[linux-block.git] / include / linux / mmu_notifier.h

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

#include <linux/list.h>
#include <linux/spinlock.h>
#include <linux/mm_types.h>
#include <linux/srcu.h>

struct mmu_notifier;
struct mmu_notifier_ops;

#ifdef CONFIG_MMU_NOTIFIER

/*
 * The mmu notifier_mm structure is allocated and installed in
 * mm->mmu_notifier_mm inside the mm_take_all_locks() protected
 * critical section and it's released only when mm_count reaches zero
 * in mmdrop().
 */
struct mmu_notifier_mm {
	/* all mmu notifiers registerd in this mm are queued in this list */
	struct hlist_head list;
	/* to serialize the list modifications and hlist_unhashed */
	spinlock_t lock;
};

struct mmu_notifier_ops {
	/*
	 * Called either by mmu_notifier_unregister or when the mm is
	 * being destroyed by exit_mmap, always before all pages are
	 * freed. This can run concurrently with other mmu notifier
	 * methods (the ones invoked outside the mm context) and it
	 * should tear down all secondary mmu mappings and freeze the
	 * secondary mmu. If this method isn't implemented you've to
	 * be sure that nothing could possibly write to the pages
	 * through the secondary mmu by the time the last thread with
	 * tsk->mm == mm exits.
	 *
	 * As side note: the pages freed after ->release returns could
	 * be immediately reallocated by the gart at an alias physical
	 * address with a different cache model, so if ->release isn't
	 * implemented because all _software_ driven memory accesses
	 * through the secondary mmu are terminated by the time the
	 * last thread of this mm quits, you've also to be sure that
	 * speculative _hardware_ operations can't allocate dirty
	 * cachelines in the cpu that could not be snooped and made
	 * coherent with the other read and write operations happening
	 * through the gart alias address, so leading to memory
	 * corruption.
	 */
	void (*release)(struct mmu_notifier *mn,
			struct mm_struct *mm);

	/*
	 * clear_flush_young is called after the VM is
	 * test-and-clearing the young/accessed bitflag in the
	 * pte. This way the VM will provide proper aging to the
	 * accesses to the page through the secondary MMUs and not
	 * only to the ones through the Linux pte.
	 * Start-end is necessary in case the secondary MMU is mapping the page
	 * at a smaller granularity than the primary MMU.
	 */
	int (*clear_flush_young)(struct mmu_notifier *mn,
				 struct mm_struct *mm,
				 unsigned long start,
				 unsigned long end);

	/*
	 * clear_young is a lightweight version of clear_flush_young. Like the
	 * latter, it is supposed to test-and-clear the young/accessed bitflag
	 * in the secondary pte, but it may omit flushing the secondary tlb.
	 */
	int (*clear_young)(struct mmu_notifier *mn,
			   struct mm_struct *mm,
			   unsigned long start,
			   unsigned long end);

	/*
	 * test_young is called to check the young/accessed bitflag in
	 * the secondary pte. This is used to know if the page is
	 * frequently used without actually clearing the flag or tearing
	 * down the secondary mapping on the page.
	 */
	int (*test_young)(struct mmu_notifier *mn,
			  struct mm_struct *mm,
			  unsigned long address);

	/*
	 * change_pte is called in cases that pte mapping to page is changed:
	 * for example, when ksm remaps pte to point to a new shared page.
	 */
	void (*change_pte)(struct mmu_notifier *mn,
			   struct mm_struct *mm,
			   unsigned long address,
			   pte_t pte);

	/*
	 * invalidate_range_start() and invalidate_range_end() must be
	 * paired and are called only when the mmap_sem and/or the
	 * locks protecting the reverse maps are held. If the subsystem
	 * can't guarantee that no additional references are taken to
	 * the pages in the range, it has to implement the
	 * invalidate_range() notifier to remove any references taken
	 * after invalidate_range_start().
	 *
	 * Invalidation of multiple concurrent ranges may be
	 * optionally permitted by the driver. Either way the
	 * establishment of sptes is forbidden in the range passed to
	 * invalidate_range_begin/end for the whole duration of the
	 * invalidate_range_begin/end critical section.
	 *
	 * invalidate_range_start() is called when all pages in the
	 * range are still mapped and have at least a refcount of one.
	 *
	 * invalidate_range_end() is called when all pages in the
	 * range have been unmapped and the pages have been freed by
	 * the VM.
	 *
	 * The VM will remove the page table entries and potentially
	 * the page between invalidate_range_start() and
	 * invalidate_range_end(). If the page must not be freed
	 * because of pending I/O or other circumstances then the
	 * invalidate_range_start() callback (or the initial mapping
	 * by the driver) must make sure that the refcount is kept
	 * elevated.
	 *
	 * If the driver increases the refcount when the pages are
	 * initially mapped into an address space then either
	 * invalidate_range_start() or invalidate_range_end() may
	 * decrease the refcount. If the refcount is decreased on
	 * invalidate_range_start() then the VM can free pages as page
	 * table entries are removed.  If the refcount is only
	 * droppped on invalidate_range_end() then the driver itself
	 * will drop the last refcount but it must take care to flush
	 * any secondary tlb before doing the final free on the
	 * page. Pages will no longer be referenced by the linux
	 * address space but may still be referenced by sptes until
	 * the last refcount is dropped.
	 *
	 * If blockable argument is set to false then the callback cannot
	 * sleep and has to return with -EAGAIN. 0 should be returned
	 * otherwise. Please note that if invalidate_range_start approves
	 * a non-blocking behavior then the same applies to
	 * invalidate_range_end.
	 *
	 */
	int (*invalidate_range_start)(struct mmu_notifier *mn,
				       struct mm_struct *mm,
				       unsigned long start, unsigned long end,
				       bool blockable);
	void (*invalidate_range_end)(struct mmu_notifier *mn,
				     struct mm_struct *mm,
				     unsigned long start, unsigned long end);

	/*
	 * invalidate_range() is either called between
	 * invalidate_range_start() and invalidate_range_end() when the
	 * VM has to free pages that where unmapped, but before the
	 * pages are actually freed, or outside of _start()/_end() when
	 * a (remote) TLB is necessary.
	 *
	 * If invalidate_range() is used to manage a non-CPU TLB with
	 * shared page-tables, it not necessary to implement the
	 * invalidate_range_start()/end() notifiers, as
	 * invalidate_range() alread catches the points in time when an
	 * external TLB range needs to be flushed. For more in depth
	 * discussion on this see Documentation/vm/mmu_notifier.rst
	 *
	 * Note that this function might be called with just a sub-range
	 * of what was passed to invalidate_range_start()/end(), if
	 * called between those functions.
	 */
	void (*invalidate_range)(struct mmu_notifier *mn, struct mm_struct *mm,
				 unsigned long start, unsigned long end);
};

/*
 * The notifier chains are protected by mmap_sem and/or the reverse map
 * semaphores. Notifier chains are only changed when all reverse maps and
 * the mmap_sem locks are taken.
 *
 * Therefore notifier chains can only be traversed when either
 *
 * 1. mmap_sem is held.
 * 2. One of the reverse map locks is held (i_mmap_rwsem or anon_vma->rwsem).
 * 3. No other concurrent thread can access the list (release)
 */
struct mmu_notifier {
	struct hlist_node hlist;
	const struct mmu_notifier_ops *ops;
};

static inline int mm_has_notifiers(struct mm_struct *mm)
{
	return unlikely(mm->mmu_notifier_mm);
}

extern int mmu_notifier_register(struct mmu_notifier *mn,
				 struct mm_struct *mm);
extern int __mmu_notifier_register(struct mmu_notifier *mn,
				   struct mm_struct *mm);
extern void mmu_notifier_unregister(struct mmu_notifier *mn,
				    struct mm_struct *mm);
extern void mmu_notifier_unregister_no_release(struct mmu_notifier *mn,
					       struct mm_struct *mm);
extern void __mmu_notifier_mm_destroy(struct mm_struct *mm);
extern void __mmu_notifier_release(struct mm_struct *mm);
extern int __mmu_notifier_clear_flush_young(struct mm_struct *mm,
					  unsigned long start,
					  unsigned long end);
extern int __mmu_notifier_clear_young(struct mm_struct *mm,
				      unsigned long start,
				      unsigned long end);
extern int __mmu_notifier_test_young(struct mm_struct *mm,
				     unsigned long address);
extern void __mmu_notifier_change_pte(struct mm_struct *mm,
				      unsigned long address, pte_t pte);
extern int __mmu_notifier_invalidate_range_start(struct mm_struct *mm,
				  unsigned long start, unsigned long end,
				  bool blockable);
extern void __mmu_notifier_invalidate_range_end(struct mm_struct *mm,
				  unsigned long start, unsigned long end,
				  bool only_end);
extern void __mmu_notifier_invalidate_range(struct mm_struct *mm,
				  unsigned long start, unsigned long end);

static inline void mmu_notifier_release(struct mm_struct *mm)
{
	if (mm_has_notifiers(mm))
		__mmu_notifier_release(mm);
}

static inline int mmu_notifier_clear_flush_young(struct mm_struct *mm,
					  unsigned long start,
					  unsigned long end)
{
	if (mm_has_notifiers(mm))
		return __mmu_notifier_clear_flush_young(mm, start, end);
	return 0;
}

static inline int mmu_notifier_clear_young(struct mm_struct *mm,
					   unsigned long start,
					   unsigned long end)
{
	if (mm_has_notifiers(mm))
		return __mmu_notifier_clear_young(mm, start, end);
	return 0;
}

static inline int mmu_notifier_test_young(struct mm_struct *mm,
					  unsigned long address)
{
	if (mm_has_notifiers(mm))
		return __mmu_notifier_test_young(mm, address);
	return 0;
}

static inline void mmu_notifier_change_pte(struct mm_struct *mm,
					   unsigned long address, pte_t pte)
{
	if (mm_has_notifiers(mm))
		__mmu_notifier_change_pte(mm, address, pte);
}

static inline void mmu_notifier_invalidate_range_start(struct mm_struct *mm,
				  unsigned long start, unsigned long end)
{
	if (mm_has_notifiers(mm))
		__mmu_notifier_invalidate_range_start(mm, start, end, true);
}

static inline int mmu_notifier_invalidate_range_start_nonblock(struct mm_struct *mm,
				  unsigned long start, unsigned long end)
{
	if (mm_has_notifiers(mm))
		return __mmu_notifier_invalidate_range_start(mm, start, end, false);
	return 0;
}

static inline void mmu_notifier_invalidate_range_end(struct mm_struct *mm,
				  unsigned long start, unsigned long end)
{
	if (mm_has_notifiers(mm))
		__mmu_notifier_invalidate_range_end(mm, start, end, false);
}

static inline void mmu_notifier_invalidate_range_only_end(struct mm_struct *mm,
				  unsigned long start, unsigned long end)
{
	if (mm_has_notifiers(mm))
		__mmu_notifier_invalidate_range_end(mm, start, end, true);
}

static inline void mmu_notifier_invalidate_range(struct mm_struct *mm,
				  unsigned long start, unsigned long end)
{
	if (mm_has_notifiers(mm))
		__mmu_notifier_invalidate_range(mm, start, end);
}

static inline void mmu_notifier_mm_init(struct mm_struct *mm)
{
	mm->mmu_notifier_mm = NULL;
}

static inline void mmu_notifier_mm_destroy(struct mm_struct *mm)
{
	if (mm_has_notifiers(mm))
		__mmu_notifier_mm_destroy(mm);
}

#define ptep_clear_flush_young_notify(__vma, __address, __ptep)		\
({									\
	int __young;							\
	struct vm_area_struct *___vma = __vma;				\
	unsigned long ___address = __address;				\
	__young = ptep_clear_flush_young(___vma, ___address, __ptep);	\
	__young |= mmu_notifier_clear_flush_young(___vma->vm_mm,	\
						  ___address,		\
						  ___address +		\
							PAGE_SIZE);	\
	__young;							\
})

#define pmdp_clear_flush_young_notify(__vma, __address, __pmdp)		\
({									\
	int __young;							\
	struct vm_area_struct *___vma = __vma;				\
	unsigned long ___address = __address;				\
	__young = pmdp_clear_flush_young(___vma, ___address, __pmdp);	\
	__young |= mmu_notifier_clear_flush_young(___vma->vm_mm,	\
						  ___address,		\
						  ___address +		\
							PMD_SIZE);	\
	__young;							\
})

#define ptep_clear_young_notify(__vma, __address, __ptep)		\
({									\
	int __young;							\
	struct vm_area_struct *___vma = __vma;				\
	unsigned long ___address = __address;				\
	__young = ptep_test_and_clear_young(___vma, ___address, __ptep);\
	__young |= mmu_notifier_clear_young(___vma->vm_mm, ___address,	\
					    ___address + PAGE_SIZE);	\
	__young;							\
})

#define pmdp_clear_young_notify(__vma, __address, __pmdp)		\
({									\
	int __young;							\
	struct vm_area_struct *___vma = __vma;				\
	unsigned long ___address = __address;				\
	__young = pmdp_test_and_clear_young(___vma, ___address, __pmdp);\
	__young |= mmu_notifier_clear_young(___vma->vm_mm, ___address,	\
					    ___address + PMD_SIZE);	\
	__young;							\
})

#define	ptep_clear_flush_notify(__vma, __address, __ptep)		\
({									\
	unsigned long ___addr = __address & PAGE_MASK;			\
	struct mm_struct *___mm = (__vma)->vm_mm;			\
	pte_t ___pte;							\
									\
	___pte = ptep_clear_flush(__vma, __address, __ptep);		\
	mmu_notifier_invalidate_range(___mm, ___addr,			\
					___addr + PAGE_SIZE);		\
									\
	___pte;								\
})

#define pmdp_huge_clear_flush_notify(__vma, __haddr, __pmd)		\
({									\
	unsigned long ___haddr = __haddr & HPAGE_PMD_MASK;		\
	struct mm_struct *___mm = (__vma)->vm_mm;			\
	pmd_t ___pmd;							\
									\
	___pmd = pmdp_huge_clear_flush(__vma, __haddr, __pmd);		\
	mmu_notifier_invalidate_range(___mm, ___haddr,			\
				      ___haddr + HPAGE_PMD_SIZE);	\
									\
	___pmd;								\
})

#define pudp_huge_clear_flush_notify(__vma, __haddr, __pud)		\
({									\
	unsigned long ___haddr = __haddr & HPAGE_PUD_MASK;		\
	struct mm_struct *___mm = (__vma)->vm_mm;			\
	pud_t ___pud;							\
									\
	___pud = pudp_huge_clear_flush(__vma, __haddr, __pud);		\
	mmu_notifier_invalidate_range(___mm, ___haddr,			\
				      ___haddr + HPAGE_PUD_SIZE);	\
									\
	___pud;								\
})

/*
 * set_pte_at_notify() sets the pte _after_ running the notifier.
 * This is safe to start by updating the secondary MMUs, because the primary MMU
 * pte invalidate must have already happened with a ptep_clear_flush() before
 * set_pte_at_notify() has been invoked.  Updating the secondary MMUs first is
 * required when we change both the protection of the mapping from read-only to
 * read-write and the pfn (like during copy on write page faults). Otherwise the
 * old page would remain mapped readonly in the secondary MMUs after the new
 * page is already writable by some CPU through the primary MMU.
 */
#define set_pte_at_notify(__mm, __address, __ptep, __pte)		\
({									\
	struct mm_struct *___mm = __mm;					\
	unsigned long ___address = __address;				\
	pte_t ___pte = __pte;						\
									\
	mmu_notifier_change_pte(___mm, ___address, ___pte);		\
	set_pte_at(___mm, ___address, __ptep, ___pte);			\
})

extern void mmu_notifier_call_srcu(struct rcu_head *rcu,
				   void (*func)(struct rcu_head *rcu));
extern void mmu_notifier_synchronize(void);

#else /* CONFIG_MMU_NOTIFIER */

static inline int mm_has_notifiers(struct mm_struct *mm)
{
	return 0;
}

static inline void mmu_notifier_release(struct mm_struct *mm)
{
}

static inline int mmu_notifier_clear_flush_young(struct mm_struct *mm,
					  unsigned long start,
					  unsigned long end)
{
	return 0;
}

static inline int mmu_notifier_test_young(struct mm_struct *mm,
					  unsigned long address)
{
	return 0;
}

static inline void mmu_notifier_change_pte(struct mm_struct *mm,
					   unsigned long address, pte_t pte)
{
}

static inline void mmu_notifier_invalidate_range_start(struct mm_struct *mm,
				  unsigned long start, unsigned long end)
{
}

static inline int mmu_notifier_invalidate_range_start_nonblock(struct mm_struct *mm,
				  unsigned long start, unsigned long end)
{
	return 0;
}

static inline void mmu_notifier_invalidate_range_end(struct mm_struct *mm,
				  unsigned long start, unsigned long end)
{
}

static inline void mmu_notifier_invalidate_range_only_end(struct mm_struct *mm,
				  unsigned long start, unsigned long end)
{
}

static inline void mmu_notifier_invalidate_range(struct mm_struct *mm,
				  unsigned long start, unsigned long end)
{
}

static inline void mmu_notifier_mm_init(struct mm_struct *mm)
{
}

static inline void mmu_notifier_mm_destroy(struct mm_struct *mm)
{
}

#define ptep_clear_flush_young_notify ptep_clear_flush_young
#define pmdp_clear_flush_young_notify pmdp_clear_flush_young
#define ptep_clear_young_notify ptep_test_and_clear_young
#define pmdp_clear_young_notify pmdp_test_and_clear_young
#define	ptep_clear_flush_notify ptep_clear_flush
#define pmdp_huge_clear_flush_notify pmdp_huge_clear_flush
#define pudp_huge_clear_flush_notify pudp_huge_clear_flush
#define set_pte_at_notify set_pte_at

#endif /* CONFIG_MMU_NOTIFIER */

#endif /* _LINUX_MMU_NOTIFIER_H */
Commit	Line	Data
b2441318	1	/* SPDX-License-Identifier: GPL-2.0 */
cddb8a5c AA	2	#ifndef _LINUX_MMU_NOTIFIER_H
	3	#define _LINUX_MMU_NOTIFIER_H
	4
	5	#include <linux/list.h>
	6	#include <linux/spinlock.h>
	7	#include <linux/mm_types.h>
21a92735	8	#include <linux/srcu.h>
cddb8a5c AA	9
	10	struct mmu_notifier;
	11	struct mmu_notifier_ops;
	12
	13	#ifdef CONFIG_MMU_NOTIFIER
	14
	15	/*
	16	* The mmu notifier_mm structure is allocated and installed in
	17	* mm->mmu_notifier_mm inside the mm_take_all_locks() protected
	18	* critical section and it's released only when mm_count reaches zero
	19	* in mmdrop().
	20	*/
	21	struct mmu_notifier_mm {
	22	/* all mmu notifiers registerd in this mm are queued in this list */
	23	struct hlist_head list;
	24	/* to serialize the list modifications and hlist_unhashed */
	25	spinlock_t lock;
	26	};
	27
	28	struct mmu_notifier_ops {
	29	/*
	30	* Called either by mmu_notifier_unregister or when the mm is
	31	* being destroyed by exit_mmap, always before all pages are
	32	* freed. This can run concurrently with other mmu notifier
	33	* methods (the ones invoked outside the mm context) and it
	34	* should tear down all secondary mmu mappings and freeze the
	35	* secondary mmu. If this method isn't implemented you've to
	36	* be sure that nothing could possibly write to the pages
	37	* through the secondary mmu by the time the last thread with
	38	* tsk->mm == mm exits.
	39	*
	40	* As side note: the pages freed after ->release returns could
	41	* be immediately reallocated by the gart at an alias physical
	42	* address with a different cache model, so if ->release isn't
	43	* implemented because all _software_ driven memory accesses
	44	* through the secondary mmu are terminated by the time the
	45	* last thread of this mm quits, you've also to be sure that
	46	* speculative _hardware_ operations can't allocate dirty
	47	* cachelines in the cpu that could not be snooped and made
	48	* coherent with the other read and write operations happening
	49	* through the gart alias address, so leading to memory
	50	* corruption.
	51	*/
	52	void (release)(struct mmu_notifier mn,
	53	struct mm_struct *mm);
	54
	55	/*
	56	* clear_flush_young is called after the VM is
	57	* test-and-clearing the young/accessed bitflag in the
	58	* pte. This way the VM will provide proper aging to the
	59	* accesses to the page through the secondary MMUs and not
	60	* only to the ones through the Linux pte.
57128468 ALC	61	* Start-end is necessary in case the secondary MMU is mapping the page
57128468 ALC	62	* at a smaller granularity than the primary MMU.
cddb8a5c AA	63	*/
	64	int (clear_flush_young)(struct mmu_notifier mn,
	65	struct mm_struct *mm,
57128468 ALC	66	unsigned long start,
57128468 ALC	67	unsigned long end);
cddb8a5c	68
1d7715c6 VD	69	/*
	70	* clear_young is a lightweight version of clear_flush_young. Like the
	71	* latter, it is supposed to test-and-clear the young/accessed bitflag
	72	* in the secondary pte, but it may omit flushing the secondary tlb.
	73	*/
	74	int (clear_young)(struct mmu_notifier mn,
	75	struct mm_struct *mm,
	76	unsigned long start,
	77	unsigned long end);
	78
8ee53820 AA	79	/*
	80	* test_young is called to check the young/accessed bitflag in
	81	* the secondary pte. This is used to know if the page is
	82	* frequently used without actually clearing the flag or tearing
	83	* down the secondary mapping on the page.
	84	*/
	85	int (test_young)(struct mmu_notifier mn,
	86	struct mm_struct *mm,
	87	unsigned long address);
	88
828502d3 IE	89	/*
	90	* change_pte is called in cases that pte mapping to page is changed:
	91	* for example, when ksm remaps pte to point to a new shared page.
	92	*/
	93	void (change_pte)(struct mmu_notifier mn,
	94	struct mm_struct *mm,
	95	unsigned long address,
	96	pte_t pte);
	97
cddb8a5c AA	98	/*
	99	* invalidate_range_start() and invalidate_range_end() must be
	100	* paired and are called only when the mmap_sem and/or the
0f0a327f JR	101	* locks protecting the reverse maps are held. If the subsystem
	102	* can't guarantee that no additional references are taken to
	103	* the pages in the range, it has to implement the
	104	* invalidate_range() notifier to remove any references taken
	105	* after invalidate_range_start().
cddb8a5c AA	106	*
	107	* Invalidation of multiple concurrent ranges may be
	108	* optionally permitted by the driver. Either way the
	109	* establishment of sptes is forbidden in the range passed to
	110	* invalidate_range_begin/end for the whole duration of the
	111	* invalidate_range_begin/end critical section.
	112	*
	113	* invalidate_range_start() is called when all pages in the
	114	* range are still mapped and have at least a refcount of one.
	115	*
	116	* invalidate_range_end() is called when all pages in the
	117	* range have been unmapped and the pages have been freed by
	118	* the VM.
	119	*
	120	* The VM will remove the page table entries and potentially
	121	* the page between invalidate_range_start() and
	122	* invalidate_range_end(). If the page must not be freed
	123	* because of pending I/O or other circumstances then the
	124	* invalidate_range_start() callback (or the initial mapping
	125	* by the driver) must make sure that the refcount is kept
	126	* elevated.
	127	*
	128	* If the driver increases the refcount when the pages are
	129	* initially mapped into an address space then either
	130	* invalidate_range_start() or invalidate_range_end() may
	131	* decrease the refcount. If the refcount is decreased on
	132	* invalidate_range_start() then the VM can free pages as page
	133	* table entries are removed. If the refcount is only
	134	* droppped on invalidate_range_end() then the driver itself
	135	* will drop the last refcount but it must take care to flush
	136	* any secondary tlb before doing the final free on the
	137	* page. Pages will no longer be referenced by the linux
	138	* address space but may still be referenced by sptes until
	139	* the last refcount is dropped.
5ff7091f	140	*
93065ac7 MH	141	* If blockable argument is set to false then the callback cannot
93065ac7 MH	142	* sleep and has to return with -EAGAIN. 0 should be returned
33490af3 MH	143	* otherwise. Please note that if invalidate_range_start approves
	144	* a non-blocking behavior then the same applies to
	145	* invalidate_range_end.
93065ac7	146	*
cddb8a5c	147	*/
93065ac7	148	int (invalidate_range_start)(struct mmu_notifier mn,
cddb8a5c	149	struct mm_struct *mm,
93065ac7 MH	150	unsigned long start, unsigned long end,
93065ac7 MH	151	bool blockable);
cddb8a5c AA	152	void (invalidate_range_end)(struct mmu_notifier mn,
	153	struct mm_struct *mm,
	154	unsigned long start, unsigned long end);
0f0a327f JR	155
	156	/*
	157	* invalidate_range() is either called between
	158	* invalidate_range_start() and invalidate_range_end() when the
	159	* VM has to free pages that where unmapped, but before the
	160	* pages are actually freed, or outside of _start()/_end() when
	161	* a (remote) TLB is necessary.
	162	*
	163	* If invalidate_range() is used to manage a non-CPU TLB with
	164	* shared page-tables, it not necessary to implement the
	165	* invalidate_range_start()/end() notifiers, as
	166	* invalidate_range() alread catches the points in time when an
0f10851e	167	* external TLB range needs to be flushed. For more in depth
ad56b738	168	* discussion on this see Documentation/vm/mmu_notifier.rst
0f0a327f	169	*
0f0a327f JR	170	* Note that this function might be called with just a sub-range
	171	* of what was passed to invalidate_range_start()/end(), if
	172	* called between those functions.
	173	*/
	174	void (invalidate_range)(struct mmu_notifier mn, struct mm_struct *mm,
	175	unsigned long start, unsigned long end);
cddb8a5c AA	176	};
	177
	178	/*
	179	* The notifier chains are protected by mmap_sem and/or the reverse map
	180	* semaphores. Notifier chains are only changed when all reverse maps and
	181	* the mmap_sem locks are taken.
	182	*
	183	* Therefore notifier chains can only be traversed when either
	184	*
	185	* 1. mmap_sem is held.
c8c06efa	186	* 2. One of the reverse map locks is held (i_mmap_rwsem or anon_vma->rwsem).
cddb8a5c AA	187	* 3. No other concurrent thread can access the list (release)
	188	*/
	189	struct mmu_notifier {
	190	struct hlist_node hlist;
	191	const struct mmu_notifier_ops *ops;
	192	};
	193
	194	static inline int mm_has_notifiers(struct mm_struct *mm)
	195	{
	196	return unlikely(mm->mmu_notifier_mm);
	197	}
	198
	199	extern int mmu_notifier_register(struct mmu_notifier *mn,
	200	struct mm_struct *mm);
	201	extern int __mmu_notifier_register(struct mmu_notifier *mn,
	202	struct mm_struct *mm);
	203	extern void mmu_notifier_unregister(struct mmu_notifier *mn,
	204	struct mm_struct *mm);
b972216e PZ	205	extern void mmu_notifier_unregister_no_release(struct mmu_notifier *mn,
b972216e PZ	206	struct mm_struct *mm);
cddb8a5c AA	207	extern void __mmu_notifier_mm_destroy(struct mm_struct *mm);
	208	extern void __mmu_notifier_release(struct mm_struct *mm);
	209	extern int __mmu_notifier_clear_flush_young(struct mm_struct *mm,
57128468 ALC	210	unsigned long start,
57128468 ALC	211	unsigned long end);
1d7715c6 VD	212	extern int __mmu_notifier_clear_young(struct mm_struct *mm,
	213	unsigned long start,
	214	unsigned long end);
8ee53820 AA	215	extern int __mmu_notifier_test_young(struct mm_struct *mm,
8ee53820 AA	216	unsigned long address);
828502d3 IE	217	extern void __mmu_notifier_change_pte(struct mm_struct *mm,
828502d3 IE	218	unsigned long address, pte_t pte);
93065ac7 MH	219	extern int __mmu_notifier_invalidate_range_start(struct mm_struct *mm,
	220	unsigned long start, unsigned long end,
	221	bool blockable);
cddb8a5c	222	extern void __mmu_notifier_invalidate_range_end(struct mm_struct *mm,
4645b9fe JG	223	unsigned long start, unsigned long end,
4645b9fe JG	224	bool only_end);
0f0a327f JR	225	extern void __mmu_notifier_invalidate_range(struct mm_struct *mm,
0f0a327f JR	226	unsigned long start, unsigned long end);
cddb8a5c AA	227
	228	static inline void mmu_notifier_release(struct mm_struct *mm)
	229	{
	230	if (mm_has_notifiers(mm))
	231	__mmu_notifier_release(mm);
	232	}
	233
	234	static inline int mmu_notifier_clear_flush_young(struct mm_struct *mm,
57128468 ALC	235	unsigned long start,
57128468 ALC	236	unsigned long end)
cddb8a5c AA	237	{
cddb8a5c AA	238	if (mm_has_notifiers(mm))
57128468	239	return __mmu_notifier_clear_flush_young(mm, start, end);
cddb8a5c AA	240	return 0;
	241	}
	242
1d7715c6 VD	243	static inline int mmu_notifier_clear_young(struct mm_struct *mm,
	244	unsigned long start,
	245	unsigned long end)
	246	{
	247	if (mm_has_notifiers(mm))
	248	return __mmu_notifier_clear_young(mm, start, end);
	249	return 0;
	250	}
	251
8ee53820 AA	252	static inline int mmu_notifier_test_young(struct mm_struct *mm,
	253	unsigned long address)
	254	{
	255	if (mm_has_notifiers(mm))
	256	return __mmu_notifier_test_young(mm, address);
	257	return 0;
	258	}
	259
828502d3 IE	260	static inline void mmu_notifier_change_pte(struct mm_struct *mm,
	261	unsigned long address, pte_t pte)
	262	{
	263	if (mm_has_notifiers(mm))
	264	__mmu_notifier_change_pte(mm, address, pte);
	265	}
	266
cddb8a5c AA	267	static inline void mmu_notifier_invalidate_range_start(struct mm_struct *mm,
	268	unsigned long start, unsigned long end)
	269	{
	270	if (mm_has_notifiers(mm))
93065ac7 MH	271	__mmu_notifier_invalidate_range_start(mm, start, end, true);
	272	}
	273
	274	static inline int mmu_notifier_invalidate_range_start_nonblock(struct mm_struct *mm,
	275	unsigned long start, unsigned long end)
	276	{
	277	if (mm_has_notifiers(mm))
	278	return __mmu_notifier_invalidate_range_start(mm, start, end, false);
	279	return 0;
cddb8a5c AA	280	}
	281
	282	static inline void mmu_notifier_invalidate_range_end(struct mm_struct *mm,
	283	unsigned long start, unsigned long end)
	284	{
	285	if (mm_has_notifiers(mm))
4645b9fe JG	286	__mmu_notifier_invalidate_range_end(mm, start, end, false);
	287	}
	288
	289	static inline void mmu_notifier_invalidate_range_only_end(struct mm_struct *mm,
	290	unsigned long start, unsigned long end)
	291	{
	292	if (mm_has_notifiers(mm))
	293	__mmu_notifier_invalidate_range_end(mm, start, end, true);
cddb8a5c AA	294	}
cddb8a5c AA	295
1897bdc4 JR	296	static inline void mmu_notifier_invalidate_range(struct mm_struct *mm,
	297	unsigned long start, unsigned long end)
	298	{
0f0a327f JR	299	if (mm_has_notifiers(mm))
0f0a327f JR	300	__mmu_notifier_invalidate_range(mm, start, end);
1897bdc4 JR	301	}
1897bdc4 JR	302
cddb8a5c AA	303	static inline void mmu_notifier_mm_init(struct mm_struct *mm)
	304	{
	305	mm->mmu_notifier_mm = NULL;
	306	}
	307
	308	static inline void mmu_notifier_mm_destroy(struct mm_struct *mm)
	309	{
	310	if (mm_has_notifiers(mm))
	311	__mmu_notifier_mm_destroy(mm);
	312	}
	313
cddb8a5c AA	314	#define ptep_clear_flush_young_notify(__vma, __address, __ptep) \
	315	({ \
	316	int __young; \
	317	struct vm_area_struct *___vma = __vma; \
	318	unsigned long ___address = __address; \
	319	__young = ptep_clear_flush_young(___vma, ___address, __ptep); \
	320	__young \|= mmu_notifier_clear_flush_young(___vma->vm_mm, \
57128468 ALC	321	___address, \
	322	___address + \
	323	PAGE_SIZE); \
cddb8a5c AA	324	__young; \
	325	})
	326
91a4ee26 AA	327	#define pmdp_clear_flush_young_notify(__vma, __address, __pmdp) \
	328	({ \
	329	int __young; \
	330	struct vm_area_struct *___vma = __vma; \
	331	unsigned long ___address = __address; \
	332	__young = pmdp_clear_flush_young(___vma, ___address, __pmdp); \
	333	__young \|= mmu_notifier_clear_flush_young(___vma->vm_mm, \
57128468 ALC	334	___address, \
	335	___address + \
	336	PMD_SIZE); \
91a4ee26 AA	337	__young; \
	338	})
	339
1d7715c6 VD	340	#define ptep_clear_young_notify(__vma, __address, __ptep) \
	341	({ \
	342	int __young; \
	343	struct vm_area_struct *___vma = __vma; \
	344	unsigned long ___address = __address; \
	345	__young = ptep_test_and_clear_young(___vma, ___address, __ptep);\
	346	__young \|= mmu_notifier_clear_young(___vma->vm_mm, ___address, \
	347	___address + PAGE_SIZE); \
	348	__young; \
	349	})
	350
	351	#define pmdp_clear_young_notify(__vma, __address, __pmdp) \
	352	({ \
	353	int __young; \
	354	struct vm_area_struct *___vma = __vma; \
	355	unsigned long ___address = __address; \
	356	__young = pmdp_test_and_clear_young(___vma, ___address, __pmdp);\
	357	__young \|= mmu_notifier_clear_young(___vma->vm_mm, ___address, \
	358	___address + PMD_SIZE); \
	359	__young; \
	360	})
	361
34ee645e JR	362	#define ptep_clear_flush_notify(__vma, __address, __ptep) \
	363	({ \
	364	unsigned long ___addr = __address & PAGE_MASK; \
	365	struct mm_struct *___mm = (__vma)->vm_mm; \
	366	pte_t ___pte; \
	367	\
	368	___pte = ptep_clear_flush(__vma, __address, __ptep); \
	369	mmu_notifier_invalidate_range(___mm, ___addr, \
	370	___addr + PAGE_SIZE); \
	371	\
	372	___pte; \
	373	})
	374
8809aa2d	375	#define pmdp_huge_clear_flush_notify(__vma, __haddr, __pmd) \
34ee645e JR	376	({ \
	377	unsigned long ___haddr = __haddr & HPAGE_PMD_MASK; \
	378	struct mm_struct *___mm = (__vma)->vm_mm; \
	379	pmd_t ___pmd; \
	380	\
8809aa2d	381	___pmd = pmdp_huge_clear_flush(__vma, __haddr, __pmd); \
34ee645e JR	382	mmu_notifier_invalidate_range(___mm, ___haddr, \
	383	___haddr + HPAGE_PMD_SIZE); \
	384	\
	385	___pmd; \
	386	})
	387
a00cc7d9 MW	388	#define pudp_huge_clear_flush_notify(__vma, __haddr, __pud) \
	389	({ \
	390	unsigned long ___haddr = __haddr & HPAGE_PUD_MASK; \
	391	struct mm_struct *___mm = (__vma)->vm_mm; \
	392	pud_t ___pud; \
	393	\
	394	___pud = pudp_huge_clear_flush(__vma, __haddr, __pud); \
	395	mmu_notifier_invalidate_range(___mm, ___haddr, \
	396	___haddr + HPAGE_PUD_SIZE); \
	397	\
	398	___pud; \
	399	})
	400
48af0d7c XG	401	/*
	402	* set_pte_at_notify() sets the pte _after_ running the notifier.
	403	* This is safe to start by updating the secondary MMUs, because the primary MMU
	404	* pte invalidate must have already happened with a ptep_clear_flush() before
	405	* set_pte_at_notify() has been invoked. Updating the secondary MMUs first is
	406	* required when we change both the protection of the mapping from read-only to
	407	* read-write and the pfn (like during copy on write page faults). Otherwise the
	408	* old page would remain mapped readonly in the secondary MMUs after the new
	409	* page is already writable by some CPU through the primary MMU.
	410	*/
828502d3 IE	411	#define set_pte_at_notify(__mm, __address, __ptep, __pte) \
	412	({ \
	413	struct mm_struct *___mm = __mm; \
	414	unsigned long ___address = __address; \
	415	pte_t ___pte = __pte; \
	416	\
828502d3	417	mmu_notifier_change_pte(___mm, ___address, ___pte); \
48af0d7c	418	set_pte_at(___mm, ___address, __ptep, ___pte); \
828502d3 IE	419	})
828502d3 IE	420
b972216e PZ	421	extern void mmu_notifier_call_srcu(struct rcu_head *rcu,
	422	void (func)(struct rcu_head rcu));
	423	extern void mmu_notifier_synchronize(void);
	424
cddb8a5c AA	425	#else /* CONFIG_MMU_NOTIFIER */
cddb8a5c AA	426
4d4bbd85 MH	427	static inline int mm_has_notifiers(struct mm_struct *mm)
	428	{
	429	return 0;
	430	}
	431
cddb8a5c AA	432	static inline void mmu_notifier_release(struct mm_struct *mm)
	433	{
	434	}
	435
	436	static inline int mmu_notifier_clear_flush_young(struct mm_struct *mm,
57128468 ALC	437	unsigned long start,
57128468 ALC	438	unsigned long end)
8ee53820 AA	439	{
	440	return 0;
	441	}
	442
	443	static inline int mmu_notifier_test_young(struct mm_struct *mm,
	444	unsigned long address)
cddb8a5c AA	445	{
	446	return 0;
	447	}
	448
828502d3 IE	449	static inline void mmu_notifier_change_pte(struct mm_struct *mm,
	450	unsigned long address, pte_t pte)
	451	{
	452	}
	453
cddb8a5c AA	454	static inline void mmu_notifier_invalidate_range_start(struct mm_struct *mm,
	455	unsigned long start, unsigned long end)
	456	{
	457	}
	458
93065ac7 MH	459	static inline int mmu_notifier_invalidate_range_start_nonblock(struct mm_struct *mm,
	460	unsigned long start, unsigned long end)
	461	{
	462	return 0;
	463	}
	464
cddb8a5c AA	465	static inline void mmu_notifier_invalidate_range_end(struct mm_struct *mm,
	466	unsigned long start, unsigned long end)
	467	{
	468	}
	469
4645b9fe JG	470	static inline void mmu_notifier_invalidate_range_only_end(struct mm_struct *mm,
	471	unsigned long start, unsigned long end)
	472	{
	473	}
	474
1897bdc4 JR	475	static inline void mmu_notifier_invalidate_range(struct mm_struct *mm,
	476	unsigned long start, unsigned long end)
	477	{
	478	}
	479
cddb8a5c AA	480	static inline void mmu_notifier_mm_init(struct mm_struct *mm)
	481	{
	482	}
	483
	484	static inline void mmu_notifier_mm_destroy(struct mm_struct *mm)
	485	{
	486	}
	487
	488	#define ptep_clear_flush_young_notify ptep_clear_flush_young
91a4ee26	489	#define pmdp_clear_flush_young_notify pmdp_clear_flush_young
33c3fc71 VD	490	#define ptep_clear_young_notify ptep_test_and_clear_young
33c3fc71 VD	491	#define pmdp_clear_young_notify pmdp_test_and_clear_young
34ee645e	492	#define ptep_clear_flush_notify ptep_clear_flush
8809aa2d	493	#define pmdp_huge_clear_flush_notify pmdp_huge_clear_flush
a00cc7d9	494	#define pudp_huge_clear_flush_notify pudp_huge_clear_flush
828502d3	495	#define set_pte_at_notify set_pte_at
cddb8a5c AA	496
	497	#endif /* CONFIG_MMU_NOTIFIER */
	498
	499	#endif /* _LINUX_MMU_NOTIFIER_H */