Blame view
mm/fadvise.c
4.87 KB
b24413180
|
1 |
// SPDX-License-Identifier: GPL-2.0 |
1da177e4c
|
2 3 4 5 6 |
/* * mm/fadvise.c * * Copyright (C) 2002, Linus Torvalds * |
e1f8e8744
|
7 |
* 11Jan2003 Andrew Morton |
1da177e4c
|
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> |
ebcf28e1c
|
19 |
#include <linux/writeback.h> |
1da177e4c
|
20 |
#include <linux/syscalls.h> |
67d46b296
|
21 |
#include <linux/swap.h> |
1da177e4c
|
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. */ |
4a0fd5bf0
|
29 |
SYSCALL_DEFINE4(fadvise64_64, int, fd, loff_t, offset, loff_t, len, int, advice) |
1da177e4c
|
30 |
{ |
2903ff019
|
31 |
struct fd f = fdget(fd); |
e748dcd09
|
32 |
struct inode *inode; |
1da177e4c
|
33 34 |
struct address_space *mapping; struct backing_dev_info *bdi; |
ebcf28e1c
|
35 |
loff_t endbyte; /* inclusive */ |
1da177e4c
|
36 37 38 39 |
pgoff_t start_index; pgoff_t end_index; unsigned long nrpages; int ret = 0; |
2903ff019
|
40 |
if (!f.file) |
1da177e4c
|
41 |
return -EBADF; |
e748dcd09
|
42 43 |
inode = file_inode(f.file); if (S_ISFIFO(inode->i_mode)) { |
87ba81dba
|
44 45 46 |
ret = -ESPIPE; goto out; } |
2903ff019
|
47 |
mapping = f.file->f_mapping; |
1da177e4c
|
48 49 50 51 |
if (!mapping || len < 0) { ret = -EINVAL; goto out; } |
3a77d2148
|
52 53 54 |
bdi = inode_to_bdi(mapping->host); if (IS_DAX(inode) || (bdi == &noop_backing_dev_info)) { |
b5beb1caf
|
55 56 57 58 59 60 61 62 63 64 65 66 |
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; } |
fe77ba6f4
|
67 |
goto out; |
b5beb1caf
|
68 |
} |
fe77ba6f4
|
69 |
|
4570403f6
|
70 71 72 73 74 75 |
/* * 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; |
1da177e4c
|
76 77 |
if (!len || endbyte < len) endbyte = -1; |
ebcf28e1c
|
78 79 |
else endbyte--; /* inclusive */ |
1da177e4c
|
80 |
|
1da177e4c
|
81 82 |
switch (advice) { case POSIX_FADV_NORMAL: |
2903ff019
|
83 84 85 86 |
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); |
1da177e4c
|
87 88 |
break; case POSIX_FADV_RANDOM: |
2903ff019
|
89 90 91 |
spin_lock(&f.file->f_lock); f.file->f_mode |= FMODE_RANDOM; spin_unlock(&f.file->f_lock); |
1da177e4c
|
92 93 |
break; case POSIX_FADV_SEQUENTIAL: |
2903ff019
|
94 95 96 97 |
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); |
1da177e4c
|
98 99 |
break; case POSIX_FADV_WILLNEED: |
1da177e4c
|
100 |
/* First and last PARTIAL page! */ |
09cbfeaf1
|
101 102 |
start_index = offset >> PAGE_SHIFT; end_index = endbyte >> PAGE_SHIFT; |
1da177e4c
|
103 104 105 106 107 |
/* Careful about overflow on the "+1" */ nrpages = end_index - start_index + 1; if (!nrpages) nrpages = ~0UL; |
3d3727cdb
|
108 109 110 111 112 |
/* * Ignore return value because fadvise() shall return * success even if filesystem can't retrieve a hint, */ |
2903ff019
|
113 |
force_page_cache_readahead(mapping, f.file, start_index, |
3d3727cdb
|
114 |
nrpages); |
1da177e4c
|
115 |
break; |
60c371bc7
|
116 117 |
case POSIX_FADV_NOREUSE: break; |
1da177e4c
|
118 |
case POSIX_FADV_DONTNEED: |
703c27088
|
119 |
if (!inode_write_congested(mapping->host)) |
ad8a1b558
|
120 121 |
__filemap_fdatawrite_range(mapping, offset, endbyte, WB_SYNC_NONE); |
1da177e4c
|
122 |
|
441c228f8
|
123 124 125 126 127 |
/* * 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. */ |
09cbfeaf1
|
128 129 |
start_index = (offset+(PAGE_SIZE-1)) >> PAGE_SHIFT; end_index = (endbyte >> PAGE_SHIFT); |
3b1d9626f
|
130 131 132 133 134 135 136 137 138 |
/* * 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) { |
18aba41cb
|
139 140 141 142 143 144 145 146 147 148 |
/* 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--; } |
1da177e4c
|
149 |
|
67d46b296
|
150 |
if (end_index >= start_index) { |
4dd72b4a4
|
151 152 153 154 155 156 157 158 159 160 161 162 163 164 |
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, |
67d46b296
|
165 166 167 168 169 170 171 172 173 174 175 |
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, |
ebcf28e1c
|
176 |
end_index); |
67d46b296
|
177 178 |
} } |
ebcf28e1c
|
179 |
break; |
1da177e4c
|
180 181 182 183 |
default: ret = -EINVAL; } out: |
2903ff019
|
184 |
fdput(f); |
1da177e4c
|
185 186 187 188 |
return ret; } #ifdef __ARCH_WANT_SYS_FADVISE64 |
4a0fd5bf0
|
189 |
SYSCALL_DEFINE4(fadvise64, int, fd, loff_t, offset, size_t, len, int, advice) |
1da177e4c
|
190 191 192 193 194 |
{ return sys_fadvise64_64(fd, offset, len, advice); } #endif |