[thirdparty/kernel/stable.git] / mm / page_vma_mapped.c

// SPDX-License-Identifier: GPL-2.0
#include <linux/mm.h>
#include <linux/rmap.h>
#include <linux/hugetlb.h>
#include <linux/swap.h>
#include <linux/swapops.h>

#include "internal.h"

static inline bool not_found(struct page_vma_mapped_walk *pvmw)
{
	page_vma_mapped_walk_done(pvmw);
	return false;
}

static bool map_pte(struct page_vma_mapped_walk *pvmw)
{
	pvmw->pte = pte_offset_map(pvmw->pmd, pvmw->address);
	if (!(pvmw->flags & PVMW_SYNC)) {
		if (pvmw->flags & PVMW_MIGRATION) {
			if (!is_swap_pte(*pvmw->pte))
				return false;
		} else {
			/*
			 * We get here when we are trying to unmap a private
			 * device page from the process address space. Such
			 * page is not CPU accessible and thus is mapped as
			 * a special swap entry, nonetheless it still does
			 * count as a valid regular mapping for the page (and
			 * is accounted as such in page maps count).
			 *
			 * So handle this special case as if it was a normal
			 * page mapping ie lock CPU page table and returns
			 * true.
			 *
			 * For more details on device private memory see HMM
			 * (include/linux/hmm.h or mm/hmm.c).
			 */
			if (is_swap_pte(*pvmw->pte)) {
				swp_entry_t entry;

				/* Handle un-addressable ZONE_DEVICE memory */
				entry = pte_to_swp_entry(*pvmw->pte);
				if (!is_device_private_entry(entry))
					return false;
			} else if (!pte_present(*pvmw->pte))
				return false;
		}
	}
	pvmw->ptl = pte_lockptr(pvmw->vma->vm_mm, pvmw->pmd);
	spin_lock(pvmw->ptl);
	return true;
}

static inline bool pfn_in_hpage(struct page *hpage, unsigned long pfn)
{
	unsigned long hpage_pfn = page_to_pfn(hpage);

	/* THP can be referenced by any subpage */
	return pfn >= hpage_pfn && pfn - hpage_pfn < hpage_nr_pages(hpage);
}

/**
 * check_pte - check if @pvmw->page is mapped at the @pvmw->pte
 *
 * page_vma_mapped_walk() found a place where @pvmw->page is *potentially*
 * mapped. check_pte() has to validate this.
 *
 * @pvmw->pte may point to empty PTE, swap PTE or PTE pointing to arbitrary
 * page.
 *
 * If PVMW_MIGRATION flag is set, returns true if @pvmw->pte contains migration
 * entry that points to @pvmw->page or any subpage in case of THP.
 *
 * If PVMW_MIGRATION flag is not set, returns true if @pvmw->pte points to
 * @pvmw->page or any subpage in case of THP.
 *
 * Otherwise, return false.
 *
 */
static bool check_pte(struct page_vma_mapped_walk *pvmw)
{
	unsigned long pfn;

	if (pvmw->flags & PVMW_MIGRATION) {
		swp_entry_t entry;
		if (!is_swap_pte(*pvmw->pte))
			return false;
		entry = pte_to_swp_entry(*pvmw->pte);

		if (!is_migration_entry(entry))
			return false;

		pfn = migration_entry_to_pfn(entry);
	} else if (is_swap_pte(*pvmw->pte)) {
		swp_entry_t entry;

		/* Handle un-addressable ZONE_DEVICE memory */
		entry = pte_to_swp_entry(*pvmw->pte);
		if (!is_device_private_entry(entry))
			return false;

		pfn = device_private_entry_to_pfn(entry);
	} else {
		if (!pte_present(*pvmw->pte))
			return false;

		pfn = pte_pfn(*pvmw->pte);
	}

	return pfn_in_hpage(pvmw->page, pfn);
}

/**
 * page_vma_mapped_walk - check if @pvmw->page is mapped in @pvmw->vma at
 * @pvmw->address
 * @pvmw: pointer to struct page_vma_mapped_walk. page, vma, address and flags
 * must be set. pmd, pte and ptl must be NULL.
 *
 * Returns true if the page is mapped in the vma. @pvmw->pmd and @pvmw->pte point
 * to relevant page table entries. @pvmw->ptl is locked. @pvmw->address is
 * adjusted if needed (for PTE-mapped THPs).
 *
 * If @pvmw->pmd is set but @pvmw->pte is not, you have found PMD-mapped page
 * (usually THP). For PTE-mapped THP, you should run page_vma_mapped_walk() in
 * a loop to find all PTEs that map the THP.
 *
 * For HugeTLB pages, @pvmw->pte is set to the relevant page table entry
 * regardless of which page table level the page is mapped at. @pvmw->pmd is
 * NULL.
 *
 * Retruns false if there are no more page table entries for the page in
 * the vma. @pvmw->ptl is unlocked and @pvmw->pte is unmapped.
 *
 * If you need to stop the walk before page_vma_mapped_walk() returned false,
 * use page_vma_mapped_walk_done(). It will do the housekeeping.
 */
bool page_vma_mapped_walk(struct page_vma_mapped_walk *pvmw)
{
	struct mm_struct *mm = pvmw->vma->vm_mm;
	struct page *page = pvmw->page;
	pgd_t *pgd;
	p4d_t *p4d;
	pud_t *pud;
	pmd_t pmde;

	/* The only possible pmd mapping has been handled on last iteration */
	if (pvmw->pmd && !pvmw->pte)
		return not_found(pvmw);

	if (pvmw->pte)
		goto next_pte;

	if (unlikely(PageHuge(pvmw->page))) {
		/* when pud is not present, pte will be NULL */
		pvmw->pte = huge_pte_offset(mm, pvmw->address,
					    PAGE_SIZE << compound_order(page));
		if (!pvmw->pte)
			return false;

		pvmw->ptl = huge_pte_lockptr(page_hstate(page), mm, pvmw->pte);
		spin_lock(pvmw->ptl);
		if (!check_pte(pvmw))
			return not_found(pvmw);
		return true;
	}
restart:
	pgd = pgd_offset(mm, pvmw->address);
	if (!pgd_present(*pgd))
		return false;
	p4d = p4d_offset(pgd, pvmw->address);
	if (!p4d_present(*p4d))
		return false;
	pud = pud_offset(p4d, pvmw->address);
	if (!pud_present(*pud))
		return false;
	pvmw->pmd = pmd_offset(pud, pvmw->address);
	/*
	 * Make sure the pmd value isn't cached in a register by the
	 * compiler and used as a stale value after we've observed a
	 * subsequent update.
	 */
	pmde = READ_ONCE(*pvmw->pmd);
	if (pmd_trans_huge(pmde) || is_pmd_migration_entry(pmde)) {
		pvmw->ptl = pmd_lock(mm, pvmw->pmd);
		if (likely(pmd_trans_huge(*pvmw->pmd))) {
			if (pvmw->flags & PVMW_MIGRATION)
				return not_found(pvmw);
			if (pmd_page(*pvmw->pmd) != page)
				return not_found(pvmw);
			return true;
		} else if (!pmd_present(*pvmw->pmd)) {
			if (thp_migration_supported()) {
				if (!(pvmw->flags & PVMW_MIGRATION))
					return not_found(pvmw);
				if (is_migration_entry(pmd_to_swp_entry(*pvmw->pmd))) {
					swp_entry_t entry = pmd_to_swp_entry(*pvmw->pmd);

					if (migration_entry_to_page(entry) != page)
						return not_found(pvmw);
					return true;
				}
			}
			return not_found(pvmw);
		} else {
			/* THP pmd was split under us: handle on pte level */
			spin_unlock(pvmw->ptl);
			pvmw->ptl = NULL;
		}
	} else if (!pmd_present(pmde)) {
		return false;
	}
	if (!map_pte(pvmw))
		goto next_pte;
	while (1) {
		if (check_pte(pvmw))
			return true;
next_pte:
		/* Seek to next pte only makes sense for THP */
		if (!PageTransHuge(pvmw->page) || PageHuge(pvmw->page))
			return not_found(pvmw);
		do {
			pvmw->address += PAGE_SIZE;
			if (pvmw->address >= pvmw->vma->vm_end ||
			    pvmw->address >=
					__vma_address(pvmw->page, pvmw->vma) +
					hpage_nr_pages(pvmw->page) * PAGE_SIZE)
				return not_found(pvmw);
			/* Did we cross page table boundary? */
			if (pvmw->address % PMD_SIZE == 0) {
				pte_unmap(pvmw->pte);
				if (pvmw->ptl) {
					spin_unlock(pvmw->ptl);
					pvmw->ptl = NULL;
				}
				goto restart;
			} else {
				pvmw->pte++;
			}
		} while (pte_none(*pvmw->pte));

		if (!pvmw->ptl) {
			pvmw->ptl = pte_lockptr(mm, pvmw->pmd);
			spin_lock(pvmw->ptl);
		}
	}
}

/**
 * page_mapped_in_vma - check whether a page is really mapped in a VMA
 * @page: the page to test
 * @vma: the VMA to test
 *
 * Returns 1 if the page is mapped into the page tables of the VMA, 0
 * if the page is not mapped into the page tables of this VMA.  Only
 * valid for normal file or anonymous VMAs.
 */
int page_mapped_in_vma(struct page *page, struct vm_area_struct *vma)
{
	struct page_vma_mapped_walk pvmw = {
		.page = page,
		.vma = vma,
		.flags = PVMW_SYNC,
	};
	unsigned long start, end;

	start = __vma_address(page, vma);
	end = start + PAGE_SIZE * (hpage_nr_pages(page) - 1);

	if (unlikely(end < vma->vm_start || start >= vma->vm_end))
		return 0;
	pvmw.address = max(start, vma->vm_start);
	if (!page_vma_mapped_walk(&pvmw))
		return 0;
	page_vma_mapped_walk_done(&pvmw);
	return 1;
}
Commit	Line	Data
b2441318	1	// SPDX-License-Identifier: GPL-2.0
ace71a19 KS	2	#include <linux/mm.h>
	3	#include <linux/rmap.h>
	4	#include <linux/hugetlb.h>
	5	#include <linux/swap.h>
	6	#include <linux/swapops.h>
	7
	8	#include "internal.h"
	9
ace71a19 KS	10	static inline bool not_found(struct page_vma_mapped_walk *pvmw)
	11	{
	12	page_vma_mapped_walk_done(pvmw);
	13	return false;
	14	}
	15
	16	static bool map_pte(struct page_vma_mapped_walk *pvmw)
	17	{
	18	pvmw->pte = pte_offset_map(pvmw->pmd, pvmw->address);
	19	if (!(pvmw->flags & PVMW_SYNC)) {
	20	if (pvmw->flags & PVMW_MIGRATION) {
	21	if (!is_swap_pte(*pvmw->pte))
	22	return false;
	23	} else {
aab8d052 RC	24	/*
	25	* We get here when we are trying to unmap a private
	26	* device page from the process address space. Such
	27	* page is not CPU accessible and thus is mapped as
	28	* a special swap entry, nonetheless it still does
	29	* count as a valid regular mapping for the page (and
	30	* is accounted as such in page maps count).
	31	*
	32	* So handle this special case as if it was a normal
	33	* page mapping ie lock CPU page table and returns
	34	* true.
	35	*
	36	* For more details on device private memory see HMM
	37	* (include/linux/hmm.h or mm/hmm.c).
	38	*/
	39	if (is_swap_pte(*pvmw->pte)) {
	40	swp_entry_t entry;
	41
	42	/* Handle un-addressable ZONE_DEVICE memory */
	43	entry = pte_to_swp_entry(*pvmw->pte);
	44	if (!is_device_private_entry(entry))
	45	return false;
	46	} else if (!pte_present(*pvmw->pte))
ace71a19 KS	47	return false;
	48	}
	49	}
	50	pvmw->ptl = pte_lockptr(pvmw->vma->vm_mm, pvmw->pmd);
	51	spin_lock(pvmw->ptl);
	52	return true;
	53	}
	54
7222708e KS	55	static inline bool pfn_in_hpage(struct page *hpage, unsigned long pfn)
	56	{
	57	unsigned long hpage_pfn = page_to_pfn(hpage);
	58
	59	/* THP can be referenced by any subpage */
	60	return pfn >= hpage_pfn && pfn - hpage_pfn < hpage_nr_pages(hpage);
	61	}
	62
0d665e7b KS	63	/**
	64	* check_pte - check if @pvmw->page is mapped at the @pvmw->pte
	65	*
	66	* page_vma_mapped_walk() found a place where @pvmw->page is potentially
	67	* mapped. check_pte() has to validate this.
	68	*
	69	* @pvmw->pte may point to empty PTE, swap PTE or PTE pointing to arbitrary
	70	* page.
	71	*
	72	* If PVMW_MIGRATION flag is set, returns true if @pvmw->pte contains migration
	73	* entry that points to @pvmw->page or any subpage in case of THP.
	74	*
	75	* If PVMW_MIGRATION flag is not set, returns true if @pvmw->pte points to
	76	* @pvmw->page or any subpage in case of THP.
	77	*
	78	* Otherwise, return false.
	79	*
	80	*/
ace71a19 KS	81	static bool check_pte(struct page_vma_mapped_walk *pvmw)
ace71a19 KS	82	{
0d665e7b KS	83	unsigned long pfn;
0d665e7b KS	84
ace71a19	85	if (pvmw->flags & PVMW_MIGRATION) {
ace71a19 KS	86	swp_entry_t entry;
	87	if (!is_swap_pte(*pvmw->pte))
	88	return false;
	89	entry = pte_to_swp_entry(*pvmw->pte);
a5430dda	90
ace71a19 KS	91	if (!is_migration_entry(entry))
ace71a19 KS	92	return false;
a5430dda	93
0d665e7b KS	94	pfn = migration_entry_to_pfn(entry);
	95	} else if (is_swap_pte(*pvmw->pte)) {
	96	swp_entry_t entry;
a5430dda	97
0d665e7b KS	98	/* Handle un-addressable ZONE_DEVICE memory */
	99	entry = pte_to_swp_entry(*pvmw->pte);
	100	if (!is_device_private_entry(entry))
ace71a19 KS	101	return false;
ace71a19 KS	102
0d665e7b KS	103	pfn = device_private_entry_to_pfn(entry);
	104	} else {
	105	if (!pte_present(*pvmw->pte))
ace71a19	106	return false;
0d665e7b KS	107
0d665e7b KS	108	pfn = pte_pfn(*pvmw->pte);
ace71a19 KS	109	}
ace71a19 KS	110
7222708e	111	return pfn_in_hpage(pvmw->page, pfn);
ace71a19 KS	112	}
	113
	114	/**
	115	* page_vma_mapped_walk - check if @pvmw->page is mapped in @pvmw->vma at
	116	* @pvmw->address
	117	* @pvmw: pointer to struct page_vma_mapped_walk. page, vma, address and flags
	118	* must be set. pmd, pte and ptl must be NULL.
	119	*
	120	* Returns true if the page is mapped in the vma. @pvmw->pmd and @pvmw->pte point
	121	* to relevant page table entries. @pvmw->ptl is locked. @pvmw->address is
	122	* adjusted if needed (for PTE-mapped THPs).
	123	*
	124	* If @pvmw->pmd is set but @pvmw->pte is not, you have found PMD-mapped page
	125	* (usually THP). For PTE-mapped THP, you should run page_vma_mapped_walk() in
	126	* a loop to find all PTEs that map the THP.
	127	*
	128	* For HugeTLB pages, @pvmw->pte is set to the relevant page table entry
	129	* regardless of which page table level the page is mapped at. @pvmw->pmd is
	130	* NULL.
	131	*
	132	* Retruns false if there are no more page table entries for the page in
	133	* the vma. @pvmw->ptl is unlocked and @pvmw->pte is unmapped.
	134	*
	135	* If you need to stop the walk before page_vma_mapped_walk() returned false,
	136	* use page_vma_mapped_walk_done(). It will do the housekeeping.
	137	*/
	138	bool page_vma_mapped_walk(struct page_vma_mapped_walk *pvmw)
	139	{
	140	struct mm_struct *mm = pvmw->vma->vm_mm;
	141	struct page *page = pvmw->page;
	142	pgd_t *pgd;
c2febafc	143	p4d_t *p4d;
ace71a19	144	pud_t *pud;
a7b10095	145	pmd_t pmde;
ace71a19 KS	146
	147	/* The only possible pmd mapping has been handled on last iteration */
	148	if (pvmw->pmd && !pvmw->pte)
	149	return not_found(pvmw);
	150
d75450ff	151	if (pvmw->pte)
ace71a19	152	goto next_pte;
ace71a19 KS	153
	154	if (unlikely(PageHuge(pvmw->page))) {
	155	/* when pud is not present, pte will be NULL */
7868a208 PA	156	pvmw->pte = huge_pte_offset(mm, pvmw->address,
7868a208 PA	157	PAGE_SIZE << compound_order(page));
ace71a19 KS	158	if (!pvmw->pte)
	159	return false;
	160
	161	pvmw->ptl = huge_pte_lockptr(page_hstate(page), mm, pvmw->pte);
	162	spin_lock(pvmw->ptl);
	163	if (!check_pte(pvmw))
	164	return not_found(pvmw);
	165	return true;
	166	}
	167	restart:
	168	pgd = pgd_offset(mm, pvmw->address);
	169	if (!pgd_present(*pgd))
	170	return false;
c2febafc KS	171	p4d = p4d_offset(pgd, pvmw->address);
	172	if (!p4d_present(*p4d))
	173	return false;
	174	pud = pud_offset(p4d, pvmw->address);
ace71a19 KS	175	if (!pud_present(*pud))
	176	return false;
	177	pvmw->pmd = pmd_offset(pud, pvmw->address);
a7b10095 WD	178	/*
	179	* Make sure the pmd value isn't cached in a register by the
	180	* compiler and used as a stale value after we've observed a
	181	* subsequent update.
	182	*/
	183	pmde = READ_ONCE(*pvmw->pmd);
	184	if (pmd_trans_huge(pmde) \|\| is_pmd_migration_entry(pmde)) {
ace71a19	185	pvmw->ptl = pmd_lock(mm, pvmw->pmd);
ace71a19 KS	186	if (likely(pmd_trans_huge(*pvmw->pmd))) {
	187	if (pvmw->flags & PVMW_MIGRATION)
	188	return not_found(pvmw);
	189	if (pmd_page(*pvmw->pmd) != page)
	190	return not_found(pvmw);
	191	return true;
616b8371 ZY	192	} else if (!pmd_present(*pvmw->pmd)) {
	193	if (thp_migration_supported()) {
	194	if (!(pvmw->flags & PVMW_MIGRATION))
	195	return not_found(pvmw);
	196	if (is_migration_entry(pmd_to_swp_entry(*pvmw->pmd))) {
	197	swp_entry_t entry = pmd_to_swp_entry(*pvmw->pmd);
	198
	199	if (migration_entry_to_page(entry) != page)
	200	return not_found(pvmw);
	201	return true;
	202	}
af0db981	203	}
616b8371	204	return not_found(pvmw);
ace71a19 KS	205	} else {
	206	/* THP pmd was split under us: handle on pte level */
	207	spin_unlock(pvmw->ptl);
	208	pvmw->ptl = NULL;
	209	}
a7b10095 WD	210	} else if (!pmd_present(pmde)) {
a7b10095 WD	211	return false;
ace71a19 KS	212	}
	213	if (!map_pte(pvmw))
	214	goto next_pte;
	215	while (1) {
	216	if (check_pte(pvmw))
	217	return true;
d75450ff HD	218	next_pte:
	219	/* Seek to next pte only makes sense for THP */
	220	if (!PageTransHuge(pvmw->page) \|\| PageHuge(pvmw->page))
	221	return not_found(pvmw);
	222	do {
ace71a19	223	pvmw->address += PAGE_SIZE;
d75450ff HD	224	if (pvmw->address >= pvmw->vma->vm_end \|\|
d75450ff HD	225	pvmw->address >=
ace71a19 KS	226	__vma_address(pvmw->page, pvmw->vma) +
	227	hpage_nr_pages(pvmw->page) * PAGE_SIZE)
	228	return not_found(pvmw);
	229	/* Did we cross page table boundary? */
	230	if (pvmw->address % PMD_SIZE == 0) {
	231	pte_unmap(pvmw->pte);
	232	if (pvmw->ptl) {
	233	spin_unlock(pvmw->ptl);
	234	pvmw->ptl = NULL;
	235	}
	236	goto restart;
	237	} else {
	238	pvmw->pte++;
	239	}
	240	} while (pte_none(*pvmw->pte));
	241
	242	if (!pvmw->ptl) {
	243	pvmw->ptl = pte_lockptr(mm, pvmw->pmd);
	244	spin_lock(pvmw->ptl);
	245	}
	246	}
	247	}
6a328a62 KS	248
	249	/**
	250	* page_mapped_in_vma - check whether a page is really mapped in a VMA
	251	* @page: the page to test
	252	* @vma: the VMA to test
	253	*
	254	* Returns 1 if the page is mapped into the page tables of the VMA, 0
	255	* if the page is not mapped into the page tables of this VMA. Only
	256	* valid for normal file or anonymous VMAs.
	257	*/
	258	int page_mapped_in_vma(struct page page, struct vm_area_struct vma)
	259	{
	260	struct page_vma_mapped_walk pvmw = {
	261	.page = page,
	262	.vma = vma,
	263	.flags = PVMW_SYNC,
	264	};
	265	unsigned long start, end;
	266
	267	start = __vma_address(page, vma);
	268	end = start + PAGE_SIZE * (hpage_nr_pages(page) - 1);
	269
	270	if (unlikely(end < vma->vm_start \|\| start >= vma->vm_end))
	271	return 0;
	272	pvmw.address = max(start, vma->vm_start);
	273	if (!page_vma_mapped_walk(&pvmw))
	274	return 0;
	275	page_vma_mapped_walk_done(&pvmw);
	276	return 1;
	277	}