[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 <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;
	spinlock_t *ptl;

	pte = pte_offset_map_lock(mm, pmd, addr, &ptl);
	do {
		if (pte_present(*pte)) {
			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);
		}
	} 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;
	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|VM_HUGETLB))) {
			charged = nrpages;
			if (security_vm_enough_memory(charged))
				return -ENOMEM;
			newflags |= VM_ACCOUNT;
		}
	}

	newprot = protection_map[newflags & 0xf];

	/*
	 * 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 = newprot;
	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 (unlikely((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. */

		if (is_vm_hugetlb_page(vma)) {
			error = -EACCES;
			goto out;
		}

		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>
	22
	23	#include <asm/uaccess.h>
	24	#include <asm/pgtable.h>
	25	#include <asm/cacheflush.h>
	26	#include <asm/tlbflush.h>
	27
	28	static void change_pte_range(struct mm_struct mm, pmd_t pmd,
	29	unsigned long addr, unsigned long end, pgprot_t newprot)
	30	{
	31	pte_t *pte;
705e87c0	32	spinlock_t *ptl;
1da177e4	33
705e87c0	34	pte = pte_offset_map_lock(mm, pmd, addr, &ptl);
1da177e4 LT	35	do {
	36	if (pte_present(*pte)) {
	37	pte_t ptent;
	38
	39	/* Avoid an SMP race with hardware updated dirty/clean
	40	* bits by wiping the pte and then setting the new pte
	41	* into place.
	42	*/
	43	ptent = pte_modify(ptep_get_and_clear(mm, addr, pte), newprot);
	44	set_pte_at(mm, addr, pte, ptent);
	45	lazy_mmu_prot_update(ptent);
	46	}
	47	} while (pte++, addr += PAGE_SIZE, addr != end);
705e87c0	48	pte_unmap_unlock(pte - 1, ptl);
1da177e4 LT	49	}
	50
	51	static inline void change_pmd_range(struct mm_struct mm, pud_t pud,
	52	unsigned long addr, unsigned long end, pgprot_t newprot)
	53	{
	54	pmd_t *pmd;
	55	unsigned long next;
	56
	57	pmd = pmd_offset(pud, addr);
	58	do {
	59	next = pmd_addr_end(addr, end);
	60	if (pmd_none_or_clear_bad(pmd))
	61	continue;
	62	change_pte_range(mm, pmd, addr, next, newprot);
	63	} while (pmd++, addr = next, addr != end);
	64	}
	65
	66	static inline void change_pud_range(struct mm_struct mm, pgd_t pgd,
	67	unsigned long addr, unsigned long end, pgprot_t newprot)
	68	{
	69	pud_t *pud;
	70	unsigned long next;
	71
	72	pud = pud_offset(pgd, addr);
	73	do {
	74	next = pud_addr_end(addr, end);
	75	if (pud_none_or_clear_bad(pud))
	76	continue;
	77	change_pmd_range(mm, pud, addr, next, newprot);
	78	} while (pud++, addr = next, addr != end);
	79	}
	80
	81	static void change_protection(struct vm_area_struct *vma,
	82	unsigned long addr, unsigned long end, pgprot_t newprot)
	83	{
	84	struct mm_struct *mm = vma->vm_mm;
	85	pgd_t *pgd;
	86	unsigned long next;
	87	unsigned long start = addr;
	88
	89	BUG_ON(addr >= end);
	90	pgd = pgd_offset(mm, addr);
	91	flush_cache_range(vma, addr, end);
1da177e4 LT	92	do {
	93	next = pgd_addr_end(addr, end);
	94	if (pgd_none_or_clear_bad(pgd))
	95	continue;
	96	change_pud_range(mm, pgd, addr, next, newprot);
	97	} while (pgd++, addr = next, addr != end);
	98	flush_tlb_range(vma, start, end);
1da177e4 LT	99	}
	100
	101	static int
	102	mprotect_fixup(struct vm_area_struct vma, struct vm_area_struct *pprev,
	103	unsigned long start, unsigned long end, unsigned long newflags)
	104	{
	105	struct mm_struct *mm = vma->vm_mm;
	106	unsigned long oldflags = vma->vm_flags;
	107	long nrpages = (end - start) >> PAGE_SHIFT;
	108	unsigned long charged = 0;
	109	pgprot_t newprot;
	110	pgoff_t pgoff;
	111	int error;
	112
	113	if (newflags == oldflags) {
	114	*pprev = vma;
	115	return 0;
	116	}
	117
	118	/*
	119	* If we make a private mapping writable we increase our commit;
	120	* but (without finer accounting) cannot reduce our commit if we
	121	* make it unwritable again.
	122	*
	123	* FIXME? We haven't defined a VM_NORESERVE flag, so mprotecting
	124	* a MAP_NORESERVE private mapping to writable will now reserve.
	125	*/
	126	if (newflags & VM_WRITE) {
	127	if (!(oldflags & (VM_ACCOUNT\|VM_WRITE\|VM_SHARED\|VM_HUGETLB))) {
	128	charged = nrpages;
	129	if (security_vm_enough_memory(charged))
	130	return -ENOMEM;
	131	newflags \|= VM_ACCOUNT;
	132	}
	133	}
	134
	135	newprot = protection_map[newflags & 0xf];
	136
	137	/*
	138	* First try to merge with previous and/or next vma.
	139	*/
	140	pgoff = vma->vm_pgoff + ((start - vma->vm_start) >> PAGE_SHIFT);
	141	pprev = vma_merge(mm, pprev, start, end, newflags,
	142	vma->anon_vma, vma->vm_file, pgoff, vma_policy(vma));
	143	if (*pprev) {
	144	vma = *pprev;
	145	goto success;
	146	}
	147
	148	*pprev = vma;
	149
	150	if (start != vma->vm_start) {
	151	error = split_vma(mm, vma, start, 1);
	152	if (error)
	153	goto fail;
	154	}
	155
	156	if (end != vma->vm_end) {
	157	error = split_vma(mm, vma, end, 0);
	158	if (error)
	159	goto fail;
	160	}
	161
	162	success:
163	/*
164	* vm_flags and vm_page_prot are protected by the mmap_sem
165	* held in write mode.
166	*/
167	vma->vm_flags = newflags;
168	vma->vm_page_prot = newprot;
169	change_protection(vma, start, end, newprot);
ab50b8ed HD	170	vm_stat_account(mm, oldflags, vma->vm_file, -nrpages);
ab50b8ed HD	171	vm_stat_account(mm, newflags, vma->vm_file, nrpages);
1da177e4 LT	172	return 0;
	173
	174	fail:
	175	vm_unacct_memory(charged);
	176	return error;
	177	}
	178
	179	asmlinkage long
	180	sys_mprotect(unsigned long start, size_t len, unsigned long prot)
	181	{
	182	unsigned long vm_flags, nstart, end, tmp, reqprot;
	183	struct vm_area_struct vma, prev;
	184	int error = -EINVAL;
	185	const int grows = prot & (PROT_GROWSDOWN\|PROT_GROWSUP);
	186	prot &= ~(PROT_GROWSDOWN\|PROT_GROWSUP);
	187	if (grows == (PROT_GROWSDOWN\|PROT_GROWSUP)) /* can't be both */
	188	return -EINVAL;
	189
	190	if (start & ~PAGE_MASK)
	191	return -EINVAL;
	192	if (!len)
	193	return 0;
	194	len = PAGE_ALIGN(len);
	195	end = start + len;
	196	if (end <= start)
	197	return -ENOMEM;
	198	if (prot & ~(PROT_READ \| PROT_WRITE \| PROT_EXEC \| PROT_SEM))
	199	return -EINVAL;
	200
	201	reqprot = prot;
	202	/*
	203	* Does the application expect PROT_READ to imply PROT_EXEC:
	204	*/
	205	if (unlikely((prot & PROT_READ) &&
	206	(current->personality & READ_IMPLIES_EXEC)))
	207	prot \|= PROT_EXEC;
	208
	209	vm_flags = calc_vm_prot_bits(prot);
	210
	211	down_write(&current->mm->mmap_sem);
	212
	213	vma = find_vma_prev(current->mm, start, &prev);
	214	error = -ENOMEM;
	215	if (!vma)
	216	goto out;
	217	if (unlikely(grows & PROT_GROWSDOWN)) {
	218	if (vma->vm_start >= end)
	219	goto out;
	220	start = vma->vm_start;
	221	error = -EINVAL;
	222	if (!(vma->vm_flags & VM_GROWSDOWN))
	223	goto out;
	224	}
	225	else {
	226	if (vma->vm_start > start)
	227	goto out;
	228	if (unlikely(grows & PROT_GROWSUP)) {
	229	end = vma->vm_end;
	230	error = -EINVAL;
	231	if (!(vma->vm_flags & VM_GROWSUP))
	232	goto out;
	233	}
	234	}
	235	if (start > vma->vm_start)
236	prev = vma;
237
238	for (nstart = start ; ; ) {
239	unsigned long newflags;
240
241	/* Here we know that vma->vm_start <= nstart < vma->vm_end. */
242
243	if (is_vm_hugetlb_page(vma)) {
244	error = -EACCES;
245	goto out;
246	}
247
248	newflags = vm_flags \| (vma->vm_flags & ~(VM_READ \| VM_WRITE \| VM_EXEC));
249
7e2cff42 PBG	250	/* newflags >> 4 shift VM_MAY% in place of VM_% */
7e2cff42 PBG	251	if ((newflags & ~(newflags >> 4)) & (VM_READ \| VM_WRITE \| VM_EXEC)) {
1da177e4 LT	252	error = -EACCES;
	253	goto out;
	254	}
	255
	256	error = security_file_mprotect(vma, reqprot, prot);
	257	if (error)
	258	goto out;
	259
	260	tmp = vma->vm_end;
	261	if (tmp > end)
	262	tmp = end;
	263	error = mprotect_fixup(vma, &prev, nstart, tmp, newflags);
	264	if (error)
	265	goto out;
	266	nstart = tmp;
	267
	268	if (nstart < prev->vm_end)
	269	nstart = prev->vm_end;
	270	if (nstart >= end)
	271	goto out;
	272
	273	vma = prev->vm_next;
	274	if (!vma \|\| vma->vm_start != nstart) {
	275	error = -ENOMEM;
	276	goto out;
	277	}
	278	}
	279	out:
	280	up_write(&current->mm->mmap_sem);
	281	return error;
	282	}