[net-next-2.6.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)
#define MPOL_F_RELATIVE_NODES	(1 << 14)

/*
 * 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 | MPOL_F_RELATIVE_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 */

/*
 * Internal flags that share the struct mempolicy flags word with
 * "mode flags".  These flags are allocated from bit 0 up, as they
 * are never OR'ed into the mode in mempolicy API arguments.
 */
#define MPOL_F_SHARED  (1 << 0)	/* identify shared policies */
#define MPOL_F_LOCAL   (1 << 1)	/* preferred local allocation */

#ifdef __KERNEL__

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

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_put() decrements the reference count to zero.
 *
 * Duplicating policy objects:
 * mpol_dup() 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_dup().
 */
struct mempolicy {
	atomic_t refcnt;
	unsigned short mode; 	/* 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_put(struct mempolicy *pol);
static inline void mpol_put(struct mempolicy *pol)
{
	if (pol)
		__mpol_put(pol);
}

/*
 * Does mempolicy pol need explicit unref after use?
 * Currently only needed for shared policies.
 */
static inline int mpol_needs_cond_ref(struct mempolicy *pol)
{
	return (pol && (pol->flags & MPOL_F_SHARED));
}

static inline void mpol_cond_put(struct mempolicy *pol)
{
	if (mpol_needs_cond_ref(pol))
		__mpol_put(pol);
}

extern struct mempolicy *__mpol_cond_copy(struct mempolicy *tompol,
					  struct mempolicy *frompol);
static inline struct mempolicy *mpol_cond_copy(struct mempolicy *tompol,
						struct mempolicy *frompol)
{
	if (!frompol)
		return frompol;
	return __mpol_cond_copy(tompol, frompol);
}

extern struct mempolicy *__mpol_dup(struct mempolicy *pol);
static inline struct mempolicy *mpol_dup(struct mempolicy *pol)
{
	if (pol)
		pol = __mpol_dup(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);
}

/*
 * 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 mode,
				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 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;
}

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

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

static inline struct mempolicy *mpol_cond_copy(struct mempolicy *to,
						struct mempolicy *from)
{
	return from;
}

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

static inline struct mempolicy *mpol_dup(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 mode, 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	27	#define MPOL_F_STATIC_NODES (1 << 15)
4c50bc01	28	#define MPOL_F_RELATIVE_NODES (1 << 14)
f5b087b5	29
028fec41 DR	30	/*
	31	* MPOL_MODE_FLAGS is the union of all possible optional mode flags passed to
	32	* either set_mempolicy() or mbind().
	33	*/
4c50bc01	34	#define MPOL_MODE_FLAGS (MPOL_F_STATIC_NODES \| MPOL_F_RELATIVE_NODES)
028fec41 DR	35
028fec41 DR	36	/* Flags for get_mempolicy */
1da177e4 LT	37	#define MPOL_F_NODE (1<<0) /* return next IL mode instead of node mask */
1da177e4 LT	38	#define MPOL_F_ADDR (1<<1) /* look up vma using address */
754af6f5	39	#define MPOL_F_MEMS_ALLOWED (1<<2) /* return allowed memories */
1da177e4 LT	40
	41	/* Flags for mbind */
	42	#define MPOL_MF_STRICT (1<<0) /* Verify existing pages in the mapping */
dc9aa5b9 CL	43	#define MPOL_MF_MOVE (1<<1) /* Move pages owned by this process to conform to mapping */
	44	#define MPOL_MF_MOVE_ALL (1<<2) /* Move every page to conform to mapping */
	45	#define MPOL_MF_INTERNAL (1<<3) /* Internal flags start here */
1da177e4	46
aab0b102 LS	47	/*
	48	* Internal flags that share the struct mempolicy flags word with
	49	* "mode flags". These flags are allocated from bit 0 up, as they
	50	* are never OR'ed into the mode in mempolicy API arguments.
	51	*/
	52	#define MPOL_F_SHARED (1 << 0) /* identify shared policies */
fc36b8d3	53	#define MPOL_F_LOCAL (1 << 1) /* preferred local allocation */
aab0b102	54
1da177e4 LT	55	#ifdef __KERNEL__
1da177e4 LT	56
1da177e4	57	#include <linux/mmzone.h>
1da177e4 LT	58	#include <linux/slab.h>
	59	#include <linux/rbtree.h>
	60	#include <linux/spinlock.h>
dfcd3c0d	61	#include <linux/nodemask.h>
1da177e4	62
45b35a5c	63	struct mm_struct;
1da177e4 LT	64
	65	#ifdef CONFIG_NUMA
	66
	67	/*
	68	* Describe a memory policy.
	69	*
	70	* A mempolicy can be either associated with a process or with a VMA.
	71	* For VMA related allocations the VMA policy is preferred, otherwise
	72	* the process policy is used. Interrupts ignore the memory policy
	73	* of the current process.
	74	*
	75	* Locking policy for interlave:
	76	* In process context there is no locking because only the process accesses
	77	* its own state. All vma manipulation is somewhat protected by a down_read on
b8072f09	78	* mmap_sem.
1da177e4 LT	79	*
1da177e4 LT	80	* Freeing policy:
19770b32	81	* Mempolicy objects are reference counted. A mempolicy will be freed when
f0be3d32	82	* mpol_put() decrements the reference count to zero.
1da177e4	83	*
846a16bf LS	84	* Duplicating policy objects:
846a16bf LS	85	* mpol_dup() allocates a new mempolicy and copies the specified mempolicy
19770b32	86	* to the new storage. The reference count of the new object is initialized
846a16bf	87	* to 1, representing the caller of mpol_dup().
1da177e4 LT	88	*/
	89	struct mempolicy {
	90	atomic_t refcnt;
45c4745a	91	unsigned short mode; /* See MPOL_* above */
028fec41	92	unsigned short flags; /* See set_mempolicy() MPOL_F_* above */
1da177e4	93	union {
1da177e4	94	short preferred_node; /* preferred */
19770b32	95	nodemask_t nodes; /* interleave/bind */
1da177e4 LT	96	/* undefined for default */
1da177e4 LT	97	} v;
f5b087b5 DR	98	union {
	99	nodemask_t cpuset_mems_allowed; /* relative to these nodes */
	100	nodemask_t user_nodemask; /* nodemask passed by user */
	101	} w;
1da177e4 LT	102	};
	103
	104	/*
	105	* Support for managing mempolicy data objects (clone, copy, destroy)
	106	* The default fast path of a NULL MPOL_DEFAULT policy is always inlined.
	107	*/
	108
f0be3d32 LS	109	extern void __mpol_put(struct mempolicy *pol);
f0be3d32 LS	110	static inline void mpol_put(struct mempolicy *pol)
1da177e4 LT	111	{
1da177e4 LT	112	if (pol)
f0be3d32	113	__mpol_put(pol);
1da177e4 LT	114	}
1da177e4 LT	115
52cd3b07 LS	116	/*
	117	* Does mempolicy pol need explicit unref after use?
	118	* Currently only needed for shared policies.
	119	*/
	120	static inline int mpol_needs_cond_ref(struct mempolicy *pol)
	121	{
	122	return (pol && (pol->flags & MPOL_F_SHARED));
	123	}
	124
	125	static inline void mpol_cond_put(struct mempolicy *pol)
	126	{
	127	if (mpol_needs_cond_ref(pol))
	128	__mpol_put(pol);
	129	}
	130
	131	extern struct mempolicy __mpol_cond_copy(struct mempolicy tompol,
	132	struct mempolicy *frompol);
	133	static inline struct mempolicy mpol_cond_copy(struct mempolicy tompol,
	134	struct mempolicy *frompol)
	135	{
	136	if (!frompol)
	137	return frompol;
	138	return __mpol_cond_copy(tompol, frompol);
	139	}
	140
846a16bf LS	141	extern struct mempolicy __mpol_dup(struct mempolicy pol);
846a16bf LS	142	static inline struct mempolicy mpol_dup(struct mempolicy pol)
1da177e4 LT	143	{
1da177e4 LT	144	if (pol)
846a16bf	145	pol = __mpol_dup(pol);
1da177e4 LT	146	return pol;
	147	}
	148
	149	#define vma_policy(vma) ((vma)->vm_policy)
	150	#define vma_set_policy(vma, pol) ((vma)->vm_policy = (pol))
	151
	152	static inline void mpol_get(struct mempolicy *pol)
	153	{
	154	if (pol)
	155	atomic_inc(&pol->refcnt);
	156	}
	157
	158	extern int __mpol_equal(struct mempolicy a, struct mempolicy b);
	159	static inline int mpol_equal(struct mempolicy a, struct mempolicy b)
	160	{
	161	if (a == b)
	162	return 1;
	163	return __mpol_equal(a, b);
	164	}
1da177e4	165
1da177e4 LT	166	/*
	167	* Tree of shared policies for a shared memory region.
	168	* Maintain the policies in a pseudo mm that contains vmas. The vmas
	169	* carry the policy. As a special twist the pseudo mm is indexed in pages, not
	170	* bytes, so that we can work with shared memory segments bigger than
	171	* unsigned long.
	172	*/
	173
	174	struct sp_node {
	175	struct rb_node nd;
	176	unsigned long start, end;
	177	struct mempolicy *policy;
	178	};
	179
	180	struct shared_policy {
	181	struct rb_root root;
	182	spinlock_t lock;
	183	};
	184
45c4745a	185	void mpol_shared_policy_init(struct shared_policy *info, unsigned short mode,
028fec41	186	unsigned short flags, nodemask_t *nodes);
1da177e4 LT	187	int mpol_set_shared_policy(struct shared_policy *info,
	188	struct vm_area_struct *vma,
	189	struct mempolicy *new);
	190	void mpol_free_shared_policy(struct shared_policy *p);
	191	struct mempolicy mpol_shared_policy_lookup(struct shared_policy sp,
	192	unsigned long idx);
	193
	194	extern void numa_default_policy(void);
	195	extern void numa_policy_init(void);
74cb2155 PJ	196	extern void mpol_rebind_task(struct task_struct *tsk,
74cb2155 PJ	197	const nodemask_t *new);
4225399a	198	extern void mpol_rebind_mm(struct mm_struct mm, nodemask_t new);
c61afb18	199	extern void mpol_fix_fork_child_flag(struct task_struct *p);
4225399a	200
5da7ca86	201	extern struct zonelist huge_zonelist(struct vm_area_struct vma,
19770b32 MG	202	unsigned long addr, gfp_t gfp_flags,
19770b32 MG	203	struct mempolicy mpol, nodemask_t nodemask);
dc85da15	204	extern unsigned slab_node(struct mempolicy *policy);
1da177e4	205
2f6726e5	206	extern enum zone_type policy_zone;
4be38e35	207
2f6726e5	208	static inline void check_highest_zone(enum zone_type k)
4be38e35	209	{
b377fd39	210	if (k > policy_zone && k != ZONE_MOVABLE)
4be38e35 CL	211	policy_zone = k;
	212	}
	213
39743889 CL	214	int do_migrate_pages(struct mm_struct *mm,
	215	const nodemask_t from_nodes, const nodemask_t to_nodes, int flags);
	216
1da177e4 LT	217	#else
	218
	219	struct mempolicy {};
	220
	221	static inline int mpol_equal(struct mempolicy a, struct mempolicy b)
	222	{
	223	return 1;
	224	}
1da177e4	225
f0be3d32	226	static inline void mpol_put(struct mempolicy *p)
1da177e4 LT	227	{
	228	}
	229
52cd3b07 LS	230	static inline void mpol_cond_put(struct mempolicy *pol)
	231	{
	232	}
	233
	234	static inline struct mempolicy mpol_cond_copy(struct mempolicy to,
	235	struct mempolicy *from)
	236	{
	237	return from;
	238	}
	239
1da177e4 LT	240	static inline void mpol_get(struct mempolicy *pol)
	241	{
	242	}
	243
846a16bf	244	static inline struct mempolicy mpol_dup(struct mempolicy old)
1da177e4 LT	245	{
	246	return NULL;
	247	}
	248
1da177e4 LT	249	struct shared_policy {};
	250
	251	static inline int mpol_set_shared_policy(struct shared_policy *info,
	252	struct vm_area_struct *vma,
	253	struct mempolicy *new)
	254	{
	255	return -EINVAL;
	256	}
	257
7339ff83	258	static inline void mpol_shared_policy_init(struct shared_policy *info,
45c4745a	259	unsigned short mode, unsigned short flags, nodemask_t *nodes)
1da177e4 LT	260	{
	261	}
	262
	263	static inline void mpol_free_shared_policy(struct shared_policy *p)
	264	{
	265	}
	266
	267	static inline struct mempolicy *
	268	mpol_shared_policy_lookup(struct shared_policy *sp, unsigned long idx)
	269	{
	270	return NULL;
	271	}
	272
	273	#define vma_policy(vma) NULL
	274	#define vma_set_policy(vma, pol) do {} while(0)
	275
	276	static inline void numa_policy_init(void)
	277	{
	278	}
	279
	280	static inline void numa_default_policy(void)
	281	{
	282	}
	283
74cb2155	284	static inline void mpol_rebind_task(struct task_struct *tsk,
68860ec1 PJ	285	const nodemask_t *new)
	286	{
	287	}
	288
4225399a PJ	289	static inline void mpol_rebind_mm(struct mm_struct mm, nodemask_t new)
	290	{
	291	}
	292
c61afb18 PJ	293	static inline void mpol_fix_fork_child_flag(struct task_struct *p)
	294	{
	295	}
	296
5da7ca86	297	static inline struct zonelist huge_zonelist(struct vm_area_struct vma,
19770b32 MG	298	unsigned long addr, gfp_t gfp_flags,
19770b32 MG	299	struct mempolicy mpol, nodemask_t nodemask)
5da7ca86	300	{
19770b32 MG	301	*mpol = NULL;
19770b32 MG	302	*nodemask = NULL;
0e88460d	303	return node_zonelist(0, gfp_flags);
5da7ca86 CL	304	}
5da7ca86 CL	305
45b07ef3 PJ	306	static inline int do_migrate_pages(struct mm_struct *mm,
	307	const nodemask_t *from_nodes,
	308	const nodemask_t *to_nodes, int flags)
	309	{
	310	return 0;
	311	}
	312
4be38e35 CL	313	static inline void check_highest_zone(int k)
	314	{
	315	}
1da177e4 LT	316	#endif /* CONFIG_NUMA */
	317	#endif /* __KERNEL__ */
	318
	319	#endif