summaryrefslogtreecommitdiff
path: root/sys/nfs/nfs_syscalls.c
blob: 966c0ce2af1aff778112c2e5ddab5e1ac7a60906 (plain)
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
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
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
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
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
181
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
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
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
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
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
372
373
374
375
376
377
378
379
380
381
382
383
384
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
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
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
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
/*	$OpenBSD: nfs_syscalls.c,v 1.84 2009/07/20 16:49:40 thib Exp $	*/
/*	$NetBSD: nfs_syscalls.c,v 1.19 1996/02/18 11:53:52 fvdl Exp $	*/

/*
 * Copyright (c) 1989, 1993
 *	The Regents of the University of California.  All rights reserved.
 *
 * This code is derived from software contributed to Berkeley by
 * Rick Macklem at The University of Guelph.
 *
 * 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 BY THE REGENTS AND CONTRIBUTORS ``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.
 *
 *	@(#)nfs_syscalls.c	8.5 (Berkeley) 3/30/95
 */

#include <sys/param.h>
#include <sys/systm.h>
#include <sys/kernel.h>
#include <sys/file.h>
#include <sys/stat.h>
#include <sys/vnode.h>
#include <sys/mount.h>
#include <sys/proc.h>
#include <sys/uio.h>
#include <sys/malloc.h>
#include <sys/buf.h>
#include <sys/mbuf.h>
#include <sys/socket.h>
#include <sys/socketvar.h>
#include <sys/domain.h>
#include <sys/protosw.h>
#include <sys/namei.h>
#include <sys/syslog.h>
#include <sys/filedesc.h>
#include <sys/signalvar.h>
#include <sys/kthread.h>
#include <sys/queue.h>

#include <sys/syscallargs.h>

#include <netinet/in.h>
#include <netinet/tcp.h>
#include <nfs/xdr_subs.h>
#include <nfs/rpcv2.h>
#include <nfs/nfsproto.h>
#include <nfs/nfs.h>
#include <nfs/nfsm_subs.h>
#include <nfs/nfsrvcache.h>
#include <nfs/nfsmount.h>
#include <nfs/nfsnode.h>
#include <nfs/nfs_var.h>

/* Global defs. */
extern int nfs_numasync;
extern struct nfsstats nfsstats;
extern int nfsrvw_procrastinate;
extern struct timeval nfsrvw_procrastinate_tv;
struct nfssvc_sock *nfs_udpsock;
int nfsd_waiting = 0;

#ifdef NFSSERVER
struct pool nfsrv_descript_pl;

int nfsrv_getslp(struct nfsd *nfsd);

static int nfs_numnfsd = 0;
int (*nfsrv3_procs[NFS_NPROCS])(struct nfsrv_descript *,
    struct nfssvc_sock *, struct proc *, struct mbuf **) = {
	nfsrv_null,
	nfsrv_getattr,
	nfsrv_setattr,
	nfsrv_lookup,
	nfsrv3_access,
	nfsrv_readlink,
	nfsrv_read,
	nfsrv_write,
	nfsrv_create,
	nfsrv_mkdir,
	nfsrv_symlink,
	nfsrv_mknod,
	nfsrv_remove,
	nfsrv_rmdir,
	nfsrv_rename,
	nfsrv_link,
	nfsrv_readdir,
	nfsrv_readdirplus,
	nfsrv_statfs,
	nfsrv_fsinfo,
	nfsrv_pathconf,
	nfsrv_commit,
	nfsrv_noop
};
#endif

struct nfssvc_sockhead nfssvc_sockhead;
struct nfsdhead nfsd_head;

int nfssvc_sockhead_flag;
int nfsd_head_flag;

#ifdef NFSCLIENT
struct proc *nfs_asyncdaemon[NFS_MAXASYNCDAEMON];
int nfs_niothreads = -1;
#endif

/*
 * NFS server pseudo system call for the nfsd's
 * Based on the flag value it either:
 * - adds a socket to the selection list
 * - remains in the kernel as an nfsd
 */
int
sys_nfssvc(struct proc *p, void *v, register_t *retval)
{
	int error = 0;
#ifdef NFSSERVER
	struct sys_nfssvc_args /* {
		syscallarg(int) flag;
		syscallarg(caddr_t) argp;
	} */ *uap = v;
	int flags = SCARG(uap, flag);
	struct file *fp;
	struct mbuf *nam;
	struct nfsd_args nfsdarg;
	struct nfsd_srvargs nfsd_srvargs, *nsd = &nfsd_srvargs;
	struct nfsd *nfsd;
#endif

	/* Must be super user */
	error = suser(p, 0);
	if (error)
		return (error);

#ifndef NFSSERVER
	error = ENOSYS;
#else

	while (nfssvc_sockhead_flag & SLP_INIT) {
		nfssvc_sockhead_flag |= SLP_WANTINIT;
		tsleep(&nfssvc_sockhead, PSOCK, "nfsd init", 0);
	}

	switch (flags) {
	case NFSSVC_ADDSOCK:
		error = copyin(SCARG(uap, argp), &nfsdarg, sizeof(nfsdarg));
		if (error)
			return (error);

		error = getsock(p->p_fd, nfsdarg.sock, &fp);
		if (error)
			return (error);

		/*
		 * Get the client address for connected sockets.
		 */
		if (nfsdarg.name == NULL || nfsdarg.namelen == 0)
			nam = NULL;
		else {
			error = sockargs(&nam, nfsdarg.name, nfsdarg.namelen,
				MT_SONAME);
			if (error) {
				FRELE(fp);
				return (error);
			}
		}
		error = nfssvc_addsock(fp, nam);
		FRELE(fp);
		break;
	case NFSSVC_NFSD:
		error = copyin(SCARG(uap, argp), nsd, sizeof(*nsd));
		if (error)
			return (error);

		nfsd = malloc(sizeof(*nfsd), M_NFSD, M_WAITOK|M_ZERO);
		nfsd->nfsd_procp = p;
		nfsd->nfsd_slp = NULL;

		error = nfssvc_nfsd(nfsd);
		break;
	default:
		error = EINVAL;
		break;
	}

	if (error == EINTR || error == ERESTART)
		error = 0;
#endif	/* !NFSSERVER */

	return (error);
}

#ifdef NFSSERVER
/*
 * Adds a socket to the list for servicing by nfsds.
 */
int
nfssvc_addsock(fp, mynam)
	struct file *fp;
	struct mbuf *mynam;
{
	struct mbuf *m;
	int siz;
	struct nfssvc_sock *slp;
	struct socket *so;
	struct nfssvc_sock *tslp;
	int error, s;

	so = (struct socket *)fp->f_data;
	tslp = NULL;
	/*
	 * Add it to the list, as required.
	 */
	if (so->so_proto->pr_protocol == IPPROTO_UDP) {
		tslp = nfs_udpsock;
		if (tslp->ns_flag & SLP_VALID) {
			m_freem(mynam);
			return (EPERM);
		}
	}
	if (so->so_type == SOCK_STREAM)
		siz = NFS_MAXPACKET + sizeof (u_long);
	else
		siz = NFS_MAXPACKET;
	error = soreserve(so, siz, siz); 
	if (error) {
		m_freem(mynam);
		return (error);
	}

	/*
	 * Set protocol specific options { for now TCP only } and
	 * reserve some space. For datagram sockets, this can get called
	 * repeatedly for the same socket, but that isn't harmful.
	 */
	if (so->so_type == SOCK_STREAM) {
		MGET(m, M_WAIT, MT_SOOPTS);
		*mtod(m, int32_t *) = 1;
		m->m_len = sizeof(int32_t);
		sosetopt(so, SOL_SOCKET, SO_KEEPALIVE, m);
	}
	if (so->so_proto->pr_domain->dom_family == AF_INET &&
	    so->so_proto->pr_protocol == IPPROTO_TCP) {
		MGET(m, M_WAIT, MT_SOOPTS);
		*mtod(m, int32_t *) = 1;
		m->m_len = sizeof(int32_t);
		sosetopt(so, IPPROTO_TCP, TCP_NODELAY, m);
	}
	so->so_rcv.sb_flags &= ~SB_NOINTR;
	so->so_rcv.sb_timeo = 0;
	so->so_snd.sb_flags &= ~SB_NOINTR;
	so->so_snd.sb_timeo = 0;
	if (tslp)
		slp = tslp;
	else {
		slp = malloc(sizeof(struct nfssvc_sock), M_NFSSVC,
		    M_WAITOK|M_ZERO);
		TAILQ_INSERT_TAIL(&nfssvc_sockhead, slp, ns_chain);
	}
	slp->ns_so = so;
	slp->ns_nam = mynam;
	fp->f_count++;
	slp->ns_fp = fp;
	s = splsoftnet();
	so->so_upcallarg = (caddr_t)slp;
	so->so_upcall = nfsrv_rcv;
	slp->ns_flag = (SLP_VALID | SLP_NEEDQ);
	nfsrv_wakenfsd(slp);
	splx(s);
	return (0);
}

/*
 * Called by nfssvc() for nfsds. Just loops around servicing rpc requests
 * until it is killed by a signal.
 */
int
nfssvc_nfsd(struct nfsd *nfsd)
{
	struct mbuf *m;
	int siz;
	struct nfssvc_sock *slp;
	struct socket *so;
	int *solockp;
	struct nfsrv_descript *nd = NULL;
	struct mbuf *mreq;
	int error = 0, cacherep, s, sotype, writes_todo;
	struct timeval tv;

	cacherep = RC_DOIT;
	writes_todo = 0;

	s = splsoftnet();
	TAILQ_INSERT_TAIL(&nfsd_head, nfsd, nfsd_chain);
	nfs_numnfsd++;
	/*
	 * Loop getting rpc requests until SIGKILL.
	 */
	for (;;) {
		if ((nfsd->nfsd_flag & NFSD_REQINPROG) == 0) {

			/* attach an nfssvc_sock to nfsd */
			error = nfsrv_getslp(nfsd);
			if (error)
				goto done;
			slp = nfsd->nfsd_slp;

			if (slp->ns_flag & SLP_VALID) {
				if (slp->ns_flag & SLP_DISCONN)
					nfsrv_zapsock(slp);
				else if (slp->ns_flag & SLP_NEEDQ) {
					slp->ns_flag &= ~SLP_NEEDQ;
					(void) nfs_sndlock(&slp->ns_solock,
						NULL);
					nfsrv_rcv(slp->ns_so, (caddr_t)slp,
						M_WAIT);
					nfs_sndunlock(&slp->ns_solock);
				}
				error = nfsrv_dorec(slp, nfsd, &nd);
				getmicrotime(&tv);
				if (error && LIST_FIRST(&slp->ns_tq) &&
				    timercmp(&LIST_FIRST(&slp->ns_tq)->nd_time,
				    &tv, <=)) {
					error = 0;
					cacherep = RC_DOIT;
					writes_todo = 1;
				} else
					writes_todo = 0;
				nfsd->nfsd_flag |= NFSD_REQINPROG;
			}
		} else {
			error = 0;
			slp = nfsd->nfsd_slp;
		}
		if (error || (slp->ns_flag & SLP_VALID) == 0) {
			if (nd) {
				pool_put(&nfsrv_descript_pl, nd);
				nd = NULL;
			}
			nfsd->nfsd_slp = NULL;
			nfsd->nfsd_flag &= ~NFSD_REQINPROG;
			nfsrv_slpderef(slp);
			continue;
		}
		splx(s);
		so = slp->ns_so;
		sotype = so->so_type;
		if (so->so_proto->pr_flags & PR_CONNREQUIRED)
			solockp = &slp->ns_solock;
		else
			solockp = NULL;
		if (nd) {
		    if (nd->nd_nam2)
			nd->nd_nam = nd->nd_nam2;
		    else
			nd->nd_nam = slp->ns_nam;

		    cacherep = nfsrv_getcache(nd, slp, &mreq);
		}

		/*
		 * Loop to get all the write rpc relies that have been
		 * gathered together.
		 */
		do {
		    switch (cacherep) {
		    case RC_DOIT:
			if (writes_todo || (!(nd->nd_flag & ND_NFSV3) &&
			    nd->nd_procnum == NFSPROC_WRITE &&
			    nfsrvw_procrastinate > 0))
				error = nfsrv_writegather(&nd, slp,
				    nfsd->nfsd_procp, &mreq);
			else
				error = (*(nfsrv3_procs[nd->nd_procnum]))(nd,
				    slp, nfsd->nfsd_procp, &mreq);
			if (mreq == NULL) {
				if (nd != NULL) {
					m_freem(nd->nd_nam2);
					m_freem(nd->nd_mrep);
				}
				break;
			}
			if (error) {
				nfsstats.srv_errs++;
				nfsrv_updatecache(nd, 0, mreq);
				if (nd->nd_nam2)
					m_freem(nd->nd_nam2);
				break;
			}
			nfsstats.srvrpccnt[nd->nd_procnum]++;
			nfsrv_updatecache(nd, 1, mreq);
			nd->nd_mrep = NULL;

			/* FALLTHROUGH */
		    case RC_REPLY:
			m = mreq;
			siz = 0;
			while (m) {
				siz += m->m_len;
				m = m->m_next;
			}
			if (siz <= 0 || siz > NFS_MAXPACKET) {
				printf("mbuf siz=%d\n",siz);
				panic("Bad nfs svc reply");
			}
			m = mreq;
			m->m_pkthdr.len = siz;
			m->m_pkthdr.rcvif = NULL;
			/*
			 * For stream protocols, prepend a Sun RPC
			 * Record Mark.
			 */
			if (sotype == SOCK_STREAM) {
				M_PREPEND(m, NFSX_UNSIGNED, M_WAIT);
				*mtod(m, u_int32_t *) = htonl(0x80000000 | siz);
			}
			if (solockp)
				(void) nfs_sndlock(solockp, NULL);
			if (slp->ns_flag & SLP_VALID)
			    error = nfs_send(so, nd->nd_nam2, m, NULL);
			else {
			    error = EPIPE;
			    m_freem(m);
			}
			if (nd->nd_nam2)
				m_freem(nd->nd_nam2);
			if (nd->nd_mrep)
				m_freem(nd->nd_mrep);
			if (error == EPIPE)
				nfsrv_zapsock(slp);
			if (solockp)
				nfs_sndunlock(solockp);
			if (error == EINTR || error == ERESTART) {
		    		pool_put(&nfsrv_descript_pl, nd);
				nfsrv_slpderef(slp);
				s = splsoftnet();
				goto done;
			}
			break;
		    case RC_DROPIT:
			m_freem(nd->nd_mrep);
			m_freem(nd->nd_nam2);
			break;
		    };
		    if (nd) {
		    	pool_put(&nfsrv_descript_pl, nd);
			nd = NULL;
		    }

		    /*
		     * Check to see if there are outstanding writes that
		     * need to be serviced.
		     */
		    getmicrotime(&tv);
		    s = splsoftclock();
		    if (LIST_FIRST(&slp->ns_tq) &&
			timercmp(&LIST_FIRST(&slp->ns_tq)->nd_time, &tv, <=)) {
			cacherep = RC_DOIT;
			writes_todo = 1;
		    } else
			writes_todo = 0;
		    splx(s);
		} while (writes_todo);
		s = splsoftnet();
		if (nfsrv_dorec(slp, nfsd, &nd)) {
			nfsd->nfsd_flag &= ~NFSD_REQINPROG;
			nfsd->nfsd_slp = NULL;
			nfsrv_slpderef(slp);
		}
	}
done:
	TAILQ_REMOVE(&nfsd_head, nfsd, nfsd_chain);
	splx(s);
	free((caddr_t)nfsd, M_NFSD);
	if (--nfs_numnfsd == 0)
		nfsrv_init(1);	/* Reinitialize everything */
	return (error);
}

/*
 * Shut down a socket associated with an nfssvc_sock structure.
 * Should be called with the send lock set, if required.
 * The trick here is to increment the sref at the start, so that the nfsds
 * will stop using it and clear ns_flag at the end so that it will not be
 * reassigned during cleanup.
 */
void
nfsrv_zapsock(slp)
	struct nfssvc_sock *slp;
{
	struct nfsrv_descript *nwp, *nnwp;
	struct socket *so;
	struct file *fp;
	struct mbuf *m, *n;
	int s;

	slp->ns_flag &= ~SLP_ALLFLAGS;
	fp = slp->ns_fp;
	if (fp) {
		FREF(fp);
		slp->ns_fp = NULL;
		so = slp->ns_so;
		so->so_upcall = NULL;
		soshutdown(so, SHUT_RDWR);
		closef(fp, NULL);
		if (slp->ns_nam)
			MFREE(slp->ns_nam, m);
		m_freem(slp->ns_raw);
		m = slp->ns_rec;
		while (m) {
			n = m->m_nextpkt;
			m_freem(m);
			m = n;
		}
		s = splsoftclock();
		for (nwp = LIST_FIRST(&slp->ns_tq); nwp != NULL; nwp = nnwp) {
			nnwp = LIST_NEXT(nwp, nd_tq);
			LIST_REMOVE(nwp, nd_tq);
			pool_put(&nfsrv_descript_pl, nwp);
		}
		LIST_INIT(&slp->ns_tq);
		splx(s);
	}
}

/*
 * Derefence a server socket structure. If it has no more references and
 * is no longer valid, you can throw it away.
 */
void
nfsrv_slpderef(slp)
	struct nfssvc_sock *slp;
{
	if (--(slp->ns_sref) == 0 && (slp->ns_flag & SLP_VALID) == 0) {
		TAILQ_REMOVE(&nfssvc_sockhead, slp, ns_chain);
		free((caddr_t)slp, M_NFSSVC);
	}
}

/*
 * Initialize the data structures for the server.
 * Handshake with any new nfsds starting up to avoid any chance of
 * corruption.
 */
void
nfsrv_init(terminating)
	int terminating;
{
	struct nfssvc_sock *slp, *nslp;

	if (nfssvc_sockhead_flag & SLP_INIT)
		panic("nfsd init");
	nfssvc_sockhead_flag |= SLP_INIT;
	if (terminating) {
		for (slp = TAILQ_FIRST(&nfssvc_sockhead); slp != NULL;
		    slp = nslp) {
			nslp = TAILQ_NEXT(slp, ns_chain);
			if (slp->ns_flag & SLP_VALID)
				nfsrv_zapsock(slp);
			TAILQ_REMOVE(&nfssvc_sockhead, slp, ns_chain);
			free((caddr_t)slp, M_NFSSVC);
		}
		nfsrv_cleancache();	/* And clear out server cache */
	}

	TAILQ_INIT(&nfssvc_sockhead);
	nfssvc_sockhead_flag &= ~SLP_INIT;
	if (nfssvc_sockhead_flag & SLP_WANTINIT) {
		nfssvc_sockhead_flag &= ~SLP_WANTINIT;
		wakeup((caddr_t)&nfssvc_sockhead);
	}

	TAILQ_INIT(&nfsd_head);
	nfsd_head_flag &= ~NFSD_CHECKSLP;

	nfs_udpsock =  malloc(sizeof(struct nfssvc_sock), M_NFSSVC,
	    M_WAITOK|M_ZERO);
	TAILQ_INSERT_HEAD(&nfssvc_sockhead, nfs_udpsock, ns_chain);

	pool_init(&nfsrv_descript_pl, sizeof(struct nfsrv_descript), 0, 0, 0,
	    "ndscpl", &pool_allocator_nointr);
}
#endif /* NFSSERVER */

#ifdef NFSCLIENT
/*
 * Asynchronous I/O threads for client nfs.
 * They do read-ahead and write-behind operations on the block I/O cache.
 * Never returns unless it fails or gets killed.
 */
void
nfssvc_iod(void *arg)
{
	struct proc *p = (struct proc *)arg;
	struct buf *bp, *nbp;
	int i, myiod;
	struct vnode *vp;
	int error = 0, s, bufcount;

	bufcount = 256;	/* XXX: Big enough? sysctl, constant ? */

	/* Assign my position or return error if too many already running. */
	myiod = -1;
	for (i = 0; i < NFS_MAXASYNCDAEMON; i++) {
		if (nfs_asyncdaemon[i] == NULL) {
			myiod = i;
			break;
		}
	}
	if (myiod == -1)
		kthread_exit(EBUSY);

	nfs_asyncdaemon[myiod] = p;
	nfs_numasync++;

	/* Upper limit on how many bufs we'll queue up for this iod. */
	if (nfs_bufqmax > bcstats.numbufs / 4) {
		nfs_bufqmax = bcstats.numbufs / 4; /* limit to 1/4 of bufs */
		bufcount = 0;
	}

	nfs_bufqmax += bufcount;

	/* Just loop around doin our stuff until SIGKILL. */
	for (;;) {
	    while (TAILQ_FIRST(&nfs_bufq) == NULL && error == 0) {
		    error = tsleep(&nfs_bufq,
			PWAIT | PCATCH, "nfsidl", 0);
	    }
	    while ((bp = TAILQ_FIRST(&nfs_bufq)) != NULL) {
		/* Take one off the front of the list */
		TAILQ_REMOVE(&nfs_bufq, bp, b_freelist);
		nfs_bufqlen--;
		if (bp->b_flags & B_READ)
		    (void) nfs_doio(bp, NULL);
		else do {
		    /*
		     * Look for a delayed write for the same vnode, so I can do 
		     * it now. We must grab it before calling nfs_doio() to
		     * avoid any risk of the vnode getting vclean()'d while
		     * we are doing the write rpc.
		     */
		    vp = bp->b_vp;
		    s = splbio();
		    LIST_FOREACH(nbp, &vp->v_dirtyblkhd, b_vnbufs) {
			if ((nbp->b_flags &
			    (B_BUSY|B_DELWRI|B_NEEDCOMMIT|B_NOCACHE))!=B_DELWRI)
			    continue;
			nbp->b_flags |= B_ASYNC;
			bremfree(nbp);
			buf_acquire(nbp);
			break;
		    }
		    /*
		     * For the delayed write, do the first part of nfs_bwrite()
		     * up to, but not including nfs_strategy().
		     */
		    if (nbp) {
			nbp->b_flags &= ~(B_READ|B_DONE|B_ERROR);
			buf_undirty(nbp);
			nbp->b_vp->v_numoutput++;
		    }
		    splx(s);

		    (void) nfs_doio(bp, NULL);
		} while ((bp = nbp) != NULL);
	    }
	    if (error) {
		nfs_asyncdaemon[myiod] = NULL;
		nfs_numasync--;
		nfs_bufqmax -= bufcount;
		kthread_exit(error);
	    }
	}
}

void
nfs_getset_niothreads(set)
	int set;
{
	struct proc *p;
	int i, have, start;
	
	for (have = 0, i = 0; i < NFS_MAXASYNCDAEMON; i++)
		if (nfs_asyncdaemon[i] != NULL)
			have++;

	if (set) {
		/* clamp to sane range */
		nfs_niothreads = max(0, min(nfs_niothreads, NFS_MAXASYNCDAEMON));

		start = nfs_niothreads - have;

		while (start > 0) {
			kthread_create(nfssvc_iod, p, &p, "nfsio");
			start--;
		}

		for (i = 0; (start < 0) && (i < NFS_MAXASYNCDAEMON); i++)
			if (nfs_asyncdaemon[i] != NULL) {
				psignal(nfs_asyncdaemon[i], SIGKILL);
				start++;
			}
	} else {
		if (nfs_niothreads >= 0)
			nfs_niothreads = have;
	}
}
#endif /* NFSCLIENT */

#ifdef NFSSERVER
/*
 * Find an nfssrv_sock for nfsd, sleeping if needed.
 */
int
nfsrv_getslp(struct nfsd *nfsd)
{
	struct nfssvc_sock *slp;
	int error;

again:
	while (nfsd->nfsd_slp == NULL &&
	    (nfsd_head_flag & NFSD_CHECKSLP) == 0) {
		nfsd->nfsd_flag |= NFSD_WAITING;
		nfsd_waiting++;
		error = tsleep(nfsd, PSOCK | PCATCH, "nfsd", 0);
		nfsd_waiting--;
		if (error)
			return (error);
	}

	if (nfsd->nfsd_slp == NULL &&
	    (nfsd_head_flag & NFSD_CHECKSLP) != 0) {
		TAILQ_FOREACH(slp, &nfssvc_sockhead, ns_chain) {
			if ((slp->ns_flag & (SLP_VALID | SLP_DOREC)) ==
			    (SLP_VALID | SLP_DOREC)) {
				slp->ns_flag &= ~SLP_DOREC;
				slp->ns_sref++;
				nfsd->nfsd_slp = slp;
				break;
			}
		}
		if (slp == NULL)
			nfsd_head_flag &= ~NFSD_CHECKSLP;
	}

	if (nfsd->nfsd_slp == NULL)
		goto again;

	return (0);
}
#endif /* NFSSERVER */