[linux.git] / mm / filemap_xip.c

/*
 *	linux/mm/filemap_xip.c
 *
 * Copyright (C) 2005 IBM Corporation
 * Author: Carsten Otte <[email protected]>
 *
 * derived from linux/mm/filemap.c - Copyright (C) Linus Torvalds
 *
 */

#include <linux/fs.h>
#include <linux/pagemap.h>
#include <linux/module.h>
#include <linux/uio.h>
#include <linux/rmap.h>
#include <asm/tlbflush.h>
#include "filemap.h"

/*
 * This is a file read routine for execute in place files, and uses
 * the mapping->a_ops->get_xip_page() function for the actual low-level
 * stuff.
 *
 * Note the struct file* is not used at all.  It may be NULL.
 */
static void
do_xip_mapping_read(struct address_space *mapping,
		    struct file_ra_state *_ra,
		    struct file *filp,
		    loff_t *ppos,
		    read_descriptor_t *desc,
		    read_actor_t actor)
{
	struct inode *inode = mapping->host;
	unsigned long index, end_index, offset;
	loff_t isize;

	BUG_ON(!mapping->a_ops->get_xip_page);

	index = *ppos >> PAGE_CACHE_SHIFT;
	offset = *ppos & ~PAGE_CACHE_MASK;

	isize = i_size_read(inode);
	if (!isize)
		goto out;

	end_index = (isize - 1) >> PAGE_CACHE_SHIFT;
	for (;;) {
		struct page *page;
		unsigned long nr, ret;

		/* nr is the maximum number of bytes to copy from this page */
		nr = PAGE_CACHE_SIZE;
		if (index >= end_index) {
			if (index > end_index)
				goto out;
			nr = ((isize - 1) & ~PAGE_CACHE_MASK) + 1;
			if (nr <= offset) {
				goto out;
			}
		}
		nr = nr - offset;

		page = mapping->a_ops->get_xip_page(mapping,
			index*(PAGE_SIZE/512), 0);
		if (!page)
			goto no_xip_page;
		if (unlikely(IS_ERR(page))) {
			if (PTR_ERR(page) == -ENODATA) {
				/* sparse */
				page = ZERO_PAGE(0);
			} else {
				desc->error = PTR_ERR(page);
				goto out;
			}
		}

		/* If users can be writing to this page using arbitrary
		 * virtual addresses, take care about potential aliasing
		 * before reading the page on the kernel side.
		 */
		if (mapping_writably_mapped(mapping))
			flush_dcache_page(page);

		/*
		 * Ok, we have the page, so now we can copy it to user space...
		 *
		 * The actor routine returns how many bytes were actually used..
		 * NOTE! This may not be the same as how much of a user buffer
		 * we filled up (we may be padding etc), so we can only update
		 * "pos" here (the actor routine has to update the user buffer
		 * pointers and the remaining count).
		 */
		ret = actor(desc, page, offset, nr);
		offset += ret;
		index += offset >> PAGE_CACHE_SHIFT;
		offset &= ~PAGE_CACHE_MASK;

		if (ret == nr && desc->count)
			continue;
		goto out;

no_xip_page:
		/* Did not get the page. Report it */
		desc->error = -EIO;
		goto out;
	}

out:
	*ppos = ((loff_t) index << PAGE_CACHE_SHIFT) + offset;
	if (filp)
		file_accessed(filp);
}

ssize_t
xip_file_read(struct file *filp, char __user *buf, size_t len, loff_t *ppos)
{
	read_descriptor_t desc;

	if (!access_ok(VERIFY_WRITE, buf, len))
		return -EFAULT;

	desc.written = 0;
	desc.arg.buf = buf;
	desc.count = len;
	desc.error = 0;

	do_xip_mapping_read(filp->f_mapping, &filp->f_ra, filp,
			    ppos, &desc, file_read_actor);

	if (desc.written)
		return desc.written;
	else
		return desc.error;
}
EXPORT_SYMBOL_GPL(xip_file_read);

ssize_t
xip_file_sendfile(struct file *in_file, loff_t *ppos,
	     size_t count, read_actor_t actor, void *target)
{
	read_descriptor_t desc;

	if (!count)
		return 0;

	desc.written = 0;
	desc.count = count;
	desc.arg.data = target;
	desc.error = 0;

	do_xip_mapping_read(in_file->f_mapping, &in_file->f_ra, in_file,
			    ppos, &desc, actor);
	if (desc.written)
		return desc.written;
	return desc.error;
}
EXPORT_SYMBOL_GPL(xip_file_sendfile);

/*
 * __xip_unmap is invoked from xip_unmap and
 * xip_write
 *
 * This function walks all vmas of the address_space and unmaps the
 * ZERO_PAGE when found at pgoff. Should it go in rmap.c?
 */
static void
__xip_unmap (struct address_space * mapping,
		     unsigned long pgoff)
{
	struct vm_area_struct *vma;
	struct mm_struct *mm;
	struct prio_tree_iter iter;
	unsigned long address;
	pte_t *pte;
	pte_t pteval;
	spinlock_t *ptl;
	struct page *page;

	spin_lock(&mapping->i_mmap_lock);
	vma_prio_tree_foreach(vma, &iter, &mapping->i_mmap, pgoff, pgoff) {
		mm = vma->vm_mm;
		address = vma->vm_start +
			((pgoff - vma->vm_pgoff) << PAGE_SHIFT);
		BUG_ON(address < vma->vm_start || address >= vma->vm_end);
		page = ZERO_PAGE(address);
		pte = page_check_address(page, mm, address, &ptl);
		if (pte) {
			/* Nuke the page table entry. */
			flush_cache_page(vma, address, pte_pfn(*pte));
			pteval = ptep_clear_flush(vma, address, pte);
			page_remove_rmap(page);
			dec_mm_counter(mm, file_rss);
			BUG_ON(pte_dirty(pteval));
			pte_unmap_unlock(pte, ptl);
			page_cache_release(page);
		}
	}
	spin_unlock(&mapping->i_mmap_lock);
}

/*
 * xip_nopage() is invoked via the vma operations vector for a
 * mapped memory region to read in file data during a page fault.
 *
 * This function is derived from filemap_nopage, but used for execute in place
 */
static struct page *
xip_file_nopage(struct vm_area_struct * area,
		   unsigned long address,
		   int *type)
{
	struct file *file = area->vm_file;
	struct address_space *mapping = file->f_mapping;
	struct inode *inode = mapping->host;
	struct page *page;
	unsigned long size, pgoff, endoff;

	pgoff = ((address - area->vm_start) >> PAGE_CACHE_SHIFT)
		+ area->vm_pgoff;
	endoff = ((area->vm_end - area->vm_start) >> PAGE_CACHE_SHIFT)
		+ area->vm_pgoff;

	size = (i_size_read(inode) + PAGE_CACHE_SIZE - 1) >> PAGE_CACHE_SHIFT;
	if (pgoff >= size) {
		return NULL;
	}

	page = mapping->a_ops->get_xip_page(mapping, pgoff*(PAGE_SIZE/512), 0);
	if (!IS_ERR(page)) {
		goto out;
	}
	if (PTR_ERR(page) != -ENODATA)
		return NULL;

	/* sparse block */
	if ((area->vm_flags & (VM_WRITE | VM_MAYWRITE)) &&
	    (area->vm_flags & (VM_SHARED| VM_MAYSHARE)) &&
	    (!(mapping->host->i_sb->s_flags & MS_RDONLY))) {
		/* maybe shared writable, allocate new block */
		page = mapping->a_ops->get_xip_page (mapping,
			pgoff*(PAGE_SIZE/512), 1);
		if (IS_ERR(page))
			return NULL;
		/* unmap page at pgoff from all other vmas */
		__xip_unmap(mapping, pgoff);
	} else {
		/* not shared and writable, use ZERO_PAGE() */
		page = ZERO_PAGE(address);
	}

out:
	page_cache_get(page);
	return page;
}

static struct vm_operations_struct xip_file_vm_ops = {
	.nopage         = xip_file_nopage,
};

int xip_file_mmap(struct file * file, struct vm_area_struct * vma)
{
	BUG_ON(!file->f_mapping->a_ops->get_xip_page);

	file_accessed(file);
	vma->vm_ops = &xip_file_vm_ops;
	return 0;
}
EXPORT_SYMBOL_GPL(xip_file_mmap);

static ssize_t
__xip_file_write(struct file *filp, const char __user *buf,
		  size_t count, loff_t pos, loff_t *ppos)
{
	struct address_space * mapping = filp->f_mapping;
	struct address_space_operations *a_ops = mapping->a_ops;
	struct inode 	*inode = mapping->host;
	long		status = 0;
	struct page	*page;
	size_t		bytes;
	ssize_t		written = 0;

	BUG_ON(!mapping->a_ops->get_xip_page);

	do {
		unsigned long index;
		unsigned long offset;
		size_t copied;

		offset = (pos & (PAGE_CACHE_SIZE -1)); /* Within page */
		index = pos >> PAGE_CACHE_SHIFT;
		bytes = PAGE_CACHE_SIZE - offset;
		if (bytes > count)
			bytes = count;

		/*
		 * Bring in the user page that we will copy from _first_.
		 * Otherwise there's a nasty deadlock on copying from the
		 * same page as we're writing to, without it being marked
		 * up-to-date.
		 */
		fault_in_pages_readable(buf, bytes);

		page = a_ops->get_xip_page(mapping,
					   index*(PAGE_SIZE/512), 0);
		if (IS_ERR(page) && (PTR_ERR(page) == -ENODATA)) {
			/* we allocate a new page unmap it */
			page = a_ops->get_xip_page(mapping,
						   index*(PAGE_SIZE/512), 1);
			if (!IS_ERR(page))
				/* unmap page at pgoff from all other vmas */
				__xip_unmap(mapping, index);
		}

		if (IS_ERR(page)) {
			status = PTR_ERR(page);
			break;
		}

		copied = filemap_copy_from_user(page, offset, buf, bytes);
		flush_dcache_page(page);
		if (likely(copied > 0)) {
			status = copied;

			if (status >= 0) {
				written += status;
				count -= status;
				pos += status;
				buf += status;
			}
		}
		if (unlikely(copied != bytes))
			if (status >= 0)
				status = -EFAULT;
		if (status < 0)
			break;
	} while (count);
	*ppos = pos;
	/*
	 * No need to use i_size_read() here, the i_size
	 * cannot change under us because we hold i_sem.
	 */
	if (pos > inode->i_size) {
		i_size_write(inode, pos);
		mark_inode_dirty(inode);
	}

	return written ? written : status;
}

ssize_t
xip_file_write(struct file *filp, const char __user *buf, size_t len,
	       loff_t *ppos)
{
	struct address_space *mapping = filp->f_mapping;
	struct inode *inode = mapping->host;
	size_t count;
	loff_t pos;
	ssize_t ret;

	down(&inode->i_sem);

	if (!access_ok(VERIFY_READ, buf, len)) {
		ret=-EFAULT;
		goto out_up;
	}

	pos = *ppos;
	count = len;

	vfs_check_frozen(inode->i_sb, SB_FREEZE_WRITE);

	/* We can write back this queue in page reclaim */
	current->backing_dev_info = mapping->backing_dev_info;

	ret = generic_write_checks(filp, &pos, &count, S_ISBLK(inode->i_mode));
	if (ret)
		goto out_backing;
	if (count == 0)
		goto out_backing;

	ret = remove_suid(filp->f_dentry);
	if (ret)
		goto out_backing;

	inode_update_time(inode, 1);

	ret = __xip_file_write (filp, buf, count, pos, ppos);

 out_backing:
	current->backing_dev_info = NULL;
 out_up:
	up(&inode->i_sem);
	return ret;
}
EXPORT_SYMBOL_GPL(xip_file_write);

/*
 * truncate a page used for execute in place
 * functionality is analog to block_truncate_page but does use get_xip_page
 * to get the page instead of page cache
 */
int
xip_truncate_page(struct address_space *mapping, loff_t from)
{
	pgoff_t index = from >> PAGE_CACHE_SHIFT;
	unsigned offset = from & (PAGE_CACHE_SIZE-1);
	unsigned blocksize;
	unsigned length;
	struct page *page;
	void *kaddr;

	BUG_ON(!mapping->a_ops->get_xip_page);

	blocksize = 1 << mapping->host->i_blkbits;
	length = offset & (blocksize - 1);

	/* Block boundary? Nothing to do */
	if (!length)
		return 0;

	length = blocksize - length;

	page = mapping->a_ops->get_xip_page(mapping,
					    index*(PAGE_SIZE/512), 0);
	if (!page)
		return -ENOMEM;
	if (unlikely(IS_ERR(page))) {
		if (PTR_ERR(page) == -ENODATA)
			/* Hole? No need to truncate */
			return 0;
		else
			return PTR_ERR(page);
	}
	kaddr = kmap_atomic(page, KM_USER0);
	memset(kaddr + offset, 0, length);
	kunmap_atomic(kaddr, KM_USER0);

	flush_dcache_page(page);
	return 0;
}
EXPORT_SYMBOL_GPL(xip_truncate_page);
Commit	Line	Data
ceffc078 CO	1	/*
	2	* linux/mm/filemap_xip.c
	3	*
	4	* Copyright (C) 2005 IBM Corporation
	5	* Author: Carsten Otte <[email protected]>
	6	*
	7	* derived from linux/mm/filemap.c - Copyright (C) Linus Torvalds
	8	*
	9	*/
	10
	11	#include <linux/fs.h>
	12	#include <linux/pagemap.h>
	13	#include <linux/module.h>
	14	#include <linux/uio.h>
	15	#include <linux/rmap.h>
	16	#include <asm/tlbflush.h>
	17	#include "filemap.h"
	18
	19	/*
	20	* This is a file read routine for execute in place files, and uses
	21	* the mapping->a_ops->get_xip_page() function for the actual low-level
	22	* stuff.
	23	*
	24	* Note the struct file* is not used at all. It may be NULL.
	25	*/
	26	static void
	27	do_xip_mapping_read(struct address_space *mapping,
	28	struct file_ra_state *_ra,
	29	struct file *filp,
	30	loff_t *ppos,
	31	read_descriptor_t *desc,
	32	read_actor_t actor)
	33	{
	34	struct inode *inode = mapping->host;
	35	unsigned long index, end_index, offset;
	36	loff_t isize;
	37
	38	BUG_ON(!mapping->a_ops->get_xip_page);
	39
	40	index = *ppos >> PAGE_CACHE_SHIFT;
	41	offset = *ppos & ~PAGE_CACHE_MASK;
	42
	43	isize = i_size_read(inode);
	44	if (!isize)
	45	goto out;
	46
	47	end_index = (isize - 1) >> PAGE_CACHE_SHIFT;
	48	for (;;) {
	49	struct page *page;
	50	unsigned long nr, ret;
	51
	52	/* nr is the maximum number of bytes to copy from this page */
	53	nr = PAGE_CACHE_SIZE;
	54	if (index >= end_index) {
	55	if (index > end_index)
	56	goto out;
	57	nr = ((isize - 1) & ~PAGE_CACHE_MASK) + 1;
	58	if (nr <= offset) {
	59	goto out;
	60	}
	61	}
	62	nr = nr - offset;
	63
	64	page = mapping->a_ops->get_xip_page(mapping,
65	index*(PAGE_SIZE/512), 0);
66	if (!page)
67	goto no_xip_page;
68	if (unlikely(IS_ERR(page))) {
69	if (PTR_ERR(page) == -ENODATA) {
70	/* sparse */
afa597ba	71	page = ZERO_PAGE(0);
ceffc078 CO	72	} else {
	73	desc->error = PTR_ERR(page);
	74	goto out;
	75	}
afa597ba	76	}
ceffc078 CO	77
	78	/* If users can be writing to this page using arbitrary
	79	* virtual addresses, take care about potential aliasing
	80	* before reading the page on the kernel side.
	81	*/
	82	if (mapping_writably_mapped(mapping))
	83	flush_dcache_page(page);
	84
	85	/*
afa597ba	86	* Ok, we have the page, so now we can copy it to user space...
ceffc078 CO	87	*
	88	* The actor routine returns how many bytes were actually used..
	89	* NOTE! This may not be the same as how much of a user buffer
	90	* we filled up (we may be padding etc), so we can only update
	91	* "pos" here (the actor routine has to update the user buffer
	92	* pointers and the remaining count).
	93	*/
	94	ret = actor(desc, page, offset, nr);
	95	offset += ret;
	96	index += offset >> PAGE_CACHE_SHIFT;
	97	offset &= ~PAGE_CACHE_MASK;
	98
	99	if (ret == nr && desc->count)
	100	continue;
	101	goto out;
	102
	103	no_xip_page:
	104	/* Did not get the page. Report it */
	105	desc->error = -EIO;
	106	goto out;
	107	}
	108
	109	out:
	110	*ppos = ((loff_t) index << PAGE_CACHE_SHIFT) + offset;
	111	if (filp)
	112	file_accessed(filp);
	113	}
	114
ceffc078	115	ssize_t
eb6fe0c3	116	xip_file_read(struct file filp, char __user buf, size_t len, loff_t *ppos)
ceffc078	117	{
eb6fe0c3	118	read_descriptor_t desc;
ceffc078	119
eb6fe0c3 CO	120	if (!access_ok(VERIFY_WRITE, buf, len))
eb6fe0c3 CO	121	return -EFAULT;
ceffc078	122
eb6fe0c3 CO	123	desc.written = 0;
	124	desc.arg.buf = buf;
	125	desc.count = len;
	126	desc.error = 0;
ceffc078	127
eb6fe0c3 CO	128	do_xip_mapping_read(filp->f_mapping, &filp->f_ra, filp,
	129	ppos, &desc, file_read_actor);
	130
	131	if (desc.written)
	132	return desc.written;
	133	else
	134	return desc.error;
ceffc078	135	}
eb6fe0c3	136	EXPORT_SYMBOL_GPL(xip_file_read);
ceffc078 CO	137
	138	ssize_t
	139	xip_file_sendfile(struct file in_file, loff_t ppos,
	140	size_t count, read_actor_t actor, void *target)
	141	{
	142	read_descriptor_t desc;
	143
	144	if (!count)
	145	return 0;
	146
	147	desc.written = 0;
	148	desc.count = count;
	149	desc.arg.data = target;
	150	desc.error = 0;
	151
	152	do_xip_mapping_read(in_file->f_mapping, &in_file->f_ra, in_file,
	153	ppos, &desc, actor);
	154	if (desc.written)
	155	return desc.written;
	156	return desc.error;
	157	}
	158	EXPORT_SYMBOL_GPL(xip_file_sendfile);
	159
	160	/*
	161	* __xip_unmap is invoked from xip_unmap and
	162	* xip_write
	163	*
	164	* This function walks all vmas of the address_space and unmaps the
afa597ba	165	* ZERO_PAGE when found at pgoff. Should it go in rmap.c?
ceffc078 CO	166	*/
	167	static void
	168	__xip_unmap (struct address_space * mapping,
	169	unsigned long pgoff)
	170	{
	171	struct vm_area_struct *vma;
	172	struct mm_struct *mm;
	173	struct prio_tree_iter iter;
	174	unsigned long address;
	175	pte_t *pte;
	176	pte_t pteval;
c0718806	177	spinlock_t *ptl;
67b02f11	178	struct page *page;
ceffc078 CO	179
	180	spin_lock(&mapping->i_mmap_lock);
	181	vma_prio_tree_foreach(vma, &iter, &mapping->i_mmap, pgoff, pgoff) {
	182	mm = vma->vm_mm;
	183	address = vma->vm_start +
	184	((pgoff - vma->vm_pgoff) << PAGE_SHIFT);
	185	BUG_ON(address < vma->vm_start \|\| address >= vma->vm_end);
67b02f11	186	page = ZERO_PAGE(address);
c0718806 HD	187	pte = page_check_address(page, mm, address, &ptl);
c0718806 HD	188	if (pte) {
ceffc078	189	/* Nuke the page table entry. */
082ff0a9	190	flush_cache_page(vma, address, pte_pfn(*pte));
ceffc078	191	pteval = ptep_clear_flush(vma, address, pte);
b5810039 NP	192	page_remove_rmap(page);
b5810039 NP	193	dec_mm_counter(mm, file_rss);
ceffc078	194	BUG_ON(pte_dirty(pteval));
c0718806	195	pte_unmap_unlock(pte, ptl);
b5810039	196	page_cache_release(page);
ceffc078 CO	197	}
	198	}
	199	spin_unlock(&mapping->i_mmap_lock);
	200	}
	201
	202	/*
	203	* xip_nopage() is invoked via the vma operations vector for a
	204	* mapped memory region to read in file data during a page fault.
	205	*
	206	* This function is derived from filemap_nopage, but used for execute in place
	207	*/
	208	static struct page *
	209	xip_file_nopage(struct vm_area_struct * area,
	210	unsigned long address,
	211	int *type)
	212	{
	213	struct file *file = area->vm_file;
	214	struct address_space *mapping = file->f_mapping;
	215	struct inode *inode = mapping->host;
	216	struct page *page;
	217	unsigned long size, pgoff, endoff;
	218
	219	pgoff = ((address - area->vm_start) >> PAGE_CACHE_SHIFT)
	220	+ area->vm_pgoff;
	221	endoff = ((area->vm_end - area->vm_start) >> PAGE_CACHE_SHIFT)
	222	+ area->vm_pgoff;
	223
	224	size = (i_size_read(inode) + PAGE_CACHE_SIZE - 1) >> PAGE_CACHE_SHIFT;
	225	if (pgoff >= size) {
	226	return NULL;
	227	}
	228
	229	page = mapping->a_ops->get_xip_page(mapping, pgoff*(PAGE_SIZE/512), 0);
	230	if (!IS_ERR(page)) {
b5810039	231	goto out;
ceffc078 CO	232	}
	233	if (PTR_ERR(page) != -ENODATA)
	234	return NULL;
	235
	236	/* sparse block */
	237	if ((area->vm_flags & (VM_WRITE \| VM_MAYWRITE)) &&
	238	(area->vm_flags & (VM_SHARED\| VM_MAYSHARE)) &&
	239	(!(mapping->host->i_sb->s_flags & MS_RDONLY))) {
	240	/* maybe shared writable, allocate new block */
	241	page = mapping->a_ops->get_xip_page (mapping,
	242	pgoff*(PAGE_SIZE/512), 1);
	243	if (IS_ERR(page))
	244	return NULL;
ceffc078 CO	245	/* unmap page at pgoff from all other vmas */
	246	__xip_unmap(mapping, pgoff);
	247	} else {
afa597ba CO	248	/* not shared and writable, use ZERO_PAGE() */
afa597ba CO	249	page = ZERO_PAGE(address);
ceffc078 CO	250	}
ceffc078 CO	251
b5810039 NP	252	out:
b5810039 NP	253	page_cache_get(page);
ceffc078 CO	254	return page;
	255	}
	256
	257	static struct vm_operations_struct xip_file_vm_ops = {
	258	.nopage = xip_file_nopage,
	259	};
	260
	261	int xip_file_mmap(struct file * file, struct vm_area_struct * vma)
	262	{
	263	BUG_ON(!file->f_mapping->a_ops->get_xip_page);
	264
	265	file_accessed(file);
	266	vma->vm_ops = &xip_file_vm_ops;
	267	return 0;
	268	}
	269	EXPORT_SYMBOL_GPL(xip_file_mmap);
	270
	271	static ssize_t
eb6fe0c3 CO	272	__xip_file_write(struct file filp, const char __user buf,
eb6fe0c3 CO	273	size_t count, loff_t pos, loff_t *ppos)
ceffc078	274	{
eb6fe0c3	275	struct address_space * mapping = filp->f_mapping;
ceffc078 CO	276	struct address_space_operations *a_ops = mapping->a_ops;
	277	struct inode *inode = mapping->host;
	278	long status = 0;
	279	struct page *page;
	280	size_t bytes;
ceffc078 CO	281	ssize_t written = 0;
	282
	283	BUG_ON(!mapping->a_ops->get_xip_page);
	284
ceffc078 CO	285	do {
	286	unsigned long index;
	287	unsigned long offset;
	288	size_t copied;
	289
	290	offset = (pos & (PAGE_CACHE_SIZE -1)); /* Within page */
	291	index = pos >> PAGE_CACHE_SHIFT;
	292	bytes = PAGE_CACHE_SIZE - offset;
	293	if (bytes > count)
	294	bytes = count;
	295
	296	/*
	297	* Bring in the user page that we will copy from _first_.
	298	* Otherwise there's a nasty deadlock on copying from the
	299	* same page as we're writing to, without it being marked
	300	* up-to-date.
	301	*/
	302	fault_in_pages_readable(buf, bytes);
	303
	304	page = a_ops->get_xip_page(mapping,
eb6fe0c3	305	index*(PAGE_SIZE/512), 0);
ceffc078 CO	306	if (IS_ERR(page) && (PTR_ERR(page) == -ENODATA)) {
	307	/* we allocate a new page unmap it */
	308	page = a_ops->get_xip_page(mapping,
eb6fe0c3	309	index*(PAGE_SIZE/512), 1);
ceffc078	310	if (!IS_ERR(page))
eb6fe0c3 CO	311	/* unmap page at pgoff from all other vmas */
eb6fe0c3 CO	312	__xip_unmap(mapping, index);
ceffc078 CO	313	}
	314
	315	if (IS_ERR(page)) {
	316	status = PTR_ERR(page);
	317	break;
	318	}
	319
eb6fe0c3	320	copied = filemap_copy_from_user(page, offset, buf, bytes);
ceffc078 CO	321	flush_dcache_page(page);
	322	if (likely(copied > 0)) {
	323	status = copied;
	324
	325	if (status >= 0) {
	326	written += status;
	327	count -= status;
	328	pos += status;
	329	buf += status;
ceffc078 CO	330	}
	331	}
	332	if (unlikely(copied != bytes))
	333	if (status >= 0)
	334	status = -EFAULT;
	335	if (status < 0)
	336	break;
	337	} while (count);
	338	*ppos = pos;
	339	/*
	340	* No need to use i_size_read() here, the i_size
	341	* cannot change under us because we hold i_sem.
	342	*/
	343	if (pos > inode->i_size) {
	344	i_size_write(inode, pos);
	345	mark_inode_dirty(inode);
	346	}
	347
	348	return written ? written : status;
	349	}
	350
eb6fe0c3 CO	351	ssize_t
	352	xip_file_write(struct file filp, const char __user buf, size_t len,
	353	loff_t *ppos)
ceffc078	354	{
eb6fe0c3 CO	355	struct address_space *mapping = filp->f_mapping;
	356	struct inode *inode = mapping->host;
	357	size_t count;
	358	loff_t pos;
	359	ssize_t ret;
ceffc078	360
eb6fe0c3	361	down(&inode->i_sem);
ceffc078	362
eb6fe0c3 CO	363	if (!access_ok(VERIFY_READ, buf, len)) {
	364	ret=-EFAULT;
	365	goto out_up;
ceffc078 CO	366	}
ceffc078 CO	367
ceffc078	368	pos = *ppos;
eb6fe0c3	369	count = len;
ceffc078 CO	370
	371	vfs_check_frozen(inode->i_sb, SB_FREEZE_WRITE);
	372
eb6fe0c3 CO	373	/* We can write back this queue in page reclaim */
eb6fe0c3 CO	374	current->backing_dev_info = mapping->backing_dev_info;
ceffc078	375
eb6fe0c3 CO	376	ret = generic_write_checks(filp, &pos, &count, S_ISBLK(inode->i_mode));
	377	if (ret)
	378	goto out_backing;
ceffc078	379	if (count == 0)
eb6fe0c3	380	goto out_backing;
ceffc078	381
eb6fe0c3 CO	382	ret = remove_suid(filp->f_dentry);
	383	if (ret)
	384	goto out_backing;
ceffc078 CO	385
	386	inode_update_time(inode, 1);
	387
eb6fe0c3	388	ret = __xip_file_write (filp, buf, count, pos, ppos);
ceffc078	389
eb6fe0c3 CO	390	out_backing:
	391	current->backing_dev_info = NULL;
	392	out_up:
ceffc078 CO	393	up(&inode->i_sem);
	394	return ret;
	395	}
eb6fe0c3	396	EXPORT_SYMBOL_GPL(xip_file_write);
ceffc078 CO	397
	398	/*
	399	* truncate a page used for execute in place
	400	* functionality is analog to block_truncate_page but does use get_xip_page
	401	* to get the page instead of page cache
	402	*/
	403	int
	404	xip_truncate_page(struct address_space *mapping, loff_t from)
	405	{
	406	pgoff_t index = from >> PAGE_CACHE_SHIFT;
	407	unsigned offset = from & (PAGE_CACHE_SIZE-1);
	408	unsigned blocksize;
	409	unsigned length;
	410	struct page *page;
	411	void *kaddr;
ceffc078 CO	412
	413	BUG_ON(!mapping->a_ops->get_xip_page);
	414
	415	blocksize = 1 << mapping->host->i_blkbits;
	416	length = offset & (blocksize - 1);
	417
	418	/* Block boundary? Nothing to do */
	419	if (!length)
	420	return 0;
	421
	422	length = blocksize - length;
	423
	424	page = mapping->a_ops->get_xip_page(mapping,
	425	index*(PAGE_SIZE/512), 0);
ceffc078	426	if (!page)
eb6fe0c3	427	return -ENOMEM;
ceffc078	428	if (unlikely(IS_ERR(page))) {
eb6fe0c3	429	if (PTR_ERR(page) == -ENODATA)
ceffc078 CO	430	/* Hole? No need to truncate */
ceffc078 CO	431	return 0;
eb6fe0c3 CO	432	else
eb6fe0c3 CO	433	return PTR_ERR(page);
afa597ba	434	}
ceffc078 CO	435	kaddr = kmap_atomic(page, KM_USER0);
	436	memset(kaddr + offset, 0, length);
	437	kunmap_atomic(kaddr, KM_USER0);
	438
	439	flush_dcache_page(page);
eb6fe0c3	440	return 0;
ceffc078 CO	441	}
ceffc078 CO	442	EXPORT_SYMBOL_GPL(xip_truncate_page);