[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 <linux/ksm.h>
#include <asm/uaccess.h>
#include <asm/pgtable.h>
#include <asm/cacheflush.h>
#include <asm/tlbflush.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 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) &&
				    (pte_soft_dirty(ptent) ||
				     !(vma->vm_flags & VM_SOFTDIRTY)))
					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))
			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)) {
			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 */
					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) {
		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;
	dirty_accountable = vma_wants_writenotify(vma);
	vma_set_page_prot(vma);

	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
1ad9f620 MG	32	/*
	33	* For a prot_numa update we only hold mmap_sem for read so there is a
	34	* potential race with faulting where a pmd was temporarily none. This
	35	* function checks for a transhuge pmd under the appropriate lock. It
	36	* returns a pte if it was successfully locked or NULL if it raced with
	37	* a transhuge insertion.
	38	*/
	39	static pte_t lock_pte_protection(struct vm_area_struct vma, pmd_t *pmd,
	40	unsigned long addr, int prot_numa, spinlock_t **ptl)
	41	{
	42	pte_t *pte;
	43	spinlock_t *pmdl;
	44
	45	/* !prot_numa is protected by mmap_sem held for write */
	46	if (!prot_numa)
	47	return pte_offset_map_lock(vma->vm_mm, pmd, addr, ptl);
	48
	49	pmdl = pmd_lock(vma->vm_mm, pmd);
	50	if (unlikely(pmd_trans_huge(pmd) \|\| pmd_none(pmd))) {
	51	spin_unlock(pmdl);
	52	return NULL;
	53	}
	54
	55	pte = pte_offset_map_lock(vma->vm_mm, pmd, addr, ptl);
	56	spin_unlock(pmdl);
	57	return pte;
	58	}
	59
4b10e7d5	60	static unsigned long change_pte_range(struct vm_area_struct vma, pmd_t pmd,
c1e6098b	61	unsigned long addr, unsigned long end, pgprot_t newprot,
0f19c179	62	int dirty_accountable, int prot_numa)
1da177e4	63	{
4b10e7d5	64	struct mm_struct *mm = vma->vm_mm;
0697212a	65	pte_t *pte, oldpte;
705e87c0	66	spinlock_t *ptl;
7da4d641	67	unsigned long pages = 0;
1da177e4	68
1ad9f620 MG	69	pte = lock_pte_protection(vma, pmd, addr, prot_numa, &ptl);
	70	if (!pte)
	71	return 0;
	72
6606c3e0	73	arch_enter_lazy_mmu_mode();
1da177e4	74	do {
0697212a CL	75	oldpte = *pte;
0697212a CL	76	if (pte_present(oldpte)) {
1da177e4	77	pte_t ptent;
4b10e7d5	78	bool updated = false;
1da177e4	79
4b10e7d5	80	if (!prot_numa) {
0c5f83c2	81	ptent = ptep_modify_prot_start(mm, addr, pte);
1667918b MG	82	if (pte_numa(ptent))
1667918b MG	83	ptent = pte_mknonnuma(ptent);
4b10e7d5	84	ptent = pte_modify(ptent, newprot);
9d85d586 AK	85	/*
	86	* Avoid taking write faults for pages we
	87	* know to be dirty.
	88	*/
64e45507 PF	89	if (dirty_accountable && pte_dirty(ptent) &&
	90	(pte_soft_dirty(ptent) \|\|
	91	!(vma->vm_flags & VM_SOFTDIRTY)))
9d85d586 AK	92	ptent = pte_mkwrite(ptent);
9d85d586 AK	93	ptep_modify_prot_commit(mm, addr, pte, ptent);
4b10e7d5 MG	94	updated = true;
	95	} else {
	96	struct page *page;
	97
	98	page = vm_normal_page(vma, addr, oldpte);
64a9a34e	99	if (page && !PageKsm(page)) {
1bc115d8	100	if (!pte_numa(oldpte)) {
56eecdb9	101	ptep_set_numa(mm, addr, pte);
4b10e7d5 MG	102	updated = true;
	103	}
	104	}
	105	}
4b10e7d5 MG	106	if (updated)
4b10e7d5 MG	107	pages++;
0661a336	108	} else if (IS_ENABLED(CONFIG_MIGRATION)) {
0697212a CL	109	swp_entry_t entry = pte_to_swp_entry(oldpte);
	110
	111	if (is_write_migration_entry(entry)) {
c3d16e16	112	pte_t newpte;
0697212a CL	113	/*
	114	* A protection check is difficult so
	115	* just be safe and disable write
	116	*/
	117	make_migration_entry_read(&entry);
c3d16e16 CG	118	newpte = swp_entry_to_pte(entry);
	119	if (pte_swp_soft_dirty(oldpte))
	120	newpte = pte_swp_mksoft_dirty(newpte);
	121	set_pte_at(mm, addr, pte, newpte);
e920e14c MG	122
e920e14c MG	123	pages++;
0697212a	124	}
1da177e4 LT	125	}
1da177e4 LT	126	} while (pte++, addr += PAGE_SIZE, addr != end);
6606c3e0	127	arch_leave_lazy_mmu_mode();
705e87c0	128	pte_unmap_unlock(pte - 1, ptl);
7da4d641 PZ	129
7da4d641 PZ	130	return pages;
1da177e4 LT	131	}
1da177e4 LT	132
7d12efae AM	133	static inline unsigned long change_pmd_range(struct vm_area_struct *vma,
	134	pud_t *pud, unsigned long addr, unsigned long end,
	135	pgprot_t newprot, int dirty_accountable, int prot_numa)
1da177e4 LT	136	{
1da177e4 LT	137	pmd_t *pmd;
a5338093	138	struct mm_struct *mm = vma->vm_mm;
1da177e4	139	unsigned long next;
7da4d641	140	unsigned long pages = 0;
72403b4a	141	unsigned long nr_huge_updates = 0;
a5338093	142	unsigned long mni_start = 0;
1da177e4 LT	143
	144	pmd = pmd_offset(pud, addr);
	145	do {
25cbbef1 MG	146	unsigned long this_pages;
25cbbef1 MG	147
1da177e4	148	next = pmd_addr_end(addr, end);
88a9ab6e RR	149	if (!pmd_trans_huge(*pmd) && pmd_none_or_clear_bad(pmd))
88a9ab6e RR	150	continue;
a5338093 RR	151
	152	/* invoke the mmu notifier if the pmd is populated */
	153	if (!mni_start) {
	154	mni_start = addr;
	155	mmu_notifier_invalidate_range_start(mm, mni_start, end);
	156	}
	157
cd7548ab JW	158	if (pmd_trans_huge(*pmd)) {
cd7548ab JW	159	if (next - addr != HPAGE_PMD_SIZE)
e180377f	160	split_huge_page_pmd(vma, addr, pmd);
f123d74a MG	161	else {
	162	int nr_ptes = change_huge_pmd(vma, pmd, addr,
	163	newprot, prot_numa);
	164
	165	if (nr_ptes) {
72403b4a MG	166	if (nr_ptes == HPAGE_PMD_NR) {
	167	pages += HPAGE_PMD_NR;
	168	nr_huge_updates++;
	169	}
1ad9f620 MG	170
1ad9f620 MG	171	/* huge pmd was handled */
f123d74a MG	172	continue;
f123d74a MG	173	}
7da4d641	174	}
88a9ab6e	175	/* fall through, the trans huge pmd just split */
cd7548ab	176	}
25cbbef1	177	this_pages = change_pte_range(vma, pmd, addr, next, newprot,
0f19c179	178	dirty_accountable, prot_numa);
25cbbef1	179	pages += this_pages;
1da177e4	180	} while (pmd++, addr = next, addr != end);
7da4d641	181
a5338093 RR	182	if (mni_start)
	183	mmu_notifier_invalidate_range_end(mm, mni_start, end);
	184
72403b4a MG	185	if (nr_huge_updates)
72403b4a MG	186	count_vm_numa_events(NUMA_HUGE_PTE_UPDATES, nr_huge_updates);
7da4d641	187	return pages;
1da177e4 LT	188	}
1da177e4 LT	189
7d12efae AM	190	static inline unsigned long change_pud_range(struct vm_area_struct *vma,
	191	pgd_t *pgd, unsigned long addr, unsigned long end,
	192	pgprot_t newprot, int dirty_accountable, int prot_numa)
1da177e4 LT	193	{
	194	pud_t *pud;
	195	unsigned long next;
7da4d641	196	unsigned long pages = 0;
1da177e4 LT	197
	198	pud = pud_offset(pgd, addr);
	199	do {
	200	next = pud_addr_end(addr, end);
	201	if (pud_none_or_clear_bad(pud))
	202	continue;
7da4d641	203	pages += change_pmd_range(vma, pud, addr, next, newprot,
4b10e7d5	204	dirty_accountable, prot_numa);
1da177e4	205	} while (pud++, addr = next, addr != end);
7da4d641 PZ	206
7da4d641 PZ	207	return pages;
1da177e4 LT	208	}
1da177e4 LT	209
7da4d641	210	static unsigned long change_protection_range(struct vm_area_struct *vma,
c1e6098b	211	unsigned long addr, unsigned long end, pgprot_t newprot,
4b10e7d5	212	int dirty_accountable, int prot_numa)
1da177e4 LT	213	{
	214	struct mm_struct *mm = vma->vm_mm;
	215	pgd_t *pgd;
	216	unsigned long next;
	217	unsigned long start = addr;
7da4d641	218	unsigned long pages = 0;
1da177e4 LT	219
	220	BUG_ON(addr >= end);
	221	pgd = pgd_offset(mm, addr);
	222	flush_cache_range(vma, addr, end);
20841405	223	set_tlb_flush_pending(mm);
1da177e4 LT	224	do {
	225	next = pgd_addr_end(addr, end);
	226	if (pgd_none_or_clear_bad(pgd))
	227	continue;
7da4d641	228	pages += change_pud_range(vma, pgd, addr, next, newprot,
4b10e7d5	229	dirty_accountable, prot_numa);
1da177e4	230	} while (pgd++, addr = next, addr != end);
7da4d641	231
1233d588 IM	232	/* Only flush the TLB if we actually modified any entries: */
	233	if (pages)
	234	flush_tlb_range(vma, start, end);
20841405	235	clear_tlb_flush_pending(mm);
7da4d641 PZ	236
	237	return pages;
	238	}
	239
	240	unsigned long change_protection(struct vm_area_struct *vma, unsigned long start,
	241	unsigned long end, pgprot_t newprot,
4b10e7d5	242	int dirty_accountable, int prot_numa)
7da4d641	243	{
7da4d641 PZ	244	unsigned long pages;
7da4d641 PZ	245
7da4d641 PZ	246	if (is_vm_hugetlb_page(vma))
	247	pages = hugetlb_change_protection(vma, start, end, newprot);
	248	else
4b10e7d5	249	pages = change_protection_range(vma, start, end, newprot, dirty_accountable, prot_numa);
7da4d641 PZ	250
7da4d641 PZ	251	return pages;
1da177e4 LT	252	}
1da177e4 LT	253
b6a2fea3	254	int
1da177e4 LT	255	mprotect_fixup(struct vm_area_struct vma, struct vm_area_struct *pprev,
	256	unsigned long start, unsigned long end, unsigned long newflags)
	257	{
	258	struct mm_struct *mm = vma->vm_mm;
	259	unsigned long oldflags = vma->vm_flags;
	260	long nrpages = (end - start) >> PAGE_SHIFT;
	261	unsigned long charged = 0;
1da177e4 LT	262	pgoff_t pgoff;
1da177e4 LT	263	int error;
c1e6098b	264	int dirty_accountable = 0;
1da177e4 LT	265
	266	if (newflags == oldflags) {
	267	*pprev = vma;
	268	return 0;
	269	}
	270
	271	/*
	272	* If we make a private mapping writable we increase our commit;
	273	* but (without finer accounting) cannot reduce our commit if we
5a6fe125 MG	274	* make it unwritable again. hugetlb mapping were accounted for
5a6fe125 MG	275	* even if read-only so there is no need to account for them here
1da177e4 LT	276	*/
1da177e4 LT	277	if (newflags & VM_WRITE) {
5a6fe125	278	if (!(oldflags & (VM_ACCOUNT\|VM_WRITE\|VM_HUGETLB\|
cdfd4325	279	VM_SHARED\|VM_NORESERVE))) {
1da177e4	280	charged = nrpages;
191c5424	281	if (security_vm_enough_memory_mm(mm, charged))
1da177e4 LT	282	return -ENOMEM;
	283	newflags \|= VM_ACCOUNT;
	284	}
	285	}
	286
1da177e4 LT	287	/*
	288	* First try to merge with previous and/or next vma.
	289	*/
	290	pgoff = vma->vm_pgoff + ((start - vma->vm_start) >> PAGE_SHIFT);
	291	pprev = vma_merge(mm, pprev, start, end, newflags,
	292	vma->anon_vma, vma->vm_file, pgoff, vma_policy(vma));
	293	if (*pprev) {
	294	vma = *pprev;
	295	goto success;
	296	}
	297
	298	*pprev = vma;
	299
	300	if (start != vma->vm_start) {
	301	error = split_vma(mm, vma, start, 1);
	302	if (error)
	303	goto fail;
	304	}
	305
	306	if (end != vma->vm_end) {
	307	error = split_vma(mm, vma, end, 0);
	308	if (error)
	309	goto fail;
	310	}
	311
	312	success:
	313	/*
	314	* vm_flags and vm_page_prot are protected by the mmap_sem
	315	* held in write mode.
	316	*/
	317	vma->vm_flags = newflags;
64e45507 PF	318	dirty_accountable = vma_wants_writenotify(vma);
64e45507 PF	319	vma_set_page_prot(vma);
d08b3851	320
7d12efae AM	321	change_protection(vma, start, end, vma->vm_page_prot,
7d12efae AM	322	dirty_accountable, 0);
7da4d641	323
ab50b8ed HD	324	vm_stat_account(mm, oldflags, vma->vm_file, -nrpages);
ab50b8ed HD	325	vm_stat_account(mm, newflags, vma->vm_file, nrpages);
63bfd738	326	perf_event_mmap(vma);
1da177e4 LT	327	return 0;
	328
	329	fail:
	330	vm_unacct_memory(charged);
	331	return error;
	332	}
	333
6a6160a7 HC	334	SYSCALL_DEFINE3(mprotect, unsigned long, start, size_t, len,
6a6160a7 HC	335	unsigned long, prot)
1da177e4 LT	336	{
	337	unsigned long vm_flags, nstart, end, tmp, reqprot;
	338	struct vm_area_struct vma, prev;
	339	int error = -EINVAL;
	340	const int grows = prot & (PROT_GROWSDOWN\|PROT_GROWSUP);
	341	prot &= ~(PROT_GROWSDOWN\|PROT_GROWSUP);
	342	if (grows == (PROT_GROWSDOWN\|PROT_GROWSUP)) /* can't be both */
	343	return -EINVAL;
	344
	345	if (start & ~PAGE_MASK)
	346	return -EINVAL;
	347	if (!len)
	348	return 0;
	349	len = PAGE_ALIGN(len);
	350	end = start + len;
	351	if (end <= start)
	352	return -ENOMEM;
b845f313	353	if (!arch_validate_prot(prot))
1da177e4 LT	354	return -EINVAL;
	355
	356	reqprot = prot;
	357	/*
	358	* Does the application expect PROT_READ to imply PROT_EXEC:
	359	*/
b344e05c	360	if ((prot & PROT_READ) && (current->personality & READ_IMPLIES_EXEC))
1da177e4 LT	361	prot \|= PROT_EXEC;
	362
	363	vm_flags = calc_vm_prot_bits(prot);
	364
	365	down_write(&current->mm->mmap_sem);
	366
097d5910	367	vma = find_vma(current->mm, start);
1da177e4 LT	368	error = -ENOMEM;
	369	if (!vma)
	370	goto out;
097d5910	371	prev = vma->vm_prev;
1da177e4 LT	372	if (unlikely(grows & PROT_GROWSDOWN)) {
	373	if (vma->vm_start >= end)
	374	goto out;
	375	start = vma->vm_start;
	376	error = -EINVAL;
	377	if (!(vma->vm_flags & VM_GROWSDOWN))
	378	goto out;
7d12efae	379	} else {
1da177e4 LT	380	if (vma->vm_start > start)
	381	goto out;
	382	if (unlikely(grows & PROT_GROWSUP)) {
	383	end = vma->vm_end;
	384	error = -EINVAL;
	385	if (!(vma->vm_flags & VM_GROWSUP))
	386	goto out;
	387	}
	388	}
	389	if (start > vma->vm_start)
	390	prev = vma;
	391
	392	for (nstart = start ; ; ) {
	393	unsigned long newflags;
	394
7d12efae	395	/* Here we know that vma->vm_start <= nstart < vma->vm_end. */
1da177e4	396
7d12efae AM	397	newflags = vm_flags;
7d12efae AM	398	newflags \|= (vma->vm_flags & ~(VM_READ \| VM_WRITE \| VM_EXEC));
1da177e4	399
7e2cff42 PBG	400	/* newflags >> 4 shift VM_MAY% in place of VM_% */
7e2cff42 PBG	401	if ((newflags & ~(newflags >> 4)) & (VM_READ \| VM_WRITE \| VM_EXEC)) {
1da177e4 LT	402	error = -EACCES;
	403	goto out;
	404	}
	405
	406	error = security_file_mprotect(vma, reqprot, prot);
	407	if (error)
	408	goto out;
	409
	410	tmp = vma->vm_end;
	411	if (tmp > end)
	412	tmp = end;
	413	error = mprotect_fixup(vma, &prev, nstart, tmp, newflags);
	414	if (error)
	415	goto out;
	416	nstart = tmp;
	417
	418	if (nstart < prev->vm_end)
	419	nstart = prev->vm_end;
	420	if (nstart >= end)
	421	goto out;
	422
	423	vma = prev->vm_next;
	424	if (!vma \|\| vma->vm_start != nstart) {
	425	error = -ENOMEM;
	426	goto out;
	427	}
	428	}
	429	out:
	430	up_write(&current->mm->mmap_sem);
	431	return error;
	432	}