nfs4state.c 95.9 KB
Newer Older
L
Linus Torvalds 已提交
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43
/*
*  linux/fs/nfsd/nfs4state.c
*
*  Copyright (c) 2001 The Regents of the University of Michigan.
*  All rights reserved.
*
*  Kendrick Smith <kmsmith@umich.edu>
*  Andy Adamson <kandros@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/param.h>
#include <linux/major.h>
#include <linux/slab.h>

#include <linux/sunrpc/svc.h>
#include <linux/nfsd/nfsd.h>
#include <linux/nfsd/cache.h>
44
#include <linux/file.h>
L
Linus Torvalds 已提交
45 46 47 48 49 50 51
#include <linux/mount.h>
#include <linux/workqueue.h>
#include <linux/smp_lock.h>
#include <linux/kthread.h>
#include <linux/nfs4.h>
#include <linux/nfsd/state.h>
#include <linux/nfsd/xdr4.h>
52
#include <linux/namei.h>
53
#include <linux/swap.h>
I
Ingo Molnar 已提交
54
#include <linux/mutex.h>
55
#include <linux/lockd/bind.h>
56
#include <linux/module.h>
57
#include <linux/sunrpc/svcauth_gss.h>
L
Linus Torvalds 已提交
58 59 60 61 62

#define NFSDDBG_FACILITY                NFSDDBG_PROC

/* Globals */
static time_t lease_time = 90;     /* default lease time */
63
static time_t user_lease_time = 90;
64
static time_t boot_time;
L
Linus Torvalds 已提交
65 66 67 68
static u32 current_ownerid = 1;
static u32 current_fileid = 1;
static u32 current_delegid = 1;
static u32 nfs4_init;
69 70 71 72 73
static stateid_t zerostateid;             /* bits all 0 */
static stateid_t onestateid;              /* bits all 1 */

#define ZERO_STATEID(stateid) (!memcmp((stateid), &zerostateid, sizeof(stateid_t)))
#define ONE_STATEID(stateid)  (!memcmp((stateid), &onestateid, sizeof(stateid_t)))
L
Linus Torvalds 已提交
74 75

/* forward declarations */
76
static struct nfs4_stateid * find_stateid(stateid_t *stid, int flags);
L
Linus Torvalds 已提交
77
static struct nfs4_delegation * find_delegation_stateid(struct inode *ino, stateid_t *stid);
78 79
static char user_recovery_dirname[PATH_MAX] = "/var/lib/nfs/v4recovery";
static void nfs4_set_recdir(char *recdir);
L
Linus Torvalds 已提交
80

81 82 83
/* Locking: */

/* Currently used for almost all code touching nfsv4 state: */
I
Ingo Molnar 已提交
84
static DEFINE_MUTEX(client_mutex);
L
Linus Torvalds 已提交
85

86 87 88 89 90 91 92
/*
 * Currently used for the del_recall_lru and file hash table.  In an
 * effort to decrease the scope of the client_mutex, this spinlock may
 * eventually cover more:
 */
static DEFINE_SPINLOCK(recall_lock);

93 94 95 96
static struct kmem_cache *stateowner_slab = NULL;
static struct kmem_cache *file_slab = NULL;
static struct kmem_cache *stateid_slab = NULL;
static struct kmem_cache *deleg_slab = NULL;
N
NeilBrown 已提交
97

L
Linus Torvalds 已提交
98 99 100
void
nfs4_lock_state(void)
{
I
Ingo Molnar 已提交
101
	mutex_lock(&client_mutex);
L
Linus Torvalds 已提交
102 103 104 105 106
}

void
nfs4_unlock_state(void)
{
I
Ingo Molnar 已提交
107
	mutex_unlock(&client_mutex);
L
Linus Torvalds 已提交
108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124
}

static inline u32
opaque_hashval(const void *ptr, int nbytes)
{
	unsigned char *cptr = (unsigned char *) ptr;

	u32 x = 0;
	while (nbytes--) {
		x *= 37;
		x += *cptr++;
	}
	return x;
}

static struct list_head del_recall_lru;

125 126 127
static inline void
put_nfs4_file(struct nfs4_file *fi)
{
128 129 130 131 132 133
	if (atomic_dec_and_lock(&fi->fi_ref, &recall_lock)) {
		list_del(&fi->fi_hash);
		spin_unlock(&recall_lock);
		iput(fi->fi_inode);
		kmem_cache_free(file_slab, fi);
	}
134 135 136 137 138
}

static inline void
get_nfs4_file(struct nfs4_file *fi)
{
139
	atomic_inc(&fi->fi_ref);
140 141
}

142
static int num_delegations;
143
unsigned int max_delegations;
144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178

/*
 * Open owner state (share locks)
 */

/* hash tables for nfs4_stateowner */
#define OWNER_HASH_BITS              8
#define OWNER_HASH_SIZE             (1 << OWNER_HASH_BITS)
#define OWNER_HASH_MASK             (OWNER_HASH_SIZE - 1)

#define ownerid_hashval(id) \
        ((id) & OWNER_HASH_MASK)
#define ownerstr_hashval(clientid, ownername) \
        (((clientid) + opaque_hashval((ownername.data), (ownername.len))) & OWNER_HASH_MASK)

static struct list_head	ownerid_hashtbl[OWNER_HASH_SIZE];
static struct list_head	ownerstr_hashtbl[OWNER_HASH_SIZE];

/* hash table for nfs4_file */
#define FILE_HASH_BITS                   8
#define FILE_HASH_SIZE                  (1 << FILE_HASH_BITS)
#define FILE_HASH_MASK                  (FILE_HASH_SIZE - 1)
/* hash table for (open)nfs4_stateid */
#define STATEID_HASH_BITS              10
#define STATEID_HASH_SIZE              (1 << STATEID_HASH_BITS)
#define STATEID_HASH_MASK              (STATEID_HASH_SIZE - 1)

#define file_hashval(x) \
        hash_ptr(x, FILE_HASH_BITS)
#define stateid_hashval(owner_id, file_id)  \
        (((owner_id) + (file_id)) & STATEID_HASH_MASK)

static struct list_head file_hashtbl[FILE_HASH_SIZE];
static struct list_head stateid_hashtbl[STATEID_HASH_SIZE];

L
Linus Torvalds 已提交
179 180 181 182 183 184 185 186
static struct nfs4_delegation *
alloc_init_deleg(struct nfs4_client *clp, struct nfs4_stateid *stp, struct svc_fh *current_fh, u32 type)
{
	struct nfs4_delegation *dp;
	struct nfs4_file *fp = stp->st_file;
	struct nfs4_callback *cb = &stp->st_stateowner->so_client->cl_callback;

	dprintk("NFSD alloc_init_deleg\n");
187 188
	if (fp->fi_had_conflict)
		return NULL;
189
	if (num_delegations > max_delegations)
190
		return NULL;
N
NeilBrown 已提交
191 192
	dp = kmem_cache_alloc(deleg_slab, GFP_KERNEL);
	if (dp == NULL)
L
Linus Torvalds 已提交
193
		return dp;
194
	num_delegations++;
195 196
	INIT_LIST_HEAD(&dp->dl_perfile);
	INIT_LIST_HEAD(&dp->dl_perclnt);
L
Linus Torvalds 已提交
197 198
	INIT_LIST_HEAD(&dp->dl_recall_lru);
	dp->dl_client = clp;
199
	get_nfs4_file(fp);
L
Linus Torvalds 已提交
200 201 202 203 204 205 206 207 208 209 210 211
	dp->dl_file = fp;
	dp->dl_flock = NULL;
	get_file(stp->st_vfs_file);
	dp->dl_vfs_file = stp->st_vfs_file;
	dp->dl_type = type;
	dp->dl_recall.cbr_dp = NULL;
	dp->dl_recall.cbr_ident = cb->cb_ident;
	dp->dl_recall.cbr_trunc = 0;
	dp->dl_stateid.si_boot = boot_time;
	dp->dl_stateid.si_stateownerid = current_delegid++;
	dp->dl_stateid.si_fileid = 0;
	dp->dl_stateid.si_generation = 0;
212
	fh_copy_shallow(&dp->dl_fh, &current_fh->fh_handle);
L
Linus Torvalds 已提交
213 214
	dp->dl_time = 0;
	atomic_set(&dp->dl_count, 1);
215 216
	list_add(&dp->dl_perfile, &fp->fi_delegations);
	list_add(&dp->dl_perclnt, &clp->cl_delegations);
L
Linus Torvalds 已提交
217 218 219 220 221 222 223 224
	return dp;
}

void
nfs4_put_delegation(struct nfs4_delegation *dp)
{
	if (atomic_dec_and_test(&dp->dl_count)) {
		dprintk("NFSD: freeing dp %p\n",dp);
225
		put_nfs4_file(dp->dl_file);
N
NeilBrown 已提交
226
		kmem_cache_free(deleg_slab, dp);
227
		num_delegations--;
L
Linus Torvalds 已提交
228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244
	}
}

/* Remove the associated file_lock first, then remove the delegation.
 * lease_modify() is called to remove the FS_LEASE file_lock from
 * the i_flock list, eventually calling nfsd's lock_manager
 * fl_release_callback.
 */
static void
nfs4_close_delegation(struct nfs4_delegation *dp)
{
	struct file *filp = dp->dl_vfs_file;

	dprintk("NFSD: close_delegation dp %p\n",dp);
	dp->dl_vfs_file = NULL;
	/* The following nfsd_close may not actually close the file,
	 * but we want to remove the lease in any case. */
N
NeilBrown 已提交
245
	if (dp->dl_flock)
246
		vfs_setlease(filp, F_UNLCK, &dp->dl_flock);
L
Linus Torvalds 已提交
247 248 249 250 251 252 253
	nfsd_close(filp);
}

/* Called under the state lock. */
static void
unhash_delegation(struct nfs4_delegation *dp)
{
254 255
	list_del_init(&dp->dl_perfile);
	list_del_init(&dp->dl_perclnt);
L
Linus Torvalds 已提交
256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273
	spin_lock(&recall_lock);
	list_del_init(&dp->dl_recall_lru);
	spin_unlock(&recall_lock);
	nfs4_close_delegation(dp);
	nfs4_put_delegation(dp);
}

/* 
 * SETCLIENTID state 
 */

/* Hash tables for nfs4_clientid state */
#define CLIENT_HASH_BITS                 4
#define CLIENT_HASH_SIZE                (1 << CLIENT_HASH_BITS)
#define CLIENT_HASH_MASK                (CLIENT_HASH_SIZE - 1)

#define clientid_hashval(id) \
	((id) & CLIENT_HASH_MASK)
N
NeilBrown 已提交
274 275
#define clientstr_hashval(name) \
	(opaque_hashval((name), 8) & CLIENT_HASH_MASK)
L
Linus Torvalds 已提交
276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300
/*
 * reclaim_str_hashtbl[] holds known client info from previous reset/reboot
 * used in reboot/reset lease grace period processing
 *
 * conf_id_hashtbl[], and conf_str_hashtbl[] hold confirmed
 * setclientid_confirmed info. 
 *
 * unconf_str_hastbl[] and unconf_id_hashtbl[] hold unconfirmed 
 * setclientid info.
 *
 * client_lru holds client queue ordered by nfs4_client.cl_time
 * for lease renewal.
 *
 * close_lru holds (open) stateowner queue ordered by nfs4_stateowner.so_time
 * for last close replay.
 */
static struct list_head	reclaim_str_hashtbl[CLIENT_HASH_SIZE];
static int reclaim_str_hashtbl_size = 0;
static struct list_head	conf_id_hashtbl[CLIENT_HASH_SIZE];
static struct list_head	conf_str_hashtbl[CLIENT_HASH_SIZE];
static struct list_head	unconf_str_hashtbl[CLIENT_HASH_SIZE];
static struct list_head	unconf_id_hashtbl[CLIENT_HASH_SIZE];
static struct list_head client_lru;
static struct list_head close_lru;

301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320
static void unhash_generic_stateid(struct nfs4_stateid *stp)
{
	list_del(&stp->st_hash);
	list_del(&stp->st_perfile);
	list_del(&stp->st_perstateowner);
}

static void free_generic_stateid(struct nfs4_stateid *stp)
{
	put_nfs4_file(stp->st_file);
	kmem_cache_free(stateid_slab, stp);
}

static void release_lock_stateid(struct nfs4_stateid *stp)
{
	unhash_generic_stateid(stp);
	locks_remove_posix(stp->st_vfs_file, (fl_owner_t)stp->st_stateowner);
	free_generic_stateid(stp);
}

321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340
static void unhash_lockowner(struct nfs4_stateowner *sop)
{
	struct nfs4_stateid *stp;

	list_del(&sop->so_idhash);
	list_del(&sop->so_strhash);
	list_del(&sop->so_perstateid);
	while (!list_empty(&sop->so_stateids)) {
		stp = list_first_entry(&sop->so_stateids,
				struct nfs4_stateid, st_perstateowner);
		release_lock_stateid(stp);
	}
}

static void release_lockowner(struct nfs4_stateowner *sop)
{
	unhash_lockowner(sop);
	nfs4_put_stateowner(sop);
}

341 342 343 344 345 346 347 348 349 350
static void
release_stateid_lockowners(struct nfs4_stateid *open_stp)
{
	struct nfs4_stateowner *lock_sop;

	while (!list_empty(&open_stp->st_lockowners)) {
		lock_sop = list_entry(open_stp->st_lockowners.next,
				struct nfs4_stateowner, so_perstateid);
		/* list_del(&open_stp->st_lockowners);  */
		BUG_ON(lock_sop->so_is_open_owner);
351
		release_lockowner(lock_sop);
352 353 354
	}
}

355 356 357 358 359 360 361 362
static void release_open_stateid(struct nfs4_stateid *stp)
{
	unhash_generic_stateid(stp);
	release_stateid_lockowners(stp);
	nfsd_close(stp->st_vfs_file);
	free_generic_stateid(stp);
}

363
static void unhash_openowner(struct nfs4_stateowner *sop)
364 365 366 367 368
{
	struct nfs4_stateid *stp;

	list_del(&sop->so_idhash);
	list_del(&sop->so_strhash);
369 370
	list_del(&sop->so_perclient);
	list_del(&sop->so_perstateid); /* XXX: necessary? */
371
	while (!list_empty(&sop->so_stateids)) {
372 373 374
		stp = list_first_entry(&sop->so_stateids,
				struct nfs4_stateid, st_perstateowner);
		release_open_stateid(stp);
375 376 377
	}
}

378
static void release_openowner(struct nfs4_stateowner *sop)
379
{
380
	unhash_openowner(sop);
381 382 383 384
	list_del(&sop->so_close_lru);
	nfs4_put_stateowner(sop);
}

M
Marc Eshel 已提交
385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427
static DEFINE_SPINLOCK(sessionid_lock);
#define SESSION_HASH_SIZE	512
static struct list_head sessionid_hashtbl[SESSION_HASH_SIZE];

static inline int
hash_sessionid(struct nfs4_sessionid *sessionid)
{
	struct nfsd4_sessionid *sid = (struct nfsd4_sessionid *)sessionid;

	return sid->sequence % SESSION_HASH_SIZE;
}

static inline void
dump_sessionid(const char *fn, struct nfs4_sessionid *sessionid)
{
	u32 *ptr = (u32 *)(&sessionid->data[0]);
	dprintk("%s: %u:%u:%u:%u\n", fn, ptr[0], ptr[1], ptr[2], ptr[3]);
}

/* caller must hold sessionid_lock */
static struct nfsd4_session *
find_in_sessionid_hashtbl(struct nfs4_sessionid *sessionid)
{
	struct nfsd4_session *elem;
	int idx;

	dump_sessionid(__func__, sessionid);
	idx = hash_sessionid(sessionid);
	dprintk("%s: idx is %d\n", __func__, idx);
	/* Search in the appropriate list */
	list_for_each_entry(elem, &sessionid_hashtbl[idx], se_hash) {
		dump_sessionid("list traversal", &elem->se_sessionid);
		if (!memcmp(elem->se_sessionid.data, sessionid->data,
			    NFS4_MAX_SESSIONID_LEN)) {
			return elem;
		}
	}

	dprintk("%s: session not found\n", __func__);
	return NULL;
}

/* caller must hold sessionid_lock */
A
Andy Adamson 已提交
428
static void
M
Marc Eshel 已提交
429
unhash_session(struct nfsd4_session *ses)
A
Andy Adamson 已提交
430 431 432
{
	list_del(&ses->se_hash);
	list_del(&ses->se_perclnt);
M
Marc Eshel 已提交
433 434 435 436 437 438 439 440
}

static void
release_session(struct nfsd4_session *ses)
{
	spin_lock(&sessionid_lock);
	unhash_session(ses);
	spin_unlock(&sessionid_lock);
A
Andy Adamson 已提交
441 442 443 444 445 446 447 448 449 450 451 452 453
	nfsd4_put_session(ses);
}

void
free_session(struct kref *kref)
{
	struct nfsd4_session *ses;

	ses = container_of(kref, struct nfsd4_session, se_ref);
	kfree(ses->se_slots);
	kfree(ses);
}

L
Linus Torvalds 已提交
454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482
static inline void
renew_client(struct nfs4_client *clp)
{
	/*
	* Move client to the end to the LRU list.
	*/
	dprintk("renewing client (clientid %08x/%08x)\n", 
			clp->cl_clientid.cl_boot, 
			clp->cl_clientid.cl_id);
	list_move_tail(&clp->cl_lru, &client_lru);
	clp->cl_time = get_seconds();
}

/* SETCLIENTID and SETCLIENTID_CONFIRM Helper functions */
static int
STALE_CLIENTID(clientid_t *clid)
{
	if (clid->cl_boot == boot_time)
		return 0;
	dprintk("NFSD stale clientid (%08x/%08x)\n", 
			clid->cl_boot, clid->cl_id);
	return 1;
}

/* 
 * XXX Should we use a slab cache ?
 * This type of memory management is somewhat inefficient, but we use it
 * anyway since SETCLIENTID is not a common operation.
 */
483
static struct nfs4_client *alloc_client(struct xdr_netobj name)
L
Linus Torvalds 已提交
484 485 486
{
	struct nfs4_client *clp;

487 488 489 490 491 492 493
	clp = kzalloc(sizeof(struct nfs4_client), GFP_KERNEL);
	if (clp == NULL)
		return NULL;
	clp->cl_name.data = kmalloc(name.len, GFP_KERNEL);
	if (clp->cl_name.data == NULL) {
		kfree(clp);
		return NULL;
L
Linus Torvalds 已提交
494
	}
495 496
	memcpy(clp->cl_name.data, name.data, name.len);
	clp->cl_name.len = name.len;
L
Linus Torvalds 已提交
497 498 499
	return clp;
}

500 501 502 503 504 505
static void
shutdown_callback_client(struct nfs4_client *clp)
{
	struct rpc_clnt *clnt = clp->cl_callback.cb_client;

	if (clnt) {
506 507 508 509
		/*
		 * Callback threads take a reference on the client, so there
		 * should be no outstanding callbacks at this point.
		 */
510 511 512 513 514
		clp->cl_callback.cb_client = NULL;
		rpc_shutdown_client(clnt);
	}
}

L
Linus Torvalds 已提交
515 516 517
static inline void
free_client(struct nfs4_client *clp)
{
518
	shutdown_callback_client(clp);
L
Linus Torvalds 已提交
519 520
	if (clp->cl_cred.cr_group_info)
		put_group_info(clp->cl_cred.cr_group_info);
521
	kfree(clp->cl_principal);
L
Linus Torvalds 已提交
522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544
	kfree(clp->cl_name.data);
	kfree(clp);
}

void
put_nfs4_client(struct nfs4_client *clp)
{
	if (atomic_dec_and_test(&clp->cl_count))
		free_client(clp);
}

static void
expire_client(struct nfs4_client *clp)
{
	struct nfs4_stateowner *sop;
	struct nfs4_delegation *dp;
	struct list_head reaplist;

	dprintk("NFSD: expire_client cl_count %d\n",
	                    atomic_read(&clp->cl_count));

	INIT_LIST_HEAD(&reaplist);
	spin_lock(&recall_lock);
545 546
	while (!list_empty(&clp->cl_delegations)) {
		dp = list_entry(clp->cl_delegations.next, struct nfs4_delegation, dl_perclnt);
L
Linus Torvalds 已提交
547 548
		dprintk("NFSD: expire client. dp %p, fp %p\n", dp,
				dp->dl_flock);
549
		list_del_init(&dp->dl_perclnt);
L
Linus Torvalds 已提交
550 551 552 553 554 555 556 557 558 559 560
		list_move(&dp->dl_recall_lru, &reaplist);
	}
	spin_unlock(&recall_lock);
	while (!list_empty(&reaplist)) {
		dp = list_entry(reaplist.next, struct nfs4_delegation, dl_recall_lru);
		list_del_init(&dp->dl_recall_lru);
		unhash_delegation(dp);
	}
	list_del(&clp->cl_idhash);
	list_del(&clp->cl_strhash);
	list_del(&clp->cl_lru);
561 562
	while (!list_empty(&clp->cl_openowners)) {
		sop = list_entry(clp->cl_openowners.next, struct nfs4_stateowner, so_perclient);
563
		release_openowner(sop);
L
Linus Torvalds 已提交
564
	}
565 566 567 568 569 570
	while (!list_empty(&clp->cl_sessions)) {
		struct nfsd4_session  *ses;
		ses = list_entry(clp->cl_sessions.next, struct nfsd4_session,
				 se_perclnt);
		release_session(ses);
	}
L
Linus Torvalds 已提交
571 572 573
	put_nfs4_client(clp);
}

574 575
static struct nfs4_client *create_client(struct xdr_netobj name, char *recdir)
{
L
Linus Torvalds 已提交
576 577
	struct nfs4_client *clp;

578 579 580
	clp = alloc_client(name);
	if (clp == NULL)
		return NULL;
N
NeilBrown 已提交
581
	memcpy(clp->cl_recdir, recdir, HEXDIR_LEN);
L
Linus Torvalds 已提交
582 583 584 585
	atomic_set(&clp->cl_count, 1);
	atomic_set(&clp->cl_callback.cb_set, 0);
	INIT_LIST_HEAD(&clp->cl_idhash);
	INIT_LIST_HEAD(&clp->cl_strhash);
586 587
	INIT_LIST_HEAD(&clp->cl_openowners);
	INIT_LIST_HEAD(&clp->cl_delegations);
588
	INIT_LIST_HEAD(&clp->cl_sessions);
L
Linus Torvalds 已提交
589 590 591 592
	INIT_LIST_HEAD(&clp->cl_lru);
	return clp;
}

593 594 595 596
static void copy_verf(struct nfs4_client *target, nfs4_verifier *source)
{
	memcpy(target->cl_verifier.data, source->data,
			sizeof(target->cl_verifier.data));
L
Linus Torvalds 已提交
597 598
}

599 600
static void copy_clid(struct nfs4_client *target, struct nfs4_client *source)
{
L
Linus Torvalds 已提交
601 602 603 604
	target->cl_clientid.cl_boot = source->cl_clientid.cl_boot; 
	target->cl_clientid.cl_id = source->cl_clientid.cl_id; 
}

605 606
static void copy_cred(struct svc_cred *target, struct svc_cred *source)
{
L
Linus Torvalds 已提交
607 608 609 610 611 612
	target->cr_uid = source->cr_uid;
	target->cr_gid = source->cr_gid;
	target->cr_group_info = source->cr_group_info;
	get_group_info(target->cr_group_info);
}

613
static int same_name(const char *n1, const char *n2)
614
{
N
NeilBrown 已提交
615
	return 0 == memcmp(n1, n2, HEXDIR_LEN);
L
Linus Torvalds 已提交
616 617 618
}

static int
619 620 621
same_verf(nfs4_verifier *v1, nfs4_verifier *v2)
{
	return 0 == memcmp(v1->data, v2->data, sizeof(v1->data));
L
Linus Torvalds 已提交
622 623 624
}

static int
625 626 627
same_clid(clientid_t *cl1, clientid_t *cl2)
{
	return (cl1->cl_boot == cl2->cl_boot) && (cl1->cl_id == cl2->cl_id);
L
Linus Torvalds 已提交
628 629 630 631
}

/* XXX what about NGROUP */
static int
632 633 634
same_creds(struct svc_cred *cr1, struct svc_cred *cr2)
{
	return cr1->cr_uid == cr2->cr_uid;
L
Linus Torvalds 已提交
635 636
}

637 638 639 640
static void gen_clid(struct nfs4_client *clp)
{
	static u32 current_clientid = 1;

L
Linus Torvalds 已提交
641 642 643 644
	clp->cl_clientid.cl_boot = boot_time;
	clp->cl_clientid.cl_id = current_clientid++; 
}

645 646 647 648
static void gen_confirm(struct nfs4_client *clp)
{
	static u32 i;
	u32 *p;
L
Linus Torvalds 已提交
649 650

	p = (u32 *)clp->cl_confirm.data;
651 652
	*p++ = get_seconds();
	*p++ = i++;
L
Linus Torvalds 已提交
653 654
}

655 656
static int check_name(struct xdr_netobj name)
{
L
Linus Torvalds 已提交
657 658 659
	if (name.len == 0) 
		return 0;
	if (name.len > NFS4_OPAQUE_LIMIT) {
660
		dprintk("NFSD: check_name: name too long(%d)!\n", name.len);
L
Linus Torvalds 已提交
661 662 663 664 665
		return 0;
	}
	return 1;
}

666
static void
L
Linus Torvalds 已提交
667 668 669 670 671 672 673 674 675 676 677
add_to_unconfirmed(struct nfs4_client *clp, unsigned int strhashval)
{
	unsigned int idhashval;

	list_add(&clp->cl_strhash, &unconf_str_hashtbl[strhashval]);
	idhashval = clientid_hashval(clp->cl_clientid.cl_id);
	list_add(&clp->cl_idhash, &unconf_id_hashtbl[idhashval]);
	list_add_tail(&clp->cl_lru, &client_lru);
	clp->cl_time = get_seconds();
}

678
static void
L
Linus Torvalds 已提交
679 680 681 682 683 684 685
move_to_confirmed(struct nfs4_client *clp)
{
	unsigned int idhashval = clientid_hashval(clp->cl_clientid.cl_id);
	unsigned int strhashval;

	dprintk("NFSD: move_to_confirm nfs4_client %p\n", clp);
	list_del_init(&clp->cl_strhash);
A
Akinobu Mita 已提交
686
	list_move(&clp->cl_idhash, &conf_id_hashtbl[idhashval]);
N
NeilBrown 已提交
687
	strhashval = clientstr_hashval(clp->cl_recdir);
L
Linus Torvalds 已提交
688 689 690 691 692 693 694 695 696 697 698
	list_add(&clp->cl_strhash, &conf_str_hashtbl[strhashval]);
	renew_client(clp);
}

static struct nfs4_client *
find_confirmed_client(clientid_t *clid)
{
	struct nfs4_client *clp;
	unsigned int idhashval = clientid_hashval(clid->cl_id);

	list_for_each_entry(clp, &conf_id_hashtbl[idhashval], cl_idhash) {
699
		if (same_clid(&clp->cl_clientid, clid))
L
Linus Torvalds 已提交
700 701 702 703 704 705 706 707 708 709 710 711
			return clp;
	}
	return NULL;
}

static struct nfs4_client *
find_unconfirmed_client(clientid_t *clid)
{
	struct nfs4_client *clp;
	unsigned int idhashval = clientid_hashval(clid->cl_id);

	list_for_each_entry(clp, &unconf_id_hashtbl[idhashval], cl_idhash) {
712
		if (same_clid(&clp->cl_clientid, clid))
L
Linus Torvalds 已提交
713 714 715 716 717
			return clp;
	}
	return NULL;
}

718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733
/*
 * Return 1 iff clp's clientid establishment method matches the use_exchange_id
 * parameter. Matching is based on the fact the at least one of the
 * EXCHGID4_FLAG_USE_{NON_PNFS,PNFS_MDS,PNFS_DS} flags must be set for v4.1
 *
 * FIXME: we need to unify the clientid namespaces for nfsv4.x
 * and correctly deal with client upgrade/downgrade in EXCHANGE_ID
 * and SET_CLIENTID{,_CONFIRM}
 */
static inline int
match_clientid_establishment(struct nfs4_client *clp, bool use_exchange_id)
{
	bool has_exchange_flags = (clp->cl_exchange_flags != 0);
	return use_exchange_id == has_exchange_flags;
}

734
static struct nfs4_client *
735 736
find_confirmed_client_by_str(const char *dname, unsigned int hashval,
			     bool use_exchange_id)
737 738 739 740
{
	struct nfs4_client *clp;

	list_for_each_entry(clp, &conf_str_hashtbl[hashval], cl_strhash) {
741 742
		if (same_name(clp->cl_recdir, dname) &&
		    match_clientid_establishment(clp, use_exchange_id))
743 744 745 746 747 748
			return clp;
	}
	return NULL;
}

static struct nfs4_client *
749 750
find_unconfirmed_client_by_str(const char *dname, unsigned int hashval,
			       bool use_exchange_id)
751 752 753 754
{
	struct nfs4_client *clp;

	list_for_each_entry(clp, &unconf_str_hashtbl[hashval], cl_strhash) {
755 756
		if (same_name(clp->cl_recdir, dname) &&
		    match_clientid_establishment(clp, use_exchange_id))
757 758 759 760 761
			return clp;
	}
	return NULL;
}

L
Linus Torvalds 已提交
762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795
/* a helper function for parse_callback */
static int
parse_octet(unsigned int *lenp, char **addrp)
{
	unsigned int len = *lenp;
	char *p = *addrp;
	int n = -1;
	char c;

	for (;;) {
		if (!len)
			break;
		len--;
		c = *p++;
		if (c == '.')
			break;
		if ((c < '0') || (c > '9')) {
			n = -1;
			break;
		}
		if (n < 0)
			n = 0;
		n = (n * 10) + (c - '0');
		if (n > 255) {
			n = -1;
			break;
		}
	}
	*lenp = len;
	*addrp = p;
	return n;
}

/* parse and set the setclientid ipv4 callback address */
796
static int
L
Linus Torvalds 已提交
797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831
parse_ipv4(unsigned int addr_len, char *addr_val, unsigned int *cbaddrp, unsigned short *cbportp)
{
	int temp = 0;
	u32 cbaddr = 0;
	u16 cbport = 0;
	u32 addrlen = addr_len;
	char *addr = addr_val;
	int i, shift;

	/* ipaddress */
	shift = 24;
	for(i = 4; i > 0  ; i--) {
		if ((temp = parse_octet(&addrlen, &addr)) < 0) {
			return 0;
		}
		cbaddr |= (temp << shift);
		if (shift > 0)
		shift -= 8;
	}
	*cbaddrp = cbaddr;

	/* port */
	shift = 8;
	for(i = 2; i > 0  ; i--) {
		if ((temp = parse_octet(&addrlen, &addr)) < 0) {
			return 0;
		}
		cbport |= (temp << shift);
		if (shift > 0)
			shift -= 8;
	}
	*cbportp = cbport;
	return 1;
}

832
static void
L
Linus Torvalds 已提交
833 834 835 836 837 838 839 840 841 842 843 844 845 846 847
gen_callback(struct nfs4_client *clp, struct nfsd4_setclientid *se)
{
	struct nfs4_callback *cb = &clp->cl_callback;

	/* Currently, we only support tcp for the callback channel */
	if ((se->se_callback_netid_len != 3) || memcmp((char *)se->se_callback_netid_val, "tcp", 3))
		goto out_err;

	if ( !(parse_ipv4(se->se_callback_addr_len, se->se_callback_addr_val,
	                 &cb->cb_addr, &cb->cb_port)))
		goto out_err;
	cb->cb_prog = se->se_callback_prog;
	cb->cb_ident = se->se_callback_ident;
	return;
out_err:
N
Neil Brown 已提交
848
	dprintk(KERN_INFO "NFSD: this client (clientid %08x/%08x) "
L
Linus Torvalds 已提交
849 850 851 852 853 854
		"will not receive delegations\n",
		clp->cl_clientid.cl_boot, clp->cl_clientid.cl_id);

	return;
}

A
Andy Adamson 已提交
855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870
/*
 * Set the exchange_id flags returned by the server.
 */
static void
nfsd4_set_ex_flags(struct nfs4_client *new, struct nfsd4_exchange_id *clid)
{
	/* pNFS is not supported */
	new->cl_exchange_flags |= EXCHGID4_FLAG_USE_NON_PNFS;

	/* Referrals are supported, Migration is not. */
	new->cl_exchange_flags |= EXCHGID4_FLAG_SUPP_MOVED_REFER;

	/* set the wire flags to return to client. */
	clid->flags = new->cl_exchange_flags;
}

A
Andy Adamson 已提交
871 872 873 874 875
__be32
nfsd4_exchange_id(struct svc_rqst *rqstp,
		  struct nfsd4_compound_state *cstate,
		  struct nfsd4_exchange_id *exid)
{
A
Andy Adamson 已提交
876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912
	struct nfs4_client *unconf, *conf, *new;
	int status;
	unsigned int		strhashval;
	char			dname[HEXDIR_LEN];
	nfs4_verifier		verf = exid->verifier;
	u32			ip_addr = svc_addr_in(rqstp)->sin_addr.s_addr;

	dprintk("%s rqstp=%p exid=%p clname.len=%u clname.data=%p "
		" ip_addr=%u flags %x, spa_how %d\n",
		__func__, rqstp, exid, exid->clname.len, exid->clname.data,
		ip_addr, exid->flags, exid->spa_how);

	if (!check_name(exid->clname) || (exid->flags & ~EXCHGID4_FLAG_MASK_A))
		return nfserr_inval;

	/* Currently only support SP4_NONE */
	switch (exid->spa_how) {
	case SP4_NONE:
		break;
	case SP4_SSV:
		return nfserr_encr_alg_unsupp;
	default:
		BUG();				/* checked by xdr code */
	case SP4_MACH_CRED:
		return nfserr_serverfault;	/* no excuse :-/ */
	}

	status = nfs4_make_rec_clidname(dname, &exid->clname);

	if (status)
		goto error;

	strhashval = clientstr_hashval(dname);

	nfs4_lock_state();
	status = nfs_ok;

913
	conf = find_confirmed_client_by_str(dname, strhashval, true);
A
Andy Adamson 已提交
914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958 959 960
	if (conf) {
		if (!same_verf(&verf, &conf->cl_verifier)) {
			/* 18.35.4 case 8 */
			if (exid->flags & EXCHGID4_FLAG_UPD_CONFIRMED_REC_A) {
				status = nfserr_not_same;
				goto out;
			}
			/* Client reboot: destroy old state */
			expire_client(conf);
			goto out_new;
		}
		if (!same_creds(&conf->cl_cred, &rqstp->rq_cred)) {
			/* 18.35.4 case 9 */
			if (exid->flags & EXCHGID4_FLAG_UPD_CONFIRMED_REC_A) {
				status = nfserr_perm;
				goto out;
			}
			expire_client(conf);
			goto out_new;
		}
		if (ip_addr != conf->cl_addr &&
		    !(exid->flags & EXCHGID4_FLAG_UPD_CONFIRMED_REC_A)) {
			/* Client collision. 18.35.4 case 3 */
			status = nfserr_clid_inuse;
			goto out;
		}
		/*
		 * Set bit when the owner id and verifier map to an already
		 * confirmed client id (18.35.3).
		 */
		exid->flags |= EXCHGID4_FLAG_CONFIRMED_R;

		/*
		 * Falling into 18.35.4 case 2, possible router replay.
		 * Leave confirmed record intact and return same result.
		 */
		copy_verf(conf, &verf);
		new = conf;
		goto out_copy;
	} else {
		/* 18.35.4 case 7 */
		if (exid->flags & EXCHGID4_FLAG_UPD_CONFIRMED_REC_A) {
			status = nfserr_noent;
			goto out;
		}
	}

961
	unconf  = find_unconfirmed_client_by_str(dname, strhashval, true);
A
Andy Adamson 已提交
962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000
	if (unconf) {
		/*
		 * Possible retry or client restart.  Per 18.35.4 case 4,
		 * a new unconfirmed record should be generated regardless
		 * of whether any properties have changed.
		 */
		expire_client(unconf);
	}

out_new:
	/* Normal case */
	new = create_client(exid->clname, dname);
	if (new == NULL) {
		status = nfserr_resource;
		goto out;
	}

	copy_verf(new, &verf);
	copy_cred(&new->cl_cred, &rqstp->rq_cred);
	new->cl_addr = ip_addr;
	gen_clid(new);
	gen_confirm(new);
	add_to_unconfirmed(new, strhashval);
out_copy:
	exid->clientid.cl_boot = new->cl_clientid.cl_boot;
	exid->clientid.cl_id = new->cl_clientid.cl_id;

	new->cl_seqid = exid->seqid = 1;
	nfsd4_set_ex_flags(new, exid);

	dprintk("nfsd4_exchange_id seqid %d flags %x\n",
		new->cl_seqid, new->cl_exchange_flags);
	status = nfs_ok;

out:
	nfs4_unlock_state();
error:
	dprintk("nfsd4_exchange_id returns %d\n", ntohl(status));
	return status;
A
Andy Adamson 已提交
1001 1002
}

B
Benny Halevy 已提交
1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028
static int
check_slot_seqid(u32 seqid, struct nfsd4_slot *slot)
{
	dprintk("%s enter. seqid %d slot->sl_seqid %d\n", __func__, seqid,
		slot->sl_seqid);

	/* The slot is in use, and no response has been sent. */
	if (slot->sl_inuse) {
		if (seqid == slot->sl_seqid)
			return nfserr_jukebox;
		else
			return nfserr_seq_misordered;
	}
	/* Normal */
	if (likely(seqid == slot->sl_seqid + 1))
		return nfs_ok;
	/* Replay */
	if (seqid == slot->sl_seqid)
		return nfserr_replay_cache;
	/* Wraparound */
	if (seqid == 1 && (slot->sl_seqid + 1) == 0)
		return nfs_ok;
	/* Misordered replay or misordered new request */
	return nfserr_seq_misordered;
}

A
Andy Adamson 已提交
1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045
__be32
nfsd4_create_session(struct svc_rqst *rqstp,
		     struct nfsd4_compound_state *cstate,
		     struct nfsd4_create_session *cr_ses)
{
	return -1;	/* stub */
}

__be32
nfsd4_destroy_session(struct svc_rqst *r,
		      struct nfsd4_compound_state *cstate,
		      struct nfsd4_destroy_session *sessionid)
{
	return -1;	/* stub */
}

__be32
B
Benny Halevy 已提交
1046
nfsd4_sequence(struct svc_rqst *rqstp,
A
Andy Adamson 已提交
1047 1048 1049
	       struct nfsd4_compound_state *cstate,
	       struct nfsd4_sequence *seq)
{
1050
	struct nfsd4_compoundres *resp = rqstp->rq_resp;
B
Benny Halevy 已提交
1051 1052 1053 1054
	struct nfsd4_session *session;
	struct nfsd4_slot *slot;
	int status;

1055 1056 1057
	if (resp->opcnt != 1)
		return nfserr_sequence_pos;

B
Benny Halevy 已提交
1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097
	spin_lock(&sessionid_lock);
	status = nfserr_badsession;
	session = find_in_sessionid_hashtbl(&seq->sessionid);
	if (!session)
		goto out;

	status = nfserr_badslot;
	if (seq->slotid >= session->se_fnumslots)
		goto out;

	slot = &session->se_slots[seq->slotid];
	dprintk("%s: slotid %d\n", __func__, seq->slotid);

	status = check_slot_seqid(seq->seqid, slot);
	if (status == nfserr_replay_cache) {
		cstate->slot = slot;
		cstate->session = session;
		goto replay_cache;
	}
	if (status)
		goto out;

	/* Success! bump slot seqid */
	slot->sl_inuse = true;
	slot->sl_seqid = seq->seqid;

	cstate->slot = slot;
	cstate->session = session;

replay_cache:
	/* Renew the clientid on success and on replay.
	 * Hold a session reference until done processing the compound:
	 * nfsd4_put_session called only if the cstate slot is set.
	 */
	renew_client(session->se_client);
	nfsd4_get_session(session);
out:
	spin_unlock(&sessionid_lock);
	dprintk("%s: return %d\n", __func__, ntohl(status));
	return status;
A
Andy Adamson 已提交
1098 1099
}

1100
__be32
1101 1102
nfsd4_setclientid(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate,
		  struct nfsd4_setclientid *setclid)
L
Linus Torvalds 已提交
1103
{
1104
	struct sockaddr_in	*sin = svc_addr_in(rqstp);
L
Linus Torvalds 已提交
1105 1106 1107 1108 1109 1110
	struct xdr_netobj 	clname = { 
		.len = setclid->se_namelen,
		.data = setclid->se_name,
	};
	nfs4_verifier		clverifier = setclid->se_verf;
	unsigned int 		strhashval;
1111
	struct nfs4_client	*conf, *unconf, *new;
1112
	__be32 			status;
1113
	char			*princ;
N
NeilBrown 已提交
1114
	char                    dname[HEXDIR_LEN];
L
Linus Torvalds 已提交
1115 1116
	
	if (!check_name(clname))
1117
		return nfserr_inval;
L
Linus Torvalds 已提交
1118

N
NeilBrown 已提交
1119 1120
	status = nfs4_make_rec_clidname(dname, &clname);
	if (status)
1121
		return status;
N
NeilBrown 已提交
1122

L
Linus Torvalds 已提交
1123 1124 1125 1126 1127
	/* 
	 * XXX The Duplicate Request Cache (DRC) has been checked (??)
	 * We get here on a DRC miss.
	 */

N
NeilBrown 已提交
1128
	strhashval = clientstr_hashval(dname);
L
Linus Torvalds 已提交
1129 1130

	nfs4_lock_state();
1131
	conf = find_confirmed_client_by_str(dname, strhashval, false);
1132
	if (conf) {
1133
		/* RFC 3530 14.2.33 CASE 0: */
L
Linus Torvalds 已提交
1134
		status = nfserr_clid_inuse;
1135 1136 1137
		if (!same_creds(&conf->cl_cred, &rqstp->rq_cred)) {
			dprintk("NFSD: setclientid: string in use by client"
				" at %pI4\n", &conf->cl_addr);
L
Linus Torvalds 已提交
1138 1139 1140
			goto out;
		}
	}
1141 1142 1143 1144 1145
	/*
	 * section 14.2.33 of RFC 3530 (under the heading "IMPLEMENTATION")
	 * has a description of SETCLIENTID request processing consisting
	 * of 5 bullet points, labeled as CASE0 - CASE4 below.
	 */
1146
	unconf = find_unconfirmed_client_by_str(dname, strhashval, false);
L
Linus Torvalds 已提交
1147 1148
	status = nfserr_resource;
	if (!conf) {
1149 1150 1151
		/*
		 * RFC 3530 14.2.33 CASE 4:
		 * placed first, because it is the normal case
L
Linus Torvalds 已提交
1152 1153 1154
		 */
		if (unconf)
			expire_client(unconf);
N
NeilBrown 已提交
1155 1156
		new = create_client(clname, dname);
		if (new == NULL)
L
Linus Torvalds 已提交
1157 1158
			goto out;
		gen_clid(new);
1159
	} else if (same_verf(&conf->cl_verifier, &clverifier)) {
L
Linus Torvalds 已提交
1160
		/*
1161 1162
		 * RFC 3530 14.2.33 CASE 1:
		 * probable callback update
L
Linus Torvalds 已提交
1163
		 */
1164 1165 1166 1167 1168 1169 1170
		if (unconf) {
			/* Note this is removing unconfirmed {*x***},
			 * which is stronger than RFC recommended {vxc**}.
			 * This has the advantage that there is at most
			 * one {*x***} in either list at any time.
			 */
			expire_client(unconf);
L
Linus Torvalds 已提交
1171
		}
N
NeilBrown 已提交
1172 1173
		new = create_client(clname, dname);
		if (new == NULL)
L
Linus Torvalds 已提交
1174 1175 1176 1177
			goto out;
		copy_clid(new, conf);
	} else if (!unconf) {
		/*
1178 1179 1180
		 * RFC 3530 14.2.33 CASE 2:
		 * probable client reboot; state will be removed if
		 * confirmed.
L
Linus Torvalds 已提交
1181
		 */
N
NeilBrown 已提交
1182 1183
		new = create_client(clname, dname);
		if (new == NULL)
L
Linus Torvalds 已提交
1184 1185
			goto out;
		gen_clid(new);
1186
	} else {
1187 1188 1189 1190
		/*
		 * RFC 3530 14.2.33 CASE 3:
		 * probable client reboot; state will be removed if
		 * confirmed.
L
Linus Torvalds 已提交
1191 1192
		 */
		expire_client(unconf);
N
NeilBrown 已提交
1193 1194
		new = create_client(clname, dname);
		if (new == NULL)
L
Linus Torvalds 已提交
1195 1196 1197
			goto out;
		gen_clid(new);
	}
1198 1199
	copy_verf(new, &clverifier);
	new->cl_addr = sin->sin_addr.s_addr;
1200
	new->cl_flavor = rqstp->rq_flavor;
1201 1202 1203 1204 1205 1206 1207 1208
	princ = svc_gss_principal(rqstp);
	if (princ) {
		new->cl_principal = kstrdup(princ, GFP_KERNEL);
		if (new->cl_principal == NULL) {
			free_client(new);
			goto out;
		}
	}
1209 1210 1211 1212
	copy_cred(&new->cl_cred, &rqstp->rq_cred);
	gen_confirm(new);
	gen_callback(new, setclid);
	add_to_unconfirmed(new, strhashval);
L
Linus Torvalds 已提交
1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223
	setclid->se_clientid.cl_boot = new->cl_clientid.cl_boot;
	setclid->se_clientid.cl_id = new->cl_clientid.cl_id;
	memcpy(setclid->se_confirm.data, new->cl_confirm.data, sizeof(setclid->se_confirm.data));
	status = nfs_ok;
out:
	nfs4_unlock_state();
	return status;
}


/*
1224 1225 1226
 * Section 14.2.34 of RFC 3530 (under the heading "IMPLEMENTATION") has
 * a description of SETCLIENTID_CONFIRM request processing consisting of 4
 * bullets, labeled as CASE1 - CASE4 below.
L
Linus Torvalds 已提交
1227
 */
1228
__be32
1229 1230 1231
nfsd4_setclientid_confirm(struct svc_rqst *rqstp,
			 struct nfsd4_compound_state *cstate,
			 struct nfsd4_setclientid_confirm *setclientid_confirm)
L
Linus Torvalds 已提交
1232
{
1233
	struct sockaddr_in *sin = svc_addr_in(rqstp);
1234
	struct nfs4_client *conf, *unconf;
L
Linus Torvalds 已提交
1235 1236
	nfs4_verifier confirm = setclientid_confirm->sc_confirm; 
	clientid_t * clid = &setclientid_confirm->sc_clientid;
1237
	__be32 status;
L
Linus Torvalds 已提交
1238 1239 1240 1241 1242 1243 1244 1245 1246

	if (STALE_CLIENTID(clid))
		return nfserr_stale_clientid;
	/* 
	 * XXX The Duplicate Request Cache (DRC) has been checked (??)
	 * We get here on a DRC miss.
	 */

	nfs4_lock_state();
1247 1248 1249 1250 1251

	conf = find_confirmed_client(clid);
	unconf = find_unconfirmed_client(clid);

	status = nfserr_clid_inuse;
1252
	if (conf && conf->cl_addr != sin->sin_addr.s_addr)
1253
		goto out;
1254
	if (unconf && unconf->cl_addr != sin->sin_addr.s_addr)
1255 1256
		goto out;

1257 1258 1259 1260 1261
	/*
	 * section 14.2.34 of RFC 3530 has a description of
	 * SETCLIENTID_CONFIRM request processing consisting
	 * of 4 bullet points, labeled as CASE1 - CASE4 below.
	 */
1262
	if (conf && unconf && same_verf(&confirm, &unconf->cl_confirm)) {
1263 1264 1265 1266
		/*
		 * RFC 3530 14.2.34 CASE 1:
		 * callback update
		 */
1267
		if (!same_creds(&conf->cl_cred, &unconf->cl_cred))
L
Linus Torvalds 已提交
1268 1269
			status = nfserr_clid_inuse;
		else {
1270 1271
			/* XXX: We just turn off callbacks until we can handle
			  * change request correctly. */
1272
			atomic_set(&conf->cl_callback.cb_set, 0);
1273
			gen_confirm(conf);
N
NeilBrown 已提交
1274
			nfsd4_remove_clid_dir(unconf);
1275
			expire_client(unconf);
L
Linus Torvalds 已提交
1276
			status = nfs_ok;
1277

L
Linus Torvalds 已提交
1278
		}
1279
	} else if (conf && !unconf) {
1280 1281 1282 1283
		/*
		 * RFC 3530 14.2.34 CASE 2:
		 * probable retransmitted request; play it safe and
		 * do nothing.
1284
		 */
1285
		if (!same_creds(&conf->cl_cred, &rqstp->rq_cred))
L
Linus Torvalds 已提交
1286
			status = nfserr_clid_inuse;
1287
		else
L
Linus Torvalds 已提交
1288
			status = nfs_ok;
1289
	} else if (!conf && unconf
1290
			&& same_verf(&unconf->cl_confirm, &confirm)) {
1291 1292 1293
		/*
		 * RFC 3530 14.2.34 CASE 3:
		 * Normal case; new or rebooted client:
1294
		 */
1295
		if (!same_creds(&unconf->cl_cred, &rqstp->rq_cred)) {
L
Linus Torvalds 已提交
1296 1297
			status = nfserr_clid_inuse;
		} else {
1298 1299 1300
			unsigned int hash =
				clientstr_hashval(unconf->cl_recdir);
			conf = find_confirmed_client_by_str(unconf->cl_recdir,
1301
							    hash, false);
1302
			if (conf) {
1303
				nfsd4_remove_clid_dir(conf);
1304 1305
				expire_client(conf);
			}
L
Linus Torvalds 已提交
1306
			move_to_confirmed(unconf);
1307
			conf = unconf;
1308
			nfsd4_probe_callback(conf);
1309
			status = nfs_ok;
L
Linus Torvalds 已提交
1310
		}
1311 1312
	} else if ((!conf || (conf && !same_verf(&conf->cl_confirm, &confirm)))
	    && (!unconf || (unconf && !same_verf(&unconf->cl_confirm,
1313
				    				&confirm)))) {
1314 1315 1316
		/*
		 * RFC 3530 14.2.34 CASE 4:
		 * Client probably hasn't noticed that we rebooted yet.
1317
		 */
L
Linus Torvalds 已提交
1318
		status = nfserr_stale_clientid;
1319
	} else {
1320 1321 1322
		/* check that we have hit one of the cases...*/
		status = nfserr_clid_inuse;
	}
L
Linus Torvalds 已提交
1323 1324 1325 1326 1327 1328 1329 1330 1331 1332 1333 1334
out:
	nfs4_unlock_state();
	return status;
}

/* OPEN Share state helper functions */
static inline struct nfs4_file *
alloc_init_file(struct inode *ino)
{
	struct nfs4_file *fp;
	unsigned int hashval = file_hashval(ino);

N
NeilBrown 已提交
1335 1336
	fp = kmem_cache_alloc(file_slab, GFP_KERNEL);
	if (fp) {
1337
		atomic_set(&fp->fi_ref, 1);
L
Linus Torvalds 已提交
1338
		INIT_LIST_HEAD(&fp->fi_hash);
1339 1340
		INIT_LIST_HEAD(&fp->fi_stateids);
		INIT_LIST_HEAD(&fp->fi_delegations);
1341
		spin_lock(&recall_lock);
L
Linus Torvalds 已提交
1342
		list_add(&fp->fi_hash, &file_hashtbl[hashval]);
1343
		spin_unlock(&recall_lock);
L
Linus Torvalds 已提交
1344 1345
		fp->fi_inode = igrab(ino);
		fp->fi_id = current_fileid++;
1346
		fp->fi_had_conflict = false;
L
Linus Torvalds 已提交
1347 1348 1349 1350 1351
		return fp;
	}
	return NULL;
}

N
NeilBrown 已提交
1352
static void
1353
nfsd4_free_slab(struct kmem_cache **slab)
L
Linus Torvalds 已提交
1354
{
N
NeilBrown 已提交
1355 1356
	if (*slab == NULL)
		return;
1357
	kmem_cache_destroy(*slab);
N
NeilBrown 已提交
1358
	*slab = NULL;
L
Linus Torvalds 已提交
1359 1360
}

1361
void
L
Linus Torvalds 已提交
1362 1363
nfsd4_free_slabs(void)
{
N
NeilBrown 已提交
1364 1365
	nfsd4_free_slab(&stateowner_slab);
	nfsd4_free_slab(&file_slab);
N
NeilBrown 已提交
1366
	nfsd4_free_slab(&stateid_slab);
N
NeilBrown 已提交
1367
	nfsd4_free_slab(&deleg_slab);
N
NeilBrown 已提交
1368
}
L
Linus Torvalds 已提交
1369

N
NeilBrown 已提交
1370 1371 1372 1373
static int
nfsd4_init_slabs(void)
{
	stateowner_slab = kmem_cache_create("nfsd4_stateowners",
1374
			sizeof(struct nfs4_stateowner), 0, 0, NULL);
N
NeilBrown 已提交
1375 1376 1377
	if (stateowner_slab == NULL)
		goto out_nomem;
	file_slab = kmem_cache_create("nfsd4_files",
1378
			sizeof(struct nfs4_file), 0, 0, NULL);
N
NeilBrown 已提交
1379 1380
	if (file_slab == NULL)
		goto out_nomem;
N
NeilBrown 已提交
1381
	stateid_slab = kmem_cache_create("nfsd4_stateids",
1382
			sizeof(struct nfs4_stateid), 0, 0, NULL);
N
NeilBrown 已提交
1383 1384
	if (stateid_slab == NULL)
		goto out_nomem;
N
NeilBrown 已提交
1385
	deleg_slab = kmem_cache_create("nfsd4_delegations",
1386
			sizeof(struct nfs4_delegation), 0, 0, NULL);
N
NeilBrown 已提交
1387 1388
	if (deleg_slab == NULL)
		goto out_nomem;
N
NeilBrown 已提交
1389 1390 1391 1392 1393
	return 0;
out_nomem:
	nfsd4_free_slabs();
	dprintk("nfsd4: out of memory while initializing nfsv4\n");
	return -ENOMEM;
L
Linus Torvalds 已提交
1394 1395 1396 1397 1398 1399 1400 1401 1402 1403 1404 1405 1406 1407 1408 1409 1410 1411 1412 1413 1414 1415 1416 1417 1418 1419 1420 1421 1422 1423 1424 1425 1426 1427 1428 1429 1430 1431 1432 1433
}

void
nfs4_free_stateowner(struct kref *kref)
{
	struct nfs4_stateowner *sop =
		container_of(kref, struct nfs4_stateowner, so_ref);
	kfree(sop->so_owner.data);
	kmem_cache_free(stateowner_slab, sop);
}

static inline struct nfs4_stateowner *
alloc_stateowner(struct xdr_netobj *owner)
{
	struct nfs4_stateowner *sop;

	if ((sop = kmem_cache_alloc(stateowner_slab, GFP_KERNEL))) {
		if ((sop->so_owner.data = kmalloc(owner->len, GFP_KERNEL))) {
			memcpy(sop->so_owner.data, owner->data, owner->len);
			sop->so_owner.len = owner->len;
			kref_init(&sop->so_ref);
			return sop;
		} 
		kmem_cache_free(stateowner_slab, sop);
	}
	return NULL;
}

static struct nfs4_stateowner *
alloc_init_open_stateowner(unsigned int strhashval, struct nfs4_client *clp, struct nfsd4_open *open) {
	struct nfs4_stateowner *sop;
	struct nfs4_replay *rp;
	unsigned int idhashval;

	if (!(sop = alloc_stateowner(&open->op_owner)))
		return NULL;
	idhashval = ownerid_hashval(current_ownerid);
	INIT_LIST_HEAD(&sop->so_idhash);
	INIT_LIST_HEAD(&sop->so_strhash);
	INIT_LIST_HEAD(&sop->so_perclient);
1434 1435
	INIT_LIST_HEAD(&sop->so_stateids);
	INIT_LIST_HEAD(&sop->so_perstateid);  /* not used */
L
Linus Torvalds 已提交
1436 1437 1438 1439
	INIT_LIST_HEAD(&sop->so_close_lru);
	sop->so_time = 0;
	list_add(&sop->so_idhash, &ownerid_hashtbl[idhashval]);
	list_add(&sop->so_strhash, &ownerstr_hashtbl[strhashval]);
1440
	list_add(&sop->so_perclient, &clp->cl_openowners);
L
Linus Torvalds 已提交
1441 1442 1443 1444 1445 1446
	sop->so_is_open_owner = 1;
	sop->so_id = current_ownerid++;
	sop->so_client = clp;
	sop->so_seqid = open->op_seqid;
	sop->so_confirmed = 0;
	rp = &sop->so_replay;
1447
	rp->rp_status = nfserr_serverfault;
L
Linus Torvalds 已提交
1448 1449 1450 1451 1452 1453 1454 1455 1456 1457 1458
	rp->rp_buflen = 0;
	rp->rp_buf = rp->rp_ibuf;
	return sop;
}

static inline void
init_stateid(struct nfs4_stateid *stp, struct nfs4_file *fp, struct nfsd4_open *open) {
	struct nfs4_stateowner *sop = open->op_stateowner;
	unsigned int hashval = stateid_hashval(sop->so_id, fp->fi_id);

	INIT_LIST_HEAD(&stp->st_hash);
1459 1460
	INIT_LIST_HEAD(&stp->st_perstateowner);
	INIT_LIST_HEAD(&stp->st_lockowners);
L
Linus Torvalds 已提交
1461 1462
	INIT_LIST_HEAD(&stp->st_perfile);
	list_add(&stp->st_hash, &stateid_hashtbl[hashval]);
1463
	list_add(&stp->st_perstateowner, &sop->so_stateids);
1464
	list_add(&stp->st_perfile, &fp->fi_stateids);
L
Linus Torvalds 已提交
1465
	stp->st_stateowner = sop;
1466
	get_nfs4_file(fp);
L
Linus Torvalds 已提交
1467 1468 1469 1470 1471 1472 1473 1474 1475
	stp->st_file = fp;
	stp->st_stateid.si_boot = boot_time;
	stp->st_stateid.si_stateownerid = sop->so_id;
	stp->st_stateid.si_fileid = fp->fi_id;
	stp->st_stateid.si_generation = 0;
	stp->st_access_bmap = 0;
	stp->st_deny_bmap = 0;
	__set_bit(open->op_share_access, &stp->st_access_bmap);
	__set_bit(open->op_share_deny, &stp->st_deny_bmap);
1476
	stp->st_openstp = NULL;
L
Linus Torvalds 已提交
1477 1478
}

1479
static void
L
Linus Torvalds 已提交
1480 1481 1482 1483
move_to_close_lru(struct nfs4_stateowner *sop)
{
	dprintk("NFSD: move_to_close_lru nfs4_stateowner %p\n", sop);

1484
	list_move_tail(&sop->so_close_lru, &close_lru);
L
Linus Torvalds 已提交
1485 1486 1487 1488
	sop->so_time = get_seconds();
}

static int
1489 1490 1491 1492 1493 1494
same_owner_str(struct nfs4_stateowner *sop, struct xdr_netobj *owner,
							clientid_t *clid)
{
	return (sop->so_owner.len == owner->len) &&
		0 == memcmp(sop->so_owner.data, owner->data, owner->len) &&
		(sop->so_client->cl_clientid.cl_id == clid->cl_id);
L
Linus Torvalds 已提交
1495 1496 1497 1498 1499 1500 1501 1502
}

static struct nfs4_stateowner *
find_openstateowner_str(unsigned int hashval, struct nfsd4_open *open)
{
	struct nfs4_stateowner *so = NULL;

	list_for_each_entry(so, &ownerstr_hashtbl[hashval], so_strhash) {
1503
		if (same_owner_str(so, &open->op_owner, &open->op_clientid))
L
Linus Torvalds 已提交
1504 1505 1506 1507 1508 1509 1510 1511 1512 1513 1514 1515
			return so;
	}
	return NULL;
}

/* search file_hashtbl[] for file */
static struct nfs4_file *
find_file(struct inode *ino)
{
	unsigned int hashval = file_hashval(ino);
	struct nfs4_file *fp;

1516
	spin_lock(&recall_lock);
L
Linus Torvalds 已提交
1517
	list_for_each_entry(fp, &file_hashtbl[hashval], fi_hash) {
1518 1519
		if (fp->fi_inode == ino) {
			get_nfs4_file(fp);
1520
			spin_unlock(&recall_lock);
L
Linus Torvalds 已提交
1521
			return fp;
1522
		}
L
Linus Torvalds 已提交
1523
	}
1524
	spin_unlock(&recall_lock);
L
Linus Torvalds 已提交
1525 1526 1527
	return NULL;
}

1528
static inline int access_valid(u32 x)
1529
{
1530 1531 1532 1533 1534
	if (x < NFS4_SHARE_ACCESS_READ)
		return 0;
	if (x > NFS4_SHARE_ACCESS_BOTH)
		return 0;
	return 1;
1535 1536
}

1537
static inline int deny_valid(u32 x)
1538
{
1539 1540
	/* Note: unlike access bits, deny bits may be zero. */
	return x <= NFS4_SHARE_DENY_BOTH;
1541
}
L
Linus Torvalds 已提交
1542

1543 1544 1545 1546 1547 1548 1549 1550 1551 1552 1553 1554 1555 1556 1557 1558 1559 1560
/*
 * We store the NONE, READ, WRITE, and BOTH bits separately in the
 * st_{access,deny}_bmap field of the stateid, in order to track not
 * only what share bits are currently in force, but also what
 * combinations of share bits previous opens have used.  This allows us
 * to enforce the recommendation of rfc 3530 14.2.19 that the server
 * return an error if the client attempt to downgrade to a combination
 * of share bits not explicable by closing some of its previous opens.
 *
 * XXX: This enforcement is actually incomplete, since we don't keep
 * track of access/deny bit combinations; so, e.g., we allow:
 *
 *	OPEN allow read, deny write
 *	OPEN allow both, deny none
 *	DOWNGRADE allow read, deny none
 *
 * which we should reject.
 */
1561
static void
L
Linus Torvalds 已提交
1562 1563 1564 1565 1566 1567 1568 1569 1570 1571
set_access(unsigned int *access, unsigned long bmap) {
	int i;

	*access = 0;
	for (i = 1; i < 4; i++) {
		if (test_bit(i, &bmap))
			*access |= i;
	}
}

1572
static void
L
Linus Torvalds 已提交
1573 1574 1575 1576 1577 1578 1579 1580 1581 1582 1583 1584 1585 1586 1587 1588 1589 1590 1591 1592 1593 1594 1595 1596 1597
set_deny(unsigned int *deny, unsigned long bmap) {
	int i;

	*deny = 0;
	for (i = 0; i < 4; i++) {
		if (test_bit(i, &bmap))
			*deny |= i ;
	}
}

static int
test_share(struct nfs4_stateid *stp, struct nfsd4_open *open) {
	unsigned int access, deny;

	set_access(&access, stp->st_access_bmap);
	set_deny(&deny, stp->st_deny_bmap);
	if ((access & open->op_share_deny) || (deny & open->op_share_access))
		return 0;
	return 1;
}

/*
 * Called to check deny when READ with all zero stateid or
 * WRITE with all zero or all one stateid
 */
1598
static __be32
L
Linus Torvalds 已提交
1599 1600 1601 1602 1603
nfs4_share_conflict(struct svc_fh *current_fh, unsigned int deny_type)
{
	struct inode *ino = current_fh->fh_dentry->d_inode;
	struct nfs4_file *fp;
	struct nfs4_stateid *stp;
1604
	__be32 ret;
L
Linus Torvalds 已提交
1605 1606 1607 1608

	dprintk("NFSD: nfs4_share_conflict\n");

	fp = find_file(ino);
1609 1610
	if (!fp)
		return nfs_ok;
1611
	ret = nfserr_locked;
L
Linus Torvalds 已提交
1612
	/* Search for conflicting share reservations */
1613 1614 1615 1616
	list_for_each_entry(stp, &fp->fi_stateids, st_perfile) {
		if (test_bit(deny_type, &stp->st_deny_bmap) ||
		    test_bit(NFS4_SHARE_DENY_BOTH, &stp->st_deny_bmap))
			goto out;
L
Linus Torvalds 已提交
1617
	}
1618 1619 1620 1621
	ret = nfs_ok;
out:
	put_nfs4_file(fp);
	return ret;
L
Linus Torvalds 已提交
1622 1623 1624 1625 1626 1627
}

static inline void
nfs4_file_downgrade(struct file *filp, unsigned int share_access)
{
	if (share_access & NFS4_SHARE_ACCESS_WRITE) {
1628
		drop_file_write_access(filp);
L
Linus Torvalds 已提交
1629 1630 1631 1632 1633 1634 1635 1636 1637 1638 1639 1640
		filp->f_mode = (filp->f_mode | FMODE_READ) & ~FMODE_WRITE;
	}
}

/*
 * Recall a delegation
 */
static int
do_recall(void *__dp)
{
	struct nfs4_delegation *dp = __dp;

1641
	dp->dl_file->fi_had_conflict = true;
L
Linus Torvalds 已提交
1642 1643 1644 1645 1646 1647 1648 1649 1650 1651 1652 1653 1654 1655 1656 1657 1658 1659 1660 1661 1662 1663 1664 1665 1666 1667 1668 1669
	nfsd4_cb_recall(dp);
	return 0;
}

/*
 * Spawn a thread to perform a recall on the delegation represented
 * by the lease (file_lock)
 *
 * Called from break_lease() with lock_kernel() held.
 * Note: we assume break_lease will only call this *once* for any given
 * lease.
 */
static
void nfsd_break_deleg_cb(struct file_lock *fl)
{
	struct nfs4_delegation *dp=  (struct nfs4_delegation *)fl->fl_owner;
	struct task_struct *t;

	dprintk("NFSD nfsd_break_deleg_cb: dp %p fl %p\n",dp,fl);
	if (!dp)
		return;

	/* We're assuming the state code never drops its reference
	 * without first removing the lease.  Since we're in this lease
	 * callback (and since the lease code is serialized by the kernel
	 * lock) we know the server hasn't removed the lease yet, we know
	 * it's safe to take a reference: */
	atomic_inc(&dp->dl_count);
1670
	atomic_inc(&dp->dl_client->cl_count);
L
Linus Torvalds 已提交
1671 1672 1673 1674 1675 1676 1677 1678

	spin_lock(&recall_lock);
	list_add_tail(&dp->dl_recall_lru, &del_recall_lru);
	spin_unlock(&recall_lock);

	/* only place dl_time is set. protected by lock_kernel*/
	dp->dl_time = get_seconds();

1679 1680 1681 1682 1683 1684
	/*
	 * We don't want the locks code to timeout the lease for us;
	 * we'll remove it ourself if the delegation isn't returned
	 * in time.
	 */
	fl->fl_break_time = 0;
L
Linus Torvalds 已提交
1685 1686 1687 1688 1689 1690 1691 1692

	t = kthread_run(do_recall, dp, "%s", "nfs4_cb_recall");
	if (IS_ERR(t)) {
		struct nfs4_client *clp = dp->dl_client;

		printk(KERN_INFO "NFSD: Callback thread failed for "
			"for client (clientid %08x/%08x)\n",
			clp->cl_clientid.cl_boot, clp->cl_clientid.cl_id);
1693
		put_nfs4_client(dp->dl_client);
L
Linus Torvalds 已提交
1694 1695 1696 1697 1698 1699 1700 1701 1702 1703 1704 1705 1706 1707 1708 1709 1710 1711 1712 1713 1714 1715 1716 1717
		nfs4_put_delegation(dp);
	}
}

/*
 * The file_lock is being reapd.
 *
 * Called by locks_free_lock() with lock_kernel() held.
 */
static
void nfsd_release_deleg_cb(struct file_lock *fl)
{
	struct nfs4_delegation *dp = (struct nfs4_delegation *)fl->fl_owner;

	dprintk("NFSD nfsd_release_deleg_cb: fl %p dp %p dl_count %d\n", fl,dp, atomic_read(&dp->dl_count));

	if (!(fl->fl_flags & FL_LEASE) || !dp)
		return;
	dp->dl_flock = NULL;
}

/*
 * Set the delegation file_lock back pointer.
 *
1718
 * Called from setlease() with lock_kernel() held.
L
Linus Torvalds 已提交
1719 1720 1721 1722 1723 1724 1725 1726 1727 1728 1729 1730 1731
 */
static
void nfsd_copy_lock_deleg_cb(struct file_lock *new, struct file_lock *fl)
{
	struct nfs4_delegation *dp = (struct nfs4_delegation *)new->fl_owner;

	dprintk("NFSD: nfsd_copy_lock_deleg_cb: new fl %p dp %p\n", new, dp);
	if (!dp)
		return;
	dp->dl_flock = new;
}

/*
1732
 * Called from setlease() with lock_kernel() held
L
Linus Torvalds 已提交
1733 1734 1735 1736 1737 1738 1739 1740 1741 1742 1743 1744 1745 1746 1747 1748 1749 1750 1751 1752 1753 1754 1755 1756 1757
 */
static
int nfsd_same_client_deleg_cb(struct file_lock *onlist, struct file_lock *try)
{
	struct nfs4_delegation *onlistd =
		(struct nfs4_delegation *)onlist->fl_owner;
	struct nfs4_delegation *tryd =
		(struct nfs4_delegation *)try->fl_owner;

	if (onlist->fl_lmops != try->fl_lmops)
		return 0;

	return onlistd->dl_client == tryd->dl_client;
}


static
int nfsd_change_deleg_cb(struct file_lock **onlist, int arg)
{
	if (arg & F_UNLCK)
		return lease_modify(onlist, arg);
	else
		return -EAGAIN;
}

1758
static struct lock_manager_operations nfsd_lease_mng_ops = {
L
Linus Torvalds 已提交
1759 1760 1761 1762 1763 1764 1765 1766
	.fl_break = nfsd_break_deleg_cb,
	.fl_release_private = nfsd_release_deleg_cb,
	.fl_copy_lock = nfsd_copy_lock_deleg_cb,
	.fl_mylease = nfsd_same_client_deleg_cb,
	.fl_change = nfsd_change_deleg_cb,
};


1767
__be32
L
Linus Torvalds 已提交
1768 1769 1770 1771 1772 1773 1774 1775
nfsd4_process_open1(struct nfsd4_open *open)
{
	clientid_t *clientid = &open->op_clientid;
	struct nfs4_client *clp = NULL;
	unsigned int strhashval;
	struct nfs4_stateowner *sop = NULL;

	if (!check_name(open->op_owner))
1776
		return nfserr_inval;
L
Linus Torvalds 已提交
1777 1778 1779 1780 1781 1782

	if (STALE_CLIENTID(&open->op_clientid))
		return nfserr_stale_clientid;

	strhashval = ownerstr_hashval(clientid->cl_id, open->op_owner);
	sop = find_openstateowner_str(strhashval, open);
1783 1784 1785
	open->op_stateowner = sop;
	if (!sop) {
		/* Make sure the client's lease hasn't expired. */
L
Linus Torvalds 已提交
1786 1787
		clp = find_confirmed_client(clientid);
		if (clp == NULL)
1788 1789
			return nfserr_expired;
		goto renew;
L
Linus Torvalds 已提交
1790
	}
1791 1792 1793
	if (!sop->so_confirmed) {
		/* Replace unconfirmed owners without checking for replay. */
		clp = sop->so_client;
1794
		release_openowner(sop);
1795 1796 1797 1798 1799
		open->op_stateowner = NULL;
		goto renew;
	}
	if (open->op_seqid == sop->so_seqid - 1) {
		if (sop->so_replay.rp_buflen)
A
Al Viro 已提交
1800
			return nfserr_replay_me;
1801 1802 1803 1804 1805 1806 1807 1808 1809 1810 1811
		/* The original OPEN failed so spectacularly
		 * that we don't even have replay data saved!
		 * Therefore, we have no choice but to continue
		 * processing this OPEN; presumably, we'll
		 * fail again for the same reason.
		 */
		dprintk("nfsd4_process_open1: replay with no replay cache\n");
		goto renew;
	}
	if (open->op_seqid != sop->so_seqid)
		return nfserr_bad_seqid;
L
Linus Torvalds 已提交
1812
renew:
1813 1814 1815 1816 1817 1818 1819
	if (open->op_stateowner == NULL) {
		sop = alloc_init_open_stateowner(strhashval, clp, open);
		if (sop == NULL)
			return nfserr_resource;
		open->op_stateowner = sop;
	}
	list_del_init(&sop->so_close_lru);
L
Linus Torvalds 已提交
1820
	renew_client(sop->so_client);
1821
	return nfs_ok;
L
Linus Torvalds 已提交
1822 1823
}

1824
static inline __be32
N
NeilBrown 已提交
1825 1826 1827 1828 1829 1830 1831 1832
nfs4_check_delegmode(struct nfs4_delegation *dp, int flags)
{
	if ((flags & WR_STATE) && (dp->dl_type == NFS4_OPEN_DELEGATE_READ))
		return nfserr_openmode;
	else
		return nfs_ok;
}

1833 1834 1835 1836 1837
static struct nfs4_delegation *
find_delegation_file(struct nfs4_file *fp, stateid_t *stid)
{
	struct nfs4_delegation *dp;

1838
	list_for_each_entry(dp, &fp->fi_delegations, dl_perfile) {
1839 1840 1841 1842 1843 1844
		if (dp->dl_stateid.si_stateownerid == stid->si_stateownerid)
			return dp;
	}
	return NULL;
}

1845
static __be32
1846 1847 1848 1849
nfs4_check_deleg(struct nfs4_file *fp, struct nfsd4_open *open,
		struct nfs4_delegation **dp)
{
	int flags;
1850
	__be32 status = nfserr_bad_stateid;
1851 1852 1853

	*dp = find_delegation_file(fp, &open->op_delegate_stateid);
	if (*dp == NULL)
1854
		goto out;
1855 1856 1857 1858 1859
	flags = open->op_share_access == NFS4_SHARE_ACCESS_READ ?
						RD_STATE : WR_STATE;
	status = nfs4_check_delegmode(*dp, flags);
	if (status)
		*dp = NULL;
1860 1861 1862 1863 1864 1865 1866
out:
	if (open->op_claim_type != NFS4_OPEN_CLAIM_DELEGATE_CUR)
		return nfs_ok;
	if (status)
		return status;
	open->op_stateowner->so_confirmed = 1;
	return nfs_ok;
1867 1868
}

1869
static __be32
L
Linus Torvalds 已提交
1870 1871 1872
nfs4_check_open(struct nfs4_file *fp, struct nfsd4_open *open, struct nfs4_stateid **stpp)
{
	struct nfs4_stateid *local;
1873
	__be32 status = nfserr_share_denied;
L
Linus Torvalds 已提交
1874 1875
	struct nfs4_stateowner *sop = open->op_stateowner;

1876
	list_for_each_entry(local, &fp->fi_stateids, st_perfile) {
L
Linus Torvalds 已提交
1877 1878 1879 1880 1881 1882 1883 1884 1885 1886 1887 1888 1889 1890 1891
		/* ignore lock owners */
		if (local->st_stateowner->so_is_open_owner == 0)
			continue;
		/* remember if we have seen this open owner */
		if (local->st_stateowner == sop)
			*stpp = local;
		/* check for conflicting share reservations */
		if (!test_share(local, open))
			goto out;
	}
	status = 0;
out:
	return status;
}

N
NeilBrown 已提交
1892 1893 1894 1895 1896 1897
static inline struct nfs4_stateid *
nfs4_alloc_stateid(void)
{
	return kmem_cache_alloc(stateid_slab, GFP_KERNEL);
}

1898
static __be32
L
Linus Torvalds 已提交
1899
nfs4_new_open(struct svc_rqst *rqstp, struct nfs4_stateid **stpp,
1900
		struct nfs4_delegation *dp,
L
Linus Torvalds 已提交
1901 1902 1903 1904
		struct svc_fh *cur_fh, int flags)
{
	struct nfs4_stateid *stp;

N
NeilBrown 已提交
1905
	stp = nfs4_alloc_stateid();
L
Linus Torvalds 已提交
1906 1907 1908
	if (stp == NULL)
		return nfserr_resource;

1909 1910 1911 1912
	if (dp) {
		get_file(dp->dl_vfs_file);
		stp->st_vfs_file = dp->dl_vfs_file;
	} else {
1913
		__be32 status;
1914 1915 1916 1917 1918
		status = nfsd_open(rqstp, cur_fh, S_IFREG, flags,
				&stp->st_vfs_file);
		if (status) {
			if (status == nfserr_dropit)
				status = nfserr_jukebox;
N
NeilBrown 已提交
1919
			kmem_cache_free(stateid_slab, stp);
1920 1921
			return status;
		}
L
Linus Torvalds 已提交
1922 1923 1924 1925 1926
	}
	*stpp = stp;
	return 0;
}

1927
static inline __be32
L
Linus Torvalds 已提交
1928 1929 1930 1931 1932 1933 1934 1935 1936 1937
nfsd4_truncate(struct svc_rqst *rqstp, struct svc_fh *fh,
		struct nfsd4_open *open)
{
	struct iattr iattr = {
		.ia_valid = ATTR_SIZE,
		.ia_size = 0,
	};
	if (!open->op_truncate)
		return 0;
	if (!(open->op_share_access & NFS4_SHARE_ACCESS_WRITE))
1938
		return nfserr_inval;
L
Linus Torvalds 已提交
1939 1940 1941
	return nfsd_setattr(rqstp, fh, &iattr, 0, (time_t)0);
}

1942
static __be32
L
Linus Torvalds 已提交
1943 1944 1945
nfs4_upgrade_open(struct svc_rqst *rqstp, struct svc_fh *cur_fh, struct nfs4_stateid *stp, struct nfsd4_open *open)
{
	struct file *filp = stp->st_vfs_file;
1946
	struct inode *inode = filp->f_path.dentry->d_inode;
1947
	unsigned int share_access, new_writer;
1948
	__be32 status;
L
Linus Torvalds 已提交
1949 1950

	set_access(&share_access, stp->st_access_bmap);
1951 1952
	new_writer = (~share_access) & open->op_share_access
			& NFS4_SHARE_ACCESS_WRITE;
L
Linus Torvalds 已提交
1953

1954
	if (new_writer) {
1955 1956 1957
		int err = get_write_access(inode);
		if (err)
			return nfserrno(err);
1958 1959 1960 1961
		err = mnt_want_write(cur_fh->fh_export->ex_path.mnt);
		if (err)
			return nfserrno(err);
		file_take_write(filp);
1962
	}
L
Linus Torvalds 已提交
1963 1964
	status = nfsd4_truncate(rqstp, cur_fh, open);
	if (status) {
1965 1966
		if (new_writer)
			put_write_access(inode);
L
Linus Torvalds 已提交
1967 1968 1969
		return status;
	}
	/* remember the open */
1970
	filp->f_mode |= open->op_share_access;
1971 1972
	__set_bit(open->op_share_access, &stp->st_access_bmap);
	__set_bit(open->op_share_deny, &stp->st_deny_bmap);
L
Linus Torvalds 已提交
1973 1974 1975 1976 1977 1978

	return nfs_ok;
}


static void
1979
nfs4_set_claim_prev(struct nfsd4_open *open)
L
Linus Torvalds 已提交
1980
{
1981 1982
	open->op_stateowner->so_confirmed = 1;
	open->op_stateowner->so_client->cl_firststate = 1;
L
Linus Torvalds 已提交
1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997
}

/*
 * Attempt to hand out a delegation.
 */
static void
nfs4_open_delegation(struct svc_fh *fh, struct nfsd4_open *open, struct nfs4_stateid *stp)
{
	struct nfs4_delegation *dp;
	struct nfs4_stateowner *sop = stp->st_stateowner;
	struct nfs4_callback *cb = &sop->so_client->cl_callback;
	struct file_lock fl, *flp = &fl;
	int status, flag = 0;

	flag = NFS4_OPEN_DELEGATE_NONE;
1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009
	open->op_recall = 0;
	switch (open->op_claim_type) {
		case NFS4_OPEN_CLAIM_PREVIOUS:
			if (!atomic_read(&cb->cb_set))
				open->op_recall = 1;
			flag = open->op_delegate_type;
			if (flag == NFS4_OPEN_DELEGATE_NONE)
				goto out;
			break;
		case NFS4_OPEN_CLAIM_NULL:
			/* Let's not give out any delegations till everyone's
			 * had the chance to reclaim theirs.... */
2010
			if (locks_in_grace())
2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021
				goto out;
			if (!atomic_read(&cb->cb_set) || !sop->so_confirmed)
				goto out;
			if (open->op_share_access & NFS4_SHARE_ACCESS_WRITE)
				flag = NFS4_OPEN_DELEGATE_WRITE;
			else
				flag = NFS4_OPEN_DELEGATE_READ;
			break;
		default:
			goto out;
	}
L
Linus Torvalds 已提交
2022 2023 2024 2025 2026 2027 2028 2029 2030

	dp = alloc_init_deleg(sop->so_client, stp, fh, flag);
	if (dp == NULL) {
		flag = NFS4_OPEN_DELEGATE_NONE;
		goto out;
	}
	locks_init_lock(&fl);
	fl.fl_lmops = &nfsd_lease_mng_ops;
	fl.fl_flags = FL_LEASE;
2031
	fl.fl_type = flag == NFS4_OPEN_DELEGATE_READ? F_RDLCK: F_WRLCK;
L
Linus Torvalds 已提交
2032 2033 2034 2035 2036
	fl.fl_end = OFFSET_MAX;
	fl.fl_owner =  (fl_owner_t)dp;
	fl.fl_file = stp->st_vfs_file;
	fl.fl_pid = current->tgid;

2037
	/* vfs_setlease checks to see if delegation should be handed out.
L
Linus Torvalds 已提交
2038 2039
	 * the lock_manager callbacks fl_mylease and fl_change are used
	 */
2040
	if ((status = vfs_setlease(stp->st_vfs_file, fl.fl_type, &flp))) {
L
Linus Torvalds 已提交
2041
		dprintk("NFSD: setlease failed [%d], no delegation\n", status);
N
NeilBrown 已提交
2042
		unhash_delegation(dp);
L
Linus Torvalds 已提交
2043 2044 2045 2046 2047 2048 2049 2050 2051 2052 2053 2054
		flag = NFS4_OPEN_DELEGATE_NONE;
		goto out;
	}

	memcpy(&open->op_delegate_stateid, &dp->dl_stateid, sizeof(dp->dl_stateid));

	dprintk("NFSD: delegation stateid=(%08x/%08x/%08x/%08x)\n\n",
	             dp->dl_stateid.si_boot,
	             dp->dl_stateid.si_stateownerid,
	             dp->dl_stateid.si_fileid,
	             dp->dl_stateid.si_generation);
out:
2055 2056 2057
	if (open->op_claim_type == NFS4_OPEN_CLAIM_PREVIOUS
			&& flag == NFS4_OPEN_DELEGATE_NONE
			&& open->op_delegate_type != NFS4_OPEN_DELEGATE_NONE)
2058
		dprintk("NFSD: WARNING: refusing delegation reclaim\n");
L
Linus Torvalds 已提交
2059 2060 2061 2062 2063 2064
	open->op_delegate_type = flag;
}

/*
 * called with nfs4_lock_state() held.
 */
2065
__be32
L
Linus Torvalds 已提交
2066 2067 2068 2069 2070
nfsd4_process_open2(struct svc_rqst *rqstp, struct svc_fh *current_fh, struct nfsd4_open *open)
{
	struct nfs4_file *fp = NULL;
	struct inode *ino = current_fh->fh_dentry->d_inode;
	struct nfs4_stateid *stp = NULL;
2071
	struct nfs4_delegation *dp = NULL;
2072
	__be32 status;
L
Linus Torvalds 已提交
2073 2074

	status = nfserr_inval;
2075 2076
	if (!access_valid(open->op_share_access)
			|| !deny_valid(open->op_share_deny))
L
Linus Torvalds 已提交
2077 2078 2079 2080 2081 2082 2083 2084 2085 2086
		goto out;
	/*
	 * Lookup file; if found, lookup stateid and check open request,
	 * and check for delegations in the process of being recalled.
	 * If not found, create the nfs4_file struct
	 */
	fp = find_file(ino);
	if (fp) {
		if ((status = nfs4_check_open(fp, open, &stp)))
			goto out;
2087 2088 2089
		status = nfs4_check_deleg(fp, open, &dp);
		if (status)
			goto out;
L
Linus Torvalds 已提交
2090
	} else {
2091 2092 2093
		status = nfserr_bad_stateid;
		if (open->op_claim_type == NFS4_OPEN_CLAIM_DELEGATE_CUR)
			goto out;
L
Linus Torvalds 已提交
2094 2095 2096 2097 2098 2099 2100 2101 2102 2103 2104 2105 2106 2107 2108
		status = nfserr_resource;
		fp = alloc_init_file(ino);
		if (fp == NULL)
			goto out;
	}

	/*
	 * OPEN the file, or upgrade an existing OPEN.
	 * If truncate fails, the OPEN fails.
	 */
	if (stp) {
		/* Stateid was found, this is an OPEN upgrade */
		status = nfs4_upgrade_open(rqstp, current_fh, stp, open);
		if (status)
			goto out;
2109
		update_stateid(&stp->st_stateid);
L
Linus Torvalds 已提交
2110 2111 2112
	} else {
		/* Stateid was not found, this is a new OPEN */
		int flags = 0;
2113
		if (open->op_share_access & NFS4_SHARE_ACCESS_READ)
M
Miklos Szeredi 已提交
2114
			flags |= NFSD_MAY_READ;
L
Linus Torvalds 已提交
2115
		if (open->op_share_access & NFS4_SHARE_ACCESS_WRITE)
M
Miklos Szeredi 已提交
2116
			flags |= NFSD_MAY_WRITE;
2117 2118
		status = nfs4_new_open(rqstp, &stp, dp, current_fh, flags);
		if (status)
L
Linus Torvalds 已提交
2119 2120 2121 2122
			goto out;
		init_stateid(stp, fp, open);
		status = nfsd4_truncate(rqstp, current_fh, open);
		if (status) {
2123
			release_open_stateid(stp);
L
Linus Torvalds 已提交
2124 2125 2126 2127 2128 2129 2130 2131 2132 2133 2134 2135 2136 2137 2138 2139 2140
			goto out;
		}
	}
	memcpy(&open->op_stateid, &stp->st_stateid, sizeof(stateid_t));

	/*
	* Attempt to hand out a delegation. No error return, because the
	* OPEN succeeds even if we fail.
	*/
	nfs4_open_delegation(current_fh, open, stp);

	status = nfs_ok;

	dprintk("nfs4_process_open2: stateid=(%08x/%08x/%08x/%08x)\n",
	            stp->st_stateid.si_boot, stp->st_stateid.si_stateownerid,
	            stp->st_stateid.si_fileid, stp->st_stateid.si_generation);
out:
2141 2142
	if (fp)
		put_nfs4_file(fp);
2143 2144
	if (status == 0 && open->op_claim_type == NFS4_OPEN_CLAIM_PREVIOUS)
		nfs4_set_claim_prev(open);
L
Linus Torvalds 已提交
2145 2146 2147 2148 2149 2150 2151 2152 2153 2154
	/*
	* To finish the open response, we just need to set the rflags.
	*/
	open->op_rflags = NFS4_OPEN_RESULT_LOCKTYPE_POSIX;
	if (!open->op_stateowner->so_confirmed)
		open->op_rflags |= NFS4_OPEN_RESULT_CONFIRM;

	return status;
}

2155
__be32
2156 2157
nfsd4_renew(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate,
	    clientid_t *clid)
L
Linus Torvalds 已提交
2158 2159
{
	struct nfs4_client *clp;
2160
	__be32 status;
L
Linus Torvalds 已提交
2161 2162 2163 2164 2165 2166 2167 2168 2169 2170 2171 2172 2173 2174 2175 2176

	nfs4_lock_state();
	dprintk("process_renew(%08x/%08x): starting\n", 
			clid->cl_boot, clid->cl_id);
	status = nfserr_stale_clientid;
	if (STALE_CLIENTID(clid))
		goto out;
	clp = find_confirmed_client(clid);
	status = nfserr_expired;
	if (clp == NULL) {
		/* We assume the client took too long to RENEW. */
		dprintk("nfsd4_renew: clientid not found!\n");
		goto out;
	}
	renew_client(clp);
	status = nfserr_cb_path_down;
2177
	if (!list_empty(&clp->cl_delegations)
L
Linus Torvalds 已提交
2178 2179 2180 2181 2182 2183 2184 2185
			&& !atomic_read(&clp->cl_callback.cb_set))
		goto out;
	status = nfs_ok;
out:
	nfs4_unlock_state();
	return status;
}

2186 2187 2188
struct lock_manager nfsd4_manager = {
};

2189
static void
2190
nfsd4_end_grace(void)
2191 2192
{
	dprintk("NFSD: end of grace period\n");
2193
	nfsd4_recdir_purge_old();
2194
	locks_end_grace(&nfsd4_manager);
2195 2196
}

2197
static time_t
L
Linus Torvalds 已提交
2198 2199 2200 2201 2202 2203 2204 2205 2206 2207 2208 2209 2210
nfs4_laundromat(void)
{
	struct nfs4_client *clp;
	struct nfs4_stateowner *sop;
	struct nfs4_delegation *dp;
	struct list_head *pos, *next, reaplist;
	time_t cutoff = get_seconds() - NFSD_LEASE_TIME;
	time_t t, clientid_val = NFSD_LEASE_TIME;
	time_t u, test_val = NFSD_LEASE_TIME;

	nfs4_lock_state();

	dprintk("NFSD: laundromat service - starting\n");
2211 2212
	if (locks_in_grace())
		nfsd4_end_grace();
L
Linus Torvalds 已提交
2213 2214 2215 2216 2217 2218 2219 2220 2221 2222
	list_for_each_safe(pos, next, &client_lru) {
		clp = list_entry(pos, struct nfs4_client, cl_lru);
		if (time_after((unsigned long)clp->cl_time, (unsigned long)cutoff)) {
			t = clp->cl_time - cutoff;
			if (clientid_val > t)
				clientid_val = t;
			break;
		}
		dprintk("NFSD: purging unused client (clientid %08x)\n",
			clp->cl_clientid.cl_id);
2223
		nfsd4_remove_clid_dir(clp);
L
Linus Torvalds 已提交
2224 2225 2226 2227 2228 2229 2230 2231 2232 2233 2234 2235 2236 2237 2238 2239 2240 2241 2242 2243 2244 2245 2246 2247 2248 2249 2250 2251 2252 2253 2254 2255 2256
		expire_client(clp);
	}
	INIT_LIST_HEAD(&reaplist);
	spin_lock(&recall_lock);
	list_for_each_safe(pos, next, &del_recall_lru) {
		dp = list_entry (pos, struct nfs4_delegation, dl_recall_lru);
		if (time_after((unsigned long)dp->dl_time, (unsigned long)cutoff)) {
			u = dp->dl_time - cutoff;
			if (test_val > u)
				test_val = u;
			break;
		}
		dprintk("NFSD: purging unused delegation dp %p, fp %p\n",
			            dp, dp->dl_flock);
		list_move(&dp->dl_recall_lru, &reaplist);
	}
	spin_unlock(&recall_lock);
	list_for_each_safe(pos, next, &reaplist) {
		dp = list_entry (pos, struct nfs4_delegation, dl_recall_lru);
		list_del_init(&dp->dl_recall_lru);
		unhash_delegation(dp);
	}
	test_val = NFSD_LEASE_TIME;
	list_for_each_safe(pos, next, &close_lru) {
		sop = list_entry(pos, struct nfs4_stateowner, so_close_lru);
		if (time_after((unsigned long)sop->so_time, (unsigned long)cutoff)) {
			u = sop->so_time - cutoff;
			if (test_val > u)
				test_val = u;
			break;
		}
		dprintk("NFSD: purging unused open stateowner (so_id %d)\n",
			sop->so_id);
2257
		release_openowner(sop);
L
Linus Torvalds 已提交
2258 2259 2260 2261 2262 2263 2264
	}
	if (clientid_val < NFSD_LAUNDROMAT_MINTIMEOUT)
		clientid_val = NFSD_LAUNDROMAT_MINTIMEOUT;
	nfs4_unlock_state();
	return clientid_val;
}

H
Harvey Harrison 已提交
2265 2266 2267 2268 2269
static struct workqueue_struct *laundry_wq;
static void laundromat_main(struct work_struct *);
static DECLARE_DELAYED_WORK(laundromat_work, laundromat_main);

static void
D
David Howells 已提交
2270
laundromat_main(struct work_struct *not_used)
L
Linus Torvalds 已提交
2271 2272 2273 2274 2275
{
	time_t t;

	t = nfs4_laundromat();
	dprintk("NFSD: laundromat_main - sleeping for %ld seconds\n", t);
2276
	queue_delayed_work(laundry_wq, &laundromat_work, t*HZ);
L
Linus Torvalds 已提交
2277 2278
}

2279
static struct nfs4_stateowner *
2280 2281
search_close_lru(u32 st_id, int flags)
{
L
Linus Torvalds 已提交
2282 2283 2284 2285 2286 2287 2288 2289 2290 2291 2292 2293 2294 2295
	struct nfs4_stateowner *local = NULL;

	if (flags & CLOSE_STATE) {
		list_for_each_entry(local, &close_lru, so_close_lru) {
			if (local->so_id == st_id)
				return local;
		}
	}
	return NULL;
}

static inline int
nfs4_check_fh(struct svc_fh *fhp, struct nfs4_stateid *stp)
{
2296
	return fhp->fh_dentry->d_inode != stp->st_vfs_file->f_path.dentry->d_inode;
L
Linus Torvalds 已提交
2297 2298 2299 2300 2301 2302 2303
}

static int
STALE_STATEID(stateid_t *stateid)
{
	if (stateid->si_boot == boot_time)
		return 0;
N
Neil Brown 已提交
2304
	dprintk("NFSD: stale stateid (%08x/%08x/%08x/%08x)!\n",
L
Linus Torvalds 已提交
2305 2306 2307 2308 2309 2310 2311 2312 2313 2314 2315 2316 2317 2318 2319 2320 2321 2322 2323 2324 2325
		stateid->si_boot, stateid->si_stateownerid, stateid->si_fileid,
		stateid->si_generation);
	return 1;
}

static inline int
access_permit_read(unsigned long access_bmap)
{
	return test_bit(NFS4_SHARE_ACCESS_READ, &access_bmap) ||
		test_bit(NFS4_SHARE_ACCESS_BOTH, &access_bmap) ||
		test_bit(NFS4_SHARE_ACCESS_WRITE, &access_bmap);
}

static inline int
access_permit_write(unsigned long access_bmap)
{
	return test_bit(NFS4_SHARE_ACCESS_WRITE, &access_bmap) ||
		test_bit(NFS4_SHARE_ACCESS_BOTH, &access_bmap);
}

static
2326
__be32 nfs4_check_openmode(struct nfs4_stateid *stp, int flags)
L
Linus Torvalds 已提交
2327
{
2328
        __be32 status = nfserr_openmode;
L
Linus Torvalds 已提交
2329 2330 2331 2332 2333 2334 2335 2336 2337 2338

	if ((flags & WR_STATE) && (!access_permit_write(stp->st_access_bmap)))
                goto out;
	if ((flags & RD_STATE) && (!access_permit_read(stp->st_access_bmap)))
                goto out;
	status = nfs_ok;
out:
	return status;
}

2339
static inline __be32
L
Linus Torvalds 已提交
2340 2341
check_special_stateids(svc_fh *current_fh, stateid_t *stateid, int flags)
{
2342
	if (ONE_STATEID(stateid) && (flags & RD_STATE))
L
Linus Torvalds 已提交
2343
		return nfs_ok;
2344
	else if (locks_in_grace()) {
L
Linus Torvalds 已提交
2345 2346 2347 2348 2349 2350 2351 2352 2353 2354 2355 2356 2357 2358 2359 2360
		/* Answer in remaining cases depends on existance of
		 * conflicting state; so we must wait out the grace period. */
		return nfserr_grace;
	} else if (flags & WR_STATE)
		return nfs4_share_conflict(current_fh,
				NFS4_SHARE_DENY_WRITE);
	else /* (flags & RD_STATE) && ZERO_STATEID(stateid) */
		return nfs4_share_conflict(current_fh,
				NFS4_SHARE_DENY_READ);
}

/*
 * Allow READ/WRITE during grace period on recovered state only for files
 * that are not able to provide mandatory locking.
 */
static inline int
2361
grace_disallows_io(struct inode *inode)
L
Linus Torvalds 已提交
2362
{
2363
	return locks_in_grace() && mandatory_lock(inode);
L
Linus Torvalds 已提交
2364 2365
}

2366 2367 2368 2369 2370 2371 2372 2373 2374 2375 2376 2377 2378 2379 2380 2381 2382 2383 2384 2385
static int check_stateid_generation(stateid_t *in, stateid_t *ref)
{
	/* If the client sends us a stateid from the future, it's buggy: */
	if (in->si_generation > ref->si_generation)
		return nfserr_bad_stateid;
	/*
	 * The following, however, can happen.  For example, if the
	 * client sends an open and some IO at the same time, the open
	 * may bump si_generation while the IO is still in flight.
	 * Thanks to hard links and renames, the client never knows what
	 * file an open will affect.  So it could avoid that situation
	 * only by serializing all opens and IO from the same open
	 * owner.  To recover from the old_stateid error, the client
	 * will just have to retry the IO:
	 */
	if (in->si_generation < ref->si_generation)
		return nfserr_old_stateid;
	return nfs_ok;
}

2386 2387 2388 2389 2390
static int is_delegation_stateid(stateid_t *stateid)
{
	return stateid->si_fileid == 0;
}

L
Linus Torvalds 已提交
2391 2392 2393
/*
* Checks for stateid operations
*/
2394
__be32
L
Linus Torvalds 已提交
2395 2396 2397 2398 2399
nfs4_preprocess_stateid_op(struct svc_fh *current_fh, stateid_t *stateid, int flags, struct file **filpp)
{
	struct nfs4_stateid *stp = NULL;
	struct nfs4_delegation *dp = NULL;
	struct inode *ino = current_fh->fh_dentry->d_inode;
2400
	__be32 status;
L
Linus Torvalds 已提交
2401 2402 2403 2404

	if (filpp)
		*filpp = NULL;

2405
	if (grace_disallows_io(ino))
L
Linus Torvalds 已提交
2406 2407 2408 2409 2410 2411 2412 2413 2414 2415
		return nfserr_grace;

	if (ZERO_STATEID(stateid) || ONE_STATEID(stateid))
		return check_special_stateids(current_fh, stateid, flags);

	status = nfserr_stale_stateid;
	if (STALE_STATEID(stateid)) 
		goto out;

	status = nfserr_bad_stateid;
2416
	if (is_delegation_stateid(stateid)) {
2417
		dp = find_delegation_stateid(ino, stateid);
J
J. Bruce Fields 已提交
2418
		if (!dp)
L
Linus Torvalds 已提交
2419
			goto out;
2420
		status = check_stateid_generation(stateid, &dp->dl_stateid);
2421 2422
		if (status)
			goto out;
2423 2424 2425 2426 2427 2428
		status = nfs4_check_delegmode(dp, flags);
		if (status)
			goto out;
		renew_client(dp->dl_client);
		if (filpp)
			*filpp = dp->dl_vfs_file;
L
Linus Torvalds 已提交
2429
	} else { /* open or lock stateid */
2430
		stp = find_stateid(stateid, flags);
J
J. Bruce Fields 已提交
2431
		if (!stp)
L
Linus Torvalds 已提交
2432
			goto out;
2433
		if (nfs4_check_fh(current_fh, stp))
L
Linus Torvalds 已提交
2434 2435 2436
			goto out;
		if (!stp->st_stateowner->so_confirmed)
			goto out;
2437
		status = check_stateid_generation(stateid, &stp->st_stateid);
2438 2439
		if (status)
			goto out;
2440 2441
		status = nfs4_check_openmode(stp, flags);
		if (status)
L
Linus Torvalds 已提交
2442 2443 2444 2445 2446 2447 2448 2449 2450 2451
			goto out;
		renew_client(stp->st_stateowner->so_client);
		if (filpp)
			*filpp = stp->st_vfs_file;
	}
	status = nfs_ok;
out:
	return status;
}

2452 2453 2454 2455 2456 2457
static inline int
setlkflg (int type)
{
	return (type == NFS4_READW_LT || type == NFS4_READ_LT) ?
		RD_STATE : WR_STATE;
}
L
Linus Torvalds 已提交
2458 2459 2460 2461

/* 
 * Checks for sequence id mutating operations. 
 */
2462
static __be32
2463
nfs4_preprocess_seqid_op(struct svc_fh *current_fh, u32 seqid, stateid_t *stateid, int flags, struct nfs4_stateowner **sopp, struct nfs4_stateid **stpp, struct nfsd4_lock *lock)
L
Linus Torvalds 已提交
2464 2465 2466
{
	struct nfs4_stateid *stp;
	struct nfs4_stateowner *sop;
2467
	__be32 status;
L
Linus Torvalds 已提交
2468 2469 2470 2471 2472

	dprintk("NFSD: preprocess_seqid_op: seqid=%d " 
			"stateid = (%08x/%08x/%08x/%08x)\n", seqid,
		stateid->si_boot, stateid->si_stateownerid, stateid->si_fileid,
		stateid->si_generation);
2473

L
Linus Torvalds 已提交
2474 2475 2476 2477
	*stpp = NULL;
	*sopp = NULL;

	if (ZERO_STATEID(stateid) || ONE_STATEID(stateid)) {
2478
		dprintk("NFSD: preprocess_seqid_op: magic stateid!\n");
2479
		return nfserr_bad_stateid;
L
Linus Torvalds 已提交
2480 2481 2482
	}

	if (STALE_STATEID(stateid))
2483
		return nfserr_stale_stateid;
L
Linus Torvalds 已提交
2484 2485 2486 2487 2488
	/*
	* We return BAD_STATEID if filehandle doesn't match stateid, 
	* the confirmed flag is incorrecly set, or the generation 
	* number is incorrect.  
	*/
2489 2490 2491 2492 2493 2494 2495 2496 2497 2498 2499 2500
	stp = find_stateid(stateid, flags);
	if (stp == NULL) {
		/*
		 * Also, we should make sure this isn't just the result of
		 * a replayed close:
		 */
		sop = search_close_lru(stateid->si_stateownerid, flags);
		if (sop == NULL)
			return nfserr_bad_stateid;
		*sopp = sop;
		goto check_replay;
	}
L
Linus Torvalds 已提交
2501

2502 2503 2504
	*stpp = stp;
	*sopp = sop = stp->st_stateowner;

2505 2506
	if (lock) {
		clientid_t *lockclid = &lock->v.new.clientid;
L
Linus Torvalds 已提交
2507
		struct nfs4_client *clp = sop->so_client;
2508
		int lkflg = 0;
2509
		__be32 status;
2510 2511 2512 2513

		lkflg = setlkflg(lock->lk_type);

		if (lock->lk_is_new) {
2514 2515 2516
			if (!sop->so_is_open_owner)
				return nfserr_bad_stateid;
			if (!same_clid(&clp->cl_clientid, lockclid))
2517
			       return nfserr_bad_stateid;
2518 2519 2520 2521 2522 2523 2524 2525 2526
			/* stp is the open stateid */
			status = nfs4_check_openmode(stp, lkflg);
			if (status)
				return status;
		} else {
			/* stp is the lock stateid */
			status = nfs4_check_openmode(stp->st_openstp, lkflg);
			if (status)
				return status;
2527
               }
L
Linus Torvalds 已提交
2528 2529
	}

2530
	if (nfs4_check_fh(current_fh, stp)) {
2531
		dprintk("NFSD: preprocess_seqid_op: fh-stateid mismatch!\n");
2532
		return nfserr_bad_stateid;
L
Linus Torvalds 已提交
2533 2534 2535 2536 2537 2538 2539
	}

	/*
	*  We now validate the seqid and stateid generation numbers.
	*  For the moment, we ignore the possibility of 
	*  generation number wraparound.
	*/
2540
	if (seqid != sop->so_seqid)
L
Linus Torvalds 已提交
2541 2542
		goto check_replay;

2543
	if (sop->so_confirmed && flags & CONFIRM) {
2544
		dprintk("NFSD: preprocess_seqid_op: expected"
2545 2546
				" unconfirmed stateowner!\n");
		return nfserr_bad_stateid;
L
Linus Torvalds 已提交
2547
	}
2548
	if (!sop->so_confirmed && !(flags & CONFIRM)) {
2549
		dprintk("NFSD: preprocess_seqid_op: stateowner not"
2550 2551
				" confirmed yet!\n");
		return nfserr_bad_stateid;
L
Linus Torvalds 已提交
2552
	}
2553 2554 2555
	status = check_stateid_generation(stateid, &stp->st_stateid);
	if (status)
		return status;
2556
	renew_client(sop->so_client);
2557
	return nfs_ok;
L
Linus Torvalds 已提交
2558 2559

check_replay:
2560
	if (seqid == sop->so_seqid - 1) {
N
Neil Brown 已提交
2561
		dprintk("NFSD: preprocess_seqid_op: retransmission?\n");
L
Linus Torvalds 已提交
2562
		/* indicate replay to calling function */
A
Al Viro 已提交
2563
		return nfserr_replay_me;
L
Linus Torvalds 已提交
2564
	}
2565
	dprintk("NFSD: preprocess_seqid_op: bad seqid (expected %d, got %d)\n",
2566 2567 2568
			sop->so_seqid, seqid);
	*sopp = NULL;
	return nfserr_bad_seqid;
L
Linus Torvalds 已提交
2569 2570
}

2571
__be32
2572
nfsd4_open_confirm(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate,
2573
		   struct nfsd4_open_confirm *oc)
L
Linus Torvalds 已提交
2574
{
2575
	__be32 status;
L
Linus Torvalds 已提交
2576 2577 2578 2579
	struct nfs4_stateowner *sop;
	struct nfs4_stateid *stp;

	dprintk("NFSD: nfsd4_open_confirm on file %.*s\n",
2580 2581
			(int)cstate->current_fh.fh_dentry->d_name.len,
			cstate->current_fh.fh_dentry->d_name.name);
L
Linus Torvalds 已提交
2582

2583
	status = fh_verify(rqstp, &cstate->current_fh, S_IFREG, 0);
2584 2585
	if (status)
		return status;
L
Linus Torvalds 已提交
2586 2587 2588

	nfs4_lock_state();

2589 2590
	if ((status = nfs4_preprocess_seqid_op(&cstate->current_fh,
					oc->oc_seqid, &oc->oc_req_stateid,
2591
					CONFIRM | OPEN_STATE,
L
Linus Torvalds 已提交
2592 2593 2594 2595 2596 2597 2598 2599 2600 2601 2602 2603 2604
					&oc->oc_stateowner, &stp, NULL)))
		goto out; 

	sop = oc->oc_stateowner;
	sop->so_confirmed = 1;
	update_stateid(&stp->st_stateid);
	memcpy(&oc->oc_resp_stateid, &stp->st_stateid, sizeof(stateid_t));
	dprintk("NFSD: nfsd4_open_confirm: success, seqid=%d " 
		"stateid=(%08x/%08x/%08x/%08x)\n", oc->oc_seqid,
		         stp->st_stateid.si_boot,
		         stp->st_stateid.si_stateownerid,
		         stp->st_stateid.si_fileid,
		         stp->st_stateid.si_generation);
2605 2606

	nfsd4_create_clid_dir(sop->so_client);
L
Linus Torvalds 已提交
2607
out:
N
Neil Brown 已提交
2608
	if (oc->oc_stateowner) {
L
Linus Torvalds 已提交
2609
		nfs4_get_stateowner(oc->oc_stateowner);
2610
		cstate->replay_owner = oc->oc_stateowner;
N
Neil Brown 已提交
2611
	}
L
Linus Torvalds 已提交
2612 2613 2614 2615 2616 2617 2618 2619 2620 2621 2622 2623 2624 2625 2626 2627 2628 2629 2630 2631 2632 2633 2634 2635 2636 2637 2638 2639 2640
	nfs4_unlock_state();
	return status;
}


/*
 * unset all bits in union bitmap (bmap) that
 * do not exist in share (from successful OPEN_DOWNGRADE)
 */
static void
reset_union_bmap_access(unsigned long access, unsigned long *bmap)
{
	int i;
	for (i = 1; i < 4; i++) {
		if ((i & access) != i)
			__clear_bit(i, bmap);
	}
}

static void
reset_union_bmap_deny(unsigned long deny, unsigned long *bmap)
{
	int i;
	for (i = 0; i < 4; i++) {
		if ((i & deny) != i)
			__clear_bit(i, bmap);
	}
}

2641
__be32
2642 2643
nfsd4_open_downgrade(struct svc_rqst *rqstp,
		     struct nfsd4_compound_state *cstate,
2644
		     struct nfsd4_open_downgrade *od)
L
Linus Torvalds 已提交
2645
{
2646
	__be32 status;
L
Linus Torvalds 已提交
2647 2648 2649 2650
	struct nfs4_stateid *stp;
	unsigned int share_access;

	dprintk("NFSD: nfsd4_open_downgrade on file %.*s\n", 
2651 2652
			(int)cstate->current_fh.fh_dentry->d_name.len,
			cstate->current_fh.fh_dentry->d_name.name);
L
Linus Torvalds 已提交
2653

2654 2655
	if (!access_valid(od->od_share_access)
			|| !deny_valid(od->od_share_deny))
L
Linus Torvalds 已提交
2656 2657 2658
		return nfserr_inval;

	nfs4_lock_state();
2659 2660
	if ((status = nfs4_preprocess_seqid_op(&cstate->current_fh,
					od->od_seqid,
L
Linus Torvalds 已提交
2661
					&od->od_stateid, 
2662
					OPEN_STATE,
L
Linus Torvalds 已提交
2663 2664 2665 2666 2667 2668 2669 2670 2671 2672 2673 2674 2675 2676 2677 2678 2679 2680 2681 2682 2683 2684 2685 2686 2687
					&od->od_stateowner, &stp, NULL)))
		goto out; 

	status = nfserr_inval;
	if (!test_bit(od->od_share_access, &stp->st_access_bmap)) {
		dprintk("NFSD:access not a subset current bitmap: 0x%lx, input access=%08x\n",
			stp->st_access_bmap, od->od_share_access);
		goto out;
	}
	if (!test_bit(od->od_share_deny, &stp->st_deny_bmap)) {
		dprintk("NFSD:deny not a subset current bitmap: 0x%lx, input deny=%08x\n",
			stp->st_deny_bmap, od->od_share_deny);
		goto out;
	}
	set_access(&share_access, stp->st_access_bmap);
	nfs4_file_downgrade(stp->st_vfs_file,
	                    share_access & ~od->od_share_access);

	reset_union_bmap_access(od->od_share_access, &stp->st_access_bmap);
	reset_union_bmap_deny(od->od_share_deny, &stp->st_deny_bmap);

	update_stateid(&stp->st_stateid);
	memcpy(&od->od_stateid, &stp->st_stateid, sizeof(stateid_t));
	status = nfs_ok;
out:
N
Neil Brown 已提交
2688
	if (od->od_stateowner) {
L
Linus Torvalds 已提交
2689
		nfs4_get_stateowner(od->od_stateowner);
2690
		cstate->replay_owner = od->od_stateowner;
N
Neil Brown 已提交
2691
	}
L
Linus Torvalds 已提交
2692 2693 2694 2695 2696 2697 2698
	nfs4_unlock_state();
	return status;
}

/*
 * nfs4_unlock_state() called after encode
 */
2699
__be32
2700
nfsd4_close(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate,
2701
	    struct nfsd4_close *close)
L
Linus Torvalds 已提交
2702
{
2703
	__be32 status;
L
Linus Torvalds 已提交
2704 2705 2706
	struct nfs4_stateid *stp;

	dprintk("NFSD: nfsd4_close on file %.*s\n", 
2707 2708
			(int)cstate->current_fh.fh_dentry->d_name.len,
			cstate->current_fh.fh_dentry->d_name.name);
L
Linus Torvalds 已提交
2709 2710 2711

	nfs4_lock_state();
	/* check close_lru for replay */
2712 2713
	if ((status = nfs4_preprocess_seqid_op(&cstate->current_fh,
					close->cl_seqid,
L
Linus Torvalds 已提交
2714
					&close->cl_stateid, 
2715
					OPEN_STATE | CLOSE_STATE,
L
Linus Torvalds 已提交
2716 2717 2718 2719 2720 2721
					&close->cl_stateowner, &stp, NULL)))
		goto out; 
	status = nfs_ok;
	update_stateid(&stp->st_stateid);
	memcpy(&close->cl_stateid, &stp->st_stateid, sizeof(stateid_t));

2722
	/* release_stateid() calls nfsd_close() if needed */
2723
	release_open_stateid(stp);
2724 2725 2726 2727 2728 2729 2730

	/* place unused nfs4_stateowners on so_close_lru list to be
	 * released by the laundromat service after the lease period
	 * to enable us to handle CLOSE replay
	 */
	if (list_empty(&close->cl_stateowner->so_stateids))
		move_to_close_lru(close->cl_stateowner);
L
Linus Torvalds 已提交
2731
out:
N
Neil Brown 已提交
2732
	if (close->cl_stateowner) {
L
Linus Torvalds 已提交
2733
		nfs4_get_stateowner(close->cl_stateowner);
2734
		cstate->replay_owner = close->cl_stateowner;
N
Neil Brown 已提交
2735
	}
L
Linus Torvalds 已提交
2736 2737 2738 2739
	nfs4_unlock_state();
	return status;
}

2740
__be32
2741 2742
nfsd4_delegreturn(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate,
		  struct nfsd4_delegreturn *dr)
L
Linus Torvalds 已提交
2743
{
2744 2745 2746
	struct nfs4_delegation *dp;
	stateid_t *stateid = &dr->dr_stateid;
	struct inode *inode;
2747
	__be32 status;
L
Linus Torvalds 已提交
2748

2749
	if ((status = fh_verify(rqstp, &cstate->current_fh, S_IFREG, 0)))
2750 2751
		return status;
	inode = cstate->current_fh.fh_dentry->d_inode;
L
Linus Torvalds 已提交
2752 2753

	nfs4_lock_state();
2754 2755 2756 2757 2758 2759
	status = nfserr_bad_stateid;
	if (ZERO_STATEID(stateid) || ONE_STATEID(stateid))
		goto out;
	status = nfserr_stale_stateid;
	if (STALE_STATEID(stateid))
		goto out;
2760
	status = nfserr_bad_stateid;
2761 2762 2763 2764 2765 2766 2767 2768 2769 2770 2771
	if (!is_delegation_stateid(stateid))
		goto out;
	dp = find_delegation_stateid(inode, stateid);
	if (!dp)
		goto out;
	status = check_stateid_generation(stateid, &dp->dl_stateid);
	if (status)
		goto out;
	renew_client(dp->dl_client);

	unhash_delegation(dp);
L
Linus Torvalds 已提交
2772
out:
2773 2774
	nfs4_unlock_state();

L
Linus Torvalds 已提交
2775 2776 2777 2778 2779 2780 2781 2782 2783 2784 2785 2786
	return status;
}


/* 
 * Lock owner state (byte-range locks)
 */
#define LOFF_OVERFLOW(start, len)      ((u64)(len) > ~(u64)(start))
#define LOCK_HASH_BITS              8
#define LOCK_HASH_SIZE             (1 << LOCK_HASH_BITS)
#define LOCK_HASH_MASK             (LOCK_HASH_SIZE - 1)

B
Benny Halevy 已提交
2787 2788 2789 2790 2791 2792 2793 2794 2795 2796 2797 2798 2799 2800 2801 2802 2803 2804 2805 2806
static inline u64
end_offset(u64 start, u64 len)
{
	u64 end;

	end = start + len;
	return end >= start ? end: NFS4_MAX_UINT64;
}

/* last octet in a range */
static inline u64
last_byte_offset(u64 start, u64 len)
{
	u64 end;

	BUG_ON(!len);
	end = start + len;
	return end > start ? end - 1: NFS4_MAX_UINT64;
}

L
Linus Torvalds 已提交
2807 2808 2809 2810 2811 2812 2813 2814 2815 2816 2817 2818 2819 2820 2821 2822
#define lockownerid_hashval(id) \
        ((id) & LOCK_HASH_MASK)

static inline unsigned int
lock_ownerstr_hashval(struct inode *inode, u32 cl_id,
		struct xdr_netobj *ownername)
{
	return (file_hashval(inode) + cl_id
			+ opaque_hashval(ownername->data, ownername->len))
		& LOCK_HASH_MASK;
}

static struct list_head lock_ownerid_hashtbl[LOCK_HASH_SIZE];
static struct list_head	lock_ownerstr_hashtbl[LOCK_HASH_SIZE];
static struct list_head lockstateid_hashtbl[STATEID_HASH_SIZE];

2823
static struct nfs4_stateid *
L
Linus Torvalds 已提交
2824 2825
find_stateid(stateid_t *stid, int flags)
{
2826
	struct nfs4_stateid *local;
L
Linus Torvalds 已提交
2827 2828 2829 2830 2831
	u32 st_id = stid->si_stateownerid;
	u32 f_id = stid->si_fileid;
	unsigned int hashval;

	dprintk("NFSD: find_stateid flags 0x%x\n",flags);
2832
	if (flags & (LOCK_STATE | RD_STATE | WR_STATE)) {
L
Linus Torvalds 已提交
2833 2834 2835 2836 2837 2838 2839
		hashval = stateid_hashval(st_id, f_id);
		list_for_each_entry(local, &lockstateid_hashtbl[hashval], st_hash) {
			if ((local->st_stateid.si_stateownerid == st_id) &&
			    (local->st_stateid.si_fileid == f_id))
				return local;
		}
	} 
2840 2841

	if (flags & (OPEN_STATE | RD_STATE | WR_STATE)) {
L
Linus Torvalds 已提交
2842 2843 2844 2845 2846 2847
		hashval = stateid_hashval(st_id, f_id);
		list_for_each_entry(local, &stateid_hashtbl[hashval], st_hash) {
			if ((local->st_stateid.si_stateownerid == st_id) &&
			    (local->st_stateid.si_fileid == f_id))
				return local;
		}
N
Neil Brown 已提交
2848
	}
L
Linus Torvalds 已提交
2849 2850 2851 2852 2853 2854
	return NULL;
}

static struct nfs4_delegation *
find_delegation_stateid(struct inode *ino, stateid_t *stid)
{
2855 2856
	struct nfs4_file *fp;
	struct nfs4_delegation *dl;
L
Linus Torvalds 已提交
2857 2858 2859 2860 2861 2862

	dprintk("NFSD:find_delegation_stateid stateid=(%08x/%08x/%08x/%08x)\n",
                    stid->si_boot, stid->si_stateownerid,
                    stid->si_fileid, stid->si_generation);

	fp = find_file(ino);
2863 2864 2865 2866 2867
	if (!fp)
		return NULL;
	dl = find_delegation_file(fp, stid);
	put_nfs4_file(fp);
	return dl;
L
Linus Torvalds 已提交
2868 2869 2870 2871 2872 2873 2874 2875 2876 2877 2878 2879 2880 2881 2882 2883 2884 2885 2886
}

/*
 * TODO: Linux file offsets are _signed_ 64-bit quantities, which means that
 * we can't properly handle lock requests that go beyond the (2^63 - 1)-th
 * byte, because of sign extension problems.  Since NFSv4 calls for 64-bit
 * locking, this prevents us from being completely protocol-compliant.  The
 * real solution to this problem is to start using unsigned file offsets in
 * the VFS, but this is a very deep change!
 */
static inline void
nfs4_transform_lock_offset(struct file_lock *lock)
{
	if (lock->fl_start < 0)
		lock->fl_start = OFFSET_MAX;
	if (lock->fl_end < 0)
		lock->fl_end = OFFSET_MAX;
}

2887 2888
/* Hack!: For now, we're defining this just so we can use a pointer to it
 * as a unique cookie to identify our (NFSv4's) posix locks. */
2889
static struct lock_manager_operations nfsd_posix_mng_ops  = {
2890
};
L
Linus Torvalds 已提交
2891 2892 2893 2894

static inline void
nfs4_set_lock_denied(struct file_lock *fl, struct nfsd4_lock_denied *deny)
{
2895 2896
	struct nfs4_stateowner *sop;
	unsigned int hval;
L
Linus Torvalds 已提交
2897

2898 2899 2900
	if (fl->fl_lmops == &nfsd_posix_mng_ops) {
		sop = (struct nfs4_stateowner *) fl->fl_owner;
		hval = lockownerid_hashval(sop->so_id);
L
Linus Torvalds 已提交
2901 2902 2903
		kref_get(&sop->so_ref);
		deny->ld_sop = sop;
		deny->ld_clientid = sop->so_client->cl_clientid;
2904 2905 2906 2907
	} else {
		deny->ld_sop = NULL;
		deny->ld_clientid.cl_boot = 0;
		deny->ld_clientid.cl_id = 0;
L
Linus Torvalds 已提交
2908 2909
	}
	deny->ld_start = fl->fl_start;
B
Benny Halevy 已提交
2910 2911
	deny->ld_length = NFS4_MAX_UINT64;
	if (fl->fl_end != NFS4_MAX_UINT64)
L
Linus Torvalds 已提交
2912 2913 2914 2915 2916 2917 2918 2919 2920 2921 2922 2923 2924 2925
		deny->ld_length = fl->fl_end - fl->fl_start + 1;        
	deny->ld_type = NFS4_READ_LT;
	if (fl->fl_type != F_RDLCK)
		deny->ld_type = NFS4_WRITE_LT;
}

static struct nfs4_stateowner *
find_lockstateowner_str(struct inode *inode, clientid_t *clid,
		struct xdr_netobj *owner)
{
	unsigned int hashval = lock_ownerstr_hashval(inode, clid->cl_id, owner);
	struct nfs4_stateowner *op;

	list_for_each_entry(op, &lock_ownerstr_hashtbl[hashval], so_strhash) {
2926
		if (same_owner_str(op, owner, clid))
L
Linus Torvalds 已提交
2927 2928 2929 2930 2931 2932 2933 2934 2935 2936 2937 2938 2939 2940 2941 2942 2943 2944 2945 2946 2947 2948 2949 2950 2951
			return op;
	}
	return NULL;
}

/*
 * Alloc a lock owner structure.
 * Called in nfsd4_lock - therefore, OPEN and OPEN_CONFIRM (if needed) has 
 * occured. 
 *
 * strhashval = lock_ownerstr_hashval 
 */

static struct nfs4_stateowner *
alloc_init_lock_stateowner(unsigned int strhashval, struct nfs4_client *clp, struct nfs4_stateid *open_stp, struct nfsd4_lock *lock) {
	struct nfs4_stateowner *sop;
	struct nfs4_replay *rp;
	unsigned int idhashval;

	if (!(sop = alloc_stateowner(&lock->lk_new_owner)))
		return NULL;
	idhashval = lockownerid_hashval(current_ownerid);
	INIT_LIST_HEAD(&sop->so_idhash);
	INIT_LIST_HEAD(&sop->so_strhash);
	INIT_LIST_HEAD(&sop->so_perclient);
2952 2953
	INIT_LIST_HEAD(&sop->so_stateids);
	INIT_LIST_HEAD(&sop->so_perstateid);
L
Linus Torvalds 已提交
2954 2955 2956 2957
	INIT_LIST_HEAD(&sop->so_close_lru); /* not used */
	sop->so_time = 0;
	list_add(&sop->so_idhash, &lock_ownerid_hashtbl[idhashval]);
	list_add(&sop->so_strhash, &lock_ownerstr_hashtbl[strhashval]);
2958
	list_add(&sop->so_perstateid, &open_stp->st_lockowners);
L
Linus Torvalds 已提交
2959 2960 2961
	sop->so_is_open_owner = 0;
	sop->so_id = current_ownerid++;
	sop->so_client = clp;
2962 2963 2964
	/* It is the openowner seqid that will be incremented in encode in the
	 * case of new lockowners; so increment the lock seqid manually: */
	sop->so_seqid = lock->lk_new_lock_seqid + 1;
L
Linus Torvalds 已提交
2965 2966
	sop->so_confirmed = 1;
	rp = &sop->so_replay;
2967
	rp->rp_status = nfserr_serverfault;
L
Linus Torvalds 已提交
2968 2969 2970 2971 2972
	rp->rp_buflen = 0;
	rp->rp_buf = rp->rp_ibuf;
	return sop;
}

2973
static struct nfs4_stateid *
L
Linus Torvalds 已提交
2974 2975 2976 2977 2978
alloc_init_lock_stateid(struct nfs4_stateowner *sop, struct nfs4_file *fp, struct nfs4_stateid *open_stp)
{
	struct nfs4_stateid *stp;
	unsigned int hashval = stateid_hashval(sop->so_id, fp->fi_id);

N
NeilBrown 已提交
2979 2980
	stp = nfs4_alloc_stateid();
	if (stp == NULL)
L
Linus Torvalds 已提交
2981 2982 2983
		goto out;
	INIT_LIST_HEAD(&stp->st_hash);
	INIT_LIST_HEAD(&stp->st_perfile);
2984 2985
	INIT_LIST_HEAD(&stp->st_perstateowner);
	INIT_LIST_HEAD(&stp->st_lockowners); /* not used */
L
Linus Torvalds 已提交
2986
	list_add(&stp->st_hash, &lockstateid_hashtbl[hashval]);
2987
	list_add(&stp->st_perfile, &fp->fi_stateids);
2988
	list_add(&stp->st_perstateowner, &sop->so_stateids);
L
Linus Torvalds 已提交
2989
	stp->st_stateowner = sop;
2990
	get_nfs4_file(fp);
L
Linus Torvalds 已提交
2991 2992 2993 2994 2995 2996 2997 2998
	stp->st_file = fp;
	stp->st_stateid.si_boot = boot_time;
	stp->st_stateid.si_stateownerid = sop->so_id;
	stp->st_stateid.si_fileid = fp->fi_id;
	stp->st_stateid.si_generation = 0;
	stp->st_vfs_file = open_stp->st_vfs_file; /* FIXME refcount?? */
	stp->st_access_bmap = open_stp->st_access_bmap;
	stp->st_deny_bmap = open_stp->st_deny_bmap;
2999
	stp->st_openstp = open_stp;
L
Linus Torvalds 已提交
3000 3001 3002 3003 3004

out:
	return stp;
}

3005
static int
L
Linus Torvalds 已提交
3006 3007
check_lock_length(u64 offset, u64 length)
{
B
Benny Halevy 已提交
3008
	return ((length == 0)  || ((length != NFS4_MAX_UINT64) &&
L
Linus Torvalds 已提交
3009 3010 3011 3012 3013 3014
	     LOFF_OVERFLOW(offset, length)));
}

/*
 *  LOCK operation 
 */
3015
__be32
3016
nfsd4_lock(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate,
3017
	   struct nfsd4_lock *lock)
L
Linus Torvalds 已提交
3018
{
3019
	struct nfs4_stateowner *open_sop = NULL;
3020
	struct nfs4_stateowner *lock_sop = NULL;
L
Linus Torvalds 已提交
3021 3022 3023
	struct nfs4_stateid *lock_stp;
	struct file *filp;
	struct file_lock file_lock;
3024
	struct file_lock conflock;
3025
	__be32 status = 0;
L
Linus Torvalds 已提交
3026
	unsigned int strhashval;
3027
	unsigned int cmd;
3028
	int err;
L
Linus Torvalds 已提交
3029 3030 3031 3032 3033 3034 3035 3036

	dprintk("NFSD: nfsd4_lock: start=%Ld length=%Ld\n",
		(long long) lock->lk_offset,
		(long long) lock->lk_length);

	if (check_lock_length(lock->lk_offset, lock->lk_length))
		 return nfserr_inval;

3037
	if ((status = fh_verify(rqstp, &cstate->current_fh,
M
Miklos Szeredi 已提交
3038
				S_IFREG, NFSD_MAY_LOCK))) {
A
Andy Adamson 已提交
3039 3040 3041 3042
		dprintk("NFSD: nfsd4_lock: permission denied!\n");
		return status;
	}

L
Linus Torvalds 已提交
3043 3044 3045
	nfs4_lock_state();

	if (lock->lk_is_new) {
N
NeilBrown 已提交
3046 3047 3048 3049 3050
		/*
		 * Client indicates that this is a new lockowner.
		 * Use open owner and open stateid to create lock owner and
		 * lock stateid.
		 */
L
Linus Torvalds 已提交
3051 3052 3053 3054
		struct nfs4_stateid *open_stp = NULL;
		struct nfs4_file *fp;
		
		status = nfserr_stale_clientid;
N
Neil Brown 已提交
3055
		if (STALE_CLIENTID(&lock->lk_new_clientid))
L
Linus Torvalds 已提交
3056 3057 3058
			goto out;

		/* validate and update open stateid and open seqid */
3059
		status = nfs4_preprocess_seqid_op(&cstate->current_fh,
L
Linus Torvalds 已提交
3060 3061
				        lock->lk_new_open_seqid,
		                        &lock->lk_new_open_stateid,
3062
					OPEN_STATE,
3063
		                        &lock->lk_replay_owner, &open_stp,
3064
					lock);
3065
		if (status)
L
Linus Torvalds 已提交
3066
			goto out;
3067
		open_sop = lock->lk_replay_owner;
L
Linus Torvalds 已提交
3068 3069 3070 3071 3072
		/* create lockowner and lock stateid */
		fp = open_stp->st_file;
		strhashval = lock_ownerstr_hashval(fp->fi_inode, 
				open_sop->so_client->cl_clientid.cl_id, 
				&lock->v.new.owner);
3073 3074 3075
		/* XXX: Do we need to check for duplicate stateowners on
		 * the same file, or should they just be allowed (and
		 * create new stateids)? */
L
Linus Torvalds 已提交
3076
		status = nfserr_resource;
3077 3078 3079
		lock_sop = alloc_init_lock_stateowner(strhashval,
				open_sop->so_client, open_stp, lock);
		if (lock_sop == NULL)
L
Linus Torvalds 已提交
3080
			goto out;
3081
		lock_stp = alloc_init_lock_stateid(lock_sop, fp, open_stp);
3082
		if (lock_stp == NULL)
L
Linus Torvalds 已提交
3083 3084 3085
			goto out;
	} else {
		/* lock (lock owner + lock stateid) already exists */
3086
		status = nfs4_preprocess_seqid_op(&cstate->current_fh,
L
Linus Torvalds 已提交
3087 3088
				       lock->lk_old_lock_seqid, 
				       &lock->lk_old_lock_stateid, 
3089
				       LOCK_STATE,
3090
				       &lock->lk_replay_owner, &lock_stp, lock);
L
Linus Torvalds 已提交
3091 3092
		if (status)
			goto out;
3093
		lock_sop = lock->lk_replay_owner;
L
Linus Torvalds 已提交
3094
	}
3095
	/* lock->lk_replay_owner and lock_stp have been created or found */
L
Linus Torvalds 已提交
3096 3097
	filp = lock_stp->st_vfs_file;

3098
	status = nfserr_grace;
3099
	if (locks_in_grace() && !lock->lk_reclaim)
3100 3101
		goto out;
	status = nfserr_no_grace;
3102
	if (!locks_in_grace() && lock->lk_reclaim)
3103 3104
		goto out;

L
Linus Torvalds 已提交
3105 3106 3107 3108 3109
	locks_init_lock(&file_lock);
	switch (lock->lk_type) {
		case NFS4_READ_LT:
		case NFS4_READW_LT:
			file_lock.fl_type = F_RDLCK;
3110
			cmd = F_SETLK;
L
Linus Torvalds 已提交
3111 3112 3113 3114
		break;
		case NFS4_WRITE_LT:
		case NFS4_WRITEW_LT:
			file_lock.fl_type = F_WRLCK;
3115
			cmd = F_SETLK;
L
Linus Torvalds 已提交
3116 3117 3118 3119 3120
		break;
		default:
			status = nfserr_inval;
		goto out;
	}
3121
	file_lock.fl_owner = (fl_owner_t)lock_sop;
L
Linus Torvalds 已提交
3122 3123 3124
	file_lock.fl_pid = current->tgid;
	file_lock.fl_file = filp;
	file_lock.fl_flags = FL_POSIX;
3125
	file_lock.fl_lmops = &nfsd_posix_mng_ops;
L
Linus Torvalds 已提交
3126 3127

	file_lock.fl_start = lock->lk_offset;
B
Benny Halevy 已提交
3128
	file_lock.fl_end = last_byte_offset(lock->lk_offset, lock->lk_length);
L
Linus Torvalds 已提交
3129 3130 3131 3132 3133 3134 3135
	nfs4_transform_lock_offset(&file_lock);

	/*
	* Try to lock the file in the VFS.
	* Note: locks.c uses the BKL to protect the inode's lock list.
	*/

3136
	err = vfs_lock_file(filp, cmd, &file_lock, &conflock);
3137
	switch (-err) {
L
Linus Torvalds 已提交
3138 3139 3140 3141
	case 0: /* success! */
		update_stateid(&lock_stp->st_stateid);
		memcpy(&lock->lk_resp_stateid, &lock_stp->st_stateid, 
				sizeof(stateid_t));
3142
		status = 0;
3143 3144 3145 3146 3147 3148
		break;
	case (EAGAIN):		/* conflock holds conflicting lock */
		status = nfserr_denied;
		dprintk("NFSD: nfsd4_lock: conflicting lock found!\n");
		nfs4_set_lock_denied(&conflock, &lock->lk_denied);
		break;
L
Linus Torvalds 已提交
3149 3150
	case (EDEADLK):
		status = nfserr_deadlock;
3151
		break;
L
Linus Torvalds 已提交
3152
	default:        
3153
		dprintk("NFSD: nfsd4_lock: vfs_lock_file() failed! status %d\n",err);
3154 3155
		status = nfserr_resource;
		break;
L
Linus Torvalds 已提交
3156 3157
	}
out:
3158
	if (status && lock->lk_is_new && lock_sop)
3159
		release_lockowner(lock_sop);
3160 3161
	if (lock->lk_replay_owner) {
		nfs4_get_stateowner(lock->lk_replay_owner);
3162
		cstate->replay_owner = lock->lk_replay_owner;
N
Neil Brown 已提交
3163
	}
L
Linus Torvalds 已提交
3164 3165 3166 3167
	nfs4_unlock_state();
	return status;
}

3168 3169 3170 3171 3172 3173 3174 3175 3176 3177 3178 3179 3180 3181 3182 3183 3184 3185 3186
/*
 * The NFSv4 spec allows a client to do a LOCKT without holding an OPEN,
 * so we do a temporary open here just to get an open file to pass to
 * vfs_test_lock.  (Arguably perhaps test_lock should be done with an
 * inode operation.)
 */
static int nfsd_test_lock(struct svc_rqst *rqstp, struct svc_fh *fhp, struct file_lock *lock)
{
	struct file *file;
	int err;

	err = nfsd_open(rqstp, fhp, S_IFREG, NFSD_MAY_READ, &file);
	if (err)
		return err;
	err = vfs_test_lock(file, lock);
	nfsd_close(file);
	return err;
}

L
Linus Torvalds 已提交
3187 3188 3189
/*
 * LOCKT operation
 */
3190
__be32
3191 3192
nfsd4_lockt(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate,
	    struct nfsd4_lockt *lockt)
L
Linus Torvalds 已提交
3193 3194 3195
{
	struct inode *inode;
	struct file_lock file_lock;
3196
	int error;
3197
	__be32 status;
L
Linus Torvalds 已提交
3198

3199
	if (locks_in_grace())
L
Linus Torvalds 已提交
3200 3201 3202 3203 3204 3205 3206 3207 3208
		return nfserr_grace;

	if (check_lock_length(lockt->lt_offset, lockt->lt_length))
		 return nfserr_inval;

	lockt->lt_stateowner = NULL;
	nfs4_lock_state();

	status = nfserr_stale_clientid;
N
Neil Brown 已提交
3209
	if (STALE_CLIENTID(&lockt->lt_clientid))
L
Linus Torvalds 已提交
3210 3211
		goto out;

3212
	if ((status = fh_verify(rqstp, &cstate->current_fh, S_IFREG, 0))) {
N
Neil Brown 已提交
3213
		dprintk("NFSD: nfsd4_lockt: fh_verify() failed!\n");
L
Linus Torvalds 已提交
3214 3215 3216 3217 3218
		if (status == nfserr_symlink)
			status = nfserr_inval;
		goto out;
	}

3219
	inode = cstate->current_fh.fh_dentry->d_inode;
L
Linus Torvalds 已提交
3220 3221 3222 3223 3224 3225 3226 3227 3228 3229 3230
	locks_init_lock(&file_lock);
	switch (lockt->lt_type) {
		case NFS4_READ_LT:
		case NFS4_READW_LT:
			file_lock.fl_type = F_RDLCK;
		break;
		case NFS4_WRITE_LT:
		case NFS4_WRITEW_LT:
			file_lock.fl_type = F_WRLCK;
		break;
		default:
3231
			dprintk("NFSD: nfs4_lockt: bad lock type!\n");
L
Linus Torvalds 已提交
3232 3233 3234 3235 3236 3237 3238 3239 3240 3241 3242 3243
			status = nfserr_inval;
		goto out;
	}

	lockt->lt_stateowner = find_lockstateowner_str(inode,
			&lockt->lt_clientid, &lockt->lt_owner);
	if (lockt->lt_stateowner)
		file_lock.fl_owner = (fl_owner_t)lockt->lt_stateowner;
	file_lock.fl_pid = current->tgid;
	file_lock.fl_flags = FL_POSIX;

	file_lock.fl_start = lockt->lt_offset;
B
Benny Halevy 已提交
3244
	file_lock.fl_end = last_byte_offset(lockt->lt_offset, lockt->lt_length);
L
Linus Torvalds 已提交
3245 3246 3247 3248

	nfs4_transform_lock_offset(&file_lock);

	status = nfs_ok;
3249
	error = nfsd_test_lock(rqstp, &cstate->current_fh, &file_lock);
3250 3251 3252 3253
	if (error) {
		status = nfserrno(error);
		goto out;
	}
3254
	if (file_lock.fl_type != F_UNLCK) {
L
Linus Torvalds 已提交
3255
		status = nfserr_denied;
3256
		nfs4_set_lock_denied(&file_lock, &lockt->lt_denied);
L
Linus Torvalds 已提交
3257 3258 3259 3260 3261 3262
	}
out:
	nfs4_unlock_state();
	return status;
}

3263
__be32
3264
nfsd4_locku(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate,
3265
	    struct nfsd4_locku *locku)
L
Linus Torvalds 已提交
3266 3267 3268 3269
{
	struct nfs4_stateid *stp;
	struct file *filp = NULL;
	struct file_lock file_lock;
3270
	__be32 status;
3271
	int err;
L
Linus Torvalds 已提交
3272 3273 3274 3275 3276 3277 3278 3279 3280 3281
						        
	dprintk("NFSD: nfsd4_locku: start=%Ld length=%Ld\n",
		(long long) locku->lu_offset,
		(long long) locku->lu_length);

	if (check_lock_length(locku->lu_offset, locku->lu_length))
		 return nfserr_inval;

	nfs4_lock_state();
									        
3282
	if ((status = nfs4_preprocess_seqid_op(&cstate->current_fh,
L
Linus Torvalds 已提交
3283 3284
					locku->lu_seqid, 
					&locku->lu_stateid, 
3285
					LOCK_STATE,
L
Linus Torvalds 已提交
3286 3287 3288 3289 3290 3291 3292 3293 3294 3295 3296
					&locku->lu_stateowner, &stp, NULL)))
		goto out;

	filp = stp->st_vfs_file;
	BUG_ON(!filp);
	locks_init_lock(&file_lock);
	file_lock.fl_type = F_UNLCK;
	file_lock.fl_owner = (fl_owner_t) locku->lu_stateowner;
	file_lock.fl_pid = current->tgid;
	file_lock.fl_file = filp;
	file_lock.fl_flags = FL_POSIX; 
3297
	file_lock.fl_lmops = &nfsd_posix_mng_ops;
L
Linus Torvalds 已提交
3298 3299
	file_lock.fl_start = locku->lu_offset;

B
Benny Halevy 已提交
3300
	file_lock.fl_end = last_byte_offset(locku->lu_offset, locku->lu_length);
L
Linus Torvalds 已提交
3301 3302 3303 3304 3305
	nfs4_transform_lock_offset(&file_lock);

	/*
	*  Try to unlock the file in the VFS.
	*/
3306
	err = vfs_lock_file(filp, F_SETLK, &file_lock, NULL);
3307
	if (err) {
3308
		dprintk("NFSD: nfs4_locku: vfs_lock_file failed!\n");
L
Linus Torvalds 已提交
3309 3310 3311 3312 3313 3314 3315 3316 3317
		goto out_nfserr;
	}
	/*
	* OK, unlock succeeded; the only thing left to do is update the stateid.
	*/
	update_stateid(&stp->st_stateid);
	memcpy(&locku->lu_stateid, &stp->st_stateid, sizeof(stateid_t));

out:
N
Neil Brown 已提交
3318
	if (locku->lu_stateowner) {
L
Linus Torvalds 已提交
3319
		nfs4_get_stateowner(locku->lu_stateowner);
3320
		cstate->replay_owner = locku->lu_stateowner;
N
Neil Brown 已提交
3321
	}
L
Linus Torvalds 已提交
3322 3323 3324 3325
	nfs4_unlock_state();
	return status;

out_nfserr:
3326
	status = nfserrno(err);
L
Linus Torvalds 已提交
3327 3328 3329 3330 3331 3332 3333 3334 3335 3336 3337 3338
	goto out;
}

/*
 * returns
 * 	1: locks held by lockowner
 * 	0: no locks held by lockowner
 */
static int
check_for_locks(struct file *filp, struct nfs4_stateowner *lowner)
{
	struct file_lock **flpp;
3339
	struct inode *inode = filp->f_path.dentry->d_inode;
L
Linus Torvalds 已提交
3340 3341 3342 3343
	int status = 0;

	lock_kernel();
	for (flpp = &inode->i_flock; *flpp != NULL; flpp = &(*flpp)->fl_next) {
3344
		if ((*flpp)->fl_owner == (fl_owner_t)lowner) {
L
Linus Torvalds 已提交
3345 3346
			status = 1;
			goto out;
3347
		}
L
Linus Torvalds 已提交
3348 3349 3350 3351 3352 3353
	}
out:
	unlock_kernel();
	return status;
}

3354
__be32
3355 3356 3357
nfsd4_release_lockowner(struct svc_rqst *rqstp,
			struct nfsd4_compound_state *cstate,
			struct nfsd4_release_lockowner *rlockowner)
L
Linus Torvalds 已提交
3358 3359
{
	clientid_t *clid = &rlockowner->rl_clientid;
3360 3361
	struct nfs4_stateowner *sop;
	struct nfs4_stateid *stp;
L
Linus Torvalds 已提交
3362
	struct xdr_netobj *owner = &rlockowner->rl_owner;
3363 3364
	struct list_head matches;
	int i;
3365
	__be32 status;
L
Linus Torvalds 已提交
3366 3367 3368 3369 3370 3371 3372

	dprintk("nfsd4_release_lockowner clientid: (%08x/%08x):\n",
		clid->cl_boot, clid->cl_id);

	/* XXX check for lease expiration */

	status = nfserr_stale_clientid;
N
Neil Brown 已提交
3373
	if (STALE_CLIENTID(clid))
L
Linus Torvalds 已提交
3374 3375 3376 3377
		return status;

	nfs4_lock_state();

3378 3379 3380 3381 3382 3383 3384 3385
	status = nfserr_locks_held;
	/* XXX: we're doing a linear search through all the lockowners.
	 * Yipes!  For now we'll just hope clients aren't really using
	 * release_lockowner much, but eventually we have to fix these
	 * data structures. */
	INIT_LIST_HEAD(&matches);
	for (i = 0; i < LOCK_HASH_SIZE; i++) {
		list_for_each_entry(sop, &lock_ownerid_hashtbl[i], so_idhash) {
3386
			if (!same_owner_str(sop, owner, clid))
3387 3388 3389 3390 3391 3392 3393 3394 3395
				continue;
			list_for_each_entry(stp, &sop->so_stateids,
					st_perstateowner) {
				if (check_for_locks(stp->st_vfs_file, sop))
					goto out;
				/* Note: so_perclient unused for lockowners,
				 * so it's OK to fool with here. */
				list_add(&sop->so_perclient, &matches);
			}
L
Linus Torvalds 已提交
3396
		}
3397 3398 3399 3400 3401
	}
	/* Clients probably won't expect us to return with some (but not all)
	 * of the lockowner state released; so don't release any until all
	 * have been checked. */
	status = nfs_ok;
N
NeilBrown 已提交
3402 3403 3404 3405 3406 3407
	while (!list_empty(&matches)) {
		sop = list_entry(matches.next, struct nfs4_stateowner,
								so_perclient);
		/* unhash_stateowner deletes so_perclient only
		 * for openowners. */
		list_del(&sop->so_perclient);
3408
		release_lockowner(sop);
L
Linus Torvalds 已提交
3409 3410 3411 3412 3413 3414 3415
	}
out:
	nfs4_unlock_state();
	return status;
}

static inline struct nfs4_client_reclaim *
N
NeilBrown 已提交
3416
alloc_reclaim(void)
L
Linus Torvalds 已提交
3417
{
N
NeilBrown 已提交
3418
	return kmalloc(sizeof(struct nfs4_client_reclaim), GFP_KERNEL);
L
Linus Torvalds 已提交
3419 3420
}

3421
int
3422
nfs4_has_reclaimed_state(const char *name, bool use_exchange_id)
3423 3424 3425 3426
{
	unsigned int strhashval = clientstr_hashval(name);
	struct nfs4_client *clp;

3427
	clp = find_confirmed_client_by_str(name, strhashval, use_exchange_id);
3428 3429 3430
	return clp ? 1 : 0;
}

L
Linus Torvalds 已提交
3431 3432 3433
/*
 * failure => all reset bets are off, nfserr_no_grace...
 */
3434 3435
int
nfs4_client_to_reclaim(const char *name)
L
Linus Torvalds 已提交
3436 3437 3438 3439
{
	unsigned int strhashval;
	struct nfs4_client_reclaim *crp = NULL;

N
NeilBrown 已提交
3440 3441
	dprintk("NFSD nfs4_client_to_reclaim NAME: %.*s\n", HEXDIR_LEN, name);
	crp = alloc_reclaim();
L
Linus Torvalds 已提交
3442 3443
	if (!crp)
		return 0;
N
NeilBrown 已提交
3444
	strhashval = clientstr_hashval(name);
L
Linus Torvalds 已提交
3445 3446
	INIT_LIST_HEAD(&crp->cr_strhash);
	list_add(&crp->cr_strhash, &reclaim_str_hashtbl[strhashval]);
N
NeilBrown 已提交
3447
	memcpy(crp->cr_recdir, name, HEXDIR_LEN);
L
Linus Torvalds 已提交
3448 3449 3450 3451 3452 3453 3454 3455 3456 3457 3458 3459 3460 3461 3462 3463 3464 3465 3466 3467 3468 3469 3470 3471
	reclaim_str_hashtbl_size++;
	return 1;
}

static void
nfs4_release_reclaim(void)
{
	struct nfs4_client_reclaim *crp = NULL;
	int i;

	for (i = 0; i < CLIENT_HASH_SIZE; i++) {
		while (!list_empty(&reclaim_str_hashtbl[i])) {
			crp = list_entry(reclaim_str_hashtbl[i].next,
			                struct nfs4_client_reclaim, cr_strhash);
			list_del(&crp->cr_strhash);
			kfree(crp);
			reclaim_str_hashtbl_size--;
		}
	}
	BUG_ON(reclaim_str_hashtbl_size);
}

/*
 * called from OPEN, CLAIM_PREVIOUS with a new clientid. */
3472
static struct nfs4_client_reclaim *
L
Linus Torvalds 已提交
3473 3474 3475 3476 3477 3478 3479 3480 3481 3482 3483 3484
nfs4_find_reclaim_client(clientid_t *clid)
{
	unsigned int strhashval;
	struct nfs4_client *clp;
	struct nfs4_client_reclaim *crp = NULL;


	/* find clientid in conf_id_hashtbl */
	clp = find_confirmed_client(clid);
	if (clp == NULL)
		return NULL;

N
NeilBrown 已提交
3485 3486 3487
	dprintk("NFSD: nfs4_find_reclaim_client for %.*s with recdir %s\n",
		            clp->cl_name.len, clp->cl_name.data,
			    clp->cl_recdir);
L
Linus Torvalds 已提交
3488 3489

	/* find clp->cl_name in reclaim_str_hashtbl */
N
NeilBrown 已提交
3490
	strhashval = clientstr_hashval(clp->cl_recdir);
L
Linus Torvalds 已提交
3491
	list_for_each_entry(crp, &reclaim_str_hashtbl[strhashval], cr_strhash) {
N
NeilBrown 已提交
3492
		if (same_name(crp->cr_recdir, clp->cl_recdir)) {
L
Linus Torvalds 已提交
3493 3494 3495 3496 3497 3498 3499 3500 3501
			return crp;
		}
	}
	return NULL;
}

/*
* Called from OPEN. Look for clientid in reclaim list.
*/
3502
__be32
L
Linus Torvalds 已提交
3503 3504
nfs4_check_open_reclaim(clientid_t *clid)
{
3505
	return nfs4_find_reclaim_client(clid) ? nfs_ok : nfserr_reclaim_bad;
L
Linus Torvalds 已提交
3506 3507
}

3508
/* initialization to perform at module load time: */
L
Linus Torvalds 已提交
3509

3510
int
3511
nfs4_state_init(void)
L
Linus Torvalds 已提交
3512
{
3513
	int i, status;
L
Linus Torvalds 已提交
3514

3515 3516 3517
	status = nfsd4_init_slabs();
	if (status)
		return status;
L
Linus Torvalds 已提交
3518 3519 3520 3521 3522 3523
	for (i = 0; i < CLIENT_HASH_SIZE; i++) {
		INIT_LIST_HEAD(&conf_id_hashtbl[i]);
		INIT_LIST_HEAD(&conf_str_hashtbl[i]);
		INIT_LIST_HEAD(&unconf_str_hashtbl[i]);
		INIT_LIST_HEAD(&unconf_id_hashtbl[i]);
	}
M
Marc Eshel 已提交
3524 3525
	for (i = 0; i < SESSION_HASH_SIZE; i++)
		INIT_LIST_HEAD(&sessionid_hashtbl[i]);
L
Linus Torvalds 已提交
3526 3527 3528 3529 3530 3531 3532 3533 3534 3535 3536 3537 3538 3539 3540 3541 3542 3543 3544
	for (i = 0; i < FILE_HASH_SIZE; i++) {
		INIT_LIST_HEAD(&file_hashtbl[i]);
	}
	for (i = 0; i < OWNER_HASH_SIZE; i++) {
		INIT_LIST_HEAD(&ownerstr_hashtbl[i]);
		INIT_LIST_HEAD(&ownerid_hashtbl[i]);
	}
	for (i = 0; i < STATEID_HASH_SIZE; i++) {
		INIT_LIST_HEAD(&stateid_hashtbl[i]);
		INIT_LIST_HEAD(&lockstateid_hashtbl[i]);
	}
	for (i = 0; i < LOCK_HASH_SIZE; i++) {
		INIT_LIST_HEAD(&lock_ownerid_hashtbl[i]);
		INIT_LIST_HEAD(&lock_ownerstr_hashtbl[i]);
	}
	memset(&onestateid, ~0, sizeof(stateid_t));
	INIT_LIST_HEAD(&close_lru);
	INIT_LIST_HEAD(&client_lru);
	INIT_LIST_HEAD(&del_recall_lru);
3545 3546 3547
	for (i = 0; i < CLIENT_HASH_SIZE; i++)
		INIT_LIST_HEAD(&reclaim_str_hashtbl[i]);
	reclaim_str_hashtbl_size = 0;
3548
	return 0;
3549 3550
}

3551 3552 3553 3554 3555
static void
nfsd4_load_reboot_recovery_data(void)
{
	int status;

3556 3557
	nfs4_lock_state();
	nfsd4_init_recdir(user_recovery_dirname);
3558
	status = nfsd4_recdir_load();
3559
	nfs4_unlock_state();
3560 3561 3562 3563
	if (status)
		printk("NFSD: Failure reading reboot recovery data\n");
}

3564 3565 3566 3567 3568 3569
unsigned long
get_nfs4_grace_period(void)
{
	return max(user_lease_time, lease_time) * HZ;
}

3570 3571 3572 3573 3574 3575 3576 3577 3578 3579 3580 3581 3582 3583 3584 3585 3586 3587 3588 3589 3590
/*
 * Since the lifetime of a delegation isn't limited to that of an open, a
 * client may quite reasonably hang on to a delegation as long as it has
 * the inode cached.  This becomes an obvious problem the first time a
 * client's inode cache approaches the size of the server's total memory.
 *
 * For now we avoid this problem by imposing a hard limit on the number
 * of delegations, which varies according to the server's memory size.
 */
static void
set_max_delegations(void)
{
	/*
	 * Allow at most 4 delegations per megabyte of RAM.  Quick
	 * estimates suggest that in the worst case (where every delegation
	 * is for a different inode), a delegation could take about 1.5K,
	 * giving a worst case usage of about 6% of memory.
	 */
	max_delegations = nr_free_buffer_pages() >> (20 - 2 - PAGE_SHIFT);
}

3591 3592 3593 3594 3595
/* initialization to perform when the nfsd service is started: */

static void
__nfs4_state_start(void)
{
3596
	unsigned long grace_time;
3597

L
Linus Torvalds 已提交
3598
	boot_time = get_seconds();
3599
	grace_time = get_nfs4_grace_period();
3600
	lease_time = user_lease_time;
3601
	locks_start_grace(&nfsd4_manager);
3602 3603
	printk(KERN_INFO "NFSD: starting %ld-second grace period\n",
	       grace_time/HZ);
3604
	laundry_wq = create_singlethread_workqueue("nfsd4");
3605
	queue_delayed_work(laundry_wq, &laundromat_work, grace_time);
3606
	set_max_delegations();
L
Linus Torvalds 已提交
3607 3608
}

3609
void
3610
nfs4_state_start(void)
L
Linus Torvalds 已提交
3611 3612
{
	if (nfs4_init)
3613
		return;
3614
	nfsd4_load_reboot_recovery_data();
3615
	__nfs4_state_start();
L
Linus Torvalds 已提交
3616
	nfs4_init = 1;
3617
	return;
L
Linus Torvalds 已提交
3618 3619 3620 3621 3622 3623 3624 3625 3626 3627 3628 3629 3630 3631 3632 3633 3634 3635 3636 3637 3638 3639 3640 3641 3642 3643 3644 3645 3646 3647 3648 3649 3650 3651 3652 3653 3654 3655 3656
}

time_t
nfs4_lease_time(void)
{
	return lease_time;
}

static void
__nfs4_state_shutdown(void)
{
	int i;
	struct nfs4_client *clp = NULL;
	struct nfs4_delegation *dp = NULL;
	struct list_head *pos, *next, reaplist;

	for (i = 0; i < CLIENT_HASH_SIZE; i++) {
		while (!list_empty(&conf_id_hashtbl[i])) {
			clp = list_entry(conf_id_hashtbl[i].next, struct nfs4_client, cl_idhash);
			expire_client(clp);
		}
		while (!list_empty(&unconf_str_hashtbl[i])) {
			clp = list_entry(unconf_str_hashtbl[i].next, struct nfs4_client, cl_strhash);
			expire_client(clp);
		}
	}
	INIT_LIST_HEAD(&reaplist);
	spin_lock(&recall_lock);
	list_for_each_safe(pos, next, &del_recall_lru) {
		dp = list_entry (pos, struct nfs4_delegation, dl_recall_lru);
		list_move(&dp->dl_recall_lru, &reaplist);
	}
	spin_unlock(&recall_lock);
	list_for_each_safe(pos, next, &reaplist) {
		dp = list_entry (pos, struct nfs4_delegation, dl_recall_lru);
		list_del_init(&dp->dl_recall_lru);
		unhash_delegation(dp);
	}

3657
	nfsd4_shutdown_recdir();
L
Linus Torvalds 已提交
3658 3659 3660 3661 3662 3663
	nfs4_init = 0;
}

void
nfs4_state_shutdown(void)
{
3664 3665
	cancel_rearming_delayed_workqueue(laundry_wq, &laundromat_work);
	destroy_workqueue(laundry_wq);
3666
	locks_end_grace(&nfsd4_manager);
L
Linus Torvalds 已提交
3667 3668 3669 3670 3671 3672
	nfs4_lock_state();
	nfs4_release_reclaim();
	__nfs4_state_shutdown();
	nfs4_unlock_state();
}

3673 3674 3675 3676
/*
 * user_recovery_dirname is protected by the nfsd_mutex since it's only
 * accessed when nfsd is starting.
 */
3677 3678 3679 3680 3681 3682 3683 3684 3685 3686 3687 3688 3689
static void
nfs4_set_recdir(char *recdir)
{
	strcpy(user_recovery_dirname, recdir);
}

/*
 * Change the NFSv4 recovery directory to recdir.
 */
int
nfs4_reset_recoverydir(char *recdir)
{
	int status;
A
Al Viro 已提交
3690
	struct path path;
3691

A
Al Viro 已提交
3692
	status = kern_path(recdir, LOOKUP_FOLLOW, &path);
3693 3694 3695
	if (status)
		return status;
	status = -ENOTDIR;
A
Al Viro 已提交
3696
	if (S_ISDIR(path.dentry->d_inode->i_mode)) {
3697 3698 3699
		nfs4_set_recdir(recdir);
		status = 0;
	}
A
Al Viro 已提交
3700
	path_put(&path);
3701 3702 3703
	return status;
}

3704 3705 3706 3707 3708 3709
char *
nfs4_recoverydir(void)
{
	return user_recovery_dirname;
}

L
Linus Torvalds 已提交
3710 3711 3712
/*
 * Called when leasetime is changed.
 *
3713 3714 3715 3716 3717
 * The only way the protocol gives us to handle on-the-fly lease changes is to
 * simulate a reboot.  Instead of doing that, we just wait till the next time
 * we start to register any changes in lease time.  If the administrator
 * really wants to change the lease time *now*, they can go ahead and bring
 * nfsd down and then back up again after changing the lease time.
3718 3719 3720
 *
 * user_lease_time is protected by nfsd_mutex since it's only really accessed
 * when nfsd is starting
L
Linus Torvalds 已提交
3721 3722 3723 3724
 */
void
nfs4_reset_lease(time_t leasetime)
{
3725
	user_lease_time = leasetime;
L
Linus Torvalds 已提交
3726
}