[net-next-2.6.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;

			/* 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 = 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 (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
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;
	49
	50	/* Avoid an SMP race with hardware updated dirty/clean
	51	* bits by wiping the pte and then setting the new pte
	52	* into place.
	53	*/
c1e6098b PZ	54	ptent = ptep_get_and_clear(mm, addr, pte);
	55	ptent = pte_modify(ptent, newprot);
	56	/*
	57	* Avoid taking write faults for pages we know to be
	58	* dirty.
	59	*/
	60	if (dirty_accountable && pte_dirty(ptent))
	61	ptent = pte_mkwrite(ptent);
1da177e4	62	set_pte_at(mm, addr, pte, ptent);
0697212a CL	63	#ifdef CONFIG_MIGRATION
	64	} else if (!pte_file(oldpte)) {
	65	swp_entry_t entry = pte_to_swp_entry(oldpte);
	66
	67	if (is_write_migration_entry(entry)) {
	68	/*
	69	* A protection check is difficult so
	70	* just be safe and disable write
	71	*/
	72	make_migration_entry_read(&entry);
	73	set_pte_at(mm, addr, pte,
	74	swp_entry_to_pte(entry));
	75	}
	76	#endif
1da177e4	77	}
0697212a	78
1da177e4	79	} while (pte++, addr += PAGE_SIZE, addr != end);
6606c3e0	80	arch_leave_lazy_mmu_mode();
705e87c0	81	pte_unmap_unlock(pte - 1, ptl);
1da177e4 LT	82	}
	83
	84	static inline void change_pmd_range(struct mm_struct mm, pud_t pud,
c1e6098b PZ	85	unsigned long addr, unsigned long end, pgprot_t newprot,
c1e6098b PZ	86	int dirty_accountable)
1da177e4 LT	87	{
	88	pmd_t *pmd;
	89	unsigned long next;
	90
	91	pmd = pmd_offset(pud, addr);
	92	do {
	93	next = pmd_addr_end(addr, end);
	94	if (pmd_none_or_clear_bad(pmd))
	95	continue;
c1e6098b	96	change_pte_range(mm, pmd, addr, next, newprot, dirty_accountable);
1da177e4 LT	97	} while (pmd++, addr = next, addr != end);
	98	}
	99
	100	static inline void change_pud_range(struct mm_struct mm, pgd_t pgd,
c1e6098b PZ	101	unsigned long addr, unsigned long end, pgprot_t newprot,
c1e6098b PZ	102	int dirty_accountable)
1da177e4 LT	103	{
	104	pud_t *pud;
	105	unsigned long next;
	106
	107	pud = pud_offset(pgd, addr);
	108	do {
	109	next = pud_addr_end(addr, end);
	110	if (pud_none_or_clear_bad(pud))
	111	continue;
c1e6098b	112	change_pmd_range(mm, pud, addr, next, newprot, dirty_accountable);
1da177e4 LT	113	} while (pud++, addr = next, addr != end);
	114	}
	115
	116	static void change_protection(struct vm_area_struct *vma,
c1e6098b PZ	117	unsigned long addr, unsigned long end, pgprot_t newprot,
c1e6098b PZ	118	int dirty_accountable)
1da177e4 LT	119	{
	120	struct mm_struct *mm = vma->vm_mm;
	121	pgd_t *pgd;
	122	unsigned long next;
	123	unsigned long start = addr;
	124
	125	BUG_ON(addr >= end);
	126	pgd = pgd_offset(mm, addr);
	127	flush_cache_range(vma, addr, end);
1da177e4 LT	128	do {
	129	next = pgd_addr_end(addr, end);
	130	if (pgd_none_or_clear_bad(pgd))
	131	continue;
c1e6098b	132	change_pud_range(mm, pgd, addr, next, newprot, dirty_accountable);
1da177e4 LT	133	} while (pgd++, addr = next, addr != end);
1da177e4 LT	134	flush_tlb_range(vma, start, end);
1da177e4 LT	135	}
1da177e4 LT	136
b6a2fea3	137	int
1da177e4 LT	138	mprotect_fixup(struct vm_area_struct vma, struct vm_area_struct *pprev,
	139	unsigned long start, unsigned long end, unsigned long newflags)
	140	{
	141	struct mm_struct *mm = vma->vm_mm;
	142	unsigned long oldflags = vma->vm_flags;
	143	long nrpages = (end - start) >> PAGE_SHIFT;
	144	unsigned long charged = 0;
1da177e4 LT	145	pgoff_t pgoff;
1da177e4 LT	146	int error;
c1e6098b	147	int dirty_accountable = 0;
1da177e4 LT	148
	149	if (newflags == oldflags) {
	150	*pprev = vma;
	151	return 0;
	152	}
	153
	154	/*
	155	* If we make a private mapping writable we increase our commit;
	156	* but (without finer accounting) cannot reduce our commit if we
	157	* make it unwritable again.
	158	*
	159	* FIXME? We haven't defined a VM_NORESERVE flag, so mprotecting
	160	* a MAP_NORESERVE private mapping to writable will now reserve.
	161	*/
	162	if (newflags & VM_WRITE) {
8f860591	163	if (!(oldflags & (VM_ACCOUNT\|VM_WRITE\|VM_SHARED))) {
1da177e4 LT	164	charged = nrpages;
	165	if (security_vm_enough_memory(charged))
	166	return -ENOMEM;
	167	newflags \|= VM_ACCOUNT;
	168	}
	169	}
	170
1da177e4 LT	171	/*
	172	* First try to merge with previous and/or next vma.
	173	*/
	174	pgoff = vma->vm_pgoff + ((start - vma->vm_start) >> PAGE_SHIFT);
	175	pprev = vma_merge(mm, pprev, start, end, newflags,
	176	vma->anon_vma, vma->vm_file, pgoff, vma_policy(vma));
	177	if (*pprev) {
	178	vma = *pprev;
	179	goto success;
	180	}
	181
	182	*pprev = vma;
	183
	184	if (start != vma->vm_start) {
	185	error = split_vma(mm, vma, start, 1);
	186	if (error)
	187	goto fail;
	188	}
	189
	190	if (end != vma->vm_end) {
	191	error = split_vma(mm, vma, end, 0);
	192	if (error)
	193	goto fail;
	194	}
	195
	196	success:
	197	/*
	198	* vm_flags and vm_page_prot are protected by the mmap_sem
	199	* held in write mode.
	200	*/
	201	vma->vm_flags = newflags;
1c12c4cf VP	202	vma->vm_page_prot = pgprot_modify(vma->vm_page_prot,
	203	vm_get_page_prot(newflags));
	204
c1e6098b	205	if (vma_wants_writenotify(vma)) {
1ddd439e	206	vma->vm_page_prot = vm_get_page_prot(newflags & ~VM_SHARED);
c1e6098b PZ	207	dirty_accountable = 1;
c1e6098b PZ	208	}
d08b3851	209
8f860591	210	if (is_vm_hugetlb_page(vma))
d08b3851	211	hugetlb_change_protection(vma, start, end, vma->vm_page_prot);
8f860591	212	else
c1e6098b	213	change_protection(vma, start, end, vma->vm_page_prot, dirty_accountable);
ab50b8ed HD	214	vm_stat_account(mm, oldflags, vma->vm_file, -nrpages);
ab50b8ed HD	215	vm_stat_account(mm, newflags, vma->vm_file, nrpages);
1da177e4 LT	216	return 0;
	217
	218	fail:
	219	vm_unacct_memory(charged);
	220	return error;
	221	}
	222
	223	asmlinkage long
	224	sys_mprotect(unsigned long start, size_t len, unsigned long prot)
	225	{
	226	unsigned long vm_flags, nstart, end, tmp, reqprot;
	227	struct vm_area_struct vma, prev;
	228	int error = -EINVAL;
	229	const int grows = prot & (PROT_GROWSDOWN\|PROT_GROWSUP);
	230	prot &= ~(PROT_GROWSDOWN\|PROT_GROWSUP);
	231	if (grows == (PROT_GROWSDOWN\|PROT_GROWSUP)) /* can't be both */
	232	return -EINVAL;
	233
	234	if (start & ~PAGE_MASK)
	235	return -EINVAL;
	236	if (!len)
	237	return 0;
	238	len = PAGE_ALIGN(len);
	239	end = start + len;
	240	if (end <= start)
	241	return -ENOMEM;
	242	if (prot & ~(PROT_READ \| PROT_WRITE \| PROT_EXEC \| PROT_SEM))
	243	return -EINVAL;
	244
	245	reqprot = prot;
	246	/*
	247	* Does the application expect PROT_READ to imply PROT_EXEC:
	248	*/
b344e05c	249	if ((prot & PROT_READ) && (current->personality & READ_IMPLIES_EXEC))
1da177e4 LT	250	prot \|= PROT_EXEC;
	251
	252	vm_flags = calc_vm_prot_bits(prot);
	253
	254	down_write(&current->mm->mmap_sem);
	255
	256	vma = find_vma_prev(current->mm, start, &prev);
	257	error = -ENOMEM;
	258	if (!vma)
	259	goto out;
	260	if (unlikely(grows & PROT_GROWSDOWN)) {
	261	if (vma->vm_start >= end)
	262	goto out;
	263	start = vma->vm_start;
	264	error = -EINVAL;
	265	if (!(vma->vm_flags & VM_GROWSDOWN))
	266	goto out;
	267	}
	268	else {
	269	if (vma->vm_start > start)
	270	goto out;
	271	if (unlikely(grows & PROT_GROWSUP)) {
	272	end = vma->vm_end;
	273	error = -EINVAL;
	274	if (!(vma->vm_flags & VM_GROWSUP))
	275	goto out;
	276	}
	277	}
	278	if (start > vma->vm_start)
	279	prev = vma;
	280
	281	for (nstart = start ; ; ) {
	282	unsigned long newflags;
	283
	284	/* Here we know that vma->vm_start <= nstart < vma->vm_end. */
	285
1da177e4 LT	286	newflags = vm_flags \| (vma->vm_flags & ~(VM_READ \| VM_WRITE \| VM_EXEC));
1da177e4 LT	287
7e2cff42 PBG	288	/* newflags >> 4 shift VM_MAY% in place of VM_% */
7e2cff42 PBG	289	if ((newflags & ~(newflags >> 4)) & (VM_READ \| VM_WRITE \| VM_EXEC)) {
1da177e4 LT	290	error = -EACCES;
	291	goto out;
	292	}
	293
	294	error = security_file_mprotect(vma, reqprot, prot);
	295	if (error)
	296	goto out;
	297
	298	tmp = vma->vm_end;
	299	if (tmp > end)
	300	tmp = end;
	301	error = mprotect_fixup(vma, &prev, nstart, tmp, newflags);
	302	if (error)
	303	goto out;
	304	nstart = tmp;
	305
	306	if (nstart < prev->vm_end)
	307	nstart = prev->vm_end;
	308	if (nstart >= end)
	309	goto out;
	310
	311	vma = prev->vm_next;
	312	if (!vma \|\| vma->vm_start != nstart) {
	313	error = -ENOMEM;
	314	goto out;
	315	}
	316	}
	317	out:
	318	up_write(&current->mm->mmap_sem);
	319	return error;
	320	}