nfssvc.c 20.4 KB
Newer Older
L
Linus Torvalds 已提交
1 2 3 4 5 6 7 8
/*
 * Central processing for nfsd.
 *
 * Authors:	Olaf Kirch (okir@monad.swb.de)
 *
 * Copyright (C) 1995, 1996, 1997 Olaf Kirch <okir@monad.swb.de>
 */

9
#include <linux/sched/signal.h>
10
#include <linux/freezer.h>
11
#include <linux/module.h>
L
Linus Torvalds 已提交
12
#include <linux/fs_struct.h>
A
Andy Adamson 已提交
13
#include <linux/swap.h>
L
Linus Torvalds 已提交
14 15 16

#include <linux/sunrpc/stats.h>
#include <linux/sunrpc/svcsock.h>
17
#include <linux/sunrpc/svc_xprt.h>
L
Linus Torvalds 已提交
18
#include <linux/lockd/bind.h>
19
#include <linux/nfsacl.h>
20
#include <linux/seq_file.h>
21 22 23
#include <linux/inetdevice.h>
#include <net/addrconf.h>
#include <net/ipv6.h>
24
#include <net/net_namespace.h>
25 26
#include "nfsd.h"
#include "cache.h"
27
#include "vfs.h"
28
#include "netns.h"
L
Linus Torvalds 已提交
29 30 31 32

#define NFSDDBG_FACILITY	NFSDDBG_SVC

extern struct svc_program	nfsd_program;
33
static int			nfsd(void *vrqstp);
L
Linus Torvalds 已提交
34

35
/*
36
 * nfsd_mutex protects nn->nfsd_serv -- both the pointer itself and the members
37 38 39
 * of the svc_serv struct. In particular, ->sv_nrthreads but also to some
 * extent ->sv_temp_socks and ->sv_permsocks. It also protects nfsdstats.th_cnt
 *
40
 * If (out side the lock) nn->nfsd_serv is non-NULL, then it must point to a
41 42 43 44 45 46 47
 * properly initialised 'struct svc_serv' with ->sv_nrthreads > 0. That number
 * of nfsd threads must exist and each must listed in ->sp_all_threads in each
 * entry of ->sv_pools[].
 *
 * Transitions of the thread count between zero and non-zero are of particular
 * interest since the svc_serv needs to be created and initialized at that
 * point, or freed.
48 49 50 51 52 53 54 55
 *
 * Finally, the nfsd_mutex also protects some of the global variables that are
 * accessed when nfsd starts and that are settable via the write_* routines in
 * nfsctl.c. In particular:
 *
 *	user_recovery_dirname
 *	user_lease_time
 *	nfsd_versions
56 57 58
 */
DEFINE_MUTEX(nfsd_mutex);

59 60 61 62 63 64 65
/*
 * nfsd_drc_lock protects nfsd_drc_max_pages and nfsd_drc_pages_used.
 * nfsd_drc_max_pages limits the total amount of memory available for
 * version 4.1 DRC caches.
 * nfsd_drc_pages_used tracks the current version 4.1 DRC memory usage.
 */
spinlock_t	nfsd_drc_lock;
66 67
unsigned long	nfsd_drc_max_mem;
unsigned long	nfsd_drc_mem_used;
68

69 70 71 72 73 74 75 76
#if defined(CONFIG_NFSD_V2_ACL) || defined(CONFIG_NFSD_V3_ACL)
static struct svc_stat	nfsd_acl_svcstats;
static struct svc_version *	nfsd_acl_version[] = {
	[2] = &nfsd_acl_version2,
	[3] = &nfsd_acl_version3,
};

#define NFSD_ACL_MINVERS            2
77
#define NFSD_ACL_NRVERS		ARRAY_SIZE(nfsd_acl_version)
78 79 80 81 82 83
static struct svc_version *nfsd_acl_versions[NFSD_ACL_NRVERS];

static struct svc_program	nfsd_acl_program = {
	.pg_prog		= NFS_ACL_PROGRAM,
	.pg_nvers		= NFSD_ACL_NRVERS,
	.pg_vers		= nfsd_acl_versions,
84
	.pg_name		= "nfsacl",
85 86 87 88 89 90 91 92 93 94
	.pg_class		= "nfsd",
	.pg_stats		= &nfsd_acl_svcstats,
	.pg_authenticate	= &svc_set_client,
};

static struct svc_stat	nfsd_acl_svcstats = {
	.program	= &nfsd_acl_program,
};
#endif /* defined(CONFIG_NFSD_V2_ACL) || defined(CONFIG_NFSD_V3_ACL) */

95 96 97 98 99 100 101 102 103 104 105
static struct svc_version *	nfsd_version[] = {
	[2] = &nfsd_version2,
#if defined(CONFIG_NFSD_V3)
	[3] = &nfsd_version3,
#endif
#if defined(CONFIG_NFSD_V4)
	[4] = &nfsd_version4,
#endif
};

#define NFSD_MINVERS    	2
106
#define NFSD_NRVERS		ARRAY_SIZE(nfsd_version)
107 108 109
static struct svc_version *nfsd_versions[NFSD_NRVERS];

struct svc_program		nfsd_program = {
110 111 112
#if defined(CONFIG_NFSD_V2_ACL) || defined(CONFIG_NFSD_V3_ACL)
	.pg_next		= &nfsd_acl_program,
#endif
113 114 115 116 117 118 119 120 121 122
	.pg_prog		= NFS_PROGRAM,		/* program number */
	.pg_nvers		= NFSD_NRVERS,		/* nr of entries in nfsd_version */
	.pg_vers		= nfsd_versions,	/* version table */
	.pg_name		= "nfsd",		/* program name */
	.pg_class		= "nfsd",		/* authentication class */
	.pg_stats		= &nfsd_svcstats,	/* version table */
	.pg_authenticate	= &svc_set_client,	/* export authentication */

};

123 124 125
static bool nfsd_supported_minorversions[NFSD_SUPPORTED_MINOR_VERSION + 1] = {
	[0] = 1,
	[1] = 1,
J
J. Bruce Fields 已提交
126
	[2] = 1,
127
};
128

129 130 131
int nfsd_vers(int vers, enum vers_op change)
{
	if (vers < NFSD_MINVERS || vers >= NFSD_NRVERS)
132
		return 0;
133 134 135 136 137
	switch(change) {
	case NFSD_SET:
		nfsd_versions[vers] = nfsd_version[vers];
#if defined(CONFIG_NFSD_V2_ACL) || defined(CONFIG_NFSD_V3_ACL)
		if (vers < NFSD_ACL_NRVERS)
138
			nfsd_acl_versions[vers] = nfsd_acl_version[vers];
139
#endif
140
		break;
141 142 143 144
	case NFSD_CLEAR:
		nfsd_versions[vers] = NULL;
#if defined(CONFIG_NFSD_V2_ACL) || defined(CONFIG_NFSD_V3_ACL)
		if (vers < NFSD_ACL_NRVERS)
145
			nfsd_acl_versions[vers] = NULL;
146 147 148 149 150 151 152 153 154
#endif
		break;
	case NFSD_TEST:
		return nfsd_versions[vers] != NULL;
	case NFSD_AVAIL:
		return nfsd_version[vers] != NULL;
	}
	return 0;
}
155

156 157 158 159 160 161 162 163 164 165 166 167
static void
nfsd_adjust_nfsd_versions4(void)
{
	unsigned i;

	for (i = 0; i <= NFSD_SUPPORTED_MINOR_VERSION; i++) {
		if (nfsd_supported_minorversions[i])
			return;
	}
	nfsd_vers(4, NFSD_CLEAR);
}

168 169
int nfsd_minorversion(u32 minorversion, enum vers_op change)
{
170 171
	if (minorversion > NFSD_SUPPORTED_MINOR_VERSION &&
	    change != NFSD_AVAIL)
172 173 174
		return -1;
	switch(change) {
	case NFSD_SET:
175
		nfsd_supported_minorversions[minorversion] = true;
176
		nfsd_vers(4, NFSD_SET);
177 178
		break;
	case NFSD_CLEAR:
179
		nfsd_supported_minorversions[minorversion] = false;
180
		nfsd_adjust_nfsd_versions4();
181 182
		break;
	case NFSD_TEST:
183
		return nfsd_supported_minorversions[minorversion];
184 185 186 187 188 189
	case NFSD_AVAIL:
		return minorversion <= NFSD_SUPPORTED_MINOR_VERSION;
	}
	return 0;
}

L
Linus Torvalds 已提交
190 191 192 193 194
/*
 * Maximum number of nfsd processes
 */
#define	NFSD_MAXSERVS		8192

195
int nfsd_nrthreads(struct net *net)
L
Linus Torvalds 已提交
196
{
N
Neil Brown 已提交
197
	int rv = 0;
198 199
	struct nfsd_net *nn = net_generic(net, nfsd_net_id);

N
Neil Brown 已提交
200
	mutex_lock(&nfsd_mutex);
201 202
	if (nn->nfsd_serv)
		rv = nn->nfsd_serv->sv_nrthreads;
N
Neil Brown 已提交
203 204
	mutex_unlock(&nfsd_mutex);
	return rv;
L
Linus Torvalds 已提交
205 206
}

207
static int nfsd_init_socks(struct net *net)
208 209
{
	int error;
210 211 212
	struct nfsd_net *nn = net_generic(net, nfsd_net_id);

	if (!list_empty(&nn->nfsd_serv->sv_permsocks))
213 214
		return 0;

215
	error = svc_create_xprt(nn->nfsd_serv, "udp", net, PF_INET, NFS_PORT,
216 217 218 219
					SVC_SOCK_DEFAULTS);
	if (error < 0)
		return error;

220
	error = svc_create_xprt(nn->nfsd_serv, "tcp", net, PF_INET, NFS_PORT,
221 222 223 224 225 226 227
					SVC_SOCK_DEFAULTS);
	if (error < 0)
		return error;

	return 0;
}

228
static int nfsd_users = 0;
229

230 231 232 233
static int nfsd_startup_generic(int nrservs)
{
	int ret;

234
	if (nfsd_users++)
235 236 237 238 239 240 241 242 243
		return 0;

	/*
	 * Readahead param cache - will no-op if it already exists.
	 * (Note therefore results will be suboptimal if number of
	 * threads is modified after nfsd start.)
	 */
	ret = nfsd_racache_init(2*nrservs);
	if (ret)
244 245
		goto dec_users;

246 247 248 249 250 251 252
	ret = nfs4_state_start();
	if (ret)
		goto out_racache;
	return 0;

out_racache:
	nfsd_racache_shutdown();
253 254
dec_users:
	nfsd_users--;
255 256 257 258 259
	return ret;
}

static void nfsd_shutdown_generic(void)
{
260 261 262
	if (--nfsd_users)
		return;

263 264 265 266
	nfs4_state_shutdown();
	nfsd_racache_shutdown();
}

267 268
static bool nfsd_needs_lockd(void)
{
269
#if defined(CONFIG_NFSD_V3)
270
	return (nfsd_versions[2] != NULL) || (nfsd_versions[3] != NULL);
271 272 273
#else
	return (nfsd_versions[2] != NULL);
#endif
274 275
}

276
static int nfsd_startup_net(int nrservs, struct net *net)
277
{
278
	struct nfsd_net *nn = net_generic(net, nfsd_net_id);
279 280
	int ret;

281 282 283
	if (nn->nfsd_net_up)
		return 0;

284
	ret = nfsd_startup_generic(nrservs);
285 286
	if (ret)
		return ret;
287 288 289
	ret = nfsd_init_socks(net);
	if (ret)
		goto out_socks;
290 291 292 293 294 295 296 297

	if (nfsd_needs_lockd() && !nn->lockd_up) {
		ret = lockd_up(net);
		if (ret)
			goto out_socks;
		nn->lockd_up = 1;
	}

298 299 300 301
	ret = nfs4_state_start_net(net);
	if (ret)
		goto out_lockd;

302
	nn->nfsd_net_up = true;
303 304 305
	return 0;

out_lockd:
306 307 308 309
	if (nn->lockd_up) {
		lockd_down(net);
		nn->lockd_up = 0;
	}
310
out_socks:
311
	nfsd_shutdown_generic();
312 313 314
	return ret;
}

315 316
static void nfsd_shutdown_net(struct net *net)
{
317 318
	struct nfsd_net *nn = net_generic(net, nfsd_net_id);

319
	nfs4_state_shutdown_net(net);
320 321 322 323
	if (nn->lockd_up) {
		lockd_down(net);
		nn->lockd_up = 0;
	}
324
	nn->nfsd_net_up = false;
325
	nfsd_shutdown_generic();
326 327
}

328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371
static int nfsd_inetaddr_event(struct notifier_block *this, unsigned long event,
	void *ptr)
{
	struct in_ifaddr *ifa = (struct in_ifaddr *)ptr;
	struct net_device *dev = ifa->ifa_dev->dev;
	struct net *net = dev_net(dev);
	struct nfsd_net *nn = net_generic(net, nfsd_net_id);
	struct sockaddr_in sin;

	if (event != NETDEV_DOWN)
		goto out;

	if (nn->nfsd_serv) {
		dprintk("nfsd_inetaddr_event: removed %pI4\n", &ifa->ifa_local);
		sin.sin_family = AF_INET;
		sin.sin_addr.s_addr = ifa->ifa_local;
		svc_age_temp_xprts_now(nn->nfsd_serv, (struct sockaddr *)&sin);
	}

out:
	return NOTIFY_DONE;
}

static struct notifier_block nfsd_inetaddr_notifier = {
	.notifier_call = nfsd_inetaddr_event,
};

#if IS_ENABLED(CONFIG_IPV6)
static int nfsd_inet6addr_event(struct notifier_block *this,
	unsigned long event, void *ptr)
{
	struct inet6_ifaddr *ifa = (struct inet6_ifaddr *)ptr;
	struct net_device *dev = ifa->idev->dev;
	struct net *net = dev_net(dev);
	struct nfsd_net *nn = net_generic(net, nfsd_net_id);
	struct sockaddr_in6 sin6;

	if (event != NETDEV_DOWN)
		goto out;

	if (nn->nfsd_serv) {
		dprintk("nfsd_inet6addr_event: removed %pI6\n", &ifa->addr);
		sin6.sin6_family = AF_INET6;
		sin6.sin6_addr = ifa->addr;
372 373
		if (ipv6_addr_type(&sin6.sin6_addr) & IPV6_ADDR_LINKLOCAL)
			sin6.sin6_scope_id = ifa->idev->dev->ifindex;
374 375 376 377 378 379 380 381 382 383 384 385
		svc_age_temp_xprts_now(nn->nfsd_serv, (struct sockaddr *)&sin6);
	}

out:
	return NOTIFY_DONE;
}

static struct notifier_block nfsd_inet6addr_notifier = {
	.notifier_call = nfsd_inet6addr_event,
};
#endif

386 387 388
/* Only used under nfsd_mutex, so this atomic may be overkill: */
static atomic_t nfsd_notifier_refcount = ATOMIC_INIT(0);

389
static void nfsd_last_thread(struct svc_serv *serv, struct net *net)
390
{
391 392
	struct nfsd_net *nn = net_generic(net, nfsd_net_id);

393 394 395
	/* check if the notifier still has clients */
	if (atomic_dec_return(&nfsd_notifier_refcount) == 0) {
		unregister_inetaddr_notifier(&nfsd_inetaddr_notifier);
396
#if IS_ENABLED(CONFIG_IPV6)
397
		unregister_inet6addr_notifier(&nfsd_inet6addr_notifier);
398
#endif
399 400
	}

401 402 403 404
	/*
	 * write_ports can create the server without actually starting
	 * any threads--if we get shut down before any threads are
	 * started, then nfsd_last_thread will be run before any of this
405
	 * other initialization has been done except the rpcb information.
406
	 */
407
	svc_rpcb_cleanup(serv, net);
408
	if (!nn->nfsd_net_up)
409
		return;
410

411
	nfsd_shutdown_net(net);
412 413
	printk(KERN_WARNING "nfsd: last server has exited, flushing export "
			    "cache\n");
414
	nfsd_export_flush(net);
415
}
416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437

void nfsd_reset_versions(void)
{
	int found_one = 0;
	int i;

	for (i = NFSD_MINVERS; i < NFSD_NRVERS; i++) {
		if (nfsd_program.pg_vers[i])
			found_one = 1;
	}

	if (!found_one) {
		for (i = NFSD_MINVERS; i < NFSD_NRVERS; i++)
			nfsd_program.pg_vers[i] = nfsd_version[i];
#if defined(CONFIG_NFSD_V2_ACL) || defined(CONFIG_NFSD_V3_ACL)
		for (i = NFSD_ACL_MINVERS; i < NFSD_ACL_NRVERS; i++)
			nfsd_acl_program.pg_vers[i] =
				nfsd_acl_version[i];
#endif
	}
}

A
Andy Adamson 已提交
438 439 440 441 442 443 444 445 446 447 448 449 450 451
/*
 * Each session guarantees a negotiated per slot memory cache for replies
 * which in turn consumes memory beyond the v2/v3/v4.0 server. A dedicated
 * NFSv4.1 server might want to use more memory for a DRC than a machine
 * with mutiple services.
 *
 * Impose a hard limit on the number of pages for the DRC which varies
 * according to the machines free pages. This is of course only a default.
 *
 * For now this is a #defined shift which could be under admin control
 * in the future.
 */
static void set_max_drc(void)
{
452
	#define NFSD_DRC_SIZE_SHIFT	10
453 454 455
	nfsd_drc_max_mem = (nr_free_buffer_pages()
					>> NFSD_DRC_SIZE_SHIFT) * PAGE_SIZE;
	nfsd_drc_mem_used = 0;
456
	spin_lock_init(&nfsd_drc_lock);
457
	dprintk("%s nfsd_drc_max_mem %lu \n", __func__, nfsd_drc_max_mem);
A
Andy Adamson 已提交
458
}
459

460
static int nfsd_get_default_max_blksize(void)
461
{
462 463 464
	struct sysinfo i;
	unsigned long long target;
	unsigned long ret;
465

466
	si_meminfo(&i);
467
	target = (i.totalram - i.totalhigh) << PAGE_SHIFT;
468 469 470 471 472 473 474 475 476 477 478 479 480
	/*
	 * Aim for 1/4096 of memory per thread This gives 1MB on 4Gig
	 * machines, but only uses 32K on 128M machines.  Bottom out at
	 * 8K on 32M and smaller.  Of course, this is only a default.
	 */
	target >>= 12;

	ret = NFSSVC_MAXBLKSIZE;
	while (ret > target && ret >= 8*1024*2)
		ret /= 2;
	return ret;
}

481 482 483 484
static struct svc_serv_ops nfsd_thread_sv_ops = {
	.svo_shutdown		= nfsd_last_thread,
	.svo_function		= nfsd,
	.svo_enqueue_xprt	= svc_xprt_do_enqueue,
485
	.svo_setup		= svc_set_num_threads,
486
	.svo_module		= THIS_MODULE,
487 488
};

489
int nfsd_create_serv(struct net *net)
490
{
491
	int error;
492
	struct nfsd_net *nn = net_generic(net, nfsd_net_id);
493

494
	WARN_ON(!mutex_is_locked(&nfsd_mutex));
495 496
	if (nn->nfsd_serv) {
		svc_get(nn->nfsd_serv);
497 498
		return 0;
	}
499 500
	if (nfsd_max_blksize == 0)
		nfsd_max_blksize = nfsd_get_default_max_blksize();
501
	nfsd_reset_versions();
502
	nn->nfsd_serv = svc_create_pooled(&nfsd_program, nfsd_max_blksize,
503
						&nfsd_thread_sv_ops);
504
	if (nn->nfsd_serv == NULL)
505
		return -ENOMEM;
506

507
	nn->nfsd_serv->sv_maxconn = nn->max_connections;
508
	error = svc_bind(nn->nfsd_serv, net);
509
	if (error < 0) {
510
		svc_destroy(nn->nfsd_serv);
511 512 513
		return error;
	}

514
	set_max_drc();
515 516 517
	/* check if the notifier is already set */
	if (atomic_inc_return(&nfsd_notifier_refcount) == 1) {
		register_inetaddr_notifier(&nfsd_inetaddr_notifier);
518
#if IS_ENABLED(CONFIG_IPV6)
519
		register_inet6addr_notifier(&nfsd_inet6addr_notifier);
520
#endif
521
	}
522
	do_gettimeofday(&nn->nfssvc_boot);		/* record boot time */
523
	return 0;
524 525
}

526
int nfsd_nrpools(struct net *net)
527
{
528 529 530
	struct nfsd_net *nn = net_generic(net, nfsd_net_id);

	if (nn->nfsd_serv == NULL)
531 532
		return 0;
	else
533
		return nn->nfsd_serv->sv_nrpools;
534 535
}

536
int nfsd_get_nrthreads(int n, int *nthreads, struct net *net)
537 538
{
	int i = 0;
539
	struct nfsd_net *nn = net_generic(net, nfsd_net_id);
540

541 542 543
	if (nn->nfsd_serv != NULL) {
		for (i = 0; i < nn->nfsd_serv->sv_nrpools && i < n; i++)
			nthreads[i] = nn->nfsd_serv->sv_pools[i].sp_nrthreads;
544 545 546 547 548
	}

	return 0;
}

549 550 551 552 553 554 555 556 557 558 559 560
void nfsd_destroy(struct net *net)
{
	struct nfsd_net *nn = net_generic(net, nfsd_net_id);
	int destroy = (nn->nfsd_serv->sv_nrthreads == 1);

	if (destroy)
		svc_shutdown_net(nn->nfsd_serv, net);
	svc_destroy(nn->nfsd_serv);
	if (destroy)
		nn->nfsd_serv = NULL;
}

561
int nfsd_set_nrthreads(int n, int *nthreads, struct net *net)
562 563 564 565
{
	int i = 0;
	int tot = 0;
	int err = 0;
566
	struct nfsd_net *nn = net_generic(net, nfsd_net_id);
567

568 569
	WARN_ON(!mutex_is_locked(&nfsd_mutex));

570
	if (nn->nfsd_serv == NULL || n <= 0)
571 572
		return 0;

573 574
	if (n > nn->nfsd_serv->sv_nrpools)
		n = nn->nfsd_serv->sv_nrpools;
575 576 577 578

	/* enforce a global maximum number of threads */
	tot = 0;
	for (i = 0; i < n; i++) {
579
		nthreads[i] = min(nthreads[i], NFSD_MAXSERVS);
580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602
		tot += nthreads[i];
	}
	if (tot > NFSD_MAXSERVS) {
		/* total too large: scale down requested numbers */
		for (i = 0; i < n && tot > 0; i++) {
		    	int new = nthreads[i] * NFSD_MAXSERVS / tot;
			tot -= (nthreads[i] - new);
			nthreads[i] = new;
		}
		for (i = 0; i < n && tot > 0; i++) {
			nthreads[i]--;
			tot--;
		}
	}

	/*
	 * There must always be a thread in pool 0; the admin
	 * can't shut down NFS completely using pool_threads.
	 */
	if (nthreads[0] == 0)
		nthreads[0] = 1;

	/* apply the new numbers */
603
	svc_get(nn->nfsd_serv);
604
	for (i = 0; i < n; i++) {
605 606
		err = nn->nfsd_serv->sv_ops->svo_setup(nn->nfsd_serv,
				&nn->nfsd_serv->sv_pools[i], nthreads[i]);
607 608 609
		if (err)
			break;
	}
610
	nfsd_destroy(net);
611 612 613
	return err;
}

614 615 616 617 618
/*
 * Adjust the number of threads and return the new number of threads.
 * This is also the function that starts the server if necessary, if
 * this is the first time nrservs is nonzero.
 */
L
Linus Torvalds 已提交
619
int
620
nfsd_svc(int nrservs, struct net *net)
L
Linus Torvalds 已提交
621 622
{
	int	error;
623
	bool	nfsd_up_before;
624
	struct nfsd_net *nn = net_generic(net, nfsd_net_id);
625 626

	mutex_lock(&nfsd_mutex);
627
	dprintk("nfsd: creating service\n");
628 629 630

	nrservs = max(nrservs, 0);
	nrservs = min(nrservs, NFSD_MAXSERVS);
631
	error = 0;
632

633
	if (nrservs == 0 && nn->nfsd_serv == NULL)
634 635
		goto out;

636
	error = nfsd_create_serv(net);
637
	if (error)
638 639
		goto out;

640
	nfsd_up_before = nn->nfsd_net_up;
641

642
	error = nfsd_startup_net(nrservs, net);
J
J. Bruce Fields 已提交
643 644
	if (error)
		goto out_destroy;
645 646
	error = nn->nfsd_serv->sv_ops->svo_setup(nn->nfsd_serv,
			NULL, nrservs);
647 648
	if (error)
		goto out_shutdown;
649
	/* We are holding a reference to nn->nfsd_serv which
J
J. Bruce Fields 已提交
650 651 652
	 * we don't want to count in the return value,
	 * so subtract 1
	 */
653
	error = nn->nfsd_serv->sv_nrthreads - 1;
654
out_shutdown:
655
	if (error < 0 && !nfsd_up_before)
656
		nfsd_shutdown_net(net);
657
out_destroy:
658
	nfsd_destroy(net);		/* Release server */
659
out:
660
	mutex_unlock(&nfsd_mutex);
L
Linus Torvalds 已提交
661 662 663 664 665 666 667
	return error;
}


/*
 * This is the NFS server kernel thread
 */
668 669
static int
nfsd(void *vrqstp)
L
Linus Torvalds 已提交
670
{
671
	struct svc_rqst *rqstp = (struct svc_rqst *) vrqstp;
672 673
	struct svc_xprt *perm_sock = list_entry(rqstp->rq_server->sv_permsocks.next, typeof(struct svc_xprt), xpt_list);
	struct net *net = perm_sock->xpt_net;
674
	struct nfsd_net *nn = net_generic(net, nfsd_net_id);
675
	int err;
L
Linus Torvalds 已提交
676 677

	/* Lock module and set up kernel thread */
678
	mutex_lock(&nfsd_mutex);
L
Linus Torvalds 已提交
679

680
	/* At this point, the thread shares current->fs
681 682
	 * with the init process. We need to create files with the
	 * umask as defined by the client instead of init's umask. */
683
	if (unshare_fs_struct() < 0) {
L
Linus Torvalds 已提交
684 685 686
		printk("Unable to start nfsd thread: out of memory\n");
		goto out;
	}
687

L
Linus Torvalds 已提交
688 689
	current->fs->umask = 0;

690 691
	/*
	 * thread is spawned with all signals set to SIG_IGN, re-enable
692
	 * the ones that will bring down the thread
693
	 */
694 695 696 697
	allow_signal(SIGKILL);
	allow_signal(SIGHUP);
	allow_signal(SIGINT);
	allow_signal(SIGQUIT);
698

L
Linus Torvalds 已提交
699
	nfsdstats.th_cnt++;
700 701
	mutex_unlock(&nfsd_mutex);

702
	set_freezable();
L
Linus Torvalds 已提交
703 704 705 706 707

	/*
	 * The main request loop
	 */
	for (;;) {
708 709 710
		/* Update sv_maxconn if it has changed */
		rqstp->rq_server->sv_maxconn = nn->max_connections;

L
Linus Torvalds 已提交
711 712 713 714
		/*
		 * Find a socket with data available and call its
		 * recvfrom routine.
		 */
715
		while ((err = svc_recv(rqstp, 60*60*HZ)) == -EAGAIN)
L
Linus Torvalds 已提交
716
			;
717
		if (err == -EINTR)
L
Linus Torvalds 已提交
718
			break;
719
		validate_process_creds();
720
		svc_process(rqstp);
721
		validate_process_creds();
L
Linus Torvalds 已提交
722 723
	}

724
	/* Clear signals before calling svc_exit_thread() */
725
	flush_signals(current);
L
Linus Torvalds 已提交
726

727
	mutex_lock(&nfsd_mutex);
L
Linus Torvalds 已提交
728 729 730
	nfsdstats.th_cnt --;

out:
731
	rqstp->rq_server = NULL;
732

L
Linus Torvalds 已提交
733 734 735
	/* Release the thread */
	svc_exit_thread(rqstp);

736
	nfsd_destroy(net);
737

L
Linus Torvalds 已提交
738
	/* Release module */
739
	mutex_unlock(&nfsd_mutex);
L
Linus Torvalds 已提交
740
	module_put_and_exit(0);
741
	return 0;
L
Linus Torvalds 已提交
742 743
}

744 745 746 747 748 749 750 751 752
static __be32 map_new_errors(u32 vers, __be32 nfserr)
{
	if (nfserr == nfserr_jukebox && vers == 2)
		return nfserr_dropit;
	if (nfserr == nfserr_wrongsec && vers < 4)
		return nfserr_acces;
	return nfserr;
}

L
Linus Torvalds 已提交
753
int
754
nfsd_dispatch(struct svc_rqst *rqstp, __be32 *statp)
L
Linus Torvalds 已提交
755 756 757
{
	struct svc_procedure	*proc;
	kxdrproc_t		xdr;
A
Al Viro 已提交
758 759
	__be32			nfserr;
	__be32			*nfserrp;
L
Linus Torvalds 已提交
760 761 762 763 764

	dprintk("nfsd_dispatch: vers %d proc %d\n",
				rqstp->rq_vers, rqstp->rq_proc);
	proc = rqstp->rq_procinfo;

765 766 767 768 769 770 771 772 773 774 775 776 777 778
	/*
	 * Give the xdr decoder a chance to change this if it wants
	 * (necessary in the NFSv4.0 compound case)
	 */
	rqstp->rq_cachetype = proc->pc_cachetype;
	/* Decode arguments */
	xdr = proc->pc_decode;
	if (xdr && !xdr(rqstp, (__be32*)rqstp->rq_arg.head[0].iov_base,
			rqstp->rq_argp)) {
		dprintk("nfsd: failed to decode arguments!\n");
		*statp = rpc_garbage_args;
		return 1;
	}

L
Linus Torvalds 已提交
779
	/* Check whether we have this call in the cache. */
780
	switch (nfsd_cache_lookup(rqstp)) {
L
Linus Torvalds 已提交
781 782 783 784 785 786 787 788 789 790 791 792 793
	case RC_DROPIT:
		return 0;
	case RC_REPLY:
		return 1;
	case RC_DOIT:;
		/* do it */
	}

	/* need to grab the location to store the status, as
	 * nfsv4 does some encoding while processing 
	 */
	nfserrp = rqstp->rq_res.head[0].iov_base
		+ rqstp->rq_res.head[0].iov_len;
A
Al Viro 已提交
794
	rqstp->rq_res.head[0].iov_len += sizeof(__be32);
L
Linus Torvalds 已提交
795 796 797

	/* Now call the procedure handler, and encode NFS status. */
	nfserr = proc->pc_func(rqstp, rqstp->rq_argp, rqstp->rq_resp);
798
	nfserr = map_new_errors(rqstp->rq_vers, nfserr);
799
	if (nfserr == nfserr_dropit || test_bit(RQ_DROPME, &rqstp->rq_flags)) {
800
		dprintk("nfsd: Dropping request; may be revisited later\n");
L
Linus Torvalds 已提交
801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823
		nfsd_cache_update(rqstp, RC_NOCACHE, NULL);
		return 0;
	}

	if (rqstp->rq_proc != 0)
		*nfserrp++ = nfserr;

	/* Encode result.
	 * For NFSv2, additional info is never returned in case of an error.
	 */
	if (!(nfserr && rqstp->rq_vers == 2)) {
		xdr = proc->pc_encode;
		if (xdr && !xdr(rqstp, nfserrp,
				rqstp->rq_resp)) {
			/* Failed to encode result. Release cache entry */
			dprintk("nfsd: failed to encode result!\n");
			nfsd_cache_update(rqstp, RC_NOCACHE, NULL);
			*statp = rpc_system_err;
			return 1;
		}
	}

	/* Store reply in cache. */
J
J. Bruce Fields 已提交
824
	nfsd_cache_update(rqstp, rqstp->rq_cachetype, statp + 1);
L
Linus Torvalds 已提交
825 826
	return 1;
}
827 828 829

int nfsd_pool_stats_open(struct inode *inode, struct file *file)
{
830
	int ret;
831
	struct nfsd_net *nn = net_generic(inode->i_sb->s_fs_info, nfsd_net_id);
832

833
	mutex_lock(&nfsd_mutex);
834
	if (nn->nfsd_serv == NULL) {
835
		mutex_unlock(&nfsd_mutex);
836
		return -ENODEV;
837 838
	}
	/* bump up the psudo refcount while traversing */
839 840
	svc_get(nn->nfsd_serv);
	ret = svc_pool_stats_open(nn->nfsd_serv, file);
841 842 843 844 845 846 847
	mutex_unlock(&nfsd_mutex);
	return ret;
}

int nfsd_pool_stats_release(struct inode *inode, struct file *file)
{
	int ret = seq_release(inode, file);
848
	struct net *net = inode->i_sb->s_fs_info;
849

850 851
	mutex_lock(&nfsd_mutex);
	/* this function really, really should have been called svc_put() */
852
	nfsd_destroy(net);
853 854
	mutex_unlock(&nfsd_mutex);
	return ret;
855
}