[linux.git] / mm / mprotect.c

/*
 *  mm/mprotect.c
 *
 *  (C) Copyright 1994 Linus Torvalds
 *  (C) Copyright 2002 Christoph Hellwig
 *
 *  Address space accounting code	<[email protected]>
 *  (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>

#include "internal.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 preserve_write = prot_numa && pte_write(oldpte);

			/*
			 * Avoid trapping faults against the zero or KSM
			 * pages. See similar comment in change_huge_pmd.
			 */
			if (prot_numa) {
				struct page *page;

				page = vm_normal_page(vma, addr, oldpte);
				if (!page || PageKsm(page))
					continue;

				/* Avoid TLB flush if possible */
				if (pte_protnone(oldpte))
					continue;
			}

			ptent = ptep_modify_prot_start(mm, addr, pte);
			ptent = pte_modify(ptent, newprot);
			if (preserve_write)
				ptent = pte_mkwrite(ptent);

			/* Avoid taking write faults for known dirty pages */
			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);
			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_devmap(*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) || pmd_devmap(*pmd)) {
			if (next - addr != HPAGE_PMD_SIZE)
				split_huge_pmd(vma, pmd, addr);
			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) {
		/* Check space limits when area turns into data. */
		if (!may_expand_vm(mm, newflags, nrpages) &&
				may_expand_vm(mm, oldflags, nrpages))
			return -ENOMEM;
		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),
			   vma->vm_userfaultfd_ctx);
	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);

	/*
	 * Private VM_LOCKED VMA becoming writable: trigger COW to avoid major
	 * fault on access.
	 */
	if ((oldflags & (VM_WRITE | VM_SHARED | VM_LOCKED)) == VM_LOCKED &&
			(newflags & VM_WRITE)) {
		populate_vma_page_range(vma, start, end, NULL);
	}

	vm_stat_account(mm, oldflags, -nrpages);
	vm_stat_account(mm, newflags, 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 <[email protected]>
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
36f88188 KS	32	#include "internal.h"
36f88188 KS	33
1ad9f620 MG	34	/*
	35	* For a prot_numa update we only hold mmap_sem for read so there is a
	36	* potential race with faulting where a pmd was temporarily none. This
	37	* function checks for a transhuge pmd under the appropriate lock. It
	38	* returns a pte if it was successfully locked or NULL if it raced with
	39	* a transhuge insertion.
	40	*/
	41	static pte_t lock_pte_protection(struct vm_area_struct vma, pmd_t *pmd,
	42	unsigned long addr, int prot_numa, spinlock_t **ptl)
	43	{
	44	pte_t *pte;
	45	spinlock_t *pmdl;
	46
	47	/* !prot_numa is protected by mmap_sem held for write */
	48	if (!prot_numa)
	49	return pte_offset_map_lock(vma->vm_mm, pmd, addr, ptl);
	50
	51	pmdl = pmd_lock(vma->vm_mm, pmd);
	52	if (unlikely(pmd_trans_huge(pmd) \|\| pmd_none(pmd))) {
	53	spin_unlock(pmdl);
	54	return NULL;
	55	}
	56
	57	pte = pte_offset_map_lock(vma->vm_mm, pmd, addr, ptl);
	58	spin_unlock(pmdl);
	59	return pte;
	60	}
	61
4b10e7d5	62	static unsigned long change_pte_range(struct vm_area_struct vma, pmd_t pmd,
c1e6098b	63	unsigned long addr, unsigned long end, pgprot_t newprot,
0f19c179	64	int dirty_accountable, int prot_numa)
1da177e4	65	{
4b10e7d5	66	struct mm_struct *mm = vma->vm_mm;
0697212a	67	pte_t *pte, oldpte;
705e87c0	68	spinlock_t *ptl;
7da4d641	69	unsigned long pages = 0;
1da177e4	70
1ad9f620 MG	71	pte = lock_pte_protection(vma, pmd, addr, prot_numa, &ptl);
	72	if (!pte)
	73	return 0;
	74
6606c3e0	75	arch_enter_lazy_mmu_mode();
1da177e4	76	do {
0697212a CL	77	oldpte = *pte;
0697212a CL	78	if (pte_present(oldpte)) {
1da177e4	79	pte_t ptent;
b191f9b1	80	bool preserve_write = prot_numa && pte_write(oldpte);
1da177e4	81
e944fd67 MG	82	/*
	83	* Avoid trapping faults against the zero or KSM
	84	* pages. See similar comment in change_huge_pmd.
	85	*/
	86	if (prot_numa) {
	87	struct page *page;
	88
	89	page = vm_normal_page(vma, addr, oldpte);
	90	if (!page \|\| PageKsm(page))
	91	continue;
10c1045f MG	92
	93	/* Avoid TLB flush if possible */
	94	if (pte_protnone(oldpte))
	95	continue;
e944fd67 MG	96	}
e944fd67 MG	97
8a0516ed MG	98	ptent = ptep_modify_prot_start(mm, addr, pte);
8a0516ed MG	99	ptent = pte_modify(ptent, newprot);
b191f9b1 MG	100	if (preserve_write)
b191f9b1 MG	101	ptent = pte_mkwrite(ptent);
4b10e7d5	102
8a0516ed MG	103	/* Avoid taking write faults for known dirty pages */
	104	if (dirty_accountable && pte_dirty(ptent) &&
	105	(pte_soft_dirty(ptent) \|\|
	106	!(vma->vm_flags & VM_SOFTDIRTY))) {
	107	ptent = pte_mkwrite(ptent);
4b10e7d5	108	}
8a0516ed MG	109	ptep_modify_prot_commit(mm, addr, pte, ptent);
8a0516ed MG	110	pages++;
0661a336	111	} else if (IS_ENABLED(CONFIG_MIGRATION)) {
0697212a CL	112	swp_entry_t entry = pte_to_swp_entry(oldpte);
	113
	114	if (is_write_migration_entry(entry)) {
c3d16e16	115	pte_t newpte;
0697212a CL	116	/*
	117	* A protection check is difficult so
	118	* just be safe and disable write
	119	*/
	120	make_migration_entry_read(&entry);
c3d16e16 CG	121	newpte = swp_entry_to_pte(entry);
	122	if (pte_swp_soft_dirty(oldpte))
	123	newpte = pte_swp_mksoft_dirty(newpte);
	124	set_pte_at(mm, addr, pte, newpte);
e920e14c MG	125
e920e14c MG	126	pages++;
0697212a	127	}
1da177e4 LT	128	}
1da177e4 LT	129	} while (pte++, addr += PAGE_SIZE, addr != end);
6606c3e0	130	arch_leave_lazy_mmu_mode();
705e87c0	131	pte_unmap_unlock(pte - 1, ptl);
7da4d641 PZ	132
7da4d641 PZ	133	return pages;
1da177e4 LT	134	}
1da177e4 LT	135
7d12efae AM	136	static inline unsigned long change_pmd_range(struct vm_area_struct *vma,
	137	pud_t *pud, unsigned long addr, unsigned long end,
	138	pgprot_t newprot, int dirty_accountable, int prot_numa)
1da177e4 LT	139	{
1da177e4 LT	140	pmd_t *pmd;
a5338093	141	struct mm_struct *mm = vma->vm_mm;
1da177e4	142	unsigned long next;
7da4d641	143	unsigned long pages = 0;
72403b4a	144	unsigned long nr_huge_updates = 0;
a5338093	145	unsigned long mni_start = 0;
1da177e4 LT	146
	147	pmd = pmd_offset(pud, addr);
	148	do {
25cbbef1 MG	149	unsigned long this_pages;
25cbbef1 MG	150
1da177e4	151	next = pmd_addr_end(addr, end);
5c7fb56e DW	152	if (!pmd_trans_huge(pmd) && !pmd_devmap(pmd)
5c7fb56e DW	153	&& pmd_none_or_clear_bad(pmd))
88a9ab6e	154	continue;
a5338093 RR	155
	156	/* invoke the mmu notifier if the pmd is populated */
	157	if (!mni_start) {
	158	mni_start = addr;
	159	mmu_notifier_invalidate_range_start(mm, mni_start, end);
	160	}
	161
5c7fb56e	162	if (pmd_trans_huge(pmd) \|\| pmd_devmap(pmd)) {
cd7548ab	163	if (next - addr != HPAGE_PMD_SIZE)
78ddc534	164	split_huge_pmd(vma, pmd, addr);
f123d74a MG	165	else {
f123d74a MG	166	int nr_ptes = change_huge_pmd(vma, pmd, addr,
e944fd67	167	newprot, prot_numa);
f123d74a MG	168
f123d74a MG	169	if (nr_ptes) {
72403b4a MG	170	if (nr_ptes == HPAGE_PMD_NR) {
	171	pages += HPAGE_PMD_NR;
	172	nr_huge_updates++;
	173	}
1ad9f620 MG	174
1ad9f620 MG	175	/* huge pmd was handled */
f123d74a MG	176	continue;
f123d74a MG	177	}
7da4d641	178	}
88a9ab6e	179	/* fall through, the trans huge pmd just split */
cd7548ab	180	}
25cbbef1	181	this_pages = change_pte_range(vma, pmd, addr, next, newprot,
0f19c179	182	dirty_accountable, prot_numa);
25cbbef1	183	pages += this_pages;
1da177e4	184	} while (pmd++, addr = next, addr != end);
7da4d641	185
a5338093 RR	186	if (mni_start)
	187	mmu_notifier_invalidate_range_end(mm, mni_start, end);
	188
72403b4a MG	189	if (nr_huge_updates)
72403b4a MG	190	count_vm_numa_events(NUMA_HUGE_PTE_UPDATES, nr_huge_updates);
7da4d641	191	return pages;
1da177e4 LT	192	}
1da177e4 LT	193
7d12efae AM	194	static inline unsigned long change_pud_range(struct vm_area_struct *vma,
	195	pgd_t *pgd, unsigned long addr, unsigned long end,
	196	pgprot_t newprot, int dirty_accountable, int prot_numa)
1da177e4 LT	197	{
	198	pud_t *pud;
	199	unsigned long next;
7da4d641	200	unsigned long pages = 0;
1da177e4 LT	201
	202	pud = pud_offset(pgd, addr);
	203	do {
	204	next = pud_addr_end(addr, end);
	205	if (pud_none_or_clear_bad(pud))
	206	continue;
7da4d641	207	pages += change_pmd_range(vma, pud, addr, next, newprot,
4b10e7d5	208	dirty_accountable, prot_numa);
1da177e4	209	} while (pud++, addr = next, addr != end);
7da4d641 PZ	210
7da4d641 PZ	211	return pages;
1da177e4 LT	212	}
1da177e4 LT	213
7da4d641	214	static unsigned long change_protection_range(struct vm_area_struct *vma,
c1e6098b	215	unsigned long addr, unsigned long end, pgprot_t newprot,
4b10e7d5	216	int dirty_accountable, int prot_numa)
1da177e4 LT	217	{
	218	struct mm_struct *mm = vma->vm_mm;
	219	pgd_t *pgd;
	220	unsigned long next;
	221	unsigned long start = addr;
7da4d641	222	unsigned long pages = 0;
1da177e4 LT	223
	224	BUG_ON(addr >= end);
	225	pgd = pgd_offset(mm, addr);
	226	flush_cache_range(vma, addr, end);
20841405	227	set_tlb_flush_pending(mm);
1da177e4 LT	228	do {
	229	next = pgd_addr_end(addr, end);
	230	if (pgd_none_or_clear_bad(pgd))
	231	continue;
7da4d641	232	pages += change_pud_range(vma, pgd, addr, next, newprot,
4b10e7d5	233	dirty_accountable, prot_numa);
1da177e4	234	} while (pgd++, addr = next, addr != end);
7da4d641	235
1233d588 IM	236	/* Only flush the TLB if we actually modified any entries: */
	237	if (pages)
	238	flush_tlb_range(vma, start, end);
20841405	239	clear_tlb_flush_pending(mm);
7da4d641 PZ	240
	241	return pages;
	242	}
	243
	244	unsigned long change_protection(struct vm_area_struct *vma, unsigned long start,
	245	unsigned long end, pgprot_t newprot,
4b10e7d5	246	int dirty_accountable, int prot_numa)
7da4d641	247	{
7da4d641 PZ	248	unsigned long pages;
7da4d641 PZ	249
7da4d641 PZ	250	if (is_vm_hugetlb_page(vma))
	251	pages = hugetlb_change_protection(vma, start, end, newprot);
	252	else
4b10e7d5	253	pages = change_protection_range(vma, start, end, newprot, dirty_accountable, prot_numa);
7da4d641 PZ	254
7da4d641 PZ	255	return pages;
1da177e4 LT	256	}
1da177e4 LT	257
b6a2fea3	258	int
1da177e4 LT	259	mprotect_fixup(struct vm_area_struct vma, struct vm_area_struct *pprev,
	260	unsigned long start, unsigned long end, unsigned long newflags)
	261	{
	262	struct mm_struct *mm = vma->vm_mm;
	263	unsigned long oldflags = vma->vm_flags;
	264	long nrpages = (end - start) >> PAGE_SHIFT;
	265	unsigned long charged = 0;
1da177e4 LT	266	pgoff_t pgoff;
1da177e4 LT	267	int error;
c1e6098b	268	int dirty_accountable = 0;
1da177e4 LT	269
	270	if (newflags == oldflags) {
	271	*pprev = vma;
	272	return 0;
	273	}
	274
	275	/*
	276	* If we make a private mapping writable we increase our commit;
	277	* but (without finer accounting) cannot reduce our commit if we
5a6fe125 MG	278	* make it unwritable again. hugetlb mapping were accounted for
5a6fe125 MG	279	* even if read-only so there is no need to account for them here
1da177e4 LT	280	*/
1da177e4 LT	281	if (newflags & VM_WRITE) {
84638335 KK	282	/* Check space limits when area turns into data. */
	283	if (!may_expand_vm(mm, newflags, nrpages) &&
	284	may_expand_vm(mm, oldflags, nrpages))
	285	return -ENOMEM;
5a6fe125	286	if (!(oldflags & (VM_ACCOUNT\|VM_WRITE\|VM_HUGETLB\|
cdfd4325	287	VM_SHARED\|VM_NORESERVE))) {
1da177e4	288	charged = nrpages;
191c5424	289	if (security_vm_enough_memory_mm(mm, charged))
1da177e4 LT	290	return -ENOMEM;
	291	newflags \|= VM_ACCOUNT;
	292	}
	293	}
	294
1da177e4 LT	295	/*
	296	* First try to merge with previous and/or next vma.
	297	*/
	298	pgoff = vma->vm_pgoff + ((start - vma->vm_start) >> PAGE_SHIFT);
	299	pprev = vma_merge(mm, pprev, start, end, newflags,
19a809af AA	300	vma->anon_vma, vma->vm_file, pgoff, vma_policy(vma),
19a809af AA	301	vma->vm_userfaultfd_ctx);
1da177e4 LT	302	if (*pprev) {
	303	vma = *pprev;
	304	goto success;
	305	}
	306
	307	*pprev = vma;
	308
	309	if (start != vma->vm_start) {
	310	error = split_vma(mm, vma, start, 1);
	311	if (error)
	312	goto fail;
	313	}
	314
	315	if (end != vma->vm_end) {
	316	error = split_vma(mm, vma, end, 0);
	317	if (error)
	318	goto fail;
	319	}
	320
	321	success:
	322	/*
	323	* vm_flags and vm_page_prot are protected by the mmap_sem
	324	* held in write mode.
	325	*/
	326	vma->vm_flags = newflags;
64e45507 PF	327	dirty_accountable = vma_wants_writenotify(vma);
64e45507 PF	328	vma_set_page_prot(vma);
d08b3851	329
7d12efae AM	330	change_protection(vma, start, end, vma->vm_page_prot,
7d12efae AM	331	dirty_accountable, 0);
7da4d641	332
36f88188 KS	333	/*
	334	* Private VM_LOCKED VMA becoming writable: trigger COW to avoid major
	335	* fault on access.
	336	*/
	337	if ((oldflags & (VM_WRITE \| VM_SHARED \| VM_LOCKED)) == VM_LOCKED &&
	338	(newflags & VM_WRITE)) {
	339	populate_vma_page_range(vma, start, end, NULL);
	340	}
	341
84638335 KK	342	vm_stat_account(mm, oldflags, -nrpages);
84638335 KK	343	vm_stat_account(mm, newflags, nrpages);
63bfd738	344	perf_event_mmap(vma);
1da177e4 LT	345	return 0;
	346
	347	fail:
	348	vm_unacct_memory(charged);
	349	return error;
	350	}
	351
6a6160a7 HC	352	SYSCALL_DEFINE3(mprotect, unsigned long, start, size_t, len,
6a6160a7 HC	353	unsigned long, prot)
1da177e4 LT	354	{
	355	unsigned long vm_flags, nstart, end, tmp, reqprot;
	356	struct vm_area_struct vma, prev;
	357	int error = -EINVAL;
	358	const int grows = prot & (PROT_GROWSDOWN\|PROT_GROWSUP);
	359	prot &= ~(PROT_GROWSDOWN\|PROT_GROWSUP);
	360	if (grows == (PROT_GROWSDOWN\|PROT_GROWSUP)) /* can't be both */
	361	return -EINVAL;
	362
	363	if (start & ~PAGE_MASK)
	364	return -EINVAL;
	365	if (!len)
	366	return 0;
	367	len = PAGE_ALIGN(len);
	368	end = start + len;
	369	if (end <= start)
	370	return -ENOMEM;
b845f313	371	if (!arch_validate_prot(prot))
1da177e4 LT	372	return -EINVAL;
	373
	374	reqprot = prot;
	375	/*
	376	* Does the application expect PROT_READ to imply PROT_EXEC:
	377	*/
b344e05c	378	if ((prot & PROT_READ) && (current->personality & READ_IMPLIES_EXEC))
1da177e4 LT	379	prot \|= PROT_EXEC;
	380
	381	vm_flags = calc_vm_prot_bits(prot);
	382
	383	down_write(&current->mm->mmap_sem);
	384
097d5910	385	vma = find_vma(current->mm, start);
1da177e4 LT	386	error = -ENOMEM;
	387	if (!vma)
	388	goto out;
097d5910	389	prev = vma->vm_prev;
1da177e4 LT	390	if (unlikely(grows & PROT_GROWSDOWN)) {
	391	if (vma->vm_start >= end)
	392	goto out;
	393	start = vma->vm_start;
	394	error = -EINVAL;
	395	if (!(vma->vm_flags & VM_GROWSDOWN))
	396	goto out;
7d12efae	397	} else {
1da177e4 LT	398	if (vma->vm_start > start)
	399	goto out;
	400	if (unlikely(grows & PROT_GROWSUP)) {
	401	end = vma->vm_end;
	402	error = -EINVAL;
	403	if (!(vma->vm_flags & VM_GROWSUP))
	404	goto out;
	405	}
	406	}
	407	if (start > vma->vm_start)
	408	prev = vma;
	409
	410	for (nstart = start ; ; ) {
	411	unsigned long newflags;
	412
7d12efae	413	/* Here we know that vma->vm_start <= nstart < vma->vm_end. */
1da177e4	414
7d12efae AM	415	newflags = vm_flags;
7d12efae AM	416	newflags \|= (vma->vm_flags & ~(VM_READ \| VM_WRITE \| VM_EXEC));
1da177e4	417
7e2cff42 PBG	418	/* newflags >> 4 shift VM_MAY% in place of VM_% */
7e2cff42 PBG	419	if ((newflags & ~(newflags >> 4)) & (VM_READ \| VM_WRITE \| VM_EXEC)) {
1da177e4 LT	420	error = -EACCES;
	421	goto out;
	422	}
	423
	424	error = security_file_mprotect(vma, reqprot, prot);
	425	if (error)
	426	goto out;
	427
	428	tmp = vma->vm_end;
	429	if (tmp > end)
	430	tmp = end;
	431	error = mprotect_fixup(vma, &prev, nstart, tmp, newflags);
	432	if (error)
	433	goto out;
	434	nstart = tmp;
	435
	436	if (nstart < prev->vm_end)
	437	nstart = prev->vm_end;
	438	if (nstart >= end)
	439	goto out;
	440
	441	vma = prev->vm_next;
	442	if (!vma \|\| vma->vm_start != nstart) {
	443	error = -ENOMEM;
	444	goto out;
	445	}
	446	}
	447	out:
	448	up_write(&current->mm->mmap_sem);
	449	return error;
	450	}