summaryrefslogtreecommitdiffstats
path: root/sys/vm/vm_glue.c
blob: 81451c12fd78802ead0fce0ebb0bc5efc45c0357 (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
/*
 * Copyright (c) 1991, 1993
 *	The Regents of the University of California.  All rights reserved.
 *
 * This code is derived from software contributed to Berkeley by
 * The Mach Operating System project at Carnegie-Mellon University.
 *
 * 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. All advertising materials mentioning features or use of this software
 *    must display the following acknowledgement:
 *	This product includes software developed by the University of
 *	California, Berkeley and its contributors.
 * 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.
 *
 *	from: @(#)vm_glue.c	8.6 (Berkeley) 1/5/94
 *
 *
 * Copyright (c) 1987, 1990 Carnegie-Mellon University.
 * All rights reserved.
 *
 * Permission to use, copy, modify and distribute this software and
 * its documentation is hereby granted, provided that both the copyright
 * notice and this permission notice appear in all copies of the
 * software, derivative works or modified versions, and any portions
 * thereof, and that both notices appear in supporting documentation.
 *
 * CARNEGIE MELLON ALLOWS FREE USE OF THIS SOFTWARE IN ITS "AS IS"
 * CONDITION.  CARNEGIE MELLON DISCLAIMS ANY LIABILITY OF ANY KIND
 * FOR ANY DAMAGES WHATSOEVER RESULTING FROM THE USE OF THIS SOFTWARE.
 *
 * Carnegie Mellon requests users of this software to return to
 *
 *  Software Distribution Coordinator  or  Software.Distribution@CS.CMU.EDU
 *  School of Computer Science
 *  Carnegie Mellon University
 *  Pittsburgh PA 15213-3890
 *
 * any improvements or extensions that they make and grant Carnegie the
 * rights to redistribute these changes.
 *
 * $FreeBSD$
 */

#include "opt_vm.h"

#include <sys/param.h>
#include <sys/systm.h>
#include <sys/lock.h>
#include <sys/mutex.h>
#include <sys/proc.h>
#include <sys/resourcevar.h>
#include <sys/shm.h>
#include <sys/vmmeter.h>
#include <sys/sx.h>
#include <sys/sysctl.h>

#include <sys/kernel.h>
#include <sys/ktr.h>
#include <sys/unistd.h>

#include <machine/limits.h>

#include <vm/vm.h>
#include <vm/vm_param.h>
#include <vm/pmap.h>
#include <vm/vm_map.h>
#include <vm/vm_page.h>
#include <vm/vm_pageout.h>
#include <vm/vm_object.h>
#include <vm/vm_kern.h>
#include <vm/vm_extern.h>
#include <vm/vm_pager.h>
#include <vm/swap_pager.h>

#include <sys/user.h>

extern int maxslp;

/*
 * System initialization
 *
 * Note: proc0 from proc.h
 */
static void vm_init_limits(void *);
SYSINIT(vm_limits, SI_SUB_VM_CONF, SI_ORDER_FIRST, vm_init_limits, &proc0)

/*
 * THIS MUST BE THE LAST INITIALIZATION ITEM!!!
 *
 * Note: run scheduling should be divorced from the vm system.
 */
static void scheduler(void *);
SYSINIT(scheduler, SI_SUB_RUN_SCHEDULER, SI_ORDER_FIRST, scheduler, NULL)

#ifndef NO_SWAPPING
static void swapout(struct proc *);
static void vm_proc_swapin(struct proc *p);
static void vm_proc_swapout(struct proc *p);
#endif

/*
 * MPSAFE
 */
int
kernacc(addr, len, rw)
	caddr_t addr;
	int len, rw;
{
	boolean_t rv;
	vm_offset_t saddr, eaddr;
	vm_prot_t prot;

	KASSERT((rw & ~VM_PROT_ALL) == 0,
	    ("illegal ``rw'' argument to kernacc (%x)\n", rw));
	prot = rw;
	saddr = trunc_page((vm_offset_t)addr);
	eaddr = round_page((vm_offset_t)addr + len);
	rv = vm_map_check_protection(kernel_map, saddr, eaddr, prot);
	return (rv == TRUE);
}

/*
 * MPSAFE
 */
int
useracc(addr, len, rw)
	caddr_t addr;
	int len, rw;
{
	boolean_t rv;
	vm_prot_t prot;
	vm_map_t map;

	KASSERT((rw & ~VM_PROT_ALL) == 0,
	    ("illegal ``rw'' argument to useracc (%x)\n", rw));
	prot = rw;
	map = &curproc->p_vmspace->vm_map;
	if ((vm_offset_t)addr + len > vm_map_max(map) ||
	    (vm_offset_t)addr + len < (vm_offset_t)addr) {
		return (FALSE);
	}
	rv = vm_map_check_protection(map, trunc_page((vm_offset_t)addr),
	    round_page((vm_offset_t)addr + len), prot);
	return (rv == TRUE);
}

/*
 * MPSAFE
 */
void
vslock(addr, len)
	caddr_t addr;
	u_int len;
{

	vm_map_wire(&curproc->p_vmspace->vm_map, trunc_page((vm_offset_t)addr),
	    round_page((vm_offset_t)addr + len), FALSE);
}

/*
 * MPSAFE
 */
void
vsunlock(addr, len)
	caddr_t addr;
	u_int len;
{

	vm_map_unwire(&curproc->p_vmspace->vm_map,
	    trunc_page((vm_offset_t)addr),
	    round_page((vm_offset_t)addr + len), FALSE);
}

/*
 * Create the U area for a new process.
 * This routine directly affects the fork perf for a process.
 */
void
vm_proc_new(struct proc *p)
{
	vm_page_t ma[UAREA_PAGES];
	vm_object_t upobj;
	vm_offset_t up;
	vm_page_t m;
	u_int i;

	/*
	 * Allocate object for the upage.
	 */
	upobj = vm_object_allocate(OBJT_DEFAULT, UAREA_PAGES);
	p->p_upages_obj = upobj;

	/*
	 * Get a kernel virtual address for the U area for this process.
	 */
	up = kmem_alloc_nofault(kernel_map, UAREA_PAGES * PAGE_SIZE);
	if (up == 0)
		panic("vm_proc_new: upage allocation failed");
	p->p_uarea = (struct user *)up;

	for (i = 0; i < UAREA_PAGES; i++) {
		/*
		 * Get a uarea page.
		 */
		m = vm_page_grab(upobj, i,
		    VM_ALLOC_NORMAL | VM_ALLOC_RETRY | VM_ALLOC_WIRED);
		ma[i] = m;

		vm_page_lock_queues();
		vm_page_wakeup(m);
		vm_page_flag_clear(m, PG_ZERO);
		m->valid = VM_PAGE_BITS_ALL;
		vm_page_unlock_queues();
	}

	/*
	 * Enter the pages into the kernel address space.
	 */
	pmap_qenter(up, ma, UAREA_PAGES);
}

/*
 * Dispose the U area for a process that has exited.
 * This routine directly impacts the exit perf of a process.
 * XXX proc_zone is marked UMA_ZONE_NOFREE, so this should never be called.
 */
void
vm_proc_dispose(struct proc *p)
{
	vm_object_t upobj;
	vm_offset_t up;
	vm_page_t m;

	upobj = p->p_upages_obj;
	if (upobj->resident_page_count != UAREA_PAGES)
		panic("vm_proc_dispose: incorrect number of pages in upobj");
	vm_page_lock_queues();
	while ((m = TAILQ_FIRST(&upobj->memq)) != NULL) {
		vm_page_busy(m);
		vm_page_unwire(m, 0);
		vm_page_free(m);
	}
	vm_page_unlock_queues();
	up = (vm_offset_t)p->p_uarea;
	pmap_qremove(up, UAREA_PAGES);
	kmem_free(kernel_map, up, UAREA_PAGES * PAGE_SIZE);
	vm_object_deallocate(upobj);
}

#ifndef NO_SWAPPING
/*
 * Allow the U area for a process to be prejudicially paged out.
 */
static void
vm_proc_swapout(struct proc *p)
{
	vm_object_t upobj;
	vm_offset_t up;
	vm_page_t m;

	upobj = p->p_upages_obj;
	if (upobj->resident_page_count != UAREA_PAGES)
		panic("vm_proc_dispose: incorrect number of pages in upobj");
	vm_page_lock_queues();
	TAILQ_FOREACH(m, &upobj->memq, listq) {
		vm_page_dirty(m);
		vm_page_unwire(m, 0);
	}
	vm_page_unlock_queues();
	up = (vm_offset_t)p->p_uarea;
	pmap_qremove(up, UAREA_PAGES);
}

/*
 * Bring the U area for a specified process back in.
 */
static void
vm_proc_swapin(struct proc *p)
{
	vm_page_t ma[UAREA_PAGES];
	vm_object_t upobj;
	vm_offset_t up;
	vm_page_t m;
	int rv;
	int i;

	upobj = p->p_upages_obj;
	for (i = 0; i < UAREA_PAGES; i++) {
		m = vm_page_grab(upobj, i, VM_ALLOC_NORMAL | VM_ALLOC_RETRY);
		if (m->valid != VM_PAGE_BITS_ALL) {
			rv = vm_pager_get_pages(upobj, &m, 1, 0);
			if (rv != VM_PAGER_OK)
				panic("vm_proc_swapin: cannot get upage");
		}
		ma[i] = m;
	}
	if (upobj->resident_page_count != UAREA_PAGES)
		panic("vm_proc_swapin: lost pages from upobj");
	vm_page_lock_queues();
	TAILQ_FOREACH(m, &upobj->memq, listq) {
		m->valid = VM_PAGE_BITS_ALL;
		vm_page_wire(m);
		vm_page_wakeup(m);
	}
	vm_page_unlock_queues();
	up = (vm_offset_t)p->p_uarea;
	pmap_qenter(up, ma, UAREA_PAGES);
}

/*
 * Swap in the UAREAs of all processes swapped out to the given device.
 * The pages in the UAREA are marked dirty and their swap metadata is freed.
 */
void
vm_proc_swapin_all(int devidx)
{
	struct proc *p;
	vm_object_t object;
	vm_page_t m;

retry:
	sx_slock(&allproc_lock);
	FOREACH_PROC_IN_SYSTEM(p) {
		PROC_LOCK(p);
		mtx_lock_spin(&sched_lock);

		object = p->p_upages_obj;
		if (object != NULL &&
		    swap_pager_isswapped(p->p_upages_obj, devidx)) {
			sx_sunlock(&allproc_lock);
			faultin(p);
			mtx_unlock_spin(&sched_lock);
			PROC_UNLOCK(p);
			vm_page_lock_queues();
			TAILQ_FOREACH(m, &object->memq, listq)
				vm_page_dirty(m);
			vm_page_unlock_queues();
			swap_pager_freespace(object, 0,
			    object->un_pager.swp.swp_bcount);
			goto retry;
		}

		mtx_unlock_spin(&sched_lock);
		PROC_UNLOCK(p);
	}
	sx_sunlock(&allproc_lock);
}
#endif

/*
 * Implement fork's actions on an address space.
 * Here we arrange for the address space to be copied or referenced,
 * allocate a user struct (pcb and kernel stack), then call the
 * machine-dependent layer to fill those in and make the new process
 * ready to run.  The new process is set up so that it returns directly
 * to user mode to avoid stack copying and relocation problems.
 */
void
vm_forkproc(td, p2, td2, flags)
	struct thread *td;
	struct proc *p2;
	struct thread *td2;
	int flags;
{
	struct proc *p1 = td->td_proc;
	struct user *up;

	GIANT_REQUIRED;

	if ((flags & RFPROC) == 0) {
		/*
		 * Divorce the memory, if it is shared, essentially
		 * this changes shared memory amongst threads, into
		 * COW locally.
		 */
		if ((flags & RFMEM) == 0) {
			if (p1->p_vmspace->vm_refcnt > 1) {
				vmspace_unshare(p1);
			}
		}
		cpu_fork(td, p2, td2, flags);
		return;
	}

	if (flags & RFMEM) {
		p2->p_vmspace = p1->p_vmspace;
		p1->p_vmspace->vm_refcnt++;
	}

	while (vm_page_count_severe()) {
		VM_WAIT;
	}

	if ((flags & RFMEM) == 0) {
		p2->p_vmspace = vmspace_fork(p1->p_vmspace);

		pmap_pinit2(vmspace_pmap(p2->p_vmspace));

		if (p1->p_vmspace->vm_shm)
			shmfork(p1, p2);
	}

	/* XXXKSE this is unsatisfactory but should be adequate */
	up = p2->p_uarea;

	/*
	 * p_stats currently points at fields in the user struct
	 * but not at &u, instead at p_addr. Copy parts of
	 * p_stats; zero the rest of p_stats (statistics).
	 *
	 * If procsig->ps_refcnt is 1 and p2->p_sigacts is NULL we dont' need
	 * to share sigacts, so we use the up->u_sigacts.
	 */
	p2->p_stats = &up->u_stats;
	if (p2->p_sigacts == NULL) {
		if (p2->p_procsig->ps_refcnt != 1)
			printf ("PID:%d NULL sigacts with refcnt not 1!\n",p2->p_pid);
		p2->p_sigacts = &up->u_sigacts;
		up->u_sigacts = *p1->p_sigacts;
	}

	bzero(&up->u_stats.pstat_startzero,
	    (unsigned) ((caddr_t) &up->u_stats.pstat_endzero -
		(caddr_t) &up->u_stats.pstat_startzero));
	bcopy(&p1->p_stats->pstat_startcopy, &up->u_stats.pstat_startcopy,
	    ((caddr_t) &up->u_stats.pstat_endcopy -
		(caddr_t) &up->u_stats.pstat_startcopy));


	/*
	 * cpu_fork will copy and update the pcb, set up the kernel stack,
	 * and make the child ready to run.
	 */
	cpu_fork(td, p2, td2, flags);
}

/*
 * Called after process has been wait(2)'ed apon and is being reaped.
 * The idea is to reclaim resources that we could not reclaim while
 * the process was still executing.
 */
void
vm_waitproc(p)
	struct proc *p;
{

	GIANT_REQUIRED;
	cpu_wait(p);
	vmspace_exitfree(p);		/* and clean-out the vmspace */
}

/*
 * Set default limits for VM system.
 * Called for proc 0, and then inherited by all others.
 *
 * XXX should probably act directly on proc0.
 */
static void
vm_init_limits(udata)
	void *udata;
{
	struct proc *p = udata;
	int rss_limit;

	/*
	 * Set up the initial limits on process VM. Set the maximum resident
	 * set size to be half of (reasonably) available memory.  Since this
	 * is a soft limit, it comes into effect only when the system is out
	 * of memory - half of main memory helps to favor smaller processes,
	 * and reduces thrashing of the object cache.
	 */
	p->p_rlimit[RLIMIT_STACK].rlim_cur = dflssiz;
	p->p_rlimit[RLIMIT_STACK].rlim_max = maxssiz;
	p->p_rlimit[RLIMIT_DATA].rlim_cur = dfldsiz;
	p->p_rlimit[RLIMIT_DATA].rlim_max = maxdsiz;
	/* limit the limit to no less than 2MB */
	rss_limit = max(cnt.v_free_count, 512);
	p->p_rlimit[RLIMIT_RSS].rlim_cur = ptoa(rss_limit);
	p->p_rlimit[RLIMIT_RSS].rlim_max = RLIM_INFINITY;
}

void
faultin(p)
	struct proc *p;
{

	GIANT_REQUIRED;
	PROC_LOCK_ASSERT(p, MA_OWNED);
	mtx_assert(&sched_lock, MA_OWNED);
#ifdef NO_SWAPPING
	if ((p->p_sflag & PS_INMEM) == 0)
		panic("faultin: proc swapped out with NO_SWAPPING!");
#else
	if ((p->p_sflag & PS_INMEM) == 0) {
		struct thread *td;

		++p->p_lock;
		/*
		 * If another process is swapping in this process,
		 * just wait until it finishes.
		 */
		if (p->p_sflag & PS_SWAPPINGIN) {
			mtx_unlock_spin(&sched_lock);
			msleep(&p->p_sflag, &p->p_mtx, PVM, "faultin", 0);
			mtx_lock_spin(&sched_lock);
			--p->p_lock;
			return;
		}

		p->p_sflag |= PS_SWAPPINGIN;
		mtx_unlock_spin(&sched_lock);
		PROC_UNLOCK(p);

		vm_proc_swapin(p);
		FOREACH_THREAD_IN_PROC (p, td) {
			pmap_swapin_thread(td);
			TD_CLR_SWAPPED(td);
		}

		PROC_LOCK(p);
		mtx_lock_spin(&sched_lock);
		p->p_sflag &= ~PS_SWAPPINGIN;
		p->p_sflag |= PS_INMEM;
		FOREACH_THREAD_IN_PROC (p, td)
			if (TD_CAN_RUN(td))
				setrunnable(td);

		wakeup(&p->p_sflag);

		/* undo the effect of setting SLOCK above */
		--p->p_lock;
	}
#endif
}

/*
 * This swapin algorithm attempts to swap-in processes only if there
 * is enough space for them.  Of course, if a process waits for a long
 * time, it will be swapped in anyway.
 *
 *  XXXKSE - process with the thread with highest priority counts..
 *
 * Giant is still held at this point, to be released in tsleep.
 */
/* ARGSUSED*/
static void
scheduler(dummy)
	void *dummy;
{
	struct proc *p;
	struct thread *td;
	int pri;
	struct proc *pp;
	int ppri;

	mtx_assert(&Giant, MA_OWNED | MA_NOTRECURSED);
	/* GIANT_REQUIRED */

loop:
	if (vm_page_count_min()) {
		VM_WAIT;
		goto loop;
	}

	pp = NULL;
	ppri = INT_MIN;
	sx_slock(&allproc_lock);
	FOREACH_PROC_IN_SYSTEM(p) {
		struct ksegrp *kg;
		if (p->p_sflag & (PS_INMEM | PS_SWAPPING | PS_SWAPPINGIN)) {
			continue;
		}
		mtx_lock_spin(&sched_lock);
		FOREACH_THREAD_IN_PROC(p, td) {
			/*
			 * An otherwise runnable thread of a process
			 * swapped out has only the TDI_SWAPPED bit set.
			 * 
			 */
			if (td->td_inhibitors == TDI_SWAPPED) {
				kg = td->td_ksegrp;
				pri = p->p_swtime + kg->kg_slptime;
				if ((p->p_sflag & PS_SWAPINREQ) == 0) {
					pri -= kg->kg_nice * 8;
				}

				/*
				 * if this ksegrp is higher priority
				 * and there is enough space, then select
				 * this process instead of the previous
				 * selection.
				 */
				if (pri > ppri) {
					pp = p;
					ppri = pri;
				}
			}
		}
		mtx_unlock_spin(&sched_lock);
	}
	sx_sunlock(&allproc_lock);

	/*
	 * Nothing to do, back to sleep.
	 */
	if ((p = pp) == NULL) {
		tsleep(&proc0, PVM, "sched", maxslp * hz / 2);
		goto loop;
	}
	PROC_LOCK(p);
	mtx_lock_spin(&sched_lock);

	/*
	 * Another process may be bringing or may have already
	 * brought this process in while we traverse all threads.
	 * Or, this process may even be being swapped out again.
	 */
	if (p->p_sflag & (PS_INMEM|PS_SWAPPING|PS_SWAPPINGIN)) {
		mtx_unlock_spin(&sched_lock);
		PROC_UNLOCK(p);
		goto loop;
	}

	p->p_sflag &= ~PS_SWAPINREQ;

	/*
	 * We would like to bring someone in. (only if there is space).
	 * [What checks the space? ]
	 */
	faultin(p);
	PROC_UNLOCK(p);
	p->p_swtime = 0;
	mtx_unlock_spin(&sched_lock);
	goto loop;
}

#ifndef NO_SWAPPING

/*
 * Swap_idle_threshold1 is the guaranteed swapped in time for a process
 */
static int swap_idle_threshold1 = 2;
SYSCTL_INT(_vm, OID_AUTO, swap_idle_threshold1,
	CTLFLAG_RW, &swap_idle_threshold1, 0, "");

/*
 * Swap_idle_threshold2 is the time that a process can be idle before
 * it will be swapped out, if idle swapping is enabled.
 */
static int swap_idle_threshold2 = 10;
SYSCTL_INT(_vm, OID_AUTO, swap_idle_threshold2,
	CTLFLAG_RW, &swap_idle_threshold2, 0, "");

/*
 * Swapout is driven by the pageout daemon.  Very simple, we find eligible
 * procs and unwire their u-areas.  We try to always "swap" at least one
 * process in case we need the room for a swapin.
 * If any procs have been sleeping/stopped for at least maxslp seconds,
 * they are swapped.  Else, we swap the longest-sleeping or stopped process,
 * if any, otherwise the longest-resident process.
 */
void
swapout_procs(action)
int action;
{
	struct proc *p;
	struct thread *td;
	struct ksegrp *kg;
	struct proc *outp, *outp2;
	int outpri, outpri2;
	int didswap = 0;

	GIANT_REQUIRED;

	outp = outp2 = NULL;
	outpri = outpri2 = INT_MIN;
retry:
	sx_slock(&allproc_lock);
	FOREACH_PROC_IN_SYSTEM(p) {
		struct vmspace *vm;
		int minslptime = 100000;
		
		/*
		 * Watch out for a process in
		 * creation.  It may have no
		 * address space or lock yet.
		 */
		mtx_lock_spin(&sched_lock);
		if (p->p_state == PRS_NEW) {
			mtx_unlock_spin(&sched_lock);
			continue;
		}
		mtx_unlock_spin(&sched_lock);

		/*
		 * An aio daemon switches its
		 * address space while running.
		 * Perform a quick check whether
		 * a process has P_SYSTEM.
		 */
		PROC_LOCK(p);
		if ((p->p_flag & P_SYSTEM) != 0) {
			PROC_UNLOCK(p);
			continue;
		}

		/*
		 * Do not swapout a process that
		 * is waiting for VM data
		 * structures as there is a possible
		 * deadlock.  Test this first as
		 * this may block.
		 *
		 * Lock the map until swapout
		 * finishes, or a thread of this
		 * process may attempt to alter
		 * the map.
		 */
		vm = p->p_vmspace;
		KASSERT(vm != NULL,
			("swapout_procs: a process has no address space"));
		++vm->vm_refcnt;
		PROC_UNLOCK(p);
		if (!vm_map_trylock(&vm->vm_map))
			goto nextproc1;

		PROC_LOCK(p);
		if (p->p_lock != 0 ||
		    (p->p_flag & (P_STOPPED_SINGLE|P_TRACED|P_SYSTEM|P_WEXIT)
		    ) != 0) {
			goto nextproc2;
		}
		/*
		 * only aiod changes vmspace, however it will be
		 * skipped because of the if statement above checking 
		 * for P_SYSTEM
		 */
		mtx_lock_spin(&sched_lock);
		if ((p->p_sflag & (PS_INMEM|PS_SWAPPING|PS_SWAPPINGIN)) != PS_INMEM)
			goto nextproc;

		switch (p->p_state) {
		default:
			/* Don't swap out processes in any sort
			 * of 'special' state. */
			goto nextproc;

		case PRS_NORMAL:
			/*
			 * do not swapout a realtime process
			 * Check all the thread groups..
			 */
			FOREACH_KSEGRP_IN_PROC(p, kg) {
				if (PRI_IS_REALTIME(kg->kg_pri_class))
					goto nextproc;

				/*
				 * Guarantee swap_idle_threshold1
				 * time in memory.
				 */
				if (kg->kg_slptime < swap_idle_threshold1)
					goto nextproc;

				/*
				 * Do not swapout a process if it is
				 * waiting on a critical event of some
				 * kind or there is a thread whose
				 * pageable memory may be accessed.
				 *
				 * This could be refined to support
				 * swapping out a thread.
				 */
				FOREACH_THREAD_IN_GROUP(kg, td) {
					if ((td->td_priority) < PSOCK ||
					    !thread_safetoswapout(td))
						goto nextproc;
				}
				/*
				 * If the system is under memory stress,
				 * or if we are swapping
				 * idle processes >= swap_idle_threshold2,
				 * then swap the process out.
				 */
				if (((action & VM_SWAP_NORMAL) == 0) &&
				    (((action & VM_SWAP_IDLE) == 0) ||
				    (kg->kg_slptime < swap_idle_threshold2)))
					goto nextproc;

				if (minslptime > kg->kg_slptime)
					minslptime = kg->kg_slptime;
			}

			/*
			 * If the process has been asleep for awhile and had
			 * most of its pages taken away already, swap it out.
			 */
			if ((action & VM_SWAP_NORMAL) ||
				((action & VM_SWAP_IDLE) &&
				 (minslptime > swap_idle_threshold2))) {
				swapout(p);
				didswap++;

				/*
				 * swapout() unlocks a proc lock. This is
				 * ugly, but avoids superfluous lock.
				 */
				mtx_unlock_spin(&sched_lock);
				vm_map_unlock(&vm->vm_map);
				vmspace_free(vm);
				sx_sunlock(&allproc_lock);
				goto retry;
			}
		}
nextproc:
		mtx_unlock_spin(&sched_lock);
nextproc2:
		PROC_UNLOCK(p);
		vm_map_unlock(&vm->vm_map);
nextproc1:
		vmspace_free(vm);
		continue;
	}
	sx_sunlock(&allproc_lock);
	/*
	 * If we swapped something out, and another process needed memory,
	 * then wakeup the sched process.
	 */
	if (didswap)
		wakeup(&proc0);
}

static void
swapout(p)
	struct proc *p;
{
	struct thread *td;

	PROC_LOCK_ASSERT(p, MA_OWNED);
	mtx_assert(&sched_lock, MA_OWNED | MA_NOTRECURSED);
#if defined(SWAP_DEBUG)
	printf("swapping out %d\n", p->p_pid);
#endif

	/*
	 * The states of this process and its threads may have changed
	 * by now.  Assuming that there is only one pageout daemon thread,
	 * this process should still be in memory.
	 */
	KASSERT((p->p_sflag & (PS_INMEM|PS_SWAPPING|PS_SWAPPINGIN)) == PS_INMEM,
		("swapout: lost a swapout race?"));

#if defined(INVARIANTS)
	/*
	 * Make sure that all threads are safe to be swapped out.
	 *
	 * Alternatively, we could swap out only safe threads.
	 */
	FOREACH_THREAD_IN_PROC(p, td) {
		KASSERT(thread_safetoswapout(td),
			("swapout: there is a thread not safe for swapout"));
	}
#endif /* INVARIANTS */

	++p->p_stats->p_ru.ru_nswap;
	/*
	 * remember the process resident count
	 */
	p->p_vmspace->vm_swrss = vmspace_resident_count(p->p_vmspace);

	PROC_UNLOCK(p);
	p->p_sflag &= ~PS_INMEM;
	p->p_sflag |= PS_SWAPPING;
	mtx_unlock_spin(&sched_lock);

	vm_proc_swapout(p);
	FOREACH_THREAD_IN_PROC(p, td) {
		pmap_swapout_thread(td);
		TD_SET_SWAPPED(td);
	}
	mtx_lock_spin(&sched_lock);
	p->p_sflag &= ~PS_SWAPPING;
	p->p_swtime = 0;
}
#endif /* !NO_SWAPPING */
OpenPOWER on IntegriCloud