[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, 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++;
				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));
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	111
9532fec1	112	*ret_all_same_node = all_same_node;
7da4d641	113	return pages;
1da177e4 LT	114	}
1da177e4 LT	115
4b10e7d5 MG	116	#ifdef CONFIG_NUMA_BALANCING
4b10e7d5 MG	117	static inline void change_pmd_protnuma(struct mm_struct *mm, unsigned long addr,
7d12efae	118	pmd_t *pmd)
4b10e7d5 MG	119	{
	120	spin_lock(&mm->page_table_lock);
	121	set_pmd_at(mm, addr & PMD_MASK, pmd, pmd_mknuma(*pmd));
	122	spin_unlock(&mm->page_table_lock);
	123	}
	124	#else
	125	static inline void change_pmd_protnuma(struct mm_struct *mm, unsigned long addr,
7d12efae	126	pmd_t *pmd)
4b10e7d5 MG	127	{
4b10e7d5 MG	128	BUG();
1da177e4	129	}
4b10e7d5	130	#endif /* CONFIG_NUMA_BALANCING */
1da177e4	131
7d12efae AM	132	static inline unsigned long change_pmd_range(struct vm_area_struct *vma,
	133	pud_t *pud, unsigned long addr, unsigned long end,
	134	pgprot_t newprot, int dirty_accountable, int prot_numa)
1da177e4 LT	135	{
	136	pmd_t *pmd;
	137	unsigned long next;
7da4d641	138	unsigned long pages = 0;
9532fec1	139	bool all_same_node;
1da177e4 LT	140
	141	pmd = pmd_offset(pud, addr);
	142	do {
	143	next = pmd_addr_end(addr, end);
cd7548ab JW	144	if (pmd_trans_huge(*pmd)) {
cd7548ab JW	145	if (next - addr != HPAGE_PMD_SIZE)
e180377f	146	split_huge_page_pmd(vma, addr, pmd);
7d12efae AM	147	else if (change_huge_pmd(vma, pmd, addr, newprot,
7d12efae AM	148	prot_numa)) {
afcae265	149	pages++;
cd7548ab	150	continue;
7da4d641	151	}
cd7548ab JW	152	/* fall through */
cd7548ab JW	153	}
1da177e4 LT	154	if (pmd_none_or_clear_bad(pmd))
1da177e4 LT	155	continue;
4b10e7d5	156	pages += change_pte_range(vma, pmd, addr, next, newprot,
9532fec1	157	dirty_accountable, prot_numa, &all_same_node);
4b10e7d5	158
9532fec1 MG	159	/*
	160	* If we are changing protections for NUMA hinting faults then
	161	* set pmd_numa if the examined pages were all on the same
	162	* node. This allows a regular PMD to be handled as one fault
	163	* and effectively batches the taking of the PTL
	164	*/
	165	if (prot_numa && all_same_node)
4b10e7d5	166	change_pmd_protnuma(vma->vm_mm, addr, pmd);
1da177e4	167	} while (pmd++, addr = next, addr != end);
7da4d641 PZ	168
7da4d641 PZ	169	return pages;
1da177e4 LT	170	}
1da177e4 LT	171
7d12efae AM	172	static inline unsigned long change_pud_range(struct vm_area_struct *vma,
	173	pgd_t *pgd, unsigned long addr, unsigned long end,
	174	pgprot_t newprot, int dirty_accountable, int prot_numa)
1da177e4 LT	175	{
	176	pud_t *pud;
	177	unsigned long next;
7da4d641	178	unsigned long pages = 0;
1da177e4 LT	179
	180	pud = pud_offset(pgd, addr);
	181	do {
	182	next = pud_addr_end(addr, end);
	183	if (pud_none_or_clear_bad(pud))
	184	continue;
7da4d641	185	pages += change_pmd_range(vma, pud, addr, next, newprot,
4b10e7d5	186	dirty_accountable, prot_numa);
1da177e4	187	} while (pud++, addr = next, addr != end);
7da4d641 PZ	188
7da4d641 PZ	189	return pages;
1da177e4 LT	190	}
1da177e4 LT	191
7da4d641	192	static unsigned long change_protection_range(struct vm_area_struct *vma,
c1e6098b	193	unsigned long addr, unsigned long end, pgprot_t newprot,
4b10e7d5	194	int dirty_accountable, int prot_numa)
1da177e4 LT	195	{
	196	struct mm_struct *mm = vma->vm_mm;
	197	pgd_t *pgd;
	198	unsigned long next;
	199	unsigned long start = addr;
7da4d641	200	unsigned long pages = 0;
1da177e4 LT	201
	202	BUG_ON(addr >= end);
	203	pgd = pgd_offset(mm, addr);
	204	flush_cache_range(vma, addr, end);
1da177e4 LT	205	do {
	206	next = pgd_addr_end(addr, end);
	207	if (pgd_none_or_clear_bad(pgd))
	208	continue;
7da4d641	209	pages += change_pud_range(vma, pgd, addr, next, newprot,
4b10e7d5	210	dirty_accountable, prot_numa);
1da177e4	211	} while (pgd++, addr = next, addr != end);
7da4d641	212
1233d588 IM	213	/* Only flush the TLB if we actually modified any entries: */
	214	if (pages)
	215	flush_tlb_range(vma, start, end);
7da4d641 PZ	216
	217	return pages;
	218	}
	219
	220	unsigned long change_protection(struct vm_area_struct *vma, unsigned long start,
	221	unsigned long end, pgprot_t newprot,
4b10e7d5	222	int dirty_accountable, int prot_numa)
7da4d641 PZ	223	{
	224	struct mm_struct *mm = vma->vm_mm;
	225	unsigned long pages;
	226
	227	mmu_notifier_invalidate_range_start(mm, start, end);
	228	if (is_vm_hugetlb_page(vma))
	229	pages = hugetlb_change_protection(vma, start, end, newprot);
	230	else
4b10e7d5	231	pages = change_protection_range(vma, start, end, newprot, dirty_accountable, prot_numa);
7da4d641 PZ	232	mmu_notifier_invalidate_range_end(mm, start, end);
	233
	234	return pages;
1da177e4 LT	235	}
1da177e4 LT	236
b6a2fea3	237	int
1da177e4 LT	238	mprotect_fixup(struct vm_area_struct vma, struct vm_area_struct *pprev,
	239	unsigned long start, unsigned long end, unsigned long newflags)
	240	{
	241	struct mm_struct *mm = vma->vm_mm;
	242	unsigned long oldflags = vma->vm_flags;
	243	long nrpages = (end - start) >> PAGE_SHIFT;
	244	unsigned long charged = 0;
1da177e4 LT	245	pgoff_t pgoff;
1da177e4 LT	246	int error;
c1e6098b	247	int dirty_accountable = 0;
1da177e4 LT	248
	249	if (newflags == oldflags) {
	250	*pprev = vma;
	251	return 0;
	252	}
	253
	254	/*
	255	* If we make a private mapping writable we increase our commit;
	256	* but (without finer accounting) cannot reduce our commit if we
5a6fe125 MG	257	* make it unwritable again. hugetlb mapping were accounted for
5a6fe125 MG	258	* even if read-only so there is no need to account for them here
1da177e4 LT	259	*/
1da177e4 LT	260	if (newflags & VM_WRITE) {
5a6fe125	261	if (!(oldflags & (VM_ACCOUNT\|VM_WRITE\|VM_HUGETLB\|
cdfd4325	262	VM_SHARED\|VM_NORESERVE))) {
1da177e4	263	charged = nrpages;
191c5424	264	if (security_vm_enough_memory_mm(mm, charged))
1da177e4 LT	265	return -ENOMEM;
	266	newflags \|= VM_ACCOUNT;
	267	}
	268	}
	269
1da177e4 LT	270	/*
	271	* First try to merge with previous and/or next vma.
	272	*/
	273	pgoff = vma->vm_pgoff + ((start - vma->vm_start) >> PAGE_SHIFT);
	274	pprev = vma_merge(mm, pprev, start, end, newflags,
	275	vma->anon_vma, vma->vm_file, pgoff, vma_policy(vma));
	276	if (*pprev) {
	277	vma = *pprev;
	278	goto success;
	279	}
	280
	281	*pprev = vma;
	282
	283	if (start != vma->vm_start) {
	284	error = split_vma(mm, vma, start, 1);
	285	if (error)
	286	goto fail;
	287	}
	288
	289	if (end != vma->vm_end) {
	290	error = split_vma(mm, vma, end, 0);
	291	if (error)
	292	goto fail;
	293	}
	294
	295	success:
	296	/*
	297	* vm_flags and vm_page_prot are protected by the mmap_sem
	298	* held in write mode.
	299	*/
	300	vma->vm_flags = newflags;
1c12c4cf VP	301	vma->vm_page_prot = pgprot_modify(vma->vm_page_prot,
	302	vm_get_page_prot(newflags));
	303
c1e6098b	304	if (vma_wants_writenotify(vma)) {
1ddd439e	305	vma->vm_page_prot = vm_get_page_prot(newflags & ~VM_SHARED);
c1e6098b PZ	306	dirty_accountable = 1;
c1e6098b PZ	307	}
d08b3851	308
7d12efae AM	309	change_protection(vma, start, end, vma->vm_page_prot,
7d12efae AM	310	dirty_accountable, 0);
7da4d641	311
ab50b8ed HD	312	vm_stat_account(mm, oldflags, vma->vm_file, -nrpages);
ab50b8ed HD	313	vm_stat_account(mm, newflags, vma->vm_file, nrpages);
63bfd738	314	perf_event_mmap(vma);
1da177e4 LT	315	return 0;
	316
	317	fail:
	318	vm_unacct_memory(charged);
	319	return error;
	320	}
	321
6a6160a7 HC	322	SYSCALL_DEFINE3(mprotect, unsigned long, start, size_t, len,
6a6160a7 HC	323	unsigned long, prot)
1da177e4 LT	324	{
	325	unsigned long vm_flags, nstart, end, tmp, reqprot;
	326	struct vm_area_struct vma, prev;
	327	int error = -EINVAL;
	328	const int grows = prot & (PROT_GROWSDOWN\|PROT_GROWSUP);
	329	prot &= ~(PROT_GROWSDOWN\|PROT_GROWSUP);
	330	if (grows == (PROT_GROWSDOWN\|PROT_GROWSUP)) /* can't be both */
	331	return -EINVAL;
	332
	333	if (start & ~PAGE_MASK)
	334	return -EINVAL;
	335	if (!len)
	336	return 0;
	337	len = PAGE_ALIGN(len);
	338	end = start + len;
	339	if (end <= start)
	340	return -ENOMEM;
b845f313	341	if (!arch_validate_prot(prot))
1da177e4 LT	342	return -EINVAL;
	343
	344	reqprot = prot;
	345	/*
	346	* Does the application expect PROT_READ to imply PROT_EXEC:
	347	*/
b344e05c	348	if ((prot & PROT_READ) && (current->personality & READ_IMPLIES_EXEC))
1da177e4 LT	349	prot \|= PROT_EXEC;
	350
	351	vm_flags = calc_vm_prot_bits(prot);
	352
	353	down_write(&current->mm->mmap_sem);
	354
097d5910	355	vma = find_vma(current->mm, start);
1da177e4 LT	356	error = -ENOMEM;
	357	if (!vma)
	358	goto out;
097d5910	359	prev = vma->vm_prev;
1da177e4 LT	360	if (unlikely(grows & PROT_GROWSDOWN)) {
	361	if (vma->vm_start >= end)
	362	goto out;
	363	start = vma->vm_start;
	364	error = -EINVAL;
	365	if (!(vma->vm_flags & VM_GROWSDOWN))
	366	goto out;
7d12efae	367	} else {
1da177e4 LT	368	if (vma->vm_start > start)
	369	goto out;
	370	if (unlikely(grows & PROT_GROWSUP)) {
	371	end = vma->vm_end;
	372	error = -EINVAL;
	373	if (!(vma->vm_flags & VM_GROWSUP))
	374	goto out;
	375	}
	376	}
	377	if (start > vma->vm_start)
	378	prev = vma;
	379
	380	for (nstart = start ; ; ) {
	381	unsigned long newflags;
	382
7d12efae	383	/* Here we know that vma->vm_start <= nstart < vma->vm_end. */
1da177e4	384
7d12efae AM	385	newflags = vm_flags;
7d12efae AM	386	newflags \|= (vma->vm_flags & ~(VM_READ \| VM_WRITE \| VM_EXEC));
1da177e4	387
7e2cff42 PBG	388	/* newflags >> 4 shift VM_MAY% in place of VM_% */
7e2cff42 PBG	389	if ((newflags & ~(newflags >> 4)) & (VM_READ \| VM_WRITE \| VM_EXEC)) {
1da177e4 LT	390	error = -EACCES;
	391	goto out;
	392	}
	393
	394	error = security_file_mprotect(vma, reqprot, prot);
	395	if (error)
	396	goto out;
	397
	398	tmp = vma->vm_end;
	399	if (tmp > end)
	400	tmp = end;
	401	error = mprotect_fixup(vma, &prev, nstart, tmp, newflags);
	402	if (error)
	403	goto out;
	404	nstart = tmp;
	405
	406	if (nstart < prev->vm_end)
	407	nstart = prev->vm_end;
	408	if (nstart >= end)
	409	goto out;
	410
	411	vma = prev->vm_next;
	412	if (!vma \|\| vma->vm_start != nstart) {
	413	error = -ENOMEM;
	414	goto out;
	415	}
	416	}
	417	out:
	418	up_write(&current->mm->mmap_sem);
	419	return error;
	420	}