summaryrefslogtreecommitdiffstats
path: root/sys/vm/vm_glue.c
blob: c8c97fe7a45279b1eaebc952a1805c4430045284 (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
/*
 * 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.
 *
 * $Id: vm_glue.c,v 1.17 1995/02/21 12:06:06 davidg Exp $
 */

#include <sys/param.h>
#include <sys/systm.h>
#include <sys/proc.h>
#include <sys/resourcevar.h>
#include <sys/buf.h>
#include <sys/shm.h>
#include <sys/user.h>

#include <sys/kernel.h>
#include <sys/dkstat.h>

#include <vm/vm.h>
#include <vm/vm_page.h>
#include <vm/vm_pageout.h>
#include <vm/vm_kern.h>

#include <machine/stdarg.h>
#include <machine/cpu.h>

extern char kstack[];
int avefree = 0;		/* XXX */
int readbuffers = 0;		/* XXX allow kgdb to read kernel buffer pool */

/* vm_map_t upages_map; */

int
kernacc(addr, len, rw)
	caddr_t addr;
	int len, rw;
{
	boolean_t rv;
	vm_offset_t saddr, eaddr;
	vm_prot_t prot = rw == B_READ ? VM_PROT_READ : VM_PROT_WRITE;

	saddr = trunc_page(addr);
	eaddr = round_page(addr + len);
	rv = vm_map_check_protection(kernel_map, saddr, eaddr, prot);
	return (rv == TRUE);
}

int
useracc(addr, len, rw)
	caddr_t addr;
	int len, rw;
{
	boolean_t rv;
	vm_prot_t prot = rw == B_READ ? VM_PROT_READ : VM_PROT_WRITE;

	/*
	 * XXX - check separately to disallow access to user area and user
	 * page tables - they are in the map.
	 * 
	 * XXX - VM_MAXUSER_ADDRESS is an end address, not a max.  It was once
	 * only used (as an end address) in trap.c.  Use it as an end address
	 * here too.  This bogusness has spread.  I just fixed where it was
	 * used as a max in vm_mmap.c.
	 */
	if ((vm_offset_t) addr + len > /* XXX */ VM_MAXUSER_ADDRESS
	    || (vm_offset_t) addr + len < (vm_offset_t) addr) {
		return (FALSE);
	}
	rv = vm_map_check_protection(&curproc->p_vmspace->vm_map,
	    trunc_page(addr), round_page(addr + len), prot);
	return (rv == TRUE);
}

#ifdef KGDB
/*
 * Change protections on kernel pages from addr to addr+len
 * (presumably so debugger can plant a breakpoint).
 * All addresses are assumed to reside in the Sysmap,
 */
chgkprot(addr, len, rw)
	register caddr_t addr;
	int len, rw;
{
	vm_prot_t prot = rw == B_READ ? VM_PROT_READ : VM_PROT_WRITE;

	vm_map_protect(kernel_map, trunc_page(addr),
	    round_page(addr + len), prot, FALSE);
}
#endif
void
vslock(addr, len)
	caddr_t addr;
	u_int len;
{
	vm_map_pageable(&curproc->p_vmspace->vm_map, trunc_page(addr),
	    round_page(addr + len), FALSE);
}

void
vsunlock(addr, len, dirtied)
	caddr_t addr;
	u_int len;
	int dirtied;
{
#ifdef	lint
	dirtied++;
#endif	/* lint */
	vm_map_pageable(&curproc->p_vmspace->vm_map, trunc_page(addr),
	    round_page(addr + len), TRUE);
}

/*
 * 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.
 * NOTE: the kernel stack may be at a different location in the child
 * process, and thus addresses of automatic variables may be invalid
 * after cpu_fork returns in the child process.  We do nothing here
 * after cpu_fork returns.
 */
int
vm_fork(p1, p2, isvfork)
	register struct proc *p1, *p2;
	int isvfork;
{
	register struct user *up;
	vm_offset_t addr, ptaddr;
	int i;
	struct vm_map *vp;

	while ((cnt.v_free_count + cnt.v_cache_count) < cnt.v_free_min) {
		VM_WAIT;
	}

	/*
	 * avoid copying any of the parent's pagetables or other per-process
	 * objects that reside in the map by marking all of them
	 * non-inheritable
	 */
	(void) vm_map_inherit(&p1->p_vmspace->vm_map,
	    UPT_MIN_ADDRESS - UPAGES * NBPG, VM_MAX_ADDRESS, VM_INHERIT_NONE);
	p2->p_vmspace = vmspace_fork(p1->p_vmspace);

#ifdef SYSVSHM
	if (p1->p_vmspace->vm_shm)
		shmfork(p1, p2, isvfork);
#endif

	/*
	 * Allocate a wired-down (for now) pcb and kernel stack for the
	 * process
	 */

	addr = (vm_offset_t) kstack;

	vp = &p2->p_vmspace->vm_map;

	/* get new pagetables and kernel stack */
	(void) vm_map_find(vp, NULL, 0, &addr, UPT_MAX_ADDRESS - addr, FALSE);

	/* force in the page table encompassing the UPAGES */
	ptaddr = trunc_page((u_int) vtopte(addr));
	vm_map_pageable(vp, ptaddr, ptaddr + NBPG, FALSE);

	/* and force in (demand-zero) the UPAGES */
	vm_map_pageable(vp, addr, addr + UPAGES * NBPG, FALSE);

	/* get a kernel virtual address for the UPAGES for this proc */
	up = (struct user *) kmem_alloc_pageable(u_map, UPAGES * NBPG);
	if (up == NULL)
		panic("vm_fork: u_map allocation failed");

	/* and force-map the upages into the kernel pmap */
	for (i = 0; i < UPAGES; i++)
		pmap_enter(vm_map_pmap(u_map),
		    ((vm_offset_t) up) + NBPG * i,
		    pmap_extract(vp->pmap, addr + NBPG * i),
		    VM_PROT_READ | VM_PROT_WRITE, 1);

	p2->p_addr = up;

	/*
	 * p_stats and p_sigacts currently point at fields in the user struct
	 * but not at &u, instead at p_addr. Copy p_sigacts and parts of
	 * p_stats; zero the rest of p_stats (statistics).
	 */
	p2->p_stats = &up->u_stats;
	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 kernel stack and pcb, and make
	 * the child ready to run.  It marks the child so that it can return
	 * differently than the parent. It returns twice, once in the parent
	 * process and once in the child.
	 */
	return (cpu_fork(p1, p2));
}

/*
 * Set default limits for VM system.
 * Called for proc 0, and then inherited by all others.
 */
void
vm_init_limits(p)
	register struct proc *p;
{
	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 / 2, 512);
	p->p_rlimit[RLIMIT_RSS].rlim_cur = ptoa(rss_limit);
	p->p_rlimit[RLIMIT_RSS].rlim_max = RLIM_INFINITY;
}

#ifdef DEBUG
int enableswap = 1;
int swapdebug = 0;

#define	SDB_FOLLOW	1
#define SDB_SWAPIN	2
#define SDB_SWAPOUT	4
#endif

void
faultin(p)
	struct proc *p;
{
	vm_offset_t i;
	vm_offset_t ptaddr;
	int s;

	if ((p->p_flag & P_INMEM) == 0) {
		vm_map_t map;

		++p->p_lock;

		map = &p->p_vmspace->vm_map;
		/* force the page table encompassing the kernel stack (upages) */
		ptaddr = trunc_page((u_int) vtopte(kstack));
		vm_map_pageable(map, ptaddr, ptaddr + NBPG, FALSE);

		/* wire in the UPAGES */
		vm_map_pageable(map, (vm_offset_t) kstack,
		    (vm_offset_t) kstack + UPAGES * NBPG, FALSE);

		/* and map them nicely into the kernel pmap */
		for (i = 0; i < UPAGES; i++) {
			vm_offset_t off = i * NBPG;
			vm_offset_t pa = (vm_offset_t)
			pmap_extract(&p->p_vmspace->vm_pmap,
			    (vm_offset_t) kstack + off);

			pmap_enter(vm_map_pmap(u_map),
			    ((vm_offset_t) p->p_addr) + off,
			    pa, VM_PROT_READ | VM_PROT_WRITE, 1);
		}

		s = splhigh();

		if (p->p_stat == SRUN)
			setrunqueue(p);

		p->p_flag |= P_INMEM;

		/* undo the effect of setting SLOCK above */
		--p->p_lock;
		splx(s);

	}
}

/*
 * 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.
 */
void
scheduler()
{
	register struct proc *p;
	register int pri;
	struct proc *pp;
	int ppri;

loop:
	while ((cnt.v_free_count + cnt.v_cache_count) < (cnt.v_free_reserved + UPAGES + 2)) {
		VM_WAIT;
		tsleep((caddr_t) &proc0, PVM, "schedm", 0);
	}

	pp = NULL;
	ppri = INT_MIN;
	for (p = (struct proc *) allproc; p != NULL; p = p->p_next) {
		if (p->p_stat == SRUN && (p->p_flag & (P_INMEM | P_SWAPPING)) == 0) {
			int mempri;

			pri = p->p_swtime + p->p_slptime - p->p_nice * 8;
			mempri = pri > 0 ? pri : 0;
			/*
			 * if this process is higher priority and there is
			 * enough space, then select this process instead of
			 * the previous selection.
			 */
			if (pri > ppri) {
				pp = p;
				ppri = pri;
			}
		}
	}

	/*
	 * Nothing to do, back to sleep
	 */
	if ((p = pp) == NULL) {
		tsleep((caddr_t) &proc0, PVM, "sched", 0);
		goto loop;
	}
	/*
	 * We would like to bring someone in. (only if there is space).
	 */
	faultin(p);
	p->p_swtime = 0;
	goto loop;
}

#define	swappable(p) \
	(((p)->p_lock == 0) && \
		((p)->p_flag & (P_TRACED|P_NOSWAP|P_SYSTEM|P_INMEM|P_WEXIT|P_PHYSIO|P_SWAPPING)) == P_INMEM)

extern int vm_pageout_free_min;

/*
 * 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_threads()
{
	register struct proc *p;
	struct proc *outp, *outp2;
	int outpri, outpri2;
	int didswap = 0;

	outp = outp2 = NULL;
	outpri = outpri2 = INT_MIN;
retry:
	for (p = (struct proc *) allproc; p != NULL; p = p->p_next) {
		if (!swappable(p))
			continue;
		switch (p->p_stat) {
		default:
			continue;

		case SSLEEP:
		case SSTOP:
			/*
			 * do not swapout a realtime process
			 */
			if (p->p_rtprio.type == RTP_PRIO_REALTIME)
				continue;

			/*
			 * do not swapout a process waiting on a critical
			 * event of some kind
			 */
			if ((p->p_priority & 0x7f) < PSOCK)
				continue;

			vm_map_reference(&p->p_vmspace->vm_map);
			/*
			 * do not swapout a process that is waiting for VM
			 * datastructures there is a possible deadlock.
			 */
			if (!lock_try_write(&p->p_vmspace->vm_map.lock)) {
				vm_map_deallocate(&p->p_vmspace->vm_map);
				continue;
			}
			vm_map_unlock(&p->p_vmspace->vm_map);
			/*
			 * If the process has been asleep for awhile and had
			 * most of its pages taken away already, swap it out.
			 */
			if (p->p_slptime > 4) {
				swapout(p);
				vm_map_deallocate(&p->p_vmspace->vm_map);
				didswap++;
				goto retry;
			}
			vm_map_deallocate(&p->p_vmspace->vm_map);
		}
	}
	/*
	 * If we swapped something out, and another process needed memory,
	 * then wakeup the sched process.
	 */
	if (didswap)
		wakeup((caddr_t) &proc0);
}

void
swapout(p)
	register struct proc *p;
{
	vm_map_t map = &p->p_vmspace->vm_map;
	vm_offset_t ptaddr;

	++p->p_stats->p_ru.ru_nswap;
	/*
	 * remember the process resident count
	 */
	p->p_vmspace->vm_swrss =
	    p->p_vmspace->vm_pmap.pm_stats.resident_count;

	(void) splhigh();
	p->p_flag &= ~P_INMEM;
	if (p->p_stat == SRUN)
		remrq(p);
	(void) spl0();

	p->p_flag |= P_SWAPPING;
	/*
	 * let the upages be paged
	 */
	pmap_remove(vm_map_pmap(u_map),
	    (vm_offset_t) p->p_addr, ((vm_offset_t) p->p_addr) + UPAGES * NBPG);

	vm_map_pageable(map, (vm_offset_t) kstack,
	    (vm_offset_t) kstack + UPAGES * NBPG, TRUE);

	ptaddr = trunc_page((u_int) vtopte(kstack));
	vm_map_pageable(map, ptaddr, ptaddr + NBPG, TRUE);

	p->p_flag &= ~P_SWAPPING;
	p->p_swtime = 0;
}

/*
 * The rest of these routines fake thread handling
 */

#ifndef assert_wait
void
assert_wait(event, ruptible)
	int event;
	boolean_t ruptible;
{
#ifdef lint
	ruptible++;
#endif
	curproc->p_thread = event;
}
#endif

void
thread_block(char *msg)
{
	if (curproc->p_thread)
		tsleep((caddr_t) curproc->p_thread, PVM, msg, 0);
}


void
thread_sleep_(event, lock, wmesg)
	int event;
	simple_lock_t lock;
	char *wmesg;
{

	curproc->p_thread = event;
	simple_unlock(lock);
	if (curproc->p_thread) {
		tsleep((caddr_t) event, PVM, wmesg, 0);
	}
}

#ifndef thread_wakeup
void
thread_wakeup(event)
	int event;
{
	wakeup((caddr_t) event);
}
#endif

/*
 * DEBUG stuff
 */

int indent = 0;

#include <machine/stdarg.h>	/* see subr_prf.c */

/*ARGSUSED2*/
void
#if __STDC__
iprintf(const char *fmt,...)
#else
iprintf(fmt /* , va_alist */ )
	char *fmt;

 /* va_dcl */
#endif
{
	register int i;
	va_list ap;

	for (i = indent; i >= 8; i -= 8)
		printf("\t");
	while (--i >= 0)
		printf(" ");
	va_start(ap, fmt);
	printf("%r", fmt, ap);
	va_end(ap);
}
OpenPOWER on IntegriCloud