1MIGRATE_PAGES(2) Linux Programmer's Manual MIGRATE_PAGES(2)
2
6 migrate_pages - move all pages in a process to another set of nodes
7
9 #include <numaif.h>
10 long migrate_pages(int pid, unsigned long maxnode,
12 const unsigned long *old_nodes,
13 const unsigned long *new_nodes);
14 Note: There is no glibc wrapper for this system call; see NOTES.
16 Link with -lnuma.
18
20 migrate_pages() attempts to move all pages of the process pid that are
21 in memory nodes old_nodes to the memory nodes in new_nodes. Pages not
22 located in any node in old_nodes will not be migrated. As far as pos‐
23 sible, the kernel maintains the relative topology relationship inside
24 old_nodes during the migration to new_nodes.
25 The old_nodes and new_nodes arguments are pointers to bit masks of node
27 numbers, with up to maxnode bits in each mask. These masks are main‐
28 tained as arrays of unsigned long integers (in the last long integer,
29 the bits beyond those specified by maxnode are ignored). The maxnode
30 argument is the maximum node number in the bit mask plus one (this is
31 the same as in mbind(2), but different from select(2)).
32 The pid argument is the ID of the process whose pages are to be moved.
34 To move pages in another process, the caller must be privileged
35 (CAP_SYS_NICE) or the real or effective user ID of the calling process
36 must match the real or saved-set user ID of the target process. If pid
37 is 0, then migrate_pages() moves pages of the calling process.
38 Pages shared with another process will be moved only if the initiating
40 process has the CAP_SYS_NICE privilege.
41
43 On success migrate_pages() returns the number of pages that could not
44 be moved (i.e., a return of zero means that all pages were successfully
45 moved). On error, it returns -1, and sets errno to indicate the error.
46
48 EFAULT Part or all of the memory range specified by old_nodes/new_nodes
49 and maxnode points outside your accessible address space.
50 EINVAL The value specified by maxnode exceeds a kernel-imposed limit.
52 Or, old_nodes or new_nodes specifies one or more node IDs that
53 are greater than the maximum supported node ID. Or, none of the
54 node IDs specified by new_nodes are on-line and allowed by the
55 process's current cpuset context, or none of the specified nodes
56 contain memory.
57 EPERM Insufficient privilege (CAP_SYS_NICE) to move pages of the
59 process specified by pid, or insufficient privilege
60 (CAP_SYS_NICE) to access the specified target nodes.
61 ESRCH No process matching pid could be found.
63
65 The migrate_pages() system call first appeared on Linux in version
66 2.6.16.
67
69 This system call is Linux-specific.
70
72 Glibc does not provide a wrapper for this system call. For information
73 on library support, see numa(7).
74 Use get_mempolicy(2) with the MPOL_F_MEMS_ALLOWED flag to obtain the
76 set of nodes that are allowed by the calling process's cpuset. Note
77 that this information is subject to change at any time by manual or au‐
78 tomatic reconfiguration of the cpuset.
79 Use of migrate_pages() may result in pages whose location (node) vio‐
81 lates the memory policy established for the specified addresses (see
82 mbind(2)) and/or the specified process (see set_mempolicy(2)). That
83 is, memory policy does not constrain the destination nodes used by mi‐
84 grate_pages().
85 The <numaif.h> header is not included with glibc, but requires in‐
87 stalling libnuma-devel or a similar package.
88
90 get_mempolicy(2), mbind(2), set_mempolicy(2), numa(3), numa_maps(5),
91 cpuset(7), numa(7), migratepages(8), numastat(8)
92 Documentation/vm/page_migration.rst in the Linux kernel source tree
94
96 This page is part of release 5.13 of the Linux man-pages project. A
97 description of the project, information about reporting bugs, and the
98 latest version of this page, can be found at
99 https://www.kernel.org/doc/man-pages/.
100Linux 2021-03-22 MIGRATE_PAGES(2)