2 .\" Copyright (c) 2001 Dag-Erling Coïdan Smørgrav
3 .\" All rights reserved.
5 .\" Redistribution and use in source and binary forms, with or without
6 .\" modification, are permitted provided that the following conditions
8 .\" 1. Redistributions of source code must retain the above copyright
9 .\" notice, this list of conditions and the following disclaimer.
10 .\" 2. Redistributions in binary form must reproduce the above copyright
11 .\" notice, this list of conditions and the following disclaimer in the
12 .\" documentation and/or other materials provided with the distribution.
14 .\" THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
15 .\" ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
16 .\" IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
17 .\" ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
18 .\" FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
19 .\" DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
20 .\" OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
21 .\" HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
22 .\" LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
23 .\" OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
42 .Nm uma_zone_set_warning,
43 .Nm uma_zone_set_maxaction
51 .Fa "char *name" "int size"
52 .Fa "uma_ctor ctor" "uma_dtor dtor" "uma_init uminit" "uma_fini fini"
53 .Fa "int align" "uint16_t flags"
56 .Fn uma_zalloc "uma_zone_t zone" "int flags"
58 .Fn uma_zalloc_arg "uma_zone_t zone" "void *arg" "int flags"
60 .Fn uma_zfree "uma_zone_t zone" "void *item"
62 .Fn uma_zfree_arg "uma_zone_t zone" "void *item" "void *arg"
64 .Fn uma_find_refcnt "uma_zone_t zone" "void *item"
66 .Fn uma_zdestroy "uma_zone_t zone"
68 .Fn uma_zone_set_max "uma_zone_t zone" "int nitems"
70 .Fn uma_zone_get_max "uma_zone_t zone"
72 .Fn uma_zone_get_cur "uma_zone_t zone"
74 .Fn uma_zone_set_warning "uma_zone_t zone" "const char *warning"
76 .Fn uma_zone_set_maxaction "uma_zone_t zone" "void (*maxaction)(uma_zone_t)"
78 .Fn SYSCTL_UMA_MAX parent nbr name access zone descr
79 .Fn SYSCTL_ADD_UMA_MAX ctx parent nbr name access zone descr
80 .Fn SYSCTL_UMA_CUR parent nbr name access zone descr
81 .Fn SYSCTL_ADD_UMA_CUR ctx parent nbr name access zone descr
83 The zone allocator provides an efficient interface for managing
84 dynamically-sized collections of items of similar size.
85 The zone allocator can work with preallocated zones as well as with
86 runtime-allocated ones, and is therefore available much earlier in the
87 boot process than other memory management routines.
89 A zone is an extensible collection of items of identical size.
90 The zone allocator keeps track of which items are in use and which
91 are not, and provides functions for allocating items from the zone and
92 for releasing them back (which makes them available for later use).
94 After the first allocation of an item,
95 it will have been cleared to zeroes, however subsequent allocations
96 will retain the contents as of the last free.
100 function creates a new zone from which items may then be allocated from.
103 argument is a text name of the zone for debugging and stats; this memory
104 should not be freed until the zone has been deallocated.
110 arguments are callback functions that are called by
111 the uma subsystem at the time of the call to
116 Their purpose is to provide hooks for initializing or
117 destroying things that need to be done at the time of the allocation
118 or release of a resource.
124 might be to adjust a global count of the number of objects allocated.
130 arguments are used to optimize the allocation of
131 objects from the zone.
132 They are called by the uma subsystem whenever
133 it needs to allocate or free several items to satisfy requests or memory
139 callbacks might be to
140 initialize and destroy mutexes contained within the object.
142 allow one to re-use already initialized mutexes when an object is returned
143 from the uma subsystem's object cache.
144 They are not called on each call to
148 but rather in a batch mode on several objects.
154 is a subset of the following flags:
155 .Bl -tag -width "foo"
156 .It Dv UMA_ZONE_NOFREE
157 Slabs of the zone are never returned back to VM.
158 .It Dv UMA_ZONE_REFCNT
159 Each item in the zone would have internal reference counter associated with it.
161 .Fn uma_find_refcnt .
162 .It Dv UMA_ZONE_NODUMP
163 Pages belonging to the zone will not be included into mini-dumps.
165 An allocation from zone would have
167 shadow copies, that are privately assigned to CPUs.
168 A CPU can address its private copy using base allocation address plus
169 multiple of current CPU id and
170 .Fn sizeof "struct pcpu" :
171 .Bd -literal -offset indent
172 foo_zone = uma_zcreate(..., UMA_ZONE_PCPU);
174 foo_base = uma_zalloc(foo_zone, ...);
177 foo_pcpu = (foo_t *)zpcpu_get(foo_base);
178 /* do something with foo_pcpu */
181 .It Dv UMA_ZONE_OFFPAGE
182 By default book-keeping of items within a slab is done in the slab page itself.
183 This flag explicitly tells subsystem that book-keeping structure should be
184 allocated separately from special internal zone.
185 This flag requires either
189 since subsystem requires a mechanism to find a book-keeping structure
190 to an item beeing freed.
191 The subsystem may choose to prefer offpage book-keeping for certain zones
193 .It Dv UMA_ZONE_ZINIT
194 The zone will have its
196 method set to internal method that initializes a new allocated slab
204 flag would not return zeroed memory on every
207 The zone should use an internal hash table to find slab book-keeping
208 structure where an allocation being freed belongs to.
209 .It Dv UMA_ZONE_VTOSLAB
210 The zone should use special field of
212 to find slab book-keeping structure where an allocation being freed belongs to.
213 .It Dv UMA_ZONE_MALLOC
218 The zone is for the VM subsystem.
221 To allocate an item from a zone, simply call
223 with a pointer to that zone
226 argument to selected flags as documented in
228 It will return a pointer to an item if successful,
231 in the rare case where all items in the zone are in use and the
232 allocator is unable to grow the zone
237 Items are released back to the zone from which they were allocated by
240 with a pointer to the zone and a pointer to the item.
254 specify an argument for the
258 functions, respectively.
260 If zone was created with
262 flag, then pointer to reference counter for an item can be retrieved with
269 can be destroyed using
271 freeing all memory that was allocated for the zone.
272 All items allocated from the zone with
274 must have been freed with
280 function limits the number of items
281 .Pq and therefore memory
282 that can be allocated to
286 argument specifies the requested upper limit number of items.
287 The effective limit is returned to the caller, as it may end up being higher
288 than requested due to the implementation rounding up to ensure all memory pages
289 allocated to the zone are utilised to capacity.
290 The limit applies to the total number of items in the zone, which includes
291 allocated items, free items and free items in the per-cpu caches.
292 On systems with more than one CPU it may not be possible to allocate
293 the specified number of items even when there is no shortage of memory,
294 because all of the remaining free items may be in the caches of the
295 other CPUs when the limit is hit.
299 function returns the effective upper limit number of items for a zone.
303 function returns the approximate current occupancy of the zone.
304 The returned value is approximate because appropriate synchronisation to
305 determine an exact value is not performed by the implementation.
306 This ensures low overhead at the expense of potentially stale data being used
310 .Fn uma_zone_set_warning
311 function sets a warning that will be printed on the system console when the
312 given zone becomes full and fails to allocate an item.
313 The warning will be printed no more often than every five minutes.
314 Warnings can be turned off globally by setting the
320 .Fn uma_zone_set_maxaction
321 function sets a function that will be called when the given zone becomes full
322 and fails to allocate an item.
323 The function will be called with the zone locked. Also, the function
324 that called the allocation function may have held additional locks. Therefore,
325 this function should do very little work (similar to a signal handler).
328 .Fn SYSCTL_UMA_MAX parent nbr name access zone descr
329 macro declares a static
331 oid that exports the effective upper limit number of items for a zone.
334 argument should be a pointer to
336 A read of the oid returns value obtained through
337 .Fn uma_zone_get_max .
338 A write to the oid sets new value via
339 .Fn uma_zone_set_max .
341 .Fn SYSCTL_ADD_UMA_MAX ctx parent nbr name access zone descr
342 macro is provided to create this type of oid dynamically.
345 .Fn SYSCTL_UMA_CUR parent nbr name access zone descr
346 macro declares a static read only
348 oid that exports the approximate current occupancy of the zone.
351 argument should be a pointer to
353 A read of the oid returns value obtained through
354 .Fn uma_zone_get_cur .
356 .Fn SYSCTL_ADD_UMA_CUR ctx parent nbr name zone descr
357 macro is provided to create this type of oid dynamically.
361 function returns a pointer to an item, or
363 if the zone ran out of unused items
370 The zone allocator first appeared in
372 It was radically changed in
374 to function as a slab allocator.
377 The zone allocator was written by
379 The zone allocator was rewritten in large parts by
380 .An Jeff Roberson Aq Mt jeff@FreeBSD.org
381 to function as a slab allocator.
383 This manual page was written by
384 .An Dag-Erling Sm\(/orgrav Aq Mt des@FreeBSD.org .
386 .An Jeroen Ruigrok van der Werven Aq Mt asmodai@FreeBSD.org .