[mirror_ubuntu-bionic-kernel.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 <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
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>
1da177e4 LT	30	#include <asm/pgtable.h>
1da177e4 LT	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
4b10e7d5	37	static unsigned long change_pte_range(struct vm_area_struct vma, pmd_t pmd,
c1e6098b	38	unsigned long addr, unsigned long end, pgprot_t newprot,
0f19c179	39	int dirty_accountable, int prot_numa)
1da177e4	40	{
4b10e7d5	41	struct mm_struct *mm = vma->vm_mm;
0697212a	42	pte_t *pte, oldpte;
705e87c0	43	spinlock_t *ptl;
7da4d641	44	unsigned long pages = 0;
3e321587	45	int target_node = NUMA_NO_NODE;
1da177e4	46
175ad4f1 AA	47	/*
	48	* Can be called with only the mmap_sem for reading by
	49	* prot_numa so we must check the pmd isn't constantly
	50	* changing from under us from pmd_none to pmd_trans_huge
	51	* and/or the other way around.
	52	*/
	53	if (pmd_trans_unstable(pmd))
	54	return 0;
	55
	56	/*
	57	* The pmd points to a regular pte so the pmd can't change
	58	* from under us even if the mmap_sem is only hold for
	59	* reading.
	60	*/
	61	pte = pte_offset_map_lock(vma->vm_mm, pmd, addr, &ptl);
1ad9f620	62
3e321587 AK	63	/* Get target node for single threaded private VMAs */
	64	if (prot_numa && !(vma->vm_flags & VM_SHARED) &&
	65	atomic_read(&vma->vm_mm->mm_users) == 1)
	66	target_node = numa_node_id();
	67
3ea27719	68	flush_tlb_batched_pending(vma->vm_mm);
6606c3e0	69	arch_enter_lazy_mmu_mode();
1da177e4	70	do {
0697212a CL	71	oldpte = *pte;
0697212a CL	72	if (pte_present(oldpte)) {
1da177e4	73	pte_t ptent;
b191f9b1	74	bool preserve_write = prot_numa && pte_write(oldpte);
1da177e4	75
e944fd67 MG	76	/*
	77	* Avoid trapping faults against the zero or KSM
	78	* pages. See similar comment in change_huge_pmd.
	79	*/
	80	if (prot_numa) {
	81	struct page *page;
	82
	83	page = vm_normal_page(vma, addr, oldpte);
	84	if (!page \|\| PageKsm(page))
	85	continue;
10c1045f MG	86
	87	/* Avoid TLB flush if possible */
	88	if (pte_protnone(oldpte))
	89	continue;
3e321587 AK	90
	91	/*
	92	* Don't mess with PTEs if page is already on the node
	93	* a single-threaded process is running on.
	94	*/
	95	if (target_node == page_to_nid(page))
	96	continue;
e944fd67 MG	97	}
e944fd67 MG	98
8a0516ed MG	99	ptent = ptep_modify_prot_start(mm, addr, pte);
8a0516ed MG	100	ptent = pte_modify(ptent, newprot);
b191f9b1	101	if (preserve_write)
288bc549	102	ptent = pte_mk_savedwrite(ptent);
4b10e7d5	103
8a0516ed MG	104	/* Avoid taking write faults for known dirty pages */
	105	if (dirty_accountable && pte_dirty(ptent) &&
	106	(pte_soft_dirty(ptent) \|\|
	107	!(vma->vm_flags & VM_SOFTDIRTY))) {
	108	ptent = pte_mkwrite(ptent);
4b10e7d5	109	}
8a0516ed MG	110	ptep_modify_prot_commit(mm, addr, pte, ptent);
8a0516ed MG	111	pages++;
0661a336	112	} else if (IS_ENABLED(CONFIG_MIGRATION)) {
0697212a CL	113	swp_entry_t entry = pte_to_swp_entry(oldpte);
	114
	115	if (is_write_migration_entry(entry)) {
c3d16e16	116	pte_t newpte;
0697212a CL	117	/*
	118	* A protection check is difficult so
	119	* just be safe and disable write
	120	*/
	121	make_migration_entry_read(&entry);
c3d16e16 CG	122	newpte = swp_entry_to_pte(entry);
	123	if (pte_swp_soft_dirty(oldpte))
	124	newpte = pte_swp_mksoft_dirty(newpte);
	125	set_pte_at(mm, addr, pte, newpte);
e920e14c MG	126
e920e14c MG	127	pages++;
0697212a	128	}
5042db43 JG	129
	130	if (is_write_device_private_entry(entry)) {
	131	pte_t newpte;
	132
	133	/*
	134	* We do not preserve soft-dirtiness. See
	135	* copy_one_pte() for explanation.
	136	*/
	137	make_device_private_entry_read(&entry);
	138	newpte = swp_entry_to_pte(entry);
	139	set_pte_at(mm, addr, pte, newpte);
	140
	141	pages++;
	142	}
1da177e4 LT	143	}
1da177e4 LT	144	} while (pte++, addr += PAGE_SIZE, addr != end);
6606c3e0	145	arch_leave_lazy_mmu_mode();
705e87c0	146	pte_unmap_unlock(pte - 1, ptl);
7da4d641 PZ	147
7da4d641 PZ	148	return pages;
1da177e4 LT	149	}
1da177e4 LT	150
7d12efae AM	151	static inline unsigned long change_pmd_range(struct vm_area_struct *vma,
	152	pud_t *pud, unsigned long addr, unsigned long end,
	153	pgprot_t newprot, int dirty_accountable, int prot_numa)
1da177e4 LT	154	{
1da177e4 LT	155	pmd_t *pmd;
a5338093	156	struct mm_struct *mm = vma->vm_mm;
1da177e4	157	unsigned long next;
7da4d641	158	unsigned long pages = 0;
72403b4a	159	unsigned long nr_huge_updates = 0;
a5338093	160	unsigned long mni_start = 0;
1da177e4 LT	161
	162	pmd = pmd_offset(pud, addr);
	163	do {
25cbbef1 MG	164	unsigned long this_pages;
25cbbef1 MG	165
1da177e4	166	next = pmd_addr_end(addr, end);
84c3fc4e	167	if (!is_swap_pmd(pmd) && !pmd_trans_huge(pmd) && !pmd_devmap(*pmd)
5c7fb56e	168	&& pmd_none_or_clear_bad(pmd))
88a9ab6e	169	continue;
a5338093 RR	170
	171	/* invoke the mmu notifier if the pmd is populated */
	172	if (!mni_start) {
	173	mni_start = addr;
	174	mmu_notifier_invalidate_range_start(mm, mni_start, end);
	175	}
	176
84c3fc4e	177	if (is_swap_pmd(pmd) \|\| pmd_trans_huge(pmd) \|\| pmd_devmap(*pmd)) {
6b9116a6	178	if (next - addr != HPAGE_PMD_SIZE) {
fd60775a	179	__split_huge_pmd(vma, pmd, addr, false, NULL);
6b9116a6	180	} else {
f123d74a	181	int nr_ptes = change_huge_pmd(vma, pmd, addr,
e944fd67	182	newprot, prot_numa);
f123d74a MG	183
f123d74a MG	184	if (nr_ptes) {
72403b4a MG	185	if (nr_ptes == HPAGE_PMD_NR) {
	186	pages += HPAGE_PMD_NR;
	187	nr_huge_updates++;
	188	}
1ad9f620 MG	189
1ad9f620 MG	190	/* huge pmd was handled */
f123d74a MG	191	continue;
f123d74a MG	192	}
7da4d641	193	}
88a9ab6e	194	/* fall through, the trans huge pmd just split */
cd7548ab	195	}
25cbbef1	196	this_pages = change_pte_range(vma, pmd, addr, next, newprot,
0f19c179	197	dirty_accountable, prot_numa);
25cbbef1	198	pages += this_pages;
1da177e4	199	} while (pmd++, addr = next, addr != end);
7da4d641	200
a5338093 RR	201	if (mni_start)
	202	mmu_notifier_invalidate_range_end(mm, mni_start, end);
	203
72403b4a MG	204	if (nr_huge_updates)
72403b4a MG	205	count_vm_numa_events(NUMA_HUGE_PTE_UPDATES, nr_huge_updates);
7da4d641	206	return pages;
1da177e4 LT	207	}
1da177e4 LT	208
7d12efae	209	static inline unsigned long change_pud_range(struct vm_area_struct *vma,
c2febafc	210	p4d_t *p4d, unsigned long addr, unsigned long end,
7d12efae	211	pgprot_t newprot, int dirty_accountable, int prot_numa)
1da177e4 LT	212	{
	213	pud_t *pud;
	214	unsigned long next;
7da4d641	215	unsigned long pages = 0;
1da177e4	216
c2febafc	217	pud = pud_offset(p4d, addr);
1da177e4 LT	218	do {
	219	next = pud_addr_end(addr, end);
	220	if (pud_none_or_clear_bad(pud))
	221	continue;
7da4d641	222	pages += change_pmd_range(vma, pud, addr, next, newprot,
4b10e7d5	223	dirty_accountable, prot_numa);
1da177e4	224	} while (pud++, addr = next, addr != end);
7da4d641 PZ	225
7da4d641 PZ	226	return pages;
1da177e4 LT	227	}
1da177e4 LT	228
c2febafc KS	229	static inline unsigned long change_p4d_range(struct vm_area_struct *vma,
	230	pgd_t *pgd, unsigned long addr, unsigned long end,
	231	pgprot_t newprot, int dirty_accountable, int prot_numa)
	232	{
	233	p4d_t *p4d;
	234	unsigned long next;
	235	unsigned long pages = 0;
	236
	237	p4d = p4d_offset(pgd, addr);
	238	do {
	239	next = p4d_addr_end(addr, end);
	240	if (p4d_none_or_clear_bad(p4d))
	241	continue;
	242	pages += change_pud_range(vma, p4d, addr, next, newprot,
	243	dirty_accountable, prot_numa);
	244	} while (p4d++, addr = next, addr != end);
	245
	246	return pages;
	247	}
	248
7da4d641	249	static unsigned long change_protection_range(struct vm_area_struct *vma,
c1e6098b	250	unsigned long addr, unsigned long end, pgprot_t newprot,
4b10e7d5	251	int dirty_accountable, int prot_numa)
1da177e4 LT	252	{
	253	struct mm_struct *mm = vma->vm_mm;
	254	pgd_t *pgd;
	255	unsigned long next;
	256	unsigned long start = addr;
7da4d641	257	unsigned long pages = 0;
1da177e4 LT	258
	259	BUG_ON(addr >= end);
	260	pgd = pgd_offset(mm, addr);
	261	flush_cache_range(vma, addr, end);
16af97dc	262	inc_tlb_flush_pending(mm);
1da177e4 LT	263	do {
	264	next = pgd_addr_end(addr, end);
	265	if (pgd_none_or_clear_bad(pgd))
	266	continue;
c2febafc	267	pages += change_p4d_range(vma, pgd, addr, next, newprot,
4b10e7d5	268	dirty_accountable, prot_numa);
1da177e4	269	} while (pgd++, addr = next, addr != end);
7da4d641	270
1233d588 IM	271	/* Only flush the TLB if we actually modified any entries: */
	272	if (pages)
	273	flush_tlb_range(vma, start, end);
16af97dc	274	dec_tlb_flush_pending(mm);
7da4d641 PZ	275
	276	return pages;
	277	}
	278
	279	unsigned long change_protection(struct vm_area_struct *vma, unsigned long start,
	280	unsigned long end, pgprot_t newprot,
4b10e7d5	281	int dirty_accountable, int prot_numa)
7da4d641	282	{
7da4d641 PZ	283	unsigned long pages;
7da4d641 PZ	284
7da4d641 PZ	285	if (is_vm_hugetlb_page(vma))
	286	pages = hugetlb_change_protection(vma, start, end, newprot);
	287	else
4b10e7d5	288	pages = change_protection_range(vma, start, end, newprot, dirty_accountable, prot_numa);
7da4d641 PZ	289
7da4d641 PZ	290	return pages;
1da177e4 LT	291	}
1da177e4 LT	292
b6a2fea3	293	int
1da177e4 LT	294	mprotect_fixup(struct vm_area_struct vma, struct vm_area_struct *pprev,
	295	unsigned long start, unsigned long end, unsigned long newflags)
	296	{
	297	struct mm_struct *mm = vma->vm_mm;
	298	unsigned long oldflags = vma->vm_flags;
	299	long nrpages = (end - start) >> PAGE_SHIFT;
	300	unsigned long charged = 0;
1da177e4 LT	301	pgoff_t pgoff;
1da177e4 LT	302	int error;
c1e6098b	303	int dirty_accountable = 0;
1da177e4 LT	304
	305	if (newflags == oldflags) {
	306	*pprev = vma;
	307	return 0;
	308	}
	309
	310	/*
	311	* If we make a private mapping writable we increase our commit;
	312	* but (without finer accounting) cannot reduce our commit if we
5a6fe125 MG	313	* make it unwritable again. hugetlb mapping were accounted for
5a6fe125 MG	314	* even if read-only so there is no need to account for them here
1da177e4 LT	315	*/
1da177e4 LT	316	if (newflags & VM_WRITE) {
84638335 KK	317	/* Check space limits when area turns into data. */
	318	if (!may_expand_vm(mm, newflags, nrpages) &&
	319	may_expand_vm(mm, oldflags, nrpages))
	320	return -ENOMEM;
5a6fe125	321	if (!(oldflags & (VM_ACCOUNT\|VM_WRITE\|VM_HUGETLB\|
cdfd4325	322	VM_SHARED\|VM_NORESERVE))) {
1da177e4	323	charged = nrpages;
191c5424	324	if (security_vm_enough_memory_mm(mm, charged))
1da177e4 LT	325	return -ENOMEM;
	326	newflags \|= VM_ACCOUNT;
	327	}
	328	}
	329
1da177e4 LT	330	/*
	331	* First try to merge with previous and/or next vma.
	332	*/
	333	pgoff = vma->vm_pgoff + ((start - vma->vm_start) >> PAGE_SHIFT);
	334	pprev = vma_merge(mm, pprev, start, end, newflags,
19a809af AA	335	vma->anon_vma, vma->vm_file, pgoff, vma_policy(vma),
19a809af AA	336	vma->vm_userfaultfd_ctx);
1da177e4 LT	337	if (*pprev) {
1da177e4 LT	338	vma = *pprev;
e86f15ee	339	VM_WARN_ON((vma->vm_flags ^ newflags) & ~VM_SOFTDIRTY);
1da177e4 LT	340	goto success;
	341	}
	342
	343	*pprev = vma;
	344
	345	if (start != vma->vm_start) {
	346	error = split_vma(mm, vma, start, 1);
	347	if (error)
	348	goto fail;
	349	}
	350
	351	if (end != vma->vm_end) {
	352	error = split_vma(mm, vma, end, 0);
	353	if (error)
	354	goto fail;
	355	}
	356
	357	success:
	358	/*
	359	* vm_flags and vm_page_prot are protected by the mmap_sem
	360	* held in write mode.
	361	*/
	362	vma->vm_flags = newflags;
6d2329f8	363	dirty_accountable = vma_wants_writenotify(vma, vma->vm_page_prot);
64e45507	364	vma_set_page_prot(vma);
d08b3851	365
7d12efae AM	366	change_protection(vma, start, end, vma->vm_page_prot,
7d12efae AM	367	dirty_accountable, 0);
7da4d641	368
36f88188 KS	369	/*
	370	* Private VM_LOCKED VMA becoming writable: trigger COW to avoid major
	371	* fault on access.
	372	*/
	373	if ((oldflags & (VM_WRITE \| VM_SHARED \| VM_LOCKED)) == VM_LOCKED &&
	374	(newflags & VM_WRITE)) {
	375	populate_vma_page_range(vma, start, end, NULL);
	376	}
	377
84638335 KK	378	vm_stat_account(mm, oldflags, -nrpages);
84638335 KK	379	vm_stat_account(mm, newflags, nrpages);
63bfd738	380	perf_event_mmap(vma);
1da177e4 LT	381	return 0;
	382
	383	fail:
	384	vm_unacct_memory(charged);
	385	return error;
	386	}
	387
7d06d9c9 DH	388	/*
	389	* pkey==-1 when doing a legacy mprotect()
	390	*/
	391	static int do_mprotect_pkey(unsigned long start, size_t len,
	392	unsigned long prot, int pkey)
1da177e4	393	{
62b5f7d0	394	unsigned long nstart, end, tmp, reqprot;
1da177e4 LT	395	struct vm_area_struct vma, prev;
	396	int error = -EINVAL;
	397	const int grows = prot & (PROT_GROWSDOWN\|PROT_GROWSUP);
f138556d PK	398	const bool rier = (current->personality & READ_IMPLIES_EXEC) &&
	399	(prot & PROT_READ);
	400
1da177e4 LT	401	prot &= ~(PROT_GROWSDOWN\|PROT_GROWSUP);
	402	if (grows == (PROT_GROWSDOWN\|PROT_GROWSUP)) /* can't be both */
	403	return -EINVAL;
	404
	405	if (start & ~PAGE_MASK)
	406	return -EINVAL;
	407	if (!len)
	408	return 0;
	409	len = PAGE_ALIGN(len);
	410	end = start + len;
	411	if (end <= start)
	412	return -ENOMEM;
b845f313	413	if (!arch_validate_prot(prot))
1da177e4 LT	414	return -EINVAL;
	415
	416	reqprot = prot;
1da177e4	417
dc0ef0df MH	418	if (down_write_killable(&current->mm->mmap_sem))
dc0ef0df MH	419	return -EINTR;
1da177e4	420
e8c24d3a DH	421	/*
	422	* If userspace did not allocate the pkey, do not let
	423	* them use it here.
	424	*/
	425	error = -EINVAL;
	426	if ((pkey != -1) && !mm_pkey_is_allocated(current->mm, pkey))
	427	goto out;
	428
097d5910	429	vma = find_vma(current->mm, start);
1da177e4 LT	430	error = -ENOMEM;
	431	if (!vma)
	432	goto out;
097d5910	433	prev = vma->vm_prev;
1da177e4 LT	434	if (unlikely(grows & PROT_GROWSDOWN)) {
	435	if (vma->vm_start >= end)
	436	goto out;
	437	start = vma->vm_start;
	438	error = -EINVAL;
	439	if (!(vma->vm_flags & VM_GROWSDOWN))
	440	goto out;
7d12efae	441	} else {
1da177e4 LT	442	if (vma->vm_start > start)
	443	goto out;
	444	if (unlikely(grows & PROT_GROWSUP)) {
	445	end = vma->vm_end;
	446	error = -EINVAL;
	447	if (!(vma->vm_flags & VM_GROWSUP))
	448	goto out;
	449	}
	450	}
	451	if (start > vma->vm_start)
	452	prev = vma;
	453
	454	for (nstart = start ; ; ) {
a8502b67	455	unsigned long mask_off_old_flags;
1da177e4	456	unsigned long newflags;
7d06d9c9	457	int new_vma_pkey;
1da177e4	458
7d12efae	459	/* Here we know that vma->vm_start <= nstart < vma->vm_end. */
1da177e4	460
f138556d PK	461	/* Does the application expect PROT_READ to imply PROT_EXEC */
	462	if (rier && (vma->vm_flags & VM_MAYEXEC))
	463	prot \|= PROT_EXEC;
	464
a8502b67 DH	465	/*
	466	* Each mprotect() call explicitly passes r/w/x permissions.
	467	* If a permission is not passed to mprotect(), it must be
	468	* cleared from the VMA.
	469	*/
	470	mask_off_old_flags = VM_READ \| VM_WRITE \| VM_EXEC \|
	471	ARCH_VM_PKEY_FLAGS;
	472
7d06d9c9 DH	473	new_vma_pkey = arch_override_mprotect_pkey(vma, prot, pkey);
7d06d9c9 DH	474	newflags = calc_vm_prot_bits(prot, new_vma_pkey);
a8502b67	475	newflags \|= (vma->vm_flags & ~mask_off_old_flags);
1da177e4	476
7e2cff42 PBG	477	/* newflags >> 4 shift VM_MAY% in place of VM_% */
7e2cff42 PBG	478	if ((newflags & ~(newflags >> 4)) & (VM_READ \| VM_WRITE \| VM_EXEC)) {
1da177e4 LT	479	error = -EACCES;
	480	goto out;
	481	}
	482
	483	error = security_file_mprotect(vma, reqprot, prot);
	484	if (error)
	485	goto out;
	486
	487	tmp = vma->vm_end;
	488	if (tmp > end)
	489	tmp = end;
	490	error = mprotect_fixup(vma, &prev, nstart, tmp, newflags);
	491	if (error)
	492	goto out;
	493	nstart = tmp;
	494
	495	if (nstart < prev->vm_end)
	496	nstart = prev->vm_end;
	497	if (nstart >= end)
	498	goto out;
	499
	500	vma = prev->vm_next;
	501	if (!vma \|\| vma->vm_start != nstart) {
	502	error = -ENOMEM;
	503	goto out;
	504	}
f138556d	505	prot = reqprot;
1da177e4 LT	506	}
	507	out:
	508	up_write(&current->mm->mmap_sem);
	509	return error;
	510	}
7d06d9c9 DH	511
	512	SYSCALL_DEFINE3(mprotect, unsigned long, start, size_t, len,
	513	unsigned long, prot)
	514	{
	515	return do_mprotect_pkey(start, len, prot, -1);
	516	}
	517
c7142aea HC	518	#ifdef CONFIG_ARCH_HAS_PKEYS
c7142aea HC	519
7d06d9c9 DH	520	SYSCALL_DEFINE4(pkey_mprotect, unsigned long, start, size_t, len,
	521	unsigned long, prot, int, pkey)
	522	{
	523	return do_mprotect_pkey(start, len, prot, pkey);
	524	}
e8c24d3a DH	525
	526	SYSCALL_DEFINE2(pkey_alloc, unsigned long, flags, unsigned long, init_val)
	527	{
	528	int pkey;
	529	int ret;
	530
	531	/* No flags supported yet. */
	532	if (flags)
	533	return -EINVAL;
	534	/* check for unsupported init values */
	535	if (init_val & ~PKEY_ACCESS_MASK)
	536	return -EINVAL;
	537
	538	down_write(&current->mm->mmap_sem);
	539	pkey = mm_pkey_alloc(current->mm);
	540
	541	ret = -ENOSPC;
	542	if (pkey == -1)
	543	goto out;
	544
	545	ret = arch_set_user_pkey_access(current, pkey, init_val);
	546	if (ret) {
	547	mm_pkey_free(current->mm, pkey);
	548	goto out;
	549	}
	550	ret = pkey;
	551	out:
	552	up_write(&current->mm->mmap_sem);
	553	return ret;
	554	}
	555
	556	SYSCALL_DEFINE1(pkey_free, int, pkey)
	557	{
	558	int ret;
	559
	560	down_write(&current->mm->mmap_sem);
	561	ret = mm_pkey_free(current->mm, pkey);
	562	up_write(&current->mm->mmap_sem);
	563
	564	/*
	565	* We could provie warnings or errors if any VMA still
	566	* has the pkey set here.
	567	*/
	568	return ret;
	569	}
c7142aea HC	570
c7142aea HC	571	#endif /* CONFIG_ARCH_HAS_PKEYS */