[mirror_ubuntu-zesty-kernel.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;
	unsigned long 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:
		if (vma->vm_flags & VM_IO) {
			error = -EINVAL;
			goto out;
		}
		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;

	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, 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)
{
	switch (behavior) {
	case MADV_REMOVE:
		return madvise_remove(vma, prev, start, end);
	case MADV_WILLNEED:
		return madvise_willneed(vma, prev, start, end);
	case MADV_DONTNEED:
		return madvise_dontneed(vma, prev, start, end);
	default:
		return madvise_behavior(vma, prev, start, end, behavior);
	}
}

static int
madvise_behavior_valid(int behavior)
{
	switch (behavior) {
	case MADV_DOFORK:
	case MADV_DONTFORK:
	case MADV_NORMAL:
	case MADV_SEQUENTIAL:
	case MADV_RANDOM:
	case MADV_REMOVE:
	case MADV_WILLNEED:
	case MADV_DONTNEED:
		return 1;

	default:
		return 0;
	}
}

/*
 * 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.
 *  MADV_DONTFORK - omit this area from child's address space when forking:
 *		typically, to avoid COWing pages pinned by get_user_pages().
 *  MADV_DOFORK - cancel MADV_DONTFORK: no longer omit this area when forking.
 *
 * 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;

	if (!madvise_behavior_valid(behavior))
		return error;

	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;
3866ea90	44	unsigned long 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:
3866ea90 HD	60	if (vma->vm_flags & VM_IO) {
	61	error = -EINVAL;
	62	goto out;
	63	}
f8225661	64	new_flags &= ~VM_DONTCOPY;
e798c6e8 PM	65	break;
	66	}
	67
05b74384 PM	68	if (new_flags == vma->vm_flags) {
05b74384 PM	69	*prev = vma;
836d5ffd	70	goto out;
05b74384 PM	71	}
	72
	73	pgoff = vma->vm_pgoff + ((start - vma->vm_start) >> PAGE_SHIFT);
	74	prev = vma_merge(mm, prev, start, end, new_flags, vma->anon_vma,
	75	vma->vm_file, pgoff, vma_policy(vma));
	76	if (*prev) {
	77	vma = *prev;
	78	goto success;
	79	}
	80
	81	*prev = vma;
1da177e4 LT	82
	83	if (start != vma->vm_start) {
	84	error = split_vma(mm, vma, start, 1);
	85	if (error)
	86	goto out;
	87	}
	88
	89	if (end != vma->vm_end) {
	90	error = split_vma(mm, vma, end, 0);
	91	if (error)
	92	goto out;
	93	}
	94
836d5ffd	95	success:
1da177e4 LT	96	/*
	97	* vm_flags is protected by the mmap_sem held in write mode.
	98	*/
e798c6e8	99	vma->vm_flags = new_flags;
1da177e4 LT	100
	101	out:
	102	if (error == -ENOMEM)
	103	error = -EAGAIN;
	104	return error;
	105	}
	106
	107	/*
	108	* Schedule all required I/O operations. Do not wait for completion.
	109	*/
	110	static long madvise_willneed(struct vm_area_struct * vma,
05b74384	111	struct vm_area_struct ** prev,
1da177e4 LT	112	unsigned long start, unsigned long end)
	113	{
	114	struct file *file = vma->vm_file;
	115
1bef4003 S	116	if (!file)
	117	return -EBADF;
	118
70688e4d	119	if (file->f_mapping->a_ops->get_xip_mem) {
fe77ba6f CO	120	/* no bad return value, but ignore advice */
	121	return 0;
	122	}
	123
05b74384	124	*prev = vma;
1da177e4 LT	125	start = ((start - vma->vm_start) >> PAGE_SHIFT) + vma->vm_pgoff;
	126	if (end > vma->vm_end)
	127	end = vma->vm_end;
	128	end = ((end - vma->vm_start) >> PAGE_SHIFT) + vma->vm_pgoff;
	129
f7e839dd	130	force_page_cache_readahead(file->f_mapping, file, start, end - start);
1da177e4 LT	131	return 0;
	132	}
	133
	134	/*
	135	* Application no longer needs these pages. If the pages are dirty,
	136	* it's OK to just throw them away. The app will be more careful about
	137	* data it wants to keep. Be sure to free swap resources too. The
7e6cbea3	138	* zap_page_range call sets things up for shrink_active_list to actually free
1da177e4 LT	139	* these pages later if no one else has touched them in the meantime,
1da177e4 LT	140	* although we could add these pages to a global reuse list for
7e6cbea3	141	* shrink_active_list to pick up before reclaiming other pages.
1da177e4 LT	142	*
	143	* NB: This interface discards data rather than pushes it out to swap,
	144	* as some implementations do. This has performance implications for
	145	* applications like large transactional databases which want to discard
	146	* pages in anonymous maps after committing to backing store the data
	147	* that was kept in them. There is no reason to write this data out to
	148	* the swap area if the application is discarding it.
	149	*
	150	* An interface that causes the system to free clean pages and flush
	151	* dirty pages is already available as msync(MS_INVALIDATE).
	152	*/
	153	static long madvise_dontneed(struct vm_area_struct * vma,
05b74384	154	struct vm_area_struct ** prev,
1da177e4 LT	155	unsigned long start, unsigned long end)
1da177e4 LT	156	{
05b74384	157	*prev = vma;
6aab341e	158	if (vma->vm_flags & (VM_LOCKED\|VM_HUGETLB\|VM_PFNMAP))
1da177e4 LT	159	return -EINVAL;
	160
	161	if (unlikely(vma->vm_flags & VM_NONLINEAR)) {
	162	struct zap_details details = {
	163	.nonlinear_vma = vma,
	164	.last_index = ULONG_MAX,
	165	};
	166	zap_page_range(vma, start, end - start, &details);
	167	} else
	168	zap_page_range(vma, start, end - start, NULL);
	169	return 0;
	170	}
	171
f6b3ec23 BP	172	/*
	173	* Application wants to free up the pages and associated backing store.
	174	* This is effectively punching a hole into the middle of a file.
	175	*
	176	* NOTE: Currently, only shmfs/tmpfs is supported for this operation.
	177	* Other filesystems return -ENOSYS.
	178	*/
	179	static long madvise_remove(struct vm_area_struct *vma,
00e9fa2d	180	struct vm_area_struct **prev,
f6b3ec23 BP	181	unsigned long start, unsigned long end)
	182	{
	183	struct address_space *mapping;
90ed52eb HD	184	loff_t offset, endoff;
90ed52eb HD	185	int error;
f6b3ec23	186
90ed52eb	187	prev = NULL; / tell sys_madvise we drop mmap_sem */
00e9fa2d	188
f6b3ec23 BP	189	if (vma->vm_flags & (VM_LOCKED\|VM_NONLINEAR\|VM_HUGETLB))
	190	return -EINVAL;
	191
	192	if (!vma->vm_file \|\| !vma->vm_file->f_mapping
	193	\|\| !vma->vm_file->f_mapping->host) {
	194	return -EINVAL;
	195	}
	196
69cf0fac HD	197	if ((vma->vm_flags & (VM_SHARED\|VM_WRITE)) != (VM_SHARED\|VM_WRITE))
	198	return -EACCES;
	199
f6b3ec23 BP	200	mapping = vma->vm_file->f_mapping;
	201
	202	offset = (loff_t)(start - vma->vm_start)
	203	+ ((loff_t)vma->vm_pgoff << PAGE_SHIFT);
	204	endoff = (loff_t)(end - vma->vm_start - 1)
	205	+ ((loff_t)vma->vm_pgoff << PAGE_SHIFT);
90ed52eb HD	206
90ed52eb HD	207	/* vmtruncate_range needs to take i_mutex and i_alloc_sem */
0a27a14a	208	up_read(&current->mm->mmap_sem);
90ed52eb	209	error = vmtruncate_range(mapping->host, offset, endoff);
0a27a14a	210	down_read(&current->mm->mmap_sem);
90ed52eb	211	return error;
f6b3ec23 BP	212	}
f6b3ec23 BP	213
165cd402	214	static long
	215	madvise_vma(struct vm_area_struct vma, struct vm_area_struct *prev,
	216	unsigned long start, unsigned long end, int behavior)
1da177e4	217	{
1da177e4	218	switch (behavior) {
f6b3ec23	219	case MADV_REMOVE:
3866ea90	220	return madvise_remove(vma, prev, start, end);
1da177e4	221	case MADV_WILLNEED:
3866ea90	222	return madvise_willneed(vma, prev, start, end);
1da177e4	223	case MADV_DONTNEED:
3866ea90	224	return madvise_dontneed(vma, prev, start, end);
1da177e4	225	default:
3866ea90	226	return madvise_behavior(vma, prev, start, end, behavior);
1da177e4	227	}
1da177e4 LT	228	}
1da177e4 LT	229
75927af8 NP	230	static int
	231	madvise_behavior_valid(int behavior)
	232	{
	233	switch (behavior) {
	234	case MADV_DOFORK:
	235	case MADV_DONTFORK:
	236	case MADV_NORMAL:
	237	case MADV_SEQUENTIAL:
	238	case MADV_RANDOM:
	239	case MADV_REMOVE:
	240	case MADV_WILLNEED:
	241	case MADV_DONTNEED:
	242	return 1;
	243
	244	default:
	245	return 0;
	246	}
	247	}
3866ea90	248
1da177e4 LT	249	/*
	250	* The madvise(2) system call.
	251	*
	252	* Applications can use madvise() to advise the kernel how it should
	253	* handle paging I/O in this VM area. The idea is to help the kernel
	254	* use appropriate read-ahead and caching techniques. The information
	255	* provided is advisory only, and can be safely disregarded by the
	256	* kernel without affecting the correct operation of the application.
	257	*
	258	* behavior values:
	259	* MADV_NORMAL - the default behavior is to read clusters. This
	260	* results in some read-ahead and read-behind.
	261	* MADV_RANDOM - the system should read the minimum amount of data
	262	* on any access, since it is unlikely that the appli-
	263	* cation will need more than what it asks for.
	264	* MADV_SEQUENTIAL - pages in the given range will probably be accessed
	265	* once, so they can be aggressively read ahead, and
	266	* can be freed soon after they are accessed.
	267	* MADV_WILLNEED - the application is notifying the system to read
	268	* some pages ahead.
	269	* MADV_DONTNEED - the application is finished with the given range,
	270	* so the kernel can free resources associated with it.
f6b3ec23 BP	271	* MADV_REMOVE - the application wants to free up the given range of
f6b3ec23 BP	272	* pages and associated backing store.
3866ea90 HD	273	* MADV_DONTFORK - omit this area from child's address space when forking:
	274	* typically, to avoid COWing pages pinned by get_user_pages().
	275	* MADV_DOFORK - cancel MADV_DONTFORK: no longer omit this area when forking.
1da177e4 LT	276	*
	277	* return values:
	278	* zero - success
	279	* -EINVAL - start + len < 0, start is not page-aligned,
	280	* "behavior" is not a valid value, or application
	281	* is attempting to release locked or shared pages.
	282	* -ENOMEM - addresses in the specified range are not currently
	283	* mapped, or are outside the AS of the process.
	284	* -EIO - an I/O error occurred while paging in data.
	285	* -EBADF - map exists, but area maps something that isn't a file.
	286	* -EAGAIN - a kernel resource was temporarily unavailable.
	287	*/
3480b257	288	SYSCALL_DEFINE3(madvise, unsigned long, start, size_t, len_in, int, behavior)
1da177e4	289	{
05b74384 PM	290	unsigned long end, tmp;
05b74384 PM	291	struct vm_area_struct * vma, *prev;
1da177e4 LT	292	int unmapped_error = 0;
1da177e4 LT	293	int error = -EINVAL;
f7977793	294	int write;
1da177e4 LT	295	size_t len;
1da177e4 LT	296
75927af8 NP	297	if (!madvise_behavior_valid(behavior))
	298	return error;
	299
f7977793 JB	300	write = madvise_need_mmap_write(behavior);
f7977793 JB	301	if (write)
0a27a14a NP	302	down_write(&current->mm->mmap_sem);
	303	else
	304	down_read(&current->mm->mmap_sem);
1da177e4 LT	305
	306	if (start & ~PAGE_MASK)
	307	goto out;
	308	len = (len_in + ~PAGE_MASK) & PAGE_MASK;
	309
	310	/* Check to see whether len was rounded up from small -ve to zero */
	311	if (len_in && !len)
	312	goto out;
	313
	314	end = start + len;
	315	if (end < start)
	316	goto out;
	317
	318	error = 0;
	319	if (end == start)
	320	goto out;
	321
	322	/*
	323	* If the interval [start,end) covers some unmapped address
	324	* ranges, just ignore them, but return -ENOMEM at the end.
05b74384	325	* - different from the way of handling in mlock etc.
1da177e4	326	*/
05b74384	327	vma = find_vma_prev(current->mm, start, &prev);
836d5ffd HD	328	if (vma && start > vma->vm_start)
	329	prev = vma;
	330
1da177e4 LT	331	for (;;) {
	332	/* Still start < end. */
	333	error = -ENOMEM;
	334	if (!vma)
	335	goto out;
	336
05b74384	337	/* Here start < (end\|vma->vm_end). */
1da177e4 LT	338	if (start < vma->vm_start) {
	339	unmapped_error = -ENOMEM;
	340	start = vma->vm_start;
05b74384 PM	341	if (start >= end)
05b74384 PM	342	goto out;
1da177e4 LT	343	}
1da177e4 LT	344
05b74384 PM	345	/* Here vma->vm_start <= start < (end\|vma->vm_end) */
	346	tmp = vma->vm_end;
	347	if (end < tmp)
	348	tmp = end;
1da177e4	349
05b74384 PM	350	/* Here vma->vm_start <= start < tmp <= (end\|vma->vm_end). */
05b74384 PM	351	error = madvise_vma(vma, &prev, start, tmp, behavior);
1da177e4 LT	352	if (error)
1da177e4 LT	353	goto out;
05b74384	354	start = tmp;
90ed52eb	355	if (prev && start < prev->vm_end)
05b74384 PM	356	start = prev->vm_end;
	357	error = unmapped_error;
	358	if (start >= end)
	359	goto out;
90ed52eb HD	360	if (prev)
	361	vma = prev->vm_next;
	362	else /* madvise_remove dropped mmap_sem */
	363	vma = find_vma(current->mm, start);
1da177e4	364	}
1da177e4	365	out:
f7977793	366	if (write)
0a27a14a NP	367	up_write(&current->mm->mmap_sem);
	368	else
	369	up_read(&current->mm->mmap_sem);
	370
1da177e4 LT	371	return error;
1da177e4 LT	372	}