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


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

FreeBSD/Linux Kernel Cross Reference
sys/vm/uma_int.h

Version: -  FREEBSD  -  FREEBSD-13-STABLE  -  FREEBSD-13-0  -  FREEBSD-12-STABLE  -  FREEBSD-12-0  -  FREEBSD-11-STABLE  -  FREEBSD-11-0  -  FREEBSD-10-STABLE  -  FREEBSD-10-0  -  FREEBSD-9-STABLE  -  FREEBSD-9-0  -  FREEBSD-8-STABLE  -  FREEBSD-8-0  -  FREEBSD-7-STABLE  -  FREEBSD-7-0  -  FREEBSD-6-STABLE  -  FREEBSD-6-0  -  FREEBSD-5-STABLE  -  FREEBSD-5-0  -  FREEBSD-4-STABLE  -  FREEBSD-3-STABLE  -  FREEBSD22  -  l41  -  OPENBSD  -  linux-2.6  -  MK84  -  PLAN9  -  xnu-8792 
SearchContext: -  none  -  3  -  10 

    1 /*-
    2  * SPDX-License-Identifier: BSD-2-Clause-FreeBSD
    3  *
    4  * Copyright (c) 2002-2005, 2009, 2013 Jeffrey Roberson <jeff@FreeBSD.org>
    5  * Copyright (c) 2004, 2005 Bosko Milekic <bmilekic@FreeBSD.org>
    6  * All rights reserved.
    7  *
    8  * Redistribution and use in source and binary forms, with or without
    9  * modification, are permitted provided that the following conditions
   10  * are met:
   11  * 1. Redistributions of source code must retain the above copyright
   12  *    notice unmodified, this list of conditions, and the following
   13  *    disclaimer.
   14  * 2. Redistributions in binary form must reproduce the above copyright
   15  *    notice, this list of conditions and the following disclaimer in the
   16  *    documentation and/or other materials provided with the distribution.
   17  *
   18  * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR
   19  * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
   20  * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
   21  * IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT,
   22  * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
   23  * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
   24  * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
   25  * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
   26  * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
   27  * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
   28  *
   29  * $FreeBSD: releng/12.0/sys/vm/uma_int.h 340011 2018-11-01 17:36:42Z markj $
   30  *
   31  */
   32 
   33 #include <sys/_bitset.h>
   34 #include <sys/_domainset.h>
   35 #include <sys/_task.h>
   36 
   37 /* 
   38  * This file includes definitions, structures, prototypes, and inlines that
   39  * should not be used outside of the actual implementation of UMA.
   40  */
   41 
   42 /* 
   43  * The brief summary;  Zones describe unique allocation types.  Zones are
   44  * organized into per-CPU caches which are filled by buckets.  Buckets are
   45  * organized according to memory domains.  Buckets are filled from kegs which
   46  * are also organized according to memory domains.  Kegs describe a unique
   47  * allocation type, backend memory provider, and layout.  Kegs are associated
   48  * with one or more zones and zones reference one or more kegs.  Kegs provide
   49  * slabs which are virtually contiguous collections of pages.  Each slab is
   50  * broken down int one or more items that will satisfy an individual allocation.
   51  *
   52  * Allocation is satisfied in the following order:
   53  * 1) Per-CPU cache
   54  * 2) Per-domain cache of buckets
   55  * 3) Slab from any of N kegs
   56  * 4) Backend page provider
   57  *
   58  * More detail on individual objects is contained below:
   59  *
   60  * Kegs contain lists of slabs which are stored in either the full bin, empty
   61  * bin, or partially allocated bin, to reduce fragmentation.  They also contain
   62  * the user supplied value for size, which is adjusted for alignment purposes
   63  * and rsize is the result of that.  The Keg also stores information for
   64  * managing a hash of page addresses that maps pages to uma_slab_t structures
   65  * for pages that don't have embedded uma_slab_t's.
   66  *
   67  * Keg slab lists are organized by memory domain to support NUMA allocation
   68  * policies.  By default allocations are spread across domains to reduce the
   69  * potential for hotspots.  Special keg creation flags may be specified to
   70  * prefer location allocation.  However there is no strict enforcement as frees
   71  * may happen on any CPU and these are returned to the CPU-local cache
   72  * regardless of the originating domain.
   73  *  
   74  * The uma_slab_t may be embedded in a UMA_SLAB_SIZE chunk of memory or it may
   75  * be allocated off the page from a special slab zone.  The free list within a
   76  * slab is managed with a bitmask.  For item sizes that would yield more than
   77  * 10% memory waste we potentially allocate a separate uma_slab_t if this will
   78  * improve the number of items per slab that will fit.  
   79  *
   80  * The only really gross cases, with regards to memory waste, are for those
   81  * items that are just over half the page size.   You can get nearly 50% waste,
   82  * so you fall back to the memory footprint of the power of two allocator. I
   83  * have looked at memory allocation sizes on many of the machines available to
   84  * me, and there does not seem to be an abundance of allocations at this range
   85  * so at this time it may not make sense to optimize for it.  This can, of 
   86  * course, be solved with dynamic slab sizes.
   87  *
   88  * Kegs may serve multiple Zones but by far most of the time they only serve
   89  * one.  When a Zone is created, a Keg is allocated and setup for it.  While
   90  * the backing Keg stores slabs, the Zone caches Buckets of items allocated
   91  * from the slabs.  Each Zone is equipped with an init/fini and ctor/dtor
   92  * pair, as well as with its own set of small per-CPU caches, layered above
   93  * the Zone's general Bucket cache.
   94  *
   95  * The PCPU caches are protected by critical sections, and may be accessed
   96  * safely only from their associated CPU, while the Zones backed by the same
   97  * Keg all share a common Keg lock (to coalesce contention on the backing
   98  * slabs).  The backing Keg typically only serves one Zone but in the case of
   99  * multiple Zones, one of the Zones is considered the Master Zone and all
  100  * Zone-related stats from the Keg are done in the Master Zone.  For an
  101  * example of a Multi-Zone setup, refer to the Mbuf allocation code.
  102  */
  103 
  104 /*
  105  *      This is the representation for normal (Non OFFPAGE slab)
  106  *
  107  *      i == item
  108  *      s == slab pointer
  109  *
  110  *      <----------------  Page (UMA_SLAB_SIZE) ------------------>
  111  *      ___________________________________________________________
  112  *     | _  _  _  _  _  _  _  _  _  _  _  _  _  _  _   ___________ |
  113  *     ||i||i||i||i||i||i||i||i||i||i||i||i||i||i||i| |slab header||
  114  *     ||_||_||_||_||_||_||_||_||_||_||_||_||_||_||_| |___________|| 
  115  *     |___________________________________________________________|
  116  *
  117  *
  118  *      This is an OFFPAGE slab. These can be larger than UMA_SLAB_SIZE.
  119  *
  120  *      ___________________________________________________________
  121  *     | _  _  _  _  _  _  _  _  _  _  _  _  _  _  _  _  _  _  _   |
  122  *     ||i||i||i||i||i||i||i||i||i||i||i||i||i||i||i||i||i||i||i|  |
  123  *     ||_||_||_||_||_||_||_||_||_||_||_||_||_||_||_||_||_||_||_|  |
  124  *     |___________________________________________________________|
  125  *       ___________    ^
  126  *      |slab header|   |
  127  *      |___________|---*
  128  *
  129  */
  130 
  131 #ifndef VM_UMA_INT_H
  132 #define VM_UMA_INT_H
  133 
  134 #define UMA_SLAB_SIZE   PAGE_SIZE       /* How big are our slabs? */
  135 #define UMA_SLAB_MASK   (PAGE_SIZE - 1) /* Mask to get back to the page */
  136 #define UMA_SLAB_SHIFT  PAGE_SHIFT      /* Number of bits PAGE_MASK */
  137 
  138 /* Max waste percentage before going to off page slab management */
  139 #define UMA_MAX_WASTE   10
  140 
  141 /*
  142  * Size of memory in a not offpage slab available for actual items.
  143  */
  144 #define UMA_SLAB_SPACE  (UMA_SLAB_SIZE - sizeof(struct uma_slab))
  145 
  146 /*
  147  * I doubt there will be many cases where this is exceeded. This is the initial
  148  * size of the hash table for uma_slabs that are managed off page. This hash
  149  * does expand by powers of two.  Currently it doesn't get smaller.
  150  */
  151 #define UMA_HASH_SIZE_INIT      32              
  152 
  153 /* 
  154  * I should investigate other hashing algorithms.  This should yield a low
  155  * number of collisions if the pages are relatively contiguous.
  156  */
  157 
  158 #define UMA_HASH(h, s) ((((uintptr_t)s) >> UMA_SLAB_SHIFT) & (h)->uh_hashmask)
  159 
  160 #define UMA_HASH_INSERT(h, s, mem)                                      \
  161                 SLIST_INSERT_HEAD(&(h)->uh_slab_hash[UMA_HASH((h),      \
  162                     (mem))], (s), us_hlink)
  163 #define UMA_HASH_REMOVE(h, s, mem)                                      \
  164                 SLIST_REMOVE(&(h)->uh_slab_hash[UMA_HASH((h),           \
  165                     (mem))], (s), uma_slab, us_hlink)
  166 
  167 /* Hash table for freed address -> slab translation */
  168 
  169 SLIST_HEAD(slabhead, uma_slab);
  170 
  171 struct uma_hash {
  172         struct slabhead *uh_slab_hash;  /* Hash table for slabs */
  173         int             uh_hashsize;    /* Current size of the hash table */
  174         int             uh_hashmask;    /* Mask used during hashing */
  175 };
  176 
  177 /*
  178  * align field or structure to cache line
  179  */
  180 #if defined(__amd64__) || defined(__powerpc64__)
  181 #define UMA_ALIGN       __aligned(128)
  182 #else
  183 #define UMA_ALIGN
  184 #endif
  185 
  186 /*
  187  * Structures for per cpu queues.
  188  */
  189 
  190 struct uma_bucket {
  191         LIST_ENTRY(uma_bucket)  ub_link;        /* Link into the zone */
  192         int16_t ub_cnt;                         /* Count of items in bucket. */
  193         int16_t ub_entries;                     /* Max items. */
  194         void    *ub_bucket[];                   /* actual allocation storage */
  195 };
  196 
  197 typedef struct uma_bucket * uma_bucket_t;
  198 
  199 struct uma_cache {
  200         uma_bucket_t    uc_freebucket;  /* Bucket we're freeing to */
  201         uma_bucket_t    uc_allocbucket; /* Bucket to allocate from */
  202         uint64_t        uc_allocs;      /* Count of allocations */
  203         uint64_t        uc_frees;       /* Count of frees */
  204 } UMA_ALIGN;
  205 
  206 typedef struct uma_cache * uma_cache_t;
  207 
  208 /*
  209  * Per-domain memory list.  Embedded in the kegs.
  210  */
  211 struct uma_domain {
  212         LIST_HEAD(,uma_slab)    ud_part_slab;   /* partially allocated slabs */
  213         LIST_HEAD(,uma_slab)    ud_free_slab;   /* empty slab list */
  214         LIST_HEAD(,uma_slab)    ud_full_slab;   /* full slabs */
  215 };
  216 
  217 typedef struct uma_domain * uma_domain_t;
  218 
  219 /*
  220  * Keg management structure
  221  *
  222  * TODO: Optimize for cache line size
  223  *
  224  */
  225 struct uma_keg {
  226         struct mtx      uk_lock;        /* Lock for the keg */
  227         struct uma_hash uk_hash;
  228         LIST_HEAD(,uma_zone)    uk_zones;       /* Keg's zones */
  229 
  230         struct domainset_ref uk_dr;     /* Domain selection policy. */
  231         uint32_t        uk_align;       /* Alignment mask */
  232         uint32_t        uk_pages;       /* Total page count */
  233         uint32_t        uk_free;        /* Count of items free in slabs */
  234         uint32_t        uk_reserve;     /* Number of reserved items. */
  235         uint32_t        uk_size;        /* Requested size of each item */
  236         uint32_t        uk_rsize;       /* Real size of each item */
  237         uint32_t        uk_maxpages;    /* Maximum number of pages to alloc */
  238 
  239         uma_init        uk_init;        /* Keg's init routine */
  240         uma_fini        uk_fini;        /* Keg's fini routine */
  241         uma_alloc       uk_allocf;      /* Allocation function */
  242         uma_free        uk_freef;       /* Free routine */
  243 
  244         u_long          uk_offset;      /* Next free offset from base KVA */
  245         vm_offset_t     uk_kva;         /* Zone base KVA */
  246         uma_zone_t      uk_slabzone;    /* Slab zone backing us, if OFFPAGE */
  247 
  248         uint32_t        uk_pgoff;       /* Offset to uma_slab struct */
  249         uint16_t        uk_ppera;       /* pages per allocation from backend */
  250         uint16_t        uk_ipers;       /* Items per slab */
  251         uint32_t        uk_flags;       /* Internal flags */
  252 
  253         /* Least used fields go to the last cache line. */
  254         const char      *uk_name;               /* Name of creating zone. */
  255         LIST_ENTRY(uma_keg)     uk_link;        /* List of all kegs */
  256 
  257         /* Must be last, variable sized. */
  258         struct uma_domain       uk_domain[];    /* Keg's slab lists. */
  259 };
  260 typedef struct uma_keg  * uma_keg_t;
  261 
  262 /*
  263  * Free bits per-slab.
  264  */
  265 #define SLAB_SETSIZE    (PAGE_SIZE / UMA_SMALLEST_UNIT)
  266 BITSET_DEFINE(slabbits, SLAB_SETSIZE);
  267 
  268 /*
  269  * The slab structure manages a single contiguous allocation from backing
  270  * store and subdivides it into individually allocatable items.
  271  */
  272 struct uma_slab {
  273         uma_keg_t       us_keg;                 /* Keg we live in */
  274         union {
  275                 LIST_ENTRY(uma_slab)    _us_link;       /* slabs in zone */
  276                 unsigned long   _us_size;       /* Size of allocation */
  277         } us_type;
  278         SLIST_ENTRY(uma_slab)   us_hlink;       /* Link for hash table */
  279         uint8_t         *us_data;               /* First item */
  280         struct slabbits us_free;                /* Free bitmask. */
  281 #ifdef INVARIANTS
  282         struct slabbits us_debugfree;           /* Debug bitmask. */
  283 #endif
  284         uint16_t        us_freecount;           /* How many are free? */
  285         uint8_t         us_flags;               /* Page flags see uma.h */
  286         uint8_t         us_domain;              /* Backing NUMA domain. */
  287 };
  288 
  289 #define us_link us_type._us_link
  290 #define us_size us_type._us_size
  291 
  292 #if MAXMEMDOM >= 255
  293 #error "Slab domain type insufficient"
  294 #endif
  295 
  296 typedef struct uma_slab * uma_slab_t;
  297 typedef uma_slab_t (*uma_slaballoc)(uma_zone_t, uma_keg_t, int, int);
  298 
  299 struct uma_klink {
  300         LIST_ENTRY(uma_klink)   kl_link;
  301         uma_keg_t               kl_keg;
  302 };
  303 typedef struct uma_klink *uma_klink_t;
  304 
  305 struct uma_zone_domain {
  306         LIST_HEAD(,uma_bucket)  uzd_buckets;    /* full buckets */
  307 };
  308 
  309 typedef struct uma_zone_domain * uma_zone_domain_t;
  310 
  311 /*
  312  * Zone management structure 
  313  *
  314  * TODO: Optimize for cache line size
  315  *
  316  */
  317 struct uma_zone {
  318         /* Offset 0, used in alloc/free fast/medium fast path and const. */
  319         struct mtx      *uz_lockptr;
  320         const char      *uz_name;       /* Text name of the zone */
  321         struct uma_zone_domain  *uz_domain;     /* per-domain buckets */
  322         uint32_t        uz_flags;       /* Flags inherited from kegs */
  323         uint32_t        uz_size;        /* Size inherited from kegs */
  324         uma_ctor        uz_ctor;        /* Constructor for each allocation */
  325         uma_dtor        uz_dtor;        /* Destructor */
  326         uma_init        uz_init;        /* Initializer for each item */
  327         uma_fini        uz_fini;        /* Finalizer for each item. */
  328 
  329         /* Offset 64, used in bucket replenish. */
  330         uma_import      uz_import;      /* Import new memory to cache. */
  331         uma_release     uz_release;     /* Release memory from cache. */
  332         void            *uz_arg;        /* Import/release argument. */
  333         uma_slaballoc   uz_slab;        /* Allocate a slab from the backend. */
  334         uint16_t        uz_count;       /* Amount of items in full bucket */
  335         uint16_t        uz_count_min;   /* Minimal amount of items there */
  336         /* 32bit pad on 64bit. */
  337         LIST_ENTRY(uma_zone)    uz_link;        /* List of all zones in keg */
  338         LIST_HEAD(,uma_klink)   uz_kegs;        /* List of kegs. */
  339 
  340         /* Offset 128 Rare. */
  341         /*
  342          * The lock is placed here to avoid adjacent line prefetcher
  343          * in fast paths and to take up space near infrequently accessed
  344          * members to reduce alignment overhead.
  345          */
  346         struct mtx      uz_lock;        /* Lock for the zone */
  347         struct uma_klink        uz_klink;       /* klink for first keg. */
  348         /* The next two fields are used to print a rate-limited warnings. */
  349         const char      *uz_warning;    /* Warning to print on failure */
  350         struct timeval  uz_ratecheck;   /* Warnings rate-limiting */
  351         struct task     uz_maxaction;   /* Task to run when at limit */
  352 
  353         /* 16 bytes of pad. */
  354 
  355         /* Offset 256, atomic stats. */
  356         volatile u_long uz_allocs UMA_ALIGN; /* Total number of allocations */
  357         volatile u_long uz_fails;       /* Total number of alloc failures */
  358         volatile u_long uz_frees;       /* Total number of frees */
  359         uint64_t        uz_sleeps;      /* Total number of alloc sleeps */
  360 
  361         /*
  362          * This HAS to be the last item because we adjust the zone size
  363          * based on NCPU and then allocate the space for the zones.
  364          */
  365         struct uma_cache        uz_cpu[]; /* Per cpu caches */
  366 
  367         /* uz_domain follows here. */
  368 };
  369 
  370 /*
  371  * These flags must not overlap with the UMA_ZONE flags specified in uma.h.
  372  */
  373 #define UMA_ZFLAG_MULTI         0x04000000      /* Multiple kegs in the zone. */
  374 #define UMA_ZFLAG_DRAINING      0x08000000      /* Running zone_drain. */
  375 #define UMA_ZFLAG_BUCKET        0x10000000      /* Bucket zone. */
  376 #define UMA_ZFLAG_INTERNAL      0x20000000      /* No offpage no PCPU. */
  377 #define UMA_ZFLAG_FULL          0x40000000      /* Reached uz_maxpages */
  378 #define UMA_ZFLAG_CACHEONLY     0x80000000      /* Don't ask VM for buckets. */
  379 
  380 #define UMA_ZFLAG_INHERIT                                               \
  381     (UMA_ZFLAG_INTERNAL | UMA_ZFLAG_CACHEONLY | UMA_ZFLAG_BUCKET)
  382 
  383 static inline uma_keg_t
  384 zone_first_keg(uma_zone_t zone)
  385 {
  386         uma_klink_t klink;
  387 
  388         klink = LIST_FIRST(&zone->uz_kegs);
  389         return (klink != NULL) ? klink->kl_keg : NULL;
  390 }
  391 
  392 #undef UMA_ALIGN
  393 
  394 #ifdef _KERNEL
  395 /* Internal prototypes */
  396 static __inline uma_slab_t hash_sfind(struct uma_hash *hash, uint8_t *data);
  397 void *uma_large_malloc(vm_size_t size, int wait);
  398 void *uma_large_malloc_domain(vm_size_t size, int domain, int wait);
  399 void uma_large_free(uma_slab_t slab);
  400 
  401 /* Lock Macros */
  402 
  403 #define KEG_LOCK_INIT(k, lc)                                    \
  404         do {                                                    \
  405                 if ((lc))                                       \
  406                         mtx_init(&(k)->uk_lock, (k)->uk_name,   \
  407                             (k)->uk_name, MTX_DEF | MTX_DUPOK); \
  408                 else                                            \
  409                         mtx_init(&(k)->uk_lock, (k)->uk_name,   \
  410                             "UMA zone", MTX_DEF | MTX_DUPOK);   \
  411         } while (0)
  412 
  413 #define KEG_LOCK_FINI(k)        mtx_destroy(&(k)->uk_lock)
  414 #define KEG_LOCK(k)     mtx_lock(&(k)->uk_lock)
  415 #define KEG_UNLOCK(k)   mtx_unlock(&(k)->uk_lock)
  416 
  417 #define ZONE_LOCK_INIT(z, lc)                                   \
  418         do {                                                    \
  419                 if ((lc))                                       \
  420                         mtx_init(&(z)->uz_lock, (z)->uz_name,   \
  421                             (z)->uz_name, MTX_DEF | MTX_DUPOK); \
  422                 else                                            \
  423                         mtx_init(&(z)->uz_lock, (z)->uz_name,   \
  424                             "UMA zone", MTX_DEF | MTX_DUPOK);   \
  425         } while (0)
  426             
  427 #define ZONE_LOCK(z)    mtx_lock((z)->uz_lockptr)
  428 #define ZONE_TRYLOCK(z) mtx_trylock((z)->uz_lockptr)
  429 #define ZONE_UNLOCK(z)  mtx_unlock((z)->uz_lockptr)
  430 #define ZONE_LOCK_FINI(z)       mtx_destroy(&(z)->uz_lock)
  431 
  432 /*
  433  * Find a slab within a hash table.  This is used for OFFPAGE zones to lookup
  434  * the slab structure.
  435  *
  436  * Arguments:
  437  *      hash  The hash table to search.
  438  *      data  The base page of the item.
  439  *
  440  * Returns:
  441  *      A pointer to a slab if successful, else NULL.
  442  */
  443 static __inline uma_slab_t
  444 hash_sfind(struct uma_hash *hash, uint8_t *data)
  445 {
  446         uma_slab_t slab;
  447         int hval;
  448 
  449         hval = UMA_HASH(hash, data);
  450 
  451         SLIST_FOREACH(slab, &hash->uh_slab_hash[hval], us_hlink) {
  452                 if ((uint8_t *)slab->us_data == data)
  453                         return (slab);
  454         }
  455         return (NULL);
  456 }
  457 
  458 static __inline uma_slab_t
  459 vtoslab(vm_offset_t va)
  460 {
  461         vm_page_t p;
  462 
  463         p = PHYS_TO_VM_PAGE(pmap_kextract(va));
  464         return ((uma_slab_t)p->plinks.s.pv);
  465 }
  466 
  467 static __inline void
  468 vsetslab(vm_offset_t va, uma_slab_t slab)
  469 {
  470         vm_page_t p;
  471 
  472         p = PHYS_TO_VM_PAGE(pmap_kextract(va));
  473         p->plinks.s.pv = slab;
  474 }
  475 
  476 /*
  477  * The following two functions may be defined by architecture specific code
  478  * if they can provide more efficient allocation functions.  This is useful
  479  * for using direct mapped addresses.
  480  */
  481 void *uma_small_alloc(uma_zone_t zone, vm_size_t bytes, int domain,
  482     uint8_t *pflag, int wait);
  483 void uma_small_free(void *mem, vm_size_t size, uint8_t flags);
  484 
  485 /* Set a global soft limit on UMA managed memory. */
  486 void uma_set_limit(unsigned long limit);
  487 #endif /* _KERNEL */
  488 
  489 #endif /* VM_UMA_INT_H */

Cache object: 0fea6ed6969d423b5ddd3b896732af6b


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


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