[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/pkeys.h>
#include <linux/ksm.h>
#include <linux/pkeys.h>
#include <asm/uaccess.h>
#include <asm/pgtable.h>
#include <asm/cacheflush.h>
#include <asm/mmu_context.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);
				if (pmd_trans_unstable(pmd))
					continue;
			} 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;
		VM_WARN_ON((vma->vm_flags ^ newflags) & ~VM_SOFTDIRTY);
		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->vm_page_prot);
	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;
}

/*
 * pkey==-1 when doing a legacy mprotect()
 */
static int do_mprotect_pkey(unsigned long start, size_t len,
		unsigned long prot, int pkey)
{
	unsigned long nstart, end, tmp, reqprot;
	struct vm_area_struct *vma, *prev;
	int error = -EINVAL;
	const int grows = prot & (PROT_GROWSDOWN|PROT_GROWSUP);
	const bool rier = (current->personality & READ_IMPLIES_EXEC) &&
				(prot & PROT_READ);

	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;

	if (down_write_killable(&current->mm->mmap_sem))
		return -EINTR;

	/*
	 * If userspace did not allocate the pkey, do not let
	 * them use it here.
	 */
	error = -EINVAL;
	if ((pkey != -1) && !mm_pkey_is_allocated(current->mm, pkey))
		goto out;

	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 mask_off_old_flags;
		unsigned long newflags;
		int new_vma_pkey;

		/* Here we know that vma->vm_start <= nstart < vma->vm_end. */

		/* Does the application expect PROT_READ to imply PROT_EXEC */
		if (rier && (vma->vm_flags & VM_MAYEXEC))
			prot |= PROT_EXEC;

		/*
		 * Each mprotect() call explicitly passes r/w/x permissions.
		 * If a permission is not passed to mprotect(), it must be
		 * cleared from the VMA.
		 */
		mask_off_old_flags = VM_READ | VM_WRITE | VM_EXEC |
					ARCH_VM_PKEY_FLAGS;

		new_vma_pkey = arch_override_mprotect_pkey(vma, prot, pkey);
		newflags = calc_vm_prot_bits(prot, new_vma_pkey);
		newflags |= (vma->vm_flags & ~mask_off_old_flags);

		/* 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;
		}
		prot = reqprot;
	}
out:
	up_write(&current->mm->mmap_sem);
	return error;
}

SYSCALL_DEFINE3(mprotect, unsigned long, start, size_t, len,
		unsigned long, prot)
{
	return do_mprotect_pkey(start, len, prot, -1);
}

SYSCALL_DEFINE4(pkey_mprotect, unsigned long, start, size_t, len,
		unsigned long, prot, int, pkey)
{
	return do_mprotect_pkey(start, len, prot, pkey);
}

SYSCALL_DEFINE2(pkey_alloc, unsigned long, flags, unsigned long, init_val)
{
	int pkey;
	int ret;

	/* No flags supported yet. */
	if (flags)
		return -EINVAL;
	/* check for unsupported init values */
	if (init_val & ~PKEY_ACCESS_MASK)
		return -EINVAL;

	down_write(&current->mm->mmap_sem);
	pkey = mm_pkey_alloc(current->mm);

	ret = -ENOSPC;
	if (pkey == -1)
		goto out;

	ret = arch_set_user_pkey_access(current, pkey, init_val);
	if (ret) {
		mm_pkey_free(current->mm, pkey);
		goto out;
	}
	ret = pkey;
out:
	up_write(&current->mm->mmap_sem);
	return ret;
}

SYSCALL_DEFINE1(pkey_free, int, pkey)
{
	int ret;

	down_write(&current->mm->mmap_sem);
	ret = mm_pkey_free(current->mm, pkey);
	up_write(&current->mm->mmap_sem);

	/*
	 * We could provie warnings or errors if any VMA still
	 * has the pkey set here.
	 */
	return ret;
}
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>
e8c24d3a	26	#include <linux/pkeys.h>
64a9a34e	27	#include <linux/ksm.h>
62b5f7d0	28	#include <linux/pkeys.h>
1da177e4 LT	29	#include <asm/uaccess.h>
	30	#include <asm/pgtable.h>
	31	#include <asm/cacheflush.h>
e8c24d3a	32	#include <asm/mmu_context.h>
1da177e4 LT	33	#include <asm/tlbflush.h>
1da177e4 LT	34
36f88188 KS	35	#include "internal.h"
36f88188 KS	36
1ad9f620 MG	37	/*
	38	* For a prot_numa update we only hold mmap_sem for read so there is a
	39	* potential race with faulting where a pmd was temporarily none. This
	40	* function checks for a transhuge pmd under the appropriate lock. It
	41	* returns a pte if it was successfully locked or NULL if it raced with
	42	* a transhuge insertion.
	43	*/
	44	static pte_t lock_pte_protection(struct vm_area_struct vma, pmd_t *pmd,
	45	unsigned long addr, int prot_numa, spinlock_t **ptl)
	46	{
	47	pte_t *pte;
	48	spinlock_t *pmdl;
	49
	50	/* !prot_numa is protected by mmap_sem held for write */
	51	if (!prot_numa)
	52	return pte_offset_map_lock(vma->vm_mm, pmd, addr, ptl);
	53
	54	pmdl = pmd_lock(vma->vm_mm, pmd);
	55	if (unlikely(pmd_trans_huge(pmd) \|\| pmd_none(pmd))) {
	56	spin_unlock(pmdl);
	57	return NULL;
	58	}
	59
	60	pte = pte_offset_map_lock(vma->vm_mm, pmd, addr, ptl);
	61	spin_unlock(pmdl);
	62	return pte;
	63	}
	64
4b10e7d5	65	static unsigned long change_pte_range(struct vm_area_struct vma, pmd_t pmd,
c1e6098b	66	unsigned long addr, unsigned long end, pgprot_t newprot,
0f19c179	67	int dirty_accountable, int prot_numa)
1da177e4	68	{
4b10e7d5	69	struct mm_struct *mm = vma->vm_mm;
0697212a	70	pte_t *pte, oldpte;
705e87c0	71	spinlock_t *ptl;
7da4d641	72	unsigned long pages = 0;
1da177e4	73
1ad9f620 MG	74	pte = lock_pte_protection(vma, pmd, addr, prot_numa, &ptl);
	75	if (!pte)
	76	return 0;
	77
6606c3e0	78	arch_enter_lazy_mmu_mode();
1da177e4	79	do {
0697212a CL	80	oldpte = *pte;
0697212a CL	81	if (pte_present(oldpte)) {
1da177e4	82	pte_t ptent;
b191f9b1	83	bool preserve_write = prot_numa && pte_write(oldpte);
1da177e4	84
e944fd67 MG	85	/*
	86	* Avoid trapping faults against the zero or KSM
	87	* pages. See similar comment in change_huge_pmd.
	88	*/
	89	if (prot_numa) {
	90	struct page *page;
	91
	92	page = vm_normal_page(vma, addr, oldpte);
	93	if (!page \|\| PageKsm(page))
	94	continue;
10c1045f MG	95
	96	/* Avoid TLB flush if possible */
	97	if (pte_protnone(oldpte))
	98	continue;
e944fd67 MG	99	}
e944fd67 MG	100
8a0516ed MG	101	ptent = ptep_modify_prot_start(mm, addr, pte);
8a0516ed MG	102	ptent = pte_modify(ptent, newprot);
b191f9b1 MG	103	if (preserve_write)
b191f9b1 MG	104	ptent = pte_mkwrite(ptent);
4b10e7d5	105
8a0516ed MG	106	/* Avoid taking write faults for known dirty pages */
	107	if (dirty_accountable && pte_dirty(ptent) &&
	108	(pte_soft_dirty(ptent) \|\|
	109	!(vma->vm_flags & VM_SOFTDIRTY))) {
	110	ptent = pte_mkwrite(ptent);
4b10e7d5	111	}
8a0516ed MG	112	ptep_modify_prot_commit(mm, addr, pte, ptent);
8a0516ed MG	113	pages++;
0661a336	114	} else if (IS_ENABLED(CONFIG_MIGRATION)) {
0697212a CL	115	swp_entry_t entry = pte_to_swp_entry(oldpte);
	116
	117	if (is_write_migration_entry(entry)) {
c3d16e16	118	pte_t newpte;
0697212a CL	119	/*
	120	* A protection check is difficult so
	121	* just be safe and disable write
	122	*/
	123	make_migration_entry_read(&entry);
c3d16e16 CG	124	newpte = swp_entry_to_pte(entry);
	125	if (pte_swp_soft_dirty(oldpte))
	126	newpte = pte_swp_mksoft_dirty(newpte);
	127	set_pte_at(mm, addr, pte, newpte);
e920e14c MG	128
e920e14c MG	129	pages++;
0697212a	130	}
1da177e4 LT	131	}
1da177e4 LT	132	} while (pte++, addr += PAGE_SIZE, addr != end);
6606c3e0	133	arch_leave_lazy_mmu_mode();
705e87c0	134	pte_unmap_unlock(pte - 1, ptl);
7da4d641 PZ	135
7da4d641 PZ	136	return pages;
1da177e4 LT	137	}
1da177e4 LT	138
7d12efae AM	139	static inline unsigned long change_pmd_range(struct vm_area_struct *vma,
	140	pud_t *pud, unsigned long addr, unsigned long end,
	141	pgprot_t newprot, int dirty_accountable, int prot_numa)
1da177e4 LT	142	{
1da177e4 LT	143	pmd_t *pmd;
a5338093	144	struct mm_struct *mm = vma->vm_mm;
1da177e4	145	unsigned long next;
7da4d641	146	unsigned long pages = 0;
72403b4a	147	unsigned long nr_huge_updates = 0;
a5338093	148	unsigned long mni_start = 0;
1da177e4 LT	149
	150	pmd = pmd_offset(pud, addr);
	151	do {
25cbbef1 MG	152	unsigned long this_pages;
25cbbef1 MG	153
1da177e4	154	next = pmd_addr_end(addr, end);
5c7fb56e DW	155	if (!pmd_trans_huge(pmd) && !pmd_devmap(pmd)
5c7fb56e DW	156	&& pmd_none_or_clear_bad(pmd))
88a9ab6e	157	continue;
a5338093 RR	158
	159	/* invoke the mmu notifier if the pmd is populated */
	160	if (!mni_start) {
	161	mni_start = addr;
	162	mmu_notifier_invalidate_range_start(mm, mni_start, end);
	163	}
	164
5c7fb56e	165	if (pmd_trans_huge(pmd) \|\| pmd_devmap(pmd)) {
6b9116a6	166	if (next - addr != HPAGE_PMD_SIZE) {
78ddc534	167	split_huge_pmd(vma, pmd, addr);
337d9abf	168	if (pmd_trans_unstable(pmd))
6b9116a6 KS	169	continue;
6b9116a6 KS	170	} else {
f123d74a	171	int nr_ptes = change_huge_pmd(vma, pmd, addr,
e944fd67	172	newprot, prot_numa);
f123d74a MG	173
f123d74a MG	174	if (nr_ptes) {
72403b4a MG	175	if (nr_ptes == HPAGE_PMD_NR) {
	176	pages += HPAGE_PMD_NR;
	177	nr_huge_updates++;
	178	}
1ad9f620 MG	179
1ad9f620 MG	180	/* huge pmd was handled */
f123d74a MG	181	continue;
f123d74a MG	182	}
7da4d641	183	}
88a9ab6e	184	/* fall through, the trans huge pmd just split */
cd7548ab	185	}
25cbbef1	186	this_pages = change_pte_range(vma, pmd, addr, next, newprot,
0f19c179	187	dirty_accountable, prot_numa);
25cbbef1	188	pages += this_pages;
1da177e4	189	} while (pmd++, addr = next, addr != end);
7da4d641	190
a5338093 RR	191	if (mni_start)
	192	mmu_notifier_invalidate_range_end(mm, mni_start, end);
	193
72403b4a MG	194	if (nr_huge_updates)
72403b4a MG	195	count_vm_numa_events(NUMA_HUGE_PTE_UPDATES, nr_huge_updates);
7da4d641	196	return pages;
1da177e4 LT	197	}
1da177e4 LT	198
7d12efae AM	199	static inline unsigned long change_pud_range(struct vm_area_struct *vma,
	200	pgd_t *pgd, unsigned long addr, unsigned long end,
	201	pgprot_t newprot, int dirty_accountable, int prot_numa)
1da177e4 LT	202	{
	203	pud_t *pud;
	204	unsigned long next;
7da4d641	205	unsigned long pages = 0;
1da177e4 LT	206
	207	pud = pud_offset(pgd, addr);
	208	do {
	209	next = pud_addr_end(addr, end);
	210	if (pud_none_or_clear_bad(pud))
	211	continue;
7da4d641	212	pages += change_pmd_range(vma, pud, addr, next, newprot,
4b10e7d5	213	dirty_accountable, prot_numa);
1da177e4	214	} while (pud++, addr = next, addr != end);
7da4d641 PZ	215
7da4d641 PZ	216	return pages;
1da177e4 LT	217	}
1da177e4 LT	218
7da4d641	219	static unsigned long change_protection_range(struct vm_area_struct *vma,
c1e6098b	220	unsigned long addr, unsigned long end, pgprot_t newprot,
4b10e7d5	221	int dirty_accountable, int prot_numa)
1da177e4 LT	222	{
	223	struct mm_struct *mm = vma->vm_mm;
	224	pgd_t *pgd;
	225	unsigned long next;
	226	unsigned long start = addr;
7da4d641	227	unsigned long pages = 0;
1da177e4 LT	228
	229	BUG_ON(addr >= end);
	230	pgd = pgd_offset(mm, addr);
	231	flush_cache_range(vma, addr, end);
20841405	232	set_tlb_flush_pending(mm);
1da177e4 LT	233	do {
	234	next = pgd_addr_end(addr, end);
	235	if (pgd_none_or_clear_bad(pgd))
	236	continue;
7da4d641	237	pages += change_pud_range(vma, pgd, addr, next, newprot,
4b10e7d5	238	dirty_accountable, prot_numa);
1da177e4	239	} while (pgd++, addr = next, addr != end);
7da4d641	240
1233d588 IM	241	/* Only flush the TLB if we actually modified any entries: */
	242	if (pages)
	243	flush_tlb_range(vma, start, end);
20841405	244	clear_tlb_flush_pending(mm);
7da4d641 PZ	245
	246	return pages;
	247	}
	248
	249	unsigned long change_protection(struct vm_area_struct *vma, unsigned long start,
	250	unsigned long end, pgprot_t newprot,
4b10e7d5	251	int dirty_accountable, int prot_numa)
7da4d641	252	{
7da4d641 PZ	253	unsigned long pages;
7da4d641 PZ	254
7da4d641 PZ	255	if (is_vm_hugetlb_page(vma))
	256	pages = hugetlb_change_protection(vma, start, end, newprot);
	257	else
4b10e7d5	258	pages = change_protection_range(vma, start, end, newprot, dirty_accountable, prot_numa);
7da4d641 PZ	259
7da4d641 PZ	260	return pages;
1da177e4 LT	261	}
1da177e4 LT	262
b6a2fea3	263	int
1da177e4 LT	264	mprotect_fixup(struct vm_area_struct vma, struct vm_area_struct *pprev,
	265	unsigned long start, unsigned long end, unsigned long newflags)
	266	{
	267	struct mm_struct *mm = vma->vm_mm;
	268	unsigned long oldflags = vma->vm_flags;
	269	long nrpages = (end - start) >> PAGE_SHIFT;
	270	unsigned long charged = 0;
1da177e4 LT	271	pgoff_t pgoff;
1da177e4 LT	272	int error;
c1e6098b	273	int dirty_accountable = 0;
1da177e4 LT	274
	275	if (newflags == oldflags) {
	276	*pprev = vma;
	277	return 0;
	278	}
	279
	280	/*
	281	* If we make a private mapping writable we increase our commit;
	282	* but (without finer accounting) cannot reduce our commit if we
5a6fe125 MG	283	* make it unwritable again. hugetlb mapping were accounted for
5a6fe125 MG	284	* even if read-only so there is no need to account for them here
1da177e4 LT	285	*/
1da177e4 LT	286	if (newflags & VM_WRITE) {
84638335 KK	287	/* Check space limits when area turns into data. */
	288	if (!may_expand_vm(mm, newflags, nrpages) &&
	289	may_expand_vm(mm, oldflags, nrpages))
	290	return -ENOMEM;
5a6fe125	291	if (!(oldflags & (VM_ACCOUNT\|VM_WRITE\|VM_HUGETLB\|
cdfd4325	292	VM_SHARED\|VM_NORESERVE))) {
1da177e4	293	charged = nrpages;
191c5424	294	if (security_vm_enough_memory_mm(mm, charged))
1da177e4 LT	295	return -ENOMEM;
	296	newflags \|= VM_ACCOUNT;
	297	}
	298	}
	299
1da177e4 LT	300	/*
	301	* First try to merge with previous and/or next vma.
	302	*/
	303	pgoff = vma->vm_pgoff + ((start - vma->vm_start) >> PAGE_SHIFT);
	304	pprev = vma_merge(mm, pprev, start, end, newflags,
19a809af AA	305	vma->anon_vma, vma->vm_file, pgoff, vma_policy(vma),
19a809af AA	306	vma->vm_userfaultfd_ctx);
1da177e4 LT	307	if (*pprev) {
1da177e4 LT	308	vma = *pprev;
e86f15ee	309	VM_WARN_ON((vma->vm_flags ^ newflags) & ~VM_SOFTDIRTY);
1da177e4 LT	310	goto success;
	311	}
	312
	313	*pprev = vma;
	314
	315	if (start != vma->vm_start) {
	316	error = split_vma(mm, vma, start, 1);
	317	if (error)
	318	goto fail;
	319	}
	320
	321	if (end != vma->vm_end) {
	322	error = split_vma(mm, vma, end, 0);
	323	if (error)
	324	goto fail;
	325	}
	326
	327	success:
	328	/*
	329	* vm_flags and vm_page_prot are protected by the mmap_sem
	330	* held in write mode.
	331	*/
	332	vma->vm_flags = newflags;
6d2329f8	333	dirty_accountable = vma_wants_writenotify(vma, vma->vm_page_prot);
64e45507	334	vma_set_page_prot(vma);
d08b3851	335
7d12efae AM	336	change_protection(vma, start, end, vma->vm_page_prot,
7d12efae AM	337	dirty_accountable, 0);
7da4d641	338
36f88188 KS	339	/*
	340	* Private VM_LOCKED VMA becoming writable: trigger COW to avoid major
	341	* fault on access.
	342	*/
	343	if ((oldflags & (VM_WRITE \| VM_SHARED \| VM_LOCKED)) == VM_LOCKED &&
	344	(newflags & VM_WRITE)) {
	345	populate_vma_page_range(vma, start, end, NULL);
	346	}
	347
84638335 KK	348	vm_stat_account(mm, oldflags, -nrpages);
84638335 KK	349	vm_stat_account(mm, newflags, nrpages);
63bfd738	350	perf_event_mmap(vma);
1da177e4 LT	351	return 0;
	352
	353	fail:
	354	vm_unacct_memory(charged);
	355	return error;
	356	}
	357
7d06d9c9 DH	358	/*
	359	* pkey==-1 when doing a legacy mprotect()
	360	*/
	361	static int do_mprotect_pkey(unsigned long start, size_t len,
	362	unsigned long prot, int pkey)
1da177e4	363	{
62b5f7d0	364	unsigned long nstart, end, tmp, reqprot;
1da177e4 LT	365	struct vm_area_struct vma, prev;
	366	int error = -EINVAL;
	367	const int grows = prot & (PROT_GROWSDOWN\|PROT_GROWSUP);
f138556d PK	368	const bool rier = (current->personality & READ_IMPLIES_EXEC) &&
	369	(prot & PROT_READ);
	370
1da177e4 LT	371	prot &= ~(PROT_GROWSDOWN\|PROT_GROWSUP);
	372	if (grows == (PROT_GROWSDOWN\|PROT_GROWSUP)) /* can't be both */
	373	return -EINVAL;
	374
	375	if (start & ~PAGE_MASK)
	376	return -EINVAL;
	377	if (!len)
	378	return 0;
	379	len = PAGE_ALIGN(len);
	380	end = start + len;
	381	if (end <= start)
	382	return -ENOMEM;
b845f313	383	if (!arch_validate_prot(prot))
1da177e4 LT	384	return -EINVAL;
	385
	386	reqprot = prot;
1da177e4	387
dc0ef0df MH	388	if (down_write_killable(&current->mm->mmap_sem))
dc0ef0df MH	389	return -EINTR;
1da177e4	390
e8c24d3a DH	391	/*
	392	* If userspace did not allocate the pkey, do not let
	393	* them use it here.
	394	*/
	395	error = -EINVAL;
	396	if ((pkey != -1) && !mm_pkey_is_allocated(current->mm, pkey))
	397	goto out;
	398
097d5910	399	vma = find_vma(current->mm, start);
1da177e4 LT	400	error = -ENOMEM;
	401	if (!vma)
	402	goto out;
097d5910	403	prev = vma->vm_prev;
1da177e4 LT	404	if (unlikely(grows & PROT_GROWSDOWN)) {
	405	if (vma->vm_start >= end)
	406	goto out;
	407	start = vma->vm_start;
	408	error = -EINVAL;
	409	if (!(vma->vm_flags & VM_GROWSDOWN))
	410	goto out;
7d12efae	411	} else {
1da177e4 LT	412	if (vma->vm_start > start)
	413	goto out;
	414	if (unlikely(grows & PROT_GROWSUP)) {
	415	end = vma->vm_end;
	416	error = -EINVAL;
	417	if (!(vma->vm_flags & VM_GROWSUP))
	418	goto out;
	419	}
	420	}
	421	if (start > vma->vm_start)
	422	prev = vma;
	423
	424	for (nstart = start ; ; ) {
a8502b67	425	unsigned long mask_off_old_flags;
1da177e4	426	unsigned long newflags;
7d06d9c9	427	int new_vma_pkey;
1da177e4	428
7d12efae	429	/* Here we know that vma->vm_start <= nstart < vma->vm_end. */
1da177e4	430
f138556d PK	431	/* Does the application expect PROT_READ to imply PROT_EXEC */
	432	if (rier && (vma->vm_flags & VM_MAYEXEC))
	433	prot \|= PROT_EXEC;
	434
a8502b67 DH	435	/*
	436	* Each mprotect() call explicitly passes r/w/x permissions.
	437	* If a permission is not passed to mprotect(), it must be
	438	* cleared from the VMA.
	439	*/
	440	mask_off_old_flags = VM_READ \| VM_WRITE \| VM_EXEC \|
	441	ARCH_VM_PKEY_FLAGS;
	442
7d06d9c9 DH	443	new_vma_pkey = arch_override_mprotect_pkey(vma, prot, pkey);
7d06d9c9 DH	444	newflags = calc_vm_prot_bits(prot, new_vma_pkey);
a8502b67	445	newflags \|= (vma->vm_flags & ~mask_off_old_flags);
1da177e4	446
7e2cff42 PBG	447	/* newflags >> 4 shift VM_MAY% in place of VM_% */
7e2cff42 PBG	448	if ((newflags & ~(newflags >> 4)) & (VM_READ \| VM_WRITE \| VM_EXEC)) {
1da177e4 LT	449	error = -EACCES;
	450	goto out;
	451	}
	452
	453	error = security_file_mprotect(vma, reqprot, prot);
	454	if (error)
	455	goto out;
	456
	457	tmp = vma->vm_end;
	458	if (tmp > end)
	459	tmp = end;
	460	error = mprotect_fixup(vma, &prev, nstart, tmp, newflags);
	461	if (error)
	462	goto out;
	463	nstart = tmp;
	464
	465	if (nstart < prev->vm_end)
	466	nstart = prev->vm_end;
	467	if (nstart >= end)
	468	goto out;
	469
	470	vma = prev->vm_next;
	471	if (!vma \|\| vma->vm_start != nstart) {
	472	error = -ENOMEM;
	473	goto out;
	474	}
f138556d	475	prot = reqprot;
1da177e4 LT	476	}
	477	out:
	478	up_write(&current->mm->mmap_sem);
	479	return error;
	480	}
7d06d9c9 DH	481
	482	SYSCALL_DEFINE3(mprotect, unsigned long, start, size_t, len,
	483	unsigned long, prot)
	484	{
	485	return do_mprotect_pkey(start, len, prot, -1);
	486	}
	487
	488	SYSCALL_DEFINE4(pkey_mprotect, unsigned long, start, size_t, len,
	489	unsigned long, prot, int, pkey)
	490	{
	491	return do_mprotect_pkey(start, len, prot, pkey);
	492	}
e8c24d3a DH	493
	494	SYSCALL_DEFINE2(pkey_alloc, unsigned long, flags, unsigned long, init_val)
	495	{
	496	int pkey;
	497	int ret;
	498
	499	/* No flags supported yet. */
	500	if (flags)
	501	return -EINVAL;
	502	/* check for unsupported init values */
	503	if (init_val & ~PKEY_ACCESS_MASK)
	504	return -EINVAL;
	505
	506	down_write(&current->mm->mmap_sem);
	507	pkey = mm_pkey_alloc(current->mm);
	508
	509	ret = -ENOSPC;
	510	if (pkey == -1)
	511	goto out;
	512
	513	ret = arch_set_user_pkey_access(current, pkey, init_val);
	514	if (ret) {
	515	mm_pkey_free(current->mm, pkey);
	516	goto out;
	517	}
	518	ret = pkey;
	519	out:
	520	up_write(&current->mm->mmap_sem);
	521	return ret;
	522	}
	523
	524	SYSCALL_DEFINE1(pkey_free, int, pkey)
	525	{
	526	int ret;
	527
	528	down_write(&current->mm->mmap_sem);
	529	ret = mm_pkey_free(current->mm, pkey);
	530	up_write(&current->mm->mmap_sem);
	531
	532	/*
	533	* We could provie warnings or errors if any VMA still
	534	* has the pkey set here.
	535	*/
	536	return ret;
	537	}