1MOVE_PAGES(2) Linux Programmer's Manual MOVE_PAGES(2)
2
3
4
6 move_pages - move individual pages of a process to another node
7
9 #include <numaif.h>
10
11 long move_pages(int pid, unsigned long count, void **pages,
12 const int *nodes, int *status, int flags);
13
15 move_pages() moves the specified pages of the process pid to the memory
16 nodes specified by nodes. The result of the move is reflected in sta‐
17 tus. The flags indicate constraints on the pages to be moved.
18
19 pid is the ID of the process in which pages are to be moved. To move
20 pages in another process, the caller must be privileged (CAP_SYS_NICE)
21 or the real or effective user ID of the calling process must match the
22 real or saved-set user ID of the target process. If pid is 0 then
23 move_pages() moves pages of the calling process.
24
25 count is the number of pages to move. It defines the size of the three
26 arrays pages, nodes, and status.
27
28 pages is an array of pointers to the pages that should be moved. These
29 are pointers that should be aligned to page boundaries. Addresses are
30 specified as seen by the process specified by pid.
31
32 nodes is an array of integers that specify the desired location for
33 each page. Each element in the array is a node number. nodes can also
34 be NULL, in which case move_pages() does not move any pages but instead
35 will return the node where each page currently resides, in the status
36 array. Obtaining the status of each page may be necessary to determine
37 pages that need to be moved.
38
39 status is an array of integers that return the status of each page.
40 The array only contains valid values if move_pages() did not return an
41 error.
42
43 flags specify what types of pages to move. MPOL_MF_MOVE means that
44 only pages that are in exclusive use by the process are to be moved.
45 MPOL_MF_MOVE_ALL means that pages shared between multiple processes can
46 also be moved. The process must be privileged (CAP_SYS_NICE) to use
47 MPOL_MF_MOVE_ALL.
48
50 The following values can be returned in each element of the status
51 array.
52
53 0..MAX_NUMNODES
54 Identifies the node on which the page resides.
55
56 -EACCES
57 The page is mapped by multiple processes and can only be moved
58 if MPOL_MF_MOVE_ALL is specified.
59
60 -EBUSY The page is currently busy and cannot be moved. Try again
61 later. This occurs if a page is undergoing I/O or another ker‐
62 nel subsystem is holding a reference to the page.
63
64 -EFAULT
65 This is a zero page or the memory area is not mapped by the
66 process.
67
68 -EIO Unable to write back a page. The page has to be written back in
69 order to move it since the page is dirty and the file system
70 does not provide a migration function that would allow the move
71 of dirty pages.
72
73 -EINVAL
74 A dirty page cannot be moved. The file system does not provide
75 a migration function and has no ability to write back pages.
76
77 -ENOENT
78 The page is not present.
79
80 -ENOMEM
81 Unable to allocate memory on target node.
82
84 On success move_pages() returns zero. On error, it returns -1, and
85 sets errno to indicate the error.
86
88 E2BIG Too many pages to move.
89
90 EACCES One of the target nodes is not allowed by the current cpuset.
91
92 EFAULT Parameter array could not be accessed.
93
94 EINVAL Flags other than MPOL_MF_MOVE and MPOL_MF_MOVE_ALL was specified
95 or an attempt was made to migrate pages of a kernel thread.
96
97 ENODEV One of the target nodes is not online.
98
99 ENOENT No pages were found that require moving. All pages are either
100 already on the target node, not present, had an invalid address
101 or could not be moved because they were mapped by multiple pro‐
102 cesses.
103
104 EPERM The caller specified MPOL_MF_MOVE_ALL without sufficient privi‐
105 leges (CAP_SYS_NICE). Or, the caller attempted to move pages of
106 a process belonging to another user but did not have privilege
107 to do so (CAP_SYS_NICE).
108
109 ESRCH Process does not exist.
110
112 move_pages() first appeared on Linux in version 2.6.18.
113
115 This system call is Linux-specific.
116
118 For information on library support, see numa(7).
119
120 Use get_mempolicy(2) with the MPOL_F_MEMS_ALLOWED flag to obtain the
121 set of nodes that are allowed by the current cpuset. Note that this
122 information is subject to change at any time by manual or automatic
123 reconfiguration of the cpuset.
124
125 Use of this function may result in pages whose location [node] violates
126 the memory policy established for the specified addresses [See
127 mbind(2)] and/or the specified process [See set_mempolicy(2)]. That
128 is, memory policy does not constrain the destination nodes used by
129 move_pages().
130
132 get_mempolicy(2), mbind(2), set_mempolicy(2), numa(3), numa_maps(5),
133 cpuset(7), numa(7), migratepages(8), numa_stat(8)
134
136 This page is part of release 3.22 of the Linux man-pages project. A
137 description of the project, and information about reporting bugs, can
138 be found at http://www.kernel.org/doc/man-pages/.
139
140
141
142Linux 2008-10-06 MOVE_PAGES(2)