[thirdparty/kernel/stable.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 <asm/uaccess.h>
#include <asm/pgtable.h>
#include <asm/cacheflush.h>
#include <asm/tlbflush.h>

#ifndef pgprot_modify
static inline pgprot_t pgprot_modify(pgprot_t oldprot, pgprot_t newprot)
{
	return newprot;
}
#endif

/*
 * 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 updated = false;

			if (!prot_numa) {
				ptent = ptep_modify_prot_start(mm, addr, pte);
				if (pte_numa(ptent))
					ptent = pte_mknonnuma(ptent);
				ptent = pte_modify(ptent, newprot);
				/*
				 * Avoid taking write faults for pages we
				 * know to be dirty.
				 */
				if (dirty_accountable && pte_dirty(ptent))
					ptent = pte_mkwrite(ptent);
				ptep_modify_prot_commit(mm, addr, pte, ptent);
				updated = true;
			} else {
				struct page *page;

				page = vm_normal_page(vma, addr, oldpte);
				if (page && !PageKsm(page)) {
					if (!pte_numa(oldpte)) {
						ptep_set_numa(mm, addr, pte);
						updated = true;
					}
				}
			}
			if (updated)
				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_none_or_clear_bad(pmd))
			goto next;

		/* 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)) {
			if (next - addr != HPAGE_PMD_SIZE)
				split_huge_page_pmd(vma, addr, pmd);
			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 */
					goto next;
				}
			}
			/* 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;
next:
		cond_resched();
	} 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);
	inc_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);
	dec_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;
}

static int prot_none_pte_entry(pte_t *pte, unsigned long addr,
			       unsigned long next, struct mm_walk *walk)
{
	return pfn_modify_allowed(pte_pfn(*pte), *(pgprot_t *)(walk->private)) ?
		0 : -EACCES;
}

static int prot_none_hugetlb_entry(pte_t *pte, unsigned long hmask,
				   unsigned long addr, unsigned long next,
				   struct mm_walk *walk)
{
	return pfn_modify_allowed(pte_pfn(*pte), *(pgprot_t *)(walk->private)) ?
		0 : -EACCES;
}

static int prot_none_test(unsigned long addr, unsigned long next,
			  struct mm_walk *walk)
{
	return 0;
}

static int prot_none_walk(struct vm_area_struct *vma, unsigned long start,
			   unsigned long end, unsigned long newflags)
{
	pgprot_t new_pgprot = vm_get_page_prot(newflags);
	struct mm_walk prot_none_walk = {
		.pte_entry = prot_none_pte_entry,
		.hugetlb_entry = prot_none_hugetlb_entry,
		.test_walk = prot_none_test,
		.mm = current->mm,
		.private = &new_pgprot,
	};

	return walk_page_range(start, end, &prot_none_walk);
}

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;
	}

	/*
	 * Do PROT_NONE PFN permission checks here when we can still
	 * bail out without undoing a lot of state. This is a rather
	 * uncommon case, so doesn't need to be very optimized.
	 */
	if (arch_has_pfn_modify_check() &&
	    (vma->vm_flags & (VM_PFNMAP|VM_MIXEDMAP)) &&
	    (newflags & (VM_READ|VM_WRITE|VM_EXEC)) == 0) {
		error = prot_none_walk(vma, start, end, newflags);
		if (error)
			return error;
	}

	/*
	 * 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) {
		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));
	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;
	vma->vm_page_prot = pgprot_modify(vma->vm_page_prot,
					  vm_get_page_prot(newflags));

	if (vma_wants_writenotify(vma)) {
		vma->vm_page_prot = vm_get_page_prot(newflags & ~VM_SHARED);
		dirty_accountable = 1;
	}

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

	vm_stat_account(mm, oldflags, vma->vm_file, -nrpages);
	vm_stat_account(mm, newflags, vma->vm_file, 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 vm_flags, nstart, end, tmp, reqprot;
	struct vm_area_struct *vma, *prev;
	int error = -EINVAL;
	const int grows = prot & (PROT_GROWSDOWN|PROT_GROWSUP);
	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;
	/*
	 * Does the application expect PROT_READ to imply PROT_EXEC:
	 */
	if ((prot & PROT_READ) && (current->personality & READ_IMPLIES_EXEC))
		prot |= PROT_EXEC;

	vm_flags = calc_vm_prot_bits(prot);

	down_write(&current->mm->mmap_sem);

	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;

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

		newflags = vm_flags;
		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;
		}
	}
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>
1da177e4 LT	27	#include <asm/uaccess.h>
	28	#include <asm/pgtable.h>
	29	#include <asm/cacheflush.h>
	30	#include <asm/tlbflush.h>
	31
1c12c4cf VP	32	#ifndef pgprot_modify
	33	static inline pgprot_t pgprot_modify(pgprot_t oldprot, pgprot_t newprot)
	34	{
	35	return newprot;
	36	}
	37	#endif
	38
1ad9f620 MG	39	/*
	40	* For a prot_numa update we only hold mmap_sem for read so there is a
	41	* potential race with faulting where a pmd was temporarily none. This
	42	* function checks for a transhuge pmd under the appropriate lock. It
	43	* returns a pte if it was successfully locked or NULL if it raced with
	44	* a transhuge insertion.
	45	*/
	46	static pte_t lock_pte_protection(struct vm_area_struct vma, pmd_t *pmd,
	47	unsigned long addr, int prot_numa, spinlock_t **ptl)
	48	{
	49	pte_t *pte;
	50	spinlock_t *pmdl;
	51
	52	/* !prot_numa is protected by mmap_sem held for write */
	53	if (!prot_numa)
	54	return pte_offset_map_lock(vma->vm_mm, pmd, addr, ptl);
	55
	56	pmdl = pmd_lock(vma->vm_mm, pmd);
	57	if (unlikely(pmd_trans_huge(pmd) \|\| pmd_none(pmd))) {
	58	spin_unlock(pmdl);
	59	return NULL;
	60	}
	61
	62	pte = pte_offset_map_lock(vma->vm_mm, pmd, addr, ptl);
	63	spin_unlock(pmdl);
	64	return pte;
	65	}
	66
4b10e7d5	67	static unsigned long change_pte_range(struct vm_area_struct vma, pmd_t pmd,
c1e6098b	68	unsigned long addr, unsigned long end, pgprot_t newprot,
0f19c179	69	int dirty_accountable, int prot_numa)
1da177e4	70	{
4b10e7d5	71	struct mm_struct *mm = vma->vm_mm;
0697212a	72	pte_t *pte, oldpte;
705e87c0	73	spinlock_t *ptl;
7da4d641	74	unsigned long pages = 0;
1da177e4	75
1ad9f620 MG	76	pte = lock_pte_protection(vma, pmd, addr, prot_numa, &ptl);
	77	if (!pte)
	78	return 0;
	79
6606c3e0	80	arch_enter_lazy_mmu_mode();
1da177e4	81	do {
0697212a CL	82	oldpte = *pte;
0697212a CL	83	if (pte_present(oldpte)) {
1da177e4	84	pte_t ptent;
4b10e7d5	85	bool updated = false;
1da177e4	86
4b10e7d5	87	if (!prot_numa) {
0c5f83c2	88	ptent = ptep_modify_prot_start(mm, addr, pte);
1667918b MG	89	if (pte_numa(ptent))
1667918b MG	90	ptent = pte_mknonnuma(ptent);
4b10e7d5	91	ptent = pte_modify(ptent, newprot);
9d85d586 AK	92	/*
	93	* Avoid taking write faults for pages we
	94	* know to be dirty.
	95	*/
	96	if (dirty_accountable && pte_dirty(ptent))
	97	ptent = pte_mkwrite(ptent);
	98	ptep_modify_prot_commit(mm, addr, pte, ptent);
4b10e7d5 MG	99	updated = true;
	100	} else {
	101	struct page *page;
	102
	103	page = vm_normal_page(vma, addr, oldpte);
64a9a34e	104	if (page && !PageKsm(page)) {
1bc115d8	105	if (!pte_numa(oldpte)) {
56eecdb9	106	ptep_set_numa(mm, addr, pte);
4b10e7d5 MG	107	updated = true;
	108	}
	109	}
	110	}
4b10e7d5 MG	111	if (updated)
4b10e7d5 MG	112	pages++;
38ae3752	113	} else if (IS_ENABLED(CONFIG_MIGRATION)) {
0697212a CL	114	swp_entry_t entry = pte_to_swp_entry(oldpte);
	115
	116	if (is_write_migration_entry(entry)) {
c3d16e16	117	pte_t newpte;
0697212a CL	118	/*
	119	* A protection check is difficult so
	120	* just be safe and disable write
	121	*/
	122	make_migration_entry_read(&entry);
c3d16e16 CG	123	newpte = swp_entry_to_pte(entry);
	124	if (pte_swp_soft_dirty(oldpte))
	125	newpte = pte_swp_mksoft_dirty(newpte);
	126	set_pte_at(mm, addr, pte, newpte);
e920e14c MG	127
e920e14c MG	128	pages++;
0697212a	129	}
1da177e4 LT	130	}
1da177e4 LT	131	} while (pte++, addr += PAGE_SIZE, addr != end);
6606c3e0	132	arch_leave_lazy_mmu_mode();
705e87c0	133	pte_unmap_unlock(pte - 1, ptl);
7da4d641 PZ	134
7da4d641 PZ	135	return pages;
1da177e4 LT	136	}
1da177e4 LT	137
7d12efae AM	138	static inline unsigned long change_pmd_range(struct vm_area_struct *vma,
	139	pud_t *pud, unsigned long addr, unsigned long end,
	140	pgprot_t newprot, int dirty_accountable, int prot_numa)
1da177e4 LT	141	{
1da177e4 LT	142	pmd_t *pmd;
a5338093	143	struct mm_struct *mm = vma->vm_mm;
1da177e4	144	unsigned long next;
7da4d641	145	unsigned long pages = 0;
72403b4a	146	unsigned long nr_huge_updates = 0;
a5338093	147	unsigned long mni_start = 0;
1da177e4 LT	148
	149	pmd = pmd_offset(pud, addr);
	150	do {
25cbbef1 MG	151	unsigned long this_pages;
25cbbef1 MG	152
1da177e4	153	next = pmd_addr_end(addr, end);
88a9ab6e	154	if (!pmd_trans_huge(*pmd) && pmd_none_or_clear_bad(pmd))
8ceeebad	155	goto next;
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
cd7548ab JW	163	if (pmd_trans_huge(*pmd)) {
cd7548ab JW	164	if (next - addr != HPAGE_PMD_SIZE)
e180377f	165	split_huge_page_pmd(vma, addr, pmd);
f123d74a MG	166	else {
	167	int nr_ptes = change_huge_pmd(vma, pmd, addr,
	168	newprot, prot_numa);
	169
	170	if (nr_ptes) {
72403b4a MG	171	if (nr_ptes == HPAGE_PMD_NR) {
	172	pages += HPAGE_PMD_NR;
	173	nr_huge_updates++;
	174	}
1ad9f620 MG	175
1ad9f620 MG	176	/* huge pmd was handled */
8ceeebad	177	goto next;
f123d74a	178	}
7da4d641	179	}
88a9ab6e	180	/* fall through, the trans huge pmd just split */
cd7548ab	181	}
25cbbef1	182	this_pages = change_pte_range(vma, pmd, addr, next, newprot,
0f19c179	183	dirty_accountable, prot_numa);
25cbbef1	184	pages += this_pages;
8ceeebad AK	185	next:
8ceeebad AK	186	cond_resched();
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);
7b590b41	230	inc_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);
7b590b41	242	dec_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
74a43011 AK	261	static int prot_none_pte_entry(pte_t *pte, unsigned long addr,
	262	unsigned long next, struct mm_walk *walk)
	263	{
	264	return pfn_modify_allowed(pte_pfn(pte), (pgprot_t *)(walk->private)) ?
	265	0 : -EACCES;
	266	}
	267
	268	static int prot_none_hugetlb_entry(pte_t *pte, unsigned long hmask,
	269	unsigned long addr, unsigned long next,
	270	struct mm_walk *walk)
	271	{
	272	return pfn_modify_allowed(pte_pfn(pte), (pgprot_t *)(walk->private)) ?
	273	0 : -EACCES;
	274	}
	275
	276	static int prot_none_test(unsigned long addr, unsigned long next,
	277	struct mm_walk *walk)
	278	{
	279	return 0;
	280	}
	281
	282	static int prot_none_walk(struct vm_area_struct *vma, unsigned long start,
	283	unsigned long end, unsigned long newflags)
	284	{
	285	pgprot_t new_pgprot = vm_get_page_prot(newflags);
	286	struct mm_walk prot_none_walk = {
	287	.pte_entry = prot_none_pte_entry,
	288	.hugetlb_entry = prot_none_hugetlb_entry,
	289	.test_walk = prot_none_test,
	290	.mm = current->mm,
	291	.private = &new_pgprot,
	292	};
	293
	294	return walk_page_range(start, end, &prot_none_walk);
	295	}
	296
b6a2fea3	297	int
1da177e4 LT	298	mprotect_fixup(struct vm_area_struct vma, struct vm_area_struct *pprev,
	299	unsigned long start, unsigned long end, unsigned long newflags)
	300	{
	301	struct mm_struct *mm = vma->vm_mm;
	302	unsigned long oldflags = vma->vm_flags;
	303	long nrpages = (end - start) >> PAGE_SHIFT;
	304	unsigned long charged = 0;
1da177e4 LT	305	pgoff_t pgoff;
1da177e4 LT	306	int error;
c1e6098b	307	int dirty_accountable = 0;
1da177e4 LT	308
	309	if (newflags == oldflags) {
	310	*pprev = vma;
	311	return 0;
	312	}
	313
74a43011 AK	314	/*
	315	* Do PROT_NONE PFN permission checks here when we can still
	316	* bail out without undoing a lot of state. This is a rather
	317	* uncommon case, so doesn't need to be very optimized.
	318	*/
	319	if (arch_has_pfn_modify_check() &&
	320	(vma->vm_flags & (VM_PFNMAP\|VM_MIXEDMAP)) &&
	321	(newflags & (VM_READ\|VM_WRITE\|VM_EXEC)) == 0) {
	322	error = prot_none_walk(vma, start, end, newflags);
	323	if (error)
	324	return error;
	325	}
	326
1da177e4 LT	327	/*
	328	* If we make a private mapping writable we increase our commit;
	329	* but (without finer accounting) cannot reduce our commit if we
5a6fe125 MG	330	* make it unwritable again. hugetlb mapping were accounted for
5a6fe125 MG	331	* even if read-only so there is no need to account for them here
1da177e4 LT	332	*/
1da177e4 LT	333	if (newflags & VM_WRITE) {
5a6fe125	334	if (!(oldflags & (VM_ACCOUNT\|VM_WRITE\|VM_HUGETLB\|
cdfd4325	335	VM_SHARED\|VM_NORESERVE))) {
1da177e4	336	charged = nrpages;
191c5424	337	if (security_vm_enough_memory_mm(mm, charged))
1da177e4 LT	338	return -ENOMEM;
	339	newflags \|= VM_ACCOUNT;
	340	}
	341	}
	342
1da177e4 LT	343	/*
	344	* First try to merge with previous and/or next vma.
	345	*/
	346	pgoff = vma->vm_pgoff + ((start - vma->vm_start) >> PAGE_SHIFT);
	347	pprev = vma_merge(mm, pprev, start, end, newflags,
	348	vma->anon_vma, vma->vm_file, pgoff, vma_policy(vma));
	349	if (*pprev) {
	350	vma = *pprev;
	351	goto success;
	352	}
	353
	354	*pprev = vma;
	355
	356	if (start != vma->vm_start) {
	357	error = split_vma(mm, vma, start, 1);
	358	if (error)
	359	goto fail;
	360	}
	361
	362	if (end != vma->vm_end) {
	363	error = split_vma(mm, vma, end, 0);
	364	if (error)
	365	goto fail;
	366	}
	367
	368	success:
	369	/*
	370	* vm_flags and vm_page_prot are protected by the mmap_sem
	371	* held in write mode.
	372	*/
	373	vma->vm_flags = newflags;
1c12c4cf VP	374	vma->vm_page_prot = pgprot_modify(vma->vm_page_prot,
	375	vm_get_page_prot(newflags));
	376
c1e6098b	377	if (vma_wants_writenotify(vma)) {
1ddd439e	378	vma->vm_page_prot = vm_get_page_prot(newflags & ~VM_SHARED);
c1e6098b PZ	379	dirty_accountable = 1;
c1e6098b PZ	380	}
d08b3851	381
7d12efae AM	382	change_protection(vma, start, end, vma->vm_page_prot,
7d12efae AM	383	dirty_accountable, 0);
7da4d641	384
ab50b8ed HD	385	vm_stat_account(mm, oldflags, vma->vm_file, -nrpages);
ab50b8ed HD	386	vm_stat_account(mm, newflags, vma->vm_file, nrpages);
63bfd738	387	perf_event_mmap(vma);
1da177e4 LT	388	return 0;
	389
	390	fail:
	391	vm_unacct_memory(charged);
	392	return error;
	393	}
	394
6a6160a7 HC	395	SYSCALL_DEFINE3(mprotect, unsigned long, start, size_t, len,
6a6160a7 HC	396	unsigned long, prot)
1da177e4 LT	397	{
	398	unsigned long vm_flags, nstart, end, tmp, reqprot;
	399	struct vm_area_struct vma, prev;
	400	int error = -EINVAL;
	401	const int grows = prot & (PROT_GROWSDOWN\|PROT_GROWSUP);
	402	prot &= ~(PROT_GROWSDOWN\|PROT_GROWSUP);
	403	if (grows == (PROT_GROWSDOWN\|PROT_GROWSUP)) /* can't be both */
	404	return -EINVAL;
	405
	406	if (start & ~PAGE_MASK)
	407	return -EINVAL;
	408	if (!len)
	409	return 0;
	410	len = PAGE_ALIGN(len);
	411	end = start + len;
	412	if (end <= start)
	413	return -ENOMEM;
b845f313	414	if (!arch_validate_prot(prot))
1da177e4 LT	415	return -EINVAL;
	416
	417	reqprot = prot;
	418	/*
	419	* Does the application expect PROT_READ to imply PROT_EXEC:
	420	*/
b344e05c	421	if ((prot & PROT_READ) && (current->personality & READ_IMPLIES_EXEC))
1da177e4 LT	422	prot \|= PROT_EXEC;
	423
	424	vm_flags = calc_vm_prot_bits(prot);
	425
	426	down_write(&current->mm->mmap_sem);
	427
097d5910	428	vma = find_vma(current->mm, start);
1da177e4 LT	429	error = -ENOMEM;
	430	if (!vma)
	431	goto out;
097d5910	432	prev = vma->vm_prev;
1da177e4 LT	433	if (unlikely(grows & PROT_GROWSDOWN)) {
	434	if (vma->vm_start >= end)
	435	goto out;
	436	start = vma->vm_start;
	437	error = -EINVAL;
	438	if (!(vma->vm_flags & VM_GROWSDOWN))
	439	goto out;
7d12efae	440	} else {
1da177e4 LT	441	if (vma->vm_start > start)
	442	goto out;
	443	if (unlikely(grows & PROT_GROWSUP)) {
	444	end = vma->vm_end;
	445	error = -EINVAL;
	446	if (!(vma->vm_flags & VM_GROWSUP))
	447	goto out;
	448	}
	449	}
	450	if (start > vma->vm_start)
	451	prev = vma;
	452
	453	for (nstart = start ; ; ) {
	454	unsigned long newflags;
	455
7d12efae	456	/* Here we know that vma->vm_start <= nstart < vma->vm_end. */
1da177e4	457
7d12efae AM	458	newflags = vm_flags;
7d12efae AM	459	newflags \|= (vma->vm_flags & ~(VM_READ \| VM_WRITE \| VM_EXEC));
1da177e4	460
7e2cff42 PBG	461	/* newflags >> 4 shift VM_MAY% in place of VM_% */
7e2cff42 PBG	462	if ((newflags & ~(newflags >> 4)) & (VM_READ \| VM_WRITE \| VM_EXEC)) {
1da177e4 LT	463	error = -EACCES;
	464	goto out;
	465	}
	466
	467	error = security_file_mprotect(vma, reqprot, prot);
	468	if (error)
	469	goto out;
	470
	471	tmp = vma->vm_end;
	472	if (tmp > end)
	473	tmp = end;
	474	error = mprotect_fixup(vma, &prev, nstart, tmp, newflags);
	475	if (error)
	476	goto out;
	477	nstart = tmp;
	478
	479	if (nstart < prev->vm_end)
	480	nstart = prev->vm_end;
	481	if (nstart >= end)
	482	goto out;
	483
	484	vma = prev->vm_next;
	485	if (!vma \|\| vma->vm_start != nstart) {
	486	error = -ENOMEM;
	487	goto out;
	488	}
	489	}
	490	out:
	491	up_write(&current->mm->mmap_sem);
	492	return error;
	493	}