The Design and Implementation of the FreeBSD Operating System, Second Edition
Now available: The Design and Implementation of the FreeBSD Operating System (Second Edition)


[ source navigation ] [ diff markup ] [ identifier search ] [ freetext search ] [ file search ] [ list types ] [ track identifier ]

FreeBSD/Linux Kernel Cross Reference
sys/vm/vm_pager.c

Version: -  FREEBSD  -  FREEBSD-12-STABLE  -  FREEBSD-12-0  -  FREEBSD-11-STABLE  -  FREEBSD-11-2  -  FREEBSD-11-1  -  FREEBSD-11-0  -  FREEBSD-10-STABLE  -  FREEBSD-10-4  -  FREEBSD-10-3  -  FREEBSD-10-2  -  FREEBSD-10-1  -  FREEBSD-10-0  -  FREEBSD-9-STABLE  -  FREEBSD-9-3  -  FREEBSD-9-2  -  FREEBSD-9-1  -  FREEBSD-9-0  -  FREEBSD-8-STABLE  -  FREEBSD-8-4  -  FREEBSD-8-3  -  FREEBSD-8-2  -  FREEBSD-8-1  -  FREEBSD-8-0  -  FREEBSD-7-STABLE  -  FREEBSD-7-4  -  FREEBSD-7-3  -  FREEBSD-7-2  -  FREEBSD-7-1  -  FREEBSD-7-0  -  FREEBSD-6-STABLE  -  FREEBSD-6-4  -  FREEBSD-6-3  -  FREEBSD-6-2  -  FREEBSD-6-1  -  FREEBSD-6-0  -  FREEBSD-5-STABLE  -  FREEBSD-5-5  -  FREEBSD-5-4  -  FREEBSD-5-3  -  FREEBSD-5-2  -  FREEBSD-5-1  -  FREEBSD-5-0  -  FREEBSD-4-STABLE  -  FREEBSD-3-STABLE  -  FREEBSD22  -  linux-2.6  -  linux-2.4.22  -  MK83  -  MK84  -  PLAN9  -  DFBSD  -  NETBSD  -  NETBSD5  -  NETBSD4  -  NETBSD3  -  NETBSD20  -  OPENBSD  -  xnu-517  -  xnu-792  -  xnu-792.6.70  -  xnu-1228  -  xnu-1456.1.26  -  xnu-1699.24.8  -  xnu-2050.18.24  -  OPENSOLARIS  -  minix-3-1-1 
SearchContext: -  none  -  3  -  10 

    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  * 4. Neither the name of the University nor the names of its contributors
   17  *    may be used to endorse or promote products derived from this software
   18  *    without specific prior written permission.
   19  *
   20  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
   21  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
   22  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
   23  * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
   24  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
   25  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
   26  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
   27  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
   28  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
   29  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
   30  * SUCH DAMAGE.
   31  *
   32  *      from: @(#)vm_pager.c    8.6 (Berkeley) 1/12/94
   33  *
   34  *
   35  * Copyright (c) 1987, 1990 Carnegie-Mellon University.
   36  * All rights reserved.
   37  *
   38  * Authors: Avadis Tevanian, Jr., Michael Wayne Young
   39  *
   40  * Permission to use, copy, modify and distribute this software and
   41  * its documentation is hereby granted, provided that both the copyright
   42  * notice and this permission notice appear in all copies of the
   43  * software, derivative works or modified versions, and any portions
   44  * thereof, and that both notices appear in supporting documentation.
   45  *
   46  * CARNEGIE MELLON ALLOWS FREE USE OF THIS SOFTWARE IN ITS "AS IS"
   47  * CONDITION.  CARNEGIE MELLON DISCLAIMS ANY LIABILITY OF ANY KIND
   48  * FOR ANY DAMAGES WHATSOEVER RESULTING FROM THE USE OF THIS SOFTWARE.
   49  *
   50  * Carnegie Mellon requests users of this software to return to
   51  *
   52  *  Software Distribution Coordinator  or  Software.Distribution@CS.CMU.EDU
   53  *  School of Computer Science
   54  *  Carnegie Mellon University
   55  *  Pittsburgh PA 15213-3890
   56  *
   57  * any improvements or extensions that they make and grant Carnegie the
   58  * rights to redistribute these changes.
   59  */
   60 
   61 /*
   62  *      Paging space routine stubs.  Emulates a matchmaker-like interface
   63  *      for builtin pagers.
   64  */
   65 
   66 #include <sys/cdefs.h>
   67 __FBSDID("$FreeBSD: src/sys/vm/vm_pager.c,v 1.95.2.3 2005/08/16 15:09:44 sobomax Exp $");
   68 
   69 #include <sys/param.h>
   70 #include <sys/systm.h>
   71 #include <sys/kernel.h>
   72 #include <sys/vnode.h>
   73 #include <sys/bio.h>
   74 #include <sys/buf.h>
   75 #include <sys/ucred.h>
   76 #include <sys/malloc.h>
   77 
   78 #include <vm/vm.h>
   79 #include <vm/vm_param.h>
   80 #include <vm/vm_object.h>
   81 #include <vm/vm_page.h>
   82 #include <vm/vm_pager.h>
   83 #include <vm/vm_extern.h>
   84 
   85 MALLOC_DEFINE(M_VMPGDATA, "VM pgdata", "XXX: VM pager private data");
   86 
   87 int cluster_pbuf_freecnt = -1;  /* unlimited to begin with */
   88 
   89 static int dead_pager_getpages(vm_object_t, vm_page_t *, int, int);
   90 static vm_object_t dead_pager_alloc(void *, vm_ooffset_t, vm_prot_t,
   91         vm_ooffset_t);
   92 static void dead_pager_putpages(vm_object_t, vm_page_t *, int, int, int *);
   93 static boolean_t dead_pager_haspage(vm_object_t, vm_pindex_t, int *, int *);
   94 static void dead_pager_dealloc(vm_object_t);
   95 
   96 static int
   97 dead_pager_getpages(obj, ma, count, req)
   98         vm_object_t obj;
   99         vm_page_t *ma;
  100         int count;
  101         int req;
  102 {
  103         return VM_PAGER_FAIL;
  104 }
  105 
  106 static vm_object_t
  107 dead_pager_alloc(handle, size, prot, off)
  108         void *handle;
  109         vm_ooffset_t size;
  110         vm_prot_t prot;
  111         vm_ooffset_t off;
  112 {
  113         return NULL;
  114 }
  115 
  116 static void
  117 dead_pager_putpages(object, m, count, flags, rtvals)
  118         vm_object_t object;
  119         vm_page_t *m;
  120         int count;
  121         int flags;
  122         int *rtvals;
  123 {
  124         int i;
  125 
  126         for (i = 0; i < count; i++) {
  127                 rtvals[i] = VM_PAGER_AGAIN;
  128         }
  129 }
  130 
  131 static int
  132 dead_pager_haspage(object, pindex, prev, next)
  133         vm_object_t object;
  134         vm_pindex_t pindex;
  135         int *prev;
  136         int *next;
  137 {
  138         if (prev)
  139                 *prev = 0;
  140         if (next)
  141                 *next = 0;
  142         return FALSE;
  143 }
  144 
  145 static void
  146 dead_pager_dealloc(object)
  147         vm_object_t object;
  148 {
  149         return;
  150 }
  151 
  152 static struct pagerops deadpagerops = {
  153         .pgo_alloc =    dead_pager_alloc,
  154         .pgo_dealloc =  dead_pager_dealloc,
  155         .pgo_getpages = dead_pager_getpages,
  156         .pgo_putpages = dead_pager_putpages,
  157         .pgo_haspage =  dead_pager_haspage,
  158 };
  159 
  160 struct pagerops *pagertab[] = {
  161         &defaultpagerops,       /* OBJT_DEFAULT */
  162         &swappagerops,          /* OBJT_SWAP */
  163         &vnodepagerops,         /* OBJT_VNODE */
  164         &devicepagerops,        /* OBJT_DEVICE */
  165         &physpagerops,          /* OBJT_PHYS */
  166         &deadpagerops           /* OBJT_DEAD */
  167 };
  168 
  169 static const int npagers = sizeof(pagertab) / sizeof(pagertab[0]);
  170 
  171 /*
  172  * Kernel address space for mapping pages.
  173  * Used by pagers where KVAs are needed for IO.
  174  *
  175  * XXX needs to be large enough to support the number of pending async
  176  * cleaning requests (NPENDINGIO == 64) * the maximum swap cluster size
  177  * (MAXPHYS == 64k) if you want to get the most efficiency.
  178  */
  179 vm_map_t pager_map;
  180 static int bswneeded;
  181 static vm_offset_t swapbkva;            /* swap buffers kva */
  182 struct mtx pbuf_mtx;
  183 static TAILQ_HEAD(swqueue, buf) bswlist;
  184 
  185 void
  186 vm_pager_init()
  187 {
  188         struct pagerops **pgops;
  189 
  190         TAILQ_INIT(&bswlist);
  191         /*
  192          * Initialize known pagers
  193          */
  194         for (pgops = pagertab; pgops < &pagertab[npagers]; pgops++)
  195                 if (pgops && ((*pgops)->pgo_init != NULL))
  196                         (*(*pgops)->pgo_init) ();
  197 }
  198 
  199 void
  200 vm_pager_bufferinit()
  201 {
  202         struct buf *bp;
  203         int i;
  204 
  205         mtx_init(&pbuf_mtx, "pbuf mutex", NULL, MTX_DEF);
  206         bp = swbuf;
  207         /*
  208          * Now set up swap and physical I/O buffer headers.
  209          */
  210         for (i = 0; i < nswbuf; i++, bp++) {
  211                 TAILQ_INSERT_HEAD(&bswlist, bp, b_freelist);
  212                 BUF_LOCKINIT(bp);
  213                 LIST_INIT(&bp->b_dep);
  214                 bp->b_rcred = bp->b_wcred = NOCRED;
  215                 bp->b_xflags = 0;
  216         }
  217 
  218         cluster_pbuf_freecnt = nswbuf / 2;
  219         vnode_pbuf_freecnt = nswbuf / 2 + 1;
  220 
  221         swapbkva = kmem_alloc_nofault(pager_map, nswbuf * MAXPHYS);
  222         if (!swapbkva)
  223                 panic("Not enough pager_map VM space for physical buffers");
  224 }
  225 
  226 /*
  227  * Allocate an instance of a pager of the given type.
  228  * Size, protection and offset parameters are passed in for pagers that
  229  * need to perform page-level validation (e.g. the device pager).
  230  */
  231 vm_object_t
  232 vm_pager_allocate(objtype_t type, void *handle, vm_ooffset_t size,
  233                   vm_prot_t prot, vm_ooffset_t off)
  234 {
  235         vm_object_t ret;
  236         struct pagerops *ops;
  237 
  238         ops = pagertab[type];
  239         if (ops)
  240                 ret = (*ops->pgo_alloc) (handle, size, prot, off);
  241         else
  242                 ret = NULL;
  243         return (ret);
  244 }
  245 
  246 /*
  247  *      The object must be locked.
  248  */
  249 void
  250 vm_pager_deallocate(object)
  251         vm_object_t object;
  252 {
  253 
  254         VM_OBJECT_LOCK_ASSERT(object, MA_OWNED);
  255         (*pagertab[object->type]->pgo_dealloc) (object);
  256 }
  257 
  258 /*
  259  * vm_pager_get_pages() - inline, see vm/vm_pager.h
  260  * vm_pager_put_pages() - inline, see vm/vm_pager.h
  261  * vm_pager_has_page() - inline, see vm/vm_pager.h
  262  */
  263 
  264 vm_object_t
  265 vm_pager_object_lookup(pg_list, handle)
  266         struct pagerlst *pg_list;
  267         void *handle;
  268 {
  269         vm_object_t object;
  270 
  271         TAILQ_FOREACH(object, pg_list, pager_object_list)
  272                 if (object->handle == handle)
  273                         return (object);
  274         return (NULL);
  275 }
  276 
  277 /*
  278  * initialize a physical buffer
  279  */
  280 
  281 /*
  282  * XXX This probably belongs in vfs_bio.c
  283  */
  284 static void
  285 initpbuf(struct buf *bp)
  286 {
  287         bp->b_rcred = NOCRED;
  288         bp->b_wcred = NOCRED;
  289         bp->b_qindex = 0;       /* On no queue (QUEUE_NONE) */
  290         bp->b_saveaddr = (caddr_t) (MAXPHYS * (bp - swbuf)) + swapbkva;
  291         bp->b_data = bp->b_saveaddr;
  292         bp->b_kvabase = bp->b_saveaddr;
  293         bp->b_kvasize = MAXPHYS;
  294         bp->b_xflags = 0;
  295         bp->b_flags = 0;
  296         bp->b_ioflags = 0;
  297         bp->b_iodone = NULL;
  298         bp->b_error = 0;
  299         bp->b_magic = B_MAGIC_BIO;
  300         bp->b_op = &buf_ops_bio;
  301         BUF_LOCK(bp, LK_EXCLUSIVE, NULL);
  302 }
  303 
  304 /*
  305  * allocate a physical buffer
  306  *
  307  *      There are a limited number (nswbuf) of physical buffers.  We need
  308  *      to make sure that no single subsystem is able to hog all of them,
  309  *      so each subsystem implements a counter which is typically initialized
  310  *      to 1/2 nswbuf.  getpbuf() decrements this counter in allocation and
  311  *      increments it on release, and blocks if the counter hits zero.  A
  312  *      subsystem may initialize the counter to -1 to disable the feature,
  313  *      but it must still be sure to match up all uses of getpbuf() with 
  314  *      relpbuf() using the same variable.
  315  *
  316  *      NOTE: pfreecnt can be NULL, but this 'feature' will be removed
  317  *      relatively soon when the rest of the subsystems get smart about it. XXX
  318  */
  319 struct buf *
  320 getpbuf(int *pfreecnt)
  321 {
  322         int s;
  323         struct buf *bp;
  324 
  325         s = splvm();
  326         mtx_lock(&pbuf_mtx);
  327 
  328         for (;;) {
  329                 if (pfreecnt) {
  330                         while (*pfreecnt == 0) {
  331                                 msleep(pfreecnt, &pbuf_mtx, PVM, "wswbuf0", 0);
  332                         }
  333                 }
  334 
  335                 /* get a bp from the swap buffer header pool */
  336                 if ((bp = TAILQ_FIRST(&bswlist)) != NULL)
  337                         break;
  338 
  339                 bswneeded = 1;
  340                 msleep(&bswneeded, &pbuf_mtx, PVM, "wswbuf1", 0);
  341                 /* loop in case someone else grabbed one */
  342         }
  343         TAILQ_REMOVE(&bswlist, bp, b_freelist);
  344         if (pfreecnt)
  345                 --*pfreecnt;
  346         mtx_unlock(&pbuf_mtx);
  347         splx(s);
  348 
  349         initpbuf(bp);
  350         return bp;
  351 }
  352 
  353 /*
  354  * allocate a physical buffer, if one is available.
  355  *
  356  *      Note that there is no NULL hack here - all subsystems using this
  357  *      call understand how to use pfreecnt.
  358  */
  359 struct buf *
  360 trypbuf(int *pfreecnt)
  361 {
  362         int s;
  363         struct buf *bp;
  364 
  365         s = splvm();
  366         mtx_lock(&pbuf_mtx);
  367         if (*pfreecnt == 0 || (bp = TAILQ_FIRST(&bswlist)) == NULL) {
  368                 mtx_unlock(&pbuf_mtx);
  369                 splx(s);
  370                 return NULL;
  371         }
  372         TAILQ_REMOVE(&bswlist, bp, b_freelist);
  373 
  374         --*pfreecnt;
  375 
  376         mtx_unlock(&pbuf_mtx);
  377         splx(s);
  378 
  379         initpbuf(bp);
  380 
  381         return bp;
  382 }
  383 
  384 /*
  385  * release a physical buffer
  386  *
  387  *      NOTE: pfreecnt can be NULL, but this 'feature' will be removed
  388  *      relatively soon when the rest of the subsystems get smart about it. XXX
  389  */
  390 void
  391 relpbuf(struct buf *bp, int *pfreecnt)
  392 {
  393         int s;
  394 
  395         s = splvm();
  396 
  397         if (bp->b_rcred != NOCRED) {
  398                 crfree(bp->b_rcred);
  399                 bp->b_rcred = NOCRED;
  400         }
  401         if (bp->b_wcred != NOCRED) {
  402                 crfree(bp->b_wcred);
  403                 bp->b_wcred = NOCRED;
  404         }
  405 
  406         if (bp->b_vp)
  407                 pbrelvp(bp);
  408 
  409         BUF_UNLOCK(bp);
  410 
  411         mtx_lock(&pbuf_mtx);
  412         TAILQ_INSERT_HEAD(&bswlist, bp, b_freelist);
  413 
  414         if (bswneeded) {
  415                 bswneeded = 0;
  416                 wakeup(&bswneeded);
  417         }
  418         if (pfreecnt) {
  419                 if (++*pfreecnt == 1)
  420                         wakeup(pfreecnt);
  421         }
  422         mtx_unlock(&pbuf_mtx);
  423         splx(s);
  424 }

Cache object: 8fc6902425cdd664bdd008cfa329a428


[ source navigation ] [ diff markup ] [ identifier search ] [ freetext search ] [ file search ] [ list types ] [ track identifier ]


This page is part of the FreeBSD/Linux Linux Kernel Cross-Reference, and was automatically generated using a modified version of the LXR engine.