[mirror_ubuntu-bionic-kernel.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/uaccess.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;

				/* 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))
			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 (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 */
					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,
		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))
		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);
}

#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
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>
7c0f6ba6	28	#include <linux/uaccess.h>
1da177e4 LT	29	#include <asm/pgtable.h>
1da177e4 LT	30	#include <asm/cacheflush.h>
e8c24d3a	31	#include <asm/mmu_context.h>
1da177e4 LT	32	#include <asm/tlbflush.h>
1da177e4 LT	33
36f88188 KS	34	#include "internal.h"
36f88188 KS	35
4b10e7d5	36	static unsigned long change_pte_range(struct vm_area_struct vma, pmd_t pmd,
c1e6098b	37	unsigned long addr, unsigned long end, pgprot_t newprot,
0f19c179	38	int dirty_accountable, int prot_numa)
1da177e4	39	{
4b10e7d5	40	struct mm_struct *mm = vma->vm_mm;
0697212a	41	pte_t *pte, oldpte;
705e87c0	42	spinlock_t *ptl;
7da4d641	43	unsigned long pages = 0;
3e321587	44	int target_node = NUMA_NO_NODE;
1da177e4	45
175ad4f1 AA	46	/*
	47	* Can be called with only the mmap_sem for reading by
	48	* prot_numa so we must check the pmd isn't constantly
	49	* changing from under us from pmd_none to pmd_trans_huge
	50	* and/or the other way around.
	51	*/
	52	if (pmd_trans_unstable(pmd))
	53	return 0;
	54
	55	/*
	56	* The pmd points to a regular pte so the pmd can't change
	57	* from under us even if the mmap_sem is only hold for
	58	* reading.
	59	*/
	60	pte = pte_offset_map_lock(vma->vm_mm, pmd, addr, &ptl);
1ad9f620	61
3e321587 AK	62	/* Get target node for single threaded private VMAs */
	63	if (prot_numa && !(vma->vm_flags & VM_SHARED) &&
	64	atomic_read(&vma->vm_mm->mm_users) == 1)
	65	target_node = numa_node_id();
	66
3ea27719	67	flush_tlb_batched_pending(vma->vm_mm);
6606c3e0	68	arch_enter_lazy_mmu_mode();
1da177e4	69	do {
0697212a CL	70	oldpte = *pte;
0697212a CL	71	if (pte_present(oldpte)) {
1da177e4	72	pte_t ptent;
b191f9b1	73	bool preserve_write = prot_numa && pte_write(oldpte);
1da177e4	74
e944fd67 MG	75	/*
	76	* Avoid trapping faults against the zero or KSM
	77	* pages. See similar comment in change_huge_pmd.
	78	*/
	79	if (prot_numa) {
	80	struct page *page;
	81
	82	page = vm_normal_page(vma, addr, oldpte);
	83	if (!page \|\| PageKsm(page))
	84	continue;
10c1045f MG	85
	86	/* Avoid TLB flush if possible */
	87	if (pte_protnone(oldpte))
	88	continue;
3e321587 AK	89
	90	/*
	91	* Don't mess with PTEs if page is already on the node
	92	* a single-threaded process is running on.
	93	*/
	94	if (target_node == page_to_nid(page))
	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	100	if (preserve_write)
288bc549	101	ptent = pte_mk_savedwrite(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	}
5042db43 JG	128
	129	if (is_write_device_private_entry(entry)) {
	130	pte_t newpte;
	131
	132	/*
	133	* We do not preserve soft-dirtiness. See
	134	* copy_one_pte() for explanation.
	135	*/
	136	make_device_private_entry_read(&entry);
	137	newpte = swp_entry_to_pte(entry);
	138	set_pte_at(mm, addr, pte, newpte);
	139
	140	pages++;
	141	}
1da177e4 LT	142	}
1da177e4 LT	143	} while (pte++, addr += PAGE_SIZE, addr != end);
6606c3e0	144	arch_leave_lazy_mmu_mode();
705e87c0	145	pte_unmap_unlock(pte - 1, ptl);
7da4d641 PZ	146
7da4d641 PZ	147	return pages;
1da177e4 LT	148	}
1da177e4 LT	149
7d12efae AM	150	static inline unsigned long change_pmd_range(struct vm_area_struct *vma,
	151	pud_t *pud, unsigned long addr, unsigned long end,
	152	pgprot_t newprot, int dirty_accountable, int prot_numa)
1da177e4 LT	153	{
1da177e4 LT	154	pmd_t *pmd;
a5338093	155	struct mm_struct *mm = vma->vm_mm;
1da177e4	156	unsigned long next;
7da4d641	157	unsigned long pages = 0;
72403b4a	158	unsigned long nr_huge_updates = 0;
a5338093	159	unsigned long mni_start = 0;
1da177e4 LT	160
	161	pmd = pmd_offset(pud, addr);
	162	do {
25cbbef1 MG	163	unsigned long this_pages;
25cbbef1 MG	164
1da177e4	165	next = pmd_addr_end(addr, end);
84c3fc4e	166	if (!is_swap_pmd(pmd) && !pmd_trans_huge(pmd) && !pmd_devmap(*pmd)
5c7fb56e	167	&& pmd_none_or_clear_bad(pmd))
88a9ab6e	168	continue;
a5338093 RR	169
	170	/* invoke the mmu notifier if the pmd is populated */
	171	if (!mni_start) {
	172	mni_start = addr;
	173	mmu_notifier_invalidate_range_start(mm, mni_start, end);
	174	}
	175
84c3fc4e	176	if (is_swap_pmd(pmd) \|\| pmd_trans_huge(pmd) \|\| pmd_devmap(*pmd)) {
6b9116a6	177	if (next - addr != HPAGE_PMD_SIZE) {
fd60775a	178	__split_huge_pmd(vma, pmd, addr, false, NULL);
6b9116a6	179	} else {
f123d74a	180	int nr_ptes = change_huge_pmd(vma, pmd, addr,
e944fd67	181	newprot, prot_numa);
f123d74a MG	182
f123d74a MG	183	if (nr_ptes) {
72403b4a MG	184	if (nr_ptes == HPAGE_PMD_NR) {
	185	pages += HPAGE_PMD_NR;
	186	nr_huge_updates++;
	187	}
1ad9f620 MG	188
1ad9f620 MG	189	/* huge pmd was handled */
f123d74a MG	190	continue;
f123d74a MG	191	}
7da4d641	192	}
88a9ab6e	193	/* fall through, the trans huge pmd just split */
cd7548ab	194	}
25cbbef1	195	this_pages = change_pte_range(vma, pmd, addr, next, newprot,
0f19c179	196	dirty_accountable, prot_numa);
25cbbef1	197	pages += this_pages;
1da177e4	198	} while (pmd++, addr = next, addr != end);
7da4d641	199
a5338093 RR	200	if (mni_start)
	201	mmu_notifier_invalidate_range_end(mm, mni_start, end);
	202
72403b4a MG	203	if (nr_huge_updates)
72403b4a MG	204	count_vm_numa_events(NUMA_HUGE_PTE_UPDATES, nr_huge_updates);
7da4d641	205	return pages;
1da177e4 LT	206	}
1da177e4 LT	207
7d12efae	208	static inline unsigned long change_pud_range(struct vm_area_struct *vma,
c2febafc	209	p4d_t *p4d, unsigned long addr, unsigned long end,
7d12efae	210	pgprot_t newprot, int dirty_accountable, int prot_numa)
1da177e4 LT	211	{
	212	pud_t *pud;
	213	unsigned long next;
7da4d641	214	unsigned long pages = 0;
1da177e4	215
c2febafc	216	pud = pud_offset(p4d, addr);
1da177e4 LT	217	do {
	218	next = pud_addr_end(addr, end);
	219	if (pud_none_or_clear_bad(pud))
	220	continue;
7da4d641	221	pages += change_pmd_range(vma, pud, addr, next, newprot,
4b10e7d5	222	dirty_accountable, prot_numa);
1da177e4	223	} while (pud++, addr = next, addr != end);
7da4d641 PZ	224
7da4d641 PZ	225	return pages;
1da177e4 LT	226	}
1da177e4 LT	227
c2febafc KS	228	static inline unsigned long change_p4d_range(struct vm_area_struct *vma,
	229	pgd_t *pgd, unsigned long addr, unsigned long end,
	230	pgprot_t newprot, int dirty_accountable, int prot_numa)
	231	{
	232	p4d_t *p4d;
	233	unsigned long next;
	234	unsigned long pages = 0;
	235
	236	p4d = p4d_offset(pgd, addr);
	237	do {
	238	next = p4d_addr_end(addr, end);
	239	if (p4d_none_or_clear_bad(p4d))
	240	continue;
	241	pages += change_pud_range(vma, p4d, addr, next, newprot,
	242	dirty_accountable, prot_numa);
	243	} while (p4d++, addr = next, addr != end);
	244
	245	return pages;
	246	}
	247
7da4d641	248	static unsigned long change_protection_range(struct vm_area_struct *vma,
c1e6098b	249	unsigned long addr, unsigned long end, pgprot_t newprot,
4b10e7d5	250	int dirty_accountable, int prot_numa)
1da177e4 LT	251	{
	252	struct mm_struct *mm = vma->vm_mm;
	253	pgd_t *pgd;
	254	unsigned long next;
	255	unsigned long start = addr;
7da4d641	256	unsigned long pages = 0;
1da177e4 LT	257
	258	BUG_ON(addr >= end);
	259	pgd = pgd_offset(mm, addr);
	260	flush_cache_range(vma, addr, end);
16af97dc	261	inc_tlb_flush_pending(mm);
1da177e4 LT	262	do {
	263	next = pgd_addr_end(addr, end);
	264	if (pgd_none_or_clear_bad(pgd))
	265	continue;
c2febafc	266	pages += change_p4d_range(vma, pgd, addr, next, newprot,
4b10e7d5	267	dirty_accountable, prot_numa);
1da177e4	268	} while (pgd++, addr = next, addr != end);
7da4d641	269
1233d588 IM	270	/* Only flush the TLB if we actually modified any entries: */
	271	if (pages)
	272	flush_tlb_range(vma, start, end);
16af97dc	273	dec_tlb_flush_pending(mm);
7da4d641 PZ	274
	275	return pages;
	276	}
	277
	278	unsigned long change_protection(struct vm_area_struct *vma, unsigned long start,
	279	unsigned long end, pgprot_t newprot,
4b10e7d5	280	int dirty_accountable, int prot_numa)
7da4d641	281	{
7da4d641 PZ	282	unsigned long pages;
7da4d641 PZ	283
7da4d641 PZ	284	if (is_vm_hugetlb_page(vma))
	285	pages = hugetlb_change_protection(vma, start, end, newprot);
	286	else
4b10e7d5	287	pages = change_protection_range(vma, start, end, newprot, dirty_accountable, prot_numa);
7da4d641 PZ	288
7da4d641 PZ	289	return pages;
1da177e4 LT	290	}
1da177e4 LT	291
b6a2fea3	292	int
1da177e4 LT	293	mprotect_fixup(struct vm_area_struct vma, struct vm_area_struct *pprev,
	294	unsigned long start, unsigned long end, unsigned long newflags)
	295	{
	296	struct mm_struct *mm = vma->vm_mm;
	297	unsigned long oldflags = vma->vm_flags;
	298	long nrpages = (end - start) >> PAGE_SHIFT;
	299	unsigned long charged = 0;
1da177e4 LT	300	pgoff_t pgoff;
1da177e4 LT	301	int error;
c1e6098b	302	int dirty_accountable = 0;
1da177e4 LT	303
	304	if (newflags == oldflags) {
	305	*pprev = vma;
	306	return 0;
	307	}
	308
	309	/*
	310	* If we make a private mapping writable we increase our commit;
	311	* but (without finer accounting) cannot reduce our commit if we
5a6fe125 MG	312	* make it unwritable again. hugetlb mapping were accounted for
5a6fe125 MG	313	* even if read-only so there is no need to account for them here
1da177e4 LT	314	*/
1da177e4 LT	315	if (newflags & VM_WRITE) {
84638335 KK	316	/* Check space limits when area turns into data. */
	317	if (!may_expand_vm(mm, newflags, nrpages) &&
	318	may_expand_vm(mm, oldflags, nrpages))
	319	return -ENOMEM;
5a6fe125	320	if (!(oldflags & (VM_ACCOUNT\|VM_WRITE\|VM_HUGETLB\|
cdfd4325	321	VM_SHARED\|VM_NORESERVE))) {
1da177e4	322	charged = nrpages;
191c5424	323	if (security_vm_enough_memory_mm(mm, charged))
1da177e4 LT	324	return -ENOMEM;
	325	newflags \|= VM_ACCOUNT;
	326	}
	327	}
	328
1da177e4 LT	329	/*
	330	* First try to merge with previous and/or next vma.
	331	*/
	332	pgoff = vma->vm_pgoff + ((start - vma->vm_start) >> PAGE_SHIFT);
	333	pprev = vma_merge(mm, pprev, start, end, newflags,
19a809af AA	334	vma->anon_vma, vma->vm_file, pgoff, vma_policy(vma),
19a809af AA	335	vma->vm_userfaultfd_ctx);
1da177e4 LT	336	if (*pprev) {
1da177e4 LT	337	vma = *pprev;
e86f15ee	338	VM_WARN_ON((vma->vm_flags ^ newflags) & ~VM_SOFTDIRTY);
1da177e4 LT	339	goto success;
	340	}
	341
	342	*pprev = vma;
	343
	344	if (start != vma->vm_start) {
	345	error = split_vma(mm, vma, start, 1);
	346	if (error)
	347	goto fail;
	348	}
	349
	350	if (end != vma->vm_end) {
	351	error = split_vma(mm, vma, end, 0);
	352	if (error)
	353	goto fail;
	354	}
	355
	356	success:
	357	/*
	358	* vm_flags and vm_page_prot are protected by the mmap_sem
	359	* held in write mode.
	360	*/
	361	vma->vm_flags = newflags;
6d2329f8	362	dirty_accountable = vma_wants_writenotify(vma, vma->vm_page_prot);
64e45507	363	vma_set_page_prot(vma);
d08b3851	364
7d12efae AM	365	change_protection(vma, start, end, vma->vm_page_prot,
7d12efae AM	366	dirty_accountable, 0);
7da4d641	367
36f88188 KS	368	/*
	369	* Private VM_LOCKED VMA becoming writable: trigger COW to avoid major
	370	* fault on access.
	371	*/
	372	if ((oldflags & (VM_WRITE \| VM_SHARED \| VM_LOCKED)) == VM_LOCKED &&
	373	(newflags & VM_WRITE)) {
	374	populate_vma_page_range(vma, start, end, NULL);
	375	}
	376
84638335 KK	377	vm_stat_account(mm, oldflags, -nrpages);
84638335 KK	378	vm_stat_account(mm, newflags, nrpages);
63bfd738	379	perf_event_mmap(vma);
1da177e4 LT	380	return 0;
	381
	382	fail:
	383	vm_unacct_memory(charged);
	384	return error;
	385	}
	386
7d06d9c9 DH	387	/*
	388	* pkey==-1 when doing a legacy mprotect()
	389	*/
	390	static int do_mprotect_pkey(unsigned long start, size_t len,
	391	unsigned long prot, int pkey)
1da177e4	392	{
62b5f7d0	393	unsigned long nstart, end, tmp, reqprot;
1da177e4 LT	394	struct vm_area_struct vma, prev;
	395	int error = -EINVAL;
	396	const int grows = prot & (PROT_GROWSDOWN\|PROT_GROWSUP);
f138556d PK	397	const bool rier = (current->personality & READ_IMPLIES_EXEC) &&
	398	(prot & PROT_READ);
	399
1da177e4 LT	400	prot &= ~(PROT_GROWSDOWN\|PROT_GROWSUP);
	401	if (grows == (PROT_GROWSDOWN\|PROT_GROWSUP)) /* can't be both */
	402	return -EINVAL;
	403
	404	if (start & ~PAGE_MASK)
	405	return -EINVAL;
	406	if (!len)
	407	return 0;
	408	len = PAGE_ALIGN(len);
	409	end = start + len;
	410	if (end <= start)
	411	return -ENOMEM;
b845f313	412	if (!arch_validate_prot(prot))
1da177e4 LT	413	return -EINVAL;
	414
	415	reqprot = prot;
1da177e4	416
dc0ef0df MH	417	if (down_write_killable(&current->mm->mmap_sem))
dc0ef0df MH	418	return -EINTR;
1da177e4	419
e8c24d3a DH	420	/*
	421	* If userspace did not allocate the pkey, do not let
	422	* them use it here.
	423	*/
	424	error = -EINVAL;
	425	if ((pkey != -1) && !mm_pkey_is_allocated(current->mm, pkey))
	426	goto out;
	427
097d5910	428	vma = find_vma(current->mm, start);
1da177e4 LT	429	error = -ENOMEM;
	430	if (!vma)
	431	goto out;
097d5910	432	prev = vma->vm_prev;
1da177e4 LT	433	if (unlikely(grows & PROT_GROWSDOWN)) {
	434	if (vma->vm_start >= end)
	435	goto out;
	436	start = vma->vm_start;
	437	error = -EINVAL;
	438	if (!(vma->vm_flags & VM_GROWSDOWN))
	439	goto out;
7d12efae	440	} else {
1da177e4 LT	441	if (vma->vm_start > start)
	442	goto out;
	443	if (unlikely(grows & PROT_GROWSUP)) {
	444	end = vma->vm_end;
	445	error = -EINVAL;
	446	if (!(vma->vm_flags & VM_GROWSUP))
	447	goto out;
	448	}
	449	}
	450	if (start > vma->vm_start)
	451	prev = vma;
	452
	453	for (nstart = start ; ; ) {
a8502b67	454	unsigned long mask_off_old_flags;
1da177e4	455	unsigned long newflags;
7d06d9c9	456	int new_vma_pkey;
1da177e4	457
7d12efae	458	/* Here we know that vma->vm_start <= nstart < vma->vm_end. */
1da177e4	459
f138556d PK	460	/* Does the application expect PROT_READ to imply PROT_EXEC */
	461	if (rier && (vma->vm_flags & VM_MAYEXEC))
	462	prot \|= PROT_EXEC;
	463
a8502b67 DH	464	/*
	465	* Each mprotect() call explicitly passes r/w/x permissions.
	466	* If a permission is not passed to mprotect(), it must be
	467	* cleared from the VMA.
	468	*/
	469	mask_off_old_flags = VM_READ \| VM_WRITE \| VM_EXEC \|
	470	ARCH_VM_PKEY_FLAGS;
	471
7d06d9c9 DH	472	new_vma_pkey = arch_override_mprotect_pkey(vma, prot, pkey);
7d06d9c9 DH	473	newflags = calc_vm_prot_bits(prot, new_vma_pkey);
a8502b67	474	newflags \|= (vma->vm_flags & ~mask_off_old_flags);
1da177e4	475
7e2cff42 PBG	476	/* newflags >> 4 shift VM_MAY% in place of VM_% */
7e2cff42 PBG	477	if ((newflags & ~(newflags >> 4)) & (VM_READ \| VM_WRITE \| VM_EXEC)) {
1da177e4 LT	478	error = -EACCES;
	479	goto out;
	480	}
	481
	482	error = security_file_mprotect(vma, reqprot, prot);
	483	if (error)
	484	goto out;
	485
	486	tmp = vma->vm_end;
	487	if (tmp > end)
	488	tmp = end;
	489	error = mprotect_fixup(vma, &prev, nstart, tmp, newflags);
	490	if (error)
	491	goto out;
	492	nstart = tmp;
	493
	494	if (nstart < prev->vm_end)
	495	nstart = prev->vm_end;
	496	if (nstart >= end)
	497	goto out;
	498
	499	vma = prev->vm_next;
	500	if (!vma \|\| vma->vm_start != nstart) {
	501	error = -ENOMEM;
	502	goto out;
	503	}
f138556d	504	prot = reqprot;
1da177e4 LT	505	}
	506	out:
	507	up_write(&current->mm->mmap_sem);
	508	return error;
	509	}
7d06d9c9 DH	510
	511	SYSCALL_DEFINE3(mprotect, unsigned long, start, size_t, len,
	512	unsigned long, prot)
	513	{
	514	return do_mprotect_pkey(start, len, prot, -1);
	515	}
	516
c7142aea HC	517	#ifdef CONFIG_ARCH_HAS_PKEYS
c7142aea HC	518
7d06d9c9 DH	519	SYSCALL_DEFINE4(pkey_mprotect, unsigned long, start, size_t, len,
	520	unsigned long, prot, int, pkey)
	521	{
	522	return do_mprotect_pkey(start, len, prot, pkey);
	523	}
e8c24d3a DH	524
	525	SYSCALL_DEFINE2(pkey_alloc, unsigned long, flags, unsigned long, init_val)
	526	{
	527	int pkey;
	528	int ret;
	529
	530	/* No flags supported yet. */
	531	if (flags)
	532	return -EINVAL;
	533	/* check for unsupported init values */
	534	if (init_val & ~PKEY_ACCESS_MASK)
	535	return -EINVAL;
	536
	537	down_write(&current->mm->mmap_sem);
	538	pkey = mm_pkey_alloc(current->mm);
	539
	540	ret = -ENOSPC;
	541	if (pkey == -1)
	542	goto out;
	543
	544	ret = arch_set_user_pkey_access(current, pkey, init_val);
	545	if (ret) {
	546	mm_pkey_free(current->mm, pkey);
	547	goto out;
	548	}
	549	ret = pkey;
	550	out:
	551	up_write(&current->mm->mmap_sem);
	552	return ret;
	553	}
	554
	555	SYSCALL_DEFINE1(pkey_free, int, pkey)
	556	{
	557	int ret;
	558
	559	down_write(&current->mm->mmap_sem);
	560	ret = mm_pkey_free(current->mm, pkey);
	561	up_write(&current->mm->mmap_sem);
	562
	563	/*
	564	* We could provie warnings or errors if any VMA still
	565	* has the pkey set here.
	566	*/
	567	return ret;
	568	}
c7142aea HC	569
c7142aea HC	570	#endif /* CONFIG_ARCH_HAS_PKEYS */