svc.c 31.9 KB
Newer Older
L
Linus Torvalds 已提交
1 2 3 4 5 6
/*
 * linux/net/sunrpc/svc.c
 *
 * High-level RPC service routines
 *
 * Copyright (C) 1995, 1996 Olaf Kirch <okir@monad.swb.de>
7 8 9 10
 *
 * Multiple threads pools and NUMAisation
 * Copyright (c) 2006 Silicon Graphics, Inc.
 * by Greg Banks <gnb@melbourne.sgi.com>
L
Linus Torvalds 已提交
11 12 13 14 15 16 17 18
 */

#include <linux/linkage.h>
#include <linux/sched.h>
#include <linux/errno.h>
#include <linux/net.h>
#include <linux/in.h>
#include <linux/mm.h>
19 20
#include <linux/interrupt.h>
#include <linux/module.h>
21
#include <linux/kthread.h>
22
#include <linux/slab.h>
L
Linus Torvalds 已提交
23 24 25 26 27 28

#include <linux/sunrpc/types.h>
#include <linux/sunrpc/xdr.h>
#include <linux/sunrpc/stats.h>
#include <linux/sunrpc/svcsock.h>
#include <linux/sunrpc/clnt.h>
29
#include <linux/sunrpc/bc_xprt.h>
L
Linus Torvalds 已提交
30 31 32

#define RPCDBG_FACILITY	RPCDBG_SVCDSP

33 34
static void svc_unregister(const struct svc_serv *serv);

35 36
#define svc_serv_is_pooled(serv)    ((serv)->sv_function)

37 38 39 40
/*
 * Mode for mapping cpus to pools.
 */
enum {
41
	SVC_POOL_AUTO = -1,	/* choose one of the others */
42 43 44 45 46
	SVC_POOL_GLOBAL,	/* no mapping, just a single global pool
				 * (legacy & UP mode) */
	SVC_POOL_PERCPU,	/* one pool per cpu */
	SVC_POOL_PERNODE	/* one pool per numa node */
};
47
#define SVC_POOL_DEFAULT	SVC_POOL_GLOBAL
48 49 50 51 52 53

/*
 * Structure for mapping cpus to pools and vice versa.
 * Setup once during sunrpc initialisation.
 */
static struct svc_pool_map {
54
	int count;			/* How many svc_servs use us */
55 56 57 58 59 60 61
	int mode;			/* Note: int not enum to avoid
					 * warnings about "enumeration value
					 * not handled in switch" */
	unsigned int npools;
	unsigned int *pool_to;		/* maps pool id to cpu or node */
	unsigned int *to_pool;		/* maps cpu or node to pool id */
} svc_pool_map = {
62 63
	.count = 0,
	.mode = SVC_POOL_DEFAULT
64
};
65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115
static DEFINE_MUTEX(svc_pool_map_mutex);/* protects svc_pool_map.count only */

static int
param_set_pool_mode(const char *val, struct kernel_param *kp)
{
	int *ip = (int *)kp->arg;
	struct svc_pool_map *m = &svc_pool_map;
	int err;

	mutex_lock(&svc_pool_map_mutex);

	err = -EBUSY;
	if (m->count)
		goto out;

	err = 0;
	if (!strncmp(val, "auto", 4))
		*ip = SVC_POOL_AUTO;
	else if (!strncmp(val, "global", 6))
		*ip = SVC_POOL_GLOBAL;
	else if (!strncmp(val, "percpu", 6))
		*ip = SVC_POOL_PERCPU;
	else if (!strncmp(val, "pernode", 7))
		*ip = SVC_POOL_PERNODE;
	else
		err = -EINVAL;

out:
	mutex_unlock(&svc_pool_map_mutex);
	return err;
}

static int
param_get_pool_mode(char *buf, struct kernel_param *kp)
{
	int *ip = (int *)kp->arg;

	switch (*ip)
	{
	case SVC_POOL_AUTO:
		return strlcpy(buf, "auto", 20);
	case SVC_POOL_GLOBAL:
		return strlcpy(buf, "global", 20);
	case SVC_POOL_PERCPU:
		return strlcpy(buf, "percpu", 20);
	case SVC_POOL_PERNODE:
		return strlcpy(buf, "pernode", 20);
	default:
		return sprintf(buf, "%d", *ip);
	}
}
116

117 118
module_param_call(pool_mode, param_set_pool_mode, param_get_pool_mode,
		 &svc_pool_map.mode, 0644);
119 120 121 122 123 124 125 126 127 128

/*
 * Detect best pool mapping mode heuristically,
 * according to the machine's topology.
 */
static int
svc_pool_map_choose_mode(void)
{
	unsigned int node;

129
	if (nr_online_nodes > 1) {
130 131 132 133 134 135 136
		/*
		 * Actually have multiple NUMA nodes,
		 * so split pools on NUMA node boundaries
		 */
		return SVC_POOL_PERNODE;
	}

137
	node = first_online_node;
138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180
	if (nr_cpus_node(node) > 2) {
		/*
		 * Non-trivial SMP, or CONFIG_NUMA on
		 * non-NUMA hardware, e.g. with a generic
		 * x86_64 kernel on Xeons.  In this case we
		 * want to divide the pools on cpu boundaries.
		 */
		return SVC_POOL_PERCPU;
	}

	/* default: one global pool */
	return SVC_POOL_GLOBAL;
}

/*
 * Allocate the to_pool[] and pool_to[] arrays.
 * Returns 0 on success or an errno.
 */
static int
svc_pool_map_alloc_arrays(struct svc_pool_map *m, unsigned int maxpools)
{
	m->to_pool = kcalloc(maxpools, sizeof(unsigned int), GFP_KERNEL);
	if (!m->to_pool)
		goto fail;
	m->pool_to = kcalloc(maxpools, sizeof(unsigned int), GFP_KERNEL);
	if (!m->pool_to)
		goto fail_free;

	return 0;

fail_free:
	kfree(m->to_pool);
fail:
	return -ENOMEM;
}

/*
 * Initialise the pool map for SVC_POOL_PERCPU mode.
 * Returns number of pools or <0 on error.
 */
static int
svc_pool_map_init_percpu(struct svc_pool_map *m)
{
181
	unsigned int maxpools = nr_cpu_ids;
182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208
	unsigned int pidx = 0;
	unsigned int cpu;
	int err;

	err = svc_pool_map_alloc_arrays(m, maxpools);
	if (err)
		return err;

	for_each_online_cpu(cpu) {
		BUG_ON(pidx > maxpools);
		m->to_pool[cpu] = pidx;
		m->pool_to[pidx] = cpu;
		pidx++;
	}
	/* cpus brought online later all get mapped to pool0, sorry */

	return pidx;
};


/*
 * Initialise the pool map for SVC_POOL_PERNODE mode.
 * Returns number of pools or <0 on error.
 */
static int
svc_pool_map_init_pernode(struct svc_pool_map *m)
{
209
	unsigned int maxpools = nr_node_ids;
210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231
	unsigned int pidx = 0;
	unsigned int node;
	int err;

	err = svc_pool_map_alloc_arrays(m, maxpools);
	if (err)
		return err;

	for_each_node_with_cpus(node) {
		/* some architectures (e.g. SN2) have cpuless nodes */
		BUG_ON(pidx > maxpools);
		m->to_pool[node] = pidx;
		m->pool_to[pidx] = node;
		pidx++;
	}
	/* nodes brought online later all get mapped to pool0, sorry */

	return pidx;
}


/*
232 233 234
 * Add a reference to the global map of cpus to pools (and
 * vice versa).  Initialise the map if we're the first user.
 * Returns the number of pools.
235 236
 */
static unsigned int
237
svc_pool_map_get(void)
238 239 240 241
{
	struct svc_pool_map *m = &svc_pool_map;
	int npools = -1;

242 243 244 245
	mutex_lock(&svc_pool_map_mutex);

	if (m->count++) {
		mutex_unlock(&svc_pool_map_mutex);
246
		return m->npools;
247
	}
248

249 250
	if (m->mode == SVC_POOL_AUTO)
		m->mode = svc_pool_map_choose_mode();
251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267

	switch (m->mode) {
	case SVC_POOL_PERCPU:
		npools = svc_pool_map_init_percpu(m);
		break;
	case SVC_POOL_PERNODE:
		npools = svc_pool_map_init_pernode(m);
		break;
	}

	if (npools < 0) {
		/* default, or memory allocation failure */
		npools = 1;
		m->mode = SVC_POOL_GLOBAL;
	}
	m->npools = npools;

268
	mutex_unlock(&svc_pool_map_mutex);
269 270 271
	return m->npools;
}

272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297

/*
 * Drop a reference to the global map of cpus to pools.
 * When the last reference is dropped, the map data is
 * freed; this allows the sysadmin to change the pool
 * mode using the pool_mode module option without
 * rebooting or re-loading sunrpc.ko.
 */
static void
svc_pool_map_put(void)
{
	struct svc_pool_map *m = &svc_pool_map;

	mutex_lock(&svc_pool_map_mutex);

	if (!--m->count) {
		m->mode = SVC_POOL_DEFAULT;
		kfree(m->to_pool);
		kfree(m->pool_to);
		m->npools = 0;
	}

	mutex_unlock(&svc_pool_map_mutex);
}


298 299 300 301 302 303 304 305 306 307 308 309
static int svc_pool_map_get_node(unsigned int pidx)
{
	const struct svc_pool_map *m = &svc_pool_map;

	if (m->count) {
		if (m->mode == SVC_POOL_PERCPU)
			return cpu_to_node(m->pool_to[pidx]);
		if (m->mode == SVC_POOL_PERNODE)
			return m->pool_to[pidx];
	}
	return NUMA_NO_NODE;
}
310
/*
311
 * Set the given thread's cpus_allowed mask so that it
312 313
 * will only run on cpus in the given pool.
 */
314 315
static inline void
svc_pool_map_set_cpumask(struct task_struct *task, unsigned int pidx)
316 317
{
	struct svc_pool_map *m = &svc_pool_map;
318
	unsigned int node = m->pool_to[pidx];
319 320 321

	/*
	 * The caller checks for sv_nrpools > 1, which
322
	 * implies that we've been initialized.
323
	 */
324
	BUG_ON(m->count == 0);
325

326
	switch (m->mode) {
327
	case SVC_POOL_PERCPU:
328
	{
329
		set_cpus_allowed_ptr(task, cpumask_of(node));
330
		break;
331
	}
332
	case SVC_POOL_PERNODE:
333
	{
334
		set_cpus_allowed_ptr(task, cpumask_of_node(node));
335
		break;
336
	}
337
	}
338 339 340 341 342 343 344 345 346 347 348 349 350 351
}

/*
 * Use the mapping mode to choose a pool for a given CPU.
 * Used when enqueueing an incoming RPC.  Always returns
 * a non-NULL pool pointer.
 */
struct svc_pool *
svc_pool_for_cpu(struct svc_serv *serv, int cpu)
{
	struct svc_pool_map *m = &svc_pool_map;
	unsigned int pidx = 0;

	/*
352
	 * An uninitialised map happens in a pure client when
353 354 355
	 * lockd is brought up, so silently treat it the
	 * same as SVC_POOL_GLOBAL.
	 */
356 357 358 359 360 361 362 363 364
	if (svc_serv_is_pooled(serv)) {
		switch (m->mode) {
		case SVC_POOL_PERCPU:
			pidx = m->to_pool[cpu];
			break;
		case SVC_POOL_PERNODE:
			pidx = m->to_pool[cpu_to_node(cpu)];
			break;
		}
365 366 367 368
	}
	return &serv->sv_pools[pidx % serv->sv_nrpools];
}

369 370 371 372 373 374 375 376 377 378 379 380 381
static int svc_rpcb_setup(struct svc_serv *serv)
{
	int err;

	err = rpcb_create_local();
	if (err)
		return err;

	/* Remove any stale portmap registrations */
	svc_unregister(serv);
	return 0;
}

382
void svc_rpcb_cleanup(struct svc_serv *serv)
383 384 385 386
{
	svc_unregister(serv);
	rpcb_put_local();
}
387
EXPORT_SYMBOL_GPL(svc_rpcb_cleanup);
388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404

static int svc_uses_rpcbind(struct svc_serv *serv)
{
	struct svc_program	*progp;
	unsigned int		i;

	for (progp = serv->sv_program; progp; progp = progp->pg_next) {
		for (i = 0; i < progp->pg_nvers; i++) {
			if (progp->pg_vers[i] == NULL)
				continue;
			if (progp->pg_vers[i]->vs_hidden == 0)
				return 1;
		}
	}

	return 0;
}
405

L
Linus Torvalds 已提交
406 407 408
/*
 * Create an RPC service
 */
409 410
static struct svc_serv *
__svc_create(struct svc_program *prog, unsigned int bufsize, int npools,
411
	     void (*shutdown)(struct svc_serv *serv))
L
Linus Torvalds 已提交
412 413
{
	struct svc_serv	*serv;
414
	unsigned int vers;
L
Linus Torvalds 已提交
415
	unsigned int xdrsize;
416
	unsigned int i;
L
Linus Torvalds 已提交
417

418
	if (!(serv = kzalloc(sizeof(*serv), GFP_KERNEL)))
L
Linus Torvalds 已提交
419
		return NULL;
420
	serv->sv_name      = prog->pg_name;
L
Linus Torvalds 已提交
421 422 423
	serv->sv_program   = prog;
	serv->sv_nrthreads = 1;
	serv->sv_stats     = prog->pg_stats;
424 425 426 427
	if (bufsize > RPCSVC_MAXPAYLOAD)
		bufsize = RPCSVC_MAXPAYLOAD;
	serv->sv_max_payload = bufsize? bufsize : 4096;
	serv->sv_max_mesg  = roundup(serv->sv_max_payload + PAGE_SIZE, PAGE_SIZE);
428
	serv->sv_shutdown  = shutdown;
L
Linus Torvalds 已提交
429
	xdrsize = 0;
430 431 432 433 434 435 436 437 438 439 440 441
	while (prog) {
		prog->pg_lovers = prog->pg_nvers-1;
		for (vers=0; vers<prog->pg_nvers ; vers++)
			if (prog->pg_vers[vers]) {
				prog->pg_hivers = vers;
				if (prog->pg_lovers > vers)
					prog->pg_lovers = vers;
				if (prog->pg_vers[vers]->vs_xdrsize > xdrsize)
					xdrsize = prog->pg_vers[vers]->vs_xdrsize;
			}
		prog = prog->pg_next;
	}
L
Linus Torvalds 已提交
442 443 444
	serv->sv_xdrsize   = xdrsize;
	INIT_LIST_HEAD(&serv->sv_tempsocks);
	INIT_LIST_HEAD(&serv->sv_permsocks);
445
	init_timer(&serv->sv_temptimer);
L
Linus Torvalds 已提交
446 447
	spin_lock_init(&serv->sv_lock);

448
	serv->sv_nrpools = npools;
449
	serv->sv_pools =
450
		kcalloc(serv->sv_nrpools, sizeof(struct svc_pool),
451 452 453 454 455 456 457 458 459
			GFP_KERNEL);
	if (!serv->sv_pools) {
		kfree(serv);
		return NULL;
	}

	for (i = 0; i < serv->sv_nrpools; i++) {
		struct svc_pool *pool = &serv->sv_pools[i];

460
		dprintk("svc: initialising pool %u for %s\n",
461 462 463 464 465
				i, serv->sv_name);

		pool->sp_id = i;
		INIT_LIST_HEAD(&pool->sp_threads);
		INIT_LIST_HEAD(&pool->sp_sockets);
466
		INIT_LIST_HEAD(&pool->sp_all_threads);
467 468 469
		spin_lock_init(&pool->sp_lock);
	}

470 471 472 473 474 475 476 477 478
	if (svc_uses_rpcbind(serv)) {
	       	if (svc_rpcb_setup(serv) < 0) {
			kfree(serv->sv_pools);
			kfree(serv);
			return NULL;
		}
		if (!serv->sv_shutdown)
			serv->sv_shutdown = svc_rpcb_cleanup;
	}
L
Linus Torvalds 已提交
479 480 481 482

	return serv;
}

483 484
struct svc_serv *
svc_create(struct svc_program *prog, unsigned int bufsize,
485
	   void (*shutdown)(struct svc_serv *serv))
486
{
487
	return __svc_create(prog, bufsize, /*npools*/1, shutdown);
488
}
489
EXPORT_SYMBOL_GPL(svc_create);
490 491 492

struct svc_serv *
svc_create_pooled(struct svc_program *prog, unsigned int bufsize,
493
		  void (*shutdown)(struct svc_serv *serv),
494
		  svc_thread_fn func, struct module *mod)
495 496
{
	struct svc_serv *serv;
497
	unsigned int npools = svc_pool_map_get();
498

499
	serv = __svc_create(prog, bufsize, npools, shutdown);
500 501 502 503 504 505 506 507

	if (serv != NULL) {
		serv->sv_function = func;
		serv->sv_module = mod;
	}

	return serv;
}
508
EXPORT_SYMBOL_GPL(svc_create_pooled);
509

L
Linus Torvalds 已提交
510
/*
511 512
 * Destroy an RPC service. Should be called with appropriate locking to
 * protect the sv_nrthreads, sv_permsocks and sv_tempsocks.
L
Linus Torvalds 已提交
513 514 515 516
 */
void
svc_destroy(struct svc_serv *serv)
{
517
	dprintk("svc: svc_destroy(%s, %d)\n",
L
Linus Torvalds 已提交
518 519 520 521 522 523 524 525 526 527 528
				serv->sv_program->pg_name,
				serv->sv_nrthreads);

	if (serv->sv_nrthreads) {
		if (--(serv->sv_nrthreads) != 0) {
			svc_sock_update_bufs(serv);
			return;
		}
	} else
		printk("svc_destroy: no threads for serv=%p!\n", serv);

529 530
	del_timer_sync(&serv->sv_temptimer);

531
	svc_close_all(serv);
532

533 534 535
	if (serv->sv_shutdown)
		serv->sv_shutdown(serv);

L
Linus Torvalds 已提交
536 537
	cache_clean_deferred(serv);

538 539 540
	if (svc_serv_is_pooled(serv))
		svc_pool_map_put();

541
	kfree(serv->sv_pools);
L
Linus Torvalds 已提交
542 543
	kfree(serv);
}
544
EXPORT_SYMBOL_GPL(svc_destroy);
L
Linus Torvalds 已提交
545 546 547 548 549 550

/*
 * Allocate an RPC server's buffer space.
 * We allocate pages and place them in rq_argpages.
 */
static int
551
svc_init_buffer(struct svc_rqst *rqstp, unsigned int size, int node)
L
Linus Torvalds 已提交
552
{
553
	unsigned int pages, arghi;
554

555 556 557 558
	/* bc_xprt uses fore channel allocated buffers */
	if (svc_is_backchannel(rqstp))
		return 1;

559 560 561
	pages = size / PAGE_SIZE + 1; /* extra page as we hold both request and reply.
				       * We assume one is at most one page
				       */
L
Linus Torvalds 已提交
562
	arghi = 0;
563
	BUG_ON(pages > RPCSVC_MAXPAGES);
L
Linus Torvalds 已提交
564
	while (pages) {
565
		struct page *p = alloc_pages_node(node, GFP_KERNEL, 0);
L
Linus Torvalds 已提交
566 567
		if (!p)
			break;
568
		rqstp->rq_pages[arghi++] = p;
L
Linus Torvalds 已提交
569 570
		pages--;
	}
571
	return pages == 0;
L
Linus Torvalds 已提交
572 573 574 575 576 577 578 579
}

/*
 * Release an RPC server buffer
 */
static void
svc_release_buffer(struct svc_rqst *rqstp)
{
580 581 582
	unsigned int i;

	for (i = 0; i < ARRAY_SIZE(rqstp->rq_pages); i++)
583 584
		if (rqstp->rq_pages[i])
			put_page(rqstp->rq_pages[i]);
L
Linus Torvalds 已提交
585 586
}

587
struct svc_rqst *
588
svc_prepare_thread(struct svc_serv *serv, struct svc_pool *pool, int node)
L
Linus Torvalds 已提交
589 590 591
{
	struct svc_rqst	*rqstp;

592
	rqstp = kzalloc_node(sizeof(*rqstp), GFP_KERNEL, node);
L
Linus Torvalds 已提交
593
	if (!rqstp)
594
		goto out_enomem;
L
Linus Torvalds 已提交
595 596 597 598

	init_waitqueue_head(&rqstp->rq_wait);

	serv->sv_nrthreads++;
599 600
	spin_lock_bh(&pool->sp_lock);
	pool->sp_nrthreads++;
601
	list_add(&rqstp->rq_all, &pool->sp_all_threads);
602
	spin_unlock_bh(&pool->sp_lock);
L
Linus Torvalds 已提交
603
	rqstp->rq_server = serv;
604
	rqstp->rq_pool = pool;
605

606
	rqstp->rq_argp = kmalloc_node(serv->sv_xdrsize, GFP_KERNEL, node);
607 608 609
	if (!rqstp->rq_argp)
		goto out_thread;

610
	rqstp->rq_resp = kmalloc_node(serv->sv_xdrsize, GFP_KERNEL, node);
611 612 613
	if (!rqstp->rq_resp)
		goto out_thread;

614
	if (!svc_init_buffer(rqstp, serv->sv_max_mesg, node))
615 616 617 618 619 620 621 622
		goto out_thread;

	return rqstp;
out_thread:
	svc_exit_thread(rqstp);
out_enomem:
	return ERR_PTR(-ENOMEM);
}
623
EXPORT_SYMBOL_GPL(svc_prepare_thread);
624

625 626 627 628 629 630 631 632 633
/*
 * Choose a pool in which to create a new thread, for svc_set_num_threads
 */
static inline struct svc_pool *
choose_pool(struct svc_serv *serv, struct svc_pool *pool, unsigned int *state)
{
	if (pool != NULL)
		return pool;

634
	return &serv->sv_pools[(*state)++ % serv->sv_nrpools];
635 636 637 638 639 640 641 642 643 644 645 646 647 648 649
}

/*
 * Choose a thread to kill, for svc_set_num_threads
 */
static inline struct task_struct *
choose_victim(struct svc_serv *serv, struct svc_pool *pool, unsigned int *state)
{
	unsigned int i;
	struct task_struct *task = NULL;

	if (pool != NULL) {
		spin_lock_bh(&pool->sp_lock);
	} else {
		/* choose a pool in round-robin fashion */
650 651
		for (i = 0; i < serv->sv_nrpools; i++) {
			pool = &serv->sv_pools[--(*state) % serv->sv_nrpools];
652
			spin_lock_bh(&pool->sp_lock);
653 654
			if (!list_empty(&pool->sp_all_threads))
				goto found_pool;
655
			spin_unlock_bh(&pool->sp_lock);
656
		}
657 658 659 660 661 662 663 664 665 666 667 668 669 670
		return NULL;
	}

found_pool:
	if (!list_empty(&pool->sp_all_threads)) {
		struct svc_rqst *rqstp;

		/*
		 * Remove from the pool->sp_all_threads list
		 * so we don't try to kill it again.
		 */
		rqstp = list_entry(pool->sp_all_threads.next, struct svc_rqst, rq_all);
		list_del_init(&rqstp->rq_all);
		task = rqstp->rq_task;
671
	}
672 673 674 675 676 677 678 679 680 681
	spin_unlock_bh(&pool->sp_lock);

	return task;
}

/*
 * Create or destroy enough new threads to make the number
 * of threads the given number.  If `pool' is non-NULL, applies
 * only to threads in that pool, otherwise round-robins between
 * all pools.  Must be called with a svc_get() reference and
682
 * the BKL or another lock to protect access to svc_serv fields.
683 684 685 686 687 688 689 690 691 692 693
 *
 * Destroying threads relies on the service threads filling in
 * rqstp->rq_task, which only the nfs ones do.  Assumes the serv
 * has been created using svc_create_pooled().
 *
 * Based on code that used to be in nfsd_svc() but tweaked
 * to be pool-aware.
 */
int
svc_set_num_threads(struct svc_serv *serv, struct svc_pool *pool, int nrservs)
{
694 695 696
	struct svc_rqst	*rqstp;
	struct task_struct *task;
	struct svc_pool *chosen_pool;
697 698
	int error = 0;
	unsigned int state = serv->sv_nrthreads-1;
699
	int node;
700 701 702 703 704 705 706 707 708 709 710 711 712

	if (pool == NULL) {
		/* The -1 assumes caller has done a svc_get() */
		nrservs -= (serv->sv_nrthreads-1);
	} else {
		spin_lock_bh(&pool->sp_lock);
		nrservs -= pool->sp_nrthreads;
		spin_unlock_bh(&pool->sp_lock);
	}

	/* create new threads */
	while (nrservs > 0) {
		nrservs--;
713 714
		chosen_pool = choose_pool(serv, pool, &state);

715 716
		node = svc_pool_map_get_node(chosen_pool->sp_id);
		rqstp = svc_prepare_thread(serv, chosen_pool, node);
717 718 719 720 721
		if (IS_ERR(rqstp)) {
			error = PTR_ERR(rqstp);
			break;
		}

722
		__module_get(serv->sv_module);
723 724
		task = kthread_create_on_node(serv->sv_function, rqstp,
					      node, serv->sv_name);
725 726
		if (IS_ERR(task)) {
			error = PTR_ERR(task);
727
			module_put(serv->sv_module);
728
			svc_exit_thread(rqstp);
729 730
			break;
		}
731 732 733 734 735 736 737

		rqstp->rq_task = task;
		if (serv->sv_nrpools > 1)
			svc_pool_map_set_cpumask(task, chosen_pool->sp_id);

		svc_sock_update_bufs(serv);
		wake_up_process(task);
738 739 740
	}
	/* destroy old threads */
	while (nrservs < 0 &&
741
	       (task = choose_victim(serv, pool, &state)) != NULL) {
742
		send_sig(SIGINT, task, 1);
743 744 745 746 747
		nrservs++;
	}

	return error;
}
748
EXPORT_SYMBOL_GPL(svc_set_num_threads);
749

750
/*
751 752
 * Called from a server thread as it's exiting. Caller must hold the BKL or
 * the "service mutex", whichever is appropriate for the service.
L
Linus Torvalds 已提交
753 754 755 756 757
 */
void
svc_exit_thread(struct svc_rqst *rqstp)
{
	struct svc_serv	*serv = rqstp->rq_server;
758
	struct svc_pool	*pool = rqstp->rq_pool;
L
Linus Torvalds 已提交
759 760

	svc_release_buffer(rqstp);
J
Jesper Juhl 已提交
761 762 763
	kfree(rqstp->rq_resp);
	kfree(rqstp->rq_argp);
	kfree(rqstp->rq_auth_data);
764 765 766

	spin_lock_bh(&pool->sp_lock);
	pool->sp_nrthreads--;
767
	list_del(&rqstp->rq_all);
768 769
	spin_unlock_bh(&pool->sp_lock);

L
Linus Torvalds 已提交
770 771 772 773 774 775
	kfree(rqstp);

	/* Release the server */
	if (serv)
		svc_destroy(serv);
}
776
EXPORT_SYMBOL_GPL(svc_exit_thread);
L
Linus Torvalds 已提交
777 778

/*
779 780
 * Register an "inet" protocol family netid with the local
 * rpcbind daemon via an rpcbind v4 SET request.
781
 *
782 783
 * No netconfig infrastructure is available in the kernel, so
 * we map IP_ protocol numbers to netids by hand.
784
 *
785 786
 * Returns zero on success; a negative errno value is returned
 * if any error occurs.
L
Linus Torvalds 已提交
787
 */
788 789 790
static int __svc_rpcb_register4(const u32 program, const u32 version,
				const unsigned short protocol,
				const unsigned short port)
791
{
792
	const struct sockaddr_in sin = {
793 794 795 796
		.sin_family		= AF_INET,
		.sin_addr.s_addr	= htonl(INADDR_ANY),
		.sin_port		= htons(port),
	};
797 798
	const char *netid;
	int error;
799 800 801 802 803 804 805 806 807

	switch (protocol) {
	case IPPROTO_UDP:
		netid = RPCBIND_NETID_UDP;
		break;
	case IPPROTO_TCP:
		netid = RPCBIND_NETID_TCP;
		break;
	default:
808
		return -ENOPROTOOPT;
809 810
	}

811 812 813 814 815 816 817 818 819 820 821
	error = rpcb_v4_register(program, version,
					(const struct sockaddr *)&sin, netid);

	/*
	 * User space didn't support rpcbind v4, so retry this
	 * registration request with the legacy rpcbind v2 protocol.
	 */
	if (error == -EPROTONOSUPPORT)
		error = rpcb_register(program, version, protocol, port);

	return error;
822 823
}

824
#if defined(CONFIG_IPV6) || defined(CONFIG_IPV6_MODULE)
825 826 827 828 829 830 831 832 833 834 835 836 837 838
/*
 * Register an "inet6" protocol family netid with the local
 * rpcbind daemon via an rpcbind v4 SET request.
 *
 * No netconfig infrastructure is available in the kernel, so
 * we map IP_ protocol numbers to netids by hand.
 *
 * Returns zero on success; a negative errno value is returned
 * if any error occurs.
 */
static int __svc_rpcb_register6(const u32 program, const u32 version,
				const unsigned short protocol,
				const unsigned short port)
{
839
	const struct sockaddr_in6 sin6 = {
840 841 842 843
		.sin6_family		= AF_INET6,
		.sin6_addr		= IN6ADDR_ANY_INIT,
		.sin6_port		= htons(port),
	};
844 845
	const char *netid;
	int error;
846

847 848 849
	switch (protocol) {
	case IPPROTO_UDP:
		netid = RPCBIND_NETID_UDP6;
850
		break;
851
	case IPPROTO_TCP:
852 853 854
		netid = RPCBIND_NETID_TCP6;
		break;
	default:
855
		return -ENOPROTOOPT;
856 857
	}

858 859 860 861 862 863 864 865 866 867 868
	error = rpcb_v4_register(program, version,
					(const struct sockaddr *)&sin6, netid);

	/*
	 * User space didn't support rpcbind version 4, so we won't
	 * use a PF_INET6 listener.
	 */
	if (error == -EPROTONOSUPPORT)
		error = -EAFNOSUPPORT;

	return error;
869
}
870
#endif	/* defined(CONFIG_IPV6) || defined(CONFIG_IPV6_MODULE) */
871 872 873 874 875 876 877

/*
 * Register a kernel RPC service via rpcbind version 4.
 *
 * Returns zero on success; a negative errno value is returned
 * if any error occurs.
 */
878 879
static int __svc_register(const char *progname,
			  const u32 program, const u32 version,
880
			  const int family,
881 882 883
			  const unsigned short protocol,
			  const unsigned short port)
{
884
	int error = -EAFNOSUPPORT;
885 886

	switch (family) {
887
	case PF_INET:
888
		error = __svc_rpcb_register4(program, version,
889
						protocol, port);
890 891
		break;
#if defined(CONFIG_IPV6) || defined(CONFIG_IPV6_MODULE)
892
	case PF_INET6:
893 894
		error = __svc_rpcb_register6(program, version,
						protocol, port);
895
#endif	/* defined(CONFIG_IPV6) || defined(CONFIG_IPV6_MODULE) */
896 897
	}

898 899 900 901
	if (error < 0)
		printk(KERN_WARNING "svc: failed to register %sv%u RPC "
			"service (errno %d).\n", progname, version, -error);
	return error;
902
}
903

904 905 906
/**
 * svc_register - register an RPC service with the local portmapper
 * @serv: svc_serv struct for the service to register
907
 * @family: protocol family of service's listener socket
908 909 910
 * @proto: transport protocol number to advertise
 * @port: port to advertise
 *
911
 * Service is registered for any address in the passed-in protocol family
912
 */
913 914
int svc_register(const struct svc_serv *serv, const int family,
		 const unsigned short proto, const unsigned short port)
L
Linus Torvalds 已提交
915 916
{
	struct svc_program	*progp;
917
	unsigned int		i;
918
	int			error = 0;
L
Linus Torvalds 已提交
919

920
	BUG_ON(proto == 0 && port == 0);
L
Linus Torvalds 已提交
921

922 923 924 925 926
	for (progp = serv->sv_program; progp; progp = progp->pg_next) {
		for (i = 0; i < progp->pg_nvers; i++) {
			if (progp->pg_vers[i] == NULL)
				continue;

927
			dprintk("svc: svc_register(%sv%d, %s, %u, %u)%s\n",
928
					progp->pg_name,
929
					i,
930 931
					proto == IPPROTO_UDP?  "udp" : "tcp",
					port,
932
					family,
933 934 935 936 937 938
					progp->pg_vers[i]->vs_hidden?
						" (but not telling portmap)" : "");

			if (progp->pg_vers[i]->vs_hidden)
				continue;

939 940
			error = __svc_register(progp->pg_name, progp->pg_prog,
						i, family, proto, port);
941 942
			if (error < 0)
				break;
L
Linus Torvalds 已提交
943 944 945
		}
	}

946 947 948
	return error;
}

C
Chuck Lever 已提交
949 950 951 952 953 954 955
/*
 * If user space is running rpcbind, it should take the v4 UNSET
 * and clear everything for this [program, version].  If user space
 * is running portmap, it will reject the v4 UNSET, but won't have
 * any "inet6" entries anyway.  So a PMAP_UNSET should be sufficient
 * in this case to clear all existing entries for [program, version].
 */
C
Chuck Lever 已提交
956 957 958 959 960
static void __svc_unregister(const u32 program, const u32 version,
			     const char *progname)
{
	int error;

C
Chuck Lever 已提交
961
	error = rpcb_v4_register(program, version, NULL, "");
C
Chuck Lever 已提交
962

C
Chuck Lever 已提交
963 964 965 966 967 968
	/*
	 * User space didn't support rpcbind v4, so retry this
	 * request with the legacy rpcbind v2 protocol.
	 */
	if (error == -EPROTONOSUPPORT)
		error = rpcb_register(program, version, 0, 0);
C
Chuck Lever 已提交
969 970 971 972 973

	dprintk("svc: %s(%sv%u), error %d\n",
			__func__, progname, version, error);
}

974
/*
C
Chuck Lever 已提交
975 976 977
 * All netids, bind addresses and ports registered for [program, version]
 * are removed from the local rpcbind database (if the service is not
 * hidden) to make way for a new instance of the service.
978
 *
C
Chuck Lever 已提交
979 980
 * The result of unregistration is reported via dprintk for those who want
 * verification of the result, but is otherwise not important.
981 982 983 984 985 986 987 988 989 990 991 992 993 994 995 996
 */
static void svc_unregister(const struct svc_serv *serv)
{
	struct svc_program *progp;
	unsigned long flags;
	unsigned int i;

	clear_thread_flag(TIF_SIGPENDING);

	for (progp = serv->sv_program; progp; progp = progp->pg_next) {
		for (i = 0; i < progp->pg_nvers; i++) {
			if (progp->pg_vers[i] == NULL)
				continue;
			if (progp->pg_vers[i]->vs_hidden)
				continue;

997 998
			dprintk("svc: attempting to unregister %sv%u\n",
				progp->pg_name, i);
C
Chuck Lever 已提交
999
			__svc_unregister(progp->pg_prog, i, progp->pg_name);
1000
		}
L
Linus Torvalds 已提交
1001 1002
	}

1003 1004 1005
	spin_lock_irqsave(&current->sighand->siglock, flags);
	recalc_sigpending();
	spin_unlock_irqrestore(&current->sighand->siglock, flags);
L
Linus Torvalds 已提交
1006 1007
}

1008 1009 1010
/*
 * Printk the given error with the address of the client that caused it.
 */
1011 1012
static __printf(2, 3)
int svc_printk(struct svc_rqst *rqstp, const char *fmt, ...)
1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030
{
	va_list args;
	int 	r;
	char 	buf[RPC_MAX_ADDRBUFLEN];

	if (!net_ratelimit())
		return 0;

	printk(KERN_WARNING "svc: %s: ",
		svc_print_addr(rqstp, buf, sizeof(buf)));

	va_start(args, fmt);
	r = vprintk(fmt, args);
	va_end(args);

	return r;
}

L
Linus Torvalds 已提交
1031
/*
R
Ricardo Labiaga 已提交
1032
 * Common routine for processing the RPC request.
L
Linus Torvalds 已提交
1033
 */
R
Ricardo Labiaga 已提交
1034 1035
static int
svc_process_common(struct svc_rqst *rqstp, struct kvec *argv, struct kvec *resv)
L
Linus Torvalds 已提交
1036 1037 1038 1039
{
	struct svc_program	*progp;
	struct svc_version	*versp = NULL;	/* compiler food */
	struct svc_procedure	*procp = NULL;
1040
	struct svc_serv		*serv = rqstp->rq_server;
L
Linus Torvalds 已提交
1041
	kxdrproc_t		xdr;
1042
	__be32			*statp;
R
Ricardo Labiaga 已提交
1043
	u32			prog, vers, proc;
1044
	__be32			auth_stat, rpc_stat;
L
Linus Torvalds 已提交
1045
	int			auth_res;
1046
	__be32			*reply_statp;
L
Linus Torvalds 已提交
1047 1048 1049 1050 1051 1052

	rpc_stat = rpc_success;

	if (argv->iov_len < 6*4)
		goto err_short_len;

1053
	/* Will be turned off only in gss privacy case: */
1054
	rqstp->rq_splice_ok = 1;
1055 1056
	/* Will be turned off only when NFSv4 Sessions are used */
	rqstp->rq_usedeferral = 1;
1057
	rqstp->rq_dropme = false;
T
Tom Tucker 已提交
1058 1059 1060

	/* Setup reply header */
	rqstp->rq_xprt->xpt_ops->xpo_prep_reply_hdr(rqstp);
L
Linus Torvalds 已提交
1061 1062 1063

	svc_putu32(resv, rqstp->rq_xid);

A
Alexey Dobriyan 已提交
1064
	vers = svc_getnl(argv);
L
Linus Torvalds 已提交
1065 1066

	/* First words of reply: */
A
Alexey Dobriyan 已提交
1067
	svc_putnl(resv, 1);		/* REPLY */
L
Linus Torvalds 已提交
1068 1069 1070 1071 1072

	if (vers != 2)		/* RPC version number */
		goto err_bad_rpc;

	/* Save position in case we later decide to reject: */
1073
	reply_statp = resv->iov_base + resv->iov_len;
L
Linus Torvalds 已提交
1074

A
Alexey Dobriyan 已提交
1075
	svc_putnl(resv, 0);		/* ACCEPT */
L
Linus Torvalds 已提交
1076

A
Alexey Dobriyan 已提交
1077 1078 1079
	rqstp->rq_prog = prog = svc_getnl(argv);	/* program number */
	rqstp->rq_vers = vers = svc_getnl(argv);	/* version number */
	rqstp->rq_proc = proc = svc_getnl(argv);	/* procedure number */
L
Linus Torvalds 已提交
1080 1081

	progp = serv->sv_program;
1082 1083 1084 1085 1086

	for (progp = serv->sv_program; progp; progp = progp->pg_next)
		if (prog == progp->pg_prog)
			break;

L
Linus Torvalds 已提交
1087 1088 1089 1090 1091 1092 1093
	/*
	 * Decode auth data, and add verifier to reply buffer.
	 * We do this before anything else in order to get a decent
	 * auth verifier.
	 */
	auth_res = svc_authenticate(rqstp, &auth_stat);
	/* Also give the program a chance to reject this call: */
1094
	if (auth_res == SVC_OK && progp) {
L
Linus Torvalds 已提交
1095 1096 1097 1098 1099 1100 1101
		auth_stat = rpc_autherr_badcred;
		auth_res = progp->pg_authenticate(rqstp);
	}
	switch (auth_res) {
	case SVC_OK:
		break;
	case SVC_GARBAGE:
1102
		goto err_garbage;
L
Linus Torvalds 已提交
1103 1104 1105 1106 1107
	case SVC_SYSERR:
		rpc_stat = rpc_system_err;
		goto err_bad;
	case SVC_DENIED:
		goto err_bad_auth;
1108 1109 1110
	case SVC_CLOSE:
		if (test_bit(XPT_TEMP, &rqstp->rq_xprt->xpt_flags))
			svc_close_xprt(rqstp->rq_xprt);
L
Linus Torvalds 已提交
1111 1112 1113 1114 1115
	case SVC_DROP:
		goto dropit;
	case SVC_COMPLETE:
		goto sendit;
	}
1116

1117
	if (progp == NULL)
L
Linus Torvalds 已提交
1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133
		goto err_bad_prog;

	if (vers >= progp->pg_nvers ||
	  !(versp = progp->pg_vers[vers]))
		goto err_bad_vers;

	procp = versp->vs_proc + proc;
	if (proc >= versp->vs_nproc || !procp->pc_func)
		goto err_bad_proc;
	rqstp->rq_procinfo = procp;

	/* Syntactic check complete */
	serv->sv_stats->rpccnt++;

	/* Build the reply header. */
	statp = resv->iov_base +resv->iov_len;
A
Alexey Dobriyan 已提交
1134
	svc_putnl(resv, RPC_SUCCESS);
L
Linus Torvalds 已提交
1135 1136 1137 1138 1139 1140 1141 1142

	/* Bump per-procedure stats counter */
	procp->pc_count++;

	/* Initialize storage for argp and resp */
	memset(rqstp->rq_argp, 0, procp->pc_argsize);
	memset(rqstp->rq_resp, 0, procp->pc_ressize);

1143
	/* un-reserve some of the out-queue now that we have a
L
Linus Torvalds 已提交
1144 1145 1146
	 * better idea of reply size
	 */
	if (procp->pc_xdrressize)
1147
		svc_reserve_auth(rqstp, procp->pc_xdrressize<<2);
L
Linus Torvalds 已提交
1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158

	/* Call the function that processes the request. */
	if (!versp->vs_dispatch) {
		/* Decode arguments */
		xdr = procp->pc_decode;
		if (xdr && !xdr(rqstp, argv->iov_base, rqstp->rq_argp))
			goto err_garbage;

		*statp = procp->pc_func(rqstp, rqstp->rq_argp, rqstp->rq_resp);

		/* Encode reply */
1159
		if (rqstp->rq_dropme) {
1160 1161 1162 1163
			if (procp->pc_release)
				procp->pc_release(rqstp, NULL, rqstp->rq_resp);
			goto dropit;
		}
1164 1165 1166
		if (*statp == rpc_success &&
		    (xdr = procp->pc_encode) &&
		    !xdr(rqstp, resv->iov_base+resv->iov_len, rqstp->rq_resp)) {
L
Linus Torvalds 已提交
1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 1191 1192 1193 1194
			dprintk("svc: failed to encode reply\n");
			/* serv->sv_stats->rpcsystemerr++; */
			*statp = rpc_system_err;
		}
	} else {
		dprintk("svc: calling dispatcher\n");
		if (!versp->vs_dispatch(rqstp, statp)) {
			/* Release reply info */
			if (procp->pc_release)
				procp->pc_release(rqstp, NULL, rqstp->rq_resp);
			goto dropit;
		}
	}

	/* Check RPC status result */
	if (*statp != rpc_success)
		resv->iov_len = ((void*)statp)  - resv->iov_base + 4;

	/* Release reply info */
	if (procp->pc_release)
		procp->pc_release(rqstp, NULL, rqstp->rq_resp);

	if (procp->pc_encode == NULL)
		goto dropit;

 sendit:
	if (svc_authorise(rqstp))
		goto dropit;
R
Ricardo Labiaga 已提交
1195
	return 1;		/* Caller can now send it */
L
Linus Torvalds 已提交
1196 1197 1198 1199 1200 1201 1202

 dropit:
	svc_authorise(rqstp);	/* doesn't hurt to call this twice */
	dprintk("svc: svc_process dropit\n");
	return 0;

err_short_len:
1203 1204
	svc_printk(rqstp, "short len %Zd, dropping request\n",
			argv->iov_len);
1205

L
Linus Torvalds 已提交
1206 1207 1208 1209
	goto dropit;			/* drop request */

err_bad_rpc:
	serv->sv_stats->rpcbadfmt++;
A
Alexey Dobriyan 已提交
1210 1211 1212 1213
	svc_putnl(resv, 1);	/* REJECT */
	svc_putnl(resv, 0);	/* RPC_MISMATCH */
	svc_putnl(resv, 2);	/* Only RPCv2 supported */
	svc_putnl(resv, 2);
L
Linus Torvalds 已提交
1214 1215 1216 1217 1218 1219
	goto sendit;

err_bad_auth:
	dprintk("svc: authentication failed (%d)\n", ntohl(auth_stat));
	serv->sv_stats->rpcbadauth++;
	/* Restore write pointer to location of accept status: */
1220
	xdr_ressize_check(rqstp, reply_statp);
A
Alexey Dobriyan 已提交
1221 1222 1223
	svc_putnl(resv, 1);	/* REJECT */
	svc_putnl(resv, 1);	/* AUTH_ERROR */
	svc_putnl(resv, ntohl(auth_stat));	/* status */
L
Linus Torvalds 已提交
1224 1225 1226
	goto sendit;

err_bad_prog:
1227
	dprintk("svc: unknown program %d\n", prog);
L
Linus Torvalds 已提交
1228
	serv->sv_stats->rpcbadfmt++;
A
Alexey Dobriyan 已提交
1229
	svc_putnl(resv, RPC_PROG_UNAVAIL);
L
Linus Torvalds 已提交
1230 1231 1232
	goto sendit;

err_bad_vers:
1233
	svc_printk(rqstp, "unknown version (%d for prog %d, %s)\n",
1234 1235
		       vers, prog, progp->pg_name);

L
Linus Torvalds 已提交
1236
	serv->sv_stats->rpcbadfmt++;
A
Alexey Dobriyan 已提交
1237 1238 1239
	svc_putnl(resv, RPC_PROG_MISMATCH);
	svc_putnl(resv, progp->pg_lovers);
	svc_putnl(resv, progp->pg_hivers);
L
Linus Torvalds 已提交
1240 1241 1242
	goto sendit;

err_bad_proc:
1243
	svc_printk(rqstp, "unknown procedure (%d)\n", proc);
1244

L
Linus Torvalds 已提交
1245
	serv->sv_stats->rpcbadfmt++;
A
Alexey Dobriyan 已提交
1246
	svc_putnl(resv, RPC_PROC_UNAVAIL);
L
Linus Torvalds 已提交
1247 1248 1249
	goto sendit;

err_garbage:
1250
	svc_printk(rqstp, "failed to decode args\n");
1251

L
Linus Torvalds 已提交
1252 1253 1254
	rpc_stat = rpc_garbage_args;
err_bad:
	serv->sv_stats->rpcbadfmt++;
A
Alexey Dobriyan 已提交
1255
	svc_putnl(resv, ntohl(rpc_stat));
L
Linus Torvalds 已提交
1256 1257
	goto sendit;
}
1258
EXPORT_SYMBOL_GPL(svc_process);
1259

R
Ricardo Labiaga 已提交
1260 1261 1262 1263 1264 1265 1266 1267 1268 1269 1270 1271 1272 1273 1274 1275 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285 1286 1287 1288 1289 1290 1291 1292 1293 1294 1295 1296
/*
 * Process the RPC request.
 */
int
svc_process(struct svc_rqst *rqstp)
{
	struct kvec		*argv = &rqstp->rq_arg.head[0];
	struct kvec		*resv = &rqstp->rq_res.head[0];
	struct svc_serv		*serv = rqstp->rq_server;
	u32			dir;

	/*
	 * Setup response xdr_buf.
	 * Initially it has just one page
	 */
	rqstp->rq_resused = 1;
	resv->iov_base = page_address(rqstp->rq_respages[0]);
	resv->iov_len = 0;
	rqstp->rq_res.pages = rqstp->rq_respages + 1;
	rqstp->rq_res.len = 0;
	rqstp->rq_res.page_base = 0;
	rqstp->rq_res.page_len = 0;
	rqstp->rq_res.buflen = PAGE_SIZE;
	rqstp->rq_res.tail[0].iov_base = NULL;
	rqstp->rq_res.tail[0].iov_len = 0;

	rqstp->rq_xid = svc_getu32(argv);

	dir  = svc_getnl(argv);
	if (dir != 0) {
		/* direction != CALL */
		svc_printk(rqstp, "bad direction %d, dropping request\n", dir);
		serv->sv_stats->rpcbadfmt++;
		svc_drop(rqstp);
		return 0;
	}

1297 1298 1299 1300 1301 1302 1303
	/* Returns 1 for send, 0 for drop */
	if (svc_process_common(rqstp, argv, resv))
		return svc_send(rqstp);
	else {
		svc_drop(rqstp);
		return 0;
	}
R
Ricardo Labiaga 已提交
1304 1305
}

1306
#if defined(CONFIG_SUNRPC_BACKCHANNEL)
1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318
/*
 * Process a backchannel RPC request that arrived over an existing
 * outbound connection
 */
int
bc_svc_process(struct svc_serv *serv, struct rpc_rqst *req,
	       struct svc_rqst *rqstp)
{
	struct kvec	*argv = &rqstp->rq_arg.head[0];
	struct kvec	*resv = &rqstp->rq_res.head[0];

	/* Build the svc_rqst used by the common processing routine */
1319
	rqstp->rq_xprt = serv->sv_bc_xprt;
1320 1321 1322 1323 1324 1325 1326 1327 1328 1329 1330 1331 1332 1333 1334 1335 1336 1337 1338 1339 1340 1341 1342 1343
	rqstp->rq_xid = req->rq_xid;
	rqstp->rq_prot = req->rq_xprt->prot;
	rqstp->rq_server = serv;

	rqstp->rq_addrlen = sizeof(req->rq_xprt->addr);
	memcpy(&rqstp->rq_addr, &req->rq_xprt->addr, rqstp->rq_addrlen);
	memcpy(&rqstp->rq_arg, &req->rq_rcv_buf, sizeof(rqstp->rq_arg));
	memcpy(&rqstp->rq_res, &req->rq_snd_buf, sizeof(rqstp->rq_res));

	/* reset result send buffer "put" position */
	resv->iov_len = 0;

	if (rqstp->rq_prot != IPPROTO_TCP) {
		printk(KERN_ERR "No support for Non-TCP transports!\n");
		BUG();
	}

	/*
	 * Skip the next two words because they've already been
	 * processed in the trasport
	 */
	svc_getu32(argv);	/* XID */
	svc_getnl(argv);	/* CALLDIR */

1344 1345 1346 1347 1348 1349 1350 1351 1352
	/* Returns 1 for send, 0 for drop */
	if (svc_process_common(rqstp, argv, resv)) {
		memcpy(&req->rq_snd_buf, &rqstp->rq_res,
						sizeof(req->rq_snd_buf));
		return bc_send(req);
	} else {
		/* Nothing to do to drop request */
		return 0;
	}
1353
}
1354
EXPORT_SYMBOL_GPL(bc_svc_process);
1355
#endif /* CONFIG_SUNRPC_BACKCHANNEL */
1356

1357 1358 1359 1360 1361
/*
 * Return (transport-specific) limit on the rpc payload.
 */
u32 svc_max_payload(const struct svc_rqst *rqstp)
{
1362
	u32 max = rqstp->rq_xprt->xpt_class->xcl_max_payload;
1363

1364 1365
	if (rqstp->rq_server->sv_max_payload < max)
		max = rqstp->rq_server->sv_max_payload;
1366 1367 1368
	return max;
}
EXPORT_SYMBOL_GPL(svc_max_payload);