[BACK]Return to subr_pool.c CVS log [TXT][DIR] Up to [cvs.NetBSD.org] / src / sys / kern

Annotation of src/sys/kern/subr_pool.c, Revision 1.112.2.1

1.112.2.1! yamt        1: /*     $NetBSD: subr_pool.c,v 1.112 2006/02/24 11:46:20 bjh21 Exp $    */
1.1       pk          2:
                      3: /*-
1.43      thorpej     4:  * Copyright (c) 1997, 1999, 2000 The NetBSD Foundation, Inc.
1.1       pk          5:  * All rights reserved.
                      6:  *
                      7:  * This code is derived from software contributed to The NetBSD Foundation
1.20      thorpej     8:  * by Paul Kranenburg; by Jason R. Thorpe of the Numerical Aerospace
                      9:  * Simulation Facility, NASA Ames Research Center.
1.1       pk         10:  *
                     11:  * Redistribution and use in source and binary forms, with or without
                     12:  * modification, are permitted provided that the following conditions
                     13:  * are met:
                     14:  * 1. Redistributions of source code must retain the above copyright
                     15:  *    notice, this list of conditions and the following disclaimer.
                     16:  * 2. Redistributions in binary form must reproduce the above copyright
                     17:  *    notice, this list of conditions and the following disclaimer in the
                     18:  *    documentation and/or other materials provided with the distribution.
                     19:  * 3. All advertising materials mentioning features or use of this software
                     20:  *    must display the following acknowledgement:
1.13      christos   21:  *     This product includes software developed by the NetBSD
                     22:  *     Foundation, Inc. and its contributors.
1.1       pk         23:  * 4. Neither the name of The NetBSD Foundation nor the names of its
                     24:  *    contributors may be used to endorse or promote products derived
                     25:  *    from this software without specific prior written permission.
                     26:  *
                     27:  * THIS SOFTWARE IS PROVIDED BY THE NETBSD FOUNDATION, INC. AND CONTRIBUTORS
                     28:  * ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED
                     29:  * TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
                     30:  * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE FOUNDATION OR CONTRIBUTORS
                     31:  * BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
                     32:  * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
                     33:  * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
                     34:  * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
                     35:  * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
                     36:  * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
                     37:  * POSSIBILITY OF SUCH DAMAGE.
                     38:  */
1.64      lukem      39:
                     40: #include <sys/cdefs.h>
1.112.2.1! yamt       41: __KERNEL_RCSID(0, "$NetBSD: subr_pool.c,v 1.112 2006/02/24 11:46:20 bjh21 Exp $");
1.24      scottr     42:
1.25      thorpej    43: #include "opt_pool.h"
1.24      scottr     44: #include "opt_poollog.h"
1.28      thorpej    45: #include "opt_lockdebug.h"
1.1       pk         46:
                     47: #include <sys/param.h>
                     48: #include <sys/systm.h>
                     49: #include <sys/proc.h>
                     50: #include <sys/errno.h>
                     51: #include <sys/kernel.h>
                     52: #include <sys/malloc.h>
                     53: #include <sys/lock.h>
                     54: #include <sys/pool.h>
1.20      thorpej    55: #include <sys/syslog.h>
1.3       pk         56:
                     57: #include <uvm/uvm.h>
                     58:
1.1       pk         59: /*
                     60:  * Pool resource management utility.
1.3       pk         61:  *
1.88      chs        62:  * Memory is allocated in pages which are split into pieces according to
                     63:  * the pool item size. Each page is kept on one of three lists in the
                     64:  * pool structure: `pr_emptypages', `pr_fullpages' and `pr_partpages',
                     65:  * for empty, full and partially-full pages respectively. The individual
                     66:  * pool items are on a linked list headed by `ph_itemlist' in each page
                     67:  * header. The memory for building the page list is either taken from
                     68:  * the allocated pages themselves (for small pool items) or taken from
                     69:  * an internal pool of page headers (`phpool').
1.1       pk         70:  */
                     71:
1.3       pk         72: /* List of all pools */
1.102     chs        73: LIST_HEAD(,pool) pool_head = LIST_HEAD_INITIALIZER(pool_head);
1.3       pk         74:
                     75: /* Private pool for page header structures */
1.97      yamt       76: #define        PHPOOL_MAX      8
                     77: static struct pool phpool[PHPOOL_MAX];
                     78: #define        PHPOOL_FREELIST_NELEM(idx)      (((idx) == 0) ? 0 : (1 << (idx)))
1.3       pk         79:
1.62      bjh21      80: #ifdef POOL_SUBPAGE
                     81: /* Pool of subpages for use by normal pools. */
                     82: static struct pool psppool;
                     83: #endif
                     84:
1.98      yamt       85: static void *pool_page_alloc_meta(struct pool *, int);
                     86: static void pool_page_free_meta(struct pool *, void *);
                     87:
                     88: /* allocator for pool metadata */
                     89: static struct pool_allocator pool_allocator_meta = {
                     90:        pool_page_alloc_meta, pool_page_free_meta
                     91: };
                     92:
1.3       pk         93: /* # of seconds to retain page after last use */
                     94: int pool_inactive_time = 10;
                     95:
                     96: /* Next candidate for drainage (see pool_drain()) */
1.23      thorpej    97: static struct pool     *drainpp;
                     98:
                     99: /* This spin lock protects both pool_head and drainpp. */
                    100: struct simplelock pool_head_slock = SIMPLELOCK_INITIALIZER;
1.3       pk        101:
1.99      yamt      102: typedef uint8_t pool_item_freelist_t;
                    103:
1.3       pk        104: struct pool_item_header {
                    105:        /* Page headers */
1.88      chs       106:        LIST_ENTRY(pool_item_header)
1.3       pk        107:                                ph_pagelist;    /* pool page list */
1.88      chs       108:        SPLAY_ENTRY(pool_item_header)
                    109:                                ph_node;        /* Off-page page headers */
1.3       pk        110:        caddr_t                 ph_page;        /* this page's address */
                    111:        struct timeval          ph_time;        /* last referenced */
1.97      yamt      112:        union {
                    113:                /* !PR_NOTOUCH */
                    114:                struct {
1.102     chs       115:                        LIST_HEAD(, pool_item)
1.97      yamt      116:                                phu_itemlist;   /* chunk list for this page */
                    117:                } phu_normal;
                    118:                /* PR_NOTOUCH */
                    119:                struct {
                    120:                        uint16_t
                    121:                                phu_off;        /* start offset in page */
1.99      yamt      122:                        pool_item_freelist_t
1.97      yamt      123:                                phu_firstfree;  /* first free item */
1.99      yamt      124:                        /*
                    125:                         * XXX it might be better to use
                    126:                         * a simple bitmap and ffs(3)
                    127:                         */
1.97      yamt      128:                } phu_notouch;
                    129:        } ph_u;
                    130:        uint16_t                ph_nmissing;    /* # of chunks in use */
1.3       pk        131: };
1.97      yamt      132: #define        ph_itemlist     ph_u.phu_normal.phu_itemlist
                    133: #define        ph_off          ph_u.phu_notouch.phu_off
                    134: #define        ph_firstfree    ph_u.phu_notouch.phu_firstfree
1.3       pk        135:
1.1       pk        136: struct pool_item {
1.3       pk        137: #ifdef DIAGNOSTIC
1.82      thorpej   138:        u_int pi_magic;
1.33      chs       139: #endif
1.82      thorpej   140: #define        PI_MAGIC 0xdeadbeefU
1.3       pk        141:        /* Other entries use only this list entry */
1.102     chs       142:        LIST_ENTRY(pool_item)   pi_list;
1.3       pk        143: };
                    144:
1.53      thorpej   145: #define        POOL_NEEDS_CATCHUP(pp)                                          \
                    146:        ((pp)->pr_nitems < (pp)->pr_minitems)
                    147:
1.43      thorpej   148: /*
                    149:  * Pool cache management.
                    150:  *
                    151:  * Pool caches provide a way for constructed objects to be cached by the
                    152:  * pool subsystem.  This can lead to performance improvements by avoiding
                    153:  * needless object construction/destruction; it is deferred until absolutely
                    154:  * necessary.
                    155:  *
                    156:  * Caches are grouped into cache groups.  Each cache group references
                    157:  * up to 16 constructed objects.  When a cache allocates an object
                    158:  * from the pool, it calls the object's constructor and places it into
                    159:  * a cache group.  When a cache group frees an object back to the pool,
                    160:  * it first calls the object's destructor.  This allows the object to
                    161:  * persist in constructed form while freed to the cache.
                    162:  *
                    163:  * Multiple caches may exist for each pool.  This allows a single
                    164:  * object type to have multiple constructed forms.  The pool references
                    165:  * each cache, so that when a pool is drained by the pagedaemon, it can
                    166:  * drain each individual cache as well.  Each time a cache is drained,
                    167:  * the most idle cache group is freed to the pool in its entirety.
                    168:  *
                    169:  * Pool caches are layed on top of pools.  By layering them, we can avoid
                    170:  * the complexity of cache management for pools which would not benefit
                    171:  * from it.
                    172:  */
                    173:
                    174: /* The cache group pool. */
                    175: static struct pool pcgpool;
1.3       pk        176:
1.102     chs       177: static void    pool_cache_reclaim(struct pool_cache *, struct pool_pagelist *,
                    178:                                   struct pool_cache_grouplist *);
                    179: static void    pcg_grouplist_free(struct pool_cache_grouplist *);
1.3       pk        180:
1.42      thorpej   181: static int     pool_catchup(struct pool *);
1.55      thorpej   182: static void    pool_prime_page(struct pool *, caddr_t,
                    183:                    struct pool_item_header *);
1.88      chs       184: static void    pool_update_curpage(struct pool *);
1.66      thorpej   185:
1.112.2.1! yamt      186: static int     pool_grow(struct pool *, int);
1.66      thorpej   187: void           *pool_allocator_alloc(struct pool *, int);
                    188: void           pool_allocator_free(struct pool *, void *);
1.3       pk        189:
1.97      yamt      190: static void pool_print_pagelist(struct pool *, struct pool_pagelist *,
1.88      chs       191:        void (*)(const char *, ...));
1.42      thorpej   192: static void pool_print1(struct pool *, const char *,
                    193:        void (*)(const char *, ...));
1.3       pk        194:
1.88      chs       195: static int pool_chk_page(struct pool *, const char *,
                    196:                         struct pool_item_header *);
                    197:
1.3       pk        198: /*
1.52      thorpej   199:  * Pool log entry. An array of these is allocated in pool_init().
1.3       pk        200:  */
                    201: struct pool_log {
                    202:        const char      *pl_file;
                    203:        long            pl_line;
                    204:        int             pl_action;
1.25      thorpej   205: #define        PRLOG_GET       1
                    206: #define        PRLOG_PUT       2
1.3       pk        207:        void            *pl_addr;
1.1       pk        208: };
                    209:
1.86      matt      210: #ifdef POOL_DIAGNOSTIC
1.3       pk        211: /* Number of entries in pool log buffers */
1.17      thorpej   212: #ifndef POOL_LOGSIZE
                    213: #define        POOL_LOGSIZE    10
                    214: #endif
                    215:
                    216: int pool_logsize = POOL_LOGSIZE;
1.1       pk        217:
1.110     perry     218: static inline void
1.42      thorpej   219: pr_log(struct pool *pp, void *v, int action, const char *file, long line)
1.3       pk        220: {
                    221:        int n = pp->pr_curlogentry;
                    222:        struct pool_log *pl;
                    223:
1.20      thorpej   224:        if ((pp->pr_roflags & PR_LOGGING) == 0)
1.3       pk        225:                return;
                    226:
                    227:        /*
                    228:         * Fill in the current entry. Wrap around and overwrite
                    229:         * the oldest entry if necessary.
                    230:         */
                    231:        pl = &pp->pr_log[n];
                    232:        pl->pl_file = file;
                    233:        pl->pl_line = line;
                    234:        pl->pl_action = action;
                    235:        pl->pl_addr = v;
                    236:        if (++n >= pp->pr_logsize)
                    237:                n = 0;
                    238:        pp->pr_curlogentry = n;
                    239: }
                    240:
                    241: static void
1.42      thorpej   242: pr_printlog(struct pool *pp, struct pool_item *pi,
                    243:     void (*pr)(const char *, ...))
1.3       pk        244: {
                    245:        int i = pp->pr_logsize;
                    246:        int n = pp->pr_curlogentry;
                    247:
1.20      thorpej   248:        if ((pp->pr_roflags & PR_LOGGING) == 0)
1.3       pk        249:                return;
                    250:
                    251:        /*
                    252:         * Print all entries in this pool's log.
                    253:         */
                    254:        while (i-- > 0) {
                    255:                struct pool_log *pl = &pp->pr_log[n];
                    256:                if (pl->pl_action != 0) {
1.25      thorpej   257:                        if (pi == NULL || pi == pl->pl_addr) {
                    258:                                (*pr)("\tlog entry %d:\n", i);
                    259:                                (*pr)("\t\taction = %s, addr = %p\n",
                    260:                                    pl->pl_action == PRLOG_GET ? "get" : "put",
                    261:                                    pl->pl_addr);
                    262:                                (*pr)("\t\tfile: %s at line %lu\n",
                    263:                                    pl->pl_file, pl->pl_line);
                    264:                        }
1.3       pk        265:                }
                    266:                if (++n >= pp->pr_logsize)
                    267:                        n = 0;
                    268:        }
                    269: }
1.25      thorpej   270:
1.110     perry     271: static inline void
1.42      thorpej   272: pr_enter(struct pool *pp, const char *file, long line)
1.25      thorpej   273: {
                    274:
1.34      thorpej   275:        if (__predict_false(pp->pr_entered_file != NULL)) {
1.25      thorpej   276:                printf("pool %s: reentrancy at file %s line %ld\n",
                    277:                    pp->pr_wchan, file, line);
                    278:                printf("         previous entry at file %s line %ld\n",
                    279:                    pp->pr_entered_file, pp->pr_entered_line);
                    280:                panic("pr_enter");
                    281:        }
                    282:
                    283:        pp->pr_entered_file = file;
                    284:        pp->pr_entered_line = line;
                    285: }
                    286:
1.110     perry     287: static inline void
1.42      thorpej   288: pr_leave(struct pool *pp)
1.25      thorpej   289: {
                    290:
1.34      thorpej   291:        if (__predict_false(pp->pr_entered_file == NULL)) {
1.25      thorpej   292:                printf("pool %s not entered?\n", pp->pr_wchan);
                    293:                panic("pr_leave");
                    294:        }
                    295:
                    296:        pp->pr_entered_file = NULL;
                    297:        pp->pr_entered_line = 0;
                    298: }
                    299:
1.110     perry     300: static inline void
1.42      thorpej   301: pr_enter_check(struct pool *pp, void (*pr)(const char *, ...))
1.25      thorpej   302: {
                    303:
                    304:        if (pp->pr_entered_file != NULL)
                    305:                (*pr)("\n\tcurrently entered from file %s line %ld\n",
                    306:                    pp->pr_entered_file, pp->pr_entered_line);
                    307: }
1.3       pk        308: #else
1.25      thorpej   309: #define        pr_log(pp, v, action, file, line)
                    310: #define        pr_printlog(pp, pi, pr)
                    311: #define        pr_enter(pp, file, line)
                    312: #define        pr_leave(pp)
                    313: #define        pr_enter_check(pp, pr)
1.59      thorpej   314: #endif /* POOL_DIAGNOSTIC */
1.3       pk        315:
1.110     perry     316: static inline int
1.97      yamt      317: pr_item_notouch_index(const struct pool *pp, const struct pool_item_header *ph,
                    318:     const void *v)
                    319: {
                    320:        const char *cp = v;
                    321:        int idx;
                    322:
                    323:        KASSERT(pp->pr_roflags & PR_NOTOUCH);
                    324:        idx = (cp - ph->ph_page - ph->ph_off) / pp->pr_size;
                    325:        KASSERT(idx < pp->pr_itemsperpage);
                    326:        return idx;
                    327: }
                    328:
1.99      yamt      329: #define        PR_FREELIST_ALIGN(p) \
                    330:        roundup((uintptr_t)(p), sizeof(pool_item_freelist_t))
                    331: #define        PR_FREELIST(ph) ((pool_item_freelist_t *)PR_FREELIST_ALIGN((ph) + 1))
                    332: #define        PR_INDEX_USED   ((pool_item_freelist_t)-1)
                    333: #define        PR_INDEX_EOL    ((pool_item_freelist_t)-2)
1.97      yamt      334:
1.110     perry     335: static inline void
1.97      yamt      336: pr_item_notouch_put(const struct pool *pp, struct pool_item_header *ph,
                    337:     void *obj)
                    338: {
                    339:        int idx = pr_item_notouch_index(pp, ph, obj);
1.99      yamt      340:        pool_item_freelist_t *freelist = PR_FREELIST(ph);
1.97      yamt      341:
                    342:        KASSERT(freelist[idx] == PR_INDEX_USED);
                    343:        freelist[idx] = ph->ph_firstfree;
                    344:        ph->ph_firstfree = idx;
                    345: }
                    346:
1.110     perry     347: static inline void *
1.97      yamt      348: pr_item_notouch_get(const struct pool *pp, struct pool_item_header *ph)
                    349: {
                    350:        int idx = ph->ph_firstfree;
1.99      yamt      351:        pool_item_freelist_t *freelist = PR_FREELIST(ph);
1.97      yamt      352:
                    353:        KASSERT(freelist[idx] != PR_INDEX_USED);
                    354:        ph->ph_firstfree = freelist[idx];
                    355:        freelist[idx] = PR_INDEX_USED;
                    356:
                    357:        return ph->ph_page + ph->ph_off + idx * pp->pr_size;
                    358: }
                    359:
1.110     perry     360: static inline int
1.88      chs       361: phtree_compare(struct pool_item_header *a, struct pool_item_header *b)
                    362: {
                    363:        if (a->ph_page < b->ph_page)
                    364:                return (-1);
                    365:        else if (a->ph_page > b->ph_page)
                    366:                return (1);
                    367:        else
                    368:                return (0);
                    369: }
                    370:
                    371: SPLAY_PROTOTYPE(phtree, pool_item_header, ph_node, phtree_compare);
                    372: SPLAY_GENERATE(phtree, pool_item_header, ph_node, phtree_compare);
                    373:
1.3       pk        374: /*
                    375:  * Return the pool page header based on page address.
                    376:  */
1.110     perry     377: static inline struct pool_item_header *
1.42      thorpej   378: pr_find_pagehead(struct pool *pp, caddr_t page)
1.3       pk        379: {
1.88      chs       380:        struct pool_item_header *ph, tmp;
1.3       pk        381:
1.20      thorpej   382:        if ((pp->pr_roflags & PR_PHINPAGE) != 0)
1.3       pk        383:                return ((struct pool_item_header *)(page + pp->pr_phoffset));
                    384:
1.88      chs       385:        tmp.ph_page = page;
                    386:        ph = SPLAY_FIND(phtree, &pp->pr_phtree, &tmp);
                    387:        return ph;
1.3       pk        388: }
                    389:
1.101     thorpej   390: static void
                    391: pr_pagelist_free(struct pool *pp, struct pool_pagelist *pq)
                    392: {
                    393:        struct pool_item_header *ph;
                    394:        int s;
                    395:
                    396:        while ((ph = LIST_FIRST(pq)) != NULL) {
                    397:                LIST_REMOVE(ph, ph_pagelist);
                    398:                pool_allocator_free(pp, ph->ph_page);
                    399:                if ((pp->pr_roflags & PR_PHINPAGE) == 0) {
                    400:                        s = splvm();
                    401:                        pool_put(pp->pr_phpool, ph);
                    402:                        splx(s);
                    403:                }
                    404:        }
                    405: }
                    406:
1.3       pk        407: /*
                    408:  * Remove a page from the pool.
                    409:  */
1.110     perry     410: static inline void
1.61      chs       411: pr_rmpage(struct pool *pp, struct pool_item_header *ph,
                    412:      struct pool_pagelist *pq)
1.3       pk        413: {
                    414:
1.101     thorpej   415:        LOCK_ASSERT(simple_lock_held(&pp->pr_slock));
1.91      yamt      416:
1.3       pk        417:        /*
1.7       thorpej   418:         * If the page was idle, decrement the idle page count.
1.3       pk        419:         */
1.6       thorpej   420:        if (ph->ph_nmissing == 0) {
                    421: #ifdef DIAGNOSTIC
                    422:                if (pp->pr_nidle == 0)
                    423:                        panic("pr_rmpage: nidle inconsistent");
1.20      thorpej   424:                if (pp->pr_nitems < pp->pr_itemsperpage)
                    425:                        panic("pr_rmpage: nitems inconsistent");
1.6       thorpej   426: #endif
                    427:                pp->pr_nidle--;
                    428:        }
1.7       thorpej   429:
1.20      thorpej   430:        pp->pr_nitems -= pp->pr_itemsperpage;
                    431:
1.7       thorpej   432:        /*
1.101     thorpej   433:         * Unlink the page from the pool and queue it for release.
1.7       thorpej   434:         */
1.88      chs       435:        LIST_REMOVE(ph, ph_pagelist);
1.91      yamt      436:        if ((pp->pr_roflags & PR_PHINPAGE) == 0)
                    437:                SPLAY_REMOVE(phtree, &pp->pr_phtree, ph);
1.101     thorpej   438:        LIST_INSERT_HEAD(pq, ph, ph_pagelist);
                    439:
1.7       thorpej   440:        pp->pr_npages--;
                    441:        pp->pr_npagefree++;
1.6       thorpej   442:
1.88      chs       443:        pool_update_curpage(pp);
1.3       pk        444: }
                    445:
                    446: /*
1.94      simonb    447:  * Initialize all the pools listed in the "pools" link set.
                    448:  */
                    449: void
                    450: link_pool_init(void)
                    451: {
                    452:        __link_set_decl(pools, struct link_pool_init);
                    453:        struct link_pool_init * const *pi;
                    454:
                    455:        __link_set_foreach(pi, pools)
                    456:                pool_init((*pi)->pp, (*pi)->size, (*pi)->align,
                    457:                    (*pi)->align_offset, (*pi)->flags, (*pi)->wchan,
                    458:                    (*pi)->palloc);
                    459: }
                    460:
                    461: /*
1.3       pk        462:  * Initialize the given pool resource structure.
                    463:  *
                    464:  * We export this routine to allow other kernel parts to declare
                    465:  * static pools that must be initialized before malloc() is available.
                    466:  */
                    467: void
1.42      thorpej   468: pool_init(struct pool *pp, size_t size, u_int align, u_int ioff, int flags,
1.66      thorpej   469:     const char *wchan, struct pool_allocator *palloc)
1.3       pk        470: {
1.88      chs       471:        int off, slack;
1.92      enami     472:        size_t trysize, phsize;
1.93      dbj       473:        int s;
1.3       pk        474:
1.99      yamt      475:        KASSERT((1UL << (CHAR_BIT * sizeof(pool_item_freelist_t))) - 2 >=
                    476:            PHPOOL_FREELIST_NELEM(PHPOOL_MAX - 1));
                    477:
1.25      thorpej   478: #ifdef POOL_DIAGNOSTIC
                    479:        /*
                    480:         * Always log if POOL_DIAGNOSTIC is defined.
                    481:         */
                    482:        if (pool_logsize != 0)
                    483:                flags |= PR_LOGGING;
                    484: #endif
                    485:
1.66      thorpej   486:        if (palloc == NULL)
                    487:                palloc = &pool_allocator_kmem;
1.112     bjh21     488: #ifdef POOL_SUBPAGE
                    489:        if (size > palloc->pa_pagesz) {
                    490:                if (palloc == &pool_allocator_kmem)
                    491:                        palloc = &pool_allocator_kmem_fullpage;
                    492:                else if (palloc == &pool_allocator_nointr)
                    493:                        palloc = &pool_allocator_nointr_fullpage;
                    494:        }
1.66      thorpej   495: #endif /* POOL_SUBPAGE */
                    496:        if ((palloc->pa_flags & PA_INITIALIZED) == 0) {
1.112     bjh21     497:                if (palloc->pa_pagesz == 0)
1.66      thorpej   498:                        palloc->pa_pagesz = PAGE_SIZE;
                    499:
                    500:                TAILQ_INIT(&palloc->pa_list);
                    501:
                    502:                simple_lock_init(&palloc->pa_slock);
                    503:                palloc->pa_pagemask = ~(palloc->pa_pagesz - 1);
                    504:                palloc->pa_pageshift = ffs(palloc->pa_pagesz) - 1;
                    505:                palloc->pa_flags |= PA_INITIALIZED;
1.4       thorpej   506:        }
1.3       pk        507:
                    508:        if (align == 0)
                    509:                align = ALIGN(1);
1.14      thorpej   510:
                    511:        if (size < sizeof(struct pool_item))
                    512:                size = sizeof(struct pool_item);
1.3       pk        513:
1.78      thorpej   514:        size = roundup(size, align);
1.66      thorpej   515: #ifdef DIAGNOSTIC
                    516:        if (size > palloc->pa_pagesz)
1.35      pk        517:                panic("pool_init: pool item size (%lu) too large",
                    518:                      (u_long)size);
1.66      thorpej   519: #endif
1.35      pk        520:
1.3       pk        521:        /*
                    522:         * Initialize the pool structure.
                    523:         */
1.88      chs       524:        LIST_INIT(&pp->pr_emptypages);
                    525:        LIST_INIT(&pp->pr_fullpages);
                    526:        LIST_INIT(&pp->pr_partpages);
1.102     chs       527:        LIST_INIT(&pp->pr_cachelist);
1.3       pk        528:        pp->pr_curpage = NULL;
                    529:        pp->pr_npages = 0;
                    530:        pp->pr_minitems = 0;
                    531:        pp->pr_minpages = 0;
                    532:        pp->pr_maxpages = UINT_MAX;
1.20      thorpej   533:        pp->pr_roflags = flags;
                    534:        pp->pr_flags = 0;
1.35      pk        535:        pp->pr_size = size;
1.3       pk        536:        pp->pr_align = align;
                    537:        pp->pr_wchan = wchan;
1.66      thorpej   538:        pp->pr_alloc = palloc;
1.20      thorpej   539:        pp->pr_nitems = 0;
                    540:        pp->pr_nout = 0;
                    541:        pp->pr_hardlimit = UINT_MAX;
                    542:        pp->pr_hardlimit_warning = NULL;
1.31      thorpej   543:        pp->pr_hardlimit_ratecap.tv_sec = 0;
                    544:        pp->pr_hardlimit_ratecap.tv_usec = 0;
                    545:        pp->pr_hardlimit_warning_last.tv_sec = 0;
                    546:        pp->pr_hardlimit_warning_last.tv_usec = 0;
1.68      thorpej   547:        pp->pr_drain_hook = NULL;
                    548:        pp->pr_drain_hook_arg = NULL;
1.3       pk        549:
                    550:        /*
                    551:         * Decide whether to put the page header off page to avoid
1.92      enami     552:         * wasting too large a part of the page or too big item.
                    553:         * Off-page page headers go on a hash table, so we can match
                    554:         * a returned item with its header based on the page address.
                    555:         * We use 1/16 of the page size and about 8 times of the item
                    556:         * size as the threshold (XXX: tune)
                    557:         *
                    558:         * However, we'll put the header into the page if we can put
                    559:         * it without wasting any items.
                    560:         *
                    561:         * Silently enforce `0 <= ioff < align'.
1.3       pk        562:         */
1.92      enami     563:        pp->pr_itemoffset = ioff %= align;
                    564:        /* See the comment below about reserved bytes. */
                    565:        trysize = palloc->pa_pagesz - ((align - ioff) % align);
                    566:        phsize = ALIGN(sizeof(struct pool_item_header));
1.97      yamt      567:        if ((pp->pr_roflags & PR_NOTOUCH) == 0 &&
                    568:            (pp->pr_size < MIN(palloc->pa_pagesz / 16, phsize << 3) ||
                    569:            trysize / pp->pr_size == (trysize - phsize) / pp->pr_size)) {
1.3       pk        570:                /* Use the end of the page for the page header */
1.20      thorpej   571:                pp->pr_roflags |= PR_PHINPAGE;
1.92      enami     572:                pp->pr_phoffset = off = palloc->pa_pagesz - phsize;
1.2       pk        573:        } else {
1.3       pk        574:                /* The page header will be taken from our page header pool */
                    575:                pp->pr_phoffset = 0;
1.66      thorpej   576:                off = palloc->pa_pagesz;
1.88      chs       577:                SPLAY_INIT(&pp->pr_phtree);
1.2       pk        578:        }
1.1       pk        579:
1.3       pk        580:        /*
                    581:         * Alignment is to take place at `ioff' within the item. This means
                    582:         * we must reserve up to `align - 1' bytes on the page to allow
                    583:         * appropriate positioning of each item.
                    584:         */
                    585:        pp->pr_itemsperpage = (off - ((align - ioff) % align)) / pp->pr_size;
1.43      thorpej   586:        KASSERT(pp->pr_itemsperpage != 0);
1.97      yamt      587:        if ((pp->pr_roflags & PR_NOTOUCH)) {
                    588:                int idx;
                    589:
                    590:                for (idx = 0; pp->pr_itemsperpage > PHPOOL_FREELIST_NELEM(idx);
                    591:                    idx++) {
                    592:                        /* nothing */
                    593:                }
                    594:                if (idx >= PHPOOL_MAX) {
                    595:                        /*
                    596:                         * if you see this panic, consider to tweak
                    597:                         * PHPOOL_MAX and PHPOOL_FREELIST_NELEM.
                    598:                         */
                    599:                        panic("%s: too large itemsperpage(%d) for PR_NOTOUCH",
                    600:                            pp->pr_wchan, pp->pr_itemsperpage);
                    601:                }
                    602:                pp->pr_phpool = &phpool[idx];
                    603:        } else if ((pp->pr_roflags & PR_PHINPAGE) == 0) {
                    604:                pp->pr_phpool = &phpool[0];
                    605:        }
                    606: #if defined(DIAGNOSTIC)
                    607:        else {
                    608:                pp->pr_phpool = NULL;
                    609:        }
                    610: #endif
1.3       pk        611:
                    612:        /*
                    613:         * Use the slack between the chunks and the page header
                    614:         * for "cache coloring".
                    615:         */
                    616:        slack = off - pp->pr_itemsperpage * pp->pr_size;
                    617:        pp->pr_maxcolor = (slack / align) * align;
                    618:        pp->pr_curcolor = 0;
                    619:
                    620:        pp->pr_nget = 0;
                    621:        pp->pr_nfail = 0;
                    622:        pp->pr_nput = 0;
                    623:        pp->pr_npagealloc = 0;
                    624:        pp->pr_npagefree = 0;
1.1       pk        625:        pp->pr_hiwat = 0;
1.8       thorpej   626:        pp->pr_nidle = 0;
1.3       pk        627:
1.59      thorpej   628: #ifdef POOL_DIAGNOSTIC
1.25      thorpej   629:        if (flags & PR_LOGGING) {
                    630:                if (kmem_map == NULL ||
                    631:                    (pp->pr_log = malloc(pool_logsize * sizeof(struct pool_log),
                    632:                     M_TEMP, M_NOWAIT)) == NULL)
1.20      thorpej   633:                        pp->pr_roflags &= ~PR_LOGGING;
1.3       pk        634:                pp->pr_curlogentry = 0;
                    635:                pp->pr_logsize = pool_logsize;
                    636:        }
1.59      thorpej   637: #endif
1.25      thorpej   638:
                    639:        pp->pr_entered_file = NULL;
                    640:        pp->pr_entered_line = 0;
1.3       pk        641:
1.21      thorpej   642:        simple_lock_init(&pp->pr_slock);
1.1       pk        643:
1.3       pk        644:        /*
1.43      thorpej   645:         * Initialize private page header pool and cache magazine pool if we
                    646:         * haven't done so yet.
1.23      thorpej   647:         * XXX LOCKING.
1.3       pk        648:         */
1.97      yamt      649:        if (phpool[0].pr_size == 0) {
                    650:                int idx;
                    651:                for (idx = 0; idx < PHPOOL_MAX; idx++) {
                    652:                        static char phpool_names[PHPOOL_MAX][6+1+6+1];
                    653:                        int nelem;
                    654:                        size_t sz;
                    655:
                    656:                        nelem = PHPOOL_FREELIST_NELEM(idx);
                    657:                        snprintf(phpool_names[idx], sizeof(phpool_names[idx]),
                    658:                            "phpool-%d", nelem);
                    659:                        sz = sizeof(struct pool_item_header);
                    660:                        if (nelem) {
                    661:                                sz = PR_FREELIST_ALIGN(sz)
1.99      yamt      662:                                    + nelem * sizeof(pool_item_freelist_t);
1.97      yamt      663:                        }
                    664:                        pool_init(&phpool[idx], sz, 0, 0, 0,
1.98      yamt      665:                            phpool_names[idx], &pool_allocator_meta);
1.97      yamt      666:                }
1.62      bjh21     667: #ifdef POOL_SUBPAGE
                    668:                pool_init(&psppool, POOL_SUBPAGE, POOL_SUBPAGE, 0,
1.98      yamt      669:                    PR_RECURSIVE, "psppool", &pool_allocator_meta);
1.62      bjh21     670: #endif
1.43      thorpej   671:                pool_init(&pcgpool, sizeof(struct pool_cache_group), 0, 0,
1.98      yamt      672:                    0, "pcgpool", &pool_allocator_meta);
1.1       pk        673:        }
                    674:
1.23      thorpej   675:        /* Insert into the list of all pools. */
                    676:        simple_lock(&pool_head_slock);
1.102     chs       677:        LIST_INSERT_HEAD(&pool_head, pp, pr_poollist);
1.23      thorpej   678:        simple_unlock(&pool_head_slock);
1.66      thorpej   679:
                    680:        /* Insert this into the list of pools using this allocator. */
1.93      dbj       681:        s = splvm();
1.66      thorpej   682:        simple_lock(&palloc->pa_slock);
                    683:        TAILQ_INSERT_TAIL(&palloc->pa_list, pp, pr_alloc_list);
                    684:        simple_unlock(&palloc->pa_slock);
1.93      dbj       685:        splx(s);
1.1       pk        686: }
                    687:
                    688: /*
                    689:  * De-commision a pool resource.
                    690:  */
                    691: void
1.42      thorpej   692: pool_destroy(struct pool *pp)
1.1       pk        693: {
1.101     thorpej   694:        struct pool_pagelist pq;
1.3       pk        695:        struct pool_item_header *ph;
1.93      dbj       696:        int s;
1.43      thorpej   697:
1.101     thorpej   698:        /* Remove from global pool list */
                    699:        simple_lock(&pool_head_slock);
1.102     chs       700:        LIST_REMOVE(pp, pr_poollist);
1.101     thorpej   701:        if (drainpp == pp)
                    702:                drainpp = NULL;
                    703:        simple_unlock(&pool_head_slock);
                    704:
                    705:        /* Remove this pool from its allocator's list of pools. */
1.93      dbj       706:        s = splvm();
1.66      thorpej   707:        simple_lock(&pp->pr_alloc->pa_slock);
                    708:        TAILQ_REMOVE(&pp->pr_alloc->pa_list, pp, pr_alloc_list);
                    709:        simple_unlock(&pp->pr_alloc->pa_slock);
1.93      dbj       710:        splx(s);
1.66      thorpej   711:
1.101     thorpej   712:        s = splvm();
                    713:        simple_lock(&pp->pr_slock);
                    714:
1.102     chs       715:        KASSERT(LIST_EMPTY(&pp->pr_cachelist));
1.3       pk        716:
                    717: #ifdef DIAGNOSTIC
1.20      thorpej   718:        if (pp->pr_nout != 0) {
1.25      thorpej   719:                pr_printlog(pp, NULL, printf);
1.80      provos    720:                panic("pool_destroy: pool busy: still out: %u",
1.20      thorpej   721:                    pp->pr_nout);
1.3       pk        722:        }
                    723: #endif
1.1       pk        724:
1.101     thorpej   725:        KASSERT(LIST_EMPTY(&pp->pr_fullpages));
                    726:        KASSERT(LIST_EMPTY(&pp->pr_partpages));
                    727:
1.3       pk        728:        /* Remove all pages */
1.101     thorpej   729:        LIST_INIT(&pq);
1.88      chs       730:        while ((ph = LIST_FIRST(&pp->pr_emptypages)) != NULL)
1.101     thorpej   731:                pr_rmpage(pp, ph, &pq);
                    732:
                    733:        simple_unlock(&pp->pr_slock);
                    734:        splx(s);
1.3       pk        735:
1.101     thorpej   736:        pr_pagelist_free(pp, &pq);
1.3       pk        737:
1.59      thorpej   738: #ifdef POOL_DIAGNOSTIC
1.20      thorpej   739:        if ((pp->pr_roflags & PR_LOGGING) != 0)
1.3       pk        740:                free(pp->pr_log, M_TEMP);
1.59      thorpej   741: #endif
1.1       pk        742: }
                    743:
1.68      thorpej   744: void
                    745: pool_set_drain_hook(struct pool *pp, void (*fn)(void *, int), void *arg)
                    746: {
                    747:
                    748:        /* XXX no locking -- must be used just after pool_init() */
                    749: #ifdef DIAGNOSTIC
                    750:        if (pp->pr_drain_hook != NULL)
                    751:                panic("pool_set_drain_hook(%s): already set", pp->pr_wchan);
                    752: #endif
                    753:        pp->pr_drain_hook = fn;
                    754:        pp->pr_drain_hook_arg = arg;
                    755: }
                    756:
1.88      chs       757: static struct pool_item_header *
1.55      thorpej   758: pool_alloc_item_header(struct pool *pp, caddr_t storage, int flags)
                    759: {
                    760:        struct pool_item_header *ph;
                    761:        int s;
                    762:
                    763:        LOCK_ASSERT(simple_lock_held(&pp->pr_slock) == 0);
                    764:
                    765:        if ((pp->pr_roflags & PR_PHINPAGE) != 0)
                    766:                ph = (struct pool_item_header *) (storage + pp->pr_phoffset);
                    767:        else {
1.85      pk        768:                s = splvm();
1.97      yamt      769:                ph = pool_get(pp->pr_phpool, flags);
1.55      thorpej   770:                splx(s);
                    771:        }
                    772:
                    773:        return (ph);
                    774: }
1.1       pk        775:
                    776: /*
1.3       pk        777:  * Grab an item from the pool; must be called at appropriate spl level
1.1       pk        778:  */
1.3       pk        779: void *
1.59      thorpej   780: #ifdef POOL_DIAGNOSTIC
1.42      thorpej   781: _pool_get(struct pool *pp, int flags, const char *file, long line)
1.56      sommerfe  782: #else
                    783: pool_get(struct pool *pp, int flags)
                    784: #endif
1.1       pk        785: {
                    786:        struct pool_item *pi;
1.3       pk        787:        struct pool_item_header *ph;
1.55      thorpej   788:        void *v;
1.1       pk        789:
1.2       pk        790: #ifdef DIAGNOSTIC
1.95      atatat    791:        if (__predict_false(pp->pr_itemsperpage == 0))
                    792:                panic("pool_get: pool %p: pr_itemsperpage is zero, "
                    793:                    "pool not initialized?", pp);
1.84      thorpej   794:        if (__predict_false(curlwp == NULL && doing_shutdown == 0 &&
1.37      sommerfe  795:                            (flags & PR_WAITOK) != 0))
1.77      matt      796:                panic("pool_get: %s: must have NOWAIT", pp->pr_wchan);
1.58      thorpej   797:
1.102     chs       798: #endif /* DIAGNOSTIC */
1.58      thorpej   799: #ifdef LOCKDEBUG
                    800:        if (flags & PR_WAITOK)
                    801:                simple_lock_only_held(NULL, "pool_get(PR_WAITOK)");
1.102     chs       802:        SCHED_ASSERT_UNLOCKED();
1.56      sommerfe  803: #endif
1.1       pk        804:
1.21      thorpej   805:        simple_lock(&pp->pr_slock);
1.25      thorpej   806:        pr_enter(pp, file, line);
1.20      thorpej   807:
                    808:  startover:
                    809:        /*
                    810:         * Check to see if we've reached the hard limit.  If we have,
                    811:         * and we can wait, then wait until an item has been returned to
                    812:         * the pool.
                    813:         */
                    814: #ifdef DIAGNOSTIC
1.34      thorpej   815:        if (__predict_false(pp->pr_nout > pp->pr_hardlimit)) {
1.25      thorpej   816:                pr_leave(pp);
1.21      thorpej   817:                simple_unlock(&pp->pr_slock);
1.20      thorpej   818:                panic("pool_get: %s: crossed hard limit", pp->pr_wchan);
                    819:        }
                    820: #endif
1.34      thorpej   821:        if (__predict_false(pp->pr_nout == pp->pr_hardlimit)) {
1.68      thorpej   822:                if (pp->pr_drain_hook != NULL) {
                    823:                        /*
                    824:                         * Since the drain hook is going to free things
                    825:                         * back to the pool, unlock, call the hook, re-lock,
                    826:                         * and check the hardlimit condition again.
                    827:                         */
                    828:                        pr_leave(pp);
                    829:                        simple_unlock(&pp->pr_slock);
                    830:                        (*pp->pr_drain_hook)(pp->pr_drain_hook_arg, flags);
                    831:                        simple_lock(&pp->pr_slock);
                    832:                        pr_enter(pp, file, line);
                    833:                        if (pp->pr_nout < pp->pr_hardlimit)
                    834:                                goto startover;
                    835:                }
                    836:
1.29      sommerfe  837:                if ((flags & PR_WAITOK) && !(flags & PR_LIMITFAIL)) {
1.20      thorpej   838:                        /*
                    839:                         * XXX: A warning isn't logged in this case.  Should
                    840:                         * it be?
                    841:                         */
                    842:                        pp->pr_flags |= PR_WANTED;
1.25      thorpej   843:                        pr_leave(pp);
1.40      sommerfe  844:                        ltsleep(pp, PSWP, pp->pr_wchan, 0, &pp->pr_slock);
1.25      thorpej   845:                        pr_enter(pp, file, line);
1.20      thorpej   846:                        goto startover;
                    847:                }
1.31      thorpej   848:
                    849:                /*
                    850:                 * Log a message that the hard limit has been hit.
                    851:                 */
                    852:                if (pp->pr_hardlimit_warning != NULL &&
                    853:                    ratecheck(&pp->pr_hardlimit_warning_last,
                    854:                              &pp->pr_hardlimit_ratecap))
                    855:                        log(LOG_ERR, "%s\n", pp->pr_hardlimit_warning);
1.21      thorpej   856:
                    857:                pp->pr_nfail++;
                    858:
1.25      thorpej   859:                pr_leave(pp);
1.21      thorpej   860:                simple_unlock(&pp->pr_slock);
1.20      thorpej   861:                return (NULL);
                    862:        }
                    863:
1.3       pk        864:        /*
                    865:         * The convention we use is that if `curpage' is not NULL, then
                    866:         * it points at a non-empty bucket. In particular, `curpage'
                    867:         * never points at a page header which has PR_PHINPAGE set and
                    868:         * has no items in its bucket.
                    869:         */
1.20      thorpej   870:        if ((ph = pp->pr_curpage) == NULL) {
1.112.2.1! yamt      871:                int error;
        !           872:
1.20      thorpej   873: #ifdef DIAGNOSTIC
                    874:                if (pp->pr_nitems != 0) {
1.21      thorpej   875:                        simple_unlock(&pp->pr_slock);
1.20      thorpej   876:                        printf("pool_get: %s: curpage NULL, nitems %u\n",
                    877:                            pp->pr_wchan, pp->pr_nitems);
1.80      provos    878:                        panic("pool_get: nitems inconsistent");
1.20      thorpej   879:                }
                    880: #endif
                    881:
1.21      thorpej   882:                /*
                    883:                 * Call the back-end page allocator for more memory.
                    884:                 * Release the pool lock, as the back-end page allocator
                    885:                 * may block.
                    886:                 */
1.25      thorpej   887:                pr_leave(pp);
1.112.2.1! yamt      888:                error = pool_grow(pp, flags);
        !           889:                pr_enter(pp, file, line);
        !           890:                if (error != 0) {
1.21      thorpej   891:                        /*
1.55      thorpej   892:                         * We were unable to allocate a page or item
                    893:                         * header, but we released the lock during
                    894:                         * allocation, so perhaps items were freed
                    895:                         * back to the pool.  Check for this case.
1.21      thorpej   896:                         */
                    897:                        if (pp->pr_curpage != NULL)
                    898:                                goto startover;
1.15      pk        899:
1.3       pk        900:                        if ((flags & PR_WAITOK) == 0) {
                    901:                                pp->pr_nfail++;
1.25      thorpej   902:                                pr_leave(pp);
1.21      thorpej   903:                                simple_unlock(&pp->pr_slock);
1.1       pk        904:                                return (NULL);
1.3       pk        905:                        }
                    906:
1.15      pk        907:                        /*
                    908:                         * Wait for items to be returned to this pool.
1.21      thorpej   909:                         *
1.109     christos  910:                         * wake up once a second and try again,
                    911:                         * as the check in pool_cache_put_paddr() is racy.
1.15      pk        912:                         */
1.1       pk        913:                        pp->pr_flags |= PR_WANTED;
1.66      thorpej   914:                        /* PA_WANTED is already set on the allocator. */
1.25      thorpej   915:                        pr_leave(pp);
1.109     christos  916:                        ltsleep(pp, PSWP, pp->pr_wchan, hz, &pp->pr_slock);
1.25      thorpej   917:                        pr_enter(pp, file, line);
1.1       pk        918:                }
1.3       pk        919:
1.20      thorpej   920:                /* Start the allocation process over. */
                    921:                goto startover;
1.3       pk        922:        }
1.97      yamt      923:        if (pp->pr_roflags & PR_NOTOUCH) {
                    924: #ifdef DIAGNOSTIC
                    925:                if (__predict_false(ph->ph_nmissing == pp->pr_itemsperpage)) {
                    926:                        pr_leave(pp);
                    927:                        simple_unlock(&pp->pr_slock);
                    928:                        panic("pool_get: %s: page empty", pp->pr_wchan);
                    929:                }
                    930: #endif
                    931:                v = pr_item_notouch_get(pp, ph);
                    932: #ifdef POOL_DIAGNOSTIC
                    933:                pr_log(pp, v, PRLOG_GET, file, line);
                    934: #endif
                    935:        } else {
1.102     chs       936:                v = pi = LIST_FIRST(&ph->ph_itemlist);
1.97      yamt      937:                if (__predict_false(v == NULL)) {
                    938:                        pr_leave(pp);
                    939:                        simple_unlock(&pp->pr_slock);
                    940:                        panic("pool_get: %s: page empty", pp->pr_wchan);
                    941:                }
1.20      thorpej   942: #ifdef DIAGNOSTIC
1.97      yamt      943:                if (__predict_false(pp->pr_nitems == 0)) {
                    944:                        pr_leave(pp);
                    945:                        simple_unlock(&pp->pr_slock);
                    946:                        printf("pool_get: %s: items on itemlist, nitems %u\n",
                    947:                            pp->pr_wchan, pp->pr_nitems);
                    948:                        panic("pool_get: nitems inconsistent");
                    949:                }
1.65      enami     950: #endif
1.56      sommerfe  951:
1.65      enami     952: #ifdef POOL_DIAGNOSTIC
1.97      yamt      953:                pr_log(pp, v, PRLOG_GET, file, line);
1.65      enami     954: #endif
1.3       pk        955:
1.65      enami     956: #ifdef DIAGNOSTIC
1.97      yamt      957:                if (__predict_false(pi->pi_magic != PI_MAGIC)) {
                    958:                        pr_printlog(pp, pi, printf);
                    959:                        panic("pool_get(%s): free list modified: "
                    960:                            "magic=%x; page %p; item addr %p\n",
                    961:                            pp->pr_wchan, pi->pi_magic, ph->ph_page, pi);
                    962:                }
1.3       pk        963: #endif
                    964:
1.97      yamt      965:                /*
                    966:                 * Remove from item list.
                    967:                 */
1.102     chs       968:                LIST_REMOVE(pi, pi_list);
1.97      yamt      969:        }
1.20      thorpej   970:        pp->pr_nitems--;
                    971:        pp->pr_nout++;
1.6       thorpej   972:        if (ph->ph_nmissing == 0) {
                    973: #ifdef DIAGNOSTIC
1.34      thorpej   974:                if (__predict_false(pp->pr_nidle == 0))
1.6       thorpej   975:                        panic("pool_get: nidle inconsistent");
                    976: #endif
                    977:                pp->pr_nidle--;
1.88      chs       978:
                    979:                /*
                    980:                 * This page was previously empty.  Move it to the list of
                    981:                 * partially-full pages.  This page is already curpage.
                    982:                 */
                    983:                LIST_REMOVE(ph, ph_pagelist);
                    984:                LIST_INSERT_HEAD(&pp->pr_partpages, ph, ph_pagelist);
1.6       thorpej   985:        }
1.3       pk        986:        ph->ph_nmissing++;
1.97      yamt      987:        if (ph->ph_nmissing == pp->pr_itemsperpage) {
1.21      thorpej   988: #ifdef DIAGNOSTIC
1.97      yamt      989:                if (__predict_false((pp->pr_roflags & PR_NOTOUCH) == 0 &&
1.102     chs       990:                    !LIST_EMPTY(&ph->ph_itemlist))) {
1.25      thorpej   991:                        pr_leave(pp);
1.21      thorpej   992:                        simple_unlock(&pp->pr_slock);
                    993:                        panic("pool_get: %s: nmissing inconsistent",
                    994:                            pp->pr_wchan);
                    995:                }
                    996: #endif
1.3       pk        997:                /*
1.88      chs       998:                 * This page is now full.  Move it to the full list
                    999:                 * and select a new current page.
1.3       pk       1000:                 */
1.88      chs      1001:                LIST_REMOVE(ph, ph_pagelist);
                   1002:                LIST_INSERT_HEAD(&pp->pr_fullpages, ph, ph_pagelist);
                   1003:                pool_update_curpage(pp);
1.1       pk       1004:        }
1.3       pk       1005:
                   1006:        pp->pr_nget++;
1.111     christos 1007:        pr_leave(pp);
1.20      thorpej  1008:
                   1009:        /*
                   1010:         * If we have a low water mark and we are now below that low
                   1011:         * water mark, add more items to the pool.
                   1012:         */
1.53      thorpej  1013:        if (POOL_NEEDS_CATCHUP(pp) && pool_catchup(pp) != 0) {
1.20      thorpej  1014:                /*
                   1015:                 * XXX: Should we log a warning?  Should we set up a timeout
                   1016:                 * to try again in a second or so?  The latter could break
                   1017:                 * a caller's assumptions about interrupt protection, etc.
                   1018:                 */
                   1019:        }
                   1020:
1.21      thorpej  1021:        simple_unlock(&pp->pr_slock);
1.1       pk       1022:        return (v);
                   1023: }
                   1024:
                   1025: /*
1.43      thorpej  1026:  * Internal version of pool_put().  Pool is already locked/entered.
1.1       pk       1027:  */
1.43      thorpej  1028: static void
1.101     thorpej  1029: pool_do_put(struct pool *pp, void *v, struct pool_pagelist *pq)
1.1       pk       1030: {
                   1031:        struct pool_item *pi = v;
1.3       pk       1032:        struct pool_item_header *ph;
                   1033:        caddr_t page;
1.21      thorpej  1034:        int s;
1.3       pk       1035:
1.61      chs      1036:        LOCK_ASSERT(simple_lock_held(&pp->pr_slock));
1.102     chs      1037:        SCHED_ASSERT_UNLOCKED();
1.61      chs      1038:
1.66      thorpej  1039:        page = (caddr_t)((u_long)v & pp->pr_alloc->pa_pagemask);
1.1       pk       1040:
1.30      thorpej  1041: #ifdef DIAGNOSTIC
1.34      thorpej  1042:        if (__predict_false(pp->pr_nout == 0)) {
1.30      thorpej  1043:                printf("pool %s: putting with none out\n",
                   1044:                    pp->pr_wchan);
                   1045:                panic("pool_put");
                   1046:        }
                   1047: #endif
1.3       pk       1048:
1.34      thorpej  1049:        if (__predict_false((ph = pr_find_pagehead(pp, page)) == NULL)) {
1.25      thorpej  1050:                pr_printlog(pp, NULL, printf);
1.3       pk       1051:                panic("pool_put: %s: page header missing", pp->pr_wchan);
                   1052:        }
1.28      thorpej  1053:
                   1054: #ifdef LOCKDEBUG
                   1055:        /*
                   1056:         * Check if we're freeing a locked simple lock.
                   1057:         */
                   1058:        simple_lock_freecheck((caddr_t)pi, ((caddr_t)pi) + pp->pr_size);
                   1059: #endif
1.3       pk       1060:
                   1061:        /*
                   1062:         * Return to item list.
                   1063:         */
1.97      yamt     1064:        if (pp->pr_roflags & PR_NOTOUCH) {
                   1065:                pr_item_notouch_put(pp, ph, v);
                   1066:        } else {
1.2       pk       1067: #ifdef DIAGNOSTIC
1.97      yamt     1068:                pi->pi_magic = PI_MAGIC;
1.3       pk       1069: #endif
1.32      chs      1070: #ifdef DEBUG
1.97      yamt     1071:                {
                   1072:                        int i, *ip = v;
1.32      chs      1073:
1.97      yamt     1074:                        for (i = 0; i < pp->pr_size / sizeof(int); i++) {
                   1075:                                *ip++ = PI_MAGIC;
                   1076:                        }
1.32      chs      1077:                }
                   1078: #endif
                   1079:
1.102     chs      1080:                LIST_INSERT_HEAD(&ph->ph_itemlist, pi, pi_list);
1.97      yamt     1081:        }
1.79      thorpej  1082:        KDASSERT(ph->ph_nmissing != 0);
1.3       pk       1083:        ph->ph_nmissing--;
                   1084:        pp->pr_nput++;
1.20      thorpej  1085:        pp->pr_nitems++;
                   1086:        pp->pr_nout--;
1.3       pk       1087:
                   1088:        /* Cancel "pool empty" condition if it exists */
                   1089:        if (pp->pr_curpage == NULL)
                   1090:                pp->pr_curpage = ph;
                   1091:
                   1092:        if (pp->pr_flags & PR_WANTED) {
                   1093:                pp->pr_flags &= ~PR_WANTED;
1.15      pk       1094:                if (ph->ph_nmissing == 0)
                   1095:                        pp->pr_nidle++;
1.3       pk       1096:                wakeup((caddr_t)pp);
                   1097:                return;
                   1098:        }
                   1099:
                   1100:        /*
1.88      chs      1101:         * If this page is now empty, do one of two things:
1.21      thorpej  1102:         *
1.88      chs      1103:         *      (1) If we have more pages than the page high water mark,
1.96      thorpej  1104:         *          free the page back to the system.  ONLY CONSIDER
1.90      thorpej  1105:         *          FREEING BACK A PAGE IF WE HAVE MORE THAN OUR MINIMUM PAGE
                   1106:         *          CLAIM.
1.21      thorpej  1107:         *
1.88      chs      1108:         *      (2) Otherwise, move the page to the empty page list.
                   1109:         *
                   1110:         * Either way, select a new current page (so we use a partially-full
                   1111:         * page if one is available).
1.3       pk       1112:         */
                   1113:        if (ph->ph_nmissing == 0) {
1.6       thorpej  1114:                pp->pr_nidle++;
1.90      thorpej  1115:                if (pp->pr_npages > pp->pr_minpages &&
                   1116:                    (pp->pr_npages > pp->pr_maxpages ||
                   1117:                     (pp->pr_alloc->pa_flags & PA_WANT) != 0)) {
1.101     thorpej  1118:                        pr_rmpage(pp, ph, pq);
1.3       pk       1119:                } else {
1.88      chs      1120:                        LIST_REMOVE(ph, ph_pagelist);
                   1121:                        LIST_INSERT_HEAD(&pp->pr_emptypages, ph, ph_pagelist);
1.3       pk       1122:
1.21      thorpej  1123:                        /*
                   1124:                         * Update the timestamp on the page.  A page must
                   1125:                         * be idle for some period of time before it can
                   1126:                         * be reclaimed by the pagedaemon.  This minimizes
                   1127:                         * ping-pong'ing for memory.
                   1128:                         */
                   1129:                        s = splclock();
                   1130:                        ph->ph_time = mono_time;
                   1131:                        splx(s);
1.1       pk       1132:                }
1.88      chs      1133:                pool_update_curpage(pp);
1.1       pk       1134:        }
1.88      chs      1135:
1.21      thorpej  1136:        /*
1.88      chs      1137:         * If the page was previously completely full, move it to the
                   1138:         * partially-full list and make it the current page.  The next
                   1139:         * allocation will get the item from this page, instead of
                   1140:         * further fragmenting the pool.
1.21      thorpej  1141:         */
                   1142:        else if (ph->ph_nmissing == (pp->pr_itemsperpage - 1)) {
1.88      chs      1143:                LIST_REMOVE(ph, ph_pagelist);
                   1144:                LIST_INSERT_HEAD(&pp->pr_partpages, ph, ph_pagelist);
1.21      thorpej  1145:                pp->pr_curpage = ph;
                   1146:        }
1.43      thorpej  1147: }
                   1148:
                   1149: /*
                   1150:  * Return resource to the pool; must be called at appropriate spl level
                   1151:  */
1.59      thorpej  1152: #ifdef POOL_DIAGNOSTIC
1.43      thorpej  1153: void
                   1154: _pool_put(struct pool *pp, void *v, const char *file, long line)
                   1155: {
1.101     thorpej  1156:        struct pool_pagelist pq;
                   1157:
                   1158:        LIST_INIT(&pq);
1.43      thorpej  1159:
                   1160:        simple_lock(&pp->pr_slock);
                   1161:        pr_enter(pp, file, line);
                   1162:
1.56      sommerfe 1163:        pr_log(pp, v, PRLOG_PUT, file, line);
                   1164:
1.101     thorpej  1165:        pool_do_put(pp, v, &pq);
1.21      thorpej  1166:
1.25      thorpej  1167:        pr_leave(pp);
1.21      thorpej  1168:        simple_unlock(&pp->pr_slock);
1.101     thorpej  1169:
1.102     chs      1170:        pr_pagelist_free(pp, &pq);
1.1       pk       1171: }
1.57      sommerfe 1172: #undef pool_put
1.59      thorpej  1173: #endif /* POOL_DIAGNOSTIC */
1.1       pk       1174:
1.56      sommerfe 1175: void
                   1176: pool_put(struct pool *pp, void *v)
                   1177: {
1.101     thorpej  1178:        struct pool_pagelist pq;
                   1179:
                   1180:        LIST_INIT(&pq);
1.56      sommerfe 1181:
                   1182:        simple_lock(&pp->pr_slock);
1.101     thorpej  1183:        pool_do_put(pp, v, &pq);
                   1184:        simple_unlock(&pp->pr_slock);
1.56      sommerfe 1185:
1.102     chs      1186:        pr_pagelist_free(pp, &pq);
1.56      sommerfe 1187: }
1.57      sommerfe 1188:
1.59      thorpej  1189: #ifdef POOL_DIAGNOSTIC
1.57      sommerfe 1190: #define                pool_put(h, v)  _pool_put((h), (v), __FILE__, __LINE__)
1.56      sommerfe 1191: #endif
1.74      thorpej  1192:
                   1193: /*
1.112.2.1! yamt     1194:  * pool_grow: grow a pool by a page.
        !          1195:  *
        !          1196:  * => called with pool locked.
        !          1197:  * => unlock and relock the pool.
        !          1198:  * => return with pool locked.
        !          1199:  */
        !          1200:
        !          1201: static int
        !          1202: pool_grow(struct pool *pp, int flags)
        !          1203: {
        !          1204:        struct pool_item_header *ph = NULL;
        !          1205:        char *cp;
        !          1206:
        !          1207:        simple_unlock(&pp->pr_slock);
        !          1208:        cp = pool_allocator_alloc(pp, flags);
        !          1209:        if (__predict_true(cp != NULL)) {
        !          1210:                ph = pool_alloc_item_header(pp, cp, flags);
        !          1211:        }
        !          1212:        if (__predict_false(cp == NULL || ph == NULL)) {
        !          1213:                if (cp != NULL) {
        !          1214:                        pool_allocator_free(pp, cp);
        !          1215:                }
        !          1216:                simple_lock(&pp->pr_slock);
        !          1217:                return ENOMEM;
        !          1218:        }
        !          1219:
        !          1220:        simple_lock(&pp->pr_slock);
        !          1221:        pool_prime_page(pp, cp, ph);
        !          1222:        pp->pr_npagealloc++;
        !          1223:        pp->pr_minpages++;
        !          1224:        return 0;
        !          1225: }
        !          1226:
        !          1227: /*
1.74      thorpej  1228:  * Add N items to the pool.
                   1229:  */
                   1230: int
                   1231: pool_prime(struct pool *pp, int n)
                   1232: {
1.75      simonb   1233:        int newpages;
1.112.2.1! yamt     1234:        int error = 0;
1.74      thorpej  1235:
                   1236:        simple_lock(&pp->pr_slock);
                   1237:
                   1238:        newpages = roundup(n, pp->pr_itemsperpage) / pp->pr_itemsperpage;
                   1239:
                   1240:        while (newpages-- > 0) {
1.112.2.1! yamt     1241:                error = pool_grow(pp, PR_NOWAIT);
        !          1242:                if (error) {
1.74      thorpej  1243:                        break;
                   1244:                }
                   1245:                pp->pr_minpages++;
                   1246:        }
                   1247:
                   1248:        if (pp->pr_minpages >= pp->pr_maxpages)
                   1249:                pp->pr_maxpages = pp->pr_minpages + 1;  /* XXX */
                   1250:
                   1251:        simple_unlock(&pp->pr_slock);
1.112.2.1! yamt     1252:        return error;
1.74      thorpej  1253: }
1.55      thorpej  1254:
                   1255: /*
1.3       pk       1256:  * Add a page worth of items to the pool.
1.21      thorpej  1257:  *
                   1258:  * Note, we must be called with the pool descriptor LOCKED.
1.3       pk       1259:  */
1.55      thorpej  1260: static void
                   1261: pool_prime_page(struct pool *pp, caddr_t storage, struct pool_item_header *ph)
1.3       pk       1262: {
                   1263:        struct pool_item *pi;
                   1264:        caddr_t cp = storage;
                   1265:        unsigned int align = pp->pr_align;
                   1266:        unsigned int ioff = pp->pr_itemoffset;
1.55      thorpej  1267:        int n;
1.89      yamt     1268:        int s;
1.36      pk       1269:
1.91      yamt     1270:        LOCK_ASSERT(simple_lock_held(&pp->pr_slock));
                   1271:
1.66      thorpej  1272: #ifdef DIAGNOSTIC
                   1273:        if (((u_long)cp & (pp->pr_alloc->pa_pagesz - 1)) != 0)
1.36      pk       1274:                panic("pool_prime_page: %s: unaligned page", pp->pr_wchan);
1.66      thorpej  1275: #endif
1.3       pk       1276:
                   1277:        /*
                   1278:         * Insert page header.
                   1279:         */
1.88      chs      1280:        LIST_INSERT_HEAD(&pp->pr_emptypages, ph, ph_pagelist);
1.102     chs      1281:        LIST_INIT(&ph->ph_itemlist);
1.3       pk       1282:        ph->ph_page = storage;
                   1283:        ph->ph_nmissing = 0;
1.89      yamt     1284:        s = splclock();
                   1285:        ph->ph_time = mono_time;
                   1286:        splx(s);
1.88      chs      1287:        if ((pp->pr_roflags & PR_PHINPAGE) == 0)
                   1288:                SPLAY_INSERT(phtree, &pp->pr_phtree, ph);
1.3       pk       1289:
1.6       thorpej  1290:        pp->pr_nidle++;
                   1291:
1.3       pk       1292:        /*
                   1293:         * Color this page.
                   1294:         */
                   1295:        cp = (caddr_t)(cp + pp->pr_curcolor);
                   1296:        if ((pp->pr_curcolor += align) > pp->pr_maxcolor)
                   1297:                pp->pr_curcolor = 0;
                   1298:
                   1299:        /*
                   1300:         * Adjust storage to apply aligment to `pr_itemoffset' in each item.
                   1301:         */
                   1302:        if (ioff != 0)
                   1303:                cp = (caddr_t)(cp + (align - ioff));
                   1304:
                   1305:        /*
                   1306:         * Insert remaining chunks on the bucket list.
                   1307:         */
                   1308:        n = pp->pr_itemsperpage;
1.20      thorpej  1309:        pp->pr_nitems += n;
1.3       pk       1310:
1.97      yamt     1311:        if (pp->pr_roflags & PR_NOTOUCH) {
1.99      yamt     1312:                pool_item_freelist_t *freelist = PR_FREELIST(ph);
1.97      yamt     1313:                int i;
                   1314:
1.99      yamt     1315:                ph->ph_off = cp - storage;
1.97      yamt     1316:                ph->ph_firstfree = 0;
                   1317:                for (i = 0; i < n - 1; i++)
                   1318:                        freelist[i] = i + 1;
                   1319:                freelist[n - 1] = PR_INDEX_EOL;
                   1320:        } else {
                   1321:                while (n--) {
                   1322:                        pi = (struct pool_item *)cp;
1.78      thorpej  1323:
1.97      yamt     1324:                        KASSERT(((((vaddr_t)pi) + ioff) & (align - 1)) == 0);
1.3       pk       1325:
1.97      yamt     1326:                        /* Insert on page list */
1.102     chs      1327:                        LIST_INSERT_HEAD(&ph->ph_itemlist, pi, pi_list);
1.3       pk       1328: #ifdef DIAGNOSTIC
1.97      yamt     1329:                        pi->pi_magic = PI_MAGIC;
1.3       pk       1330: #endif
1.97      yamt     1331:                        cp = (caddr_t)(cp + pp->pr_size);
                   1332:                }
1.3       pk       1333:        }
                   1334:
                   1335:        /*
                   1336:         * If the pool was depleted, point at the new page.
                   1337:         */
                   1338:        if (pp->pr_curpage == NULL)
                   1339:                pp->pr_curpage = ph;
                   1340:
                   1341:        if (++pp->pr_npages > pp->pr_hiwat)
                   1342:                pp->pr_hiwat = pp->pr_npages;
                   1343: }
                   1344:
1.20      thorpej  1345: /*
1.52      thorpej  1346:  * Used by pool_get() when nitems drops below the low water mark.  This
1.88      chs      1347:  * is used to catch up pr_nitems with the low water mark.
1.20      thorpej  1348:  *
1.21      thorpej  1349:  * Note 1, we never wait for memory here, we let the caller decide what to do.
1.20      thorpej  1350:  *
1.73      thorpej  1351:  * Note 2, we must be called with the pool already locked, and we return
1.20      thorpej  1352:  * with it locked.
                   1353:  */
                   1354: static int
1.42      thorpej  1355: pool_catchup(struct pool *pp)
1.20      thorpej  1356: {
                   1357:        int error = 0;
                   1358:
1.54      thorpej  1359:        while (POOL_NEEDS_CATCHUP(pp)) {
1.112.2.1! yamt     1360:                error = pool_grow(pp, PR_NOWAIT);
        !          1361:                if (error) {
1.20      thorpej  1362:                        break;
                   1363:                }
                   1364:        }
1.112.2.1! yamt     1365:        return error;
1.20      thorpej  1366: }
                   1367:
1.88      chs      1368: static void
                   1369: pool_update_curpage(struct pool *pp)
                   1370: {
                   1371:
                   1372:        pp->pr_curpage = LIST_FIRST(&pp->pr_partpages);
                   1373:        if (pp->pr_curpage == NULL) {
                   1374:                pp->pr_curpage = LIST_FIRST(&pp->pr_emptypages);
                   1375:        }
                   1376: }
                   1377:
1.3       pk       1378: void
1.42      thorpej  1379: pool_setlowat(struct pool *pp, int n)
1.3       pk       1380: {
1.15      pk       1381:
1.21      thorpej  1382:        simple_lock(&pp->pr_slock);
                   1383:
1.3       pk       1384:        pp->pr_minitems = n;
1.15      pk       1385:        pp->pr_minpages = (n == 0)
                   1386:                ? 0
1.18      thorpej  1387:                : roundup(n, pp->pr_itemsperpage) / pp->pr_itemsperpage;
1.20      thorpej  1388:
                   1389:        /* Make sure we're caught up with the newly-set low water mark. */
1.75      simonb   1390:        if (POOL_NEEDS_CATCHUP(pp) && pool_catchup(pp) != 0) {
1.20      thorpej  1391:                /*
                   1392:                 * XXX: Should we log a warning?  Should we set up a timeout
                   1393:                 * to try again in a second or so?  The latter could break
                   1394:                 * a caller's assumptions about interrupt protection, etc.
                   1395:                 */
                   1396:        }
1.21      thorpej  1397:
                   1398:        simple_unlock(&pp->pr_slock);
1.3       pk       1399: }
                   1400:
                   1401: void
1.42      thorpej  1402: pool_sethiwat(struct pool *pp, int n)
1.3       pk       1403: {
1.15      pk       1404:
1.21      thorpej  1405:        simple_lock(&pp->pr_slock);
                   1406:
1.15      pk       1407:        pp->pr_maxpages = (n == 0)
                   1408:                ? 0
1.18      thorpej  1409:                : roundup(n, pp->pr_itemsperpage) / pp->pr_itemsperpage;
1.21      thorpej  1410:
                   1411:        simple_unlock(&pp->pr_slock);
1.3       pk       1412: }
                   1413:
1.20      thorpej  1414: void
1.42      thorpej  1415: pool_sethardlimit(struct pool *pp, int n, const char *warnmess, int ratecap)
1.20      thorpej  1416: {
                   1417:
1.21      thorpej  1418:        simple_lock(&pp->pr_slock);
1.20      thorpej  1419:
                   1420:        pp->pr_hardlimit = n;
                   1421:        pp->pr_hardlimit_warning = warnmess;
1.31      thorpej  1422:        pp->pr_hardlimit_ratecap.tv_sec = ratecap;
                   1423:        pp->pr_hardlimit_warning_last.tv_sec = 0;
                   1424:        pp->pr_hardlimit_warning_last.tv_usec = 0;
1.20      thorpej  1425:
                   1426:        /*
1.21      thorpej  1427:         * In-line version of pool_sethiwat(), because we don't want to
                   1428:         * release the lock.
1.20      thorpej  1429:         */
                   1430:        pp->pr_maxpages = (n == 0)
                   1431:                ? 0
                   1432:                : roundup(n, pp->pr_itemsperpage) / pp->pr_itemsperpage;
1.21      thorpej  1433:
                   1434:        simple_unlock(&pp->pr_slock);
1.20      thorpej  1435: }
1.3       pk       1436:
                   1437: /*
                   1438:  * Release all complete pages that have not been used recently.
                   1439:  */
1.66      thorpej  1440: int
1.59      thorpej  1441: #ifdef POOL_DIAGNOSTIC
1.42      thorpej  1442: _pool_reclaim(struct pool *pp, const char *file, long line)
1.56      sommerfe 1443: #else
                   1444: pool_reclaim(struct pool *pp)
                   1445: #endif
1.3       pk       1446: {
                   1447:        struct pool_item_header *ph, *phnext;
1.43      thorpej  1448:        struct pool_cache *pc;
1.61      chs      1449:        struct pool_pagelist pq;
1.102     chs      1450:        struct pool_cache_grouplist pcgl;
                   1451:        struct timeval curtime, diff;
1.21      thorpej  1452:        int s;
1.3       pk       1453:
1.68      thorpej  1454:        if (pp->pr_drain_hook != NULL) {
                   1455:                /*
                   1456:                 * The drain hook must be called with the pool unlocked.
                   1457:                 */
                   1458:                (*pp->pr_drain_hook)(pp->pr_drain_hook_arg, PR_NOWAIT);
                   1459:        }
                   1460:
1.21      thorpej  1461:        if (simple_lock_try(&pp->pr_slock) == 0)
1.66      thorpej  1462:                return (0);
1.25      thorpej  1463:        pr_enter(pp, file, line);
1.68      thorpej  1464:
1.88      chs      1465:        LIST_INIT(&pq);
1.102     chs      1466:        LIST_INIT(&pcgl);
1.3       pk       1467:
1.43      thorpej  1468:        /*
                   1469:         * Reclaim items from the pool's caches.
                   1470:         */
1.102     chs      1471:        LIST_FOREACH(pc, &pp->pr_cachelist, pc_poollist)
                   1472:                pool_cache_reclaim(pc, &pq, &pcgl);
1.43      thorpej  1473:
1.21      thorpej  1474:        s = splclock();
                   1475:        curtime = mono_time;
                   1476:        splx(s);
                   1477:
1.88      chs      1478:        for (ph = LIST_FIRST(&pp->pr_emptypages); ph != NULL; ph = phnext) {
                   1479:                phnext = LIST_NEXT(ph, ph_pagelist);
1.3       pk       1480:
                   1481:                /* Check our minimum page claim */
                   1482:                if (pp->pr_npages <= pp->pr_minpages)
                   1483:                        break;
                   1484:
1.88      chs      1485:                KASSERT(ph->ph_nmissing == 0);
                   1486:                timersub(&curtime, &ph->ph_time, &diff);
                   1487:                if (diff.tv_sec < pool_inactive_time)
                   1488:                        continue;
1.21      thorpej  1489:
1.88      chs      1490:                /*
                   1491:                 * If freeing this page would put us below
                   1492:                 * the low water mark, stop now.
                   1493:                 */
                   1494:                if ((pp->pr_nitems - pp->pr_itemsperpage) <
                   1495:                    pp->pr_minitems)
                   1496:                        break;
1.21      thorpej  1497:
1.88      chs      1498:                pr_rmpage(pp, ph, &pq);
1.3       pk       1499:        }
                   1500:
1.25      thorpej  1501:        pr_leave(pp);
1.21      thorpej  1502:        simple_unlock(&pp->pr_slock);
1.102     chs      1503:        if (LIST_EMPTY(&pq) && LIST_EMPTY(&pcgl))
                   1504:                return 0;
1.66      thorpej  1505:
1.101     thorpej  1506:        pr_pagelist_free(pp, &pq);
1.102     chs      1507:        pcg_grouplist_free(&pcgl);
1.66      thorpej  1508:        return (1);
1.3       pk       1509: }
                   1510:
                   1511: /*
                   1512:  * Drain pools, one at a time.
1.21      thorpej  1513:  *
                   1514:  * Note, we must never be called from an interrupt context.
1.3       pk       1515:  */
                   1516: void
1.42      thorpej  1517: pool_drain(void *arg)
1.3       pk       1518: {
                   1519:        struct pool *pp;
1.23      thorpej  1520:        int s;
1.3       pk       1521:
1.61      chs      1522:        pp = NULL;
1.49      thorpej  1523:        s = splvm();
1.23      thorpej  1524:        simple_lock(&pool_head_slock);
1.61      chs      1525:        if (drainpp == NULL) {
1.102     chs      1526:                drainpp = LIST_FIRST(&pool_head);
1.61      chs      1527:        }
                   1528:        if (drainpp) {
                   1529:                pp = drainpp;
1.102     chs      1530:                drainpp = LIST_NEXT(pp, pr_poollist);
1.61      chs      1531:        }
                   1532:        simple_unlock(&pool_head_slock);
1.63      chs      1533:        pool_reclaim(pp);
1.61      chs      1534:        splx(s);
1.3       pk       1535: }
                   1536:
                   1537: /*
                   1538:  * Diagnostic helpers.
                   1539:  */
                   1540: void
1.42      thorpej  1541: pool_print(struct pool *pp, const char *modif)
1.21      thorpej  1542: {
                   1543:        int s;
                   1544:
1.49      thorpej  1545:        s = splvm();
1.25      thorpej  1546:        if (simple_lock_try(&pp->pr_slock) == 0) {
                   1547:                printf("pool %s is locked; try again later\n",
                   1548:                    pp->pr_wchan);
                   1549:                splx(s);
                   1550:                return;
                   1551:        }
                   1552:        pool_print1(pp, modif, printf);
1.21      thorpej  1553:        simple_unlock(&pp->pr_slock);
                   1554:        splx(s);
                   1555: }
                   1556:
1.25      thorpej  1557: void
1.108     yamt     1558: pool_printall(const char *modif, void (*pr)(const char *, ...))
                   1559: {
                   1560:        struct pool *pp;
                   1561:
                   1562:        if (simple_lock_try(&pool_head_slock) == 0) {
                   1563:                (*pr)("WARNING: pool_head_slock is locked\n");
                   1564:        } else {
                   1565:                simple_unlock(&pool_head_slock);
                   1566:        }
                   1567:
                   1568:        LIST_FOREACH(pp, &pool_head, pr_poollist) {
                   1569:                pool_printit(pp, modif, pr);
                   1570:        }
                   1571: }
                   1572:
                   1573: void
1.42      thorpej  1574: pool_printit(struct pool *pp, const char *modif, void (*pr)(const char *, ...))
1.25      thorpej  1575: {
                   1576:
                   1577:        if (pp == NULL) {
                   1578:                (*pr)("Must specify a pool to print.\n");
                   1579:                return;
                   1580:        }
                   1581:
                   1582:        /*
                   1583:         * Called from DDB; interrupts should be blocked, and all
                   1584:         * other processors should be paused.  We can skip locking
                   1585:         * the pool in this case.
                   1586:         *
                   1587:         * We do a simple_lock_try() just to print the lock
                   1588:         * status, however.
                   1589:         */
                   1590:
                   1591:        if (simple_lock_try(&pp->pr_slock) == 0)
                   1592:                (*pr)("WARNING: pool %s is locked\n", pp->pr_wchan);
                   1593:        else
1.107     yamt     1594:                simple_unlock(&pp->pr_slock);
1.25      thorpej  1595:
                   1596:        pool_print1(pp, modif, pr);
                   1597: }
                   1598:
1.21      thorpej  1599: static void
1.97      yamt     1600: pool_print_pagelist(struct pool *pp, struct pool_pagelist *pl,
                   1601:     void (*pr)(const char *, ...))
1.88      chs      1602: {
                   1603:        struct pool_item_header *ph;
                   1604: #ifdef DIAGNOSTIC
                   1605:        struct pool_item *pi;
                   1606: #endif
                   1607:
                   1608:        LIST_FOREACH(ph, pl, ph_pagelist) {
                   1609:                (*pr)("\t\tpage %p, nmissing %d, time %lu,%lu\n",
                   1610:                    ph->ph_page, ph->ph_nmissing,
                   1611:                    (u_long)ph->ph_time.tv_sec,
                   1612:                    (u_long)ph->ph_time.tv_usec);
                   1613: #ifdef DIAGNOSTIC
1.97      yamt     1614:                if (!(pp->pr_roflags & PR_NOTOUCH)) {
1.102     chs      1615:                        LIST_FOREACH(pi, &ph->ph_itemlist, pi_list) {
1.97      yamt     1616:                                if (pi->pi_magic != PI_MAGIC) {
                   1617:                                        (*pr)("\t\t\titem %p, magic 0x%x\n",
                   1618:                                            pi, pi->pi_magic);
                   1619:                                }
1.88      chs      1620:                        }
                   1621:                }
                   1622: #endif
                   1623:        }
                   1624: }
                   1625:
                   1626: static void
1.42      thorpej  1627: pool_print1(struct pool *pp, const char *modif, void (*pr)(const char *, ...))
1.3       pk       1628: {
1.25      thorpej  1629:        struct pool_item_header *ph;
1.44      thorpej  1630:        struct pool_cache *pc;
                   1631:        struct pool_cache_group *pcg;
                   1632:        int i, print_log = 0, print_pagelist = 0, print_cache = 0;
1.25      thorpej  1633:        char c;
                   1634:
                   1635:        while ((c = *modif++) != '\0') {
                   1636:                if (c == 'l')
                   1637:                        print_log = 1;
                   1638:                if (c == 'p')
                   1639:                        print_pagelist = 1;
1.44      thorpej  1640:                if (c == 'c')
                   1641:                        print_cache = 1;
1.25      thorpej  1642:        }
                   1643:
                   1644:        (*pr)("POOL %s: size %u, align %u, ioff %u, roflags 0x%08x\n",
                   1645:            pp->pr_wchan, pp->pr_size, pp->pr_align, pp->pr_itemoffset,
                   1646:            pp->pr_roflags);
1.66      thorpej  1647:        (*pr)("\talloc %p\n", pp->pr_alloc);
1.25      thorpej  1648:        (*pr)("\tminitems %u, minpages %u, maxpages %u, npages %u\n",
                   1649:            pp->pr_minitems, pp->pr_minpages, pp->pr_maxpages, pp->pr_npages);
                   1650:        (*pr)("\titemsperpage %u, nitems %u, nout %u, hardlimit %u\n",
                   1651:            pp->pr_itemsperpage, pp->pr_nitems, pp->pr_nout, pp->pr_hardlimit);
                   1652:
                   1653:        (*pr)("\n\tnget %lu, nfail %lu, nput %lu\n",
                   1654:            pp->pr_nget, pp->pr_nfail, pp->pr_nput);
                   1655:        (*pr)("\tnpagealloc %lu, npagefree %lu, hiwat %u, nidle %lu\n",
                   1656:            pp->pr_npagealloc, pp->pr_npagefree, pp->pr_hiwat, pp->pr_nidle);
                   1657:
                   1658:        if (print_pagelist == 0)
                   1659:                goto skip_pagelist;
                   1660:
1.88      chs      1661:        if ((ph = LIST_FIRST(&pp->pr_emptypages)) != NULL)
                   1662:                (*pr)("\n\tempty page list:\n");
1.97      yamt     1663:        pool_print_pagelist(pp, &pp->pr_emptypages, pr);
1.88      chs      1664:        if ((ph = LIST_FIRST(&pp->pr_fullpages)) != NULL)
                   1665:                (*pr)("\n\tfull page list:\n");
1.97      yamt     1666:        pool_print_pagelist(pp, &pp->pr_fullpages, pr);
1.88      chs      1667:        if ((ph = LIST_FIRST(&pp->pr_partpages)) != NULL)
                   1668:                (*pr)("\n\tpartial-page list:\n");
1.97      yamt     1669:        pool_print_pagelist(pp, &pp->pr_partpages, pr);
1.88      chs      1670:
1.25      thorpej  1671:        if (pp->pr_curpage == NULL)
                   1672:                (*pr)("\tno current page\n");
                   1673:        else
                   1674:                (*pr)("\tcurpage %p\n", pp->pr_curpage->ph_page);
                   1675:
                   1676:  skip_pagelist:
                   1677:        if (print_log == 0)
                   1678:                goto skip_log;
                   1679:
                   1680:        (*pr)("\n");
                   1681:        if ((pp->pr_roflags & PR_LOGGING) == 0)
                   1682:                (*pr)("\tno log\n");
                   1683:        else
                   1684:                pr_printlog(pp, NULL, pr);
1.3       pk       1685:
1.25      thorpej  1686:  skip_log:
1.44      thorpej  1687:        if (print_cache == 0)
                   1688:                goto skip_cache;
                   1689:
1.102     chs      1690: #define PR_GROUPLIST(pcg)                                              \
                   1691:        (*pr)("\t\tgroup %p: avail %d\n", pcg, pcg->pcg_avail);         \
                   1692:        for (i = 0; i < PCG_NOBJECTS; i++) {                            \
                   1693:                if (pcg->pcg_objects[i].pcgo_pa !=                      \
                   1694:                    POOL_PADDR_INVALID) {                               \
                   1695:                        (*pr)("\t\t\t%p, 0x%llx\n",                     \
                   1696:                            pcg->pcg_objects[i].pcgo_va,                \
                   1697:                            (unsigned long long)                        \
                   1698:                            pcg->pcg_objects[i].pcgo_pa);               \
                   1699:                } else {                                                \
                   1700:                        (*pr)("\t\t\t%p\n",                             \
                   1701:                            pcg->pcg_objects[i].pcgo_va);               \
                   1702:                }                                                       \
                   1703:        }
                   1704:
                   1705:        LIST_FOREACH(pc, &pp->pr_cachelist, pc_poollist) {
1.103     chs      1706:                (*pr)("\tcache %p\n", pc);
1.48      thorpej  1707:                (*pr)("\t    hits %lu misses %lu ngroups %lu nitems %lu\n",
                   1708:                    pc->pc_hits, pc->pc_misses, pc->pc_ngroups, pc->pc_nitems);
1.102     chs      1709:                (*pr)("\t    full groups:\n");
1.103     chs      1710:                LIST_FOREACH(pcg, &pc->pc_fullgroups, pcg_list) {
1.102     chs      1711:                        PR_GROUPLIST(pcg);
1.103     chs      1712:                }
1.102     chs      1713:                (*pr)("\t    partial groups:\n");
1.103     chs      1714:                LIST_FOREACH(pcg, &pc->pc_partgroups, pcg_list) {
1.102     chs      1715:                        PR_GROUPLIST(pcg);
1.103     chs      1716:                }
1.102     chs      1717:                (*pr)("\t    empty groups:\n");
1.103     chs      1718:                LIST_FOREACH(pcg, &pc->pc_emptygroups, pcg_list) {
1.102     chs      1719:                        PR_GROUPLIST(pcg);
1.103     chs      1720:                }
1.44      thorpej  1721:        }
1.102     chs      1722: #undef PR_GROUPLIST
1.44      thorpej  1723:
                   1724:  skip_cache:
1.88      chs      1725:        pr_enter_check(pp, pr);
                   1726: }
                   1727:
                   1728: static int
                   1729: pool_chk_page(struct pool *pp, const char *label, struct pool_item_header *ph)
                   1730: {
                   1731:        struct pool_item *pi;
                   1732:        caddr_t page;
                   1733:        int n;
                   1734:
                   1735:        page = (caddr_t)((u_long)ph & pp->pr_alloc->pa_pagemask);
                   1736:        if (page != ph->ph_page &&
                   1737:            (pp->pr_roflags & PR_PHINPAGE) != 0) {
                   1738:                if (label != NULL)
                   1739:                        printf("%s: ", label);
                   1740:                printf("pool(%p:%s): page inconsistency: page %p;"
                   1741:                       " at page head addr %p (p %p)\n", pp,
                   1742:                        pp->pr_wchan, ph->ph_page,
                   1743:                        ph, page);
                   1744:                return 1;
                   1745:        }
1.3       pk       1746:
1.97      yamt     1747:        if ((pp->pr_roflags & PR_NOTOUCH) != 0)
                   1748:                return 0;
                   1749:
1.102     chs      1750:        for (pi = LIST_FIRST(&ph->ph_itemlist), n = 0;
1.88      chs      1751:             pi != NULL;
1.102     chs      1752:             pi = LIST_NEXT(pi,pi_list), n++) {
1.88      chs      1753:
                   1754: #ifdef DIAGNOSTIC
                   1755:                if (pi->pi_magic != PI_MAGIC) {
                   1756:                        if (label != NULL)
                   1757:                                printf("%s: ", label);
                   1758:                        printf("pool(%s): free list modified: magic=%x;"
                   1759:                               " page %p; item ordinal %d;"
                   1760:                               " addr %p (p %p)\n",
                   1761:                                pp->pr_wchan, pi->pi_magic, ph->ph_page,
                   1762:                                n, pi, page);
                   1763:                        panic("pool");
                   1764:                }
                   1765: #endif
                   1766:                page =
                   1767:                    (caddr_t)((u_long)pi & pp->pr_alloc->pa_pagemask);
                   1768:                if (page == ph->ph_page)
                   1769:                        continue;
                   1770:
                   1771:                if (label != NULL)
                   1772:                        printf("%s: ", label);
                   1773:                printf("pool(%p:%s): page inconsistency: page %p;"
                   1774:                       " item ordinal %d; addr %p (p %p)\n", pp,
                   1775:                        pp->pr_wchan, ph->ph_page,
                   1776:                        n, pi, page);
                   1777:                return 1;
                   1778:        }
                   1779:        return 0;
1.3       pk       1780: }
                   1781:
1.88      chs      1782:
1.3       pk       1783: int
1.42      thorpej  1784: pool_chk(struct pool *pp, const char *label)
1.3       pk       1785: {
                   1786:        struct pool_item_header *ph;
                   1787:        int r = 0;
                   1788:
1.21      thorpej  1789:        simple_lock(&pp->pr_slock);
1.88      chs      1790:        LIST_FOREACH(ph, &pp->pr_emptypages, ph_pagelist) {
                   1791:                r = pool_chk_page(pp, label, ph);
                   1792:                if (r) {
                   1793:                        goto out;
                   1794:                }
                   1795:        }
                   1796:        LIST_FOREACH(ph, &pp->pr_fullpages, ph_pagelist) {
                   1797:                r = pool_chk_page(pp, label, ph);
                   1798:                if (r) {
1.3       pk       1799:                        goto out;
                   1800:                }
1.88      chs      1801:        }
                   1802:        LIST_FOREACH(ph, &pp->pr_partpages, ph_pagelist) {
                   1803:                r = pool_chk_page(pp, label, ph);
                   1804:                if (r) {
1.3       pk       1805:                        goto out;
                   1806:                }
                   1807:        }
1.88      chs      1808:
1.3       pk       1809: out:
1.21      thorpej  1810:        simple_unlock(&pp->pr_slock);
1.3       pk       1811:        return (r);
1.43      thorpej  1812: }
                   1813:
                   1814: /*
                   1815:  * pool_cache_init:
                   1816:  *
                   1817:  *     Initialize a pool cache.
                   1818:  *
                   1819:  *     NOTE: If the pool must be protected from interrupts, we expect
                   1820:  *     to be called at the appropriate interrupt priority level.
                   1821:  */
                   1822: void
                   1823: pool_cache_init(struct pool_cache *pc, struct pool *pp,
                   1824:     int (*ctor)(void *, void *, int),
                   1825:     void (*dtor)(void *, void *),
                   1826:     void *arg)
                   1827: {
                   1828:
1.102     chs      1829:        LIST_INIT(&pc->pc_emptygroups);
                   1830:        LIST_INIT(&pc->pc_fullgroups);
                   1831:        LIST_INIT(&pc->pc_partgroups);
1.43      thorpej  1832:        simple_lock_init(&pc->pc_slock);
                   1833:
                   1834:        pc->pc_pool = pp;
                   1835:
                   1836:        pc->pc_ctor = ctor;
                   1837:        pc->pc_dtor = dtor;
                   1838:        pc->pc_arg  = arg;
                   1839:
1.48      thorpej  1840:        pc->pc_hits   = 0;
                   1841:        pc->pc_misses = 0;
                   1842:
                   1843:        pc->pc_ngroups = 0;
                   1844:
                   1845:        pc->pc_nitems = 0;
                   1846:
1.43      thorpej  1847:        simple_lock(&pp->pr_slock);
1.102     chs      1848:        LIST_INSERT_HEAD(&pp->pr_cachelist, pc, pc_poollist);
1.43      thorpej  1849:        simple_unlock(&pp->pr_slock);
                   1850: }
                   1851:
                   1852: /*
                   1853:  * pool_cache_destroy:
                   1854:  *
                   1855:  *     Destroy a pool cache.
                   1856:  */
                   1857: void
                   1858: pool_cache_destroy(struct pool_cache *pc)
                   1859: {
                   1860:        struct pool *pp = pc->pc_pool;
                   1861:
                   1862:        /* First, invalidate the entire cache. */
                   1863:        pool_cache_invalidate(pc);
                   1864:
                   1865:        /* ...and remove it from the pool's cache list. */
                   1866:        simple_lock(&pp->pr_slock);
1.102     chs      1867:        LIST_REMOVE(pc, pc_poollist);
1.43      thorpej  1868:        simple_unlock(&pp->pr_slock);
                   1869: }
                   1870:
1.110     perry    1871: static inline void *
1.87      thorpej  1872: pcg_get(struct pool_cache_group *pcg, paddr_t *pap)
1.43      thorpej  1873: {
                   1874:        void *object;
                   1875:        u_int idx;
                   1876:
                   1877:        KASSERT(pcg->pcg_avail <= PCG_NOBJECTS);
1.45      thorpej  1878:        KASSERT(pcg->pcg_avail != 0);
1.43      thorpej  1879:        idx = --pcg->pcg_avail;
                   1880:
1.87      thorpej  1881:        KASSERT(pcg->pcg_objects[idx].pcgo_va != NULL);
                   1882:        object = pcg->pcg_objects[idx].pcgo_va;
                   1883:        if (pap != NULL)
                   1884:                *pap = pcg->pcg_objects[idx].pcgo_pa;
                   1885:        pcg->pcg_objects[idx].pcgo_va = NULL;
1.43      thorpej  1886:
                   1887:        return (object);
                   1888: }
                   1889:
1.110     perry    1890: static inline void
1.87      thorpej  1891: pcg_put(struct pool_cache_group *pcg, void *object, paddr_t pa)
1.43      thorpej  1892: {
                   1893:        u_int idx;
                   1894:
                   1895:        KASSERT(pcg->pcg_avail < PCG_NOBJECTS);
                   1896:        idx = pcg->pcg_avail++;
                   1897:
1.87      thorpej  1898:        KASSERT(pcg->pcg_objects[idx].pcgo_va == NULL);
                   1899:        pcg->pcg_objects[idx].pcgo_va = object;
                   1900:        pcg->pcg_objects[idx].pcgo_pa = pa;
1.43      thorpej  1901: }
                   1902:
1.102     chs      1903: static void
                   1904: pcg_grouplist_free(struct pool_cache_grouplist *pcgl)
                   1905: {
                   1906:        struct pool_cache_group *pcg;
                   1907:        int s;
                   1908:
                   1909:        s = splvm();
                   1910:        while ((pcg = LIST_FIRST(pcgl)) != NULL) {
                   1911:                LIST_REMOVE(pcg, pcg_list);
                   1912:                pool_put(&pcgpool, pcg);
                   1913:        }
                   1914:        splx(s);
                   1915: }
                   1916:
1.43      thorpej  1917: /*
1.87      thorpej  1918:  * pool_cache_get{,_paddr}:
1.43      thorpej  1919:  *
1.87      thorpej  1920:  *     Get an object from a pool cache (optionally returning
                   1921:  *     the physical address of the object).
1.43      thorpej  1922:  */
                   1923: void *
1.87      thorpej  1924: pool_cache_get_paddr(struct pool_cache *pc, int flags, paddr_t *pap)
1.43      thorpej  1925: {
                   1926:        struct pool_cache_group *pcg;
                   1927:        void *object;
1.58      thorpej  1928:
                   1929: #ifdef LOCKDEBUG
                   1930:        if (flags & PR_WAITOK)
                   1931:                simple_lock_only_held(NULL, "pool_cache_get(PR_WAITOK)");
                   1932: #endif
1.43      thorpej  1933:
                   1934:        simple_lock(&pc->pc_slock);
                   1935:
1.102     chs      1936:        pcg = LIST_FIRST(&pc->pc_partgroups);
                   1937:        if (pcg == NULL) {
                   1938:                pcg = LIST_FIRST(&pc->pc_fullgroups);
                   1939:                if (pcg != NULL) {
                   1940:                        LIST_REMOVE(pcg, pcg_list);
                   1941:                        LIST_INSERT_HEAD(&pc->pc_partgroups, pcg, pcg_list);
1.43      thorpej  1942:                }
1.102     chs      1943:        }
                   1944:        if (pcg == NULL) {
1.43      thorpej  1945:
                   1946:                /*
                   1947:                 * No groups with any available objects.  Allocate
                   1948:                 * a new object, construct it, and return it to
                   1949:                 * the caller.  We will allocate a group, if necessary,
                   1950:                 * when the object is freed back to the cache.
                   1951:                 */
1.48      thorpej  1952:                pc->pc_misses++;
1.43      thorpej  1953:                simple_unlock(&pc->pc_slock);
                   1954:                object = pool_get(pc->pc_pool, flags);
                   1955:                if (object != NULL && pc->pc_ctor != NULL) {
                   1956:                        if ((*pc->pc_ctor)(pc->pc_arg, object, flags) != 0) {
                   1957:                                pool_put(pc->pc_pool, object);
                   1958:                                return (NULL);
                   1959:                        }
                   1960:                }
1.87      thorpej  1961:                if (object != NULL && pap != NULL) {
                   1962: #ifdef POOL_VTOPHYS
                   1963:                        *pap = POOL_VTOPHYS(object);
                   1964: #else
                   1965:                        *pap = POOL_PADDR_INVALID;
                   1966: #endif
                   1967:                }
1.43      thorpej  1968:                return (object);
                   1969:        }
                   1970:
1.48      thorpej  1971:        pc->pc_hits++;
                   1972:        pc->pc_nitems--;
1.87      thorpej  1973:        object = pcg_get(pcg, pap);
1.43      thorpej  1974:
1.102     chs      1975:        if (pcg->pcg_avail == 0) {
                   1976:                LIST_REMOVE(pcg, pcg_list);
                   1977:                LIST_INSERT_HEAD(&pc->pc_emptygroups, pcg, pcg_list);
                   1978:        }
1.43      thorpej  1979:        simple_unlock(&pc->pc_slock);
                   1980:
                   1981:        return (object);
                   1982: }
                   1983:
                   1984: /*
1.87      thorpej  1985:  * pool_cache_put{,_paddr}:
1.43      thorpej  1986:  *
1.87      thorpej  1987:  *     Put an object back to the pool cache (optionally caching the
                   1988:  *     physical address of the object).
1.43      thorpej  1989:  */
                   1990: void
1.87      thorpej  1991: pool_cache_put_paddr(struct pool_cache *pc, void *object, paddr_t pa)
1.43      thorpej  1992: {
                   1993:        struct pool_cache_group *pcg;
1.60      thorpej  1994:        int s;
1.43      thorpej  1995:
1.109     christos 1996:        if (__predict_false((pc->pc_pool->pr_flags & PR_WANTED) != 0)) {
                   1997:                goto destruct;
                   1998:        }
                   1999:
1.43      thorpej  2000:        simple_lock(&pc->pc_slock);
                   2001:
1.102     chs      2002:        pcg = LIST_FIRST(&pc->pc_partgroups);
                   2003:        if (pcg == NULL) {
                   2004:                pcg = LIST_FIRST(&pc->pc_emptygroups);
                   2005:                if (pcg != NULL) {
                   2006:                        LIST_REMOVE(pcg, pcg_list);
                   2007:                        LIST_INSERT_HEAD(&pc->pc_partgroups, pcg, pcg_list);
1.43      thorpej  2008:                }
1.102     chs      2009:        }
                   2010:        if (pcg == NULL) {
1.43      thorpej  2011:
                   2012:                /*
                   2013:                 * No empty groups to free the object to.  Attempt to
1.47      thorpej  2014:                 * allocate one.
1.43      thorpej  2015:                 */
1.47      thorpej  2016:                simple_unlock(&pc->pc_slock);
1.60      thorpej  2017:                s = splvm();
1.43      thorpej  2018:                pcg = pool_get(&pcgpool, PR_NOWAIT);
1.60      thorpej  2019:                splx(s);
1.102     chs      2020:                if (pcg == NULL) {
1.109     christos 2021: destruct:
1.102     chs      2022:
                   2023:                        /*
                   2024:                         * Unable to allocate a cache group; destruct the object
                   2025:                         * and free it back to the pool.
                   2026:                         */
                   2027:                        pool_cache_destruct_object(pc, object);
                   2028:                        return;
1.43      thorpej  2029:                }
1.102     chs      2030:                memset(pcg, 0, sizeof(*pcg));
                   2031:                simple_lock(&pc->pc_slock);
                   2032:                pc->pc_ngroups++;
                   2033:                LIST_INSERT_HEAD(&pc->pc_partgroups, pcg, pcg_list);
1.43      thorpej  2034:        }
                   2035:
1.48      thorpej  2036:        pc->pc_nitems++;
1.87      thorpej  2037:        pcg_put(pcg, object, pa);
1.43      thorpej  2038:
1.102     chs      2039:        if (pcg->pcg_avail == PCG_NOBJECTS) {
                   2040:                LIST_REMOVE(pcg, pcg_list);
                   2041:                LIST_INSERT_HEAD(&pc->pc_fullgroups, pcg, pcg_list);
                   2042:        }
1.43      thorpej  2043:        simple_unlock(&pc->pc_slock);
1.51      thorpej  2044: }
                   2045:
                   2046: /*
                   2047:  * pool_cache_destruct_object:
                   2048:  *
                   2049:  *     Force destruction of an object and its release back into
                   2050:  *     the pool.
                   2051:  */
                   2052: void
                   2053: pool_cache_destruct_object(struct pool_cache *pc, void *object)
                   2054: {
                   2055:
                   2056:        if (pc->pc_dtor != NULL)
                   2057:                (*pc->pc_dtor)(pc->pc_arg, object);
                   2058:        pool_put(pc->pc_pool, object);
1.43      thorpej  2059: }
                   2060:
1.102     chs      2061: static void
1.106     christos 2062: pool_do_cache_invalidate_grouplist(struct pool_cache_grouplist *pcgsl,
1.105     christos 2063:     struct pool_cache *pc, struct pool_pagelist *pq,
1.106     christos 2064:     struct pool_cache_grouplist *pcgdl)
1.102     chs      2065: {
1.106     christos 2066:        struct pool_cache_group *pcg, *npcg;
1.102     chs      2067:        void *object;
                   2068:
1.106     christos 2069:        for (pcg = LIST_FIRST(pcgsl); pcg != NULL; pcg = npcg) {
1.102     chs      2070:                npcg = LIST_NEXT(pcg, pcg_list);
                   2071:                while (pcg->pcg_avail != 0) {
                   2072:                        pc->pc_nitems--;
                   2073:                        object = pcg_get(pcg, NULL);
                   2074:                        if (pc->pc_dtor != NULL)
                   2075:                                (*pc->pc_dtor)(pc->pc_arg, object);
                   2076:                        pool_do_put(pc->pc_pool, object, pq);
                   2077:                }
1.103     chs      2078:                pc->pc_ngroups--;
1.102     chs      2079:                LIST_REMOVE(pcg, pcg_list);
1.106     christos 2080:                LIST_INSERT_HEAD(pcgdl, pcg, pcg_list);
1.102     chs      2081:        }
1.105     christos 2082: }
                   2083:
                   2084: static void
                   2085: pool_do_cache_invalidate(struct pool_cache *pc, struct pool_pagelist *pq,
                   2086:     struct pool_cache_grouplist *pcgl)
                   2087: {
                   2088:
                   2089:        LOCK_ASSERT(simple_lock_held(&pc->pc_slock));
                   2090:        LOCK_ASSERT(simple_lock_held(&pc->pc_pool->pr_slock));
                   2091:
1.106     christos 2092:        pool_do_cache_invalidate_grouplist(&pc->pc_fullgroups, pc, pq, pcgl);
                   2093:        pool_do_cache_invalidate_grouplist(&pc->pc_partgroups, pc, pq, pcgl);
1.103     chs      2094:
                   2095:        KASSERT(LIST_EMPTY(&pc->pc_partgroups));
                   2096:        KASSERT(LIST_EMPTY(&pc->pc_fullgroups));
                   2097:        KASSERT(pc->pc_nitems == 0);
1.102     chs      2098: }
                   2099:
1.43      thorpej  2100: /*
1.101     thorpej  2101:  * pool_cache_invalidate:
1.43      thorpej  2102:  *
1.101     thorpej  2103:  *     Invalidate a pool cache (destruct and release all of the
                   2104:  *     cached objects).
1.43      thorpej  2105:  */
1.101     thorpej  2106: void
                   2107: pool_cache_invalidate(struct pool_cache *pc)
1.43      thorpej  2108: {
1.101     thorpej  2109:        struct pool_pagelist pq;
1.102     chs      2110:        struct pool_cache_grouplist pcgl;
1.101     thorpej  2111:
                   2112:        LIST_INIT(&pq);
1.102     chs      2113:        LIST_INIT(&pcgl);
1.101     thorpej  2114:
                   2115:        simple_lock(&pc->pc_slock);
                   2116:        simple_lock(&pc->pc_pool->pr_slock);
1.43      thorpej  2117:
1.102     chs      2118:        pool_do_cache_invalidate(pc, &pq, &pcgl);
1.43      thorpej  2119:
1.101     thorpej  2120:        simple_unlock(&pc->pc_pool->pr_slock);
                   2121:        simple_unlock(&pc->pc_slock);
1.43      thorpej  2122:
1.102     chs      2123:        pr_pagelist_free(pc->pc_pool, &pq);
                   2124:        pcg_grouplist_free(&pcgl);
1.43      thorpej  2125: }
                   2126:
                   2127: /*
                   2128:  * pool_cache_reclaim:
                   2129:  *
                   2130:  *     Reclaim a pool cache for pool_reclaim().
                   2131:  */
                   2132: static void
1.102     chs      2133: pool_cache_reclaim(struct pool_cache *pc, struct pool_pagelist *pq,
                   2134:     struct pool_cache_grouplist *pcgl)
1.43      thorpej  2135: {
1.101     thorpej  2136:
                   2137:        /*
                   2138:         * We're locking in the wrong order (normally pool_cache -> pool,
                   2139:         * but the pool is already locked when we get here), so we have
                   2140:         * to use trylock.  If we can't lock the pool_cache, it's not really
                   2141:         * a big deal here.
                   2142:         */
                   2143:        if (simple_lock_try(&pc->pc_slock) == 0)
                   2144:                return;
                   2145:
1.102     chs      2146:        pool_do_cache_invalidate(pc, pq, pcgl);
1.43      thorpej  2147:
                   2148:        simple_unlock(&pc->pc_slock);
1.3       pk       2149: }
1.66      thorpej  2150:
                   2151: /*
                   2152:  * Pool backend allocators.
                   2153:  *
                   2154:  * Each pool has a backend allocator that handles allocation, deallocation,
                   2155:  * and any additional draining that might be needed.
                   2156:  *
                   2157:  * We provide two standard allocators:
                   2158:  *
                   2159:  *     pool_allocator_kmem - the default when no allocator is specified
                   2160:  *
                   2161:  *     pool_allocator_nointr - used for pools that will not be accessed
                   2162:  *     in interrupt context.
                   2163:  */
                   2164: void   *pool_page_alloc(struct pool *, int);
                   2165: void   pool_page_free(struct pool *, void *);
                   2166:
1.112     bjh21    2167: #ifdef POOL_SUBPAGE
                   2168: struct pool_allocator pool_allocator_kmem_fullpage = {
                   2169:        pool_page_alloc, pool_page_free, 0,
                   2170: };
                   2171: #else
1.66      thorpej  2172: struct pool_allocator pool_allocator_kmem = {
                   2173:        pool_page_alloc, pool_page_free, 0,
                   2174: };
1.112     bjh21    2175: #endif
1.66      thorpej  2176:
                   2177: void   *pool_page_alloc_nointr(struct pool *, int);
                   2178: void   pool_page_free_nointr(struct pool *, void *);
                   2179:
1.112     bjh21    2180: #ifdef POOL_SUBPAGE
                   2181: struct pool_allocator pool_allocator_nointr_fullpage = {
                   2182:        pool_page_alloc_nointr, pool_page_free_nointr, 0,
                   2183: };
                   2184: #else
1.66      thorpej  2185: struct pool_allocator pool_allocator_nointr = {
                   2186:        pool_page_alloc_nointr, pool_page_free_nointr, 0,
                   2187: };
1.112     bjh21    2188: #endif
1.66      thorpej  2189:
                   2190: #ifdef POOL_SUBPAGE
                   2191: void   *pool_subpage_alloc(struct pool *, int);
                   2192: void   pool_subpage_free(struct pool *, void *);
                   2193:
1.112     bjh21    2194: struct pool_allocator pool_allocator_kmem = {
                   2195:        pool_subpage_alloc, pool_subpage_free, POOL_SUBPAGE,
                   2196: };
                   2197:
                   2198: void   *pool_subpage_alloc_nointr(struct pool *, int);
                   2199: void   pool_subpage_free_nointr(struct pool *, void *);
                   2200:
                   2201: struct pool_allocator pool_allocator_nointr = {
                   2202:        pool_subpage_alloc, pool_subpage_free, POOL_SUBPAGE,
1.66      thorpej  2203: };
                   2204: #endif /* POOL_SUBPAGE */
                   2205:
                   2206: /*
                   2207:  * We have at least three different resources for the same allocation and
                   2208:  * each resource can be depleted.  First, we have the ready elements in the
                   2209:  * pool.  Then we have the resource (typically a vm_map) for this allocator.
                   2210:  * Finally, we have physical memory.  Waiting for any of these can be
                   2211:  * unnecessary when any other is freed, but the kernel doesn't support
                   2212:  * sleeping on multiple wait channels, so we have to employ another strategy.
                   2213:  *
                   2214:  * The caller sleeps on the pool (so that it can be awakened when an item
                   2215:  * is returned to the pool), but we set PA_WANT on the allocator.  When a
                   2216:  * page is returned to the allocator and PA_WANT is set, pool_allocator_free
                   2217:  * will wake up all sleeping pools belonging to this allocator.
                   2218:  *
                   2219:  * XXX Thundering herd.
                   2220:  */
                   2221: void *
                   2222: pool_allocator_alloc(struct pool *org, int flags)
                   2223: {
                   2224:        struct pool_allocator *pa = org->pr_alloc;
                   2225:        struct pool *pp, *start;
                   2226:        int s, freed;
                   2227:        void *res;
                   2228:
1.91      yamt     2229:        LOCK_ASSERT(!simple_lock_held(&org->pr_slock));
                   2230:
1.66      thorpej  2231:        do {
                   2232:                if ((res = (*pa->pa_alloc)(org, flags)) != NULL)
                   2233:                        return (res);
1.68      thorpej  2234:                if ((flags & PR_WAITOK) == 0) {
                   2235:                        /*
                   2236:                         * We only run the drain hookhere if PR_NOWAIT.
                   2237:                         * In other cases, the hook will be run in
                   2238:                         * pool_reclaim().
                   2239:                         */
                   2240:                        if (org->pr_drain_hook != NULL) {
                   2241:                                (*org->pr_drain_hook)(org->pr_drain_hook_arg,
                   2242:                                    flags);
                   2243:                                if ((res = (*pa->pa_alloc)(org, flags)) != NULL)
                   2244:                                        return (res);
                   2245:                        }
1.66      thorpej  2246:                        break;
1.68      thorpej  2247:                }
1.66      thorpej  2248:
                   2249:                /*
1.109     christos 2250:                 * Drain all pools, that use this allocator.
                   2251:                 * We do this to reclaim VA space.
1.66      thorpej  2252:                 * pa_alloc is responsible for waiting for
                   2253:                 * physical memory.
                   2254:                 *
                   2255:                 * XXX We risk looping forever if start if someone
                   2256:                 * calls pool_destroy on "start".  But there is no
                   2257:                 * other way to have potentially sleeping pool_reclaim,
                   2258:                 * non-sleeping locks on pool_allocator, and some
                   2259:                 * stirring of drained pools in the allocator.
1.68      thorpej  2260:                 *
                   2261:                 * XXX Maybe we should use pool_head_slock for locking
                   2262:                 * the allocators?
1.66      thorpej  2263:                 */
                   2264:                freed = 0;
                   2265:
                   2266:                s = splvm();
                   2267:                simple_lock(&pa->pa_slock);
                   2268:                pp = start = TAILQ_FIRST(&pa->pa_list);
                   2269:                do {
                   2270:                        TAILQ_REMOVE(&pa->pa_list, pp, pr_alloc_list);
                   2271:                        TAILQ_INSERT_TAIL(&pa->pa_list, pp, pr_alloc_list);
1.73      thorpej  2272:                        simple_unlock(&pa->pa_slock);
1.66      thorpej  2273:                        freed = pool_reclaim(pp);
1.73      thorpej  2274:                        simple_lock(&pa->pa_slock);
1.66      thorpej  2275:                } while ((pp = TAILQ_FIRST(&pa->pa_list)) != start &&
                   2276:                         freed == 0);
                   2277:
                   2278:                if (freed == 0) {
                   2279:                        /*
                   2280:                         * We set PA_WANT here, the caller will most likely
                   2281:                         * sleep waiting for pages (if not, this won't hurt
                   2282:                         * that much), and there is no way to set this in
                   2283:                         * the caller without violating locking order.
                   2284:                         */
                   2285:                        pa->pa_flags |= PA_WANT;
                   2286:                }
                   2287:                simple_unlock(&pa->pa_slock);
                   2288:                splx(s);
                   2289:        } while (freed);
                   2290:        return (NULL);
                   2291: }
                   2292:
                   2293: void
                   2294: pool_allocator_free(struct pool *pp, void *v)
                   2295: {
                   2296:        struct pool_allocator *pa = pp->pr_alloc;
                   2297:        int s;
                   2298:
1.91      yamt     2299:        LOCK_ASSERT(!simple_lock_held(&pp->pr_slock));
                   2300:
1.66      thorpej  2301:        (*pa->pa_free)(pp, v);
                   2302:
                   2303:        s = splvm();
                   2304:        simple_lock(&pa->pa_slock);
                   2305:        if ((pa->pa_flags & PA_WANT) == 0) {
                   2306:                simple_unlock(&pa->pa_slock);
                   2307:                splx(s);
                   2308:                return;
                   2309:        }
                   2310:
                   2311:        TAILQ_FOREACH(pp, &pa->pa_list, pr_alloc_list) {
                   2312:                simple_lock(&pp->pr_slock);
                   2313:                if ((pp->pr_flags & PR_WANTED) != 0) {
                   2314:                        pp->pr_flags &= ~PR_WANTED;
                   2315:                        wakeup(pp);
                   2316:                }
1.69      thorpej  2317:                simple_unlock(&pp->pr_slock);
1.66      thorpej  2318:        }
                   2319:        pa->pa_flags &= ~PA_WANT;
                   2320:        simple_unlock(&pa->pa_slock);
                   2321:        splx(s);
                   2322: }
                   2323:
                   2324: void *
                   2325: pool_page_alloc(struct pool *pp, int flags)
                   2326: {
                   2327:        boolean_t waitok = (flags & PR_WAITOK) ? TRUE : FALSE;
                   2328:
1.100     yamt     2329:        return ((void *) uvm_km_alloc_poolpage_cache(kmem_map, waitok));
1.66      thorpej  2330: }
                   2331:
                   2332: void
                   2333: pool_page_free(struct pool *pp, void *v)
                   2334: {
                   2335:
1.98      yamt     2336:        uvm_km_free_poolpage_cache(kmem_map, (vaddr_t) v);
                   2337: }
                   2338:
                   2339: static void *
                   2340: pool_page_alloc_meta(struct pool *pp, int flags)
                   2341: {
                   2342:        boolean_t waitok = (flags & PR_WAITOK) ? TRUE : FALSE;
                   2343:
1.100     yamt     2344:        return ((void *) uvm_km_alloc_poolpage(kmem_map, waitok));
1.98      yamt     2345: }
                   2346:
                   2347: static void
                   2348: pool_page_free_meta(struct pool *pp, void *v)
                   2349: {
                   2350:
1.100     yamt     2351:        uvm_km_free_poolpage(kmem_map, (vaddr_t) v);
1.66      thorpej  2352: }
                   2353:
                   2354: #ifdef POOL_SUBPAGE
                   2355: /* Sub-page allocator, for machines with large hardware pages. */
                   2356: void *
                   2357: pool_subpage_alloc(struct pool *pp, int flags)
                   2358: {
1.93      dbj      2359:        void *v;
                   2360:        int s;
                   2361:        s = splvm();
                   2362:        v = pool_get(&psppool, flags);
                   2363:        splx(s);
                   2364:        return v;
1.66      thorpej  2365: }
                   2366:
                   2367: void
                   2368: pool_subpage_free(struct pool *pp, void *v)
                   2369: {
1.93      dbj      2370:        int s;
                   2371:        s = splvm();
1.66      thorpej  2372:        pool_put(&psppool, v);
1.93      dbj      2373:        splx(s);
1.66      thorpej  2374: }
                   2375:
                   2376: /* We don't provide a real nointr allocator.  Maybe later. */
                   2377: void *
1.112     bjh21    2378: pool_subpage_alloc_nointr(struct pool *pp, int flags)
1.66      thorpej  2379: {
                   2380:
                   2381:        return (pool_subpage_alloc(pp, flags));
                   2382: }
                   2383:
                   2384: void
1.112     bjh21    2385: pool_subpage_free_nointr(struct pool *pp, void *v)
1.66      thorpej  2386: {
                   2387:
                   2388:        pool_subpage_free(pp, v);
                   2389: }
1.112     bjh21    2390: #endif /* POOL_SUBPAGE */
1.66      thorpej  2391: void *
                   2392: pool_page_alloc_nointr(struct pool *pp, int flags)
                   2393: {
                   2394:        boolean_t waitok = (flags & PR_WAITOK) ? TRUE : FALSE;
                   2395:
1.100     yamt     2396:        return ((void *) uvm_km_alloc_poolpage_cache(kernel_map, waitok));
1.66      thorpej  2397: }
                   2398:
                   2399: void
                   2400: pool_page_free_nointr(struct pool *pp, void *v)
                   2401: {
                   2402:
1.98      yamt     2403:        uvm_km_free_poolpage_cache(kernel_map, (vaddr_t) v);
1.66      thorpej  2404: }

CVSweb <webmaster@jp.NetBSD.org>