summaryrefslogtreecommitdiffstats
path: root/sys/kern/kern_intr.c
blob: 7693954f5175ae7846b7263545cac2dfcc152693 (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
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
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
/*-
 * Copyright (c) 1997, Stefan Esser <se@freebsd.org>
 * 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 unmodified, 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.
 *
 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``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 AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT,
 * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
 * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
 * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
 * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
 * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
 * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 */

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

#include "opt_ddb.h"

#include <sys/param.h>
#include <sys/bus.h>
#include <sys/conf.h>
#include <sys/rtprio.h>
#include <sys/systm.h>
#include <sys/interrupt.h>
#include <sys/kernel.h>
#include <sys/kthread.h>
#include <sys/ktr.h>
#include <sys/limits.h>
#include <sys/lock.h>
#include <sys/malloc.h>
#include <sys/mutex.h>
#include <sys/proc.h>
#include <sys/random.h>
#include <sys/resourcevar.h>
#include <sys/sched.h>
#include <sys/sysctl.h>
#include <sys/unistd.h>
#include <sys/vmmeter.h>
#include <machine/atomic.h>
#include <machine/cpu.h>
#include <machine/md_var.h>
#include <machine/stdarg.h>
#ifdef DDB
#include <ddb/ddb.h>
#include <ddb/db_sym.h>
#endif

/*
 * Describe an interrupt thread.  There is one of these per interrupt event.
 */
struct intr_thread {
	struct intr_event *it_event;
	struct thread *it_thread;	/* Kernel thread. */
	int	it_flags;		/* (j) IT_* flags. */
	int	it_need;		/* Needs service. */
};

/* Interrupt thread flags kept in it_flags */
#define	IT_DEAD		0x000001	/* Thread is waiting to exit. */

struct	intr_entropy {
	struct	thread *td;
	uintptr_t event;
};

struct	intr_event *clk_intr_event;
struct	intr_event *tty_intr_event;
void	*softclock_ih;
void	*vm_ih;

static MALLOC_DEFINE(M_ITHREAD, "ithread", "Interrupt Threads");

static int intr_storm_threshold = 500;
TUNABLE_INT("hw.intr_storm_threshold", &intr_storm_threshold);
SYSCTL_INT(_hw, OID_AUTO, intr_storm_threshold, CTLFLAG_RW,
    &intr_storm_threshold, 0,
    "Number of consecutive interrupts before storm protection is enabled");
static TAILQ_HEAD(, intr_event) event_list =
    TAILQ_HEAD_INITIALIZER(event_list);

static void	intr_event_update(struct intr_event *ie);
static struct intr_thread *ithread_create(const char *name);
static void	ithread_destroy(struct intr_thread *ithread);
static void	ithread_execute_handlers(struct proc *p, struct intr_event *ie);
static void	ithread_loop(void *);
static void	ithread_update(struct intr_thread *ithd);
static void	start_softintr(void *);

u_char
intr_priority(enum intr_type flags)
{
	u_char pri;

	flags &= (INTR_TYPE_TTY | INTR_TYPE_BIO | INTR_TYPE_NET |
	    INTR_TYPE_CAM | INTR_TYPE_MISC | INTR_TYPE_CLK | INTR_TYPE_AV);
	switch (flags) {
	case INTR_TYPE_TTY:
		pri = PI_TTYLOW;
		break;
	case INTR_TYPE_BIO:
		/*
		 * XXX We need to refine this.  BSD/OS distinguishes
		 * between tape and disk priorities.
		 */
		pri = PI_DISK;
		break;
	case INTR_TYPE_NET:
		pri = PI_NET;
		break;
	case INTR_TYPE_CAM:
		pri = PI_DISK;          /* XXX or PI_CAM? */
		break;
	case INTR_TYPE_AV:		/* Audio/video */
		pri = PI_AV;
		break;
	case INTR_TYPE_CLK:
		pri = PI_REALTIME;
		break;
	case INTR_TYPE_MISC:
		pri = PI_DULL;          /* don't care */
		break;
	default:
		/* We didn't specify an interrupt level. */
		panic("intr_priority: no interrupt type in flags");
	}

	return pri;
}

/*
 * Update an ithread based on the associated intr_event.
 */
static void
ithread_update(struct intr_thread *ithd)
{
	struct intr_event *ie;
	struct thread *td;
	u_char pri;

	ie = ithd->it_event;
	td = ithd->it_thread;

	/* Determine the overall priority of this event. */
	if (TAILQ_EMPTY(&ie->ie_handlers))
		pri = PRI_MAX_ITHD;
	else
		pri = TAILQ_FIRST(&ie->ie_handlers)->ih_pri;

	/* Update name and priority. */
	strlcpy(td->td_proc->p_comm, ie->ie_fullname,
	    sizeof(td->td_proc->p_comm));
	mtx_lock_spin(&sched_lock);
	sched_prio(td, pri);
	mtx_unlock_spin(&sched_lock);
}

/*
 * Regenerate the full name of an interrupt event and update its priority.
 */
static void
intr_event_update(struct intr_event *ie)
{
	struct intr_handler *ih;
	char *last;
	int missed, space;

	/* Start off with no entropy and just the name of the event. */
	mtx_assert(&ie->ie_lock, MA_OWNED);
	strlcpy(ie->ie_fullname, ie->ie_name, sizeof(ie->ie_fullname));
	ie->ie_flags &= ~IE_ENTROPY;
	missed = 0;
	space = 1;

	/* Run through all the handlers updating values. */
	TAILQ_FOREACH(ih, &ie->ie_handlers, ih_next) {
		if (strlen(ie->ie_fullname) + strlen(ih->ih_name) + 1 <
		    sizeof(ie->ie_fullname)) {
			strcat(ie->ie_fullname, " ");
			strcat(ie->ie_fullname, ih->ih_name);
			space = 0;
		} else
			missed++;
		if (ih->ih_flags & IH_ENTROPY)
			ie->ie_flags |= IE_ENTROPY;
	}

	/*
	 * If the handler names were too long, add +'s to indicate missing
	 * names. If we run out of room and still have +'s to add, change
	 * the last character from a + to a *.
	 */
	last = &ie->ie_fullname[sizeof(ie->ie_fullname) - 2];
	while (missed-- > 0) {
		if (strlen(ie->ie_fullname) + 1 == sizeof(ie->ie_fullname)) {
			if (*last == '+') {
				*last = '*';
				break;
			} else
				*last = '+';
		} else if (space) {
			strcat(ie->ie_fullname, " +");
			space = 0;
		} else
			strcat(ie->ie_fullname, "+");
	}

	/*
	 * If this event has an ithread, update it's priority and
	 * name.
	 */
	if (ie->ie_thread != NULL)
		ithread_update(ie->ie_thread);
	CTR2(KTR_INTR, "%s: updated %s", __func__, ie->ie_fullname);
}

int
intr_event_create(struct intr_event **event, void *source, int flags,
    void (*enable)(void *), const char *fmt, ...)
{
	struct intr_event *ie;
	va_list ap;

	/* The only valid flag during creation is IE_SOFT. */
	if ((flags & ~IE_SOFT) != 0)
		return (EINVAL);
	ie = malloc(sizeof(struct intr_event), M_ITHREAD, M_WAITOK | M_ZERO);
	ie->ie_source = source;
	ie->ie_enable = enable;
	ie->ie_flags = flags;
	TAILQ_INIT(&ie->ie_handlers);
	mtx_init(&ie->ie_lock, "intr event", NULL, MTX_DEF);

	va_start(ap, fmt);
	vsnprintf(ie->ie_name, sizeof(ie->ie_name), fmt, ap);
	va_end(ap);
	strlcpy(ie->ie_fullname, ie->ie_name, sizeof(ie->ie_fullname));
	mtx_pool_lock(mtxpool_sleep, &event_list);
	TAILQ_INSERT_TAIL(&event_list, ie, ie_list);
	mtx_pool_unlock(mtxpool_sleep, &event_list);
	if (event != NULL)
		*event = ie;
	CTR2(KTR_INTR, "%s: created %s", __func__, ie->ie_name);
	return (0);
}

int
intr_event_destroy(struct intr_event *ie)
{

	mtx_lock(&ie->ie_lock);
	if (!TAILQ_EMPTY(&ie->ie_handlers)) {
		mtx_unlock(&ie->ie_lock);
		return (EBUSY);
	}
	mtx_pool_lock(mtxpool_sleep, &event_list);
	TAILQ_REMOVE(&event_list, ie, ie_list);
	mtx_pool_unlock(mtxpool_sleep, &event_list);
#ifndef notyet
	if (ie->ie_thread != NULL) {
		ithread_destroy(ie->ie_thread);
		ie->ie_thread = NULL;
	}
#endif
	mtx_unlock(&ie->ie_lock);
	mtx_destroy(&ie->ie_lock);
	free(ie, M_ITHREAD);
	return (0);
}

static struct intr_thread *
ithread_create(const char *name)
{
	struct intr_thread *ithd;
	struct thread *td;
	struct proc *p;
	int error;

	ithd = malloc(sizeof(struct intr_thread), M_ITHREAD, M_WAITOK | M_ZERO);

	error = kthread_create(ithread_loop, ithd, &p, RFSTOPPED | RFHIGHPID,
	    0, "%s", name);
	if (error)
		panic("kthread_create() failed with %d", error);
	td = FIRST_THREAD_IN_PROC(p);	/* XXXKSE */
	mtx_lock_spin(&sched_lock);
	td->td_ksegrp->kg_pri_class = PRI_ITHD;
	TD_SET_IWAIT(td);
	mtx_unlock_spin(&sched_lock);
	td->td_pflags |= TDP_ITHREAD;
	ithd->it_thread = td;
	CTR2(KTR_INTR, "%s: created %s", __func__, name);
	return (ithd);
}

static void
ithread_destroy(struct intr_thread *ithread)
{
	struct thread *td;

	CTR2(KTR_INTR, "%s: killing %s", __func__, ithread->it_event->ie_name);
	td = ithread->it_thread;
	mtx_lock_spin(&sched_lock);
	ithread->it_flags |= IT_DEAD;
	if (TD_AWAITING_INTR(td)) {
		TD_CLR_IWAIT(td);
		setrunqueue(td, SRQ_INTR);
	}
	mtx_unlock_spin(&sched_lock);
}

int
intr_event_add_handler(struct intr_event *ie, const char *name,
    driver_intr_t handler, void *arg, u_char pri, enum intr_type flags,
    void **cookiep)
{
	struct intr_handler *ih, *temp_ih;
	struct intr_thread *it;

	if (ie == NULL || name == NULL || handler == NULL)
		return (EINVAL);

	/* Allocate and populate an interrupt handler structure. */
	ih = malloc(sizeof(struct intr_handler), M_ITHREAD, M_WAITOK | M_ZERO);
	ih->ih_handler = handler;
	ih->ih_argument = arg;
	ih->ih_name = name;
	ih->ih_event = ie;
	ih->ih_pri = pri;
	if (flags & INTR_FAST)
		ih->ih_flags = IH_FAST;
	else if (flags & INTR_EXCL)
		ih->ih_flags = IH_EXCLUSIVE;
	if (flags & INTR_MPSAFE)
		ih->ih_flags |= IH_MPSAFE;
	if (flags & INTR_ENTROPY)
		ih->ih_flags |= IH_ENTROPY;

	/* We can only have one exclusive handler in a event. */
	mtx_lock(&ie->ie_lock);
	if (!TAILQ_EMPTY(&ie->ie_handlers)) {
		if ((flags & INTR_EXCL) ||
		    (TAILQ_FIRST(&ie->ie_handlers)->ih_flags & IH_EXCLUSIVE)) {
			mtx_unlock(&ie->ie_lock);
			free(ih, M_ITHREAD);
			return (EINVAL);
		}
	}

	/* Add the new handler to the event in priority order. */
	TAILQ_FOREACH(temp_ih, &ie->ie_handlers, ih_next) {
		if (temp_ih->ih_pri > ih->ih_pri)
			break;
	}
	if (temp_ih == NULL)
		TAILQ_INSERT_TAIL(&ie->ie_handlers, ih, ih_next);
	else
		TAILQ_INSERT_BEFORE(temp_ih, ih, ih_next);
	intr_event_update(ie);

	/* Create a thread if we need one. */
	while (ie->ie_thread == NULL && !(flags & INTR_FAST)) {
		if (ie->ie_flags & IE_ADDING_THREAD)
			msleep(ie, &ie->ie_lock, 0, "ithread", 0);
		else {
			ie->ie_flags |= IE_ADDING_THREAD;
			mtx_unlock(&ie->ie_lock);
			it = ithread_create("intr: newborn");
			mtx_lock(&ie->ie_lock);
			ie->ie_flags &= ~IE_ADDING_THREAD;
			ie->ie_thread = it;
			it->it_event = ie;
			ithread_update(it);
			wakeup(ie);
		}
	}
	CTR3(KTR_INTR, "%s: added %s to %s", __func__, ih->ih_name,
	    ie->ie_name);
	mtx_unlock(&ie->ie_lock);

	if (cookiep != NULL)
		*cookiep = ih;
	return (0);
}

int
intr_event_remove_handler(void *cookie)
{
	struct intr_handler *handler = (struct intr_handler *)cookie;
	struct intr_event *ie;
#ifdef INVARIANTS
	struct intr_handler *ih;
#endif
#ifdef notyet
	int dead;
#endif

	if (handler == NULL)
		return (EINVAL);
	ie = handler->ih_event;
	KASSERT(ie != NULL,
	    ("interrupt handler \"%s\" has a NULL interrupt event",
		handler->ih_name));
	mtx_lock(&ie->ie_lock);
	CTR3(KTR_INTR, "%s: removing %s from %s", __func__, handler->ih_name,
	    ie->ie_name);
#ifdef INVARIANTS
	TAILQ_FOREACH(ih, &ie->ie_handlers, ih_next)
		if (ih == handler)
			goto ok;
	mtx_unlock(&ie->ie_lock);
	panic("interrupt handler \"%s\" not found in interrupt event \"%s\"",
	    ih->ih_name, ie->ie_name);
ok:
#endif
	/*
	 * If there is no ithread, then just remove the handler and return.
	 * XXX: Note that an INTR_FAST handler might be running on another
	 * CPU!
	 */
	if (ie->ie_thread == NULL) {
		TAILQ_REMOVE(&ie->ie_handlers, handler, ih_next);
		mtx_unlock(&ie->ie_lock);
		free(handler, M_ITHREAD);
		return (0);
	}

	/*
	 * If the interrupt thread is already running, then just mark this
	 * handler as being dead and let the ithread do the actual removal.
	 *
	 * During a cold boot while cold is set, msleep() does not sleep,
	 * so we have to remove the handler here rather than letting the
	 * thread do it.
	 */
	mtx_lock_spin(&sched_lock);
	if (!TD_AWAITING_INTR(ie->ie_thread->it_thread) && !cold) {
		handler->ih_flags |= IH_DEAD;

		/*
		 * Ensure that the thread will process the handler list
		 * again and remove this handler if it has already passed
		 * it on the list.
		 */
		ie->ie_thread->it_need = 1;
	} else
		TAILQ_REMOVE(&ie->ie_handlers, handler, ih_next);
	mtx_unlock_spin(&sched_lock);
	while (handler->ih_flags & IH_DEAD)
		msleep(handler, &ie->ie_lock, 0, "iev_rmh", 0);
	intr_event_update(ie);
#ifdef notyet
	/*
	 * XXX: This could be bad in the case of ppbus(8).  Also, I think
	 * this could lead to races of stale data when servicing an
	 * interrupt.
	 */
	dead = 1;
	TAILQ_FOREACH(ih, &ie->ie_handlers, ih_next) {
		if (!(ih->ih_flags & IH_FAST)) {
			dead = 0;
			break;
		}
	}
	if (dead) {
		ithread_destroy(ie->ie_thread);
		ie->ie_thread = NULL;
	}
#endif
	mtx_unlock(&ie->ie_lock);
	free(handler, M_ITHREAD);
	return (0);
}

int
intr_event_schedule_thread(struct intr_event *ie)
{
	struct intr_entropy entropy;
	struct intr_thread *it;
	struct thread *td;
	struct thread *ctd;
	struct proc *p;

	/*
	 * If no ithread or no handlers, then we have a stray interrupt.
	 */
	if (ie == NULL || TAILQ_EMPTY(&ie->ie_handlers) ||
	    ie->ie_thread == NULL)
		return (EINVAL);

	ctd = curthread;
	it = ie->ie_thread;
	td = it->it_thread;
	p = td->td_proc;

	/*
	 * If any of the handlers for this ithread claim to be good
	 * sources of entropy, then gather some.
	 */
	if (harvest.interrupt && ie->ie_flags & IE_ENTROPY) {
		CTR3(KTR_INTR, "%s: pid %d (%s) gathering entropy", __func__,
		    p->p_pid, p->p_comm);
		entropy.event = (uintptr_t)ie;
		entropy.td = ctd;
		random_harvest(&entropy, sizeof(entropy), 2, 0,
		    RANDOM_INTERRUPT);
	}

	KASSERT(p != NULL, ("ithread %s has no process", ie->ie_name));

	/*
	 * Set it_need to tell the thread to keep running if it is already
	 * running.  Then, grab sched_lock and see if we actually need to
	 * put this thread on the runqueue.
	 */
	it->it_need = 1;
	mtx_lock_spin(&sched_lock);
	if (TD_AWAITING_INTR(td)) {
		CTR3(KTR_INTR, "%s: schedule pid %d (%s)", __func__, p->p_pid,
		    p->p_comm);
		TD_CLR_IWAIT(td);
		setrunqueue(td, SRQ_INTR);
	} else {
		CTR5(KTR_INTR, "%s: pid %d (%s): it_need %d, state %d",
		    __func__, p->p_pid, p->p_comm, it->it_need, td->td_state);
	}
	mtx_unlock_spin(&sched_lock);

	return (0);
}

/*
 * Add a software interrupt handler to a specified event.  If a given event
 * is not specified, then a new event is created.
 */
int
swi_add(struct intr_event **eventp, const char *name, driver_intr_t handler,
	    void *arg, int pri, enum intr_type flags, void **cookiep)
{
	struct intr_event *ie;
	int error;

	if (flags & (INTR_FAST | INTR_ENTROPY))
		return (EINVAL);

	ie = (eventp != NULL) ? *eventp : NULL;

	if (ie != NULL) {
		if (!(ie->ie_flags & IE_SOFT))
			return (EINVAL);
	} else {
		error = intr_event_create(&ie, NULL, IE_SOFT, NULL,
		    "swi%d:", pri);
		if (error)
			return (error);
		if (eventp != NULL)
			*eventp = ie;
	}
	return (intr_event_add_handler(ie, name, handler, arg,
		    (pri * RQ_PPQ) + PI_SOFT, flags, cookiep));
		    /* XXKSE.. think of a better way to get separate queues */
}

/*
 * Schedule a software interrupt thread.
 */
void
swi_sched(void *cookie, int flags)
{
	struct intr_handler *ih = (struct intr_handler *)cookie;
	struct intr_event *ie = ih->ih_event;
	int error;

	PCPU_LAZY_INC(cnt.v_intr);

	CTR3(KTR_INTR, "swi_sched: %s %s need=%d", ie->ie_name, ih->ih_name,
	    ih->ih_need);

	/*
	 * Set ih_need for this handler so that if the ithread is already
	 * running it will execute this handler on the next pass.  Otherwise,
	 * it will execute it the next time it runs.
	 */
	atomic_store_rel_int(&ih->ih_need, 1);
	if (!(flags & SWI_DELAY)) {
		error = intr_event_schedule_thread(ie);
		KASSERT(error == 0, ("stray software interrupt"));
	}
}

/*
 * Remove a software interrupt handler.  Currently this code does not
 * remove the associated interrupt event if it becomes empty.  Calling code
 * may do so manually via intr_event_destroy(), but that's not really
 * an optimal interface.
 */
int
swi_remove(void *cookie)
{

	return (intr_event_remove_handler(cookie));
}

static void
ithread_execute_handlers(struct proc *p, struct intr_event *ie)
{
	struct intr_handler *ih, *ihn;

	/* Interrupt handlers should not sleep. */
	if (!(ie->ie_flags & IE_SOFT))
		THREAD_NO_SLEEPING();
	TAILQ_FOREACH_SAFE(ih, &ie->ie_handlers, ih_next, ihn) {

		/*
		 * If this handler is marked for death, remove it from
		 * the list of handlers and wake up the sleeper.
		 */
		if (ih->ih_flags & IH_DEAD) {
			mtx_lock(&ie->ie_lock);
			TAILQ_REMOVE(&ie->ie_handlers, ih, ih_next);
			ih->ih_flags &= ~IH_DEAD;
			wakeup(ih);
			mtx_unlock(&ie->ie_lock);
			continue;
		}

		/*
		 * For software interrupt threads, we only execute
		 * handlers that have their need flag set.  Hardware
		 * interrupt threads always invoke all of their handlers.
		 */
		if (ie->ie_flags & IE_SOFT) {
			if (!ih->ih_need)
				continue;
			else
				atomic_store_rel_int(&ih->ih_need, 0);
		}

		/* Fast handlers are handled in primary interrupt context. */
		if (ih->ih_flags & IH_FAST)
			continue;

		/* Execute this handler. */
		CTR6(KTR_INTR, "%s: pid %d exec %p(%p) for %s flg=%x",
		    __func__, p->p_pid, (void *)ih->ih_handler, ih->ih_argument,
		    ih->ih_name, ih->ih_flags);

		if (!(ih->ih_flags & IH_MPSAFE))
			mtx_lock(&Giant);
		ih->ih_handler(ih->ih_argument);
		if (!(ih->ih_flags & IH_MPSAFE))
			mtx_unlock(&Giant);
	}
	if (!(ie->ie_flags & IE_SOFT))
		THREAD_SLEEPING_OK();

	/*
	 * Interrupt storm handling:
	 *
	 * If this interrupt source is currently storming, then throttle
	 * it to only fire the handler once  per clock tick.
	 *
	 * If this interrupt source is not currently storming, but the
	 * number of back to back interrupts exceeds the storm threshold,
	 * then enter storming mode.
	 */
	if (intr_storm_threshold != 0 && ie->ie_count >= intr_storm_threshold) {
		if (ie->ie_warned == 0) {
			printf(
	"Interrupt storm detected on \"%s\"; throttling interrupt source\n",
			    ie->ie_name);
			ie->ie_warned = 1;
		}
		tsleep(&ie->ie_count, 0, "istorm", 1);
	} else
		ie->ie_count++;

	/*
	 * Now that all the handlers have had a chance to run, reenable
	 * the interrupt source.
	 */
	if (ie->ie_enable != NULL)
		ie->ie_enable(ie->ie_source);
}

/*
 * This is the main code for interrupt threads.
 */
static void
ithread_loop(void *arg)
{
	struct intr_thread *ithd;
	struct intr_event *ie;
	struct thread *td;
	struct proc *p;

	td = curthread;
	p = td->td_proc;
	ithd = (struct intr_thread *)arg;
	KASSERT(ithd->it_thread == td,
	    ("%s: ithread and proc linkage out of sync", __func__));
	ie = ithd->it_event;
	ie->ie_count = 0;

	/*
	 * As long as we have interrupts outstanding, go through the
	 * list of handlers, giving each one a go at it.
	 */
	for (;;) {
		/*
		 * If we are an orphaned thread, then just die.
		 */
		if (ithd->it_flags & IT_DEAD) {
			CTR3(KTR_INTR, "%s: pid %d (%s) exiting", __func__,
			    p->p_pid, p->p_comm);
			free(ithd, M_ITHREAD);
			kthread_exit(0);
		}

		/*
		 * Service interrupts.  If another interrupt arrives while
		 * we are running, it will set it_need to note that we
		 * should make another pass.
		 */
		while (ithd->it_need) {
			/*
			 * This might need a full read and write barrier
			 * to make sure that this write posts before any
			 * of the memory or device accesses in the
			 * handlers.
			 */
			atomic_store_rel_int(&ithd->it_need, 0);
			ithread_execute_handlers(p, ie);
		}
		WITNESS_WARN(WARN_PANIC, NULL, "suspending ithread");
		mtx_assert(&Giant, MA_NOTOWNED);

		/*
		 * Processed all our interrupts.  Now get the sched
		 * lock.  This may take a while and it_need may get
		 * set again, so we have to check it again.
		 */
		mtx_lock_spin(&sched_lock);
		if (!ithd->it_need && !(ithd->it_flags & IT_DEAD)) {
			TD_SET_IWAIT(td);
			ie->ie_count = 0;
			mi_switch(SW_VOL, NULL);
		}
		mtx_unlock_spin(&sched_lock);
	}
}

#ifdef DDB
/*
 * Dump details about an interrupt handler
 */
static void
db_dump_intrhand(struct intr_handler *ih)
{
	int comma;

	db_printf("\t%-10s ", ih->ih_name);
	switch (ih->ih_pri) {
	case PI_REALTIME:
		db_printf("CLK ");
		break;
	case PI_AV:
		db_printf("AV  ");
		break;
	case PI_TTYHIGH:
	case PI_TTYLOW:
		db_printf("TTY ");
		break;
	case PI_TAPE:
		db_printf("TAPE");
		break;
	case PI_NET:
		db_printf("NET ");
		break;
	case PI_DISK:
	case PI_DISKLOW:
		db_printf("DISK");
		break;
	case PI_DULL:
		db_printf("DULL");
		break;
	default:
		if (ih->ih_pri >= PI_SOFT)
			db_printf("SWI ");
		else
			db_printf("%4u", ih->ih_pri);
		break;
	}
	db_printf(" ");
	db_printsym((uintptr_t)ih->ih_handler, DB_STGY_PROC);
	db_printf("(%p)", ih->ih_argument);
	if (ih->ih_need ||
	    (ih->ih_flags & (IH_FAST | IH_EXCLUSIVE | IH_ENTROPY | IH_DEAD |
	    IH_MPSAFE)) != 0) {
		db_printf(" {");
		comma = 0;
		if (ih->ih_flags & IH_FAST) {
			db_printf("FAST");
			comma = 1;
		}
		if (ih->ih_flags & IH_EXCLUSIVE) {
			if (comma)
				db_printf(", ");
			db_printf("EXCL");
			comma = 1;
		}
		if (ih->ih_flags & IH_ENTROPY) {
			if (comma)
				db_printf(", ");
			db_printf("ENTROPY");
			comma = 1;
		}
		if (ih->ih_flags & IH_DEAD) {
			if (comma)
				db_printf(", ");
			db_printf("DEAD");
			comma = 1;
		}
		if (ih->ih_flags & IH_MPSAFE) {
			if (comma)
				db_printf(", ");
			db_printf("MPSAFE");
			comma = 1;
		}
		if (ih->ih_need) {
			if (comma)
				db_printf(", ");
			db_printf("NEED");
		}
		db_printf("}");
	}
	db_printf("\n");
}

/*
 * Dump details about a event.
 */
void
db_dump_intr_event(struct intr_event *ie, int handlers)
{
	struct intr_handler *ih;
	struct intr_thread *it;
	int comma;

	db_printf("%s ", ie->ie_fullname);
	it = ie->ie_thread;
	if (it != NULL)
		db_printf("(pid %d)", it->it_thread->td_proc->p_pid);
	else
		db_printf("(no thread)");
	if ((ie->ie_flags & (IE_SOFT | IE_ENTROPY | IE_ADDING_THREAD)) != 0 ||
	    (it != NULL && it->it_need)) {
		db_printf(" {");
		comma = 0;
		if (ie->ie_flags & IE_SOFT) {
			db_printf("SOFT");
			comma = 1;
		}
		if (ie->ie_flags & IE_ENTROPY) {
			if (comma)
				db_printf(", ");
			db_printf("ENTROPY");
			comma = 1;
		}
		if (ie->ie_flags & IE_ADDING_THREAD) {
			if (comma)
				db_printf(", ");
			db_printf("ADDING_THREAD");
			comma = 1;
		}
		if (it != NULL && it->it_need) {
			if (comma)
				db_printf(", ");
			db_printf("NEED");
		}
		db_printf("}");
	}
	db_printf("\n");

	if (handlers)
		TAILQ_FOREACH(ih, &ie->ie_handlers, ih_next)
		    db_dump_intrhand(ih);
}

/*
 * Dump data about interrupt handlers
 */
DB_SHOW_COMMAND(intr, db_show_intr)
{
	struct intr_event *ie;
	int all, verbose;

	verbose = index(modif, 'v') != NULL;
	all = index(modif, 'a') != NULL;
	TAILQ_FOREACH(ie, &event_list, ie_list) {
		if (!all && TAILQ_EMPTY(&ie->ie_handlers))
			continue;
		db_dump_intr_event(ie, verbose);
		if (db_pager_quit)
			break;
	}
}
#endif /* DDB */

/*
 * Start standard software interrupt threads
 */
static void
start_softintr(void *dummy)
{
	struct proc *p;

	if (swi_add(&clk_intr_event, "clock", softclock, NULL, SWI_CLOCK,
		INTR_MPSAFE, &softclock_ih) ||
	    swi_add(NULL, "vm", swi_vm, NULL, SWI_VM, INTR_MPSAFE, &vm_ih))
		panic("died while creating standard software ithreads");

	p = clk_intr_event->ie_thread->it_thread->td_proc;
	PROC_LOCK(p);
	p->p_flag |= P_NOLOAD;
	PROC_UNLOCK(p);
}
SYSINIT(start_softintr, SI_SUB_SOFTINTR, SI_ORDER_FIRST, start_softintr, NULL)

/*
 * Sysctls used by systat and others: hw.intrnames and hw.intrcnt.
 * The data for this machine dependent, and the declarations are in machine
 * dependent code.  The layout of intrnames and intrcnt however is machine
 * independent.
 *
 * We do not know the length of intrcnt and intrnames at compile time, so
 * calculate things at run time.
 */
static int
sysctl_intrnames(SYSCTL_HANDLER_ARGS)
{
	return (sysctl_handle_opaque(oidp, intrnames, eintrnames - intrnames,
	   req));
}

SYSCTL_PROC(_hw, OID_AUTO, intrnames, CTLTYPE_OPAQUE | CTLFLAG_RD,
    NULL, 0, sysctl_intrnames, "", "Interrupt Names");

static int
sysctl_intrcnt(SYSCTL_HANDLER_ARGS)
{
	return (sysctl_handle_opaque(oidp, intrcnt,
	    (char *)eintrcnt - (char *)intrcnt, req));
}

SYSCTL_PROC(_hw, OID_AUTO, intrcnt, CTLTYPE_OPAQUE | CTLFLAG_RD,
    NULL, 0, sysctl_intrcnt, "", "Interrupt Counts");

#ifdef DDB
/*
 * DDB command to dump the interrupt statistics.
 */
DB_SHOW_COMMAND(intrcnt, db_show_intrcnt)
{
	u_long *i;
	char *cp;

	cp = intrnames;
	for (i = intrcnt; i != eintrcnt && !db_pager_quit; i++) {
		if (*cp == '\0')
			break;
		if (*i != 0)
			db_printf("%s\t%lu\n", cp, *i);
		cp += strlen(cp) + 1;
	}
}
#endif
OpenPOWER on IntegriCloud