[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>

#ifndef pgprot_modify
static inline pgprot_t pgprot_modify(pgprot_t oldprot, pgprot_t newprot)
{
	return newprot;
}
#endif

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;

			ptent = ptep_modify_prot_start(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);

			ptep_modify_prot_commit(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 = pgprot_modify(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 (!arch_validate_prot(prot))
		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
1c12c4cf VP	29	#ifndef pgprot_modify
	30	static inline pgprot_t pgprot_modify(pgprot_t oldprot, pgprot_t newprot)
	31	{
	32	return newprot;
	33	}
	34	#endif
	35
1da177e4	36	static void change_pte_range(struct mm_struct mm, pmd_t pmd,
c1e6098b PZ	37	unsigned long addr, unsigned long end, pgprot_t newprot,
c1e6098b PZ	38	int dirty_accountable)
1da177e4	39	{
0697212a	40	pte_t *pte, oldpte;
705e87c0	41	spinlock_t *ptl;
1da177e4	42
705e87c0	43	pte = pte_offset_map_lock(mm, pmd, addr, &ptl);
6606c3e0	44	arch_enter_lazy_mmu_mode();
1da177e4	45	do {
0697212a CL	46	oldpte = *pte;
0697212a CL	47	if (pte_present(oldpte)) {
1da177e4 LT	48	pte_t ptent;
1da177e4 LT	49
1ea0704e	50	ptent = ptep_modify_prot_start(mm, addr, pte);
c1e6098b	51	ptent = pte_modify(ptent, newprot);
1ea0704e	52
c1e6098b PZ	53	/*
	54	* Avoid taking write faults for pages we know to be
	55	* dirty.
	56	*/
	57	if (dirty_accountable && pte_dirty(ptent))
	58	ptent = pte_mkwrite(ptent);
1ea0704e JF	59
1ea0704e JF	60	ptep_modify_prot_commit(mm, addr, pte, ptent);
0697212a CL	61	#ifdef CONFIG_MIGRATION
	62	} else if (!pte_file(oldpte)) {
	63	swp_entry_t entry = pte_to_swp_entry(oldpte);
	64
	65	if (is_write_migration_entry(entry)) {
	66	/*
	67	* A protection check is difficult so
	68	* just be safe and disable write
	69	*/
	70	make_migration_entry_read(&entry);
	71	set_pte_at(mm, addr, pte,
	72	swp_entry_to_pte(entry));
	73	}
	74	#endif
1da177e4	75	}
0697212a	76
1da177e4	77	} while (pte++, addr += PAGE_SIZE, addr != end);
6606c3e0	78	arch_leave_lazy_mmu_mode();
705e87c0	79	pte_unmap_unlock(pte - 1, ptl);
1da177e4 LT	80	}
	81
	82	static inline void change_pmd_range(struct mm_struct mm, pud_t pud,
c1e6098b PZ	83	unsigned long addr, unsigned long end, pgprot_t newprot,
c1e6098b PZ	84	int dirty_accountable)
1da177e4 LT	85	{
	86	pmd_t *pmd;
	87	unsigned long next;
	88
	89	pmd = pmd_offset(pud, addr);
	90	do {
	91	next = pmd_addr_end(addr, end);
	92	if (pmd_none_or_clear_bad(pmd))
	93	continue;
c1e6098b	94	change_pte_range(mm, pmd, addr, next, newprot, dirty_accountable);
1da177e4 LT	95	} while (pmd++, addr = next, addr != end);
	96	}
	97
	98	static inline void change_pud_range(struct mm_struct mm, pgd_t pgd,
c1e6098b PZ	99	unsigned long addr, unsigned long end, pgprot_t newprot,
c1e6098b PZ	100	int dirty_accountable)
1da177e4 LT	101	{
	102	pud_t *pud;
	103	unsigned long next;
	104
	105	pud = pud_offset(pgd, addr);
	106	do {
	107	next = pud_addr_end(addr, end);
	108	if (pud_none_or_clear_bad(pud))
	109	continue;
c1e6098b	110	change_pmd_range(mm, pud, addr, next, newprot, dirty_accountable);
1da177e4 LT	111	} while (pud++, addr = next, addr != end);
	112	}
	113
	114	static void change_protection(struct vm_area_struct *vma,
c1e6098b PZ	115	unsigned long addr, unsigned long end, pgprot_t newprot,
c1e6098b PZ	116	int dirty_accountable)
1da177e4 LT	117	{
	118	struct mm_struct *mm = vma->vm_mm;
	119	pgd_t *pgd;
	120	unsigned long next;
	121	unsigned long start = addr;
	122
	123	BUG_ON(addr >= end);
	124	pgd = pgd_offset(mm, addr);
	125	flush_cache_range(vma, addr, end);
1da177e4 LT	126	do {
	127	next = pgd_addr_end(addr, end);
	128	if (pgd_none_or_clear_bad(pgd))
	129	continue;
c1e6098b	130	change_pud_range(mm, pgd, addr, next, newprot, dirty_accountable);
1da177e4 LT	131	} while (pgd++, addr = next, addr != end);
1da177e4 LT	132	flush_tlb_range(vma, start, end);
1da177e4 LT	133	}
1da177e4 LT	134
b6a2fea3	135	int
1da177e4 LT	136	mprotect_fixup(struct vm_area_struct vma, struct vm_area_struct *pprev,
	137	unsigned long start, unsigned long end, unsigned long newflags)
	138	{
	139	struct mm_struct *mm = vma->vm_mm;
	140	unsigned long oldflags = vma->vm_flags;
	141	long nrpages = (end - start) >> PAGE_SHIFT;
	142	unsigned long charged = 0;
1da177e4 LT	143	pgoff_t pgoff;
1da177e4 LT	144	int error;
c1e6098b	145	int dirty_accountable = 0;
1da177e4 LT	146
	147	if (newflags == oldflags) {
	148	*pprev = vma;
	149	return 0;
	150	}
	151
	152	/*
	153	* If we make a private mapping writable we increase our commit;
	154	* but (without finer accounting) cannot reduce our commit if we
	155	* make it unwritable again.
	156	*
	157	* FIXME? We haven't defined a VM_NORESERVE flag, so mprotecting
	158	* a MAP_NORESERVE private mapping to writable will now reserve.
	159	*/
	160	if (newflags & VM_WRITE) {
8f860591	161	if (!(oldflags & (VM_ACCOUNT\|VM_WRITE\|VM_SHARED))) {
1da177e4 LT	162	charged = nrpages;
	163	if (security_vm_enough_memory(charged))
	164	return -ENOMEM;
	165	newflags \|= VM_ACCOUNT;
	166	}
	167	}
	168
1da177e4 LT	169	/*
	170	* First try to merge with previous and/or next vma.
	171	*/
	172	pgoff = vma->vm_pgoff + ((start - vma->vm_start) >> PAGE_SHIFT);
	173	pprev = vma_merge(mm, pprev, start, end, newflags,
	174	vma->anon_vma, vma->vm_file, pgoff, vma_policy(vma));
	175	if (*pprev) {
	176	vma = *pprev;
	177	goto success;
	178	}
	179
	180	*pprev = vma;
	181
	182	if (start != vma->vm_start) {
	183	error = split_vma(mm, vma, start, 1);
	184	if (error)
	185	goto fail;
	186	}
	187
	188	if (end != vma->vm_end) {
	189	error = split_vma(mm, vma, end, 0);
	190	if (error)
	191	goto fail;
	192	}
	193
	194	success:
	195	/*
	196	* vm_flags and vm_page_prot are protected by the mmap_sem
	197	* held in write mode.
	198	*/
	199	vma->vm_flags = newflags;
1c12c4cf VP	200	vma->vm_page_prot = pgprot_modify(vma->vm_page_prot,
	201	vm_get_page_prot(newflags));
	202
c1e6098b	203	if (vma_wants_writenotify(vma)) {
1ddd439e	204	vma->vm_page_prot = vm_get_page_prot(newflags & ~VM_SHARED);
c1e6098b PZ	205	dirty_accountable = 1;
c1e6098b PZ	206	}
d08b3851	207
8f860591	208	if (is_vm_hugetlb_page(vma))
d08b3851	209	hugetlb_change_protection(vma, start, end, vma->vm_page_prot);
8f860591	210	else
c1e6098b	211	change_protection(vma, start, end, vma->vm_page_prot, dirty_accountable);
ab50b8ed HD	212	vm_stat_account(mm, oldflags, vma->vm_file, -nrpages);
ab50b8ed HD	213	vm_stat_account(mm, newflags, vma->vm_file, nrpages);
1da177e4 LT	214	return 0;
	215
	216	fail:
	217	vm_unacct_memory(charged);
	218	return error;
	219	}
	220
	221	asmlinkage long
	222	sys_mprotect(unsigned long start, size_t len, unsigned long prot)
	223	{
	224	unsigned long vm_flags, nstart, end, tmp, reqprot;
	225	struct vm_area_struct vma, prev;
	226	int error = -EINVAL;
	227	const int grows = prot & (PROT_GROWSDOWN\|PROT_GROWSUP);
	228	prot &= ~(PROT_GROWSDOWN\|PROT_GROWSUP);
	229	if (grows == (PROT_GROWSDOWN\|PROT_GROWSUP)) /* can't be both */
	230	return -EINVAL;
	231
	232	if (start & ~PAGE_MASK)
	233	return -EINVAL;
	234	if (!len)
	235	return 0;
	236	len = PAGE_ALIGN(len);
	237	end = start + len;
	238	if (end <= start)
	239	return -ENOMEM;
b845f313	240	if (!arch_validate_prot(prot))
1da177e4 LT	241	return -EINVAL;
	242
	243	reqprot = prot;
	244	/*
	245	* Does the application expect PROT_READ to imply PROT_EXEC:
	246	*/
b344e05c	247	if ((prot & PROT_READ) && (current->personality & READ_IMPLIES_EXEC))
1da177e4 LT	248	prot \|= PROT_EXEC;
	249
	250	vm_flags = calc_vm_prot_bits(prot);
	251
	252	down_write(&current->mm->mmap_sem);
	253
	254	vma = find_vma_prev(current->mm, start, &prev);
	255	error = -ENOMEM;
	256	if (!vma)
	257	goto out;
	258	if (unlikely(grows & PROT_GROWSDOWN)) {
	259	if (vma->vm_start >= end)
	260	goto out;
	261	start = vma->vm_start;
	262	error = -EINVAL;
	263	if (!(vma->vm_flags & VM_GROWSDOWN))
	264	goto out;
	265	}
	266	else {
	267	if (vma->vm_start > start)
	268	goto out;
	269	if (unlikely(grows & PROT_GROWSUP)) {
	270	end = vma->vm_end;
	271	error = -EINVAL;
	272	if (!(vma->vm_flags & VM_GROWSUP))
	273	goto out;
	274	}
	275	}
	276	if (start > vma->vm_start)
	277	prev = vma;
	278
	279	for (nstart = start ; ; ) {
	280	unsigned long newflags;
	281
	282	/* Here we know that vma->vm_start <= nstart < vma->vm_end. */
	283
1da177e4 LT	284	newflags = vm_flags \| (vma->vm_flags & ~(VM_READ \| VM_WRITE \| VM_EXEC));
1da177e4 LT	285
7e2cff42 PBG	286	/* newflags >> 4 shift VM_MAY% in place of VM_% */
7e2cff42 PBG	287	if ((newflags & ~(newflags >> 4)) & (VM_READ \| VM_WRITE \| VM_EXEC)) {
1da177e4 LT	288	error = -EACCES;
	289	goto out;
	290	}
	291
	292	error = security_file_mprotect(vma, reqprot, prot);
	293	if (error)
	294	goto out;
	295
	296	tmp = vma->vm_end;
	297	if (tmp > end)
	298	tmp = end;
	299	error = mprotect_fixup(vma, &prev, nstart, tmp, newflags);
	300	if (error)
	301	goto out;
	302	nstart = tmp;
	303
	304	if (nstart < prev->vm_end)
	305	nstart = prev->vm_end;
	306	if (nstart >= end)
	307	goto out;
	308
	309	vma = prev->vm_next;
	310	if (!vma \|\| vma->vm_start != nstart) {
	311	error = -ENOMEM;
	312	goto out;
	313	}
	314	}
	315	out:
	316	up_write(&current->mm->mmap_sem);
	317	return error;
	318	}