本文共 60482 字，大约阅读时间需要 201 分钟。
目录
GET_MEMPOLICY(2) Linux Programmer's Manual GET_MEMPOLICY(2)
NAME
get_mempolicy - retrieve NUMA memory policy for a thread
SYNOPSIS
#includelong get_mempolicy(int *mode, unsigned long *nodemask, unsigned long maxnode, void *addr, unsigned long flags);
Link with -lnuma. Note: There is no glibc wrapper for this system call; see NOTES.
DESCRIPTION
get_mempolicy() retrieves the NUMA policy of the calling thread or of a memory address, depending on the setting of flags. A NUMA machine has different memory controllers with different distances to specific CPUs. The memory policy defines from which node memory is allocated for the thread. If flags is specified as 0, then information about the calling thread's default policy (as set by ) is returned, in the buffers pointed to by mode and nodemask. The value returned in these arguments may be used to restore the thread's policy to its state at the time of the call to get_mempolicy() using . When flags is 0, addr must be specified as NULL. If flags specifies MPOL_F_MEMS_ALLOWED (available since Linux 2.6.24), the mode argument is ignored and the set of nodes (memories) that the thread is allowed to specify in subsequent calls to or (in the absence of any mode flags) is returned in nodemask. It is not permitted to combine MPOL_F_MEMS_ALLOWED with either MPOL_F_ADDR or MPOL_F_NODE. If flags specifies MPOL_F_ADDR, then information is returned about the policy governing the memory address given in addr. This policy may be different from the thread's default policy if or one of the helper functions described in has been used to establish a policy for the memory range containing addr. If the mode argument is not NULL, then get_mempolicy() will store the policy mode and any optional mode flags of the requested NUMA policy in the location pointed to by this argument. If nodemask is not NULL, then the nodemask associated with the policy will be stored in the location pointed to by this argument. maxnode specifies the number of node IDs that can be stored into nodemask—that is, the maximum node ID plus one. The value specified by maxnode is always rounded to a multiple of sizeof(unsigned long)*8. If flags specifies both MPOL_F_NODE and MPOL_F_ADDR, get_mempolicy() will return the node ID of the node on which the address addr is allocated into the location pointed to by mode. If no page has yet been allocated for the specified address, get_mempolicy() will allocate a page as if the thread had performed a read (load) access to that address, and return the ID of the node where that page was allocated. If flags specifies MPOL_F_NODE, but not MPOL_F_ADDR, and the thread's current policy is MPOL_INTERLEAVE, then get_mempolicy() will return in the location pointed to by a non-NULL mode argument, the node ID of the next node that will be used for interleaving of internal kernel pages allocated on behalf of the thread. These allocations include pages for memory-mapped files in process memory ranges mapped using the call with the MAP_PRIVATE flag for read accesses, and in memory ranges mapped with the MAP_SHARED flag for all accesses. Other flag values are reserved. For an overview of the possible policies see .
RETURN VALUE
On success, get_mempolicy() returns 0; on error, -1 is returned and is set to indicate the error.
ERRORS
EFAULT Part of all of the memory range specified by nodemask and maxnode points outside your accessible address space. EINVAL The value specified by maxnode is less than the number of node IDs supported by the system. Or flags specified values other than MPOL_F_NODE or MPOL_F_ADDR; or flags specified MPOL_F_ADDR and addr is NULL, or flags did not specify MPOL_F_ADDR and addr is not NULL. Or, flags specified MPOL_F_NODE but not MPOL_F_ADDR and the current thread policy is not MPOL_INTERLEAVE. Or, flags specified MPOL_F_MEMS_ALLOWED with either MPOL_F_ADDR or MPOL_F_NODE. (And there are other EINVAL cases.)
VERSIONS
The get_mempolicy() system call was added to the Linux kernel in version 2.6.7.
CONFORMING TO
This system call is Linux-specific.
NOTES
Glibc does not provide a wrapper for this system call. For information on library support, see .
SEE ALSO
, , , , , ,
COLOPHON
This page is part of release 5.11 of the Linux man-pages project. A description of the project, information about reporting bugs, and the latest version of this page, can be found at .Linux 2021-03-22 GET_MEMPOLICY(2)
Pages that refer to this page: , , , , , , , , ,
DEMO
intrte_vhost_get_numa_node(int vid){#ifdef RTE_LIBRTE_VHOST_NUMA struct virtio_net *dev = get_device(vid); int numa_node; int ret; if (dev == NULL || numa_available() != 0) return -1; ret = get_mempolicy(&numa_node, NULL, 0, dev, MPOL_F_NODE | MPOL_F_ADDR); if (ret < 0) { VHOST_LOG_CONFIG(ERR, "(%d) failed to query numa node: %s\n", vid, rte_strerror(errno)); return -1; } return numa_node;#else RTE_SET_USED(vid); return -1;#endif}
SET_MEMPOLICY(2) Linux Programmer's Manual SET_MEMPOLICY(2)
NAME
set_mempolicy - set default NUMA memory policy for a thread and its children
SYNOPSIS
#includelong set_mempolicy(int mode, const unsigned long *nodemask, unsigned long maxnode);
Link with -lnuma.
DESCRIPTION
set_mempolicy() sets the NUMA memory policy of the calling thread, which consists of a policy mode and zero or more nodes, to the values specified by the mode, nodemask, and maxnode arguments. A NUMA machine has different memory controllers with different distances to specific CPUs. The memory policy defines from which node memory is allocated for the thread. This system call defines the default policy for the thread. The thread policy governs allocation of pages in the process's address space outside of memory ranges controlled by a more specific policy set by . The thread default policy also controls allocation of any pages for memory-mapped files mapped using the call with the MAP_PRIVATE flag and that are only read (loaded) from by the thread and of memory-mapped files mapped using the call with the MAP_SHARED flag, regardless of the access type. The policy is applied only when a new page is allocated for the thread. For anonymous memory this is when the page is first touched by the thread. The mode argument must specify one of MPOL_DEFAULT, MPOL_BIND, MPOL_INTERLEAVE, MPOL_PREFERRED, or MPOL_LOCAL (which are described in detail below). All modes except MPOL_DEFAULT require the caller to specify the node or nodes to which the mode applies, via the nodemask argument. The mode argument may also include an optional mode flag. The supported mode flags are: MPOL_F_STATIC_NODES (since Linux 2.6.26) A nonempty nodemask specifies physical node IDs. Linux will not remap the nodemask when the process moves to a different cpuset context, nor when the set of nodes allowed by the process's current cpuset context changes. MPOL_F_RELATIVE_NODES (since Linux 2.6.26) A nonempty nodemask specifies node IDs that are relative to the set of node IDs allowed by the process's current cpuset. nodemask points to a bit mask of node IDs that contains up to maxnode bits. The bit mask size is rounded to the next multiple of sizeof(unsigned long), but the kernel will use bits only up to maxnode. A NULL value of nodemask or a maxnode value of zero specifies the empty set of nodes. If the value of maxnode is zero, the nodemask argument is ignored. Where a nodemask is required, it must contain at least one node that is on-line, allowed by the process's current cpuset context, (unless the MPOL_F_STATIC_NODES mode flag is specified), and contains memory. If the MPOL_F_STATIC_NODES is set in mode and a required nodemask contains no nodes that are allowed by the process's current cpuset context, the memory policy reverts to local allocation. This effectively overrides the specified policy until the process's cpuset context includes one or more of the nodes specified by nodemask. The mode argument must include one of the following values: MPOL_DEFAULT This mode specifies that any nondefault thread memory policy be removed, so that the memory policy "falls back" to the system default policy. The system default policy is "local allocation"—that is, allocate memory on the node of the CPU that triggered the allocation. nodemask must be specified as NULL. If the "local node" contains no free memory, the system will attempt to allocate memory from a "near by" node. MPOL_BIND This mode defines a strict policy that restricts memory allocation to the nodes specified in nodemask. If nodemask specifies more than one node, page allocations will come from the node with the lowest numeric node ID first, until that node contains no free memory. Allocations will then come from the node with the next highest node ID specified in nodemask and so forth, until none of the specified nodes contain free memory. Pages will not be allocated from any node not specified in the nodemask. MPOL_INTERLEAVE This mode interleaves page allocations across the nodes specified in nodemask in numeric node ID order. This optimizes for bandwidth instead of latency by spreading out pages and memory accesses to those pages across multiple nodes. However, accesses to a single page will still be limited to the memory bandwidth of a single node. MPOL_PREFERRED This mode sets the preferred node for allocation. The kernel will try to allocate pages from this node first and fall back to "near by" nodes if the preferred node is low on free memory. If nodemask specifies more than one node ID, the first node in the mask will be selected as the preferred node. If the nodemask and maxnode arguments specify the empty set, then the policy specifies "local allocation" (like the system default policy discussed above). MPOL_LOCAL (since Linux 3.8) This mode specifies "local allocation"; the memory is allocated on the node of the CPU that triggered the allocation (the "local node"). The nodemask and maxnode arguments must specify the empty set. If the "local node" is low on free memory, the kernel will try to allocate memory from other nodes. The kernel will allocate memory from the "local node" whenever memory for this node is available. If the "local node" is not allowed by the process's current cpuset context, the kernel will try to allocate memory from other nodes. The kernel will allocate memory from the "local node" whenever it becomes allowed by the process's current cpuset context. The thread memory policy is preserved across an , and is inherited by child threads created using or .
RETURN VALUE
On success, set_mempolicy() returns 0; on error, -1 is returned and is set to indicate the error.
ERRORS
EFAULT Part of all of the memory range specified by nodemask and maxnode points outside your accessible address space. EINVAL mode is invalid. Or, mode is MPOL_DEFAULT and nodemask is nonempty, or mode is MPOL_BIND or MPOL_INTERLEAVE and nodemask is empty. Or, maxnode specifies more than a page worth of bits. Or, nodemask specifies one or more node IDs that are greater than the maximum supported node ID. Or, none of the node IDs specified by nodemask are on-line and allowed by the process's current cpuset context, or none of the specified nodes contain memory. Or, the mode argument specified both MPOL_F_STATIC_NODES and MPOL_F_RELATIVE_NODES. ENOMEM Insufficient kernel memory was available.
VERSIONS
The set_mempolicy() system call was added to the Linux kernel in version 2.6.7.
CONFORMING TO
This system call is Linux-specific.
NOTES
Memory policy is not remembered if the page is swapped out. When such a page is paged back in, it will use the policy of the thread or memory range that is in effect at the time the page is allocated. For information on library support, see .
SEE ALSO
, , , , , , ,
COLOPHON
This page is part of release 5.11 of the Linux man-pages project. A description of the project, information about reporting bugs, and the latest version of this page, can be found at .Linux 2020-12-21 SET_MEMPOLICY(2)
Pages that refer to this page: , , , , , , , , , , , ,
MBIND(2) Linux Programmer's Manual MBIND(2)
NAME
mbind - set memory policy for a memory range
SYNOPSIS
#includelong mbind(void *addr, unsigned long len, int mode, const unsigned long *nodemask, unsigned long maxnode, unsigned int flags);
Link with -lnuma. Note: There is no glibc wrapper for this system call; see NOTES.
DESCRIPTION
mbind() sets the NUMA memory policy, which consists of a policy mode and zero or more nodes, for the memory range starting with addr and continuing for len bytes. The memory policy defines from which node memory is allocated. If the memory range specified by the addr and len arguments includes an "anonymous" region of memory—that is a region of memory created using the system call with the MAP_ANONYMOUS—or a memory-mapped file, mapped using the system call with the MAP_PRIVATE flag, pages will be allocated only according to the specified policy when the application writes (stores) to the page. For anonymous regions, an initial read access will use a shared page in the kernel containing all zeros. For a file mapped with MAP_PRIVATE, an initial read access will allocate pages according to the memory policy of the thread that causes the page to be allocated. This may not be the thread that called mbind(). The specified policy will be ignored for any MAP_SHARED mappings in the specified memory range. Rather the pages will be allocated according to the memory policy of the thread that caused the page to be allocated. Again, this may not be the thread that called mbind(). If the specified memory range includes a shared memory region created using the system call and attached using the system call, pages allocated for the anonymous or shared memory region will be allocated according to the policy specified, regardless of which process attached to the shared memory segment causes the allocation. If, however, the shared memory region was created with the SHM_HUGETLB flag, the huge pages will be allocated according to the policy specified only if the page allocation is caused by the process that calls mbind() for that region. By default, mbind() has an effect only for new allocations; if the pages inside the range have been already touched before setting the policy, then the policy has no effect. This default behavior may be overridden by the MPOL_MF_MOVE and MPOL_MF_MOVE_ALL flags described below. The mode argument must specify one of MPOL_DEFAULT, MPOL_BIND, MPOL_INTERLEAVE, MPOL_PREFERRED, or MPOL_LOCAL (which are described in detail below). All policy modes except MPOL_DEFAULT require the caller to specify the node or nodes to which the mode applies, via the nodemask argument. The mode argument may also include an optional mode flag. The supported mode flags are: MPOL_F_STATIC_NODES (since Linux-2.6.26) A nonempty nodemask specifies physical node IDs. Linux does not remap the nodemask when the thread moves to a different cpuset context, nor when the set of nodes allowed by the thread's current cpuset context changes. MPOL_F_RELATIVE_NODES (since Linux-2.6.26) A nonempty nodemask specifies node IDs that are relative to the set of node IDs allowed by the thread's current cpuset. nodemask points to a bit mask of nodes containing up to maxnode bits. The bit mask size is rounded to the next multiple of sizeof(unsigned long), but the kernel will use bits only up to maxnode. A NULL value of nodemask or a maxnode value of zero specifies the empty set of nodes. If the value of maxnode is zero, the nodemask argument is ignored. Where a nodemask is required, it must contain at least one node that is on-line, allowed by the thread's current cpuset context (unless the MPOL_F_STATIC_NODES mode flag is specified), and contains memory. The mode argument must include one of the following values: MPOL_DEFAULT This mode requests that any nondefault policy be removed, restoring default behavior. When applied to a range of memory via mbind(), this means to use the thread memory policy, which may have been set with . If the mode of the thread memory policy is also MPOL_DEFAULT, the system-wide default policy will be used. The system- wide default policy allocates pages on the node of the CPU that triggers the allocation. For MPOL_DEFAULT, the nodemask and maxnode arguments must be specify the empty set of nodes. MPOL_BIND This mode specifies a strict policy that restricts memory allocation to the nodes specified in nodemask. If nodemask specifies more than one node, page allocations will come from the node with sufficient free memory that is closest to the node where the allocation takes place. Pages will not be allocated from any node not specified in the IR nodemask . (Before Linux 2.6.26, page allocations came from the node with the lowest numeric node ID first, until that node contained no free memory. Allocations then came from the node with the next highest node ID specified in nodemask and so forth, until none of the specified nodes contained free memory.) MPOL_INTERLEAVE(交错) This mode specifies that page allocations be interleaved across the set of nodes specified in nodemask. This optimizes for bandwidth instead of latency by spreading out pages and memory accesses to those pages across multiple nodes. To be effective the memory area should be fairly large, at least 1 MB or bigger with a fairly uniform access pattern. Accesses to a single page of the area will still be limited to the memory bandwidth of a single node. MPOL_PREFERRED(首选的；优先；更喜欢) This mode sets the preferred node for allocation. The kernel will try to allocate pages from this node first and fall back to other nodes if the preferred nodes is low on free memory. If nodemask specifies more than one node ID, the first node in the mask will be selected as the preferred node. If the nodemask and maxnode arguments specify the empty set, then the memory is allocated on the node of the CPU that triggered the allocation. MPOL_LOCAL (since Linux 3.8) This mode specifies "local allocation"; the memory is allocated on the node of the CPU that triggered the allocation (the "local node"). The nodemask and maxnode arguments must specify the empty set. If the "local node" is low on free memory, the kernel will try to allocate memory from other nodes. The kernel will allocate memory from the "local node" whenever memory for this node is available. If the "local node" is not allowed by the thread's current cpuset context, the kernel will try to allocate memory from other nodes. The kernel will allocate memory from the "local node" whenever it becomes allowed by the thread's current cpuset context. By contrast, MPOL_DEFAULT reverts to the memory policy of the thread (which may be set via ); that policy may be something other than "local allocation". If MPOL_MF_STRICT is passed in flags and mode is not MPOL_DEFAULT, then the call fails with the error EIO if the existing pages in the memory range don't follow the policy. If MPOL_MF_MOVE is specified in flags, then the kernel will attempt to move all the existing pages in the memory range so that they follow the policy. Pages that are shared with other processes will not be moved. If MPOL_MF_STRICT is also specified, then the call fails with the error EIO if some pages could not be moved. If MPOL_MF_MOVE_ALL is passed in flags, then the kernel will attempt to move all existing pages in the memory range regardless of whether other processes use the pages. The calling thread must be privileged (CAP_SYS_NICE) to use this flag. If MPOL_MF_STRICT is also specified, then the call fails with the error EIO if some pages could not be moved.
RETURN VALUE
On success, mbind() returns 0; on error, -1 is returned and is set to indicate the error.
ERRORS
EFAULT Part or all of the memory range specified by nodemask and maxnode points outside your accessible address space. Or, there was an unmapped hole in the specified memory range specified by addr and len. EINVAL An invalid value was specified for flags or mode; or addr + len was less than addr; or addr is not a multiple of the system page size. Or, mode is MPOL_DEFAULT and nodemask specified a nonempty set; or mode is MPOL_BIND or MPOL_INTERLEAVE and nodemask is empty. Or, maxnode exceeds a kernel-imposed limit. Or, nodemask specifies one or more node IDs that are greater than the maximum supported node ID. Or, none of the node IDs specified by nodemask are on-line and allowed by the thread's current cpuset context, or none of the specified nodes contain memory. Or, the mode argument specified both MPOL_F_STATIC_NODES and MPOL_F_RELATIVE_NODES. EIO MPOL_MF_STRICT was specified and an existing page was already on a node that does not follow the policy; or MPOL_MF_MOVE or MPOL_MF_MOVE_ALL was specified and the kernel was unable to move all existing pages in the range. ENOMEM Insufficient kernel memory was available. EPERM The flags argument included the MPOL_MF_MOVE_ALL flag and the caller does not have the CAP_SYS_NICE privilege.
VERSIONS
The mbind() system call was added to the Linux kernel in version 2.6.7.
CONFORMING TO
This system call is Linux-specific.
NOTES
Glibc does not provide a wrapper for this system call. For information on library support, see . NUMA policy is not supported on a memory-mapped file range that was mapped with the MAP_SHARED flag. The MPOL_DEFAULT mode can have different effects for mbind() and . When MPOL_DEFAULT is specified for , the thread's memory policy reverts to the system default policy or local allocation. When MPOL_DEFAULT is specified for a range of memory using mbind(), any pages subsequently allocated for that range will use the thread's memory policy, as set by . This effectively removes the explicit policy from the specified range, "falling back" to a possibly nondefault policy. To select explicit "local allocation" for a memory range, specify a mode of MPOL_LOCAL or MPOL_PREFERRED with an empty set of nodes. This method will work for , as well. Support for huge page policy was added with 2.6.16. For interleave policy to be effective on huge page mappings the policied memory needs to be tens of megabytes or larger. Before Linux 5.7. MPOL_MF_STRICT was ignored on huge page mappings. MPOL_MF_MOVE and MPOL_MF_MOVE_ALL are available only on Linux 2.6.16 and later.
SEE ALSO
, , , , , , , , ,
COLOPHON
This page is part of release 5.11 of the Linux man-pages project. A description of the project, information about reporting bugs, and the latest version of this page, can be found at .Linux 2021-03-22 MBIND(2)
Pages that refer to this page: , , , , , , , , , , , , ,
NUMA(3) Linux Programmer's Manual NUMA(3)
NAME
numa - NUMA policy library
SYNOPSIS
#includecc ... -lnuma int numa_available(void); int numa_max_possible_node(void); int numa_num_possible_nodes(); int numa_max_node(void); int numa_num_configured_nodes(); struct bitmask *numa_get_mems_allowed(void); int numa_num_configured_cpus(void); struct bitmask *numa_all_nodes_ptr; struct bitmask *numa_no_nodes_ptr; struct bitmask *numa_all_cpus_ptr; int numa_num_task_cpus(); int numa_num_task_nodes(); int numa_parse_bitmap(char *line , struct bitmask *mask); struct bitmask *numa_parse_nodestring(const char *string); struct bitmask *numa_parse_nodestring_all(const char *string); struct bitmask *numa_parse_cpustring(const char *string); struct bitmask *numa_parse_cpustring_all(const char *string); long numa_node_size(int node, long *freep); long long numa_node_size64(int node, long long *freep); int numa_preferred(void); void numa_set_preferred(int node); int numa_get_interleave_node(void); struct bitmask *numa_get_interleave_mask(void); void numa_set_interleave_mask(struct bitmask *nodemask); void numa_interleave_memory(void *start, size_t size, struct bitmask *nodemask); void numa_bind(struct bitmask *nodemask); void numa_set_localalloc(void); void numa_set_membind(struct bitmask *nodemask); void numa_set_membind_balancing(struct bitmask *nodemask); struct bitmask *numa_get_membind(void); void *numa_alloc_onnode(size_t size, int node); void *numa_alloc_local(size_t size); void *numa_alloc_interleaved(size_t size); void *numa_alloc_interleaved_subset(size_t size, struct bitmask *nodemask); void *numa_alloc(size_t size); void *numa_realloc(void *old_addr, size_t old_size, size_t new_size); void numa_free(void *start, size_t size); int numa_run_on_node(int node); int numa_run_on_node_mask(struct bitmask *nodemask); int numa_run_on_node_mask_all(struct bitmask *nodemask); struct bitmask *numa_get_run_node_mask(void); void numa_tonode_memory(void *start, size_t size, int node); void numa_tonodemask_memory(void *start, size_t size, struct bitmask *nodemask); void numa_setlocal_memory(void *start, size_t size); void numa_police_memory(void *start, size_t size); void numa_set_bind_policy(int strict); void numa_set_strict(int strict); int numa_distance(int node1, int node2); int numa_sched_getaffinity(pid_t pid, struct bitmask *mask); int numa_sched_setaffinity(pid_t pid, struct bitmask *mask); int numa_node_to_cpus(int node, struct bitmask *mask); void numa_node_to_cpu_update(); int numa_node_of_cpu(int cpu); struct bitmask *numa_allocate_cpumask(); void numa_free_cpumask(); struct bitmask *numa_allocate_nodemask(); void numa_free_nodemask(); struct bitmask *numa_bitmask_alloc(unsigned int n); struct bitmask *numa_bitmask_clearall(struct bitmask *bmp); struct bitmask *numa_bitmask_clearbit(struct bitmask *bmp, unsigned int n); int numa_bitmask_equal(const struct bitmask *bmp1, const struct bitmask *bmp2); void numa_bitmask_free(struct bitmask *bmp); int numa_bitmask_isbitset(const struct bitmask *bmp, unsigned int n); unsigned int numa_bitmask_nbytes(struct bitmask *bmp); struct bitmask *numa_bitmask_setall(struct bitmask *bmp); struct bitmask *numa_bitmask_setbit(struct bitmask *bmp, unsigned int n); void copy_bitmask_to_nodemask(struct bitmask *bmp, nodemask_t *nodemask) void copy_nodemask_to_bitmask(nodemask_t *nodemask, struct bitmask *bmp) void copy_bitmask_to_bitmask(struct bitmask *bmpfrom, struct bitmask *bmpto) unsigned int numa_bitmask_weight(const struct bitmask *bmp ) int numa_move_pages(int pid, unsigned long count, void **pages, const int *nodes, int *status, int flags); int numa_migrate_pages(int pid, struct bitmask *fromnodes, struct bitmask *tonodes); void numa_error(char *where); extern int numa_exit_on_error; extern int numa_exit_on_warn; void numa_warn(int number, char *where, ...);
DESCRIPTION
The libnuma library offers a simple programming interface to the NUMA (Non Uniform Memory Access) policy supported by the Linux kernel. On a NUMA architecture some memory areas have different latency or bandwidth than others. Available policies are page interleaving (i.e., allocate in a round-robin fashion from all, or a subset, of the nodes on the system), preferred node allocation (i.e., preferably allocate on a particular node), local allocation (i.e., allocate on the node on which the task is currently executing), or allocation only on specific nodes (i.e., allocate on some subset of the available nodes). It is also possible to bind tasks to specific nodes. Numa memory allocation policy may be specified as a per-task attribute, that is inherited by children tasks and processes, or as an attribute of a range of process virtual address space. Numa memory policies specified for a range of virtual address space are shared by all tasks in the process. Furthermore, memory policies specified for a range of a shared memory attached using or from shmfs/hugetlbfs are shared by all processes that attach to that region. Memory policies for shared disk backed file mappings are currently ignored. The default memory allocation policy for tasks and all memory range is local allocation. This assumes that no ancestor has installed a non-default policy. For setting a specific policy globally for all memory allocations in a process and its children it is easiest to start it with the utility. For more finegrained policy inside an application this library can be used. All numa memory allocation policy only takes effect when a page is actually faulted into the address space of a process by accessing it. The numa_alloc_* functions take care of this automatically. A node is defined as an area where all memory has the same speed as seen from a particular CPU. A node can contain multiple CPUs. Caches are ignored for this definition. Most functions in this library are only concerned about numa nodes and their memory. The exceptions to this are: numa_node_to_cpus(), numa_node_to_cpu_update(), numa_node_of_cpu(), numa_bind(), numa_run_on_node(), numa_run_on_node_mask(), numa_run_on_node_mask_all(), and numa_get_run_node_mask(). These functions deal with the CPUs associated with numa nodes. See the descriptions below for more information. Some of these functions accept or return a pointer to struct bitmask. A struct bitmask controls a bit map of arbitrary length containing a bit representation of nodes. The predefined variable numa_all_nodes_ptr points to a bit mask that has all available nodes set; numa_no_nodes_ptr points to the empty set. Before any other calls in this library can be used numa_available() must be called. If it returns -1, all other functions in this library are undefined. numa_max_possible_node() returns the number of the highest possible node in a system. In other words, the size of a kernel type nodemask_t (in bits) minus 1. This number can be gotten by calling numa_num_possible_nodes() and subtracting 1. numa_num_possible_nodes() returns the size of kernel's node mask (kernel type nodemask_t). In other words, large enough to represent the maximum number of nodes that the kernel can handle. This will match the kernel's MAX_NUMNODES value. This count is derived from /proc/self/status, field Mems_allowed. numa_max_node() returns the highest node number available on the current system. (See the node numbers in /sys/devices/system/node/ ). Also see numa_num_configured_nodes(). numa_num_configured_nodes() returns the number of memory nodes in the system. This count includes any nodes that are currently disabled. This count is derived from the node numbers in /sys/devices/system/node. (Depends on the kernel being configured with /sys (CONFIG_SYSFS)). numa_get_mems_allowed() returns the mask of nodes from which the process is allowed to allocate memory in it's current cpuset context. Any nodes that are not included in the returned bitmask will be ignored in any of the following libnuma memory policy calls. numa_num_configured_cpus() returns the number of cpus in the system. This count includes any cpus that are currently disabled. This count is derived from the cpu numbers in /sys/devices/system/cpu. If the kernel is configured without /sys (CONFIG_SYSFS=n) then it falls back to using the number of online cpus. numa_all_nodes_ptr points to a bitmask that is allocated by the library with bits representing all nodes on which the calling task may allocate memory. This set may be up to all nodes on the system, or up to the nodes in the current cpuset. The bitmask is allocated by a call to numa_allocate_nodemask() using size numa_max_possible_node(). The set of nodes to record is derived from /proc/self/status, field "Mems_allowed". The user should not alter this bitmask. numa_no_nodes_ptr points to a bitmask that is allocated by the library and left all zeroes. The bitmask is allocated by a call to numa_allocate_nodemask() using size numa_max_possible_node(). The user should not alter this bitmask. numa_all_cpus_ptr points to a bitmask that is allocated by the library with bits representing all cpus on which the calling task may execute. This set may be up to all cpus on the system, or up to the cpus in the current cpuset. The bitmask is allocated by a call to numa_allocate_cpumask() using size numa_num_possible_cpus(). The set of cpus to record is derived from /proc/self/status, field "Cpus_allowed". The user should not alter this bitmask. numa_num_task_cpus() returns the number of cpus that the calling task is allowed to use. This count is derived from the map /proc/self/status, field "Cpus_allowed". Also see the bitmask numa_all_cpus_ptr. numa_num_task_nodes() returns the number of nodes on which the calling task is allowed to allocate memory. This count is derived from the map /proc/self/status, field "Mems_allowed". Also see the bitmask numa_all_nodes_ptr. numa_parse_bitmap() parses line , which is a character string such as found in /sys/devices/system/node/nodeN/cpumap into a bitmask structure. The string contains the hexadecimal representation of a bit map. The bitmask may be allocated with numa_allocate_cpumask(). Returns 0 on success. Returns -1 on failure. This function is probably of little use to a user application, but it is used by libnuma internally. numa_parse_nodestring() parses a character string list of nodes into a bit mask. The bit mask is allocated by numa_allocate_nodemask(). The string is a comma-separated list of node numbers or node ranges. A leading ! can be used to indicate "not" this list (in other words, all nodes except this list), and a leading + can be used to indicate that the node numbers in the list are relative to the task's cpuset. The string can be "all" to specify all ( numa_num_task_nodes() ) nodes. Node numbers are limited by the number in the system. See numa_max_node() and numa_num_configured_nodes(). Examples: 1-5,7,10 !4-5 +0-3 If the string is of 0 length, bitmask numa_no_nodes_ptr is returned. Returns 0 if the string is invalid. numa_parse_nodestring_all() is similar to numa_parse_nodestring , but can parse all possible nodes, not only current nodeset. numa_parse_cpustring() parses a character string list of cpus into a bit mask. The bit mask is allocated by numa_allocate_cpumask(). The string is a comma-separated list of cpu numbers or cpu ranges. A leading ! can be used to indicate "not" this list (in other words, all cpus except this list), and a leading + can be used to indicate that the cpu numbers in the list are relative to the task's cpuset. The string can be "all" to specify all ( numa_num_task_cpus() ) cpus. Cpu numbers are limited by the number in the system. See numa_num_task_cpus() and numa_num_configured_cpus(). Examples: 1-5,7,10 !4-5 +0-3 Returns 0 if the string is invalid. numa_parse_cpustring_all() is similar to numa_parse_cpustring , but can parse all possible cpus, not only current cpuset. numa_node_size() returns the memory size of a node. If the argument freep is not NULL, it used to return the amount of free memory on the node. On error it returns -1. numa_node_size64() works the same as numa_node_size() except that it returns values as long long instead of long. This is useful on 32-bit architectures with large nodes. numa_preferred() returns the preferred node of the current task. This is the node on which the kernel preferably allocates memory, unless some other policy overrides this. numa_set_preferred() sets the preferred node for the current task to node. The system will attempt to allocate memory from the preferred node, but will fall back to other nodes if no memory is available on the the preferred node. Passing a node of -1 argument specifies local allocation and is equivalent to calling numa_set_localalloc(). numa_get_interleave_mask() returns the current interleave mask if the task's memory allocation policy is page interleaved. Otherwise, this function returns an empty mask. numa_set_interleave_mask() sets the memory interleave mask for the current task to nodemask. All new memory allocations are page interleaved over all nodes in the interleave mask. Interleaving can be turned off again by passing an empty mask (numa_no_nodes). The page interleaving only occurs on the actual page fault that puts a new page into the current address space. It is also only a hint: the kernel will fall back to other nodes if no memory is available on the interleave target. numa_interleave_memory() interleaves size bytes of memory page by page from start on nodes specified in nodemask. The size argument will be rounded up to a multiple of the system page size. If nodemask contains nodes that are externally denied to this process, this call will fail. This is a lower level function to interleave allocated but not yet faulted in memory. Not yet faulted in means the memory is allocated using or , but has not been accessed by the current process yet. The memory is page interleaved to all nodes specified in nodemask. Normally numa_alloc_interleaved() should be used for private memory instead, but this function is useful to handle shared memory areas. To be useful the memory area should be several megabytes at least (or tens of megabytes of hugetlbfs mappings) If the numa_set_strict() flag is true then the operation will cause a numa_error if there were already pages in the mapping that do not follow the policy. numa_bind() binds the current task and its children to the nodes specified in nodemask. They will only run on the CPUs of the specified nodes and only be able to allocate memory from them. This function is equivalent to calling numa_run_on_node_mask(nodemask) followed by numa_set_membind(nodemask). If tasks should be bound to individual CPUs inside nodes consider using numa_node_to_cpus and the syscall. numa_set_localalloc() sets the memory allocation policy for the calling task to local allocation. In this mode, the preferred node for memory allocation is effectively the node where the task is executing at the time of a page allocation. numa_set_membind() sets the memory allocation mask. The task will only allocate memory from the nodes set in nodemask. Passing an empty nodemask or a nodemask that contains nodes other than those in the mask returned by numa_get_mems_allowed() will result in an error. numa_set_membind_balancing() sets the memory allocation mask and enable the Linux kernel NUMA balancing for the task if the feature is supported by the kernel. The task will only allocate memory from the nodes set in nodemask. Passing an empty nodemask or a nodemask that contains nodes other than those in the mask returned by numa_get_mems_allowed() will result in an error. numa_get_membind() returns the mask of nodes from which memory can currently be allocated. If the returned mask is equal to numa_all_nodes, then memory allocation is allowed from all nodes. numa_alloc_onnode() allocates memory on a specific node. The size argument will be rounded up to a multiple of the system page size. if the specified node is externally denied to this process, this call will fail. This function is relatively slow compared to the , family of functions. The memory must be freed with numa_free(). On errors NULL is returned. numa_alloc_local() allocates size bytes of memory on the local node. The size argument will be rounded up to a multiple of the system page size. This function is relatively slow compared to the family of functions. The memory must be freed with numa_free(). On errors NULL is returned. numa_alloc_interleaved() allocates size bytes of memory page interleaved on all nodes. This function is relatively slow and should only be used for large areas consisting of multiple pages. The interleaving works at page level and will only show an effect when the area is large. The allocated memory must be freed with numa_free(). On error, NULL is returned. numa_alloc_interleaved_subset() attempts to allocate size bytes of memory page interleaved on all nodes. The size argument will be rounded up to a multiple of the system page size. The nodes on which a process is allowed to allocate memory may be constrained externally. If this is the case, this function may fail. This function is relatively slow compare to , family of functions and should only be used for large areas consisting of multiple pages. The interleaving works at page level and will only show an effect when the area is large. The allocated memory must be freed with numa_free(). On error, NULL is returned. numa_alloc() allocates size bytes of memory with the current NUMA policy. The size argument will be rounded up to a multiple of the system page size. This function is relatively slow compare to the family of functions. The memory must be freed with numa_free(). On errors NULL is returned. numa_realloc() changes the size of the memory area pointed to by old_addr from old_size to new_size. The memory area pointed to by old_addr must have been allocated with one of the numa_alloc* functions. The new_size will be rounded up to a multiple of the system page size. The contents of the memory area will be unchanged to the minimum of the old and new sizes; newly allocated memory will be uninitialized. The memory policy (and node bindings) associated with the original memory area will be preserved in the resized area. For example, if the initial area was allocated with a call to numa_alloc_onnode(), then the new pages (if the area is enlarged) will be allocated on the same node. However, if no memory policy was set for the original area, then numa_realloc() cannot guarantee that the new pages will be allocated on the same node. On success, the address of the resized area is returned (which might be different from that of the initial area), otherwise NULL is returned and is set to indicate the error. The pointer returned by numa_realloc() is suitable for passing to numa_free(). numa_free() frees size bytes of memory starting at start, allocated by the numa_alloc_* functions above. The size argument will be rounded up to a multiple of the system page size. numa_run_on_node() runs the current task and its children on a specific node. They will not migrate to CPUs of other nodes until the node affinity is reset with a new call to numa_run_on_node_mask(). Passing -1 permits the kernel to schedule on all nodes again. On success, 0 is returned; on error -1 is returned, and is set to indicate the error. numa_run_on_node_mask() runs the current task and its children only on nodes specified in nodemask. They will not migrate to CPUs of other nodes until the node affinity is reset with a new call to numa_run_on_node_mask() or numa_run_on_node(). Passing numa_all_nodes permits the kernel to schedule on all nodes again. On success, 0 is returned; on error -1 is returned, and is set to indicate the error. numa_run_on_node_mask_all() runs the current task and its children only on nodes specified in nodemask like numa_run_on_node_mask but without any cpuset awareness. numa_get_run_node_mask() returns a mask of CPUs on which the current task is allowed to run. numa_tonode_memory() put memory on a specific node. The constraints described for numa_interleave_memory() apply here too. numa_tonodemask_memory() put memory on a specific set of nodes. The constraints described for numa_interleave_memory() apply here too. numa_setlocal_memory() locates memory on the current node. The constraints described for numa_interleave_memory() apply here too. numa_police_memory() locates memory with the current NUMA policy. The constraints described for numa_interleave_memory() apply here too. numa_distance() reports the distance in the machine topology between two nodes. The factors are a multiple of 10. It returns 0 when the distance cannot be determined. A node has distance 10 to itself. Reporting the distance requires a Linux kernel version of 2.6.10 or newer. numa_set_bind_policy() specifies whether calls that bind memory to a specific node should use the preferred policy or a strict policy. The preferred policy allows the kernel to allocate memory on other nodes when there isn't enough free on the target node. strict will fail the allocation in that case. Setting the argument to specifies strict, 0 preferred. Note that specifying more than one node non strict may only use the first node in some kernel versions. numa_set_strict() sets a flag that says whether the functions allocating on specific nodes should use use a strict policy. Strict means the allocation will fail if the memory cannot be allocated on the target node. Default operation is to fall back to other nodes. This doesn't apply to interleave and default. numa_get_interleave_node() is used by libnuma internally. It is probably not useful for user applications. It uses the MPOL_F_NODE flag of the get_mempolicy system call, which is not intended for application use (its operation may change or be removed altogether in future kernel versions). See get_mempolicy(2). numa_pagesize() returns the number of bytes in page. This function is simply a fast alternative to repeated calls to the getpagesize system call. See getpagesize(2). numa_sched_getaffinity() retrieves a bitmask of the cpus on which a task may run. The task is specified by pid. Returns the return value of the sched_getaffinity system call. See sched_getaffinity(2). The bitmask must be at least the size of the kernel's cpu mask structure. Use numa_allocate_cpumask() to allocate it. Test the bits in the mask by calling numa_bitmask_isbitset(). numa_sched_setaffinity() sets a task's allowed cpu's to those cpu's specified in mask. The task is specified by pid. Returns the return value of the sched_setaffinity system call. See sched_setaffinity(2). You may allocate the bitmask with numa_allocate_cpumask(). Or the bitmask may be smaller than the kernel's cpu mask structure. For example, call numa_bitmask_alloc() using a maximum number of cpus from numa_num_configured_cpus(). Set the bits in the mask by calling numa_bitmask_setbit(). numa_node_to_cpus() converts a node number to a bitmask of CPUs. The user must pass a bitmask structure with a mask buffer long enough to represent all possible cpu's. Use numa_allocate_cpumask() to create it. If the bitmask is not long enough will be set to ERANGE and -1 returned. On success 0 is returned. numa_node_to_cpu_update() Mark cpus bitmask of all nodes stale, then get the latest bitmask by calling numa_node_to_cpus() This allows to update the libnuma state after a CPU hotplug event. The application is in charge of detecting CPU hotplug events. numa_node_of_cpu() returns the node that a cpu belongs to. If the user supplies an invalid cpu will be set to EINVAL and -1 will be returned. numa_allocate_cpumask () returns a bitmask of a size equal to the kernel's cpu mask (kernel type cpumask_t). In other words, large enough to represent NR_CPUS cpus. This number of cpus can be gotten by calling numa_num_possible_cpus(). The bitmask is zero- filled. numa_free_cpumask frees a cpumask previously allocate by numa_allocate_cpumask. numa_allocate_nodemask() returns a bitmask of a size equal to the kernel's node mask (kernel type nodemask_t). In other words, large enough to represent MAX_NUMNODES nodes. This number of nodes can be gotten by calling numa_num_possible_nodes(). The bitmask is zero-filled. numa_free_nodemask() frees a nodemask previous allocated by numa_allocate_nodemask(). numa_bitmask_alloc() allocates a bitmask structure and its associated bit mask. The memory allocated for the bit mask contains enough words (type unsigned long) to contain n bits. The bit mask is zero-filled. The bitmask structure points to the bit mask and contains the n value. numa_bitmask_clearall() sets all bits in the bit mask to 0. The bitmask structure points to the bit mask and contains its size ( bmp ->size). The value of bmp is always returned. Note that numa_bitmask_alloc() creates a zero-filled bit mask. numa_bitmask_clearbit() sets a specified bit in a bit mask to 0. Nothing is done if the n value is greater than the size of the bitmask (and no error is returned). The value of bmp is always returned. numa_bitmask_equal() returns 1 if two bitmasks are equal. It returns 0 if they are not equal. If the bitmask structures control bit masks of different sizes, the "missing" trailing bits of the smaller bit mask are considered to be 0. numa_bitmask_free() deallocates the memory of both the bitmask structure pointed to by bmp and the bit mask. It is an error to attempt to free this bitmask twice. numa_bitmask_isbitset() returns the value of a specified bit in a bit mask. If the n value is greater than the size of the bit map, 0 is returned. numa_bitmask_nbytes() returns the size (in bytes) of the bit mask controlled by bmp. The bit masks are always full words (type unsigned long), and the returned size is the actual size of all those words. numa_bitmask_setall() sets all bits in the bit mask to 1. The bitmask structure points to the bit mask and contains its size ( bmp ->size). The value of bmp is always returned. numa_bitmask_setbit() sets a specified bit in a bit mask to 1. Nothing is done if n is greater than the size of the bitmask (and no error is returned). The value of bmp is always returned. copy_bitmask_to_nodemask() copies the body (the bit map itself) of the bitmask structure pointed to by bmp to the nodemask_t structure pointed to by the nodemask pointer. If the two areas differ in size, the copy is truncated to the size of the receiving field or zero-filled. copy_nodemask_to_bitmask() copies the nodemask_t structure pointed to by the nodemask pointer to the body (the bit map itself) of the bitmask structure pointed to by the bmp pointer. If the two areas differ in size, the copy is truncated to the size of the receiving field or zero-filled. copy_bitmask_to_bitmask() copies the body (the bit map itself) of the bitmask structure pointed to by the bmpfrom pointer to the body of the bitmask structure pointed to by the bmpto pointer. If the two areas differ in size, the copy is truncated to the size of the receiving field or zero-filled. numa_bitmask_weight() returns a count of the bits that are set in the body of the bitmask pointed to by the bmp argument. numa_move_pages() moves a list of pages in the address space of the currently executing or current process. It simply uses the move_pages system call. pid - ID of task. If not valid, use the current task. count - Number of pages. pages - List of pages to move. nodes - List of nodes to which pages can be moved. status - Field to which status is to be returned. flags - MPOL_MF_MOVE or MPOL_MF_MOVE_ALL See move_pages(2). numa_migrate_pages() simply uses the migrate_pages system call to cause the pages of the calling task, or a specified task, to be migated from one set of nodes to another. See migrate_pages(2). The bit masks representing the nodes should be allocated with numa_allocate_nodemask() , or with numa_bitmask_alloc() using an n value returned from numa_num_possible_nodes(). A task's current node set can be gotten by calling numa_get_membind(). Bits in the tonodes mask can be set by calls to numa_bitmask_setbit(). numa_error() is a libnuma internal function that can be overridden by the user program. This function is called with a char * argument when a libnuma function fails. Overriding the library internal definition makes it possible to specify a different error handling strategy when a libnuma function fails. It does not affect numa_available(). The numa_error() function defined in libnuma prints an error on stderr and terminates the program if numa_exit_on_error is set to a non-zero value. The default value of numa_exit_on_error is zero. numa_warn() is a libnuma internal function that can be also overridden by the user program. It is called to warn the user when a libnuma function encounters a non-fatal error. The default implementation prints a warning to stderr. The first argument is a unique number identifying each warning. After that there is a -style format string and a variable number of arguments. numa_warn exits the program when numa_exit_on_warn is set to a non-zero value. The default value of numa_exit_on_warn is zero.
Compatibility with libnuma version 1
Binaries that were compiled for libnuma version 1 need not be re- compiled to run with libnuma version 2. Source codes written for libnuma version 1 may be re-compiled without change with version 2 installed. To do so, in the code's Makefile add this option to CFLAGS: -DNUMA_VERSION1_COMPATIBILITY
THREAD SAFETY
numa_set_bind_policy and numa_exit_on_error are process global. The other calls are thread safe.
COPYRIGHT
Copyright 2002, 2004, 2007, 2008 Andi Kleen, SuSE Labs. libnuma is under the GNU Lesser General Public License, v2.1.
SEE ALSO
, , , , , , ,
COLOPHON
This page is part of the numactl (NUMA commands) project. Information about the project can be found at 〈〉. If you have a bug report for this manual page, send it to linux-numa@vger.kernel.org. This page was obtained from the project's upstream Git repository 〈〉 on 2021-04-01. (At that time, the date of the most recent commit that was found in the repository was 2021-03-31.) If you discover any rendering problems in this HTML version of the page, or you believe there is a better or more up-to-date source for the page, or you have corrections or improvements to the information in this COLOPHON (which is not part of the original manual page), send a mail to man-pages@man7.orgSuSE Labs December 2007 NUMA(3)
Pages that refer to this page: , , , , , ,
转载地址：http://qjvaf.baihongyu.com/