[linux-2.6-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/types.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;

/* mmu_notifier_ops flags */
#define MMU_INVALIDATE_DOES_NOT_BLOCK	(0x01)

#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 {
	/*
	 * Flags to specify behavior of callbacks for this MMU notifier.
	 * Used to determine which context an operation may be called.
	 *
	 * MMU_INVALIDATE_DOES_NOT_BLOCK: invalidate_range_* callbacks do not
	 *	block
	 */
	int flags;

	/*
	 * 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 both of these callbacks cannot block, and invalidate_range
	 * cannot block, mmu_notifier_ops.flags should have
	 * MMU_INVALIDATE_DOES_NOT_BLOCK set.
	 */
	void (*invalidate_range_start)(struct mmu_notifier *mn,
				       struct mm_struct *mm,
				       unsigned long start, unsigned long end);
	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.txt
	 *
	 * 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.
	 *
	 * If this callback cannot block, and invalidate_range_{start,end}
	 * cannot block, mmu_notifier_ops.flags should have
	 * MMU_INVALIDATE_DOES_NOT_BLOCK set.
	 */
	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 void __mmu_notifier_invalidate_range_start(struct mm_struct *mm,
				  unsigned long start, unsigned long end);
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);
extern bool mm_has_blockable_invalidate_notifiers(struct mm_struct *mm);

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);
}

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 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 bool mm_has_blockable_invalidate_notifiers(struct mm_struct *mm)
{
	return false;
}

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
5ff7091f	5	#include <linux/types.h>
cddb8a5c AA	6	#include <linux/list.h>
	7	#include <linux/spinlock.h>
	8	#include <linux/mm_types.h>
21a92735	9	#include <linux/srcu.h>
cddb8a5c AA	10
	11	struct mmu_notifier;
	12	struct mmu_notifier_ops;
	13
5ff7091f DR	14	/* mmu_notifier_ops flags */
	15	#define MMU_INVALIDATE_DOES_NOT_BLOCK (0x01)
	16
cddb8a5c AA	17	#ifdef CONFIG_MMU_NOTIFIER
	18
	19	/*
	20	* The mmu notifier_mm structure is allocated and installed in
	21	* mm->mmu_notifier_mm inside the mm_take_all_locks() protected
	22	* critical section and it's released only when mm_count reaches zero
	23	* in mmdrop().
	24	*/
	25	struct mmu_notifier_mm {
	26	/* all mmu notifiers registerd in this mm are queued in this list */
	27	struct hlist_head list;
	28	/* to serialize the list modifications and hlist_unhashed */
	29	spinlock_t lock;
	30	};
	31
	32	struct mmu_notifier_ops {
5ff7091f DR	33	/*
	34	* Flags to specify behavior of callbacks for this MMU notifier.
	35	* Used to determine which context an operation may be called.
	36	*
	37	* MMU_INVALIDATE_DOES_NOT_BLOCK: invalidate_range_* callbacks do not
	38	* block
	39	*/
	40	int flags;
	41
cddb8a5c AA	42	/*
	43	* Called either by mmu_notifier_unregister or when the mm is
	44	* being destroyed by exit_mmap, always before all pages are
	45	* freed. This can run concurrently with other mmu notifier
	46	* methods (the ones invoked outside the mm context) and it
	47	* should tear down all secondary mmu mappings and freeze the
	48	* secondary mmu. If this method isn't implemented you've to
	49	* be sure that nothing could possibly write to the pages
	50	* through the secondary mmu by the time the last thread with
	51	* tsk->mm == mm exits.
	52	*
	53	* As side note: the pages freed after ->release returns could
	54	* be immediately reallocated by the gart at an alias physical
	55	* address with a different cache model, so if ->release isn't
	56	* implemented because all _software_ driven memory accesses
	57	* through the secondary mmu are terminated by the time the
	58	* last thread of this mm quits, you've also to be sure that
	59	* speculative _hardware_ operations can't allocate dirty
	60	* cachelines in the cpu that could not be snooped and made
	61	* coherent with the other read and write operations happening
	62	* through the gart alias address, so leading to memory
	63	* corruption.
	64	*/
	65	void (release)(struct mmu_notifier mn,
	66	struct mm_struct *mm);
	67
	68	/*
	69	* clear_flush_young is called after the VM is
	70	* test-and-clearing the young/accessed bitflag in the
	71	* pte. This way the VM will provide proper aging to the
	72	* accesses to the page through the secondary MMUs and not
	73	* only to the ones through the Linux pte.
57128468 ALC	74	* Start-end is necessary in case the secondary MMU is mapping the page
57128468 ALC	75	* at a smaller granularity than the primary MMU.
cddb8a5c AA	76	*/
	77	int (clear_flush_young)(struct mmu_notifier mn,
	78	struct mm_struct *mm,
57128468 ALC	79	unsigned long start,
57128468 ALC	80	unsigned long end);
cddb8a5c	81
1d7715c6 VD	82	/*
	83	* clear_young is a lightweight version of clear_flush_young. Like the
	84	* latter, it is supposed to test-and-clear the young/accessed bitflag
	85	* in the secondary pte, but it may omit flushing the secondary tlb.
	86	*/
	87	int (clear_young)(struct mmu_notifier mn,
	88	struct mm_struct *mm,
	89	unsigned long start,
	90	unsigned long end);
	91
8ee53820 AA	92	/*
	93	* test_young is called to check the young/accessed bitflag in
	94	* the secondary pte. This is used to know if the page is
	95	* frequently used without actually clearing the flag or tearing
	96	* down the secondary mapping on the page.
	97	*/
	98	int (test_young)(struct mmu_notifier mn,
	99	struct mm_struct *mm,
	100	unsigned long address);
	101
828502d3 IE	102	/*
	103	* change_pte is called in cases that pte mapping to page is changed:
	104	* for example, when ksm remaps pte to point to a new shared page.
	105	*/
	106	void (change_pte)(struct mmu_notifier mn,
	107	struct mm_struct *mm,
	108	unsigned long address,
	109	pte_t pte);
	110
cddb8a5c AA	111	/*
	112	* invalidate_range_start() and invalidate_range_end() must be
	113	* paired and are called only when the mmap_sem and/or the
0f0a327f JR	114	* locks protecting the reverse maps are held. If the subsystem
	115	* can't guarantee that no additional references are taken to
	116	* the pages in the range, it has to implement the
	117	* invalidate_range() notifier to remove any references taken
	118	* after invalidate_range_start().
cddb8a5c AA	119	*
	120	* Invalidation of multiple concurrent ranges may be
	121	* optionally permitted by the driver. Either way the
	122	* establishment of sptes is forbidden in the range passed to
	123	* invalidate_range_begin/end for the whole duration of the
	124	* invalidate_range_begin/end critical section.
	125	*
	126	* invalidate_range_start() is called when all pages in the
	127	* range are still mapped and have at least a refcount of one.
	128	*
	129	* invalidate_range_end() is called when all pages in the
	130	* range have been unmapped and the pages have been freed by
	131	* the VM.
	132	*
	133	* The VM will remove the page table entries and potentially
	134	* the page between invalidate_range_start() and
	135	* invalidate_range_end(). If the page must not be freed
	136	* because of pending I/O or other circumstances then the
	137	* invalidate_range_start() callback (or the initial mapping
	138	* by the driver) must make sure that the refcount is kept
	139	* elevated.
	140	*
	141	* If the driver increases the refcount when the pages are
	142	* initially mapped into an address space then either
	143	* invalidate_range_start() or invalidate_range_end() may
	144	* decrease the refcount. If the refcount is decreased on
	145	* invalidate_range_start() then the VM can free pages as page
	146	* table entries are removed. If the refcount is only
	147	* droppped on invalidate_range_end() then the driver itself
	148	* will drop the last refcount but it must take care to flush
	149	* any secondary tlb before doing the final free on the
	150	* page. Pages will no longer be referenced by the linux
	151	* address space but may still be referenced by sptes until
	152	* the last refcount is dropped.
5ff7091f DR	153	*
	154	* If both of these callbacks cannot block, and invalidate_range
	155	* cannot block, mmu_notifier_ops.flags should have
	156	* MMU_INVALIDATE_DOES_NOT_BLOCK set.
cddb8a5c AA	157	*/
	158	void (invalidate_range_start)(struct mmu_notifier mn,
	159	struct mm_struct *mm,
	160	unsigned long start, unsigned long end);
	161	void (invalidate_range_end)(struct mmu_notifier mn,
	162	struct mm_struct *mm,
	163	unsigned long start, unsigned long end);
0f0a327f JR	164
	165	/*
	166	* invalidate_range() is either called between
	167	* invalidate_range_start() and invalidate_range_end() when the
	168	* VM has to free pages that where unmapped, but before the
	169	* pages are actually freed, or outside of _start()/_end() when
	170	* a (remote) TLB is necessary.
	171	*
	172	* If invalidate_range() is used to manage a non-CPU TLB with
	173	* shared page-tables, it not necessary to implement the
	174	* invalidate_range_start()/end() notifiers, as
	175	* invalidate_range() alread catches the points in time when an
0f10851e JG	176	* external TLB range needs to be flushed. For more in depth
0f10851e JG	177	* discussion on this see Documentation/vm/mmu_notifier.txt
0f0a327f	178	*
0f0a327f JR	179	* Note that this function might be called with just a sub-range
	180	* of what was passed to invalidate_range_start()/end(), if
	181	* called between those functions.
5ff7091f DR	182	*
	183	* If this callback cannot block, and invalidate_range_{start,end}
	184	* cannot block, mmu_notifier_ops.flags should have
	185	* MMU_INVALIDATE_DOES_NOT_BLOCK set.
0f0a327f JR	186	*/
	187	void (invalidate_range)(struct mmu_notifier mn, struct mm_struct *mm,
	188	unsigned long start, unsigned long end);
cddb8a5c AA	189	};
	190
	191	/*
	192	* The notifier chains are protected by mmap_sem and/or the reverse map
	193	* semaphores. Notifier chains are only changed when all reverse maps and
	194	* the mmap_sem locks are taken.
	195	*
	196	* Therefore notifier chains can only be traversed when either
	197	*
	198	* 1. mmap_sem is held.
c8c06efa	199	* 2. One of the reverse map locks is held (i_mmap_rwsem or anon_vma->rwsem).
cddb8a5c AA	200	* 3. No other concurrent thread can access the list (release)
	201	*/
	202	struct mmu_notifier {
	203	struct hlist_node hlist;
	204	const struct mmu_notifier_ops *ops;
	205	};
	206
	207	static inline int mm_has_notifiers(struct mm_struct *mm)
	208	{
	209	return unlikely(mm->mmu_notifier_mm);
	210	}
	211
	212	extern int mmu_notifier_register(struct mmu_notifier *mn,
	213	struct mm_struct *mm);
	214	extern int __mmu_notifier_register(struct mmu_notifier *mn,
	215	struct mm_struct *mm);
	216	extern void mmu_notifier_unregister(struct mmu_notifier *mn,
	217	struct mm_struct *mm);
b972216e PZ	218	extern void mmu_notifier_unregister_no_release(struct mmu_notifier *mn,
b972216e PZ	219	struct mm_struct *mm);
cddb8a5c AA	220	extern void __mmu_notifier_mm_destroy(struct mm_struct *mm);
	221	extern void __mmu_notifier_release(struct mm_struct *mm);
	222	extern int __mmu_notifier_clear_flush_young(struct mm_struct *mm,
57128468 ALC	223	unsigned long start,
57128468 ALC	224	unsigned long end);
1d7715c6 VD	225	extern int __mmu_notifier_clear_young(struct mm_struct *mm,
	226	unsigned long start,
	227	unsigned long end);
8ee53820 AA	228	extern int __mmu_notifier_test_young(struct mm_struct *mm,
8ee53820 AA	229	unsigned long address);
828502d3 IE	230	extern void __mmu_notifier_change_pte(struct mm_struct *mm,
828502d3 IE	231	unsigned long address, pte_t pte);
cddb8a5c AA	232	extern void __mmu_notifier_invalidate_range_start(struct mm_struct *mm,
	233	unsigned long start, unsigned long end);
	234	extern void __mmu_notifier_invalidate_range_end(struct mm_struct *mm,
4645b9fe JG	235	unsigned long start, unsigned long end,
4645b9fe JG	236	bool only_end);
0f0a327f JR	237	extern void __mmu_notifier_invalidate_range(struct mm_struct *mm,
0f0a327f JR	238	unsigned long start, unsigned long end);
5ff7091f	239	extern bool mm_has_blockable_invalidate_notifiers(struct mm_struct *mm);
cddb8a5c AA	240
	241	static inline void mmu_notifier_release(struct mm_struct *mm)
	242	{
	243	if (mm_has_notifiers(mm))
	244	__mmu_notifier_release(mm);
	245	}
	246
	247	static inline int mmu_notifier_clear_flush_young(struct mm_struct *mm,
57128468 ALC	248	unsigned long start,
57128468 ALC	249	unsigned long end)
cddb8a5c AA	250	{
cddb8a5c AA	251	if (mm_has_notifiers(mm))
57128468	252	return __mmu_notifier_clear_flush_young(mm, start, end);
cddb8a5c AA	253	return 0;
	254	}
	255
1d7715c6 VD	256	static inline int mmu_notifier_clear_young(struct mm_struct *mm,
	257	unsigned long start,
	258	unsigned long end)
	259	{
	260	if (mm_has_notifiers(mm))
	261	return __mmu_notifier_clear_young(mm, start, end);
	262	return 0;
	263	}
	264
8ee53820 AA	265	static inline int mmu_notifier_test_young(struct mm_struct *mm,
	266	unsigned long address)
	267	{
	268	if (mm_has_notifiers(mm))
	269	return __mmu_notifier_test_young(mm, address);
	270	return 0;
	271	}
	272
828502d3 IE	273	static inline void mmu_notifier_change_pte(struct mm_struct *mm,
	274	unsigned long address, pte_t pte)
	275	{
	276	if (mm_has_notifiers(mm))
	277	__mmu_notifier_change_pte(mm, address, pte);
	278	}
	279
cddb8a5c AA	280	static inline void mmu_notifier_invalidate_range_start(struct mm_struct *mm,
	281	unsigned long start, unsigned long end)
	282	{
	283	if (mm_has_notifiers(mm))
	284	__mmu_notifier_invalidate_range_start(mm, start, end);
	285	}
	286
	287	static inline void mmu_notifier_invalidate_range_end(struct mm_struct *mm,
	288	unsigned long start, unsigned long end)
	289	{
	290	if (mm_has_notifiers(mm))
4645b9fe JG	291	__mmu_notifier_invalidate_range_end(mm, start, end, false);
	292	}
	293
	294	static inline void mmu_notifier_invalidate_range_only_end(struct mm_struct *mm,
	295	unsigned long start, unsigned long end)
	296	{
	297	if (mm_has_notifiers(mm))
	298	__mmu_notifier_invalidate_range_end(mm, start, end, true);
cddb8a5c AA	299	}
cddb8a5c AA	300
1897bdc4 JR	301	static inline void mmu_notifier_invalidate_range(struct mm_struct *mm,
	302	unsigned long start, unsigned long end)
	303	{
0f0a327f JR	304	if (mm_has_notifiers(mm))
0f0a327f JR	305	__mmu_notifier_invalidate_range(mm, start, end);
1897bdc4 JR	306	}
1897bdc4 JR	307
cddb8a5c AA	308	static inline void mmu_notifier_mm_init(struct mm_struct *mm)
	309	{
	310	mm->mmu_notifier_mm = NULL;
	311	}
	312
	313	static inline void mmu_notifier_mm_destroy(struct mm_struct *mm)
	314	{
	315	if (mm_has_notifiers(mm))
	316	__mmu_notifier_mm_destroy(mm);
	317	}
	318
cddb8a5c AA	319	#define ptep_clear_flush_young_notify(__vma, __address, __ptep) \
	320	({ \
	321	int __young; \
	322	struct vm_area_struct *___vma = __vma; \
	323	unsigned long ___address = __address; \
	324	__young = ptep_clear_flush_young(___vma, ___address, __ptep); \
	325	__young \|= mmu_notifier_clear_flush_young(___vma->vm_mm, \
57128468 ALC	326	___address, \
	327	___address + \
	328	PAGE_SIZE); \
cddb8a5c AA	329	__young; \
	330	})
	331
91a4ee26 AA	332	#define pmdp_clear_flush_young_notify(__vma, __address, __pmdp) \
	333	({ \
	334	int __young; \
	335	struct vm_area_struct *___vma = __vma; \
	336	unsigned long ___address = __address; \
	337	__young = pmdp_clear_flush_young(___vma, ___address, __pmdp); \
	338	__young \|= mmu_notifier_clear_flush_young(___vma->vm_mm, \
57128468 ALC	339	___address, \
	340	___address + \
	341	PMD_SIZE); \
91a4ee26 AA	342	__young; \
	343	})
	344
1d7715c6 VD	345	#define ptep_clear_young_notify(__vma, __address, __ptep) \
	346	({ \
	347	int __young; \
	348	struct vm_area_struct *___vma = __vma; \
	349	unsigned long ___address = __address; \
	350	__young = ptep_test_and_clear_young(___vma, ___address, __ptep);\
	351	__young \|= mmu_notifier_clear_young(___vma->vm_mm, ___address, \
	352	___address + PAGE_SIZE); \
	353	__young; \
	354	})
	355
	356	#define pmdp_clear_young_notify(__vma, __address, __pmdp) \
	357	({ \
	358	int __young; \
	359	struct vm_area_struct *___vma = __vma; \
	360	unsigned long ___address = __address; \
	361	__young = pmdp_test_and_clear_young(___vma, ___address, __pmdp);\
	362	__young \|= mmu_notifier_clear_young(___vma->vm_mm, ___address, \
	363	___address + PMD_SIZE); \
	364	__young; \
	365	})
	366
34ee645e JR	367	#define ptep_clear_flush_notify(__vma, __address, __ptep) \
	368	({ \
	369	unsigned long ___addr = __address & PAGE_MASK; \
	370	struct mm_struct *___mm = (__vma)->vm_mm; \
	371	pte_t ___pte; \
	372	\
	373	___pte = ptep_clear_flush(__vma, __address, __ptep); \
	374	mmu_notifier_invalidate_range(___mm, ___addr, \
	375	___addr + PAGE_SIZE); \
	376	\
	377	___pte; \
	378	})
	379
8809aa2d	380	#define pmdp_huge_clear_flush_notify(__vma, __haddr, __pmd) \
34ee645e JR	381	({ \
	382	unsigned long ___haddr = __haddr & HPAGE_PMD_MASK; \
	383	struct mm_struct *___mm = (__vma)->vm_mm; \
	384	pmd_t ___pmd; \
	385	\
8809aa2d	386	___pmd = pmdp_huge_clear_flush(__vma, __haddr, __pmd); \
34ee645e JR	387	mmu_notifier_invalidate_range(___mm, ___haddr, \
	388	___haddr + HPAGE_PMD_SIZE); \
	389	\
	390	___pmd; \
	391	})
	392
a00cc7d9 MW	393	#define pudp_huge_clear_flush_notify(__vma, __haddr, __pud) \
	394	({ \
	395	unsigned long ___haddr = __haddr & HPAGE_PUD_MASK; \
	396	struct mm_struct *___mm = (__vma)->vm_mm; \
	397	pud_t ___pud; \
	398	\
	399	___pud = pudp_huge_clear_flush(__vma, __haddr, __pud); \
	400	mmu_notifier_invalidate_range(___mm, ___haddr, \
	401	___haddr + HPAGE_PUD_SIZE); \
	402	\
	403	___pud; \
	404	})
	405
48af0d7c XG	406	/*
	407	* set_pte_at_notify() sets the pte _after_ running the notifier.
	408	* This is safe to start by updating the secondary MMUs, because the primary MMU
	409	* pte invalidate must have already happened with a ptep_clear_flush() before
	410	* set_pte_at_notify() has been invoked. Updating the secondary MMUs first is
	411	* required when we change both the protection of the mapping from read-only to
	412	* read-write and the pfn (like during copy on write page faults). Otherwise the
	413	* old page would remain mapped readonly in the secondary MMUs after the new
	414	* page is already writable by some CPU through the primary MMU.
	415	*/
828502d3 IE	416	#define set_pte_at_notify(__mm, __address, __ptep, __pte) \
	417	({ \
	418	struct mm_struct *___mm = __mm; \
	419	unsigned long ___address = __address; \
	420	pte_t ___pte = __pte; \
	421	\
828502d3	422	mmu_notifier_change_pte(___mm, ___address, ___pte); \
48af0d7c	423	set_pte_at(___mm, ___address, __ptep, ___pte); \
828502d3 IE	424	})
828502d3 IE	425
b972216e PZ	426	extern void mmu_notifier_call_srcu(struct rcu_head *rcu,
	427	void (func)(struct rcu_head rcu));
	428	extern void mmu_notifier_synchronize(void);
	429
cddb8a5c AA	430	#else /* CONFIG_MMU_NOTIFIER */
cddb8a5c AA	431
4d4bbd85 MH	432	static inline int mm_has_notifiers(struct mm_struct *mm)
	433	{
	434	return 0;
	435	}
	436
cddb8a5c AA	437	static inline void mmu_notifier_release(struct mm_struct *mm)
	438	{
	439	}
	440
	441	static inline int mmu_notifier_clear_flush_young(struct mm_struct *mm,
57128468 ALC	442	unsigned long start,
57128468 ALC	443	unsigned long end)
8ee53820 AA	444	{
	445	return 0;
	446	}
	447
	448	static inline int mmu_notifier_test_young(struct mm_struct *mm,
	449	unsigned long address)
cddb8a5c AA	450	{
	451	return 0;
	452	}
	453
828502d3 IE	454	static inline void mmu_notifier_change_pte(struct mm_struct *mm,
	455	unsigned long address, pte_t pte)
	456	{
	457	}
	458
cddb8a5c AA	459	static inline void mmu_notifier_invalidate_range_start(struct mm_struct *mm,
	460	unsigned long start, unsigned long end)
	461	{
	462	}
	463
	464	static inline void mmu_notifier_invalidate_range_end(struct mm_struct *mm,
	465	unsigned long start, unsigned long end)
	466	{
	467	}
	468
4645b9fe JG	469	static inline void mmu_notifier_invalidate_range_only_end(struct mm_struct *mm,
	470	unsigned long start, unsigned long end)
	471	{
	472	}
	473
1897bdc4 JR	474	static inline void mmu_notifier_invalidate_range(struct mm_struct *mm,
	475	unsigned long start, unsigned long end)
	476	{
	477	}
	478
5ff7091f DR	479	static inline bool mm_has_blockable_invalidate_notifiers(struct mm_struct *mm)
	480	{
	481	return false;
	482	}
	483
cddb8a5c AA	484	static inline void mmu_notifier_mm_init(struct mm_struct *mm)
	485	{
	486	}
	487
	488	static inline void mmu_notifier_mm_destroy(struct mm_struct *mm)
	489	{
	490	}
	491
	492	#define ptep_clear_flush_young_notify ptep_clear_flush_young
91a4ee26	493	#define pmdp_clear_flush_young_notify pmdp_clear_flush_young
33c3fc71 VD	494	#define ptep_clear_young_notify ptep_test_and_clear_young
33c3fc71 VD	495	#define pmdp_clear_young_notify pmdp_test_and_clear_young
34ee645e	496	#define ptep_clear_flush_notify ptep_clear_flush
8809aa2d	497	#define pmdp_huge_clear_flush_notify pmdp_huge_clear_flush
a00cc7d9	498	#define pudp_huge_clear_flush_notify pudp_huge_clear_flush
828502d3	499	#define set_pte_at_notify set_pte_at
cddb8a5c AA	500
	501	#endif /* CONFIG_MMU_NOTIFIER */
	502
	503	#endif /* _LINUX_MMU_NOTIFIER_H */