[linux-2.6-block.git] / mm / fadvise.c

/*
 * mm/fadvise.c
 *
 * Copyright (C) 2002, Linus Torvalds
 *
 * 11Jan2003	akpm@digeo.com
 *		Initial version.
 */

#include <linux/kernel.h>
#include <linux/file.h>
#include <linux/fs.h>
#include <linux/mm.h>
#include <linux/pagemap.h>
#include <linux/backing-dev.h>
#include <linux/pagevec.h>
#include <linux/fadvise.h>
#include <linux/writeback.h>
#include <linux/syscalls.h>

#include <asm/unistd.h>

/*
 * POSIX_FADV_WILLNEED could set PG_Referenced, and POSIX_FADV_NOREUSE could
 * deactivate the pages and clear PG_Referenced.
 *
 * LINUX_FADV_ASYNC_WRITE: start async writeout of any dirty pages between file
 * offsets `offset' and `offset+len' inclusive.  Any pages which are currently
 * under writeout are skipped, whether or not they are dirty.
 *
 * LINUX_FADV_WRITE_WAIT: wait upon writeout of any dirty pages between file
 * offsets `offset' and `offset+len'.
 *
 * By combining these two operations the application may do several things:
 *
 * LINUX_FADV_ASYNC_WRITE: push some or all of the dirty pages at the disk.
 *
 * LINUX_FADV_WRITE_WAIT, LINUX_FADV_ASYNC_WRITE: push all of the currently
 * dirty pages at the disk.
 *
 * LINUX_FADV_WRITE_WAIT, LINUX_FADV_ASYNC_WRITE, LINUX_FADV_WRITE_WAIT: push
 * all of the currently dirty pages at the disk, wait until they have been
 * written.
 *
 * It should be noted that none of these operations write out the file's
 * metadata.  So unless the application is strictly performing overwrites of
 * already-instantiated disk blocks, there are no guarantees here that the data
 * will be available after a crash.
 */
asmlinkage long sys_fadvise64_64(int fd, loff_t offset, loff_t len, int advice)
{
	struct file *file = fget(fd);
	struct address_space *mapping;
	struct backing_dev_info *bdi;
	loff_t endbyte;			/* inclusive */
	pgoff_t start_index;
	pgoff_t end_index;
	unsigned long nrpages;
	int ret = 0;

	if (!file)
		return -EBADF;

	if (S_ISFIFO(file->f_dentry->d_inode->i_mode)) {
		ret = -ESPIPE;
		goto out;
	}

	mapping = file->f_mapping;
	if (!mapping || len < 0) {
		ret = -EINVAL;
		goto out;
	}

	if (mapping->a_ops->get_xip_page)
		/* no bad return value, but ignore advice */
		goto out;

	/* Careful about overflows. Len == 0 means "as much as possible" */
	endbyte = offset + len;
	if (!len || endbyte < len)
		endbyte = -1;
	else
		endbyte--;		/* inclusive */

	bdi = mapping->backing_dev_info;

	switch (advice) {
	case POSIX_FADV_NORMAL:
		file->f_ra.ra_pages = bdi->ra_pages;
		break;
	case POSIX_FADV_RANDOM:
		file->f_ra.ra_pages = 0;
		break;
	case POSIX_FADV_SEQUENTIAL:
		file->f_ra.ra_pages = bdi->ra_pages * 2;
		break;
	case POSIX_FADV_WILLNEED:
	case POSIX_FADV_NOREUSE:
		if (!mapping->a_ops->readpage) {
			ret = -EINVAL;
			break;
		}

		/* First and last PARTIAL page! */
		start_index = offset >> PAGE_CACHE_SHIFT;
		end_index = endbyte >> PAGE_CACHE_SHIFT;

		/* Careful about overflow on the "+1" */
		nrpages = end_index - start_index + 1;
		if (!nrpages)
			nrpages = ~0UL;
		
		ret = force_page_cache_readahead(mapping, file,
				start_index,
				max_sane_readahead(nrpages));
		if (ret > 0)
			ret = 0;
		break;
	case POSIX_FADV_DONTNEED:
		if (!bdi_write_congested(mapping->backing_dev_info))
			filemap_flush(mapping);

		/* First and last FULL page! */
		start_index = (offset+(PAGE_CACHE_SIZE-1)) >> PAGE_CACHE_SHIFT;
		end_index = (endbyte >> PAGE_CACHE_SHIFT);

		if (end_index >= start_index)
			invalidate_mapping_pages(mapping, start_index,
						end_index);
		break;
	case LINUX_FADV_ASYNC_WRITE:
		ret = __filemap_fdatawrite_range(mapping, offset, endbyte,
						WB_SYNC_NONE);
		break;
	case LINUX_FADV_WRITE_WAIT:
		ret = wait_on_page_writeback_range(mapping,
					offset >> PAGE_CACHE_SHIFT,
					endbyte >> PAGE_CACHE_SHIFT);
		break;
	default:
		ret = -EINVAL;
	}
out:
	fput(file);
	return ret;
}

#ifdef __ARCH_WANT_SYS_FADVISE64

asmlinkage long sys_fadvise64(int fd, loff_t offset, size_t len, int advice)
{
	return sys_fadvise64_64(fd, offset, len, advice);
}

#endif
Commit	Line	Data
1da177e4 LT	1	/*
	2	* mm/fadvise.c
	3	*
	4	* Copyright (C) 2002, Linus Torvalds
	5	*
	6	* 11Jan2003 akpm@digeo.com
	7	* Initial version.
	8	*/
	9
	10	#include <linux/kernel.h>
	11	#include <linux/file.h>
	12	#include <linux/fs.h>
	13	#include <linux/mm.h>
	14	#include <linux/pagemap.h>
	15	#include <linux/backing-dev.h>
	16	#include <linux/pagevec.h>
	17	#include <linux/fadvise.h>
ebcf28e1	18	#include <linux/writeback.h>
1da177e4 LT	19	#include <linux/syscalls.h>
	20
	21	#include <asm/unistd.h>
	22
	23	/*
	24	* POSIX_FADV_WILLNEED could set PG_Referenced, and POSIX_FADV_NOREUSE could
	25	* deactivate the pages and clear PG_Referenced.
ebcf28e1 AM	26	*
	27	* LINUX_FADV_ASYNC_WRITE: start async writeout of any dirty pages between file
	28	* offsets `offset' and `offset+len' inclusive. Any pages which are currently
	29	* under writeout are skipped, whether or not they are dirty.
	30	*
	31	* LINUX_FADV_WRITE_WAIT: wait upon writeout of any dirty pages between file
	32	* offsets `offset' and `offset+len'.
	33	*
	34	* By combining these two operations the application may do several things:
	35	*
	36	* LINUX_FADV_ASYNC_WRITE: push some or all of the dirty pages at the disk.
	37	*
	38	* LINUX_FADV_WRITE_WAIT, LINUX_FADV_ASYNC_WRITE: push all of the currently
	39	* dirty pages at the disk.
	40	*
	41	* LINUX_FADV_WRITE_WAIT, LINUX_FADV_ASYNC_WRITE, LINUX_FADV_WRITE_WAIT: push
	42	* all of the currently dirty pages at the disk, wait until they have been
	43	* written.
	44	*
	45	* It should be noted that none of these operations write out the file's
	46	* metadata. So unless the application is strictly performing overwrites of
	47	* already-instantiated disk blocks, there are no guarantees here that the data
	48	* will be available after a crash.
1da177e4 LT	49	*/
	50	asmlinkage long sys_fadvise64_64(int fd, loff_t offset, loff_t len, int advice)
	51	{
	52	struct file *file = fget(fd);
	53	struct address_space *mapping;
	54	struct backing_dev_info *bdi;
ebcf28e1	55	loff_t endbyte; /* inclusive */
1da177e4 LT	56	pgoff_t start_index;
	57	pgoff_t end_index;
	58	unsigned long nrpages;
	59	int ret = 0;
	60
	61	if (!file)
	62	return -EBADF;
	63
87ba81db VB	64	if (S_ISFIFO(file->f_dentry->d_inode->i_mode)) {
	65	ret = -ESPIPE;
	66	goto out;
	67	}
	68
1da177e4 LT	69	mapping = file->f_mapping;
	70	if (!mapping \|\| len < 0) {
	71	ret = -EINVAL;
	72	goto out;
	73	}
	74
fe77ba6f CO	75	if (mapping->a_ops->get_xip_page)
	76	/* no bad return value, but ignore advice */
	77	goto out;
	78
1da177e4 LT	79	/* Careful about overflows. Len == 0 means "as much as possible" */
	80	endbyte = offset + len;
	81	if (!len \|\| endbyte < len)
	82	endbyte = -1;
ebcf28e1 AM	83	else
ebcf28e1 AM	84	endbyte--; /* inclusive */
1da177e4 LT	85
	86	bdi = mapping->backing_dev_info;
	87
	88	switch (advice) {
	89	case POSIX_FADV_NORMAL:
	90	file->f_ra.ra_pages = bdi->ra_pages;
	91	break;
	92	case POSIX_FADV_RANDOM:
	93	file->f_ra.ra_pages = 0;
	94	break;
	95	case POSIX_FADV_SEQUENTIAL:
	96	file->f_ra.ra_pages = bdi->ra_pages * 2;
	97	break;
	98	case POSIX_FADV_WILLNEED:
	99	case POSIX_FADV_NOREUSE:
	100	if (!mapping->a_ops->readpage) {
	101	ret = -EINVAL;
	102	break;
	103	}
	104
	105	/* First and last PARTIAL page! */
	106	start_index = offset >> PAGE_CACHE_SHIFT;
ebcf28e1	107	end_index = endbyte >> PAGE_CACHE_SHIFT;
1da177e4 LT	108
	109	/* Careful about overflow on the "+1" */
	110	nrpages = end_index - start_index + 1;
	111	if (!nrpages)
	112	nrpages = ~0UL;
	113
	114	ret = force_page_cache_readahead(mapping, file,
	115	start_index,
	116	max_sane_readahead(nrpages));
	117	if (ret > 0)
	118	ret = 0;
	119	break;
	120	case POSIX_FADV_DONTNEED:
	121	if (!bdi_write_congested(mapping->backing_dev_info))
	122	filemap_flush(mapping);
	123
	124	/* First and last FULL page! */
ebcf28e1	125	start_index = (offset+(PAGE_CACHE_SIZE-1)) >> PAGE_CACHE_SHIFT;
1da177e4 LT	126	end_index = (endbyte >> PAGE_CACHE_SHIFT);
1da177e4 LT	127
ebcf28e1 AM	128	if (end_index >= start_index)
	129	invalidate_mapping_pages(mapping, start_index,
	130	end_index);
	131	break;
	132	case LINUX_FADV_ASYNC_WRITE:
	133	ret = __filemap_fdatawrite_range(mapping, offset, endbyte,
	134	WB_SYNC_NONE);
	135	break;
	136	case LINUX_FADV_WRITE_WAIT:
	137	ret = wait_on_page_writeback_range(mapping,
	138	offset >> PAGE_CACHE_SHIFT,
	139	endbyte >> PAGE_CACHE_SHIFT);
1da177e4 LT	140	break;
	141	default:
	142	ret = -EINVAL;
	143	}
	144	out:
	145	fput(file);
	146	return ret;
	147	}
	148
	149	#ifdef __ARCH_WANT_SYS_FADVISE64
	150
	151	asmlinkage long sys_fadvise64(int fd, loff_t offset, size_t len, int advice)
	152	{
	153	return sys_fadvise64_64(fd, offset, len, advice);
	154	}
	155
	156	#endif