1/*
2 * Copyright (c) 1991, 1993
3 * The Regents of the University of California. All rights reserved.
4 *
5 * This code is derived from software contributed to Berkeley by
6 * The Mach Operating System project at Carnegie-Mellon University.
7 *
8 * Redistribution and use in source and binary forms, with or without
9 * modification, are permitted provided that the following conditions
10 * are met:
11 * 1. Redistributions of source code must retain the above copyright
12 * notice, this list of conditions and the following disclaimer.
13 * 2. Redistributions in binary form must reproduce the above copyright
14 * notice, this list of conditions and the following disclaimer in the
15 * documentation and/or other materials provided with the distribution.
16 * 3. All advertising materials mentioning features or use of this software
17 * must display the following acknowledgement:
18 * This product includes software developed by the University of
19 * California, Berkeley and its contributors.
20 * 4. Neither the name of the University nor the names of its contributors
21 * may be used to endorse or promote products derived from this software
22 * without specific prior written permission.
23 *
24 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
25 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
26 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
27 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
28 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
29 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
30 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
31 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
32 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
33 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
34 * SUCH DAMAGE.
35 *
36 * from: @(#)vm_page.h 8.2 (Berkeley) 12/13/93
37 *
38 *
39 * Copyright (c) 1987, 1990 Carnegie-Mellon University.
40 * All rights reserved.
41 *
42 * Authors: Avadis Tevanian, Jr., Michael Wayne Young
43 *
44 * Permission to use, copy, modify and distribute this software and
45 * its documentation is hereby granted, provided that both the copyright
46 * notice and this permission notice appear in all copies of the
47 * software, derivative works or modified versions, and any portions
48 * thereof, and that both notices appear in supporting documentation.
49 *
50 * CARNEGIE MELLON ALLOWS FREE USE OF THIS SOFTWARE IN ITS "AS IS"
51 * CONDITION. CARNEGIE MELLON DISCLAIMS ANY LIABILITY OF ANY KIND
52 * FOR ANY DAMAGES WHATSOEVER RESULTING FROM THE USE OF THIS SOFTWARE.
53 *
54 * Carnegie Mellon requests users of this software to return to
55 *
56 * Software Distribution Coordinator or Software.Distribution@CS.CMU.EDU
57 * School of Computer Science
58 * Carnegie Mellon University
59 * Pittsburgh PA 15213-3890
60 *
61 * any improvements or extensions that they make and grant Carnegie the
62 * rights to redistribute these changes.
63 *
|
64 * $FreeBSD: head/sys/vm/vm_page.h 102382 2002-08-25 00:22:31Z alc $
|
| 64 * $FreeBSD: head/sys/vm/vm_page.h 103531 2002-09-18 08:26:30Z jeff $ |
65 */
66
67/*
68 * Resident memory system definitions.
69 */
70
71#ifndef _VM_PAGE_
72#define _VM_PAGE_
73
74#if !defined(KLD_MODULE)
75#include "opt_vmpage.h"
76#endif
77
78#include <vm/pmap.h>
79
80/*
81 * Management of resident (logical) pages.
82 *
83 * A small structure is kept for each resident
84 * page, indexed by page number. Each structure
85 * is an element of several lists:
86 *
87 * A hash table bucket used to quickly
88 * perform object/offset lookups
89 *
90 * A list of all pages for a given object,
91 * so they can be quickly deactivated at
92 * time of deallocation.
93 *
94 * An ordered list of pages due for pageout.
95 *
96 * In addition, the structure contains the object
97 * and offset to which this page belongs (for pageout),
98 * and sundry status bits.
99 *
100 * Fields in this structure are locked either by the lock on the
101 * object that the page belongs to (O) or by the lock on the page
102 * queues (P).
103 *
104 * The 'valid' and 'dirty' fields are distinct. A page may have dirty
105 * bits set without having associated valid bits set. This is used by
106 * NFS to implement piecemeal writes.
107 */
108
109TAILQ_HEAD(pglist, vm_page);
110
111struct vm_page {
112 TAILQ_ENTRY(vm_page) pageq; /* queue info for FIFO queue or free list (P) */
113 struct vm_page *hnext; /* hash table link (O,P) */
114 TAILQ_ENTRY(vm_page) listq; /* pages in same object (O) */
115
116 vm_object_t object; /* which object am I in (O,P)*/
117 vm_pindex_t pindex; /* offset into object (O,P) */
118 vm_offset_t phys_addr; /* physical address of page */
119 struct md_page md; /* machine dependant stuff */
120 u_short queue; /* page queue index */
121 u_short flags, /* see below */
122 pc; /* page color */
123 u_short wire_count; /* wired down maps refs (P) */
124 short hold_count; /* page hold count */
125 u_char act_count; /* page usage count */
126 u_char busy; /* page busy count */
127 /* NOTE that these must support one bit per DEV_BSIZE in a page!!! */
128 /* so, on normal X86 kernels, they must be at least 8 bits wide */
129#if PAGE_SIZE == 4096
130 u_char valid; /* map of valid DEV_BSIZE chunks */
131 u_char dirty; /* map of dirty DEV_BSIZE chunks */
132#elif PAGE_SIZE == 8192
133 u_short valid; /* map of valid DEV_BSIZE chunks */
134 u_short dirty; /* map of dirty DEV_BSIZE chunks */
135#endif
136 u_int cow; /* page cow mapping count */
137};
138
139/*
140 * note: currently use SWAPBLK_NONE as an absolute value rather then
141 * a flag bit.
142 */
143
144#define SWAPBLK_MASK ((daddr_t)((u_daddr_t)-1 >> 1)) /* mask */
145#define SWAPBLK_NONE ((daddr_t)((u_daddr_t)SWAPBLK_MASK + 1))/* flag */
146
147#if !defined(KLD_MODULE)
148/*
149 * Page coloring parameters
150 */
151/* Each of PQ_FREE, and PQ_CACHE have PQ_HASH_SIZE entries */
152
153/* Backward compatibility for existing PQ_*CACHE config options. */
154#if !defined(PQ_CACHESIZE)
155#if defined(PQ_HUGECACHE)
156#define PQ_CACHESIZE 1024
157#elif defined(PQ_LARGECACHE)
158#define PQ_CACHESIZE 512
159#elif defined(PQ_MEDIUMCACHE)
160#define PQ_CACHESIZE 256
161#elif defined(PQ_NORMALCACHE)
162#define PQ_CACHESIZE 64
163#elif defined(PQ_NOOPT)
164#define PQ_CACHESIZE 0
165#else
166#define PQ_CACHESIZE 128
167#endif
168#endif /* !defined(PQ_CACHESIZE) */
169
170#if PQ_CACHESIZE >= 1024
171#define PQ_PRIME1 31 /* Prime number somewhat less than PQ_HASH_SIZE */
172#define PQ_PRIME2 23 /* Prime number somewhat less than PQ_HASH_SIZE */
173#define PQ_L2_SIZE 256 /* A number of colors opt for 1M cache */
174
175#elif PQ_CACHESIZE >= 512
176#define PQ_PRIME1 31 /* Prime number somewhat less than PQ_HASH_SIZE */
177#define PQ_PRIME2 23 /* Prime number somewhat less than PQ_HASH_SIZE */
178#define PQ_L2_SIZE 128 /* A number of colors opt for 512K cache */
179
180#elif PQ_CACHESIZE >= 256
181#define PQ_PRIME1 13 /* Prime number somewhat less than PQ_HASH_SIZE */
182#define PQ_PRIME2 7 /* Prime number somewhat less than PQ_HASH_SIZE */
183#define PQ_L2_SIZE 64 /* A number of colors opt for 256K cache */
184
185#elif PQ_CACHESIZE >= 128
186#define PQ_PRIME1 9 /* Produces a good PQ_L2_SIZE/3 + PQ_PRIME1 */
187#define PQ_PRIME2 5 /* Prime number somewhat less than PQ_HASH_SIZE */
188#define PQ_L2_SIZE 32 /* A number of colors opt for 128k cache */
189
190#elif PQ_CACHESIZE >= 64
191#define PQ_PRIME1 5 /* Prime number somewhat less than PQ_HASH_SIZE */
192#define PQ_PRIME2 3 /* Prime number somewhat less than PQ_HASH_SIZE */
193#define PQ_L2_SIZE 16 /* A reasonable number of colors (opt for 64K cache) */
194
195#else
196#define PQ_PRIME1 1 /* Disable page coloring. */
197#define PQ_PRIME2 1
198#define PQ_L2_SIZE 1
199
200#endif
201
202#define PQ_L2_MASK (PQ_L2_SIZE - 1)
203
204#define PQ_NONE 0
205#define PQ_FREE 1
206#define PQ_INACTIVE (1 + 1*PQ_L2_SIZE)
207#define PQ_ACTIVE (2 + 1*PQ_L2_SIZE)
208#define PQ_CACHE (3 + 1*PQ_L2_SIZE)
209#define PQ_HOLD (3 + 2*PQ_L2_SIZE)
210#define PQ_COUNT (4 + 2*PQ_L2_SIZE)
211
212struct vpgqueues {
213 struct pglist pl;
214 int *cnt;
215 int lcnt;
216};
217
218extern struct vpgqueues vm_page_queues[PQ_COUNT];
219extern struct mtx vm_page_queue_free_mtx;
220
221#endif /* !defined(KLD_MODULE) */
222
223/*
224 * These are the flags defined for vm_page.
225 *
226 * Note: PG_FILLED and PG_DIRTY are added for the filesystems.
227 *
228 * Note: PG_UNMANAGED (used by OBJT_PHYS) indicates that the page is
229 * not under PV management but otherwise should be treated as a
230 * normal page. Pages not under PV management cannot be paged out
231 * via the object/vm_page_t because there is no knowledge of their
232 * pte mappings, nor can they be removed from their objects via
233 * the object, and such pages are also not on any PQ queue.
234 */
235#define PG_BUSY 0x0001 /* page is in transit (O) */
236#define PG_WANTED 0x0002 /* someone is waiting for page (O) */
237#define PG_WINATCFLS 0x0004 /* flush dirty page on inactive q */
238#define PG_FICTITIOUS 0x0008 /* physical page doesn't exist (O) */
239#define PG_WRITEABLE 0x0010 /* page is mapped writeable */
240#define PG_ZERO 0x0040 /* page is zeroed */
241#define PG_REFERENCED 0x0080 /* page has been referenced */
242#define PG_CLEANCHK 0x0100 /* page will be checked for cleaning */
243#define PG_SWAPINPROG 0x0200 /* swap I/O in progress on page */
244#define PG_NOSYNC 0x0400 /* do not collect for syncer */
245#define PG_UNMANAGED 0x0800 /* No PV management for page */
246#define PG_MARKER 0x1000 /* special queue marker page */
|
| 247#define PG_SLAB 0x2000 /* object pointer is actually a slab */ |
248
249/*
250 * Misc constants.
251 */
252#define ACT_DECLINE 1
253#define ACT_ADVANCE 3
254#define ACT_INIT 5
255#define ACT_MAX 64
256#define PFCLUSTER_BEHIND 3
257#define PFCLUSTER_AHEAD 3
258
259#ifdef _KERNEL
260/*
261 * Each pageable resident page falls into one of four lists:
262 *
263 * free
264 * Available for allocation now.
265 *
266 * The following are all LRU sorted:
267 *
268 * cache
269 * Almost available for allocation. Still in an
270 * object, but clean and immediately freeable at
271 * non-interrupt times.
272 *
273 * inactive
274 * Low activity, candidates for reclamation.
275 * This is the list of pages that should be
276 * paged out next.
277 *
278 * active
279 * Pages that are "active" i.e. they have been
280 * recently referenced.
281 *
282 * zero
283 * Pages that are really free and have been pre-zeroed
284 *
285 */
286
287extern int vm_page_zero_count;
288
289extern vm_page_t vm_page_array; /* First resident page in table */
290extern int vm_page_array_size; /* number of vm_page_t's */
291extern long first_page; /* first physical page number */
292
293#define VM_PAGE_TO_PHYS(entry) ((entry)->phys_addr)
294
295#define PHYS_TO_VM_PAGE(pa) \
296 (&vm_page_array[atop(pa) - first_page ])
297
298extern struct mtx vm_page_queue_mtx;
299#define vm_page_lock_queues() mtx_lock(&vm_page_queue_mtx)
300#define vm_page_unlock_queues() mtx_unlock(&vm_page_queue_mtx)
301
302#if PAGE_SIZE == 4096
303#define VM_PAGE_BITS_ALL 0xff
304#endif
305
306#if PAGE_SIZE == 8192
307#define VM_PAGE_BITS_ALL 0xffff
308#endif
309
310/* page allocation classes: */
311#define VM_ALLOC_NORMAL 0
312#define VM_ALLOC_INTERRUPT 1
313#define VM_ALLOC_SYSTEM 2
314#define VM_ALLOC_CLASS_MASK 3
315/* page allocation flags: */
316#define VM_ALLOC_WIRED 0x20
317#define VM_ALLOC_ZERO 0x40
318#define VM_ALLOC_RETRY 0x80 /* vm_page_grab() only */
319
320void vm_page_flag_set(vm_page_t m, unsigned short bits);
321void vm_page_flag_clear(vm_page_t m, unsigned short bits);
322void vm_page_busy(vm_page_t m);
323void vm_page_flash(vm_page_t m);
324void vm_page_io_start(vm_page_t m);
325void vm_page_io_finish(vm_page_t m);
326void vm_page_hold(vm_page_t mem);
327void vm_page_unhold(vm_page_t mem);
328void vm_page_protect(vm_page_t mem, int prot);
329void vm_page_copy(vm_page_t src_m, vm_page_t dest_m);
330void vm_page_free(vm_page_t m);
331void vm_page_free_zero(vm_page_t m);
332int vm_page_sleep_busy(vm_page_t m, int also_m_busy, const char *msg);
333int vm_page_sleep_if_busy(vm_page_t m, int also_m_busy, const char *msg);
334void vm_page_dirty(vm_page_t m);
335void vm_page_undirty(vm_page_t m);
336void vm_page_wakeup(vm_page_t m);
337
338void vm_pageq_init(void);
339vm_page_t vm_pageq_add_new_page(vm_offset_t pa);
340void vm_pageq_enqueue(int queue, vm_page_t m);
341void vm_pageq_remove_nowakeup(vm_page_t m);
342void vm_pageq_remove(vm_page_t m);
343vm_page_t vm_pageq_find(int basequeue, int index, boolean_t prefer_zero);
344void vm_pageq_requeue(vm_page_t m);
345
346void vm_page_activate (vm_page_t);
347vm_page_t vm_page_alloc (vm_object_t, vm_pindex_t, int);
348vm_page_t vm_page_grab (vm_object_t, vm_pindex_t, int);
349void vm_page_cache (register vm_page_t);
350int vm_page_try_to_cache (vm_page_t);
351int vm_page_try_to_free (vm_page_t);
352void vm_page_dontneed (register vm_page_t);
353void vm_page_deactivate (vm_page_t);
354void vm_page_insert (vm_page_t, vm_object_t, vm_pindex_t);
355vm_page_t vm_page_lookup (vm_object_t, vm_pindex_t);
356void vm_page_remove (vm_page_t);
357void vm_page_rename (vm_page_t, vm_object_t, vm_pindex_t);
358vm_offset_t vm_page_startup (vm_offset_t, vm_offset_t, vm_offset_t);
359void vm_page_unmanage (vm_page_t);
360void vm_page_unwire (vm_page_t, int);
361void vm_page_wire (vm_page_t);
362void vm_page_set_validclean (vm_page_t, int, int);
363void vm_page_set_dirty (vm_page_t, int, int);
364void vm_page_clear_dirty (vm_page_t, int, int);
365void vm_page_set_invalid (vm_page_t, int, int);
366int vm_page_is_valid (vm_page_t, int, int);
367void vm_page_test_dirty (vm_page_t);
368int vm_page_bits (int, int);
369void vm_page_zero_invalid(vm_page_t m, boolean_t setvalid);
370void vm_page_free_toq(vm_page_t m);
371void vm_page_zero_idle_wakeup(void);
372void vm_page_cowfault (vm_page_t);
373void vm_page_cowsetup (vm_page_t);
374void vm_page_cowclear (vm_page_t);
375
376#endif /* _KERNEL */
377#endif /* !_VM_PAGE_ */
|