[linux.git] / mm / mprotect.c

/*
 *  mm/mprotect.c
 *
 *  (C) Copyright 1994 Linus Torvalds
 *  (C) Copyright 2002 Christoph Hellwig
 *
 *  Address space accounting code	<[email protected]>
 *  (C) Copyright 2002 Red Hat Inc, All Rights Reserved
 */

#include <linux/mm.h>
#include <linux/hugetlb.h>
#include <linux/slab.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 <asm/uaccess.h>
#include <asm/pgtable.h>
#include <asm/cacheflush.h>
#include <asm/tlbflush.h>

static void change_pte_range(struct mm_struct *mm, pmd_t *pmd,
		unsigned long addr, unsigned long end, pgprot_t newprot,
		int dirty_accountable)
{
	pte_t *pte, oldpte;
	spinlock_t *ptl;

	pte = pte_offset_map_lock(mm, pmd, addr, &ptl);
	arch_enter_lazy_mmu_mode();
	do {
		oldpte = *pte;
		if (pte_present(oldpte)) {
			pte_t ptent;

			/* Avoid an SMP race with hardware updated dirty/clean
			 * bits by wiping the pte and then setting the new pte
			 * into place.
			 */
			ptent = ptep_get_and_clear(mm, addr, pte);
			ptent = pte_modify(ptent, newprot);
			/*
			 * Avoid taking write faults for pages we know to be
			 * dirty.
			 */
			if (dirty_accountable && pte_dirty(ptent))
				ptent = pte_mkwrite(ptent);
			set_pte_at(mm, addr, pte, ptent);
#ifdef CONFIG_MIGRATION
		} else if (!pte_file(oldpte)) {
			swp_entry_t entry = pte_to_swp_entry(oldpte);

			if (is_write_migration_entry(entry)) {
				/*
				 * A protection check is difficult so
				 * just be safe and disable write
				 */
				make_migration_entry_read(&entry);
				set_pte_at(mm, addr, pte,
					swp_entry_to_pte(entry));
			}
#endif
		}

	} while (pte++, addr += PAGE_SIZE, addr != end);
	arch_leave_lazy_mmu_mode();
	pte_unmap_unlock(pte - 1, ptl);
}

static inline void change_pmd_range(struct mm_struct *mm, pud_t *pud,
		unsigned long addr, unsigned long end, pgprot_t newprot,
		int dirty_accountable)
{
	pmd_t *pmd;
	unsigned long next;

	pmd = pmd_offset(pud, addr);
	do {
		next = pmd_addr_end(addr, end);
		if (pmd_none_or_clear_bad(pmd))
			continue;
		change_pte_range(mm, pmd, addr, next, newprot, dirty_accountable);
	} while (pmd++, addr = next, addr != end);
}

static inline void change_pud_range(struct mm_struct *mm, pgd_t *pgd,
		unsigned long addr, unsigned long end, pgprot_t newprot,
		int dirty_accountable)
{
	pud_t *pud;
	unsigned long next;

	pud = pud_offset(pgd, addr);
	do {
		next = pud_addr_end(addr, end);
		if (pud_none_or_clear_bad(pud))
			continue;
		change_pmd_range(mm, pud, addr, next, newprot, dirty_accountable);
	} while (pud++, addr = next, addr != end);
}

static void change_protection(struct vm_area_struct *vma,
		unsigned long addr, unsigned long end, pgprot_t newprot,
		int dirty_accountable)
{
	struct mm_struct *mm = vma->vm_mm;
	pgd_t *pgd;
	unsigned long next;
	unsigned long start = addr;

	BUG_ON(addr >= end);
	pgd = pgd_offset(mm, addr);
	flush_cache_range(vma, addr, end);
	do {
		next = pgd_addr_end(addr, end);
		if (pgd_none_or_clear_bad(pgd))
			continue;
		change_pud_range(mm, pgd, addr, next, newprot, dirty_accountable);
	} while (pgd++, addr = next, addr != end);
	flush_tlb_range(vma, start, end);
}

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.
	 *
	 * FIXME? We haven't defined a VM_NORESERVE flag, so mprotecting
	 * a MAP_NORESERVE private mapping to writable will now reserve.
	 */
	if (newflags & VM_WRITE) {
		if (!(oldflags & (VM_ACCOUNT|VM_WRITE|VM_SHARED))) {
			charged = nrpages;
			if (security_vm_enough_memory(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));
	if (*pprev) {
		vma = *pprev;
		goto success;
	}

	*pprev = vma;

	if (start != vma->vm_start) {
		error = split_vma(mm, vma, start, 1);
		if (error)
			goto fail;
	}

	if (end != vma->vm_end) {
		error = split_vma(mm, vma, end, 0);
		if (error)
			goto fail;
	}

success:
	/*
	 * vm_flags and vm_page_prot are protected by the mmap_sem
	 * held in write mode.
	 */
	vma->vm_flags = newflags;
	vma->vm_page_prot = vm_get_page_prot(newflags);
	if (vma_wants_writenotify(vma)) {
		vma->vm_page_prot = vm_get_page_prot(newflags & ~VM_SHARED);
		dirty_accountable = 1;
	}

	if (is_vm_hugetlb_page(vma))
		hugetlb_change_protection(vma, start, end, vma->vm_page_prot);
	else
		change_protection(vma, start, end, vma->vm_page_prot, dirty_accountable);
	vm_stat_account(mm, oldflags, vma->vm_file, -nrpages);
	vm_stat_account(mm, newflags, vma->vm_file, nrpages);
	return 0;

fail:
	vm_unacct_memory(charged);
	return error;
}

asmlinkage long
sys_mprotect(unsigned long start, size_t len, unsigned long prot)
{
	unsigned long vm_flags, nstart, end, tmp, reqprot;
	struct vm_area_struct *vma, *prev;
	int error = -EINVAL;
	const int grows = prot & (PROT_GROWSDOWN|PROT_GROWSUP);
	prot &= ~(PROT_GROWSDOWN|PROT_GROWSUP);
	if (grows == (PROT_GROWSDOWN|PROT_GROWSUP)) /* can't be both */
		return -EINVAL;

	if (start & ~PAGE_MASK)
		return -EINVAL;
	if (!len)
		return 0;
	len = PAGE_ALIGN(len);
	end = start + len;
	if (end <= start)
		return -ENOMEM;
	if (prot & ~(PROT_READ | PROT_WRITE | PROT_EXEC | PROT_SEM))
		return -EINVAL;

	reqprot = prot;
	/*
	 * Does the application expect PROT_READ to imply PROT_EXEC:
	 */
	if ((prot & PROT_READ) && (current->personality & READ_IMPLIES_EXEC))
		prot |= PROT_EXEC;

	vm_flags = calc_vm_prot_bits(prot);

	down_write(&current->mm->mmap_sem);

	vma = find_vma_prev(current->mm, start, &prev);
	error = -ENOMEM;
	if (!vma)
		goto out;
	if (unlikely(grows & PROT_GROWSDOWN)) {
		if (vma->vm_start >= end)
			goto out;
		start = vma->vm_start;
		error = -EINVAL;
		if (!(vma->vm_flags & VM_GROWSDOWN))
			goto out;
	}
	else {
		if (vma->vm_start > start)
			goto out;
		if (unlikely(grows & PROT_GROWSUP)) {
			end = vma->vm_end;
			error = -EINVAL;
			if (!(vma->vm_flags & VM_GROWSUP))
				goto out;
		}
	}
	if (start > vma->vm_start)
		prev = vma;

	for (nstart = start ; ; ) {
		unsigned long newflags;

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

		newflags = vm_flags | (vma->vm_flags & ~(VM_READ | VM_WRITE | VM_EXEC));

		/* newflags >> 4 shift VM_MAY% in place of VM_% */
		if ((newflags & ~(newflags >> 4)) & (VM_READ | VM_WRITE | VM_EXEC)) {
			error = -EACCES;
			goto out;
		}

		error = security_file_mprotect(vma, reqprot, prot);
		if (error)
			goto out;

		tmp = vma->vm_end;
		if (tmp > end)
			tmp = end;
		error = mprotect_fixup(vma, &prev, nstart, tmp, newflags);
		if (error)
			goto out;
		nstart = tmp;

		if (nstart < prev->vm_end)
			nstart = prev->vm_end;
		if (nstart >= end)
			goto out;

		vma = prev->vm_next;
		if (!vma || vma->vm_start != nstart) {
			error = -ENOMEM;
			goto out;
		}
	}
out:
	up_write(&current->mm->mmap_sem);
	return error;
}
Commit	Line	Data
1da177e4 LT	1	/*
	2	* mm/mprotect.c
	3	*
	4	* (C) Copyright 1994 Linus Torvalds
	5	* (C) Copyright 2002 Christoph Hellwig
	6	*
	7	* Address space accounting code <[email protected]>
	8	* (C) Copyright 2002 Red Hat Inc, All Rights Reserved
	9	*/
	10
	11	#include <linux/mm.h>
	12	#include <linux/hugetlb.h>
	13	#include <linux/slab.h>
	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>
1da177e4 LT	24	#include <asm/uaccess.h>
	25	#include <asm/pgtable.h>
	26	#include <asm/cacheflush.h>
	27	#include <asm/tlbflush.h>
	28
	29	static void change_pte_range(struct mm_struct mm, pmd_t pmd,
c1e6098b PZ	30	unsigned long addr, unsigned long end, pgprot_t newprot,
c1e6098b PZ	31	int dirty_accountable)
1da177e4	32	{
0697212a	33	pte_t *pte, oldpte;
705e87c0	34	spinlock_t *ptl;
1da177e4	35
705e87c0	36	pte = pte_offset_map_lock(mm, pmd, addr, &ptl);
6606c3e0	37	arch_enter_lazy_mmu_mode();
1da177e4	38	do {
0697212a CL	39	oldpte = *pte;
0697212a CL	40	if (pte_present(oldpte)) {
1da177e4 LT	41	pte_t ptent;
	42
	43	/* Avoid an SMP race with hardware updated dirty/clean
	44	* bits by wiping the pte and then setting the new pte
	45	* into place.
	46	*/
c1e6098b PZ	47	ptent = ptep_get_and_clear(mm, addr, pte);
	48	ptent = pte_modify(ptent, newprot);
	49	/*
	50	* Avoid taking write faults for pages we know to be
	51	* dirty.
	52	*/
	53	if (dirty_accountable && pte_dirty(ptent))
	54	ptent = pte_mkwrite(ptent);
1da177e4	55	set_pte_at(mm, addr, pte, ptent);
0697212a CL	56	#ifdef CONFIG_MIGRATION
	57	} else if (!pte_file(oldpte)) {
	58	swp_entry_t entry = pte_to_swp_entry(oldpte);
	59
	60	if (is_write_migration_entry(entry)) {
	61	/*
	62	* A protection check is difficult so
	63	* just be safe and disable write
	64	*/
	65	make_migration_entry_read(&entry);
	66	set_pte_at(mm, addr, pte,
	67	swp_entry_to_pte(entry));
	68	}
	69	#endif
1da177e4	70	}
0697212a	71
1da177e4	72	} while (pte++, addr += PAGE_SIZE, addr != end);
6606c3e0	73	arch_leave_lazy_mmu_mode();
705e87c0	74	pte_unmap_unlock(pte - 1, ptl);
1da177e4 LT	75	}
	76
	77	static inline void change_pmd_range(struct mm_struct mm, pud_t pud,
c1e6098b PZ	78	unsigned long addr, unsigned long end, pgprot_t newprot,
c1e6098b PZ	79	int dirty_accountable)
1da177e4 LT	80	{
	81	pmd_t *pmd;
	82	unsigned long next;
	83
	84	pmd = pmd_offset(pud, addr);
	85	do {
	86	next = pmd_addr_end(addr, end);
	87	if (pmd_none_or_clear_bad(pmd))
	88	continue;
c1e6098b	89	change_pte_range(mm, pmd, addr, next, newprot, dirty_accountable);
1da177e4 LT	90	} while (pmd++, addr = next, addr != end);
	91	}
	92
	93	static inline void change_pud_range(struct mm_struct mm, pgd_t pgd,
c1e6098b PZ	94	unsigned long addr, unsigned long end, pgprot_t newprot,
c1e6098b PZ	95	int dirty_accountable)
1da177e4 LT	96	{
	97	pud_t *pud;
	98	unsigned long next;
	99
	100	pud = pud_offset(pgd, addr);
	101	do {
	102	next = pud_addr_end(addr, end);
	103	if (pud_none_or_clear_bad(pud))
	104	continue;
c1e6098b	105	change_pmd_range(mm, pud, addr, next, newprot, dirty_accountable);
1da177e4 LT	106	} while (pud++, addr = next, addr != end);
	107	}
	108
	109	static void change_protection(struct vm_area_struct *vma,
c1e6098b PZ	110	unsigned long addr, unsigned long end, pgprot_t newprot,
c1e6098b PZ	111	int dirty_accountable)
1da177e4 LT	112	{
	113	struct mm_struct *mm = vma->vm_mm;
	114	pgd_t *pgd;
	115	unsigned long next;
	116	unsigned long start = addr;
	117
	118	BUG_ON(addr >= end);
	119	pgd = pgd_offset(mm, addr);
	120	flush_cache_range(vma, addr, end);
1da177e4 LT	121	do {
	122	next = pgd_addr_end(addr, end);
	123	if (pgd_none_or_clear_bad(pgd))
	124	continue;
c1e6098b	125	change_pud_range(mm, pgd, addr, next, newprot, dirty_accountable);
1da177e4 LT	126	} while (pgd++, addr = next, addr != end);
1da177e4 LT	127	flush_tlb_range(vma, start, end);
1da177e4 LT	128	}
1da177e4 LT	129
b6a2fea3	130	int
1da177e4 LT	131	mprotect_fixup(struct vm_area_struct vma, struct vm_area_struct *pprev,
	132	unsigned long start, unsigned long end, unsigned long newflags)
	133	{
	134	struct mm_struct *mm = vma->vm_mm;
	135	unsigned long oldflags = vma->vm_flags;
	136	long nrpages = (end - start) >> PAGE_SHIFT;
	137	unsigned long charged = 0;
1da177e4 LT	138	pgoff_t pgoff;
1da177e4 LT	139	int error;
c1e6098b	140	int dirty_accountable = 0;
1da177e4 LT	141
	142	if (newflags == oldflags) {
	143	*pprev = vma;
	144	return 0;
	145	}
	146
	147	/*
	148	* If we make a private mapping writable we increase our commit;
	149	* but (without finer accounting) cannot reduce our commit if we
	150	* make it unwritable again.
	151	*
	152	* FIXME? We haven't defined a VM_NORESERVE flag, so mprotecting
	153	* a MAP_NORESERVE private mapping to writable will now reserve.
	154	*/
	155	if (newflags & VM_WRITE) {
8f860591	156	if (!(oldflags & (VM_ACCOUNT\|VM_WRITE\|VM_SHARED))) {
1da177e4 LT	157	charged = nrpages;
	158	if (security_vm_enough_memory(charged))
	159	return -ENOMEM;
	160	newflags \|= VM_ACCOUNT;
	161	}
	162	}
	163
1da177e4 LT	164	/*
	165	* First try to merge with previous and/or next vma.
	166	*/
	167	pgoff = vma->vm_pgoff + ((start - vma->vm_start) >> PAGE_SHIFT);
	168	pprev = vma_merge(mm, pprev, start, end, newflags,
	169	vma->anon_vma, vma->vm_file, pgoff, vma_policy(vma));
	170	if (*pprev) {
	171	vma = *pprev;
	172	goto success;
	173	}
	174
	175	*pprev = vma;
	176
	177	if (start != vma->vm_start) {
	178	error = split_vma(mm, vma, start, 1);
	179	if (error)
	180	goto fail;
	181	}
	182
	183	if (end != vma->vm_end) {
	184	error = split_vma(mm, vma, end, 0);
	185	if (error)
	186	goto fail;
	187	}
	188
	189	success:
	190	/*
	191	* vm_flags and vm_page_prot are protected by the mmap_sem
	192	* held in write mode.
	193	*/
	194	vma->vm_flags = newflags;
3ed75eb8	195	vma->vm_page_prot = vm_get_page_prot(newflags);
c1e6098b	196	if (vma_wants_writenotify(vma)) {
1ddd439e	197	vma->vm_page_prot = vm_get_page_prot(newflags & ~VM_SHARED);
c1e6098b PZ	198	dirty_accountable = 1;
c1e6098b PZ	199	}
d08b3851	200
8f860591	201	if (is_vm_hugetlb_page(vma))
d08b3851	202	hugetlb_change_protection(vma, start, end, vma->vm_page_prot);
8f860591	203	else
c1e6098b	204	change_protection(vma, start, end, vma->vm_page_prot, dirty_accountable);
ab50b8ed HD	205	vm_stat_account(mm, oldflags, vma->vm_file, -nrpages);
ab50b8ed HD	206	vm_stat_account(mm, newflags, vma->vm_file, nrpages);
1da177e4 LT	207	return 0;
	208
	209	fail:
	210	vm_unacct_memory(charged);
	211	return error;
	212	}
	213
	214	asmlinkage long
	215	sys_mprotect(unsigned long start, size_t len, unsigned long prot)
	216	{
	217	unsigned long vm_flags, nstart, end, tmp, reqprot;
	218	struct vm_area_struct vma, prev;
	219	int error = -EINVAL;
	220	const int grows = prot & (PROT_GROWSDOWN\|PROT_GROWSUP);
	221	prot &= ~(PROT_GROWSDOWN\|PROT_GROWSUP);
	222	if (grows == (PROT_GROWSDOWN\|PROT_GROWSUP)) /* can't be both */
	223	return -EINVAL;
	224
	225	if (start & ~PAGE_MASK)
	226	return -EINVAL;
	227	if (!len)
	228	return 0;
	229	len = PAGE_ALIGN(len);
	230	end = start + len;
	231	if (end <= start)
	232	return -ENOMEM;
	233	if (prot & ~(PROT_READ \| PROT_WRITE \| PROT_EXEC \| PROT_SEM))
	234	return -EINVAL;
	235
	236	reqprot = prot;
	237	/*
	238	* Does the application expect PROT_READ to imply PROT_EXEC:
	239	*/
b344e05c	240	if ((prot & PROT_READ) && (current->personality & READ_IMPLIES_EXEC))
1da177e4 LT	241	prot \|= PROT_EXEC;
	242
	243	vm_flags = calc_vm_prot_bits(prot);
	244
	245	down_write(&current->mm->mmap_sem);
	246
	247	vma = find_vma_prev(current->mm, start, &prev);
	248	error = -ENOMEM;
	249	if (!vma)
	250	goto out;
	251	if (unlikely(grows & PROT_GROWSDOWN)) {
	252	if (vma->vm_start >= end)
	253	goto out;
	254	start = vma->vm_start;
	255	error = -EINVAL;
	256	if (!(vma->vm_flags & VM_GROWSDOWN))
	257	goto out;
	258	}
	259	else {
	260	if (vma->vm_start > start)
	261	goto out;
	262	if (unlikely(grows & PROT_GROWSUP)) {
	263	end = vma->vm_end;
	264	error = -EINVAL;
	265	if (!(vma->vm_flags & VM_GROWSUP))
	266	goto out;
	267	}
	268	}
	269	if (start > vma->vm_start)
	270	prev = vma;
	271
	272	for (nstart = start ; ; ) {
	273	unsigned long newflags;
	274
	275	/* Here we know that vma->vm_start <= nstart < vma->vm_end. */
	276
1da177e4 LT	277	newflags = vm_flags \| (vma->vm_flags & ~(VM_READ \| VM_WRITE \| VM_EXEC));
1da177e4 LT	278
7e2cff42 PBG	279	/* newflags >> 4 shift VM_MAY% in place of VM_% */
7e2cff42 PBG	280	if ((newflags & ~(newflags >> 4)) & (VM_READ \| VM_WRITE \| VM_EXEC)) {
1da177e4 LT	281	error = -EACCES;
	282	goto out;
	283	}
	284
	285	error = security_file_mprotect(vma, reqprot, prot);
	286	if (error)
	287	goto out;
	288
	289	tmp = vma->vm_end;
	290	if (tmp > end)
	291	tmp = end;
	292	error = mprotect_fixup(vma, &prev, nstart, tmp, newflags);
	293	if (error)
	294	goto out;
	295	nstart = tmp;
	296
	297	if (nstart < prev->vm_end)
	298	nstart = prev->vm_end;
	299	if (nstart >= end)
	300	goto out;
	301
	302	vma = prev->vm_next;
	303	if (!vma \|\| vma->vm_start != nstart) {
	304	error = -ENOMEM;
	305	goto out;
	306	}
	307	}
	308	out:
	309	up_write(&current->mm->mmap_sem);
	310	return error;
	311	}