nfs4proc.c 169.3 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
struct nfs4_opendata;
76
static int _nfs4_proc_open(struct nfs4_opendata *data);
77
static int _nfs4_recover_proc_open(struct nfs4_opendata *data);
L
Linus Torvalds 已提交
78
static int nfs4_do_fsinfo(struct nfs_server *, struct nfs_fh *, struct nfs_fsinfo *);
79
static int nfs4_async_handle_error(struct rpc_task *, const struct nfs_server *, struct nfs4_state *);
80
static int _nfs4_proc_getattr(struct nfs_server *server, struct nfs_fh *fhandle, struct nfs_fattr *fattr);
81 82 83
static int nfs4_do_setattr(struct inode *inode, struct rpc_cred *cred,
			    struct nfs_fattr *fattr, struct iattr *sattr,
			    struct nfs4_state *state);
84 85 86 87
#ifdef CONFIG_NFS_V4_1
static int nfs41_test_stateid(struct nfs_server *, struct nfs4_state *);
static int nfs41_free_stateid(struct nfs_server *, struct nfs4_state *);
#endif
L
Linus Torvalds 已提交
88
/* Prevent leaks of NFSv4 errors into userland */
89
static int nfs4_map_errors(int err)
L
Linus Torvalds 已提交
90
{
91 92 93 94 95
	if (err >= -1000)
		return err;
	switch (err) {
	case -NFS4ERR_RESOURCE:
		return -EREMOTEIO;
96 97
	case -NFS4ERR_WRONGSEC:
		return -EPERM;
98 99 100
	case -NFS4ERR_BADOWNER:
	case -NFS4ERR_BADNAME:
		return -EINVAL;
101
	default:
L
Linus Torvalds 已提交
102
		dprintk("%s could not handle NFSv4 error %d\n",
103
				__func__, -err);
104
		break;
L
Linus Torvalds 已提交
105
	}
106
	return -EIO;
L
Linus Torvalds 已提交
107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137
}

/*
 * 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
};

138
const u32 nfs4_pathconf_bitmap[2] = {
L
Linus Torvalds 已提交
139 140 141 142 143
	FATTR4_WORD0_MAXLINK
	| FATTR4_WORD0_MAXNAME,
	0
};

144
const u32 nfs4_fsinfo_bitmap[3] = { FATTR4_WORD0_MAXFILESIZE
L
Linus Torvalds 已提交
145 146 147
			| FATTR4_WORD0_MAXREAD
			| FATTR4_WORD0_MAXWRITE
			| FATTR4_WORD0_LEASE_TIME,
148
			FATTR4_WORD1_TIME_DELTA
149 150
			| FATTR4_WORD1_FS_LAYOUT_TYPES,
			FATTR4_WORD2_LAYOUT_BLKSIZE
L
Linus Torvalds 已提交
151 152
};

153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171
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
172
static void nfs4_setup_readdir(u64 cookie, __be32 *verifier, struct dentry *dentry,
L
Linus Torvalds 已提交
173 174
		struct nfs4_readdir_arg *readdir)
{
175
	__be32 *start, *p;
L
Linus Torvalds 已提交
176 177 178

	BUG_ON(readdir->count < 80);
	if (cookie > 2) {
179
		readdir->cookie = cookie;
L
Linus Torvalds 已提交
180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195
		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.
	 */
196
	start = p = kmap_atomic(*readdir->pages, KM_USER0);
L
Linus Torvalds 已提交
197 198 199 200 201 202 203 204 205 206 207
	
	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
208
		p = xdr_encode_hyper(p, NFS_FILEID(dentry->d_inode));
L
Linus Torvalds 已提交
209 210 211 212 213 214 215 216 217 218 219
	}
	
	*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
220
	p = xdr_encode_hyper(p, NFS_FILEID(dentry->d_parent->d_inode));
L
Linus Torvalds 已提交
221 222 223 224 225 226

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

227 228 229 230 231 232
static int nfs4_wait_clnt_recover(struct nfs_client *clp)
{
	int res;

	might_sleep();

233
	res = wait_on_bit(&clp->cl_state, NFS4CLNT_MANAGER_RUNNING,
234
			nfs_wait_bit_killable, TASK_KILLABLE);
235 236 237 238 239 240 241 242 243 244 245 246 247
	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;
248
	freezable_schedule_timeout_killable(*timeout);
249 250 251 252 253 254 255 256 257
	if (fatal_signal_pending(current))
		res = -ERESTARTSYS;
	*timeout <<= 1;
	return res;
}

/* This is the error handling routine for processes that are allowed
 * to sleep.
 */
258
static int nfs4_handle_exception(struct nfs_server *server, int errorcode, struct nfs4_exception *exception)
259 260
{
	struct nfs_client *clp = server->nfs_client;
261
	struct nfs4_state *state = exception->state;
262 263 264 265 266 267
	int ret = errorcode;

	exception->retry = 0;
	switch(errorcode) {
		case 0:
			return 0;
268 269 270 271 272
		case -NFS4ERR_ADMIN_REVOKED:
		case -NFS4ERR_BAD_STATEID:
		case -NFS4ERR_OPENMODE:
			if (state == NULL)
				break;
273 274
			nfs4_schedule_stateid_recovery(server, state);
			goto wait_on_recovery;
275 276 277
		case -NFS4ERR_EXPIRED:
			if (state != NULL)
				nfs4_schedule_stateid_recovery(server, state);
278
		case -NFS4ERR_STALE_STATEID:
279
		case -NFS4ERR_STALE_CLIENTID:
280 281
			nfs4_schedule_lease_recovery(clp);
			goto wait_on_recovery;
282
#if defined(CONFIG_NFS_V4_1)
283 284 285 286 287 288 289 290 291
		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);
292
			nfs4_schedule_session_recovery(clp->cl_session);
293
			exception->retry = 1;
294
			break;
295
#endif /* defined(CONFIG_NFS_V4_1) */
296
		case -NFS4ERR_FILE_OPEN:
297 298 299 300 301 302 303
			if (exception->timeout > HZ) {
				/* We have retried a decent amount, time to
				 * fail
				 */
				ret = -EBUSY;
				break;
			}
304 305
		case -NFS4ERR_GRACE:
		case -NFS4ERR_DELAY:
306
		case -EKEYEXPIRED:
307 308 309
			ret = nfs4_delay(server->client, &exception->timeout);
			if (ret != 0)
				break;
310
		case -NFS4ERR_RETRY_UNCACHED_REP:
311 312
		case -NFS4ERR_OLD_STATEID:
			exception->retry = 1;
313 314 315 316 317 318 319 320 321 322 323 324 325
			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);
			}
326 327 328
	}
	/* We failed to handle the error */
	return nfs4_map_errors(ret);
329
wait_on_recovery:
330 331 332 333
	ret = nfs4_wait_clnt_recover(clp);
	if (ret == 0)
		exception->retry = 1;
	return ret;
334 335 336
}


337
static void do_renew_lease(struct nfs_client *clp, unsigned long timestamp)
L
Linus Torvalds 已提交
338 339 340 341 342 343 344
{
	spin_lock(&clp->cl_lock);
	if (time_before(clp->cl_last_renewal,timestamp))
		clp->cl_last_renewal = timestamp;
	spin_unlock(&clp->cl_lock);
}

345 346 347 348 349
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
350 351
#if defined(CONFIG_NFS_V4_1)

A
Andy Adamson 已提交
352 353 354 355 356 357 358 359 360 361 362 363
/*
 * 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.
 * If none found, highest_used_slotid is set to -1.
364 365
 *
 * Must be called while holding tbl->slot_tbl_lock
A
Andy Adamson 已提交
366 367
 */
static void
368
nfs4_free_slot(struct nfs4_slot_table *tbl, u8 free_slotid)
A
Andy Adamson 已提交
369 370 371
{
	int slotid = free_slotid;

372
	BUG_ON(slotid < 0 || 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
380 381 382 383 384 385 386 387
			tbl->highest_used_slotid = slotid;
		else
			tbl->highest_used_slotid = -1;
	}
	dprintk("%s: free_slotid %u highest_used_slotid %d\n", __func__,
		free_slotid, tbl->highest_used_slotid);
}

388
/*
389
 * Signal state manager thread if session fore channel is drained
390
 */
391
static void nfs4_check_drain_fc_complete(struct nfs4_session *ses)
392
{
393 394
	struct rpc_task *task;

395
	if (!test_bit(NFS4_SESSION_DRAINING, &ses->session_state)) {
396 397 398
		task = rpc_wake_up_next(&ses->fc_slot_table.slot_tbl_waitq);
		if (task)
			rpc_task_set_priority(task, RPC_PRIORITY_PRIVILEGED);
399 400 401 402 403 404
		return;
	}

	if (ses->fc_slot_table.highest_used_slotid != -1)
		return;

405 406 407 408 409 410 411 412 413 414 415 416 417 418
	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) ||
	    ses->bc_slot_table.highest_used_slotid != -1)
		return;
	dprintk("%s COMPLETE: Session Back Channel Drained\n", __func__);
	complete(&ses->bc_slot_table.complete);
419 420
}

421
static void nfs41_sequence_free_slot(struct nfs4_sequence_res *res)
A
Andy Adamson 已提交
422 423 424
{
	struct nfs4_slot_table *tbl;

425
	tbl = &res->sr_session->fc_slot_table;
426
	if (!res->sr_slot) {
A
Andy Adamson 已提交
427 428
		/* just wake up the next guy waiting since
		 * we may have not consumed a slot after all */
A
Andy Adamson 已提交
429
		dprintk("%s: No slot\n", __func__);
430
		return;
A
Andy Adamson 已提交
431
	}
432

433
	spin_lock(&tbl->slot_tbl_lock);
434
	nfs4_free_slot(tbl, res->sr_slot - tbl->slots);
435
	nfs4_check_drain_fc_complete(res->sr_session);
436
	spin_unlock(&tbl->slot_tbl_lock);
437
	res->sr_slot = NULL;
A
Andy Adamson 已提交
438 439
}

440
static int nfs41_sequence_done(struct rpc_task *task, struct nfs4_sequence_res *res)
A
Andy Adamson 已提交
441 442
{
	unsigned long timestamp;
443
	struct nfs_client *clp;
A
Andy Adamson 已提交
444 445 446 447 448 449 450 451 452 453

	/*
	 * 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;

454 455
	/* don't increment the sequence number if the task wasn't sent */
	if (!RPC_WAS_SENT(task))
A
Andy Adamson 已提交
456 457
		goto out;

A
Andy Adamson 已提交
458
	/* Check the SEQUENCE operation status */
459 460
	switch (res->sr_status) {
	case 0:
A
Andy Adamson 已提交
461
		/* Update the slot's sequence and clientid lease timer */
462
		++res->sr_slot->seq_nr;
A
Andy Adamson 已提交
463
		timestamp = res->sr_renewal_time;
464
		clp = res->sr_session->clp;
465
		do_renew_lease(clp, timestamp);
466
		/* Check sequence flags */
467 468
		if (res->sr_status_flags != 0)
			nfs4_schedule_lease_recovery(clp);
469 470 471 472 473 474
		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.
		 */
475
		dprintk("%s: slot=%td seq=%d: Operation in progress\n",
476 477 478
			__func__,
			res->sr_slot - res->sr_session->fc_slot_table.slots,
			res->sr_slot->seq_nr);
479 480 481
		goto out_retry;
	default:
		/* Just update the slot sequence no. */
482
		++res->sr_slot->seq_nr;
A
Andy Adamson 已提交
483 484 485 486
	}
out:
	/* The session may be reset by one of the error handlers. */
	dprintk("%s: Error %d free the slot \n", __func__, res->sr_status);
487
	nfs41_sequence_free_slot(res);
488 489
	return 1;
out_retry:
490
	if (!rpc_restart_call(task))
491 492 493
		goto out;
	rpc_delay(task, NFS4_POLL_RETRY_MAX);
	return 0;
A
Andy Adamson 已提交
494 495
}

496 497
static int nfs4_sequence_done(struct rpc_task *task,
			       struct nfs4_sequence_res *res)
498
{
499 500 501
	if (res->sr_session == NULL)
		return 1;
	return nfs41_sequence_done(task, res);
502 503
}

B
Benny Halevy 已提交
504 505 506 507 508 509 510
/*
 * 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.
 * The slot number is returned if found, or NFS4_MAX_SLOT_TABLE otherwise.
A
Andy Adamson 已提交
511 512
 *
 * Note: must be called with under the slot_tbl_lock.
B
Benny Halevy 已提交
513 514
 */
static u8
515
nfs4_find_slot(struct nfs4_slot_table *tbl)
B
Benny Halevy 已提交
516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536
{
	int slotid;
	u8 ret_id = NFS4_MAX_SLOT_TABLE;
	BUILD_BUG_ON((u8)NFS4_MAX_SLOT_TABLE != (int)NFS4_MAX_SLOT_TABLE);

	dprintk("--> %s used_slots=%04lx highest_used=%d max_slots=%d\n",
		__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);
	if (slotid > tbl->highest_used_slotid)
		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;
}

A
Andy Adamson 已提交
537
int nfs41_setup_sequence(struct nfs4_session *session,
A
Andy Adamson 已提交
538 539 540 541 542
				struct nfs4_sequence_args *args,
				struct nfs4_sequence_res *res,
				int cache_reply,
				struct rpc_task *task)
{
A
Andy Adamson 已提交
543 544 545 546 547
	struct nfs4_slot *slot;
	struct nfs4_slot_table *tbl;
	u8 slotid;

	dprintk("--> %s\n", __func__);
A
Andy Adamson 已提交
548
	/* slot already allocated? */
549
	if (res->sr_slot != NULL)
A
Andy Adamson 已提交
550 551
		return 0;

A
Andy Adamson 已提交
552 553 554
	tbl = &session->fc_slot_table;

	spin_lock(&tbl->slot_tbl_lock);
555
	if (test_bit(NFS4_SESSION_DRAINING, &session->session_state) &&
556
	    !rpc_task_has_priority(task, RPC_PRIORITY_PRIVILEGED)) {
557
		/* The state manager will wait until the slot table is empty */
A
Andy Adamson 已提交
558
		rpc_sleep_on(&tbl->slot_tbl_waitq, task, NULL);
559
		spin_unlock(&tbl->slot_tbl_lock);
560
		dprintk("%s session is draining\n", __func__);
A
Andy Adamson 已提交
561
		return -EAGAIN;
562 563
	}

564 565 566 567 568 569 570 571
	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;
	}

572
	slotid = nfs4_find_slot(tbl);
A
Andy Adamson 已提交
573 574 575 576 577 578 579 580
	if (slotid == NFS4_MAX_SLOT_TABLE) {
		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);

581
	rpc_task_set_priority(task, RPC_PRIORITY_NORMAL);
A
Andy Adamson 已提交
582
	slot = tbl->slots + slotid;
B
Benny Halevy 已提交
583
	args->sa_session = session;
A
Andy Adamson 已提交
584 585 586 587 588
	args->sa_slotid = slotid;
	args->sa_cache_this = cache_reply;

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

B
Benny Halevy 已提交
589
	res->sr_session = session;
590
	res->sr_slot = slot;
A
Andy Adamson 已提交
591
	res->sr_renewal_time = jiffies;
592
	res->sr_status_flags = 0;
A
Andy Adamson 已提交
593 594 595 596 597
	/*
	 * 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
598 599
	return 0;
}
A
Andy Adamson 已提交
600
EXPORT_SYMBOL_GPL(nfs41_setup_sequence);
A
Andy Adamson 已提交
601

602
int nfs4_setup_sequence(const struct nfs_server *server,
A
Andy Adamson 已提交
603 604 605 606 607
			struct nfs4_sequence_args *args,
			struct nfs4_sequence_res *res,
			int cache_reply,
			struct rpc_task *task)
{
608
	struct nfs4_session *session = nfs4_get_session(server);
A
Andy Adamson 已提交
609 610
	int ret = 0;

611 612 613 614 615 616
	if (session == NULL) {
		args->sa_session = NULL;
		res->sr_session = NULL;
		goto out;
	}

617
	dprintk("--> %s clp %p session %p sr_slot %td\n",
618 619
		__func__, session->clp, session, res->sr_slot ?
			res->sr_slot - session->fc_slot_table.slots : -1);
A
Andy Adamson 已提交
620

621
	ret = nfs41_setup_sequence(session, args, res, cache_reply,
A
Andy Adamson 已提交
622 623 624 625 626 627 628
				   task);
out:
	dprintk("<-- %s status=%d\n", __func__, ret);
	return ret;
}

struct nfs41_call_sync_data {
629
	const struct nfs_server *seq_server;
A
Andy Adamson 已提交
630 631 632 633 634 635 636 637 638
	struct nfs4_sequence_args *seq_args;
	struct nfs4_sequence_res *seq_res;
	int cache_reply;
};

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

639 640 641
	dprintk("--> %s data->seq_server %p\n", __func__, data->seq_server);

	if (nfs4_setup_sequence(data->seq_server, data->seq_args,
A
Andy Adamson 已提交
642 643 644 645 646
				data->seq_res, data->cache_reply, task))
		return;
	rpc_call_start(task);
}

647 648 649 650 651 652
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
653 654 655 656
static void nfs41_call_sync_done(struct rpc_task *task, void *calldata)
{
	struct nfs41_call_sync_data *data = calldata;

657
	nfs41_sequence_done(task, data->seq_res);
A
Andy Adamson 已提交
658 659
}

A
Andy Adamson 已提交
660 661
struct rpc_call_ops nfs41_call_sync_ops = {
	.rpc_call_prepare = nfs41_call_sync_prepare,
A
Andy Adamson 已提交
662
	.rpc_call_done = nfs41_call_sync_done,
A
Andy Adamson 已提交
663 664
};

665 666 667 668 669
struct rpc_call_ops nfs41_call_priv_sync_ops = {
	.rpc_call_prepare = nfs41_call_priv_sync_prepare,
	.rpc_call_done = nfs41_call_sync_done,
};

670 671
static int nfs4_call_sync_sequence(struct rpc_clnt *clnt,
				   struct nfs_server *server,
A
Andy Adamson 已提交
672 673 674
				   struct rpc_message *msg,
				   struct nfs4_sequence_args *args,
				   struct nfs4_sequence_res *res,
675 676
				   int cache_reply,
				   int privileged)
A
Andy Adamson 已提交
677 678 679 680
{
	int ret;
	struct rpc_task *task;
	struct nfs41_call_sync_data data = {
681
		.seq_server = server,
A
Andy Adamson 已提交
682 683 684 685 686
		.seq_args = args,
		.seq_res = res,
		.cache_reply = cache_reply,
	};
	struct rpc_task_setup task_setup = {
687
		.rpc_client = clnt,
A
Andy Adamson 已提交
688 689 690 691 692
		.rpc_message = msg,
		.callback_ops = &nfs41_call_sync_ops,
		.callback_data = &data
	};

693
	res->sr_slot = NULL;
694 695
	if (privileged)
		task_setup.callback_ops = &nfs41_call_priv_sync_ops;
A
Andy Adamson 已提交
696 697 698 699 700 701 702 703 704 705
	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;
}

706 707
int _nfs4_call_sync_session(struct rpc_clnt *clnt,
			    struct nfs_server *server,
A
Andy Adamson 已提交
708 709 710 711 712
			    struct rpc_message *msg,
			    struct nfs4_sequence_args *args,
			    struct nfs4_sequence_res *res,
			    int cache_reply)
{
713
	return nfs4_call_sync_sequence(clnt, server, msg, args, res, cache_reply, 0);
A
Andy Adamson 已提交
714 715
}

716
#else
717 718
static int nfs4_sequence_done(struct rpc_task *task,
			       struct nfs4_sequence_res *res)
719
{
720
	return 1;
721
}
A
Andy Adamson 已提交
722 723
#endif /* CONFIG_NFS_V4_1 */

724 725
int _nfs4_call_sync(struct rpc_clnt *clnt,
		    struct nfs_server *server,
A
Andy Adamson 已提交
726 727 728 729 730
		    struct rpc_message *msg,
		    struct nfs4_sequence_args *args,
		    struct nfs4_sequence_res *res,
		    int cache_reply)
{
731
	args->sa_session = res->sr_session = NULL;
732
	return rpc_call_sync(clnt, msg, 0);
A
Andy Adamson 已提交
733 734
}

735
static inline
736 737
int nfs4_call_sync(struct rpc_clnt *clnt,
		   struct nfs_server *server,
738 739 740 741 742
		   struct rpc_message *msg,
		   struct nfs4_sequence_args *args,
		   struct nfs4_sequence_res *res,
		   int cache_reply)
{
743 744
	return server->nfs_client->cl_mvops->call_sync(clnt, server, msg,
						args, res, cache_reply);
745
}
A
Andy Adamson 已提交
746

747
static void update_changeattr(struct inode *dir, struct nfs4_change_info *cinfo)
L
Linus Torvalds 已提交
748
{
749
	struct nfs_inode *nfsi = NFS_I(dir);
L
Linus Torvalds 已提交
750

751
	spin_lock(&dir->i_lock);
752
	nfsi->cache_validity |= NFS_INO_INVALID_ATTR|NFS_INO_REVAL_PAGECACHE|NFS_INO_INVALID_DATA;
753
	if (!cinfo->atomic || cinfo->before != dir->i_version)
754
		nfs_force_lookup_revalidate(dir);
755
	dir->i_version = cinfo->after;
756
	spin_unlock(&dir->i_lock);
L
Linus Torvalds 已提交
757 758
}

759
struct nfs4_opendata {
760
	struct kref kref;
761 762
	struct nfs_openargs o_arg;
	struct nfs_openres o_res;
763 764
	struct nfs_open_confirmargs c_arg;
	struct nfs_open_confirmres c_res;
765 766
	struct nfs4_string owner_name;
	struct nfs4_string group_name;
767 768 769
	struct nfs_fattr f_attr;
	struct nfs_fattr dir_attr;
	struct dentry *dir;
770
	struct dentry *dentry;
771
	struct nfs4_state_owner *owner;
772
	struct nfs4_state *state;
773
	struct iattr attrs;
774
	unsigned long timestamp;
775
	unsigned int rpc_done : 1;
776 777
	int rpc_status;
	int cancelled;
778 779
};

780 781 782 783 784

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;
785 786
	p->o_res.seqid = p->o_arg.seqid;
	p->c_res.seqid = p->c_arg.seqid;
787 788 789
	p->o_res.server = p->o_arg.server;
	nfs_fattr_init(&p->f_attr);
	nfs_fattr_init(&p->dir_attr);
790
	nfs_fattr_init_names(&p->f_attr, &p->owner_name, &p->group_name);
791 792
}

793
static struct nfs4_opendata *nfs4_opendata_alloc(struct dentry *dentry,
794
		struct nfs4_state_owner *sp, fmode_t fmode, int flags,
795 796
		const struct iattr *attrs,
		gfp_t gfp_mask)
797
{
798
	struct dentry *parent = dget_parent(dentry);
799 800 801 802
	struct inode *dir = parent->d_inode;
	struct nfs_server *server = NFS_SERVER(dir);
	struct nfs4_opendata *p;

803
	p = kzalloc(sizeof(*p), gfp_mask);
804 805
	if (p == NULL)
		goto err;
806
	p->o_arg.seqid = nfs_alloc_seqid(&sp->so_seqid, gfp_mask);
807 808
	if (p->o_arg.seqid == NULL)
		goto err_free;
809 810
	nfs_sb_active(dentry->d_sb);
	p->dentry = dget(dentry);
811 812 813 814
	p->dir = parent;
	p->owner = sp;
	atomic_inc(&sp->so_count);
	p->o_arg.fh = NFS_FH(dir);
815 816
	p->o_arg.open_flags = flags;
	p->o_arg.fmode = fmode & (FMODE_READ|FMODE_WRITE);
817
	p->o_arg.clientid = server->nfs_client->cl_clientid;
818
	p->o_arg.id = sp->so_owner_id;
819
	p->o_arg.name = &dentry->d_name;
820 821
	p->o_arg.server = server;
	p->o_arg.bitmask = server->attr_bitmask;
822
	p->o_arg.dir_bitmask = server->cache_consistency_bitmask;
823
	p->o_arg.claim = NFS4_OPEN_CLAIM_NULL;
824 825 826
	if (flags & O_CREAT) {
		u32 *s;

827 828
		p->o_arg.u.attrs = &p->attrs;
		memcpy(&p->attrs, attrs, sizeof(p->attrs));
829 830 831
		s = (u32 *) p->o_arg.u.verifier.data;
		s[0] = jiffies;
		s[1] = current->pid;
832
	}
833 834 835
	p->c_arg.fh = &p->o_res.fh;
	p->c_arg.stateid = &p->o_res.stateid;
	p->c_arg.seqid = p->o_arg.seqid;
836
	nfs4_init_opendata_res(p);
837
	kref_init(&p->kref);
838 839 840 841 842 843 844 845
	return p;
err_free:
	kfree(p);
err:
	dput(parent);
	return NULL;
}

846
static void nfs4_opendata_free(struct kref *kref)
847
{
848 849
	struct nfs4_opendata *p = container_of(kref,
			struct nfs4_opendata, kref);
850
	struct super_block *sb = p->dentry->d_sb;
851 852

	nfs_free_seqid(p->o_arg.seqid);
853 854
	if (p->state != NULL)
		nfs4_put_open_state(p->state);
855 856
	nfs4_put_state_owner(p->owner);
	dput(p->dir);
857 858
	dput(p->dentry);
	nfs_sb_deactive(sb);
859
	nfs_fattr_free_names(&p->f_attr);
860 861 862 863 864 865 866
	kfree(p);
}

static void nfs4_opendata_put(struct nfs4_opendata *p)
{
	if (p != NULL)
		kref_put(&p->kref, nfs4_opendata_free);
867 868
}

869 870 871 872 873 874 875 876
static int nfs4_wait_for_completion_rpc_task(struct rpc_task *task)
{
	int ret;

	ret = rpc_wait_for_completion_task(task);
	return ret;
}

877
static int can_open_cached(struct nfs4_state *state, fmode_t mode, int open_mode)
878 879
{
	int ret = 0;
880 881 882 883

	if (open_mode & O_EXCL)
		goto out;
	switch (mode & (FMODE_READ|FMODE_WRITE)) {
884
		case FMODE_READ:
885 886
			ret |= test_bit(NFS_O_RDONLY_STATE, &state->flags) != 0
				&& state->n_rdonly != 0;
887 888
			break;
		case FMODE_WRITE:
889 890
			ret |= test_bit(NFS_O_WRONLY_STATE, &state->flags) != 0
				&& state->n_wronly != 0;
891 892
			break;
		case FMODE_READ|FMODE_WRITE:
893 894
			ret |= test_bit(NFS_O_RDWR_STATE, &state->flags) != 0
				&& state->n_rdwr != 0;
895
	}
896
out:
897 898 899
	return ret;
}

900
static int can_open_delegated(struct nfs_delegation *delegation, fmode_t fmode)
901
{
902 903
	if (delegation == NULL)
		return 0;
904
	if ((delegation->type & fmode) != fmode)
905
		return 0;
906
	if (test_bit(NFS_DELEGATION_NEED_RECLAIM, &delegation->flags))
907
		return 0;
908
	nfs_mark_delegation_referenced(delegation);
909 910 911
	return 1;
}

912
static void update_open_stateflags(struct nfs4_state *state, fmode_t fmode)
913
{
914
	switch (fmode) {
915 916 917 918 919 920 921 922 923
		case FMODE_WRITE:
			state->n_wronly++;
			break;
		case FMODE_READ:
			state->n_rdonly++;
			break;
		case FMODE_READ|FMODE_WRITE:
			state->n_rdwr++;
	}
924
	nfs4_state_set_mode_locked(state, state->state | fmode);
925 926
}

927
static void nfs_set_open_stateid_locked(struct nfs4_state *state, nfs4_stateid *stateid, fmode_t fmode)
928 929 930 931
{
	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));
932
	switch (fmode) {
933 934 935 936 937 938 939 940 941
		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);
	}
942 943
}

944
static void nfs_set_open_stateid(struct nfs4_state *state, nfs4_stateid *stateid, fmode_t fmode)
945
{
946
	write_seqlock(&state->seqlock);
947
	nfs_set_open_stateid_locked(state, stateid, fmode);
948
	write_sequnlock(&state->seqlock);
949 950
}

951
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
952
{
953 954 955 956 957
	/*
	 * Protect the call to nfs4_state_set_mode_locked and
	 * serialise the stateid update
	 */
	write_seqlock(&state->seqlock);
958 959 960 961 962
	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)
963
		nfs_set_open_stateid_locked(state, open_stateid, fmode);
964 965
	write_sequnlock(&state->seqlock);
	spin_lock(&state->owner->so_lock);
966
	update_open_stateflags(state, fmode);
967
	spin_unlock(&state->owner->so_lock);
L
Linus Torvalds 已提交
968 969
}

970
static int update_open_stateid(struct nfs4_state *state, nfs4_stateid *open_stateid, nfs4_stateid *delegation, fmode_t fmode)
971 972 973 974 975
{
	struct nfs_inode *nfsi = NFS_I(state->inode);
	struct nfs_delegation *deleg_cur;
	int ret = 0;

976
	fmode &= (FMODE_READ|FMODE_WRITE);
977 978 979 980 981 982 983 984

	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 ||
985
	    (deleg_cur->type & fmode) != fmode)
986 987 988 989 990 991 992
		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;

993
	nfs_mark_delegation_referenced(deleg_cur);
994
	__update_open_stateid(state, open_stateid, &deleg_cur->stateid, fmode);
995 996 997 998 999 1000 1001
	ret = 1;
no_delegation_unlock:
	spin_unlock(&deleg_cur->lock);
no_delegation:
	rcu_read_unlock();

	if (!ret && open_stateid != NULL) {
1002
		__update_open_stateid(state, open_stateid, NULL, fmode);
1003 1004 1005 1006 1007 1008 1009
		ret = 1;
	}

	return ret;
}


1010
static void nfs4_return_incompatible_delegation(struct inode *inode, fmode_t fmode)
1011 1012 1013 1014 1015
{
	struct nfs_delegation *delegation;

	rcu_read_lock();
	delegation = rcu_dereference(NFS_I(inode)->delegation);
1016
	if (delegation == NULL || (delegation->type & fmode) == fmode) {
1017 1018 1019 1020 1021 1022 1023
		rcu_read_unlock();
		return;
	}
	rcu_read_unlock();
	nfs_inode_return_delegation(inode);
}

1024
static struct nfs4_state *nfs4_try_open_cached(struct nfs4_opendata *opendata)
1025 1026 1027 1028
{
	struct nfs4_state *state = opendata->state;
	struct nfs_inode *nfsi = NFS_I(state->inode);
	struct nfs_delegation *delegation;
1029 1030
	int open_mode = opendata->o_arg.open_flags & O_EXCL;
	fmode_t fmode = opendata->o_arg.fmode;
1031 1032 1033 1034
	nfs4_stateid stateid;
	int ret = -EAGAIN;

	for (;;) {
1035
		if (can_open_cached(state, fmode, open_mode)) {
1036
			spin_lock(&state->owner->so_lock);
1037 1038
			if (can_open_cached(state, fmode, open_mode)) {
				update_open_stateflags(state, fmode);
1039 1040 1041 1042 1043
				spin_unlock(&state->owner->so_lock);
				goto out_return_state;
			}
			spin_unlock(&state->owner->so_lock);
		}
1044 1045
		rcu_read_lock();
		delegation = rcu_dereference(nfsi->delegation);
1046
		if (!can_open_delegated(delegation, fmode)) {
1047
			rcu_read_unlock();
1048
			break;
1049
		}
1050 1051 1052
		/* Save the delegation */
		memcpy(stateid.data, delegation->stateid.data, sizeof(stateid.data));
		rcu_read_unlock();
1053
		ret = nfs_may_open(state->inode, state->owner->so_cred, open_mode);
1054 1055 1056
		if (ret != 0)
			goto out;
		ret = -EAGAIN;
1057 1058

		/* Try to update the stateid using the delegation */
1059
		if (update_open_stateid(state, NULL, &stateid, fmode))
1060
			goto out_return_state;
1061 1062 1063 1064 1065 1066 1067 1068
	}
out:
	return ERR_PTR(ret);
out_return_state:
	atomic_inc(&state->count);
	return state;
}

1069 1070 1071 1072
static struct nfs4_state *nfs4_opendata_to_nfs4_state(struct nfs4_opendata *data)
{
	struct inode *inode;
	struct nfs4_state *state = NULL;
1073
	struct nfs_delegation *delegation;
1074
	int ret;
1075

1076
	if (!data->rpc_done) {
1077
		state = nfs4_try_open_cached(data);
1078 1079 1080
		goto out;
	}

1081
	ret = -EAGAIN;
1082
	if (!(data->f_attr.valid & NFS_ATTR_FATTR))
1083
		goto err;
1084
	inode = nfs_fhget(data->dir->d_sb, &data->o_res.fh, &data->f_attr);
1085
	ret = PTR_ERR(inode);
1086
	if (IS_ERR(inode))
1087 1088
		goto err;
	ret = -ENOMEM;
1089 1090
	state = nfs4_get_open_state(inode, data->owner);
	if (state == NULL)
1091
		goto err_put_inode;
1092 1093 1094
	if (data->o_res.delegation_type != 0) {
		int delegation_flags = 0;

1095 1096 1097 1098 1099
		rcu_read_lock();
		delegation = rcu_dereference(NFS_I(inode)->delegation);
		if (delegation)
			delegation_flags = delegation->flags;
		rcu_read_unlock();
1100 1101 1102 1103 1104 1105
		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)
1106 1107 1108 1109 1110 1111 1112 1113
			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);
	}
1114 1115

	update_open_stateid(state, &data->o_res.stateid, NULL,
1116
			data->o_arg.fmode);
1117
	iput(inode);
1118
out:
1119
	return state;
1120 1121 1122 1123
err_put_inode:
	iput(inode);
err:
	return ERR_PTR(ret);
1124 1125
}

1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142
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
1143 1144 1145 1146
static struct nfs4_opendata *nfs4_open_recoverdata_alloc(struct nfs_open_context *ctx, struct nfs4_state *state)
{
	struct nfs4_opendata *opendata;

1147
	opendata = nfs4_opendata_alloc(ctx->dentry, state->owner, 0, 0, NULL, GFP_NOFS);
T
Trond Myklebust 已提交
1148 1149 1150 1151 1152 1153 1154
	if (opendata == NULL)
		return ERR_PTR(-ENOMEM);
	opendata->state = state;
	atomic_inc(&state->count);
	return opendata;
}

1155
static int nfs4_open_recover_helper(struct nfs4_opendata *opendata, fmode_t fmode, struct nfs4_state **res)
1156
{
1157
	struct nfs4_state *newstate;
1158 1159
	int ret;

1160 1161
	opendata->o_arg.open_flags = 0;
	opendata->o_arg.fmode = fmode;
1162 1163 1164
	memset(&opendata->o_res, 0, sizeof(opendata->o_res));
	memset(&opendata->c_res, 0, sizeof(opendata->c_res));
	nfs4_init_opendata_res(opendata);
1165
	ret = _nfs4_recover_proc_open(opendata);
1166 1167
	if (ret != 0)
		return ret; 
1168
	newstate = nfs4_opendata_to_nfs4_state(opendata);
1169 1170
	if (IS_ERR(newstate))
		return PTR_ERR(newstate);
1171
	nfs4_close_state(newstate, fmode);
1172
	*res = newstate;
1173 1174 1175 1176 1177 1178 1179 1180 1181
	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_* */
1182
	clear_bit(NFS_DELEGATED_STATE, &state->flags);
1183 1184
	smp_rmb();
	if (state->n_rdwr != 0) {
T
Trond Myklebust 已提交
1185
		clear_bit(NFS_O_RDWR_STATE, &state->flags);
1186
		ret = nfs4_open_recover_helper(opendata, FMODE_READ|FMODE_WRITE, &newstate);
1187 1188
		if (ret != 0)
			return ret;
1189 1190
		if (newstate != state)
			return -ESTALE;
1191 1192
	}
	if (state->n_wronly != 0) {
T
Trond Myklebust 已提交
1193
		clear_bit(NFS_O_WRONLY_STATE, &state->flags);
1194
		ret = nfs4_open_recover_helper(opendata, FMODE_WRITE, &newstate);
1195 1196
		if (ret != 0)
			return ret;
1197 1198
		if (newstate != state)
			return -ESTALE;
1199 1200
	}
	if (state->n_rdonly != 0) {
T
Trond Myklebust 已提交
1201
		clear_bit(NFS_O_RDONLY_STATE, &state->flags);
1202
		ret = nfs4_open_recover_helper(opendata, FMODE_READ, &newstate);
1203 1204
		if (ret != 0)
			return ret;
1205 1206
		if (newstate != state)
			return -ESTALE;
1207
	}
1208 1209 1210 1211 1212 1213
	/*
	 * 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) {
1214
		write_seqlock(&state->seqlock);
1215 1216
		if (test_bit(NFS_DELEGATED_STATE, &state->flags) == 0)
			memcpy(state->stateid.data, state->open_stateid.data, sizeof(state->stateid.data));
1217
		write_sequnlock(&state->seqlock);
1218
	}
1219 1220 1221
	return 0;
}

L
Linus Torvalds 已提交
1222 1223 1224 1225
/*
 * OPEN_RECLAIM:
 * 	reclaim state on the server after a reboot.
 */
1226
static int _nfs4_do_open_reclaim(struct nfs_open_context *ctx, struct nfs4_state *state)
L
Linus Torvalds 已提交
1227
{
1228
	struct nfs_delegation *delegation;
1229
	struct nfs4_opendata *opendata;
1230
	fmode_t delegation_type = 0;
L
Linus Torvalds 已提交
1231 1232
	int status;

T
Trond Myklebust 已提交
1233 1234 1235
	opendata = nfs4_open_recoverdata_alloc(ctx, state);
	if (IS_ERR(opendata))
		return PTR_ERR(opendata);
1236 1237
	opendata->o_arg.claim = NFS4_OPEN_CLAIM_PREVIOUS;
	opendata->o_arg.fh = NFS_FH(state->inode);
1238 1239
	rcu_read_lock();
	delegation = rcu_dereference(NFS_I(state->inode)->delegation);
1240
	if (delegation != NULL && test_bit(NFS_DELEGATION_NEED_RECLAIM, &delegation->flags) != 0)
1241
		delegation_type = delegation->type;
1242
	rcu_read_unlock();
1243 1244
	opendata->o_arg.u.delegation_type = delegation_type;
	status = nfs4_open_recover(opendata, state);
1245
	nfs4_opendata_put(opendata);
L
Linus Torvalds 已提交
1246 1247 1248
	return status;
}

1249
static int nfs4_do_open_reclaim(struct nfs_open_context *ctx, struct nfs4_state *state)
L
Linus Torvalds 已提交
1250 1251 1252 1253 1254
{
	struct nfs_server *server = NFS_SERVER(state->inode);
	struct nfs4_exception exception = { };
	int err;
	do {
1255
		err = _nfs4_do_open_reclaim(ctx, state);
1256
		if (err != -NFS4ERR_DELAY)
1257 1258
			break;
		nfs4_handle_exception(server, err, &exception);
L
Linus Torvalds 已提交
1259 1260 1261 1262
	} while (exception.retry);
	return err;
}

1263 1264 1265 1266 1267 1268 1269 1270
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);
1271
	ret = nfs4_do_open_reclaim(ctx, state);
1272 1273 1274 1275
	put_nfs_open_context(ctx);
	return ret;
}

1276
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
1277
{
1278
	struct nfs4_opendata *opendata;
1279
	int ret;
L
Linus Torvalds 已提交
1280

T
Trond Myklebust 已提交
1281 1282 1283
	opendata = nfs4_open_recoverdata_alloc(ctx, state);
	if (IS_ERR(opendata))
		return PTR_ERR(opendata);
1284
	opendata->o_arg.claim = NFS4_OPEN_CLAIM_DELEGATE_CUR;
1285
	memcpy(opendata->o_arg.u.delegation.data, stateid->data,
1286
			sizeof(opendata->o_arg.u.delegation.data));
1287
	ret = nfs4_open_recover(opendata, state);
1288
	nfs4_opendata_put(opendata);
1289
	return ret;
L
Linus Torvalds 已提交
1290 1291
}

1292
int nfs4_open_delegation_recall(struct nfs_open_context *ctx, struct nfs4_state *state, const nfs4_stateid *stateid)
L
Linus Torvalds 已提交
1293 1294
{
	struct nfs4_exception exception = { };
1295
	struct nfs_server *server = NFS_SERVER(state->inode);
L
Linus Torvalds 已提交
1296 1297
	int err;
	do {
1298
		err = _nfs4_open_delegation_recall(ctx, state, stateid);
L
Linus Torvalds 已提交
1299 1300
		switch (err) {
			case 0:
1301 1302 1303
			case -ENOENT:
			case -ESTALE:
				goto out;
1304 1305 1306 1307 1308
			case -NFS4ERR_BADSESSION:
			case -NFS4ERR_BADSLOT:
			case -NFS4ERR_BAD_HIGH_SLOT:
			case -NFS4ERR_CONN_NOT_BOUND_TO_SESSION:
			case -NFS4ERR_DEADSESSION:
1309
				nfs4_schedule_session_recovery(server->nfs_client->cl_session);
1310
				goto out;
L
Linus Torvalds 已提交
1311 1312 1313 1314
			case -NFS4ERR_STALE_CLIENTID:
			case -NFS4ERR_STALE_STATEID:
			case -NFS4ERR_EXPIRED:
				/* Don't recall a delegation if it was lost */
1315
				nfs4_schedule_lease_recovery(server->nfs_client);
1316 1317 1318 1319 1320 1321 1322 1323
				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:
1324
				nfs4_schedule_stateid_recovery(server, state);
1325 1326 1327 1328 1329 1330 1331
			case -EKEYEXPIRED:
				/*
				 * User RPCSEC_GSS context has expired.
				 * We cannot recover this stateid now, so
				 * skip it and allow recovery thread to
				 * proceed.
				 */
1332 1333 1334
			case -ENOMEM:
				err = 0;
				goto out;
L
Linus Torvalds 已提交
1335 1336 1337
		}
		err = nfs4_handle_exception(server, err, &exception);
	} while (exception.retry);
1338
out:
L
Linus Torvalds 已提交
1339 1340 1341
	return err;
}

1342 1343 1344 1345 1346
static void nfs4_open_confirm_done(struct rpc_task *task, void *calldata)
{
	struct nfs4_opendata *data = calldata;

	data->rpc_status = task->tk_status;
1347
	if (data->rpc_status == 0) {
1348 1349
		memcpy(data->o_res.stateid.data, data->c_res.stateid.data,
				sizeof(data->o_res.stateid.data));
1350
		nfs_confirm_seqid(&data->owner->so_seqid, 0);
1351
		renew_lease(data->o_res.server, data->timestamp);
1352
		data->rpc_done = 1;
1353
	}
1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 1364
}

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! */
1365
	if (!data->rpc_done)
1366 1367
		goto out_free;
	state = nfs4_opendata_to_nfs4_state(data);
1368
	if (!IS_ERR(state))
1369
		nfs4_close_state(state, data->o_arg.fmode);
1370
out_free:
1371
	nfs4_opendata_put(data);
1372 1373 1374 1375 1376 1377 1378 1379 1380 1381 1382 1383 1384 1385
}

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;
1386 1387 1388 1389 1390 1391
	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
1392 1393
	struct rpc_task_setup task_setup_data = {
		.rpc_client = server->client,
1394
		.rpc_message = &msg,
T
Trond Myklebust 已提交
1395 1396
		.callback_ops = &nfs4_open_confirm_ops,
		.callback_data = data,
1397
		.workqueue = nfsiod_workqueue,
T
Trond Myklebust 已提交
1398 1399
		.flags = RPC_TASK_ASYNC,
	};
L
Linus Torvalds 已提交
1400 1401
	int status;

1402
	kref_get(&data->kref);
1403 1404
	data->rpc_done = 0;
	data->rpc_status = 0;
1405
	data->timestamp = jiffies;
T
Trond Myklebust 已提交
1406
	task = rpc_run_task(&task_setup_data);
1407
	if (IS_ERR(task))
1408 1409 1410 1411 1412 1413 1414
		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;
1415
	rpc_put_task(task);
L
Linus Torvalds 已提交
1416 1417 1418
	return status;
}

1419
static void nfs4_open_prepare(struct rpc_task *task, void *calldata)
L
Linus Torvalds 已提交
1420
{
1421 1422
	struct nfs4_opendata *data = calldata;
	struct nfs4_state_owner *sp = data->owner;
1423

1424 1425
	if (nfs_wait_on_sequence(data->o_arg.seqid, task) != 0)
		return;
1426 1427 1428 1429 1430 1431 1432
	/*
	 * 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;

1433
		if (can_open_cached(data->state, data->o_arg.fmode, data->o_arg.open_flags))
1434
			goto out_no_action;
1435 1436
		rcu_read_lock();
		delegation = rcu_dereference(NFS_I(data->state->inode)->delegation);
1437 1438 1439
		if (data->o_arg.claim != NFS4_OPEN_CLAIM_DELEGATE_CUR &&
		    can_open_delegated(delegation, data->o_arg.fmode))
			goto unlock_no_action;
1440 1441
		rcu_read_unlock();
	}
1442
	/* Update sequence id. */
1443
	data->o_arg.id = sp->so_owner_id;
1444
	data->o_arg.clientid = sp->so_server->nfs_client->cl_clientid;
T
Trond Myklebust 已提交
1445
	if (data->o_arg.claim == NFS4_OPEN_CLAIM_PREVIOUS) {
1446
		task->tk_msg.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_OPEN_NOATTR];
T
Trond Myklebust 已提交
1447 1448
		nfs_copy_fh(&data->o_res.fh, data->o_arg.fh);
	}
1449
	data->timestamp = jiffies;
1450
	if (nfs4_setup_sequence(data->o_arg.server,
1451 1452 1453
				&data->o_arg.seq_args,
				&data->o_res.seq_res, 1, task))
		return;
1454
	rpc_call_start(task);
1455
	return;
1456 1457
unlock_no_action:
	rcu_read_unlock();
1458 1459 1460
out_no_action:
	task->tk_action = NULL;

1461
}
L
Linus Torvalds 已提交
1462

1463 1464 1465 1466 1467 1468
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);
}

1469 1470 1471
static void nfs4_open_done(struct rpc_task *task, void *calldata)
{
	struct nfs4_opendata *data = calldata;
L
Linus Torvalds 已提交
1472

1473
	data->rpc_status = task->tk_status;
1474

1475 1476
	if (!nfs4_sequence_done(task, &data->o_res.seq_res))
		return;
1477

1478 1479
	if (task->tk_status == 0) {
		switch (data->o_res.f_attr->mode & S_IFMT) {
1480 1481 1482
			case S_IFREG:
				break;
			case S_IFLNK:
1483
				data->rpc_status = -ELOOP;
1484 1485
				break;
			case S_IFDIR:
1486
				data->rpc_status = -EISDIR;
1487 1488
				break;
			default:
1489
				data->rpc_status = -ENOTDIR;
1490
		}
1491
		renew_lease(data->o_res.server, data->timestamp);
1492 1493
		if (!(data->o_res.rflags & NFS4_OPEN_RESULT_CONFIRM))
			nfs_confirm_seqid(&data->owner->so_seqid, 0);
1494
	}
1495
	data->rpc_done = 1;
1496
}
1497

1498 1499 1500 1501 1502 1503 1504 1505 1506
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! */
1507
	if (data->rpc_status != 0 || !data->rpc_done)
1508 1509 1510 1511 1512
		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);
1513
	if (!IS_ERR(state))
1514
		nfs4_close_state(state, data->o_arg.fmode);
1515
out_free:
1516
	nfs4_opendata_put(data);
1517 1518 1519 1520 1521 1522 1523 1524
}

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,
};

1525 1526 1527 1528 1529 1530 1531
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)
1532 1533 1534 1535 1536 1537
{
	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;
1538 1539 1540 1541 1542 1543
	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
1544 1545
	struct rpc_task_setup task_setup_data = {
		.rpc_client = server->client,
1546
		.rpc_message = &msg,
T
Trond Myklebust 已提交
1547 1548
		.callback_ops = &nfs4_open_ops,
		.callback_data = data,
1549
		.workqueue = nfsiod_workqueue,
T
Trond Myklebust 已提交
1550 1551
		.flags = RPC_TASK_ASYNC,
	};
1552 1553
	int status;

1554
	kref_get(&data->kref);
1555 1556
	data->rpc_done = 0;
	data->rpc_status = 0;
1557
	data->cancelled = 0;
1558 1559
	if (isrecover)
		task_setup_data.callback_ops = &nfs4_recover_open_ops;
T
Trond Myklebust 已提交
1560
	task = rpc_run_task(&task_setup_data);
1561 1562 1563 1564 1565 1566 1567 1568 1569 1570 1571 1572 1573 1574 1575 1576 1577 1578 1579 1580 1581 1582 1583
        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;

1584 1585
	nfs_fattr_map_and_free_names(NFS_SERVER(dir), &data->f_attr);

1586 1587 1588 1589 1590 1591 1592 1593 1594 1595 1596 1597 1598 1599 1600 1601 1602 1603 1604 1605 1606 1607 1608
	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);
1609 1610 1611 1612 1613 1614
	if (!data->rpc_done)
		return status;
	if (status != 0) {
		if (status == -NFS4ERR_BADNAME &&
				!(o_arg->open_flags & O_CREAT))
			return -ENOENT;
1615
		return status;
1616
	}
1617

1618 1619
	nfs_fattr_map_and_free_names(server, &data->f_attr);

1620 1621 1622 1623 1624
	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
1625 1626
	if ((o_res->rflags & NFS4_OPEN_RESULT_LOCKTYPE_POSIX) == 0)
		server->caps &= ~NFS_CAP_POSIX_LOCK;
L
Linus Torvalds 已提交
1627
	if(o_res->rflags & NFS4_OPEN_RESULT_CONFIRM) {
1628
		status = _nfs4_proc_open_confirm(data);
L
Linus Torvalds 已提交
1629
		if (status != 0)
1630
			return status;
L
Linus Torvalds 已提交
1631 1632
	}
	if (!(o_res->f_attr->valid & NFS_ATTR_FATTR))
1633
		_nfs4_proc_getattr(server, &o_res->fh, o_res->f_attr);
1634
	return 0;
L
Linus Torvalds 已提交
1635 1636
}

A
Andy Adamson 已提交
1637
static int nfs4_client_recover_expired_lease(struct nfs_client *clp)
1638
{
1639
	unsigned int loop;
1640
	int ret;
1641

1642
	for (loop = NFS4_MAX_LOOP_ON_RECOVER; loop != 0; loop--) {
1643
		ret = nfs4_wait_clnt_recover(clp);
1644
		if (ret != 0)
1645
			break;
1646 1647
		if (!test_bit(NFS4CLNT_LEASE_EXPIRED, &clp->cl_state) &&
		    !test_bit(NFS4CLNT_CHECK_LEASE,&clp->cl_state))
1648
			break;
1649
		nfs4_schedule_state_manager(clp);
1650
		ret = -EIO;
1651
	}
1652
	return ret;
1653 1654
}

A
Andy Adamson 已提交
1655 1656 1657 1658 1659
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
1660 1661 1662 1663 1664
/*
 * OPEN_EXPIRED:
 * 	reclaim state on the server after a network partition.
 * 	Assumes caller holds the appropriate lock
 */
1665
static int _nfs4_open_expired(struct nfs_open_context *ctx, struct nfs4_state *state)
L
Linus Torvalds 已提交
1666
{
1667
	struct nfs4_opendata *opendata;
1668
	int ret;
L
Linus Torvalds 已提交
1669

T
Trond Myklebust 已提交
1670 1671 1672
	opendata = nfs4_open_recoverdata_alloc(ctx, state);
	if (IS_ERR(opendata))
		return PTR_ERR(opendata);
1673
	ret = nfs4_open_recover(opendata, state);
1674
	if (ret == -ESTALE)
1675
		d_drop(ctx->dentry);
1676
	nfs4_opendata_put(opendata);
1677
	return ret;
L
Linus Torvalds 已提交
1678 1679
}

1680
static int nfs4_do_open_expired(struct nfs_open_context *ctx, struct nfs4_state *state)
1681
{
1682
	struct nfs_server *server = NFS_SERVER(state->inode);
1683 1684 1685 1686
	struct nfs4_exception exception = { };
	int err;

	do {
1687
		err = _nfs4_open_expired(ctx, state);
1688 1689 1690 1691 1692 1693 1694 1695
		switch (err) {
		default:
			goto out;
		case -NFS4ERR_GRACE:
		case -NFS4ERR_DELAY:
			nfs4_handle_exception(server, err, &exception);
			err = 0;
		}
1696
	} while (exception.retry);
1697
out:
1698 1699 1700
	return err;
}

L
Linus Torvalds 已提交
1701 1702 1703
static int nfs4_open_expired(struct nfs4_state_owner *sp, struct nfs4_state *state)
{
	struct nfs_open_context *ctx;
1704
	int ret;
L
Linus Torvalds 已提交
1705

1706 1707 1708
	ctx = nfs4_state_find_open_context(state);
	if (IS_ERR(ctx))
		return PTR_ERR(ctx);
1709
	ret = nfs4_do_open_expired(ctx, state);
1710 1711
	put_nfs_open_context(ctx);
	return ret;
L
Linus Torvalds 已提交
1712 1713
}

1714 1715 1716 1717 1718 1719 1720 1721 1722 1723 1724 1725 1726 1727
#if defined(CONFIG_NFS_V4_1)
static int nfs41_open_expired(struct nfs4_state_owner *sp, struct nfs4_state *state)
{
	int status;
	struct nfs_server *server = NFS_SERVER(state->inode);

	status = nfs41_test_stateid(server, state);
	if (status == NFS_OK)
		return 0;
	nfs41_free_stateid(server, state);
	return nfs4_open_expired(sp, state);
}
#endif

1728 1729 1730 1731 1732 1733 1734 1735 1736 1737 1738 1739 1740 1741 1742 1743
/*
 * 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
1744
/*
1745
 * Returns a referenced nfs4_state
L
Linus Torvalds 已提交
1746
 */
1747
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
1748 1749 1750 1751
{
	struct nfs4_state_owner  *sp;
	struct nfs4_state     *state = NULL;
	struct nfs_server       *server = NFS_SERVER(dir);
1752
	struct nfs4_opendata *opendata;
1753
	int status;
L
Linus Torvalds 已提交
1754 1755 1756

	/* Protect against reboot recovery conflicts */
	status = -ENOMEM;
1757 1758
	sp = nfs4_get_state_owner(server, cred, GFP_KERNEL);
	if (sp == NULL) {
L
Linus Torvalds 已提交
1759 1760 1761
		dprintk("nfs4_do_open: nfs4_get_state_owner failed!\n");
		goto out_err;
	}
1762 1763
	status = nfs4_recover_expired_lease(server);
	if (status != 0)
1764
		goto err_put_state_owner;
1765 1766
	if (dentry->d_inode != NULL)
		nfs4_return_incompatible_delegation(dentry->d_inode, fmode);
1767
	status = -ENOMEM;
1768
	opendata = nfs4_opendata_alloc(dentry, sp, fmode, flags, sattr, GFP_KERNEL);
1769
	if (opendata == NULL)
T
Trond Myklebust 已提交
1770
		goto err_put_state_owner;
L
Linus Torvalds 已提交
1771

1772 1773
	if (dentry->d_inode != NULL)
		opendata->state = nfs4_get_open_state(dentry->d_inode, sp);
1774

1775
	status = _nfs4_proc_open(opendata);
L
Linus Torvalds 已提交
1776
	if (status != 0)
1777
		goto err_opendata_put;
L
Linus Torvalds 已提交
1778

1779
	state = nfs4_opendata_to_nfs4_state(opendata);
1780 1781
	status = PTR_ERR(state);
	if (IS_ERR(state))
1782
		goto err_opendata_put;
T
Trond Myklebust 已提交
1783
	if (server->caps & NFS_CAP_POSIX_LOCK)
1784
		set_bit(NFS_STATE_POSIX_LOCKS, &state->flags);
1785 1786 1787 1788 1789 1790 1791 1792 1793 1794 1795 1796

	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);
	}
1797
	nfs4_opendata_put(opendata);
L
Linus Torvalds 已提交
1798 1799 1800
	nfs4_put_state_owner(sp);
	*res = state;
	return 0;
1801 1802
err_opendata_put:
	nfs4_opendata_put(opendata);
1803 1804
err_put_state_owner:
	nfs4_put_state_owner(sp);
L
Linus Torvalds 已提交
1805 1806 1807 1808 1809 1810
out_err:
	*res = NULL;
	return status;
}


1811
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
1812 1813 1814 1815 1816 1817
{
	struct nfs4_exception exception = { };
	struct nfs4_state *res;
	int status;

	do {
1818
		status = _nfs4_do_open(dir, dentry, fmode, flags, sattr, cred, &res);
L
Linus Torvalds 已提交
1819 1820 1821 1822 1823 1824 1825 1826
		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
1827
		 * doing an OPEN, we assume that nfs_increment_open_seqid() will
L
Linus Torvalds 已提交
1828 1829 1830 1831 1832
		 * 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) {
1833 1834 1835
			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
1836 1837 1838
			exception.retry = 1;
			continue;
		}
1839 1840 1841 1842 1843 1844 1845 1846 1847 1848
		/*
		 * 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;
		}
1849 1850 1851 1852 1853
		if (status == -EAGAIN) {
			/* We must have found a delegation */
			exception.retry = 1;
			continue;
		}
L
Linus Torvalds 已提交
1854 1855 1856 1857 1858 1859
		res = ERR_PTR(nfs4_handle_exception(NFS_SERVER(dir),
					status, &exception));
	} while (exception.retry);
	return res;
}

1860 1861 1862
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
1863
{
1864
	struct nfs_server *server = NFS_SERVER(inode);
L
Linus Torvalds 已提交
1865
        struct nfs_setattrargs  arg = {
1866
                .fh             = NFS_FH(inode),
L
Linus Torvalds 已提交
1867 1868 1869 1870 1871 1872 1873 1874 1875
                .iap            = sattr,
		.server		= server,
		.bitmask = server->attr_bitmask,
        };
        struct nfs_setattrres  res = {
		.fattr		= fattr,
		.server		= server,
        };
        struct rpc_message msg = {
1876 1877 1878 1879
		.rpc_proc	= &nfs4_procedures[NFSPROC4_CLNT_SETATTR],
		.rpc_argp	= &arg,
		.rpc_resp	= &res,
		.rpc_cred	= cred,
L
Linus Torvalds 已提交
1880
        };
1881
	unsigned long timestamp = jiffies;
1882
	int status;
L
Linus Torvalds 已提交
1883

1884
	nfs_fattr_init(fattr);
L
Linus Torvalds 已提交
1885

1886 1887 1888
	if (nfs4_copy_delegation_stateid(&arg.stateid, inode)) {
		/* Use that stateid */
	} else if (state != NULL) {
1889
		nfs4_copy_stateid(&arg.stateid, state, current->files, current->tgid);
1890
	} else
L
Linus Torvalds 已提交
1891 1892
		memcpy(&arg.stateid, &zero_stateid, sizeof(arg.stateid));

1893
	status = nfs4_call_sync(server->client, server, &msg, &arg.seq_args, &res.seq_res, 1);
1894 1895
	if (status == 0 && state != NULL)
		renew_lease(server, timestamp);
1896
	return status;
L
Linus Torvalds 已提交
1897 1898
}

1899 1900 1901
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
1902
{
1903
	struct nfs_server *server = NFS_SERVER(inode);
L
Linus Torvalds 已提交
1904 1905 1906 1907
	struct nfs4_exception exception = { };
	int err;
	do {
		err = nfs4_handle_exception(server,
1908
				_nfs4_do_setattr(inode, cred, fattr, sattr, state),
L
Linus Torvalds 已提交
1909 1910 1911 1912 1913 1914 1915 1916 1917 1918
				&exception);
	} while (exception.retry);
	return err;
}

struct nfs4_closedata {
	struct inode *inode;
	struct nfs4_state *state;
	struct nfs_closeargs arg;
	struct nfs_closeres res;
1919
	struct nfs_fattr fattr;
1920
	unsigned long timestamp;
F
Fred Isaman 已提交
1921 1922
	bool roc;
	u32 roc_barrier;
L
Linus Torvalds 已提交
1923 1924
};

1925
static void nfs4_free_closedata(void *data)
1926
{
1927 1928
	struct nfs4_closedata *calldata = data;
	struct nfs4_state_owner *sp = calldata->state->owner;
1929
	struct super_block *sb = calldata->state->inode->i_sb;
1930

F
Fred Isaman 已提交
1931 1932
	if (calldata->roc)
		pnfs_roc_release(calldata->state->inode);
1933 1934 1935
	nfs4_put_open_state(calldata->state);
	nfs_free_seqid(calldata->arg.seqid);
	nfs4_put_state_owner(sp);
1936
	nfs_sb_deactive(sb);
1937 1938 1939
	kfree(calldata);
}

1940 1941 1942 1943 1944 1945 1946 1947 1948 1949 1950 1951
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);
}

1952
static void nfs4_close_done(struct rpc_task *task, void *data)
L
Linus Torvalds 已提交
1953
{
1954
	struct nfs4_closedata *calldata = data;
L
Linus Torvalds 已提交
1955 1956 1957
	struct nfs4_state *state = calldata->state;
	struct nfs_server *server = NFS_SERVER(calldata->inode);

1958 1959
	if (!nfs4_sequence_done(task, &calldata->res.seq_res))
		return;
L
Linus Torvalds 已提交
1960 1961 1962 1963 1964
        /* 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
1965 1966 1967
			if (calldata->roc)
				pnfs_roc_set_barrier(state->inode,
						     calldata->roc_barrier);
1968
			nfs_set_open_stateid(state, &calldata->res.stateid, 0);
1969
			renew_lease(server, calldata->timestamp);
1970 1971
			nfs4_close_clear_stateid_flags(state,
					calldata->arg.fmode);
L
Linus Torvalds 已提交
1972 1973
			break;
		case -NFS4ERR_STALE_STATEID:
1974 1975
		case -NFS4ERR_OLD_STATEID:
		case -NFS4ERR_BAD_STATEID:
L
Linus Torvalds 已提交
1976
		case -NFS4ERR_EXPIRED:
1977
			if (calldata->arg.fmode == 0)
1978
				break;
L
Linus Torvalds 已提交
1979
		default:
1980 1981
			if (nfs4_async_handle_error(task, server, state) == -EAGAIN)
				rpc_restart_call_prepare(task);
L
Linus Torvalds 已提交
1982
	}
1983
	nfs_release_seqid(calldata->arg.seqid);
1984
	nfs_refresh_inode(calldata->inode, calldata->res.fattr);
L
Linus Torvalds 已提交
1985 1986
}

T
Trond Myklebust 已提交
1987
static void nfs4_close_prepare(struct rpc_task *task, void *data)
L
Linus Torvalds 已提交
1988
{
T
Trond Myklebust 已提交
1989
	struct nfs4_closedata *calldata = data;
1990
	struct nfs4_state *state = calldata->state;
1991
	int call_close = 0;
1992

1993
	if (nfs_wait_on_sequence(calldata->arg.seqid, task) != 0)
1994
		return;
1995

1996 1997
	task->tk_msg.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_OPEN_DOWNGRADE];
	calldata->arg.fmode = FMODE_READ|FMODE_WRITE;
1998
	spin_lock(&state->owner->so_lock);
1999
	/* Calculate the change in open mode */
2000
	if (state->n_rdwr == 0) {
2001
		if (state->n_rdonly == 0) {
2002 2003 2004
			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;
2005 2006
		}
		if (state->n_wronly == 0) {
2007 2008 2009
			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;
2010
		}
2011
	}
2012
	spin_unlock(&state->owner->so_lock);
2013 2014

	if (!call_close) {
2015 2016
		/* Note: exit _without_ calling nfs4_close_done */
		task->tk_action = NULL;
2017 2018
		return;
	}
2019

F
Fred Isaman 已提交
2020
	if (calldata->arg.fmode == 0) {
2021
		task->tk_msg.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_CLOSE];
F
Fred Isaman 已提交
2022 2023 2024 2025 2026 2027 2028
		if (calldata->roc &&
		    pnfs_roc_drain(calldata->inode, &calldata->roc_barrier)) {
			rpc_sleep_on(&NFS_SERVER(calldata->inode)->roc_rpcwaitq,
				     task, NULL);
			return;
		}
	}
2029

2030
	nfs_fattr_init(calldata->res.fattr);
2031
	calldata->timestamp = jiffies;
2032
	if (nfs4_setup_sequence(NFS_SERVER(calldata->inode),
2033 2034 2035
				&calldata->arg.seq_args, &calldata->res.seq_res,
				1, task))
		return;
2036
	rpc_call_start(task);
L
Linus Torvalds 已提交
2037 2038
}

2039
static const struct rpc_call_ops nfs4_close_ops = {
T
Trond Myklebust 已提交
2040
	.rpc_call_prepare = nfs4_close_prepare,
2041 2042 2043 2044
	.rpc_call_done = nfs4_close_done,
	.rpc_release = nfs4_free_closedata,
};

L
Linus Torvalds 已提交
2045 2046 2047 2048 2049 2050 2051 2052 2053 2054 2055
/* 
 * 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!
 */
2056
int nfs4_do_close(struct nfs4_state *state, gfp_t gfp_mask, int wait, bool roc)
L
Linus Torvalds 已提交
2057
{
2058
	struct nfs_server *server = NFS_SERVER(state->inode);
L
Linus Torvalds 已提交
2059
	struct nfs4_closedata *calldata;
2060 2061
	struct nfs4_state_owner *sp = state->owner;
	struct rpc_task *task;
2062 2063 2064 2065
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_CLOSE],
		.rpc_cred = state->owner->so_cred,
	};
T
Trond Myklebust 已提交
2066 2067
	struct rpc_task_setup task_setup_data = {
		.rpc_client = server->client,
2068
		.rpc_message = &msg,
T
Trond Myklebust 已提交
2069
		.callback_ops = &nfs4_close_ops,
2070
		.workqueue = nfsiod_workqueue,
T
Trond Myklebust 已提交
2071 2072
		.flags = RPC_TASK_ASYNC,
	};
2073
	int status = -ENOMEM;
L
Linus Torvalds 已提交
2074

2075
	calldata = kzalloc(sizeof(*calldata), gfp_mask);
L
Linus Torvalds 已提交
2076
	if (calldata == NULL)
2077
		goto out;
2078
	calldata->inode = state->inode;
L
Linus Torvalds 已提交
2079
	calldata->state = state;
2080
	calldata->arg.fh = NFS_FH(state->inode);
2081
	calldata->arg.stateid = &state->open_stateid;
L
Linus Torvalds 已提交
2082
	/* Serialization for the sequence id */
2083
	calldata->arg.seqid = nfs_alloc_seqid(&state->owner->so_seqid, gfp_mask);
2084 2085
	if (calldata->arg.seqid == NULL)
		goto out_free_calldata;
2086
	calldata->arg.fmode = 0;
2087
	calldata->arg.bitmask = server->cache_consistency_bitmask;
2088
	calldata->res.fattr = &calldata->fattr;
2089
	calldata->res.seqid = calldata->arg.seqid;
2090
	calldata->res.server = server;
F
Fred Isaman 已提交
2091
	calldata->roc = roc;
2092
	nfs_sb_active(calldata->inode->i_sb);
2093

2094 2095
	msg.rpc_argp = &calldata->arg;
	msg.rpc_resp = &calldata->res;
T
Trond Myklebust 已提交
2096 2097
	task_setup_data.callback_data = calldata;
	task = rpc_run_task(&task_setup_data);
2098 2099
	if (IS_ERR(task))
		return PTR_ERR(task);
2100 2101 2102
	status = 0;
	if (wait)
		status = rpc_wait_for_completion_task(task);
2103
	rpc_put_task(task);
2104
	return status;
2105 2106 2107
out_free_calldata:
	kfree(calldata);
out:
F
Fred Isaman 已提交
2108 2109
	if (roc)
		pnfs_roc_release(state->inode);
2110 2111
	nfs4_put_open_state(state);
	nfs4_put_state_owner(sp);
2112
	return status;
L
Linus Torvalds 已提交
2113 2114
}

2115
static struct inode *
2116
nfs4_atomic_open(struct inode *dir, struct nfs_open_context *ctx, int open_flags, struct iattr *attr)
L
Linus Torvalds 已提交
2117 2118 2119
{
	struct nfs4_state *state;

T
Trond Myklebust 已提交
2120
	/* Protect against concurrent sillydeletes */
2121
	state = nfs4_do_open(dir, ctx->dentry, ctx->mode, open_flags, attr, ctx->cred);
2122 2123
	if (IS_ERR(state))
		return ERR_CAST(state);
2124
	ctx->state = state;
2125
	return igrab(state->inode);
L
Linus Torvalds 已提交
2126 2127
}

2128
static void nfs4_close_context(struct nfs_open_context *ctx, int is_sync)
T
Trond Myklebust 已提交
2129 2130 2131 2132
{
	if (ctx->state == NULL)
		return;
	if (is_sync)
2133
		nfs4_close_sync(ctx->state, ctx->mode);
T
Trond Myklebust 已提交
2134
	else
2135
		nfs4_close_state(ctx->state, ctx->mode);
T
Trond Myklebust 已提交
2136
}
L
Linus Torvalds 已提交
2137 2138 2139

static int _nfs4_server_capabilities(struct nfs_server *server, struct nfs_fh *fhandle)
{
B
Benny Halevy 已提交
2140 2141 2142
	struct nfs4_server_caps_arg args = {
		.fhandle = fhandle,
	};
L
Linus Torvalds 已提交
2143 2144 2145
	struct nfs4_server_caps_res res = {};
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_SERVER_CAPS],
B
Benny Halevy 已提交
2146
		.rpc_argp = &args,
L
Linus Torvalds 已提交
2147 2148 2149 2150
		.rpc_resp = &res,
	};
	int status;

2151
	status = nfs4_call_sync(server->client, server, &msg, &args.seq_args, &res.seq_res, 0);
L
Linus Torvalds 已提交
2152 2153
	if (status == 0) {
		memcpy(server->attr_bitmask, res.attr_bitmask, sizeof(server->attr_bitmask));
2154 2155 2156 2157 2158
		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
2159 2160 2161 2162 2163 2164
		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;
2165 2166 2167 2168 2169 2170 2171 2172 2173 2174 2175 2176 2177 2178 2179 2180 2181
		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;

2182 2183 2184
		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
2185 2186
		server->acl_bitmask = res.acl_bitmask;
	}
A
Andy Adamson 已提交
2187

L
Linus Torvalds 已提交
2188 2189 2190
	return status;
}

2191
int nfs4_server_capabilities(struct nfs_server *server, struct nfs_fh *fhandle)
L
Linus Torvalds 已提交
2192 2193 2194 2195 2196 2197 2198 2199 2200 2201 2202 2203 2204 2205 2206 2207 2208 2209 2210
{
	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,
2211
		.fattr = info->fattr,
L
Linus Torvalds 已提交
2212 2213 2214 2215 2216 2217 2218
		.fh = fhandle,
	};
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_LOOKUP_ROOT],
		.rpc_argp = &args,
		.rpc_resp = &res,
	};
2219

2220
	nfs_fattr_init(info->fattr);
2221
	return nfs4_call_sync(server->client, server, &msg, &args.seq_args, &res.seq_res, 0);
L
Linus Torvalds 已提交
2222 2223 2224 2225 2226 2227 2228 2229
}

static int nfs4_lookup_root(struct nfs_server *server, struct nfs_fh *fhandle,
		struct nfs_fsinfo *info)
{
	struct nfs4_exception exception = { };
	int err;
	do {
2230 2231 2232 2233 2234 2235 2236 2237
		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
2238 2239 2240 2241
	} while (exception.retry);
	return err;
}

2242 2243 2244 2245 2246 2247 2248 2249 2250 2251 2252 2253 2254 2255 2256 2257
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;
}

2258
static int nfs4_find_root_sec(struct nfs_server *server, struct nfs_fh *fhandle,
2259
			      struct nfs_fsinfo *info)
L
Linus Torvalds 已提交
2260
{
2261
	int i, len, status = 0;
2262
	rpc_authflavor_t flav_array[NFS_MAX_SECFLAVORS];
L
Linus Torvalds 已提交
2263

2264 2265 2266
	len = gss_mech_list_pseudoflavors(&flav_array[0]);
	flav_array[len] = RPC_AUTH_NULL;
	len += 1;
2267 2268 2269

	for (i = 0; i < len; i++) {
		status = nfs4_lookup_root_sec(server, fhandle, info, flav_array[i]);
2270
		if (status == -NFS4ERR_WRONGSEC || status == -EACCES)
2271 2272
			continue;
		break;
2273
	}
2274 2275 2276 2277 2278 2279 2280 2281 2282
	/*
	 * -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;
2283 2284 2285 2286 2287 2288 2289 2290 2291
	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)
{
2292
	int minor_version = server->nfs_client->cl_minorversion;
2293
	int status = nfs4_lookup_root(server, fhandle, info);
2294 2295 2296 2297 2298
	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.
		 */
2299
		status = nfs_v4_minor_ops[minor_version]->find_root_sec(server, fhandle, info);
L
Linus Torvalds 已提交
2300 2301 2302 2303
	if (status == 0)
		status = nfs4_server_capabilities(server, fhandle);
	if (status == 0)
		status = nfs4_do_fsinfo(server, fhandle, info);
2304
	return nfs4_map_errors(status);
L
Linus Torvalds 已提交
2305 2306
}

2307
static void nfs_fixup_referral_attributes(struct nfs_fattr *fattr);
M
Manoj Naik 已提交
2308 2309 2310 2311 2312
/*
 * 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
 */
2313 2314
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
2315 2316 2317 2318 2319 2320 2321 2322 2323 2324 2325 2326
{
	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;

2327
	status = nfs4_proc_fs_locations(dir, name, locations, page);
M
Manoj Naik 已提交
2328 2329 2330 2331
	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)) {
2332 2333
		dprintk("%s: server did not return a different fsid for"
			" a referral at %s\n", __func__, name->name);
M
Manoj Naik 已提交
2334 2335 2336
		status = -EIO;
		goto out;
	}
2337 2338
	/* Fixup attributes for the nfs_lookup() call to nfs_fhget() */
	nfs_fixup_referral_attributes(&locations->fattr);
M
Manoj Naik 已提交
2339

2340
	/* replace the lookup nfs_fattr with the locations nfs_fattr */
M
Manoj Naik 已提交
2341 2342 2343 2344 2345
	memcpy(fattr, &locations->fattr, sizeof(struct nfs_fattr));
	memset(fhandle, 0, sizeof(struct nfs_fh));
out:
	if (page)
		__free_page(page);
2346
	kfree(locations);
M
Manoj Naik 已提交
2347 2348 2349
	return status;
}

L
Linus Torvalds 已提交
2350 2351 2352 2353 2354 2355 2356 2357 2358 2359 2360 2361 2362 2363 2364 2365
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,
	};
	
2366
	nfs_fattr_init(fattr);
2367
	return nfs4_call_sync(server->client, server, &msg, &args.seq_args, &res.seq_res, 0);
L
Linus Torvalds 已提交
2368 2369 2370 2371 2372 2373 2374 2375 2376 2377 2378 2379 2380 2381 2382 2383 2384 2385 2386 2387 2388 2389 2390 2391 2392 2393 2394 2395 2396 2397 2398 2399 2400 2401 2402
}

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)
{
2403
	struct inode *inode = dentry->d_inode;
2404
	struct rpc_cred *cred = NULL;
2405
	struct nfs4_state *state = NULL;
L
Linus Torvalds 已提交
2406 2407
	int status;

B
Benny Halevy 已提交
2408 2409 2410
	if (pnfs_ld_layoutret_on_setattr(inode))
		pnfs_return_layout(inode);

2411
	nfs_fattr_init(fattr);
L
Linus Torvalds 已提交
2412
	
2413
	/* Search for an existing open(O_WRITE) file */
2414 2415 2416 2417
	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
2418 2419 2420 2421
		if (ctx) {
			cred = ctx->cred;
			state = ctx->state;
		}
2422
	}
2423

2424
	status = nfs4_do_setattr(inode, cred, fattr, sattr, state);
2425 2426
	if (status == 0)
		nfs_setattr_update_inode(inode, sattr);
L
Linus Torvalds 已提交
2427 2428 2429
	return status;
}

2430 2431 2432
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
2433
{
2434
	struct nfs_server *server = NFS_SERVER(dir);
D
David Howells 已提交
2435 2436 2437
	int		       status;
	struct nfs4_lookup_arg args = {
		.bitmask = server->attr_bitmask,
2438
		.dir_fh = NFS_FH(dir),
D
David Howells 已提交
2439 2440 2441 2442 2443 2444 2445 2446 2447 2448 2449 2450 2451 2452 2453
		.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
2454
	dprintk("NFS call  lookup %s\n", name->name);
2455
	status = nfs4_call_sync(clnt, server, &msg, &args.seq_args, &res.seq_res, 0);
L
Linus Torvalds 已提交
2456 2457 2458 2459
	dprintk("NFS reply lookup: %d\n", status);
	return status;
}

2460 2461 2462 2463 2464 2465 2466 2467 2468 2469
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;
}

2470 2471
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
2472 2473 2474 2475
{
	struct nfs4_exception exception = { };
	int err;
	do {
2476 2477 2478 2479
		int status;

		status = _nfs4_proc_lookup(clnt, dir, name, fhandle, fattr);
		switch (status) {
2480 2481
		case -NFS4ERR_BADNAME:
			return -ENOENT;
2482
		case -NFS4ERR_MOVED:
2483
			return nfs4_get_referral(dir, name, fattr, fhandle);
2484
		case -NFS4ERR_WRONGSEC:
2485
			nfs_fixup_secinfo_attributes(fattr, fhandle);
2486 2487 2488
		}
		err = nfs4_handle_exception(NFS_SERVER(dir),
				status, &exception);
L
Linus Torvalds 已提交
2489 2490 2491 2492 2493 2494
	} while (exception.retry);
	return err;
}

static int _nfs4_proc_access(struct inode *inode, struct nfs_access_entry *entry)
{
2495
	struct nfs_server *server = NFS_SERVER(inode);
L
Linus Torvalds 已提交
2496 2497
	struct nfs4_accessargs args = {
		.fh = NFS_FH(inode),
2498 2499 2500 2501
		.bitmask = server->attr_bitmask,
	};
	struct nfs4_accessres res = {
		.server = server,
L
Linus Torvalds 已提交
2502 2503 2504 2505 2506 2507 2508 2509 2510 2511 2512 2513 2514 2515 2516 2517 2518 2519 2520 2521 2522 2523 2524 2525 2526 2527
	};
	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;
	}
2528 2529 2530 2531 2532

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

2533
	status = nfs4_call_sync(server->client, server, &msg, &args.seq_args, &res.seq_res, 0);
L
Linus Torvalds 已提交
2534 2535 2536 2537 2538 2539 2540 2541
	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;
2542
		nfs_refresh_inode(inode, res.fattr);
L
Linus Torvalds 已提交
2543
	}
2544
	nfs_free_fattr(res.fattr);
L
Linus Torvalds 已提交
2545 2546 2547 2548 2549 2550 2551 2552 2553 2554 2555 2556 2557 2558 2559 2560 2561 2562 2563 2564 2565 2566 2567 2568 2569 2570 2571 2572 2573 2574 2575 2576 2577 2578 2579 2580 2581 2582 2583 2584 2585 2586 2587 2588 2589 2590 2591 2592
	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
2593
	struct nfs4_readlink_res res;
L
Linus Torvalds 已提交
2594 2595 2596
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_READLINK],
		.rpc_argp = &args,
B
Benny Halevy 已提交
2597
		.rpc_resp = &res,
L
Linus Torvalds 已提交
2598 2599
	};

2600
	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
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 2627 2628 2629 2630 2631 2632
}

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,
2633
                 int flags, struct nfs_open_context *ctx)
L
Linus Torvalds 已提交
2634
{
2635
	struct dentry *de = dentry;
L
Linus Torvalds 已提交
2636
	struct nfs4_state *state;
2637 2638
	struct rpc_cred *cred = NULL;
	fmode_t fmode = 0;
L
Linus Torvalds 已提交
2639 2640
	int status = 0;

2641 2642
	if (ctx != NULL) {
		cred = ctx->cred;
2643
		de = ctx->dentry;
2644
		fmode = ctx->mode;
L
Linus Torvalds 已提交
2645
	}
A
Aneesh Kumar K.V 已提交
2646
	sattr->ia_mode &= ~current_umask();
2647
	state = nfs4_do_open(dir, de, fmode, flags, sattr, cred);
2648
	d_drop(dentry);
L
Linus Torvalds 已提交
2649 2650
	if (IS_ERR(state)) {
		status = PTR_ERR(state);
2651
		goto out;
L
Linus Torvalds 已提交
2652
	}
2653
	d_add(dentry, igrab(state->inode));
2654
	nfs_set_verifier(dentry, nfs_save_change_attribute(dir));
2655 2656
	if (ctx != NULL)
		ctx->state = state;
2657
	else
2658
		nfs4_close_sync(state, fmode);
L
Linus Torvalds 已提交
2659 2660 2661 2662 2663 2664
out:
	return status;
}

static int _nfs4_proc_remove(struct inode *dir, struct qstr *name)
{
2665
	struct nfs_server *server = NFS_SERVER(dir);
2666
	struct nfs_removeargs args = {
L
Linus Torvalds 已提交
2667
		.fh = NFS_FH(dir),
2668 2669
		.name.len = name->len,
		.name.name = name->name,
2670 2671
		.bitmask = server->attr_bitmask,
	};
2672
	struct nfs_removeres res = {
2673
		.server = server,
L
Linus Torvalds 已提交
2674 2675
	};
	struct rpc_message msg = {
2676 2677 2678
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_REMOVE],
		.rpc_argp = &args,
		.rpc_resp = &res,
L
Linus Torvalds 已提交
2679
	};
2680 2681 2682 2683 2684
	int status = -ENOMEM;

	res.dir_attr = nfs_alloc_fattr();
	if (res.dir_attr == NULL)
		goto out;
L
Linus Torvalds 已提交
2685

2686
	status = nfs4_call_sync(server->client, server, &msg, &args.seq_args, &res.seq_res, 1);
2687 2688
	if (status == 0) {
		update_changeattr(dir, &res.cinfo);
2689
		nfs_post_op_update_inode(dir, res.dir_attr);
2690
	}
2691 2692
	nfs_free_fattr(res.dir_attr);
out:
L
Linus Torvalds 已提交
2693 2694 2695 2696 2697 2698 2699 2700 2701 2702 2703 2704 2705 2706 2707
	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;
}

2708
static void nfs4_proc_unlink_setup(struct rpc_message *msg, struct inode *dir)
L
Linus Torvalds 已提交
2709
{
2710 2711 2712
	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
2713

2714
	args->bitmask = server->cache_consistency_bitmask;
2715
	res->server = server;
2716
	res->seq_res.sr_slot = NULL;
L
Linus Torvalds 已提交
2717 2718 2719
	msg->rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_REMOVE];
}

2720
static int nfs4_proc_unlink_done(struct rpc_task *task, struct inode *dir)
L
Linus Torvalds 已提交
2721
{
2722 2723
	struct nfs_removeres *res = task->tk_msg.rpc_resp;

2724 2725
	if (!nfs4_sequence_done(task, &res->seq_res))
		return 0;
2726
	if (nfs4_async_handle_error(task, res->server, NULL) == -EAGAIN)
2727 2728
		return 0;
	update_changeattr(dir, &res->cinfo);
2729
	nfs_post_op_update_inode(dir, res->dir_attr);
2730
	return 1;
L
Linus Torvalds 已提交
2731 2732
}

2733 2734 2735 2736 2737 2738 2739 2740 2741 2742 2743 2744 2745 2746 2747 2748 2749 2750 2751 2752 2753 2754 2755 2756 2757 2758 2759 2760
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;
}

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
2761 2762 2763
static int _nfs4_proc_rename(struct inode *old_dir, struct qstr *old_name,
		struct inode *new_dir, struct qstr *new_name)
{
2764
	struct nfs_server *server = NFS_SERVER(old_dir);
2765
	struct nfs_renameargs arg = {
L
Linus Torvalds 已提交
2766 2767 2768 2769
		.old_dir = NFS_FH(old_dir),
		.new_dir = NFS_FH(new_dir),
		.old_name = old_name,
		.new_name = new_name,
2770 2771
		.bitmask = server->attr_bitmask,
	};
2772
	struct nfs_renameres res = {
2773
		.server = server,
L
Linus Torvalds 已提交
2774 2775 2776 2777 2778 2779
	};
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_RENAME],
		.rpc_argp = &arg,
		.rpc_resp = &res,
	};
2780
	int status = -ENOMEM;
L
Linus Torvalds 已提交
2781
	
2782 2783 2784 2785
	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
2786

2787
	status = nfs4_call_sync(server->client, server, &msg, &arg.seq_args, &res.seq_res, 1);
L
Linus Torvalds 已提交
2788 2789
	if (!status) {
		update_changeattr(old_dir, &res.old_cinfo);
2790
		nfs_post_op_update_inode(old_dir, res.old_fattr);
L
Linus Torvalds 已提交
2791
		update_changeattr(new_dir, &res.new_cinfo);
2792
		nfs_post_op_update_inode(new_dir, res.new_fattr);
L
Linus Torvalds 已提交
2793
	}
2794 2795 2796
out:
	nfs_free_fattr(res.new_fattr);
	nfs_free_fattr(res.old_fattr);
L
Linus Torvalds 已提交
2797 2798 2799 2800 2801 2802 2803 2804 2805 2806 2807 2808 2809 2810 2811 2812 2813 2814 2815
	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)
{
2816
	struct nfs_server *server = NFS_SERVER(inode);
L
Linus Torvalds 已提交
2817 2818 2819 2820
	struct nfs4_link_arg arg = {
		.fh     = NFS_FH(inode),
		.dir_fh = NFS_FH(dir),
		.name   = name,
2821 2822 2823 2824
		.bitmask = server->attr_bitmask,
	};
	struct nfs4_link_res res = {
		.server = server,
L
Linus Torvalds 已提交
2825 2826 2827 2828
	};
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_LINK],
		.rpc_argp = &arg,
2829
		.rpc_resp = &res,
L
Linus Torvalds 已提交
2830
	};
2831 2832 2833 2834 2835 2836
	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
2837

2838
	status = nfs4_call_sync(server->client, server, &msg, &arg.seq_args, &res.seq_res, 1);
2839 2840 2841
	if (!status) {
		update_changeattr(dir, &res.cinfo);
		nfs_post_op_update_inode(dir, res.dir_attr);
2842
		nfs_post_op_update_inode(inode, res.fattr);
2843
	}
2844 2845 2846
out:
	nfs_free_fattr(res.dir_attr);
	nfs_free_fattr(res.fattr);
L
Linus Torvalds 已提交
2847 2848 2849 2850 2851 2852 2853 2854 2855 2856 2857 2858 2859 2860 2861
	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;
}

2862 2863 2864 2865 2866 2867 2868 2869 2870 2871 2872 2873 2874 2875 2876 2877 2878 2879 2880 2881 2882 2883 2884 2885 2886 2887 2888 2889 2890 2891 2892 2893 2894 2895 2896 2897 2898 2899 2900
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)
{
2901
	int status = nfs4_call_sync(NFS_SERVER(dir)->client, NFS_SERVER(dir), &data->msg,
2902
				    &data->arg.seq_args, &data->res.seq_res, 1);
2903 2904 2905 2906 2907 2908 2909 2910 2911 2912 2913 2914 2915
	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);
}

2916
static int _nfs4_proc_symlink(struct inode *dir, struct dentry *dentry,
2917
		struct page *page, unsigned int len, struct iattr *sattr)
L
Linus Torvalds 已提交
2918
{
2919 2920
	struct nfs4_createdata *data;
	int status = -ENAMETOOLONG;
L
Linus Torvalds 已提交
2921

2922
	if (len > NFS4_MAXPATHLEN)
2923
		goto out;
2924

2925 2926 2927 2928 2929 2930 2931 2932
	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
2933
	
2934 2935 2936 2937
	status = nfs4_do_create(dir, dentry, data);

	nfs4_free_createdata(data);
out:
L
Linus Torvalds 已提交
2938 2939 2940
	return status;
}

2941
static int nfs4_proc_symlink(struct inode *dir, struct dentry *dentry,
2942
		struct page *page, unsigned int len, struct iattr *sattr)
L
Linus Torvalds 已提交
2943 2944 2945 2946 2947
{
	struct nfs4_exception exception = { };
	int err;
	do {
		err = nfs4_handle_exception(NFS_SERVER(dir),
2948 2949
				_nfs4_proc_symlink(dir, dentry, page,
							len, sattr),
L
Linus Torvalds 已提交
2950 2951 2952 2953 2954 2955 2956 2957
				&exception);
	} while (exception.retry);
	return err;
}

static int _nfs4_proc_mkdir(struct inode *dir, struct dentry *dentry,
		struct iattr *sattr)
{
2958 2959
	struct nfs4_createdata *data;
	int status = -ENOMEM;
L
Linus Torvalds 已提交
2960

2961 2962 2963 2964 2965 2966 2967 2968
	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
2969 2970 2971 2972 2973 2974 2975 2976
	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
2977 2978

	sattr->ia_mode &= ~current_umask();
L
Linus Torvalds 已提交
2979 2980 2981 2982 2983 2984 2985 2986 2987
	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
2988
		u64 cookie, struct page **pages, unsigned int count, int plus)
L
Linus Torvalds 已提交
2989 2990 2991 2992
{
	struct inode		*dir = dentry->d_inode;
	struct nfs4_readdir_arg args = {
		.fh = NFS_FH(dir),
B
Bryan Schumaker 已提交
2993
		.pages = pages,
L
Linus Torvalds 已提交
2994 2995
		.pgbase = 0,
		.count = count,
2996
		.bitmask = NFS_SERVER(dentry->d_inode)->attr_bitmask,
B
Bryan Schumaker 已提交
2997
		.plus = plus,
L
Linus Torvalds 已提交
2998 2999 3000 3001 3002 3003 3004 3005 3006 3007
	};
	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;

3008
	dprintk("%s: dentry = %s/%s, cookie = %Lu\n", __func__,
3009 3010 3011
			dentry->d_parent->d_name.name,
			dentry->d_name.name,
			(unsigned long long)cookie);
L
Linus Torvalds 已提交
3012 3013
	nfs4_setup_readdir(cookie, NFS_COOKIEVERF(dir), dentry, &args);
	res.pgbase = args.pgbase;
3014
	status = nfs4_call_sync(NFS_SERVER(dir)->client, NFS_SERVER(dir), &msg, &args.seq_args, &res.seq_res, 0);
3015
	if (status >= 0) {
L
Linus Torvalds 已提交
3016
		memcpy(NFS_COOKIEVERF(dir), res.verifier.data, NFS4_VERIFIER_SIZE);
3017 3018
		status += args.pgbase;
	}
3019 3020 3021

	nfs_invalidate_atime(dir);

3022
	dprintk("%s: returns %d\n", __func__, status);
L
Linus Torvalds 已提交
3023 3024 3025 3026
	return status;
}

static int nfs4_proc_readdir(struct dentry *dentry, struct rpc_cred *cred,
B
Bryan Schumaker 已提交
3027
		u64 cookie, struct page **pages, unsigned int count, int plus)
L
Linus Torvalds 已提交
3028 3029 3030 3031 3032 3033
{
	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
3034
					pages, count, plus),
L
Linus Torvalds 已提交
3035 3036 3037 3038 3039 3040 3041 3042
				&exception);
	} while (exception.retry);
	return err;
}

static int _nfs4_proc_mknod(struct inode *dir, struct dentry *dentry,
		struct iattr *sattr, dev_t rdev)
{
3043 3044 3045
	struct nfs4_createdata *data;
	int mode = sattr->ia_mode;
	int status = -ENOMEM;
L
Linus Torvalds 已提交
3046 3047 3048

	BUG_ON(!(sattr->ia_valid & ATTR_MODE));
	BUG_ON(!S_ISFIFO(mode) && !S_ISBLK(mode) && !S_ISCHR(mode) && !S_ISSOCK(mode));
3049 3050 3051 3052 3053

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

L
Linus Torvalds 已提交
3054
	if (S_ISFIFO(mode))
3055
		data->arg.ftype = NF4FIFO;
L
Linus Torvalds 已提交
3056
	else if (S_ISBLK(mode)) {
3057 3058 3059
		data->arg.ftype = NF4BLK;
		data->arg.u.device.specdata1 = MAJOR(rdev);
		data->arg.u.device.specdata2 = MINOR(rdev);
L
Linus Torvalds 已提交
3060 3061
	}
	else if (S_ISCHR(mode)) {
3062 3063 3064
		data->arg.ftype = NF4CHR;
		data->arg.u.device.specdata1 = MAJOR(rdev);
		data->arg.u.device.specdata2 = MINOR(rdev);
L
Linus Torvalds 已提交
3065 3066
	}
	
3067 3068 3069 3070
	status = nfs4_do_create(dir, dentry, data);

	nfs4_free_createdata(data);
out:
L
Linus Torvalds 已提交
3071 3072 3073 3074 3075 3076 3077 3078
	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
3079 3080

	sattr->ia_mode &= ~current_umask();
L
Linus Torvalds 已提交
3081 3082 3083 3084 3085 3086 3087 3088 3089 3090 3091 3092 3093 3094 3095
	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
3096 3097 3098
	struct nfs4_statfs_res res = {
		.fsstat = fsstat,
	};
L
Linus Torvalds 已提交
3099 3100 3101
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_STATFS],
		.rpc_argp = &args,
B
Benny Halevy 已提交
3102
		.rpc_resp = &res,
L
Linus Torvalds 已提交
3103 3104
	};

3105
	nfs_fattr_init(fsstat->fattr);
3106
	return  nfs4_call_sync(server->client, server, &msg, &args.seq_args, &res.seq_res, 0);
L
Linus Torvalds 已提交
3107 3108 3109 3110 3111 3112 3113 3114 3115 3116 3117 3118 3119 3120 3121 3122 3123 3124 3125 3126 3127
}

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
3128 3129 3130
	struct nfs4_fsinfo_res res = {
		.fsinfo = fsinfo,
	};
L
Linus Torvalds 已提交
3131 3132 3133
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_FSINFO],
		.rpc_argp = &args,
B
Benny Halevy 已提交
3134
		.rpc_resp = &res,
L
Linus Torvalds 已提交
3135 3136
	};

3137
	return nfs4_call_sync(server->client, server, &msg, &args.seq_args, &res.seq_res, 0);
L
Linus Torvalds 已提交
3138 3139 3140 3141 3142 3143 3144 3145 3146 3147 3148 3149 3150 3151 3152 3153 3154
}

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)
{
3155
	nfs_fattr_init(fsinfo->fattr);
L
Linus Torvalds 已提交
3156 3157 3158 3159 3160 3161 3162 3163 3164 3165
	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
3166 3167 3168
	struct nfs4_pathconf_res res = {
		.pathconf = pathconf,
	};
L
Linus Torvalds 已提交
3169 3170 3171
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_PATHCONF],
		.rpc_argp = &args,
B
Benny Halevy 已提交
3172
		.rpc_resp = &res,
L
Linus Torvalds 已提交
3173 3174 3175 3176 3177 3178 3179 3180
	};

	/* 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;
	}

3181
	nfs_fattr_init(pathconf->fattr);
3182
	return nfs4_call_sync(server->client, server, &msg, &args.seq_args, &res.seq_res, 0);
L
Linus Torvalds 已提交
3183 3184 3185 3186 3187 3188 3189 3190 3191 3192 3193 3194 3195 3196 3197 3198
}

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;
}

3199 3200 3201 3202 3203
void __nfs4_read_done_cb(struct nfs_read_data *data)
{
	nfs_invalidate_atime(data->inode);
}

3204
static int nfs4_read_done_cb(struct rpc_task *task, struct nfs_read_data *data)
L
Linus Torvalds 已提交
3205
{
T
Trond Myklebust 已提交
3206
	struct nfs_server *server = NFS_SERVER(data->inode);
L
Linus Torvalds 已提交
3207

3208
	if (nfs4_async_handle_error(task, server, data->args.context->state) == -EAGAIN) {
3209
		rpc_restart_call_prepare(task);
T
Trond Myklebust 已提交
3210
		return -EAGAIN;
L
Linus Torvalds 已提交
3211
	}
3212

3213
	__nfs4_read_done_cb(data);
L
Linus Torvalds 已提交
3214
	if (task->tk_status > 0)
T
Trond Myklebust 已提交
3215 3216
		renew_lease(server, data->timestamp);
	return 0;
L
Linus Torvalds 已提交
3217 3218
}

3219 3220 3221 3222 3223 3224 3225 3226
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;

3227 3228
	return data->read_done_cb ? data->read_done_cb(task, data) :
				    nfs4_read_done_cb(task, data);
3229 3230
}

3231
static void nfs4_proc_read_setup(struct nfs_read_data *data, struct rpc_message *msg)
L
Linus Torvalds 已提交
3232 3233
{
	data->timestamp   = jiffies;
3234
	data->read_done_cb = nfs4_read_done_cb;
3235
	msg->rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_READ];
L
Linus Torvalds 已提交
3236 3237
}

3238 3239 3240 3241 3242 3243 3244 3245 3246 3247 3248 3249 3250 3251 3252 3253
/* 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);

3254
static int nfs4_write_done_cb(struct rpc_task *task, struct nfs_write_data *data)
L
Linus Torvalds 已提交
3255 3256 3257
{
	struct inode *inode = data->inode;
	
3258
	if (nfs4_async_handle_error(task, NFS_SERVER(inode), data->args.context->state) == -EAGAIN) {
3259
		rpc_restart_call_prepare(task);
3260
		return -EAGAIN;
L
Linus Torvalds 已提交
3261
	}
3262
	if (task->tk_status >= 0) {
L
Linus Torvalds 已提交
3263
		renew_lease(NFS_SERVER(inode), data->timestamp);
3264
		nfs_post_op_update_inode_force_wcc(inode, data->res.fattr);
3265
	}
3266
	return 0;
L
Linus Torvalds 已提交
3267 3268
}

3269 3270 3271 3272
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;
3273 3274
	return data->write_done_cb ? data->write_done_cb(task, data) :
		nfs4_write_done_cb(task, data);
3275 3276
}

3277 3278 3279 3280 3281 3282 3283 3284 3285 3286 3287 3288 3289 3290 3291 3292 3293
/* 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);

3294
static void nfs4_proc_write_setup(struct nfs_write_data *data, struct rpc_message *msg)
L
Linus Torvalds 已提交
3295
{
3296 3297
	struct nfs_server *server = NFS_SERVER(data->inode);

3298 3299 3300 3301 3302
	if (data->lseg) {
		data->args.bitmask = NULL;
		data->res.fattr = NULL;
	} else
		data->args.bitmask = server->cache_consistency_bitmask;
3303 3304
	if (!data->write_done_cb)
		data->write_done_cb = nfs4_write_done_cb;
3305
	data->res.server = server;
L
Linus Torvalds 已提交
3306 3307
	data->timestamp   = jiffies;

3308
	msg->rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_WRITE];
L
Linus Torvalds 已提交
3309 3310
}

3311
static int nfs4_commit_done_cb(struct rpc_task *task, struct nfs_write_data *data)
L
Linus Torvalds 已提交
3312 3313
{
	struct inode *inode = data->inode;
3314

3315
	if (nfs4_async_handle_error(task, NFS_SERVER(inode), NULL) == -EAGAIN) {
3316
		rpc_restart_call_prepare(task);
3317
		return -EAGAIN;
L
Linus Torvalds 已提交
3318
	}
3319
	nfs_refresh_inode(inode, data->res.fattr);
3320
	return 0;
L
Linus Torvalds 已提交
3321 3322
}

3323 3324 3325 3326 3327 3328 3329
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);
}

3330
static void nfs4_proc_commit_setup(struct nfs_write_data *data, struct rpc_message *msg)
L
Linus Torvalds 已提交
3331
{
3332
	struct nfs_server *server = NFS_SERVER(data->inode);
3333 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_commit_done_cb;
3341
	data->res.server = server;
3342
	msg->rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_COMMIT];
L
Linus Torvalds 已提交
3343 3344
}

3345 3346 3347 3348 3349
struct nfs4_renewdata {
	struct nfs_client	*client;
	unsigned long		timestamp;
};

L
Linus Torvalds 已提交
3350 3351 3352 3353
/*
 * nfs4_proc_async_renew(): This is not one of the nfs_rpc_ops; it is a special
 * standalone procedure for queueing an asynchronous RENEW.
 */
3354
static void nfs4_renew_release(void *calldata)
3355
{
3356 3357
	struct nfs4_renewdata *data = calldata;
	struct nfs_client *clp = data->client;
3358

3359 3360 3361
	if (atomic_read(&clp->cl_count) > 1)
		nfs4_schedule_state_renewal(clp);
	nfs_put_client(clp);
3362
	kfree(data);
3363 3364
}

3365
static void nfs4_renew_done(struct rpc_task *task, void *calldata)
L
Linus Torvalds 已提交
3366
{
3367 3368 3369
	struct nfs4_renewdata *data = calldata;
	struct nfs_client *clp = data->client;
	unsigned long timestamp = data->timestamp;
L
Linus Torvalds 已提交
3370 3371

	if (task->tk_status < 0) {
3372
		/* Unless we're shutting down, schedule state recovery! */
3373 3374 3375
		if (test_bit(NFS_CS_RENEWD, &clp->cl_res_state) == 0)
			return;
		if (task->tk_status != NFS4ERR_CB_PATH_DOWN) {
3376
			nfs4_schedule_lease_recovery(clp);
3377 3378 3379
			return;
		}
		nfs4_schedule_path_down_recovery(clp);
L
Linus Torvalds 已提交
3380
	}
3381
	do_renew_lease(clp, timestamp);
L
Linus Torvalds 已提交
3382 3383
}

3384 3385
static const struct rpc_call_ops nfs4_renew_ops = {
	.rpc_call_done = nfs4_renew_done,
3386
	.rpc_release = nfs4_renew_release,
3387 3388
};

3389
static int nfs4_proc_async_renew(struct nfs_client *clp, struct rpc_cred *cred, unsigned renew_flags)
L
Linus Torvalds 已提交
3390 3391 3392 3393
{
	struct rpc_message msg = {
		.rpc_proc	= &nfs4_procedures[NFSPROC4_CLNT_RENEW],
		.rpc_argp	= clp,
3394
		.rpc_cred	= cred,
L
Linus Torvalds 已提交
3395
	};
3396
	struct nfs4_renewdata *data;
L
Linus Torvalds 已提交
3397

3398 3399
	if (renew_flags == 0)
		return 0;
3400 3401
	if (!atomic_inc_not_zero(&clp->cl_count))
		return -EIO;
3402
	data = kmalloc(sizeof(*data), GFP_NOFS);
3403 3404 3405 3406
	if (data == NULL)
		return -ENOMEM;
	data->client = clp;
	data->timestamp = jiffies;
L
Linus Torvalds 已提交
3407
	return rpc_call_async(clp->cl_rpcclient, &msg, RPC_TASK_SOFT,
3408
			&nfs4_renew_ops, data);
L
Linus Torvalds 已提交
3409 3410
}

3411
static int nfs4_proc_renew(struct nfs_client *clp, struct rpc_cred *cred)
L
Linus Torvalds 已提交
3412 3413 3414 3415
{
	struct rpc_message msg = {
		.rpc_proc	= &nfs4_procedures[NFSPROC4_CLNT_RENEW],
		.rpc_argp	= clp,
3416
		.rpc_cred	= cred,
L
Linus Torvalds 已提交
3417 3418 3419 3420 3421 3422 3423
	};
	unsigned long now = jiffies;
	int status;

	status = rpc_call_sync(clp->cl_rpcclient, &msg, 0);
	if (status < 0)
		return status;
3424
	do_renew_lease(clp, now);
L
Linus Torvalds 已提交
3425 3426 3427
	return 0;
}

3428 3429 3430 3431 3432 3433 3434 3435 3436 3437 3438 3439 3440
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)

3441 3442 3443 3444 3445 3446 3447 3448 3449
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
3450
		len = min_t(size_t, PAGE_CACHE_SIZE, buflen);
3451 3452 3453 3454 3455 3456 3457 3458 3459 3460 3461 3462 3463 3464 3465 3466 3467 3468 3469
		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;
}

3470 3471 3472
struct nfs4_cached_acl {
	int cached;
	size_t len;
A
Andrew Morton 已提交
3473
	char data[0];
3474 3475 3476 3477 3478 3479 3480 3481 3482 3483 3484 3485 3486 3487 3488 3489 3490 3491 3492 3493 3494 3495 3496 3497 3498 3499 3500 3501 3502 3503 3504 3505 3506 3507 3508 3509 3510 3511 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
};

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);
}

3537 3538 3539 3540 3541 3542 3543 3544 3545 3546
/*
 * 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.
 */
3547
static ssize_t __nfs4_get_acl_uncached(struct inode *inode, void *buf, size_t buflen)
3548
{
3549
	struct page *pages[NFS4ACL_MAXPAGES] = {NULL, };
3550 3551 3552 3553 3554
	struct nfs_getaclargs args = {
		.fh = NFS_FH(inode),
		.acl_pages = pages,
		.acl_len = buflen,
	};
B
Benny Halevy 已提交
3555 3556 3557
	struct nfs_getaclres res = {
		.acl_len = buflen,
	};
3558
	void *resp_buf;
3559 3560 3561
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_GETACL],
		.rpc_argp = &args,
B
Benny Halevy 已提交
3562
		.rpc_resp = &res,
3563
	};
3564
	int ret = -ENOMEM, npages, i, acl_len = 0;
3565

3566 3567 3568 3569 3570 3571 3572 3573 3574 3575
	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;
3576
	}
3577 3578 3579 3580 3581
	if (npages > 1) {
		/* for decoding across pages */
		args.acl_scratch = alloc_page(GFP_KERNEL);
		if (!args.acl_scratch)
			goto out_free;
3582
	}
3583 3584 3585 3586 3587 3588 3589 3590
	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]);

3591
	dprintk("%s  buf %p buflen %zu npages %d args.acl_len %zu\n",
3592 3593 3594
		__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);
3595 3596
	if (ret)
		goto out_free;
3597 3598 3599 3600

	acl_len = res.acl_len - res.acl_data_offset;
	if (acl_len > args.acl_len)
		nfs4_write_cached_acl(inode, NULL, acl_len);
3601
	else
3602 3603
		nfs4_write_cached_acl(inode, resp_buf + res.acl_data_offset,
				      acl_len);
3604 3605
	if (buf) {
		ret = -ERANGE;
3606
		if (acl_len > buflen)
3607
			goto out_free;
3608 3609
		_copy_from_pages(buf, pages, res.acl_data_offset,
				res.acl_len);
3610
	}
3611
	ret = acl_len;
3612
out_free:
3613 3614 3615 3616 3617
	for (i = 0; i < npages; i++)
		if (pages[i])
			__free_page(pages[i]);
	if (args.acl_scratch)
		__free_page(args.acl_scratch);
3618 3619 3620
	return ret;
}

3621 3622 3623 3624 3625 3626 3627 3628 3629 3630 3631 3632 3633
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;
}

3634 3635 3636 3637 3638 3639 3640 3641 3642 3643
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;
3644 3645
	if (NFS_I(inode)->cache_validity & NFS_INO_INVALID_ACL)
		nfs_zap_acl_cache(inode);
3646 3647
	ret = nfs4_read_cached_acl(inode, buf, buflen);
	if (ret != -ENOENT)
3648 3649
		/* -ENOENT is returned if there is no ACL or if there is an ACL
		 * but no cached acl data, just the acl length */
3650 3651 3652 3653
		return ret;
	return nfs4_get_acl_uncached(inode, buf, buflen);
}

3654
static int __nfs4_proc_set_acl(struct inode *inode, const void *buf, size_t buflen)
3655 3656 3657 3658 3659 3660 3661 3662
{
	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
3663
	struct nfs_setaclres res;
3664 3665 3666
	struct rpc_message msg = {
		.rpc_proc	= &nfs4_procedures[NFSPROC4_CLNT_SETACL],
		.rpc_argp	= &arg,
B
Benny Halevy 已提交
3667
		.rpc_resp	= &res,
3668
	};
3669
	int ret, i;
3670 3671 3672

	if (!nfs4_server_supports_acls(server))
		return -EOPNOTSUPP;
3673 3674 3675
	i = buf_to_pages_noslab(buf, buflen, arg.acl_pages, &arg.acl_pgbase);
	if (i < 0)
		return i;
3676
	nfs_inode_return_delegation(inode);
3677
	ret = nfs4_call_sync(server->client, server, &msg, &arg.seq_args, &res.seq_res, 1);
3678 3679 3680 3681 3682 3683 3684 3685

	/*
	 * 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]);

3686 3687 3688 3689 3690 3691 3692
	/*
	 * 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);
3693 3694
	nfs_access_zap_cache(inode);
	nfs_zap_acl_cache(inode);
3695 3696 3697
	return ret;
}

3698 3699 3700 3701 3702 3703 3704 3705 3706 3707 3708 3709
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
3710
static int
3711
nfs4_async_handle_error(struct rpc_task *task, const struct nfs_server *server, struct nfs4_state *state)
L
Linus Torvalds 已提交
3712
{
3713 3714 3715
	struct nfs_client *clp = server->nfs_client;

	if (task->tk_status >= 0)
L
Linus Torvalds 已提交
3716 3717
		return 0;
	switch(task->tk_status) {
3718 3719 3720 3721 3722
		case -NFS4ERR_ADMIN_REVOKED:
		case -NFS4ERR_BAD_STATEID:
		case -NFS4ERR_OPENMODE:
			if (state == NULL)
				break;
3723 3724
			nfs4_schedule_stateid_recovery(server, state);
			goto wait_on_recovery;
3725 3726 3727
		case -NFS4ERR_EXPIRED:
			if (state != NULL)
				nfs4_schedule_stateid_recovery(server, state);
L
Linus Torvalds 已提交
3728
		case -NFS4ERR_STALE_STATEID:
3729
		case -NFS4ERR_STALE_CLIENTID:
3730 3731
			nfs4_schedule_lease_recovery(clp);
			goto wait_on_recovery;
3732 3733 3734 3735 3736 3737 3738 3739 3740 3741
#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);
3742
			nfs4_schedule_session_recovery(clp->cl_session);
3743 3744 3745
			task->tk_status = 0;
			return -EAGAIN;
#endif /* CONFIG_NFS_V4_1 */
L
Linus Torvalds 已提交
3746
		case -NFS4ERR_DELAY:
3747
			nfs_inc_server_stats(server, NFSIOS_DELAY);
3748
		case -NFS4ERR_GRACE:
3749
		case -EKEYEXPIRED:
L
Linus Torvalds 已提交
3750 3751 3752
			rpc_delay(task, NFS4_POLL_RETRY_MAX);
			task->tk_status = 0;
			return -EAGAIN;
3753
		case -NFS4ERR_RETRY_UNCACHED_REP:
L
Linus Torvalds 已提交
3754 3755 3756 3757 3758 3759
		case -NFS4ERR_OLD_STATEID:
			task->tk_status = 0;
			return -EAGAIN;
	}
	task->tk_status = nfs4_map_errors(task->tk_status);
	return 0;
3760
wait_on_recovery:
3761 3762 3763 3764 3765
	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
3766 3767
}

3768 3769 3770
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
3771 3772 3773 3774 3775
{
	nfs4_verifier sc_verifier;
	struct nfs4_setclientid setclientid = {
		.sc_verifier = &sc_verifier,
		.sc_prog = program,
A
Andy Adamson 已提交
3776
		.sc_cb_ident = clp->cl_cb_ident,
L
Linus Torvalds 已提交
3777 3778 3779 3780
	};
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_SETCLIENTID],
		.rpc_argp = &setclientid,
3781
		.rpc_resp = res,
3782
		.rpc_cred = cred,
L
Linus Torvalds 已提交
3783
	};
A
Al Viro 已提交
3784
	__be32 *p;
L
Linus Torvalds 已提交
3785 3786 3787
	int loop = 0;
	int status;

A
Al Viro 已提交
3788
	p = (__be32*)sc_verifier.data;
L
Linus Torvalds 已提交
3789 3790 3791 3792 3793
	*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,
3794
				sizeof(setclientid.sc_name), "%s/%s %s %s %u",
3795 3796 3797
				clp->cl_ipaddr,
				rpc_peeraddr2str(clp->cl_rpcclient,
							RPC_DISPLAY_ADDR),
3798 3799
				rpc_peeraddr2str(clp->cl_rpcclient,
							RPC_DISPLAY_PROTO),
3800
				clp->cl_rpcclient->cl_auth->au_ops->au_name,
L
Linus Torvalds 已提交
3801 3802
				clp->cl_id_uniquifier);
		setclientid.sc_netid_len = scnprintf(setclientid.sc_netid,
3803 3804 3805
				sizeof(setclientid.sc_netid),
				rpc_peeraddr2str(clp->cl_rpcclient,
							RPC_DISPLAY_NETID));
L
Linus Torvalds 已提交
3806
		setclientid.sc_uaddr_len = scnprintf(setclientid.sc_uaddr,
3807
				sizeof(setclientid.sc_uaddr), "%s.%u.%u",
L
Linus Torvalds 已提交
3808 3809
				clp->cl_ipaddr, port >> 8, port & 255);

3810
		status = rpc_call_sync(clp->cl_rpcclient, &msg, RPC_TASK_TIMEOUT);
L
Linus Torvalds 已提交
3811 3812
		if (status != -NFS4ERR_CLID_INUSE)
			break;
3813 3814
		if (loop != 0) {
			++clp->cl_id_uniquifier;
L
Linus Torvalds 已提交
3815
			break;
3816 3817 3818
		}
		++loop;
		ssleep(clp->cl_lease_time / HZ + 1);
L
Linus Torvalds 已提交
3819 3820 3821 3822
	}
	return status;
}

3823
int nfs4_proc_setclientid_confirm(struct nfs_client *clp,
3824 3825
		struct nfs4_setclientid_res *arg,
		struct rpc_cred *cred)
L
Linus Torvalds 已提交
3826 3827 3828 3829
{
	struct nfs_fsinfo fsinfo;
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_SETCLIENTID_CONFIRM],
3830
		.rpc_argp = arg,
L
Linus Torvalds 已提交
3831
		.rpc_resp = &fsinfo,
3832
		.rpc_cred = cred,
L
Linus Torvalds 已提交
3833 3834 3835 3836 3837
	};
	unsigned long now;
	int status;

	now = jiffies;
3838
	status = rpc_call_sync(clp->cl_rpcclient, &msg, RPC_TASK_TIMEOUT);
L
Linus Torvalds 已提交
3839 3840 3841 3842 3843 3844 3845 3846 3847
	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;
}

3848 3849
struct nfs4_delegreturndata {
	struct nfs4_delegreturnargs args;
3850
	struct nfs4_delegreturnres res;
3851 3852
	struct nfs_fh fh;
	nfs4_stateid stateid;
3853
	unsigned long timestamp;
3854
	struct nfs_fattr fattr;
3855 3856 3857 3858 3859 3860
	int rpc_status;
};

static void nfs4_delegreturn_done(struct rpc_task *task, void *calldata)
{
	struct nfs4_delegreturndata *data = calldata;
3861

3862 3863
	if (!nfs4_sequence_done(task, &data->res.seq_res))
		return;
3864

3865 3866 3867 3868
	switch (task->tk_status) {
	case -NFS4ERR_STALE_STATEID:
	case -NFS4ERR_EXPIRED:
	case 0:
3869
		renew_lease(data->res.server, data->timestamp);
3870 3871 3872 3873
		break;
	default:
		if (nfs4_async_handle_error(task, data->res.server, NULL) ==
				-EAGAIN) {
3874
			rpc_restart_call_prepare(task);
3875 3876 3877 3878
			return;
		}
	}
	data->rpc_status = task->tk_status;
3879 3880 3881 3882 3883 3884 3885
}

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

3886 3887 3888 3889 3890 3891 3892
#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;

3893
	if (nfs4_setup_sequence(d_data->res.server,
3894 3895 3896 3897 3898 3899 3900
				&d_data->args.seq_args,
				&d_data->res.seq_res, 1, task))
		return;
	rpc_call_start(task);
}
#endif /* CONFIG_NFS_V4_1 */

3901
static const struct rpc_call_ops nfs4_delegreturn_ops = {
3902 3903 3904
#if defined(CONFIG_NFS_V4_1)
	.rpc_call_prepare = nfs4_delegreturn_prepare,
#endif /* CONFIG_NFS_V4_1 */
3905 3906 3907 3908
	.rpc_call_done = nfs4_delegreturn_done,
	.rpc_release = nfs4_delegreturn_release,
};

3909
static int _nfs4_proc_delegreturn(struct inode *inode, struct rpc_cred *cred, const nfs4_stateid *stateid, int issync)
3910 3911
{
	struct nfs4_delegreturndata *data;
3912
	struct nfs_server *server = NFS_SERVER(inode);
3913
	struct rpc_task *task;
3914 3915 3916 3917
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_DELEGRETURN],
		.rpc_cred = cred,
	};
T
Trond Myklebust 已提交
3918 3919
	struct rpc_task_setup task_setup_data = {
		.rpc_client = server->client,
3920
		.rpc_message = &msg,
T
Trond Myklebust 已提交
3921 3922 3923
		.callback_ops = &nfs4_delegreturn_ops,
		.flags = RPC_TASK_ASYNC,
	};
3924
	int status = 0;
3925

3926
	data = kzalloc(sizeof(*data), GFP_NOFS);
3927 3928 3929 3930
	if (data == NULL)
		return -ENOMEM;
	data->args.fhandle = &data->fh;
	data->args.stateid = &data->stateid;
3931
	data->args.bitmask = server->attr_bitmask;
3932 3933
	nfs_copy_fh(&data->fh, NFS_FH(inode));
	memcpy(&data->stateid, stateid, sizeof(data->stateid));
3934 3935
	data->res.fattr = &data->fattr;
	data->res.server = server;
3936
	nfs_fattr_init(data->res.fattr);
3937
	data->timestamp = jiffies;
3938 3939
	data->rpc_status = 0;

T
Trond Myklebust 已提交
3940
	task_setup_data.callback_data = data;
3941 3942
	msg.rpc_argp = &data->args;
	msg.rpc_resp = &data->res;
T
Trond Myklebust 已提交
3943
	task = rpc_run_task(&task_setup_data);
3944
	if (IS_ERR(task))
3945
		return PTR_ERR(task);
3946 3947
	if (!issync)
		goto out;
3948
	status = nfs4_wait_for_completion_rpc_task(task);
3949 3950 3951 3952 3953 3954 3955
	if (status != 0)
		goto out;
	status = data->rpc_status;
	if (status != 0)
		goto out;
	nfs_refresh_inode(inode, &data->fattr);
out:
3956
	rpc_put_task(task);
3957
	return status;
L
Linus Torvalds 已提交
3958 3959
}

3960
int nfs4_proc_delegreturn(struct inode *inode, struct rpc_cred *cred, const nfs4_stateid *stateid, int issync)
L
Linus Torvalds 已提交
3961 3962 3963 3964 3965
{
	struct nfs_server *server = NFS_SERVER(inode);
	struct nfs4_exception exception = { };
	int err;
	do {
3966
		err = _nfs4_proc_delegreturn(inode, cred, stateid, issync);
L
Linus Torvalds 已提交
3967 3968 3969 3970 3971 3972 3973 3974 3975 3976 3977 3978 3979 3980 3981 3982 3983 3984 3985 3986
		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)
{
3987
	freezable_schedule_timeout_killable(timeout);
L
Linus Torvalds 已提交
3988 3989 3990 3991 3992 3993 3994 3995 3996 3997
	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);
3998
	struct nfs_client *clp = server->nfs_client;
T
Trond Myklebust 已提交
3999
	struct nfs_lockt_args arg = {
L
Linus Torvalds 已提交
4000
		.fh = NFS_FH(inode),
T
Trond Myklebust 已提交
4001
		.fl = request,
L
Linus Torvalds 已提交
4002
	};
T
Trond Myklebust 已提交
4003 4004
	struct nfs_lockt_res res = {
		.denied = request,
L
Linus Torvalds 已提交
4005 4006 4007 4008 4009 4010 4011 4012 4013 4014
	};
	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
4015
	arg.lock_owner.clientid = clp->cl_clientid;
4016 4017 4018 4019
	status = nfs4_set_lock_state(state, request);
	if (status != 0)
		goto out;
	lsp = request->fl_u.nfs4_fl.owner;
4020
	arg.lock_owner.id = lsp->ls_id;
4021
	arg.lock_owner.s_dev = server->s_dev;
4022
	status = nfs4_call_sync(server->client, server, &msg, &arg.seq_args, &res.seq_res, 1);
T
Trond Myklebust 已提交
4023 4024 4025 4026 4027 4028
	switch (status) {
		case 0:
			request->fl_type = F_UNLCK;
			break;
		case -NFS4ERR_DENIED:
			status = 0;
L
Linus Torvalds 已提交
4029
	}
4030
	request->fl_ops->fl_release_private(request);
4031
out:
L
Linus Torvalds 已提交
4032 4033 4034 4035 4036 4037 4038 4039 4040 4041 4042 4043 4044 4045 4046 4047 4048 4049 4050 4051 4052 4053 4054 4055 4056 4057 4058 4059 4060 4061 4062 4063
	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;
}

4064
struct nfs4_unlockdata {
T
Trond Myklebust 已提交
4065 4066
	struct nfs_locku_args arg;
	struct nfs_locku_res res;
4067 4068
	struct nfs4_lock_state *lsp;
	struct nfs_open_context *ctx;
T
Trond Myklebust 已提交
4069 4070
	struct file_lock fl;
	const struct nfs_server *server;
4071
	unsigned long timestamp;
4072 4073
};

T
Trond Myklebust 已提交
4074 4075 4076 4077 4078 4079 4080 4081
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;

4082
	p = kzalloc(sizeof(*p), GFP_NOFS);
T
Trond Myklebust 已提交
4083 4084 4085 4086 4087
	if (p == NULL)
		return NULL;
	p->arg.fh = NFS_FH(inode);
	p->arg.fl = &p->fl;
	p->arg.seqid = seqid;
4088
	p->res.seqid = seqid;
T
Trond Myklebust 已提交
4089 4090 4091 4092 4093 4094 4095 4096 4097 4098
	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;
}

4099
static void nfs4_locku_release_calldata(void *data)
4100
{
4101
	struct nfs4_unlockdata *calldata = data;
T
Trond Myklebust 已提交
4102
	nfs_free_seqid(calldata->arg.seqid);
4103 4104 4105
	nfs4_put_lock_state(calldata->lsp);
	put_nfs_open_context(calldata->ctx);
	kfree(calldata);
4106 4107
}

4108
static void nfs4_locku_done(struct rpc_task *task, void *data)
4109
{
4110
	struct nfs4_unlockdata *calldata = data;
4111

4112 4113
	if (!nfs4_sequence_done(task, &calldata->res.seq_res))
		return;
4114 4115 4116
	switch (task->tk_status) {
		case 0:
			memcpy(calldata->lsp->ls_stateid.data,
T
Trond Myklebust 已提交
4117
					calldata->res.stateid.data,
4118
					sizeof(calldata->lsp->ls_stateid.data));
4119
			renew_lease(calldata->server, calldata->timestamp);
4120
			break;
4121 4122
		case -NFS4ERR_BAD_STATEID:
		case -NFS4ERR_OLD_STATEID:
4123 4124 4125 4126
		case -NFS4ERR_STALE_STATEID:
		case -NFS4ERR_EXPIRED:
			break;
		default:
4127
			if (nfs4_async_handle_error(task, calldata->server, NULL) == -EAGAIN)
4128
				rpc_restart_call_prepare(task);
4129 4130 4131
	}
}

T
Trond Myklebust 已提交
4132
static void nfs4_locku_prepare(struct rpc_task *task, void *data)
4133
{
T
Trond Myklebust 已提交
4134
	struct nfs4_unlockdata *calldata = data;
4135

T
Trond Myklebust 已提交
4136
	if (nfs_wait_on_sequence(calldata->arg.seqid, task) != 0)
4137 4138
		return;
	if ((calldata->lsp->ls_flags & NFS_LOCK_INITIALIZED) == 0) {
4139 4140
		/* Note: exit _without_ running nfs4_locku_done */
		task->tk_action = NULL;
4141 4142
		return;
	}
4143
	calldata->timestamp = jiffies;
4144
	if (nfs4_setup_sequence(calldata->server,
4145 4146 4147
				&calldata->arg.seq_args,
				&calldata->res.seq_res, 1, task))
		return;
4148
	rpc_call_start(task);
4149 4150
}

4151
static const struct rpc_call_ops nfs4_locku_ops = {
T
Trond Myklebust 已提交
4152
	.rpc_call_prepare = nfs4_locku_prepare,
4153
	.rpc_call_done = nfs4_locku_done,
4154
	.rpc_release = nfs4_locku_release_calldata,
4155 4156
};

4157 4158 4159 4160 4161 4162
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;
4163 4164 4165 4166
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_LOCKU],
		.rpc_cred = ctx->cred,
	};
T
Trond Myklebust 已提交
4167 4168
	struct rpc_task_setup task_setup_data = {
		.rpc_client = NFS_CLIENT(lsp->ls_state->inode),
4169
		.rpc_message = &msg,
T
Trond Myklebust 已提交
4170
		.callback_ops = &nfs4_locku_ops,
4171
		.workqueue = nfsiod_workqueue,
T
Trond Myklebust 已提交
4172 4173
		.flags = RPC_TASK_ASYNC,
	};
4174

4175 4176 4177 4178 4179
	/* 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;

4180 4181 4182 4183 4184 4185
	data = nfs4_alloc_unlockdata(fl, ctx, lsp, seqid);
	if (data == NULL) {
		nfs_free_seqid(seqid);
		return ERR_PTR(-ENOMEM);
	}

4186 4187
	msg.rpc_argp = &data->arg;
	msg.rpc_resp = &data->res;
T
Trond Myklebust 已提交
4188 4189
	task_setup_data.callback_data = data;
	return rpc_run_task(&task_setup_data);
4190 4191
}

4192 4193
static int nfs4_proc_unlck(struct nfs4_state *state, int cmd, struct file_lock *request)
{
4194
	struct nfs_inode *nfsi = NFS_I(state->inode);
T
Trond Myklebust 已提交
4195
	struct nfs_seqid *seqid;
L
Linus Torvalds 已提交
4196
	struct nfs4_lock_state *lsp;
4197 4198
	struct rpc_task *task;
	int status = 0;
4199
	unsigned char fl_flags = request->fl_flags;
4200

4201
	status = nfs4_set_lock_state(state, request);
4202 4203
	/* Unlock _before_ we do the RPC call */
	request->fl_flags |= FL_EXISTS;
4204 4205 4206
	down_read(&nfsi->rwsem);
	if (do_vfs_lock(request->fl_file, request) == -ENOENT) {
		up_read(&nfsi->rwsem);
4207
		goto out;
4208 4209
	}
	up_read(&nfsi->rwsem);
4210
	if (status != 0)
4211 4212 4213 4214
		goto out;
	/* Is this a delegated lock? */
	if (test_bit(NFS_DELEGATED_STATE, &state->flags))
		goto out;
4215
	lsp = request->fl_u.nfs4_fl.owner;
4216
	seqid = nfs_alloc_seqid(&lsp->ls_seqid, GFP_KERNEL);
4217
	status = -ENOMEM;
T
Trond Myklebust 已提交
4218
	if (seqid == NULL)
4219
		goto out;
4220
	task = nfs4_do_unlck(request, nfs_file_open_context(request->fl_file), lsp, seqid);
4221 4222
	status = PTR_ERR(task);
	if (IS_ERR(task))
4223
		goto out;
4224
	status = nfs4_wait_for_completion_rpc_task(task);
4225
	rpc_put_task(task);
4226
out:
4227
	request->fl_flags = fl_flags;
L
Linus Torvalds 已提交
4228 4229 4230
	return status;
}

4231 4232 4233 4234 4235 4236
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;
4237
	unsigned long timestamp;
4238 4239
	int rpc_status;
	int cancelled;
4240
	struct nfs_server *server;
4241 4242 4243
};

static struct nfs4_lockdata *nfs4_alloc_lockdata(struct file_lock *fl,
4244 4245
		struct nfs_open_context *ctx, struct nfs4_lock_state *lsp,
		gfp_t gfp_mask)
L
Linus Torvalds 已提交
4246
{
4247 4248
	struct nfs4_lockdata *p;
	struct inode *inode = lsp->ls_state->inode;
L
Linus Torvalds 已提交
4249
	struct nfs_server *server = NFS_SERVER(inode);
4250

4251
	p = kzalloc(sizeof(*p), gfp_mask);
4252 4253 4254 4255 4256
	if (p == NULL)
		return NULL;

	p->arg.fh = NFS_FH(inode);
	p->arg.fl = &p->fl;
4257
	p->arg.open_seqid = nfs_alloc_seqid(&lsp->ls_state->owner->so_seqid, gfp_mask);
4258 4259
	if (p->arg.open_seqid == NULL)
		goto out_free;
4260
	p->arg.lock_seqid = nfs_alloc_seqid(&lsp->ls_seqid, gfp_mask);
4261
	if (p->arg.lock_seqid == NULL)
4262
		goto out_free_seqid;
4263
	p->arg.lock_stateid = &lsp->ls_stateid;
4264
	p->arg.lock_owner.clientid = server->nfs_client->cl_clientid;
4265
	p->arg.lock_owner.id = lsp->ls_id;
4266
	p->arg.lock_owner.s_dev = server->s_dev;
4267
	p->res.lock_seqid = p->arg.lock_seqid;
4268
	p->lsp = lsp;
4269
	p->server = server;
4270 4271 4272 4273
	atomic_inc(&lsp->ls_count);
	p->ctx = get_nfs_open_context(ctx);
	memcpy(&p->fl, fl, sizeof(p->fl));
	return p;
4274 4275
out_free_seqid:
	nfs_free_seqid(p->arg.open_seqid);
4276 4277 4278 4279 4280 4281 4282 4283 4284
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;
4285

4286
	dprintk("%s: begin!\n", __func__);
4287 4288
	if (nfs_wait_on_sequence(data->arg.lock_seqid, task) != 0)
		return;
4289 4290
	/* Do we need to do an open_to_lock_owner? */
	if (!(data->arg.lock_seqid->sequence->flags & NFS_SEQID_CONFIRMED)) {
4291 4292
		if (nfs_wait_on_sequence(data->arg.open_seqid, task) != 0)
			return;
4293 4294
		data->arg.open_stateid = &state->stateid;
		data->arg.new_lock_owner = 1;
4295
		data->res.open_seqid = data->arg.open_seqid;
4296 4297
	} else
		data->arg.new_lock_owner = 0;
4298
	data->timestamp = jiffies;
4299 4300
	if (nfs4_setup_sequence(data->server,
				&data->arg.seq_args,
4301 4302
				&data->res.seq_res, 1, task))
		return;
4303
	rpc_call_start(task);
4304
	dprintk("%s: done!, ret = %d\n", __func__, data->rpc_status);
4305 4306
}

4307 4308 4309 4310 4311 4312
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);
}

4313 4314 4315 4316
static void nfs4_lock_done(struct rpc_task *task, void *calldata)
{
	struct nfs4_lockdata *data = calldata;

4317
	dprintk("%s: begin!\n", __func__);
4318

4319 4320
	if (!nfs4_sequence_done(task, &data->res.seq_res))
		return;
4321

4322 4323 4324 4325 4326 4327 4328 4329 4330 4331 4332
	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;
4333
		renew_lease(NFS_SERVER(data->ctx->dentry->d_inode), data->timestamp);
4334 4335
	}
out:
4336
	dprintk("%s: done, ret = %d!\n", __func__, data->rpc_status);
4337 4338 4339 4340 4341 4342
}

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

4343
	dprintk("%s: begin!\n", __func__);
4344
	nfs_free_seqid(data->arg.open_seqid);
4345 4346 4347 4348 4349
	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))
4350
			rpc_put_task_async(task);
4351
		dprintk("%s: cancelling lock!\n", __func__);
4352 4353 4354 4355 4356
	} else
		nfs_free_seqid(data->arg.lock_seqid);
	nfs4_put_lock_state(data->lsp);
	put_nfs_open_context(data->ctx);
	kfree(data);
4357
	dprintk("%s: done!\n", __func__);
4358 4359 4360 4361 4362 4363 4364 4365
}

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,
};

4366 4367 4368 4369 4370 4371
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,
};

4372 4373 4374 4375 4376
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:
4377
		lsp->ls_seqid.flags &= ~NFS_SEQID_CONFIRMED;
4378 4379
		if (new_lock_owner != 0 ||
		   (lsp->ls_flags & NFS_LOCK_INITIALIZED) != 0)
4380
			nfs4_schedule_stateid_recovery(server, lsp->ls_state);
4381 4382 4383
		break;
	case -NFS4ERR_STALE_STATEID:
		lsp->ls_seqid.flags &= ~NFS_SEQID_CONFIRMED;
4384 4385
	case -NFS4ERR_EXPIRED:
		nfs4_schedule_lease_recovery(server->nfs_client);
4386 4387 4388
	};
}

4389
static int _nfs4_do_setlk(struct nfs4_state *state, int cmd, struct file_lock *fl, int recovery_type)
4390 4391 4392
{
	struct nfs4_lockdata *data;
	struct rpc_task *task;
4393 4394 4395 4396
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_LOCK],
		.rpc_cred = state->owner->so_cred,
	};
T
Trond Myklebust 已提交
4397 4398
	struct rpc_task_setup task_setup_data = {
		.rpc_client = NFS_CLIENT(state->inode),
4399
		.rpc_message = &msg,
T
Trond Myklebust 已提交
4400
		.callback_ops = &nfs4_lock_ops,
4401
		.workqueue = nfsiod_workqueue,
T
Trond Myklebust 已提交
4402 4403
		.flags = RPC_TASK_ASYNC,
	};
4404 4405
	int ret;

4406
	dprintk("%s: begin!\n", __func__);
4407
	data = nfs4_alloc_lockdata(fl, nfs_file_open_context(fl->fl_file),
4408 4409
			fl->fl_u.nfs4_fl.owner,
			recovery_type == NFS_LOCK_NEW ? GFP_KERNEL : GFP_NOFS);
4410 4411 4412 4413
	if (data == NULL)
		return -ENOMEM;
	if (IS_SETLKW(cmd))
		data->arg.block = 1;
4414 4415
	if (recovery_type > NFS_LOCK_NEW) {
		if (recovery_type == NFS_LOCK_RECLAIM)
4416
			data->arg.reclaim = NFS_LOCK_RECLAIM;
4417 4418
		task_setup_data.callback_ops = &nfs4_recover_lock_ops;
	}
4419 4420
	msg.rpc_argp = &data->arg;
	msg.rpc_resp = &data->res;
T
Trond Myklebust 已提交
4421 4422
	task_setup_data.callback_data = data;
	task = rpc_run_task(&task_setup_data);
4423
	if (IS_ERR(task))
4424 4425 4426 4427
		return PTR_ERR(task);
	ret = nfs4_wait_for_completion_rpc_task(task);
	if (ret == 0) {
		ret = data->rpc_status;
4428 4429 4430
		if (ret)
			nfs4_handle_setlk_error(data->server, data->lsp,
					data->arg.new_lock_owner, ret);
4431 4432
	} else
		data->cancelled = 1;
4433
	rpc_put_task(task);
4434
	dprintk("%s: done, ret = %d!\n", __func__, ret);
4435
	return ret;
L
Linus Torvalds 已提交
4436 4437 4438 4439
}

static int nfs4_lock_reclaim(struct nfs4_state *state, struct file_lock *request)
{
4440 4441 4442 4443 4444
	struct nfs_server *server = NFS_SERVER(state->inode);
	struct nfs4_exception exception = { };
	int err;

	do {
4445 4446 4447
		/* Cache the lock if possible... */
		if (test_bit(NFS_DELEGATED_STATE, &state->flags) != 0)
			return 0;
4448
		err = _nfs4_do_setlk(state, F_SETLK, request, NFS_LOCK_RECLAIM);
4449
		if (err != -NFS4ERR_DELAY)
4450 4451 4452 4453
			break;
		nfs4_handle_exception(server, err, &exception);
	} while (exception.retry);
	return err;
L
Linus Torvalds 已提交
4454 4455 4456 4457
}

static int nfs4_lock_expired(struct nfs4_state *state, struct file_lock *request)
{
4458 4459 4460 4461
	struct nfs_server *server = NFS_SERVER(state->inode);
	struct nfs4_exception exception = { };
	int err;

4462 4463 4464
	err = nfs4_set_lock_state(state, request);
	if (err != 0)
		return err;
4465
	do {
4466 4467
		if (test_bit(NFS_DELEGATED_STATE, &state->flags) != 0)
			return 0;
4468
		err = _nfs4_do_setlk(state, F_SETLK, request, NFS_LOCK_EXPIRED);
4469 4470 4471 4472 4473 4474 4475 4476
		switch (err) {
		default:
			goto out;
		case -NFS4ERR_GRACE:
		case -NFS4ERR_DELAY:
			nfs4_handle_exception(server, err, &exception);
			err = 0;
		}
4477
	} while (exception.retry);
4478
out:
4479
	return err;
L
Linus Torvalds 已提交
4480 4481
}

4482 4483 4484 4485 4486 4487 4488 4489 4490 4491 4492 4493 4494 4495
#if defined(CONFIG_NFS_V4_1)
static int nfs41_lock_expired(struct nfs4_state *state, struct file_lock *request)
{
	int status;
	struct nfs_server *server = NFS_SERVER(state->inode);

	status = nfs41_test_stateid(server, state);
	if (status == NFS_OK)
		return 0;
	nfs41_free_stateid(server, state);
	return nfs4_lock_expired(state, request);
}
#endif

L
Linus Torvalds 已提交
4496 4497
static int _nfs4_proc_setlk(struct nfs4_state *state, int cmd, struct file_lock *request)
{
4498
	struct nfs_inode *nfsi = NFS_I(state->inode);
4499
	unsigned char fl_flags = request->fl_flags;
4500
	int status = -ENOLCK;
L
Linus Torvalds 已提交
4501

4502 4503 4504
	if ((fl_flags & FL_POSIX) &&
			!test_bit(NFS_STATE_POSIX_LOCKS, &state->flags))
		goto out;
4505 4506 4507 4508
	/* Is this a delegated open? */
	status = nfs4_set_lock_state(state, request);
	if (status != 0)
		goto out;
4509 4510 4511 4512
	request->fl_flags |= FL_ACCESS;
	status = do_vfs_lock(request->fl_file, request);
	if (status < 0)
		goto out;
4513
	down_read(&nfsi->rwsem);
4514 4515 4516
	if (test_bit(NFS_DELEGATED_STATE, &state->flags)) {
		/* Yes: cache locks! */
		/* ...but avoid races with delegation recall... */
4517 4518 4519
		request->fl_flags = fl_flags & ~FL_SLEEP;
		status = do_vfs_lock(request->fl_file, request);
		goto out_unlock;
4520
	}
4521
	status = _nfs4_do_setlk(state, cmd, request, NFS_LOCK_NEW);
4522
	if (status != 0)
4523
		goto out_unlock;
4524
	/* Note: we always want to sleep here! */
4525
	request->fl_flags = fl_flags | FL_SLEEP;
4526
	if (do_vfs_lock(request->fl_file, request) < 0)
4527
		printk(KERN_WARNING "%s: VFS is out of sync with lock manager!\n", __func__);
4528
out_unlock:
4529
	up_read(&nfsi->rwsem);
4530 4531
out:
	request->fl_flags = fl_flags;
L
Linus Torvalds 已提交
4532 4533 4534 4535 4536 4537 4538 4539 4540
	return status;
}

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

	do {
4541 4542 4543
		err = _nfs4_proc_setlk(state, cmd, request);
		if (err == -NFS4ERR_DENIED)
			err = -EAGAIN;
L
Linus Torvalds 已提交
4544
		err = nfs4_handle_exception(NFS_SERVER(state->inode),
4545
				err, &exception);
L
Linus Torvalds 已提交
4546 4547 4548 4549 4550 4551 4552 4553 4554 4555 4556 4557 4558
	} 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 */
4559
	ctx = nfs_file_open_context(filp);
L
Linus Torvalds 已提交
4560 4561 4562 4563 4564
	state = ctx->state;

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

4565 4566 4567 4568 4569
	if (IS_GETLK(cmd)) {
		if (state != NULL)
			return nfs4_proc_getlk(state, F_GETLK, request);
		return 0;
	}
L
Linus Torvalds 已提交
4570 4571 4572 4573

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

4574 4575 4576 4577 4578
	if (request->fl_type == F_UNLCK) {
		if (state != NULL)
			return nfs4_proc_unlck(state, cmd, request);
		return 0;
	}
L
Linus Torvalds 已提交
4579

4580 4581
	if (state == NULL)
		return -ENOLCK;
L
Linus Torvalds 已提交
4582 4583 4584 4585 4586 4587 4588 4589 4590 4591 4592 4593
	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;
}

4594 4595 4596 4597 4598 4599 4600 4601 4602 4603
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 {
4604
		err = _nfs4_do_setlk(state, F_SETLK, fl, NFS_LOCK_NEW);
4605 4606 4607 4608 4609
		switch (err) {
			default:
				printk(KERN_ERR "%s: unhandled error %d.\n",
						__func__, err);
			case 0:
4610
			case -ESTALE:
4611 4612
				goto out;
			case -NFS4ERR_EXPIRED:
4613
				nfs4_schedule_stateid_recovery(server, state);
4614
			case -NFS4ERR_STALE_CLIENTID:
4615
			case -NFS4ERR_STALE_STATEID:
4616 4617
				nfs4_schedule_lease_recovery(server->nfs_client);
				goto out;
4618 4619 4620 4621 4622
			case -NFS4ERR_BADSESSION:
			case -NFS4ERR_BADSLOT:
			case -NFS4ERR_BAD_HIGH_SLOT:
			case -NFS4ERR_CONN_NOT_BOUND_TO_SESSION:
			case -NFS4ERR_DEADSESSION:
4623
				nfs4_schedule_session_recovery(server->nfs_client->cl_session);
4624
				goto out;
4625 4626 4627 4628 4629 4630 4631 4632
			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:
4633
				nfs4_schedule_stateid_recovery(server, state);
4634 4635
				err = 0;
				goto out;
4636 4637 4638 4639 4640 4641 4642 4643 4644
			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;
4645 4646 4647 4648 4649
			case -ENOMEM:
			case -NFS4ERR_DENIED:
				/* kill_proc(fl->fl_pid, SIGLOST, 1); */
				err = 0;
				goto out;
4650 4651 4652
			case -NFS4ERR_DELAY:
				break;
		}
4653 4654 4655 4656 4657
		err = nfs4_handle_exception(server, err, &exception);
	} while (exception.retry);
out:
	return err;
}
4658

4659 4660 4661 4662 4663 4664 4665 4666 4667 4668 4669 4670 4671 4672 4673 4674 4675 4676 4677 4678 4679 4680 4681
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;
4682
	args->lock_owner.id = lsp->ls_id;
4683
	args->lock_owner.s_dev = server->s_dev;
4684 4685 4686 4687
	msg.rpc_argp = args;
	rpc_call_async(server->client, &msg, 0, &nfs4_release_lockowner_ops, args);
}

4688 4689
#define XATTR_NAME_NFSV4_ACL "system.nfs4_acl"

4690 4691 4692
static int nfs4_xattr_set_nfs4_acl(struct dentry *dentry, const char *key,
				   const void *buf, size_t buflen,
				   int flags, int type)
4693
{
4694 4695
	if (strcmp(key, "") != 0)
		return -EINVAL;
4696

4697
	return nfs4_proc_set_acl(dentry->d_inode, buf, buflen);
4698 4699
}

4700 4701
static int nfs4_xattr_get_nfs4_acl(struct dentry *dentry, const char *key,
				   void *buf, size_t buflen, int type)
4702
{
4703 4704
	if (strcmp(key, "") != 0)
		return -EINVAL;
4705

4706
	return nfs4_proc_get_acl(dentry->d_inode, buf, buflen);
4707 4708
}

4709 4710 4711
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)
4712
{
4713
	size_t len = sizeof(XATTR_NAME_NFSV4_ACL);
4714

4715 4716
	if (!nfs4_server_supports_acls(NFS_SERVER(dentry->d_inode)))
		return 0;
4717 4718 4719

	if (list && len <= list_len)
		memcpy(list, XATTR_NAME_NFSV4_ACL, len);
4720
	return len;
4721 4722
}

4723 4724 4725
/*
 * nfs_fhget will use either the mounted_on_fileid or the fileid
 */
4726 4727
static void nfs_fixup_referral_attributes(struct nfs_fattr *fattr)
{
4728 4729 4730 4731
	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)))
4732 4733 4734 4735 4736 4737 4738 4739
		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;
}

4740
int nfs4_proc_fs_locations(struct inode *dir, const struct qstr *name,
4741
		struct nfs4_fs_locations *fs_locations, struct page *page)
4742 4743 4744
{
	struct nfs_server *server = NFS_SERVER(dir);
	u32 bitmask[2] = {
4745
		[0] = FATTR4_WORD0_FSID | FATTR4_WORD0_FS_LOCATIONS,
4746 4747 4748
	};
	struct nfs4_fs_locations_arg args = {
		.dir_fh = NFS_FH(dir),
4749
		.name = name,
4750 4751 4752
		.page = page,
		.bitmask = bitmask,
	};
B
Benny Halevy 已提交
4753 4754 4755
	struct nfs4_fs_locations_res res = {
		.fs_locations = fs_locations,
	};
4756 4757 4758
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_FS_LOCATIONS],
		.rpc_argp = &args,
B
Benny Halevy 已提交
4759
		.rpc_resp = &res,
4760 4761 4762
	};
	int status;

4763
	dprintk("%s: start\n", __func__);
4764 4765 4766 4767 4768 4769 4770 4771

	/* 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;

4772
	nfs_fattr_init(&fs_locations->fattr);
4773
	fs_locations->server = server;
4774
	fs_locations->nlocations = 0;
4775
	status = nfs4_call_sync(server->client, server, &msg, &args.seq_args, &res.seq_res, 0);
4776
	dprintk("%s: returned status = %d\n", __func__, status);
4777 4778 4779
	return status;
}

B
Bryan Schumaker 已提交
4780 4781 4782 4783 4784 4785 4786 4787 4788 4789 4790 4791 4792 4793 4794 4795 4796 4797 4798 4799 4800 4801 4802 4803 4804 4805 4806 4807 4808 4809 4810 4811 4812 4813
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;
}

4814
#ifdef CONFIG_NFS_V4_1
4815 4816 4817 4818 4819 4820 4821 4822 4823 4824 4825 4826 4827 4828 4829 4830 4831 4832 4833
/*
 * 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;
}

4834 4835 4836 4837 4838 4839 4840 4841 4842 4843
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
4844 4845 4846 4847 4848 4849 4850 4851
/*
 * 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.
 */
4852
int nfs4_proc_exchange_id(struct nfs_client *clp, struct rpc_cred *cred)
B
Benny Halevy 已提交
4853 4854 4855 4856
{
	nfs4_verifier verifier;
	struct nfs41_exchange_id_args args = {
		.client = clp,
4857
		.flags = EXCHGID4_FLAG_SUPP_MOVED_REFER,
B
Benny Halevy 已提交
4858 4859 4860 4861 4862 4863 4864 4865 4866 4867 4868 4869 4870 4871 4872
	};
	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);
4873

B
Benny Halevy 已提交
4874 4875 4876 4877 4878
	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;

4879 4880 4881 4882 4883 4884
	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
4885

4886 4887 4888 4889
	res.server_scope = kzalloc(sizeof(struct server_scope), GFP_KERNEL);
	if (unlikely(!res.server_scope))
		return -ENOMEM;

4890
	status = rpc_call_sync(clp->cl_rpcclient, &msg, RPC_TASK_TIMEOUT);
4891 4892
	if (!status)
		status = nfs4_check_cl_exchange_flags(clp->cl_exchange_flags);
4893 4894 4895 4896 4897 4898 4899 4900 4901 4902 4903 4904 4905 4906 4907 4908 4909 4910

	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;
		}

		if (!clp->server_scope)
			clp->server_scope = res.server_scope;
		else
			kfree(res.server_scope);
	}

B
Benny Halevy 已提交
4911 4912 4913 4914
	dprintk("<-- %s status= %d\n", __func__, status);
	return status;
}

A
Andy Adamson 已提交
4915 4916 4917 4918 4919 4920 4921 4922 4923 4924 4925 4926 4927 4928
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__);
4929
	rpc_task_set_priority(task, RPC_PRIORITY_PRIVILEGED);
A
Andy Adamson 已提交
4930 4931 4932
	/* just setup sequence, do not trigger session recovery
	   since we're invoked within one */
	ret = nfs41_setup_sequence(data->clp->cl_session,
4933 4934
				   &data->args->la_seq_args,
				   &data->res->lr_seq_res, 0, task);
A
Andy Adamson 已提交
4935 4936 4937 4938 4939 4940 4941 4942 4943 4944 4945 4946 4947 4948 4949 4950

	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__);
4951 4952
	if (!nfs41_sequence_done(task, &data->res->lr_seq_res))
		return;
A
Andy Adamson 已提交
4953 4954 4955 4956 4957 4958
	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;
4959 4960
		/* fall through */
	case -NFS4ERR_RETRY_UNCACHED_REP:
4961
		rpc_restart_call_prepare(task);
A
Andy Adamson 已提交
4962 4963 4964 4965 4966 4967 4968 4969 4970 4971 4972 4973 4974 4975 4976 4977 4978 4979 4980 4981 4982 4983 4984 4985 4986 4987 4988 4989 4990 4991 4992
		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,
4993 4994
		.callback_data = &data,
		.flags = RPC_TASK_TIMEOUT,
A
Andy Adamson 已提交
4995 4996 4997 4998 4999 5000 5001 5002 5003 5004 5005 5006 5007 5008 5009 5010 5011
	};
	int status;

	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;
}

5012 5013 5014
/*
 * Reset a slot table
 */
A
Andy Adamson 已提交
5015 5016
static int nfs4_reset_slot_table(struct nfs4_slot_table *tbl, u32 max_reqs,
				 int ivalue)
5017
{
A
Andy Adamson 已提交
5018
	struct nfs4_slot *new = NULL;
5019
	int i;
5020 5021
	int ret = 0;

A
Andy Adamson 已提交
5022 5023
	dprintk("--> %s: max_reqs=%u, tbl->max_slots %d\n", __func__,
		max_reqs, tbl->max_slots);
5024

A
Andy Adamson 已提交
5025 5026 5027 5028
	/* 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),
5029
			      GFP_NOFS);
A
Andy Adamson 已提交
5030 5031 5032 5033
		if (!new)
			goto out;
		ret = 0;
		kfree(tbl->slots);
5034 5035
	}
	spin_lock(&tbl->slot_tbl_lock);
A
Andy Adamson 已提交
5036 5037 5038 5039 5040
	if (new) {
		tbl->slots = new;
		tbl->max_slots = max_reqs;
	}
	for (i = 0; i < tbl->max_slots; ++i)
5041
		tbl->slots[i].seq_nr = ivalue;
5042 5043 5044 5045 5046 5047 5048 5049
	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;
}

5050 5051 5052 5053 5054 5055 5056 5057 5058 5059 5060 5061 5062 5063
/* 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
5064 5065 5066
/*
 * Initialize slot table
 */
5067 5068
static int nfs4_init_slot_table(struct nfs4_slot_table *tbl,
		int max_slots, int ivalue)
A
Andy Adamson 已提交
5069 5070 5071 5072 5073 5074
{
	struct nfs4_slot *slot;
	int ret = -ENOMEM;

	BUG_ON(max_slots > NFS4_MAX_SLOT_TABLE);

5075
	dprintk("--> %s: max_reqs=%u\n", __func__, max_slots);
A
Andy Adamson 已提交
5076

5077
	slot = kcalloc(max_slots, sizeof(struct nfs4_slot), GFP_NOFS);
A
Andy Adamson 已提交
5078 5079 5080 5081 5082 5083 5084 5085 5086 5087 5088 5089 5090 5091 5092 5093
	if (!slot)
		goto out;
	ret = 0;

	spin_lock(&tbl->slot_tbl_lock);
	tbl->max_slots = max_slots;
	tbl->slots = slot;
	tbl->highest_used_slotid = -1;  /* no slot is currently used */
	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;
}

5094
/*
5095
 * Initialize or reset the forechannel and backchannel tables
5096
 */
5097
static int nfs4_setup_session_slot_tables(struct nfs4_session *ses)
5098
{
5099
	struct nfs4_slot_table *tbl;
5100
	int status;
5101

5102 5103 5104
	dprintk("--> %s\n", __func__);
	/* Fore channel */
	tbl = &ses->fc_slot_table;
5105
	if (tbl->slots == NULL) {
5106 5107 5108 5109 5110
		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);
5111 5112 5113
		if (status)
			return status;
	}
5114 5115
	/* Back channel */
	tbl = &ses->bc_slot_table;
5116
	if (tbl->slots == NULL) {
5117
		status = nfs4_init_slot_table(tbl, ses->bc_attrs.max_reqs, 0);
5118
		if (status)
5119 5120 5121 5122 5123
			/* 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);
5124
	return status;
5125 5126 5127 5128 5129 5130 5131
}

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

5132
	session = kzalloc(sizeof(struct nfs4_session), GFP_NOFS);
5133 5134
	if (!session)
		return NULL;
5135

5136
	tbl = &session->fc_slot_table;
5137
	tbl->highest_used_slotid = -1;
5138
	spin_lock_init(&tbl->slot_tbl_lock);
5139
	rpc_init_priority_wait_queue(&tbl->slot_tbl_waitq, "ForeChannel Slot table");
5140
	init_completion(&tbl->complete);
5141 5142

	tbl = &session->bc_slot_table;
5143
	tbl->highest_used_slotid = -1;
5144 5145
	spin_lock_init(&tbl->slot_tbl_lock);
	rpc_init_wait_queue(&tbl->slot_tbl_waitq, "BackChannel Slot table");
5146
	init_completion(&tbl->complete);
5147

5148 5149
	session->session_state = 1<<NFS4_SESSION_INITING;

5150 5151 5152 5153 5154 5155
	session->clp = clp;
	return session;
}

void nfs4_destroy_session(struct nfs4_session *session)
{
5156 5157 5158 5159 5160
	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);
5161
	nfs4_destroy_slot_tables(session);
5162 5163 5164
	kfree(session);
}

A
Andy Adamson 已提交
5165 5166 5167 5168 5169 5170 5171 5172 5173 5174 5175 5176 5177 5178 5179 5180 5181 5182 5183 5184 5185 5186 5187 5188 5189 5190
/*
 * 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;
	args->fc_attrs.max_reqs = session->clp->cl_rpcclient->cl_xprt->max_reqs;

	dprintk("%s: Fore Channel : max_rqst_sz=%u max_resp_sz=%u "
5191
		"max_ops=%u max_reqs=%u\n",
A
Andy Adamson 已提交
5192 5193
		__func__,
		args->fc_attrs.max_rqst_sz, args->fc_attrs.max_resp_sz,
5194
		args->fc_attrs.max_ops, args->fc_attrs.max_reqs);
A
Andy Adamson 已提交
5195 5196 5197 5198 5199 5200 5201 5202 5203 5204 5205 5206 5207 5208 5209 5210

	/* 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);
}

5211
static int nfs4_verify_fore_channel_attrs(struct nfs41_create_session_args *args, struct nfs4_session *session)
5212
{
5213 5214 5215 5216 5217 5218 5219 5220 5221 5222 5223 5224 5225 5226 5227 5228
	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;
	return 0;
5229 5230
}

5231 5232 5233 5234
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;
5235

5236 5237 5238 5239 5240 5241 5242 5243 5244 5245 5246 5247 5248
	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: */
	if (rcvd->max_ops  == 0)
		return -EINVAL;
	if (rcvd->max_reqs == 0)
		return -EINVAL;
	return 0;
}
5249 5250 5251 5252

static int nfs4_verify_channel_attrs(struct nfs41_create_session_args *args,
				     struct nfs4_session *session)
{
5253
	int ret;
5254

5255 5256 5257 5258
	ret = nfs4_verify_fore_channel_attrs(args, session);
	if (ret)
		return ret;
	return nfs4_verify_back_channel_attrs(args, session);
5259 5260
}

A
Andy Adamson 已提交
5261 5262 5263 5264 5265 5266 5267 5268 5269 5270 5271 5272 5273 5274 5275 5276 5277 5278
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);
5279
	args.flags = (SESSION4_PERSIST | SESSION4_BACK_CHAN);
A
Andy Adamson 已提交
5280

5281
	status = rpc_call_sync(session->clp->cl_rpcclient, &msg, RPC_TASK_TIMEOUT);
A
Andy Adamson 已提交
5282

5283 5284 5285
	if (!status)
		/* Verify the session's negotiated channel_attrs values */
		status = nfs4_verify_channel_attrs(&args, session);
A
Andy Adamson 已提交
5286 5287 5288 5289 5290 5291 5292 5293 5294 5295 5296 5297 5298
	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.
 */
5299
int nfs4_proc_create_session(struct nfs_client *clp)
A
Andy Adamson 已提交
5300 5301 5302 5303 5304 5305 5306
{
	int status;
	unsigned *ptr;
	struct nfs4_session *session = clp->cl_session;

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

5307
	status = _nfs4_proc_create_session(clp);
A
Andy Adamson 已提交
5308 5309 5310
	if (status)
		goto out;

5311 5312 5313
	/* 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
5314 5315 5316 5317 5318 5319 5320 5321 5322 5323 5324
	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
5325 5326 5327 5328 5329 5330 5331 5332 5333 5334 5335 5336 5337 5338 5339 5340 5341 5342 5343
/*
 * 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;
5344
	status = rpc_call_sync(session->clp->cl_rpcclient, &msg, RPC_TASK_TIMEOUT);
A
Andy Adamson 已提交
5345 5346 5347 5348 5349 5350 5351 5352 5353 5354

	if (status)
		printk(KERN_WARNING
			"Got error %d from the server on DESTROY_SESSION. "
			"Session has been destroyed regardless...\n", status);

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

5355 5356 5357
int nfs4_init_session(struct nfs_server *server)
{
	struct nfs_client *clp = server->nfs_client;
5358
	struct nfs4_session *session;
5359
	unsigned int rsize, wsize;
5360 5361 5362 5363 5364
	int ret;

	if (!nfs4_has_session(clp))
		return 0;

5365 5366 5367 5368
	session = clp->cl_session;
	if (!test_and_clear_bit(NFS4_SESSION_INITING, &session->session_state))
		return 0;

5369 5370 5371 5372 5373 5374 5375 5376 5377
	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;
5378

5379 5380 5381 5382 5383 5384
	ret = nfs4_recover_expired_lease(server);
	if (!ret)
		ret = nfs4_check_client_ready(clp);
	return ret;
}

A
Andy Adamson 已提交
5385 5386 5387 5388 5389 5390 5391 5392 5393 5394 5395 5396 5397 5398 5399 5400 5401 5402 5403 5404 5405
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
5406 5407 5408
/*
 * Renew the cl_session lease.
 */
5409 5410 5411 5412 5413 5414
struct nfs4_sequence_data {
	struct nfs_client *clp;
	struct nfs4_sequence_args args;
	struct nfs4_sequence_res res;
};

5415 5416
static void nfs41_sequence_release(void *data)
{
5417 5418
	struct nfs4_sequence_data *calldata = data;
	struct nfs_client *clp = calldata->clp;
5419

5420 5421 5422
	if (atomic_read(&clp->cl_count) > 1)
		nfs4_schedule_state_renewal(clp);
	nfs_put_client(clp);
5423
	kfree(calldata);
5424 5425
}

5426 5427 5428 5429 5430 5431 5432
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:
5433
		nfs4_schedule_lease_recovery(clp);
5434 5435 5436 5437
	}
	return 0;
}

5438
static void nfs41_sequence_call_done(struct rpc_task *task, void *data)
A
Andy Adamson 已提交
5439
{
5440 5441
	struct nfs4_sequence_data *calldata = data;
	struct nfs_client *clp = calldata->clp;
A
Andy Adamson 已提交
5442

5443 5444
	if (!nfs41_sequence_done(task, task->tk_msg.rpc_resp))
		return;
A
Andy Adamson 已提交
5445 5446 5447

	if (task->tk_status < 0) {
		dprintk("%s ERROR %d\n", __func__, task->tk_status);
5448 5449
		if (atomic_read(&clp->cl_count) == 1)
			goto out;
A
Andy Adamson 已提交
5450

5451 5452
		if (nfs41_sequence_handle_errors(task, clp) == -EAGAIN) {
			rpc_restart_call_prepare(task);
A
Andy Adamson 已提交
5453 5454 5455 5456
			return;
		}
	}
	dprintk("%s rpc_cred %p\n", __func__, task->tk_msg.rpc_cred);
5457
out:
A
Andy Adamson 已提交
5458 5459 5460 5461 5462
	dprintk("<-- %s\n", __func__);
}

static void nfs41_sequence_prepare(struct rpc_task *task, void *data)
{
5463 5464
	struct nfs4_sequence_data *calldata = data;
	struct nfs_client *clp = calldata->clp;
A
Andy Adamson 已提交
5465 5466 5467 5468 5469 5470
	struct nfs4_sequence_args *args;
	struct nfs4_sequence_res *res;

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

5471
	if (nfs41_setup_sequence(clp->cl_session, args, res, 0, task))
A
Andy Adamson 已提交
5472 5473 5474 5475 5476 5477 5478
		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,
5479
	.rpc_release = nfs41_sequence_release,
A
Andy Adamson 已提交
5480 5481
};

5482
static struct rpc_task *_nfs41_proc_sequence(struct nfs_client *clp, struct rpc_cred *cred)
A
Andy Adamson 已提交
5483
{
5484
	struct nfs4_sequence_data *calldata;
A
Andy Adamson 已提交
5485 5486 5487 5488
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_SEQUENCE],
		.rpc_cred = cred,
	};
5489 5490 5491 5492 5493 5494
	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
5495

5496
	if (!atomic_inc_not_zero(&clp->cl_count))
5497
		return ERR_PTR(-EIO);
5498
	calldata = kzalloc(sizeof(*calldata), GFP_NOFS);
5499
	if (calldata == NULL) {
5500
		nfs_put_client(clp);
5501
		return ERR_PTR(-ENOMEM);
A
Andy Adamson 已提交
5502
	}
5503 5504 5505
	msg.rpc_argp = &calldata->args;
	msg.rpc_resp = &calldata->res;
	calldata->clp = clp;
5506
	task_setup_data.callback_data = calldata;
A
Andy Adamson 已提交
5507

5508 5509 5510
	return rpc_run_task(&task_setup_data);
}

5511
static int nfs41_proc_async_sequence(struct nfs_client *clp, struct rpc_cred *cred, unsigned renew_flags)
5512 5513 5514 5515
{
	struct rpc_task *task;
	int ret = 0;

5516 5517
	if ((renew_flags & NFS4_RENEW_TIMEOUT) == 0)
		return 0;
5518 5519 5520 5521
	task = _nfs41_proc_sequence(clp, cred);
	if (IS_ERR(task))
		ret = PTR_ERR(task);
	else
5522
		rpc_put_task_async(task);
5523 5524 5525 5526 5527 5528 5529 5530 5531 5532 5533 5534 5535 5536 5537
	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);
5538 5539 5540 5541 5542
	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);
5543
		ret = task->tk_status;
5544
	}
5545 5546 5547 5548
	rpc_put_task(task);
out:
	dprintk("<-- %s status=%d\n", __func__, ret);
	return ret;
A
Andy Adamson 已提交
5549 5550
}

5551 5552 5553 5554 5555 5556 5557 5558 5559 5560
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;

5561
	rpc_task_set_priority(task, RPC_PRIORITY_PRIVILEGED);
5562 5563
	if (nfs41_setup_sequence(calldata->clp->cl_session,
				&calldata->arg.seq_args,
5564 5565 5566 5567 5568 5569
				&calldata->res.seq_res, 0, task))
		return;

	rpc_call_start(task);
}

5570 5571 5572 5573 5574 5575 5576 5577 5578
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);
5579 5580
		/* fall through */
	case -NFS4ERR_RETRY_UNCACHED_REP:
5581 5582
		return -EAGAIN;
	default:
5583
		nfs4_schedule_lease_recovery(clp);
5584 5585 5586 5587
	}
	return 0;
}

5588 5589 5590 5591 5592 5593 5594
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__);
5595 5596
	if (!nfs41_sequence_done(task, res))
		return;
5597

5598 5599 5600 5601
	if (nfs41_reclaim_complete_handle_errors(task, clp) == -EAGAIN) {
		rpc_restart_call_prepare(task);
		return;
	}
5602 5603 5604 5605 5606 5607 5608 5609 5610 5611 5612 5613 5614 5615 5616 5617 5618 5619 5620 5621 5622 5623 5624 5625 5626 5627 5628 5629 5630 5631 5632 5633 5634 5635 5636
	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__);
5637
	calldata = kzalloc(sizeof(*calldata), GFP_NOFS);
5638 5639 5640 5641 5642 5643 5644 5645 5646
	if (calldata == NULL)
		goto out;
	calldata->clp = clp;
	calldata->arg.one_fs = 0;

	msg.rpc_argp = &calldata->arg;
	msg.rpc_resp = &calldata->res;
	task_setup_data.callback_data = calldata;
	task = rpc_run_task(&task_setup_data);
5647
	if (IS_ERR(task)) {
5648
		status = PTR_ERR(task);
5649 5650
		goto out;
	}
5651 5652 5653
	status = nfs4_wait_for_completion_rpc_task(task);
	if (status == 0)
		status = task->tk_status;
5654
	rpc_put_task(task);
5655
	return 0;
5656 5657 5658 5659
out:
	dprintk("<-- %s status=%d\n", __func__, status);
	return status;
}
5660 5661 5662 5663 5664

static void
nfs4_layoutget_prepare(struct rpc_task *task, void *calldata)
{
	struct nfs4_layoutget *lgp = calldata;
F
Fred Isaman 已提交
5665
	struct nfs_server *server = NFS_SERVER(lgp->args.inode);
5666 5667

	dprintk("--> %s\n", __func__);
F
Fred Isaman 已提交
5668 5669 5670 5671 5672
	/* 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.
	 */
5673 5674 5675
	if (nfs4_setup_sequence(server, &lgp->args.seq_args,
				&lgp->res.seq_res, 0, task))
		return;
5676 5677 5678 5679 5680 5681
	if (pnfs_choose_layoutget_stateid(&lgp->args.stateid,
					  NFS_I(lgp->args.inode)->layout,
					  lgp->args.ctx->state)) {
		rpc_exit(task, NFS4_OK);
		return;
	}
5682 5683 5684 5685 5686 5687 5688 5689 5690 5691 5692 5693 5694 5695 5696 5697 5698 5699 5700 5701 5702 5703 5704 5705 5706 5707 5708 5709 5710 5711 5712 5713 5714 5715 5716 5717 5718 5719 5720 5721 5722 5723 5724 5725 5726 5727 5728 5729 5730 5731 5732 5733 5734 5735 5736 5737 5738 5739 5740 5741 5742 5743 5744 5745 5746
	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__);

5747
	lgp->res.layoutp = &lgp->args.layout;
5748 5749 5750 5751 5752
	lgp->res.seq_res.sr_slot = NULL;
	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
5753 5754 5755 5756
	if (status == 0)
		status = task->tk_status;
	if (status == 0)
		status = pnfs_layout_process(lgp);
5757 5758 5759 5760 5761
	rpc_put_task(task);
	dprintk("<-- %s status=%d\n", __func__, status);
	return status;
}

B
Benny Halevy 已提交
5762 5763 5764 5765 5766 5767 5768 5769 5770 5771 5772 5773 5774 5775 5776 5777
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,
				&lrp->res.seq_res, 0, task))
		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
5778
	struct pnfs_layout_hdr *lo = lrp->args.layout;
B
Benny Halevy 已提交
5779 5780 5781 5782 5783 5784 5785 5786

	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) {
5787
		rpc_restart_call_prepare(task);
B
Benny Halevy 已提交
5788 5789
		return;
	}
5790
	spin_lock(&lo->plh_inode->i_lock);
B
Benny Halevy 已提交
5791 5792 5793 5794 5795 5796
	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));
	}
5797 5798
	lo->plh_block_lgets--;
	spin_unlock(&lo->plh_inode->i_lock);
B
Benny Halevy 已提交
5799 5800 5801 5802 5803 5804 5805 5806
	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
5807
	put_layout_hdr(lrp->args.layout);
B
Benny Halevy 已提交
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 5837 5838 5839 5840 5841 5842 5843
	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__);
	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
5844 5845 5846 5847 5848 5849 5850 5851 5852 5853 5854 5855 5856 5857 5858 5859 5860 5861 5862 5863 5864 5865 5866 5867 5868 5869 5870 5871 5872 5873 5874 5875 5876 5877 5878 5879 5880 5881 5882 5883 5884 5885 5886 5887 5888 5889 5890 5891
/*
 * 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);

5892 5893 5894 5895 5896 5897 5898 5899 5900 5901 5902 5903 5904 5905 5906 5907 5908
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__);
5909
	status = nfs4_call_sync(server->client, server, &msg, &args.seq_args, &res.seq_res, 0);
5910 5911 5912 5913 5914 5915 5916 5917 5918 5919 5920 5921 5922 5923 5924 5925 5926 5927 5928
	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
5929 5930 5931 5932 5933 5934 5935 5936 5937 5938 5939 5940 5941 5942 5943 5944 5945 5946 5947 5948 5949 5950 5951 5952 5953 5954 5955 5956 5957
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,
				&data->res.seq_res, 1, task))
		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) {
5958
		rpc_restart_call_prepare(task);
A
Andy Adamson 已提交
5959 5960 5961 5962 5963 5964 5965 5966 5967 5968 5969
		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;
5970
	struct pnfs_layout_segment *lseg, *tmp;
P
Peng Tao 已提交
5971
	unsigned long *bitlock = &NFS_I(data->args.inode)->flags;
A
Andy Adamson 已提交
5972

A
Andy Adamson 已提交
5973
	pnfs_cleanup_layoutcommit(data);
A
Andy Adamson 已提交
5974
	/* Matched by references in pnfs_set_layoutcommit */
5975 5976 5977 5978 5979 5980
	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
5981 5982 5983 5984 5985

	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
5986 5987 5988 5989 5990 5991 5992 5993 5994 5995 5996
	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
5997
nfs4_proc_layoutcommit(struct nfs4_layoutcommit_data *data, bool sync)
A
Andy Adamson 已提交
5998 5999 6000 6001 6002 6003 6004 6005 6006 6007 6008 6009 6010 6011 6012 6013 6014 6015 6016 6017 6018 6019 6020 6021 6022 6023 6024
{
	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);

	task = rpc_run_task(&task_setup_data);
	if (IS_ERR(task))
		return PTR_ERR(task);
6025
	if (sync == false)
A
Andy Adamson 已提交
6026 6027 6028 6029 6030 6031 6032 6033 6034 6035
		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;
}
6036 6037 6038 6039 6040 6041 6042 6043 6044 6045 6046 6047 6048 6049 6050 6051 6052 6053 6054 6055 6056 6057 6058 6059 6060 6061 6062 6063 6064 6065 6066 6067 6068 6069 6070 6071 6072 6073 6074 6075 6076 6077 6078 6079 6080 6081 6082 6083 6084 6085 6086 6087 6088 6089 6090 6091 6092 6093 6094 6095 6096 6097 6098 6099 6100 6101 6102 6103 6104 6105 6106 6107 6108 6109 6110 6111 6112 6113 6114

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;
}
B
Bryan Schumaker 已提交
6115 6116 6117 6118 6119 6120 6121 6122 6123 6124 6125 6126 6127 6128 6129 6130 6131 6132 6133 6134 6135 6136 6137 6138 6139 6140 6141 6142
static int _nfs41_test_stateid(struct nfs_server *server, struct nfs4_state *state)
{
	int status;
	struct nfs41_test_stateid_args args = {
		.stateid = &state->stateid,
	};
	struct nfs41_test_stateid_res res;
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_TEST_STATEID],
		.rpc_argp = &args,
		.rpc_resp = &res,
	};
	args.seq_args.sa_session = res.seq_res.sr_session = NULL;
	status = nfs4_call_sync_sequence(server->client, server, &msg, &args.seq_args, &res.seq_res, 0, 1);
	return status;
}

static int nfs41_test_stateid(struct nfs_server *server, struct nfs4_state *state)
{
	struct nfs4_exception exception = { };
	int err;
	do {
		err = nfs4_handle_exception(server,
				_nfs41_test_stateid(server, state),
				&exception);
	} while (exception.retry);
	return err;
}
B
Bryan Schumaker 已提交
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

static int _nfs4_free_stateid(struct nfs_server *server, struct nfs4_state *state)
{
	int status;
	struct nfs41_free_stateid_args args = {
		.stateid = &state->stateid,
	};
	struct nfs41_free_stateid_res res;
	struct rpc_message msg = {
		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_FREE_STATEID],
		.rpc_argp = &args,
		.rpc_resp = &res,
	};

	args.seq_args.sa_session = res.seq_res.sr_session = NULL;
	status = nfs4_call_sync_sequence(server->client, server, &msg, &args.seq_args, &res.seq_res, 0, 1);
	return status;
}

static int nfs41_free_stateid(struct nfs_server *server, struct nfs4_state *state)
{
	struct nfs4_exception exception = { };
	int err;
	do {
		err = nfs4_handle_exception(server,
				_nfs4_free_stateid(server, state),
				&exception);
	} while (exception.retry);
	return err;
}
6173 6174
#endif /* CONFIG_NFS_V4_1 */

6175
struct nfs4_state_recovery_ops nfs40_reboot_recovery_ops = {
6176
	.owner_flag_bit = NFS_OWNER_RECLAIM_REBOOT,
6177
	.state_flag_bit	= NFS_STATE_RECLAIM_REBOOT,
L
Linus Torvalds 已提交
6178 6179
	.recover_open	= nfs4_open_reclaim,
	.recover_lock	= nfs4_lock_reclaim,
6180
	.establish_clid = nfs4_init_clientid,
6181
	.get_clid_cred	= nfs4_get_setclientid_cred,
L
Linus Torvalds 已提交
6182 6183
};

6184 6185 6186 6187 6188 6189
#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,
6190
	.establish_clid = nfs41_init_clientid,
6191
	.get_clid_cred	= nfs4_get_exchange_id_cred,
6192
	.reclaim_complete = nfs41_proc_reclaim_complete,
6193 6194 6195 6196 6197 6198 6199 6200 6201
};
#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,
6202
	.get_clid_cred	= nfs4_get_setclientid_cred,
6203 6204 6205 6206
};

#if defined(CONFIG_NFS_V4_1)
struct nfs4_state_recovery_ops nfs41_nograce_recovery_ops = {
6207
	.owner_flag_bit = NFS_OWNER_RECLAIM_NOGRACE,
6208
	.state_flag_bit	= NFS_STATE_RECLAIM_NOGRACE,
6209 6210
	.recover_open	= nfs41_open_expired,
	.recover_lock	= nfs41_lock_expired,
6211
	.establish_clid = nfs41_init_clientid,
6212
	.get_clid_cred	= nfs4_get_exchange_id_cred,
L
Linus Torvalds 已提交
6213
};
6214
#endif /* CONFIG_NFS_V4_1 */
L
Linus Torvalds 已提交
6215

B
Benny Halevy 已提交
6216 6217
struct nfs4_state_maintenance_ops nfs40_state_renewal_ops = {
	.sched_state_renewal = nfs4_proc_async_renew,
6218
	.get_state_renewal_cred_locked = nfs4_get_renew_cred_locked,
6219
	.renew_lease = nfs4_proc_renew,
B
Benny Halevy 已提交
6220 6221 6222 6223 6224
};

#if defined(CONFIG_NFS_V4_1)
struct nfs4_state_maintenance_ops nfs41_state_renewal_ops = {
	.sched_state_renewal = nfs41_proc_async_sequence,
6225
	.get_state_renewal_cred_locked = nfs4_get_machine_cred_locked,
6226
	.renew_lease = nfs4_proc_sequence,
B
Benny Halevy 已提交
6227 6228 6229
};
#endif

6230 6231 6232
static const struct nfs4_minor_version_ops nfs_v4_0_minor_ops = {
	.minor_version = 0,
	.call_sync = _nfs4_call_sync,
6233
	.validate_stateid = nfs4_validate_delegation_stateid,
6234
	.find_root_sec = nfs4_find_root_sec,
6235 6236 6237
	.reboot_recovery_ops = &nfs40_reboot_recovery_ops,
	.nograce_recovery_ops = &nfs40_nograce_recovery_ops,
	.state_renewal_ops = &nfs40_state_renewal_ops,
6238 6239 6240 6241 6242 6243
};

#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,
6244
	.validate_stateid = nfs41_validate_delegation_stateid,
6245
	.find_root_sec = nfs41_find_root_sec,
6246 6247 6248
	.reboot_recovery_ops = &nfs41_reboot_recovery_ops,
	.nograce_recovery_ops = &nfs41_nograce_recovery_ops,
	.state_renewal_ops = &nfs41_state_renewal_ops,
6249 6250 6251 6252 6253 6254 6255 6256 6257 6258
};
#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
};

6259
static const struct inode_operations nfs4_file_inode_operations = {
6260 6261 6262
	.permission	= nfs_permission,
	.getattr	= nfs_getattr,
	.setattr	= nfs_setattr,
6263 6264 6265 6266
	.getxattr	= generic_getxattr,
	.setxattr	= generic_setxattr,
	.listxattr	= generic_listxattr,
	.removexattr	= generic_removexattr,
6267 6268
};

D
David Howells 已提交
6269
const struct nfs_rpc_ops nfs_v4_clientops = {
L
Linus Torvalds 已提交
6270 6271 6272
	.version	= 4,			/* protocol version */
	.dentry_ops	= &nfs4_dentry_operations,
	.dir_inode_ops	= &nfs4_dir_inode_operations,
6273
	.file_inode_ops	= &nfs4_file_inode_operations,
6274
	.file_ops	= &nfs4_file_operations,
L
Linus Torvalds 已提交
6275 6276 6277 6278 6279 6280 6281 6282 6283 6284 6285
	.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,
6286 6287
	.rename_setup	= nfs4_proc_rename_setup,
	.rename_done	= nfs4_proc_rename_done,
L
Linus Torvalds 已提交
6288 6289 6290 6291 6292 6293 6294 6295 6296
	.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,
6297
	.set_capabilities = nfs4_server_capabilities,
L
Linus Torvalds 已提交
6298 6299
	.decode_dirent	= nfs4_decode_dirent,
	.read_setup	= nfs4_proc_read_setup,
T
Trond Myklebust 已提交
6300
	.read_done	= nfs4_read_done,
L
Linus Torvalds 已提交
6301
	.write_setup	= nfs4_proc_write_setup,
6302
	.write_done	= nfs4_write_done,
L
Linus Torvalds 已提交
6303
	.commit_setup	= nfs4_proc_commit_setup,
6304
	.commit_done	= nfs4_commit_done,
L
Linus Torvalds 已提交
6305
	.lock		= nfs4_proc_lock,
6306
	.clear_acl_cache = nfs4_zap_acl_attr,
T
Trond Myklebust 已提交
6307
	.close_context  = nfs4_close_context,
6308
	.open_context	= nfs4_atomic_open,
6309
	.init_client	= nfs4_init_client,
B
Bryan Schumaker 已提交
6310
	.secinfo	= nfs4_proc_secinfo,
L
Linus Torvalds 已提交
6311 6312
};

6313 6314 6315 6316 6317 6318 6319 6320 6321 6322 6323 6324
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
};

L
Linus Torvalds 已提交
6325 6326 6327 6328 6329
/*
 * Local variables:
 *  c-basic-offset: 8
 * End:
 */