xref: /freebsd/share/man/man9/zone.9 (revision 9a14aa017b21c292740c00ee098195cd46642730)
1.\"-
2.\" Copyright (c) 2001 Dag-Erling Coïdan Smørgrav
3.\" All rights reserved.
4.\"
5.\" Redistribution and use in source and binary forms, with or without
6.\" modification, are permitted provided that the following conditions
7.\" are met:
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.
13.\"
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
24.\" SUCH DAMAGE.
25.\"
26.\" $FreeBSD$
27.\"
28.Dd October 9, 2010
29.Dt ZONE 9
30.Os
31.Sh NAME
32.Nm uma_zcreate ,
33.Nm uma_zalloc ,
34.Nm uma_zalloc_arg ,
35.Nm uma_zfree ,
36.Nm uma_zfree_arg ,
37.Nm uma_zdestroy ,
38.Nm uma_zone_set_max,
39.Nm uma_zone_get_max,
40.Nm uma_zone_get_cur
41.Nd zone allocator
42.Sh SYNOPSIS
43.In sys/param.h
44.In sys/queue.h
45.In vm/uma.h
46.Ft uma_zone_t
47.Fo uma_zcreate
48.Fa "char *name" "int size"
49.Fa "uma_ctor ctor" "uma_dtor dtor" "uma_init uminit" "uma_fini fini"
50.Fa "int align" "u_int16_t flags"
51.Fc
52.Ft "void *"
53.Fn uma_zalloc "uma_zone_t zone" "int flags"
54.Ft "void *"
55.Fn uma_zalloc_arg "uma_zone_t zone" "void *arg" "int flags"
56.Ft void
57.Fn uma_zfree "uma_zone_t zone" "void *item"
58.Ft void
59.Fn uma_zfree_arg "uma_zone_t zone" "void *item" "void *arg"
60.Ft void
61.Fn uma_zdestroy "uma_zone_t zone"
62.Ft int
63.Fn uma_zone_set_max "uma_zone_t zone" "int nitems"
64.Ft int
65.Fn uma_zone_get_max "uma_zone_t zone"
66.Ft int
67.Fn uma_zone_get_cur "uma_zone_t zone"
68.Sh DESCRIPTION
69The zone allocator provides an efficient interface for managing
70dynamically-sized collections of items of similar size.
71The zone allocator can work with preallocated zones as well as with
72runtime-allocated ones, and is therefore available much earlier in the
73boot process than other memory management routines.
74.Pp
75A zone is an extensible collection of items of identical size.
76The zone allocator keeps track of which items are in use and which
77are not, and provides functions for allocating items from the zone and
78for releasing them back (which makes them available for later use).
79.Pp
80After the first allocation of an item,
81it will have been cleared to zeroes, however subsequent allocations
82will retain the contents as of the last free.
83.Pp
84The
85.Fn uma_zcreate
86function creates a new zone from which items may then be allocated from.
87The
88.Fa name
89argument is a text name of the zone for debugging and stats; this memory
90should not be freed until the zone has been deallocated.
91.Pp
92The
93.Fa ctor
94and
95.Fa dtor
96arguments are callback functions that are called by
97the uma subsystem at the time of the call to
98.Fn uma_zalloc
99and
100.Fn uma_zfree
101respectively.
102Their purpose is to provide hooks for initializing or
103destroying things that need to be done at the time of the allocation
104or release of a resource.
105A good usage for the
106.Fa ctor
107and
108.Fa dtor
109callbacks
110might be to adjust a global count of the number of objects allocated.
111.Pp
112The
113.Fa uminit
114and
115.Fa fini
116arguments are used to optimize the allocation of
117objects from the zone.
118They are called by the uma subsystem whenever
119it needs to allocate or free several items to satisfy requests or memory
120pressure.
121A good use for the
122.Fa uminit
123and
124.Fa fini
125callbacks might be to
126initialize and destroy mutexes contained within the object.
127This would
128allow one to re-use already initialized mutexes when an object is returned
129from the uma subsystem's object cache.
130They are not called on each call to
131.Fn uma_zalloc
132and
133.Fn uma_zfree
134but rather in a batch mode on several objects.
135.Pp
136To allocate an item from a zone, simply call
137.Fn uma_zalloc
138with a pointer to that zone
139and set the
140.Fa flags
141argument to selected flags as documented in
142.Xr malloc 9 .
143It will return a pointer to an item if successful,
144or
145.Dv NULL
146in the rare case where all items in the zone are in use and the
147allocator is unable to grow the zone
148or when
149.Dv M_NOWAIT
150is specified.
151.Pp
152Items are released back to the zone from which they were allocated by
153calling
154.Fn uma_zfree
155with a pointer to the zone and a pointer to the item.
156If
157.Fa item
158is
159.Dv NULL ,
160then
161.Fn uma_zfree
162does nothing.
163.Pp
164The variations
165.Fn uma_zalloc_arg
166and
167.Fn uma_zfree_arg
168allow to
169specify an argument for the
170.Dv ctor
171and
172.Dv dtor
173functions, respectively.
174.Pp
175Created zones,
176which are empty,
177can be destroyed using
178.Fn uma_zdestroy ,
179freeing all memory that was allocated for the zone.
180All items allocated from the zone with
181.Fn uma_zalloc
182must have been freed with
183.Fn uma_zfree
184before.
185.Pp
186The
187.Fn uma_zone_set_max
188function limits the number of items
189.Pq and therefore memory
190that can be allocated to
191.Fa zone .
192The
193.Fa nitems
194argument specifies the requested upper limit number of items.
195The effective limit is returned to the caller, as it may end up being higher
196than requested due to the implementation rounding up to ensure all memory pages
197allocated to the zone are utilised to capacity.
198The limit applies to the total number of items in the zone, which includes
199allocated items, free items and free items in the per-cpu caches.
200On systems with more than one CPU it may not be possible to allocate
201the specified number of items even when there is no shortage of memory,
202because all of the remaining free items may be in the caches of the
203other CPUs when the limit is hit.
204.Pp
205The
206.Fn uma_zone_get_max
207function returns the effective upper limit number of items for a zone.
208.Pp
209The
210.Fn uma_zone_get_cur
211function returns the approximate current occupancy of the zone.
212The returned value is approximate because appropriate synchronisation to
213determine an exact value is not performend by the implementation.
214This ensures low overhead at the expense of potentially stale data being used
215in the calculation.
216.Sh RETURN VALUES
217The
218.Fn uma_zalloc
219function returns a pointer to an item, or
220.Dv NULL
221if the zone ran out of unused items and the allocator was unable to
222enlarge it.
223.Sh SEE ALSO
224.Xr malloc 9
225.Sh HISTORY
226The zone allocator first appeared in
227.Fx 3.0 .
228It was radically changed in
229.Fx 5.0
230to function as a slab allocator.
231.Sh AUTHORS
232.An -nosplit
233The zone allocator was written by
234.An John S. Dyson .
235The zone allocator was rewritten in large parts by
236.An Jeff Roberson Aq jeff@FreeBSD.org
237to function as a slab allocator.
238.Pp
239This manual page was written by
240.An Dag-Erling Sm\(/orgrav Aq des@FreeBSD.org .
241Changes for UMA by
242.An Jeroen Ruigrok van der Werven Aq asmodai@FreeBSD.org .
243