summaryrefslogtreecommitdiffstats
path: root/sys/vm/vm_fault.c
blob: 330cce138613917a34ab6a60b34df3e8c9e83719 (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
990
991
992
993
994
995
996
997
998
999
1000
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073
1074
1075
1076
1077
1078
1079
1080
1081
1082
1083
1084
1085
1086
1087
1088
1089
1090
1091
1092
1093
1094
1095
1096
1097
1098
1099
1100
1101
1102
1103
1104
1105
1106
1107
1108
1109
1110
1111
1112
1113
1114
1115
1116
1117
1118
1119
1120
1121
1122
1123
1124
1125
1126
1127
1128
1129
1130
1131
1132
1133
1134
1135
1136
1137
1138
1139
1140
1141
1142
1143
1144
1145
1146
1147
1148
1149
1150
1151
1152
1153
1154
1155
1156
1157
1158
1159
1160
1161
1162
1163
1164
1165
1166
1167
1168
1169
1170
1171
1172
1173
1174
1175
1176
1177
1178
1179
1180
1181
1182
1183
1184
1185
1186
1187
1188
1189
1190
1191
1192
1193
1194
1195
1196
1197
1198
1199
1200
1201
1202
1203
1204
1205
1206
1207
1208
1209
1210
1211
1212
1213
1214
1215
1216
1217
1218
1219
1220
1221
1222
1223
1224
1225
1226
1227
1228
1229
1230
1231
1232
1233
1234
1235
1236
1237
1238
1239
1240
1241
1242
1243
1244
1245
1246
1247
1248
1249
1250
1251
1252
1253
1254
1255
1256
1257
1258
1259
1260
1261
1262
1263
1264
1265
1266
1267
1268
1269
1270
1271
1272
/*
 * Copyright (c) 1991, 1993
 *	The Regents of the University of California.  All rights reserved.
 * Copyright (c) 1994 John S. Dyson
 * All rights reserved.
 * Copyright (c) 1994 David Greenman
 * 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_fault.c	8.4 (Berkeley) 1/12/94
 *
 *
 * Copyright (c) 1987, 1990 Carnegie-Mellon University.
 * All rights reserved.
 *
 * Authors: Avadis Tevanian, Jr., Michael Wayne Young
 *
 * 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.
 *
 * $FreeBSD$
 */

/*
 *	Page fault handling module.
 */

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

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

static int vm_fault_additional_pages __P((vm_page_t, int,
					  int, vm_page_t *, int *));

#define VM_FAULT_READ_AHEAD 8
#define VM_FAULT_READ_BEHIND 7
#define VM_FAULT_READ (VM_FAULT_READ_AHEAD+VM_FAULT_READ_BEHIND+1)

struct faultstate {
	vm_page_t m;
	vm_object_t object;
	vm_pindex_t pindex;
	vm_page_t first_m;
	vm_object_t	first_object;
	vm_pindex_t first_pindex;
	vm_map_t map;
	vm_map_entry_t entry;
	int lookup_still_valid;
	struct vnode *vp;
};

static __inline void
release_page(struct faultstate *fs)
{
	vm_page_wakeup(fs->m);
	vm_page_deactivate(fs->m);
	fs->m = NULL;
}

static __inline void
unlock_map(struct faultstate *fs)
{
	if (fs->lookup_still_valid) {
		vm_map_lookup_done(fs->map, fs->entry);
		fs->lookup_still_valid = FALSE;
	}
}

static void
_unlock_things(struct faultstate *fs, int dealloc)
{
	GIANT_REQUIRED;
	vm_object_pip_wakeup(fs->object);
	if (fs->object != fs->first_object) {
		vm_page_free(fs->first_m);
		vm_object_pip_wakeup(fs->first_object);
		fs->first_m = NULL;
	}
	if (dealloc) {
		vm_object_deallocate(fs->first_object);
	}
	unlock_map(fs);	
	if (fs->vp != NULL) { 
		vput(fs->vp);
		fs->vp = NULL;
	}
}

#define unlock_things(fs) _unlock_things(fs, 0)
#define unlock_and_deallocate(fs) _unlock_things(fs, 1)

/*
 * TRYPAGER - used by vm_fault to calculate whether the pager for the
 *	      current object *might* contain the page.
 *
 *	      default objects are zero-fill, there is no real pager.
 */

#define TRYPAGER	(fs.object->type != OBJT_DEFAULT && \
			(((fault_flags & VM_FAULT_WIRE_MASK) == 0) || wired))

/*
 *	vm_fault:
 *
 *	Handle a page fault occurring at the given address,
 *	requiring the given permissions, in the map specified.
 *	If successful, the page is inserted into the
 *	associated physical map.
 *
 *	NOTE: the given address should be truncated to the
 *	proper page address.
 *
 *	KERN_SUCCESS is returned if the page fault is handled; otherwise,
 *	a standard error specifying why the fault is fatal is returned.
 *
 *
 *	The map in question must be referenced, and remains so.
 *	Caller may hold no locks.
 */
static int vm_fault1 __P((vm_map_t, vm_offset_t, vm_prot_t, int));

int
vm_fault(vm_map_t map, vm_offset_t vaddr, vm_prot_t fault_type,
	 int fault_flags)
{
	int ret;

	mtx_lock(&Giant);
	/* GIANT_REQUIRED */

	ret = vm_fault1(map, vaddr, fault_type, fault_flags);
	mtx_unlock(&Giant);
	return (ret);
}

static int
vm_fault1(vm_map_t map, vm_offset_t vaddr, vm_prot_t fault_type,
	  int fault_flags)
{
	vm_prot_t prot;
	int result;
	boolean_t wired;
	int map_generation;
	vm_object_t next_object;
	vm_page_t marray[VM_FAULT_READ];
	int hardfault;
	int faultcount;
	struct faultstate fs;

	GIANT_REQUIRED;

	cnt.v_vm_faults++;
	hardfault = 0;

RetryFault:;

	/*
	 * Find the backing store object and offset into it to begin the
	 * search.
	 */
	fs.map = map;
	if ((result = vm_map_lookup(&fs.map, vaddr,
		fault_type, &fs.entry, &fs.first_object,
		&fs.first_pindex, &prot, &wired)) != KERN_SUCCESS) {
		if ((result != KERN_PROTECTION_FAILURE) ||
			((fault_flags & VM_FAULT_WIRE_MASK) != VM_FAULT_USER_WIRE)) {
			return result;
		}

		/*
   		 * If we are user-wiring a r/w segment, and it is COW, then
   		 * we need to do the COW operation.  Note that we don't COW
   		 * currently RO sections now, because it is NOT desirable
   		 * to COW .text.  We simply keep .text from ever being COW'ed
   		 * and take the heat that one cannot debug wired .text sections.
   		 */
		result = vm_map_lookup(&fs.map, vaddr,
			VM_PROT_READ|VM_PROT_WRITE|VM_PROT_OVERRIDE_WRITE,
			&fs.entry, &fs.first_object, &fs.first_pindex, &prot, &wired);
		if (result != KERN_SUCCESS) {
			return result;
		}

		/*
		 * If we don't COW now, on a user wire, the user will never
		 * be able to write to the mapping.  If we don't make this
		 * restriction, the bookkeeping would be nearly impossible.
		 */
		if ((fs.entry->protection & VM_PROT_WRITE) == 0)
			fs.entry->max_protection &= ~VM_PROT_WRITE;
	}

	map_generation = fs.map->timestamp;

	if (fs.entry->eflags & MAP_ENTRY_NOFAULT) {
		panic("vm_fault: fault on nofault entry, addr: %lx",
		    (u_long)vaddr);
	}

	/*
	 * Make a reference to this object to prevent its disposal while we
	 * are messing with it.  Once we have the reference, the map is free
	 * to be diddled.  Since objects reference their shadows (and copies),
	 * they will stay around as well.
	 */
	vm_object_reference(fs.first_object);
	vm_object_pip_add(fs.first_object, 1);

	fs.vp = vnode_pager_lock(fs.first_object);
	if ((fault_type & VM_PROT_WRITE) &&
		(fs.first_object->type == OBJT_VNODE)) {
		vm_freeze_copyopts(fs.first_object,
			fs.first_pindex, fs.first_pindex + 1);
	}

	fs.lookup_still_valid = TRUE;

	if (wired)
		fault_type = prot;

	fs.first_m = NULL;

	/*
	 * Search for the page at object/offset.
	 */

	fs.object = fs.first_object;
	fs.pindex = fs.first_pindex;

	while (TRUE) {
		/*
		 * If the object is dead, we stop here
		 */

		if (fs.object->flags & OBJ_DEAD) {
			unlock_and_deallocate(&fs);
			return (KERN_PROTECTION_FAILURE);
		}

		/*
		 * See if page is resident
		 */
			
		fs.m = vm_page_lookup(fs.object, fs.pindex);
		if (fs.m != NULL) {
			int queue, s;
			/*
			 * Wait/Retry if the page is busy.  We have to do this
			 * if the page is busy via either PG_BUSY or 
			 * vm_page_t->busy because the vm_pager may be using
			 * vm_page_t->busy for pageouts ( and even pageins if
			 * it is the vnode pager ), and we could end up trying
			 * to pagein and pageout the same page simultaneously.
			 *
			 * We can theoretically allow the busy case on a read
			 * fault if the page is marked valid, but since such
			 * pages are typically already pmap'd, putting that
			 * special case in might be more effort then it is 
			 * worth.  We cannot under any circumstances mess
			 * around with a vm_page_t->busy page except, perhaps,
			 * to pmap it.
			 */
			if ((fs.m->flags & PG_BUSY) || fs.m->busy) {
				unlock_things(&fs);
				(void)vm_page_sleep_busy(fs.m, TRUE, "vmpfw");
				cnt.v_intrans++;
				vm_object_deallocate(fs.first_object);
				goto RetryFault;
			}

			queue = fs.m->queue;
			s = splvm();
			vm_page_unqueue_nowakeup(fs.m);
			splx(s);

			if ((queue - fs.m->pc) == PQ_CACHE && vm_page_count_severe()) {
				vm_page_activate(fs.m);
				unlock_and_deallocate(&fs);
				VM_WAIT;
				goto RetryFault;
			}

			/*
			 * Mark page busy for other processes, and the 
			 * pagedaemon.  If it still isn't completely valid
			 * (readable), jump to readrest, else break-out ( we
			 * found the page ).
			 */

			vm_page_busy(fs.m);
			if (((fs.m->valid & VM_PAGE_BITS_ALL) != VM_PAGE_BITS_ALL) &&
				fs.m->object != kernel_object && fs.m->object != kmem_object) {
				goto readrest;
			}

			break;
		}

		/*
		 * Page is not resident, If this is the search termination
		 * or the pager might contain the page, allocate a new page.
		 */

		if (TRYPAGER || fs.object == fs.first_object) {
			if (fs.pindex >= fs.object->size) {
				unlock_and_deallocate(&fs);
				return (KERN_PROTECTION_FAILURE);
			}

			/*
			 * Allocate a new page for this object/offset pair.
			 */
			fs.m = NULL;
			if (!vm_page_count_severe()) {
				fs.m = vm_page_alloc(fs.object, fs.pindex,
				    (fs.vp || fs.object->backing_object)? VM_ALLOC_NORMAL: VM_ALLOC_ZERO);
			}
			if (fs.m == NULL) {
				unlock_and_deallocate(&fs);
				VM_WAIT;
				goto RetryFault;
			}
		}

readrest:
		/*
		 * We have found a valid page or we have allocated a new page.
		 * The page thus may not be valid or may not be entirely 
		 * valid.
		 *
		 * Attempt to fault-in the page if there is a chance that the
		 * pager has it, and potentially fault in additional pages
		 * at the same time.
		 */

		if (TRYPAGER) {
			int rv;
			int reqpage;
			int ahead, behind;
			u_char behavior = vm_map_entry_behavior(fs.entry);

			if (behavior == MAP_ENTRY_BEHAV_RANDOM) {
				ahead = 0;
				behind = 0;
			} else {
				behind = (vaddr - fs.entry->start) >> PAGE_SHIFT;
				if (behind > VM_FAULT_READ_BEHIND)
					behind = VM_FAULT_READ_BEHIND;

				ahead = ((fs.entry->end - vaddr) >> PAGE_SHIFT) - 1;
				if (ahead > VM_FAULT_READ_AHEAD)
					ahead = VM_FAULT_READ_AHEAD;
			}

			if ((fs.first_object->type != OBJT_DEVICE) &&
			    (behavior == MAP_ENTRY_BEHAV_SEQUENTIAL ||
                                (behavior != MAP_ENTRY_BEHAV_RANDOM &&
                                fs.pindex >= fs.entry->lastr &&
                                fs.pindex < fs.entry->lastr + VM_FAULT_READ))
			) {
				vm_pindex_t firstpindex, tmppindex;

				if (fs.first_pindex < 2 * VM_FAULT_READ)
					firstpindex = 0;
				else
					firstpindex = fs.first_pindex - 2 * VM_FAULT_READ;

				/*
				 * note: partially valid pages cannot be 
				 * included in the lookahead - NFS piecemeal
				 * writes will barf on it badly.
				 */

				for(tmppindex = fs.first_pindex - 1;
					tmppindex >= firstpindex;
					--tmppindex) {
					vm_page_t mt;
					mt = vm_page_lookup( fs.first_object, tmppindex);
					if (mt == NULL || (mt->valid != VM_PAGE_BITS_ALL))
						break;
					if (mt->busy ||
						(mt->flags & (PG_BUSY | PG_FICTITIOUS | PG_UNMANAGED)) ||
						mt->hold_count ||
						mt->wire_count) 
						continue;
					if (mt->dirty == 0)
						vm_page_test_dirty(mt);
					if (mt->dirty) {
						vm_page_protect(mt, VM_PROT_NONE);
						vm_page_deactivate(mt);
					} else {
						vm_page_cache(mt);
					}
				}

				ahead += behind;
				behind = 0;
			}

			/*
			 * now we find out if any other pages should be paged
			 * in at this time this routine checks to see if the
			 * pages surrounding this fault reside in the same
			 * object as the page for this fault.  If they do,
			 * then they are faulted in also into the object.  The
			 * array "marray" returned contains an array of
			 * vm_page_t structs where one of them is the
			 * vm_page_t passed to the routine.  The reqpage
			 * return value is the index into the marray for the
			 * vm_page_t passed to the routine.
			 *
			 * fs.m plus the additional pages are PG_BUSY'd.
			 */
			faultcount = vm_fault_additional_pages(
			    fs.m, behind, ahead, marray, &reqpage);

			/*
			 * update lastr imperfectly (we do not know how much
			 * getpages will actually read), but good enough.
			 */
			fs.entry->lastr = fs.pindex + faultcount - behind;

			/*
			 * Call the pager to retrieve the data, if any, after
			 * releasing the lock on the map.  We hold a ref on
			 * fs.object and the pages are PG_BUSY'd.
			 */
			unlock_map(&fs);

			rv = faultcount ?
			    vm_pager_get_pages(fs.object, marray, faultcount,
				reqpage) : VM_PAGER_FAIL;

			if (rv == VM_PAGER_OK) {
				/*
				 * Found the page. Leave it busy while we play
				 * with it.
				 */

				/*
				 * Relookup in case pager changed page. Pager
				 * is responsible for disposition of old page
				 * if moved.
				 */
				fs.m = vm_page_lookup(fs.object, fs.pindex);
				if(!fs.m) {
					unlock_and_deallocate(&fs);
					goto RetryFault;
				}

				hardfault++;
				break; /* break to PAGE HAS BEEN FOUND */
			}
			/*
			 * Remove the bogus page (which does not exist at this
			 * object/offset); before doing so, we must get back
			 * our object lock to preserve our invariant.
			 *
			 * Also wake up any other process that may want to bring
			 * in this page.
			 *
			 * If this is the top-level object, we must leave the
			 * busy page to prevent another process from rushing
			 * past us, and inserting the page in that object at
			 * the same time that we are.
			 */

			if (rv == VM_PAGER_ERROR)
				printf("vm_fault: pager read error, pid %d (%s)\n",
				    curproc->p_pid, curproc->p_comm);
			/*
			 * Data outside the range of the pager or an I/O error
			 */
			/*
			 * XXX - the check for kernel_map is a kludge to work
			 * around having the machine panic on a kernel space
			 * fault w/ I/O error.
			 */
			if (((fs.map != kernel_map) && (rv == VM_PAGER_ERROR)) ||
				(rv == VM_PAGER_BAD)) {
				vm_page_free(fs.m);
				fs.m = NULL;
				unlock_and_deallocate(&fs);
				return ((rv == VM_PAGER_ERROR) ? KERN_FAILURE : KERN_PROTECTION_FAILURE);
			}
			if (fs.object != fs.first_object) {
				vm_page_free(fs.m);
				fs.m = NULL;
				/*
				 * XXX - we cannot just fall out at this
				 * point, m has been freed and is invalid!
				 */
			}
		}

		/*
		 * We get here if the object has default pager (or unwiring) 
		 * or the pager doesn't have the page.
		 */
		if (fs.object == fs.first_object)
			fs.first_m = fs.m;

		/*
		 * Move on to the next object.  Lock the next object before
		 * unlocking the current one.
		 */

		fs.pindex += OFF_TO_IDX(fs.object->backing_object_offset);
		next_object = fs.object->backing_object;
		if (next_object == NULL) {
			/*
			 * If there's no object left, fill the page in the top
			 * object with zeros.
			 */
			if (fs.object != fs.first_object) {
				vm_object_pip_wakeup(fs.object);

				fs.object = fs.first_object;
				fs.pindex = fs.first_pindex;
				fs.m = fs.first_m;
			}
			fs.first_m = NULL;

			/*
			 * Zero the page if necessary and mark it valid.
			 */
			if ((fs.m->flags & PG_ZERO) == 0) {
				vm_page_zero_fill(fs.m);
			} else {
				cnt.v_ozfod++;
			}
			cnt.v_zfod++;
			fs.m->valid = VM_PAGE_BITS_ALL;
			break;	/* break to PAGE HAS BEEN FOUND */
		} else {
			if (fs.object != fs.first_object) {
				vm_object_pip_wakeup(fs.object);
			}
			KASSERT(fs.object != next_object, ("object loop %p", next_object));
			fs.object = next_object;
			vm_object_pip_add(fs.object, 1);
		}
	}

	KASSERT((fs.m->flags & PG_BUSY) != 0,
	    ("vm_fault: not busy after main loop"));

	/*
	 * PAGE HAS BEEN FOUND. [Loop invariant still holds -- the object lock
	 * is held.]
	 */

	/*
	 * If the page is being written, but isn't already owned by the
	 * top-level object, we have to copy it into a new page owned by the
	 * top-level object.
	 */

	if (fs.object != fs.first_object) {
		/*
		 * We only really need to copy if we want to write it.
		 */

		if (fault_type & VM_PROT_WRITE) {
			/*
			 * This allows pages to be virtually copied from a 
			 * backing_object into the first_object, where the 
			 * backing object has no other refs to it, and cannot
			 * gain any more refs.  Instead of a bcopy, we just 
			 * move the page from the backing object to the 
			 * first object.  Note that we must mark the page 
			 * dirty in the first object so that it will go out 
			 * to swap when needed.
			 */
			if (map_generation == fs.map->timestamp &&
				/*
				 * Only one shadow object
				 */
				(fs.object->shadow_count == 1) &&
				/*
				 * No COW refs, except us
				 */
				(fs.object->ref_count == 1) &&
				/*
				 * No one else can look this object up
				 */
				(fs.object->handle == NULL) &&
				/*
				 * No other ways to look the object up
				 */
				((fs.object->type == OBJT_DEFAULT) ||
				 (fs.object->type == OBJT_SWAP)) &&
				/*
				 * We don't chase down the shadow chain
				 */
				(fs.object == fs.first_object->backing_object) &&

				/*
				 * grab the lock if we need to
				 */
				(fs.lookup_still_valid ||
				 lockmgr(&fs.map->lock, LK_EXCLUSIVE|LK_NOWAIT, (void *)0, curproc) == 0)
			    ) {
				
				fs.lookup_still_valid = 1;
				/*
				 * get rid of the unnecessary page
				 */
				vm_page_protect(fs.first_m, VM_PROT_NONE);
				vm_page_free(fs.first_m);
				fs.first_m = NULL;

				/*
				 * grab the page and put it into the 
				 * process'es object.  The page is 
				 * automatically made dirty.
				 */
				vm_page_rename(fs.m, fs.first_object, fs.first_pindex);
				fs.first_m = fs.m;
				vm_page_busy(fs.first_m);
				fs.m = NULL;
				cnt.v_cow_optim++;
			} else {
				/*
				 * Oh, well, lets copy it.
				 */
				vm_page_copy(fs.m, fs.first_m);
			}

			if (fs.m) {
				/*
				 * We no longer need the old page or object.
				 */
				release_page(&fs);
			}

			/*
			 * fs.object != fs.first_object due to above 
			 * conditional
			 */

			vm_object_pip_wakeup(fs.object);

			/*
			 * Only use the new page below...
			 */

			cnt.v_cow_faults++;
			fs.m = fs.first_m;
			fs.object = fs.first_object;
			fs.pindex = fs.first_pindex;

		} else {
			prot &= ~VM_PROT_WRITE;
		}
	}

	/*
	 * We must verify that the maps have not changed since our last
	 * lookup.
	 */

	if (!fs.lookup_still_valid &&
		(fs.map->timestamp != map_generation)) {
		vm_object_t retry_object;
		vm_pindex_t retry_pindex;
		vm_prot_t retry_prot;

		/*
		 * Since map entries may be pageable, make sure we can take a
		 * page fault on them.
		 */

		/*
		 * Unlock vnode before the lookup to avoid deadlock.   E.G.
		 * avoid a deadlock between the inode and exec_map that can
		 * occur due to locks being obtained in different orders.
		 */

		if (fs.vp != NULL) {
			vput(fs.vp);
			fs.vp = NULL;
		}
		
		if (fs.map->infork) {
			release_page(&fs);
			unlock_and_deallocate(&fs);
			goto RetryFault;
		}

		/*
		 * To avoid trying to write_lock the map while another process
		 * has it read_locked (in vm_map_pageable), we do not try for
		 * write permission.  If the page is still writable, we will
		 * get write permission.  If it is not, or has been marked
		 * needs_copy, we enter the mapping without write permission,
		 * and will merely take another fault.
		 */
		result = vm_map_lookup(&fs.map, vaddr, fault_type & ~VM_PROT_WRITE,
		    &fs.entry, &retry_object, &retry_pindex, &retry_prot, &wired);
		map_generation = fs.map->timestamp;

		/*
		 * If we don't need the page any longer, put it on the active
		 * list (the easiest thing to do here).  If no one needs it,
		 * pageout will grab it eventually.
		 */

		if (result != KERN_SUCCESS) {
			release_page(&fs);
			unlock_and_deallocate(&fs);
			return (result);
		}
		fs.lookup_still_valid = TRUE;

		if ((retry_object != fs.first_object) ||
		    (retry_pindex != fs.first_pindex)) {
			release_page(&fs);
			unlock_and_deallocate(&fs);
			goto RetryFault;
		}
		/*
		 * Check whether the protection has changed or the object has
		 * been copied while we left the map unlocked. Changing from
		 * read to write permission is OK - we leave the page
		 * write-protected, and catch the write fault. Changing from
		 * write to read permission means that we can't mark the page
		 * write-enabled after all.
		 */
		prot &= retry_prot;
	}

	/*
	 * Put this page into the physical map. We had to do the unlock above
	 * because pmap_enter may cause other faults.   We don't put the page
	 * back on the active queue until later so that the page-out daemon
	 * won't find us (yet).
	 */

	if (prot & VM_PROT_WRITE) {
		vm_page_flag_set(fs.m, PG_WRITEABLE);
		vm_object_set_flag(fs.m->object,
				   OBJ_WRITEABLE|OBJ_MIGHTBEDIRTY);

		/*
		 * If the fault is a write, we know that this page is being
		 * written NOW so dirty it explicitly to save on 
		 * pmap_is_modified() calls later.
		 *
		 * If this is a NOSYNC mmap we do not want to set PG_NOSYNC
		 * if the page is already dirty to prevent data written with
		 * the expectation of being synced from not being synced.
		 * Likewise if this entry does not request NOSYNC then make
		 * sure the page isn't marked NOSYNC.  Applications sharing
		 * data should use the same flags to avoid ping ponging.
		 *
		 * Also tell the backing pager, if any, that it should remove
		 * any swap backing since the page is now dirty.
		 */
		if (fs.entry->eflags & MAP_ENTRY_NOSYNC) {
			if (fs.m->dirty == 0)
				vm_page_flag_set(fs.m, PG_NOSYNC);
		} else {
			vm_page_flag_clear(fs.m, PG_NOSYNC);
		}
		if (fault_flags & VM_FAULT_DIRTY) {
			int s;
			vm_page_dirty(fs.m);
			s = splvm();
			vm_pager_page_unswapped(fs.m);
			splx(s);
		}
	}

	/*
	 * Page had better still be busy
	 */

	KASSERT(fs.m->flags & PG_BUSY,
		("vm_fault: page %p not busy!", fs.m));

	unlock_things(&fs);

	/*
	 * Sanity check: page must be completely valid or it is not fit to
	 * map into user space.  vm_pager_get_pages() ensures this.
	 */

	if (fs.m->valid != VM_PAGE_BITS_ALL) {
		vm_page_zero_invalid(fs.m, TRUE);
		printf("Warning: page %p partially invalid on fault\n", fs.m);
	}

	pmap_enter(fs.map->pmap, vaddr, fs.m, prot, wired);

	if (((fault_flags & VM_FAULT_WIRE_MASK) == 0) && (wired == 0)) {
		pmap_prefault(fs.map->pmap, vaddr, fs.entry);
	}

	vm_page_flag_clear(fs.m, PG_ZERO);
	vm_page_flag_set(fs.m, PG_MAPPED|PG_REFERENCED);
	if (fault_flags & VM_FAULT_HOLD)
		vm_page_hold(fs.m);

	/*
	 * If the page is not wired down, then put it where the pageout daemon
	 * can find it.
	 */

	if (fault_flags & VM_FAULT_WIRE_MASK) {
		if (wired)
			vm_page_wire(fs.m);
		else
			vm_page_unwire(fs.m, 1);
	} else {
		vm_page_activate(fs.m);
	}

	mtx_lock_spin(&sched_lock);
	if (curproc && (curproc->p_sflag & PS_INMEM) && curproc->p_stats) {
		if (hardfault) {
			curproc->p_stats->p_ru.ru_majflt++;
		} else {
			curproc->p_stats->p_ru.ru_minflt++;
		}
	}
	mtx_unlock_spin(&sched_lock);

	/*
	 * Unlock everything, and return
	 */

	vm_page_wakeup(fs.m);
	vm_object_deallocate(fs.first_object);

	return (KERN_SUCCESS);

}

/*
 *	vm_fault_wire:
 *
 *	Wire down a range of virtual addresses in a map.
 */
int
vm_fault_wire(map, start, end)
	vm_map_t map;
	vm_offset_t start, end;
{

	vm_offset_t va;
	pmap_t pmap;
	int rv;

	pmap = vm_map_pmap(map);

	/*
	 * Inform the physical mapping system that the range of addresses may
	 * not fault, so that page tables and such can be locked down as well.
	 */

	pmap_pageable(pmap, start, end, FALSE);

	/*
	 * We simulate a fault to get the page and enter it in the physical
	 * map.
	 */

	for (va = start; va < end; va += PAGE_SIZE) {
		rv = vm_fault(map, va, VM_PROT_READ|VM_PROT_WRITE,
			VM_FAULT_CHANGE_WIRING);
		if (rv) {
			if (va != start)
				vm_fault_unwire(map, start, va);
			return (rv);
		}
	}
	return (KERN_SUCCESS);
}

/*
 *	vm_fault_user_wire:
 *
 *	Wire down a range of virtual addresses in a map.  This
 *	is for user mode though, so we only ask for read access
 *	on currently read only sections.
 */
int
vm_fault_user_wire(map, start, end)
	vm_map_t map;
	vm_offset_t start, end;
{

	vm_offset_t va;
	pmap_t pmap;
	int rv;

	GIANT_REQUIRED;

	pmap = vm_map_pmap(map);

	/*
	 * Inform the physical mapping system that the range of addresses may
	 * not fault, so that page tables and such can be locked down as well.
	 */

	pmap_pageable(pmap, start, end, FALSE);

	/*
	 * We simulate a fault to get the page and enter it in the physical
	 * map.
	 */
	for (va = start; va < end; va += PAGE_SIZE) {
		rv = vm_fault(map, va, VM_PROT_READ, VM_FAULT_USER_WIRE);
		if (rv) {
			if (va != start)
				vm_fault_unwire(map, start, va);
			return (rv);
		}
	}
	return (KERN_SUCCESS);
}


/*
 *	vm_fault_unwire:
 *
 *	Unwire a range of virtual addresses in a map.
 */
void
vm_fault_unwire(map, start, end)
	vm_map_t map;
	vm_offset_t start, end;
{

	vm_offset_t va, pa;
	pmap_t pmap;

	pmap = vm_map_pmap(map);

	/*
	 * Since the pages are wired down, we must be able to get their
	 * mappings from the physical map system.
	 */

	for (va = start; va < end; va += PAGE_SIZE) {
		pa = pmap_extract(pmap, va);
		if (pa != (vm_offset_t) 0) {
			pmap_change_wiring(pmap, va, FALSE);
			vm_page_unwire(PHYS_TO_VM_PAGE(pa), 1);
		}
	}

	/*
	 * Inform the physical mapping system that the range of addresses may
	 * fault, so that page tables and such may be unwired themselves.
	 */

	pmap_pageable(pmap, start, end, TRUE);

}

/*
 *	Routine:
 *		vm_fault_copy_entry
 *	Function:
 *		Copy all of the pages from a wired-down map entry to another.
 *
 *	In/out conditions:
 *		The source and destination maps must be locked for write.
 *		The source map entry must be wired down (or be a sharing map
 *		entry corresponding to a main map entry that is wired down).
 */

void
vm_fault_copy_entry(dst_map, src_map, dst_entry, src_entry)
	vm_map_t dst_map;
	vm_map_t src_map;
	vm_map_entry_t dst_entry;
	vm_map_entry_t src_entry;
{
	vm_object_t dst_object;
	vm_object_t src_object;
	vm_ooffset_t dst_offset;
	vm_ooffset_t src_offset;
	vm_prot_t prot;
	vm_offset_t vaddr;
	vm_page_t dst_m;
	vm_page_t src_m;

#ifdef	lint
	src_map++;
#endif	/* lint */

	src_object = src_entry->object.vm_object;
	src_offset = src_entry->offset;

	/*
	 * Create the top-level object for the destination entry. (Doesn't
	 * actually shadow anything - we copy the pages directly.)
	 */
	dst_object = vm_object_allocate(OBJT_DEFAULT,
	    (vm_size_t) OFF_TO_IDX(dst_entry->end - dst_entry->start));

	dst_entry->object.vm_object = dst_object;
	dst_entry->offset = 0;

	prot = dst_entry->max_protection;

	/*
	 * Loop through all of the pages in the entry's range, copying each
	 * one from the source object (it should be there) to the destination
	 * object.
	 */
	for (vaddr = dst_entry->start, dst_offset = 0;
	    vaddr < dst_entry->end;
	    vaddr += PAGE_SIZE, dst_offset += PAGE_SIZE) {

		/*
		 * Allocate a page in the destination object
		 */
		do {
			dst_m = vm_page_alloc(dst_object,
				OFF_TO_IDX(dst_offset), VM_ALLOC_NORMAL);
			if (dst_m == NULL) {
				VM_WAIT;
			}
		} while (dst_m == NULL);

		/*
		 * Find the page in the source object, and copy it in.
		 * (Because the source is wired down, the page will be in
		 * memory.)
		 */
		src_m = vm_page_lookup(src_object,
			OFF_TO_IDX(dst_offset + src_offset));
		if (src_m == NULL)
			panic("vm_fault_copy_wired: page missing");

		vm_page_copy(src_m, dst_m);

		/*
		 * Enter it in the pmap...
		 */

		vm_page_flag_clear(dst_m, PG_ZERO);
		pmap_enter(dst_map->pmap, vaddr, dst_m, prot, FALSE);
		vm_page_flag_set(dst_m, PG_WRITEABLE|PG_MAPPED);

		/*
		 * Mark it no longer busy, and put it on the active list.
		 */
		vm_page_activate(dst_m);
		vm_page_wakeup(dst_m);
	}
}


/*
 * This routine checks around the requested page for other pages that
 * might be able to be faulted in.  This routine brackets the viable
 * pages for the pages to be paged in.
 *
 * Inputs:
 *	m, rbehind, rahead
 *
 * Outputs:
 *  marray (array of vm_page_t), reqpage (index of requested page)
 *
 * Return value:
 *  number of pages in marray
 *
 * This routine can't block.
 */
static int
vm_fault_additional_pages(m, rbehind, rahead, marray, reqpage)
	vm_page_t m;
	int rbehind;
	int rahead;
	vm_page_t *marray;
	int *reqpage;
{
	int i,j;
	vm_object_t object;
	vm_pindex_t pindex, startpindex, endpindex, tpindex;
	vm_page_t rtm;
	int cbehind, cahead;

	GIANT_REQUIRED;

	object = m->object;
	pindex = m->pindex;

	/*
	 * we don't fault-ahead for device pager
	 */
	if (object->type == OBJT_DEVICE) {
		*reqpage = 0;
		marray[0] = m;
		return 1;
	}

	/*
	 * if the requested page is not available, then give up now
	 */

	if (!vm_pager_has_page(object, pindex, &cbehind, &cahead)) {
		return 0;
	}

	if ((cbehind == 0) && (cahead == 0)) {
		*reqpage = 0;
		marray[0] = m;
		return 1;
	}

	if (rahead > cahead) {
		rahead = cahead;
	}

	if (rbehind > cbehind) {
		rbehind = cbehind;
	}

	/*
	 * try to do any readahead that we might have free pages for.
	 */
	if ((rahead + rbehind) >
		((cnt.v_free_count + cnt.v_cache_count) - cnt.v_free_reserved)) {
		pagedaemon_wakeup();
		marray[0] = m;
		*reqpage = 0;
		return 1;
	}

	/*
	 * scan backward for the read behind pages -- in memory 
	 */
	if (pindex > 0) {
		if (rbehind > pindex) {
			rbehind = pindex;
			startpindex = 0;
		} else {
			startpindex = pindex - rbehind;
		}

		for ( tpindex = pindex - 1; tpindex >= startpindex; tpindex -= 1) {
			if (vm_page_lookup( object, tpindex)) {
				startpindex = tpindex + 1;
				break;
			}
			if (tpindex == 0)
				break;
		}

		for(i = 0, tpindex = startpindex; tpindex < pindex; i++, tpindex++) {

			rtm = vm_page_alloc(object, tpindex, VM_ALLOC_NORMAL);
			if (rtm == NULL) {
				for (j = 0; j < i; j++) {
					vm_page_free(marray[j]);
				}
				marray[0] = m;
				*reqpage = 0;
				return 1;
			}

			marray[i] = rtm;
		}
	} else {
		startpindex = 0;
		i = 0;
	}

	marray[i] = m;
	/* page offset of the required page */
	*reqpage = i;

	tpindex = pindex + 1;
	i++;

	/*
	 * scan forward for the read ahead pages
	 */
	endpindex = tpindex + rahead;
	if (endpindex > object->size)
		endpindex = object->size;

	for( ; tpindex < endpindex; i++, tpindex++) {

		if (vm_page_lookup(object, tpindex)) {
			break;
		}

		rtm = vm_page_alloc(object, tpindex, VM_ALLOC_NORMAL);
		if (rtm == NULL) {
			break;
		}

		marray[i] = rtm;
	}

	/* return number of bytes of pages */
	return i;
}
OpenPOWER on IntegriCloud