[deliverable/linux.git] / mm / madvise.c

/*
 *	linux/mm/madvise.c
 *
 * Copyright (C) 1999  Linus Torvalds
 * Copyright (C) 2002  Christoph Hellwig
 */

#include <linux/mman.h>
#include <linux/pagemap.h>
#include <linux/syscalls.h>
#include <linux/mempolicy.h>
#include <linux/hugetlb.h>
#include <linux/sched.h>

/*
 * Any behaviour which results in changes to the vma->vm_flags needs to
 * take mmap_sem for writing. Others, which simply traverse vmas, need
 * to only take it for reading.
 */
static int madvise_need_mmap_write(int behavior)
{
	switch (behavior) {
	case MADV_REMOVE:
	case MADV_WILLNEED:
	case MADV_DONTNEED:
		return 0;
	default:
		/* be safe, default to 1. list exceptions explicitly */
		return 1;
	}
}

/*
 * We can potentially split a vm area into separate
 * areas, each area with its own behavior.
 */
static long madvise_behavior(struct vm_area_struct * vma,
		     struct vm_area_struct **prev,
		     unsigned long start, unsigned long end, int behavior)
{
	struct mm_struct * mm = vma->vm_mm;
	int error = 0;
	pgoff_t pgoff;
	int new_flags = vma->vm_flags;

	switch (behavior) {
	case MADV_NORMAL:
		new_flags = new_flags & ~VM_RAND_READ & ~VM_SEQ_READ;
		break;
	case MADV_SEQUENTIAL:
		new_flags = (new_flags & ~VM_RAND_READ) | VM_SEQ_READ;
		break;
	case MADV_RANDOM:
		new_flags = (new_flags & ~VM_SEQ_READ) | VM_RAND_READ;
		break;
	case MADV_DONTFORK:
		new_flags |= VM_DONTCOPY;
		break;
	case MADV_DOFORK:
		new_flags &= ~VM_DONTCOPY;
		break;
	}

	if (new_flags == vma->vm_flags) {
		*prev = vma;
		goto out;
	}

	pgoff = vma->vm_pgoff + ((start - vma->vm_start) >> PAGE_SHIFT);
	*prev = vma_merge(mm, *prev, start, end, new_flags, vma->anon_vma,
				vma->vm_file, pgoff, vma_policy(vma));
	if (*prev) {
		vma = *prev;
		goto success;
	}

	*prev = vma;

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

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

success:
	/*
	 * vm_flags is protected by the mmap_sem held in write mode.
	 */
	vma->vm_flags = new_flags;

out:
	if (error == -ENOMEM)
		error = -EAGAIN;
	return error;
}

/*
 * Schedule all required I/O operations.  Do not wait for completion.
 */
static long madvise_willneed(struct vm_area_struct * vma,
			     struct vm_area_struct ** prev,
			     unsigned long start, unsigned long end)
{
	struct file *file = vma->vm_file;

	if (!file)
		return -EBADF;

	/*
	 * Page cache readahead assumes page cache pages are order-0 which
	 * is not the case for hugetlbfs. Do not give a bad return value
	 * but ignore the advice.
	 */
	if (vma->vm_flags & VM_HUGETLB)
		return 0;

	if (file->f_mapping->a_ops->get_xip_mem) {
		/* no bad return value, but ignore advice */
		return 0;
	}

	*prev = vma;
	start = ((start - vma->vm_start) >> PAGE_SHIFT) + vma->vm_pgoff;
	if (end > vma->vm_end)
		end = vma->vm_end;
	end = ((end - vma->vm_start) >> PAGE_SHIFT) + vma->vm_pgoff;

	force_page_cache_readahead(file->f_mapping,
			file, start, max_sane_readahead(end - start));
	return 0;
}

/*
 * Application no longer needs these pages.  If the pages are dirty,
 * it's OK to just throw them away.  The app will be more careful about
 * data it wants to keep.  Be sure to free swap resources too.  The
 * zap_page_range call sets things up for shrink_active_list to actually free
 * these pages later if no one else has touched them in the meantime,
 * although we could add these pages to a global reuse list for
 * shrink_active_list to pick up before reclaiming other pages.
 *
 * NB: This interface discards data rather than pushes it out to swap,
 * as some implementations do.  This has performance implications for
 * applications like large transactional databases which want to discard
 * pages in anonymous maps after committing to backing store the data
 * that was kept in them.  There is no reason to write this data out to
 * the swap area if the application is discarding it.
 *
 * An interface that causes the system to free clean pages and flush
 * dirty pages is already available as msync(MS_INVALIDATE).
 */
static long madvise_dontneed(struct vm_area_struct * vma,
			     struct vm_area_struct ** prev,
			     unsigned long start, unsigned long end)
{
	*prev = vma;
	if (vma->vm_flags & (VM_LOCKED|VM_HUGETLB|VM_PFNMAP))
		return -EINVAL;

	if (unlikely(vma->vm_flags & VM_NONLINEAR)) {
		struct zap_details details = {
			.nonlinear_vma = vma,
			.last_index = ULONG_MAX,
		};
		zap_page_range(vma, start, end - start, &details);
	} else
		zap_page_range(vma, start, end - start, NULL);
	return 0;
}

/*
 * Application wants to free up the pages and associated backing store.
 * This is effectively punching a hole into the middle of a file.
 *
 * NOTE: Currently, only shmfs/tmpfs is supported for this operation.
 * Other filesystems return -ENOSYS.
 */
static long madvise_remove(struct vm_area_struct *vma,
				struct vm_area_struct **prev,
				unsigned long start, unsigned long end)
{
	struct address_space *mapping;
	loff_t offset, endoff;
	int error;

	*prev = NULL;	/* tell sys_madvise we drop mmap_sem */

	if (vma->vm_flags & (VM_LOCKED|VM_NONLINEAR|VM_HUGETLB))
		return -EINVAL;

	if (!vma->vm_file || !vma->vm_file->f_mapping
		|| !vma->vm_file->f_mapping->host) {
			return -EINVAL;
	}

	if ((vma->vm_flags & (VM_SHARED|VM_WRITE)) != (VM_SHARED|VM_WRITE))
		return -EACCES;

	mapping = vma->vm_file->f_mapping;

	offset = (loff_t)(start - vma->vm_start)
			+ ((loff_t)vma->vm_pgoff << PAGE_SHIFT);
	endoff = (loff_t)(end - vma->vm_start - 1)
			+ ((loff_t)vma->vm_pgoff << PAGE_SHIFT);

	/* vmtruncate_range needs to take i_mutex and i_alloc_sem */
	up_read(&current->mm->mmap_sem);
	error = vmtruncate_range(mapping->host, offset, endoff);
	down_read(&current->mm->mmap_sem);
	return error;
}

static long
madvise_vma(struct vm_area_struct *vma, struct vm_area_struct **prev,
		unsigned long start, unsigned long end, int behavior)
{
	long error;

	switch (behavior) {
	case MADV_DOFORK:
		if (vma->vm_flags & VM_IO) {
			error = -EINVAL;
			break;
		}
	case MADV_DONTFORK:
	case MADV_NORMAL:
	case MADV_SEQUENTIAL:
	case MADV_RANDOM:
		error = madvise_behavior(vma, prev, start, end, behavior);
		break;
	case MADV_REMOVE:
		error = madvise_remove(vma, prev, start, end);
		break;

	case MADV_WILLNEED:
		error = madvise_willneed(vma, prev, start, end);
		break;

	case MADV_DONTNEED:
		error = madvise_dontneed(vma, prev, start, end);
		break;

	default:
		error = -EINVAL;
		break;
	}
	return error;
}

/*
 * The madvise(2) system call.
 *
 * Applications can use madvise() to advise the kernel how it should
 * handle paging I/O in this VM area.  The idea is to help the kernel
 * use appropriate read-ahead and caching techniques.  The information
 * provided is advisory only, and can be safely disregarded by the
 * kernel without affecting the correct operation of the application.
 *
 * behavior values:
 *  MADV_NORMAL - the default behavior is to read clusters.  This
 *		results in some read-ahead and read-behind.
 *  MADV_RANDOM - the system should read the minimum amount of data
 *		on any access, since it is unlikely that the appli-
 *		cation will need more than what it asks for.
 *  MADV_SEQUENTIAL - pages in the given range will probably be accessed
 *		once, so they can be aggressively read ahead, and
 *		can be freed soon after they are accessed.
 *  MADV_WILLNEED - the application is notifying the system to read
 *		some pages ahead.
 *  MADV_DONTNEED - the application is finished with the given range,
 *		so the kernel can free resources associated with it.
 *  MADV_REMOVE - the application wants to free up the given range of
 *		pages and associated backing store.
 *
 * return values:
 *  zero    - success
 *  -EINVAL - start + len < 0, start is not page-aligned,
 *		"behavior" is not a valid value, or application
 *		is attempting to release locked or shared pages.
 *  -ENOMEM - addresses in the specified range are not currently
 *		mapped, or are outside the AS of the process.
 *  -EIO    - an I/O error occurred while paging in data.
 *  -EBADF  - map exists, but area maps something that isn't a file.
 *  -EAGAIN - a kernel resource was temporarily unavailable.
 */
SYSCALL_DEFINE3(madvise, unsigned long, start, size_t, len_in, int, behavior)
{
	unsigned long end, tmp;
	struct vm_area_struct * vma, *prev;
	int unmapped_error = 0;
	int error = -EINVAL;
	int write;
	size_t len;

	write = madvise_need_mmap_write(behavior);
	if (write)
		down_write(&current->mm->mmap_sem);
	else
		down_read(&current->mm->mmap_sem);

	if (start & ~PAGE_MASK)
		goto out;
	len = (len_in + ~PAGE_MASK) & PAGE_MASK;

	/* Check to see whether len was rounded up from small -ve to zero */
	if (len_in && !len)
		goto out;

	end = start + len;
	if (end < start)
		goto out;

	error = 0;
	if (end == start)
		goto out;

	/*
	 * If the interval [start,end) covers some unmapped address
	 * ranges, just ignore them, but return -ENOMEM at the end.
	 * - different from the way of handling in mlock etc.
	 */
	vma = find_vma_prev(current->mm, start, &prev);
	if (vma && start > vma->vm_start)
		prev = vma;

	for (;;) {
		/* Still start < end. */
		error = -ENOMEM;
		if (!vma)
			goto out;

		/* Here start < (end|vma->vm_end). */
		if (start < vma->vm_start) {
			unmapped_error = -ENOMEM;
			start = vma->vm_start;
			if (start >= end)
				goto out;
		}

		/* Here vma->vm_start <= start < (end|vma->vm_end) */
		tmp = vma->vm_end;
		if (end < tmp)
			tmp = end;

		/* Here vma->vm_start <= start < tmp <= (end|vma->vm_end). */
		error = madvise_vma(vma, &prev, start, tmp, behavior);
		if (error)
			goto out;
		start = tmp;
		if (prev && start < prev->vm_end)
			start = prev->vm_end;
		error = unmapped_error;
		if (start >= end)
			goto out;
		if (prev)
			vma = prev->vm_next;
		else	/* madvise_remove dropped mmap_sem */
			vma = find_vma(current->mm, start);
	}
out:
	if (write)
		up_write(&current->mm->mmap_sem);
	else
		up_read(&current->mm->mmap_sem);

	return error;
}
Commit	Line	Data
1da177e4 LT	1	/*
	2	* linux/mm/madvise.c
	3	*
	4	* Copyright (C) 1999 Linus Torvalds
	5	* Copyright (C) 2002 Christoph Hellwig
	6	*/
	7
	8	#include <linux/mman.h>
	9	#include <linux/pagemap.h>
	10	#include <linux/syscalls.h>
05b74384	11	#include <linux/mempolicy.h>
1da177e4	12	#include <linux/hugetlb.h>
e8edc6e0	13	#include <linux/sched.h>
1da177e4	14
0a27a14a NP	15	/*
	16	* Any behaviour which results in changes to the vma->vm_flags needs to
	17	* take mmap_sem for writing. Others, which simply traverse vmas, need
	18	* to only take it for reading.
	19	*/
	20	static int madvise_need_mmap_write(int behavior)
	21	{
	22	switch (behavior) {
	23	case MADV_REMOVE:
	24	case MADV_WILLNEED:
	25	case MADV_DONTNEED:
	26	return 0;
	27	default:
	28	/* be safe, default to 1. list exceptions explicitly */
	29	return 1;
	30	}
	31	}
	32
1da177e4 LT	33	/*
	34	* We can potentially split a vm area into separate
	35	* areas, each area with its own behavior.
	36	*/
05b74384 PM	37	static long madvise_behavior(struct vm_area_struct * vma,
	38	struct vm_area_struct **prev,
	39	unsigned long start, unsigned long end, int behavior)
1da177e4 LT	40	{
	41	struct mm_struct * mm = vma->vm_mm;
	42	int error = 0;
05b74384	43	pgoff_t pgoff;
f8225661	44	int new_flags = vma->vm_flags;
e798c6e8 PM	45
e798c6e8 PM	46	switch (behavior) {
f8225661 MT	47	case MADV_NORMAL:
	48	new_flags = new_flags & ~VM_RAND_READ & ~VM_SEQ_READ;
	49	break;
e798c6e8	50	case MADV_SEQUENTIAL:
f8225661	51	new_flags = (new_flags & ~VM_RAND_READ) \| VM_SEQ_READ;
e798c6e8 PM	52	break;
e798c6e8 PM	53	case MADV_RANDOM:
f8225661	54	new_flags = (new_flags & ~VM_SEQ_READ) \| VM_RAND_READ;
e798c6e8	55	break;
f8225661 MT	56	case MADV_DONTFORK:
	57	new_flags \|= VM_DONTCOPY;
	58	break;
	59	case MADV_DOFORK:
	60	new_flags &= ~VM_DONTCOPY;
e798c6e8 PM	61	break;
	62	}
	63
05b74384 PM	64	if (new_flags == vma->vm_flags) {
05b74384 PM	65	*prev = vma;
836d5ffd	66	goto out;
05b74384 PM	67	}
	68
	69	pgoff = vma->vm_pgoff + ((start - vma->vm_start) >> PAGE_SHIFT);
	70	prev = vma_merge(mm, prev, start, end, new_flags, vma->anon_vma,
	71	vma->vm_file, pgoff, vma_policy(vma));
	72	if (*prev) {
	73	vma = *prev;
	74	goto success;
	75	}
	76
	77	*prev = vma;
1da177e4 LT	78
	79	if (start != vma->vm_start) {
	80	error = split_vma(mm, vma, start, 1);
	81	if (error)
	82	goto out;
	83	}
	84
	85	if (end != vma->vm_end) {
	86	error = split_vma(mm, vma, end, 0);
	87	if (error)
	88	goto out;
	89	}
	90
836d5ffd	91	success:
1da177e4 LT	92	/*
	93	* vm_flags is protected by the mmap_sem held in write mode.
	94	*/
e798c6e8	95	vma->vm_flags = new_flags;
1da177e4 LT	96
	97	out:
	98	if (error == -ENOMEM)
	99	error = -EAGAIN;
	100	return error;
	101	}
	102
	103	/*
	104	* Schedule all required I/O operations. Do not wait for completion.
	105	*/
	106	static long madvise_willneed(struct vm_area_struct * vma,
05b74384	107	struct vm_area_struct ** prev,
1da177e4 LT	108	unsigned long start, unsigned long end)
	109	{
	110	struct file *file = vma->vm_file;
	111
1bef4003 S	112	if (!file)
	113	return -EBADF;
	114
a425a638 MG	115	/*
	116	* Page cache readahead assumes page cache pages are order-0 which
	117	* is not the case for hugetlbfs. Do not give a bad return value
	118	* but ignore the advice.
	119	*/
	120	if (vma->vm_flags & VM_HUGETLB)
	121	return 0;
	122
70688e4d	123	if (file->f_mapping->a_ops->get_xip_mem) {
fe77ba6f CO	124	/* no bad return value, but ignore advice */
	125	return 0;
	126	}
	127
05b74384	128	*prev = vma;
1da177e4 LT	129	start = ((start - vma->vm_start) >> PAGE_SHIFT) + vma->vm_pgoff;
	130	if (end > vma->vm_end)
	131	end = vma->vm_end;
	132	end = ((end - vma->vm_start) >> PAGE_SHIFT) + vma->vm_pgoff;
	133
	134	force_page_cache_readahead(file->f_mapping,
	135	file, start, max_sane_readahead(end - start));
	136	return 0;
	137	}
	138
	139	/*
	140	* Application no longer needs these pages. If the pages are dirty,
	141	* it's OK to just throw them away. The app will be more careful about
	142	* data it wants to keep. Be sure to free swap resources too. The
7e6cbea3	143	* zap_page_range call sets things up for shrink_active_list to actually free
1da177e4 LT	144	* these pages later if no one else has touched them in the meantime,
1da177e4 LT	145	* although we could add these pages to a global reuse list for
7e6cbea3	146	* shrink_active_list to pick up before reclaiming other pages.
1da177e4 LT	147	*
	148	* NB: This interface discards data rather than pushes it out to swap,
	149	* as some implementations do. This has performance implications for
	150	* applications like large transactional databases which want to discard
	151	* pages in anonymous maps after committing to backing store the data
	152	* that was kept in them. There is no reason to write this data out to
	153	* the swap area if the application is discarding it.
	154	*
	155	* An interface that causes the system to free clean pages and flush
	156	* dirty pages is already available as msync(MS_INVALIDATE).
	157	*/
	158	static long madvise_dontneed(struct vm_area_struct * vma,
05b74384	159	struct vm_area_struct ** prev,
1da177e4 LT	160	unsigned long start, unsigned long end)
1da177e4 LT	161	{
05b74384	162	*prev = vma;
6aab341e	163	if (vma->vm_flags & (VM_LOCKED\|VM_HUGETLB\|VM_PFNMAP))
1da177e4 LT	164	return -EINVAL;
	165
	166	if (unlikely(vma->vm_flags & VM_NONLINEAR)) {
	167	struct zap_details details = {
	168	.nonlinear_vma = vma,
	169	.last_index = ULONG_MAX,
	170	};
	171	zap_page_range(vma, start, end - start, &details);
	172	} else
	173	zap_page_range(vma, start, end - start, NULL);
	174	return 0;
	175	}
	176
f6b3ec23 BP	177	/*
	178	* Application wants to free up the pages and associated backing store.
	179	* This is effectively punching a hole into the middle of a file.
	180	*
	181	* NOTE: Currently, only shmfs/tmpfs is supported for this operation.
	182	* Other filesystems return -ENOSYS.
	183	*/
	184	static long madvise_remove(struct vm_area_struct *vma,
00e9fa2d	185	struct vm_area_struct **prev,
f6b3ec23 BP	186	unsigned long start, unsigned long end)
	187	{
	188	struct address_space *mapping;
90ed52eb HD	189	loff_t offset, endoff;
90ed52eb HD	190	int error;
f6b3ec23	191
90ed52eb	192	prev = NULL; / tell sys_madvise we drop mmap_sem */
00e9fa2d	193
f6b3ec23 BP	194	if (vma->vm_flags & (VM_LOCKED\|VM_NONLINEAR\|VM_HUGETLB))
	195	return -EINVAL;
	196
	197	if (!vma->vm_file \|\| !vma->vm_file->f_mapping
	198	\|\| !vma->vm_file->f_mapping->host) {
	199	return -EINVAL;
	200	}
	201
69cf0fac HD	202	if ((vma->vm_flags & (VM_SHARED\|VM_WRITE)) != (VM_SHARED\|VM_WRITE))
	203	return -EACCES;
	204
f6b3ec23 BP	205	mapping = vma->vm_file->f_mapping;
	206
	207	offset = (loff_t)(start - vma->vm_start)
	208	+ ((loff_t)vma->vm_pgoff << PAGE_SHIFT);
	209	endoff = (loff_t)(end - vma->vm_start - 1)
	210	+ ((loff_t)vma->vm_pgoff << PAGE_SHIFT);
90ed52eb HD	211
90ed52eb HD	212	/* vmtruncate_range needs to take i_mutex and i_alloc_sem */
0a27a14a	213	up_read(&current->mm->mmap_sem);
90ed52eb	214	error = vmtruncate_range(mapping->host, offset, endoff);
0a27a14a	215	down_read(&current->mm->mmap_sem);
90ed52eb	216	return error;
f6b3ec23 BP	217	}
f6b3ec23 BP	218
165cd402	219	static long
	220	madvise_vma(struct vm_area_struct vma, struct vm_area_struct *prev,
	221	unsigned long start, unsigned long end, int behavior)
1da177e4	222	{
1bef4003	223	long error;
165cd402	224
1da177e4	225	switch (behavior) {
f8225661 MT	226	case MADV_DOFORK:
	227	if (vma->vm_flags & VM_IO) {
	228	error = -EINVAL;
	229	break;
	230	}
	231	case MADV_DONTFORK:
1da177e4 LT	232	case MADV_NORMAL:
	233	case MADV_SEQUENTIAL:
	234	case MADV_RANDOM:
05b74384	235	error = madvise_behavior(vma, prev, start, end, behavior);
1da177e4	236	break;
f6b3ec23	237	case MADV_REMOVE:
00e9fa2d	238	error = madvise_remove(vma, prev, start, end);
f6b3ec23	239	break;
1da177e4 LT	240
1da177e4 LT	241	case MADV_WILLNEED:
05b74384	242	error = madvise_willneed(vma, prev, start, end);
1da177e4 LT	243	break;
	244
	245	case MADV_DONTNEED:
05b74384	246	error = madvise_dontneed(vma, prev, start, end);
1da177e4 LT	247	break;
	248
	249	default:
	250	error = -EINVAL;
	251	break;
	252	}
1da177e4 LT	253	return error;
	254	}
	255
	256	/*
	257	* The madvise(2) system call.
	258	*
	259	* Applications can use madvise() to advise the kernel how it should
	260	* handle paging I/O in this VM area. The idea is to help the kernel
	261	* use appropriate read-ahead and caching techniques. The information
	262	* provided is advisory only, and can be safely disregarded by the
	263	* kernel without affecting the correct operation of the application.
	264	*
	265	* behavior values:
	266	* MADV_NORMAL - the default behavior is to read clusters. This
	267	* results in some read-ahead and read-behind.
	268	* MADV_RANDOM - the system should read the minimum amount of data
	269	* on any access, since it is unlikely that the appli-
	270	* cation will need more than what it asks for.
	271	* MADV_SEQUENTIAL - pages in the given range will probably be accessed
	272	* once, so they can be aggressively read ahead, and
	273	* can be freed soon after they are accessed.
	274	* MADV_WILLNEED - the application is notifying the system to read
	275	* some pages ahead.
	276	* MADV_DONTNEED - the application is finished with the given range,
	277	* so the kernel can free resources associated with it.
f6b3ec23 BP	278	* MADV_REMOVE - the application wants to free up the given range of
f6b3ec23 BP	279	* pages and associated backing store.
1da177e4 LT	280	*
	281	* return values:
	282	* zero - success
	283	* -EINVAL - start + len < 0, start is not page-aligned,
	284	* "behavior" is not a valid value, or application
	285	* is attempting to release locked or shared pages.
	286	* -ENOMEM - addresses in the specified range are not currently
	287	* mapped, or are outside the AS of the process.
	288	* -EIO - an I/O error occurred while paging in data.
	289	* -EBADF - map exists, but area maps something that isn't a file.
	290	* -EAGAIN - a kernel resource was temporarily unavailable.
	291	*/
3480b257	292	SYSCALL_DEFINE3(madvise, unsigned long, start, size_t, len_in, int, behavior)
1da177e4	293	{
05b74384 PM	294	unsigned long end, tmp;
05b74384 PM	295	struct vm_area_struct * vma, *prev;
1da177e4 LT	296	int unmapped_error = 0;
1da177e4 LT	297	int error = -EINVAL;
f7977793	298	int write;
1da177e4 LT	299	size_t len;
1da177e4 LT	300
f7977793 JB	301	write = madvise_need_mmap_write(behavior);
f7977793 JB	302	if (write)
0a27a14a NP	303	down_write(&current->mm->mmap_sem);
	304	else
	305	down_read(&current->mm->mmap_sem);
1da177e4 LT	306
	307	if (start & ~PAGE_MASK)
	308	goto out;
	309	len = (len_in + ~PAGE_MASK) & PAGE_MASK;
	310
	311	/* Check to see whether len was rounded up from small -ve to zero */
	312	if (len_in && !len)
	313	goto out;
	314
	315	end = start + len;
	316	if (end < start)
	317	goto out;
	318
	319	error = 0;
	320	if (end == start)
	321	goto out;
	322
	323	/*
	324	* If the interval [start,end) covers some unmapped address
	325	* ranges, just ignore them, but return -ENOMEM at the end.
05b74384	326	* - different from the way of handling in mlock etc.
1da177e4	327	*/
05b74384	328	vma = find_vma_prev(current->mm, start, &prev);
836d5ffd HD	329	if (vma && start > vma->vm_start)
	330	prev = vma;
	331
1da177e4 LT	332	for (;;) {
	333	/* Still start < end. */
	334	error = -ENOMEM;
	335	if (!vma)
	336	goto out;
	337
05b74384	338	/* Here start < (end\|vma->vm_end). */
1da177e4 LT	339	if (start < vma->vm_start) {
	340	unmapped_error = -ENOMEM;
	341	start = vma->vm_start;
05b74384 PM	342	if (start >= end)
05b74384 PM	343	goto out;
1da177e4 LT	344	}
1da177e4 LT	345
05b74384 PM	346	/* Here vma->vm_start <= start < (end\|vma->vm_end) */
	347	tmp = vma->vm_end;
	348	if (end < tmp)
	349	tmp = end;
1da177e4	350
05b74384 PM	351	/* Here vma->vm_start <= start < tmp <= (end\|vma->vm_end). */
05b74384 PM	352	error = madvise_vma(vma, &prev, start, tmp, behavior);
1da177e4 LT	353	if (error)
1da177e4 LT	354	goto out;
05b74384	355	start = tmp;
90ed52eb	356	if (prev && start < prev->vm_end)
05b74384 PM	357	start = prev->vm_end;
	358	error = unmapped_error;
	359	if (start >= end)
	360	goto out;
90ed52eb HD	361	if (prev)
	362	vma = prev->vm_next;
	363	else /* madvise_remove dropped mmap_sem */
	364	vma = find_vma(current->mm, start);
1da177e4	365	}
1da177e4	366	out:
f7977793	367	if (write)
0a27a14a NP	368	up_write(&current->mm->mmap_sem);
	369	else
	370	up_read(&current->mm->mmap_sem);
	371
1da177e4 LT	372	return error;
1da177e4 LT	373	}