[deliverable/linux.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, bool *ret_all_same_node)
{
	struct mm_struct *mm = vma->vm_mm;
	pte_t *pte, oldpte;
	spinlock_t *ptl;
	unsigned long pages = 0;
	bool all_same_node = true;
	int last_nid = -1;

	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;

			ptent = ptep_modify_prot_start(mm, addr, pte);
			if (!prot_numa) {
				ptent = pte_modify(ptent, newprot);
				updated = true;
			} else {
				struct page *page;

				page = vm_normal_page(vma, addr, oldpte);
				if (page) {
					int this_nid = page_to_nid(page);
					if (last_nid == -1)
						last_nid = this_nid;
					if (last_nid != this_nid)
						all_same_node = false;

					/* only check non-shared pages */
					if (!pte_numa(oldpte) &&
					    page_mapcount(page) == 1) {
						ptent = pte_mknuma(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++;
			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)) {
				/*
				 * A protection check is difficult so
				 * just be safe and disable write
				 */
				make_migration_entry_read(&entry);
				set_pte_at(mm, addr, pte,
					swp_entry_to_pte(entry));
			}
			pages++;
		}
	} while (pte++, addr += PAGE_SIZE, addr != end);
	arch_leave_lazy_mmu_mode();
	pte_unmap_unlock(pte - 1, ptl);

	*ret_all_same_node = all_same_node;
	return pages;
}

#ifdef CONFIG_NUMA_BALANCING
static inline void change_pmd_protnuma(struct mm_struct *mm, unsigned long addr,
				       pmd_t *pmd)
{
	spin_lock(&mm->page_table_lock);
	set_pmd_at(mm, addr & PMD_MASK, pmd, pmd_mknuma(*pmd));
	spin_unlock(&mm->page_table_lock);
}
#else
static inline void change_pmd_protnuma(struct mm_struct *mm, unsigned long addr,
				       pmd_t *pmd)
{
	BUG();
}
#endif /* CONFIG_NUMA_BALANCING */

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;
	bool all_same_node;

	pmd = pmd_offset(pud, addr);
	do {
		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 if (change_huge_pmd(vma, pmd, addr, newprot,
						 prot_numa)) {
				pages += HPAGE_PMD_NR;
				continue;
			}
			/* fall through */
		}
		if (pmd_none_or_clear_bad(pmd))
			continue;
		pages += change_pte_range(vma, pmd, addr, next, newprot,
				 dirty_accountable, prot_numa, &all_same_node);

		/*
		 * If we are changing protections for NUMA hinting faults then
		 * set pmd_numa if the examined pages were all on the same
		 * node. This allows a regular PMD to be handled as one fault
		 * and effectively batches the taking of the PTL
		 */
		if (prot_numa && all_same_node)
			change_pmd_protnuma(vma->vm_mm, addr, pmd);
	} while (pmd++, addr = next, addr != end);

	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);
	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);

	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,
9532fec1	40	int dirty_accountable, int prot_numa, bool *ret_all_same_node)
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;
9532fec1 MG	46	bool all_same_node = true;
9532fec1 MG	47	int last_nid = -1;
1da177e4	48
705e87c0	49	pte = pte_offset_map_lock(mm, pmd, addr, &ptl);
6606c3e0	50	arch_enter_lazy_mmu_mode();
1da177e4	51	do {
0697212a CL	52	oldpte = *pte;
0697212a CL	53	if (pte_present(oldpte)) {
1da177e4	54	pte_t ptent;
4b10e7d5	55	bool updated = false;
1da177e4	56
1ea0704e	57	ptent = ptep_modify_prot_start(mm, addr, pte);
4b10e7d5 MG	58	if (!prot_numa) {
	59	ptent = pte_modify(ptent, newprot);
	60	updated = true;
	61	} else {
	62	struct page *page;
	63
	64	page = vm_normal_page(vma, addr, oldpte);
	65	if (page) {
9532fec1 MG	66	int this_nid = page_to_nid(page);
	67	if (last_nid == -1)
	68	last_nid = this_nid;
	69	if (last_nid != this_nid)
	70	all_same_node = false;
	71
4b10e7d5 MG	72	/* only check non-shared pages */
	73	if (!pte_numa(oldpte) &&
	74	page_mapcount(page) == 1) {
	75	ptent = pte_mknuma(ptent);
	76	updated = true;
	77	}
	78	}
	79	}
1ea0704e	80
c1e6098b PZ	81	/*
	82	* Avoid taking write faults for pages we know to be
	83	* dirty.
	84	*/
4b10e7d5	85	if (dirty_accountable && pte_dirty(ptent)) {
c1e6098b	86	ptent = pte_mkwrite(ptent);
4b10e7d5 MG	87	updated = true;
4b10e7d5 MG	88	}
1ea0704e	89
4b10e7d5 MG	90	if (updated)
4b10e7d5 MG	91	pages++;
1ea0704e	92	ptep_modify_prot_commit(mm, addr, pte, ptent);
ce1744f4	93	} else if (IS_ENABLED(CONFIG_MIGRATION) && !pte_file(oldpte)) {
0697212a CL	94	swp_entry_t entry = pte_to_swp_entry(oldpte);
	95
	96	if (is_write_migration_entry(entry)) {
	97	/*
	98	* A protection check is difficult so
	99	* just be safe and disable write
	100	*/
	101	make_migration_entry_read(&entry);
	102	set_pte_at(mm, addr, pte,
	103	swp_entry_to_pte(entry));
	104	}
7da4d641	105	pages++;
1da177e4 LT	106	}
1da177e4 LT	107	} while (pte++, addr += PAGE_SIZE, addr != end);
6606c3e0	108	arch_leave_lazy_mmu_mode();
705e87c0	109	pte_unmap_unlock(pte - 1, ptl);
7da4d641	110
9532fec1	111	*ret_all_same_node = all_same_node;
7da4d641	112	return pages;
1da177e4 LT	113	}
1da177e4 LT	114
4b10e7d5 MG	115	#ifdef CONFIG_NUMA_BALANCING
4b10e7d5 MG	116	static inline void change_pmd_protnuma(struct mm_struct *mm, unsigned long addr,
7d12efae	117	pmd_t *pmd)
4b10e7d5 MG	118	{
	119	spin_lock(&mm->page_table_lock);
	120	set_pmd_at(mm, addr & PMD_MASK, pmd, pmd_mknuma(*pmd));
	121	spin_unlock(&mm->page_table_lock);
	122	}
	123	#else
	124	static inline void change_pmd_protnuma(struct mm_struct *mm, unsigned long addr,
7d12efae	125	pmd_t *pmd)
4b10e7d5 MG	126	{
4b10e7d5 MG	127	BUG();
1da177e4	128	}
4b10e7d5	129	#endif /* CONFIG_NUMA_BALANCING */
1da177e4	130
7d12efae AM	131	static inline unsigned long change_pmd_range(struct vm_area_struct *vma,
	132	pud_t *pud, unsigned long addr, unsigned long end,
	133	pgprot_t newprot, int dirty_accountable, int prot_numa)
1da177e4 LT	134	{
	135	pmd_t *pmd;
	136	unsigned long next;
7da4d641	137	unsigned long pages = 0;
9532fec1	138	bool all_same_node;
1da177e4 LT	139
	140	pmd = pmd_offset(pud, addr);
	141	do {
	142	next = pmd_addr_end(addr, end);
cd7548ab JW	143	if (pmd_trans_huge(*pmd)) {
cd7548ab JW	144	if (next - addr != HPAGE_PMD_SIZE)
e180377f	145	split_huge_page_pmd(vma, addr, pmd);
7d12efae AM	146	else if (change_huge_pmd(vma, pmd, addr, newprot,
7d12efae AM	147	prot_numa)) {
7da4d641	148	pages += HPAGE_PMD_NR;
cd7548ab	149	continue;
7da4d641	150	}
cd7548ab JW	151	/* fall through */
cd7548ab JW	152	}
1da177e4 LT	153	if (pmd_none_or_clear_bad(pmd))
1da177e4 LT	154	continue;
4b10e7d5	155	pages += change_pte_range(vma, pmd, addr, next, newprot,
9532fec1	156	dirty_accountable, prot_numa, &all_same_node);
4b10e7d5	157
9532fec1 MG	158	/*
	159	* If we are changing protections for NUMA hinting faults then
	160	* set pmd_numa if the examined pages were all on the same
	161	* node. This allows a regular PMD to be handled as one fault
	162	* and effectively batches the taking of the PTL
	163	*/
	164	if (prot_numa && all_same_node)
4b10e7d5	165	change_pmd_protnuma(vma->vm_mm, addr, pmd);
1da177e4	166	} while (pmd++, addr = next, addr != end);
7da4d641 PZ	167
7da4d641 PZ	168	return pages;
1da177e4 LT	169	}
1da177e4 LT	170
7d12efae AM	171	static inline unsigned long change_pud_range(struct vm_area_struct *vma,
	172	pgd_t *pgd, unsigned long addr, unsigned long end,
	173	pgprot_t newprot, int dirty_accountable, int prot_numa)
1da177e4 LT	174	{
	175	pud_t *pud;
	176	unsigned long next;
7da4d641	177	unsigned long pages = 0;
1da177e4 LT	178
	179	pud = pud_offset(pgd, addr);
	180	do {
	181	next = pud_addr_end(addr, end);
	182	if (pud_none_or_clear_bad(pud))
	183	continue;
7da4d641	184	pages += change_pmd_range(vma, pud, addr, next, newprot,
4b10e7d5	185	dirty_accountable, prot_numa);
1da177e4	186	} while (pud++, addr = next, addr != end);
7da4d641 PZ	187
7da4d641 PZ	188	return pages;
1da177e4 LT	189	}
1da177e4 LT	190
7da4d641	191	static unsigned long change_protection_range(struct vm_area_struct *vma,
c1e6098b	192	unsigned long addr, unsigned long end, pgprot_t newprot,
4b10e7d5	193	int dirty_accountable, int prot_numa)
1da177e4 LT	194	{
	195	struct mm_struct *mm = vma->vm_mm;
	196	pgd_t *pgd;
	197	unsigned long next;
	198	unsigned long start = addr;
7da4d641	199	unsigned long pages = 0;
1da177e4 LT	200
	201	BUG_ON(addr >= end);
	202	pgd = pgd_offset(mm, addr);
	203	flush_cache_range(vma, addr, end);
1da177e4 LT	204	do {
	205	next = pgd_addr_end(addr, end);
	206	if (pgd_none_or_clear_bad(pgd))
	207	continue;
7da4d641	208	pages += change_pud_range(vma, pgd, addr, next, newprot,
4b10e7d5	209	dirty_accountable, prot_numa);
1da177e4	210	} while (pgd++, addr = next, addr != end);
7da4d641	211
1233d588 IM	212	/* Only flush the TLB if we actually modified any entries: */
	213	if (pages)
	214	flush_tlb_range(vma, start, end);
7da4d641 PZ	215
	216	return pages;
	217	}
	218
	219	unsigned long change_protection(struct vm_area_struct *vma, unsigned long start,
	220	unsigned long end, pgprot_t newprot,
4b10e7d5	221	int dirty_accountable, int prot_numa)
7da4d641 PZ	222	{
	223	struct mm_struct *mm = vma->vm_mm;
	224	unsigned long pages;
	225
	226	mmu_notifier_invalidate_range_start(mm, start, end);
	227	if (is_vm_hugetlb_page(vma))
	228	pages = hugetlb_change_protection(vma, start, end, newprot);
	229	else
4b10e7d5	230	pages = change_protection_range(vma, start, end, newprot, dirty_accountable, prot_numa);
7da4d641 PZ	231	mmu_notifier_invalidate_range_end(mm, start, end);
	232
	233	return pages;
1da177e4 LT	234	}
1da177e4 LT	235
b6a2fea3	236	int
1da177e4 LT	237	mprotect_fixup(struct vm_area_struct vma, struct vm_area_struct *pprev,
	238	unsigned long start, unsigned long end, unsigned long newflags)
	239	{
	240	struct mm_struct *mm = vma->vm_mm;
	241	unsigned long oldflags = vma->vm_flags;
	242	long nrpages = (end - start) >> PAGE_SHIFT;
	243	unsigned long charged = 0;
1da177e4 LT	244	pgoff_t pgoff;
1da177e4 LT	245	int error;
c1e6098b	246	int dirty_accountable = 0;
1da177e4 LT	247
	248	if (newflags == oldflags) {
	249	*pprev = vma;
	250	return 0;
	251	}
	252
	253	/*
	254	* If we make a private mapping writable we increase our commit;
	255	* but (without finer accounting) cannot reduce our commit if we
5a6fe125 MG	256	* make it unwritable again. hugetlb mapping were accounted for
5a6fe125 MG	257	* even if read-only so there is no need to account for them here
1da177e4 LT	258	*/
1da177e4 LT	259	if (newflags & VM_WRITE) {
5a6fe125	260	if (!(oldflags & (VM_ACCOUNT\|VM_WRITE\|VM_HUGETLB\|
cdfd4325	261	VM_SHARED\|VM_NORESERVE))) {
1da177e4	262	charged = nrpages;
191c5424	263	if (security_vm_enough_memory_mm(mm, charged))
1da177e4 LT	264	return -ENOMEM;
	265	newflags \|= VM_ACCOUNT;
	266	}
	267	}
	268
1da177e4 LT	269	/*
	270	* First try to merge with previous and/or next vma.
	271	*/
	272	pgoff = vma->vm_pgoff + ((start - vma->vm_start) >> PAGE_SHIFT);
	273	pprev = vma_merge(mm, pprev, start, end, newflags,
	274	vma->anon_vma, vma->vm_file, pgoff, vma_policy(vma));
	275	if (*pprev) {
	276	vma = *pprev;
	277	goto success;
	278	}
	279
	280	*pprev = vma;
	281
	282	if (start != vma->vm_start) {
	283	error = split_vma(mm, vma, start, 1);
	284	if (error)
	285	goto fail;
	286	}
	287
	288	if (end != vma->vm_end) {
	289	error = split_vma(mm, vma, end, 0);
	290	if (error)
	291	goto fail;
	292	}
	293
	294	success:
	295	/*
	296	* vm_flags and vm_page_prot are protected by the mmap_sem
	297	* held in write mode.
	298	*/
	299	vma->vm_flags = newflags;
1c12c4cf VP	300	vma->vm_page_prot = pgprot_modify(vma->vm_page_prot,
	301	vm_get_page_prot(newflags));
	302
c1e6098b	303	if (vma_wants_writenotify(vma)) {
1ddd439e	304	vma->vm_page_prot = vm_get_page_prot(newflags & ~VM_SHARED);
c1e6098b PZ	305	dirty_accountable = 1;
c1e6098b PZ	306	}
d08b3851	307
7d12efae AM	308	change_protection(vma, start, end, vma->vm_page_prot,
7d12efae AM	309	dirty_accountable, 0);
7da4d641	310
ab50b8ed HD	311	vm_stat_account(mm, oldflags, vma->vm_file, -nrpages);
ab50b8ed HD	312	vm_stat_account(mm, newflags, vma->vm_file, nrpages);
63bfd738	313	perf_event_mmap(vma);
1da177e4 LT	314	return 0;
	315
	316	fail:
	317	vm_unacct_memory(charged);
	318	return error;
	319	}
	320
6a6160a7 HC	321	SYSCALL_DEFINE3(mprotect, unsigned long, start, size_t, len,
6a6160a7 HC	322	unsigned long, prot)
1da177e4 LT	323	{
	324	unsigned long vm_flags, nstart, end, tmp, reqprot;
	325	struct vm_area_struct vma, prev;
	326	int error = -EINVAL;
	327	const int grows = prot & (PROT_GROWSDOWN\|PROT_GROWSUP);
	328	prot &= ~(PROT_GROWSDOWN\|PROT_GROWSUP);
	329	if (grows == (PROT_GROWSDOWN\|PROT_GROWSUP)) /* can't be both */
	330	return -EINVAL;
	331
	332	if (start & ~PAGE_MASK)
	333	return -EINVAL;
	334	if (!len)
	335	return 0;
	336	len = PAGE_ALIGN(len);
	337	end = start + len;
	338	if (end <= start)
	339	return -ENOMEM;
b845f313	340	if (!arch_validate_prot(prot))
1da177e4 LT	341	return -EINVAL;
	342
	343	reqprot = prot;
	344	/*
	345	* Does the application expect PROT_READ to imply PROT_EXEC:
	346	*/
b344e05c	347	if ((prot & PROT_READ) && (current->personality & READ_IMPLIES_EXEC))
1da177e4 LT	348	prot \|= PROT_EXEC;
	349
	350	vm_flags = calc_vm_prot_bits(prot);
	351
	352	down_write(&current->mm->mmap_sem);
	353
097d5910	354	vma = find_vma(current->mm, start);
1da177e4 LT	355	error = -ENOMEM;
	356	if (!vma)
	357	goto out;
097d5910	358	prev = vma->vm_prev;
1da177e4 LT	359	if (unlikely(grows & PROT_GROWSDOWN)) {
	360	if (vma->vm_start >= end)
	361	goto out;
	362	start = vma->vm_start;
	363	error = -EINVAL;
	364	if (!(vma->vm_flags & VM_GROWSDOWN))
	365	goto out;
7d12efae	366	} else {
1da177e4 LT	367	if (vma->vm_start > start)
	368	goto out;
	369	if (unlikely(grows & PROT_GROWSUP)) {
	370	end = vma->vm_end;
	371	error = -EINVAL;
	372	if (!(vma->vm_flags & VM_GROWSUP))
	373	goto out;
	374	}
	375	}
	376	if (start > vma->vm_start)
	377	prev = vma;
	378
	379	for (nstart = start ; ; ) {
	380	unsigned long newflags;
	381
7d12efae	382	/* Here we know that vma->vm_start <= nstart < vma->vm_end. */
1da177e4	383
7d12efae AM	384	newflags = vm_flags;
7d12efae AM	385	newflags \|= (vma->vm_flags & ~(VM_READ \| VM_WRITE \| VM_EXEC));
1da177e4	386
7e2cff42 PBG	387	/* newflags >> 4 shift VM_MAY% in place of VM_% */
7e2cff42 PBG	388	if ((newflags & ~(newflags >> 4)) & (VM_READ \| VM_WRITE \| VM_EXEC)) {
1da177e4 LT	389	error = -EACCES;
	390	goto out;
	391	}
	392
	393	error = security_file_mprotect(vma, reqprot, prot);
	394	if (error)
	395	goto out;
	396
	397	tmp = vma->vm_end;
	398	if (tmp > end)
	399	tmp = end;
	400	error = mprotect_fixup(vma, &prev, nstart, tmp, newflags);
	401	if (error)
	402	goto out;
	403	nstart = tmp;
	404
	405	if (nstart < prev->vm_end)
	406	nstart = prev->vm_end;
	407	if (nstart >= end)
	408	goto out;
	409
	410	vma = prev->vm_next;
	411	if (!vma \|\| vma->vm_start != nstart) {
	412	error = -ENOMEM;
	413	goto out;
	414	}
	415	}
	416	out:
	417	up_write(&current->mm->mmap_sem);
	418	return error;
	419	}