland */
91
static int nfs4_map_errors(int err)
L
Linus Torvalds 已提交
92
{
93 94 95 96 97
	if (err >= -1000)
		return err;
	switch (err) {
	case -NFS4ERR_RESOURCE:
		return -EREMOTEIO;
98 99
	case -NFS4ERR_WRONGSEC:
		return -EPERM;
100 101 102
	case -NFS4ERR_BADOWNER:
	case -NFS4ERR_BADNAME:
		return -EINVAL;
103
	default:
L
Linus Torvalds 已提交
104
		dprintk("%s could not handle NFSv4 error %d\n",
105
				__func__, -err);
106
		break;
L
Linus Torvalds 已提交
107
	}
108
	return -EIO;
L
Linus Torvalds 已提交
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
}

/*
 * This is our standard bitmap for GETATTR requests.
 */
const u32 nfs4_fattr_bitmap[2] = {
	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
};

const u32 nfs4_statfs_bitmap[2] = {
	FATTR4_WORD0_FILES_AVAIL
	| FATTR4_WORD0_FILES_FREE
	| FATTR4_WORD0_FILES_TOTAL,
	FATTR4_WORD1_SPACE_AVAIL
	| FATTR4_WORD1_SPACE_FREE
	| FATTR4_WORD1_SPACE_TOTAL
};

140
const u32 nfs4_pathconf_bitmap[2] = {
L
Linus Torvalds 已提交
141 142 143 144 145
	FATTR4_WORD0_MAXLINK
	| FATTR4_WORD0_MAXNAME,
	0
};

146
const u32 nfs4_fsinfo_bitmap[3] = { FATTR4_WORD0_MAXFILESIZE
L
Linus Torvalds 已提交
147 148 149
			| FATTR4_WORD0_MAXREAD
			| FATTR4_WORD0_MAXWRITE
			| FATTR4_WORD0_LEASE_TIME,
150
			FATTR4_WORD1_TIME_DELTA
151 152
			| FATTR4_WORD1_FS_LAYOUT_TYPES,
			FATTR4_WORD2_LAYOUT_BLKSIZE
L
Linus Torvalds 已提交
153 154
};

155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173
const u32 nfs4_fs_locations_bitmap[2] = {
	FATTR4_WORD0_TYPE
	| FATTR4_WORD0_CHANGE
	| FATTR4_WORD0_SIZE
	| FATTR4_WORD0_FSID
	| FATTR4_WORD0_FILEID
	| FATTR4_WORD0_FS_LOCATIONS,
	FATTR4_WORD1_MODE
	| FATTR4_WORD1_NUMLINKS
	| FATTR4_WORD1_OWNER
	| FATTR4_WORD1_OWNER_GROUP
	| FATTR4_WORD1_RAWDEV
	| FATTR4_WORD1_SPACE_USED
	| FATTR4_WORD1_TIME_ACCESS
	| FATTR4_WORD1_TIME_METADATA
	| FATTR4_WORD1_TIME_MODIFY
	| FATTR4_WORD1_MOUNTED_ON_FILEID
};

A
Al Viro 已提交
174
static void nfs4_setup_readdir(u64 cookie, __be32 *verifier, struct dentry *dentry,
L
Linus Torvalds 已提交
175 176
		struct nfs4_readdir_arg *readdir)
{
177
	__be32 *start, *p;
L
Linus Torvalds 已提交
178 179 180

	BUG_ON(readdir->count < 80);
	if (cookie > 2) {
181
		readdir->cookie = cookie;
L
Linus Torvalds 已提交
182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197
		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.
	 */
198
	start = p = kmap_atomic(*readdir->pages, KM_USER0);
L
Linus Torvalds 已提交
199 200 201 202 203 204 205 206 207 208 209
	
	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
210
		p = xdr_encode_hyper(p, NFS_FILEID(dentry->d_inode));
L
Linus Torvalds 已提交
211 212 213 214 215 216 217 218 219 220 221
	}
	
	*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
222
	p = xdr_encode_hyper(p, NFS_FILEID(dentry->d_parent->d_inode));
L
Linus Torvalds 已提交
223 224 225 226 227 228

	readdir->pgbase = (char *)p - (char *)start;
	readdir->count -= readdir->pgbase;
	kunmap_atomic(start, KM_USER0);
}

229 230 231 232 233 234
static int nfs4_wait_clnt_recover(struct nfs_client *clp)
{
	int res;

	might_sleep();

235
	res = wait_on_bit(&clp->cl_state, NFS4CLNT_MANAGER_RUNNING,
236
			nfs_wait_bit_killable, TASK_KILLABLE);
237 238 239 240 241 242 243 244 245 246 247 248 249
	return res;
}

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;
250
	freezable_schedule_timeout_killable(*timeout);
251 252 253 254 255 256 257 258 259
	if (fatal_signal_pending(current))
		res = -ERESTARTSYS;
	*timeout <<= 1;
	return res;
}

/* This is the error handling routine for processes that are allowed
 * to sleep.
 */
260
static int nfs4_handle_exception(struct nfs_server *server, int errorcode, struct nfs4_exception *exception)
261 262
{
	struct nfs_client *clp = server->nfs_client;
263
	struct nfs4_state *state = exception->state;
264 265 266 267 268 269
	int ret = errorcode;

	exception->retry = 0;
	switch(errorcode) {
		case 0:
			return 0;
270 271 272 273 274
		case -NFS4ERR_ADMIN_REVOKED:
		case -NFS4ERR_BAD_STATEID:
		case -NFS4ERR_OPENMODE:
			if (state == NULL)
				break;
275 276
			nfs4_schedule_stateid_recovery(server, state);
			goto wait_on_recovery;
277 278 279
		case -NFS4ERR_EXPIRED:
			if (state != NULL)
				nfs4_schedule_stateid_recovery(server, state);
280
		case -NFS4ERR_STALE_STATEID:
281
		case -NFS4ERR_STALE_CLIENTID:
282 283
			nfs4_schedule_lease_recovery(clp);
			goto wait_on_recovery;
284
#if defined(CONFIG_NFS_V4_1)
285 286 287 288 289 290 291 292 293
		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);
294
			nfs4_schedule_session_recovery(clp->cl_session);
295
			exception->retry = 1;
296
			break;
297
#endif /* defined(CONFIG_NFS_V4_1) */
298
		case -NFS4ERR_FILE_OPEN:
299 300 301 302 303 304 305
			if (exception->timeout > HZ) {
				/* We have retried a decent amount, time to
				 * fail
				 */
				ret = -EBUSY;
				break;
			}
306 307
		case -NFS4ERR_GRACE:
		case -NFS4ERR_DELAY:
308
		case -EKEYEXPIRED:
309 310 311
			ret = nfs4_delay(server->client, &exception->timeout);
			if (ret != 0)
				break;
312
		case -NFS4ERR_RETRY_UNCACHED_REP:
313 314
		case -NFS4ERR_OLD_STATEID:
			exception->retry = 1;
315 316 317 318 319 320 321 322 323 324 325 326 327
			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);
			}
328 329 330
	}
	/* We failed to handle the error */
	return nfs4_map_errors(ret);
331
wait_on_recovery:
332 333 334 335
	ret = nfs4_wait_clnt_recover(clp);
	if (ret == 0)
		exception->retry = 1;
	return ret;
336 337 338
}


339
static void do_renew_lease(struct nfs_client *clp, unsigned long timestamp)
L
Linus Torvalds 已提交
340 341 342 343 344 345 346
{
	spin_lock(&clp->cl_lock);
	if (time_before(clp->cl_last_renewal,timestamp))
		clp->cl_last_renewal = timestamp;
	spin_unlock(&clp->cl_lock);
}

347 348 349 350 351
static void renew_lease(const struct nfs_server *server, unsigned long timestamp)
{
	do_renew_lease(server->nfs_client, timestamp);
}

A
Andy Adamson 已提交
352 353
#if defined(CONFIG_NFS_V4_1)

A
Andy Adamson 已提交
354 355 356 357 358 359 360 361 362 363 364
/*
 * nfs4_free_slot - free a slot and efficiently update slot table.
 *
 * freeing a slot is trivially done by clearing its respective bit
 * in the bitmap.
 * If the freed slotid equals highest_used_slotid we want to update it
 * so that the server would be able to size down the slot table if needed,
 * otherwise we know that the highest_used_slotid is still in use.
 * When updating highest_used_slotid there may be "holes" in the bitmap
 * so we need to scan down from highest_used_slotid to 0 looking for the now
 * highest slotid in use.
365
 * If none found, highest_used_slotid is set to NFS4_NO_SLOT.
366 367
 *
 * Must be called while holding tbl->slot_tbl_lock
A
Andy Adamson 已提交
368 369
 */
static void
370
nfs4_free_slot(struct nfs4_slot_table *tbl, u32 slotid)
A
Andy Adamson 已提交
371
{
372
	BUG_ON(slotid >= NFS4_MAX_SLOT_TABLE);
A
Andy Adamson 已提交
373 374 375 376 377 378
	/* clear used bit in bitmap */
	__clear_bit(slotid, tbl->used_slots);

	/* update highest_used_slotid when it is freed */
	if (slotid == tbl->highest_used_slotid) {
		slotid = find_last_bit(tbl->used_slots, tbl->max_slots);
T
Trond Myklebust 已提交
379
		if (slotid < tbl->max_slots)
A
Andy Adamson 已提交
380 381
			tbl->highest_used_slotid = slotid;
		else
382
			tbl->highest_used_slotid = NFS4_NO_SLOT;
A
Andy Adamson 已提交
383
	}
384 385
	dprintk("%s: slotid %u highest_used_slotid %d\n", __func__,
		slotid, tbl->highest_used_slotid);
A
Andy Adamson 已提交
386 387
}

388 389 390 391 392 393
bool nfs4_set_task_privileged(struct rpc_task *task, void *dummy)
{
	rpc_task_set_priority(task, RPC_PRIORITY_PRIVILEGED);
	return true;
}

394
/*
395
 * Signal state manager thread if session fore channel is drained
396
 */
397
static void nfs4_check_drain_fc_complete(struct nfs4_session *ses)
398
{
399
	if (!test_bit(NFS4_SESSION_DRAINING, &ses->session_state)) {
400 401
		rpc_wake_up_first(&ses->fc_slot_table.slot_tbl_waitq,
				nfs4_set_task_privileged, NULL);
402 403 404
		return;
	}

405
	if (ses->fc_slot_table.highest_used_slotid != NFS4_NO_SLOT)
406 407
		return;

408 409 410 411 412 413 414 415 416 417
	dprintk("%s COMPLETE: Session Fore Channel Drained\n", __func__);
	complete(&ses->fc_slot_table.complete);
}

/*
 * Signal state manager thread if session back channel is drained
 */
void nfs4_check_drain_bc_complete(struct nfs4_session *ses)
{
	if (!test_bit(NFS4_SESSION_DRAINING, &ses->session_state) ||
418
	    ses->bc_slot_table.highest_used_slotid != NFS4_NO_SLOT)
419 420 421
		return;
	dprintk("%s COMPLETE: Session Back Channel Drained\n", __func__);
	complete(&ses->bc_slot_table.complete);
422 423
}

424
static void nfs41_sequence_free_slot(struct nfs4_sequence_res *res)
A
Andy Adamson 已提交
425 426 427
{
	struct nfs4_slot_table *tbl;

428
	tbl = &res->sr_session->fc_slot_table;
429
	if (!res->sr_slot) {
A
Andy Adamson 已提交
430 431
		/* just wake up the next guy waiting since
		 * we may have not consumed a slot after all */
A
Andy Adamson 已提交
432
		dprintk("%s: No slot\n", __func__);
433
		return;
A
Andy Adamson 已提交
434
	}
435

436
	spin_lock(&tbl->slot_tbl_lock);
437
	nfs4_free_slot(tbl, res->sr_slot - tbl->slots);
438
	nfs4_check_drain_fc_complete(res->sr_session);
439
	spin_unlock(&tbl->slot_tbl_lock);
440
	res->sr_slot = NULL;
A
Andy Adamson 已提交
441 442
}

443
static int nfs41_sequence_done(struct rpc_task *task, struct nfs4_sequence_res *res)
A
Andy Adamson 已提交
444 445
{
	unsigned long timestamp;
446
	struct nfs_client *clp;
A
Andy Adamson 已提交
447 448 449 450 451 452 453 454 455 456

	/*
	 * sr_status remains 1 if an RPC level error occurred. The server
	 * may or may not have processed the sequence operation..
	 * Proceed as if the server received and processed the sequence
	 * operation.
	 */
	if (res->sr_status == 1)
		res->sr_status = NFS_OK;

457 458
	/* don't increment the sequence number if the task wasn't sent */
	if (!RPC_WAS_SENT(task))
A
Andy Adamson 已提交
459 460
		goto out;

A
Andy Adamson 已提交
461
	/* Check the SEQUENCE operation status */
462 463
	switch (res->sr_status) {
	case 0:
A
Andy Adamson 已提交
464
		/* Update the slot's sequence and clientid lease timer */
465
		++res->sr_slot->seq_nr;
A
Andy Adamson 已提交
466
		timestamp = res->sr_renewal_time;
467
		clp = res->sr_session->clp;
468
		do_renew_lease(clp, timestamp);
469
		/* Check sequence flags */
470 471
		if (res->sr_status_flags != 0)
			nfs4_schedule_lease_recovery(clp);
472 473 474 475 476 477
		break;
	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.
		 */
478
		dprintk("%s: slot=%td seq=%d: Operation in progress\n",
479 480 481
			__func__,
			res->sr_slot - res->sr_session->fc_slot_table.slots,
			res->sr_slot->seq_nr);
482 483 484
		goto out_retry;
	default:
		/* Just update the slot sequence no. */
485
		++res->sr_slot->seq_nr;
A
Andy Adamson 已提交
486 487 488 489
	}
out:
	/* The session may be reset by one of the error handlers. */
	dprintk("%s: Error %d free the slot \n", __func__, res->sr_status);
490
	nfs41_sequence_free_slot(res);
491 492
	return 1;
out_retry:
493
	if (!rpc_restart_call(task))
494 495 496
		goto out;
	rpc_delay(task, NFS4_POLL_RETRY_MAX);
	return 0;
A
Andy Adamson 已提交
497 498
}

499 500
static int nfs4_sequence_done(struct rpc_task *task,
			       struct nfs4_sequence_res *res)
501
{
502 503 504
	if (res->sr_session == NULL)
		return 1;
	return nfs41_sequence_done(task, res);
505 506
}

B
Benny Halevy 已提交
507 508 509 510 511 512
/*
 * nfs4_find_slot - efficiently look for a free slot
 *
 * nfs4_find_slot looks for an unset bit in the used_slots bitmap.
 * If found, we mark the slot as used, update the highest_used_slotid,
 * and respectively set up the sequence operation args.
513
 * The slot number is returned if found, or NFS4_NO_SLOT otherwise.
A
Andy Adamson 已提交
514 515
 *
 * Note: must be called with under the slot_tbl_lock.
B
Benny Halevy 已提交
516
 */
517
static u32
518
nfs4_find_slot(struct nfs4_slot_table *tbl)
B
Benny Halevy 已提交
519
{
520 521
	u32 slotid;
	u32 ret_id = NFS4_NO_SLOT;
B
Benny Halevy 已提交
522

523
	dprintk("--> %s used_slots=%04lx highest_used=%u max_slots=%u\n",
B
Benny Halevy 已提交
524 525 526 527 528 529
		__func__, tbl->used_slots[0], tbl->highest_used_slotid,
		tbl->max_slots);
	slotid = find_first_zero_bit(tbl->used_slots, tbl->max_slots);
	if (slotid >= tbl->max_slots)
		goto out;
	__set_bit(slotid, tbl->used_slots);
530 531
	if (slotid > tbl->highest_used_slotid ||
			tbl->highest_used_slotid == NFS4_NO_SLOT)
B
Benny Halevy 已提交
532 533 534 535 536 537 538 539
		tbl->highest_used_slotid = slotid;
	ret_id = slotid;
out:
	dprintk("<-- %s used_slots=%04lx highest_used=%d slotid=%d \n",
		__func__, tbl->used_slots[0], tbl->highest_used_slotid, ret_id);
	return ret_id;
}

540 541 542 543 544 545 546 547 548 549 550
static void nfs41_init_sequence(struct nfs4_sequence_args *args,
		struct nfs4_sequence_res *res, int cache_reply)
{
	args->sa_session = NULL;
	args->sa_cache_this = 0;
	if (cache_reply)
		args->sa_cache_this = 1;
	res->sr_session = NULL;
	res->sr_slot = NULL;
}

A
Andy Adamson 已提交
551
int nfs41_setup_sequence(struct nfs4_session *session,
A
Andy Adamson 已提交
552 553 554 555
				struct nfs4_sequence_args *args,
				struct nfs4_sequence_res *res,
				struct rpc_task *task)
{
A
Andy Adamson 已提交
556 557
	struct nfs4_slot *slot;
	struct nfs4_slot_table *tbl;
558
	u32 slotid;
A
Andy Adamson 已提交
559 560

	dprintk("--> %s\n", __func__);
A
Andy Adamson 已提交
561
	/* slot already allocated? */
562
	if (res->sr_slot != NULL)
A
Andy Adamson 已提交
563 564
		return 0;

A
Andy Adamson 已提交
565 566 567
	tbl = &session->fc_slot_table;

	spin_lock(&tbl->slot_tbl_lock);
568
	if (test_bit(NFS4_SESSION_DRAINING, &session->session_state) &&
569
	    !rpc_task_has_priority(task, RPC_PRIORITY_PRIVILEGED)) {
570
		/* The state manager will wait until the slot table is empty */
A
Andy Adamson 已提交
571
		rpc_sleep_on(&tbl->slot_tbl_waitq, task, NULL);
572
		spin_unlock(&tbl->slot_tbl_lock);
573
		dprintk("%s session is draining\n", __func__);
A
Andy Adamson 已提交
574
		return -EAGAIN;
575 576
	}

577 578 579 580 581 582 583 584
	if (!rpc_queue_empty(&tbl->slot_tbl_waitq) &&
	    !rpc_task_has_priority(task, RPC_PRIORITY_PRIVILEGED)) {
		rpc_sleep_on(&tbl->slot_tbl_waitq, task, NULL);
		spin_unlock(&tbl->slot_tbl_lock);
		dprintk("%s enforce FIFO order\n", __func__);
		return -EAGAIN;
	}

585
	slotid = nfs4_find_slot(tbl);
586
	if (slotid == NFS4_NO_SLOT) {
A
Andy Adamson 已提交
587 588 589 590 591 592 593
		rpc_sleep_on(&tbl->slot_tbl_waitq, task, NULL);
		spin_unlock(&tbl->slot_tbl_lock);
		dprintk("<-- %s: no free slots\n", __func__);
		return -EAGAIN;
	}
	spin_unlock(&tbl->slot_tbl_lock);

594
	rpc_task_set_priority(task, RPC_PRIORITY_NORMAL);
A
Andy Adamson 已提交
595
	slot = tbl->slots + slotid;
B
Benny Halevy 已提交
596
	args->sa_session = session;
A
Andy Adamson 已提交
597 598 599 600
	args->sa_slotid = slotid;

	dprintk("<-- %s slotid=%d seqid=%d\n", __func__, slotid, slot->seq_nr);

B
Benny Halevy 已提交
601
	res->sr_session = session;
602
	res->sr_slot = slot;
A
Andy Adamson 已提交
603
	res->sr_renewal_time = jiffies;
604
	res->sr_status_flags = 0;
A
Andy Adamson 已提交
605 606 607 608 609
	/*
	 * 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;
A
Andy Adamson 已提交
610 611
	return 0;
}
A
Andy Adamson 已提交
612
EXPORT_SYMBOL_GPL(nfs41_setup_sequence);
A
Andy Adamson 已提交
613

614
int nfs4_setup_sequence(const struct nfs_server *server,
A
Andy Adamson 已提交
615 616 617 618
			struct nfs4_sequence_args *args,
			struct nfs4_sequence_res *res,
			struct rpc_task *task)
{
619
	struct nfs4_session *session = nfs4_get_session(server);
A
Andy Adamson 已提交
620 621
	int ret = 0;

622
	if (session == NULL)
623 624
		goto out;

625
	dprintk("--> %s clp %p session %p sr_slot %td\n",
626 627
		__func__, session->clp, session, res->sr_slot ?
			res->sr_slot - session->fc_slot_table.slots : -1);
A
Andy Adamson 已提交
628

629
	ret = nfs41_setup_sequence(session, args, res, task);
A
Andy Adamson 已提交
630 631 632 633 634 635
out:
	dprintk("<-- %s status=%d\n", __func__, ret);
	return ret;
}

struct nfs41_call_sync_data {
636
	const struct nfs_server *seq_server;
A
Andy Adamson 已提交
637 638 639 640 641 642 643 644
	struct nfs4_sequence_args *seq_args;
	struct nfs4_sequence_res *seq_res;
};

static void nfs41_call_sync_prepare(struct rpc_task *task, void *calldata)
{
	struct nfs41_call_sync_data *data = calldata;

645 646 647
	dprintk("--> %s data->seq_server %p\n", __func__, data->seq_server);

	if (nfs4_setup_sequence(data->seq_server, data->seq_args,
648
				data->seq_res, task))
A
Andy Adamson 已提交
649 650 651 652
		return;
	rpc_call_start(task);
}

653 654 655 656 657 658
static void nfs41_call_priv_sync_prepare(struct rpc_task *task, void *calldata)
{
	rpc_task_set_priority(task, RPC_PRIORITY_PRIVILEGED);
	nfs41_call_sync_prepare(task, calldata);
}

A
Andy Adamson 已提交
659 660 661 662
static void nfs41_call_sync_done(struct rpc_task *task, void *calldata)
{
	struct nfs41_call_sync_data *data = calldata;

663
	nfs41_sequence_done(task, data->seq_res);
A
Andy Adamson 已提交
664 665
}

A
Andy Adamson 已提交
666 667
struct rpc_call_ops nfs41_call_sync_ops = {
	.rpc_call_prepare = nfs41_call_sync_prepare,
A
Andy Adamson 已提交
668
	.rpc_call_done = nfs41_call_sync_done,
A
Andy Adamson 已提交
669 670
};

671 672 673 674 675
struct rpc_call_ops nfs41_call_priv_sync_ops = {
	.rpc_call_prepare = nfs41_call_priv_sync_prepare,
	.rpc_call_done = nfs41_call_sync_done,
};

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

697 698
	if (privileged)
		task_setup.callback_ops = &nfs41_call_priv_sync_ops;
A
Andy Adamson 已提交
699 700 701 702 703 704 705 706 707 708
	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;
}

709 710
int _nfs4_call_sync_session(struct rpc_clnt *clnt,
			    struct nfs_server *server,
A
Andy Adamson 已提交
711 712 713 714 715
			    struct rpc_message *msg,
			    struct nfs4_sequence_args *args,
			    struct nfs4_sequence_res *res,
			    int cache_reply)
{
716 717
	nfs41_init_sequence(args, res, cache_reply);
	return nfs4_call_sync_sequence(clnt, server, msg, args, res, 0);
A
Andy Adamson 已提交
718 719
}

720
#else
721 722 723 724 725 726
static inline
void nfs41_init_sequence(struct nfs4_sequence_args *args,
		struct nfs4_sequence_res *res, int cache_reply)
{
}

727 728
static int nfs4_sequence_done(struct rpc_task *task,
			       struct nfs4_sequence_res *res)
729
{
730
	return 1;
731
}
A
Andy Adamson 已提交
732 733
#endif /* CONFIG_NFS_V4_1 */

734 735
int _nfs4_call_sync(struct rpc_clnt *clnt,
		    struct nfs_server *server,
A
Andy Adamson 已提交
736 737 738 739 740
		    struct rpc_message *msg,
		    struct nfs4_sequence_args *args,
		    struct nfs4_sequence_res *res,
		    int cache_reply)
{
741
	nfs41_init_sequence(args, res, cache_reply);
742
	return rpc_call_sync(clnt, msg, 0);
A
Andy Adamson 已提交
743 744
}

745
static inline
746 747
int nfs4_call_sync(struct rpc_clnt *clnt,
		   struct nfs_server *server,
748 749 750 751 752
		   struct rpc_message *msg,
		   struct nfs4_sequence_args *args,
		   struct nfs4_sequence_res *res,
		   int cache_reply)
{
753 754
	return server->nfs_client->cl_mvops->call_sync(clnt, server, msg,
						args, res, cache_reply);
755
}
A
Andy Adamson 已提交
756

757
static void update_changeattr(struct inode *dir, struct nfs4_change_info *cinfo)
L
Linus Torvalds 已提交
758
{
759
	struct nfs_inode *nfsi = NFS_I(dir);
L
Linus Torvalds 已提交
760

761
	spin_lock(&dir->i_lock);
762
	nfsi->cache_validity |= NFS_INO_INVALID_ATTR|NFS_INO_REVAL_PAGECACHE|NFS_INO_INVALID_DATA;
763
	if (!cinfo->atomic || cinfo->before != dir->i_version)
764
		nfs_force_lookup_revalidate(dir);
765
	dir->i_version = cinfo->after;
766
	spin_unlock(&dir->i_lock);
L
Linus Torvalds 已提交
767 768
}

769
struct nfs4_opendata {
770
	struct kref kref;
771 772
	struct nfs_openargs o_arg;
	struct nfs_openres o_res;
773 774
	struct nfs_open_confirmargs c_arg;
	struct nfs_open_confirmres c_res;
775 776
	struct nfs4_string owner_name;
	struct nfs4_string group_name;
777 778 779
	struct nfs_fattr f_attr;
	struct nfs_fattr dir_attr;
	struct dentry *dir;
780
	struct dentry *dentry;
781
	struct nfs4_state_owner *owner;
782
	struct nfs4_state *state;
783
	struct iattr attrs;
784
	unsigned long timestamp;
785
	unsigned int rpc_done : 1;
786 787
	int rpc_status;
	int cancelled;
788 789
};

790 791 792 793 794

static void nfs4_init_opendata_res(struct nfs4_opendata *p)
{
	p->o_res.f_attr = &p->f_attr;
	p->o_res.dir_attr = &p->dir_attr;
795 796
	p->o_res.seqid = p->o_arg.seqid;
	p->c_res.seqid = p->c_arg.seqid;
797 798 799
	p->o_res.server = p->o_arg.server;
	nfs_fattr_init(&p->f_attr);
	nfs_fattr_init(&p->dir_attr);
800
	nfs_fattr_init_names(&p->f_attr, &p->owner_name, &p->group_name);
801 802
}

803
static struct nfs4_opendata *nfs4_opendata_alloc(struct dentry *dentry,
804
		struct nfs4_state_owner *sp, fmode_t fmode, int flags,
805 806
		const struct iattr *attrs,
		gfp_t gfp_mask)
807
{
808
	struct dentry *parent = dget_parent(dentry);
809 810 811 812
	struct inode *dir = parent->d_inode;
	struct nfs_server *server = NFS_SERVER(dir);
	struct nfs4_opendata *p;

813
	p = kzalloc(sizeof(*p), gfp_mask);
814 815
	if (p == NULL)
		goto err;
816
	p->o_arg.seqid = nfs_alloc_seqid(&sp->so_seqid, gfp_mask);
817 818
	if (p->o_arg.seqid == NULL)
		goto err_free;
819 820
	nfs_sb_active(dentry->d_sb);
	p->dentry = dget(dentry);
821 822 823 824
	p->dir = parent;
	p->owner = sp;
	atomic_inc(&sp->so_count);
	p->o_arg.fh = NFS_FH(dir);
825 826
	p->o_arg.open_flags = flags;
	p->o_arg.fmode = fmode & (FMODE_READ|FMODE_WRITE);
827
	p->o_arg.clientid = server->nfs_client->cl_clientid;
828
	p->o_arg.id = sp->so_seqid.owner_id;
829
	p->o_arg.name = &dentry->d_name;
830 831
	p->o_arg.server = server;
	p->o_arg.bitmask = server->attr_bitmask;
832
	p->o_arg.dir_bitmask = server->cache_consistency_bitmask;
833
	p->o_arg.claim = NFS4_OPEN_CLAIM_NULL;
834
	if (attrs != NULL && attrs->ia_valid != 0) {
835 836
		u32 *s;

837 838
		p->o_arg.u.attrs = &p->attrs;
		memcpy(&p->attrs, attrs, sizeof(p->attrs));
839 840 841
		s = (u32 *) p->o_arg.u.verifier.data;
		s[0] = jiffies;
		s[1] = current->pid;
842
	}
843 844 845
	p->c_arg.fh = &p->o_res.fh;
	p->c_arg.stateid = &p->o_res.stateid;
	p->c_arg.seqid = p->o_arg.seqid;
846
	nfs4_init_opendata_res(p);
847
	kref_init(&p->kref);
848 849 850 851 852 853 854 855
	return p;
err_free:
	kfree(p);
err:
	dput(parent);
	return NULL;
}

856
static void nfs4_opendata_free(struct kref *kref)
857
{
858 859
	struct nfs4_opendata *p = container_of(kref,
			struct nfs4_opendata, kref);
860
	struct super_block *sb = p->dentry->d_sb;
861 862

	nfs_free_seqid(p->o_arg.seqid);
863 864
	if (p->state != NULL)
		nfs4_put_open_state(p->state);
865 866
	nfs4_put_state_owner(p->owner);
	dput(p->dir);
867 868
	dput(p->dentry);
	nfs_sb_deactive(sb);
869
	nfs_fattr_free_names(&p->f_attr);
870 871 872 873 874 875 876
	kfree(p);
}

static void nfs4_opendata_put(struct nfs4_opendata *p)
{
	if (p != NULL)
		kref_put(&p->kref, nfs4_opendata_free);
877 878
}

879 880 881 882 883 884 885 886
static int nfs4_wait_for_completion_rpc_task(struct rpc_task *task)
{
	int ret;

	ret = rpc_wait_for_completion_task(task);
	return ret;
}

887
static int can_open_cached(struct nfs4_state *state, fmode_t mode, int open_mode)
888 889
{
	int ret = 0;
890

891
	if (open_mode & (O_EXCL|O_TRUNC))
892 893
		goto out;
	switch (mode & (FMODE_READ|FMODE_WRITE)) {
894
		case FMODE_READ:
895 896
			ret |= test_bit(NFS_O_RDONLY_STATE, &state->flags) != 0
				&& state->n_rdonly != 0;
897 898
			break;
		case FMODE_WRITE:
899 900
			ret |= test_bit(NFS_O_WRONLY_STATE, &state->flags) != 0
				&& state->n_wronly != 0;
901 902
			break;
		case FMODE_READ|FMODE_WRITE:
903 904
			ret |= test_bit(NFS_O_RDWR_STATE, &state->flags) != 0
				&& state->n_rdwr != 0;
905
	}
906
out:
907 908 909
	return ret;
}

910
static int can_open_delegated(struct nfs_delegation *delegation, fmode_t fmode)
911
{
912 913
	if (delegation == NULL)
		return 0;
914
	if ((delegation->type & fmode) != fmode)
915
		return 0;
916
	if (test_bit(NFS_DELEGATION_NEED_RECLAIM, &delegation->flags))
917
		return 0;
918
	nfs_mark_delegation_referenced(delegation);
919 920 921
	return 1;
}

922
static void update_open_stateflags(struct nfs4_state *state, fmode_t fmode)
923
{
924
	switch (fmode) {
925 926 927 928 929 930 931 932 933
		case FMODE_WRITE:
			state->n_wronly++;
			break;
		case FMODE_READ:
			state->n_rdonly++;
			break;
		case FMODE_READ|FMODE_WRITE:
			state->n_rdwr++;
	}
934
	nfs4_state_set_mode_locked(state, state->state | fmode);
935 936
}

937
static void nfs_set_open_stateid_locked(struct nfs4_state *state, nfs4_stateid *stateid, fmode_t fmode)
938 939 940 941
{
	if (test_bit(NFS_DELEGATED_STATE, &state->flags) == 0)
		memcpy(state->stateid.data, stateid->data, sizeof(state->stateid.data));
	memcpy(state->open_stateid.data, stateid->data, sizeof(state->open_stateid.data));
942
	switch (fmode) {
943 944 945 946 947 948 949 950 951
		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);
	}
952 953
}

954
static void nfs_set_open_stateid(struct nfs4_state *state, nfs4_stateid *stateid, fmode_t fmode)
955
{
956
	write_seqlock(&state->seqlock);
957
	nfs_set_open_stateid_locked(state, stateid, fmode);
958
	write_sequnlock(&state->seqlock);
959 960
}

961
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
962
{
963 964 965 966 967
	/*
	 * Protect the call to nfs4_state_set_mode_locked and
	 * serialise the stateid update
	 */
	write_seqlock(&state->seqlock);
968 969 970 971 972
	if (deleg_stateid != NULL) {
		memcpy(state->stateid.data, deleg_stateid->data, sizeof(state->stateid.data));
		set_bit(NFS_DELEGATED_STATE, &state->flags);
	}
	if (open_stateid != NULL)
973
		nfs_set_open_stateid_locked(state, open_stateid, fmode);
974 975
	write_sequnlock(&state->seqlock);
	spin_lock(&state->owner->so_lock);
976
	update_open_stateflags(state, fmode);
977
	spin_unlock(&state->owner->so_lock);
L
Linus Torvalds 已提交
978 979
}

980
static int update_open_stateid(struct nfs4_state *state, nfs4_stateid *open_stateid, nfs4_stateid *delegation, fmode_t fmode)
981 982 983 984 985
{
	struct nfs_inode *nfsi = NFS_I(state->inode);
	struct nfs_delegation *deleg_cur;
	int ret = 0;

986
	fmode &= (FMODE_READ|FMODE_WRITE);
987 988 989 990 991 992 993 994

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

	spin_lock(&deleg_cur->lock);
	if (nfsi->delegation != deleg_cur ||
995
	    (deleg_cur->type & fmode) != fmode)
996 997 998 999 1000 1001 1002
		goto no_delegation_unlock;

	if (delegation == NULL)
		delegation = &deleg_cur->stateid;
	else if (memcmp(deleg_cur->stateid.data, delegation->data, NFS4_STATEID_SIZE) != 0)
		goto no_delegation_unlock;

1003
	nfs_mark_delegation_referenced(deleg_cur);
1004
	__update_open_stateid(state, open_stateid, &deleg_cur->stateid, fmode);
1005 1006 1007 1008 1009 1010 1011
	ret = 1;
no_delegation_unlock:
	spin_unlock(&deleg_cur->lock);
no_delegation:
	rcu_read_unlock();

	if (!ret && open_stateid != NULL) {
1012
		__update_open_stateid(state, open_stateid, NULL, fmode);
1013 1014 1015 1016 1017 1018 1019
		ret = 1;
	}

	return ret;
}


1020
static void nfs4_return_incompatible_delegation(struct inode *inode, fmode_t fmode)
1021 1022 1023 1024 1025
{
	struct nfs_delegation *delegation;

	rcu_read_lock();
	delegation = rcu_dereference(NFS_I(inode)->delegation);
1026
	if (delegation == NULL || (delegation->type & fmode) == fmode) {
1027 1028 1029 1030 1031 1032 1033
		rcu_read_unlock();
		return;
	}
	rcu_read_unlock();
	nfs_inode_return_delegation(inode);
}

1034
static struct nfs4_state *nfs4_try_open_cached(struct nfs4_opendata *opendata)
1035 1036 1037 1038
{
	struct nfs4_state *state = opendata->state;
	struct nfs_inode *nfsi = NFS_I(state->inode);
	struct nfs_delegation *delegation;
1039
	int open_mode = opendata->o_arg.open_flags & (O_EXCL|O_TRUNC);
1040
	fmode_t fmode = opendata->o_arg.fmode;
1041 1042 1043 1044
	nfs4_stateid stateid;
	int ret = -EAGAIN;

	for (;;) {
1045
		if (can_open_cached(state, fmode, open_mode)) {
1046
			spin_lock(&state->owner->so_lock);
1047 1048
			if (can_open_cached(state, fmode, open_mode)) {
				update_open_stateflags(state, fmode);
1049 1050 1051 1052 1053
				spin_unlock(&state->owner->so_lock);
				goto out_return_state;
			}
			spin_unlock(&state->owner->so_lock);
		}
1054 1055
		rcu_read_lock();
		delegation = rcu_dereference(nfsi->delegation);
1056
		if (!can_open_delegated(delegation, fmode)) {
1057
			rcu_read_unlock();
1058
			break;
1059
		}
1060 1061 1062
		/* Save the delegation */
		memcpy(stateid.data, delegation->stateid.data, sizeof(stateid.data));
		rcu_read_unlock();
1063
		ret = nfs_may_open(state->inode, state->owner->so_cred, open_mode);
1064 1065 1066
		if (ret != 0)
			goto out;
		ret = -EAGAIN;
1067 1068

		/* Try to update the stateid using the delegation */
1069
		if (update_open_stateid(state, NULL, &stateid, fmode))
1070
			goto out_return_state;
1071 1072 1073 1074 1075 1076 1077 1078
	}
out:
	return ERR_PTR(ret);
out_return_state:
	atomic_inc(&state->count);
	return state;
}

1079 1080 1081 1082
static struct nfs4_state *nfs4_opendata_to_nfs4_state(struct nfs4_opendata *data)
{
	struct inode *inode;
	struct nfs4_state *state = NULL;
1083
	struct nfs_delegation *delegation;
1084
	int ret;
1085

1086
	if (!data->rpc_done) {
1087
		state = nfs4_try_open_cached(data);
1088 1089 1090
		goto out;
	}

1091
	ret = -EAGAIN;
1092
	if (!(data->f_attr.valid & NFS_ATTR_FATTR))
1093
		goto err;
1094
	inode = nfs_fhget(data->dir->d_sb, &data->o_res.fh, &data->f_attr);
1095
	ret = PTR_ERR(inode);
1096
	if (IS_ERR(inode))
1097 1098
		goto err;
	ret = -ENOMEM;
1099 1100
	state = nfs4_get_open_state(inode, data->owner);
	if (state == NULL)
1101
		goto err_put_inode;
1102 1103 1104
	if (data->o_res.delegation_type != 0) {
		int delegation_flags = 0;

1105 1106 1107 1108 1109
		rcu_read_lock();
		delegation = rcu_dereference(NFS_I(inode)->delegation);
		if (delegation)
			delegation_flags = delegation->flags;
		rcu_read_unlock();
1110 1111 1112 1113 1114 1115
		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",
					NFS_CLIENT(inode)->cl_server);
		} else if ((delegation_flags & 1UL<<NFS_DELEGATION_NEED_RECLAIM) == 0)
1116 1117 1118 1119 1120 1121 1122 1123
			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);
	}
1124 1125

	update_open_stateid(state, &data->o_res.stateid, NULL,
1126
			data->o_arg.fmode);
1127
	iput(inode);
1128
out:
1129
	return state;
1130 1131 1132 1133
err_put_inode:
	iput(inode);
err:
	return ERR_PTR(ret);
1134 1135
}

1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152
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);
}

T
Trond Myklebust 已提交
1153 1154 1155 1156
static struct nfs4_opendata *nfs4_open_recoverdata_alloc(struct nfs_open_context *ctx, struct nfs4_state *state)
{
	struct nfs4_opendata *opendata;

1157
	opendata = nfs4_opendata_alloc(ctx->dentry, state->owner, 0, 0, NULL, GFP_NOFS);
T
Trond Myklebust 已提交
1158 1159 1160 1161 1162 1163 1164
	if (opendata == NULL)
		return ERR_PTR(-ENOMEM);
	opendata->state = state;
	atomic_inc(&state->count);
	return opendata;
}

1165
static int nfs4_open_recover_helper(struct nfs4_opendata *opendata, fmode_t fmode, struct nfs4_state **res)
1166
{
1167
	struct nfs4_state *newstate;
1168 1169
	int ret;

1170 1171
	opendata->o_arg.open_flags = 0;
	opendata->o_arg.fmode = fmode;
1172 1173 1174
	memset(&opendata->o_res, 0, sizeof(opendata->o_res));
	memset(&opendata->c_res, 0, sizeof(opendata->c_res));
	nfs4_init_opendata_res(opendata);
1175
	ret = _nfs4_recover_proc_open(opendata);
1176 1177
	if (ret != 0)
		return ret; 
1178
	newstate = nfs4_opendata_to_nfs4_state(opendata);
1179 1180
	if (IS_ERR(newstate))
		return PTR_ERR(newstate);
1181
	nfs4_close_state(newstate, fmode);
1182
	*res = newstate;
1183 1184 1185 1186 1187 1188 1189 1190 1191
	return 0;
}

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

	/* memory barrier prior to reading state->n_* */
1192
	clear_bit(NFS_DELEGATED_STATE, &state->flags);
1193 1194
	smp_rmb();
	if (state->n_rdwr != 0) {
T
Trond Myklebust 已提交
1195
		clear_bit(NFS_O_RDWR_STATE, &state->flags);
1196
		ret = nfs4_open_recover_helper(opendata, FMODE_READ|FMODE_WRITE, &newstate);
1197 1198
		if (ret != 0)
			return ret;
1199 1200
		if (newstate != state)
			return -ESTALE;
1201 1202
	}
	if (state->n_wronly != 0) {
T
Trond Myklebust 已提交
1203
		clear_bit(NFS_O_WRONLY_STATE, &state->flags);
1204
		ret = nfs4_open_recover_helper(opendata, FMODE_WRITE, &newstate);
1205 1206
		if (ret != 0)
			return ret;
1207 1208
		if (newstate != state)
			return -ESTALE;
1209 1210
	}
	if (state->n_rdonly != 0) {
T
Trond Myklebust 已提交
1211
		clear_bit(NFS_O_RDONLY_STATE, &state->flags);
1212
		ret = nfs4_open_recover_helper(opendata, FMODE_READ, &newstate);
1213 1214
		if (ret != 0)
			return ret;
1215 1216
		if (newstate != state)
			return -ESTALE;
1217
	}
1218 1219 1220 1221 1222 1223
	/*
	 * 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 &&
	    memcmp(state->stateid.data, state->open_stateid.data, sizeof(state->stateid.data)) != 0) {
1224
		write_seqlock(&state->seqlock);
1225 1226
		if (test_bit(NFS_DELEGATED_STATE, &state->flags) == 0)
			memcpy(state->stateid.data, state->open_stateid.data, sizeof(state->stateid.data));
1227
		write_sequnlock(&state->seqlock);
1228
	}
1229 1230 1231
	return 0;
}

L
Linus Torvalds 已提交
1232 1233 1234 1235
/*
 * OPEN_RECLAIM:
 * 	reclaim state on the server after a reboot.
 */
1236
static int _nfs4_do_open_reclaim(struct nfs_open_context *ctx, struct nfs4_state *state)
L
Linus Torvalds 已提交
1237
{
1238
	struct nfs_delegation *delegation;
1239
	struct nfs4_opendata *opendata;
1240
	fmode_t delegation_type = 0;
L
Linus Torvalds 已提交
1241 1242
	int status;

T
Trond Myklebust 已提交
1243 1244 1245
	opendata = nfs4_open_recoverdata_alloc(ctx, state);
	if (IS_ERR(opendata))
		return PTR_ERR(opendata);
1246 1247
	opendata->o_arg.claim = NFS4_OPEN_CLAIM_PREVIOUS;
	opendata->o_arg.fh = NFS_FH(state->inode);
1248 1249
	rcu_read_lock();
	delegation = rcu_dereference(NFS_I(state->inode)->delegation);
1250
	if (delegation != NULL && test_bit(NFS_DELEGATION_NEED_RECLAIM, &delegation->flags) != 0)
1251
		delegation_type = delegation->type;
1252
	rcu_read_unlock();
1253 1254
	opendata->o_arg.u.delegation_type = delegation_type;
	status = nfs4_open_recover(opendata, state);
1255
	nfs4_opendata_put(opendata);
L
Linus Torvalds 已提交
1256 1257 1258
	return status;
}

1259
static int nfs4_do_open_reclaim(struct nfs_open_context *ctx, struct nfs4_state *state)
L
Linus Torvalds 已提交
1260 1261 1262 1263 1264
{
	struct nfs_server *server = NFS_SERVER(state->inode);
	struct nfs4_exception exception = { };
	int err;
	do {
1265
		err = _nfs4_do_open_reclaim(ctx, state);
1266
		if (err != -NFS4ERR_DELAY)
1267 1268
			break;
		nfs4_handle_exception(server, err, &exception);
L
Linus Torvalds 已提交
1269 1270 1271 1272
	} while (exception.retry);
	return err;
}

1273 1274 1275 1276 1277 1278 1279 1280
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))
		return PTR_ERR(ctx);
1281
	ret = nfs4_do_open_reclaim(ctx, state);
1282 1283 1284 1285
	put_nfs_open_context(ctx);
	return ret;
}

1286
static int _nfs4_open_delegation_recall(struct nfs_open_context *ctx, struct nfs4_state *state, const nfs4_stateid *stateid)
L
Linus Torvalds 已提交
1287
{
1288
	struct nfs4_opendata *opendata;
1289
	int ret;
L
Linus Torvalds 已提交
1290

T
Trond Myklebust 已提交
1291 1292 1293
	opendata = nfs4_open_recoverdata_alloc(ctx, state);
	if (IS_ERR(opendata))
		return PTR_ERR(opendata);
1294
	opendata->o_arg.claim = NFS4_OPEN_CLAIM_DELEGATE_CUR;
1295
	memcpy(opendata->o_arg.u.delegation.data, stateid->data,
1296
			sizeof(opendata->o_arg.u.delegation.data));
1297
	ret = nfs4_open_recover(opendata, state);
1298
	nfs4_opendata_put(opendata);
1299
	return ret;
L
Linus Torvalds 已提交
1300 1301
}

1302
int nfs4_open_delegation_recall(struct nfs_open_context *ctx, struct nfs4_state *state, const nfs4_stateid *stateid)
L
Linus Torvalds 已提交
1303 1304
{
	struct nfs4_exception exception = { };
1305
	struct nfs_server *server = NFS_SERVER(state->inode);
L
Linus Torvalds 已提交
1306 1307
	int err;
	do {
1308
		err = _nfs4_open_delegation_recall(ctx, state, stateid);
L
Linus Torvalds 已提交
1309 1310
		switch (err) {
			case 0:
1311 1312 1313
			case -ENOENT:
			case -ESTALE:
				goto out;
1314 1315 1316 1317 1318
			case -NFS4ERR_BADSESSION:
			case -NFS4ERR_BADSLOT:
			case -NFS4ERR_BAD_HIGH_SLOT:
			case -NFS4ERR_CONN_NOT_BOUND_TO_SESSION:
			case -NFS4ERR_DEADSESSION:
1319
				nfs4_schedule_session_recovery(server->nfs_client->cl_session);
1320
				goto out;
L
Linus Torvalds 已提交
1321 1322 1323 1324
			case -NFS4ERR_STALE_CLIENTID:
			case -NFS4ERR_STALE_STATEID:
			case -NFS4ERR_EXPIRED:
				/* Don't recall a delegation if it was lost */
1325
				nfs4_schedule_lease_recovery(server->nfs_client);
1326 1327 1328 1329 1330 1331 1332 1333
				goto out;
			case -ERESTARTSYS:
				/*
				 * The show must go on: exit, but mark the
				 * stateid as needing recovery.
				 */
			case -NFS4ERR_ADMIN_REVOKED:
			case -NFS4ERR_BAD_STATEID:
1334
				nfs4_schedule_stateid_recovery(server, state);
1335 1336 1337 1338 1339 1340 1341
			case -EKEYEXPIRED:
				/*
				 * User RPCSEC_GSS context has expired.
				 * We cannot recover this stateid now, so
				 * skip it and allow recovery thread to
				 * proceed.
				 */
1342 1343 1344
			case -ENOMEM:
				err = 0;
				goto out;
L
Linus Torvalds 已提交
1345 1346 1347
		}
		err = nfs4_handle_exception(server, err, &exception);
	} while (exception.retry);
1348
out:
L
Linus Torvalds 已提交
1349 1350 1351
	return err;
}

1352 1353 1354 1355 1356
static void nfs4_open_confirm_done(struct rpc_task *task, void *calldata)
{
	struct nfs4_opendata *data = calldata;

	data->rpc_status = task->tk_status;
1357
	if (data->rpc_status == 0) {
1358 1359
		memcpy(data->o_res.stateid.data, data->c_res.stateid.data,
				sizeof(data->o_res.stateid.data));
1360
		nfs_confirm_seqid(&data->owner->so_seqid, 0);
1361
		renew_lease(data->o_res.server, data->timestamp);
1362
		data->rpc_done = 1;
1363
	}
1364 1365 1366 1367 1368 1369 1370 1371 1372 1373 1374
}

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! */
1375
	if (!data->rpc_done)
1376 1377
		goto out_free;
	state = nfs4_opendata_to_nfs4_state(data);
1378
	if (!IS_ERR(state))
1379
		nfs4_close_state(state, data->o_arg.fmode);
1380
out_free:
1381
	nfs4_opendata_put(data);
1382 1383 1384 1385 1386 1387 1388 1389 1390 1391 1392 1393 1394 1395
}

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

/*
 * Note: On error, nfs4_proc_open_confirm will free the struct nfs4_opendata
 */
static int _nfs4_proc_open_confirm(struct nfs4_opendata *data)
{
	struct nfs_server *server = NFS_SERVER(data->dir->d_inode);
	struct rpc_task *task;
1396 1397 1398 1399 1400 1401
	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
1402 1403
	struct rpc_task_setup task_setup_data = {
		.rpc_client = server->client,
1404
		.rpc_message = &msg,
T
Trond Myklebust 已提交
1405 1406
		.callback_ops = &nfs4_open_confirm_ops,
		.callback_data = data,
1407
		.workqueue = nfsiod_workqueue,
T
Trond Myklebust 已提交
1408 1409
		.flags = RPC_TASK_ASYNC,
	};
L
Linus Torvalds 已提交
1410 1411
	int status;

1412
	kref_get(&data->kref);
1413 1414
	data->rpc_done = 0;
	data->rpc_status = 0;
1415
	data->timestamp = jiffies;
T
Trond Myklebust 已提交
1416
	task = rpc_run_task(&task_setup_data);
1417
	if (IS_ERR(task))
1418 1419 1420 1421 1422 1423 1424
		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;
1425
	rpc_put_task(task);
L
Linus Torvalds 已提交
1426 1427 1428
	return status;
}

1429
static void nfs4_open_prepare(struct rpc_task *task, void *calldata)
L
Linus Torvalds 已提交
1430
{
1431 1432
	struct nfs4_opendata *data = calldata;
	struct nfs4_state_owner *sp = data->owner;
1433

1434 1435
	if (nfs_wait_on_sequence(data->o_arg.seqid, task) != 0)
		return;
1436 1437 1438 1439 1440 1441 1442
	/*
	 * 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;

1443
		if (can_open_cached(data->state, data->o_arg.fmode, data->o_arg.open_flags))
1444
			goto out_no_action;
1445 1446
		rcu_read_lock();
		delegation = rcu_dereference(NFS_I(data->state->inode)->delegation);
1447 1448 1449
		if (data->o_arg.claim != NFS4_OPEN_CLAIM_DELEGATE_CUR &&
		    can_open_delegated(delegation, data->o_arg.fmode))
			goto unlock_no_action;
1450 1451
		rcu_read_unlock();
	}
1452
	/* Update sequence id. */
1453
	data->o_arg.id = sp->so_seqid.owner_id;
1454
	data->o_arg.clientid = sp->so_server->nfs_client->cl_clientid;
T
Trond Myklebust 已提交
1455
	if (data->o_arg.claim == NFS4_OPEN_CLAIM_PREVIOUS) {
1456
		task->tk_msg.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_OPEN_NOATTR];
T
Trond Myklebust 已提交
1457 1458
		nfs_copy_fh(&data->o_res.fh, data->o_arg.fh);
	}
1459
	data->timestamp = jiffies;
1460
	if (nfs4_setup_sequence(data->o_arg.server,
1461
				&data->o_arg.seq_args,
1462
				&data->o_res.seq_res, task))
1463
		return;
1464
	rpc_call_start(task);
1465
	return;
1466 1467
unlock_no_action:
	rcu_read_unlock();
1468 1469 1470
out_no_action:
	task->tk_action = NULL;

1471
}
L
Linus Torvalds 已提交
1472

1473 1474 1475 1476 1477 1478
static void nfs4_recover_open_prepare(struct rpc_task *task, void *calldata)
{
	rpc_task_set_priority(task, RPC_PRIORITY_PRIVILEGED);
	nfs4_open_prepare(task, calldata);
}

1479 1480 1481
static void nfs4_open_done(struct rpc_task *task, void *calldata)
{
	struct nfs4_opendata *data = calldata;
L
Linus Torvalds 已提交
1482

1483
	data->rpc_status = task->tk_status;
1484

1485 1486
	if (!nfs4_sequence_done(task, &data->o_res.seq_res))
		return;
1487

1488 1489
	if (task->tk_status == 0) {
		switch (data->o_res.f_attr->mode & S_IFMT) {
1490 1491 1492
			case S_IFREG:
				break;
			case S_IFLNK:
1493
				data->rpc_status = -ELOOP;
1494 1495
				break;
			case S_IFDIR:
1496
				data->rpc_status = -EISDIR;
1497 1498
				break;
			default:
1499
				data->rpc_status = -ENOTDIR;
1500
		}
1501
		renew_lease(data->o_res.server, data->timestamp);
1502 1503
		if (!(data->o_res.rflags & NFS4_OPEN_RESULT_CONFIRM))
			nfs_confirm_seqid(&data->owner->so_seqid, 0);
1504
	}
1505
	data->rpc_done = 1;
1506
}
1507

1508 1509 1510 1511 1512 1513 1514 1515 1516
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! */
1517
	if (data->rpc_status != 0 || !data->rpc_done)
1518 1519 1520 1521 1522
		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);
1523
	if (!IS_ERR(state))
1524
		nfs4_close_state(state, data->o_arg.fmode);
1525
out_free:
1526
	nfs4_opendata_put(data);
1527 1528 1529 1530 1531 1532 1533 1534
}

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,
};

1535 1536 1537 1538 1539 1540 1541
static const struct rpc_call_ops nfs4_recover_open_ops = {
	.rpc_call_prepare = nfs4_recover_open_prepare,
	.rpc_call_done = nfs4_open_done,
	.rpc_release = nfs4_open_release,
};

static int nfs4_run_open_task(struct nfs4_opendata *data, int isrecover)
1542 1543 1544 1545 1546 1547
{
	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;
1548 1549 1550 1551 1552 1553
	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
1554 1555
	struct rpc_task_setup task_setup_data = {
		.rpc_client = server->client,
1556
		.rpc_message = &msg,
T
Trond Myklebust 已提交
1557 1558
		.callback_ops = &nfs4_open_ops,
		.callback_data = data,
1559
		.workqueue = nfsiod_workqueue,
T
Trond Myklebust 已提交
1560 1561
		.flags = RPC_TASK_ASYNC,
	};
1562 1563
	int status;

1564
	nfs41_init_sequence(&o_arg->seq_args, &o_res->seq_res, 1);
1565
	kref_get(&data->kref);
1566 1567
	data->rpc_done = 0;
	data->rpc_status = 0;
1568
	data->cancelled = 0;
1569 1570
	if (isrecover)
		task_setup_data.callback_ops = &nfs4_recover_open_ops;
T
Trond Myklebust 已提交
1571
	task = rpc_run_task(&task_setup_data);
1572 1573 1574 1575 1576 1577 1578 1579 1580 1581 1582 1583 1584 1585 1586 1587 1588 1589 1590 1591 1592 1593 1594
        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;

1595 1596
	nfs_fattr_map_and_free_names(NFS_SERVER(dir), &data->f_attr);

1597 1598 1599 1600 1601 1602 1603 1604 1605 1606 1607 1608 1609 1610 1611 1612 1613 1614 1615 1616 1617 1618 1619
	nfs_refresh_inode(dir, o_res->dir_attr);

	if (o_res->rflags & NFS4_OPEN_RESULT_CONFIRM) {
		status = _nfs4_proc_open_confirm(data);
		if (status != 0)
			return status;
	}

	return status;
}

/*
 * 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);
1620 1621 1622 1623 1624 1625
	if (!data->rpc_done)
		return status;
	if (status != 0) {
		if (status == -NFS4ERR_BADNAME &&
				!(o_arg->open_flags & O_CREAT))
			return -ENOENT;
1626
		return status;
1627
	}
1628

1629 1630
	nfs_fattr_map_and_free_names(server, &data->f_attr);

1631 1632 1633 1634 1635
	if (o_arg->open_flags & O_CREAT) {
		update_changeattr(dir, &o_res->cinfo);
		nfs_post_op_update_inode(dir, o_res->dir_attr);
	} else
		nfs_refresh_inode(dir, o_res->dir_attr);
T
Trond Myklebust 已提交
1636 1637
	if ((o_res->rflags & NFS4_OPEN_RESULT_LOCKTYPE_POSIX) == 0)
		server->caps &= ~NFS_CAP_POSIX_LOCK;
L
Linus Torvalds 已提交
1638
	if(o_res->rflags & NFS4_OPEN_RESULT_CONFIRM) {
1639
		status = _nfs4_proc_open_confirm(data);
L
Linus Torvalds 已提交
1640
		if (status != 0)
1641
			return status;
L
Linus Torvalds 已提交
1642 1643
	}
	if (!(o_res->f_attr->valid & NFS_ATTR_FATTR))
1644
		_nfs4_proc_getattr(server, &o_res->fh, o_res->f_attr);
1645
	return 0;
L
Linus Torvalds 已提交
1646 1647
}

A
Andy Adamson 已提交
1648
static int nfs4_client_recover_expired_lease(struct nfs_client *clp)
1649
{
1650
	unsigned int loop;
1651
	int ret;
1652

1653
	for (loop = NFS4_MAX_LOOP_ON_RECOVER; loop != 0; loop--) {
1654
		ret = nfs4_wait_clnt_recover(clp);
1655
		if (ret != 0)
1656
			break;
1657 1658
		if (!test_bit(NFS4CLNT_LEASE_EXPIRED, &clp->cl_state) &&
		    !test_bit(NFS4CLNT_CHECK_LEASE,&clp->cl_state))
1659
			break;
1660
		nfs4_schedule_state_manager(clp);
1661
		ret = -EIO;
1662
	}
1663
	return ret;
1664 1665
}

A
Andy Adamson 已提交
1666 1667 1668 1669 1670
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
1671 1672 1673 1674 1675
/*
 * OPEN_EXPIRED:
 * 	reclaim state on the server after a network partition.
 * 	Assumes caller holds the appropriate lock
 */
1676
static int _nfs4_open_expired(struct nfs_open_context *ctx, struct nfs4_state *state)
L
Linus Torvalds 已提交
1677
{
1678
	struct nfs4_opendata *opendata;
1679
	int ret;
L
Linus Torvalds 已提交
1680

T
Trond Myklebust 已提交
1681 1682 1683
	opendata = nfs4_open_recoverdata_alloc(ctx, state);
	if (IS_ERR(opendata))
		return PTR_ERR(opendata);
1684
	ret = nfs4_open_recover(opendata, state);
1685
	if (ret == -ESTALE)
1686
		d_drop(ctx->dentry);
1687
	nfs4_opendata_put(opendata);
1688
	return ret;
L
Linus Torvalds 已提交
1689 1690
}

1691
static int nfs4_do_open_expired(struct nfs_open_context *ctx, struct nfs4_state *state)
1692
{
1693
	struct nfs_server *server = NFS_SERVER(state->inode);
1694 1695 1696 1697
	struct nfs4_exception exception = { };
	int err;

	do {
1698
		err = _nfs4_open_expired(ctx, state);
1699 1700 1701 1702 1703 1704 1705 1706
		switch (err) {
		default:
			goto out;
		case -NFS4ERR_GRACE:
		case -NFS4ERR_DELAY:
			nfs4_handle_exception(server, err, &exception);
			err = 0;
		}
1707
	} while (exception.retry);
1708
out:
1709 1710 1711
	return err;
}

L
Linus Torvalds 已提交
1712 1713 1714
static int nfs4_open_expired(struct nfs4_state_owner *sp, struct nfs4_state *state)
{
	struct nfs_open_context *ctx;
1715
	int ret;
L
Linus Torvalds 已提交
1716

1717 1718 1719
	ctx = nfs4_state_find_open_context(state);
	if (IS_ERR(ctx))
		return PTR_ERR(ctx);
1720
	ret = nfs4_do_open_expired(ctx, state);
1721 1722
	put_nfs_open_context(ctx);
	return ret;
L
Linus Torvalds 已提交
1723 1724
}

1725
#if defined(CONFIG_NFS_V4_1)
1726
static int nfs41_check_expired_stateid(struct nfs4_state *state, nfs4_stateid *stateid, unsigned int flags)
1727
{
1728
	int status = NFS_OK;
1729 1730
	struct nfs_server *server = NFS_SERVER(state->inode);

1731 1732 1733 1734 1735 1736 1737 1738 1739 1740 1741 1742 1743 1744 1745 1746 1747 1748 1749 1750 1751
	if (state->flags & flags) {
		status = nfs41_test_stateid(server, stateid);
		if (status != NFS_OK) {
			nfs41_free_stateid(server, stateid);
			state->flags &= ~flags;
		}
	}
	return status;
}

static int nfs41_open_expired(struct nfs4_state_owner *sp, struct nfs4_state *state)
{
	int deleg_status, open_status;
	int deleg_flags = 1 << NFS_DELEGATED_STATE;
	int open_flags = (1 << NFS_O_RDONLY_STATE) | (1 << NFS_O_WRONLY_STATE) | (1 << NFS_O_RDWR_STATE);

	deleg_status = nfs41_check_expired_stateid(state, &state->stateid, deleg_flags);
	open_status = nfs41_check_expired_stateid(state,  &state->open_stateid, open_flags);

	if ((deleg_status == NFS_OK) && (open_status == NFS_OK))
		return NFS_OK;
1752 1753 1754 1755
	return nfs4_open_expired(sp, state);
}
#endif

1756 1757 1758 1759 1760 1761 1762 1763 1764 1765 1766 1767 1768 1769 1770 1771
/*
 * 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;
}

L
Linus Torvalds 已提交
1772
/*
1773
 * Returns a referenced nfs4_state
L
Linus Torvalds 已提交
1774
 */
1775
static int _nfs4_do_open(struct inode *dir, struct dentry *dentry, fmode_t fmode, int flags, struct iattr *sattr, struct rpc_cred *cred, struct nfs4_state **res)
L
Linus Torvalds 已提交
1776 1777 1778 1779
{
	struct nfs4_state_owner  *sp;
	struct nfs4_state     *state = NULL;
	struct nfs_server       *server = NFS_SERVER(dir);
1780
	struct nfs4_opendata *opendata;
1781
	int status;
L
Linus Torvalds 已提交
1782 1783 1784

	/* Protect against reboot recovery conflicts */
	status = -ENOMEM;
1785 1786
	sp = nfs4_get_state_owner(server, cred, GFP_KERNEL);
	if (sp == NULL) {
L
Linus Torvalds 已提交
1787 1788 1789
		dprintk("nfs4_do_open: nfs4_get_state_owner failed!\n");
		goto out_err;
	}
1790 1791
	status = nfs4_recover_expired_lease(server);
	if (status != 0)
1792
		goto err_put_state_owner;
1793 1794
	if (dentry->d_inode != NULL)
		nfs4_return_incompatible_delegation(dentry->d_inode, fmode);
1795
	status = -ENOMEM;
1796
	opendata = nfs4_opendata_alloc(dentry, sp, fmode, flags, sattr, GFP_KERNEL);
1797
	if (opendata == NULL)
T
Trond Myklebust 已提交
1798
		goto err_put_state_owner;
L
Linus Torvalds 已提交
1799

1800 1801
	if (dentry->d_inode != NULL)
		opendata->state = nfs4_get_open_state(dentry->d_inode, sp);
1802

1803
	status = _nfs4_proc_open(opendata);
L
Linus Torvalds 已提交
1804
	if (status != 0)
1805
		goto err_opendata_put;
L
Linus Torvalds 已提交
1806

1807
	state = nfs4_opendata_to_nfs4_state(opendata);
1808 1809
	status = PTR_ERR(state);
	if (IS_ERR(state))
1810
		goto err_opendata_put;
T
Trond Myklebust 已提交
1811
	if (server->caps & NFS_CAP_POSIX_LOCK)
1812
		set_bit(NFS_STATE_POSIX_LOCKS, &state->flags);
1813 1814 1815 1816 1817 1818 1819 1820 1821 1822 1823 1824

	if (opendata->o_arg.open_flags & O_EXCL) {
		nfs4_exclusive_attrset(opendata, sattr);

		nfs_fattr_init(opendata->o_res.f_attr);
		status = nfs4_do_setattr(state->inode, cred,
				opendata->o_res.f_attr, sattr,
				state);
		if (status == 0)
			nfs_setattr_update_inode(state->inode, sattr);
		nfs_post_op_update_inode(state->inode, opendata->o_res.f_attr);
	}
1825
	nfs4_opendata_put(opendata);
L
Linus Torvalds 已提交
1826 1827 1828
	nfs4_put_state_owner(sp);
	*res = state;
	return 0;
1829 1830
err_opendata_put:
	nfs4_opendata_put(opendata);
1831 1832
err_put_state_owner:
	nfs4_put_state_owner(sp);
L
Linus Torvalds 已提交
1833 1834 1835 1836 1837 1838
out_err:
	*res = NULL;
	return status;
}


1839
static struct nfs4_state *nfs4_do_open(struct inode *dir, struct dentry *dentry, fmode_t fmode, int flags, struct iattr *sattr, struct rpc_cred *cred)
L
Linus Torvalds 已提交
1840 1841 1842 1843 1844 1845
{
	struct nfs4_exception exception = { };
	struct nfs4_state *res;
	int status;

	do {
1846
		status = _nfs4_do_open(dir, dentry, fmode, flags, sattr, cred, &res);
L
Linus Torvalds 已提交
1847 1848 1849 1850 1851 1852 1853 1854
		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
1855
		 * doing an OPEN, we assume that nfs_increment_open_seqid() will
L
Linus Torvalds 已提交
1856 1857 1858 1859 1860
		 * 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) {
1861 1862 1863
			printk(KERN_WARNING "NFS: v4 server %s "
					" returned a bad sequence-id error!\n",
					NFS_SERVER(dir)->nfs_client->cl_hostname);
L
Linus Torvalds 已提交
1864 1865 1866
			exception.retry = 1;
			continue;
		}
1867 1868 1869 1870 1871 1872 1873 1874 1875 1876
		/*
		 * 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;
		}
1877 1878 1879 1880 1881
		if (status == -EAGAIN) {
			/* We must have found a delegation */
			exception.retry = 1;
			continue;
		}
L
Linus Torvalds 已提交
1882 1883 1884 1885 1886 1887
		res = ERR_PTR(nfs4_handle_exception(NFS_SERVER(dir),
					status, &exception));
	} while (exception.retry);
	return res;
}

1888 1889 1890
static int _nfs4_do_setattr(struct inode *inode, struct rpc_cred *cred,
			    struct nfs_fattr *fattr, struct iattr *sattr,
			    struct nfs4_state *state)
L
Linus Torvalds 已提交
1891
{
1892
	struct nfs_server *server = NFS_SERVER(inode);
L
Linus Torvalds 已提交
1893
        struct nfs_setattrargs  arg = {
1894
                .fh             = NFS_FH(inode),
L
Linus Torvalds 已提交
1895 1896 1897 1898 1899 1900 1901 1902 1903
                .iap            = sattr,
		.server		= server,
		.bitmask = server->attr_bitmask,
        };
        struct nfs_setattrres  res = {
		.fattr		= fattr,
		.server		= server,
        };
        struct rpc_message msg = {
1904 1905 1906 1907
		.rpc_proc	= &nfs4_procedures[NFSPROC4_CLNT_SETATTR],
		.rpc_argp	= &arg,
		.rpc_resp	= &res,
		.rpc_cred	= cred,
L
Linus Torvalds 已提交
1908
        };
1909
	unsigned long timestamp = jiffies;
1910
	int status;
L
Linus Torvalds 已提交
1911

1912
	nfs_fattr_init(fattr);
L
Linus Torvalds 已提交
1913

1914 1915 1916
	if (nfs4_copy_delegation_stateid(&arg.stateid, inode)) {
		/* Use that stateid */
	} else if (state != NULL) {
1917
		nfs4_copy_stateid(&arg.stateid, state, current->files, current->tgid);
1918
	} else
L
Linus Torvalds 已提交
1919 1920
		memcpy(&arg.stateid, &zero_stateid, sizeof(arg.stateid));

1921
	status = nfs4_call_sync(server->client, server, &msg, &arg.seq_args, &res.seq_res, 1);
1922 1923
	if (status == 0 && state != NULL)
		renew_lease(server, timestamp);
1924
	return status;
L
Linus Torvalds 已提交
1925 1926
}

1927 1928 1929
static int nfs4_do_setattr(struct inode *inode, struct rpc_cred *cred,
			   struct nfs_fattr *fattr, struct iattr *sattr,
			   struct nfs4_state *state)
L
Linus Torvalds 已提交
1930
{
1931
	struct nfs_server *server = NFS_SERVER(inode);
L
Linus Torvalds 已提交
1932 1933 1934 1935
	struct nfs4_exception exception = { };
	int err;
	do {
		err = nfs4_handle_exception(server,
1936
				_nfs4_do_setattr(inode, cred, fattr, sattr, state),
L
Linus Torvalds 已提交
1937 1938 1939 1940 1941 1942 1943 1944 1945 1946
				&exception);
	} while (exception.retry);
	return err;
}

struct nfs4_closedata {
	struct inode *inode;
	struct nfs4_state *state;
	struct nfs_closeargs arg;
	struct nfs_closeres res;
1947
	struct nfs_fattr fattr;
1948
	unsigned long timestamp;
F
Fred Isaman 已提交
1949 1950
	bool roc;
	u32 roc_barrier;
L
Linus Torvalds 已提交
1951 1952
};

1953
static void nfs4_free_closedata(void *data)
1954
{
1955 1956
	struct nfs4_closedata *calldata = data;
	struct nfs4_state_owner *sp = calldata->state->owner;
1957
	struct super_block *sb = calldata->state->inode->i_sb;
1958

F
Fred Isaman 已提交
1959 1960
	if (calldata->roc)
		pnfs_roc_release(calldata->state->inode);
1961 1962 1963
	nfs4_put_open_state(calldata->state);
	nfs_free_seqid(calldata->arg.seqid);
	nfs4_put_state_owner(sp);
1964
	nfs_sb_deactive(sb);
1965 1966 1967
	kfree(calldata);
}

1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979
static void nfs4_close_clear_stateid_flags(struct nfs4_state *state,
		fmode_t fmode)
{
	spin_lock(&state->owner->so_lock);
	if (!(fmode & FMODE_READ))
		clear_bit(NFS_O_RDONLY_STATE, &state->flags);
	if (!(fmode & FMODE_WRITE))
		clear_bit(NFS_O_WRONLY_STATE, &state->flags);
	clear_bit(NFS_O_RDWR_STATE, &state->flags);
	spin_unlock(&state->owner->so_lock);
}

1980
static void nfs4_close_done(struct rpc_task *task, void *data)
L
Linus Torvalds 已提交
1981
{
1982
	struct nfs4_closedata *calldata = data;
L
Linus Torvalds 已提交
1983 1984 1985
	struct nfs4_state *state = calldata->state;
	struct nfs_server *server = NFS_SERVER(calldata->inode);

1986 1987
	if (!nfs4_sequence_done(task, &calldata->res.seq_res))
		return;
L
Linus Torvalds 已提交
1988 1989 1990 1991 1992
        /* 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
1993 1994 1995
			if (calldata->roc)
				pnfs_roc_set_barrier(state->inode,
						     calldata->roc_barrier);
1996
			nfs_set_open_stateid(state, &calldata->res.stateid, 0);
1997
			renew_lease(server, calldata->timestamp);
1998 1999
			nfs4_close_clear_stateid_flags(state,
					calldata->arg.fmode);
L
Linus Torvalds 已提交
2000 2001
			break;
		case -NFS4ERR_STALE_STATEID:
2002 2003
		case -NFS4ERR_OLD_STATEID:
		case -NFS4ERR_BAD_STATEID:
L
Linus Torvalds 已提交
2004
		case -NFS4ERR_EXPIRED:
2005
			if (calldata->arg.fmode == 0)
2006
				break;
L
Linus Torvalds 已提交
2007
		default:
2008 2009
			if (nfs4_async_handle_error(task, server, state) == -EAGAIN)
				rpc_restart_call_prepare(task);
L
Linus Torvalds 已提交
2010
	}
2011
	nfs_release_seqid(calldata->arg.seqid);
2012
	nfs_refresh_inode(calldata->inode, calldata->res.fattr);
L
Linus Torvalds 已提交
2013 2014
}

T
Trond Myklebust 已提交
2015
static void nfs4_close_prepare(struct rpc_task *task, void *data)
L
Linus Torvalds 已提交
2016
{
T
Trond Myklebust 已提交
2017
	struct nfs4_closedata *calldata = data;
2018
	struct nfs4_state *state = calldata->state;
2019
	int call_close = 0;
2020

2021
	if (nfs_wait_on_sequence(calldata->arg.seqid, task) != 0)
2022
		return;
2023

2024 2025
	task->tk_msg.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_OPEN_DOWNGRADE];
	calldata->arg.fmode = FMODE_READ|FMODE_WRITE;
2026
	spin_lock(&state->owner->so_lock);
2027
	/* Calculate the change in open mode */
2028
	if (state->n_rdwr == 0) {
2029
		if (state->n_rdonly == 0) {
2030 2031 2032
			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;
2033 2034
		}
		if (state->n_wronly == 0) {
2035 2036 2037
			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;
2038
		}
2039
	}
2040
	spin_unlock(&state->owner->so_lock);
2041 2042

	if (!call_close) {
2043 2044
		/* Note: exit _without_ calling nfs4_close_done */
		task->tk_action = NULL;
2045 2046
		return;
	}
2047

F
Fred Isaman 已提交
2048
	if (calldata->arg.fmode == 0) {
2049
		task->tk_msg.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_CLOSE];
F
Fred Isaman 已提交
2050 2051 2052 2053 2054 2055 2056
		if (calldata->roc &&
		    pnfs_roc_drain(calldata->inode, &calldata->roc_barrier)) {
			rpc_sleep_on(&NFS_SERVER(calldata->inode)->roc_rpcwaitq,
				     task, NULL);
			return;
		}
	}
2057

2058
	nfs_fattr_init(calldata->res.fattr);
2059
	calldata->timestamp = jiffies;
2060
	if (nfs4_setup_sequence(NFS_SERVER(calldata->inode),
2061 2062 2063
				&calldata->arg.seq_args,
				&calldata->res.seq_res,
				task))
2064
		return;
2065
	rpc_call_start(task);
L
Linus Torvalds 已提交
2066 2067
}

2068
static const struct rpc_call_ops nfs4_close_ops = {
T
Trond Myklebust 已提交
2069
	.rpc_call_prepare = nfs4_close_prepare,
2070 2071 2072 2073
	.rpc_call_done = nfs4_close_done,
	.rpc_release = nfs4_free_closedata,
};

L
Linus Torvalds 已提交
2074 2075 2076 2077 2078 2079 2080 2081 2082 2083 2084
/* 
 * 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!
 */
2085
int nfs4_do_close(struct nfs4_state *state, gfp_t gfp_mask, int wait, bool roc)
L
Linus Torvalds 已提交
2086
{
2087
	struct nfs_server *server = NFS_SERVER(state->inode);
L
Linus Torvalds 已提交
2088
	struct nfs4_closedata *calldata;
2089 2090
	struct nfs4_state_owner *sp = state->owner;
	struct rpc_task *task;
2091 2092 2093 2094
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_CLOSE],
		.rpc_cred = state->owner->so_cred,
	};
T
Trond Myklebust 已提交
2095 2096
	struct rpc_task_setup task_setup_data = {
		.rpc_client = server->client,
2097
		.rpc_message = &msg,
T
Trond Myklebust 已提交
2098
		.callback_ops = &nfs4_close_ops,
2099
		.workqueue = nfsiod_workqueue,
T
Trond Myklebust 已提交
2100 2101
		.flags = RPC_TASK_ASYNC,
	};
2102
	int status = -ENOMEM;
L
Linus Torvalds 已提交
2103

2104
	calldata = kzalloc(sizeof(*calldata), gfp_mask);
L
Linus Torvalds 已提交
2105
	if (calldata == NULL)
2106
		goto out;
2107
	nfs41_init_sequence(&calldata->arg.seq_args, &calldata->res.seq_res, 1);
2108
	calldata->inode = state->inode;
L
Linus Torvalds 已提交
2109
	calldata->state = state;
2110
	calldata->arg.fh = NFS_FH(state->inode);
2111
	calldata->arg.stateid = &state->open_stateid;
L
Linus Torvalds 已提交
2112
	/* Serialization for the sequence id */
2113
	calldata->arg.seqid = nfs_alloc_seqid(&state->owner->so_seqid, gfp_mask);
2114 2115
	if (calldata->arg.seqid == NULL)
		goto out_free_calldata;
2116
	calldata->arg.fmode = 0;
2117
	calldata->arg.bitmask = server->cache_consistency_bitmask;
2118
	calldata->res.fattr = &calldata->fattr;
2119
	calldata->res.seqid = calldata->arg.seqid;
2120
	calldata->res.server = server;
F
Fred Isaman 已提交
2121
	calldata->roc = roc;
2122
	nfs_sb_active(calldata->inode->i_sb);
2123

2124 2125
	msg.rpc_argp = &calldata->arg;
	msg.rpc_resp = &calldata->res;
T
Trond Myklebust 已提交
2126 2127
	task_setup_data.callback_data = calldata;
	task = rpc_run_task(&task_setup_data);
2128 2129
	if (IS_ERR(task))
		return PTR_ERR(task);
2130 2131 2132
	status = 0;
	if (wait)
		status = rpc_wait_for_completion_task(task);
2133
	rpc_put_task(task);
2134
	return status;
2135 2136 2137
out_free_calldata:
	kfree(calldata);
out:
F
Fred Isaman 已提交
2138 2139
	if (roc)
		pnfs_roc_release(state->inode);
2140 2141
	nfs4_put_open_state(state);
	nfs4_put_state_owner(sp);
2142
	return status;
L
Linus Torvalds 已提交
2143 2144
}

2145
static struct inode *
2146
nfs4_atomic_open(struct inode *dir, struct nfs_open_context *ctx, int open_flags, struct iattr *attr)
L
Linus Torvalds 已提交
2147 2148 2149
{
	struct nfs4_state *state;

T
Trond Myklebust 已提交
2150
	/* Protect against concurrent sillydeletes */
2151
	state = nfs4_do_open(dir, ctx->dentry, ctx->mode, open_flags, attr, ctx->cred);
2152 2153
	if (IS_ERR(state))
		return ERR_CAST(state);
2154
	ctx->state = state;
2155
	return igrab(state->inode);
L
Linus Torvalds 已提交
2156 2157
}

2158
static void nfs4_close_context(struct nfs_open_context *ctx, int is_sync)
T
Trond Myklebust 已提交
2159 2160 2161 2162
{
	if (ctx->state == NULL)
		return;
	if (is_sync)
2163
		nfs4_close_sync(ctx->state, ctx->mode);
T
Trond Myklebust 已提交
2164
	else
2165
		nfs4_close_state(ctx->state, ctx->mode);
T
Trond Myklebust 已提交
2166
}
L
Linus Torvalds 已提交
2167 2168 2169

static int _nfs4_server_capabilities(struct nfs_server *server, struct nfs_fh *fhandle)
{
B
Benny Halevy 已提交
2170 2171 2172
	struct nfs4_server_caps_arg args = {
		.fhandle = fhandle,
	};
L
Linus Torvalds 已提交
2173 2174 2175
	struct nfs4_server_caps_res res = {};
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_SERVER_CAPS],
B
Benny Halevy 已提交
2176
		.rpc_argp = &args,
L
Linus Torvalds 已提交
2177 2178 2179 2180
		.rpc_resp = &res,
	};
	int status;

2181
	status = nfs4_call_sync(server->client, server, &msg, &args.seq_args, &res.seq_res, 0);
L
Linus Torvalds 已提交
2182 2183
	if (status == 0) {
		memcpy(server->attr_bitmask, res.attr_bitmask, sizeof(server->attr_bitmask));
2184 2185 2186 2187 2188
		server->caps &= ~(NFS_CAP_ACLS|NFS_CAP_HARDLINKS|
				NFS_CAP_SYMLINKS|NFS_CAP_FILEID|
				NFS_CAP_MODE|NFS_CAP_NLINK|NFS_CAP_OWNER|
				NFS_CAP_OWNER_GROUP|NFS_CAP_ATIME|
				NFS_CAP_CTIME|NFS_CAP_MTIME);
L
Linus Torvalds 已提交
2189 2190 2191 2192 2193 2194
		if (res.attr_bitmask[0] & FATTR4_WORD0_ACL)
			server->caps |= NFS_CAP_ACLS;
		if (res.has_links != 0)
			server->caps |= NFS_CAP_HARDLINKS;
		if (res.has_symlinks != 0)
			server->caps |= NFS_CAP_SYMLINKS;
2195 2196 2197 2198 2199 2200 2201 2202 2203 2204 2205 2206 2207 2208 2209 2210 2211
		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;

2212 2213 2214
		memcpy(server->cache_consistency_bitmask, res.attr_bitmask, sizeof(server->cache_consistency_bitmask));
		server->cache_consistency_bitmask[0] &= FATTR4_WORD0_CHANGE|FATTR4_WORD0_SIZE;
		server->cache_consistency_bitmask[1] &= FATTR4_WORD1_TIME_METADATA|FATTR4_WORD1_TIME_MODIFY;
L
Linus Torvalds 已提交
2215 2216
		server->acl_bitmask = res.acl_bitmask;
	}
A
Andy Adamson 已提交
2217

L
Linus Torvalds 已提交
2218 2219 2220
	return status;
}

2221
int nfs4_server_capabilities(struct nfs_server *server, struct nfs_fh *fhandle)
L
Linus Torvalds 已提交
2222 2223 2224 2225 2226 2227 2228 2229 2230 2231 2232 2233 2234 2235 2236 2237 2238 2239 2240
{
	struct nfs4_exception exception = { };
	int err;
	do {
		err = nfs4_handle_exception(server,
				_nfs4_server_capabilities(server, fhandle),
				&exception);
	} while (exception.retry);
	return err;
}

static int _nfs4_lookup_root(struct nfs_server *server, struct nfs_fh *fhandle,
		struct nfs_fsinfo *info)
{
	struct nfs4_lookup_root_arg args = {
		.bitmask = nfs4_fattr_bitmap,
	};
	struct nfs4_lookup_res res = {
		.server = server,
2241
		.fattr = info->fattr,
L
Linus Torvalds 已提交
2242 2243 2244 2245 2246 2247 2248
		.fh = fhandle,
	};
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_LOOKUP_ROOT],
		.rpc_argp = &args,
		.rpc_resp = &res,
	};
2249

2250
	nfs_fattr_init(info->fattr);
2251
	return nfs4_call_sync(server->client, server, &msg, &args.seq_args, &res.seq_res, 0);
L
Linus Torvalds 已提交
2252 2253 2254 2255 2256 2257 2258 2259
}

static int nfs4_lookup_root(struct nfs_server *server, struct nfs_fh *fhandle,
		struct nfs_fsinfo *info)
{
	struct nfs4_exception exception = { };
	int err;
	do {
2260 2261 2262 2263 2264 2265 2266 2267
		err = _nfs4_lookup_root(server, fhandle, info);
		switch (err) {
		case 0:
		case -NFS4ERR_WRONGSEC:
			break;
		default:
			err = nfs4_handle_exception(server, err, &exception);
		}
L
Linus Torvalds 已提交
2268 2269 2270 2271
	} while (exception.retry);
	return err;
}

2272 2273 2274 2275 2276 2277 2278 2279 2280 2281 2282 2283 2284 2285 2286 2287
static int nfs4_lookup_root_sec(struct nfs_server *server, struct nfs_fh *fhandle,
				struct nfs_fsinfo *info, rpc_authflavor_t flavor)
{
	struct rpc_auth *auth;
	int ret;

	auth = rpcauth_create(flavor, server->client);
	if (!auth) {
		ret = -EIO;
		goto out;
	}
	ret = nfs4_lookup_root(server, fhandle, info);
out:
	return ret;
}

2288
static int nfs4_find_root_sec(struct nfs_server *server, struct nfs_fh *fhandle,
2289
			      struct nfs_fsinfo *info)
L
Linus Torvalds 已提交
2290
{
2291
	int i, len, status = 0;
2292
	rpc_authflavor_t flav_array[NFS_MAX_SECFLAVORS];
L
Linus Torvalds 已提交
2293

2294 2295 2296
	len = gss_mech_list_pseudoflavors(&flav_array[0]);
	flav_array[len] = RPC_AUTH_NULL;
	len += 1;
2297 2298 2299

	for (i = 0; i < len; i++) {
		status = nfs4_lookup_root_sec(server, fhandle, info, flav_array[i]);
2300
		if (status == -NFS4ERR_WRONGSEC || status == -EACCES)
2301 2302
			continue;
		break;
2303
	}
2304 2305 2306 2307 2308 2309 2310 2311 2312
	/*
	 * -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;
2313 2314 2315 2316 2317 2318 2319 2320 2321
	return status;
}

/*
 * get the file handle for the "/" directory on the server
 */
static int nfs4_proc_get_root(struct nfs_server *server, struct nfs_fh *fhandle,
			      struct nfs_fsinfo *info)
{
2322
	int minor_version = server->nfs_client->cl_minorversion;
2323
	int status = nfs4_lookup_root(server, fhandle, info);
2324 2325 2326 2327 2328
	if ((status == -NFS4ERR_WRONGSEC) && !(server->flags & NFS_MOUNT_SECFLAVOUR))
		/*
		 * A status of -NFS4ERR_WRONGSEC will be mapped to -EPERM
		 * by nfs4_map_errors() as this function exits.
		 */
2329
		status = nfs_v4_minor_ops[minor_version]->find_root_sec(server, fhandle, info);
L
Linus Torvalds 已提交
2330 2331 2332 2333
	if (status == 0)
		status = nfs4_server_capabilities(server, fhandle);
	if (status == 0)
		status = nfs4_do_fsinfo(server, fhandle, info);
2334
	return nfs4_map_errors(status);
L
Linus Torvalds 已提交
2335 2336
}

2337
static void nfs_fixup_referral_attributes(struct nfs_fattr *fattr);
M
Manoj Naik 已提交
2338 2339 2340 2341 2342
/*
 * 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
 */
2343 2344
static int nfs4_get_referral(struct inode *dir, const struct qstr *name,
			     struct nfs_fattr *fattr, struct nfs_fh *fhandle)
M
Manoj Naik 已提交
2345 2346 2347 2348 2349 2350 2351 2352 2353 2354 2355 2356
{
	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;

2357
	status = nfs4_proc_fs_locations(dir, name, locations, page);
M
Manoj Naik 已提交
2358 2359 2360 2361
	if (status != 0)
		goto out;
	/* Make sure server returned a different fsid for the referral */
	if (nfs_fsid_equal(&NFS_SERVER(dir)->fsid, &locations->fattr.fsid)) {
2362 2363
		dprintk("%s: server did not return a different fsid for"
			" a referral at %s\n", __func__, name->name);
M
Manoj Naik 已提交
2364 2365 2366
		status = -EIO;
		goto out;
	}
2367 2368
	/* Fixup attributes for the nfs_lookup() call to nfs_fhget() */
	nfs_fixup_referral_attributes(&locations->fattr);
M
Manoj Naik 已提交
2369

2370
	/* replace the lookup nfs_fattr with the locations nfs_fattr */
M
Manoj Naik 已提交
2371 2372 2373 2374 2375
	memcpy(fattr, &locations->fattr, sizeof(struct nfs_fattr));
	memset(fhandle, 0, sizeof(struct nfs_fh));
out:
	if (page)
		__free_page(page);
2376
	kfree(locations);
M
Manoj Naik 已提交
2377 2378 2379
	return status;
}

L
Linus Torvalds 已提交
2380 2381 2382 2383 2384 2385 2386 2387 2388 2389 2390 2391 2392 2393 2394 2395
static int _nfs4_proc_getattr(struct nfs_server *server, struct nfs_fh *fhandle, struct nfs_fattr *fattr)
{
	struct nfs4_getattr_arg args = {
		.fh = fhandle,
		.bitmask = server->attr_bitmask,
	};
	struct nfs4_getattr_res res = {
		.fattr = fattr,
		.server = server,
	};
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_GETATTR],
		.rpc_argp = &args,
		.rpc_resp = &res,
	};
	
2396
	nfs_fattr_init(fattr);
2397
	return nfs4_call_sync(server->client, server, &msg, &args.seq_args, &res.seq_res, 0);
L
Linus Torvalds 已提交
2398 2399 2400 2401 2402 2403 2404 2405 2406 2407 2408 2409 2410 2411 2412 2413 2414 2415 2416 2417 2418 2419 2420 2421 2422 2423 2424 2425 2426 2427 2428 2429 2430 2431 2432
}

static int nfs4_proc_getattr(struct nfs_server *server, struct nfs_fh *fhandle, struct nfs_fattr *fattr)
{
	struct nfs4_exception exception = { };
	int err;
	do {
		err = nfs4_handle_exception(server,
				_nfs4_proc_getattr(server, fhandle, fattr),
				&exception);
	} while (exception.retry);
	return err;
}

/* 
 * The file is not closed if it is opened due to the a request to change
 * the size of the file. The open call will not be needed once the
 * VFS layer lookup-intents are implemented.
 *
 * Close is called when the inode is destroyed.
 * If we haven't opened the file for O_WRONLY, we
 * need to in the size_change case to obtain a stateid.
 *
 * Got race?
 * Because OPEN is always done by name in nfsv4, it is
 * possible that we opened a different file by the same
 * name.  We can recognize this race condition, but we
 * can't do anything about it besides returning an error.
 *
 * This will be fixed with VFS changes (lookup-intent).
 */
static int
nfs4_proc_setattr(struct dentry *dentry, struct nfs_fattr *fattr,
		  struct iattr *sattr)
{
2433
	struct inode *inode = dentry->d_inode;
2434
	struct rpc_cred *cred = NULL;
2435
	struct nfs4_state *state = NULL;
L
Linus Torvalds 已提交
2436 2437
	int status;

B
Benny Halevy 已提交
2438 2439 2440
	if (pnfs_ld_layoutret_on_setattr(inode))
		pnfs_return_layout(inode);

2441
	nfs_fattr_init(fattr);
L
Linus Torvalds 已提交
2442
	
2443
	/* Search for an existing open(O_WRITE) file */
2444 2445 2446 2447
	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
2448 2449 2450 2451
		if (ctx) {
			cred = ctx->cred;
			state = ctx->state;
		}
2452
	}
2453

2454 2455 2456 2457
	/* Deal with open(O_TRUNC) */
	if (sattr->ia_valid & ATTR_OPEN)
		sattr->ia_valid &= ~(ATTR_MTIME|ATTR_CTIME|ATTR_OPEN);

2458
	status = nfs4_do_setattr(inode, cred, fattr, sattr, state);
2459 2460
	if (status == 0)
		nfs_setattr_update_inode(inode, sattr);
L
Linus Torvalds 已提交
2461 2462 2463
	return status;
}

2464 2465 2466
static int _nfs4_proc_lookup(struct rpc_clnt *clnt, struct inode *dir,
		const struct qstr *name, struct nfs_fh *fhandle,
		struct nfs_fattr *fattr)
D
David Howells 已提交
2467
{
2468
	struct nfs_server *server = NFS_SERVER(dir);
D
David Howells 已提交
2469 2470 2471
	int		       status;
	struct nfs4_lookup_arg args = {
		.bitmask = server->attr_bitmask,
2472
		.dir_fh = NFS_FH(dir),
D
David Howells 已提交
2473 2474 2475 2476 2477 2478 2479 2480 2481 2482 2483 2484 2485 2486 2487
		.name = name,
	};
	struct nfs4_lookup_res res = {
		.server = server,
		.fattr = fattr,
		.fh = fhandle,
	};
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_LOOKUP],
		.rpc_argp = &args,
		.rpc_resp = &res,
	};

	nfs_fattr_init(fattr);

L
Linus Torvalds 已提交
2488
	dprintk("NFS call  lookup %s\n", name->name);
2489
	status = nfs4_call_sync(clnt, server, &msg, &args.seq_args, &res.seq_res, 0);
L
Linus Torvalds 已提交
2490 2491 2492 2493
	dprintk("NFS reply lookup: %d\n", status);
	return status;
}

2494 2495 2496 2497 2498 2499 2500 2501 2502 2503
void nfs_fixup_secinfo_attributes(struct nfs_fattr *fattr, struct nfs_fh *fh)
{
	memset(fh, 0, sizeof(struct nfs_fh));
	fattr->fsid.major = 1;
	fattr->valid |= NFS_ATTR_FATTR_TYPE | NFS_ATTR_FATTR_MODE |
		NFS_ATTR_FATTR_NLINK | NFS_ATTR_FATTR_FSID | NFS_ATTR_FATTR_MOUNTPOINT;
	fattr->mode = S_IFDIR | S_IRUGO | S_IXUGO;
	fattr->nlink = 2;
}

2504 2505
static int nfs4_proc_lookup(struct rpc_clnt *clnt, struct inode *dir, struct qstr *name,
			    struct nfs_fh *fhandle, struct nfs_fattr *fattr)
L
Linus Torvalds 已提交
2506 2507 2508 2509
{
	struct nfs4_exception exception = { };
	int err;
	do {
2510 2511 2512 2513
		int status;

		status = _nfs4_proc_lookup(clnt, dir, name, fhandle, fattr);
		switch (status) {
2514 2515
		case -NFS4ERR_BADNAME:
			return -ENOENT;
2516
		case -NFS4ERR_MOVED:
2517
			return nfs4_get_referral(dir, name, fattr, fhandle);
2518
		case -NFS4ERR_WRONGSEC:
2519
			nfs_fixup_secinfo_attributes(fattr, fhandle);
2520 2521 2522
		}
		err = nfs4_handle_exception(NFS_SERVER(dir),
				status, &exception);
L
Linus Torvalds 已提交
2523 2524 2525 2526 2527 2528
	} while (exception.retry);
	return err;
}

static int _nfs4_proc_access(struct inode *inode, struct nfs_access_entry *entry)
{
2529
	struct nfs_server *server = NFS_SERVER(inode);
L
Linus Torvalds 已提交
2530 2531
	struct nfs4_accessargs args = {
		.fh = NFS_FH(inode),
2532
		.bitmask = server->cache_consistency_bitmask,
2533 2534 2535
	};
	struct nfs4_accessres res = {
		.server = server,
L
Linus Torvalds 已提交
2536 2537 2538 2539 2540 2541 2542 2543 2544 2545 2546 2547 2548 2549 2550 2551 2552 2553 2554 2555 2556 2557 2558 2559 2560 2561
	};
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_ACCESS],
		.rpc_argp = &args,
		.rpc_resp = &res,
		.rpc_cred = entry->cred,
	};
	int mode = entry->mask;
	int status;

	/*
	 * Determine which access bits we want to ask for...
	 */
	if (mode & MAY_READ)
		args.access |= NFS4_ACCESS_READ;
	if (S_ISDIR(inode->i_mode)) {
		if (mode & MAY_WRITE)
			args.access |= NFS4_ACCESS_MODIFY | NFS4_ACCESS_EXTEND | NFS4_ACCESS_DELETE;
		if (mode & MAY_EXEC)
			args.access |= NFS4_ACCESS_LOOKUP;
	} else {
		if (mode & MAY_WRITE)
			args.access |= NFS4_ACCESS_MODIFY | NFS4_ACCESS_EXTEND;
		if (mode & MAY_EXEC)
			args.access |= NFS4_ACCESS_EXECUTE;
	}
2562 2563 2564 2565 2566

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

2567
	status = nfs4_call_sync(server->client, server, &msg, &args.seq_args, &res.seq_res, 0);
L
Linus Torvalds 已提交
2568 2569 2570 2571 2572 2573 2574 2575
	if (!status) {
		entry->mask = 0;
		if (res.access & NFS4_ACCESS_READ)
			entry->mask |= MAY_READ;
		if (res.access & (NFS4_ACCESS_MODIFY | NFS4_ACCESS_EXTEND | NFS4_ACCESS_DELETE))
			entry->mask |= MAY_WRITE;
		if (res.access & (NFS4_ACCESS_LOOKUP|NFS4_ACCESS_EXECUTE))
			entry->mask |= MAY_EXEC;
2576
		nfs_refresh_inode(inode, res.fattr);
L
Linus Torvalds 已提交
2577
	}
2578
	nfs_free_fattr(res.fattr);
L
Linus Torvalds 已提交
2579 2580 2581 2582 2583 2584 2585 2586 2587 2588 2589 2590 2591 2592 2593 2594 2595 2596 2597 2598 2599 2600 2601 2602 2603 2604 2605 2606 2607 2608 2609 2610 2611 2612 2613 2614 2615 2616 2617 2618 2619 2620 2621 2622 2623 2624 2625 2626
	return status;
}

static int nfs4_proc_access(struct inode *inode, struct nfs_access_entry *entry)
{
	struct nfs4_exception exception = { };
	int err;
	do {
		err = nfs4_handle_exception(NFS_SERVER(inode),
				_nfs4_proc_access(inode, entry),
				&exception);
	} while (exception.retry);
	return err;
}

/*
 * TODO: For the time being, we don't try to get any attributes
 * along with any of the zero-copy operations READ, READDIR,
 * READLINK, WRITE.
 *
 * In the case of the first three, we want to put the GETATTR
 * after the read-type operation -- this is because it is hard
 * to predict the length of a GETATTR response in v4, and thus
 * align the READ data correctly.  This means that the GETATTR
 * may end up partially falling into the page cache, and we should
 * shift it into the 'tail' of the xdr_buf before processing.
 * To do this efficiently, we need to know the total length
 * of data received, which doesn't seem to be available outside
 * of the RPC layer.
 *
 * In the case of WRITE, we also want to put the GETATTR after
 * the operation -- in this case because we want to make sure
 * we get the post-operation mtime and size.  This means that
 * we can't use xdr_encode_pages() as written: we need a variant
 * of it which would leave room in the 'tail' iovec.
 *
 * 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
2627
	struct nfs4_readlink_res res;
L
Linus Torvalds 已提交
2628 2629 2630
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_READLINK],
		.rpc_argp = &args,
B
Benny Halevy 已提交
2631
		.rpc_resp = &res,
L
Linus Torvalds 已提交
2632 2633
	};

2634
	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
2635 2636 2637 2638 2639 2640 2641 2642 2643 2644 2645 2646 2647 2648 2649 2650 2651 2652 2653 2654 2655 2656 2657 2658 2659 2660 2661 2662 2663 2664 2665 2666
}

static int nfs4_proc_readlink(struct inode *inode, struct page *page,
		unsigned int pgbase, unsigned int pglen)
{
	struct nfs4_exception exception = { };
	int err;
	do {
		err = nfs4_handle_exception(NFS_SERVER(inode),
				_nfs4_proc_readlink(inode, page, pgbase, pglen),
				&exception);
	} while (exception.retry);
	return err;
}

/*
 * Got race?
 * We will need to arrange for the VFS layer to provide an atomic open.
 * Until then, this create/open method is prone to inefficiency and race
 * conditions due to the lookup, create, and open VFS calls from sys_open()
 * placed on the wire.
 *
 * Given the above sorry state of affairs, I'm simply sending an OPEN.
 * The file will be opened again in the subsequent VFS open call
 * (nfs4_proc_file_open).
 *
 * The open for read will just hang around to be used by any process that
 * opens the file O_RDONLY. This will all be resolved with the VFS changes.
 */

static int
nfs4_proc_create(struct inode *dir, struct dentry *dentry, struct iattr *sattr,
2667
                 int flags, struct nfs_open_context *ctx)
L
Linus Torvalds 已提交
2668
{
2669
	struct dentry *de = dentry;
L
Linus Torvalds 已提交
2670
	struct nfs4_state *state;
2671 2672
	struct rpc_cred *cred = NULL;
	fmode_t fmode = 0;
L
Linus Torvalds 已提交
2673 2674
	int status = 0;

2675 2676
	if (ctx != NULL) {
		cred = ctx->cred;
2677
		de = ctx->dentry;
2678
		fmode = ctx->mode;
L
Linus Torvalds 已提交
2679
	}
A
Aneesh Kumar K.V 已提交
2680
	sattr->ia_mode &= ~current_umask();
2681
	state = nfs4_do_open(dir, de, fmode, flags, sattr, cred);
2682
	d_drop(dentry);
L
Linus Torvalds 已提交
2683 2684
	if (IS_ERR(state)) {
		status = PTR_ERR(state);
2685
		goto out;
L
Linus Torvalds 已提交
2686
	}
2687
	d_add(dentry, igrab(state->inode));
2688
	nfs_set_verifier(dentry, nfs_save_change_attribute(dir));
2689 2690
	if (ctx != NULL)
		ctx->state = state;
2691
	else
2692
		nfs4_close_sync(state, fmode);
L
Linus Torvalds 已提交
2693 2694 2695 2696 2697 2698
out:
	return status;
}

static int _nfs4_proc_remove(struct inode *dir, struct qstr *name)
{
2699
	struct nfs_server *server = NFS_SERVER(dir);
2700
	struct nfs_removeargs args = {
L
Linus Torvalds 已提交
2701
		.fh = NFS_FH(dir),
2702 2703
		.name.len = name->len,
		.name.name = name->name,
2704 2705
		.bitmask = server->attr_bitmask,
	};
2706
	struct nfs_removeres res = {
2707
		.server = server,
L
Linus Torvalds 已提交
2708 2709
	};
	struct rpc_message msg = {
2710 2711 2712
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_REMOVE],
		.rpc_argp = &args,
		.rpc_resp = &res,
L
Linus Torvalds 已提交
2713
	};
2714 2715 2716 2717 2718
	int status = -ENOMEM;

	res.dir_attr = nfs_alloc_fattr();
	if (res.dir_attr == NULL)
		goto out;
L
Linus Torvalds 已提交
2719

2720
	status = nfs4_call_sync(server->client, server, &msg, &args.seq_args, &res.seq_res, 1);
2721 2722
	if (status == 0) {
		update_changeattr(dir, &res.cinfo);
2723
		nfs_post_op_update_inode(dir, res.dir_attr);
2724
	}
2725 2726
	nfs_free_fattr(res.dir_attr);
out:
L
Linus Torvalds 已提交
2727 2728 2729 2730 2731 2732 2733 2734 2735 2736 2737 2738 2739 2740 2741
	return status;
}

static int nfs4_proc_remove(struct inode *dir, struct qstr *name)
{
	struct nfs4_exception exception = { };
	int err;
	do {
		err = nfs4_handle_exception(NFS_SERVER(dir),
				_nfs4_proc_remove(dir, name),
				&exception);
	} while (exception.retry);
	return err;
}

2742
static void nfs4_proc_unlink_setup(struct rpc_message *msg, struct inode *dir)
L
Linus Torvalds 已提交
2743
{
2744 2745 2746
	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
2747

2748
	args->bitmask = server->cache_consistency_bitmask;
2749
	res->server = server;
L
Linus Torvalds 已提交
2750
	msg->rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_REMOVE];
2751
	nfs41_init_sequence(&args->seq_args, &res->seq_res, 1);
L
Linus Torvalds 已提交
2752 2753
}

2754
static int nfs4_proc_unlink_done(struct rpc_task *task, struct inode *dir)
L
Linus Torvalds 已提交
2755
{
2756 2757
	struct nfs_removeres *res = task->tk_msg.rpc_resp;

2758 2759
	if (!nfs4_sequence_done(task, &res->seq_res))
		return 0;
2760
	if (nfs4_async_handle_error(task, res->server, NULL) == -EAGAIN)
2761 2762
		return 0;
	update_changeattr(dir, &res->cinfo);
2763
	nfs_post_op_update_inode(dir, res->dir_attr);
2764
	return 1;
L
Linus Torvalds 已提交
2765 2766
}

2767 2768 2769 2770 2771 2772 2773 2774 2775
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];
	arg->bitmask = server->attr_bitmask;
	res->server = server;
2776
	nfs41_init_sequence(&arg->seq_args, &res->seq_res, 1);
2777 2778 2779 2780 2781 2782 2783 2784 2785 2786 2787 2788 2789 2790 2791 2792 2793 2794 2795
}

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

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

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

L
Linus Torvalds 已提交
2796 2797 2798
static int _nfs4_proc_rename(struct inode *old_dir, struct qstr *old_name,
		struct inode *new_dir, struct qstr *new_name)
{
2799
	struct nfs_server *server = NFS_SERVER(old_dir);
2800
	struct nfs_renameargs arg = {
L
Linus Torvalds 已提交
2801 2802 2803 2804
		.old_dir = NFS_FH(old_dir),
		.new_dir = NFS_FH(new_dir),
		.old_name = old_name,
		.new_name = new_name,
2805 2806
		.bitmask = server->attr_bitmask,
	};
2807
	struct nfs_renameres res = {
2808
		.server = server,
L
Linus Torvalds 已提交
2809 2810 2811 2812 2813 2814
	};
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_RENAME],
		.rpc_argp = &arg,
		.rpc_resp = &res,
	};
2815
	int status = -ENOMEM;
L
Linus Torvalds 已提交
2816
	
2817 2818 2819 2820
	res.old_fattr = nfs_alloc_fattr();
	res.new_fattr = nfs_alloc_fattr();
	if (res.old_fattr == NULL || res.new_fattr == NULL)
		goto out;
L
Linus Torvalds 已提交
2821

2822
	status = nfs4_call_sync(server->client, server, &msg, &arg.seq_args, &res.seq_res, 1);
L
Linus Torvalds 已提交
2823 2824
	if (!status) {
		update_changeattr(old_dir, &res.old_cinfo);
2825
		nfs_post_op_update_inode(old_dir, res.old_fattr);
L
Linus Torvalds 已提交
2826
		update_changeattr(new_dir, &res.new_cinfo);
2827
		nfs_post_op_update_inode(new_dir, res.new_fattr);
L
Linus Torvalds 已提交
2828
	}
2829 2830 2831
out:
	nfs_free_fattr(res.new_fattr);
	nfs_free_fattr(res.old_fattr);
L
Linus Torvalds 已提交
2832 2833 2834 2835 2836 2837 2838 2839 2840 2841 2842 2843 2844 2845 2846 2847 2848 2849 2850
	return status;
}

static int nfs4_proc_rename(struct inode *old_dir, struct qstr *old_name,
		struct inode *new_dir, struct qstr *new_name)
{
	struct nfs4_exception exception = { };
	int err;
	do {
		err = nfs4_handle_exception(NFS_SERVER(old_dir),
				_nfs4_proc_rename(old_dir, old_name,
					new_dir, new_name),
				&exception);
	} while (exception.retry);
	return err;
}

static int _nfs4_proc_link(struct inode *inode, struct inode *dir, struct qstr *name)
{
2851
	struct nfs_server *server = NFS_SERVER(inode);
L
Linus Torvalds 已提交
2852 2853 2854 2855
	struct nfs4_link_arg arg = {
		.fh     = NFS_FH(inode),
		.dir_fh = NFS_FH(dir),
		.name   = name,
2856 2857 2858 2859
		.bitmask = server->attr_bitmask,
	};
	struct nfs4_link_res res = {
		.server = server,
L
Linus Torvalds 已提交
2860 2861 2862 2863
	};
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_LINK],
		.rpc_argp = &arg,
2864
		.rpc_resp = &res,
L
Linus Torvalds 已提交
2865
	};
2866 2867 2868 2869 2870 2871
	int status = -ENOMEM;

	res.fattr = nfs_alloc_fattr();
	res.dir_attr = nfs_alloc_fattr();
	if (res.fattr == NULL || res.dir_attr == NULL)
		goto out;
L
Linus Torvalds 已提交
2872

2873
	status = nfs4_call_sync(server->client, server, &msg, &arg.seq_args, &res.seq_res, 1);
2874 2875 2876
	if (!status) {
		update_changeattr(dir, &res.cinfo);
		nfs_post_op_update_inode(dir, res.dir_attr);
2877
		nfs_post_op_update_inode(inode, res.fattr);
2878
	}
2879 2880 2881
out:
	nfs_free_fattr(res.dir_attr);
	nfs_free_fattr(res.fattr);
L
Linus Torvalds 已提交
2882 2883 2884 2885 2886 2887 2888 2889 2890 2891 2892 2893 2894 2895 2896
	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;
}

2897 2898 2899 2900 2901 2902 2903 2904 2905 2906 2907 2908 2909 2910 2911 2912 2913 2914 2915 2916 2917 2918 2919 2920 2921 2922 2923 2924 2925 2926 2927 2928 2929 2930 2931 2932 2933 2934 2935
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;
	struct nfs_fattr dir_fattr;
};

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

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

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

static int nfs4_do_create(struct inode *dir, struct dentry *dentry, struct nfs4_createdata *data)
{
2936
	int status = nfs4_call_sync(NFS_SERVER(dir)->client, NFS_SERVER(dir), &data->msg,
2937
				    &data->arg.seq_args, &data->res.seq_res, 1);
2938 2939 2940 2941 2942 2943 2944 2945 2946 2947 2948 2949 2950
	if (status == 0) {
		update_changeattr(dir, &data->res.dir_cinfo);
		nfs_post_op_update_inode(dir, data->res.dir_fattr);
		status = nfs_instantiate(dentry, data->res.fh, data->res.fattr);
	}
	return status;
}

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

2951
static int _nfs4_proc_symlink(struct inode *dir, struct dentry *dentry,
2952
		struct page *page, unsigned int len, struct iattr *sattr)
L
Linus Torvalds 已提交
2953
{
2954 2955
	struct nfs4_createdata *data;
	int status = -ENAMETOOLONG;
L
Linus Torvalds 已提交
2956

2957
	if (len > NFS4_MAXPATHLEN)
2958
		goto out;
2959

2960 2961 2962 2963 2964 2965 2966 2967
	status = -ENOMEM;
	data = nfs4_alloc_createdata(dir, &dentry->d_name, sattr, NF4LNK);
	if (data == NULL)
		goto out;

	data->msg.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_SYMLINK];
	data->arg.u.symlink.pages = &page;
	data->arg.u.symlink.len = len;
L
Linus Torvalds 已提交
2968
	
2969 2970 2971 2972
	status = nfs4_do_create(dir, dentry, data);

	nfs4_free_createdata(data);
out:
L
Linus Torvalds 已提交
2973 2974 2975
	return status;
}

2976
static int nfs4_proc_symlink(struct inode *dir, struct dentry *dentry,
2977
		struct page *page, unsigned int len, struct iattr *sattr)
L
Linus Torvalds 已提交
2978 2979 2980 2981 2982
{
	struct nfs4_exception exception = { };
	int err;
	do {
		err = nfs4_handle_exception(NFS_SERVER(dir),
2983 2984
				_nfs4_proc_symlink(dir, dentry, page,
							len, sattr),
L
Linus Torvalds 已提交
2985 2986 2987 2988 2989 2990 2991 2992
				&exception);
	} while (exception.retry);
	return err;
}

static int _nfs4_proc_mkdir(struct inode *dir, struct dentry *dentry,
		struct iattr *sattr)
{
2993 2994
	struct nfs4_createdata *data;
	int status = -ENOMEM;
L
Linus Torvalds 已提交
2995

2996 2997 2998 2999 3000 3001 3002 3003
	data = nfs4_alloc_createdata(dir, &dentry->d_name, sattr, NF4DIR);
	if (data == NULL)
		goto out;

	status = nfs4_do_create(dir, dentry, data);

	nfs4_free_createdata(data);
out:
L
Linus Torvalds 已提交
3004 3005 3006 3007 3008 3009 3010 3011
	return status;
}

static int nfs4_proc_mkdir(struct inode *dir, struct dentry *dentry,
		struct iattr *sattr)
{
	struct nfs4_exception exception = { };
	int err;
A
Aneesh Kumar K.V 已提交
3012 3013

	sattr->ia_mode &= ~current_umask();
L
Linus Torvalds 已提交
3014 3015 3016 3017 3018 3019 3020 3021 3022
	do {
		err = nfs4_handle_exception(NFS_SERVER(dir),
				_nfs4_proc_mkdir(dir, dentry, sattr),
				&exception);
	} while (exception.retry);
	return err;
}

static int _nfs4_proc_readdir(struct dentry *dentry, struct rpc_cred *cred,
B
Bryan Schumaker 已提交
3023
		u64 cookie, struct page **pages, unsigned int count, int plus)
L
Linus Torvalds 已提交
3024 3025 3026 3027
{
	struct inode		*dir = dentry->d_inode;
	struct nfs4_readdir_arg args = {
		.fh = NFS_FH(dir),
B
Bryan Schumaker 已提交
3028
		.pages = pages,
L
Linus Torvalds 已提交
3029 3030
		.pgbase = 0,
		.count = count,
3031
		.bitmask = NFS_SERVER(dentry->d_inode)->attr_bitmask,
B
Bryan Schumaker 已提交
3032
		.plus = plus,
L
Linus Torvalds 已提交
3033 3034 3035 3036 3037 3038 3039 3040 3041 3042
	};
	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;

3043
	dprintk("%s: dentry = %s/%s, cookie = %Lu\n", __func__,
3044 3045 3046
			dentry->d_parent->d_name.name,
			dentry->d_name.name,
			(unsigned long long)cookie);
L
Linus Torvalds 已提交
3047 3048
	nfs4_setup_readdir(cookie, NFS_COOKIEVERF(dir), dentry, &args);
	res.pgbase = args.pgbase;
3049
	status = nfs4_call_sync(NFS_SERVER(dir)->client, NFS_SERVER(dir), &msg, &args.seq_args, &res.seq_res, 0);
3050
	if (status >= 0) {
L
Linus Torvalds 已提交
3051
		memcpy(NFS_COOKIEVERF(dir), res.verifier.data, NFS4_VERIFIER_SIZE);
3052 3053
		status += args.pgbase;
	}
3054 3055 3056

	nfs_invalidate_atime(dir);

3057
	dprintk("%s: returns %d\n", __func__, status);
L
Linus Torvalds 已提交
3058 3059 3060 3061
	return status;
}

static int nfs4_proc_readdir(struct dentry *dentry, struct rpc_cred *cred,
B
Bryan Schumaker 已提交
3062
		u64 cookie, struct page **pages, unsigned int count, int plus)
L
Linus Torvalds 已提交
3063 3064 3065 3066 3067 3068
{
	struct nfs4_exception exception = { };
	int err;
	do {
		err = nfs4_handle_exception(NFS_SERVER(dentry->d_inode),
				_nfs4_proc_readdir(dentry, cred, cookie,
B
Bryan Schumaker 已提交
3069
					pages, count, plus),
L
Linus Torvalds 已提交
3070 3071 3072 3073 3074 3075 3076 3077
				&exception);
	} while (exception.retry);
	return err;
}

static int _nfs4_proc_mknod(struct inode *dir, struct dentry *dentry,
		struct iattr *sattr, dev_t rdev)
{
3078 3079 3080
	struct nfs4_createdata *data;
	int mode = sattr->ia_mode;
	int status = -ENOMEM;
L
Linus Torvalds 已提交
3081 3082 3083

	BUG_ON(!(sattr->ia_valid & ATTR_MODE));
	BUG_ON(!S_ISFIFO(mode) && !S_ISBLK(mode) && !S_ISCHR(mode) && !S_ISSOCK(mode));
3084 3085 3086 3087 3088

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

L
Linus Torvalds 已提交
3089
	if (S_ISFIFO(mode))
3090
		data->arg.ftype = NF4FIFO;
L
Linus Torvalds 已提交
3091
	else if (S_ISBLK(mode)) {
3092 3093 3094
		data->arg.ftype = NF4BLK;
		data->arg.u.device.specdata1 = MAJOR(rdev);
		data->arg.u.device.specdata2 = MINOR(rdev);
L
Linus Torvalds 已提交
3095 3096
	}
	else if (S_ISCHR(mode)) {
3097 3098 3099
		data->arg.ftype = NF4CHR;
		data->arg.u.device.specdata1 = MAJOR(rdev);
		data->arg.u.device.specdata2 = MINOR(rdev);
L
Linus Torvalds 已提交
3100 3101
	}
	
3102 3103 3104 3105
	status = nfs4_do_create(dir, dentry, data);

	nfs4_free_createdata(data);
out:
L
Linus Torvalds 已提交
3106 3107 3108 3109 3110 3111 3112 3113
	return status;
}

static int nfs4_proc_mknod(struct inode *dir, struct dentry *dentry,
		struct iattr *sattr, dev_t rdev)
{
	struct nfs4_exception exception = { };
	int err;
A
Aneesh Kumar K.V 已提交
3114 3115

	sattr->ia_mode &= ~current_umask();
L
Linus Torvalds 已提交
3116 3117 3118 3119 3120 3121 3122 3123 3124 3125 3126 3127 3128 3129 3130
	do {
		err = nfs4_handle_exception(NFS_SERVER(dir),
				_nfs4_proc_mknod(dir, dentry, sattr, rdev),
				&exception);
	} while (exception.retry);
	return err;
}

static int _nfs4_proc_statfs(struct nfs_server *server, struct nfs_fh *fhandle,
		 struct nfs_fsstat *fsstat)
{
	struct nfs4_statfs_arg args = {
		.fh = fhandle,
		.bitmask = server->attr_bitmask,
	};
B
Benny Halevy 已提交
3131 3132 3133
	struct nfs4_statfs_res res = {
		.fsstat = fsstat,
	};
L
Linus Torvalds 已提交
3134 3135 3136
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_STATFS],
		.rpc_argp = &args,
B
Benny Halevy 已提交
3137
		.rpc_resp = &res,
L
Linus Torvalds 已提交
3138 3139
	};

3140
	nfs_fattr_init(fsstat->fattr);
3141
	return  nfs4_call_sync(server->client, server, &msg, &args.seq_args, &res.seq_res, 0);
L
Linus Torvalds 已提交
3142 3143 3144 3145 3146 3147 3148 3149 3150 3151 3152 3153 3154 3155 3156 3157 3158 3159 3160 3161 3162
}

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
3163 3164 3165
	struct nfs4_fsinfo_res res = {
		.fsinfo = fsinfo,
	};
L
Linus Torvalds 已提交
3166 3167 3168
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_FSINFO],
		.rpc_argp = &args,
B
Benny Halevy 已提交
3169
		.rpc_resp = &res,
L
Linus Torvalds 已提交
3170 3171
	};

3172
	return nfs4_call_sync(server->client, server, &msg, &args.seq_args, &res.seq_res, 0);
L
Linus Torvalds 已提交
3173 3174 3175 3176 3177 3178 3179 3180 3181 3182 3183 3184 3185 3186 3187 3188 3189
}

static int nfs4_do_fsinfo(struct nfs_server *server, struct nfs_fh *fhandle, struct nfs_fsinfo *fsinfo)
{
	struct nfs4_exception exception = { };
	int err;

	do {
		err = nfs4_handle_exception(server,
				_nfs4_do_fsinfo(server, fhandle, fsinfo),
				&exception);
	} while (exception.retry);
	return err;
}

static int nfs4_proc_fsinfo(struct nfs_server *server, struct nfs_fh *fhandle, struct nfs_fsinfo *fsinfo)
{
3190
	nfs_fattr_init(fsinfo->fattr);
L
Linus Torvalds 已提交
3191 3192 3193 3194 3195 3196 3197 3198 3199 3200
	return nfs4_do_fsinfo(server, fhandle, fsinfo);
}

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
3201 3202 3203
	struct nfs4_pathconf_res res = {
		.pathconf = pathconf,
	};
L
Linus Torvalds 已提交
3204 3205 3206
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_PATHCONF],
		.rpc_argp = &args,
B
Benny Halevy 已提交
3207
		.rpc_resp = &res,
L
Linus Torvalds 已提交
3208 3209 3210 3211 3212 3213 3214 3215
	};

	/* 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;
	}

3216
	nfs_fattr_init(pathconf->fattr);
3217
	return nfs4_call_sync(server->client, server, &msg, &args.seq_args, &res.seq_res, 0);
L
Linus Torvalds 已提交
3218 3219 3220 3221 3222 3223 3224 3225 3226 3227 3228 3229 3230 3231 3232 3233
}

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;
}

3234 3235 3236 3237 3238
void __nfs4_read_done_cb(struct nfs_read_data *data)
{
	nfs_invalidate_atime(data->inode);
}

3239
static int nfs4_read_done_cb(struct rpc_task *task, struct nfs_read_data *data)
L
Linus Torvalds 已提交
3240
{
T
Trond Myklebust 已提交
3241
	struct nfs_server *server = NFS_SERVER(data->inode);
L
Linus Torvalds 已提交
3242

3243
	if (nfs4_async_handle_error(task, server, data->args.context->state) == -EAGAIN) {
3244
		rpc_restart_call_prepare(task);
T
Trond Myklebust 已提交
3245
		return -EAGAIN;
L
Linus Torvalds 已提交
3246
	}
3247

3248
	__nfs4_read_done_cb(data);
L
Linus Torvalds 已提交
3249
	if (task->tk_status > 0)
T
Trond Myklebust 已提交
3250 3251
		renew_lease(server, data->timestamp);
	return 0;
L
Linus Torvalds 已提交
3252 3253
}

3254 3255 3256 3257 3258 3259 3260 3261
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;

3262 3263
	return data->read_done_cb ? data->read_done_cb(task, data) :
				    nfs4_read_done_cb(task, data);
3264 3265
}

3266
static void nfs4_proc_read_setup(struct nfs_read_data *data, struct rpc_message *msg)
L
Linus Torvalds 已提交
3267 3268
{
	data->timestamp   = jiffies;
3269
	data->read_done_cb = nfs4_read_done_cb;
3270
	msg->rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_READ];
3271
	nfs41_init_sequence(&data->args.seq_args, &data->res.seq_res, 0);
L
Linus Torvalds 已提交
3272 3273
}

3274 3275 3276 3277 3278 3279 3280 3281 3282 3283 3284 3285 3286 3287 3288 3289
/* Reset the the nfs_read_data to send the read to the MDS. */
void nfs4_reset_read(struct rpc_task *task, struct nfs_read_data *data)
{
	dprintk("%s Reset task for i/o through\n", __func__);
	put_lseg(data->lseg);
	data->lseg = NULL;
	/* offsets will differ in the dense stripe case */
	data->args.offset = data->mds_offset;
	data->ds_clp = NULL;
	data->args.fh     = NFS_FH(data->inode);
	data->read_done_cb = nfs4_read_done_cb;
	task->tk_ops = data->mds_ops;
	rpc_task_reset_client(task, NFS_CLIENT(data->inode));
}
EXPORT_SYMBOL_GPL(nfs4_reset_read);

3290
static int nfs4_write_done_cb(struct rpc_task *task, struct nfs_write_data *data)
L
Linus Torvalds 已提交
3291 3292 3293
{
	struct inode *inode = data->inode;
	
3294
	if (nfs4_async_handle_error(task, NFS_SERVER(inode), data->args.context->state) == -EAGAIN) {
3295
		rpc_restart_call_prepare(task);
3296
		return -EAGAIN;
L
Linus Torvalds 已提交
3297
	}
3298
	if (task->tk_status >= 0) {
L
Linus Torvalds 已提交
3299
		renew_lease(NFS_SERVER(inode), data->timestamp);
3300
		nfs_post_op_update_inode_force_wcc(inode, data->res.fattr);
3301
	}
3302
	return 0;
L
Linus Torvalds 已提交
3303 3304
}

3305 3306 3307 3308
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;
3309 3310
	return data->write_done_cb ? data->write_done_cb(task, data) :
		nfs4_write_done_cb(task, data);
3311 3312
}

3313 3314 3315 3316 3317 3318 3319 3320 3321 3322 3323 3324 3325 3326 3327 3328 3329
/* Reset the the nfs_write_data to send the write to the MDS. */
void nfs4_reset_write(struct rpc_task *task, struct nfs_write_data *data)
{
	dprintk("%s Reset task for i/o through\n", __func__);
	put_lseg(data->lseg);
	data->lseg          = NULL;
	data->ds_clp        = NULL;
	data->write_done_cb = nfs4_write_done_cb;
	data->args.fh       = NFS_FH(data->inode);
	data->args.bitmask  = data->res.server->cache_consistency_bitmask;
	data->args.offset   = data->mds_offset;
	data->res.fattr     = &data->fattr;
	task->tk_ops        = data->mds_ops;
	rpc_task_reset_client(task, NFS_CLIENT(data->inode));
}
EXPORT_SYMBOL_GPL(nfs4_reset_write);

3330
static void nfs4_proc_write_setup(struct nfs_write_data *data, struct rpc_message *msg)
L
Linus Torvalds 已提交
3331
{
3332 3333
	struct nfs_server *server = NFS_SERVER(data->inode);

3334 3335 3336 3337 3338
	if (data->lseg) {
		data->args.bitmask = NULL;
		data->res.fattr = NULL;
	} else
		data->args.bitmask = server->cache_consistency_bitmask;
3339 3340
	if (!data->write_done_cb)
		data->write_done_cb = nfs4_write_done_cb;
3341
	data->res.server = server;
L
Linus Torvalds 已提交
3342 3343
	data->timestamp   = jiffies;

3344
	msg->rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_WRITE];
3345
	nfs41_init_sequence(&data->args.seq_args, &data->res.seq_res, 1);
L
Linus Torvalds 已提交
3346 3347
}

3348
static int nfs4_commit_done_cb(struct rpc_task *task, struct nfs_write_data *data)
L
Linus Torvalds 已提交
3349 3350
{
	struct inode *inode = data->inode;
3351

3352
	if (nfs4_async_handle_error(task, NFS_SERVER(inode), NULL) == -EAGAIN) {
3353
		rpc_restart_call_prepare(task);
3354
		return -EAGAIN;
L
Linus Torvalds 已提交
3355
	}
3356
	nfs_refresh_inode(inode, data->res.fattr);
3357
	return 0;
L
Linus Torvalds 已提交
3358 3359
}

3360 3361 3362 3363 3364 3365 3366
static int nfs4_commit_done(struct rpc_task *task, struct nfs_write_data *data)
{
	if (!nfs4_sequence_done(task, &data->res.seq_res))
		return -EAGAIN;
	return data->write_done_cb(task, data);
}

3367
static void nfs4_proc_commit_setup(struct nfs_write_data *data, struct rpc_message *msg)
L
Linus Torvalds 已提交
3368
{
3369
	struct nfs_server *server = NFS_SERVER(data->inode);
3370 3371 3372 3373 3374 3375

	if (data->lseg) {
		data->args.bitmask = NULL;
		data->res.fattr = NULL;
	} else
		data->args.bitmask = server->cache_consistency_bitmask;
3376 3377
	if (!data->write_done_cb)
		data->write_done_cb = nfs4_commit_done_cb;
3378
	data->res.server = server;
3379
	msg->rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_COMMIT];
3380
	nfs41_init_sequence(&data->args.seq_args, &data->res.seq_res, 1);
L
Linus Torvalds 已提交
3381 3382
}

3383 3384 3385 3386 3387
struct nfs4_renewdata {
	struct nfs_client	*client;
	unsigned long		timestamp;
};

L
Linus Torvalds 已提交
3388 3389 3390 3391
/*
 * nfs4_proc_async_renew(): This is not one of the nfs_rpc_ops; it is a special
 * standalone procedure for queueing an asynchronous RENEW.
 */
3392
static void nfs4_renew_release(void *calldata)
3393
{
3394 3395
	struct nfs4_renewdata *data = calldata;
	struct nfs_client *clp = data->client;
3396

3397 3398 3399
	if (atomic_read(&clp->cl_count) > 1)
		nfs4_schedule_state_renewal(clp);
	nfs_put_client(clp);
3400
	kfree(data);
3401 3402
}

3403
static void nfs4_renew_done(struct rpc_task *task, void *calldata)
L
Linus Torvalds 已提交
3404
{
3405 3406 3407
	struct nfs4_renewdata *data = calldata;
	struct nfs_client *clp = data->client;
	unsigned long timestamp = data->timestamp;
L
Linus Torvalds 已提交
3408 3409

	if (task->tk_status < 0) {
3410
		/* Unless we're shutting down, schedule state recovery! */
3411 3412 3413
		if (test_bit(NFS_CS_RENEWD, &clp->cl_res_state) == 0)
			return;
		if (task->tk_status != NFS4ERR_CB_PATH_DOWN) {
3414
			nfs4_schedule_lease_recovery(clp);
3415 3416 3417
			return;
		}
		nfs4_schedule_path_down_recovery(clp);
L
Linus Torvalds 已提交
3418
	}
3419
	do_renew_lease(clp, timestamp);
L
Linus Torvalds 已提交
3420 3421
}

3422 3423
static const struct rpc_call_ops nfs4_renew_ops = {
	.rpc_call_done = nfs4_renew_done,
3424
	.rpc_release = nfs4_renew_release,
3425 3426
};

3427
static int nfs4_proc_async_renew(struct nfs_client *clp, struct rpc_cred *cred, unsigned renew_flags)
L
Linus Torvalds 已提交
3428 3429 3430 3431
{
	struct rpc_message msg = {
		.rpc_proc	= &nfs4_procedures[NFSPROC4_CLNT_RENEW],
		.rpc_argp	= clp,
3432
		.rpc_cred	= cred,
L
Linus Torvalds 已提交
3433
	};
3434
	struct nfs4_renewdata *data;
L
Linus Torvalds 已提交
3435

3436 3437
	if (renew_flags == 0)
		return 0;
3438 3439
	if (!atomic_inc_not_zero(&clp->cl_count))
		return -EIO;
3440
	data = kmalloc(sizeof(*data), GFP_NOFS);
3441 3442 3443 3444
	if (data == NULL)
		return -ENOMEM;
	data->client = clp;
	data->timestamp = jiffies;
L
Linus Torvalds 已提交
3445
	return rpc_call_async(clp->cl_rpcclient, &msg, RPC_TASK_SOFT,
3446
			&nfs4_renew_ops, data);
L
Linus Torvalds 已提交
3447 3448
}

3449
static int nfs4_proc_renew(struct nfs_client *clp, struct rpc_cred *cred)
L
Linus Torvalds 已提交
3450 3451 3452 3453
{
	struct rpc_message msg = {
		.rpc_proc	= &nfs4_procedures[NFSPROC4_CLNT_RENEW],
		.rpc_argp	= clp,
3454
		.rpc_cred	= cred,
L
Linus Torvalds 已提交
3455 3456 3457 3458 3459 3460 3461
	};
	unsigned long now = jiffies;
	int status;

	status = rpc_call_sync(clp->cl_rpcclient, &msg, 0);
	if (status < 0)
		return status;
3462
	do_renew_lease(clp, now);
L
Linus Torvalds 已提交
3463 3464 3465
	return 0;
}

3466 3467 3468 3469 3470 3471 3472 3473 3474 3475 3476 3477 3478
static inline int nfs4_server_supports_acls(struct nfs_server *server)
{
	return (server->caps & NFS_CAP_ACLS)
		&& (server->acl_bitmask & ACL4_SUPPORT_ALLOW_ACL)
		&& (server->acl_bitmask & ACL4_SUPPORT_DENY_ACL);
}

/* Assuming that XATTR_SIZE_MAX is a multiple of PAGE_CACHE_SIZE, and that
 * it's OK to put sizeof(void) * (XATTR_SIZE_MAX/PAGE_CACHE_SIZE) bytes on
 * the stack.
 */
#define NFS4ACL_MAXPAGES (XATTR_SIZE_MAX >> PAGE_CACHE_SHIFT)

3479 3480 3481 3482 3483 3484 3485 3486 3487
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 {
J
Jovi Zhang 已提交
3488
		len = min_t(size_t, PAGE_CACHE_SIZE, buflen);
3489 3490 3491 3492 3493 3494 3495 3496 3497 3498 3499 3500 3501 3502 3503 3504 3505 3506 3507
		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;
}

3508 3509 3510
struct nfs4_cached_acl {
	int cached;
	size_t len;
A
Andrew Morton 已提交
3511
	char data[0];
3512 3513 3514 3515 3516 3517 3518 3519 3520 3521 3522 3523 3524 3525 3526 3527 3528 3529 3530 3531 3532 3533 3534 3535 3536 3537 3538 3539 3540 3541 3542 3543 3544 3545 3546 3547 3548 3549 3550 3551 3552 3553 3554 3555 3556 3557 3558 3559 3560 3561 3562 3563 3564 3565 3566 3567 3568 3569 3570 3571 3572 3573 3574
};

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;
}

static void nfs4_write_cached_acl(struct inode *inode, const char *buf, size_t acl_len)
{
	struct nfs4_cached_acl *acl;

	if (buf && acl_len <= PAGE_SIZE) {
		acl = kmalloc(sizeof(*acl) + acl_len, GFP_KERNEL);
		if (acl == NULL)
			goto out;
		acl->cached = 1;
		memcpy(acl->data, buf, acl_len);
	} 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);
}

3575 3576 3577 3578 3579 3580 3581 3582 3583 3584
/*
 * 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.
 */
3585
static ssize_t __nfs4_get_acl_uncached(struct inode *inode, void *buf, size_t buflen)
3586
{
3587
	struct page *pages[NFS4ACL_MAXPAGES] = {NULL, };
3588 3589 3590 3591 3592
	struct nfs_getaclargs args = {
		.fh = NFS_FH(inode),
		.acl_pages = pages,
		.acl_len = buflen,
	};
B
Benny Halevy 已提交
3593 3594 3595
	struct nfs_getaclres res = {
		.acl_len = buflen,
	};
3596
	void *resp_buf;
3597 3598 3599
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_GETACL],
		.rpc_argp = &args,
B
Benny Halevy 已提交
3600
		.rpc_resp = &res,
3601
	};
3602
	int ret = -ENOMEM, npages, i, acl_len = 0;
3603

3604 3605 3606 3607 3608 3609 3610 3611 3612 3613
	npages = (buflen + PAGE_SIZE - 1) >> PAGE_SHIFT;
	/* 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;

	for (i = 0; i < npages; i++) {
		pages[i] = alloc_page(GFP_KERNEL);
		if (!pages[i])
			goto out_free;
3614
	}
3615 3616 3617 3618 3619
	if (npages > 1) {
		/* for decoding across pages */
		args.acl_scratch = alloc_page(GFP_KERNEL);
		if (!args.acl_scratch)
			goto out_free;
3620
	}
3621 3622 3623 3624 3625 3626 3627 3628
	args.acl_len = npages * PAGE_SIZE;
	args.acl_pgbase = 0;
	/* Let decode_getfacl know not to fail if the ACL data is larger than
	 * the page we send as a guess */
	if (buf == NULL)
		res.acl_flags |= NFS4_ACL_LEN_REQUEST;
	resp_buf = page_address(pages[0]);

3629
	dprintk("%s  buf %p buflen %zu npages %d args.acl_len %zu\n",
3630 3631 3632
		__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);
3633 3634
	if (ret)
		goto out_free;
3635 3636 3637 3638

	acl_len = res.acl_len - res.acl_data_offset;
	if (acl_len > args.acl_len)
		nfs4_write_cached_acl(inode, NULL, acl_len);
3639
	else
3640 3641
		nfs4_write_cached_acl(inode, resp_buf + res.acl_data_offset,
				      acl_len);
3642 3643
	if (buf) {
		ret = -ERANGE;
3644
		if (acl_len > buflen)
3645
			goto out_free;
3646 3647
		_copy_from_pages(buf, pages, res.acl_data_offset,
				res.acl_len);
3648
	}
3649
	ret = acl_len;
3650
out_free:
3651 3652 3653 3654 3655
	for (i = 0; i < npages; i++)
		if (pages[i])
			__free_page(pages[i]);
	if (args.acl_scratch)
		__free_page(args.acl_scratch);
3656 3657 3658
	return ret;
}

3659 3660 3661 3662 3663 3664 3665 3666 3667 3668 3669 3670 3671
static ssize_t nfs4_get_acl_uncached(struct inode *inode, void *buf, size_t buflen)
{
	struct nfs4_exception exception = { };
	ssize_t ret;
	do {
		ret = __nfs4_get_acl_uncached(inode, buf, buflen);
		if (ret >= 0)
			break;
		ret = nfs4_handle_exception(NFS_SERVER(inode), ret, &exception);
	} while (exception.retry);
	return ret;
}

3672 3673 3674 3675 3676 3677 3678 3679 3680 3681
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;
3682 3683
	if (NFS_I(inode)->cache_validity & NFS_INO_INVALID_ACL)
		nfs_zap_acl_cache(inode);
3684 3685
	ret = nfs4_read_cached_acl(inode, buf, buflen);
	if (ret != -ENOENT)
3686 3687
		/* -ENOENT is returned if there is no ACL or if there is an ACL
		 * but no cached acl data, just the acl length */
3688 3689 3690 3691
		return ret;
	return nfs4_get_acl_uncached(inode, buf, buflen);
}

3692
static int __nfs4_proc_set_acl(struct inode *inode, const void *buf, size_t buflen)
3693 3694 3695 3696 3697 3698 3699 3700
{
	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
3701
	struct nfs_setaclres res;
3702 3703 3704
	struct rpc_message msg = {
		.rpc_proc	= &nfs4_procedures[NFSPROC4_CLNT_SETACL],
		.rpc_argp	= &arg,
B
Benny Halevy 已提交
3705
		.rpc_resp	= &res,
3706
	};
3707
	int ret, i;
3708 3709 3710

	if (!nfs4_server_supports_acls(server))
		return -EOPNOTSUPP;
3711 3712 3713
	i = buf_to_pages_noslab(buf, buflen, arg.acl_pages, &arg.acl_pgbase);
	if (i < 0)
		return i;
3714
	nfs_inode_return_delegation(inode);
3715
	ret = nfs4_call_sync(server->client, server, &msg, &arg.seq_args, &res.seq_res, 1);
3716 3717 3718 3719 3720 3721 3722 3723

	/*
	 * 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]);

3724 3725 3726 3727 3728 3729 3730
	/*
	 * 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);
3731 3732
	nfs_access_zap_cache(inode);
	nfs_zap_acl_cache(inode);
3733 3734 3735
	return ret;
}

3736 3737 3738 3739 3740 3741 3742 3743 3744 3745 3746 3747
static int nfs4_proc_set_acl(struct inode *inode, const void *buf, size_t buflen)
{
	struct nfs4_exception exception = { };
	int err;
	do {
		err = nfs4_handle_exception(NFS_SERVER(inode),
				__nfs4_proc_set_acl(inode, buf, buflen),
				&exception);
	} while (exception.retry);
	return err;
}

L
Linus Torvalds 已提交
3748
static int
3749
nfs4_async_handle_error(struct rpc_task *task, const struct nfs_server *server, struct nfs4_state *state)
L
Linus Torvalds 已提交
3750
{
3751 3752 3753
	struct nfs_client *clp = server->nfs_client;

	if (task->tk_status >= 0)
L
Linus Torvalds 已提交
3754 3755
		return 0;
	switch(task->tk_status) {
3756 3757 3758 3759 3760
		case -NFS4ERR_ADMIN_REVOKED:
		case -NFS4ERR_BAD_STATEID:
		case -NFS4ERR_OPENMODE:
			if (state == NULL)
				break;
3761 3762
			nfs4_schedule_stateid_recovery(server, state);
			goto wait_on_recovery;
3763 3764 3765
		case -NFS4ERR_EXPIRED:
			if (state != NULL)
				nfs4_schedule_stateid_recovery(server, state);
L
Linus Torvalds 已提交
3766
		case -NFS4ERR_STALE_STATEID:
3767
		case -NFS4ERR_STALE_CLIENTID:
3768 3769
			nfs4_schedule_lease_recovery(clp);
			goto wait_on_recovery;
3770 3771 3772 3773 3774 3775 3776 3777 3778 3779
#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);
3780
			nfs4_schedule_session_recovery(clp->cl_session);
3781 3782 3783
			task->tk_status = 0;
			return -EAGAIN;
#endif /* CONFIG_NFS_V4_1 */
L
Linus Torvalds 已提交
3784
		case -NFS4ERR_DELAY:
3785
			nfs_inc_server_stats(server, NFSIOS_DELAY);
3786
		case -NFS4ERR_GRACE:
3787
		case -EKEYEXPIRED:
L
Linus Torvalds 已提交
3788 3789 3790
			rpc_delay(task, NFS4_POLL_RETRY_MAX);
			task->tk_status = 0;
			return -EAGAIN;
3791
		case -NFS4ERR_RETRY_UNCACHED_REP:
L
Linus Torvalds 已提交
3792 3793 3794 3795 3796 3797
		case -NFS4ERR_OLD_STATEID:
			task->tk_status = 0;
			return -EAGAIN;
	}
	task->tk_status = nfs4_map_errors(task->tk_status);
	return 0;
3798
wait_on_recovery:
3799 3800 3801 3802 3803
	rpc_sleep_on(&clp->cl_rpcwaitq, task, NULL);
	if (test_bit(NFS4CLNT_MANAGER_RUNNING, &clp->cl_state) == 0)
		rpc_wake_up_queued_task(&clp->cl_rpcwaitq, task);
	task->tk_status = 0;
	return -EAGAIN;
L
Linus Torvalds 已提交
3804 3805
}

3806 3807 3808
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
3809 3810 3811 3812 3813
{
	nfs4_verifier sc_verifier;
	struct nfs4_setclientid setclientid = {
		.sc_verifier = &sc_verifier,
		.sc_prog = program,
A
Andy Adamson 已提交
3814
		.sc_cb_ident = clp->cl_cb_ident,
L
Linus Torvalds 已提交
3815 3816 3817 3818
	};
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_SETCLIENTID],
		.rpc_argp = &setclientid,
3819
		.rpc_resp = res,
3820
		.rpc_cred = cred,
L
Linus Torvalds 已提交
3821
	};
A
Al Viro 已提交
3822
	__be32 *p;
L
Linus Torvalds 已提交
3823 3824 3825
	int loop = 0;
	int status;

A
Al Viro 已提交
3826
	p = (__be32*)sc_verifier.data;
L
Linus Torvalds 已提交
3827 3828 3829 3830 3831
	*p++ = htonl((u32)clp->cl_boot_time.tv_sec);
	*p = htonl((u32)clp->cl_boot_time.tv_nsec);

	for(;;) {
		setclientid.sc_name_len = scnprintf(setclientid.sc_name,
3832
				sizeof(setclientid.sc_name), "%s/%s %s %s %u",
3833 3834 3835
				clp->cl_ipaddr,
				rpc_peeraddr2str(clp->cl_rpcclient,
							RPC_DISPLAY_ADDR),
3836 3837
				rpc_peeraddr2str(clp->cl_rpcclient,
							RPC_DISPLAY_PROTO),
3838
				clp->cl_rpcclient->cl_auth->au_ops->au_name,
L
Linus Torvalds 已提交
3839 3840
				clp->cl_id_uniquifier);
		setclientid.sc_netid_len = scnprintf(setclientid.sc_netid,
3841 3842 3843
				sizeof(setclientid.sc_netid),
				rpc_peeraddr2str(clp->cl_rpcclient,
							RPC_DISPLAY_NETID));
L
Linus Torvalds 已提交
3844
		setclientid.sc_uaddr_len = scnprintf(setclientid.sc_uaddr,
3845
				sizeof(setclientid.sc_uaddr), "%s.%u.%u",
L
Linus Torvalds 已提交
3846 3847
				clp->cl_ipaddr, port >> 8, port & 255);

3848
		status = rpc_call_sync(clp->cl_rpcclient, &msg, RPC_TASK_TIMEOUT);
L
Linus Torvalds 已提交
3849 3850
		if (status != -NFS4ERR_CLID_INUSE)
			break;
3851 3852
		if (loop != 0) {
			++clp->cl_id_uniquifier;
L
Linus Torvalds 已提交
3853
			break;
3854 3855 3856
		}
		++loop;
		ssleep(clp->cl_lease_time / HZ + 1);
L
Linus Torvalds 已提交
3857 3858 3859 3860
	}
	return status;
}

3861
int nfs4_proc_setclientid_confirm(struct nfs_client *clp,
3862 3863
		struct nfs4_setclientid_res *arg,
		struct rpc_cred *cred)
L
Linus Torvalds 已提交
3864 3865 3866 3867
{
	struct nfs_fsinfo fsinfo;
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_SETCLIENTID_CONFIRM],
3868
		.rpc_argp = arg,
L
Linus Torvalds 已提交
3869
		.rpc_resp = &fsinfo,
3870
		.rpc_cred = cred,
L
Linus Torvalds 已提交
3871 3872 3873 3874 3875
	};
	unsigned long now;
	int status;

	now = jiffies;
3876
	status = rpc_call_sync(clp->cl_rpcclient, &msg, RPC_TASK_TIMEOUT);
L
Linus Torvalds 已提交
3877 3878 3879 3880 3881 3882 3883 3884 3885
	if (status == 0) {
		spin_lock(&clp->cl_lock);
		clp->cl_lease_time = fsinfo.lease_time * HZ;
		clp->cl_last_renewal = now;
		spin_unlock(&clp->cl_lock);
	}
	return status;
}

3886 3887
struct nfs4_delegreturndata {
	struct nfs4_delegreturnargs args;
3888
	struct nfs4_delegreturnres res;
3889 3890
	struct nfs_fh fh;
	nfs4_stateid stateid;
3891
	unsigned long timestamp;
3892
	struct nfs_fattr fattr;
3893 3894 3895 3896 3897 3898
	int rpc_status;
};

static void nfs4_delegreturn_done(struct rpc_task *task, void *calldata)
{
	struct nfs4_delegreturndata *data = calldata;
3899

3900 3901
	if (!nfs4_sequence_done(task, &data->res.seq_res))
		return;
3902

3903 3904 3905 3906
	switch (task->tk_status) {
	case -NFS4ERR_STALE_STATEID:
	case -NFS4ERR_EXPIRED:
	case 0:
3907
		renew_lease(data->res.server, data->timestamp);
3908 3909 3910 3911
		break;
	default:
		if (nfs4_async_handle_error(task, data->res.server, NULL) ==
				-EAGAIN) {
3912
			rpc_restart_call_prepare(task);
3913 3914 3915 3916
			return;
		}
	}
	data->rpc_status = task->tk_status;
3917 3918 3919 3920 3921 3922 3923
}

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

3924 3925 3926 3927 3928 3929 3930
#if defined(CONFIG_NFS_V4_1)
static void nfs4_delegreturn_prepare(struct rpc_task *task, void *data)
{
	struct nfs4_delegreturndata *d_data;

	d_data = (struct nfs4_delegreturndata *)data;

3931
	if (nfs4_setup_sequence(d_data->res.server,
3932
				&d_data->args.seq_args,
3933
				&d_data->res.seq_res, task))
3934 3935 3936 3937 3938
		return;
	rpc_call_start(task);
}
#endif /* CONFIG_NFS_V4_1 */

3939
static const struct rpc_call_ops nfs4_delegreturn_ops = {
3940 3941 3942
#if defined(CONFIG_NFS_V4_1)
	.rpc_call_prepare = nfs4_delegreturn_prepare,
#endif /* CONFIG_NFS_V4_1 */
3943 3944 3945 3946
	.rpc_call_done = nfs4_delegreturn_done,
	.rpc_release = nfs4_delegreturn_release,
};

3947
static int _nfs4_proc_delegreturn(struct inode *inode, struct rpc_cred *cred, const nfs4_stateid *stateid, int issync)
3948 3949
{
	struct nfs4_delegreturndata *data;
3950
	struct nfs_server *server = NFS_SERVER(inode);
3951
	struct rpc_task *task;
3952 3953 3954 3955
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_DELEGRETURN],
		.rpc_cred = cred,
	};
T
Trond Myklebust 已提交
3956 3957
	struct rpc_task_setup task_setup_data = {
		.rpc_client = server->client,
3958
		.rpc_message = &msg,
T
Trond Myklebust 已提交
3959 3960 3961
		.callback_ops = &nfs4_delegreturn_ops,
		.flags = RPC_TASK_ASYNC,
	};
3962
	int status = 0;
3963

3964
	data = kzalloc(sizeof(*data), GFP_NOFS);
3965 3966
	if (data == NULL)
		return -ENOMEM;
3967
	nfs41_init_sequence(&data->args.seq_args, &data->res.seq_res, 1);
3968 3969
	data->args.fhandle = &data->fh;
	data->args.stateid = &data->stateid;
3970
	data->args.bitmask = server->attr_bitmask;
3971 3972
	nfs_copy_fh(&data->fh, NFS_FH(inode));
	memcpy(&data->stateid, stateid, sizeof(data->stateid));
3973 3974
	data->res.fattr = &data->fattr;
	data->res.server = server;
3975
	nfs_fattr_init(data->res.fattr);
3976
	data->timestamp = jiffies;
3977 3978
	data->rpc_status = 0;

T
Trond Myklebust 已提交
3979
	task_setup_data.callback_data = data;
3980 3981
	msg.rpc_argp = &data->args;
	msg.rpc_resp = &data->res;
T
Trond Myklebust 已提交
3982
	task = rpc_run_task(&task_setup_data);
3983
	if (IS_ERR(task))
3984
		return PTR_ERR(task);
3985 3986
	if (!issync)
		goto out;
3987
	status = nfs4_wait_for_completion_rpc_task(task);
3988 3989 3990 3991 3992 3993 3994
	if (status != 0)
		goto out;
	status = data->rpc_status;
	if (status != 0)
		goto out;
	nfs_refresh_inode(inode, &data->fattr);
out:
3995
	rpc_put_task(task);
3996
	return status;
L
Linus Torvalds 已提交
3997 3998
}

3999
int nfs4_proc_delegreturn(struct inode *inode, struct rpc_cred *cred, const nfs4_stateid *stateid, int issync)
L
Linus Torvalds 已提交
4000 4001 4002 4003 4004
{
	struct nfs_server *server = NFS_SERVER(inode);
	struct nfs4_exception exception = { };
	int err;
	do {
4005
		err = _nfs4_proc_delegreturn(inode, cred, stateid, issync);
L
Linus Torvalds 已提交
4006 4007 4008 4009 4010 4011 4012 4013 4014 4015 4016 4017 4018 4019 4020 4021 4022 4023 4024 4025
		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)
{
4026
	freezable_schedule_timeout_killable(timeout);
L
Linus Torvalds 已提交
4027 4028 4029 4030 4031 4032 4033 4034 4035 4036
	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);
4037
	struct nfs_client *clp = server->nfs_client;
T
Trond Myklebust 已提交
4038
	struct nfs_lockt_args arg = {
L
Linus Torvalds 已提交
4039
		.fh = NFS_FH(inode),
T
Trond Myklebust 已提交
4040
		.fl = request,
L
Linus Torvalds 已提交
4041
	};
T
Trond Myklebust 已提交
4042 4043
	struct nfs_lockt_res res = {
		.denied = request,
L
Linus Torvalds 已提交
4044 4045 4046 4047 4048 4049 4050 4051 4052 4053
	};
	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
4054
	arg.lock_owner.clientid = clp->cl_clientid;
4055 4056 4057 4058
	status = nfs4_set_lock_state(state, request);
	if (status != 0)
		goto out;
	lsp = request->fl_u.nfs4_fl.owner;
4059
	arg.lock_owner.id = lsp->ls_seqid.owner_id;
4060
	arg.lock_owner.s_dev = server->s_dev;
4061
	status = nfs4_call_sync(server->client, server, &msg, &arg.seq_args, &res.seq_res, 1);
T
Trond Myklebust 已提交
4062 4063 4064 4065 4066 4067
	switch (status) {
		case 0:
			request->fl_type = F_UNLCK;
			break;
		case -NFS4ERR_DENIED:
			status = 0;
L
Linus Torvalds 已提交
4068
	}
4069
	request->fl_ops->fl_release_private(request);
4070
out:
L
Linus Torvalds 已提交
4071 4072 4073 4074 4075 4076 4077 4078 4079 4080 4081 4082 4083 4084 4085 4086 4087 4088 4089 4090 4091 4092 4093 4094 4095 4096 4097 4098 4099 4100 4101 4102
	return status;
}

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

	do {
		err = nfs4_handle_exception(NFS_SERVER(state->inode),
				_nfs4_proc_getlk(state, cmd, request),
				&exception);
	} while (exception.retry);
	return err;
}

static int do_vfs_lock(struct file *file, struct file_lock *fl)
{
	int res = 0;
	switch (fl->fl_flags & (FL_POSIX|FL_FLOCK)) {
		case FL_POSIX:
			res = posix_lock_file_wait(file, fl);
			break;
		case FL_FLOCK:
			res = flock_lock_file_wait(file, fl);
			break;
		default:
			BUG();
	}
	return res;
}

4103
struct nfs4_unlockdata {
T
Trond Myklebust 已提交
4104 4105
	struct nfs_locku_args arg;
	struct nfs_locku_res res;
4106 4107
	struct nfs4_lock_state *lsp;
	struct nfs_open_context *ctx;
T
Trond Myklebust 已提交
4108 4109
	struct file_lock fl;
	const struct nfs_server *server;
4110
	unsigned long timestamp;
4111 4112
};

T
Trond Myklebust 已提交
4113 4114 4115 4116 4117 4118 4119 4120
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;

4121
	p = kzalloc(sizeof(*p), GFP_NOFS);
T
Trond Myklebust 已提交
4122 4123 4124 4125 4126
	if (p == NULL)
		return NULL;
	p->arg.fh = NFS_FH(inode);
	p->arg.fl = &p->fl;
	p->arg.seqid = seqid;
4127
	p->res.seqid = seqid;
T
Trond Myklebust 已提交
4128 4129 4130 4131 4132 4133 4134 4135 4136 4137
	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;
}

4138
static void nfs4_locku_release_calldata(void *data)
4139
{
4140
	struct nfs4_unlockdata *calldata = data;
T
Trond Myklebust 已提交
4141
	nfs_free_seqid(calldata->arg.seqid);
4142 4143 4144
	nfs4_put_lock_state(calldata->lsp);
	put_nfs_open_context(calldata->ctx);
	kfree(calldata);
4145 4146
}

4147
static void nfs4_locku_done(struct rpc_task *task, void *data)
4148
{
4149
	struct nfs4_unlockdata *calldata = data;
4150

4151 4152
	if (!nfs4_sequence_done(task, &calldata->res.seq_res))
		return;
4153 4154 4155
	switch (task->tk_status) {
		case 0:
			memcpy(calldata->lsp->ls_stateid.data,
T
Trond Myklebust 已提交
4156
					calldata->res.stateid.data,
4157
					sizeof(calldata->lsp->ls_stateid.data));
4158
			renew_lease(calldata->server, calldata->timestamp);
4159
			break;
4160 4161
		case -NFS4ERR_BAD_STATEID:
		case -NFS4ERR_OLD_STATEID:
4162 4163 4164 4165
		case -NFS4ERR_STALE_STATEID:
		case -NFS4ERR_EXPIRED:
			break;
		default:
4166
			if (nfs4_async_handle_error(task, calldata->server, NULL) == -EAGAIN)
4167
				rpc_restart_call_prepare(task);
4168 4169 4170
	}
}

T
Trond Myklebust 已提交
4171
static void nfs4_locku_prepare(struct rpc_task *task, void *data)
4172
{
T
Trond Myklebust 已提交
4173
	struct nfs4_unlockdata *calldata = data;
4174

T
Trond Myklebust 已提交
4175
	if (nfs_wait_on_sequence(calldata->arg.seqid, task) != 0)
4176 4177
		return;
	if ((calldata->lsp->ls_flags & NFS_LOCK_INITIALIZED) == 0) {
4178 4179
		/* Note: exit _without_ running nfs4_locku_done */
		task->tk_action = NULL;
4180 4181
		return;
	}
4182
	calldata->timestamp = jiffies;
4183
	if (nfs4_setup_sequence(calldata->server,
4184
				&calldata->arg.seq_args,
4185
				&calldata->res.seq_res, task))
4186
		return;
4187
	rpc_call_start(task);
4188 4189
}

4190
static const struct rpc_call_ops nfs4_locku_ops = {
T
Trond Myklebust 已提交
4191
	.rpc_call_prepare = nfs4_locku_prepare,
4192
	.rpc_call_done = nfs4_locku_done,
4193
	.rpc_release = nfs4_locku_release_calldata,
4194 4195
};

4196 4197 4198 4199 4200 4201
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;
4202 4203 4204 4205
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_LOCKU],
		.rpc_cred = ctx->cred,
	};
T
Trond Myklebust 已提交
4206 4207
	struct rpc_task_setup task_setup_data = {
		.rpc_client = NFS_CLIENT(lsp->ls_state->inode),
4208
		.rpc_message = &msg,
T
Trond Myklebust 已提交
4209
		.callback_ops = &nfs4_locku_ops,
4210
		.workqueue = nfsiod_workqueue,
T
Trond Myklebust 已提交
4211 4212
		.flags = RPC_TASK_ASYNC,
	};
4213

4214 4215 4216 4217 4218
	/* 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;

4219 4220 4221 4222 4223 4224
	data = nfs4_alloc_unlockdata(fl, ctx, lsp, seqid);
	if (data == NULL) {
		nfs_free_seqid(seqid);
		return ERR_PTR(-ENOMEM);
	}

4225
	nfs41_init_sequence(&data->arg.seq_args, &data->res.seq_res, 1);
4226 4227
	msg.rpc_argp = &data->arg;
	msg.rpc_resp = &data->res;
T
Trond Myklebust 已提交
4228 4229
	task_setup_data.callback_data = data;
	return rpc_run_task(&task_setup_data);
4230 4231
}

4232 4233
static int nfs4_proc_unlck(struct nfs4_state *state, int cmd, struct file_lock *request)
{
4234
	struct nfs_inode *nfsi = NFS_I(state->inode);
T
Trond Myklebust 已提交
4235
	struct nfs_seqid *seqid;
L
Linus Torvalds 已提交
4236
	struct nfs4_lock_state *lsp;
4237 4238
	struct rpc_task *task;
	int status = 0;
4239
	unsigned char fl_flags = request->fl_flags;
4240

4241
	status = nfs4_set_lock_state(state, request);
4242 4243
	/* Unlock _before_ we do the RPC call */
	request->fl_flags |= FL_EXISTS;
4244 4245 4246
	down_read(&nfsi->rwsem);
	if (do_vfs_lock(request->fl_file, request) == -ENOENT) {
		up_read(&nfsi->rwsem);
4247
		goto out;
4248 4249
	}
	up_read(&nfsi->rwsem);
4250
	if (status != 0)
4251 4252 4253 4254
		goto out;
	/* Is this a delegated lock? */
	if (test_bit(NFS_DELEGATED_STATE, &state->flags))
		goto out;
4255
	lsp = request->fl_u.nfs4_fl.owner;
4256
	seqid = nfs_alloc_seqid(&lsp->ls_seqid, GFP_KERNEL);
4257
	status = -ENOMEM;
T
Trond Myklebust 已提交
4258
	if (seqid == NULL)
4259
		goto out;
4260
	task = nfs4_do_unlck(request, nfs_file_open_context(request->fl_file), lsp, seqid);
4261 4262
	status = PTR_ERR(task);
	if (IS_ERR(task))
4263
		goto out;
4264
	status = nfs4_wait_for_completion_rpc_task(task);
4265
	rpc_put_task(task);
4266
out:
4267
	request->fl_flags = fl_flags;
L
Linus Torvalds 已提交
4268 4269 4270
	return status;
}

4271 4272 4273 4274 4275 4276
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;
4277
	unsigned long timestamp;
4278 4279
	int rpc_status;
	int cancelled;
4280
	struct nfs_server *server;
4281 4282 4283
};

static struct nfs4_lockdata *nfs4_alloc_lockdata(struct file_lock *fl,
4284 4285
		struct nfs_open_context *ctx, struct nfs4_lock_state *lsp,
		gfp_t gfp_mask)
L
Linus Torvalds 已提交
4286
{
4287 4288
	struct nfs4_lockdata *p;
	struct inode *inode = lsp->ls_state->inode;
L
Linus Torvalds 已提交
4289
	struct nfs_server *server = NFS_SERVER(inode);
4290

4291
	p = kzalloc(sizeof(*p), gfp_mask);
4292 4293 4294 4295 4296
	if (p == NULL)
		return NULL;

	p->arg.fh = NFS_FH(inode);
	p->arg.fl = &p->fl;
4297
	p->arg.open_seqid = nfs_alloc_seqid(&lsp->ls_state->owner->so_seqid, gfp_mask);
4298 4299
	if (p->arg.open_seqid == NULL)
		goto out_free;
4300
	p->arg.lock_seqid = nfs_alloc_seqid(&lsp->ls_seqid, gfp_mask);
4301
	if (p->arg.lock_seqid == NULL)
4302
		goto out_free_seqid;
4303
	p->arg.lock_stateid = &lsp->ls_stateid;
4304
	p->arg.lock_owner.clientid = server->nfs_client->cl_clientid;
4305
	p->arg.lock_owner.id = lsp->ls_seqid.owner_id;
4306
	p->arg.lock_owner.s_dev = server->s_dev;
4307
	p->res.lock_seqid = p->arg.lock_seqid;
4308
	p->lsp = lsp;
4309
	p->server = server;
4310 4311 4312 4313
	atomic_inc(&lsp->ls_count);
	p->ctx = get_nfs_open_context(ctx);
	memcpy(&p->fl, fl, sizeof(p->fl));
	return p;
4314 4315
out_free_seqid:
	nfs_free_seqid(p->arg.open_seqid);
4316 4317 4318 4319 4320 4321 4322 4323 4324
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;
4325

4326
	dprintk("%s: begin!\n", __func__);
4327 4328
	if (nfs_wait_on_sequence(data->arg.lock_seqid, task) != 0)
		return;
4329 4330
	/* Do we need to do an open_to_lock_owner? */
	if (!(data->arg.lock_seqid->sequence->flags & NFS_SEQID_CONFIRMED)) {
4331 4332
		if (nfs_wait_on_sequence(data->arg.open_seqid, task) != 0)
			return;
4333 4334
		data->arg.open_stateid = &state->stateid;
		data->arg.new_lock_owner = 1;
4335
		data->res.open_seqid = data->arg.open_seqid;
4336 4337
	} else
		data->arg.new_lock_owner = 0;
4338
	data->timestamp = jiffies;
4339 4340
	if (nfs4_setup_sequence(data->server,
				&data->arg.seq_args,
4341
				&data->res.seq_res, task))
4342
		return;
4343
	rpc_call_start(task);
4344
	dprintk("%s: done!, ret = %d\n", __func__, data->rpc_status);
4345 4346
}

4347 4348 4349 4350 4351 4352
static void nfs4_recover_lock_prepare(struct rpc_task *task, void *calldata)
{
	rpc_task_set_priority(task, RPC_PRIORITY_PRIVILEGED);
	nfs4_lock_prepare(task, calldata);
}

4353 4354 4355 4356
static void nfs4_lock_done(struct rpc_task *task, void *calldata)
{
	struct nfs4_lockdata *data = calldata;

4357
	dprintk("%s: begin!\n", __func__);
4358

4359 4360
	if (!nfs4_sequence_done(task, &data->res.seq_res))
		return;
4361

4362 4363 4364 4365 4366 4367 4368 4369 4370 4371 4372
	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) {
		memcpy(data->lsp->ls_stateid.data, data->res.stateid.data,
					sizeof(data->lsp->ls_stateid.data));
		data->lsp->ls_flags |= NFS_LOCK_INITIALIZED;
4373
		renew_lease(NFS_SERVER(data->ctx->dentry->d_inode), data->timestamp);
4374 4375
	}
out:
4376
	dprintk("%s: done, ret = %d!\n", __func__, data->rpc_status);
4377 4378 4379 4380 4381 4382
}

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

4383
	dprintk("%s: begin!\n", __func__);
4384
	nfs_free_seqid(data->arg.open_seqid);
4385 4386 4387 4388 4389
	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))
4390
			rpc_put_task_async(task);
4391
		dprintk("%s: cancelling lock!\n", __func__);
4392 4393 4394 4395 4396
	} else
		nfs_free_seqid(data->arg.lock_seqid);
	nfs4_put_lock_state(data->lsp);
	put_nfs_open_context(data->ctx);
	kfree(data);
4397
	dprintk("%s: done!\n", __func__);
4398 4399 4400 4401 4402 4403 4404 4405
}

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,
};

4406 4407 4408 4409 4410 4411
static const struct rpc_call_ops nfs4_recover_lock_ops = {
	.rpc_call_prepare = nfs4_recover_lock_prepare,
	.rpc_call_done = nfs4_lock_done,
	.rpc_release = nfs4_lock_release,
};

4412 4413 4414 4415 4416
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:
4417
		lsp->ls_seqid.flags &= ~NFS_SEQID_CONFIRMED;
4418 4419
		if (new_lock_owner != 0 ||
		   (lsp->ls_flags & NFS_LOCK_INITIALIZED) != 0)
4420
			nfs4_schedule_stateid_recovery(server, lsp->ls_state);
4421 4422 4423
		break;
	case -NFS4ERR_STALE_STATEID:
		lsp->ls_seqid.flags &= ~NFS_SEQID_CONFIRMED;
4424 4425
	case -NFS4ERR_EXPIRED:
		nfs4_schedule_lease_recovery(server->nfs_client);
4426 4427 4428
	};
}

4429
static int _nfs4_do_setlk(struct nfs4_state *state, int cmd, struct file_lock *fl, int recovery_type)
4430 4431 4432
{
	struct nfs4_lockdata *data;
	struct rpc_task *task;
4433 4434 4435 4436
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_LOCK],
		.rpc_cred = state->owner->so_cred,
	};
T
Trond Myklebust 已提交
4437 4438
	struct rpc_task_setup task_setup_data = {
		.rpc_client = NFS_CLIENT(state->inode),
4439
		.rpc_message = &msg,
T
Trond Myklebust 已提交
4440
		.callback_ops = &nfs4_lock_ops,
4441
		.workqueue = nfsiod_workqueue,
T
Trond Myklebust 已提交
4442 4443
		.flags = RPC_TASK_ASYNC,
	};
4444 4445
	int ret;

4446
	dprintk("%s: begin!\n", __func__);
4447
	data = nfs4_alloc_lockdata(fl, nfs_file_open_context(fl->fl_file),
4448 4449
			fl->fl_u.nfs4_fl.owner,
			recovery_type == NFS_LOCK_NEW ? GFP_KERNEL : GFP_NOFS);
4450 4451 4452 4453
	if (data == NULL)
		return -ENOMEM;
	if (IS_SETLKW(cmd))
		data->arg.block = 1;
4454 4455
	if (recovery_type > NFS_LOCK_NEW) {
		if (recovery_type == NFS_LOCK_RECLAIM)
4456
			data->arg.reclaim = NFS_LOCK_RECLAIM;
4457 4458
		task_setup_data.callback_ops = &nfs4_recover_lock_ops;
	}
4459
	nfs41_init_sequence(&data->arg.seq_args, &data->res.seq_res, 1);
4460 4461
	msg.rpc_argp = &data->arg;
	msg.rpc_resp = &data->res;
T
Trond Myklebust 已提交
4462 4463
	task_setup_data.callback_data = data;
	task = rpc_run_task(&task_setup_data);
4464
	if (IS_ERR(task))
4465 4466 4467 4468
		return PTR_ERR(task);
	ret = nfs4_wait_for_completion_rpc_task(task);
	if (ret == 0) {
		ret = data->rpc_status;
4469 4470 4471
		if (ret)
			nfs4_handle_setlk_error(data->server, data->lsp,
					data->arg.new_lock_owner, ret);
4472 4473
	} else
		data->cancelled = 1;
4474
	rpc_put_task(task);
4475
	dprintk("%s: done, ret = %d!\n", __func__, ret);
4476
	return ret;
L
Linus Torvalds 已提交
4477 4478 4479 4480
}

static int nfs4_lock_reclaim(struct nfs4_state *state, struct file_lock *request)
{
4481 4482 4483 4484 4485
	struct nfs_server *server = NFS_SERVER(state->inode);
	struct nfs4_exception exception = { };
	int err;

	do {
4486 4487 4488
		/* Cache the lock if possible... */
		if (test_bit(NFS_DELEGATED_STATE, &state->flags) != 0)
			return 0;
4489
		err = _nfs4_do_setlk(state, F_SETLK, request, NFS_LOCK_RECLAIM);
4490
		if (err != -NFS4ERR_DELAY)
4491 4492 4493 4494
			break;
		nfs4_handle_exception(server, err, &exception);
	} while (exception.retry);
	return err;
L
Linus Torvalds 已提交
4495 4496 4497 4498
}

static int nfs4_lock_expired(struct nfs4_state *state, struct file_lock *request)
{
4499 4500 4501 4502
	struct nfs_server *server = NFS_SERVER(state->inode);
	struct nfs4_exception exception = { };
	int err;

4503 4504 4505
	err = nfs4_set_lock_state(state, request);
	if (err != 0)
		return err;
4506
	do {
4507 4508
		if (test_bit(NFS_DELEGATED_STATE, &state->flags) != 0)
			return 0;
4509
		err = _nfs4_do_setlk(state, F_SETLK, request, NFS_LOCK_EXPIRED);
4510 4511 4512 4513 4514 4515 4516 4517
		switch (err) {
		default:
			goto out;
		case -NFS4ERR_GRACE:
		case -NFS4ERR_DELAY:
			nfs4_handle_exception(server, err, &exception);
			err = 0;
		}
4518
	} while (exception.retry);
4519
out:
4520
	return err;
L
Linus Torvalds 已提交
4521 4522
}

4523
#if defined(CONFIG_NFS_V4_1)
4524
static int nfs41_check_expired_locks(struct nfs4_state *state)
4525
{
4526 4527
	int status, ret = NFS_OK;
	struct nfs4_lock_state *lsp;
4528 4529
	struct nfs_server *server = NFS_SERVER(state->inode);

4530 4531 4532 4533 4534 4535 4536 4537 4538 4539 4540 4541 4542 4543 4544 4545 4546 4547 4548 4549
	list_for_each_entry(lsp, &state->lock_states, ls_locks) {
		if (lsp->ls_flags & NFS_LOCK_INITIALIZED) {
			status = nfs41_test_stateid(server, &lsp->ls_stateid);
			if (status != NFS_OK) {
				nfs41_free_stateid(server, &lsp->ls_stateid);
				lsp->ls_flags &= ~NFS_LOCK_INITIALIZED;
				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);
4550
	if (status == NFS_OK)
4551
		return status;
4552 4553 4554 4555
	return nfs4_lock_expired(state, request);
}
#endif

L
Linus Torvalds 已提交
4556 4557
static int _nfs4_proc_setlk(struct nfs4_state *state, int cmd, struct file_lock *request)
{
4558
	struct nfs_inode *nfsi = NFS_I(state->inode);
4559
	unsigned char fl_flags = request->fl_flags;
4560
	int status = -ENOLCK;
L
Linus Torvalds 已提交
4561

4562 4563 4564
	if ((fl_flags & FL_POSIX) &&
			!test_bit(NFS_STATE_POSIX_LOCKS, &state->flags))
		goto out;
4565 4566 4567 4568
	/* Is this a delegated open? */
	status = nfs4_set_lock_state(state, request);
	if (status != 0)
		goto out;
4569 4570 4571 4572
	request->fl_flags |= FL_ACCESS;
	status = do_vfs_lock(request->fl_file, request);
	if (status < 0)
		goto out;
4573
	down_read(&nfsi->rwsem);
4574 4575 4576
	if (test_bit(NFS_DELEGATED_STATE, &state->flags)) {
		/* Yes: cache locks! */
		/* ...but avoid races with delegation recall... */
4577 4578 4579
		request->fl_flags = fl_flags & ~FL_SLEEP;
		status = do_vfs_lock(request->fl_file, request);
		goto out_unlock;
4580
	}
4581
	status = _nfs4_do_setlk(state, cmd, request, NFS_LOCK_NEW);
4582
	if (status != 0)
4583
		goto out_unlock;
4584
	/* Note: we always want to sleep here! */
4585
	request->fl_flags = fl_flags | FL_SLEEP;
4586
	if (do_vfs_lock(request->fl_file, request) < 0)
4587 4588
		printk(KERN_WARNING "NFS: %s: VFS is out of sync with lock "
			"manager!\n", __func__);
4589
out_unlock:
4590
	up_read(&nfsi->rwsem);
4591 4592
out:
	request->fl_flags = fl_flags;
L
Linus Torvalds 已提交
4593 4594 4595 4596 4597 4598 4599 4600 4601
	return status;
}

static int nfs4_proc_setlk(struct nfs4_state *state, int cmd, struct file_lock *request)
{
	struct nfs4_exception exception = { };
	int err;

	do {
4602 4603 4604
		err = _nfs4_proc_setlk(state, cmd, request);
		if (err == -NFS4ERR_DENIED)
			err = -EAGAIN;
L
Linus Torvalds 已提交
4605
		err = nfs4_handle_exception(NFS_SERVER(state->inode),
4606
				err, &exception);
L
Linus Torvalds 已提交
4607 4608 4609 4610 4611 4612 4613 4614 4615 4616 4617 4618 4619
	} 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 */
4620
	ctx = nfs_file_open_context(filp);
L
Linus Torvalds 已提交
4621 4622 4623 4624 4625
	state = ctx->state;

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

4626 4627 4628 4629 4630
	if (IS_GETLK(cmd)) {
		if (state != NULL)
			return nfs4_proc_getlk(state, F_GETLK, request);
		return 0;
	}
L
Linus Torvalds 已提交
4631 4632 4633 4634

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

4635 4636 4637 4638 4639
	if (request->fl_type == F_UNLCK) {
		if (state != NULL)
			return nfs4_proc_unlck(state, cmd, request);
		return 0;
	}
L
Linus Torvalds 已提交
4640

4641 4642
	if (state == NULL)
		return -ENOLCK;
L
Linus Torvalds 已提交
4643 4644 4645 4646 4647 4648 4649 4650 4651 4652 4653 4654
	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;
}

4655 4656 4657 4658 4659 4660 4661 4662 4663 4664
int nfs4_lock_delegation_recall(struct nfs4_state *state, struct file_lock *fl)
{
	struct nfs_server *server = NFS_SERVER(state->inode);
	struct nfs4_exception exception = { };
	int err;

	err = nfs4_set_lock_state(state, fl);
	if (err != 0)
		goto out;
	do {
4665
		err = _nfs4_do_setlk(state, F_SETLK, fl, NFS_LOCK_NEW);
4666 4667
		switch (err) {
			default:
4668 4669
				printk(KERN_ERR "NFS: %s: unhandled error "
					"%d.\n", __func__, err);
4670
			case 0:
4671
			case -ESTALE:
4672 4673
				goto out;
			case -NFS4ERR_EXPIRED:
4674
				nfs4_schedule_stateid_recovery(server, state);
4675
			case -NFS4ERR_STALE_CLIENTID:
4676
			case -NFS4ERR_STALE_STATEID:
4677 4678
				nfs4_schedule_lease_recovery(server->nfs_client);
				goto out;
4679 4680 4681 4682 4683
			case -NFS4ERR_BADSESSION:
			case -NFS4ERR_BADSLOT:
			case -NFS4ERR_BAD_HIGH_SLOT:
			case -NFS4ERR_CONN_NOT_BOUND_TO_SESSION:
			case -NFS4ERR_DEADSESSION:
4684
				nfs4_schedule_session_recovery(server->nfs_client->cl_session);
4685
				goto out;
4686 4687 4688 4689 4690 4691 4692 4693
			case -ERESTARTSYS:
				/*
				 * The show must go on: exit, but mark the
				 * stateid as needing recovery.
				 */
			case -NFS4ERR_ADMIN_REVOKED:
			case -NFS4ERR_BAD_STATEID:
			case -NFS4ERR_OPENMODE:
4694
				nfs4_schedule_stateid_recovery(server, state);
4695 4696
				err = 0;
				goto out;
4697 4698 4699 4700 4701 4702 4703 4704 4705
			case -EKEYEXPIRED:
				/*
				 * User RPCSEC_GSS context has expired.
				 * We cannot recover this stateid now, so
				 * skip it and allow recovery thread to
				 * proceed.
				 */
				err = 0;
				goto out;
4706 4707 4708 4709 4710
			case -ENOMEM:
			case -NFS4ERR_DENIED:
				/* kill_proc(fl->fl_pid, SIGLOST, 1); */
				err = 0;
				goto out;
4711 4712 4713
			case -NFS4ERR_DELAY:
				break;
		}
4714 4715 4716 4717 4718
		err = nfs4_handle_exception(server, err, &exception);
	} while (exception.retry);
out:
	return err;
}
4719

4720 4721 4722 4723 4724 4725 4726 4727 4728 4729 4730 4731 4732 4733 4734 4735 4736 4737 4738 4739 4740 4741 4742
static void nfs4_release_lockowner_release(void *calldata)
{
	kfree(calldata);
}

const struct rpc_call_ops nfs4_release_lockowner_ops = {
	.rpc_release = nfs4_release_lockowner_release,
};

void nfs4_release_lockowner(const struct nfs4_lock_state *lsp)
{
	struct nfs_server *server = lsp->ls_state->owner->so_server;
	struct nfs_release_lockowner_args *args;
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_RELEASE_LOCKOWNER],
	};

	if (server->nfs_client->cl_mvops->minor_version != 0)
		return;
	args = kmalloc(sizeof(*args), GFP_NOFS);
	if (!args)
		return;
	args->lock_owner.clientid = server->nfs_client->cl_clientid;
4743
	args->lock_owner.id = lsp->ls_seqid.owner_id;
4744
	args->lock_owner.s_dev = server->s_dev;
4745 4746 4747 4748
	msg.rpc_argp = args;
	rpc_call_async(server->client, &msg, 0, &nfs4_release_lockowner_ops, args);
}

4749 4750
#define XATTR_NAME_NFSV4_ACL "system.nfs4_acl"

4751 4752 4753
static int nfs4_xattr_set_nfs4_acl(struct dentry *dentry, const char *key,
				   const void *buf, size_t buflen,
				   int flags, int type)
4754
{
4755 4756
	if (strcmp(key, "") != 0)
		return -EINVAL;
4757

4758
	return nfs4_proc_set_acl(dentry->d_inode, buf, buflen);
4759 4760
}

4761 4762
static int nfs4_xattr_get_nfs4_acl(struct dentry *dentry, const char *key,
				   void *buf, size_t buflen, int type)
4763
{
4764 4765
	if (strcmp(key, "") != 0)
		return -EINVAL;
4766

4767
	return nfs4_proc_get_acl(dentry->d_inode, buf, buflen);
4768 4769
}

4770 4771 4772
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)
4773
{
4774
	size_t len = sizeof(XATTR_NAME_NFSV4_ACL);
4775

4776 4777
	if (!nfs4_server_supports_acls(NFS_SERVER(dentry->d_inode)))
		return 0;
4778 4779 4780

	if (list && len <= list_len)
		memcpy(list, XATTR_NAME_NFSV4_ACL, len);
4781
	return len;
4782 4783
}

4784 4785 4786
/*
 * nfs_fhget will use either the mounted_on_fileid or the fileid
 */
4787 4788
static void nfs_fixup_referral_attributes(struct nfs_fattr *fattr)
{
4789 4790 4791 4792
	if (!(((fattr->valid & NFS_ATTR_FATTR_MOUNTED_ON_FILEID) ||
	       (fattr->valid & NFS_ATTR_FATTR_FILEID)) &&
	      (fattr->valid & NFS_ATTR_FATTR_FSID) &&
	      (fattr->valid & NFS_ATTR_FATTR_V4_REFERRAL)))
4793 4794 4795 4796 4797 4798 4799 4800
		return;

	fattr->valid |= NFS_ATTR_FATTR_TYPE | NFS_ATTR_FATTR_MODE |
		NFS_ATTR_FATTR_NLINK;
	fattr->mode = S_IFDIR | S_IRUGO | S_IXUGO;
	fattr->nlink = 2;
}

4801
int nfs4_proc_fs_locations(struct inode *dir, const struct qstr *name,
4802
		struct nfs4_fs_locations *fs_locations, struct page *page)
4803 4804 4805
{
	struct nfs_server *server = NFS_SERVER(dir);
	u32 bitmask[2] = {
4806
		[0] = FATTR4_WORD0_FSID | FATTR4_WORD0_FS_LOCATIONS,
4807 4808 4809
	};
	struct nfs4_fs_locations_arg args = {
		.dir_fh = NFS_FH(dir),
4810
		.name = name,
4811 4812 4813
		.page = page,
		.bitmask = bitmask,
	};
B
Benny Halevy 已提交
4814 4815 4816
	struct nfs4_fs_locations_res res = {
		.fs_locations = fs_locations,
	};
4817 4818 4819
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_FS_LOCATIONS],
		.rpc_argp = &args,
B
Benny Halevy 已提交
4820
		.rpc_resp = &res,
4821 4822 4823
	};
	int status;

4824
	dprintk("%s: start\n", __func__);
4825 4826 4827 4828 4829 4830 4831 4832

	/* 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;

4833
	nfs_fattr_init(&fs_locations->fattr);
4834
	fs_locations->server = server;
4835
	fs_locations->nlocations = 0;
4836
	status = nfs4_call_sync(server->client, server, &msg, &args.seq_args, &res.seq_res, 0);
4837
	dprintk("%s: returned status = %d\n", __func__, status);
4838 4839 4840
	return status;
}

B
Bryan Schumaker 已提交
4841 4842 4843 4844 4845 4846 4847 4848 4849 4850 4851 4852 4853 4854 4855 4856 4857 4858 4859 4860 4861 4862 4863 4864 4865 4866 4867 4868 4869 4870 4871 4872 4873 4874
static int _nfs4_proc_secinfo(struct inode *dir, const struct qstr *name, struct nfs4_secinfo_flavors *flavors)
{
	int status;
	struct nfs4_secinfo_arg args = {
		.dir_fh = NFS_FH(dir),
		.name   = name,
	};
	struct nfs4_secinfo_res res = {
		.flavors     = flavors,
	};
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_SECINFO],
		.rpc_argp = &args,
		.rpc_resp = &res,
	};

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

int nfs4_proc_secinfo(struct inode *dir, const struct qstr *name, struct nfs4_secinfo_flavors *flavors)
{
	struct nfs4_exception exception = { };
	int err;
	do {
		err = nfs4_handle_exception(NFS_SERVER(dir),
				_nfs4_proc_secinfo(dir, name, flavors),
				&exception);
	} while (exception.retry);
	return err;
}

4875
#ifdef CONFIG_NFS_V4_1
4876 4877 4878 4879 4880 4881 4882 4883 4884 4885 4886 4887 4888 4889 4890 4891 4892 4893 4894
/*
 * 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;
}

4895 4896 4897 4898 4899 4900 4901 4902 4903 4904
static bool
nfs41_same_server_scope(struct server_scope *a, struct server_scope *b)
{
	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;
}

B
Benny Halevy 已提交
4905 4906 4907 4908 4909 4910 4911 4912
/*
 * nfs4_proc_exchange_id()
 *
 * 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.
 */
4913
int nfs4_proc_exchange_id(struct nfs_client *clp, struct rpc_cred *cred)
B
Benny Halevy 已提交
4914 4915 4916 4917
{
	nfs4_verifier verifier;
	struct nfs41_exchange_id_args args = {
		.client = clp,
4918
		.flags = EXCHGID4_FLAG_SUPP_MOVED_REFER,
B
Benny Halevy 已提交
4919 4920 4921 4922 4923 4924 4925 4926 4927 4928 4929 4930 4931 4932 4933
	};
	struct nfs41_exchange_id_res res = {
		.client = clp,
	};
	int status;
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_EXCHANGE_ID],
		.rpc_argp = &args,
		.rpc_resp = &res,
		.rpc_cred = cred,
	};
	__be32 *p;

	dprintk("--> %s\n", __func__);
	BUG_ON(clp == NULL);
4934

B
Benny Halevy 已提交
4935 4936 4937 4938 4939
	p = (u32 *)verifier.data;
	*p++ = htonl((u32)clp->cl_boot_time.tv_sec);
	*p = htonl((u32)clp->cl_boot_time.tv_nsec);
	args.verifier = &verifier;

4940 4941 4942 4943 4944 4945
	args.id_len = scnprintf(args.id, sizeof(args.id),
				"%s/%s.%s/%u",
				clp->cl_ipaddr,
				init_utsname()->nodename,
				init_utsname()->domainname,
				clp->cl_rpcclient->cl_auth->au_flavor);
B
Benny Halevy 已提交
4946

4947
	res.server_scope = kzalloc(sizeof(struct server_scope), GFP_KERNEL);
4948 4949 4950 4951
	if (unlikely(!res.server_scope)) {
		status = -ENOMEM;
		goto out;
	}
4952

4953 4954 4955 4956 4957 4958
	res.impl_id = kzalloc(sizeof(struct nfs41_impl_id), GFP_KERNEL);
	if (unlikely(!res.impl_id)) {
		status = -ENOMEM;
		goto out_server_scope;
	}

4959
	status = rpc_call_sync(clp->cl_rpcclient, &msg, RPC_TASK_TIMEOUT);
4960 4961
	if (!status)
		status = nfs4_check_cl_exchange_flags(clp->cl_exchange_flags);
4962

4963 4964 4965 4966 4967 4968 4969
	if (!status) {
		/* use the most recent implementation id */
		kfree(clp->impl_id);
		clp->impl_id = res.impl_id;
	} else
		kfree(res.impl_id);

4970 4971 4972 4973 4974 4975 4976 4977 4978 4979 4980
	if (!status) {
		if (clp->server_scope &&
		    !nfs41_same_server_scope(clp->server_scope,
					     res.server_scope)) {
			dprintk("%s: server_scope mismatch detected\n",
				__func__);
			set_bit(NFS4CLNT_SERVER_SCOPE_MISMATCH, &clp->cl_state);
			kfree(clp->server_scope);
			clp->server_scope = NULL;
		}

4981
		if (!clp->server_scope) {
4982
			clp->server_scope = res.server_scope;
4983 4984
			goto out;
		}
4985
	}
4986 4987

out_server_scope:
4988 4989
	kfree(res.server_scope);
out:
4990 4991 4992 4993 4994 4995
	if (clp->impl_id)
		dprintk("%s: Server Implementation ID: "
			"domain: %s, name: %s, date: %llu,%u\n",
			__func__, clp->impl_id->domain, clp->impl_id->name,
			clp->impl_id->date.seconds,
			clp->impl_id->date.nseconds);
B
Benny Halevy 已提交
4996 4997 4998 4999
	dprintk("<-- %s status= %d\n", __func__, status);
	return status;
}

A
Andy Adamson 已提交
5000 5001 5002 5003 5004 5005 5006 5007 5008 5009 5010 5011 5012 5013
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)
{
	int ret;
	struct nfs4_get_lease_time_data *data =
			(struct nfs4_get_lease_time_data *)calldata;

	dprintk("--> %s\n", __func__);
5014
	rpc_task_set_priority(task, RPC_PRIORITY_PRIVILEGED);
A
Andy Adamson 已提交
5015 5016 5017
	/* just setup sequence, do not trigger session recovery
	   since we're invoked within one */
	ret = nfs41_setup_sequence(data->clp->cl_session,
5018
				   &data->args->la_seq_args,
5019
				   &data->res->lr_seq_res, task);
A
Andy Adamson 已提交
5020 5021 5022 5023 5024 5025 5026 5027 5028 5029 5030 5031 5032 5033 5034 5035

	BUG_ON(ret == -EAGAIN);
	rpc_call_start(task);
	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__);
5036 5037
	if (!nfs41_sequence_done(task, &data->res->lr_seq_res))
		return;
A
Andy Adamson 已提交
5038 5039 5040 5041 5042 5043
	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;
5044 5045
		/* fall through */
	case -NFS4ERR_RETRY_UNCACHED_REP:
5046
		rpc_restart_call_prepare(task);
A
Andy Adamson 已提交
5047 5048 5049 5050 5051 5052 5053 5054 5055 5056 5057 5058 5059 5060 5061 5062 5063 5064 5065 5066 5067 5068 5069 5070 5071 5072 5073 5074 5075 5076 5077
		return;
	}
	dprintk("<-- %s\n", __func__);
}

struct rpc_call_ops nfs4_get_lease_time_ops = {
	.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,
5078 5079
		.callback_data = &data,
		.flags = RPC_TASK_TIMEOUT,
A
Andy Adamson 已提交
5080 5081 5082
	};
	int status;

5083
	nfs41_init_sequence(&args.la_seq_args, &res.lr_seq_res, 0);
A
Andy Adamson 已提交
5084 5085 5086 5087 5088 5089 5090 5091 5092 5093 5094 5095 5096 5097
	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;
}

5098 5099 5100
/*
 * Reset a slot table
 */
A
Andy Adamson 已提交
5101 5102
static int nfs4_reset_slot_table(struct nfs4_slot_table *tbl, u32 max_reqs,
				 int ivalue)
5103
{
A
Andy Adamson 已提交
5104
	struct nfs4_slot *new = NULL;
5105
	int i;
5106 5107
	int ret = 0;

A
Andy Adamson 已提交
5108 5109
	dprintk("--> %s: max_reqs=%u, tbl->max_slots %d\n", __func__,
		max_reqs, tbl->max_slots);
5110

A
Andy Adamson 已提交
5111 5112 5113 5114
	/* Does the newly negotiated max_reqs match the existing slot table? */
	if (max_reqs != tbl->max_slots) {
		ret = -ENOMEM;
		new = kmalloc(max_reqs * sizeof(struct nfs4_slot),
5115
			      GFP_NOFS);
A
Andy Adamson 已提交
5116 5117 5118 5119
		if (!new)
			goto out;
		ret = 0;
		kfree(tbl->slots);
5120 5121
	}
	spin_lock(&tbl->slot_tbl_lock);
A
Andy Adamson 已提交
5122 5123 5124 5125 5126
	if (new) {
		tbl->slots = new;
		tbl->max_slots = max_reqs;
	}
	for (i = 0; i < tbl->max_slots; ++i)
5127
		tbl->slots[i].seq_nr = ivalue;
5128 5129 5130 5131 5132 5133 5134 5135
	spin_unlock(&tbl->slot_tbl_lock);
	dprintk("%s: tbl=%p slots=%p max_slots=%d\n", __func__,
		tbl, tbl->slots, tbl->max_slots);
out:
	dprintk("<-- %s: return %d\n", __func__, ret);
	return ret;
}

5136 5137 5138 5139 5140 5141 5142 5143 5144 5145 5146 5147 5148 5149
/* Destroy the slot table */
static void nfs4_destroy_slot_tables(struct nfs4_session *session)
{
	if (session->fc_slot_table.slots != NULL) {
		kfree(session->fc_slot_table.slots);
		session->fc_slot_table.slots = NULL;
	}
	if (session->bc_slot_table.slots != NULL) {
		kfree(session->bc_slot_table.slots);
		session->bc_slot_table.slots = NULL;
	}
	return;
}

A
Andy Adamson 已提交
5150 5151 5152
/*
 * Initialize slot table
 */
5153 5154
static int nfs4_init_slot_table(struct nfs4_slot_table *tbl,
		int max_slots, int ivalue)
A
Andy Adamson 已提交
5155 5156 5157 5158 5159 5160
{
	struct nfs4_slot *slot;
	int ret = -ENOMEM;

	BUG_ON(max_slots > NFS4_MAX_SLOT_TABLE);

5161
	dprintk("--> %s: max_reqs=%u\n", __func__, max_slots);
A
Andy Adamson 已提交
5162

5163
	slot = kcalloc(max_slots, sizeof(struct nfs4_slot), GFP_NOFS);
A
Andy Adamson 已提交
5164 5165 5166 5167 5168 5169 5170
	if (!slot)
		goto out;
	ret = 0;

	spin_lock(&tbl->slot_tbl_lock);
	tbl->max_slots = max_slots;
	tbl->slots = slot;
5171
	tbl->highest_used_slotid = NFS4_NO_SLOT;  /* no slot is currently used */
A
Andy Adamson 已提交
5172 5173 5174 5175 5176 5177 5178 5179
	spin_unlock(&tbl->slot_tbl_lock);
	dprintk("%s: tbl=%p slots=%p max_slots=%d\n", __func__,
		tbl, tbl->slots, tbl->max_slots);
out:
	dprintk("<-- %s: return %d\n", __func__, ret);
	return ret;
}

5180
/*
5181
 * Initialize or reset the forechannel and backchannel tables
5182
 */
5183
static int nfs4_setup_session_slot_tables(struct nfs4_session *ses)
5184
{
5185
	struct nfs4_slot_table *tbl;
5186
	int status;
5187

5188 5189 5190
	dprintk("--> %s\n", __func__);
	/* Fore channel */
	tbl = &ses->fc_slot_table;
5191
	if (tbl->slots == NULL) {
5192 5193 5194 5195 5196
		status = nfs4_init_slot_table(tbl, ses->fc_attrs.max_reqs, 1);
		if (status) /* -ENOMEM */
			return status;
	} else {
		status = nfs4_reset_slot_table(tbl, ses->fc_attrs.max_reqs, 1);
5197 5198 5199
		if (status)
			return status;
	}
5200 5201
	/* Back channel */
	tbl = &ses->bc_slot_table;
5202
	if (tbl->slots == NULL) {
5203
		status = nfs4_init_slot_table(tbl, ses->bc_attrs.max_reqs, 0);
5204
		if (status)
5205 5206 5207 5208 5209
			/* Fore and back channel share a connection so get
			 * both slot tables or neither */
			nfs4_destroy_slot_tables(ses);
	} else
		status = nfs4_reset_slot_table(tbl, ses->bc_attrs.max_reqs, 0);
5210
	return status;
5211 5212 5213 5214 5215 5216 5217
}

struct nfs4_session *nfs4_alloc_session(struct nfs_client *clp)
{
	struct nfs4_session *session;
	struct nfs4_slot_table *tbl;

5218
	session = kzalloc(sizeof(struct nfs4_session), GFP_NOFS);
5219 5220
	if (!session)
		return NULL;
5221

5222
	tbl = &session->fc_slot_table;
5223
	tbl->highest_used_slotid = NFS4_NO_SLOT;
5224
	spin_lock_init(&tbl->slot_tbl_lock);
5225
	rpc_init_priority_wait_queue(&tbl->slot_tbl_waitq, "ForeChannel Slot table");
5226
	init_completion(&tbl->complete);
5227 5228

	tbl = &session->bc_slot_table;
5229
	tbl->highest_used_slotid = NFS4_NO_SLOT;
5230 5231
	spin_lock_init(&tbl->slot_tbl_lock);
	rpc_init_wait_queue(&tbl->slot_tbl_waitq, "BackChannel Slot table");
5232
	init_completion(&tbl->complete);
5233

5234 5235
	session->session_state = 1<<NFS4_SESSION_INITING;

5236 5237 5238 5239 5240 5241
	session->clp = clp;
	return session;
}

void nfs4_destroy_session(struct nfs4_session *session)
{
5242 5243 5244 5245 5246
	nfs4_proc_destroy_session(session);
	dprintk("%s Destroy backchannel for xprt %p\n",
		__func__, session->clp->cl_rpcclient->cl_xprt);
	xprt_destroy_backchannel(session->clp->cl_rpcclient->cl_xprt,
				NFS41_BC_MIN_CALLBACKS);
5247
	nfs4_destroy_slot_tables(session);
5248 5249 5250
	kfree(session);
}

A
Andy Adamson 已提交
5251 5252 5253 5254 5255 5256 5257 5258 5259 5260 5261 5262 5263 5264 5265 5266 5267 5268 5269 5270 5271 5272 5273
/*
 * 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)
{
	struct nfs4_session *session = args->client->cl_session;
	unsigned int mxrqst_sz = session->fc_attrs.max_rqst_sz,
		     mxresp_sz = session->fc_attrs.max_resp_sz;

	if (mxrqst_sz == 0)
		mxrqst_sz = NFS_MAX_FILE_IO_SIZE;
	if (mxresp_sz == 0)
		mxresp_sz = NFS_MAX_FILE_IO_SIZE;
	/* Fore channel attributes */
	args->fc_attrs.max_rqst_sz = mxrqst_sz;
	args->fc_attrs.max_resp_sz = mxresp_sz;
	args->fc_attrs.max_ops = NFS4_MAX_OPS;
5274
	args->fc_attrs.max_reqs = max_session_slots;
A
Andy Adamson 已提交
5275 5276

	dprintk("%s: Fore Channel : max_rqst_sz=%u max_resp_sz=%u "
5277
		"max_ops=%u max_reqs=%u\n",
A
Andy Adamson 已提交
5278 5279
		__func__,
		args->fc_attrs.max_rqst_sz, args->fc_attrs.max_resp_sz,
5280
		args->fc_attrs.max_ops, args->fc_attrs.max_reqs);
A
Andy Adamson 已提交
5281 5282 5283 5284 5285 5286 5287 5288 5289 5290 5291 5292 5293 5294 5295 5296

	/* 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);
}

5297
static int nfs4_verify_fore_channel_attrs(struct nfs41_create_session_args *args, struct nfs4_session *session)
5298
{
5299 5300 5301 5302 5303 5304 5305 5306 5307 5308 5309 5310 5311 5312 5313
	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;
5314 5315
	if (rcvd->max_reqs > NFS4_MAX_SLOT_TABLE)
		rcvd->max_reqs = NFS4_MAX_SLOT_TABLE;
5316
	return 0;
5317 5318
}

5319 5320 5321 5322
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;
5323

5324 5325 5326 5327 5328 5329 5330
	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: */
5331
	if (rcvd->max_ops != sent->max_ops)
5332
		return -EINVAL;
5333
	if (rcvd->max_reqs != sent->max_reqs)
5334 5335 5336
		return -EINVAL;
	return 0;
}
5337 5338 5339 5340

static int nfs4_verify_channel_attrs(struct nfs41_create_session_args *args,
				     struct nfs4_session *session)
{
5341
	int ret;
5342

5343 5344 5345 5346
	ret = nfs4_verify_fore_channel_attrs(args, session);
	if (ret)
		return ret;
	return nfs4_verify_back_channel_attrs(args, session);
5347 5348
}

A
Andy Adamson 已提交
5349 5350 5351 5352 5353 5354 5355 5356 5357 5358 5359 5360 5361 5362 5363 5364 5365 5366
static int _nfs4_proc_create_session(struct nfs_client *clp)
{
	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,
	};
	int status;

	nfs4_init_channel_attrs(&args);
5367
	args.flags = (SESSION4_PERSIST | SESSION4_BACK_CHAN);
A
Andy Adamson 已提交
5368

5369
	status = rpc_call_sync(session->clp->cl_rpcclient, &msg, RPC_TASK_TIMEOUT);
A
Andy Adamson 已提交
5370

5371 5372 5373
	if (!status)
		/* Verify the session's negotiated channel_attrs values */
		status = nfs4_verify_channel_attrs(&args, session);
A
Andy Adamson 已提交
5374 5375 5376 5377 5378 5379 5380 5381 5382 5383 5384 5385 5386
	if (!status) {
		/* 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.
 */
5387
int nfs4_proc_create_session(struct nfs_client *clp)
A
Andy Adamson 已提交
5388 5389 5390 5391 5392 5393 5394
{
	int status;
	unsigned *ptr;
	struct nfs4_session *session = clp->cl_session;

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

5395
	status = _nfs4_proc_create_session(clp);
A
Andy Adamson 已提交
5396 5397 5398
	if (status)
		goto out;

5399 5400 5401
	/* 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
5402 5403 5404 5405 5406 5407 5408 5409 5410 5411 5412
	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
5413 5414 5415 5416 5417 5418 5419 5420 5421 5422 5423 5424 5425 5426 5427 5428 5429 5430 5431
/*
 * Issue the over-the-wire RPC DESTROY_SESSION.
 * The caller must serialize access to this routine.
 */
int nfs4_proc_destroy_session(struct nfs4_session *session)
{
	int status = 0;
	struct rpc_message msg;

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

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

	msg.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_DESTROY_SESSION];
	msg.rpc_argp = session;
	msg.rpc_resp = NULL;
	msg.rpc_cred = NULL;
5432
	status = rpc_call_sync(session->clp->cl_rpcclient, &msg, RPC_TASK_TIMEOUT);
A
Andy Adamson 已提交
5433 5434 5435

	if (status)
		printk(KERN_WARNING
5436
			"NFS: Got error %d from the server on DESTROY_SESSION. "
A
Andy Adamson 已提交
5437 5438 5439 5440 5441 5442
			"Session has been destroyed regardless...\n", status);

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

5443 5444 5445
int nfs4_init_session(struct nfs_server *server)
{
	struct nfs_client *clp = server->nfs_client;
5446
	struct nfs4_session *session;
5447
	unsigned int rsize, wsize;
5448 5449 5450 5451 5452
	int ret;

	if (!nfs4_has_session(clp))
		return 0;

5453 5454 5455 5456
	session = clp->cl_session;
	if (!test_and_clear_bit(NFS4_SESSION_INITING, &session->session_state))
		return 0;

5457 5458 5459 5460 5461 5462 5463 5464 5465
	rsize = server->rsize;
	if (rsize == 0)
		rsize = NFS_MAX_FILE_IO_SIZE;
	wsize = server->wsize;
	if (wsize == 0)
		wsize = NFS_MAX_FILE_IO_SIZE;

	session->fc_attrs.max_rqst_sz = wsize + nfs41_maxwrite_overhead;
	session->fc_attrs.max_resp_sz = rsize + nfs41_maxread_overhead;
5466

5467 5468 5469 5470 5471 5472
	ret = nfs4_recover_expired_lease(server);
	if (!ret)
		ret = nfs4_check_client_ready(clp);
	return ret;
}

A
Andy Adamson 已提交
5473 5474 5475 5476 5477 5478 5479 5480 5481 5482 5483 5484 5485 5486 5487 5488 5489 5490 5491 5492 5493
int nfs4_init_ds_session(struct nfs_client *clp)
{
	struct nfs4_session *session = clp->cl_session;
	int ret;

	if (!test_and_clear_bit(NFS4_SESSION_INITING, &session->session_state))
		return 0;

	ret = nfs4_client_recover_expired_lease(clp);
	if (!ret)
		/* Test for the DS role */
		if (!is_ds_client(clp))
			ret = -ENODEV;
	if (!ret)
		ret = nfs4_check_client_ready(clp);
	return ret;

}
EXPORT_SYMBOL_GPL(nfs4_init_ds_session);


A
Andy Adamson 已提交
5494 5495 5496
/*
 * Renew the cl_session lease.
 */
5497 5498 5499 5500 5501 5502
struct nfs4_sequence_data {
	struct nfs_client *clp;
	struct nfs4_sequence_args args;
	struct nfs4_sequence_res res;
};

5503 5504
static void nfs41_sequence_release(void *data)
{
5505 5506
	struct nfs4_sequence_data *calldata = data;
	struct nfs_client *clp = calldata->clp;
5507

5508 5509 5510
	if (atomic_read(&clp->cl_count) > 1)
		nfs4_schedule_state_renewal(clp);
	nfs_put_client(clp);
5511
	kfree(calldata);
5512 5513
}

5514 5515 5516 5517 5518 5519 5520
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:
5521
		nfs4_schedule_lease_recovery(clp);
5522 5523 5524 5525
	}
	return 0;
}

5526
static void nfs41_sequence_call_done(struct rpc_task *task, void *data)
A
Andy Adamson 已提交
5527
{
5528 5529
	struct nfs4_sequence_data *calldata = data;
	struct nfs_client *clp = calldata->clp;
A
Andy Adamson 已提交
5530

5531 5532
	if (!nfs41_sequence_done(task, task->tk_msg.rpc_resp))
		return;
A
Andy Adamson 已提交
5533 5534 5535

	if (task->tk_status < 0) {
		dprintk("%s ERROR %d\n", __func__, task->tk_status);
5536 5537
		if (atomic_read(&clp->cl_count) == 1)
			goto out;
A
Andy Adamson 已提交
5538

5539 5540
		if (nfs41_sequence_handle_errors(task, clp) == -EAGAIN) {
			rpc_restart_call_prepare(task);
A
Andy Adamson 已提交
5541 5542 5543 5544
			return;
		}
	}
	dprintk("%s rpc_cred %p\n", __func__, task->tk_msg.rpc_cred);
5545
out:
A
Andy Adamson 已提交
5546 5547 5548 5549 5550
	dprintk("<-- %s\n", __func__);
}

static void nfs41_sequence_prepare(struct rpc_task *task, void *data)
{
5551 5552
	struct nfs4_sequence_data *calldata = data;
	struct nfs_client *clp = calldata->clp;
A
Andy Adamson 已提交
5553 5554 5555 5556 5557 5558
	struct nfs4_sequence_args *args;
	struct nfs4_sequence_res *res;

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

5559
	if (nfs41_setup_sequence(clp->cl_session, args, res, task))
A
Andy Adamson 已提交
5560 5561 5562 5563 5564 5565 5566
		return;
	rpc_call_start(task);
}

static const struct rpc_call_ops nfs41_sequence_ops = {
	.rpc_call_done = nfs41_sequence_call_done,
	.rpc_call_prepare = nfs41_sequence_prepare,
5567
	.rpc_release = nfs41_sequence_release,
A
Andy Adamson 已提交
5568 5569
};

5570
static struct rpc_task *_nfs41_proc_sequence(struct nfs_client *clp, struct rpc_cred *cred)
A
Andy Adamson 已提交
5571
{
5572
	struct nfs4_sequence_data *calldata;
A
Andy Adamson 已提交
5573 5574 5575 5576
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_SEQUENCE],
		.rpc_cred = cred,
	};
5577 5578 5579 5580 5581 5582
	struct rpc_task_setup task_setup_data = {
		.rpc_client = clp->cl_rpcclient,
		.rpc_message = &msg,
		.callback_ops = &nfs41_sequence_ops,
		.flags = RPC_TASK_ASYNC | RPC_TASK_SOFT,
	};
A
Andy Adamson 已提交
5583

5584
	if (!atomic_inc_not_zero(&clp->cl_count))
5585
		return ERR_PTR(-EIO);
5586
	calldata = kzalloc(sizeof(*calldata), GFP_NOFS);
5587
	if (calldata == NULL) {
5588
		nfs_put_client(clp);
5589
		return ERR_PTR(-ENOMEM);
A
Andy Adamson 已提交
5590
	}
5591
	nfs41_init_sequence(&calldata->args, &calldata->res, 0);
5592 5593 5594
	msg.rpc_argp = &calldata->args;
	msg.rpc_resp = &calldata->res;
	calldata->clp = clp;
5595
	task_setup_data.callback_data = calldata;
A
Andy Adamson 已提交
5596

5597 5598 5599
	return rpc_run_task(&task_setup_data);
}

5600
static int nfs41_proc_async_sequence(struct nfs_client *clp, struct rpc_cred *cred, unsigned renew_flags)
5601 5602 5603 5604
{
	struct rpc_task *task;
	int ret = 0;

5605 5606
	if ((renew_flags & NFS4_RENEW_TIMEOUT) == 0)
		return 0;
5607 5608 5609 5610
	task = _nfs41_proc_sequence(clp, cred);
	if (IS_ERR(task))
		ret = PTR_ERR(task);
	else
5611
		rpc_put_task_async(task);
5612 5613 5614 5615 5616 5617 5618 5619 5620 5621 5622 5623 5624 5625 5626
	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;

	task = _nfs41_proc_sequence(clp, cred);
	if (IS_ERR(task)) {
		ret = PTR_ERR(task);
		goto out;
	}
	ret = rpc_wait_for_completion_task(task);
5627 5628 5629 5630 5631
	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);
5632
		ret = task->tk_status;
5633
	}
5634 5635 5636 5637
	rpc_put_task(task);
out:
	dprintk("<-- %s status=%d\n", __func__, ret);
	return ret;
A
Andy Adamson 已提交
5638 5639
}

5640 5641 5642 5643 5644 5645 5646 5647 5648 5649
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;

5650
	rpc_task_set_priority(task, RPC_PRIORITY_PRIVILEGED);
5651 5652
	if (nfs41_setup_sequence(calldata->clp->cl_session,
				&calldata->arg.seq_args,
5653
				&calldata->res.seq_res, task))
5654 5655 5656 5657 5658
		return;

	rpc_call_start(task);
}

5659 5660 5661 5662 5663 5664 5665 5666 5667
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);
5668 5669
		/* fall through */
	case -NFS4ERR_RETRY_UNCACHED_REP:
5670 5671
		return -EAGAIN;
	default:
5672
		nfs4_schedule_lease_recovery(clp);
5673 5674 5675 5676
	}
	return 0;
}

5677 5678 5679 5680 5681 5682 5683
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__);
5684 5685
	if (!nfs41_sequence_done(task, res))
		return;
5686

5687 5688 5689 5690
	if (nfs41_reclaim_complete_handle_errors(task, clp) == -EAGAIN) {
		rpc_restart_call_prepare(task);
		return;
	}
5691 5692 5693 5694 5695 5696 5697 5698 5699 5700 5701 5702 5703 5704 5705 5706 5707 5708 5709 5710 5711 5712 5713 5714 5715 5716 5717 5718 5719 5720 5721 5722 5723 5724 5725
	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.
 */
static int nfs41_proc_reclaim_complete(struct nfs_client *clp)
{
	struct nfs4_reclaim_complete_data *calldata;
	struct rpc_task *task;
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_RECLAIM_COMPLETE],
	};
	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__);
5726
	calldata = kzalloc(sizeof(*calldata), GFP_NOFS);
5727 5728 5729 5730 5731
	if (calldata == NULL)
		goto out;
	calldata->clp = clp;
	calldata->arg.one_fs = 0;

5732
	nfs41_init_sequence(&calldata->arg.seq_args, &calldata->res.seq_res, 0);
5733 5734 5735 5736
	msg.rpc_argp = &calldata->arg;
	msg.rpc_resp = &calldata->res;
	task_setup_data.callback_data = calldata;
	task = rpc_run_task(&task_setup_data);
5737
	if (IS_ERR(task)) {
5738
		status = PTR_ERR(task);
5739 5740
		goto out;
	}
5741 5742 5743
	status = nfs4_wait_for_completion_rpc_task(task);
	if (status == 0)
		status = task->tk_status;
5744
	rpc_put_task(task);
5745
	return 0;
5746 5747 5748 5749
out:
	dprintk("<-- %s status=%d\n", __func__, status);
	return status;
}
5750 5751 5752 5753 5754

static void
nfs4_layoutget_prepare(struct rpc_task *task, void *calldata)
{
	struct nfs4_layoutget *lgp = calldata;
F
Fred Isaman 已提交
5755
	struct nfs_server *server = NFS_SERVER(lgp->args.inode);
5756 5757

	dprintk("--> %s\n", __func__);
F
Fred Isaman 已提交
5758 5759 5760 5761 5762
	/* 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.
	 */
5763
	if (nfs4_setup_sequence(server, &lgp->args.seq_args,
5764
				&lgp->res.seq_res, task))
5765
		return;
5766 5767 5768 5769 5770 5771
	if (pnfs_choose_layoutget_stateid(&lgp->args.stateid,
					  NFS_I(lgp->args.inode)->layout,
					  lgp->args.ctx->state)) {
		rpc_exit(task, NFS4_OK);
		return;
	}
5772 5773 5774 5775 5776 5777 5778 5779 5780 5781 5782 5783 5784 5785 5786 5787 5788 5789 5790 5791 5792 5793 5794 5795 5796 5797 5798 5799 5800 5801 5802 5803 5804 5805 5806 5807 5808 5809 5810 5811 5812 5813 5814 5815 5816 5817 5818 5819 5820 5821 5822 5823 5824 5825 5826 5827 5828 5829 5830 5831 5832 5833 5834 5835 5836
	rpc_call_start(task);
}

static void nfs4_layoutget_done(struct rpc_task *task, void *calldata)
{
	struct nfs4_layoutget *lgp = calldata;
	struct nfs_server *server = NFS_SERVER(lgp->args.inode);

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

	if (!nfs4_sequence_done(task, &lgp->res.seq_res))
		return;

	switch (task->tk_status) {
	case 0:
		break;
	case -NFS4ERR_LAYOUTTRYLATER:
	case -NFS4ERR_RECALLCONFLICT:
		task->tk_status = -NFS4ERR_DELAY;
		/* Fall through */
	default:
		if (nfs4_async_handle_error(task, server, NULL) == -EAGAIN) {
			rpc_restart_call_prepare(task);
			return;
		}
	}
	dprintk("<-- %s\n", __func__);
}

static void nfs4_layoutget_release(void *calldata)
{
	struct nfs4_layoutget *lgp = calldata;

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

int nfs4_proc_layoutget(struct nfs4_layoutget *lgp)
{
	struct nfs_server *server = NFS_SERVER(lgp->args.inode);
	struct rpc_task *task;
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_LAYOUTGET],
		.rpc_argp = &lgp->args,
		.rpc_resp = &lgp->res,
	};
	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,
	};
	int status = 0;

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

5837
	lgp->res.layoutp = &lgp->args.layout;
5838
	lgp->res.seq_res.sr_slot = NULL;
5839
	nfs41_init_sequence(&lgp->args.seq_args, &lgp->res.seq_res, 0);
5840 5841 5842 5843
	task = rpc_run_task(&task_setup_data);
	if (IS_ERR(task))
		return PTR_ERR(task);
	status = nfs4_wait_for_completion_rpc_task(task);
F
Fred Isaman 已提交
5844 5845 5846 5847
	if (status == 0)
		status = task->tk_status;
	if (status == 0)
		status = pnfs_layout_process(lgp);
5848 5849 5850 5851 5852
	rpc_put_task(task);
	dprintk("<-- %s status=%d\n", __func__, status);
	return status;
}

B
Benny Halevy 已提交
5853 5854 5855 5856 5857 5858 5859
static void
nfs4_layoutreturn_prepare(struct rpc_task *task, void *calldata)
{
	struct nfs4_layoutreturn *lrp = calldata;

	dprintk("--> %s\n", __func__);
	if (nfs41_setup_sequence(lrp->clp->cl_session, &lrp->args.seq_args,
5860
				&lrp->res.seq_res, task))
B
Benny Halevy 已提交
5861 5862 5863 5864 5865 5866 5867 5868
		return;
	rpc_call_start(task);
}

static void nfs4_layoutreturn_done(struct rpc_task *task, void *calldata)
{
	struct nfs4_layoutreturn *lrp = calldata;
	struct nfs_server *server;
T
Trond Myklebust 已提交
5869
	struct pnfs_layout_hdr *lo = lrp->args.layout;
B
Benny Halevy 已提交
5870 5871 5872 5873 5874 5875 5876 5877

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

	if (!nfs4_sequence_done(task, &lrp->res.seq_res))
		return;

	server = NFS_SERVER(lrp->args.inode);
	if (nfs4_async_handle_error(task, server, NULL) == -EAGAIN) {
5878
		rpc_restart_call_prepare(task);
B
Benny Halevy 已提交
5879 5880
		return;
	}
5881
	spin_lock(&lo->plh_inode->i_lock);
B
Benny Halevy 已提交
5882 5883 5884 5885 5886 5887
	if (task->tk_status == 0) {
		if (lrp->res.lrs_present) {
			pnfs_set_layout_stateid(lo, &lrp->res.stateid, true);
		} else
			BUG_ON(!list_empty(&lo->plh_segs));
	}
5888 5889
	lo->plh_block_lgets--;
	spin_unlock(&lo->plh_inode->i_lock);
B
Benny Halevy 已提交
5890 5891 5892 5893 5894 5895 5896 5897
	dprintk("<-- %s\n", __func__);
}

static void nfs4_layoutreturn_release(void *calldata)
{
	struct nfs4_layoutreturn *lrp = calldata;

	dprintk("--> %s\n", __func__);
T
Trond Myklebust 已提交
5898
	put_layout_hdr(lrp->args.layout);
B
Benny Halevy 已提交
5899 5900 5901 5902 5903 5904 5905 5906 5907 5908 5909 5910 5911 5912 5913 5914 5915 5916 5917 5918 5919 5920 5921 5922 5923 5924 5925
	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,
	};
	struct rpc_task_setup task_setup_data = {
		.rpc_client = lrp->clp->cl_rpcclient,
		.rpc_message = &msg,
		.callback_ops = &nfs4_layoutreturn_call_ops,
		.callback_data = lrp,
	};
	int status;

	dprintk("--> %s\n", __func__);
5926
	nfs41_init_sequence(&lrp->args.seq_args, &lrp->res.seq_res, 1);
B
Benny Halevy 已提交
5927 5928 5929 5930 5931 5932 5933 5934 5935
	task = rpc_run_task(&task_setup_data);
	if (IS_ERR(task))
		return PTR_ERR(task);
	status = task->tk_status;
	dprintk("<-- %s status=%d\n", __func__, status);
	rpc_put_task(task);
	return status;
}

A
Andy Adamson 已提交
5936 5937 5938 5939 5940 5941 5942 5943 5944 5945 5946 5947 5948 5949 5950 5951 5952 5953 5954 5955 5956 5957 5958 5959 5960 5961 5962 5963 5964 5965 5966 5967 5968 5969 5970 5971 5972 5973 5974 5975 5976 5977 5978 5979 5980 5981 5982 5983
/*
 * 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);

5984 5985 5986 5987 5988 5989 5990 5991 5992 5993 5994 5995 5996 5997 5998 5999 6000
static int
_nfs4_proc_getdeviceinfo(struct nfs_server *server, struct pnfs_device *pdev)
{
	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,
	};
	int status;

	dprintk("--> %s\n", __func__);
6001
	status = nfs4_call_sync(server->client, server, &msg, &args.seq_args, &res.seq_res, 0);
6002 6003 6004 6005 6006 6007 6008 6009 6010 6011 6012 6013 6014 6015 6016 6017 6018 6019 6020
	dprintk("<-- %s status=%d\n", __func__, status);

	return status;
}

int nfs4_proc_getdeviceinfo(struct nfs_server *server, struct pnfs_device *pdev)
{
	struct nfs4_exception exception = { };
	int err;

	do {
		err = nfs4_handle_exception(server,
					_nfs4_proc_getdeviceinfo(server, pdev),
					&exception);
	} while (exception.retry);
	return err;
}
EXPORT_SYMBOL_GPL(nfs4_proc_getdeviceinfo);

A
Andy Adamson 已提交
6021 6022 6023 6024 6025 6026
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);

	if (nfs4_setup_sequence(server, &data->args.seq_args,
6027
				&data->res.seq_res, task))
A
Andy Adamson 已提交
6028 6029 6030 6031 6032 6033 6034 6035 6036 6037 6038 6039 6040 6041 6042 6043 6044 6045 6046 6047 6048 6049
		return;
	rpc_call_start(task);
}

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);

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

	switch (task->tk_status) { /* Just ignore these failures */
	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 */
		task->tk_status = 0;
	}

	if (nfs4_async_handle_error(task, server, NULL) == -EAGAIN) {
6050
		rpc_restart_call_prepare(task);
A
Andy Adamson 已提交
6051 6052 6053 6054 6055 6056 6057 6058 6059 6060 6061
		return;
	}

	if (task->tk_status == 0)
		nfs_post_op_update_inode_force_wcc(data->args.inode,
						   data->res.fattr);
}

static void nfs4_layoutcommit_release(void *calldata)
{
	struct nfs4_layoutcommit_data *data = calldata;
6062
	struct pnfs_layout_segment *lseg, *tmp;
P
Peng Tao 已提交
6063
	unsigned long *bitlock = &NFS_I(data->args.inode)->flags;
A
Andy Adamson 已提交
6064

A
Andy Adamson 已提交
6065
	pnfs_cleanup_layoutcommit(data);
A
Andy Adamson 已提交
6066
	/* Matched by references in pnfs_set_layoutcommit */
6067 6068 6069 6070 6071 6072
	list_for_each_entry_safe(lseg, tmp, &data->lseg_list, pls_lc_list) {
		list_del_init(&lseg->pls_lc_list);
		if (test_and_clear_bit(NFS_LSEG_LAYOUTCOMMIT,
				       &lseg->pls_flags))
			put_lseg(lseg);
	}
P
Peng Tao 已提交
6073 6074 6075 6076 6077

	clear_bit_unlock(NFS_INO_LAYOUTCOMMITTING, bitlock);
	smp_mb__after_clear_bit();
	wake_up_bit(bitlock, NFS_INO_LAYOUTCOMMITTING);

A
Andy Adamson 已提交
6078 6079 6080 6081 6082 6083 6084 6085 6086 6087 6088
	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
6089
nfs4_proc_layoutcommit(struct nfs4_layoutcommit_data *data, bool sync)
A
Andy Adamson 已提交
6090 6091 6092 6093 6094 6095 6096 6097 6098 6099 6100 6101 6102 6103 6104 6105 6106 6107 6108 6109 6110 6111 6112 6113
{
	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);

6114
	nfs41_init_sequence(&data->args.seq_args, &data->res.seq_res, 1);
A
Andy Adamson 已提交
6115 6116 6117
	task = rpc_run_task(&task_setup_data);
	if (IS_ERR(task))
		return PTR_ERR(task);
6118
	if (sync == false)
A
Andy Adamson 已提交
6119 6120 6121 6122 6123 6124 6125 6126 6127 6128
		goto out;
	status = nfs4_wait_for_completion_rpc_task(task);
	if (status != 0)
		goto out;
	status = task->tk_status;
out:
	dprintk("%s: status %d\n", __func__, status);
	rpc_put_task(task);
	return status;
}
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

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

static int
nfs41_proc_secinfo_no_name(struct nfs_server *server, struct nfs_fh *fhandle,
			   struct nfs_fsinfo *info, struct nfs4_secinfo_flavors *flavors)
{
	struct nfs4_exception exception = { };
	int err;
	do {
		err = _nfs41_proc_secinfo_no_name(server, fhandle, info, flavors);
		switch (err) {
		case 0:
		case -NFS4ERR_WRONGSEC:
		case -NFS4ERR_NOTSUPP:
			break;
		default:
			err = nfs4_handle_exception(server, err, &exception);
		}
	} while (exception.retry);
	return err;
}

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

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

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

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

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

out_freepage:
	put_page(page);
	if (err == -EACCES)
		return -EPERM;
out:
	return err;
}
6208 6209

static int _nfs41_test_stateid(struct nfs_server *server, nfs4_stateid *stateid)
B
Bryan Schumaker 已提交
6210
{
6211
	int status;
B
Bryan Schumaker 已提交
6212
	struct nfs41_test_stateid_args args = {
6213
		.stateid = stateid,
B
Bryan Schumaker 已提交
6214 6215 6216 6217 6218 6219 6220
	};
	struct nfs41_test_stateid_res res;
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_TEST_STATEID],
		.rpc_argp = &args,
		.rpc_resp = &res,
	};
6221

6222
	nfs41_init_sequence(&args.seq_args, &res.seq_res, 0);
6223 6224 6225 6226 6227
	status = nfs4_call_sync_sequence(server->client, server, &msg, &args.seq_args, &res.seq_res, 1);

	if (status == NFS_OK)
		return res.status;
	return status;
B
Bryan Schumaker 已提交
6228 6229
}

6230
static int nfs41_test_stateid(struct nfs_server *server, nfs4_stateid *stateid)
B
Bryan Schumaker 已提交
6231 6232 6233 6234 6235
{
	struct nfs4_exception exception = { };
	int err;
	do {
		err = nfs4_handle_exception(server,
6236
				_nfs41_test_stateid(server, stateid),
B
Bryan Schumaker 已提交
6237 6238 6239 6240
				&exception);
	} while (exception.retry);
	return err;
}
B
Bryan Schumaker 已提交
6241

6242
static int _nfs4_free_stateid(struct nfs_server *server, nfs4_stateid *stateid)
B
Bryan Schumaker 已提交
6243 6244
{
	struct nfs41_free_stateid_args args = {
6245
		.stateid = stateid,
B
Bryan Schumaker 已提交
6246 6247 6248 6249 6250 6251 6252 6253
	};
	struct nfs41_free_stateid_res res;
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_FREE_STATEID],
		.rpc_argp = &args,
		.rpc_resp = &res,
	};

6254 6255
	nfs41_init_sequence(&args.seq_args, &res.seq_res, 0);
	return nfs4_call_sync_sequence(server->client, server, &msg, &args.seq_args, &res.seq_res, 1);
B
Bryan Schumaker 已提交
6256 6257
}

6258
static int nfs41_free_stateid(struct nfs_server *server, nfs4_stateid *stateid)
B
Bryan Schumaker 已提交
6259 6260 6261 6262 6263
{
	struct nfs4_exception exception = { };
	int err;
	do {
		err = nfs4_handle_exception(server,
6264
				_nfs4_free_stateid(server, stateid),
B
Bryan Schumaker 已提交
6265 6266 6267 6268
				&exception);
	} while (exception.retry);
	return err;
}
6269 6270
#endif /* CONFIG_NFS_V4_1 */

6271
struct nfs4_state_recovery_ops nfs40_reboot_recovery_ops = {
6272
	.owner_flag_bit = NFS_OWNER_RECLAIM_REBOOT,
6273
	.state_flag_bit	= NFS_STATE_RECLAIM_REBOOT,
L
Linus Torvalds 已提交
6274 6275
	.recover_open	= nfs4_open_reclaim,
	.recover_lock	= nfs4_lock_reclaim,
6276
	.establish_clid = nfs4_init_clientid,
6277
	.get_clid_cred	= nfs4_get_setclientid_cred,
L
Linus Torvalds 已提交
6278 6279
};

6280 6281 6282 6283 6284 6285
#if defined(CONFIG_NFS_V4_1)
struct nfs4_state_recovery_ops nfs41_reboot_recovery_ops = {
	.owner_flag_bit = NFS_OWNER_RECLAIM_REBOOT,
	.state_flag_bit	= NFS_STATE_RECLAIM_REBOOT,
	.recover_open	= nfs4_open_reclaim,
	.recover_lock	= nfs4_lock_reclaim,
6286
	.establish_clid = nfs41_init_clientid,
6287
	.get_clid_cred	= nfs4_get_exchange_id_cred,
6288
	.reclaim_complete = nfs41_proc_reclaim_complete,
6289 6290 6291 6292 6293 6294 6295 6296 6297
};
#endif /* CONFIG_NFS_V4_1 */

struct nfs4_state_recovery_ops nfs40_nograce_recovery_ops = {
	.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,
6298
	.get_clid_cred	= nfs4_get_setclientid_cred,
6299 6300 6301 6302
};

#if defined(CONFIG_NFS_V4_1)
struct nfs4_state_recovery_ops nfs41_nograce_recovery_ops = {
6303
	.owner_flag_bit = NFS_OWNER_RECLAIM_NOGRACE,
6304
	.state_flag_bit	= NFS_STATE_RECLAIM_NOGRACE,
6305 6306
	.recover_open	= nfs41_open_expired,
	.recover_lock	= nfs41_lock_expired,
6307
	.establish_clid = nfs41_init_clientid,
6308
	.get_clid_cred	= nfs4_get_exchange_id_cred,
L
Linus Torvalds 已提交
6309
};
6310
#endif /* CONFIG_NFS_V4_1 */
L
Linus Torvalds 已提交
6311

B
Benny Halevy 已提交
6312 6313
struct nfs4_state_maintenance_ops nfs40_state_renewal_ops = {
	.sched_state_renewal = nfs4_proc_async_renew,
6314
	.get_state_renewal_cred_locked = nfs4_get_renew_cred_locked,
6315
	.renew_lease = nfs4_proc_renew,
B
Benny Halevy 已提交
6316 6317 6318 6319 6320
};

#if defined(CONFIG_NFS_V4_1)
struct nfs4_state_maintenance_ops nfs41_state_renewal_ops = {
	.sched_state_renewal = nfs41_proc_async_sequence,
6321
	.get_state_renewal_cred_locked = nfs4_get_machine_cred_locked,
6322
	.renew_lease = nfs4_proc_sequence,
B
Benny Halevy 已提交
6323 6324 6325
};
#endif

6326 6327 6328
static const struct nfs4_minor_version_ops nfs_v4_0_minor_ops = {
	.minor_version = 0,
	.call_sync = _nfs4_call_sync,
6329
	.validate_stateid = nfs4_validate_delegation_stateid,
6330
	.find_root_sec = nfs4_find_root_sec,
6331 6332 6333
	.reboot_recovery_ops = &nfs40_reboot_recovery_ops,
	.nograce_recovery_ops = &nfs40_nograce_recovery_ops,
	.state_renewal_ops = &nfs40_state_renewal_ops,
6334 6335 6336 6337 6338 6339
};

#if defined(CONFIG_NFS_V4_1)
static const struct nfs4_minor_version_ops nfs_v4_1_minor_ops = {
	.minor_version = 1,
	.call_sync = _nfs4_call_sync_session,
6340
	.validate_stateid = nfs41_validate_delegation_stateid,
6341
	.find_root_sec = nfs41_find_root_sec,
6342 6343 6344
	.reboot_recovery_ops = &nfs41_reboot_recovery_ops,
	.nograce_recovery_ops = &nfs41_nograce_recovery_ops,
	.state_renewal_ops = &nfs41_state_renewal_ops,
6345 6346 6347 6348 6349 6350 6351 6352 6353 6354
};
#endif

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

6355
static const struct inode_operations nfs4_file_inode_operations = {
6356 6357 6358
	.permission	= nfs_permission,
	.getattr	= nfs_getattr,
	.setattr	= nfs_setattr,
6359 6360 6361 6362
	.getxattr	= generic_getxattr,
	.setxattr	= generic_setxattr,
	.listxattr	= generic_listxattr,
	.removexattr	= generic_removexattr,
6363 6364
};

D
David Howells 已提交
6365
const struct nfs_rpc_ops nfs_v4_clientops = {
L
Linus Torvalds 已提交
6366 6367 6368
	.version	= 4,			/* protocol version */
	.dentry_ops	= &nfs4_dentry_operations,
	.dir_inode_ops	= &nfs4_dir_inode_operations,
6369
	.file_inode_ops	= &nfs4_file_inode_operations,
6370
	.file_ops	= &nfs4_file_operations,
L
Linus Torvalds 已提交
6371 6372 6373 6374 6375 6376 6377 6378 6379 6380 6381
	.getroot	= nfs4_proc_get_root,
	.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,
	.unlink_done	= nfs4_proc_unlink_done,
	.rename		= nfs4_proc_rename,
6382 6383
	.rename_setup	= nfs4_proc_rename_setup,
	.rename_done	= nfs4_proc_rename_done,
L
Linus Torvalds 已提交
6384 6385 6386 6387 6388 6389 6390 6391 6392
	.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,
6393
	.set_capabilities = nfs4_server_capabilities,
L
Linus Torvalds 已提交
6394 6395
	.decode_dirent	= nfs4_decode_dirent,
	.read_setup	= nfs4_proc_read_setup,
T
Trond Myklebust 已提交
6396
	.read_done	= nfs4_read_done,
L
Linus Torvalds 已提交
6397
	.write_setup	= nfs4_proc_write_setup,
6398
	.write_done	= nfs4_write_done,
L
Linus Torvalds 已提交
6399
	.commit_setup	= nfs4_proc_commit_setup,
6400
	.commit_done	= nfs4_commit_done,
L
Linus Torvalds 已提交
6401
	.lock		= nfs4_proc_lock,
6402
	.clear_acl_cache = nfs4_zap_acl_attr,
T
Trond Myklebust 已提交
6403
	.close_context  = nfs4_close_context,
6404
	.open_context	= nfs4_atomic_open,
6405
	.init_client	= nfs4_init_client,
B
Bryan Schumaker 已提交
6406
	.secinfo	= nfs4_proc_secinfo,
L
Linus Torvalds 已提交
6407 6408
};

6409 6410 6411 6412 6413 6414 6415 6416 6417 6418 6419 6420
static const struct xattr_handler nfs4_xattr_nfs4_acl_handler = {
	.prefix	= XATTR_NAME_NFSV4_ACL,
	.list	= nfs4_xattr_list_nfs4_acl,
	.get	= nfs4_xattr_get_nfs4_acl,
	.set	= nfs4_xattr_set_nfs4_acl,
};

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

6421 6422 6423 6424
module_param(max_session_slots, ushort, 0644);
MODULE_PARM_DESC(max_session_slots, "Maximum number of outstanding NFSv4.1 "
		"requests the client will negotiate");

L
Linus Torvalds 已提交
6425 6426 6427 6428 6429
/*
 * Local variables:
 *  c-basic-offset: 8
 * End:
 */