[deliverable/linux.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
 */

/* Policies */
#define MPOL_DEFAULT	0
#define MPOL_PREFERRED	1
#define MPOL_BIND	2
#define MPOL_INTERLEAVE	3

#define MPOL_MAX MPOL_INTERLEAVE

/* Flags for get_mem_policy */
#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 */

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

struct vm_area_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:
 * When policy is MPOL_BIND v.zonelist is kmalloc'ed and must be kfree'd.
 * All other policies don't have any external state. mpol_free() handles this.
 *
 * Copying policy objects:
 * For MPOL_BIND the zonelist must be always duplicated. mpol_clone() does this.
 */
struct mempolicy {
	atomic_t refcnt;
	short policy; 	/* See MPOL_* above */
	union {
		struct zonelist  *zonelist;	/* bind */
		short 		 preferred_node; /* preferred */
		nodemask_t	 nodes;		/* interleave */
		/* undefined for default */
	} v;
	nodemask_t cpuset_mems_allowed;	/* mempolicy relative to these nodes */
};

/*
 * 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);
}
#define vma_mpol_equal(a,b) mpol_equal(vma_policy(a), vma_policy(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, int policy,
				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_policy(struct mempolicy *pol, const nodemask_t *new);
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);
#define set_cpuset_being_rebound(x) (cpuset_being_rebound = (x))

#ifdef CONFIG_CPUSET
#define current_cpuset_is_being_rebound() \
				(cpuset_being_rebound == current->cpuset)
#else
#define current_cpuset_is_being_rebound() 0
#endif

extern struct mempolicy default_policy;
extern struct zonelist *huge_zonelist(struct vm_area_struct *vma,
		unsigned long addr);
extern unsigned slab_node(struct mempolicy *policy);

extern int policy_zone;

static inline void check_highest_zone(int k)
{
	if (k > policy_zone)
		policy_zone = k;
}

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

extern void *cpuset_being_rebound;	/* Trigger mpol_copy vma rebind */

#else

struct mempolicy {};

static inline int mpol_equal(struct mempolicy *a, struct mempolicy *b)
{
	return 1;
}
#define vma_mpol_equal(a,b) 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,
					int policy, 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_policy(struct mempolicy *pol,
					const nodemask_t *new)
{
}

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)
{
}

#define set_cpuset_being_rebound(x) do {} while (0)

static inline struct zonelist *huge_zonelist(struct vm_area_struct *vma,
		unsigned long addr)
{
	return NODE_DATA(0)->node_zonelists + gfp_zone(GFP_HIGHUSER);
}

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