[linux-2.6-block.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 <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

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 = pte_offset_map_lock(mm, pmd, addr, &ptl);
	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);
				updated = true;
			} else {
				struct page *page;

				ptent = *pte;
				page = vm_normal_page(vma, addr, oldpte);
				if (page) {
					if (!pte_numa(oldpte)) {
						ptent = pte_mknuma(ptent);
						set_pte_at(mm, addr, pte, ptent);
						updated = true;
					}
				}
			}

			/*
			 * Avoid taking write faults for pages we know to be
			 * dirty.
			 */
			if (dirty_accountable && pte_dirty(ptent)) {
				ptent = pte_mkwrite(ptent);
				updated = true;
			}

			if (updated)
				pages++;

			/* Only !prot_numa always clears the pte */
			if (!prot_numa)
				ptep_modify_prot_commit(mm, addr, pte, ptent);
		} else if (IS_ENABLED(CONFIG_MIGRATION) && !pte_file(oldpte)) {
			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;
	unsigned long next;
	unsigned long pages = 0;
	unsigned long nr_huge_updates = 0;

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

		next = pmd_addr_end(addr, 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++;
					}
					continue;
				}
			}
			/* fall through */
		}
		if (pmd_none_or_clear_bad(pmd))
			continue;
		this_pages = change_pte_range(vma, pmd, addr, next, newprot,
				 dirty_accountable, prot_numa);
		pages += this_pages;
	} while (pmd++, addr = next, addr != 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)
{
	struct mm_struct *mm = vma->vm_mm;
	unsigned long pages;

	mmu_notifier_invalidate_range_start(mm, start, end);
	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);
	mmu_notifier_invalidate_range_end(mm, start, end);

	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) {
		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>
1da177e4 LT	26	#include <asm/uaccess.h>
	27	#include <asm/pgtable.h>
	28	#include <asm/cacheflush.h>
	29	#include <asm/tlbflush.h>
	30
1c12c4cf VP	31	#ifndef pgprot_modify
	32	static inline pgprot_t pgprot_modify(pgprot_t oldprot, pgprot_t newprot)
	33	{
	34	return newprot;
	35	}
	36	#endif
	37
4b10e7d5	38	static unsigned long change_pte_range(struct vm_area_struct vma, pmd_t pmd,
c1e6098b	39	unsigned long addr, unsigned long end, pgprot_t newprot,
0f19c179	40	int dirty_accountable, int prot_numa)
1da177e4	41	{
4b10e7d5	42	struct mm_struct *mm = vma->vm_mm;
0697212a	43	pte_t *pte, oldpte;
705e87c0	44	spinlock_t *ptl;
7da4d641	45	unsigned long pages = 0;
1da177e4	46
705e87c0	47	pte = pte_offset_map_lock(mm, pmd, addr, &ptl);
6606c3e0	48	arch_enter_lazy_mmu_mode();
1da177e4	49	do {
0697212a CL	50	oldpte = *pte;
0697212a CL	51	if (pte_present(oldpte)) {
1da177e4	52	pte_t ptent;
4b10e7d5	53	bool updated = false;
1da177e4	54
4b10e7d5	55	if (!prot_numa) {
0c5f83c2	56	ptent = ptep_modify_prot_start(mm, addr, pte);
1667918b MG	57	if (pte_numa(ptent))
1667918b MG	58	ptent = pte_mknonnuma(ptent);
4b10e7d5 MG	59	ptent = pte_modify(ptent, newprot);
	60	updated = true;
	61	} else {
	62	struct page *page;
	63
0c5f83c2	64	ptent = *pte;
4b10e7d5 MG	65	page = vm_normal_page(vma, addr, oldpte);
4b10e7d5 MG	66	if (page) {
1bc115d8	67	if (!pte_numa(oldpte)) {
4b10e7d5	68	ptent = pte_mknuma(ptent);
0c5f83c2	69	set_pte_at(mm, addr, pte, ptent);
4b10e7d5 MG	70	updated = true;
	71	}
	72	}
	73	}
1ea0704e	74
c1e6098b PZ	75	/*
	76	* Avoid taking write faults for pages we know to be
	77	* dirty.
	78	*/
4b10e7d5	79	if (dirty_accountable && pte_dirty(ptent)) {
c1e6098b	80	ptent = pte_mkwrite(ptent);
4b10e7d5 MG	81	updated = true;
4b10e7d5 MG	82	}
1ea0704e	83
4b10e7d5 MG	84	if (updated)
4b10e7d5 MG	85	pages++;
0c5f83c2 MG	86
	87	/* Only !prot_numa always clears the pte */
	88	if (!prot_numa)
	89	ptep_modify_prot_commit(mm, addr, pte, ptent);
ce1744f4	90	} else if (IS_ENABLED(CONFIG_MIGRATION) && !pte_file(oldpte)) {
0697212a CL	91	swp_entry_t entry = pte_to_swp_entry(oldpte);
	92
	93	if (is_write_migration_entry(entry)) {
c3d16e16	94	pte_t newpte;
0697212a CL	95	/*
	96	* A protection check is difficult so
	97	* just be safe and disable write
	98	*/
	99	make_migration_entry_read(&entry);
c3d16e16 CG	100	newpte = swp_entry_to_pte(entry);
	101	if (pte_swp_soft_dirty(oldpte))
	102	newpte = pte_swp_mksoft_dirty(newpte);
	103	set_pte_at(mm, addr, pte, newpte);
e920e14c MG	104
e920e14c MG	105	pages++;
0697212a	106	}
1da177e4 LT	107	}
1da177e4 LT	108	} while (pte++, addr += PAGE_SIZE, addr != end);
6606c3e0	109	arch_leave_lazy_mmu_mode();
705e87c0	110	pte_unmap_unlock(pte - 1, ptl);
7da4d641 PZ	111
7da4d641 PZ	112	return pages;
1da177e4 LT	113	}
1da177e4 LT	114
7d12efae AM	115	static inline unsigned long change_pmd_range(struct vm_area_struct *vma,
	116	pud_t *pud, unsigned long addr, unsigned long end,
	117	pgprot_t newprot, int dirty_accountable, int prot_numa)
1da177e4 LT	118	{
	119	pmd_t *pmd;
	120	unsigned long next;
7da4d641	121	unsigned long pages = 0;
72403b4a	122	unsigned long nr_huge_updates = 0;
1da177e4 LT	123
	124	pmd = pmd_offset(pud, addr);
	125	do {
25cbbef1 MG	126	unsigned long this_pages;
25cbbef1 MG	127
1da177e4	128	next = pmd_addr_end(addr, end);
cd7548ab JW	129	if (pmd_trans_huge(*pmd)) {
cd7548ab JW	130	if (next - addr != HPAGE_PMD_SIZE)
e180377f	131	split_huge_page_pmd(vma, addr, pmd);
f123d74a MG	132	else {
	133	int nr_ptes = change_huge_pmd(vma, pmd, addr,
	134	newprot, prot_numa);
	135
	136	if (nr_ptes) {
72403b4a MG	137	if (nr_ptes == HPAGE_PMD_NR) {
	138	pages += HPAGE_PMD_NR;
	139	nr_huge_updates++;
	140	}
f123d74a MG	141	continue;
f123d74a MG	142	}
7da4d641	143	}
cd7548ab JW	144	/* fall through */
cd7548ab JW	145	}
1da177e4 LT	146	if (pmd_none_or_clear_bad(pmd))
1da177e4 LT	147	continue;
25cbbef1	148	this_pages = change_pte_range(vma, pmd, addr, next, newprot,
0f19c179	149	dirty_accountable, prot_numa);
25cbbef1	150	pages += this_pages;
1da177e4	151	} while (pmd++, addr = next, addr != end);
7da4d641	152
72403b4a MG	153	if (nr_huge_updates)
72403b4a MG	154	count_vm_numa_events(NUMA_HUGE_PTE_UPDATES, nr_huge_updates);
7da4d641	155	return pages;
1da177e4 LT	156	}
1da177e4 LT	157
7d12efae AM	158	static inline unsigned long change_pud_range(struct vm_area_struct *vma,
	159	pgd_t *pgd, unsigned long addr, unsigned long end,
	160	pgprot_t newprot, int dirty_accountable, int prot_numa)
1da177e4 LT	161	{
	162	pud_t *pud;
	163	unsigned long next;
7da4d641	164	unsigned long pages = 0;
1da177e4 LT	165
	166	pud = pud_offset(pgd, addr);
	167	do {
	168	next = pud_addr_end(addr, end);
	169	if (pud_none_or_clear_bad(pud))
	170	continue;
7da4d641	171	pages += change_pmd_range(vma, pud, addr, next, newprot,
4b10e7d5	172	dirty_accountable, prot_numa);
1da177e4	173	} while (pud++, addr = next, addr != end);
7da4d641 PZ	174
7da4d641 PZ	175	return pages;
1da177e4 LT	176	}
1da177e4 LT	177
7da4d641	178	static unsigned long change_protection_range(struct vm_area_struct *vma,
c1e6098b	179	unsigned long addr, unsigned long end, pgprot_t newprot,
4b10e7d5	180	int dirty_accountable, int prot_numa)
1da177e4 LT	181	{
	182	struct mm_struct *mm = vma->vm_mm;
	183	pgd_t *pgd;
	184	unsigned long next;
	185	unsigned long start = addr;
7da4d641	186	unsigned long pages = 0;
1da177e4 LT	187
	188	BUG_ON(addr >= end);
	189	pgd = pgd_offset(mm, addr);
	190	flush_cache_range(vma, addr, end);
20841405	191	set_tlb_flush_pending(mm);
1da177e4 LT	192	do {
	193	next = pgd_addr_end(addr, end);
	194	if (pgd_none_or_clear_bad(pgd))
	195	continue;
7da4d641	196	pages += change_pud_range(vma, pgd, addr, next, newprot,
4b10e7d5	197	dirty_accountable, prot_numa);
1da177e4	198	} while (pgd++, addr = next, addr != end);
7da4d641	199
1233d588 IM	200	/* Only flush the TLB if we actually modified any entries: */
	201	if (pages)
	202	flush_tlb_range(vma, start, end);
20841405	203	clear_tlb_flush_pending(mm);
7da4d641 PZ	204
	205	return pages;
	206	}
	207
	208	unsigned long change_protection(struct vm_area_struct *vma, unsigned long start,
	209	unsigned long end, pgprot_t newprot,
4b10e7d5	210	int dirty_accountable, int prot_numa)
7da4d641 PZ	211	{
	212	struct mm_struct *mm = vma->vm_mm;
	213	unsigned long pages;
	214
	215	mmu_notifier_invalidate_range_start(mm, start, end);
	216	if (is_vm_hugetlb_page(vma))
	217	pages = hugetlb_change_protection(vma, start, end, newprot);
	218	else
4b10e7d5	219	pages = change_protection_range(vma, start, end, newprot, dirty_accountable, prot_numa);
7da4d641 PZ	220	mmu_notifier_invalidate_range_end(mm, start, end);
	221
	222	return pages;
1da177e4 LT	223	}
1da177e4 LT	224
b6a2fea3	225	int
1da177e4 LT	226	mprotect_fixup(struct vm_area_struct vma, struct vm_area_struct *pprev,
	227	unsigned long start, unsigned long end, unsigned long newflags)
	228	{
	229	struct mm_struct *mm = vma->vm_mm;
	230	unsigned long oldflags = vma->vm_flags;
	231	long nrpages = (end - start) >> PAGE_SHIFT;
	232	unsigned long charged = 0;
1da177e4 LT	233	pgoff_t pgoff;
1da177e4 LT	234	int error;
c1e6098b	235	int dirty_accountable = 0;
1da177e4 LT	236
	237	if (newflags == oldflags) {
	238	*pprev = vma;
	239	return 0;
	240	}
	241
	242	/*
	243	* If we make a private mapping writable we increase our commit;
	244	* but (without finer accounting) cannot reduce our commit if we
5a6fe125 MG	245	* make it unwritable again. hugetlb mapping were accounted for
5a6fe125 MG	246	* even if read-only so there is no need to account for them here
1da177e4 LT	247	*/
1da177e4 LT	248	if (newflags & VM_WRITE) {
5a6fe125	249	if (!(oldflags & (VM_ACCOUNT\|VM_WRITE\|VM_HUGETLB\|
cdfd4325	250	VM_SHARED\|VM_NORESERVE))) {
1da177e4	251	charged = nrpages;
191c5424	252	if (security_vm_enough_memory_mm(mm, charged))
1da177e4 LT	253	return -ENOMEM;
	254	newflags \|= VM_ACCOUNT;
	255	}
	256	}
	257
1da177e4 LT	258	/*
	259	* First try to merge with previous and/or next vma.
	260	*/
	261	pgoff = vma->vm_pgoff + ((start - vma->vm_start) >> PAGE_SHIFT);
	262	pprev = vma_merge(mm, pprev, start, end, newflags,
	263	vma->anon_vma, vma->vm_file, pgoff, vma_policy(vma));
	264	if (*pprev) {
	265	vma = *pprev;
	266	goto success;
	267	}
	268
	269	*pprev = vma;
	270
	271	if (start != vma->vm_start) {
	272	error = split_vma(mm, vma, start, 1);
	273	if (error)
	274	goto fail;
	275	}
	276
	277	if (end != vma->vm_end) {
	278	error = split_vma(mm, vma, end, 0);
	279	if (error)
	280	goto fail;
	281	}
	282
	283	success:
	284	/*
	285	* vm_flags and vm_page_prot are protected by the mmap_sem
	286	* held in write mode.
	287	*/
	288	vma->vm_flags = newflags;
1c12c4cf VP	289	vma->vm_page_prot = pgprot_modify(vma->vm_page_prot,
	290	vm_get_page_prot(newflags));
	291
c1e6098b	292	if (vma_wants_writenotify(vma)) {
1ddd439e	293	vma->vm_page_prot = vm_get_page_prot(newflags & ~VM_SHARED);
c1e6098b PZ	294	dirty_accountable = 1;
c1e6098b PZ	295	}
d08b3851	296
7d12efae AM	297	change_protection(vma, start, end, vma->vm_page_prot,
7d12efae AM	298	dirty_accountable, 0);
7da4d641	299
ab50b8ed HD	300	vm_stat_account(mm, oldflags, vma->vm_file, -nrpages);
ab50b8ed HD	301	vm_stat_account(mm, newflags, vma->vm_file, nrpages);
63bfd738	302	perf_event_mmap(vma);
1da177e4 LT	303	return 0;
	304
	305	fail:
	306	vm_unacct_memory(charged);
	307	return error;
	308	}
	309
6a6160a7 HC	310	SYSCALL_DEFINE3(mprotect, unsigned long, start, size_t, len,
6a6160a7 HC	311	unsigned long, prot)
1da177e4 LT	312	{
	313	unsigned long vm_flags, nstart, end, tmp, reqprot;
	314	struct vm_area_struct vma, prev;
	315	int error = -EINVAL;
	316	const int grows = prot & (PROT_GROWSDOWN\|PROT_GROWSUP);
	317	prot &= ~(PROT_GROWSDOWN\|PROT_GROWSUP);
	318	if (grows == (PROT_GROWSDOWN\|PROT_GROWSUP)) /* can't be both */
	319	return -EINVAL;
	320
	321	if (start & ~PAGE_MASK)
	322	return -EINVAL;
	323	if (!len)
	324	return 0;
	325	len = PAGE_ALIGN(len);
	326	end = start + len;
	327	if (end <= start)
	328	return -ENOMEM;
b845f313	329	if (!arch_validate_prot(prot))
1da177e4 LT	330	return -EINVAL;
	331
	332	reqprot = prot;
	333	/*
	334	* Does the application expect PROT_READ to imply PROT_EXEC:
	335	*/
b344e05c	336	if ((prot & PROT_READ) && (current->personality & READ_IMPLIES_EXEC))
1da177e4 LT	337	prot \|= PROT_EXEC;
	338
	339	vm_flags = calc_vm_prot_bits(prot);
	340
	341	down_write(&current->mm->mmap_sem);
	342
097d5910	343	vma = find_vma(current->mm, start);
1da177e4 LT	344	error = -ENOMEM;
	345	if (!vma)
	346	goto out;
097d5910	347	prev = vma->vm_prev;
1da177e4 LT	348	if (unlikely(grows & PROT_GROWSDOWN)) {
	349	if (vma->vm_start >= end)
	350	goto out;
	351	start = vma->vm_start;
	352	error = -EINVAL;
	353	if (!(vma->vm_flags & VM_GROWSDOWN))
	354	goto out;
7d12efae	355	} else {
1da177e4 LT	356	if (vma->vm_start > start)
	357	goto out;
	358	if (unlikely(grows & PROT_GROWSUP)) {
	359	end = vma->vm_end;
	360	error = -EINVAL;
	361	if (!(vma->vm_flags & VM_GROWSUP))
	362	goto out;
	363	}
	364	}
	365	if (start > vma->vm_start)
	366	prev = vma;
	367
	368	for (nstart = start ; ; ) {
	369	unsigned long newflags;
	370
7d12efae	371	/* Here we know that vma->vm_start <= nstart < vma->vm_end. */
1da177e4	372
7d12efae AM	373	newflags = vm_flags;
7d12efae AM	374	newflags \|= (vma->vm_flags & ~(VM_READ \| VM_WRITE \| VM_EXEC));
1da177e4	375
7e2cff42 PBG	376	/* newflags >> 4 shift VM_MAY% in place of VM_% */
7e2cff42 PBG	377	if ((newflags & ~(newflags >> 4)) & (VM_READ \| VM_WRITE \| VM_EXEC)) {
1da177e4 LT	378	error = -EACCES;
	379	goto out;
	380	}
	381
	382	error = security_file_mprotect(vma, reqprot, prot);
	383	if (error)
	384	goto out;
	385
	386	tmp = vma->vm_end;
	387	if (tmp > end)
	388	tmp = end;
	389	error = mprotect_fixup(vma, &prev, nstart, tmp, newflags);
	390	if (error)
	391	goto out;
	392	nstart = tmp;
	393
	394	if (nstart < prev->vm_end)
	395	nstart = prev->vm_end;
	396	if (nstart >= end)
	397	goto out;
	398
	399	vma = prev->vm_next;
	400	if (!vma \|\| vma->vm_start != nstart) {
	401	error = -ENOMEM;
	402	goto out;
	403	}
	404	}
	405	out:
	406	up_write(&current->mm->mmap_sem);
	407	return error;
	408	}