[mirror_ubuntu-jammy-kernel.git] / mm / mprotect.c

/*
 *  mm/mprotect.c
 *
 *  (C) Copyright 1994 Linus Torvalds
 *  (C) Copyright 2002 Christoph Hellwig
 *
 *  Address space accounting code	<alan@lxorguk.ukuu.org.uk>
 *  (C) Copyright 2002 Red Hat Inc, All Rights Reserved
 */

#include <linux/mm.h>
#include <linux/hugetlb.h>
#include <linux/shm.h>
#include <linux/mman.h>
#include <linux/fs.h>
#include <linux/highmem.h>
#include <linux/security.h>
#include <linux/mempolicy.h>
#include <linux/personality.h>
#include <linux/syscalls.h>
#include <linux/swap.h>
#include <linux/swapops.h>
#include <linux/mmu_notifier.h>
#include <linux/migrate.h>
#include <linux/perf_event.h>
#include <linux/ksm.h>
#include <linux/pkeys.h>
#include <asm/uaccess.h>
#include <asm/pgtable.h>
#include <asm/cacheflush.h>
#include <asm/tlbflush.h>

#include "internal.h"

/*
 * For a prot_numa update we only hold mmap_sem for read so there is a
 * potential race with faulting where a pmd was temporarily none. This
 * function checks for a transhuge pmd under the appropriate lock. It
 * returns a pte if it was successfully locked or NULL if it raced with
 * a transhuge insertion.
 */
static pte_t *lock_pte_protection(struct vm_area_struct *vma, pmd_t *pmd,
			unsigned long addr, int prot_numa, spinlock_t **ptl)
{
	pte_t *pte;
	spinlock_t *pmdl;

	/* !prot_numa is protected by mmap_sem held for write */
	if (!prot_numa)
		return pte_offset_map_lock(vma->vm_mm, pmd, addr, ptl);

	pmdl = pmd_lock(vma->vm_mm, pmd);
	if (unlikely(pmd_trans_huge(*pmd) || pmd_none(*pmd))) {
		spin_unlock(pmdl);
		return NULL;
	}

	pte = pte_offset_map_lock(vma->vm_mm, pmd, addr, ptl);
	spin_unlock(pmdl);
	return pte;
}

static unsigned long change_pte_range(struct vm_area_struct *vma, pmd_t *pmd,
		unsigned long addr, unsigned long end, pgprot_t newprot,
		int dirty_accountable, int prot_numa)
{
	struct mm_struct *mm = vma->vm_mm;
	pte_t *pte, oldpte;
	spinlock_t *ptl;
	unsigned long pages = 0;

	pte = lock_pte_protection(vma, pmd, addr, prot_numa, &ptl);
	if (!pte)
		return 0;

	arch_enter_lazy_mmu_mode();
	do {
		oldpte = *pte;
		if (pte_present(oldpte)) {
			pte_t ptent;
			bool preserve_write = prot_numa && pte_write(oldpte);

			/*
			 * Avoid trapping faults against the zero or KSM
			 * pages. See similar comment in change_huge_pmd.
			 */
			if (prot_numa) {
				struct page *page;

				page = vm_normal_page(vma, addr, oldpte);
				if (!page || PageKsm(page))
					continue;

				/* Avoid TLB flush if possible */
				if (pte_protnone(oldpte))
					continue;
			}

			ptent = ptep_modify_prot_start(mm, addr, pte);
			ptent = pte_modify(ptent, newprot);
			if (preserve_write)
				ptent = pte_mkwrite(ptent);

			/* Avoid taking write faults for known dirty pages */
			if (dirty_accountable && pte_dirty(ptent) &&
					(pte_soft_dirty(ptent) ||
					 !(vma->vm_flags & VM_SOFTDIRTY))) {
				ptent = pte_mkwrite(ptent);
			}
			ptep_modify_prot_commit(mm, addr, pte, ptent);
			pages++;
		} else if (IS_ENABLED(CONFIG_MIGRATION)) {
			swp_entry_t entry = pte_to_swp_entry(oldpte);

			if (is_write_migration_entry(entry)) {
				pte_t newpte;
				/*
				 * A protection check is difficult so
				 * just be safe and disable write
				 */
				make_migration_entry_read(&entry);
				newpte = swp_entry_to_pte(entry);
				if (pte_swp_soft_dirty(oldpte))
					newpte = pte_swp_mksoft_dirty(newpte);
				set_pte_at(mm, addr, pte, newpte);

				pages++;
			}
		}
	} while (pte++, addr += PAGE_SIZE, addr != end);
	arch_leave_lazy_mmu_mode();
	pte_unmap_unlock(pte - 1, ptl);

	return pages;
}

static inline unsigned long change_pmd_range(struct vm_area_struct *vma,
		pud_t *pud, unsigned long addr, unsigned long end,
		pgprot_t newprot, int dirty_accountable, int prot_numa)
{
	pmd_t *pmd;
	struct mm_struct *mm = vma->vm_mm;
	unsigned long next;
	unsigned long pages = 0;
	unsigned long nr_huge_updates = 0;
	unsigned long mni_start = 0;

	pmd = pmd_offset(pud, addr);
	do {
		unsigned long this_pages;

		next = pmd_addr_end(addr, end);
		if (!pmd_trans_huge(*pmd) && !pmd_devmap(*pmd)
				&& pmd_none_or_clear_bad(pmd))
			continue;

		/* invoke the mmu notifier if the pmd is populated */
		if (!mni_start) {
			mni_start = addr;
			mmu_notifier_invalidate_range_start(mm, mni_start, end);
		}

		if (pmd_trans_huge(*pmd) || pmd_devmap(*pmd)) {
			if (next - addr != HPAGE_PMD_SIZE) {
				split_huge_pmd(vma, pmd, addr);
				if (pmd_trans_unstable(pmd))
					continue;
			} else {
				int nr_ptes = change_huge_pmd(vma, pmd, addr,
						newprot, prot_numa);

				if (nr_ptes) {
					if (nr_ptes == HPAGE_PMD_NR) {
						pages += HPAGE_PMD_NR;
						nr_huge_updates++;
					}

					/* huge pmd was handled */
					continue;
				}
			}
			/* fall through, the trans huge pmd just split */
		}
		this_pages = change_pte_range(vma, pmd, addr, next, newprot,
				 dirty_accountable, prot_numa);
		pages += this_pages;
	} while (pmd++, addr = next, addr != end);

	if (mni_start)
		mmu_notifier_invalidate_range_end(mm, mni_start, end);

	if (nr_huge_updates)
		count_vm_numa_events(NUMA_HUGE_PTE_UPDATES, nr_huge_updates);
	return pages;
}

static inline unsigned long change_pud_range(struct vm_area_struct *vma,
		pgd_t *pgd, unsigned long addr, unsigned long end,
		pgprot_t newprot, int dirty_accountable, int prot_numa)
{
	pud_t *pud;
	unsigned long next;
	unsigned long pages = 0;

	pud = pud_offset(pgd, addr);
	do {
		next = pud_addr_end(addr, end);
		if (pud_none_or_clear_bad(pud))
			continue;
		pages += change_pmd_range(vma, pud, addr, next, newprot,
				 dirty_accountable, prot_numa);
	} while (pud++, addr = next, addr != end);

	return pages;
}

static unsigned long change_protection_range(struct vm_area_struct *vma,
		unsigned long addr, unsigned long end, pgprot_t newprot,
		int dirty_accountable, int prot_numa)
{
	struct mm_struct *mm = vma->vm_mm;
	pgd_t *pgd;
	unsigned long next;
	unsigned long start = addr;
	unsigned long pages = 0;

	BUG_ON(addr >= end);
	pgd = pgd_offset(mm, addr);
	flush_cache_range(vma, addr, end);
	set_tlb_flush_pending(mm);
	do {
		next = pgd_addr_end(addr, end);
		if (pgd_none_or_clear_bad(pgd))
			continue;
		pages += change_pud_range(vma, pgd, addr, next, newprot,
				 dirty_accountable, prot_numa);
	} while (pgd++, addr = next, addr != end);

	/* Only flush the TLB if we actually modified any entries: */
	if (pages)
		flush_tlb_range(vma, start, end);
	clear_tlb_flush_pending(mm);

	return pages;
}

unsigned long change_protection(struct vm_area_struct *vma, unsigned long start,
		       unsigned long end, pgprot_t newprot,
		       int dirty_accountable, int prot_numa)
{
	unsigned long pages;

	if (is_vm_hugetlb_page(vma))
		pages = hugetlb_change_protection(vma, start, end, newprot);
	else
		pages = change_protection_range(vma, start, end, newprot, dirty_accountable, prot_numa);

	return pages;
}

int
mprotect_fixup(struct vm_area_struct *vma, struct vm_area_struct **pprev,
	unsigned long start, unsigned long end, unsigned long newflags)
{
	struct mm_struct *mm = vma->vm_mm;
	unsigned long oldflags = vma->vm_flags;
	long nrpages = (end - start) >> PAGE_SHIFT;
	unsigned long charged = 0;
	pgoff_t pgoff;
	int error;
	int dirty_accountable = 0;

	if (newflags == oldflags) {
		*pprev = vma;
		return 0;
	}

	/*
	 * If we make a private mapping writable we increase our commit;
	 * but (without finer accounting) cannot reduce our commit if we
	 * make it unwritable again. hugetlb mapping were accounted for
	 * even if read-only so there is no need to account for them here
	 */
	if (newflags & VM_WRITE) {
		/* Check space limits when area turns into data. */
		if (!may_expand_vm(mm, newflags, nrpages) &&
				may_expand_vm(mm, oldflags, nrpages))
			return -ENOMEM;
		if (!(oldflags & (VM_ACCOUNT|VM_WRITE|VM_HUGETLB|
						VM_SHARED|VM_NORESERVE))) {
			charged = nrpages;
			if (security_vm_enough_memory_mm(mm, charged))
				return -ENOMEM;
			newflags |= VM_ACCOUNT;
		}
	}

	/*
	 * First try to merge with previous and/or next vma.
	 */
	pgoff = vma->vm_pgoff + ((start - vma->vm_start) >> PAGE_SHIFT);
	*pprev = vma_merge(mm, *pprev, start, end, newflags,
			   vma->anon_vma, vma->vm_file, pgoff, vma_policy(vma),
			   vma->vm_userfaultfd_ctx);
	if (*pprev) {
		vma = *pprev;
		goto success;
	}

	*pprev = vma;

	if (start != vma->vm_start) {
		error = split_vma(mm, vma, start, 1);
		if (error)
			goto fail;
	}

	if (end != vma->vm_end) {
		error = split_vma(mm, vma, end, 0);
		if (error)
			goto fail;
	}

success:
	/*
	 * vm_flags and vm_page_prot are protected by the mmap_sem
	 * held in write mode.
	 */
	vma->vm_flags = newflags;
	dirty_accountable = vma_wants_writenotify(vma);
	vma_set_page_prot(vma);

	change_protection(vma, start, end, vma->vm_page_prot,
			  dirty_accountable, 0);

	/*
	 * Private VM_LOCKED VMA becoming writable: trigger COW to avoid major
	 * fault on access.
	 */
	if ((oldflags & (VM_WRITE | VM_SHARED | VM_LOCKED)) == VM_LOCKED &&
			(newflags & VM_WRITE)) {
		populate_vma_page_range(vma, start, end, NULL);
	}

	vm_stat_account(mm, oldflags, -nrpages);
	vm_stat_account(mm, newflags, nrpages);
	perf_event_mmap(vma);
	return 0;

fail:
	vm_unacct_memory(charged);
	return error;
}

SYSCALL_DEFINE3(mprotect, unsigned long, start, size_t, len,
		unsigned long, prot)
{
	unsigned long nstart, end, tmp, reqprot;
	struct vm_area_struct *vma, *prev;
	int error = -EINVAL;
	const int grows = prot & (PROT_GROWSDOWN|PROT_GROWSUP);
	const bool rier = (current->personality & READ_IMPLIES_EXEC) &&
				(prot & PROT_READ);

	prot &= ~(PROT_GROWSDOWN|PROT_GROWSUP);
	if (grows == (PROT_GROWSDOWN|PROT_GROWSUP)) /* can't be both */
		return -EINVAL;

	if (start & ~PAGE_MASK)
		return -EINVAL;
	if (!len)
		return 0;
	len = PAGE_ALIGN(len);
	end = start + len;
	if (end <= start)
		return -ENOMEM;
	if (!arch_validate_prot(prot))
		return -EINVAL;

	reqprot = prot;

	if (down_write_killable(&current->mm->mmap_sem))
		return -EINTR;

	vma = find_vma(current->mm, start);
	error = -ENOMEM;
	if (!vma)
		goto out;
	prev = vma->vm_prev;
	if (unlikely(grows & PROT_GROWSDOWN)) {
		if (vma->vm_start >= end)
			goto out;
		start = vma->vm_start;
		error = -EINVAL;
		if (!(vma->vm_flags & VM_GROWSDOWN))
			goto out;
	} else {
		if (vma->vm_start > start)
			goto out;
		if (unlikely(grows & PROT_GROWSUP)) {
			end = vma->vm_end;
			error = -EINVAL;
			if (!(vma->vm_flags & VM_GROWSUP))
				goto out;
		}
	}
	if (start > vma->vm_start)
		prev = vma;

	for (nstart = start ; ; ) {
		unsigned long newflags;
		int pkey = arch_override_mprotect_pkey(vma, prot, -1);

		/* Here we know that vma->vm_start <= nstart < vma->vm_end. */

		/* Does the application expect PROT_READ to imply PROT_EXEC */
		if (rier && (vma->vm_flags & VM_MAYEXEC))
			prot |= PROT_EXEC;

		newflags = calc_vm_prot_bits(prot, pkey);
		newflags |= (vma->vm_flags & ~(VM_READ | VM_WRITE | VM_EXEC));

		/* newflags >> 4 shift VM_MAY% in place of VM_% */
		if ((newflags & ~(newflags >> 4)) & (VM_READ | VM_WRITE | VM_EXEC)) {
			error = -EACCES;
			goto out;
		}

		error = security_file_mprotect(vma, reqprot, prot);
		if (error)
			goto out;

		tmp = vma->vm_end;
		if (tmp > end)
			tmp = end;
		error = mprotect_fixup(vma, &prev, nstart, tmp, newflags);
		if (error)
			goto out;
		nstart = tmp;

		if (nstart < prev->vm_end)
			nstart = prev->vm_end;
		if (nstart >= end)
			goto out;

		vma = prev->vm_next;
		if (!vma || vma->vm_start != nstart) {
			error = -ENOMEM;
			goto out;
		}
		prot = reqprot;
	}
out:
	up_write(&current->mm->mmap_sem);
	return error;
}
Commit	Line	Data
1da177e4 LT	1	/*
	2	* mm/mprotect.c
	3	*
	4	* (C) Copyright 1994 Linus Torvalds
	5	* (C) Copyright 2002 Christoph Hellwig
	6	*
046c6884	7	* Address space accounting code <alan@lxorguk.ukuu.org.uk>
1da177e4 LT	8	* (C) Copyright 2002 Red Hat Inc, All Rights Reserved
	9	*/
	10
	11	#include <linux/mm.h>
	12	#include <linux/hugetlb.h>
1da177e4 LT	13	#include <linux/shm.h>
	14	#include <linux/mman.h>
	15	#include <linux/fs.h>
	16	#include <linux/highmem.h>
	17	#include <linux/security.h>
	18	#include <linux/mempolicy.h>
	19	#include <linux/personality.h>
	20	#include <linux/syscalls.h>
0697212a CL	21	#include <linux/swap.h>
0697212a CL	22	#include <linux/swapops.h>
cddb8a5c	23	#include <linux/mmu_notifier.h>
64cdd548	24	#include <linux/migrate.h>
cdd6c482	25	#include <linux/perf_event.h>
64a9a34e	26	#include <linux/ksm.h>
62b5f7d0	27	#include <linux/pkeys.h>
1da177e4 LT	28	#include <asm/uaccess.h>
	29	#include <asm/pgtable.h>
	30	#include <asm/cacheflush.h>
	31	#include <asm/tlbflush.h>
	32
36f88188 KS	33	#include "internal.h"
36f88188 KS	34
1ad9f620 MG	35	/*
	36	* For a prot_numa update we only hold mmap_sem for read so there is a
	37	* potential race with faulting where a pmd was temporarily none. This
	38	* function checks for a transhuge pmd under the appropriate lock. It
	39	* returns a pte if it was successfully locked or NULL if it raced with
	40	* a transhuge insertion.
	41	*/
	42	static pte_t lock_pte_protection(struct vm_area_struct vma, pmd_t *pmd,
	43	unsigned long addr, int prot_numa, spinlock_t **ptl)
	44	{
	45	pte_t *pte;
	46	spinlock_t *pmdl;
	47
	48	/* !prot_numa is protected by mmap_sem held for write */
	49	if (!prot_numa)
	50	return pte_offset_map_lock(vma->vm_mm, pmd, addr, ptl);
	51
	52	pmdl = pmd_lock(vma->vm_mm, pmd);
	53	if (unlikely(pmd_trans_huge(pmd) \|\| pmd_none(pmd))) {
	54	spin_unlock(pmdl);
	55	return NULL;
	56	}
	57
	58	pte = pte_offset_map_lock(vma->vm_mm, pmd, addr, ptl);
	59	spin_unlock(pmdl);
	60	return pte;
	61	}
	62
4b10e7d5	63	static unsigned long change_pte_range(struct vm_area_struct vma, pmd_t pmd,
c1e6098b	64	unsigned long addr, unsigned long end, pgprot_t newprot,
0f19c179	65	int dirty_accountable, int prot_numa)
1da177e4	66	{
4b10e7d5	67	struct mm_struct *mm = vma->vm_mm;
0697212a	68	pte_t *pte, oldpte;
705e87c0	69	spinlock_t *ptl;
7da4d641	70	unsigned long pages = 0;
1da177e4	71
1ad9f620 MG	72	pte = lock_pte_protection(vma, pmd, addr, prot_numa, &ptl);
	73	if (!pte)
	74	return 0;
	75
6606c3e0	76	arch_enter_lazy_mmu_mode();
1da177e4	77	do {
0697212a CL	78	oldpte = *pte;
0697212a CL	79	if (pte_present(oldpte)) {
1da177e4	80	pte_t ptent;
b191f9b1	81	bool preserve_write = prot_numa && pte_write(oldpte);
1da177e4	82
e944fd67 MG	83	/*
	84	* Avoid trapping faults against the zero or KSM
	85	* pages. See similar comment in change_huge_pmd.
	86	*/
	87	if (prot_numa) {
	88	struct page *page;
	89
	90	page = vm_normal_page(vma, addr, oldpte);
	91	if (!page \|\| PageKsm(page))
	92	continue;
10c1045f MG	93
	94	/* Avoid TLB flush if possible */
	95	if (pte_protnone(oldpte))
	96	continue;
e944fd67 MG	97	}
e944fd67 MG	98
8a0516ed MG	99	ptent = ptep_modify_prot_start(mm, addr, pte);
8a0516ed MG	100	ptent = pte_modify(ptent, newprot);
b191f9b1 MG	101	if (preserve_write)
b191f9b1 MG	102	ptent = pte_mkwrite(ptent);
4b10e7d5	103
8a0516ed MG	104	/* Avoid taking write faults for known dirty pages */
	105	if (dirty_accountable && pte_dirty(ptent) &&
	106	(pte_soft_dirty(ptent) \|\|
	107	!(vma->vm_flags & VM_SOFTDIRTY))) {
	108	ptent = pte_mkwrite(ptent);
4b10e7d5	109	}
8a0516ed MG	110	ptep_modify_prot_commit(mm, addr, pte, ptent);
8a0516ed MG	111	pages++;
0661a336	112	} else if (IS_ENABLED(CONFIG_MIGRATION)) {
0697212a CL	113	swp_entry_t entry = pte_to_swp_entry(oldpte);
	114
	115	if (is_write_migration_entry(entry)) {
c3d16e16	116	pte_t newpte;
0697212a CL	117	/*
	118	* A protection check is difficult so
	119	* just be safe and disable write
	120	*/
	121	make_migration_entry_read(&entry);
c3d16e16 CG	122	newpte = swp_entry_to_pte(entry);
	123	if (pte_swp_soft_dirty(oldpte))
	124	newpte = pte_swp_mksoft_dirty(newpte);
	125	set_pte_at(mm, addr, pte, newpte);
e920e14c MG	126
e920e14c MG	127	pages++;
0697212a	128	}
1da177e4 LT	129	}
1da177e4 LT	130	} while (pte++, addr += PAGE_SIZE, addr != end);
6606c3e0	131	arch_leave_lazy_mmu_mode();
705e87c0	132	pte_unmap_unlock(pte - 1, ptl);
7da4d641 PZ	133
7da4d641 PZ	134	return pages;
1da177e4 LT	135	}
1da177e4 LT	136
7d12efae AM	137	static inline unsigned long change_pmd_range(struct vm_area_struct *vma,
	138	pud_t *pud, unsigned long addr, unsigned long end,
	139	pgprot_t newprot, int dirty_accountable, int prot_numa)
1da177e4 LT	140	{
1da177e4 LT	141	pmd_t *pmd;
a5338093	142	struct mm_struct *mm = vma->vm_mm;
1da177e4	143	unsigned long next;
7da4d641	144	unsigned long pages = 0;
72403b4a	145	unsigned long nr_huge_updates = 0;
a5338093	146	unsigned long mni_start = 0;
1da177e4 LT	147
	148	pmd = pmd_offset(pud, addr);
	149	do {
25cbbef1 MG	150	unsigned long this_pages;
25cbbef1 MG	151
1da177e4	152	next = pmd_addr_end(addr, end);
5c7fb56e DW	153	if (!pmd_trans_huge(pmd) && !pmd_devmap(pmd)
5c7fb56e DW	154	&& pmd_none_or_clear_bad(pmd))
88a9ab6e	155	continue;
a5338093 RR	156
	157	/* invoke the mmu notifier if the pmd is populated */
	158	if (!mni_start) {
	159	mni_start = addr;
	160	mmu_notifier_invalidate_range_start(mm, mni_start, end);
	161	}
	162
5c7fb56e	163	if (pmd_trans_huge(pmd) \|\| pmd_devmap(pmd)) {
6b9116a6	164	if (next - addr != HPAGE_PMD_SIZE) {
78ddc534	165	split_huge_pmd(vma, pmd, addr);
337d9abf	166	if (pmd_trans_unstable(pmd))
6b9116a6 KS	167	continue;
6b9116a6 KS	168	} else {
f123d74a	169	int nr_ptes = change_huge_pmd(vma, pmd, addr,
e944fd67	170	newprot, prot_numa);
f123d74a MG	171
f123d74a MG	172	if (nr_ptes) {
72403b4a MG	173	if (nr_ptes == HPAGE_PMD_NR) {
	174	pages += HPAGE_PMD_NR;
	175	nr_huge_updates++;
	176	}
1ad9f620 MG	177
1ad9f620 MG	178	/* huge pmd was handled */
f123d74a MG	179	continue;
f123d74a MG	180	}
7da4d641	181	}
88a9ab6e	182	/* fall through, the trans huge pmd just split */
cd7548ab	183	}
25cbbef1	184	this_pages = change_pte_range(vma, pmd, addr, next, newprot,
0f19c179	185	dirty_accountable, prot_numa);
25cbbef1	186	pages += this_pages;
1da177e4	187	} while (pmd++, addr = next, addr != end);
7da4d641	188
a5338093 RR	189	if (mni_start)
	190	mmu_notifier_invalidate_range_end(mm, mni_start, end);
	191
72403b4a MG	192	if (nr_huge_updates)
72403b4a MG	193	count_vm_numa_events(NUMA_HUGE_PTE_UPDATES, nr_huge_updates);
7da4d641	194	return pages;
1da177e4 LT	195	}
1da177e4 LT	196
7d12efae AM	197	static inline unsigned long change_pud_range(struct vm_area_struct *vma,
	198	pgd_t *pgd, unsigned long addr, unsigned long end,
	199	pgprot_t newprot, int dirty_accountable, int prot_numa)
1da177e4 LT	200	{
	201	pud_t *pud;
	202	unsigned long next;
7da4d641	203	unsigned long pages = 0;
1da177e4 LT	204
	205	pud = pud_offset(pgd, addr);
	206	do {
	207	next = pud_addr_end(addr, end);
	208	if (pud_none_or_clear_bad(pud))
	209	continue;
7da4d641	210	pages += change_pmd_range(vma, pud, addr, next, newprot,
4b10e7d5	211	dirty_accountable, prot_numa);
1da177e4	212	} while (pud++, addr = next, addr != end);
7da4d641 PZ	213
7da4d641 PZ	214	return pages;
1da177e4 LT	215	}
1da177e4 LT	216
7da4d641	217	static unsigned long change_protection_range(struct vm_area_struct *vma,
c1e6098b	218	unsigned long addr, unsigned long end, pgprot_t newprot,
4b10e7d5	219	int dirty_accountable, int prot_numa)
1da177e4 LT	220	{
	221	struct mm_struct *mm = vma->vm_mm;
	222	pgd_t *pgd;
	223	unsigned long next;
	224	unsigned long start = addr;
7da4d641	225	unsigned long pages = 0;
1da177e4 LT	226
	227	BUG_ON(addr >= end);
	228	pgd = pgd_offset(mm, addr);
	229	flush_cache_range(vma, addr, end);
20841405	230	set_tlb_flush_pending(mm);
1da177e4 LT	231	do {
	232	next = pgd_addr_end(addr, end);
	233	if (pgd_none_or_clear_bad(pgd))
	234	continue;
7da4d641	235	pages += change_pud_range(vma, pgd, addr, next, newprot,
4b10e7d5	236	dirty_accountable, prot_numa);
1da177e4	237	} while (pgd++, addr = next, addr != end);
7da4d641	238
1233d588 IM	239	/* Only flush the TLB if we actually modified any entries: */
	240	if (pages)
	241	flush_tlb_range(vma, start, end);
20841405	242	clear_tlb_flush_pending(mm);
7da4d641 PZ	243
	244	return pages;
	245	}
	246
	247	unsigned long change_protection(struct vm_area_struct *vma, unsigned long start,
	248	unsigned long end, pgprot_t newprot,
4b10e7d5	249	int dirty_accountable, int prot_numa)
7da4d641	250	{
7da4d641 PZ	251	unsigned long pages;
7da4d641 PZ	252
7da4d641 PZ	253	if (is_vm_hugetlb_page(vma))
	254	pages = hugetlb_change_protection(vma, start, end, newprot);
	255	else
4b10e7d5	256	pages = change_protection_range(vma, start, end, newprot, dirty_accountable, prot_numa);
7da4d641 PZ	257
7da4d641 PZ	258	return pages;
1da177e4 LT	259	}
1da177e4 LT	260
b6a2fea3	261	int
1da177e4 LT	262	mprotect_fixup(struct vm_area_struct vma, struct vm_area_struct *pprev,
	263	unsigned long start, unsigned long end, unsigned long newflags)
	264	{
	265	struct mm_struct *mm = vma->vm_mm;
	266	unsigned long oldflags = vma->vm_flags;
	267	long nrpages = (end - start) >> PAGE_SHIFT;
	268	unsigned long charged = 0;
1da177e4 LT	269	pgoff_t pgoff;
1da177e4 LT	270	int error;
c1e6098b	271	int dirty_accountable = 0;
1da177e4 LT	272
	273	if (newflags == oldflags) {
	274	*pprev = vma;
	275	return 0;
	276	}
	277
	278	/*
	279	* If we make a private mapping writable we increase our commit;
	280	* but (without finer accounting) cannot reduce our commit if we
5a6fe125 MG	281	* make it unwritable again. hugetlb mapping were accounted for
5a6fe125 MG	282	* even if read-only so there is no need to account for them here
1da177e4 LT	283	*/
1da177e4 LT	284	if (newflags & VM_WRITE) {
84638335 KK	285	/* Check space limits when area turns into data. */
	286	if (!may_expand_vm(mm, newflags, nrpages) &&
	287	may_expand_vm(mm, oldflags, nrpages))
	288	return -ENOMEM;
5a6fe125	289	if (!(oldflags & (VM_ACCOUNT\|VM_WRITE\|VM_HUGETLB\|
cdfd4325	290	VM_SHARED\|VM_NORESERVE))) {
1da177e4	291	charged = nrpages;
191c5424	292	if (security_vm_enough_memory_mm(mm, charged))
1da177e4 LT	293	return -ENOMEM;
	294	newflags \|= VM_ACCOUNT;
	295	}
	296	}
	297
1da177e4 LT	298	/*
	299	* First try to merge with previous and/or next vma.
	300	*/
	301	pgoff = vma->vm_pgoff + ((start - vma->vm_start) >> PAGE_SHIFT);
	302	pprev = vma_merge(mm, pprev, start, end, newflags,
19a809af AA	303	vma->anon_vma, vma->vm_file, pgoff, vma_policy(vma),
19a809af AA	304	vma->vm_userfaultfd_ctx);
1da177e4 LT	305	if (*pprev) {
	306	vma = *pprev;
	307	goto success;
	308	}
	309
	310	*pprev = vma;
	311
	312	if (start != vma->vm_start) {
	313	error = split_vma(mm, vma, start, 1);
	314	if (error)
	315	goto fail;
	316	}
	317
	318	if (end != vma->vm_end) {
	319	error = split_vma(mm, vma, end, 0);
	320	if (error)
	321	goto fail;
	322	}
	323
	324	success:
	325	/*
	326	* vm_flags and vm_page_prot are protected by the mmap_sem
	327	* held in write mode.
	328	*/
	329	vma->vm_flags = newflags;
64e45507 PF	330	dirty_accountable = vma_wants_writenotify(vma);
64e45507 PF	331	vma_set_page_prot(vma);
d08b3851	332
7d12efae AM	333	change_protection(vma, start, end, vma->vm_page_prot,
7d12efae AM	334	dirty_accountable, 0);
7da4d641	335
36f88188 KS	336	/*
	337	* Private VM_LOCKED VMA becoming writable: trigger COW to avoid major
	338	* fault on access.
	339	*/
	340	if ((oldflags & (VM_WRITE \| VM_SHARED \| VM_LOCKED)) == VM_LOCKED &&
	341	(newflags & VM_WRITE)) {
	342	populate_vma_page_range(vma, start, end, NULL);
	343	}
	344
84638335 KK	345	vm_stat_account(mm, oldflags, -nrpages);
84638335 KK	346	vm_stat_account(mm, newflags, nrpages);
63bfd738	347	perf_event_mmap(vma);
1da177e4 LT	348	return 0;
	349
	350	fail:
	351	vm_unacct_memory(charged);
	352	return error;
	353	}
	354
6a6160a7 HC	355	SYSCALL_DEFINE3(mprotect, unsigned long, start, size_t, len,
6a6160a7 HC	356	unsigned long, prot)
1da177e4	357	{
62b5f7d0	358	unsigned long nstart, end, tmp, reqprot;
1da177e4 LT	359	struct vm_area_struct vma, prev;
	360	int error = -EINVAL;
	361	const int grows = prot & (PROT_GROWSDOWN\|PROT_GROWSUP);
f138556d PK	362	const bool rier = (current->personality & READ_IMPLIES_EXEC) &&
	363	(prot & PROT_READ);
	364
1da177e4 LT	365	prot &= ~(PROT_GROWSDOWN\|PROT_GROWSUP);
	366	if (grows == (PROT_GROWSDOWN\|PROT_GROWSUP)) /* can't be both */
	367	return -EINVAL;
	368
	369	if (start & ~PAGE_MASK)
	370	return -EINVAL;
	371	if (!len)
	372	return 0;
	373	len = PAGE_ALIGN(len);
	374	end = start + len;
	375	if (end <= start)
	376	return -ENOMEM;
b845f313	377	if (!arch_validate_prot(prot))
1da177e4 LT	378	return -EINVAL;
	379
	380	reqprot = prot;
1da177e4	381
dc0ef0df MH	382	if (down_write_killable(&current->mm->mmap_sem))
dc0ef0df MH	383	return -EINTR;
1da177e4	384
097d5910	385	vma = find_vma(current->mm, start);
1da177e4 LT	386	error = -ENOMEM;
	387	if (!vma)
	388	goto out;
097d5910	389	prev = vma->vm_prev;
1da177e4 LT	390	if (unlikely(grows & PROT_GROWSDOWN)) {
	391	if (vma->vm_start >= end)
	392	goto out;
	393	start = vma->vm_start;
	394	error = -EINVAL;
	395	if (!(vma->vm_flags & VM_GROWSDOWN))
	396	goto out;
7d12efae	397	} else {
1da177e4 LT	398	if (vma->vm_start > start)
	399	goto out;
	400	if (unlikely(grows & PROT_GROWSUP)) {
	401	end = vma->vm_end;
	402	error = -EINVAL;
	403	if (!(vma->vm_flags & VM_GROWSUP))
	404	goto out;
	405	}
	406	}
	407	if (start > vma->vm_start)
	408	prev = vma;
	409
	410	for (nstart = start ; ; ) {
	411	unsigned long newflags;
62b5f7d0	412	int pkey = arch_override_mprotect_pkey(vma, prot, -1);
1da177e4	413
7d12efae	414	/* Here we know that vma->vm_start <= nstart < vma->vm_end. */
1da177e4	415
f138556d PK	416	/* Does the application expect PROT_READ to imply PROT_EXEC */
	417	if (rier && (vma->vm_flags & VM_MAYEXEC))
	418	prot \|= PROT_EXEC;
	419
62b5f7d0	420	newflags = calc_vm_prot_bits(prot, pkey);
7d12efae	421	newflags \|= (vma->vm_flags & ~(VM_READ \| VM_WRITE \| VM_EXEC));
1da177e4	422
7e2cff42 PBG	423	/* newflags >> 4 shift VM_MAY% in place of VM_% */
7e2cff42 PBG	424	if ((newflags & ~(newflags >> 4)) & (VM_READ \| VM_WRITE \| VM_EXEC)) {
1da177e4 LT	425	error = -EACCES;
	426	goto out;
	427	}
	428
	429	error = security_file_mprotect(vma, reqprot, prot);
	430	if (error)
	431	goto out;
	432
	433	tmp = vma->vm_end;
	434	if (tmp > end)
	435	tmp = end;
	436	error = mprotect_fixup(vma, &prev, nstart, tmp, newflags);
	437	if (error)
	438	goto out;
	439	nstart = tmp;
	440
	441	if (nstart < prev->vm_end)
	442	nstart = prev->vm_end;
	443	if (nstart >= end)
	444	goto out;
	445
	446	vma = prev->vm_next;
	447	if (!vma \|\| vma->vm_start != nstart) {
	448	error = -ENOMEM;
	449	goto out;
	450	}
f138556d	451	prot = reqprot;
1da177e4 LT	452	}
	453	out:
	454	up_write(&current->mm->mmap_sem);
	455	return error;
	456	}