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

/*
 *  mm/mprotect.c
 *
 *  (C) Copyright 1994 Linus Torvalds
 *  (C) Copyright 2002 Christoph Hellwig
 *
 *  Address space accounting code	<alan@redhat.com>
 *  (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)
{
	pte_t *pte, oldpte;
	spinlock_t *ptl;

	pte = pte_offset_map_lock(mm, pmd, addr, &ptl);
	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 = pte_modify(ptep_get_and_clear(mm, addr, pte), newprot);
			set_pte_at(mm, addr, pte, ptent);
			lazy_mmu_prot_update(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);
	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)
{
	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);
	} 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)
{
	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);
	} while (pud++, addr = next, addr != end);
}

static void change_protection(struct vm_area_struct *vma,
		unsigned long addr, unsigned long end, pgprot_t newprot)
{
	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);
	} while (pgd++, addr = next, addr != end);
	flush_tlb_range(vma, start, end);
}

static 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;
	unsigned int mask;
	pgprot_t newprot;
	pgoff_t pgoff;
	int error;

	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:
	/* Don't make the VMA automatically writable if it's shared, but the
	 * backer wishes to know when pages are first written to */
	mask = VM_READ|VM_WRITE|VM_EXEC|VM_SHARED;
	if (vma->vm_ops && vma->vm_ops->page_mkwrite)
		mask &= ~VM_SHARED;

	newprot = protection_map[newflags & mask];

	/*
	 * vm_flags and vm_page_prot are protected by the mmap_sem
	 * held in write mode.
	 */
	vma->vm_flags = newflags;
	vma->vm_page_prot = newprot;
	if (is_vm_hugetlb_page(vma))
		hugetlb_change_protection(vma, start, end, newprot);
	else
		change_protection(vma, start, end, newprot);
	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 <alan@redhat.com>
	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,
	30	unsigned long addr, unsigned long end, pgprot_t newprot)
	31	{
0697212a	32	pte_t *pte, oldpte;
705e87c0	33	spinlock_t *ptl;
1da177e4	34
705e87c0	35	pte = pte_offset_map_lock(mm, pmd, addr, &ptl);
1da177e4	36	do {
0697212a CL	37	oldpte = *pte;
0697212a CL	38	if (pte_present(oldpte)) {
1da177e4 LT	39	pte_t ptent;
	40
	41	/* Avoid an SMP race with hardware updated dirty/clean
	42	* bits by wiping the pte and then setting the new pte
	43	* into place.
	44	*/
	45	ptent = pte_modify(ptep_get_and_clear(mm, addr, pte), newprot);
	46	set_pte_at(mm, addr, pte, ptent);
	47	lazy_mmu_prot_update(ptent);
0697212a CL	48	#ifdef CONFIG_MIGRATION
	49	} else if (!pte_file(oldpte)) {
	50	swp_entry_t entry = pte_to_swp_entry(oldpte);
	51
	52	if (is_write_migration_entry(entry)) {
	53	/*
	54	* A protection check is difficult so
	55	* just be safe and disable write
	56	*/
	57	make_migration_entry_read(&entry);
	58	set_pte_at(mm, addr, pte,
	59	swp_entry_to_pte(entry));
	60	}
	61	#endif
1da177e4	62	}
0697212a	63
1da177e4	64	} while (pte++, addr += PAGE_SIZE, addr != end);
705e87c0	65	pte_unmap_unlock(pte - 1, ptl);
1da177e4 LT	66	}
	67
	68	static inline void change_pmd_range(struct mm_struct mm, pud_t pud,
	69	unsigned long addr, unsigned long end, pgprot_t newprot)
	70	{
	71	pmd_t *pmd;
	72	unsigned long next;
	73
	74	pmd = pmd_offset(pud, addr);
	75	do {
	76	next = pmd_addr_end(addr, end);
	77	if (pmd_none_or_clear_bad(pmd))
	78	continue;
	79	change_pte_range(mm, pmd, addr, next, newprot);
	80	} while (pmd++, addr = next, addr != end);
	81	}
	82
	83	static inline void change_pud_range(struct mm_struct mm, pgd_t pgd,
	84	unsigned long addr, unsigned long end, pgprot_t newprot)
	85	{
	86	pud_t *pud;
	87	unsigned long next;
	88
	89	pud = pud_offset(pgd, addr);
	90	do {
	91	next = pud_addr_end(addr, end);
	92	if (pud_none_or_clear_bad(pud))
	93	continue;
	94	change_pmd_range(mm, pud, addr, next, newprot);
	95	} while (pud++, addr = next, addr != end);
	96	}
	97
	98	static void change_protection(struct vm_area_struct *vma,
	99	unsigned long addr, unsigned long end, pgprot_t newprot)
	100	{
	101	struct mm_struct *mm = vma->vm_mm;
	102	pgd_t *pgd;
	103	unsigned long next;
	104	unsigned long start = addr;
	105
	106	BUG_ON(addr >= end);
	107	pgd = pgd_offset(mm, addr);
	108	flush_cache_range(vma, addr, end);
1da177e4 LT	109	do {
	110	next = pgd_addr_end(addr, end);
	111	if (pgd_none_or_clear_bad(pgd))
	112	continue;
	113	change_pud_range(mm, pgd, addr, next, newprot);
	114	} while (pgd++, addr = next, addr != end);
	115	flush_tlb_range(vma, start, end);
1da177e4 LT	116	}
	117
	118	static int
	119	mprotect_fixup(struct vm_area_struct vma, struct vm_area_struct *pprev,
	120	unsigned long start, unsigned long end, unsigned long newflags)
	121	{
	122	struct mm_struct *mm = vma->vm_mm;
	123	unsigned long oldflags = vma->vm_flags;
	124	long nrpages = (end - start) >> PAGE_SHIFT;
	125	unsigned long charged = 0;
9637a5ef	126	unsigned int mask;
1da177e4 LT	127	pgprot_t newprot;
	128	pgoff_t pgoff;
	129	int error;
	130
	131	if (newflags == oldflags) {
	132	*pprev = vma;
	133	return 0;
	134	}
	135
	136	/*
	137	* If we make a private mapping writable we increase our commit;
	138	* but (without finer accounting) cannot reduce our commit if we
	139	* make it unwritable again.
	140	*
	141	* FIXME? We haven't defined a VM_NORESERVE flag, so mprotecting
	142	* a MAP_NORESERVE private mapping to writable will now reserve.
	143	*/
	144	if (newflags & VM_WRITE) {
8f860591	145	if (!(oldflags & (VM_ACCOUNT\|VM_WRITE\|VM_SHARED))) {
1da177e4 LT	146	charged = nrpages;
	147	if (security_vm_enough_memory(charged))
	148	return -ENOMEM;
	149	newflags \|= VM_ACCOUNT;
	150	}
	151	}
	152
1da177e4 LT	153	/*
	154	* First try to merge with previous and/or next vma.
	155	*/
	156	pgoff = vma->vm_pgoff + ((start - vma->vm_start) >> PAGE_SHIFT);
	157	pprev = vma_merge(mm, pprev, start, end, newflags,
	158	vma->anon_vma, vma->vm_file, pgoff, vma_policy(vma));
	159	if (*pprev) {
	160	vma = *pprev;
	161	goto success;
	162	}
	163
	164	*pprev = vma;
	165
	166	if (start != vma->vm_start) {
	167	error = split_vma(mm, vma, start, 1);
	168	if (error)
	169	goto fail;
	170	}
	171
	172	if (end != vma->vm_end) {
	173	error = split_vma(mm, vma, end, 0);
	174	if (error)
	175	goto fail;
	176	}
	177
	178	success:
9637a5ef DH	179	/* Don't make the VMA automatically writable if it's shared, but the
	180	* backer wishes to know when pages are first written to */
	181	mask = VM_READ\|VM_WRITE\|VM_EXEC\|VM_SHARED;
	182	if (vma->vm_ops && vma->vm_ops->page_mkwrite)
	183	mask &= ~VM_SHARED;
	184
	185	newprot = protection_map[newflags & mask];
	186
1da177e4 LT	187	/*
	188	* vm_flags and vm_page_prot are protected by the mmap_sem
	189	* held in write mode.
	190	*/
	191	vma->vm_flags = newflags;
	192	vma->vm_page_prot = newprot;
8f860591 ZY	193	if (is_vm_hugetlb_page(vma))
	194	hugetlb_change_protection(vma, start, end, newprot);
	195	else
	196	change_protection(vma, start, end, newprot);
ab50b8ed HD	197	vm_stat_account(mm, oldflags, vma->vm_file, -nrpages);
ab50b8ed HD	198	vm_stat_account(mm, newflags, vma->vm_file, nrpages);
1da177e4 LT	199	return 0;
	200
	201	fail:
	202	vm_unacct_memory(charged);
	203	return error;
	204	}
	205
	206	asmlinkage long
	207	sys_mprotect(unsigned long start, size_t len, unsigned long prot)
	208	{
	209	unsigned long vm_flags, nstart, end, tmp, reqprot;
	210	struct vm_area_struct vma, prev;
	211	int error = -EINVAL;
	212	const int grows = prot & (PROT_GROWSDOWN\|PROT_GROWSUP);
	213	prot &= ~(PROT_GROWSDOWN\|PROT_GROWSUP);
	214	if (grows == (PROT_GROWSDOWN\|PROT_GROWSUP)) /* can't be both */
	215	return -EINVAL;
	216
	217	if (start & ~PAGE_MASK)
	218	return -EINVAL;
	219	if (!len)
	220	return 0;
	221	len = PAGE_ALIGN(len);
	222	end = start + len;
	223	if (end <= start)
	224	return -ENOMEM;
	225	if (prot & ~(PROT_READ \| PROT_WRITE \| PROT_EXEC \| PROT_SEM))
	226	return -EINVAL;
	227
	228	reqprot = prot;
	229	/*
	230	* Does the application expect PROT_READ to imply PROT_EXEC:
	231	*/
b344e05c	232	if ((prot & PROT_READ) && (current->personality & READ_IMPLIES_EXEC))
1da177e4 LT	233	prot \|= PROT_EXEC;
	234
	235	vm_flags = calc_vm_prot_bits(prot);
	236
	237	down_write(&current->mm->mmap_sem);
	238
	239	vma = find_vma_prev(current->mm, start, &prev);
	240	error = -ENOMEM;
	241	if (!vma)
	242	goto out;
	243	if (unlikely(grows & PROT_GROWSDOWN)) {
	244	if (vma->vm_start >= end)
	245	goto out;
	246	start = vma->vm_start;
	247	error = -EINVAL;
	248	if (!(vma->vm_flags & VM_GROWSDOWN))
	249	goto out;
	250	}
	251	else {
	252	if (vma->vm_start > start)
	253	goto out;
	254	if (unlikely(grows & PROT_GROWSUP)) {
	255	end = vma->vm_end;
	256	error = -EINVAL;
	257	if (!(vma->vm_flags & VM_GROWSUP))
	258	goto out;
	259	}
	260	}
	261	if (start > vma->vm_start)
	262	prev = vma;
	263
	264	for (nstart = start ; ; ) {
	265	unsigned long newflags;
	266
	267	/* Here we know that vma->vm_start <= nstart < vma->vm_end. */
	268
1da177e4 LT	269	newflags = vm_flags \| (vma->vm_flags & ~(VM_READ \| VM_WRITE \| VM_EXEC));
1da177e4 LT	270
7e2cff42 PBG	271	/* newflags >> 4 shift VM_MAY% in place of VM_% */
7e2cff42 PBG	272	if ((newflags & ~(newflags >> 4)) & (VM_READ \| VM_WRITE \| VM_EXEC)) {
1da177e4 LT	273	error = -EACCES;
	274	goto out;
	275	}
	276
	277	error = security_file_mprotect(vma, reqprot, prot);
	278	if (error)
	279	goto out;
	280
	281	tmp = vma->vm_end;
	282	if (tmp > end)
	283	tmp = end;
	284	error = mprotect_fixup(vma, &prev, nstart, tmp, newflags);
	285	if (error)
	286	goto out;
	287	nstart = tmp;
	288
	289	if (nstart < prev->vm_end)
	290	nstart = prev->vm_end;
	291	if (nstart >= end)
	292	goto out;
	293
	294	vma = prev->vm_next;
	295	if (!vma \|\| vma->vm_start != nstart) {
	296	error = -ENOMEM;
	297	goto out;
	298	}
	299	}
	300	out:
	301	up_write(&current->mm->mmap_sem);
	302	return error;
	303	}