[linux-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.txt> 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.txt> 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);
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) {}
#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(gfp_t gfp_mask);
extern void fs_reclaim_release(gfp_t gfp_mask);
#else
static inline void fs_reclaim_acquire(gfp_t gfp_mask) { }
static inline void fs_reclaim_release(gfp_t gfp_mask) { }
#endif

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

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

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

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	*
	31	* See also <Documentation/vm/active_mm.txt> for an in-depth explanation
	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	*
	65	* See also <Documentation/vm/active_mm.txt> for an in-depth explanation
	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
	107	extern void arch_pick_mmap_layout(struct mm_struct *mm);
	108	extern unsigned long
	109	arch_get_unmapped_area(struct file *, unsigned long, unsigned long,
	110	unsigned long, unsigned long);
	111	extern unsigned long
	112	arch_get_unmapped_area_topdown(struct file *filp, unsigned long addr,
	113	unsigned long len, unsigned long pgoff,
	114	unsigned long flags);
	115	#else
	116	static inline void arch_pick_mmap_layout(struct mm_struct *mm) {}
	117	#endif
	118
d026ce79 IM	119	static inline bool in_vfork(struct task_struct *tsk)
	120	{
	121	bool ret;
	122
	123	/*
	124	* need RCU to access ->real_parent if CLONE_VM was used along with
	125	* CLONE_PARENT.
	126	*
	127	* We check real_parent->mm == tsk->mm because CLONE_VFORK does not
	128	* imply CLONE_VM
	129	*
	130	* CLONE_VFORK can be used with CLONE_PARENT/CLONE_THREAD and thus
	131	* ->real_parent is not necessarily the task doing vfork(), so in
	132	* theory we can't rely on task_lock() if we want to dereference it.
	133	*
	134	* And in this case we can't trust the real_parent->mm == tsk->mm
	135	* check, it can be false negative. But we do not care, if init or
	136	* another oom-unkillable task does this it should blame itself.
	137	*/
	138	rcu_read_lock();
	139	ret = tsk->vfork_done && tsk->real_parent->mm == tsk->mm;
	140	rcu_read_unlock();
	141
	142	return ret;
	143	}
	144
7dea19f9 MH	145	/*
	146	* Applies per-task gfp context to the given allocation flags.
	147	* PF_MEMALLOC_NOIO implies GFP_NOIO
	148	* PF_MEMALLOC_NOFS implies GFP_NOFS
74444eda	149	*/
7dea19f9	150	static inline gfp_t current_gfp_context(gfp_t flags)
74444eda	151	{
7dea19f9 MH	152	/*
	153	* NOIO implies both NOIO and NOFS and it is a weaker context
	154	* so always make sure it makes precendence
	155	*/
74444eda IM	156	if (unlikely(current->flags & PF_MEMALLOC_NOIO))
74444eda IM	157	flags &= ~(__GFP_IO \| __GFP_FS);
7dea19f9 MH	158	else if (unlikely(current->flags & PF_MEMALLOC_NOFS))
7dea19f9 MH	159	flags &= ~__GFP_FS;
74444eda IM	160	return flags;
	161	}
	162
d92a8cfc PZ	163	#ifdef CONFIG_LOCKDEP
	164	extern void fs_reclaim_acquire(gfp_t gfp_mask);
	165	extern void fs_reclaim_release(gfp_t gfp_mask);
	166	#else
	167	static inline void fs_reclaim_acquire(gfp_t gfp_mask) { }
	168	static inline void fs_reclaim_release(gfp_t gfp_mask) { }
	169	#endif
	170
74444eda IM	171	static inline unsigned int memalloc_noio_save(void)
	172	{
	173	unsigned int flags = current->flags & PF_MEMALLOC_NOIO;
	174	current->flags \|= PF_MEMALLOC_NOIO;
	175	return flags;
	176	}
	177
	178	static inline void memalloc_noio_restore(unsigned int flags)
	179	{
	180	current->flags = (current->flags & ~PF_MEMALLOC_NOIO) \| flags;
	181	}
	182
7dea19f9 MH	183	static inline unsigned int memalloc_nofs_save(void)
	184	{
	185	unsigned int flags = current->flags & PF_MEMALLOC_NOFS;
	186	current->flags \|= PF_MEMALLOC_NOFS;
	187	return flags;
	188	}
	189
	190	static inline void memalloc_nofs_restore(unsigned int flags)
	191	{
	192	current->flags = (current->flags & ~PF_MEMALLOC_NOFS) \| flags;
	193	}
	194
499118e9 VB	195	static inline unsigned int memalloc_noreclaim_save(void)
	196	{
	197	unsigned int flags = current->flags & PF_MEMALLOC;
	198	current->flags \|= PF_MEMALLOC;
	199	return flags;
	200	}
	201
	202	static inline void memalloc_noreclaim_restore(unsigned int flags)
	203	{
	204	current->flags = (current->flags & ~PF_MEMALLOC) \| flags;
	205	}
	206
a961e409 MD	207	#ifdef CONFIG_MEMBARRIER
a961e409 MD	208	enum {
c5f58bd5 MD	209	MEMBARRIER_STATE_PRIVATE_EXPEDITED_READY = (1U << 0),
	210	MEMBARRIER_STATE_PRIVATE_EXPEDITED = (1U << 1),
	211	MEMBARRIER_STATE_GLOBAL_EXPEDITED_READY = (1U << 2),
	212	MEMBARRIER_STATE_GLOBAL_EXPEDITED = (1U << 3),
70216e18 MD	213	MEMBARRIER_STATE_PRIVATE_EXPEDITED_SYNC_CORE_READY = (1U << 4),
	214	MEMBARRIER_STATE_PRIVATE_EXPEDITED_SYNC_CORE = (1U << 5),
	215	};
	216
	217	enum {
	218	MEMBARRIER_FLAG_SYNC_CORE = (1U << 0),
a961e409 MD	219	};
a961e409 MD	220
3ccfebed MD	221	#ifdef CONFIG_ARCH_HAS_MEMBARRIER_CALLBACKS
	222	#include <asm/membarrier.h>
	223	#endif
	224
70216e18 MD	225	static inline void membarrier_mm_sync_core_before_usermode(struct mm_struct *mm)
	226	{
	227	if (likely(!(atomic_read(&mm->membarrier_state) &
	228	MEMBARRIER_STATE_PRIVATE_EXPEDITED_SYNC_CORE)))
	229	return;
	230	sync_core_before_usermode();
	231	}
	232
a961e409 MD	233	static inline void membarrier_execve(struct task_struct *t)
	234	{
	235	atomic_set(&t->mm->membarrier_state, 0);
	236	}
	237	#else
3ccfebed MD	238	#ifdef CONFIG_ARCH_HAS_MEMBARRIER_CALLBACKS
	239	static inline void membarrier_arch_switch_mm(struct mm_struct *prev,
	240	struct mm_struct *next,
	241	struct task_struct *tsk)
	242	{
	243	}
	244	#endif
a961e409 MD	245	static inline void membarrier_execve(struct task_struct *t)
	246	{
	247	}
70216e18 MD	248	static inline void membarrier_mm_sync_core_before_usermode(struct mm_struct *mm)
	249	{
	250	}
a961e409 MD	251	#endif
a961e409 MD	252
6e84f315	253	#endif /* _LINUX_SCHED_MM_H */