[linux-2.6-block.git] / include / linux / sched / mm.h

/* SPDX-License-Identifier: GPL-2.0 */
#ifndef _LINUX_SCHED_MM_H
#define _LINUX_SCHED_MM_H

#include <linux/kernel.h>
#include <linux/atomic.h>
#include <linux/sched.h>
#include <linux/mm_types.h>
#include <linux/gfp.h>
#include <linux/sync_core.h>

/*
 * Routines for handling mm_structs
 */
extern struct mm_struct *mm_alloc(void);

/**
 * mmgrab() - Pin a &struct mm_struct.
 * @mm: The &struct mm_struct to pin.
 *
 * Make sure that @mm will not get freed even after the owning task
 * exits. This doesn't guarantee that the associated address space
 * will still exist later on and mmget_not_zero() has to be used before
 * accessing it.
 *
 * This is a preferred way to to pin @mm for a longer/unbounded amount
 * of time.
 *
 * Use mmdrop() to release the reference acquired by mmgrab().
 *
 * See also <Documentation/vm/active_mm.rst> for an in-depth explanation
 * of &mm_struct.mm_count vs &mm_struct.mm_users.
 */
static inline void mmgrab(struct mm_struct *mm)
{
	atomic_inc(&mm->mm_count);
}

extern void __mmdrop(struct mm_struct *mm);

static inline void mmdrop(struct mm_struct *mm)
{
	/*
	 * The implicit full barrier implied by atomic_dec_and_test() is
	 * required by the membarrier system call before returning to
	 * user-space, after storing to rq->curr.
	 */
	if (unlikely(atomic_dec_and_test(&mm->mm_count)))
		__mmdrop(mm);
}

/**
 * mmget() - Pin the address space associated with a &struct mm_struct.
 * @mm: The address space to pin.
 *
 * Make sure that the address space of the given &struct mm_struct doesn't
 * go away. This does not protect against parts of the address space being
 * modified or freed, however.
 *
 * Never use this function to pin this address space for an
 * unbounded/indefinite amount of time.
 *
 * Use mmput() to release the reference acquired by mmget().
 *
 * See also <Documentation/vm/active_mm.rst> for an in-depth explanation
 * of &mm_struct.mm_count vs &mm_struct.mm_users.
 */
static inline void mmget(struct mm_struct *mm)
{
	atomic_inc(&mm->mm_users);
}

static inline bool mmget_not_zero(struct mm_struct *mm)
{
	return atomic_inc_not_zero(&mm->mm_users);
}

/* mmput gets rid of the mappings and all user-space */
extern void mmput(struct mm_struct *);
#ifdef CONFIG_MMU
/* same as above but performs the slow path from the async context. Can
 * be called from the atomic context as well
 */
void mmput_async(struct mm_struct *);
#endif

/* Grab a reference to a task's mm, if it is not already going away */
extern struct mm_struct *get_task_mm(struct task_struct *task);
/*
 * Grab a reference to a task's mm, if it is not already going away
 * and ptrace_may_access with the mode parameter passed to it
 * succeeds.
 */
extern struct mm_struct *mm_access(struct task_struct *task, unsigned int mode);
/* Remove the current tasks stale references to the old mm_struct */
extern void mm_release(struct task_struct *, struct mm_struct *);

#ifdef CONFIG_MEMCG
extern void mm_update_next_owner(struct mm_struct *mm);
#else
static inline void mm_update_next_owner(struct mm_struct *mm)
{
}
#endif /* CONFIG_MEMCG */

#ifdef CONFIG_MMU
extern void arch_pick_mmap_layout(struct mm_struct *mm,
				  struct rlimit *rlim_stack);
extern unsigned long
arch_get_unmapped_area(struct file *, unsigned long, unsigned long,
		       unsigned long, unsigned long);
extern unsigned long
arch_get_unmapped_area_topdown(struct file *filp, unsigned long addr,
			  unsigned long len, unsigned long pgoff,
			  unsigned long flags);
#else
static inline void arch_pick_mmap_layout(struct mm_struct *mm,
					 struct rlimit *rlim_stack) {}
#endif

static inline bool in_vfork(struct task_struct *tsk)
{
	bool ret;

	/*
	 * need RCU to access ->real_parent if CLONE_VM was used along with
	 * CLONE_PARENT.
	 *
	 * We check real_parent->mm == tsk->mm because CLONE_VFORK does not
	 * imply CLONE_VM
	 *
	 * CLONE_VFORK can be used with CLONE_PARENT/CLONE_THREAD and thus
	 * ->real_parent is not necessarily the task doing vfork(), so in
	 * theory we can't rely on task_lock() if we want to dereference it.
	 *
	 * And in this case we can't trust the real_parent->mm == tsk->mm
	 * check, it can be false negative. But we do not care, if init or
	 * another oom-unkillable task does this it should blame itself.
	 */
	rcu_read_lock();
	ret = tsk->vfork_done && tsk->real_parent->mm == tsk->mm;
	rcu_read_unlock();

	return ret;
}

/*
 * Applies per-task gfp context to the given allocation flags.
 * PF_MEMALLOC_NOIO implies GFP_NOIO
 * PF_MEMALLOC_NOFS implies GFP_NOFS
 */
static inline gfp_t current_gfp_context(gfp_t flags)
{
	/*
	 * NOIO implies both NOIO and NOFS and it is a weaker context
	 * so always make sure it makes precendence
	 */
	if (unlikely(current->flags & PF_MEMALLOC_NOIO))
		flags &= ~(__GFP_IO | __GFP_FS);
	else if (unlikely(current->flags & PF_MEMALLOC_NOFS))
		flags &= ~__GFP_FS;
	return flags;
}

#ifdef CONFIG_LOCKDEP
extern void __fs_reclaim_acquire(void);
extern void __fs_reclaim_release(void);
extern void fs_reclaim_acquire(gfp_t gfp_mask);
extern void fs_reclaim_release(gfp_t gfp_mask);
#else
static inline void __fs_reclaim_acquire(void) { }
static inline void __fs_reclaim_release(void) { }
static inline void fs_reclaim_acquire(gfp_t gfp_mask) { }
static inline void fs_reclaim_release(gfp_t gfp_mask) { }
#endif

/**
 * memalloc_noio_save - Marks implicit GFP_NOIO allocation scope.
 *
 * This functions marks the beginning of the GFP_NOIO allocation scope.
 * All further allocations will implicitly drop __GFP_IO flag and so
 * they are safe for the IO critical section from the allocation recursion
 * point of view. Use memalloc_noio_restore to end the scope with flags
 * returned by this function.
 *
 * This function is safe to be used from any context.
 */
static inline unsigned int memalloc_noio_save(void)
{
	unsigned int flags = current->flags & PF_MEMALLOC_NOIO;
	current->flags |= PF_MEMALLOC_NOIO;
	return flags;
}

/**
 * memalloc_noio_restore - Ends the implicit GFP_NOIO scope.
 * @flags: Flags to restore.
 *
 * Ends the implicit GFP_NOIO scope started by memalloc_noio_save function.
 * Always make sure that that the given flags is the return value from the
 * pairing memalloc_noio_save call.
 */
static inline void memalloc_noio_restore(unsigned int flags)
{
	current->flags = (current->flags & ~PF_MEMALLOC_NOIO) | flags;
}

/**
 * memalloc_nofs_save - Marks implicit GFP_NOFS allocation scope.
 *
 * This functions marks the beginning of the GFP_NOFS allocation scope.
 * All further allocations will implicitly drop __GFP_FS flag and so
 * they are safe for the FS critical section from the allocation recursion
 * point of view. Use memalloc_nofs_restore to end the scope with flags
 * returned by this function.
 *
 * This function is safe to be used from any context.
 */
static inline unsigned int memalloc_nofs_save(void)
{
	unsigned int flags = current->flags & PF_MEMALLOC_NOFS;
	current->flags |= PF_MEMALLOC_NOFS;
	return flags;
}

/**
 * memalloc_nofs_restore - Ends the implicit GFP_NOFS scope.
 * @flags: Flags to restore.
 *
 * Ends the implicit GFP_NOFS scope started by memalloc_nofs_save function.
 * Always make sure that that the given flags is the return value from the
 * pairing memalloc_nofs_save call.
 */
static inline void memalloc_nofs_restore(unsigned int flags)
{
	current->flags = (current->flags & ~PF_MEMALLOC_NOFS) | flags;
}

static inline unsigned int memalloc_noreclaim_save(void)
{
	unsigned int flags = current->flags & PF_MEMALLOC;
	current->flags |= PF_MEMALLOC;
	return flags;
}

static inline void memalloc_noreclaim_restore(unsigned int flags)
{
	current->flags = (current->flags & ~PF_MEMALLOC) | flags;
}

#ifdef CONFIG_MEMBARRIER
enum {
	MEMBARRIER_STATE_PRIVATE_EXPEDITED_READY		= (1U << 0),
	MEMBARRIER_STATE_PRIVATE_EXPEDITED			= (1U << 1),
	MEMBARRIER_STATE_GLOBAL_EXPEDITED_READY			= (1U << 2),
	MEMBARRIER_STATE_GLOBAL_EXPEDITED			= (1U << 3),
	MEMBARRIER_STATE_PRIVATE_EXPEDITED_SYNC_CORE_READY	= (1U << 4),
	MEMBARRIER_STATE_PRIVATE_EXPEDITED_SYNC_CORE		= (1U << 5),
};

enum {
	MEMBARRIER_FLAG_SYNC_CORE	= (1U << 0),
};

#ifdef CONFIG_ARCH_HAS_MEMBARRIER_CALLBACKS
#include <asm/membarrier.h>
#endif

static inline void membarrier_mm_sync_core_before_usermode(struct mm_struct *mm)
{
	if (likely(!(atomic_read(&mm->membarrier_state) &
		     MEMBARRIER_STATE_PRIVATE_EXPEDITED_SYNC_CORE)))
		return;
	sync_core_before_usermode();
}

static inline void membarrier_execve(struct task_struct *t)
{
	atomic_set(&t->mm->membarrier_state, 0);
}
#else
#ifdef CONFIG_ARCH_HAS_MEMBARRIER_CALLBACKS
static inline void membarrier_arch_switch_mm(struct mm_struct *prev,
					     struct mm_struct *next,
					     struct task_struct *tsk)
{
}
#endif
static inline void membarrier_execve(struct task_struct *t)
{
}
static inline void membarrier_mm_sync_core_before_usermode(struct mm_struct *mm)
{
}
#endif

#endif /* _LINUX_SCHED_MM_H */
Commit	Line	Data
b2441318	1	/* SPDX-License-Identifier: GPL-2.0 */
6e84f315 IM	2	#ifndef _LINUX_SCHED_MM_H
	3	#define _LINUX_SCHED_MM_H
	4
b8d6d80b IM	5	#include <linux/kernel.h>
b8d6d80b IM	6	#include <linux/atomic.h>
6e84f315	7	#include <linux/sched.h>
589ee628	8	#include <linux/mm_types.h>
fd771233	9	#include <linux/gfp.h>
70216e18	10	#include <linux/sync_core.h>
6e84f315	11
68e21be2 IM	12	/*
	13	* Routines for handling mm_structs
	14	*/
d70f2a14	15	extern struct mm_struct *mm_alloc(void);
68e21be2 IM	16
	17	/**
	18	* mmgrab() - Pin a &struct mm_struct.
	19	* @mm: The &struct mm_struct to pin.
	20	*
	21	* Make sure that @mm will not get freed even after the owning task
	22	* exits. This doesn't guarantee that the associated address space
	23	* will still exist later on and mmget_not_zero() has to be used before
	24	* accessing it.
	25	*
	26	* This is a preferred way to to pin @mm for a longer/unbounded amount
	27	* of time.
	28	*
	29	* Use mmdrop() to release the reference acquired by mmgrab().
	30	*
ad56b738	31	* See also <Documentation/vm/active_mm.rst> for an in-depth explanation
68e21be2 IM	32	* of &mm_struct.mm_count vs &mm_struct.mm_users.
	33	*/
	34	static inline void mmgrab(struct mm_struct *mm)
	35	{
	36	atomic_inc(&mm->mm_count);
	37	}
	38
d34bc48f AM	39	extern void __mmdrop(struct mm_struct *mm);
	40
	41	static inline void mmdrop(struct mm_struct *mm)
	42	{
	43	/*
	44	* The implicit full barrier implied by atomic_dec_and_test() is
	45	* required by the membarrier system call before returning to
	46	* user-space, after storing to rq->curr.
	47	*/
	48	if (unlikely(atomic_dec_and_test(&mm->mm_count)))
	49	__mmdrop(mm);
	50	}
68e21be2 IM	51
	52	/**
	53	* mmget() - Pin the address space associated with a &struct mm_struct.
	54	* @mm: The address space to pin.
	55	*
	56	* Make sure that the address space of the given &struct mm_struct doesn't
	57	* go away. This does not protect against parts of the address space being
	58	* modified or freed, however.
	59	*
	60	* Never use this function to pin this address space for an
	61	* unbounded/indefinite amount of time.
	62	*
	63	* Use mmput() to release the reference acquired by mmget().
	64	*
ad56b738	65	* See also <Documentation/vm/active_mm.rst> for an in-depth explanation
68e21be2 IM	66	* of &mm_struct.mm_count vs &mm_struct.mm_users.
	67	*/
	68	static inline void mmget(struct mm_struct *mm)
	69	{
	70	atomic_inc(&mm->mm_users);
	71	}
	72
	73	static inline bool mmget_not_zero(struct mm_struct *mm)
	74	{
	75	return atomic_inc_not_zero(&mm->mm_users);
	76	}
	77
	78	/* mmput gets rid of the mappings and all user-space */
	79	extern void mmput(struct mm_struct *);
a1b2289c SY	80	#ifdef CONFIG_MMU
	81	/* same as above but performs the slow path from the async context. Can
	82	* be called from the atomic context as well
	83	*/
	84	void mmput_async(struct mm_struct *);
	85	#endif
68e21be2 IM	86
	87	/* Grab a reference to a task's mm, if it is not already going away */
	88	extern struct mm_struct get_task_mm(struct task_struct task);
	89	/*
	90	* Grab a reference to a task's mm, if it is not already going away
	91	* and ptrace_may_access with the mode parameter passed to it
	92	* succeeds.
	93	*/
	94	extern struct mm_struct mm_access(struct task_struct task, unsigned int mode);
	95	/* Remove the current tasks stale references to the old mm_struct */
	96	extern void mm_release(struct task_struct , struct mm_struct );
	97
4240c8bf IM	98	#ifdef CONFIG_MEMCG
	99	extern void mm_update_next_owner(struct mm_struct *mm);
	100	#else
	101	static inline void mm_update_next_owner(struct mm_struct *mm)
	102	{
	103	}
	104	#endif /* CONFIG_MEMCG */
	105
	106	#ifdef CONFIG_MMU
8f2af155 KC	107	extern void arch_pick_mmap_layout(struct mm_struct *mm,
8f2af155 KC	108	struct rlimit *rlim_stack);
4240c8bf IM	109	extern unsigned long
	110	arch_get_unmapped_area(struct file *, unsigned long, unsigned long,
	111	unsigned long, unsigned long);
	112	extern unsigned long
	113	arch_get_unmapped_area_topdown(struct file *filp, unsigned long addr,
	114	unsigned long len, unsigned long pgoff,
	115	unsigned long flags);
	116	#else
8f2af155 KC	117	static inline void arch_pick_mmap_layout(struct mm_struct *mm,
8f2af155 KC	118	struct rlimit *rlim_stack) {}
4240c8bf IM	119	#endif
4240c8bf IM	120
d026ce79 IM	121	static inline bool in_vfork(struct task_struct *tsk)
	122	{
	123	bool ret;
	124
	125	/*
	126	* need RCU to access ->real_parent if CLONE_VM was used along with
	127	* CLONE_PARENT.
	128	*
	129	* We check real_parent->mm == tsk->mm because CLONE_VFORK does not
	130	* imply CLONE_VM
	131	*
	132	* CLONE_VFORK can be used with CLONE_PARENT/CLONE_THREAD and thus
	133	* ->real_parent is not necessarily the task doing vfork(), so in
	134	* theory we can't rely on task_lock() if we want to dereference it.
	135	*
	136	* And in this case we can't trust the real_parent->mm == tsk->mm
	137	* check, it can be false negative. But we do not care, if init or
	138	* another oom-unkillable task does this it should blame itself.
	139	*/
	140	rcu_read_lock();
	141	ret = tsk->vfork_done && tsk->real_parent->mm == tsk->mm;
	142	rcu_read_unlock();
	143
	144	return ret;
	145	}
	146
7dea19f9 MH	147	/*
	148	* Applies per-task gfp context to the given allocation flags.
	149	* PF_MEMALLOC_NOIO implies GFP_NOIO
	150	* PF_MEMALLOC_NOFS implies GFP_NOFS
74444eda	151	*/
7dea19f9	152	static inline gfp_t current_gfp_context(gfp_t flags)
74444eda	153	{
7dea19f9 MH	154	/*
	155	* NOIO implies both NOIO and NOFS and it is a weaker context
	156	* so always make sure it makes precendence
	157	*/
74444eda IM	158	if (unlikely(current->flags & PF_MEMALLOC_NOIO))
74444eda IM	159	flags &= ~(__GFP_IO \| __GFP_FS);
7dea19f9 MH	160	else if (unlikely(current->flags & PF_MEMALLOC_NOFS))
7dea19f9 MH	161	flags &= ~__GFP_FS;
74444eda IM	162	return flags;
	163	}
	164
d92a8cfc	165	#ifdef CONFIG_LOCKDEP
93781325 OS	166	extern void __fs_reclaim_acquire(void);
93781325 OS	167	extern void __fs_reclaim_release(void);
d92a8cfc PZ	168	extern void fs_reclaim_acquire(gfp_t gfp_mask);
	169	extern void fs_reclaim_release(gfp_t gfp_mask);
	170	#else
93781325 OS	171	static inline void __fs_reclaim_acquire(void) { }
93781325 OS	172	static inline void __fs_reclaim_release(void) { }
d92a8cfc PZ	173	static inline void fs_reclaim_acquire(gfp_t gfp_mask) { }
	174	static inline void fs_reclaim_release(gfp_t gfp_mask) { }
	175	#endif
	176
46ca3599 MH	177	/**
	178	* memalloc_noio_save - Marks implicit GFP_NOIO allocation scope.
	179	*
	180	* This functions marks the beginning of the GFP_NOIO allocation scope.
	181	* All further allocations will implicitly drop __GFP_IO flag and so
	182	* they are safe for the IO critical section from the allocation recursion
	183	* point of view. Use memalloc_noio_restore to end the scope with flags
	184	* returned by this function.
	185	*
	186	* This function is safe to be used from any context.
	187	*/
74444eda IM	188	static inline unsigned int memalloc_noio_save(void)
	189	{
	190	unsigned int flags = current->flags & PF_MEMALLOC_NOIO;
	191	current->flags \|= PF_MEMALLOC_NOIO;
	192	return flags;
	193	}
	194
46ca3599 MH	195	/**
	196	* memalloc_noio_restore - Ends the implicit GFP_NOIO scope.
	197	* @flags: Flags to restore.
	198	*
	199	* Ends the implicit GFP_NOIO scope started by memalloc_noio_save function.
	200	* Always make sure that that the given flags is the return value from the
	201	* pairing memalloc_noio_save call.
	202	*/
74444eda IM	203	static inline void memalloc_noio_restore(unsigned int flags)
	204	{
	205	current->flags = (current->flags & ~PF_MEMALLOC_NOIO) \| flags;
	206	}
	207
46ca3599 MH	208	/**
	209	* memalloc_nofs_save - Marks implicit GFP_NOFS allocation scope.
	210	*
	211	* This functions marks the beginning of the GFP_NOFS allocation scope.
	212	* All further allocations will implicitly drop __GFP_FS flag and so
	213	* they are safe for the FS critical section from the allocation recursion
	214	* point of view. Use memalloc_nofs_restore to end the scope with flags
	215	* returned by this function.
	216	*
	217	* This function is safe to be used from any context.
	218	*/
7dea19f9 MH	219	static inline unsigned int memalloc_nofs_save(void)
	220	{
	221	unsigned int flags = current->flags & PF_MEMALLOC_NOFS;
	222	current->flags \|= PF_MEMALLOC_NOFS;
	223	return flags;
	224	}
	225
46ca3599 MH	226	/**
	227	* memalloc_nofs_restore - Ends the implicit GFP_NOFS scope.
	228	* @flags: Flags to restore.
	229	*
	230	* Ends the implicit GFP_NOFS scope started by memalloc_nofs_save function.
	231	* Always make sure that that the given flags is the return value from the
	232	* pairing memalloc_nofs_save call.
	233	*/
7dea19f9 MH	234	static inline void memalloc_nofs_restore(unsigned int flags)
	235	{
	236	current->flags = (current->flags & ~PF_MEMALLOC_NOFS) \| flags;
	237	}
	238
499118e9 VB	239	static inline unsigned int memalloc_noreclaim_save(void)
	240	{
	241	unsigned int flags = current->flags & PF_MEMALLOC;
	242	current->flags \|= PF_MEMALLOC;
	243	return flags;
	244	}
	245
	246	static inline void memalloc_noreclaim_restore(unsigned int flags)
	247	{
	248	current->flags = (current->flags & ~PF_MEMALLOC) \| flags;
	249	}
	250
a961e409 MD	251	#ifdef CONFIG_MEMBARRIER
a961e409 MD	252	enum {
c5f58bd5 MD	253	MEMBARRIER_STATE_PRIVATE_EXPEDITED_READY = (1U << 0),
	254	MEMBARRIER_STATE_PRIVATE_EXPEDITED = (1U << 1),
	255	MEMBARRIER_STATE_GLOBAL_EXPEDITED_READY = (1U << 2),
	256	MEMBARRIER_STATE_GLOBAL_EXPEDITED = (1U << 3),
70216e18 MD	257	MEMBARRIER_STATE_PRIVATE_EXPEDITED_SYNC_CORE_READY = (1U << 4),
	258	MEMBARRIER_STATE_PRIVATE_EXPEDITED_SYNC_CORE = (1U << 5),
	259	};
	260
	261	enum {
	262	MEMBARRIER_FLAG_SYNC_CORE = (1U << 0),
a961e409 MD	263	};
a961e409 MD	264
3ccfebed MD	265	#ifdef CONFIG_ARCH_HAS_MEMBARRIER_CALLBACKS
	266	#include <asm/membarrier.h>
	267	#endif
	268
70216e18 MD	269	static inline void membarrier_mm_sync_core_before_usermode(struct mm_struct *mm)
	270	{
	271	if (likely(!(atomic_read(&mm->membarrier_state) &
	272	MEMBARRIER_STATE_PRIVATE_EXPEDITED_SYNC_CORE)))
	273	return;
	274	sync_core_before_usermode();
	275	}
	276
a961e409 MD	277	static inline void membarrier_execve(struct task_struct *t)
	278	{
	279	atomic_set(&t->mm->membarrier_state, 0);
	280	}
	281	#else
3ccfebed MD	282	#ifdef CONFIG_ARCH_HAS_MEMBARRIER_CALLBACKS
	283	static inline void membarrier_arch_switch_mm(struct mm_struct *prev,
	284	struct mm_struct *next,
	285	struct task_struct *tsk)
	286	{
	287	}
	288	#endif
a961e409 MD	289	static inline void membarrier_execve(struct task_struct *t)
	290	{
	291	}
70216e18 MD	292	static inline void membarrier_mm_sync_core_before_usermode(struct mm_struct *mm)
	293	{
	294	}
a961e409 MD	295	#endif
a961e409 MD	296
6e84f315	297	#endif /* _LINUX_SCHED_MM_H */