1SET_MEMPOLICY(2) Linux Programmer's Manual SET_MEMPOLICY(2)
2
3
4
6 set_mempolicy - set default NUMA memory policy for a thread and its
7 children
8
10 #include <numaif.h>
11
12 long set_mempolicy(int mode, const unsigned long *nodemask,
13 unsigned long maxnode);
14
15 Link with -lnuma.
16
18 set_mempolicy() sets the NUMA memory policy of the calling thread,
19 which consists of a policy mode and zero or more nodes, to the values
20 specified by the mode, nodemask, and maxnode arguments.
21
22 A NUMA machine has different memory controllers with different dis‐
23 tances to specific CPUs. The memory policy defines from which node
24 memory is allocated for the thread.
25
26 This system call defines the default policy for the thread. The thread
27 policy governs allocation of pages in the process's address space out‐
28 side of memory ranges controlled by a more specific policy set by
29 mbind(2). The thread default policy also controls allocation of any
30 pages for memory-mapped files mapped using the mmap(2) call with the
31 MAP_PRIVATE flag and that are only read (loaded) from by the thread and
32 of memory-mapped files mapped using the mmap(2) call with the
33 MAP_SHARED flag, regardless of the access type. The policy is applied
34 only when a new page is allocated for the thread. For anonymous memory
35 this is when the page is first touched by the thread.
36
37 The mode argument must specify one of MPOL_DEFAULT, MPOL_BIND, MPOL_IN‐
38 TERLEAVE, MPOL_PREFERRED, or MPOL_LOCAL (which are described in detail
39 below). All modes except MPOL_DEFAULT require the caller to specify
40 the node or nodes to which the mode applies, via the nodemask argument.
41
42 The mode argument may also include an optional mode flag. The sup‐
43 ported mode flags are:
44
45 MPOL_F_STATIC_NODES (since Linux 2.6.26)
46 A nonempty nodemask specifies physical node IDs. Linux will not
47 remap the nodemask when the process moves to a different cpuset
48 context, nor when the set of nodes allowed by the process's cur‐
49 rent cpuset context changes.
50
51 MPOL_F_RELATIVE_NODES (since Linux 2.6.26)
52 A nonempty nodemask specifies node IDs that are relative to the
53 set of node IDs allowed by the process's current cpuset.
54
55 nodemask points to a bit mask of node IDs that contains up to maxnode
56 bits. The bit mask size is rounded to the next multiple of sizeof(un‐
57 signed long), but the kernel will use bits only up to maxnode. A NULL
58 value of nodemask or a maxnode value of zero specifies the empty set of
59 nodes. If the value of maxnode is zero, the nodemask argument is ig‐
60 nored.
61
62 Where a nodemask is required, it must contain at least one node that is
63 on-line, allowed by the process's current cpuset context, (unless the
64 MPOL_F_STATIC_NODES mode flag is specified), and contains memory. If
65 the MPOL_F_STATIC_NODES is set in mode and a required nodemask contains
66 no nodes that are allowed by the process's current cpuset context, the
67 memory policy reverts to local allocation. This effectively overrides
68 the specified policy until the process's cpuset context includes one or
69 more of the nodes specified by nodemask.
70
71 The mode argument must include one of the following values:
72
73 MPOL_DEFAULT
74 This mode specifies that any nondefault thread memory policy be
75 removed, so that the memory policy "falls back" to the system
76 default policy. The system default policy is "local alloca‐
77 tion"—that is, allocate memory on the node of the CPU that trig‐
78 gered the allocation. nodemask must be specified as NULL. If
79 the "local node" contains no free memory, the system will at‐
80 tempt to allocate memory from a "near by" node.
81
82 MPOL_BIND
83 This mode defines a strict policy that restricts memory alloca‐
84 tion to the nodes specified in nodemask. If nodemask specifies
85 more than one node, page allocations will come from the node
86 with the lowest numeric node ID first, until that node contains
87 no free memory. Allocations will then come from the node with
88 the next highest node ID specified in nodemask and so forth, un‐
89 til none of the specified nodes contain free memory. Pages will
90 not be allocated from any node not specified in the nodemask.
91
92 MPOL_INTERLEAVE
93 This mode interleaves page allocations across the nodes speci‐
94 fied in nodemask in numeric node ID order. This optimizes for
95 bandwidth instead of latency by spreading out pages and memory
96 accesses to those pages across multiple nodes. However, ac‐
97 cesses to a single page will still be limited to the memory
98 bandwidth of a single node.
99
100 MPOL_PREFERRED
101 This mode sets the preferred node for allocation. The kernel
102 will try to allocate pages from this node first and fall back to
103 "near by" nodes if the preferred node is low on free memory. If
104 nodemask specifies more than one node ID, the first node in the
105 mask will be selected as the preferred node. If the nodemask
106 and maxnode arguments specify the empty set, then the policy
107 specifies "local allocation" (like the system default policy
108 discussed above).
109
110 MPOL_LOCAL (since Linux 3.8)
111 This mode specifies "local allocation"; the memory is allocated
112 on the node of the CPU that triggered the allocation (the "local
113 node"). The nodemask and maxnode arguments must specify the
114 empty set. If the "local node" is low on free memory, the ker‐
115 nel will try to allocate memory from other nodes. The kernel
116 will allocate memory from the "local node" whenever memory for
117 this node is available. If the "local node" is not allowed by
118 the process's current cpuset context, the kernel will try to al‐
119 locate memory from other nodes. The kernel will allocate memory
120 from the "local node" whenever it becomes allowed by the
121 process's current cpuset context.
122
123 The thread memory policy is preserved across an execve(2), and is in‐
124 herited by child threads created using fork(2) or clone(2).
125
127 On success, set_mempolicy() returns 0; on error, -1 is returned and er‐
128 rno is set to indicate the error.
129
131 EFAULT Part of all of the memory range specified by nodemask and maxn‐
132 ode points outside your accessible address space.
133
134 EINVAL mode is invalid. Or, mode is MPOL_DEFAULT and nodemask is
135 nonempty, or mode is MPOL_BIND or MPOL_INTERLEAVE and nodemask
136 is empty. Or, maxnode specifies more than a page worth of bits.
137 Or, nodemask specifies one or more node IDs that are greater
138 than the maximum supported node ID. Or, none of the node IDs
139 specified by nodemask are on-line and allowed by the process's
140 current cpuset context, or none of the specified nodes contain
141 memory. Or, the mode argument specified both
142 MPOL_F_STATIC_NODES and MPOL_F_RELATIVE_NODES.
143
144 ENOMEM Insufficient kernel memory was available.
145
147 The set_mempolicy() system call was added to the Linux kernel in ver‐
148 sion 2.6.7.
149
151 This system call is Linux-specific.
152
154 Memory policy is not remembered if the page is swapped out. When such
155 a page is paged back in, it will use the policy of the thread or memory
156 range that is in effect at the time the page is allocated.
157
158 For information on library support, see numa(7).
159
161 get_mempolicy(2), getcpu(2), mbind(2), mmap(2), numa(3), cpuset(7),
162 numa(7), numactl(8)
163
165 This page is part of release 5.10 of the Linux man-pages project. A
166 description of the project, information about reporting bugs, and the
167 latest version of this page, can be found at
168 https://www.kernel.org/doc/man-pages/.
169
170
171
172Linux 2020-12-21 SET_MEMPOLICY(2)