[linux-2.6-block.git] / mm / mprotect.c

// SPDX-License-Identifier: GPL-2.0
/*
 *  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/uaccess.h>
#include <linux/mm_inline.h>
#include <asm/pgtable.h>
#include <asm/cacheflush.h>
#include <asm/mmu_context.h>
#include <asm/tlbflush.h>

#include "internal.h"

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;
	int target_node = NUMA_NO_NODE;

	/*
	 * Can be called with only the mmap_sem for reading by
	 * prot_numa so we must check the pmd isn't constantly
	 * changing from under us from pmd_none to pmd_trans_huge
	 * and/or the other way around.
	 */
	if (pmd_trans_unstable(pmd))
		return 0;

	/*
	 * The pmd points to a regular pte so the pmd can't change
	 * from under us even if the mmap_sem is only hold for
	 * reading.
	 */
	pte = pte_offset_map_lock(vma->vm_mm, pmd, addr, &ptl);

	/* Get target node for single threaded private VMAs */
	if (prot_numa && !(vma->vm_flags & VM_SHARED) &&
	    atomic_read(&vma->vm_mm->mm_users) == 1)
		target_node = numa_node_id();

	flush_tlb_batched_pending(vma->vm_mm);
	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;

				/* Also skip shared copy-on-write pages */
				if (is_cow_mapping(vma->vm_flags) &&
				    page_mapcount(page) != 1)
					continue;

				/*
				 * While migration can move some dirty pages,
				 * it cannot move them all from MIGRATE_ASYNC
				 * context.
				 */
				if (page_is_file_cache(page) && PageDirty(page))
					continue;

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

				/*
				 * Don't mess with PTEs if page is already on the node
				 * a single-threaded process is running on.
				 */
				if (target_node == page_to_nid(page))
					continue;
			}

			ptent = ptep_modify_prot_start(mm, addr, pte);
			ptent = pte_modify(ptent, newprot);
			if (preserve_write)
				ptent = pte_mk_savedwrite(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++;
			}

			if (is_write_device_private_entry(entry)) {
				pte_t newpte;

				/*
				 * We do not preserve soft-dirtiness. See
				 * copy_one_pte() for explanation.
				 */
				make_device_private_entry_read(&entry);
				newpte = swp_entry_to_pte(entry);
				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 (!is_swap_pmd(*pmd) && !pmd_trans_huge(*pmd) && !pmd_devmap(*pmd)
				&& pmd_none_or_clear_bad(pmd))
			goto next;

		/* 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 (is_swap_pmd(*pmd) || pmd_trans_huge(*pmd) || pmd_devmap(*pmd)) {
			if (next - addr != HPAGE_PMD_SIZE) {
				__split_huge_pmd(vma, pmd, addr, false, NULL);
			} 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 */
					goto next;
				}
			}
			/* 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;
next:
		cond_resched();
	} 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,
		p4d_t *p4d, 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(p4d, 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 inline unsigned long change_p4d_range(struct vm_area_struct *vma,
		pgd_t *pgd, unsigned long addr, unsigned long end,
		pgprot_t newprot, int dirty_accountable, int prot_numa)
{
	p4d_t *p4d;
	unsigned long next;
	unsigned long pages = 0;

	p4d = p4d_offset(pgd, addr);
	do {
		next = p4d_addr_end(addr, end);
		if (p4d_none_or_clear_bad(p4d))
			continue;
		pages += change_pud_range(vma, p4d, addr, next, newprot,
				 dirty_accountable, prot_numa);
	} while (p4d++, 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);
	inc_tlb_flush_pending(mm);
	do {
		next = pgd_addr_end(addr, end);
		if (pgd_none_or_clear_bad(pgd))
			continue;
		pages += change_p4d_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);
	dec_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, start))
		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 |
					VM_FLAGS_CLEAR;

		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);
}

#ifdef CONFIG_ARCH_HAS_PKEYS

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;
}

#endif /* CONFIG_ARCH_HAS_PKEYS */
Commit	Line	Data
b2441318	1	// SPDX-License-Identifier: GPL-2.0
1da177e4 LT	2	/*
	3	* mm/mprotect.c
	4	*
	5	* (C) Copyright 1994 Linus Torvalds
	6	* (C) Copyright 2002 Christoph Hellwig
	7	*
046c6884	8	* Address space accounting code <alan@lxorguk.ukuu.org.uk>
1da177e4 LT	9	* (C) Copyright 2002 Red Hat Inc, All Rights Reserved
	10	*/
	11
	12	#include <linux/mm.h>
	13	#include <linux/hugetlb.h>
1da177e4 LT	14	#include <linux/shm.h>
	15	#include <linux/mman.h>
	16	#include <linux/fs.h>
	17	#include <linux/highmem.h>
	18	#include <linux/security.h>
	19	#include <linux/mempolicy.h>
	20	#include <linux/personality.h>
	21	#include <linux/syscalls.h>
0697212a CL	22	#include <linux/swap.h>
0697212a CL	23	#include <linux/swapops.h>
cddb8a5c	24	#include <linux/mmu_notifier.h>
64cdd548	25	#include <linux/migrate.h>
cdd6c482	26	#include <linux/perf_event.h>
e8c24d3a	27	#include <linux/pkeys.h>
64a9a34e	28	#include <linux/ksm.h>
7c0f6ba6	29	#include <linux/uaccess.h>
09a913a7	30	#include <linux/mm_inline.h>
1da177e4 LT	31	#include <asm/pgtable.h>
1da177e4 LT	32	#include <asm/cacheflush.h>
e8c24d3a	33	#include <asm/mmu_context.h>
1da177e4 LT	34	#include <asm/tlbflush.h>
1da177e4 LT	35
36f88188 KS	36	#include "internal.h"
36f88188 KS	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,
0f19c179	40	int dirty_accountable, int prot_numa)
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;
3e321587	46	int target_node = NUMA_NO_NODE;
1da177e4	47
175ad4f1 AA	48	/*
	49	* Can be called with only the mmap_sem for reading by
	50	* prot_numa so we must check the pmd isn't constantly
	51	* changing from under us from pmd_none to pmd_trans_huge
	52	* and/or the other way around.
	53	*/
	54	if (pmd_trans_unstable(pmd))
	55	return 0;
	56
	57	/*
	58	* The pmd points to a regular pte so the pmd can't change
	59	* from under us even if the mmap_sem is only hold for
	60	* reading.
	61	*/
	62	pte = pte_offset_map_lock(vma->vm_mm, pmd, addr, &ptl);
1ad9f620	63
3e321587 AK	64	/* Get target node for single threaded private VMAs */
	65	if (prot_numa && !(vma->vm_flags & VM_SHARED) &&
	66	atomic_read(&vma->vm_mm->mm_users) == 1)
	67	target_node = numa_node_id();
	68
3ea27719	69	flush_tlb_batched_pending(vma->vm_mm);
6606c3e0	70	arch_enter_lazy_mmu_mode();
1da177e4	71	do {
0697212a CL	72	oldpte = *pte;
0697212a CL	73	if (pte_present(oldpte)) {
1da177e4	74	pte_t ptent;
b191f9b1	75	bool preserve_write = prot_numa && pte_write(oldpte);
1da177e4	76
e944fd67 MG	77	/*
	78	* Avoid trapping faults against the zero or KSM
	79	* pages. See similar comment in change_huge_pmd.
	80	*/
	81	if (prot_numa) {
	82	struct page *page;
	83
	84	page = vm_normal_page(vma, addr, oldpte);
	85	if (!page \|\| PageKsm(page))
	86	continue;
10c1045f	87
859d4adc HW	88	/* Also skip shared copy-on-write pages */
	89	if (is_cow_mapping(vma->vm_flags) &&
	90	page_mapcount(page) != 1)
	91	continue;
	92
09a913a7 MG	93	/*
	94	* While migration can move some dirty pages,
	95	* it cannot move them all from MIGRATE_ASYNC
	96	* context.
	97	*/
	98	if (page_is_file_cache(page) && PageDirty(page))
	99	continue;
	100
10c1045f MG	101	/* Avoid TLB flush if possible */
	102	if (pte_protnone(oldpte))
	103	continue;
3e321587 AK	104
	105	/*
	106	* Don't mess with PTEs if page is already on the node
	107	* a single-threaded process is running on.
	108	*/
	109	if (target_node == page_to_nid(page))
	110	continue;
e944fd67 MG	111	}
e944fd67 MG	112
8a0516ed MG	113	ptent = ptep_modify_prot_start(mm, addr, pte);
8a0516ed MG	114	ptent = pte_modify(ptent, newprot);
b191f9b1	115	if (preserve_write)
288bc549	116	ptent = pte_mk_savedwrite(ptent);
4b10e7d5	117
8a0516ed MG	118	/* Avoid taking write faults for known dirty pages */
	119	if (dirty_accountable && pte_dirty(ptent) &&
	120	(pte_soft_dirty(ptent) \|\|
	121	!(vma->vm_flags & VM_SOFTDIRTY))) {
	122	ptent = pte_mkwrite(ptent);
4b10e7d5	123	}
8a0516ed MG	124	ptep_modify_prot_commit(mm, addr, pte, ptent);
8a0516ed MG	125	pages++;
0661a336	126	} else if (IS_ENABLED(CONFIG_MIGRATION)) {
0697212a CL	127	swp_entry_t entry = pte_to_swp_entry(oldpte);
	128
	129	if (is_write_migration_entry(entry)) {
c3d16e16	130	pte_t newpte;
0697212a CL	131	/*
	132	* A protection check is difficult so
	133	* just be safe and disable write
	134	*/
	135	make_migration_entry_read(&entry);
c3d16e16 CG	136	newpte = swp_entry_to_pte(entry);
	137	if (pte_swp_soft_dirty(oldpte))
	138	newpte = pte_swp_mksoft_dirty(newpte);
	139	set_pte_at(mm, addr, pte, newpte);
e920e14c MG	140
e920e14c MG	141	pages++;
0697212a	142	}
5042db43 JG	143
	144	if (is_write_device_private_entry(entry)) {
	145	pte_t newpte;
	146
	147	/*
	148	* We do not preserve soft-dirtiness. See
	149	* copy_one_pte() for explanation.
	150	*/
	151	make_device_private_entry_read(&entry);
	152	newpte = swp_entry_to_pte(entry);
	153	set_pte_at(mm, addr, pte, newpte);
	154
	155	pages++;
	156	}
1da177e4 LT	157	}
1da177e4 LT	158	} while (pte++, addr += PAGE_SIZE, addr != end);
6606c3e0	159	arch_leave_lazy_mmu_mode();
705e87c0	160	pte_unmap_unlock(pte - 1, ptl);
7da4d641 PZ	161
7da4d641 PZ	162	return pages;
1da177e4 LT	163	}
1da177e4 LT	164
7d12efae AM	165	static inline unsigned long change_pmd_range(struct vm_area_struct *vma,
	166	pud_t *pud, unsigned long addr, unsigned long end,
	167	pgprot_t newprot, int dirty_accountable, int prot_numa)
1da177e4 LT	168	{
1da177e4 LT	169	pmd_t *pmd;
a5338093	170	struct mm_struct *mm = vma->vm_mm;
1da177e4	171	unsigned long next;
7da4d641	172	unsigned long pages = 0;
72403b4a	173	unsigned long nr_huge_updates = 0;
a5338093	174	unsigned long mni_start = 0;
1da177e4 LT	175
	176	pmd = pmd_offset(pud, addr);
	177	do {
25cbbef1 MG	178	unsigned long this_pages;
25cbbef1 MG	179
1da177e4	180	next = pmd_addr_end(addr, end);
84c3fc4e	181	if (!is_swap_pmd(pmd) && !pmd_trans_huge(pmd) && !pmd_devmap(*pmd)
5c7fb56e	182	&& pmd_none_or_clear_bad(pmd))
4991c09c	183	goto next;
a5338093 RR	184
	185	/* invoke the mmu notifier if the pmd is populated */
	186	if (!mni_start) {
	187	mni_start = addr;
	188	mmu_notifier_invalidate_range_start(mm, mni_start, end);
	189	}
	190
84c3fc4e	191	if (is_swap_pmd(pmd) \|\| pmd_trans_huge(pmd) \|\| pmd_devmap(*pmd)) {
6b9116a6	192	if (next - addr != HPAGE_PMD_SIZE) {
fd60775a	193	__split_huge_pmd(vma, pmd, addr, false, NULL);
6b9116a6	194	} else {
f123d74a	195	int nr_ptes = change_huge_pmd(vma, pmd, addr,
e944fd67	196	newprot, prot_numa);
f123d74a MG	197
f123d74a MG	198	if (nr_ptes) {
72403b4a MG	199	if (nr_ptes == HPAGE_PMD_NR) {
	200	pages += HPAGE_PMD_NR;
	201	nr_huge_updates++;
	202	}
1ad9f620 MG	203
1ad9f620 MG	204	/* huge pmd was handled */
4991c09c	205	goto next;
f123d74a	206	}
7da4d641	207	}
88a9ab6e	208	/* fall through, the trans huge pmd just split */
cd7548ab	209	}
25cbbef1	210	this_pages = change_pte_range(vma, pmd, addr, next, newprot,
0f19c179	211	dirty_accountable, prot_numa);
25cbbef1	212	pages += this_pages;
4991c09c AK	213	next:
4991c09c AK	214	cond_resched();
1da177e4	215	} while (pmd++, addr = next, addr != end);
7da4d641	216
a5338093 RR	217	if (mni_start)
	218	mmu_notifier_invalidate_range_end(mm, mni_start, end);
	219
72403b4a MG	220	if (nr_huge_updates)
72403b4a MG	221	count_vm_numa_events(NUMA_HUGE_PTE_UPDATES, nr_huge_updates);
7da4d641	222	return pages;
1da177e4 LT	223	}
1da177e4 LT	224
7d12efae	225	static inline unsigned long change_pud_range(struct vm_area_struct *vma,
c2febafc	226	p4d_t *p4d, unsigned long addr, unsigned long end,
7d12efae	227	pgprot_t newprot, int dirty_accountable, int prot_numa)
1da177e4 LT	228	{
	229	pud_t *pud;
	230	unsigned long next;
7da4d641	231	unsigned long pages = 0;
1da177e4	232
c2febafc	233	pud = pud_offset(p4d, addr);
1da177e4 LT	234	do {
	235	next = pud_addr_end(addr, end);
	236	if (pud_none_or_clear_bad(pud))
	237	continue;
7da4d641	238	pages += change_pmd_range(vma, pud, addr, next, newprot,
4b10e7d5	239	dirty_accountable, prot_numa);
1da177e4	240	} while (pud++, addr = next, addr != end);
7da4d641 PZ	241
7da4d641 PZ	242	return pages;
1da177e4 LT	243	}
1da177e4 LT	244
c2febafc KS	245	static inline unsigned long change_p4d_range(struct vm_area_struct *vma,
	246	pgd_t *pgd, unsigned long addr, unsigned long end,
	247	pgprot_t newprot, int dirty_accountable, int prot_numa)
	248	{
	249	p4d_t *p4d;
	250	unsigned long next;
	251	unsigned long pages = 0;
	252
	253	p4d = p4d_offset(pgd, addr);
	254	do {
	255	next = p4d_addr_end(addr, end);
	256	if (p4d_none_or_clear_bad(p4d))
	257	continue;
	258	pages += change_pud_range(vma, p4d, addr, next, newprot,
	259	dirty_accountable, prot_numa);
	260	} while (p4d++, addr = next, addr != end);
	261
	262	return pages;
	263	}
	264
7da4d641	265	static unsigned long change_protection_range(struct vm_area_struct *vma,
c1e6098b	266	unsigned long addr, unsigned long end, pgprot_t newprot,
4b10e7d5	267	int dirty_accountable, int prot_numa)
1da177e4 LT	268	{
	269	struct mm_struct *mm = vma->vm_mm;
	270	pgd_t *pgd;
	271	unsigned long next;
	272	unsigned long start = addr;
7da4d641	273	unsigned long pages = 0;
1da177e4 LT	274
	275	BUG_ON(addr >= end);
	276	pgd = pgd_offset(mm, addr);
	277	flush_cache_range(vma, addr, end);
16af97dc	278	inc_tlb_flush_pending(mm);
1da177e4 LT	279	do {
	280	next = pgd_addr_end(addr, end);
	281	if (pgd_none_or_clear_bad(pgd))
	282	continue;
c2febafc	283	pages += change_p4d_range(vma, pgd, addr, next, newprot,
4b10e7d5	284	dirty_accountable, prot_numa);
1da177e4	285	} while (pgd++, addr = next, addr != end);
7da4d641	286
1233d588 IM	287	/* Only flush the TLB if we actually modified any entries: */
	288	if (pages)
	289	flush_tlb_range(vma, start, end);
16af97dc	290	dec_tlb_flush_pending(mm);
7da4d641 PZ	291
	292	return pages;
	293	}
	294
	295	unsigned long change_protection(struct vm_area_struct *vma, unsigned long start,
	296	unsigned long end, pgprot_t newprot,
4b10e7d5	297	int dirty_accountable, int prot_numa)
7da4d641	298	{
7da4d641 PZ	299	unsigned long pages;
7da4d641 PZ	300
7da4d641 PZ	301	if (is_vm_hugetlb_page(vma))
	302	pages = hugetlb_change_protection(vma, start, end, newprot);
	303	else
4b10e7d5	304	pages = change_protection_range(vma, start, end, newprot, dirty_accountable, prot_numa);
7da4d641 PZ	305
7da4d641 PZ	306	return pages;
1da177e4 LT	307	}
1da177e4 LT	308
b6a2fea3	309	int
1da177e4 LT	310	mprotect_fixup(struct vm_area_struct vma, struct vm_area_struct *pprev,
	311	unsigned long start, unsigned long end, unsigned long newflags)
	312	{
	313	struct mm_struct *mm = vma->vm_mm;
	314	unsigned long oldflags = vma->vm_flags;
	315	long nrpages = (end - start) >> PAGE_SHIFT;
	316	unsigned long charged = 0;
1da177e4 LT	317	pgoff_t pgoff;
1da177e4 LT	318	int error;
c1e6098b	319	int dirty_accountable = 0;
1da177e4 LT	320
	321	if (newflags == oldflags) {
	322	*pprev = vma;
	323	return 0;
	324	}
	325
	326	/*
	327	* If we make a private mapping writable we increase our commit;
	328	* but (without finer accounting) cannot reduce our commit if we
5a6fe125 MG	329	* make it unwritable again. hugetlb mapping were accounted for
5a6fe125 MG	330	* even if read-only so there is no need to account for them here
1da177e4 LT	331	*/
1da177e4 LT	332	if (newflags & VM_WRITE) {
84638335 KK	333	/* Check space limits when area turns into data. */
	334	if (!may_expand_vm(mm, newflags, nrpages) &&
	335	may_expand_vm(mm, oldflags, nrpages))
	336	return -ENOMEM;
5a6fe125	337	if (!(oldflags & (VM_ACCOUNT\|VM_WRITE\|VM_HUGETLB\|
cdfd4325	338	VM_SHARED\|VM_NORESERVE))) {
1da177e4	339	charged = nrpages;
191c5424	340	if (security_vm_enough_memory_mm(mm, charged))
1da177e4 LT	341	return -ENOMEM;
	342	newflags \|= VM_ACCOUNT;
	343	}
	344	}
	345
1da177e4 LT	346	/*
	347	* First try to merge with previous and/or next vma.
	348	*/
	349	pgoff = vma->vm_pgoff + ((start - vma->vm_start) >> PAGE_SHIFT);
	350	pprev = vma_merge(mm, pprev, start, end, newflags,
19a809af AA	351	vma->anon_vma, vma->vm_file, pgoff, vma_policy(vma),
19a809af AA	352	vma->vm_userfaultfd_ctx);
1da177e4 LT	353	if (*pprev) {
1da177e4 LT	354	vma = *pprev;
e86f15ee	355	VM_WARN_ON((vma->vm_flags ^ newflags) & ~VM_SOFTDIRTY);
1da177e4 LT	356	goto success;
	357	}
	358
	359	*pprev = vma;
	360
	361	if (start != vma->vm_start) {
	362	error = split_vma(mm, vma, start, 1);
	363	if (error)
	364	goto fail;
	365	}
	366
	367	if (end != vma->vm_end) {
	368	error = split_vma(mm, vma, end, 0);
	369	if (error)
	370	goto fail;
	371	}
	372
	373	success:
	374	/*
	375	* vm_flags and vm_page_prot are protected by the mmap_sem
	376	* held in write mode.
	377	*/
	378	vma->vm_flags = newflags;
6d2329f8	379	dirty_accountable = vma_wants_writenotify(vma, vma->vm_page_prot);
64e45507	380	vma_set_page_prot(vma);
d08b3851	381
7d12efae AM	382	change_protection(vma, start, end, vma->vm_page_prot,
7d12efae AM	383	dirty_accountable, 0);
7da4d641	384
36f88188 KS	385	/*
	386	* Private VM_LOCKED VMA becoming writable: trigger COW to avoid major
	387	* fault on access.
	388	*/
	389	if ((oldflags & (VM_WRITE \| VM_SHARED \| VM_LOCKED)) == VM_LOCKED &&
	390	(newflags & VM_WRITE)) {
	391	populate_vma_page_range(vma, start, end, NULL);
	392	}
	393
84638335 KK	394	vm_stat_account(mm, oldflags, -nrpages);
84638335 KK	395	vm_stat_account(mm, newflags, nrpages);
63bfd738	396	perf_event_mmap(vma);
1da177e4 LT	397	return 0;
	398
	399	fail:
	400	vm_unacct_memory(charged);
	401	return error;
	402	}
	403
7d06d9c9 DH	404	/*
	405	* pkey==-1 when doing a legacy mprotect()
	406	*/
	407	static int do_mprotect_pkey(unsigned long start, size_t len,
	408	unsigned long prot, int pkey)
1da177e4	409	{
62b5f7d0	410	unsigned long nstart, end, tmp, reqprot;
1da177e4 LT	411	struct vm_area_struct vma, prev;
	412	int error = -EINVAL;
	413	const int grows = prot & (PROT_GROWSDOWN\|PROT_GROWSUP);
f138556d PK	414	const bool rier = (current->personality & READ_IMPLIES_EXEC) &&
	415	(prot & PROT_READ);
	416
1da177e4 LT	417	prot &= ~(PROT_GROWSDOWN\|PROT_GROWSUP);
	418	if (grows == (PROT_GROWSDOWN\|PROT_GROWSUP)) /* can't be both */
	419	return -EINVAL;
	420
	421	if (start & ~PAGE_MASK)
	422	return -EINVAL;
	423	if (!len)
	424	return 0;
	425	len = PAGE_ALIGN(len);
	426	end = start + len;
	427	if (end <= start)
	428	return -ENOMEM;
9035cf9a	429	if (!arch_validate_prot(prot, start))
1da177e4 LT	430	return -EINVAL;
	431
	432	reqprot = prot;
1da177e4	433
dc0ef0df MH	434	if (down_write_killable(&current->mm->mmap_sem))
dc0ef0df MH	435	return -EINTR;
1da177e4	436
e8c24d3a DH	437	/*
	438	* If userspace did not allocate the pkey, do not let
	439	* them use it here.
	440	*/
	441	error = -EINVAL;
	442	if ((pkey != -1) && !mm_pkey_is_allocated(current->mm, pkey))
	443	goto out;
	444
097d5910	445	vma = find_vma(current->mm, start);
1da177e4 LT	446	error = -ENOMEM;
	447	if (!vma)
	448	goto out;
097d5910	449	prev = vma->vm_prev;
1da177e4 LT	450	if (unlikely(grows & PROT_GROWSDOWN)) {
	451	if (vma->vm_start >= end)
	452	goto out;
	453	start = vma->vm_start;
	454	error = -EINVAL;
	455	if (!(vma->vm_flags & VM_GROWSDOWN))
	456	goto out;
7d12efae	457	} else {
1da177e4 LT	458	if (vma->vm_start > start)
	459	goto out;
	460	if (unlikely(grows & PROT_GROWSUP)) {
	461	end = vma->vm_end;
	462	error = -EINVAL;
	463	if (!(vma->vm_flags & VM_GROWSUP))
	464	goto out;
	465	}
	466	}
	467	if (start > vma->vm_start)
	468	prev = vma;
	469
	470	for (nstart = start ; ; ) {
a8502b67	471	unsigned long mask_off_old_flags;
1da177e4	472	unsigned long newflags;
7d06d9c9	473	int new_vma_pkey;
1da177e4	474
7d12efae	475	/* Here we know that vma->vm_start <= nstart < vma->vm_end. */
1da177e4	476
f138556d PK	477	/* Does the application expect PROT_READ to imply PROT_EXEC */
	478	if (rier && (vma->vm_flags & VM_MAYEXEC))
	479	prot \|= PROT_EXEC;
	480
a8502b67 DH	481	/*
	482	* Each mprotect() call explicitly passes r/w/x permissions.
	483	* If a permission is not passed to mprotect(), it must be
	484	* cleared from the VMA.
	485	*/
	486	mask_off_old_flags = VM_READ \| VM_WRITE \| VM_EXEC \|
2c2d57b5	487	VM_FLAGS_CLEAR;
a8502b67	488
7d06d9c9 DH	489	new_vma_pkey = arch_override_mprotect_pkey(vma, prot, pkey);
7d06d9c9 DH	490	newflags = calc_vm_prot_bits(prot, new_vma_pkey);
a8502b67	491	newflags \|= (vma->vm_flags & ~mask_off_old_flags);
1da177e4	492
7e2cff42 PBG	493	/* newflags >> 4 shift VM_MAY% in place of VM_% */
7e2cff42 PBG	494	if ((newflags & ~(newflags >> 4)) & (VM_READ \| VM_WRITE \| VM_EXEC)) {
1da177e4 LT	495	error = -EACCES;
	496	goto out;
	497	}
	498
	499	error = security_file_mprotect(vma, reqprot, prot);
	500	if (error)
	501	goto out;
	502
	503	tmp = vma->vm_end;
	504	if (tmp > end)
	505	tmp = end;
	506	error = mprotect_fixup(vma, &prev, nstart, tmp, newflags);
	507	if (error)
	508	goto out;
	509	nstart = tmp;
	510
	511	if (nstart < prev->vm_end)
	512	nstart = prev->vm_end;
	513	if (nstart >= end)
	514	goto out;
	515
	516	vma = prev->vm_next;
	517	if (!vma \|\| vma->vm_start != nstart) {
	518	error = -ENOMEM;
	519	goto out;
	520	}
f138556d	521	prot = reqprot;
1da177e4 LT	522	}
	523	out:
	524	up_write(&current->mm->mmap_sem);
	525	return error;
	526	}
7d06d9c9 DH	527
	528	SYSCALL_DEFINE3(mprotect, unsigned long, start, size_t, len,
	529	unsigned long, prot)
	530	{
	531	return do_mprotect_pkey(start, len, prot, -1);
	532	}
	533
c7142aea HC	534	#ifdef CONFIG_ARCH_HAS_PKEYS
c7142aea HC	535
7d06d9c9 DH	536	SYSCALL_DEFINE4(pkey_mprotect, unsigned long, start, size_t, len,
	537	unsigned long, prot, int, pkey)
	538	{
	539	return do_mprotect_pkey(start, len, prot, pkey);
	540	}
e8c24d3a DH	541
	542	SYSCALL_DEFINE2(pkey_alloc, unsigned long, flags, unsigned long, init_val)
	543	{
	544	int pkey;
	545	int ret;
	546
	547	/* No flags supported yet. */
	548	if (flags)
	549	return -EINVAL;
	550	/* check for unsupported init values */
	551	if (init_val & ~PKEY_ACCESS_MASK)
	552	return -EINVAL;
	553
	554	down_write(&current->mm->mmap_sem);
	555	pkey = mm_pkey_alloc(current->mm);
	556
	557	ret = -ENOSPC;
	558	if (pkey == -1)
	559	goto out;
	560
	561	ret = arch_set_user_pkey_access(current, pkey, init_val);
	562	if (ret) {
	563	mm_pkey_free(current->mm, pkey);
	564	goto out;
	565	}
	566	ret = pkey;
	567	out:
	568	up_write(&current->mm->mmap_sem);
	569	return ret;
	570	}
	571
	572	SYSCALL_DEFINE1(pkey_free, int, pkey)
	573	{
	574	int ret;
	575
	576	down_write(&current->mm->mmap_sem);
	577	ret = mm_pkey_free(current->mm, pkey);
	578	up_write(&current->mm->mmap_sem);
	579
	580	/*
	581	* We could provie warnings or errors if any VMA still
	582	* has the pkey set here.
	583	*/
	584	return ret;
	585	}
c7142aea HC	586
c7142aea HC	587	#endif /* CONFIG_ARCH_HAS_PKEYS */