fadvise.c 4.87 KB
Newer Older
1
// SPDX-License-Identifier: GPL-2.0
Linus Torvalds's avatar
Linus Torvalds committed
2 3 4 5 6
/*
 * mm/fadvise.c
 *
 * Copyright (C) 2002, Linus Torvalds
 *
7
 * 11Jan2003	Andrew Morton
Linus Torvalds's avatar
Linus Torvalds committed
8 9 10 11 12 13 14 15 16 17 18
 *		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>
19
#include <linux/writeback.h>
Linus Torvalds's avatar
Linus Torvalds committed
20
#include <linux/syscalls.h>
21
#include <linux/swap.h>
Linus Torvalds's avatar
Linus Torvalds committed
22 23 24 25 26 27 28

#include <asm/unistd.h>

/*
 * POSIX_FADV_WILLNEED could set PG_Referenced, and POSIX_FADV_NOREUSE could
 * deactivate the pages and clear PG_Referenced.
 */
29
SYSCALL_DEFINE4(fadvise64_64, int, fd, loff_t, offset, loff_t, len, int, advice)
Linus Torvalds's avatar
Linus Torvalds committed
30
{
31
	struct fd f = fdget(fd);
Matthew Wilcox's avatar
Matthew Wilcox committed
32
	struct inode *inode;
Linus Torvalds's avatar
Linus Torvalds committed
33 34
	struct address_space *mapping;
	struct backing_dev_info *bdi;
35
	loff_t endbyte;			/* inclusive */
Linus Torvalds's avatar
Linus Torvalds committed
36 37 38 39 40
	pgoff_t start_index;
	pgoff_t end_index;
	unsigned long nrpages;
	int ret = 0;

41
	if (!f.file)
Linus Torvalds's avatar
Linus Torvalds committed
42 43
		return -EBADF;

Matthew Wilcox's avatar
Matthew Wilcox committed
44 45
	inode = file_inode(f.file);
	if (S_ISFIFO(inode->i_mode)) {
46 47 48 49
		ret = -ESPIPE;
		goto out;
	}

50
	mapping = f.file->f_mapping;
Linus Torvalds's avatar
Linus Torvalds committed
51 52 53 54 55
	if (!mapping || len < 0) {
		ret = -EINVAL;
		goto out;
	}

56 57 58
	bdi = inode_to_bdi(mapping->host);

	if (IS_DAX(inode) || (bdi == &noop_backing_dev_info)) {
59 60 61 62 63 64 65 66 67 68 69 70
		switch (advice) {
		case POSIX_FADV_NORMAL:
		case POSIX_FADV_RANDOM:
		case POSIX_FADV_SEQUENTIAL:
		case POSIX_FADV_WILLNEED:
		case POSIX_FADV_NOREUSE:
		case POSIX_FADV_DONTNEED:
			/* no bad return value, but ignore advice */
			break;
		default:
			ret = -EINVAL;
		}
71
		goto out;
72
	}
73

74 75 76 77 78 79
	/*
	 * Careful about overflows. Len == 0 means "as much as possible".  Use
	 * unsigned math because signed overflows are undefined and UBSan
	 * complains.
	 */
	endbyte = (u64)offset + (u64)len;
Linus Torvalds's avatar
Linus Torvalds committed
80 81
	if (!len || endbyte < len)
		endbyte = -1;
82 83
	else
		endbyte--;		/* inclusive */
Linus Torvalds's avatar
Linus Torvalds committed
84 85 86

	switch (advice) {
	case POSIX_FADV_NORMAL:
87 88 89 90
		f.file->f_ra.ra_pages = bdi->ra_pages;
		spin_lock(&f.file->f_lock);
		f.file->f_mode &= ~FMODE_RANDOM;
		spin_unlock(&f.file->f_lock);
Linus Torvalds's avatar
Linus Torvalds committed
91 92
		break;
	case POSIX_FADV_RANDOM:
93 94 95
		spin_lock(&f.file->f_lock);
		f.file->f_mode |= FMODE_RANDOM;
		spin_unlock(&f.file->f_lock);
Linus Torvalds's avatar
Linus Torvalds committed
96 97
		break;
	case POSIX_FADV_SEQUENTIAL:
98 99 100 101
		f.file->f_ra.ra_pages = bdi->ra_pages * 2;
		spin_lock(&f.file->f_lock);
		f.file->f_mode &= ~FMODE_RANDOM;
		spin_unlock(&f.file->f_lock);
Linus Torvalds's avatar
Linus Torvalds committed
102 103 104
		break;
	case POSIX_FADV_WILLNEED:
		/* First and last PARTIAL page! */
105 106
		start_index = offset >> PAGE_SHIFT;
		end_index = endbyte >> PAGE_SHIFT;
Linus Torvalds's avatar
Linus Torvalds committed
107 108 109 110 111

		/* Careful about overflow on the "+1" */
		nrpages = end_index - start_index + 1;
		if (!nrpages)
			nrpages = ~0UL;
112 113 114 115 116

		/*
		 * Ignore return value because fadvise() shall return
		 * success even if filesystem can't retrieve a hint,
		 */
117
		force_page_cache_readahead(mapping, f.file, start_index,
118
					   nrpages);
Linus Torvalds's avatar
Linus Torvalds committed
119
		break;
120 121
	case POSIX_FADV_NOREUSE:
		break;
Linus Torvalds's avatar
Linus Torvalds committed
122
	case POSIX_FADV_DONTNEED:
123
		if (!inode_write_congested(mapping->host))
124 125
			__filemap_fdatawrite_range(mapping, offset, endbyte,
						   WB_SYNC_NONE);
Linus Torvalds's avatar
Linus Torvalds committed
126

127 128 129 130 131
		/*
		 * First and last FULL page! Partial pages are deliberately
		 * preserved on the expectation that it is better to preserve
		 * needed memory than to discard unneeded memory.
		 */
132 133
		start_index = (offset+(PAGE_SIZE-1)) >> PAGE_SHIFT;
		end_index = (endbyte >> PAGE_SHIFT);
134 135 136 137 138 139 140 141 142
		/*
		 * The page at end_index will be inclusively discarded according
		 * by invalidate_mapping_pages(), so subtracting 1 from
		 * end_index means we will skip the last page.  But if endbyte
		 * is page aligned or is at the end of file, we should not skip
		 * that page - discarding the last page is safe enough.
		 */
		if ((endbyte & ~PAGE_MASK) != ~PAGE_MASK &&
				endbyte != inode->i_size - 1) {
143 144 145 146 147 148 149 150 151 152
			/* First page is tricky as 0 - 1 = -1, but pgoff_t
			 * is unsigned, so the end_index >= start_index
			 * check below would be true and we'll discard the whole
			 * file cache which is not what was asked.
			 */
			if (end_index == 0)
				break;

			end_index--;
		}
Linus Torvalds's avatar
Linus Torvalds committed
153

154
		if (end_index >= start_index) {
155 156 157 158 159 160 161 162 163 164 165 166 167 168
			unsigned long count;

			/*
			 * It's common to FADV_DONTNEED right after
			 * the read or write that instantiates the
			 * pages, in which case there will be some
			 * sitting on the local LRU cache. Try to
			 * avoid the expensive remote drain and the
			 * second cache tree walk below by flushing
			 * them out right away.
			 */
			lru_add_drain();

			count = invalidate_mapping_pages(mapping,
169 170 171 172 173 174 175 176 177 178 179
						start_index, end_index);

			/*
			 * If fewer pages were invalidated than expected then
			 * it is possible that some of the pages were on
			 * a per-cpu pagevec for a remote CPU. Drain all
			 * pagevecs and try again.
			 */
			if (count < (end_index - start_index + 1)) {
				lru_add_drain_all();
				invalidate_mapping_pages(mapping, start_index,
180
						end_index);
181 182
			}
		}
183
		break;
Linus Torvalds's avatar
Linus Torvalds committed
184 185 186 187
	default:
		ret = -EINVAL;
	}
out:
188
	fdput(f);
Linus Torvalds's avatar
Linus Torvalds committed
189 190 191 192 193
	return ret;
}

#ifdef __ARCH_WANT_SYS_FADVISE64

194
SYSCALL_DEFINE4(fadvise64, int, fd, loff_t, offset, size_t, len, int, advice)
Linus Torvalds's avatar
Linus Torvalds committed
195 196 197 198 199
{
	return sys_fadvise64_64(fd, offset, len, advice);
}

#endif