aboutsummaryrefslogtreecommitdiff
path: root/sys/kern/kern_ktrace.c
blob: 98c08729731754ce9adaabbfdd727d4e64fd8905 (plain) (blame)
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
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
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
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
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
/*
 * Copyright (c) 1989, 1993
 *	The Regents of the University of California.  All rights reserved.
 *
 * 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.
 * 4. 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.
 *
 *	@(#)kern_ktrace.c	8.2 (Berkeley) 9/23/93
 */

#include <sys/cdefs.h>
__FBSDID("$FreeBSD$");

#include "opt_ktrace.h"
#include "opt_mac.h"

#include <sys/param.h>
#include <sys/systm.h>
#include <sys/fcntl.h>
#include <sys/kernel.h>
#include <sys/kthread.h>
#include <sys/lock.h>
#include <sys/mutex.h>
#include <sys/mac.h>
#include <sys/malloc.h>
#include <sys/namei.h>
#include <sys/proc.h>
#include <sys/unistd.h>
#include <sys/vnode.h>
#include <sys/ktrace.h>
#include <sys/sx.h>
#include <sys/sysctl.h>
#include <sys/syslog.h>
#include <sys/sysproto.h>

static MALLOC_DEFINE(M_KTRACE, "KTRACE", "KTRACE");

#ifdef KTRACE

#ifndef KTRACE_REQUEST_POOL
#define	KTRACE_REQUEST_POOL	100
#endif

struct ktr_request {
	struct	ktr_header ktr_header;
	struct	ucred *ktr_cred;
	struct	vnode *ktr_vp;
	union {
		struct	ktr_syscall ktr_syscall;
		struct	ktr_sysret ktr_sysret;
		struct	ktr_genio ktr_genio;
		struct	ktr_psig ktr_psig;
		struct	ktr_csw ktr_csw;
	} ktr_data;
	STAILQ_ENTRY(ktr_request) ktr_list;
};

static int data_lengths[] = {
	0,					/* none */
	offsetof(struct ktr_syscall, ktr_args),	/* KTR_SYSCALL */
	sizeof(struct ktr_sysret),		/* KTR_SYSRET */
	0,					/* KTR_NAMEI */
	sizeof(struct ktr_genio),		/* KTR_GENIO */
	sizeof(struct ktr_psig),		/* KTR_PSIG */
	sizeof(struct ktr_csw),			/* KTR_CSW */
	0					/* KTR_USER */
};

static STAILQ_HEAD(, ktr_request) ktr_todo;
static STAILQ_HEAD(, ktr_request) ktr_free;

SYSCTL_NODE(_kern, OID_AUTO, ktrace, CTLFLAG_RD, 0, "KTRACE options");

static u_int ktr_requestpool = KTRACE_REQUEST_POOL;
TUNABLE_INT("kern.ktrace.request_pool", &ktr_requestpool);

static u_int ktr_geniosize = PAGE_SIZE;
TUNABLE_INT("kern.ktrace.genio_size", &ktr_geniosize);
SYSCTL_UINT(_kern_ktrace, OID_AUTO, genio_size, CTLFLAG_RW, &ktr_geniosize,
    0, "Maximum size of genio event payload");

static int print_message = 1;
struct mtx ktrace_mtx;
static struct cv ktrace_cv;

static void ktrace_init(void *dummy);
static int sysctl_kern_ktrace_request_pool(SYSCTL_HANDLER_ARGS);
static u_int ktrace_resize_pool(u_int newsize);
static struct ktr_request *ktr_getrequest(int type);
static void ktr_submitrequest(struct ktr_request *req);
static void ktr_freerequest(struct ktr_request *req);
static void ktr_loop(void *dummy);
static void ktr_writerequest(struct ktr_request *req);
static int ktrcanset(struct thread *,struct proc *);
static int ktrsetchildren(struct thread *,struct proc *,int,int,struct vnode *);
static int ktrops(struct thread *,struct proc *,int,int,struct vnode *);

static void
ktrace_init(void *dummy)
{
	struct ktr_request *req;
	int i;

	mtx_init(&ktrace_mtx, "ktrace", NULL, MTX_DEF | MTX_QUIET);
	cv_init(&ktrace_cv, "ktrace");
	STAILQ_INIT(&ktr_todo);
	STAILQ_INIT(&ktr_free);
	for (i = 0; i < ktr_requestpool; i++) {
		req = malloc(sizeof(struct ktr_request), M_KTRACE, M_WAITOK);
		STAILQ_INSERT_HEAD(&ktr_free, req, ktr_list);
	}
	kthread_create(ktr_loop, NULL, NULL, RFHIGHPID, 0, "ktrace");
}
SYSINIT(ktrace_init, SI_SUB_KTRACE, SI_ORDER_ANY, ktrace_init, NULL);

static int
sysctl_kern_ktrace_request_pool(SYSCTL_HANDLER_ARGS)
{
	struct thread *td;
	u_int newsize, oldsize, wantsize;
	int error;

	/* Handle easy read-only case first to avoid warnings from GCC. */
	if (!req->newptr) {
		mtx_lock(&ktrace_mtx);
		oldsize = ktr_requestpool;
		mtx_unlock(&ktrace_mtx);
		return (SYSCTL_OUT(req, &oldsize, sizeof(u_int)));
	}

	error = SYSCTL_IN(req, &wantsize, sizeof(u_int));
	if (error)
		return (error);
	td = curthread;
	td->td_pflags |= TDP_INKTRACE;
	mtx_lock(&ktrace_mtx);
	oldsize = ktr_requestpool;
	newsize = ktrace_resize_pool(wantsize);
	mtx_unlock(&ktrace_mtx);
	td->td_pflags &= ~TDP_INKTRACE;
	error = SYSCTL_OUT(req, &oldsize, sizeof(u_int));
	if (error)
		return (error);
	if (wantsize > oldsize && newsize < wantsize)
		return (ENOSPC);
	return (0);
}
SYSCTL_PROC(_kern_ktrace, OID_AUTO, request_pool, CTLTYPE_UINT|CTLFLAG_RW,
    &ktr_requestpool, 0, sysctl_kern_ktrace_request_pool, "IU", "");

static u_int
ktrace_resize_pool(u_int newsize)
{
	struct ktr_request *req;
	int bound;

	mtx_assert(&ktrace_mtx, MA_OWNED);
	print_message = 1;
	bound = newsize - ktr_requestpool;
	if (bound == 0)
		return (ktr_requestpool);
	if (bound < 0)
		/* Shrink pool down to newsize if possible. */
		while (bound++ < 0) {
			req = STAILQ_FIRST(&ktr_free);
			if (req == NULL)
				return (ktr_requestpool);
			STAILQ_REMOVE_HEAD(&ktr_free, ktr_list);
			ktr_requestpool--;
			mtx_unlock(&ktrace_mtx);
			free(req, M_KTRACE);
			mtx_lock(&ktrace_mtx);
		}
	else
		/* Grow pool up to newsize. */
		while (bound-- > 0) {
			mtx_unlock(&ktrace_mtx);
			req = malloc(sizeof(struct ktr_request), M_KTRACE,
			    M_WAITOK);
			mtx_lock(&ktrace_mtx);
			STAILQ_INSERT_HEAD(&ktr_free, req, ktr_list);
			ktr_requestpool++;
		}
	return (ktr_requestpool);
}

static struct ktr_request *
ktr_getrequest(int type)
{
	struct ktr_request *req;
	struct thread *td = curthread;
	struct proc *p = td->td_proc;
	int pm;

	td->td_pflags |= TDP_INKTRACE;
	mtx_lock(&ktrace_mtx);
	if (!KTRCHECK(td, type)) {
		mtx_unlock(&ktrace_mtx);
		td->td_pflags &= ~TDP_INKTRACE;
		return (NULL);
	}
	req = STAILQ_FIRST(&ktr_free);
	if (req != NULL) {
		STAILQ_REMOVE_HEAD(&ktr_free, ktr_list);
		req->ktr_header.ktr_type = type;
		if (p->p_traceflag & KTRFAC_DROP) {
			req->ktr_header.ktr_type |= KTR_DROP;
			p->p_traceflag &= ~KTRFAC_DROP;
		}
		KASSERT(p->p_tracevp != NULL, ("ktrace: no trace vnode"));
		KASSERT(p->p_tracecred != NULL, ("ktrace: no trace cred"));
		req->ktr_vp = p->p_tracevp;
		VREF(p->p_tracevp);
		req->ktr_cred = crhold(p->p_tracecred);
		mtx_unlock(&ktrace_mtx);
		microtime(&req->ktr_header.ktr_time);
		req->ktr_header.ktr_pid = p->p_pid;
		bcopy(p->p_comm, req->ktr_header.ktr_comm, MAXCOMLEN + 1);
		req->ktr_header.ktr_buffer = NULL;
		req->ktr_header.ktr_len = 0;
	} else {
		p->p_traceflag |= KTRFAC_DROP;
		pm = print_message;
		print_message = 0;
		mtx_unlock(&ktrace_mtx);
		if (pm)
			printf("Out of ktrace request objects.\n");
		td->td_pflags &= ~TDP_INKTRACE;
	}
	return (req);
}

static void
ktr_submitrequest(struct ktr_request *req)
{

	mtx_lock(&ktrace_mtx);
	STAILQ_INSERT_TAIL(&ktr_todo, req, ktr_list);
	cv_signal(&ktrace_cv);
	mtx_unlock(&ktrace_mtx);
	curthread->td_pflags &= ~TDP_INKTRACE;
}

static void
ktr_freerequest(struct ktr_request *req)
{

	crfree(req->ktr_cred);
	if (req->ktr_vp != NULL) {
		mtx_lock(&Giant);
		vrele(req->ktr_vp);
		mtx_unlock(&Giant);
	}
	if (req->ktr_header.ktr_buffer != NULL)
		free(req->ktr_header.ktr_buffer, M_KTRACE);
	mtx_lock(&ktrace_mtx);
	STAILQ_INSERT_HEAD(&ktr_free, req, ktr_list);
	mtx_unlock(&ktrace_mtx);
}

static void
ktr_loop(void *dummy)
{
	struct ktr_request *req;
	struct thread *td;
	struct ucred *cred;

	/* Only cache these values once. */
	td = curthread;
	cred = td->td_ucred;
	for (;;) {
		mtx_lock(&ktrace_mtx);
		while (STAILQ_EMPTY(&ktr_todo))
			cv_wait(&ktrace_cv, &ktrace_mtx);
		req = STAILQ_FIRST(&ktr_todo);
		STAILQ_REMOVE_HEAD(&ktr_todo, ktr_list);
		KASSERT(req != NULL, ("got a NULL request"));
		mtx_unlock(&ktrace_mtx);
		/*
		 * It is not enough just to pass the cached cred
		 * to the VOP's in ktr_writerequest().  Some VFS
		 * operations use curthread->td_ucred, so we need
		 * to modify our thread's credentials as well.
		 * Evil.
		 */
		td->td_ucred = req->ktr_cred;
		ktr_writerequest(req);
		td->td_ucred = cred;
		ktr_freerequest(req);
	}
}

/*
 * MPSAFE
 */
void
ktrsyscall(code, narg, args)
	int code, narg;
	register_t args[];
{
	struct ktr_request *req;
	struct ktr_syscall *ktp;
	size_t buflen;
	char *buf = NULL;

	buflen = sizeof(register_t) * narg;
	if (buflen > 0) {
		buf = malloc(buflen, M_KTRACE, M_WAITOK);
		bcopy(args, buf, buflen);
	}
	req = ktr_getrequest(KTR_SYSCALL);
	if (req == NULL) {
		if (buf != NULL)
			free(buf, M_KTRACE);
		return;
	}
	ktp = &req->ktr_data.ktr_syscall;
	ktp->ktr_code = code;
	ktp->ktr_narg = narg;
	if (buflen > 0) {
		req->ktr_header.ktr_len = buflen;
		req->ktr_header.ktr_buffer = buf;
	}
	ktr_submitrequest(req);
}

/*
 * MPSAFE
 */
void
ktrsysret(code, error, retval)
	int code, error;
	register_t retval;
{
	struct ktr_request *req;
	struct ktr_sysret *ktp;

	req = ktr_getrequest(KTR_SYSRET);
	if (req == NULL)
		return;
	ktp = &req->ktr_data.ktr_sysret;
	ktp->ktr_code = code;
	ktp->ktr_error = error;
	ktp->ktr_retval = retval;		/* what about val2 ? */
	ktr_submitrequest(req);
}

void
ktrnamei(path)
	char *path;
{
	struct ktr_request *req;
	int namelen;
	char *buf = NULL;

	namelen = strlen(path);
	if (namelen > 0) {
		buf = malloc(namelen, M_KTRACE, M_WAITOK);
		bcopy(path, buf, namelen);
	}
	req = ktr_getrequest(KTR_NAMEI);
	if (req == NULL) {
		if (buf != NULL)
			free(buf, M_KTRACE);
		return;
	}
	if (namelen > 0) {
		req->ktr_header.ktr_len = namelen;
		req->ktr_header.ktr_buffer = buf;
	}
	ktr_submitrequest(req);
}

/*
 * Since the uio may not stay valid, we can not hand off this request to
 * the thread and need to process it synchronously.  However, we wish to
 * keep the relative order of records in a trace file correct, so we
 * do put this request on the queue (if it isn't empty) and then block.
 * The ktrace thread waks us back up when it is time for this event to
 * be posted and blocks until we have completed writing out the event
 * and woken it back up.
 */
void
ktrgenio(fd, rw, uio, error)
	int fd;
	enum uio_rw rw;
	struct uio *uio;
	int error;
{
	struct ktr_request *req;
	struct ktr_genio *ktg;
	int datalen;
	char *buf;

	if (error) {
		free(uio, M_IOV);
		return;
	}
	uio->uio_offset = 0;
	uio->uio_rw = UIO_WRITE;
	datalen = imin(uio->uio_resid, ktr_geniosize);
	buf = malloc(datalen, M_KTRACE, M_WAITOK);
	error = uiomove(buf, datalen, uio);
	free(uio, M_IOV);
	if (error) {
		free(buf, M_KTRACE);
		return;
	}
	req = ktr_getrequest(KTR_GENIO);
	if (req == NULL) {
		free(buf, M_KTRACE);
		return;
	}
	ktg = &req->ktr_data.ktr_genio;
	ktg->ktr_fd = fd;
	ktg->ktr_rw = rw;
	req->ktr_header.ktr_len = datalen;
	req->ktr_header.ktr_buffer = buf;
	ktr_submitrequest(req);
}

void
ktrpsig(sig, action, mask, code)
	int sig;
	sig_t action;
	sigset_t *mask;
	int code;
{
	struct ktr_request *req;
	struct ktr_psig	*kp;

	req = ktr_getrequest(KTR_PSIG);
	if (req == NULL)
		return;
	kp = &req->ktr_data.ktr_psig;
	kp->signo = (char)sig;
	kp->action = action;
	kp->mask = *mask;
	kp->code = code;
	ktr_submitrequest(req);
}

void
ktrcsw(out, user)
	int out, user;
{
	struct ktr_request *req;
	struct ktr_csw *kc;

	req = ktr_getrequest(KTR_CSW);
	if (req == NULL)
		return;
	kc = &req->ktr_data.ktr_csw;
	kc->out = out;
	kc->user = user;
	ktr_submitrequest(req);
}
#endif /* KTRACE */

/* Interface and common routines */

/*
 * ktrace system call
 *
 * MPSAFE
 */
#ifndef _SYS_SYSPROTO_H_
struct ktrace_args {
	char	*fname;
	int	ops;
	int	facs;
	int	pid;
};
#endif
/* ARGSUSED */
int
ktrace(td, uap)
	struct thread *td;
	register struct ktrace_args *uap;
{
#ifdef KTRACE
	register struct vnode *vp = NULL;
	register struct proc *p;
	struct pgrp *pg;
	int facs = uap->facs & ~KTRFAC_ROOT;
	int ops = KTROP(uap->ops);
	int descend = uap->ops & KTRFLAG_DESCEND;
	int ret = 0;
	int flags, error = 0;
	struct nameidata nd;
	struct ucred *cred;

	/*
	 * Need something to (un)trace.
	 */
	if (ops != KTROP_CLEARFILE && facs == 0)
		return (EINVAL);

	td->td_pflags |= TDP_INKTRACE;
	if (ops != KTROP_CLEAR) {
		/*
		 * an operation which requires a file argument.
		 */
		NDINIT(&nd, LOOKUP, NOFOLLOW, UIO_USERSPACE, uap->fname, td);
		flags = FREAD | FWRITE | O_NOFOLLOW;
		mtx_lock(&Giant);
		error = vn_open(&nd, &flags, 0, -1);
		if (error) {
			mtx_unlock(&Giant);
			td->td_pflags &= ~TDP_INKTRACE;
			return (error);
		}
		NDFREE(&nd, NDF_ONLY_PNBUF);
		vp = nd.ni_vp;
		VOP_UNLOCK(vp, 0, td);
		if (vp->v_type != VREG) {
			(void) vn_close(vp, FREAD|FWRITE, td->td_ucred, td);
			mtx_unlock(&Giant);
			td->td_pflags &= ~TDP_INKTRACE;
			return (EACCES);
		}
		mtx_unlock(&Giant);
	}
	/*
	 * Clear all uses of the tracefile.
	 */
	if (ops == KTROP_CLEARFILE) {
		sx_slock(&allproc_lock);
		LIST_FOREACH(p, &allproc, p_list) {
			PROC_LOCK(p);
			if (p->p_tracevp == vp) {
				if (ktrcanset(td, p)) {
					mtx_lock(&ktrace_mtx);
					cred = p->p_tracecred;
					p->p_tracecred = NULL;
					p->p_tracevp = NULL;
					p->p_traceflag = 0;
					mtx_unlock(&ktrace_mtx);
					PROC_UNLOCK(p);
					mtx_lock(&Giant);
					(void) vn_close(vp, FREAD|FWRITE,
						cred, td);
					mtx_unlock(&Giant);
					crfree(cred);
				} else {
					PROC_UNLOCK(p);
					error = EPERM;
				}
			} else
				PROC_UNLOCK(p);
		}
		sx_sunlock(&allproc_lock);
		goto done;
	}
	/*
	 * do it
	 */
	sx_slock(&proctree_lock);
	if (uap->pid < 0) {
		/*
		 * by process group
		 */
		pg = pgfind(-uap->pid);
		if (pg == NULL) {
			sx_sunlock(&proctree_lock);
			error = ESRCH;
			goto done;
		}
		/*
		 * ktrops() may call vrele(). Lock pg_members
		 * by the proctree_lock rather than pg_mtx.
		 */
		PGRP_UNLOCK(pg);
		LIST_FOREACH(p, &pg->pg_members, p_pglist)
			if (descend)
				ret |= ktrsetchildren(td, p, ops, facs, vp);
			else
				ret |= ktrops(td, p, ops, facs, vp);
	} else {
		/*
		 * by pid
		 */
		p = pfind(uap->pid);
		if (p == NULL) {
			sx_sunlock(&proctree_lock);
			error = ESRCH;
			goto done;
		}
		/*
		 * The slock of the proctree lock will keep this process
		 * from going away, so unlocking the proc here is ok.
		 */
		PROC_UNLOCK(p);
		if (descend)
			ret |= ktrsetchildren(td, p, ops, facs, vp);
		else
			ret |= ktrops(td, p, ops, facs, vp);
	}
	sx_sunlock(&proctree_lock);
	if (!ret)
		error = EPERM;
done:
	if (vp != NULL) {
		mtx_lock(&Giant);
		(void) vn_close(vp, FWRITE, td->td_ucred, td);
		mtx_unlock(&Giant);
	}
	td->td_pflags &= ~TDP_INKTRACE;
	return (error);
#else /* !KTRACE */
	return (ENOSYS);
#endif /* KTRACE */
}

/*
 * utrace system call
 *
 * MPSAFE
 */
/* ARGSUSED */
int
utrace(td, uap)
	struct thread *td;
	register struct utrace_args *uap;
{

#ifdef KTRACE
	struct ktr_request *req;
	void *cp;
	int error;

	if (!KTRPOINT(td, KTR_USER))
		return (0);
	if (uap->len > KTR_USER_MAXLEN)
		return (EINVAL);
	cp = malloc(uap->len, M_KTRACE, M_WAITOK);
	error = copyin(uap->addr, cp, uap->len);
	if (error) {
		free(cp, M_KTRACE);
		return (error);
	}
	req = ktr_getrequest(KTR_USER);
	if (req == NULL) {
		free(cp, M_KTRACE);
		return (ENOMEM);
	}
	req->ktr_header.ktr_buffer = cp;
	req->ktr_header.ktr_len = uap->len;
	ktr_submitrequest(req);
	return (0);
#else /* !KTRACE */
	return (ENOSYS);
#endif /* KTRACE */
}

#ifdef KTRACE
static int
ktrops(td, p, ops, facs, vp)
	struct thread *td;
	struct proc *p;
	int ops, facs;
	struct vnode *vp;
{
	struct vnode *tracevp = NULL;
	struct ucred *tracecred = NULL;

	PROC_LOCK(p);
	if (!ktrcanset(td, p)) {
		PROC_UNLOCK(p);
		return (0);
	}
	mtx_lock(&ktrace_mtx);
	if (ops == KTROP_SET) {
		if (p->p_tracevp != vp) {
			/*
			 * if trace file already in use, relinquish below
			 */
			tracevp = p->p_tracevp;
			VREF(vp);
			p->p_tracevp = vp;
		}
		if (p->p_tracecred != td->td_ucred) {
			tracecred = p->p_tracecred;
			p->p_tracecred = crhold(td->td_ucred);
		}
		p->p_traceflag |= facs;
		if (td->td_ucred->cr_uid == 0)
			p->p_traceflag |= KTRFAC_ROOT;
	} else {
		/* KTROP_CLEAR */
		if (((p->p_traceflag &= ~facs) & KTRFAC_MASK) == 0) {
			/* no more tracing */
			p->p_traceflag = 0;
			tracevp = p->p_tracevp;
			p->p_tracevp = NULL;
			tracecred = p->p_tracecred;
			p->p_tracecred = NULL;
		}
	}
	mtx_unlock(&ktrace_mtx);
	PROC_UNLOCK(p);
	if (tracevp != NULL) {
		mtx_lock(&Giant);
		vrele(tracevp);
		mtx_unlock(&Giant);
	}
	if (tracecred != NULL)
		crfree(tracecred);

	return (1);
}

static int
ktrsetchildren(td, top, ops, facs, vp)
	struct thread *td;
	struct proc *top;
	int ops, facs;
	struct vnode *vp;
{
	register struct proc *p;
	register int ret = 0;

	p = top;
	sx_assert(&proctree_lock, SX_LOCKED);
	for (;;) {
		ret |= ktrops(td, p, ops, facs, vp);
		/*
		 * If this process has children, descend to them next,
		 * otherwise do any siblings, and if done with this level,
		 * follow back up the tree (but not past top).
		 */
		if (!LIST_EMPTY(&p->p_children))
			p = LIST_FIRST(&p->p_children);
		else for (;;) {
			if (p == top)
				return (ret);
			if (LIST_NEXT(p, p_sibling)) {
				p = LIST_NEXT(p, p_sibling);
				break;
			}
			p = p->p_pptr;
		}
	}
	/*NOTREACHED*/
}

static void
ktr_writerequest(struct ktr_request *req)
{
	struct ktr_header *kth;
	struct vnode *vp;
	struct proc *p;
	struct thread *td;
	struct ucred *cred;
	struct uio auio;
	struct iovec aiov[3];
	struct mount *mp;
	int datalen, buflen, vrele_count;
	int error;

	vp = req->ktr_vp;
	/*
	 * If vp is NULL, the vp has been cleared out from under this
	 * request, so just drop it.
	 */
	if (vp == NULL)
		return;
	kth = &req->ktr_header;
	datalen = data_lengths[(u_short)kth->ktr_type & ~KTR_DROP];
	buflen = kth->ktr_len;
	cred = req->ktr_cred;
	td = curthread;
	auio.uio_iov = &aiov[0];
	auio.uio_offset = 0;
	auio.uio_segflg = UIO_SYSSPACE;
	auio.uio_rw = UIO_WRITE;
	aiov[0].iov_base = (caddr_t)kth;
	aiov[0].iov_len = sizeof(struct ktr_header);
	auio.uio_resid = sizeof(struct ktr_header);
	auio.uio_iovcnt = 1;
	auio.uio_td = td;
	if (datalen != 0) {
		aiov[1].iov_base = (caddr_t)&req->ktr_data;
		aiov[1].iov_len = datalen;
		auio.uio_resid += datalen;
		auio.uio_iovcnt++;
		kth->ktr_len += datalen;
	}
	if (buflen != 0) {
		KASSERT(kth->ktr_buffer != NULL, ("ktrace: nothing to write"));
		aiov[auio.uio_iovcnt].iov_base = kth->ktr_buffer;
		aiov[auio.uio_iovcnt].iov_len = buflen;
		auio.uio_resid += buflen;
		auio.uio_iovcnt++;
	}
	mtx_lock(&Giant);
	vn_start_write(vp, &mp, V_WAIT);
	vn_lock(vp, LK_EXCLUSIVE | LK_RETRY, td);
	(void)VOP_LEASE(vp, td, cred, LEASE_WRITE);
#ifdef MAC
	error = mac_check_vnode_write(cred, NOCRED, vp);
	if (error == 0)
#endif
		error = VOP_WRITE(vp, &auio, IO_UNIT | IO_APPEND, cred);
	VOP_UNLOCK(vp, 0, td);
	vn_finished_write(mp);
	mtx_unlock(&Giant);
	if (!error)
		return;
	/*
	 * If error encountered, give up tracing on this vnode.  We defer
	 * all the vrele()'s on the vnode until after we are finished walking
	 * the various lists to avoid needlessly holding locks.
	 */
	log(LOG_NOTICE, "ktrace write failed, errno %d, tracing stopped\n",
	    error);
	vrele_count = 0;
	/*
	 * First, clear this vnode from being used by any processes in the
	 * system.
	 * XXX - If one process gets an EPERM writing to the vnode, should
	 * we really do this?  Other processes might have suitable
	 * credentials for the operation.
	 */
	cred = NULL;
	sx_slock(&allproc_lock);
	LIST_FOREACH(p, &allproc, p_list) {
		PROC_LOCK(p);
		if (p->p_tracevp == vp) {
			mtx_lock(&ktrace_mtx);
			p->p_tracevp = NULL;
			p->p_traceflag = 0;
			cred = p->p_tracecred;
			p->p_tracecred = NULL;
			mtx_unlock(&ktrace_mtx);
			vrele_count++;
		}
		PROC_UNLOCK(p);
		if (cred != NULL) {
			crfree(cred);
			cred = NULL;
		}
	}
	sx_sunlock(&allproc_lock);
	/*
	 * Second, clear this vnode from any pending requests.
	 */
	mtx_lock(&ktrace_mtx);
	STAILQ_FOREACH(req, &ktr_todo, ktr_list) {
		if (req->ktr_vp == vp) {
			req->ktr_vp = NULL;
			vrele_count++;
		}
	}
	mtx_unlock(&ktrace_mtx);
	mtx_lock(&Giant);
	while (vrele_count-- > 0)
		vrele(vp);
	mtx_unlock(&Giant);
}

/*
 * Return true if caller has permission to set the ktracing state
 * of target.  Essentially, the target can't possess any
 * more permissions than the caller.  KTRFAC_ROOT signifies that
 * root previously set the tracing status on the target process, and
 * so, only root may further change it.
 */
static int
ktrcanset(td, targetp)
	struct thread *td;
	struct proc *targetp;
{

	PROC_LOCK_ASSERT(targetp, MA_OWNED);
	if (targetp->p_traceflag & KTRFAC_ROOT &&
	    suser_cred(td->td_ucred, SUSER_ALLOWJAIL))
		return (0);

	if (p_candebug(td, targetp) != 0)
		return (0);

	return (1);
}

#endif /* KTRACE */