[mirror_ubuntu-zesty-kernel.git] / include / linux / mempolicy.h

#ifndef _LINUX_MEMPOLICY_H
#define _LINUX_MEMPOLICY_H 1

#include <linux/errno.h>

/*
 * NUMA memory policies for Linux.
 * Copyright 2003,2004 Andi Kleen SuSE Labs
 */

/*
 * Both the MPOL_* mempolicy mode and the MPOL_F_* optional mode flags are
 * passed by the user to either set_mempolicy() or mbind() in an 'int' actual.
 * The MPOL_MODE_FLAGS macro determines the legal set of optional mode flags.
 */

/* Policies */
enum {
	MPOL_DEFAULT,
	MPOL_PREFERRED,
	MPOL_BIND,
	MPOL_INTERLEAVE,
	MPOL_MAX,	/* always last member of enum */
};

/* Flags for set_mempolicy */
#define MPOL_F_STATIC_NODES	(1 << 15)

/*
 * MPOL_MODE_FLAGS is the union of all possible optional mode flags passed to
 * either set_mempolicy() or mbind().
 */
#define MPOL_MODE_FLAGS	(MPOL_F_STATIC_NODES)

/* Flags for get_mempolicy */
#define MPOL_F_NODE	(1<<0)	/* return next IL mode instead of node mask */
#define MPOL_F_ADDR	(1<<1)	/* look up vma using address */
#define MPOL_F_MEMS_ALLOWED (1<<2) /* return allowed memories */

/* Flags for mbind */
#define MPOL_MF_STRICT	(1<<0)	/* Verify existing pages in the mapping */
#define MPOL_MF_MOVE	(1<<1)	/* Move pages owned by this process to conform to mapping */
#define MPOL_MF_MOVE_ALL (1<<2)	/* Move every page to conform to mapping */
#define MPOL_MF_INTERNAL (1<<3)	/* Internal flags start here */

#ifdef __KERNEL__

#include <linux/mmzone.h>
#include <linux/slab.h>
#include <linux/rbtree.h>
#include <linux/spinlock.h>
#include <linux/nodemask.h>

struct vm_area_struct;
struct mm_struct;

#ifdef CONFIG_NUMA

/*
 * Describe a memory policy.
 *
 * A mempolicy can be either associated with a process or with a VMA.
 * For VMA related allocations the VMA policy is preferred, otherwise
 * the process policy is used. Interrupts ignore the memory policy
 * of the current process.
 *
 * Locking policy for interlave:
 * In process context there is no locking because only the process accesses
 * its own state. All vma manipulation is somewhat protected by a down_read on
 * mmap_sem.
 *
 * Freeing policy:
 * Mempolicy objects are reference counted.  A mempolicy will be freed when
 * mpol_free() decrements the reference count to zero.
 *
 * Copying policy objects:
 * mpol_copy() allocates a new mempolicy and copies the specified mempolicy
 * to the new storage.  The reference count of the new object is initialized
 * to 1, representing the caller of mpol_copy().
 */
struct mempolicy {
	atomic_t refcnt;
	unsigned short policy; 	/* See MPOL_* above */
	unsigned short flags;	/* See set_mempolicy() MPOL_F_* above */
	union {
		short 		 preferred_node; /* preferred */
		nodemask_t	 nodes;		/* interleave/bind */
		/* undefined for default */
	} v;
	union {
		nodemask_t cpuset_mems_allowed;	/* relative to these nodes */
		nodemask_t user_nodemask;	/* nodemask passed by user */
	} w;
};

/*
 * Support for managing mempolicy data objects (clone, copy, destroy)
 * The default fast path of a NULL MPOL_DEFAULT policy is always inlined.
 */

extern void __mpol_free(struct mempolicy *pol);
static inline void mpol_free(struct mempolicy *pol)
{
	if (pol)
		__mpol_free(pol);
}

extern struct mempolicy *__mpol_copy(struct mempolicy *pol);
static inline struct mempolicy *mpol_copy(struct mempolicy *pol)
{
	if (pol)
		pol = __mpol_copy(pol);
	return pol;
}

#define vma_policy(vma) ((vma)->vm_policy)
#define vma_set_policy(vma, pol) ((vma)->vm_policy = (pol))

static inline void mpol_get(struct mempolicy *pol)
{
	if (pol)
		atomic_inc(&pol->refcnt);
}

extern int __mpol_equal(struct mempolicy *a, struct mempolicy *b);
static inline int mpol_equal(struct mempolicy *a, struct mempolicy *b)
{
	if (a == b)
		return 1;
	return __mpol_equal(a, b);
}

/* Could later add inheritance of the process policy here. */

#define mpol_set_vma_default(vma) ((vma)->vm_policy = NULL)

/*
 * Tree of shared policies for a shared memory region.
 * Maintain the policies in a pseudo mm that contains vmas. The vmas
 * carry the policy. As a special twist the pseudo mm is indexed in pages, not
 * bytes, so that we can work with shared memory segments bigger than
 * unsigned long.
 */

struct sp_node {
	struct rb_node nd;
	unsigned long start, end;
	struct mempolicy *policy;
};

struct shared_policy {
	struct rb_root root;
	spinlock_t lock;
};

void mpol_shared_policy_init(struct shared_policy *info, unsigned short policy,
				unsigned short flags, nodemask_t *nodes);
int mpol_set_shared_policy(struct shared_policy *info,
				struct vm_area_struct *vma,
				struct mempolicy *new);
void mpol_free_shared_policy(struct shared_policy *p);
struct mempolicy *mpol_shared_policy_lookup(struct shared_policy *sp,
					    unsigned long idx);

extern void numa_default_policy(void);
extern void numa_policy_init(void);
extern void mpol_rebind_task(struct task_struct *tsk,
					const nodemask_t *new);
extern void mpol_rebind_mm(struct mm_struct *mm, nodemask_t *new);
extern void mpol_fix_fork_child_flag(struct task_struct *p);

extern struct mempolicy default_policy;
extern struct zonelist *huge_zonelist(struct vm_area_struct *vma,
				unsigned long addr, gfp_t gfp_flags,
				struct mempolicy **mpol, nodemask_t **nodemask);
extern unsigned slab_node(struct mempolicy *policy);

extern enum zone_type policy_zone;

static inline void check_highest_zone(enum zone_type k)
{
	if (k > policy_zone && k != ZONE_MOVABLE)
		policy_zone = k;
}

int do_migrate_pages(struct mm_struct *mm,
	const nodemask_t *from_nodes, const nodemask_t *to_nodes, int flags);

#else

struct mempolicy {};

static inline int mpol_equal(struct mempolicy *a, struct mempolicy *b)
{
	return 1;
}

#define mpol_set_vma_default(vma) do {} while(0)

static inline void mpol_free(struct mempolicy *p)
{
}

static inline void mpol_get(struct mempolicy *pol)
{
}

static inline struct mempolicy *mpol_copy(struct mempolicy *old)
{
	return NULL;
}

struct shared_policy {};

static inline int mpol_set_shared_policy(struct shared_policy *info,
					struct vm_area_struct *vma,
					struct mempolicy *new)
{
	return -EINVAL;
}

static inline void mpol_shared_policy_init(struct shared_policy *info,
		unsigned short policy, unsigned short flags, nodemask_t *nodes)
{
}

static inline void mpol_free_shared_policy(struct shared_policy *p)
{
}

static inline struct mempolicy *
mpol_shared_policy_lookup(struct shared_policy *sp, unsigned long idx)
{
	return NULL;
}

#define vma_policy(vma) NULL
#define vma_set_policy(vma, pol) do {} while(0)

static inline void numa_policy_init(void)
{
}

static inline void numa_default_policy(void)
{
}

static inline void mpol_rebind_task(struct task_struct *tsk,
					const nodemask_t *new)
{
}

static inline void mpol_rebind_mm(struct mm_struct *mm, nodemask_t *new)
{
}

static inline void mpol_fix_fork_child_flag(struct task_struct *p)
{
}

static inline struct zonelist *huge_zonelist(struct vm_area_struct *vma,
				unsigned long addr, gfp_t gfp_flags,
				struct mempolicy **mpol, nodemask_t **nodemask)
{
	*mpol = NULL;
	*nodemask = NULL;
	return node_zonelist(0, gfp_flags);
}

static inline int do_migrate_pages(struct mm_struct *mm,
			const nodemask_t *from_nodes,
			const nodemask_t *to_nodes, int flags)
{
	return 0;
}

static inline void check_highest_zone(int k)
{
}
#endif /* CONFIG_NUMA */
#endif /* __KERNEL__ */

#endif
Commit	Line	Data
1da177e4 LT	1	#ifndef _LINUX_MEMPOLICY_H
	2	#define _LINUX_MEMPOLICY_H 1
	3
	4	#include <linux/errno.h>
	5
	6	/*
	7	* NUMA memory policies for Linux.
	8	* Copyright 2003,2004 Andi Kleen SuSE Labs
	9	*/
	10
028fec41 DR	11	/*
	12	* Both the MPOL_* mempolicy mode and the MPOL_F_* optional mode flags are
	13	* passed by the user to either set_mempolicy() or mbind() in an 'int' actual.
	14	* The MPOL_MODE_FLAGS macro determines the legal set of optional mode flags.
	15	*/
	16
1da177e4	17	/* Policies */
a3b51e01 DR	18	enum {
	19	MPOL_DEFAULT,
	20	MPOL_PREFERRED,
	21	MPOL_BIND,
	22	MPOL_INTERLEAVE,
	23	MPOL_MAX, /* always last member of enum */
	24	};
1da177e4	25
028fec41	26	/* Flags for set_mempolicy */
f5b087b5 DR	27	#define MPOL_F_STATIC_NODES (1 << 15)
f5b087b5 DR	28
028fec41 DR	29	/*
	30	* MPOL_MODE_FLAGS is the union of all possible optional mode flags passed to
	31	* either set_mempolicy() or mbind().
	32	*/
f5b087b5	33	#define MPOL_MODE_FLAGS (MPOL_F_STATIC_NODES)
028fec41 DR	34
028fec41 DR	35	/* Flags for get_mempolicy */
1da177e4 LT	36	#define MPOL_F_NODE (1<<0) /* return next IL mode instead of node mask */
1da177e4 LT	37	#define MPOL_F_ADDR (1<<1) /* look up vma using address */
754af6f5	38	#define MPOL_F_MEMS_ALLOWED (1<<2) /* return allowed memories */
1da177e4 LT	39
	40	/* Flags for mbind */
	41	#define MPOL_MF_STRICT (1<<0) /* Verify existing pages in the mapping */
dc9aa5b9 CL	42	#define MPOL_MF_MOVE (1<<1) /* Move pages owned by this process to conform to mapping */
	43	#define MPOL_MF_MOVE_ALL (1<<2) /* Move every page to conform to mapping */
	44	#define MPOL_MF_INTERNAL (1<<3) /* Internal flags start here */
1da177e4 LT	45
	46	#ifdef __KERNEL__
	47
1da177e4	48	#include <linux/mmzone.h>
1da177e4 LT	49	#include <linux/slab.h>
	50	#include <linux/rbtree.h>
	51	#include <linux/spinlock.h>
dfcd3c0d	52	#include <linux/nodemask.h>
1da177e4 LT	53
1da177e4 LT	54	struct vm_area_struct;
45b35a5c	55	struct mm_struct;
1da177e4 LT	56
	57	#ifdef CONFIG_NUMA
	58
	59	/*
	60	* Describe a memory policy.
	61	*
	62	* A mempolicy can be either associated with a process or with a VMA.
	63	* For VMA related allocations the VMA policy is preferred, otherwise
	64	* the process policy is used. Interrupts ignore the memory policy
	65	* of the current process.
	66	*
	67	* Locking policy for interlave:
	68	* In process context there is no locking because only the process accesses
	69	* its own state. All vma manipulation is somewhat protected by a down_read on
b8072f09	70	* mmap_sem.
1da177e4 LT	71	*
1da177e4 LT	72	* Freeing policy:
19770b32 MG	73	* Mempolicy objects are reference counted. A mempolicy will be freed when
19770b32 MG	74	* mpol_free() decrements the reference count to zero.
1da177e4 LT	75	*
1da177e4 LT	76	* Copying policy objects:
19770b32 MG	77	* mpol_copy() allocates a new mempolicy and copies the specified mempolicy
	78	* to the new storage. The reference count of the new object is initialized
	79	* to 1, representing the caller of mpol_copy().
1da177e4 LT	80	*/
	81	struct mempolicy {
	82	atomic_t refcnt;
a3b51e01	83	unsigned short policy; /* See MPOL_* above */
028fec41	84	unsigned short flags; /* See set_mempolicy() MPOL_F_* above */
1da177e4	85	union {
1da177e4	86	short preferred_node; /* preferred */
19770b32	87	nodemask_t nodes; /* interleave/bind */
1da177e4 LT	88	/* undefined for default */
1da177e4 LT	89	} v;
f5b087b5 DR	90	union {
	91	nodemask_t cpuset_mems_allowed; /* relative to these nodes */
	92	nodemask_t user_nodemask; /* nodemask passed by user */
	93	} w;
1da177e4 LT	94	};
	95
	96	/*
	97	* Support for managing mempolicy data objects (clone, copy, destroy)
	98	* The default fast path of a NULL MPOL_DEFAULT policy is always inlined.
	99	*/
	100
	101	extern void __mpol_free(struct mempolicy *pol);
	102	static inline void mpol_free(struct mempolicy *pol)
	103	{
	104	if (pol)
	105	__mpol_free(pol);
	106	}
	107
	108	extern struct mempolicy __mpol_copy(struct mempolicy pol);
	109	static inline struct mempolicy mpol_copy(struct mempolicy pol)
	110	{
	111	if (pol)
	112	pol = __mpol_copy(pol);
	113	return pol;
	114	}
	115
	116	#define vma_policy(vma) ((vma)->vm_policy)
	117	#define vma_set_policy(vma, pol) ((vma)->vm_policy = (pol))
	118
	119	static inline void mpol_get(struct mempolicy *pol)
	120	{
	121	if (pol)
	122	atomic_inc(&pol->refcnt);
	123	}
	124
	125	extern int __mpol_equal(struct mempolicy a, struct mempolicy b);
	126	static inline int mpol_equal(struct mempolicy a, struct mempolicy b)
	127	{
	128	if (a == b)
	129	return 1;
	130	return __mpol_equal(a, b);
	131	}
1da177e4 LT	132
	133	/* Could later add inheritance of the process policy here. */
	134
	135	#define mpol_set_vma_default(vma) ((vma)->vm_policy = NULL)
	136
1da177e4 LT	137	/*
	138	* Tree of shared policies for a shared memory region.
	139	* Maintain the policies in a pseudo mm that contains vmas. The vmas
	140	* carry the policy. As a special twist the pseudo mm is indexed in pages, not
	141	* bytes, so that we can work with shared memory segments bigger than
	142	* unsigned long.
	143	*/
	144
	145	struct sp_node {
	146	struct rb_node nd;
	147	unsigned long start, end;
	148	struct mempolicy *policy;
	149	};
	150
	151	struct shared_policy {
	152	struct rb_root root;
	153	spinlock_t lock;
	154	};
	155
a3b51e01	156	void mpol_shared_policy_init(struct shared_policy *info, unsigned short policy,
028fec41	157	unsigned short flags, nodemask_t *nodes);
1da177e4 LT	158	int mpol_set_shared_policy(struct shared_policy *info,
	159	struct vm_area_struct *vma,
	160	struct mempolicy *new);
	161	void mpol_free_shared_policy(struct shared_policy *p);
	162	struct mempolicy mpol_shared_policy_lookup(struct shared_policy sp,
	163	unsigned long idx);
	164
	165	extern void numa_default_policy(void);
	166	extern void numa_policy_init(void);
74cb2155 PJ	167	extern void mpol_rebind_task(struct task_struct *tsk,
74cb2155 PJ	168	const nodemask_t *new);
4225399a	169	extern void mpol_rebind_mm(struct mm_struct mm, nodemask_t new);
c61afb18	170	extern void mpol_fix_fork_child_flag(struct task_struct *p);
4225399a	171
d42c6997	172	extern struct mempolicy default_policy;
5da7ca86	173	extern struct zonelist huge_zonelist(struct vm_area_struct vma,
19770b32 MG	174	unsigned long addr, gfp_t gfp_flags,
19770b32 MG	175	struct mempolicy mpol, nodemask_t nodemask);
dc85da15	176	extern unsigned slab_node(struct mempolicy *policy);
1da177e4	177
2f6726e5	178	extern enum zone_type policy_zone;
4be38e35	179
2f6726e5	180	static inline void check_highest_zone(enum zone_type k)
4be38e35	181	{
b377fd39	182	if (k > policy_zone && k != ZONE_MOVABLE)
4be38e35 CL	183	policy_zone = k;
	184	}
	185
39743889 CL	186	int do_migrate_pages(struct mm_struct *mm,
	187	const nodemask_t from_nodes, const nodemask_t to_nodes, int flags);
	188
1da177e4 LT	189	#else
	190
	191	struct mempolicy {};
	192
	193	static inline int mpol_equal(struct mempolicy a, struct mempolicy b)
	194	{
	195	return 1;
	196	}
1da177e4 LT	197
	198	#define mpol_set_vma_default(vma) do {} while(0)
	199
	200	static inline void mpol_free(struct mempolicy *p)
	201	{
	202	}
	203
	204	static inline void mpol_get(struct mempolicy *pol)
	205	{
	206	}
	207
	208	static inline struct mempolicy mpol_copy(struct mempolicy old)
	209	{
	210	return NULL;
	211	}
	212
1da177e4 LT	213	struct shared_policy {};
	214
	215	static inline int mpol_set_shared_policy(struct shared_policy *info,
	216	struct vm_area_struct *vma,
	217	struct mempolicy *new)
	218	{
	219	return -EINVAL;
	220	}
	221
7339ff83	222	static inline void mpol_shared_policy_init(struct shared_policy *info,
028fec41	223	unsigned short policy, unsigned short flags, nodemask_t *nodes)
1da177e4 LT	224	{
	225	}
	226
	227	static inline void mpol_free_shared_policy(struct shared_policy *p)
	228	{
	229	}
	230
	231	static inline struct mempolicy *
	232	mpol_shared_policy_lookup(struct shared_policy *sp, unsigned long idx)
	233	{
	234	return NULL;
	235	}
	236
	237	#define vma_policy(vma) NULL
	238	#define vma_set_policy(vma, pol) do {} while(0)
	239
	240	static inline void numa_policy_init(void)
	241	{
	242	}
	243
	244	static inline void numa_default_policy(void)
	245	{
	246	}
	247
74cb2155	248	static inline void mpol_rebind_task(struct task_struct *tsk,
68860ec1 PJ	249	const nodemask_t *new)
	250	{
	251	}
	252
4225399a PJ	253	static inline void mpol_rebind_mm(struct mm_struct mm, nodemask_t new)
	254	{
	255	}
	256
c61afb18 PJ	257	static inline void mpol_fix_fork_child_flag(struct task_struct *p)
	258	{
	259	}
	260
5da7ca86	261	static inline struct zonelist huge_zonelist(struct vm_area_struct vma,
19770b32 MG	262	unsigned long addr, gfp_t gfp_flags,
19770b32 MG	263	struct mempolicy mpol, nodemask_t nodemask)
5da7ca86	264	{
19770b32 MG	265	*mpol = NULL;
19770b32 MG	266	*nodemask = NULL;
0e88460d	267	return node_zonelist(0, gfp_flags);
5da7ca86 CL	268	}
5da7ca86 CL	269
45b07ef3 PJ	270	static inline int do_migrate_pages(struct mm_struct *mm,
	271	const nodemask_t *from_nodes,
	272	const nodemask_t *to_nodes, int flags)
	273	{
	274	return 0;
	275	}
	276
4be38e35 CL	277	static inline void check_highest_zone(int k)
	278	{
	279	}
1da177e4 LT	280	#endif /* CONFIG_NUMA */
	281	#endif /* __KERNEL__ */
	282
	283	#endif