]>
Commit | Line | Data |
---|---|---|
b2441318 | 1 | /* SPDX-License-Identifier: GPL-2.0 */ |
1da177e4 LT |
2 | #ifndef _LINUX_FS_H |
3 | #define _LINUX_FS_H | |
4 | ||
1da177e4 | 5 | #include <linux/linkage.h> |
5dd43ce2 | 6 | #include <linux/wait_bit.h> |
1da177e4 LT |
7 | #include <linux/kdev_t.h> |
8 | #include <linux/dcache.h> | |
3f8206d4 | 9 | #include <linux/path.h> |
1da177e4 LT |
10 | #include <linux/stat.h> |
11 | #include <linux/cache.h> | |
1da177e4 | 12 | #include <linux/list.h> |
bc3b14cb | 13 | #include <linux/list_lru.h> |
4f5e65a1 | 14 | #include <linux/llist.h> |
1da177e4 | 15 | #include <linux/radix-tree.h> |
b93b0163 | 16 | #include <linux/xarray.h> |
6b2dbba8 | 17 | #include <linux/rbtree.h> |
1da177e4 | 18 | #include <linux/init.h> |
914e2637 | 19 | #include <linux/pid.h> |
187f1882 | 20 | #include <linux/bug.h> |
1b1dcc1b | 21 | #include <linux/mutex.h> |
c8c06efa | 22 | #include <linux/rwsem.h> |
baabda26 | 23 | #include <linux/mm_types.h> |
3bd858ab | 24 | #include <linux/capability.h> |
6188e10d | 25 | #include <linux/semaphore.h> |
c75b1d94 | 26 | #include <linux/fcntl.h> |
c4b929b8 | 27 | #include <linux/fiemap.h> |
ceb5bdc2 | 28 | #include <linux/rculist_bl.h> |
07b8ce1e | 29 | #include <linux/atomic.h> |
83aeeada | 30 | #include <linux/shrinker.h> |
c1aab02d | 31 | #include <linux/migrate_mode.h> |
92361636 | 32 | #include <linux/uidgid.h> |
5accdf82 | 33 | #include <linux/lockdep.h> |
c2b1ad80 | 34 | #include <linux/percpu-rwsem.h> |
853b39a7 | 35 | #include <linux/workqueue.h> |
fceef393 | 36 | #include <linux/delayed_call.h> |
85787090 | 37 | #include <linux/uuid.h> |
5660e13d | 38 | #include <linux/errseq.h> |
1da177e4 | 39 | |
1da177e4 | 40 | #include <asm/byteorder.h> |
607ca46e | 41 | #include <uapi/linux/fs.h> |
1da177e4 | 42 | |
b83ae6d4 | 43 | struct backing_dev_info; |
52ebea74 | 44 | struct bdi_writeback; |
2f8b5444 | 45 | struct bio; |
a5694255 | 46 | struct export_operations; |
a885c8c4 | 47 | struct hd_geometry; |
1da177e4 | 48 | struct iovec; |
92198f7e | 49 | struct kiocb; |
57cc7215 | 50 | struct kobject; |
1da177e4 LT |
51 | struct pipe_inode_info; |
52 | struct poll_table_struct; | |
53 | struct kstatfs; | |
54 | struct vm_area_struct; | |
55 | struct vfsmount; | |
745ca247 | 56 | struct cred; |
a509bc1a | 57 | struct swap_info_struct; |
55985dd7 | 58 | struct seq_file; |
7b7a8665 | 59 | struct workqueue_struct; |
3b93f911 | 60 | struct iov_iter; |
0b81d077 JK |
61 | struct fscrypt_info; |
62 | struct fscrypt_operations; | |
1da177e4 | 63 | |
74bf17cf | 64 | extern void __init inode_init(void); |
1da177e4 | 65 | extern void __init inode_init_early(void); |
4248b0da MG |
66 | extern void __init files_init(void); |
67 | extern void __init files_maxfiles_init(void); | |
1da177e4 | 68 | |
dded4f4d | 69 | extern struct files_stat_struct files_stat; |
518de9b3 | 70 | extern unsigned long get_max_files(void); |
9b80a184 | 71 | extern unsigned int sysctl_nr_open; |
dded4f4d JE |
72 | extern struct inodes_stat_t inodes_stat; |
73 | extern int leases_enable, lease_break_time; | |
800179c9 KC |
74 | extern int sysctl_protected_symlinks; |
75 | extern int sysctl_protected_hardlinks; | |
dded4f4d | 76 | |
ddef7ed2 CH |
77 | typedef __kernel_rwf_t rwf_t; |
78 | ||
1da177e4 LT |
79 | struct buffer_head; |
80 | typedef int (get_block_t)(struct inode *inode, sector_t iblock, | |
81 | struct buffer_head *bh_result, int create); | |
187372a3 | 82 | typedef int (dio_iodone_t)(struct kiocb *iocb, loff_t offset, |
7b7a8665 | 83 | ssize_t bytes, void *private); |
1da177e4 | 84 | |
bbc1096a DH |
85 | #define MAY_EXEC 0x00000001 |
86 | #define MAY_WRITE 0x00000002 | |
87 | #define MAY_READ 0x00000004 | |
88 | #define MAY_APPEND 0x00000008 | |
89 | #define MAY_ACCESS 0x00000010 | |
90 | #define MAY_OPEN 0x00000020 | |
91 | #define MAY_CHDIR 0x00000040 | |
92 | /* called from RCU mode, don't block */ | |
93 | #define MAY_NOT_BLOCK 0x00000080 | |
94 | ||
95 | /* | |
96 | * flags in file.f_mode. Note that FMODE_READ and FMODE_WRITE must correspond | |
97 | * to O_WRONLY and O_RDWR via the strange trick in __dentry_open() | |
98 | */ | |
99 | ||
100 | /* file is open for reading */ | |
101 | #define FMODE_READ ((__force fmode_t)0x1) | |
102 | /* file is open for writing */ | |
103 | #define FMODE_WRITE ((__force fmode_t)0x2) | |
104 | /* file is seekable */ | |
105 | #define FMODE_LSEEK ((__force fmode_t)0x4) | |
106 | /* file can be accessed using pread */ | |
107 | #define FMODE_PREAD ((__force fmode_t)0x8) | |
108 | /* file can be accessed using pwrite */ | |
109 | #define FMODE_PWRITE ((__force fmode_t)0x10) | |
110 | /* File is opened for execution with sys_execve / sys_uselib */ | |
111 | #define FMODE_EXEC ((__force fmode_t)0x20) | |
112 | /* File is opened with O_NDELAY (only set for block devices) */ | |
113 | #define FMODE_NDELAY ((__force fmode_t)0x40) | |
114 | /* File is opened with O_EXCL (only set for block devices) */ | |
115 | #define FMODE_EXCL ((__force fmode_t)0x80) | |
116 | /* File is opened using open(.., 3, ..) and is writeable only for ioctls | |
117 | (specialy hack for floppy.c) */ | |
118 | #define FMODE_WRITE_IOCTL ((__force fmode_t)0x100) | |
119 | /* 32bit hashes as llseek() offset (for directories) */ | |
120 | #define FMODE_32BITHASH ((__force fmode_t)0x200) | |
121 | /* 64bit hashes as llseek() offset (for directories) */ | |
122 | #define FMODE_64BITHASH ((__force fmode_t)0x400) | |
123 | ||
124 | /* | |
125 | * Don't update ctime and mtime. | |
126 | * | |
127 | * Currently a special hack for the XFS open_by_handle ioctl, but we'll | |
128 | * hopefully graduate it to a proper O_CMTIME flag supported by open(2) soon. | |
129 | */ | |
130 | #define FMODE_NOCMTIME ((__force fmode_t)0x800) | |
131 | ||
132 | /* Expect random access pattern */ | |
133 | #define FMODE_RANDOM ((__force fmode_t)0x1000) | |
134 | ||
135 | /* File is huge (eg. /dev/kmem): treat loff_t as unsigned */ | |
136 | #define FMODE_UNSIGNED_OFFSET ((__force fmode_t)0x2000) | |
137 | ||
138 | /* File is opened with O_PATH; almost nothing can be done with it */ | |
139 | #define FMODE_PATH ((__force fmode_t)0x4000) | |
140 | ||
9c225f26 LT |
141 | /* File needs atomic accesses to f_pos */ |
142 | #define FMODE_ATOMIC_POS ((__force fmode_t)0x8000) | |
83f936c7 AV |
143 | /* Write access to underlying fs */ |
144 | #define FMODE_WRITER ((__force fmode_t)0x10000) | |
7f7f25e8 AV |
145 | /* Has read method(s) */ |
146 | #define FMODE_CAN_READ ((__force fmode_t)0x20000) | |
147 | /* Has write method(s) */ | |
148 | #define FMODE_CAN_WRITE ((__force fmode_t)0x40000) | |
9c225f26 | 149 | |
bbc1096a | 150 | /* File was opened by fanotify and shouldn't generate fanotify events */ |
75069f2b | 151 | #define FMODE_NONOTIFY ((__force fmode_t)0x4000000) |
bbc1096a | 152 | |
91f9943e CH |
153 | /* File is capable of returning -EAGAIN if I/O will block */ |
154 | #define FMODE_NOWAIT ((__force fmode_t)0x8000000) | |
b745fafa | 155 | |
bbc1096a DH |
156 | /* |
157 | * Flag for rw_copy_check_uvector and compat_rw_copy_check_uvector | |
158 | * that indicates that they should check the contents of the iovec are | |
159 | * valid, but not check the memory that the iovec elements | |
160 | * points too. | |
161 | */ | |
162 | #define CHECK_IOVEC_ONLY -1 | |
163 | ||
1da177e4 LT |
164 | /* |
165 | * Attribute flags. These should be or-ed together to figure out what | |
166 | * has been changed! | |
167 | */ | |
9767d749 MS |
168 | #define ATTR_MODE (1 << 0) |
169 | #define ATTR_UID (1 << 1) | |
170 | #define ATTR_GID (1 << 2) | |
171 | #define ATTR_SIZE (1 << 3) | |
172 | #define ATTR_ATIME (1 << 4) | |
173 | #define ATTR_MTIME (1 << 5) | |
174 | #define ATTR_CTIME (1 << 6) | |
175 | #define ATTR_ATIME_SET (1 << 7) | |
176 | #define ATTR_MTIME_SET (1 << 8) | |
177 | #define ATTR_FORCE (1 << 9) /* Not a change, but a change it */ | |
178 | #define ATTR_ATTR_FLAG (1 << 10) | |
179 | #define ATTR_KILL_SUID (1 << 11) | |
180 | #define ATTR_KILL_SGID (1 << 12) | |
181 | #define ATTR_FILE (1 << 13) | |
182 | #define ATTR_KILL_PRIV (1 << 14) | |
183 | #define ATTR_OPEN (1 << 15) /* Truncating from open(O_TRUNC) */ | |
184 | #define ATTR_TIMES_SET (1 << 16) | |
f2b20f6e | 185 | #define ATTR_TOUCH (1 << 17) |
1da177e4 | 186 | |
787fb6bc MS |
187 | /* |
188 | * Whiteout is represented by a char device. The following constants define the | |
189 | * mode and device number to use. | |
190 | */ | |
191 | #define WHITEOUT_MODE 0 | |
192 | #define WHITEOUT_DEV 0 | |
193 | ||
1da177e4 LT |
194 | /* |
195 | * This is the Inode Attributes structure, used for notify_change(). It | |
196 | * uses the above definitions as flags, to know which values have changed. | |
197 | * Also, in this manner, a Filesystem can look at only the values it cares | |
198 | * about. Basically, these are the attributes that the VFS layer can | |
199 | * request to change from the FS layer. | |
200 | * | |
201 | * Derek Atkins <warlord@MIT.EDU> 94-10-20 | |
202 | */ | |
203 | struct iattr { | |
204 | unsigned int ia_valid; | |
205 | umode_t ia_mode; | |
92361636 EB |
206 | kuid_t ia_uid; |
207 | kgid_t ia_gid; | |
1da177e4 LT |
208 | loff_t ia_size; |
209 | struct timespec ia_atime; | |
210 | struct timespec ia_mtime; | |
211 | struct timespec ia_ctime; | |
cc4e69de MS |
212 | |
213 | /* | |
25985edc | 214 | * Not an attribute, but an auxiliary info for filesystems wanting to |
cc4e69de MS |
215 | * implement an ftruncate() like method. NOTE: filesystem should |
216 | * check for (ia_valid & ATTR_FILE), and not for (ia_file != NULL). | |
217 | */ | |
218 | struct file *ia_file; | |
1da177e4 LT |
219 | }; |
220 | ||
1da177e4 LT |
221 | /* |
222 | * Includes for diskquotas. | |
223 | */ | |
224 | #include <linux/quota.h> | |
225 | ||
69c433ed MS |
226 | /* |
227 | * Maximum number of layers of fs stack. Needs to be limited to | |
228 | * prevent kernel stack overflow | |
229 | */ | |
230 | #define FILESYSTEM_MAX_STACK_DEPTH 2 | |
231 | ||
994fc28c ZB |
232 | /** |
233 | * enum positive_aop_returns - aop return codes with specific semantics | |
234 | * | |
235 | * @AOP_WRITEPAGE_ACTIVATE: Informs the caller that page writeback has | |
236 | * completed, that the page is still locked, and | |
237 | * should be considered active. The VM uses this hint | |
238 | * to return the page to the active list -- it won't | |
239 | * be a candidate for writeback again in the near | |
240 | * future. Other callers must be careful to unlock | |
241 | * the page if they get this return. Returned by | |
242 | * writepage(); | |
243 | * | |
244 | * @AOP_TRUNCATED_PAGE: The AOP method that was handed a locked page has | |
245 | * unlocked it and the page might have been truncated. | |
246 | * The caller should back up to acquiring a new page and | |
247 | * trying again. The aop will be taking reasonable | |
248 | * precautions not to livelock. If the caller held a page | |
249 | * reference, it should drop it before retrying. Returned | |
55144768 | 250 | * by readpage(). |
994fc28c ZB |
251 | * |
252 | * address_space_operation functions return these large constants to indicate | |
253 | * special semantics to the caller. These are much larger than the bytes in a | |
254 | * page to allow for functions that return the number of bytes operated on in a | |
255 | * given page. | |
256 | */ | |
257 | ||
258 | enum positive_aop_returns { | |
259 | AOP_WRITEPAGE_ACTIVATE = 0x80000, | |
260 | AOP_TRUNCATED_PAGE = 0x80001, | |
261 | }; | |
262 | ||
c718a975 TH |
263 | #define AOP_FLAG_CONT_EXPAND 0x0001 /* called from cont_expand */ |
264 | #define AOP_FLAG_NOFS 0x0002 /* used by filesystem to direct | |
54566b2c NP |
265 | * helper code (eg buffer layer) |
266 | * to clear GFP_FS from alloc */ | |
afddba49 | 267 | |
1da177e4 LT |
268 | /* |
269 | * oh the beauties of C type declarations. | |
270 | */ | |
271 | struct page; | |
272 | struct address_space; | |
273 | struct writeback_control; | |
1da177e4 | 274 | |
c75b1d94 JA |
275 | /* |
276 | * Write life time hint values. | |
277 | */ | |
278 | enum rw_hint { | |
279 | WRITE_LIFE_NOT_SET = 0, | |
280 | WRITE_LIFE_NONE = RWH_WRITE_LIFE_NONE, | |
281 | WRITE_LIFE_SHORT = RWH_WRITE_LIFE_SHORT, | |
282 | WRITE_LIFE_MEDIUM = RWH_WRITE_LIFE_MEDIUM, | |
283 | WRITE_LIFE_LONG = RWH_WRITE_LIFE_LONG, | |
284 | WRITE_LIFE_EXTREME = RWH_WRITE_LIFE_EXTREME, | |
285 | }; | |
286 | ||
e2e40f2c | 287 | #define IOCB_EVENTFD (1 << 0) |
2ba48ce5 AV |
288 | #define IOCB_APPEND (1 << 1) |
289 | #define IOCB_DIRECT (1 << 2) | |
97be7ebe | 290 | #define IOCB_HIPRI (1 << 3) |
dde0c2e7 CH |
291 | #define IOCB_DSYNC (1 << 4) |
292 | #define IOCB_SYNC (1 << 5) | |
70fe2f48 | 293 | #define IOCB_WRITE (1 << 6) |
b745fafa | 294 | #define IOCB_NOWAIT (1 << 7) |
e2e40f2c CH |
295 | |
296 | struct kiocb { | |
297 | struct file *ki_filp; | |
298 | loff_t ki_pos; | |
299 | void (*ki_complete)(struct kiocb *iocb, long ret, long ret2); | |
300 | void *private; | |
301 | int ki_flags; | |
c75b1d94 | 302 | enum rw_hint ki_hint; |
3859a271 | 303 | } __randomize_layout; |
e2e40f2c CH |
304 | |
305 | static inline bool is_sync_kiocb(struct kiocb *kiocb) | |
306 | { | |
307 | return kiocb->ki_complete == NULL; | |
308 | } | |
309 | ||
8ab22b9a HH |
310 | /* |
311 | * "descriptor" for what we're up to with a read. | |
312 | * This allows us to use the same read code yet | |
313 | * have multiple different users of the data that | |
314 | * we read from a file. | |
315 | * | |
316 | * The simplest case just copies the data to user | |
317 | * mode. | |
318 | */ | |
319 | typedef struct { | |
320 | size_t written; | |
321 | size_t count; | |
322 | union { | |
323 | char __user *buf; | |
324 | void *data; | |
325 | } arg; | |
326 | int error; | |
327 | } read_descriptor_t; | |
328 | ||
329 | typedef int (*read_actor_t)(read_descriptor_t *, struct page *, | |
330 | unsigned long, unsigned long); | |
2f718ffc | 331 | |
1da177e4 LT |
332 | struct address_space_operations { |
333 | int (*writepage)(struct page *page, struct writeback_control *wbc); | |
334 | int (*readpage)(struct file *, struct page *); | |
1da177e4 LT |
335 | |
336 | /* Write back some dirty pages from this mapping. */ | |
337 | int (*writepages)(struct address_space *, struct writeback_control *); | |
338 | ||
4741c9fd | 339 | /* Set a page dirty. Return true if this dirtied it */ |
1da177e4 LT |
340 | int (*set_page_dirty)(struct page *page); |
341 | ||
342 | int (*readpages)(struct file *filp, struct address_space *mapping, | |
343 | struct list_head *pages, unsigned nr_pages); | |
344 | ||
afddba49 NP |
345 | int (*write_begin)(struct file *, struct address_space *mapping, |
346 | loff_t pos, unsigned len, unsigned flags, | |
347 | struct page **pagep, void **fsdata); | |
348 | int (*write_end)(struct file *, struct address_space *mapping, | |
349 | loff_t pos, unsigned len, unsigned copied, | |
350 | struct page *page, void *fsdata); | |
351 | ||
1da177e4 LT |
352 | /* Unfortunately this kludge is needed for FIBMAP. Don't use it */ |
353 | sector_t (*bmap)(struct address_space *, sector_t); | |
d47992f8 | 354 | void (*invalidatepage) (struct page *, unsigned int, unsigned int); |
27496a8c | 355 | int (*releasepage) (struct page *, gfp_t); |
6072d13c | 356 | void (*freepage)(struct page *); |
c8b8e32d | 357 | ssize_t (*direct_IO)(struct kiocb *, struct iov_iter *iter); |
b969c4ab | 358 | /* |
ef277c73 TC |
359 | * migrate the contents of a page to the specified target. If |
360 | * migrate_mode is MIGRATE_ASYNC, it must not block. | |
b969c4ab | 361 | */ |
2d1db3b1 | 362 | int (*migratepage) (struct address_space *, |
a6bc32b8 | 363 | struct page *, struct page *, enum migrate_mode); |
bda807d4 MK |
364 | bool (*isolate_page)(struct page *, isolate_mode_t); |
365 | void (*putback_page)(struct page *); | |
e3db7691 | 366 | int (*launder_page) (struct page *); |
c186afb4 | 367 | int (*is_partially_uptodate) (struct page *, unsigned long, |
8ab22b9a | 368 | unsigned long); |
b4597226 | 369 | void (*is_dirty_writeback) (struct page *, bool *, bool *); |
25718736 | 370 | int (*error_remove_page)(struct address_space *, struct page *); |
62c230bc MG |
371 | |
372 | /* swapfile support */ | |
a509bc1a MG |
373 | int (*swap_activate)(struct swap_info_struct *sis, struct file *file, |
374 | sector_t *span); | |
375 | void (*swap_deactivate)(struct file *file); | |
1da177e4 LT |
376 | }; |
377 | ||
7dcda1c9 JA |
378 | extern const struct address_space_operations empty_aops; |
379 | ||
afddba49 NP |
380 | /* |
381 | * pagecache_write_begin/pagecache_write_end must be used by general code | |
382 | * to write into the pagecache. | |
383 | */ | |
384 | int pagecache_write_begin(struct file *, struct address_space *mapping, | |
385 | loff_t pos, unsigned len, unsigned flags, | |
386 | struct page **pagep, void **fsdata); | |
387 | ||
388 | int pagecache_write_end(struct file *, struct address_space *mapping, | |
389 | loff_t pos, unsigned len, unsigned copied, | |
390 | struct page *page, void *fsdata); | |
391 | ||
1da177e4 LT |
392 | struct address_space { |
393 | struct inode *host; /* owner: inode, block_device */ | |
b93b0163 | 394 | struct radix_tree_root i_pages; /* cached pages */ |
4bb5f5d9 | 395 | atomic_t i_mmap_writable;/* count VM_SHARED mappings */ |
f808c13f | 396 | struct rb_root_cached i_mmap; /* tree of private and shared mappings */ |
c8c06efa | 397 | struct rw_semaphore i_mmap_rwsem; /* protect tree, count, list */ |
b93b0163 | 398 | /* Protected by the i_pages lock */ |
1da177e4 | 399 | unsigned long nrpages; /* number of total pages */ |
f9fe48be RZ |
400 | /* number of shadow or DAX exceptional entries */ |
401 | unsigned long nrexceptional; | |
1da177e4 | 402 | pgoff_t writeback_index;/* writeback starts here */ |
f5e54d6e | 403 | const struct address_space_operations *a_ops; /* methods */ |
9c5d760b | 404 | unsigned long flags; /* error bits */ |
1da177e4 | 405 | spinlock_t private_lock; /* for use by the address_space */ |
9c5d760b | 406 | gfp_t gfp_mask; /* implicit gfp mask for allocations */ |
32e57c29 | 407 | struct list_head private_list; /* for use by the address_space */ |
252aa6f5 | 408 | void *private_data; /* ditto */ |
5660e13d | 409 | errseq_t wb_err; |
3859a271 | 410 | } __attribute__((aligned(sizeof(long)))) __randomize_layout; |
1da177e4 LT |
411 | /* |
412 | * On most architectures that alignment is already the case; but | |
25985edc | 413 | * must be enforced here for CRIS, to let the least significant bit |
1da177e4 LT |
414 | * of struct page's "mapping" pointer be used for PAGE_MAPPING_ANON. |
415 | */ | |
87192a2a | 416 | struct request_queue; |
1da177e4 LT |
417 | |
418 | struct block_device { | |
419 | dev_t bd_dev; /* not a kdev_t - it's a search key */ | |
f3ccfcda | 420 | int bd_openers; |
1da177e4 | 421 | struct inode * bd_inode; /* will die */ |
87d8fe1e | 422 | struct super_block * bd_super; |
c039e313 | 423 | struct mutex bd_mutex; /* open/close mutex */ |
6b4517a7 | 424 | void * bd_claiming; |
1da177e4 LT |
425 | void * bd_holder; |
426 | int bd_holders; | |
77ea887e | 427 | bool bd_write_holder; |
641dc636 | 428 | #ifdef CONFIG_SYSFS |
49731baa | 429 | struct list_head bd_holder_disks; |
641dc636 | 430 | #endif |
1da177e4 LT |
431 | struct block_device * bd_contains; |
432 | unsigned bd_block_size; | |
c2ee070f | 433 | u8 bd_partno; |
1da177e4 LT |
434 | struct hd_struct * bd_part; |
435 | /* number of times partitions within this device have been opened. */ | |
436 | unsigned bd_part_count; | |
437 | int bd_invalidated; | |
438 | struct gendisk * bd_disk; | |
87192a2a | 439 | struct request_queue * bd_queue; |
b1d2dc56 | 440 | struct backing_dev_info *bd_bdi; |
1da177e4 | 441 | struct list_head bd_list; |
1da177e4 LT |
442 | /* |
443 | * Private data. You must have bd_claim'ed the block_device | |
444 | * to use this. NOTE: bd_claim allows an owner to claim | |
445 | * the same device multiple times, the owner must take special | |
446 | * care to not mess up bd_private for that case. | |
447 | */ | |
448 | unsigned long bd_private; | |
fcccf502 TS |
449 | |
450 | /* The counter of freeze processes */ | |
451 | int bd_fsfreeze_count; | |
452 | /* Mutex for freeze */ | |
453 | struct mutex bd_fsfreeze_mutex; | |
3859a271 | 454 | } __randomize_layout; |
1da177e4 LT |
455 | |
456 | /* | |
457 | * Radix-tree tags, for tagging dirty and writeback pages within the pagecache | |
458 | * radix trees | |
459 | */ | |
460 | #define PAGECACHE_TAG_DIRTY 0 | |
461 | #define PAGECACHE_TAG_WRITEBACK 1 | |
f446daae | 462 | #define PAGECACHE_TAG_TOWRITE 2 |
1da177e4 LT |
463 | |
464 | int mapping_tagged(struct address_space *mapping, int tag); | |
465 | ||
8b28f621 DB |
466 | static inline void i_mmap_lock_write(struct address_space *mapping) |
467 | { | |
c8c06efa | 468 | down_write(&mapping->i_mmap_rwsem); |
8b28f621 DB |
469 | } |
470 | ||
471 | static inline void i_mmap_unlock_write(struct address_space *mapping) | |
472 | { | |
c8c06efa | 473 | up_write(&mapping->i_mmap_rwsem); |
8b28f621 DB |
474 | } |
475 | ||
3dec0ba0 DB |
476 | static inline void i_mmap_lock_read(struct address_space *mapping) |
477 | { | |
478 | down_read(&mapping->i_mmap_rwsem); | |
479 | } | |
480 | ||
481 | static inline void i_mmap_unlock_read(struct address_space *mapping) | |
482 | { | |
483 | up_read(&mapping->i_mmap_rwsem); | |
484 | } | |
485 | ||
1da177e4 LT |
486 | /* |
487 | * Might pages of this file be mapped into userspace? | |
488 | */ | |
489 | static inline int mapping_mapped(struct address_space *mapping) | |
490 | { | |
f808c13f | 491 | return !RB_EMPTY_ROOT(&mapping->i_mmap.rb_root); |
1da177e4 LT |
492 | } |
493 | ||
494 | /* | |
495 | * Might pages of this file have been modified in userspace? | |
496 | * Note that i_mmap_writable counts all VM_SHARED vmas: do_mmap_pgoff | |
497 | * marks vma as VM_SHARED if it is shared, and the file was opened for | |
498 | * writing i.e. vma may be mprotected writable even if now readonly. | |
4bb5f5d9 DH |
499 | * |
500 | * If i_mmap_writable is negative, no new writable mappings are allowed. You | |
501 | * can only deny writable mappings, if none exists right now. | |
1da177e4 LT |
502 | */ |
503 | static inline int mapping_writably_mapped(struct address_space *mapping) | |
504 | { | |
4bb5f5d9 DH |
505 | return atomic_read(&mapping->i_mmap_writable) > 0; |
506 | } | |
507 | ||
508 | static inline int mapping_map_writable(struct address_space *mapping) | |
509 | { | |
510 | return atomic_inc_unless_negative(&mapping->i_mmap_writable) ? | |
511 | 0 : -EPERM; | |
512 | } | |
513 | ||
514 | static inline void mapping_unmap_writable(struct address_space *mapping) | |
515 | { | |
516 | atomic_dec(&mapping->i_mmap_writable); | |
517 | } | |
518 | ||
519 | static inline int mapping_deny_writable(struct address_space *mapping) | |
520 | { | |
521 | return atomic_dec_unless_positive(&mapping->i_mmap_writable) ? | |
522 | 0 : -EBUSY; | |
523 | } | |
524 | ||
525 | static inline void mapping_allow_writable(struct address_space *mapping) | |
526 | { | |
527 | atomic_inc(&mapping->i_mmap_writable); | |
1da177e4 LT |
528 | } |
529 | ||
530 | /* | |
531 | * Use sequence counter to get consistent i_size on 32-bit processors. | |
532 | */ | |
533 | #if BITS_PER_LONG==32 && defined(CONFIG_SMP) | |
534 | #include <linux/seqlock.h> | |
535 | #define __NEED_I_SIZE_ORDERED | |
536 | #define i_size_ordered_init(inode) seqcount_init(&inode->i_size_seqcount) | |
537 | #else | |
538 | #define i_size_ordered_init(inode) do { } while (0) | |
539 | #endif | |
540 | ||
f19d4a8f AV |
541 | struct posix_acl; |
542 | #define ACL_NOT_CACHED ((void *)(-1)) | |
2a3a2a3f | 543 | #define ACL_DONT_CACHE ((void *)(-3)) |
f19d4a8f | 544 | |
b8a7a3a6 AG |
545 | static inline struct posix_acl * |
546 | uncached_acl_sentinel(struct task_struct *task) | |
547 | { | |
548 | return (void *)task + 1; | |
549 | } | |
550 | ||
551 | static inline bool | |
552 | is_uncached_acl(struct posix_acl *acl) | |
553 | { | |
554 | return (long)acl & 1; | |
555 | } | |
556 | ||
3ddcd056 LT |
557 | #define IOP_FASTPERM 0x0001 |
558 | #define IOP_LOOKUP 0x0002 | |
559 | #define IOP_NOFOLLOW 0x0004 | |
d0a5b995 | 560 | #define IOP_XATTR 0x0008 |
76fca90e | 561 | #define IOP_DEFAULT_READLINK 0x0010 |
3ddcd056 | 562 | |
9dd813c1 JK |
563 | struct fsnotify_mark_connector; |
564 | ||
3ddcd056 LT |
565 | /* |
566 | * Keep mostly read-only and often accessed (especially for | |
567 | * the RCU path lookup and 'stat' data) fields at the beginning | |
568 | * of the 'struct inode' | |
569 | */ | |
1da177e4 | 570 | struct inode { |
44a7d7a8 | 571 | umode_t i_mode; |
3ddcd056 | 572 | unsigned short i_opflags; |
92361636 EB |
573 | kuid_t i_uid; |
574 | kgid_t i_gid; | |
3ddcd056 LT |
575 | unsigned int i_flags; |
576 | ||
577 | #ifdef CONFIG_FS_POSIX_ACL | |
578 | struct posix_acl *i_acl; | |
579 | struct posix_acl *i_default_acl; | |
580 | #endif | |
581 | ||
44a7d7a8 NP |
582 | const struct inode_operations *i_op; |
583 | struct super_block *i_sb; | |
3ddcd056 | 584 | struct address_space *i_mapping; |
44a7d7a8 | 585 | |
13e12d14 LT |
586 | #ifdef CONFIG_SECURITY |
587 | void *i_security; | |
588 | #endif | |
44a7d7a8 | 589 | |
3ddcd056 LT |
590 | /* Stat data, not accessed from path walking */ |
591 | unsigned long i_ino; | |
a78ef704 MS |
592 | /* |
593 | * Filesystems may only read i_nlink directly. They shall use the | |
594 | * following functions for modification: | |
595 | * | |
596 | * (set|clear|inc|drop)_nlink | |
597 | * inode_(inc|dec)_link_count | |
598 | */ | |
599 | union { | |
600 | const unsigned int i_nlink; | |
601 | unsigned int __i_nlink; | |
602 | }; | |
3ddcd056 | 603 | dev_t i_rdev; |
2f9d3df8 | 604 | loff_t i_size; |
3ddcd056 LT |
605 | struct timespec i_atime; |
606 | struct timespec i_mtime; | |
607 | struct timespec i_ctime; | |
6cdbb0ef TT |
608 | spinlock_t i_lock; /* i_blocks, i_bytes, maybe i_size */ |
609 | unsigned short i_bytes; | |
2f9d3df8 | 610 | unsigned int i_blkbits; |
c75b1d94 | 611 | enum rw_hint i_write_hint; |
3ddcd056 LT |
612 | blkcnt_t i_blocks; |
613 | ||
614 | #ifdef __NEED_I_SIZE_ORDERED | |
615 | seqcount_t i_size_seqcount; | |
616 | #endif | |
617 | ||
618 | /* Misc */ | |
619 | unsigned long i_state; | |
9902af79 | 620 | struct rw_semaphore i_rwsem; |
13e12d14 | 621 | |
44a7d7a8 | 622 | unsigned long dirtied_when; /* jiffies of first dirtying */ |
a2f48706 | 623 | unsigned long dirtied_time_when; |
44a7d7a8 | 624 | |
1da177e4 | 625 | struct hlist_node i_hash; |
c7f54084 | 626 | struct list_head i_io_list; /* backing dev IO list */ |
52ebea74 TH |
627 | #ifdef CONFIG_CGROUP_WRITEBACK |
628 | struct bdi_writeback *i_wb; /* the associated cgroup wb */ | |
2a814908 TH |
629 | |
630 | /* foreign inode detection, see wbc_detach_inode() */ | |
631 | int i_wb_frn_winner; | |
632 | u16 i_wb_frn_avg_time; | |
633 | u16 i_wb_frn_history; | |
52ebea74 | 634 | #endif |
7ccf19a8 | 635 | struct list_head i_lru; /* inode LRU list */ |
1da177e4 | 636 | struct list_head i_sb_list; |
6c60d2b5 | 637 | struct list_head i_wb_list; /* backing dev writeback list */ |
fa0d7e3d | 638 | union { |
b3d9b7a3 | 639 | struct hlist_head i_dentry; |
fa0d7e3d NP |
640 | struct rcu_head i_rcu; |
641 | }; | |
f02a9ad1 | 642 | atomic64_t i_version; |
2f9d3df8 | 643 | atomic_t i_count; |
bd5fe6c5 | 644 | atomic_t i_dio_count; |
6cdbb0ef | 645 | atomic_t i_writecount; |
d984ea60 MZ |
646 | #ifdef CONFIG_IMA |
647 | atomic_t i_readcount; /* struct files open RO */ | |
648 | #endif | |
99ac48f5 | 649 | const struct file_operations *i_fop; /* former ->i_op->default_file_ops */ |
4a075e39 | 650 | struct file_lock_context *i_flctx; |
1da177e4 | 651 | struct address_space i_data; |
1da177e4 | 652 | struct list_head i_devices; |
4c154168 TT |
653 | union { |
654 | struct pipe_inode_info *i_pipe; | |
eaf796e7 | 655 | struct block_device *i_bdev; |
577c4eb0 | 656 | struct cdev *i_cdev; |
61ba64fc | 657 | char *i_link; |
84e710da | 658 | unsigned i_dir_seq; |
4c154168 | 659 | }; |
1da177e4 LT |
660 | |
661 | __u32 i_generation; | |
662 | ||
3be25f49 EP |
663 | #ifdef CONFIG_FSNOTIFY |
664 | __u32 i_fsnotify_mask; /* all events this inode cares about */ | |
08991e83 | 665 | struct fsnotify_mark_connector __rcu *i_fsnotify_marks; |
0eeca283 RL |
666 | #endif |
667 | ||
0b81d077 JK |
668 | #if IS_ENABLED(CONFIG_FS_ENCRYPTION) |
669 | struct fscrypt_info *i_crypt_info; | |
670 | #endif | |
671 | ||
8e18e294 | 672 | void *i_private; /* fs or device private pointer */ |
3859a271 | 673 | } __randomize_layout; |
1da177e4 | 674 | |
93407472 FF |
675 | static inline unsigned int i_blocksize(const struct inode *node) |
676 | { | |
677 | return (1 << node->i_blkbits); | |
678 | } | |
679 | ||
1d3382cb AV |
680 | static inline int inode_unhashed(struct inode *inode) |
681 | { | |
682 | return hlist_unhashed(&inode->i_hash); | |
683 | } | |
684 | ||
f2eace23 IM |
685 | /* |
686 | * inode->i_mutex nesting subclasses for the lock validator: | |
687 | * | |
688 | * 0: the object of the current VFS operation | |
689 | * 1: parent | |
690 | * 2: child/target | |
40bd22c9 BF |
691 | * 3: xattr |
692 | * 4: second non-directory | |
d1b72cc6 MS |
693 | * 5: second parent (when locking independent directories in rename) |
694 | * | |
695 | * I_MUTEX_NONDIR2 is for certain operations (such as rename) which lock two | |
40bd22c9 | 696 | * non-directories at once. |
f2eace23 IM |
697 | * |
698 | * The locking order between these classes is | |
d1b72cc6 | 699 | * parent[2] -> child -> grandchild -> normal -> xattr -> second non-directory |
f2eace23 IM |
700 | */ |
701 | enum inode_i_mutex_lock_class | |
702 | { | |
703 | I_MUTEX_NORMAL, | |
704 | I_MUTEX_PARENT, | |
705 | I_MUTEX_CHILD, | |
4df46240 | 706 | I_MUTEX_XATTR, |
d1b72cc6 MS |
707 | I_MUTEX_NONDIR2, |
708 | I_MUTEX_PARENT2, | |
f2eace23 IM |
709 | }; |
710 | ||
5955102c AV |
711 | static inline void inode_lock(struct inode *inode) |
712 | { | |
9902af79 | 713 | down_write(&inode->i_rwsem); |
5955102c AV |
714 | } |
715 | ||
716 | static inline void inode_unlock(struct inode *inode) | |
717 | { | |
9902af79 AV |
718 | up_write(&inode->i_rwsem); |
719 | } | |
720 | ||
721 | static inline void inode_lock_shared(struct inode *inode) | |
722 | { | |
723 | down_read(&inode->i_rwsem); | |
724 | } | |
725 | ||
726 | static inline void inode_unlock_shared(struct inode *inode) | |
727 | { | |
728 | up_read(&inode->i_rwsem); | |
5955102c AV |
729 | } |
730 | ||
731 | static inline int inode_trylock(struct inode *inode) | |
732 | { | |
9902af79 AV |
733 | return down_write_trylock(&inode->i_rwsem); |
734 | } | |
735 | ||
736 | static inline int inode_trylock_shared(struct inode *inode) | |
737 | { | |
738 | return down_read_trylock(&inode->i_rwsem); | |
5955102c AV |
739 | } |
740 | ||
741 | static inline int inode_is_locked(struct inode *inode) | |
742 | { | |
9902af79 | 743 | return rwsem_is_locked(&inode->i_rwsem); |
5955102c AV |
744 | } |
745 | ||
746 | static inline void inode_lock_nested(struct inode *inode, unsigned subclass) | |
747 | { | |
9902af79 | 748 | down_write_nested(&inode->i_rwsem, subclass); |
5955102c AV |
749 | } |
750 | ||
01c2e13d DW |
751 | static inline void inode_lock_shared_nested(struct inode *inode, unsigned subclass) |
752 | { | |
753 | down_read_nested(&inode->i_rwsem, subclass); | |
754 | } | |
755 | ||
375e289e BF |
756 | void lock_two_nondirectories(struct inode *, struct inode*); |
757 | void unlock_two_nondirectories(struct inode *, struct inode*); | |
758 | ||
1da177e4 LT |
759 | /* |
760 | * NOTE: in a 32bit arch with a preemptable kernel and | |
761 | * an UP compile the i_size_read/write must be atomic | |
762 | * with respect to the local cpu (unlike with preempt disabled), | |
763 | * but they don't need to be atomic with respect to other cpus like in | |
764 | * true SMP (so they need either to either locally disable irq around | |
765 | * the read or for example on x86 they can be still implemented as a | |
766 | * cmpxchg8b without the need of the lock prefix). For SMP compiles | |
767 | * and 64bit archs it makes no difference if preempt is enabled or not. | |
768 | */ | |
48ed214d | 769 | static inline loff_t i_size_read(const struct inode *inode) |
1da177e4 LT |
770 | { |
771 | #if BITS_PER_LONG==32 && defined(CONFIG_SMP) | |
772 | loff_t i_size; | |
773 | unsigned int seq; | |
774 | ||
775 | do { | |
776 | seq = read_seqcount_begin(&inode->i_size_seqcount); | |
777 | i_size = inode->i_size; | |
778 | } while (read_seqcount_retry(&inode->i_size_seqcount, seq)); | |
779 | return i_size; | |
780 | #elif BITS_PER_LONG==32 && defined(CONFIG_PREEMPT) | |
781 | loff_t i_size; | |
782 | ||
783 | preempt_disable(); | |
784 | i_size = inode->i_size; | |
785 | preempt_enable(); | |
786 | return i_size; | |
787 | #else | |
788 | return inode->i_size; | |
789 | #endif | |
790 | } | |
791 | ||
7762f5a0 MS |
792 | /* |
793 | * NOTE: unlike i_size_read(), i_size_write() does need locking around it | |
794 | * (normally i_mutex), otherwise on 32bit/SMP an update of i_size_seqcount | |
795 | * can be lost, resulting in subsequent i_size_read() calls spinning forever. | |
796 | */ | |
1da177e4 LT |
797 | static inline void i_size_write(struct inode *inode, loff_t i_size) |
798 | { | |
799 | #if BITS_PER_LONG==32 && defined(CONFIG_SMP) | |
74e3d1e1 | 800 | preempt_disable(); |
1da177e4 LT |
801 | write_seqcount_begin(&inode->i_size_seqcount); |
802 | inode->i_size = i_size; | |
803 | write_seqcount_end(&inode->i_size_seqcount); | |
74e3d1e1 | 804 | preempt_enable(); |
1da177e4 LT |
805 | #elif BITS_PER_LONG==32 && defined(CONFIG_PREEMPT) |
806 | preempt_disable(); | |
807 | inode->i_size = i_size; | |
808 | preempt_enable(); | |
809 | #else | |
810 | inode->i_size = i_size; | |
811 | #endif | |
812 | } | |
813 | ||
48ed214d | 814 | static inline unsigned iminor(const struct inode *inode) |
1da177e4 LT |
815 | { |
816 | return MINOR(inode->i_rdev); | |
817 | } | |
818 | ||
48ed214d | 819 | static inline unsigned imajor(const struct inode *inode) |
1da177e4 LT |
820 | { |
821 | return MAJOR(inode->i_rdev); | |
822 | } | |
823 | ||
824 | extern struct block_device *I_BDEV(struct inode *inode); | |
825 | ||
826 | struct fown_struct { | |
827 | rwlock_t lock; /* protects pid, uid, euid fields */ | |
609d7fa9 EB |
828 | struct pid *pid; /* pid or -pgrp where SIGIO should be sent */ |
829 | enum pid_type pid_type; /* Kind of process group SIGIO should be sent to */ | |
92361636 | 830 | kuid_t uid, euid; /* uid/euid of process setting the owner */ |
1da177e4 LT |
831 | int signum; /* posix.1b rt signal to be delivered on IO */ |
832 | }; | |
833 | ||
834 | /* | |
835 | * Track a single file's readahead state | |
836 | */ | |
837 | struct file_ra_state { | |
937085aa FW |
838 | pgoff_t start; /* where readahead started */ |
839 | unsigned int size; /* # of readahead pages */ | |
840 | unsigned int async_size; /* do asynchronous readahead when | |
f9acc8c7 | 841 | there are only # of pages ahead */ |
5ce1110b | 842 | |
937085aa | 843 | unsigned int ra_pages; /* Maximum readahead window */ |
1ebf26a9 | 844 | unsigned int mmap_miss; /* Cache miss stat for mmap accesses */ |
f4e6b498 | 845 | loff_t prev_pos; /* Cache last read() position */ |
1da177e4 | 846 | }; |
1da177e4 | 847 | |
5ce1110b FW |
848 | /* |
849 | * Check if @index falls in the readahead windows. | |
850 | */ | |
851 | static inline int ra_has_index(struct file_ra_state *ra, pgoff_t index) | |
852 | { | |
f9acc8c7 FW |
853 | return (index >= ra->start && |
854 | index < ra->start + ra->size); | |
5ce1110b FW |
855 | } |
856 | ||
1da177e4 | 857 | struct file { |
2f512016 | 858 | union { |
4f5e65a1 | 859 | struct llist_node fu_llist; |
2f512016 ED |
860 | struct rcu_head fu_rcuhead; |
861 | } f_u; | |
0f7fc9e4 | 862 | struct path f_path; |
dd37978c | 863 | struct inode *f_inode; /* cached value */ |
99ac48f5 | 864 | const struct file_operations *f_op; |
ef3d0fd2 AK |
865 | |
866 | /* | |
9c225f26 | 867 | * Protects f_ep_links, f_flags. |
ef3d0fd2 AK |
868 | * Must not be taken from IRQ context. |
869 | */ | |
870 | spinlock_t f_lock; | |
c75b1d94 | 871 | enum rw_hint f_write_hint; |
516e0cc5 | 872 | atomic_long_t f_count; |
1da177e4 | 873 | unsigned int f_flags; |
aeb5d727 | 874 | fmode_t f_mode; |
9c225f26 | 875 | struct mutex f_pos_lock; |
1da177e4 LT |
876 | loff_t f_pos; |
877 | struct fown_struct f_owner; | |
d76b0d9b | 878 | const struct cred *f_cred; |
1da177e4 LT |
879 | struct file_ra_state f_ra; |
880 | ||
2b47c361 | 881 | u64 f_version; |
50462062 | 882 | #ifdef CONFIG_SECURITY |
1da177e4 | 883 | void *f_security; |
50462062 | 884 | #endif |
1da177e4 LT |
885 | /* needed for tty driver, and maybe others */ |
886 | void *private_data; | |
887 | ||
888 | #ifdef CONFIG_EPOLL | |
889 | /* Used by fs/eventpoll.c to link all the hooks to this file */ | |
890 | struct list_head f_ep_links; | |
28d82dc1 | 891 | struct list_head f_tfile_llink; |
1da177e4 LT |
892 | #endif /* #ifdef CONFIG_EPOLL */ |
893 | struct address_space *f_mapping; | |
5660e13d | 894 | errseq_t f_wb_err; |
3859a271 KC |
895 | } __randomize_layout |
896 | __attribute__((aligned(4))); /* lest something weird decides that 2 is OK */ | |
1da177e4 | 897 | |
990d6c2d AK |
898 | struct file_handle { |
899 | __u32 handle_bytes; | |
900 | int handle_type; | |
901 | /* file identifier */ | |
902 | unsigned char f_handle[0]; | |
903 | }; | |
904 | ||
cb0942b8 AV |
905 | static inline struct file *get_file(struct file *f) |
906 | { | |
907 | atomic_long_inc(&f->f_count); | |
908 | return f; | |
909 | } | |
90f31d0e | 910 | #define get_file_rcu(x) atomic_long_inc_not_zero(&(x)->f_count) |
d7065da0 | 911 | #define fput_atomic(x) atomic_long_add_unless(&(x)->f_count, -1, 1) |
516e0cc5 | 912 | #define file_count(x) atomic_long_read(&(x)->f_count) |
1da177e4 LT |
913 | |
914 | #define MAX_NON_LFS ((1UL<<31) - 1) | |
915 | ||
916 | /* Page cache limit. The filesystems should put that into their s_maxbytes | |
917 | limits, otherwise bad things can happen in VM. */ | |
918 | #if BITS_PER_LONG==32 | |
0cc3b0ec | 919 | #define MAX_LFS_FILESIZE ((loff_t)ULONG_MAX << PAGE_SHIFT) |
1da177e4 | 920 | #elif BITS_PER_LONG==64 |
0cc3b0ec | 921 | #define MAX_LFS_FILESIZE ((loff_t)LLONG_MAX) |
1da177e4 LT |
922 | #endif |
923 | ||
924 | #define FL_POSIX 1 | |
925 | #define FL_FLOCK 2 | |
617588d5 | 926 | #define FL_DELEG 4 /* NFSv4 delegation */ |
1da177e4 | 927 | #define FL_ACCESS 8 /* not trying to lock, just looking */ |
f475ae95 | 928 | #define FL_EXISTS 16 /* when unlocking, test for existence */ |
1da177e4 | 929 | #define FL_LEASE 32 /* lease held on this file */ |
75e1fcc0 | 930 | #define FL_CLOSE 64 /* unlock on close */ |
1da177e4 | 931 | #define FL_SLEEP 128 /* A blocking lock */ |
778fc546 BF |
932 | #define FL_DOWNGRADE_PENDING 256 /* Lease is being downgraded */ |
933 | #define FL_UNLOCK_PENDING 512 /* Lease is being broken */ | |
cff2fce5 | 934 | #define FL_OFDLCK 1024 /* lock is "owned" by struct file */ |
11afe9f7 | 935 | #define FL_LAYOUT 2048 /* outstanding pNFS layout */ |
1da177e4 | 936 | |
50f2112c BC |
937 | #define FL_CLOSE_POSIX (FL_POSIX | FL_CLOSE) |
938 | ||
bde74e4b MS |
939 | /* |
940 | * Special return value from posix_lock_file() and vfs_lock_file() for | |
941 | * asynchronous locking. | |
942 | */ | |
943 | #define FILE_LOCK_DEFERRED 1 | |
944 | ||
7ca76311 | 945 | /* legacy typedef, should eventually be removed */ |
17fa388d | 946 | typedef void *fl_owner_t; |
1da177e4 | 947 | |
a7231a97 JL |
948 | struct file_lock; |
949 | ||
1da177e4 | 950 | struct file_lock_operations { |
1da177e4 LT |
951 | void (*fl_copy_lock)(struct file_lock *, struct file_lock *); |
952 | void (*fl_release_private)(struct file_lock *); | |
953 | }; | |
954 | ||
955 | struct lock_manager_operations { | |
8fb47a4f | 956 | int (*lm_compare_owner)(struct file_lock *, struct file_lock *); |
3999e493 | 957 | unsigned long (*lm_owner_key)(struct file_lock *); |
cae80b30 JL |
958 | fl_owner_t (*lm_get_owner)(fl_owner_t); |
959 | void (*lm_put_owner)(fl_owner_t); | |
8fb47a4f | 960 | void (*lm_notify)(struct file_lock *); /* unblock callback */ |
d0449b90 | 961 | int (*lm_grant)(struct file_lock *, int); |
4d01b7f5 | 962 | bool (*lm_break)(struct file_lock *); |
7448cc37 | 963 | int (*lm_change)(struct file_lock *, int, struct list_head *); |
1c7dd2ff | 964 | void (*lm_setup)(struct file_lock *, void **); |
1da177e4 LT |
965 | }; |
966 | ||
af558e33 BF |
967 | struct lock_manager { |
968 | struct list_head list; | |
c87fb4a3 BF |
969 | /* |
970 | * NFSv4 and up also want opens blocked during the grace period; | |
971 | * NLM doesn't care: | |
972 | */ | |
973 | bool block_opens; | |
af558e33 BF |
974 | }; |
975 | ||
5ccb0066 SK |
976 | struct net; |
977 | void locks_start_grace(struct net *, struct lock_manager *); | |
af558e33 | 978 | void locks_end_grace(struct lock_manager *); |
003278e4 CL |
979 | bool locks_in_grace(struct net *); |
980 | bool opens_in_grace(struct net *); | |
af558e33 | 981 | |
1da177e4 LT |
982 | /* that will die - we need it for nfs_lock_info */ |
983 | #include <linux/nfs_fs_i.h> | |
984 | ||
1cb36012 JL |
985 | /* |
986 | * struct file_lock represents a generic "file lock". It's used to represent | |
987 | * POSIX byte range locks, BSD (flock) locks, and leases. It's important to | |
988 | * note that the same struct is used to represent both a request for a lock and | |
989 | * the lock itself, but the same object is never used for both. | |
990 | * | |
991 | * FIXME: should we create a separate "struct lock_request" to help distinguish | |
992 | * these two uses? | |
993 | * | |
8116bf4c | 994 | * The varous i_flctx lists are ordered by: |
1cb36012 | 995 | * |
8116bf4c JL |
996 | * 1) lock owner |
997 | * 2) lock range start | |
998 | * 3) lock range end | |
1cb36012 JL |
999 | * |
1000 | * Obviously, the last two criteria only matter for POSIX locks. | |
1001 | */ | |
1da177e4 LT |
1002 | struct file_lock { |
1003 | struct file_lock *fl_next; /* singly linked list for this inode */ | |
6dee60f6 | 1004 | struct list_head fl_list; /* link into file_lock_context */ |
139ca04e | 1005 | struct hlist_node fl_link; /* node in global lists */ |
1da177e4 LT |
1006 | struct list_head fl_block; /* circular list of blocked processes */ |
1007 | fl_owner_t fl_owner; | |
710b7216 | 1008 | unsigned int fl_flags; |
afc1246f | 1009 | unsigned char fl_type; |
1da177e4 | 1010 | unsigned int fl_pid; |
7012b02a | 1011 | int fl_link_cpu; /* what cpu's list is this on? */ |
1da177e4 LT |
1012 | wait_queue_head_t fl_wait; |
1013 | struct file *fl_file; | |
1da177e4 LT |
1014 | loff_t fl_start; |
1015 | loff_t fl_end; | |
1016 | ||
1017 | struct fasync_struct * fl_fasync; /* for lease break notifications */ | |
778fc546 BF |
1018 | /* for lease breaks: */ |
1019 | unsigned long fl_break_time; | |
1020 | unsigned long fl_downgrade_time; | |
1da177e4 | 1021 | |
6aed6285 | 1022 | const struct file_lock_operations *fl_ops; /* Callbacks for filesystems */ |
7b021967 | 1023 | const struct lock_manager_operations *fl_lmops; /* Callbacks for lockmanagers */ |
1da177e4 LT |
1024 | union { |
1025 | struct nfs_lock_info nfs_fl; | |
8d0a8a9d | 1026 | struct nfs4_lock_info nfs4_fl; |
e8d6c554 DH |
1027 | struct { |
1028 | struct list_head link; /* link in AFS vnode's pending_locks list */ | |
1029 | int state; /* state of grant or error if -ve */ | |
1030 | } afs; | |
1da177e4 | 1031 | } fl_u; |
3859a271 | 1032 | } __randomize_layout; |
1da177e4 | 1033 | |
4a075e39 | 1034 | struct file_lock_context { |
6109c850 | 1035 | spinlock_t flc_lock; |
4a075e39 | 1036 | struct list_head flc_flock; |
bd61e0a9 | 1037 | struct list_head flc_posix; |
8634b51f | 1038 | struct list_head flc_lease; |
4a075e39 JL |
1039 | }; |
1040 | ||
1da177e4 LT |
1041 | /* The following constant reflects the upper bound of the file/locking space */ |
1042 | #ifndef OFFSET_MAX | |
1043 | #define INT_LIMIT(x) (~((x)1 << (sizeof(x)*8 - 1))) | |
1044 | #define OFFSET_MAX INT_LIMIT(loff_t) | |
1045 | #define OFFT_OFFSET_MAX INT_LIMIT(off_t) | |
1046 | #endif | |
1047 | ||
bfcd17a6 TP |
1048 | extern void send_sigio(struct fown_struct *fown, int fd, int band); |
1049 | ||
c568d683 MS |
1050 | /* |
1051 | * Return the inode to use for locking | |
1052 | * | |
1053 | * For overlayfs this should be the overlay inode, not the real inode returned | |
1054 | * by file_inode(). For any other fs file_inode(filp) and locks_inode(filp) are | |
1055 | * equal. | |
1056 | */ | |
1057 | static inline struct inode *locks_inode(const struct file *f) | |
1058 | { | |
1059 | return f->f_path.dentry->d_inode; | |
1060 | } | |
1061 | ||
bfcd17a6 | 1062 | #ifdef CONFIG_FILE_LOCKING |
a75d30c7 | 1063 | extern int fcntl_getlk(struct file *, unsigned int, struct flock *); |
c293621b | 1064 | extern int fcntl_setlk(unsigned int, struct file *, unsigned int, |
a75d30c7 | 1065 | struct flock *); |
1da177e4 LT |
1066 | |
1067 | #if BITS_PER_LONG == 32 | |
a75d30c7 | 1068 | extern int fcntl_getlk64(struct file *, unsigned int, struct flock64 *); |
c293621b | 1069 | extern int fcntl_setlk64(unsigned int, struct file *, unsigned int, |
a75d30c7 | 1070 | struct flock64 *); |
1da177e4 LT |
1071 | #endif |
1072 | ||
1da177e4 LT |
1073 | extern int fcntl_setlease(unsigned int fd, struct file *filp, long arg); |
1074 | extern int fcntl_getlease(struct file *filp); | |
1075 | ||
1076 | /* fs/locks.c */ | |
f27a0fe0 | 1077 | void locks_free_lock_context(struct inode *inode); |
05fa3135 | 1078 | void locks_free_lock(struct file_lock *fl); |
1da177e4 | 1079 | extern void locks_init_lock(struct file_lock *); |
c5b1f0d9 | 1080 | extern struct file_lock * locks_alloc_lock(void); |
1da177e4 | 1081 | extern void locks_copy_lock(struct file_lock *, struct file_lock *); |
3fe0fff1 | 1082 | extern void locks_copy_conflock(struct file_lock *, struct file_lock *); |
1da177e4 | 1083 | extern void locks_remove_posix(struct file *, fl_owner_t); |
78ed8a13 | 1084 | extern void locks_remove_file(struct file *); |
a9e61e25 | 1085 | extern void locks_release_private(struct file_lock *); |
6d34ac19 | 1086 | extern void posix_test_lock(struct file *, struct file_lock *); |
150b3934 | 1087 | extern int posix_lock_file(struct file *, struct file_lock *, struct file_lock *); |
f891a29f | 1088 | extern int posix_unblock_lock(struct file_lock *); |
3ee17abd | 1089 | extern int vfs_test_lock(struct file *, struct file_lock *); |
150b3934 | 1090 | extern int vfs_lock_file(struct file *, unsigned int, struct file_lock *, struct file_lock *); |
9b9d2ab4 | 1091 | extern int vfs_cancel_lock(struct file *filp, struct file_lock *fl); |
e55c34a6 | 1092 | extern int locks_lock_inode_wait(struct inode *inode, struct file_lock *fl); |
df4e8d2c | 1093 | extern int __break_lease(struct inode *inode, unsigned int flags, unsigned int type); |
1da177e4 | 1094 | extern void lease_get_mtime(struct inode *, struct timespec *time); |
e6f5c789 JL |
1095 | extern int generic_setlease(struct file *, long, struct file_lock **, void **priv); |
1096 | extern int vfs_setlease(struct file *, long, struct file_lock **, void **); | |
7448cc37 | 1097 | extern int lease_modify(struct file_lock *, int, struct list_head *); |
6c8c9031 AV |
1098 | struct files_struct; |
1099 | extern void show_fd_locks(struct seq_file *f, | |
1100 | struct file *filp, struct files_struct *files); | |
bfcd17a6 | 1101 | #else /* !CONFIG_FILE_LOCKING */ |
c1e62b8f JL |
1102 | static inline int fcntl_getlk(struct file *file, unsigned int cmd, |
1103 | struct flock __user *user) | |
c2aca5e5 SW |
1104 | { |
1105 | return -EINVAL; | |
1106 | } | |
1107 | ||
1108 | static inline int fcntl_setlk(unsigned int fd, struct file *file, | |
1109 | unsigned int cmd, struct flock __user *user) | |
1110 | { | |
1111 | return -EACCES; | |
1112 | } | |
1113 | ||
bfcd17a6 | 1114 | #if BITS_PER_LONG == 32 |
c1e62b8f JL |
1115 | static inline int fcntl_getlk64(struct file *file, unsigned int cmd, |
1116 | struct flock64 __user *user) | |
c2aca5e5 SW |
1117 | { |
1118 | return -EINVAL; | |
1119 | } | |
1120 | ||
1121 | static inline int fcntl_setlk64(unsigned int fd, struct file *file, | |
1122 | unsigned int cmd, struct flock64 __user *user) | |
1123 | { | |
1124 | return -EACCES; | |
1125 | } | |
bfcd17a6 | 1126 | #endif |
c2aca5e5 SW |
1127 | static inline int fcntl_setlease(unsigned int fd, struct file *filp, long arg) |
1128 | { | |
1b2b32dc | 1129 | return -EINVAL; |
c2aca5e5 SW |
1130 | } |
1131 | ||
1132 | static inline int fcntl_getlease(struct file *filp) | |
1133 | { | |
1b2b32dc | 1134 | return F_UNLCK; |
c2aca5e5 SW |
1135 | } |
1136 | ||
4a075e39 | 1137 | static inline void |
f27a0fe0 | 1138 | locks_free_lock_context(struct inode *inode) |
4a075e39 JL |
1139 | { |
1140 | } | |
1141 | ||
c2aca5e5 SW |
1142 | static inline void locks_init_lock(struct file_lock *fl) |
1143 | { | |
1144 | return; | |
1145 | } | |
1146 | ||
3fe0fff1 | 1147 | static inline void locks_copy_conflock(struct file_lock *new, struct file_lock *fl) |
c2aca5e5 SW |
1148 | { |
1149 | return; | |
1150 | } | |
1151 | ||
1152 | static inline void locks_copy_lock(struct file_lock *new, struct file_lock *fl) | |
1153 | { | |
1154 | return; | |
1155 | } | |
1156 | ||
1157 | static inline void locks_remove_posix(struct file *filp, fl_owner_t owner) | |
1158 | { | |
1159 | return; | |
1160 | } | |
1161 | ||
78ed8a13 | 1162 | static inline void locks_remove_file(struct file *filp) |
c2aca5e5 SW |
1163 | { |
1164 | return; | |
1165 | } | |
1166 | ||
1167 | static inline void posix_test_lock(struct file *filp, struct file_lock *fl) | |
1168 | { | |
1169 | return; | |
1170 | } | |
1171 | ||
1172 | static inline int posix_lock_file(struct file *filp, struct file_lock *fl, | |
1173 | struct file_lock *conflock) | |
1174 | { | |
1175 | return -ENOLCK; | |
1176 | } | |
1177 | ||
f891a29f | 1178 | static inline int posix_unblock_lock(struct file_lock *waiter) |
c2aca5e5 SW |
1179 | { |
1180 | return -ENOENT; | |
1181 | } | |
1182 | ||
1183 | static inline int vfs_test_lock(struct file *filp, struct file_lock *fl) | |
1184 | { | |
1185 | return 0; | |
1186 | } | |
1187 | ||
1188 | static inline int vfs_lock_file(struct file *filp, unsigned int cmd, | |
1189 | struct file_lock *fl, struct file_lock *conf) | |
1190 | { | |
1191 | return -ENOLCK; | |
1192 | } | |
1193 | ||
1194 | static inline int vfs_cancel_lock(struct file *filp, struct file_lock *fl) | |
1195 | { | |
1196 | return 0; | |
1197 | } | |
1198 | ||
e55c34a6 BC |
1199 | static inline int locks_lock_inode_wait(struct inode *inode, struct file_lock *fl) |
1200 | { | |
1201 | return -ENOLCK; | |
1202 | } | |
1203 | ||
df4e8d2c | 1204 | static inline int __break_lease(struct inode *inode, unsigned int mode, unsigned int type) |
c2aca5e5 SW |
1205 | { |
1206 | return 0; | |
1207 | } | |
1208 | ||
1209 | static inline void lease_get_mtime(struct inode *inode, struct timespec *time) | |
1210 | { | |
1211 | return; | |
1212 | } | |
1213 | ||
1214 | static inline int generic_setlease(struct file *filp, long arg, | |
e6f5c789 | 1215 | struct file_lock **flp, void **priv) |
c2aca5e5 SW |
1216 | { |
1217 | return -EINVAL; | |
1218 | } | |
1219 | ||
1220 | static inline int vfs_setlease(struct file *filp, long arg, | |
e6f5c789 | 1221 | struct file_lock **lease, void **priv) |
c2aca5e5 SW |
1222 | { |
1223 | return -EINVAL; | |
1224 | } | |
1225 | ||
7448cc37 | 1226 | static inline int lease_modify(struct file_lock *fl, int arg, |
c45198ed | 1227 | struct list_head *dispose) |
c2aca5e5 SW |
1228 | { |
1229 | return -EINVAL; | |
1230 | } | |
6c8c9031 AV |
1231 | |
1232 | struct files_struct; | |
1233 | static inline void show_fd_locks(struct seq_file *f, | |
1234 | struct file *filp, struct files_struct *files) {} | |
bfcd17a6 TP |
1235 | #endif /* !CONFIG_FILE_LOCKING */ |
1236 | ||
ee296d7c JL |
1237 | static inline struct inode *file_inode(const struct file *f) |
1238 | { | |
1239 | return f->f_inode; | |
1240 | } | |
1241 | ||
d101a125 MS |
1242 | static inline struct dentry *file_dentry(const struct file *file) |
1243 | { | |
495e6429 | 1244 | return d_real(file->f_path.dentry, file_inode(file), 0, 0); |
d101a125 MS |
1245 | } |
1246 | ||
e55c34a6 BC |
1247 | static inline int locks_lock_file_wait(struct file *filp, struct file_lock *fl) |
1248 | { | |
c568d683 | 1249 | return locks_lock_inode_wait(locks_inode(filp), fl); |
e55c34a6 BC |
1250 | } |
1251 | ||
1da177e4 | 1252 | struct fasync_struct { |
989a2979 ED |
1253 | spinlock_t fa_lock; |
1254 | int magic; | |
1255 | int fa_fd; | |
1256 | struct fasync_struct *fa_next; /* singly linked list */ | |
1257 | struct file *fa_file; | |
1258 | struct rcu_head fa_rcu; | |
1da177e4 LT |
1259 | }; |
1260 | ||
1261 | #define FASYNC_MAGIC 0x4601 | |
1262 | ||
1263 | /* SMP safe fasync helpers: */ | |
1264 | extern int fasync_helper(int, struct file *, int, struct fasync_struct **); | |
f7347ce4 LT |
1265 | extern struct fasync_struct *fasync_insert_entry(int, struct file *, struct fasync_struct **, struct fasync_struct *); |
1266 | extern int fasync_remove_entry(struct file *, struct fasync_struct **); | |
1267 | extern struct fasync_struct *fasync_alloc(void); | |
1268 | extern void fasync_free(struct fasync_struct *); | |
1269 | ||
1da177e4 LT |
1270 | /* can be called from interrupts */ |
1271 | extern void kill_fasync(struct fasync_struct **, int, int); | |
1da177e4 | 1272 | |
e0b93edd | 1273 | extern void __f_setown(struct file *filp, struct pid *, enum pid_type, int force); |
393cc3f5 | 1274 | extern int f_setown(struct file *filp, unsigned long arg, int force); |
1da177e4 | 1275 | extern void f_delown(struct file *filp); |
609d7fa9 | 1276 | extern pid_t f_getown(struct file *filp); |
1da177e4 LT |
1277 | extern int send_sigurg(struct fown_struct *fown); |
1278 | ||
e462ec50 DH |
1279 | /* |
1280 | * sb->s_flags. Note that these mirror the equivalent MS_* flags where | |
1281 | * represented in both. | |
1282 | */ | |
1283 | #define SB_RDONLY 1 /* Mount read-only */ | |
1284 | #define SB_NOSUID 2 /* Ignore suid and sgid bits */ | |
1285 | #define SB_NODEV 4 /* Disallow access to device special files */ | |
1286 | #define SB_NOEXEC 8 /* Disallow program execution */ | |
1287 | #define SB_SYNCHRONOUS 16 /* Writes are synced at once */ | |
1288 | #define SB_MANDLOCK 64 /* Allow mandatory locks on an FS */ | |
1289 | #define SB_DIRSYNC 128 /* Directory modifications are synchronous */ | |
1290 | #define SB_NOATIME 1024 /* Do not update access times. */ | |
1291 | #define SB_NODIRATIME 2048 /* Do not update directory access times */ | |
1292 | #define SB_SILENT 32768 | |
1293 | #define SB_POSIXACL (1<<16) /* VFS does not apply the umask */ | |
1294 | #define SB_KERNMOUNT (1<<22) /* this is a kern_mount call */ | |
1295 | #define SB_I_VERSION (1<<23) /* Update inode I_version field */ | |
1296 | #define SB_LAZYTIME (1<<25) /* Update the on-disk [acm]times lazily */ | |
1297 | ||
1298 | /* These sb flags are internal to the kernel */ | |
1299 | #define SB_SUBMOUNT (1<<26) | |
1300 | #define SB_NOREMOTELOCK (1<<27) | |
1301 | #define SB_NOSEC (1<<28) | |
1302 | #define SB_BORN (1<<29) | |
1303 | #define SB_ACTIVE (1<<30) | |
1304 | #define SB_NOUSER (1<<31) | |
1305 | ||
1da177e4 LT |
1306 | /* |
1307 | * Umount options | |
1308 | */ | |
1309 | ||
1310 | #define MNT_FORCE 0x00000001 /* Attempt to forcibily umount */ | |
1311 | #define MNT_DETACH 0x00000002 /* Just detach from the tree */ | |
1312 | #define MNT_EXPIRE 0x00000004 /* Mark for expiry */ | |
db1f05bb MS |
1313 | #define UMOUNT_NOFOLLOW 0x00000008 /* Don't follow symlink on umount */ |
1314 | #define UMOUNT_UNUSED 0x80000000 /* Flag guaranteed to be unused */ | |
1da177e4 | 1315 | |
46b15caa TH |
1316 | /* sb->s_iflags */ |
1317 | #define SB_I_CGROUPWB 0x00000001 /* cgroup-aware writeback enabled */ | |
90f8572b | 1318 | #define SB_I_NOEXEC 0x00000002 /* Ignore executables on this fs */ |
a2982cc9 | 1319 | #define SB_I_NODEV 0x00000004 /* Ignore devices on this fs */ |
95dd7758 | 1320 | #define SB_I_MULTIROOT 0x00000008 /* Multiple roots to the dentry tree */ |
1da177e4 | 1321 | |
8654df4e EB |
1322 | /* sb->s_iflags to limit user namespace mounts */ |
1323 | #define SB_I_USERNS_VISIBLE 0x00000010 /* fstype already mounted */ | |
57b56ac6 MZ |
1324 | #define SB_I_IMA_UNVERIFIABLE_SIGNATURE 0x00000020 |
1325 | #define SB_I_UNTRUSTED_MOUNTER 0x00000040 | |
1da177e4 | 1326 | |
5accdf82 JK |
1327 | /* Possible states of 'frozen' field */ |
1328 | enum { | |
1329 | SB_UNFROZEN = 0, /* FS is unfrozen */ | |
1330 | SB_FREEZE_WRITE = 1, /* Writes, dir ops, ioctls frozen */ | |
5accdf82 JK |
1331 | SB_FREEZE_PAGEFAULT = 2, /* Page faults stopped as well */ |
1332 | SB_FREEZE_FS = 3, /* For internal FS use (e.g. to stop | |
1333 | * internal threads if needed) */ | |
1334 | SB_FREEZE_COMPLETE = 4, /* ->freeze_fs finished successfully */ | |
1335 | }; | |
1336 | ||
1337 | #define SB_FREEZE_LEVELS (SB_FREEZE_COMPLETE - 1) | |
1338 | ||
1339 | struct sb_writers { | |
8129ed29 ON |
1340 | int frozen; /* Is sb frozen? */ |
1341 | wait_queue_head_t wait_unfrozen; /* for get_super_thawed() */ | |
1342 | struct percpu_rw_semaphore rw_sem[SB_FREEZE_LEVELS]; | |
5accdf82 JK |
1343 | }; |
1344 | ||
1da177e4 LT |
1345 | struct super_block { |
1346 | struct list_head s_list; /* Keep this first */ | |
1347 | dev_t s_dev; /* search index; _not_ kdev_t */ | |
270ba5f7 RK |
1348 | unsigned char s_blocksize_bits; |
1349 | unsigned long s_blocksize; | |
42cb56ae | 1350 | loff_t s_maxbytes; /* Max file size */ |
1da177e4 | 1351 | struct file_system_type *s_type; |
ee9b6d61 | 1352 | const struct super_operations *s_op; |
61e225dc | 1353 | const struct dquot_operations *dq_op; |
0d54b217 | 1354 | const struct quotactl_ops *s_qcop; |
39655164 | 1355 | const struct export_operations *s_export_op; |
1da177e4 | 1356 | unsigned long s_flags; |
46b15caa | 1357 | unsigned long s_iflags; /* internal SB_I_* flags */ |
1da177e4 LT |
1358 | unsigned long s_magic; |
1359 | struct dentry *s_root; | |
1360 | struct rw_semaphore s_umount; | |
1da177e4 | 1361 | int s_count; |
1da177e4 | 1362 | atomic_t s_active; |
50462062 | 1363 | #ifdef CONFIG_SECURITY |
1da177e4 | 1364 | void *s_security; |
50462062 | 1365 | #endif |
bb435453 | 1366 | const struct xattr_handler **s_xattr; |
1da177e4 | 1367 | |
0b81d077 JK |
1368 | const struct fscrypt_operations *s_cop; |
1369 | ||
f1ee6162 | 1370 | struct hlist_bl_head s_roots; /* alternate root dentries for NFS */ |
39f7c4db | 1371 | struct list_head s_mounts; /* list of mounts; _not_ for fs use */ |
1da177e4 | 1372 | struct block_device *s_bdev; |
32a88aa1 | 1373 | struct backing_dev_info *s_bdi; |
acaebfd8 | 1374 | struct mtd_info *s_mtd; |
a5166169 | 1375 | struct hlist_node s_instances; |
2c5f648a | 1376 | unsigned int s_quota_types; /* Bitmask of supported quota types */ |
1da177e4 LT |
1377 | struct quota_info s_dquot; /* Diskquota specific options */ |
1378 | ||
5accdf82 | 1379 | struct sb_writers s_writers; |
1da177e4 | 1380 | |
85787090 CH |
1381 | char s_id[32]; /* Informational name */ |
1382 | uuid_t s_uuid; /* UUID */ | |
1da177e4 LT |
1383 | |
1384 | void *s_fs_info; /* Filesystem private info */ | |
8de52778 | 1385 | unsigned int s_max_links; |
30c40d2c | 1386 | fmode_t s_mode; |
1da177e4 | 1387 | |
270ba5f7 RK |
1388 | /* Granularity of c/m/atime in ns. |
1389 | Cannot be worse than a second */ | |
1390 | u32 s_time_gran; | |
1391 | ||
1da177e4 LT |
1392 | /* |
1393 | * The next field is for VFS *only*. No filesystems have any business | |
1394 | * even looking at it. You had been warned. | |
1395 | */ | |
a11f3a05 | 1396 | struct mutex s_vfs_rename_mutex; /* Kludge */ |
1da177e4 | 1397 | |
79c0b2df MS |
1398 | /* |
1399 | * Filesystem subtype. If non-empty the filesystem type field | |
1400 | * in /proc/mounts will be "type.subtype" | |
1401 | */ | |
1402 | char *s_subtype; | |
b3b304a2 | 1403 | |
c8aebb0c | 1404 | const struct dentry_operations *s_d_op; /* default d_op for dentries */ |
9fdfdcf1 DM |
1405 | |
1406 | /* | |
1407 | * Saved pool identifier for cleancache (-1 means none) | |
1408 | */ | |
1409 | int cleancache_poolid; | |
b0d40c92 DC |
1410 | |
1411 | struct shrinker s_shrink; /* per-sb shrinker handle */ | |
4ed5e82f | 1412 | |
7ada4db8 MS |
1413 | /* Number of inodes with nlink == 0 but still referenced */ |
1414 | atomic_long_t s_remove_count; | |
1415 | ||
4ed5e82f MS |
1416 | /* Being remounted read-only */ |
1417 | int s_readonly_remount; | |
7b7a8665 CH |
1418 | |
1419 | /* AIO completions deferred from interrupt context */ | |
1420 | struct workqueue_struct *s_dio_done_wq; | |
215752fc | 1421 | struct hlist_head s_pins; |
f6041567 | 1422 | |
6e4eab57 EB |
1423 | /* |
1424 | * Owning user namespace and default context in which to | |
1425 | * interpret filesystem uids, gids, quotas, device nodes, | |
1426 | * xattrs and security labels. | |
1427 | */ | |
1428 | struct user_namespace *s_user_ns; | |
1429 | ||
f6041567 DC |
1430 | /* |
1431 | * Keep the lru lists last in the structure so they always sit on their | |
1432 | * own individual cachelines. | |
1433 | */ | |
1434 | struct list_lru s_dentry_lru ____cacheline_aligned_in_smp; | |
1435 | struct list_lru s_inode_lru ____cacheline_aligned_in_smp; | |
e2fec7c3 | 1436 | struct rcu_head rcu; |
853b39a7 | 1437 | struct work_struct destroy_work; |
69c433ed | 1438 | |
e97fedb9 | 1439 | struct mutex s_sync_lock; /* sync serialisation lock */ |
69c433ed MS |
1440 | |
1441 | /* | |
1442 | * Indicates how deep in a filesystem stack this SB is | |
1443 | */ | |
1444 | int s_stack_depth; | |
74278da9 DC |
1445 | |
1446 | /* s_inode_list_lock protects s_inodes */ | |
1447 | spinlock_t s_inode_list_lock ____cacheline_aligned_in_smp; | |
1448 | struct list_head s_inodes; /* all inodes */ | |
6c60d2b5 DC |
1449 | |
1450 | spinlock_t s_inode_wblist_lock; | |
1451 | struct list_head s_inodes_wb; /* writeback inodes */ | |
3859a271 | 1452 | } __randomize_layout; |
1da177e4 | 1453 | |
81754357 SF |
1454 | /* Helper functions so that in most cases filesystems will |
1455 | * not need to deal directly with kuid_t and kgid_t and can | |
1456 | * instead deal with the raw numeric values that are stored | |
1457 | * in the filesystem. | |
1458 | */ | |
1459 | static inline uid_t i_uid_read(const struct inode *inode) | |
1460 | { | |
1461 | return from_kuid(inode->i_sb->s_user_ns, inode->i_uid); | |
1462 | } | |
1463 | ||
1464 | static inline gid_t i_gid_read(const struct inode *inode) | |
1465 | { | |
1466 | return from_kgid(inode->i_sb->s_user_ns, inode->i_gid); | |
1467 | } | |
1468 | ||
1469 | static inline void i_uid_write(struct inode *inode, uid_t uid) | |
1470 | { | |
1471 | inode->i_uid = make_kuid(inode->i_sb->s_user_ns, uid); | |
1472 | } | |
1473 | ||
1474 | static inline void i_gid_write(struct inode *inode, gid_t gid) | |
1475 | { | |
1476 | inode->i_gid = make_kgid(inode->i_sb->s_user_ns, gid); | |
1477 | } | |
1478 | ||
3cd88666 | 1479 | extern struct timespec current_time(struct inode *inode); |
1da177e4 LT |
1480 | |
1481 | /* | |
1482 | * Snapshotting support. | |
1483 | */ | |
1da177e4 | 1484 | |
5accdf82 JK |
1485 | void __sb_end_write(struct super_block *sb, int level); |
1486 | int __sb_start_write(struct super_block *sb, int level, bool wait); | |
1487 | ||
bee9182d | 1488 | #define __sb_writers_acquired(sb, lev) \ |
8129ed29 | 1489 | percpu_rwsem_acquire(&(sb)->s_writers.rw_sem[(lev)-1], 1, _THIS_IP_) |
bee9182d | 1490 | #define __sb_writers_release(sb, lev) \ |
8129ed29 | 1491 | percpu_rwsem_release(&(sb)->s_writers.rw_sem[(lev)-1], 1, _THIS_IP_) |
bee9182d | 1492 | |
5accdf82 JK |
1493 | /** |
1494 | * sb_end_write - drop write access to a superblock | |
1495 | * @sb: the super we wrote to | |
1496 | * | |
1497 | * Decrement number of writers to the filesystem. Wake up possible waiters | |
1498 | * wanting to freeze the filesystem. | |
1499 | */ | |
1500 | static inline void sb_end_write(struct super_block *sb) | |
1501 | { | |
1502 | __sb_end_write(sb, SB_FREEZE_WRITE); | |
1503 | } | |
1504 | ||
1505 | /** | |
1506 | * sb_end_pagefault - drop write access to a superblock from a page fault | |
1507 | * @sb: the super we wrote to | |
1508 | * | |
1509 | * Decrement number of processes handling write page fault to the filesystem. | |
1510 | * Wake up possible waiters wanting to freeze the filesystem. | |
1511 | */ | |
1512 | static inline void sb_end_pagefault(struct super_block *sb) | |
1513 | { | |
1514 | __sb_end_write(sb, SB_FREEZE_PAGEFAULT); | |
1515 | } | |
1516 | ||
1517 | /** | |
1518 | * sb_end_intwrite - drop write access to a superblock for internal fs purposes | |
1519 | * @sb: the super we wrote to | |
1520 | * | |
1521 | * Decrement fs-internal number of writers to the filesystem. Wake up possible | |
1522 | * waiters wanting to freeze the filesystem. | |
1523 | */ | |
1524 | static inline void sb_end_intwrite(struct super_block *sb) | |
1525 | { | |
1526 | __sb_end_write(sb, SB_FREEZE_FS); | |
1527 | } | |
1528 | ||
1529 | /** | |
1530 | * sb_start_write - get write access to a superblock | |
1531 | * @sb: the super we write to | |
1532 | * | |
1533 | * When a process wants to write data or metadata to a file system (i.e. dirty | |
1534 | * a page or an inode), it should embed the operation in a sb_start_write() - | |
1535 | * sb_end_write() pair to get exclusion against file system freezing. This | |
1536 | * function increments number of writers preventing freezing. If the file | |
1537 | * system is already frozen, the function waits until the file system is | |
1538 | * thawed. | |
1539 | * | |
1540 | * Since freeze protection behaves as a lock, users have to preserve | |
1541 | * ordering of freeze protection and other filesystem locks. Generally, | |
1542 | * freeze protection should be the outermost lock. In particular, we have: | |
1543 | * | |
1544 | * sb_start_write | |
1545 | * -> i_mutex (write path, truncate, directory ops, ...) | |
1546 | * -> s_umount (freeze_super, thaw_super) | |
1547 | */ | |
1548 | static inline void sb_start_write(struct super_block *sb) | |
1549 | { | |
1550 | __sb_start_write(sb, SB_FREEZE_WRITE, true); | |
1551 | } | |
1552 | ||
1553 | static inline int sb_start_write_trylock(struct super_block *sb) | |
1554 | { | |
1555 | return __sb_start_write(sb, SB_FREEZE_WRITE, false); | |
1556 | } | |
1557 | ||
1558 | /** | |
1559 | * sb_start_pagefault - get write access to a superblock from a page fault | |
1560 | * @sb: the super we write to | |
1561 | * | |
1562 | * When a process starts handling write page fault, it should embed the | |
1563 | * operation into sb_start_pagefault() - sb_end_pagefault() pair to get | |
1564 | * exclusion against file system freezing. This is needed since the page fault | |
1565 | * is going to dirty a page. This function increments number of running page | |
1566 | * faults preventing freezing. If the file system is already frozen, the | |
1567 | * function waits until the file system is thawed. | |
1568 | * | |
1569 | * Since page fault freeze protection behaves as a lock, users have to preserve | |
1570 | * ordering of freeze protection and other filesystem locks. It is advised to | |
1571 | * put sb_start_pagefault() close to mmap_sem in lock ordering. Page fault | |
1572 | * handling code implies lock dependency: | |
1573 | * | |
1574 | * mmap_sem | |
1575 | * -> sb_start_pagefault | |
1576 | */ | |
1577 | static inline void sb_start_pagefault(struct super_block *sb) | |
1578 | { | |
1579 | __sb_start_write(sb, SB_FREEZE_PAGEFAULT, true); | |
1580 | } | |
1581 | ||
1582 | /* | |
1583 | * sb_start_intwrite - get write access to a superblock for internal fs purposes | |
1584 | * @sb: the super we write to | |
1585 | * | |
1586 | * This is the third level of protection against filesystem freezing. It is | |
1587 | * free for use by a filesystem. The only requirement is that it must rank | |
1588 | * below sb_start_pagefault. | |
1589 | * | |
1590 | * For example filesystem can call sb_start_intwrite() when starting a | |
1591 | * transaction which somewhat eases handling of freezing for internal sources | |
1592 | * of filesystem changes (internal fs threads, discarding preallocation on file | |
1593 | * close, etc.). | |
1594 | */ | |
1595 | static inline void sb_start_intwrite(struct super_block *sb) | |
1596 | { | |
1597 | __sb_start_write(sb, SB_FREEZE_FS, true); | |
1598 | } | |
1da177e4 | 1599 | |
1da177e4 | 1600 | |
2e149670 | 1601 | extern bool inode_owner_or_capable(const struct inode *inode); |
3bd858ab | 1602 | |
1da177e4 LT |
1603 | /* |
1604 | * VFS helper functions.. | |
1605 | */ | |
312b63fb | 1606 | extern int vfs_create(struct inode *, struct dentry *, umode_t, bool); |
18bb1db3 | 1607 | extern int vfs_mkdir(struct inode *, struct dentry *, umode_t); |
1a67aafb | 1608 | extern int vfs_mknod(struct inode *, struct dentry *, umode_t, dev_t); |
db2e747b | 1609 | extern int vfs_symlink(struct inode *, struct dentry *, const char *); |
146a8595 | 1610 | extern int vfs_link(struct dentry *, struct inode *, struct dentry *, struct inode **); |
1da177e4 | 1611 | extern int vfs_rmdir(struct inode *, struct dentry *); |
b21996e3 | 1612 | extern int vfs_unlink(struct inode *, struct dentry *, struct inode **); |
520c8b16 | 1613 | extern int vfs_rename(struct inode *, struct dentry *, struct inode *, struct dentry *, struct inode **, unsigned int); |
787fb6bc | 1614 | extern int vfs_whiteout(struct inode *, struct dentry *); |
1da177e4 | 1615 | |
af7bd4dc AG |
1616 | extern struct dentry *vfs_tmpfile(struct dentry *dentry, umode_t mode, |
1617 | int open_flag); | |
1618 | ||
8e6c848e AV |
1619 | int vfs_mkobj(struct dentry *, umode_t, |
1620 | int (*f)(struct dentry *, umode_t, void *), | |
1621 | void *); | |
1622 | ||
8c744fb8 CH |
1623 | /* |
1624 | * VFS file helper functions. | |
1625 | */ | |
a1bd120d | 1626 | extern void inode_init_owner(struct inode *inode, const struct inode *dir, |
62bb1091 | 1627 | umode_t mode); |
a2982cc9 | 1628 | extern bool may_open_dev(const struct path *path); |
c4b929b8 MF |
1629 | /* |
1630 | * VFS FS_IOC_FIEMAP helper definitions. | |
1631 | */ | |
1632 | struct fiemap_extent_info { | |
1633 | unsigned int fi_flags; /* Flags as passed from user */ | |
1634 | unsigned int fi_extents_mapped; /* Number of mapped extents */ | |
1635 | unsigned int fi_extents_max; /* Size of fiemap_extent array */ | |
ecf5632d NK |
1636 | struct fiemap_extent __user *fi_extents_start; /* Start of |
1637 | fiemap_extent array */ | |
c4b929b8 MF |
1638 | }; |
1639 | int fiemap_fill_next_extent(struct fiemap_extent_info *info, u64 logical, | |
1640 | u64 phys, u64 len, u32 flags); | |
1641 | int fiemap_check_flags(struct fiemap_extent_info *fieinfo, u32 fs_flags); | |
1642 | ||
1da177e4 LT |
1643 | /* |
1644 | * File types | |
1645 | * | |
1646 | * NOTE! These match bits 12..15 of stat.st_mode | |
1647 | * (ie "(i_mode >> 12) & 15"). | |
1648 | */ | |
1649 | #define DT_UNKNOWN 0 | |
1650 | #define DT_FIFO 1 | |
1651 | #define DT_CHR 2 | |
1652 | #define DT_DIR 4 | |
1653 | #define DT_BLK 6 | |
1654 | #define DT_REG 8 | |
1655 | #define DT_LNK 10 | |
1656 | #define DT_SOCK 12 | |
1657 | #define DT_WHT 14 | |
1658 | ||
1da177e4 LT |
1659 | /* |
1660 | * This is the "filldir" function type, used by readdir() to let | |
1661 | * the kernel specify what kind of dirent layout it wants to have. | |
1662 | * This allows the kernel to read directories into kernel space or | |
1663 | * to have different dirent layouts depending on the binary type. | |
1664 | */ | |
ac7576f4 MS |
1665 | struct dir_context; |
1666 | typedef int (*filldir_t)(struct dir_context *, const char *, int, loff_t, u64, | |
1667 | unsigned); | |
1668 | ||
5c0ba4e0 | 1669 | struct dir_context { |
a09acf4b | 1670 | filldir_t actor; |
bb6f619b | 1671 | loff_t pos; |
5c0ba4e0 | 1672 | }; |
bb6f619b | 1673 | |
08f85851 | 1674 | struct block_device_operations; |
1da177e4 | 1675 | |
1da177e4 LT |
1676 | /* These macros are for out of kernel modules to test that |
1677 | * the kernel supports the unlocked_ioctl and compat_ioctl | |
1678 | * fields in struct file_operations. */ | |
1679 | #define HAVE_COMPAT_IOCTL 1 | |
1680 | #define HAVE_UNLOCKED_IOCTL 1 | |
1681 | ||
b4caecd4 CH |
1682 | /* |
1683 | * These flags let !MMU mmap() govern direct device mapping vs immediate | |
1684 | * copying more easily for MAP_PRIVATE, especially for ROM filesystems. | |
1685 | * | |
1686 | * NOMMU_MAP_COPY: Copy can be mapped (MAP_PRIVATE) | |
1687 | * NOMMU_MAP_DIRECT: Can be mapped directly (MAP_SHARED) | |
1688 | * NOMMU_MAP_READ: Can be mapped for reading | |
1689 | * NOMMU_MAP_WRITE: Can be mapped for writing | |
1690 | * NOMMU_MAP_EXEC: Can be mapped for execution | |
1691 | */ | |
1692 | #define NOMMU_MAP_COPY 0x00000001 | |
1693 | #define NOMMU_MAP_DIRECT 0x00000008 | |
1694 | #define NOMMU_MAP_READ VM_MAYREAD | |
1695 | #define NOMMU_MAP_WRITE VM_MAYWRITE | |
1696 | #define NOMMU_MAP_EXEC VM_MAYEXEC | |
1697 | ||
1698 | #define NOMMU_VMFLAGS \ | |
1699 | (NOMMU_MAP_READ | NOMMU_MAP_WRITE | NOMMU_MAP_EXEC) | |
1700 | ||
1701 | ||
293bc982 AV |
1702 | struct iov_iter; |
1703 | ||
1da177e4 LT |
1704 | struct file_operations { |
1705 | struct module *owner; | |
1706 | loff_t (*llseek) (struct file *, loff_t, int); | |
1707 | ssize_t (*read) (struct file *, char __user *, size_t, loff_t *); | |
1da177e4 | 1708 | ssize_t (*write) (struct file *, const char __user *, size_t, loff_t *); |
293bc982 AV |
1709 | ssize_t (*read_iter) (struct kiocb *, struct iov_iter *); |
1710 | ssize_t (*write_iter) (struct kiocb *, struct iov_iter *); | |
bb6f619b | 1711 | int (*iterate) (struct file *, struct dir_context *); |
61922694 | 1712 | int (*iterate_shared) (struct file *, struct dir_context *); |
a3f8683b | 1713 | __poll_t (*poll) (struct file *, struct poll_table_struct *); |
1da177e4 LT |
1714 | long (*unlocked_ioctl) (struct file *, unsigned int, unsigned long); |
1715 | long (*compat_ioctl) (struct file *, unsigned int, unsigned long); | |
1716 | int (*mmap) (struct file *, struct vm_area_struct *); | |
1c972597 | 1717 | unsigned long mmap_supported_flags; |
1da177e4 | 1718 | int (*open) (struct inode *, struct file *); |
75e1fcc0 | 1719 | int (*flush) (struct file *, fl_owner_t id); |
1da177e4 | 1720 | int (*release) (struct inode *, struct file *); |
02c24a82 | 1721 | int (*fsync) (struct file *, loff_t, loff_t, int datasync); |
1da177e4 LT |
1722 | int (*fasync) (int, struct file *, int); |
1723 | int (*lock) (struct file *, int, struct file_lock *); | |
1da177e4 LT |
1724 | ssize_t (*sendpage) (struct file *, struct page *, int, size_t, loff_t *, int); |
1725 | unsigned long (*get_unmapped_area)(struct file *, unsigned long, unsigned long, unsigned long, unsigned long); | |
1726 | int (*check_flags)(int); | |
1da177e4 | 1727 | int (*flock) (struct file *, int, struct file_lock *); |
cbb7e577 JA |
1728 | ssize_t (*splice_write)(struct pipe_inode_info *, struct file *, loff_t *, size_t, unsigned int); |
1729 | ssize_t (*splice_read)(struct file *, loff_t *, struct pipe_inode_info *, size_t, unsigned int); | |
e6f5c789 | 1730 | int (*setlease)(struct file *, long, struct file_lock **, void **); |
2fe17c10 CH |
1731 | long (*fallocate)(struct file *file, int mode, loff_t offset, |
1732 | loff_t len); | |
a3816ab0 | 1733 | void (*show_fdinfo)(struct seq_file *m, struct file *f); |
b4caecd4 CH |
1734 | #ifndef CONFIG_MMU |
1735 | unsigned (*mmap_capabilities)(struct file *); | |
1736 | #endif | |
04b38d60 CH |
1737 | ssize_t (*copy_file_range)(struct file *, loff_t, struct file *, |
1738 | loff_t, size_t, unsigned int); | |
1739 | int (*clone_file_range)(struct file *, loff_t, struct file *, loff_t, | |
1740 | u64); | |
54dbc151 DW |
1741 | ssize_t (*dedupe_file_range)(struct file *, u64, u64, struct file *, |
1742 | u64); | |
3859a271 | 1743 | } __randomize_layout; |
1da177e4 LT |
1744 | |
1745 | struct inode_operations { | |
00cd8dd3 | 1746 | struct dentry * (*lookup) (struct inode *,struct dentry *, unsigned int); |
fceef393 | 1747 | const char * (*get_link) (struct dentry *, struct inode *, struct delayed_call *); |
10556cb2 | 1748 | int (*permission) (struct inode *, int); |
4e34e719 | 1749 | struct posix_acl * (*get_acl)(struct inode *, int); |
44a7d7a8 NP |
1750 | |
1751 | int (*readlink) (struct dentry *, char __user *,int); | |
44a7d7a8 | 1752 | |
ebfc3b49 | 1753 | int (*create) (struct inode *,struct dentry *, umode_t, bool); |
1da177e4 LT |
1754 | int (*link) (struct dentry *,struct inode *,struct dentry *); |
1755 | int (*unlink) (struct inode *,struct dentry *); | |
1756 | int (*symlink) (struct inode *,struct dentry *,const char *); | |
18bb1db3 | 1757 | int (*mkdir) (struct inode *,struct dentry *,umode_t); |
1da177e4 | 1758 | int (*rmdir) (struct inode *,struct dentry *); |
1a67aafb | 1759 | int (*mknod) (struct inode *,struct dentry *,umode_t,dev_t); |
1da177e4 | 1760 | int (*rename) (struct inode *, struct dentry *, |
520c8b16 | 1761 | struct inode *, struct dentry *, unsigned int); |
1da177e4 | 1762 | int (*setattr) (struct dentry *, struct iattr *); |
a528d35e | 1763 | int (*getattr) (const struct path *, struct kstat *, u32, unsigned int); |
1da177e4 | 1764 | ssize_t (*listxattr) (struct dentry *, char *, size_t); |
c4b929b8 MF |
1765 | int (*fiemap)(struct inode *, struct fiemap_extent_info *, u64 start, |
1766 | u64 len); | |
c3b2da31 | 1767 | int (*update_time)(struct inode *, struct timespec *, int); |
d9585277 | 1768 | int (*atomic_open)(struct inode *, struct dentry *, |
30d90494 | 1769 | struct file *, unsigned open_flag, |
d9585277 | 1770 | umode_t create_mode, int *opened); |
60545d0d | 1771 | int (*tmpfile) (struct inode *, struct dentry *, umode_t); |
893d46e4 | 1772 | int (*set_acl)(struct inode *, struct posix_acl *, int); |
44a7d7a8 | 1773 | } ____cacheline_aligned; |
1da177e4 | 1774 | |
bb7462b6 MS |
1775 | static inline ssize_t call_read_iter(struct file *file, struct kiocb *kio, |
1776 | struct iov_iter *iter) | |
1777 | { | |
1778 | return file->f_op->read_iter(kio, iter); | |
1779 | } | |
1780 | ||
1781 | static inline ssize_t call_write_iter(struct file *file, struct kiocb *kio, | |
1782 | struct iov_iter *iter) | |
1783 | { | |
1784 | return file->f_op->write_iter(kio, iter); | |
1785 | } | |
1786 | ||
f74ac015 MS |
1787 | static inline int call_mmap(struct file *file, struct vm_area_struct *vma) |
1788 | { | |
1789 | return file->f_op->mmap(file, vma); | |
1790 | } | |
1791 | ||
eed4e51f | 1792 | ssize_t rw_copy_check_uvector(int type, const struct iovec __user * uvector, |
fcf63409 CY |
1793 | unsigned long nr_segs, unsigned long fast_segs, |
1794 | struct iovec *fast_pointer, | |
ac34ebb3 | 1795 | struct iovec **ret_pointer); |
eed4e51f | 1796 | |
6fb5032e | 1797 | extern ssize_t __vfs_read(struct file *, char __user *, size_t, loff_t *); |
1da177e4 LT |
1798 | extern ssize_t vfs_read(struct file *, char __user *, size_t, loff_t *); |
1799 | extern ssize_t vfs_write(struct file *, const char __user *, size_t, loff_t *); | |
1800 | extern ssize_t vfs_readv(struct file *, const struct iovec __user *, | |
ddef7ed2 | 1801 | unsigned long, loff_t *, rwf_t); |
29732938 ZB |
1802 | extern ssize_t vfs_copy_file_range(struct file *, loff_t , struct file *, |
1803 | loff_t, size_t, unsigned int); | |
876bec6f DW |
1804 | extern int vfs_clone_file_prep_inodes(struct inode *inode_in, loff_t pos_in, |
1805 | struct inode *inode_out, loff_t pos_out, | |
1806 | u64 *len, bool is_dedupe); | |
04b38d60 CH |
1807 | extern int vfs_clone_file_range(struct file *file_in, loff_t pos_in, |
1808 | struct file *file_out, loff_t pos_out, u64 len); | |
876bec6f DW |
1809 | extern int vfs_dedupe_file_range_compare(struct inode *src, loff_t srcoff, |
1810 | struct inode *dest, loff_t destoff, | |
1811 | loff_t len, bool *is_same); | |
54dbc151 DW |
1812 | extern int vfs_dedupe_file_range(struct file *file, |
1813 | struct file_dedupe_range *same); | |
1da177e4 | 1814 | |
1da177e4 LT |
1815 | struct super_operations { |
1816 | struct inode *(*alloc_inode)(struct super_block *sb); | |
1817 | void (*destroy_inode)(struct inode *); | |
1818 | ||
aa385729 | 1819 | void (*dirty_inode) (struct inode *, int flags); |
a9185b41 | 1820 | int (*write_inode) (struct inode *, struct writeback_control *wbc); |
45321ac5 | 1821 | int (*drop_inode) (struct inode *); |
be7ce416 | 1822 | void (*evict_inode) (struct inode *); |
1da177e4 | 1823 | void (*put_super) (struct super_block *); |
1da177e4 | 1824 | int (*sync_fs)(struct super_block *sb, int wait); |
48b6bca6 | 1825 | int (*freeze_super) (struct super_block *); |
c4be0c1d | 1826 | int (*freeze_fs) (struct super_block *); |
48b6bca6 | 1827 | int (*thaw_super) (struct super_block *); |
c4be0c1d | 1828 | int (*unfreeze_fs) (struct super_block *); |
726c3342 | 1829 | int (*statfs) (struct dentry *, struct kstatfs *); |
1da177e4 | 1830 | int (*remount_fs) (struct super_block *, int *, char *); |
42faad99 | 1831 | void (*umount_begin) (struct super_block *); |
1da177e4 | 1832 | |
34c80b1d | 1833 | int (*show_options)(struct seq_file *, struct dentry *); |
d861c630 | 1834 | int (*show_devname)(struct seq_file *, struct dentry *); |
a6322de6 | 1835 | int (*show_path)(struct seq_file *, struct dentry *); |
64132379 | 1836 | int (*show_stats)(struct seq_file *, struct dentry *); |
0e51a720 | 1837 | #ifdef CONFIG_QUOTA |
1da177e4 LT |
1838 | ssize_t (*quota_read)(struct super_block *, int, char *, size_t, loff_t); |
1839 | ssize_t (*quota_write)(struct super_block *, int, const char *, size_t, loff_t); | |
2d0fa467 | 1840 | struct dquot **(*get_dquots)(struct inode *); |
0e51a720 | 1841 | #endif |
87d8fe1e | 1842 | int (*bdev_try_to_free_page)(struct super_block*, struct page*, gfp_t); |
4101b624 VD |
1843 | long (*nr_cached_objects)(struct super_block *, |
1844 | struct shrink_control *); | |
1845 | long (*free_cached_objects)(struct super_block *, | |
1846 | struct shrink_control *); | |
1da177e4 LT |
1847 | }; |
1848 | ||
bbc1096a DH |
1849 | /* |
1850 | * Inode flags - they have no relation to superblock flags now | |
1851 | */ | |
1852 | #define S_SYNC 1 /* Writes are synced at once */ | |
1853 | #define S_NOATIME 2 /* Do not update access times */ | |
1854 | #define S_APPEND 4 /* Append-only file */ | |
1855 | #define S_IMMUTABLE 8 /* Immutable file */ | |
1856 | #define S_DEAD 16 /* removed, but still open directory */ | |
1857 | #define S_NOQUOTA 32 /* Inode is not counted to quota */ | |
1858 | #define S_DIRSYNC 64 /* Directory modifications are synchronous */ | |
1859 | #define S_NOCMTIME 128 /* Do not update file c/mtime */ | |
1860 | #define S_SWAPFILE 256 /* Do not truncate: swapon got its bmaps */ | |
1861 | #define S_PRIVATE 512 /* Inode is fs-internal */ | |
1862 | #define S_IMA 1024 /* Inode has an associated IMA struct */ | |
1863 | #define S_AUTOMOUNT 2048 /* Automount/referral quasi-directory */ | |
1864 | #define S_NOSEC 4096 /* no suid or xattr security attributes */ | |
6cd176a5 | 1865 | #ifdef CONFIG_FS_DAX |
fbbbad4b MW |
1866 | #define S_DAX 8192 /* Direct Access, avoiding the page cache */ |
1867 | #else | |
1868 | #define S_DAX 0 /* Make all the DAX code disappear */ | |
1869 | #endif | |
2ee6a576 | 1870 | #define S_ENCRYPTED 16384 /* Encrypted file (using fs/crypto/) */ |
bbc1096a DH |
1871 | |
1872 | /* | |
1873 | * Note that nosuid etc flags are inode-specific: setting some file-system | |
1874 | * flags just means all the inodes inherit those flags by default. It might be | |
1875 | * possible to override it selectively if you really wanted to with some | |
1876 | * ioctl() that is not currently implemented. | |
1877 | * | |
e462ec50 | 1878 | * Exception: SB_RDONLY is always applied to the entire file system. |
bbc1096a DH |
1879 | * |
1880 | * Unfortunately, it is possible to change a filesystems flags with it mounted | |
1881 | * with files in use. This means that all of the inodes will not have their | |
1882 | * i_flags updated. Hence, i_flags no longer inherit the superblock mount | |
1883 | * flags, so these have to be checked separately. -- rmk@arm.uk.linux.org | |
1884 | */ | |
1885 | #define __IS_FLG(inode, flg) ((inode)->i_sb->s_flags & (flg)) | |
1886 | ||
1751e8a6 | 1887 | static inline bool sb_rdonly(const struct super_block *sb) { return sb->s_flags & SB_RDONLY; } |
94e92e7a | 1888 | #define IS_RDONLY(inode) sb_rdonly((inode)->i_sb) |
e462ec50 | 1889 | #define IS_SYNC(inode) (__IS_FLG(inode, SB_SYNCHRONOUS) || \ |
bbc1096a | 1890 | ((inode)->i_flags & S_SYNC)) |
e462ec50 | 1891 | #define IS_DIRSYNC(inode) (__IS_FLG(inode, SB_SYNCHRONOUS|SB_DIRSYNC) || \ |
bbc1096a | 1892 | ((inode)->i_flags & (S_SYNC|S_DIRSYNC))) |
e462ec50 DH |
1893 | #define IS_MANDLOCK(inode) __IS_FLG(inode, SB_MANDLOCK) |
1894 | #define IS_NOATIME(inode) __IS_FLG(inode, SB_RDONLY|SB_NOATIME) | |
1895 | #define IS_I_VERSION(inode) __IS_FLG(inode, SB_I_VERSION) | |
bbc1096a DH |
1896 | |
1897 | #define IS_NOQUOTA(inode) ((inode)->i_flags & S_NOQUOTA) | |
1898 | #define IS_APPEND(inode) ((inode)->i_flags & S_APPEND) | |
1899 | #define IS_IMMUTABLE(inode) ((inode)->i_flags & S_IMMUTABLE) | |
e462ec50 | 1900 | #define IS_POSIXACL(inode) __IS_FLG(inode, SB_POSIXACL) |
bbc1096a DH |
1901 | |
1902 | #define IS_DEADDIR(inode) ((inode)->i_flags & S_DEAD) | |
1903 | #define IS_NOCMTIME(inode) ((inode)->i_flags & S_NOCMTIME) | |
1904 | #define IS_SWAPFILE(inode) ((inode)->i_flags & S_SWAPFILE) | |
1905 | #define IS_PRIVATE(inode) ((inode)->i_flags & S_PRIVATE) | |
1906 | #define IS_IMA(inode) ((inode)->i_flags & S_IMA) | |
1907 | #define IS_AUTOMOUNT(inode) ((inode)->i_flags & S_AUTOMOUNT) | |
1908 | #define IS_NOSEC(inode) ((inode)->i_flags & S_NOSEC) | |
fbbbad4b | 1909 | #define IS_DAX(inode) ((inode)->i_flags & S_DAX) |
2ee6a576 | 1910 | #define IS_ENCRYPTED(inode) ((inode)->i_flags & S_ENCRYPTED) |
bbc1096a | 1911 | |
787fb6bc MS |
1912 | #define IS_WHITEOUT(inode) (S_ISCHR(inode->i_mode) && \ |
1913 | (inode)->i_rdev == WHITEOUT_DEV) | |
1914 | ||
0bd23d09 EB |
1915 | static inline bool HAS_UNMAPPED_ID(struct inode *inode) |
1916 | { | |
1917 | return !uid_valid(inode->i_uid) || !gid_valid(inode->i_gid); | |
1918 | } | |
1919 | ||
c75b1d94 JA |
1920 | static inline enum rw_hint file_write_hint(struct file *file) |
1921 | { | |
1922 | if (file->f_write_hint != WRITE_LIFE_NOT_SET) | |
1923 | return file->f_write_hint; | |
1924 | ||
1925 | return file_inode(file)->i_write_hint; | |
1926 | } | |
1927 | ||
1928 | static inline int iocb_flags(struct file *file); | |
1929 | ||
1930 | static inline void init_sync_kiocb(struct kiocb *kiocb, struct file *filp) | |
1931 | { | |
1932 | *kiocb = (struct kiocb) { | |
1933 | .ki_filp = filp, | |
1934 | .ki_flags = iocb_flags(filp), | |
1935 | .ki_hint = file_write_hint(filp), | |
1936 | }; | |
1937 | } | |
1938 | ||
1c0eeaf5 | 1939 | /* |
250df6ed | 1940 | * Inode state bits. Protected by inode->i_lock |
1c0eeaf5 JE |
1941 | * |
1942 | * Three bits determine the dirty state of the inode, I_DIRTY_SYNC, | |
1943 | * I_DIRTY_DATASYNC and I_DIRTY_PAGES. | |
1944 | * | |
1945 | * Four bits define the lifetime of an inode. Initially, inodes are I_NEW, | |
1946 | * until that flag is cleared. I_WILL_FREE, I_FREEING and I_CLEAR are set at | |
1947 | * various stages of removing an inode. | |
1948 | * | |
eaff8079 | 1949 | * Two bits are used for locking and completion notification, I_NEW and I_SYNC. |
1c0eeaf5 | 1950 | * |
e7ca2d41 JE |
1951 | * I_DIRTY_SYNC Inode is dirty, but doesn't have to be written on |
1952 | * fdatasync(). i_atime is the usual cause. | |
000cb48e JK |
1953 | * I_DIRTY_DATASYNC Data-related inode changes pending. We keep track of |
1954 | * these changes separately from I_DIRTY_SYNC so that we | |
1955 | * don't have to write inode on fdatasync() when only | |
1956 | * mtime has changed in it. | |
1c0eeaf5 | 1957 | * I_DIRTY_PAGES Inode has dirty pages. Inode itself may be clean. |
eaff8079 CH |
1958 | * I_NEW Serves as both a mutex and completion notification. |
1959 | * New inodes set I_NEW. If two processes both create | |
1960 | * the same inode, one of them will release its inode and | |
1961 | * wait for I_NEW to be released before returning. | |
1962 | * Inodes in I_WILL_FREE, I_FREEING or I_CLEAR state can | |
1963 | * also cause waiting on I_NEW, without I_NEW actually | |
1964 | * being set. find_inode() uses this to prevent returning | |
1965 | * nearly-dead inodes. | |
1c0eeaf5 JE |
1966 | * I_WILL_FREE Must be set when calling write_inode_now() if i_count |
1967 | * is zero. I_FREEING must be set when I_WILL_FREE is | |
1968 | * cleared. | |
1969 | * I_FREEING Set when inode is about to be freed but still has dirty | |
1970 | * pages or buffers attached or the inode itself is still | |
1971 | * dirty. | |
dbd5768f JK |
1972 | * I_CLEAR Added by clear_inode(). In this state the inode is |
1973 | * clean and can be destroyed. Inode keeps I_FREEING. | |
1c0eeaf5 JE |
1974 | * |
1975 | * Inodes that are I_WILL_FREE, I_FREEING or I_CLEAR are | |
1976 | * prohibited for many purposes. iget() must wait for | |
1977 | * the inode to be completely released, then create it | |
1978 | * anew. Other functions will just ignore such inodes, | |
eaff8079 | 1979 | * if appropriate. I_NEW is used for waiting. |
1c0eeaf5 | 1980 | * |
169ebd90 JK |
1981 | * I_SYNC Writeback of inode is running. The bit is set during |
1982 | * data writeback, and cleared with a wakeup on the bit | |
1983 | * address once it is done. The bit is also used to pin | |
1984 | * the inode in memory for flusher thread. | |
1c0eeaf5 | 1985 | * |
bd5fe6c5 CH |
1986 | * I_REFERENCED Marks the inode as recently references on the LRU list. |
1987 | * | |
1988 | * I_DIO_WAKEUP Never set. Only used as a key for wait_on_bit(). | |
1989 | * | |
682aa8e1 TH |
1990 | * I_WB_SWITCH Cgroup bdi_writeback switching in progress. Used to |
1991 | * synchronize competing switching instances and to tell | |
b93b0163 | 1992 | * wb stat updates to grab the i_pages lock. See |
682aa8e1 TH |
1993 | * inode_switch_wb_work_fn() for details. |
1994 | * | |
ad0af710 AG |
1995 | * I_OVL_INUSE Used by overlayfs to get exclusive ownership on upper |
1996 | * and work dirs among overlayfs mounts. | |
1997 | * | |
1c0eeaf5 | 1998 | * Q: What is the difference between I_WILL_FREE and I_FREEING? |
1c0eeaf5 | 1999 | */ |
9e38d86f NP |
2000 | #define I_DIRTY_SYNC (1 << 0) |
2001 | #define I_DIRTY_DATASYNC (1 << 1) | |
2002 | #define I_DIRTY_PAGES (1 << 2) | |
eaff8079 CH |
2003 | #define __I_NEW 3 |
2004 | #define I_NEW (1 << __I_NEW) | |
9e38d86f NP |
2005 | #define I_WILL_FREE (1 << 4) |
2006 | #define I_FREEING (1 << 5) | |
2007 | #define I_CLEAR (1 << 6) | |
eaff8079 | 2008 | #define __I_SYNC 7 |
1c0eeaf5 | 2009 | #define I_SYNC (1 << __I_SYNC) |
9e38d86f | 2010 | #define I_REFERENCED (1 << 8) |
bd5fe6c5 | 2011 | #define __I_DIO_WAKEUP 9 |
ac74d8d6 | 2012 | #define I_DIO_WAKEUP (1 << __I_DIO_WAKEUP) |
f4e0c30c | 2013 | #define I_LINKABLE (1 << 10) |
0ae45f63 TT |
2014 | #define I_DIRTY_TIME (1 << 11) |
2015 | #define __I_DIRTY_TIME_EXPIRED 12 | |
2016 | #define I_DIRTY_TIME_EXPIRED (1 << __I_DIRTY_TIME_EXPIRED) | |
682aa8e1 | 2017 | #define I_WB_SWITCH (1 << 13) |
ad0af710 | 2018 | #define I_OVL_INUSE (1 << 14) |
1da177e4 | 2019 | |
0e11f644 CH |
2020 | #define I_DIRTY_INODE (I_DIRTY_SYNC | I_DIRTY_DATASYNC) |
2021 | #define I_DIRTY (I_DIRTY_INODE | I_DIRTY_PAGES) | |
0ae45f63 | 2022 | #define I_DIRTY_ALL (I_DIRTY | I_DIRTY_TIME) |
1da177e4 LT |
2023 | |
2024 | extern void __mark_inode_dirty(struct inode *, int); | |
2025 | static inline void mark_inode_dirty(struct inode *inode) | |
2026 | { | |
2027 | __mark_inode_dirty(inode, I_DIRTY); | |
2028 | } | |
2029 | ||
2030 | static inline void mark_inode_dirty_sync(struct inode *inode) | |
2031 | { | |
2032 | __mark_inode_dirty(inode, I_DIRTY_SYNC); | |
2033 | } | |
2034 | ||
7ada4db8 MS |
2035 | extern void inc_nlink(struct inode *inode); |
2036 | extern void drop_nlink(struct inode *inode); | |
2037 | extern void clear_nlink(struct inode *inode); | |
2038 | extern void set_nlink(struct inode *inode, unsigned int nlink); | |
d8c76e6f DH |
2039 | |
2040 | static inline void inode_inc_link_count(struct inode *inode) | |
2041 | { | |
2042 | inc_nlink(inode); | |
66d21730 AD |
2043 | mark_inode_dirty(inode); |
2044 | } | |
2045 | ||
9a53c3a7 DH |
2046 | static inline void inode_dec_link_count(struct inode *inode) |
2047 | { | |
2048 | drop_nlink(inode); | |
66d21730 AD |
2049 | mark_inode_dirty(inode); |
2050 | } | |
2051 | ||
c3b2da31 JB |
2052 | enum file_time_flags { |
2053 | S_ATIME = 1, | |
2054 | S_MTIME = 2, | |
2055 | S_CTIME = 4, | |
2056 | S_VERSION = 8, | |
2057 | }; | |
2058 | ||
badcf2b7 | 2059 | extern void touch_atime(const struct path *); |
1da177e4 LT |
2060 | static inline void file_accessed(struct file *file) |
2061 | { | |
2062 | if (!(file->f_flags & O_NOATIME)) | |
68ac1234 | 2063 | touch_atime(&file->f_path); |
1da177e4 LT |
2064 | } |
2065 | ||
2066 | int sync_inode(struct inode *inode, struct writeback_control *wbc); | |
c3765016 | 2067 | int sync_inode_metadata(struct inode *inode, int wait); |
1da177e4 | 2068 | |
1da177e4 LT |
2069 | struct file_system_type { |
2070 | const char *name; | |
2071 | int fs_flags; | |
bbc1096a DH |
2072 | #define FS_REQUIRES_DEV 1 |
2073 | #define FS_BINARY_MOUNTDATA 2 | |
2074 | #define FS_HAS_SUBTYPE 4 | |
0c55cfc4 | 2075 | #define FS_USERNS_MOUNT 8 /* Can be mounted by userns root */ |
bbc1096a | 2076 | #define FS_RENAME_DOES_D_MOVE 32768 /* FS will handle d_move() during rename() internally. */ |
c96e41e9 AV |
2077 | struct dentry *(*mount) (struct file_system_type *, int, |
2078 | const char *, void *); | |
1da177e4 LT |
2079 | void (*kill_sb) (struct super_block *); |
2080 | struct module *owner; | |
2081 | struct file_system_type * next; | |
a5166169 | 2082 | struct hlist_head fs_supers; |
d475fd42 | 2083 | |
cf516249 | 2084 | struct lock_class_key s_lock_key; |
897c6ff9 | 2085 | struct lock_class_key s_umount_key; |
51ee049e | 2086 | struct lock_class_key s_vfs_rename_key; |
5accdf82 | 2087 | struct lock_class_key s_writers_key[SB_FREEZE_LEVELS]; |
d475fd42 PZ |
2088 | |
2089 | struct lock_class_key i_lock_key; | |
2090 | struct lock_class_key i_mutex_key; | |
14358e6d | 2091 | struct lock_class_key i_mutex_dir_key; |
1da177e4 LT |
2092 | }; |
2093 | ||
7f78e035 EB |
2094 | #define MODULE_ALIAS_FS(NAME) MODULE_ALIAS("fs-" NAME) |
2095 | ||
d91ee87d EB |
2096 | extern struct dentry *mount_ns(struct file_system_type *fs_type, |
2097 | int flags, void *data, void *ns, struct user_namespace *user_ns, | |
2098 | int (*fill_super)(struct super_block *, void *, int)); | |
f175f307 | 2099 | #ifdef CONFIG_BLOCK |
152a0836 AV |
2100 | extern struct dentry *mount_bdev(struct file_system_type *fs_type, |
2101 | int flags, const char *dev_name, void *data, | |
2102 | int (*fill_super)(struct super_block *, void *, int)); | |
f175f307 AV |
2103 | #else |
2104 | static inline struct dentry *mount_bdev(struct file_system_type *fs_type, | |
2105 | int flags, const char *dev_name, void *data, | |
2106 | int (*fill_super)(struct super_block *, void *, int)) | |
2107 | { | |
2108 | return ERR_PTR(-ENODEV); | |
2109 | } | |
2110 | #endif | |
fc14f2fe AV |
2111 | extern struct dentry *mount_single(struct file_system_type *fs_type, |
2112 | int flags, void *data, | |
2113 | int (*fill_super)(struct super_block *, void *, int)); | |
3c26ff6e AV |
2114 | extern struct dentry *mount_nodev(struct file_system_type *fs_type, |
2115 | int flags, void *data, | |
2116 | int (*fill_super)(struct super_block *, void *, int)); | |
ea441d11 | 2117 | extern struct dentry *mount_subtree(struct vfsmount *mnt, const char *path); |
1da177e4 | 2118 | void generic_shutdown_super(struct super_block *sb); |
f175f307 | 2119 | #ifdef CONFIG_BLOCK |
1da177e4 | 2120 | void kill_block_super(struct super_block *sb); |
f175f307 AV |
2121 | #else |
2122 | static inline void kill_block_super(struct super_block *sb) | |
2123 | { | |
2124 | BUG(); | |
2125 | } | |
2126 | #endif | |
1da177e4 LT |
2127 | void kill_anon_super(struct super_block *sb); |
2128 | void kill_litter_super(struct super_block *sb); | |
2129 | void deactivate_super(struct super_block *sb); | |
74dbbdd7 | 2130 | void deactivate_locked_super(struct super_block *sb); |
1da177e4 | 2131 | int set_anon_super(struct super_block *s, void *data); |
0ee5dc67 AV |
2132 | int get_anon_bdev(dev_t *); |
2133 | void free_anon_bdev(dev_t); | |
6e4eab57 EB |
2134 | struct super_block *sget_userns(struct file_system_type *type, |
2135 | int (*test)(struct super_block *,void *), | |
2136 | int (*set)(struct super_block *,void *), | |
2137 | int flags, struct user_namespace *user_ns, | |
2138 | void *data); | |
1da177e4 LT |
2139 | struct super_block *sget(struct file_system_type *type, |
2140 | int (*test)(struct super_block *,void *), | |
2141 | int (*set)(struct super_block *,void *), | |
9249e17f | 2142 | int flags, void *data); |
bba0bd31 AG |
2143 | extern struct dentry *mount_pseudo_xattr(struct file_system_type *, char *, |
2144 | const struct super_operations *ops, | |
2145 | const struct xattr_handler **xattr, | |
2146 | const struct dentry_operations *dops, | |
2147 | unsigned long); | |
2148 | ||
2149 | static inline struct dentry * | |
2150 | mount_pseudo(struct file_system_type *fs_type, char *name, | |
2151 | const struct super_operations *ops, | |
2152 | const struct dentry_operations *dops, unsigned long magic) | |
2153 | { | |
2154 | return mount_pseudo_xattr(fs_type, name, ops, NULL, dops, magic); | |
2155 | } | |
1da177e4 LT |
2156 | |
2157 | /* Alas, no aliases. Too much hassle with bringing module.h everywhere */ | |
2158 | #define fops_get(fops) \ | |
2159 | (((fops) && try_module_get((fops)->owner) ? (fops) : NULL)) | |
2160 | #define fops_put(fops) \ | |
2161 | do { if (fops) module_put((fops)->owner); } while(0) | |
e84f9e57 AV |
2162 | /* |
2163 | * This one is to be used *ONLY* from ->open() instances. | |
2164 | * fops must be non-NULL, pinned down *and* module dependencies | |
2165 | * should be sufficient to pin the caller down as well. | |
2166 | */ | |
2167 | #define replace_fops(f, fops) \ | |
2168 | do { \ | |
2169 | struct file *__file = (f); \ | |
2170 | fops_put(__file->f_op); \ | |
2171 | BUG_ON(!(__file->f_op = (fops))); \ | |
2172 | } while(0) | |
1da177e4 LT |
2173 | |
2174 | extern int register_filesystem(struct file_system_type *); | |
2175 | extern int unregister_filesystem(struct file_system_type *); | |
8bf9725c PE |
2176 | extern struct vfsmount *kern_mount_data(struct file_system_type *, void *data); |
2177 | #define kern_mount(type) kern_mount_data(type, NULL) | |
423e0ab0 | 2178 | extern void kern_unmount(struct vfsmount *mnt); |
1da177e4 LT |
2179 | extern int may_umount_tree(struct vfsmount *); |
2180 | extern int may_umount(struct vfsmount *); | |
5e6123f3 SL |
2181 | extern long do_mount(const char *, const char __user *, |
2182 | const char *, unsigned long, void *); | |
ca71cf71 | 2183 | extern struct vfsmount *collect_mounts(const struct path *); |
8aec0809 | 2184 | extern void drop_collected_mounts(struct vfsmount *); |
1f707137 AV |
2185 | extern int iterate_mounts(int (*)(struct vfsmount *, void *), void *, |
2186 | struct vfsmount *); | |
f0bb5aaf | 2187 | extern int vfs_statfs(const struct path *, struct kstatfs *); |
c8b91acc AV |
2188 | extern int user_statfs(const char __user *, struct kstatfs *); |
2189 | extern int fd_statfs(int, struct kstatfs *); | |
18e9e510 JB |
2190 | extern int freeze_super(struct super_block *super); |
2191 | extern int thaw_super(struct super_block *super); | |
02125a82 | 2192 | extern bool our_mnt(struct vfsmount *mnt); |
fca39346 JK |
2193 | extern __printf(2, 3) |
2194 | int super_setup_bdi_name(struct super_block *sb, char *fmt, ...); | |
2195 | extern int super_setup_bdi(struct super_block *sb); | |
1da177e4 | 2196 | |
ce3b0f8d AV |
2197 | extern int current_umask(void); |
2198 | ||
5a14696c BF |
2199 | extern void ihold(struct inode * inode); |
2200 | extern void iput(struct inode *); | |
0ae45f63 | 2201 | extern int generic_update_time(struct inode *, struct timespec *, int); |
5a14696c | 2202 | |
f87fd4c2 | 2203 | /* /sys/fs */ |
00d26666 | 2204 | extern struct kobject *fs_kobj; |
f87fd4c2 | 2205 | |
09cbfeaf | 2206 | #define MAX_RW_COUNT (INT_MAX & PAGE_MASK) |
bfcd17a6 | 2207 | |
9e8925b6 | 2208 | #ifdef CONFIG_MANDATORY_FILE_LOCKING |
d7a06983 | 2209 | extern int locks_mandatory_locked(struct file *); |
acc15575 | 2210 | extern int locks_mandatory_area(struct inode *, struct file *, loff_t, loff_t, unsigned char); |
1da177e4 LT |
2211 | |
2212 | /* | |
2213 | * Candidates for mandatory locking have the setgid bit set | |
2214 | * but no group execute bit - an otherwise meaningless combination. | |
2215 | */ | |
a16877ca PE |
2216 | |
2217 | static inline int __mandatory_lock(struct inode *ino) | |
2218 | { | |
2219 | return (ino->i_mode & (S_ISGID | S_IXGRP)) == S_ISGID; | |
2220 | } | |
2221 | ||
2222 | /* | |
e462ec50 | 2223 | * ... and these candidates should be on SB_MANDLOCK mounted fs, |
a16877ca PE |
2224 | * otherwise these will be advisory locks |
2225 | */ | |
2226 | ||
2227 | static inline int mandatory_lock(struct inode *ino) | |
2228 | { | |
2229 | return IS_MANDLOCK(ino) && __mandatory_lock(ino); | |
2230 | } | |
1da177e4 | 2231 | |
d7a06983 | 2232 | static inline int locks_verify_locked(struct file *file) |
1da177e4 | 2233 | { |
c568d683 | 2234 | if (mandatory_lock(locks_inode(file))) |
d7a06983 | 2235 | return locks_mandatory_locked(file); |
1da177e4 LT |
2236 | return 0; |
2237 | } | |
2238 | ||
1da177e4 | 2239 | static inline int locks_verify_truncate(struct inode *inode, |
acc15575 | 2240 | struct file *f, |
1da177e4 LT |
2241 | loff_t size) |
2242 | { | |
acc15575 CH |
2243 | if (!inode->i_flctx || !mandatory_lock(inode)) |
2244 | return 0; | |
2245 | ||
2246 | if (size < inode->i_size) { | |
2247 | return locks_mandatory_area(inode, f, size, inode->i_size - 1, | |
2248 | F_WRLCK); | |
2249 | } else { | |
2250 | return locks_mandatory_area(inode, f, inode->i_size, size - 1, | |
2251 | F_WRLCK); | |
2252 | } | |
1da177e4 LT |
2253 | } |
2254 | ||
9e8925b6 JL |
2255 | #else /* !CONFIG_MANDATORY_FILE_LOCKING */ |
2256 | ||
2257 | static inline int locks_mandatory_locked(struct file *file) | |
2258 | { | |
2259 | return 0; | |
2260 | } | |
2261 | ||
fce205e9 LT |
2262 | static inline int locks_mandatory_area(struct inode *inode, struct file *filp, |
2263 | loff_t start, loff_t end, unsigned char type) | |
9e8925b6 JL |
2264 | { |
2265 | return 0; | |
2266 | } | |
2267 | ||
2268 | static inline int __mandatory_lock(struct inode *inode) | |
2269 | { | |
2270 | return 0; | |
2271 | } | |
2272 | ||
2273 | static inline int mandatory_lock(struct inode *inode) | |
2274 | { | |
2275 | return 0; | |
2276 | } | |
2277 | ||
2278 | static inline int locks_verify_locked(struct file *file) | |
2279 | { | |
2280 | return 0; | |
2281 | } | |
2282 | ||
2283 | static inline int locks_verify_truncate(struct inode *inode, struct file *filp, | |
2284 | size_t size) | |
2285 | { | |
2286 | return 0; | |
2287 | } | |
2288 | ||
2289 | #endif /* CONFIG_MANDATORY_FILE_LOCKING */ | |
2290 | ||
2291 | ||
2292 | #ifdef CONFIG_FILE_LOCKING | |
1da177e4 LT |
2293 | static inline int break_lease(struct inode *inode, unsigned int mode) |
2294 | { | |
24cbe784 JL |
2295 | /* |
2296 | * Since this check is lockless, we must ensure that any refcounts | |
8116bf4c JL |
2297 | * taken are done before checking i_flctx->flc_lease. Otherwise, we |
2298 | * could end up racing with tasks trying to set a new lease on this | |
2299 | * file. | |
24cbe784 JL |
2300 | */ |
2301 | smp_mb(); | |
8634b51f | 2302 | if (inode->i_flctx && !list_empty_careful(&inode->i_flctx->flc_lease)) |
df4e8d2c | 2303 | return __break_lease(inode, mode, FL_LEASE); |
1da177e4 LT |
2304 | return 0; |
2305 | } | |
df4e8d2c BF |
2306 | |
2307 | static inline int break_deleg(struct inode *inode, unsigned int mode) | |
2308 | { | |
962bd40b JL |
2309 | /* |
2310 | * Since this check is lockless, we must ensure that any refcounts | |
8116bf4c JL |
2311 | * taken are done before checking i_flctx->flc_lease. Otherwise, we |
2312 | * could end up racing with tasks trying to set a new lease on this | |
2313 | * file. | |
962bd40b JL |
2314 | */ |
2315 | smp_mb(); | |
8634b51f | 2316 | if (inode->i_flctx && !list_empty_careful(&inode->i_flctx->flc_lease)) |
df4e8d2c | 2317 | return __break_lease(inode, mode, FL_DELEG); |
1da177e4 LT |
2318 | return 0; |
2319 | } | |
df4e8d2c | 2320 | |
5a14696c BF |
2321 | static inline int try_break_deleg(struct inode *inode, struct inode **delegated_inode) |
2322 | { | |
2323 | int ret; | |
2324 | ||
2325 | ret = break_deleg(inode, O_WRONLY|O_NONBLOCK); | |
2326 | if (ret == -EWOULDBLOCK && delegated_inode) { | |
2327 | *delegated_inode = inode; | |
2328 | ihold(inode); | |
2329 | } | |
2330 | return ret; | |
2331 | } | |
2332 | ||
2333 | static inline int break_deleg_wait(struct inode **delegated_inode) | |
2334 | { | |
2335 | int ret; | |
2336 | ||
2337 | ret = break_deleg(*delegated_inode, O_WRONLY); | |
2338 | iput(*delegated_inode); | |
2339 | *delegated_inode = NULL; | |
2340 | return ret; | |
2341 | } | |
2342 | ||
11afe9f7 CH |
2343 | static inline int break_layout(struct inode *inode, bool wait) |
2344 | { | |
2345 | smp_mb(); | |
2346 | if (inode->i_flctx && !list_empty_careful(&inode->i_flctx->flc_lease)) | |
2347 | return __break_lease(inode, | |
2348 | wait ? O_WRONLY : O_WRONLY | O_NONBLOCK, | |
2349 | FL_LAYOUT); | |
2350 | return 0; | |
2351 | } | |
2352 | ||
bfcd17a6 | 2353 | #else /* !CONFIG_FILE_LOCKING */ |
af5df566 SW |
2354 | static inline int break_lease(struct inode *inode, unsigned int mode) |
2355 | { | |
2356 | return 0; | |
2357 | } | |
2358 | ||
df4e8d2c BF |
2359 | static inline int break_deleg(struct inode *inode, unsigned int mode) |
2360 | { | |
2361 | return 0; | |
2362 | } | |
5a14696c BF |
2363 | |
2364 | static inline int try_break_deleg(struct inode *inode, struct inode **delegated_inode) | |
2365 | { | |
2366 | return 0; | |
2367 | } | |
2368 | ||
2369 | static inline int break_deleg_wait(struct inode **delegated_inode) | |
2370 | { | |
2371 | BUG(); | |
2372 | return 0; | |
2373 | } | |
2374 | ||
11afe9f7 CH |
2375 | static inline int break_layout(struct inode *inode, bool wait) |
2376 | { | |
2377 | return 0; | |
2378 | } | |
2379 | ||
bfcd17a6 | 2380 | #endif /* CONFIG_FILE_LOCKING */ |
1da177e4 LT |
2381 | |
2382 | /* fs/open.c */ | |
adb5c247 | 2383 | struct audit_names; |
91a27b2a | 2384 | struct filename { |
adb5c247 JL |
2385 | const char *name; /* pointer to actual string */ |
2386 | const __user char *uptr; /* original userland pointer */ | |
55422d0b | 2387 | int refcnt; |
1c949843 | 2388 | struct audit_names *aname; |
fd2f7cb5 | 2389 | const char iname[]; |
91a27b2a | 2390 | }; |
1da177e4 | 2391 | |
7df818b2 | 2392 | extern long vfs_truncate(const struct path *, loff_t); |
4a30131e N |
2393 | extern int do_truncate(struct dentry *, loff_t start, unsigned int time_attrs, |
2394 | struct file *filp); | |
72c72bdf | 2395 | extern int vfs_fallocate(struct file *file, int mode, loff_t offset, |
3e63cbb1 | 2396 | loff_t len); |
8e8a1407 | 2397 | extern long do_sys_open(int dfd, const char __user *filename, int flags, |
a218d0fd | 2398 | umode_t mode); |
669abf4e | 2399 | extern struct file *file_open_name(struct filename *, int, umode_t); |
a218d0fd | 2400 | extern struct file *filp_open(const char *, int, umode_t); |
73d049a4 | 2401 | extern struct file *file_open_root(struct dentry *, struct vfsmount *, |
378c6520 | 2402 | const char *, int, umode_t); |
765927b2 | 2403 | extern struct file * dentry_open(const struct path *, int, const struct cred *); |
1da177e4 | 2404 | extern int filp_close(struct file *, fl_owner_t id); |
91a27b2a | 2405 | |
51f39a1f | 2406 | extern struct filename *getname_flags(const char __user *, int, int *); |
91a27b2a | 2407 | extern struct filename *getname(const char __user *); |
c4ad8f98 | 2408 | extern struct filename *getname_kernel(const char *); |
55422d0b | 2409 | extern void putname(struct filename *name); |
91a27b2a | 2410 | |
47237687 AV |
2411 | enum { |
2412 | FILE_CREATED = 1, | |
2413 | FILE_OPENED = 2 | |
2414 | }; | |
30d90494 AV |
2415 | extern int finish_open(struct file *file, struct dentry *dentry, |
2416 | int (*open)(struct inode *, struct file *), | |
2417 | int *opened); | |
e45198a6 | 2418 | extern int finish_no_open(struct file *file, struct dentry *dentry); |
1da177e4 | 2419 | |
3e63cbb1 AJ |
2420 | /* fs/ioctl.c */ |
2421 | ||
2422 | extern int ioctl_preallocate(struct file *filp, void __user *argp); | |
2423 | ||
1da177e4 LT |
2424 | /* fs/dcache.c */ |
2425 | extern void __init vfs_caches_init_early(void); | |
4248b0da | 2426 | extern void __init vfs_caches_init(void); |
1da177e4 | 2427 | |
b86c089b CL |
2428 | extern struct kmem_cache *names_cachep; |
2429 | ||
a608ca21 | 2430 | #define __getname() kmem_cache_alloc(names_cachep, GFP_KERNEL) |
3446a8aa | 2431 | #define __putname(name) kmem_cache_free(names_cachep, (void *)(name)) |
1da177e4 | 2432 | |
9361401e | 2433 | #ifdef CONFIG_BLOCK |
1da177e4 | 2434 | extern int register_blkdev(unsigned int, const char *); |
f4480240 | 2435 | extern void unregister_blkdev(unsigned int, const char *); |
f44f1ab5 | 2436 | extern void bdev_unhash_inode(dev_t dev); |
1da177e4 | 2437 | extern struct block_device *bdget(dev_t); |
dddac6a7 | 2438 | extern struct block_device *bdgrab(struct block_device *bdev); |
1da177e4 LT |
2439 | extern void bd_set_size(struct block_device *, loff_t size); |
2440 | extern void bd_forget(struct inode *inode); | |
2441 | extern void bdput(struct block_device *); | |
585d3bc0 | 2442 | extern void invalidate_bdev(struct block_device *); |
5c0d6b60 | 2443 | extern void iterate_bdevs(void (*)(struct block_device *, void *), void *); |
585d3bc0 | 2444 | extern int sync_blockdev(struct block_device *bdev); |
ff01bb48 | 2445 | extern void kill_bdev(struct block_device *); |
585d3bc0 | 2446 | extern struct super_block *freeze_bdev(struct block_device *); |
c2d75438 | 2447 | extern void emergency_thaw_all(void); |
08fdc8a0 | 2448 | extern void emergency_thaw_bdev(struct super_block *sb); |
585d3bc0 NP |
2449 | extern int thaw_bdev(struct block_device *bdev, struct super_block *sb); |
2450 | extern int fsync_bdev(struct block_device *); | |
a212b105 TH |
2451 | |
2452 | extern struct super_block *blockdev_superblock; | |
2453 | ||
2454 | static inline bool sb_is_blkdev_sb(struct super_block *sb) | |
2455 | { | |
2456 | return sb == blockdev_superblock; | |
2457 | } | |
9361401e DH |
2458 | #else |
2459 | static inline void bd_forget(struct inode *inode) {} | |
47e4491b | 2460 | static inline int sync_blockdev(struct block_device *bdev) { return 0; } |
ff01bb48 | 2461 | static inline void kill_bdev(struct block_device *bdev) {} |
47e4491b AV |
2462 | static inline void invalidate_bdev(struct block_device *bdev) {} |
2463 | ||
2464 | static inline struct super_block *freeze_bdev(struct block_device *sb) | |
2465 | { | |
2466 | return NULL; | |
2467 | } | |
2468 | ||
2469 | static inline int thaw_bdev(struct block_device *bdev, struct super_block *sb) | |
2470 | { | |
2471 | return 0; | |
2472 | } | |
5c0d6b60 | 2473 | |
08fdc8a0 MG |
2474 | static inline int emergency_thaw_bdev(struct super_block *sb) |
2475 | { | |
2476 | return 0; | |
2477 | } | |
2478 | ||
5c0d6b60 JK |
2479 | static inline void iterate_bdevs(void (*f)(struct block_device *, void *), void *arg) |
2480 | { | |
2481 | } | |
a8855990 | 2482 | |
3e004eea | 2483 | static inline bool sb_is_blkdev_sb(struct super_block *sb) |
a8855990 | 2484 | { |
3e004eea | 2485 | return false; |
a8855990 | 2486 | } |
9361401e | 2487 | #endif |
60b0680f | 2488 | extern int sync_filesystem(struct super_block *); |
9361401e | 2489 | extern const struct file_operations def_blk_fops; |
4b6f5d20 | 2490 | extern const struct file_operations def_chr_fops; |
9361401e | 2491 | #ifdef CONFIG_BLOCK |
1da177e4 | 2492 | extern int ioctl_by_bdev(struct block_device *, unsigned, unsigned long); |
56b26add | 2493 | extern int blkdev_ioctl(struct block_device *, fmode_t, unsigned, unsigned long); |
1da177e4 | 2494 | extern long compat_blkdev_ioctl(struct file *, unsigned, unsigned long); |
e525fd89 | 2495 | extern int blkdev_get(struct block_device *bdev, fmode_t mode, void *holder); |
d4d77629 TH |
2496 | extern struct block_device *blkdev_get_by_path(const char *path, fmode_t mode, |
2497 | void *holder); | |
2498 | extern struct block_device *blkdev_get_by_dev(dev_t dev, fmode_t mode, | |
2499 | void *holder); | |
4385bab1 | 2500 | extern void blkdev_put(struct block_device *bdev, fmode_t mode); |
be324177 JW |
2501 | extern int __blkdev_reread_part(struct block_device *bdev); |
2502 | extern int blkdev_reread_part(struct block_device *bdev); | |
2503 | ||
641dc636 | 2504 | #ifdef CONFIG_SYSFS |
e09b457b | 2505 | extern int bd_link_disk_holder(struct block_device *bdev, struct gendisk *disk); |
49731baa TH |
2506 | extern void bd_unlink_disk_holder(struct block_device *bdev, |
2507 | struct gendisk *disk); | |
641dc636 | 2508 | #else |
e09b457b TH |
2509 | static inline int bd_link_disk_holder(struct block_device *bdev, |
2510 | struct gendisk *disk) | |
2511 | { | |
2512 | return 0; | |
2513 | } | |
49731baa TH |
2514 | static inline void bd_unlink_disk_holder(struct block_device *bdev, |
2515 | struct gendisk *disk) | |
2516 | { | |
2517 | } | |
641dc636 | 2518 | #endif |
9361401e | 2519 | #endif |
1da177e4 LT |
2520 | |
2521 | /* fs/char_dev.c */ | |
8a932f73 | 2522 | #define CHRDEV_MAJOR_MAX 512 |
49db08c3 LW |
2523 | /* Marks the bottom of the first segment of free char majors */ |
2524 | #define CHRDEV_MAJOR_DYN_END 234 | |
a5d31a3f LG |
2525 | /* Marks the top and bottom of the second segment of free char majors */ |
2526 | #define CHRDEV_MAJOR_DYN_EXT_START 511 | |
2527 | #define CHRDEV_MAJOR_DYN_EXT_END 384 | |
2528 | ||
1da177e4 LT |
2529 | extern int alloc_chrdev_region(dev_t *, unsigned, unsigned, const char *); |
2530 | extern int register_chrdev_region(dev_t, unsigned, const char *); | |
1905b1bf TH |
2531 | extern int __register_chrdev(unsigned int major, unsigned int baseminor, |
2532 | unsigned int count, const char *name, | |
2533 | const struct file_operations *fops); | |
2534 | extern void __unregister_chrdev(unsigned int major, unsigned int baseminor, | |
2535 | unsigned int count, const char *name); | |
1da177e4 | 2536 | extern void unregister_chrdev_region(dev_t, unsigned); |
68eef3b4 | 2537 | extern void chrdev_show(struct seq_file *,off_t); |
1da177e4 | 2538 | |
1905b1bf TH |
2539 | static inline int register_chrdev(unsigned int major, const char *name, |
2540 | const struct file_operations *fops) | |
2541 | { | |
2542 | return __register_chrdev(major, 0, 256, name, fops); | |
2543 | } | |
2544 | ||
2545 | static inline void unregister_chrdev(unsigned int major, const char *name) | |
2546 | { | |
2547 | __unregister_chrdev(major, 0, 256, name); | |
2548 | } | |
2549 | ||
1da177e4 LT |
2550 | /* fs/block_dev.c */ |
2551 | #define BDEVNAME_SIZE 32 /* Largest string for a blockdev identifier */ | |
1f014290 | 2552 | #define BDEVT_SIZE 10 /* Largest string for MAJ:MIN for blkdev */ |
9361401e DH |
2553 | |
2554 | #ifdef CONFIG_BLOCK | |
133d55cd | 2555 | #define BLKDEV_MAJOR_MAX 512 |
1da177e4 LT |
2556 | extern const char *__bdevname(dev_t, char *buffer); |
2557 | extern const char *bdevname(struct block_device *bdev, char *buffer); | |
2558 | extern struct block_device *lookup_bdev(const char *); | |
68eef3b4 | 2559 | extern void blkdev_show(struct seq_file *,off_t); |
30c40d2c | 2560 | |
9361401e | 2561 | #else |
133d55cd | 2562 | #define BLKDEV_MAJOR_MAX 0 |
9361401e | 2563 | #endif |
1da177e4 LT |
2564 | |
2565 | extern void init_special_inode(struct inode *, umode_t, dev_t); | |
2566 | ||
2567 | /* Invalid inode operations -- fs/bad_inode.c */ | |
2568 | extern void make_bad_inode(struct inode *); | |
0e3ef1fe | 2569 | extern bool is_bad_inode(struct inode *); |
1da177e4 | 2570 | |
c11f0c0b | 2571 | #ifdef CONFIG_BLOCK |
c3279d14 AP |
2572 | extern void check_disk_size_change(struct gendisk *disk, |
2573 | struct block_device *bdev); | |
0c002c2f | 2574 | extern int revalidate_disk(struct gendisk *); |
1da177e4 | 2575 | extern int check_disk_change(struct block_device *); |
93b270f7 | 2576 | extern int __invalidate_device(struct block_device *, bool); |
1da177e4 | 2577 | extern int invalidate_partition(struct gendisk *, int); |
9361401e | 2578 | #endif |
1da177e4 LT |
2579 | unsigned long invalidate_mapping_pages(struct address_space *mapping, |
2580 | pgoff_t start, pgoff_t end); | |
54bc4855 | 2581 | |
1da177e4 LT |
2582 | static inline void invalidate_remote_inode(struct inode *inode) |
2583 | { | |
2584 | if (S_ISREG(inode->i_mode) || S_ISDIR(inode->i_mode) || | |
2585 | S_ISLNK(inode->i_mode)) | |
fc0ecff6 | 2586 | invalidate_mapping_pages(inode->i_mapping, 0, -1); |
1da177e4 LT |
2587 | } |
2588 | extern int invalidate_inode_pages2(struct address_space *mapping); | |
2589 | extern int invalidate_inode_pages2_range(struct address_space *mapping, | |
2590 | pgoff_t start, pgoff_t end); | |
2591 | extern int write_inode_now(struct inode *, int); | |
2592 | extern int filemap_fdatawrite(struct address_space *); | |
2593 | extern int filemap_flush(struct address_space *); | |
76341cab | 2594 | extern int filemap_fdatawait_keep_errors(struct address_space *mapping); |
d3bccb6f JK |
2595 | extern int filemap_fdatawait_range(struct address_space *, loff_t lstart, |
2596 | loff_t lend); | |
ffb959bb JL |
2597 | |
2598 | static inline int filemap_fdatawait(struct address_space *mapping) | |
2599 | { | |
2600 | return filemap_fdatawait_range(mapping, 0, LLONG_MAX); | |
2601 | } | |
2602 | ||
7fc9e472 GR |
2603 | extern bool filemap_range_has_page(struct address_space *, loff_t lstart, |
2604 | loff_t lend); | |
a823e458 JL |
2605 | extern int __must_check file_fdatawait_range(struct file *file, loff_t lstart, |
2606 | loff_t lend); | |
1da177e4 LT |
2607 | extern int filemap_write_and_wait(struct address_space *mapping); |
2608 | extern int filemap_write_and_wait_range(struct address_space *mapping, | |
2609 | loff_t lstart, loff_t lend); | |
ebcf28e1 AM |
2610 | extern int __filemap_fdatawrite_range(struct address_space *mapping, |
2611 | loff_t start, loff_t end, int sync_mode); | |
f4c0a0fd JK |
2612 | extern int filemap_fdatawrite_range(struct address_space *mapping, |
2613 | loff_t start, loff_t end); | |
d72d9e2a | 2614 | extern int filemap_check_errors(struct address_space *mapping); |
5660e13d | 2615 | extern void __filemap_set_wb_err(struct address_space *mapping, int err); |
a823e458 JL |
2616 | |
2617 | extern int __must_check file_fdatawait_range(struct file *file, loff_t lstart, | |
2618 | loff_t lend); | |
5660e13d JL |
2619 | extern int __must_check file_check_and_advance_wb_err(struct file *file); |
2620 | extern int __must_check file_write_and_wait_range(struct file *file, | |
2621 | loff_t start, loff_t end); | |
2622 | ||
a823e458 JL |
2623 | static inline int file_write_and_wait(struct file *file) |
2624 | { | |
2625 | return file_write_and_wait_range(file, 0, LLONG_MAX); | |
2626 | } | |
2627 | ||
5660e13d JL |
2628 | /** |
2629 | * filemap_set_wb_err - set a writeback error on an address_space | |
2630 | * @mapping: mapping in which to set writeback error | |
2631 | * @err: error to be set in mapping | |
2632 | * | |
2633 | * When writeback fails in some way, we must record that error so that | |
2634 | * userspace can be informed when fsync and the like are called. We endeavor | |
2635 | * to report errors on any file that was open at the time of the error. Some | |
2636 | * internal callers also need to know when writeback errors have occurred. | |
2637 | * | |
2638 | * When a writeback error occurs, most filesystems will want to call | |
2639 | * filemap_set_wb_err to record the error in the mapping so that it will be | |
2640 | * automatically reported whenever fsync is called on the file. | |
5660e13d JL |
2641 | */ |
2642 | static inline void filemap_set_wb_err(struct address_space *mapping, int err) | |
2643 | { | |
2644 | /* Fastpath for common case of no error */ | |
2645 | if (unlikely(err)) | |
2646 | __filemap_set_wb_err(mapping, err); | |
2647 | } | |
2648 | ||
2649 | /** | |
2650 | * filemap_check_wb_error - has an error occurred since the mark was sampled? | |
2651 | * @mapping: mapping to check for writeback errors | |
2652 | * @since: previously-sampled errseq_t | |
2653 | * | |
2654 | * Grab the errseq_t value from the mapping, and see if it has changed "since" | |
2655 | * the given value was sampled. | |
2656 | * | |
2657 | * If it has then report the latest error set, otherwise return 0. | |
2658 | */ | |
2659 | static inline int filemap_check_wb_err(struct address_space *mapping, | |
2660 | errseq_t since) | |
2661 | { | |
2662 | return errseq_check(&mapping->wb_err, since); | |
2663 | } | |
2664 | ||
2665 | /** | |
2666 | * filemap_sample_wb_err - sample the current errseq_t to test for later errors | |
2667 | * @mapping: mapping to be sampled | |
2668 | * | |
2669 | * Writeback errors are always reported relative to a particular sample point | |
2670 | * in the past. This function provides those sample points. | |
2671 | */ | |
2672 | static inline errseq_t filemap_sample_wb_err(struct address_space *mapping) | |
2673 | { | |
2674 | return errseq_sample(&mapping->wb_err); | |
2675 | } | |
2676 | ||
8018ab05 CH |
2677 | extern int vfs_fsync_range(struct file *file, loff_t start, loff_t end, |
2678 | int datasync); | |
2679 | extern int vfs_fsync(struct file *file, int datasync); | |
e2592217 CH |
2680 | |
2681 | /* | |
2682 | * Sync the bytes written if this was a synchronous write. Expect ki_pos | |
2683 | * to already be updated for the write, and will return either the amount | |
2684 | * of bytes passed in, or an error if syncing the file failed. | |
2685 | */ | |
2686 | static inline ssize_t generic_write_sync(struct kiocb *iocb, ssize_t count) | |
2687 | { | |
2688 | if (iocb->ki_flags & IOCB_DSYNC) { | |
2689 | int ret = vfs_fsync_range(iocb->ki_filp, | |
2690 | iocb->ki_pos - count, iocb->ki_pos - 1, | |
2691 | (iocb->ki_flags & IOCB_SYNC) ? 0 : 1); | |
2692 | if (ret) | |
2693 | return ret; | |
2694 | } | |
2695 | ||
2696 | return count; | |
d311d79d | 2697 | } |
e2592217 | 2698 | |
1da177e4 LT |
2699 | extern void emergency_sync(void); |
2700 | extern void emergency_remount(void); | |
9361401e | 2701 | #ifdef CONFIG_BLOCK |
1da177e4 | 2702 | extern sector_t bmap(struct inode *, sector_t); |
9361401e | 2703 | #endif |
27ac0ffe | 2704 | extern int notify_change(struct dentry *, struct iattr *, struct inode **); |
f419a2e3 | 2705 | extern int inode_permission(struct inode *, int); |
2830ba7f | 2706 | extern int generic_permission(struct inode *, int); |
cbdf35bc | 2707 | extern int __check_sticky(struct inode *dir, struct inode *inode); |
1da177e4 | 2708 | |
f696a365 MS |
2709 | static inline bool execute_ok(struct inode *inode) |
2710 | { | |
2711 | return (inode->i_mode & S_IXUGO) || S_ISDIR(inode->i_mode); | |
2712 | } | |
2713 | ||
8d71db4f AV |
2714 | static inline void file_start_write(struct file *file) |
2715 | { | |
2716 | if (!S_ISREG(file_inode(file)->i_mode)) | |
2717 | return; | |
2718 | __sb_start_write(file_inode(file)->i_sb, SB_FREEZE_WRITE, true); | |
2719 | } | |
2720 | ||
5ae98f15 JK |
2721 | static inline bool file_start_write_trylock(struct file *file) |
2722 | { | |
2723 | if (!S_ISREG(file_inode(file)->i_mode)) | |
2724 | return true; | |
2725 | return __sb_start_write(file_inode(file)->i_sb, SB_FREEZE_WRITE, false); | |
2726 | } | |
2727 | ||
8d71db4f AV |
2728 | static inline void file_end_write(struct file *file) |
2729 | { | |
2730 | if (!S_ISREG(file_inode(file)->i_mode)) | |
2731 | return; | |
2732 | __sb_end_write(file_inode(file)->i_sb, SB_FREEZE_WRITE); | |
2733 | } | |
2734 | ||
bfe219d3 AG |
2735 | static inline int do_clone_file_range(struct file *file_in, loff_t pos_in, |
2736 | struct file *file_out, loff_t pos_out, | |
2737 | u64 len) | |
2738 | { | |
2739 | int ret; | |
2740 | ||
2741 | file_start_write(file_out); | |
2742 | ret = vfs_clone_file_range(file_in, pos_in, file_out, pos_out, len); | |
2743 | file_end_write(file_out); | |
2744 | ||
2745 | return ret; | |
2746 | } | |
2747 | ||
07b8ce1e AV |
2748 | /* |
2749 | * get_write_access() gets write permission for a file. | |
2750 | * put_write_access() releases this write permission. | |
2751 | * This is used for regular files. | |
2752 | * We cannot support write (and maybe mmap read-write shared) accesses and | |
2753 | * MAP_DENYWRITE mmappings simultaneously. The i_writecount field of an inode | |
2754 | * can have the following values: | |
2755 | * 0: no writers, no VM_DENYWRITE mappings | |
2756 | * < 0: (-i_writecount) vm_area_structs with VM_DENYWRITE set exist | |
2757 | * > 0: (i_writecount) users are writing to the file. | |
2758 | * | |
2759 | * Normally we operate on that counter with atomic_{inc,dec} and it's safe | |
2760 | * except for the cases where we don't hold i_writecount yet. Then we need to | |
2761 | * use {get,deny}_write_access() - these functions check the sign and refuse | |
2762 | * to do the change if sign is wrong. | |
2763 | */ | |
2764 | static inline int get_write_access(struct inode *inode) | |
2765 | { | |
2766 | return atomic_inc_unless_negative(&inode->i_writecount) ? 0 : -ETXTBSY; | |
2767 | } | |
2768 | static inline int deny_write_access(struct file *file) | |
2769 | { | |
496ad9aa | 2770 | struct inode *inode = file_inode(file); |
07b8ce1e AV |
2771 | return atomic_dec_unless_positive(&inode->i_writecount) ? 0 : -ETXTBSY; |
2772 | } | |
1da177e4 LT |
2773 | static inline void put_write_access(struct inode * inode) |
2774 | { | |
2775 | atomic_dec(&inode->i_writecount); | |
2776 | } | |
2777 | static inline void allow_write_access(struct file *file) | |
2778 | { | |
2779 | if (file) | |
496ad9aa | 2780 | atomic_inc(&file_inode(file)->i_writecount); |
1da177e4 | 2781 | } |
f1fe29b4 DH |
2782 | static inline bool inode_is_open_for_write(const struct inode *inode) |
2783 | { | |
2784 | return atomic_read(&inode->i_writecount) > 0; | |
2785 | } | |
2786 | ||
a5c96ebf MZ |
2787 | #ifdef CONFIG_IMA |
2788 | static inline void i_readcount_dec(struct inode *inode) | |
2789 | { | |
2790 | BUG_ON(!atomic_read(&inode->i_readcount)); | |
2791 | atomic_dec(&inode->i_readcount); | |
2792 | } | |
2793 | static inline void i_readcount_inc(struct inode *inode) | |
2794 | { | |
2795 | atomic_inc(&inode->i_readcount); | |
2796 | } | |
2797 | #else | |
2798 | static inline void i_readcount_dec(struct inode *inode) | |
2799 | { | |
2800 | return; | |
2801 | } | |
2802 | static inline void i_readcount_inc(struct inode *inode) | |
2803 | { | |
2804 | return; | |
2805 | } | |
2806 | #endif | |
ed8cae8b | 2807 | extern int do_pipe_flags(int *, int); |
1da177e4 | 2808 | |
1284ab5b MZ |
2809 | #define __kernel_read_file_id(id) \ |
2810 | id(UNKNOWN, unknown) \ | |
2811 | id(FIRMWARE, firmware) \ | |
a098ecd2 | 2812 | id(FIRMWARE_PREALLOC_BUFFER, firmware) \ |
1284ab5b MZ |
2813 | id(MODULE, kernel-module) \ |
2814 | id(KEXEC_IMAGE, kexec-image) \ | |
2815 | id(KEXEC_INITRAMFS, kexec-initramfs) \ | |
2816 | id(POLICY, security-policy) \ | |
a7d3d039 | 2817 | id(X509_CERTIFICATE, x509-certificate) \ |
1284ab5b MZ |
2818 | id(MAX_ID, ) |
2819 | ||
2820 | #define __fid_enumify(ENUM, dummy) READING_ ## ENUM, | |
2821 | #define __fid_stringify(dummy, str) #str, | |
2822 | ||
bc8ca5b9 | 2823 | enum kernel_read_file_id { |
1284ab5b MZ |
2824 | __kernel_read_file_id(__fid_enumify) |
2825 | }; | |
2826 | ||
2827 | static const char * const kernel_read_file_str[] = { | |
2828 | __kernel_read_file_id(__fid_stringify) | |
bc8ca5b9 MZ |
2829 | }; |
2830 | ||
da20dfe6 | 2831 | static inline const char *kernel_read_file_id_str(enum kernel_read_file_id id) |
1284ab5b | 2832 | { |
cbfd0c10 | 2833 | if ((unsigned)id >= READING_MAX_ID) |
1284ab5b MZ |
2834 | return kernel_read_file_str[READING_UNKNOWN]; |
2835 | ||
2836 | return kernel_read_file_str[id]; | |
2837 | } | |
2838 | ||
bc8ca5b9 MZ |
2839 | extern int kernel_read_file(struct file *, void **, loff_t *, loff_t, |
2840 | enum kernel_read_file_id); | |
711aab1d | 2841 | extern int kernel_read_file_from_path(const char *, void **, loff_t *, loff_t, |
09596b94 | 2842 | enum kernel_read_file_id); |
b844f0ec MZ |
2843 | extern int kernel_read_file_from_fd(int, void **, loff_t *, loff_t, |
2844 | enum kernel_read_file_id); | |
bdd1d2d3 | 2845 | extern ssize_t kernel_read(struct file *, void *, size_t, loff_t *); |
e13ec939 | 2846 | extern ssize_t kernel_write(struct file *, const void *, size_t, loff_t *); |
73e18f7c | 2847 | extern ssize_t __kernel_write(struct file *, const void *, size_t, loff_t *); |
1da177e4 LT |
2848 | extern struct file * open_exec(const char *); |
2849 | ||
2850 | /* fs/dcache.c -- generic fs support functions */ | |
a6e5787f | 2851 | extern bool is_subdir(struct dentry *, struct dentry *); |
640eb7e7 | 2852 | extern bool path_is_under(const struct path *, const struct path *); |
1da177e4 | 2853 | |
9bf39ab2 MS |
2854 | extern char *file_path(struct file *, char *, int); |
2855 | ||
1da177e4 LT |
2856 | #include <linux/err.h> |
2857 | ||
2858 | /* needed for stackable file system support */ | |
965c8e59 | 2859 | extern loff_t default_llseek(struct file *file, loff_t offset, int whence); |
1da177e4 | 2860 | |
965c8e59 | 2861 | extern loff_t vfs_llseek(struct file *file, loff_t offset, int whence); |
1da177e4 | 2862 | |
54e34621 | 2863 | extern int inode_init_always(struct super_block *, struct inode *); |
1da177e4 | 2864 | extern void inode_init_once(struct inode *); |
2aa15890 | 2865 | extern void address_space_init_once(struct address_space *mapping); |
1da177e4 LT |
2866 | extern struct inode * igrab(struct inode *); |
2867 | extern ino_t iunique(struct super_block *, ino_t); | |
2868 | extern int inode_needs_sync(struct inode *inode); | |
45321ac5 | 2869 | extern int generic_delete_inode(struct inode *inode); |
9bcb4b73 AV |
2870 | static inline int generic_drop_inode(struct inode *inode) |
2871 | { | |
2872 | return !inode->i_nlink || inode_unhashed(inode); | |
2873 | } | |
1da177e4 | 2874 | |
88bd5121 AA |
2875 | extern struct inode *ilookup5_nowait(struct super_block *sb, |
2876 | unsigned long hashval, int (*test)(struct inode *, void *), | |
2877 | void *data); | |
1da177e4 LT |
2878 | extern struct inode *ilookup5(struct super_block *sb, unsigned long hashval, |
2879 | int (*test)(struct inode *, void *), void *data); | |
2880 | extern struct inode *ilookup(struct super_block *sb, unsigned long ino); | |
2881 | ||
2882 | extern struct inode * iget5_locked(struct super_block *, unsigned long, int (*test)(struct inode *, void *), int (*set)(struct inode *, void *), void *); | |
2883 | extern struct inode * iget_locked(struct super_block *, unsigned long); | |
fe032c42 TT |
2884 | extern struct inode *find_inode_nowait(struct super_block *, |
2885 | unsigned long, | |
2886 | int (*match)(struct inode *, | |
2887 | unsigned long, void *), | |
2888 | void *data); | |
261bca86 AV |
2889 | extern int insert_inode_locked4(struct inode *, unsigned long, int (*test)(struct inode *, void *), void *); |
2890 | extern int insert_inode_locked(struct inode *); | |
e096d0c7 JB |
2891 | #ifdef CONFIG_DEBUG_LOCK_ALLOC |
2892 | extern void lockdep_annotate_inode_mutex_key(struct inode *inode); | |
2893 | #else | |
2894 | static inline void lockdep_annotate_inode_mutex_key(struct inode *inode) { }; | |
2895 | #endif | |
1da177e4 | 2896 | extern void unlock_new_inode(struct inode *); |
85fe4025 | 2897 | extern unsigned int get_next_ino(void); |
799ea9e9 | 2898 | extern void evict_inodes(struct super_block *sb); |
1da177e4 | 2899 | |
1da177e4 | 2900 | extern void __iget(struct inode * inode); |
b46980fe | 2901 | extern void iget_failed(struct inode *); |
dbd5768f | 2902 | extern void clear_inode(struct inode *); |
2e00c97e | 2903 | extern void __destroy_inode(struct inode *); |
a209dfc7 ED |
2904 | extern struct inode *new_inode_pseudo(struct super_block *sb); |
2905 | extern struct inode *new_inode(struct super_block *sb); | |
ff0c7d15 | 2906 | extern void free_inode_nonrcu(struct inode *inode); |
01de85e0 | 2907 | extern int should_remove_suid(struct dentry *); |
5fa8e0a1 | 2908 | extern int file_remove_privs(struct file *); |
1da177e4 LT |
2909 | |
2910 | extern void __insert_inode_hash(struct inode *, unsigned long hashval); | |
646ec461 CH |
2911 | static inline void insert_inode_hash(struct inode *inode) |
2912 | { | |
1da177e4 LT |
2913 | __insert_inode_hash(inode, inode->i_ino); |
2914 | } | |
f2ee7abf ED |
2915 | |
2916 | extern void __remove_inode_hash(struct inode *); | |
2917 | static inline void remove_inode_hash(struct inode *inode) | |
2918 | { | |
cbedaac6 | 2919 | if (!inode_unhashed(inode) && !hlist_fake(&inode->i_hash)) |
f2ee7abf ED |
2920 | __remove_inode_hash(inode); |
2921 | } | |
2922 | ||
646ec461 | 2923 | extern void inode_sb_list_add(struct inode *inode); |
1da177e4 | 2924 | |
9361401e | 2925 | #ifdef CONFIG_BLOCK |
1da177e4 | 2926 | extern int bdev_read_only(struct block_device *); |
9361401e | 2927 | #endif |
1da177e4 LT |
2928 | extern int set_blocksize(struct block_device *, int); |
2929 | extern int sb_set_blocksize(struct super_block *, int); | |
2930 | extern int sb_min_blocksize(struct super_block *, int); | |
2931 | ||
2932 | extern int generic_file_mmap(struct file *, struct vm_area_struct *); | |
2933 | extern int generic_file_readonly_mmap(struct file *, struct vm_area_struct *); | |
3309dd04 | 2934 | extern ssize_t generic_write_checks(struct kiocb *, struct iov_iter *); |
ed978a81 | 2935 | extern ssize_t generic_file_read_iter(struct kiocb *, struct iov_iter *); |
8174202b | 2936 | extern ssize_t __generic_file_write_iter(struct kiocb *, struct iov_iter *); |
8174202b | 2937 | extern ssize_t generic_file_write_iter(struct kiocb *, struct iov_iter *); |
1af5bb49 | 2938 | extern ssize_t generic_file_direct_write(struct kiocb *, struct iov_iter *); |
3b93f911 | 2939 | extern ssize_t generic_perform_write(struct file *, struct iov_iter *, loff_t); |
88e6faef | 2940 | |
18e9710e | 2941 | ssize_t vfs_iter_read(struct file *file, struct iov_iter *iter, loff_t *ppos, |
ddef7ed2 | 2942 | rwf_t flags); |
abbb6589 | 2943 | ssize_t vfs_iter_write(struct file *file, struct iov_iter *iter, loff_t *ppos, |
ddef7ed2 | 2944 | rwf_t flags); |
dbe4e192 | 2945 | |
eef99380 | 2946 | /* fs/block_dev.c */ |
b2de525f | 2947 | extern ssize_t blkdev_read_iter(struct kiocb *iocb, struct iov_iter *to); |
1456c0a8 | 2948 | extern ssize_t blkdev_write_iter(struct kiocb *iocb, struct iov_iter *from); |
02c24a82 JB |
2949 | extern int blkdev_fsync(struct file *filp, loff_t start, loff_t end, |
2950 | int datasync); | |
ff01bb48 | 2951 | extern void block_sync_page(struct page *page); |
eef99380 | 2952 | |
88e6faef | 2953 | /* fs/splice.c */ |
cbb7e577 | 2954 | extern ssize_t generic_file_splice_read(struct file *, loff_t *, |
88e6faef | 2955 | struct pipe_inode_info *, size_t, unsigned int); |
8d020765 | 2956 | extern ssize_t iter_file_splice_write(struct pipe_inode_info *, |
cbb7e577 | 2957 | struct file *, loff_t *, size_t, unsigned int); |
88e6faef | 2958 | extern ssize_t generic_splice_sendpage(struct pipe_inode_info *pipe, |
cbb7e577 | 2959 | struct file *out, loff_t *, size_t len, unsigned int flags); |
1c118596 MS |
2960 | extern long do_splice_direct(struct file *in, loff_t *ppos, struct file *out, |
2961 | loff_t *opos, size_t len, unsigned int flags); | |
2962 | ||
88e6faef | 2963 | |
1da177e4 LT |
2964 | extern void |
2965 | file_ra_state_init(struct file_ra_state *ra, struct address_space *mapping); | |
965c8e59 AM |
2966 | extern loff_t noop_llseek(struct file *file, loff_t offset, int whence); |
2967 | extern loff_t no_llseek(struct file *file, loff_t offset, int whence); | |
46a1c2c7 | 2968 | extern loff_t vfs_setpos(struct file *file, loff_t offset, loff_t maxsize); |
965c8e59 | 2969 | extern loff_t generic_file_llseek(struct file *file, loff_t offset, int whence); |
5760495a | 2970 | extern loff_t generic_file_llseek_size(struct file *file, loff_t offset, |
965c8e59 | 2971 | int whence, loff_t maxsize, loff_t eof); |
1bf9d14d AV |
2972 | extern loff_t fixed_size_llseek(struct file *file, loff_t offset, |
2973 | int whence, loff_t size); | |
b25472f9 AV |
2974 | extern loff_t no_seek_end_llseek_size(struct file *, loff_t, int, loff_t); |
2975 | extern loff_t no_seek_end_llseek(struct file *, loff_t, int); | |
1da177e4 LT |
2976 | extern int generic_file_open(struct inode * inode, struct file * filp); |
2977 | extern int nonseekable_open(struct inode * inode, struct file * filp); | |
2978 | ||
9361401e | 2979 | #ifdef CONFIG_BLOCK |
8a4c1e42 | 2980 | typedef void (dio_submit_t)(struct bio *bio, struct inode *inode, |
facd07b0 | 2981 | loff_t file_offset); |
1da177e4 LT |
2982 | |
2983 | enum { | |
1e431f5c CH |
2984 | /* need locking between buffered and direct access */ |
2985 | DIO_LOCKING = 0x01, | |
2986 | ||
2987 | /* filesystem does not support filling holes */ | |
2988 | DIO_SKIP_HOLES = 0x02, | |
1da177e4 LT |
2989 | }; |
2990 | ||
4055351c | 2991 | void dio_end_io(struct bio *bio); |
5a9d929d | 2992 | void dio_warn_stale_pagecache(struct file *filp); |
eafdc7d1 | 2993 | |
17f8c842 OS |
2994 | ssize_t __blockdev_direct_IO(struct kiocb *iocb, struct inode *inode, |
2995 | struct block_device *bdev, struct iov_iter *iter, | |
c8b8e32d | 2996 | get_block_t get_block, |
17f8c842 OS |
2997 | dio_iodone_t end_io, dio_submit_t submit_io, |
2998 | int flags); | |
7bb46a67 | 2999 | |
17f8c842 OS |
3000 | static inline ssize_t blockdev_direct_IO(struct kiocb *iocb, |
3001 | struct inode *inode, | |
c8b8e32d | 3002 | struct iov_iter *iter, |
17f8c842 | 3003 | get_block_t get_block) |
1da177e4 | 3004 | { |
17f8c842 | 3005 | return __blockdev_direct_IO(iocb, inode, inode->i_sb->s_bdev, iter, |
c8b8e32d | 3006 | get_block, NULL, NULL, DIO_LOCKING | DIO_SKIP_HOLES); |
1da177e4 | 3007 | } |
9361401e | 3008 | #endif |
1da177e4 | 3009 | |
1d59d61f | 3010 | void inode_dio_wait(struct inode *inode); |
fe0f07d0 JA |
3011 | |
3012 | /* | |
3013 | * inode_dio_begin - signal start of a direct I/O requests | |
3014 | * @inode: inode the direct I/O happens on | |
3015 | * | |
3016 | * This is called once we've finished processing a direct I/O request, | |
3017 | * and is used to wake up callers waiting for direct I/O to be quiesced. | |
3018 | */ | |
3019 | static inline void inode_dio_begin(struct inode *inode) | |
3020 | { | |
3021 | atomic_inc(&inode->i_dio_count); | |
3022 | } | |
3023 | ||
3024 | /* | |
3025 | * inode_dio_end - signal finish of a direct I/O requests | |
3026 | * @inode: inode the direct I/O happens on | |
3027 | * | |
3028 | * This is called once we've finished processing a direct I/O request, | |
3029 | * and is used to wake up callers waiting for direct I/O to be quiesced. | |
3030 | */ | |
3031 | static inline void inode_dio_end(struct inode *inode) | |
3032 | { | |
3033 | if (atomic_dec_and_test(&inode->i_dio_count)) | |
3034 | wake_up_bit(&inode->i_state, __I_DIO_WAKEUP); | |
3035 | } | |
1d59d61f | 3036 | |
5f16f322 TT |
3037 | extern void inode_set_flags(struct inode *inode, unsigned int flags, |
3038 | unsigned int mask); | |
3039 | ||
4b6f5d20 | 3040 | extern const struct file_operations generic_ro_fops; |
1da177e4 LT |
3041 | |
3042 | #define special_file(m) (S_ISCHR(m)||S_ISBLK(m)||S_ISFIFO(m)||S_ISSOCK(m)) | |
3043 | ||
5d826c84 | 3044 | extern int readlink_copy(char __user *, int, const char *); |
1da177e4 | 3045 | extern int page_readlink(struct dentry *, char __user *, int); |
fceef393 AV |
3046 | extern const char *page_get_link(struct dentry *, struct inode *, |
3047 | struct delayed_call *); | |
3048 | extern void page_put_link(void *); | |
0adb25d2 | 3049 | extern int __page_symlink(struct inode *inode, const char *symname, int len, |
54566b2c | 3050 | int nofs); |
1da177e4 | 3051 | extern int page_symlink(struct inode *inode, const char *symname, int len); |
c5ef1c42 | 3052 | extern const struct inode_operations page_symlink_inode_operations; |
fceef393 | 3053 | extern void kfree_link(void *); |
1da177e4 | 3054 | extern void generic_fillattr(struct inode *, struct kstat *); |
a528d35e DH |
3055 | extern int vfs_getattr_nosec(const struct path *, struct kstat *, u32, unsigned int); |
3056 | extern int vfs_getattr(const struct path *, struct kstat *, u32, unsigned int); | |
b462707e | 3057 | void __inode_add_bytes(struct inode *inode, loff_t bytes); |
1da177e4 | 3058 | void inode_add_bytes(struct inode *inode, loff_t bytes); |
1c8924eb | 3059 | void __inode_sub_bytes(struct inode *inode, loff_t bytes); |
1da177e4 | 3060 | void inode_sub_bytes(struct inode *inode, loff_t bytes); |
f4a8116a JK |
3061 | static inline loff_t __inode_get_bytes(struct inode *inode) |
3062 | { | |
3063 | return (((loff_t)inode->i_blocks) << 9) + inode->i_bytes; | |
3064 | } | |
1da177e4 LT |
3065 | loff_t inode_get_bytes(struct inode *inode); |
3066 | void inode_set_bytes(struct inode *inode, loff_t bytes); | |
fceef393 AV |
3067 | const char *simple_get_link(struct dentry *, struct inode *, |
3068 | struct delayed_call *); | |
61ba64fc | 3069 | extern const struct inode_operations simple_symlink_inode_operations; |
1da177e4 | 3070 | |
5c0ba4e0 | 3071 | extern int iterate_dir(struct file *, struct dir_context *); |
1da177e4 | 3072 | |
a528d35e DH |
3073 | extern int vfs_statx(int, const char __user *, int, struct kstat *, u32); |
3074 | extern int vfs_statx_fd(unsigned int, struct kstat *, u32, unsigned int); | |
3075 | ||
3076 | static inline int vfs_stat(const char __user *filename, struct kstat *stat) | |
3077 | { | |
deccf497 DH |
3078 | return vfs_statx(AT_FDCWD, filename, AT_NO_AUTOMOUNT, |
3079 | stat, STATX_BASIC_STATS); | |
a528d35e DH |
3080 | } |
3081 | static inline int vfs_lstat(const char __user *name, struct kstat *stat) | |
3082 | { | |
deccf497 | 3083 | return vfs_statx(AT_FDCWD, name, AT_SYMLINK_NOFOLLOW | AT_NO_AUTOMOUNT, |
a528d35e DH |
3084 | stat, STATX_BASIC_STATS); |
3085 | } | |
3086 | static inline int vfs_fstatat(int dfd, const char __user *filename, | |
3087 | struct kstat *stat, int flags) | |
3088 | { | |
5d38f049 IK |
3089 | return vfs_statx(dfd, filename, flags | AT_NO_AUTOMOUNT, |
3090 | stat, STATX_BASIC_STATS); | |
a528d35e DH |
3091 | } |
3092 | static inline int vfs_fstat(int fd, struct kstat *stat) | |
3093 | { | |
3094 | return vfs_statx_fd(fd, stat, STATX_BASIC_STATS, 0); | |
3095 | } | |
3096 | ||
3097 | ||
d60874cd | 3098 | extern const char *vfs_get_link(struct dentry *, struct delayed_call *); |
fd4a0edf | 3099 | extern int vfs_readlink(struct dentry *, char __user *, int); |
1da177e4 | 3100 | |
e9079cce | 3101 | extern int __generic_block_fiemap(struct inode *inode, |
3a3076f4 JB |
3102 | struct fiemap_extent_info *fieinfo, |
3103 | loff_t start, loff_t len, | |
3104 | get_block_t *get_block); | |
68c9d702 JB |
3105 | extern int generic_block_fiemap(struct inode *inode, |
3106 | struct fiemap_extent_info *fieinfo, u64 start, | |
3107 | u64 len, get_block_t *get_block); | |
1da177e4 | 3108 | |
ee416bcd | 3109 | extern struct file_system_type *get_filesystem(struct file_system_type *fs); |
c18479fe | 3110 | extern void put_filesystem(struct file_system_type *fs); |
1da177e4 LT |
3111 | extern struct file_system_type *get_fs_type(const char *name); |
3112 | extern struct super_block *get_super(struct block_device *); | |
6b6dc836 | 3113 | extern struct super_block *get_super_thawed(struct block_device *); |
ba6379f7 | 3114 | extern struct super_block *get_super_exclusive_thawed(struct block_device *bdev); |
4504230a | 3115 | extern struct super_block *get_active_super(struct block_device *bdev); |
1da177e4 | 3116 | extern void drop_super(struct super_block *sb); |
ba6379f7 | 3117 | extern void drop_super_exclusive(struct super_block *sb); |
01a05b33 | 3118 | extern void iterate_supers(void (*)(struct super_block *, void *), void *); |
43e15cdb AV |
3119 | extern void iterate_supers_type(struct file_system_type *, |
3120 | void (*)(struct super_block *, void *), void *); | |
1da177e4 LT |
3121 | |
3122 | extern int dcache_dir_open(struct inode *, struct file *); | |
3123 | extern int dcache_dir_close(struct inode *, struct file *); | |
3124 | extern loff_t dcache_dir_lseek(struct file *, loff_t, int); | |
5f99f4e7 | 3125 | extern int dcache_readdir(struct file *, struct dir_context *); |
7bb46a67 | 3126 | extern int simple_setattr(struct dentry *, struct iattr *); |
a528d35e | 3127 | extern int simple_getattr(const struct path *, struct kstat *, u32, unsigned int); |
726c3342 | 3128 | extern int simple_statfs(struct dentry *, struct kstatfs *); |
20955e89 | 3129 | extern int simple_open(struct inode *inode, struct file *file); |
1da177e4 LT |
3130 | extern int simple_link(struct dentry *, struct inode *, struct dentry *); |
3131 | extern int simple_unlink(struct inode *, struct dentry *); | |
3132 | extern int simple_rmdir(struct inode *, struct dentry *); | |
e0e0be8a MS |
3133 | extern int simple_rename(struct inode *, struct dentry *, |
3134 | struct inode *, struct dentry *, unsigned int); | |
02c24a82 | 3135 | extern int noop_fsync(struct file *, loff_t, loff_t, int); |
f44c7763 DW |
3136 | extern int noop_set_page_dirty(struct page *page); |
3137 | extern void noop_invalidatepage(struct page *page, unsigned int offset, | |
3138 | unsigned int length); | |
3139 | extern ssize_t noop_direct_IO(struct kiocb *iocb, struct iov_iter *iter); | |
1da177e4 LT |
3140 | extern int simple_empty(struct dentry *); |
3141 | extern int simple_readpage(struct file *file, struct page *page); | |
afddba49 NP |
3142 | extern int simple_write_begin(struct file *file, struct address_space *mapping, |
3143 | loff_t pos, unsigned len, unsigned flags, | |
3144 | struct page **pagep, void **fsdata); | |
3145 | extern int simple_write_end(struct file *file, struct address_space *mapping, | |
3146 | loff_t pos, unsigned len, unsigned copied, | |
3147 | struct page *page, void *fsdata); | |
b26d4cd3 | 3148 | extern int always_delete_dentry(const struct dentry *); |
6987843f | 3149 | extern struct inode *alloc_anon_inode(struct super_block *); |
e6f5c789 | 3150 | extern int simple_nosetlease(struct file *, long, struct file_lock **, void **); |
b26d4cd3 | 3151 | extern const struct dentry_operations simple_dentry_operations; |
1da177e4 | 3152 | |
00cd8dd3 | 3153 | extern struct dentry *simple_lookup(struct inode *, struct dentry *, unsigned int flags); |
1da177e4 | 3154 | extern ssize_t generic_read_dir(struct file *, char __user *, size_t, loff_t *); |
4b6f5d20 | 3155 | extern const struct file_operations simple_dir_operations; |
c5ef1c42 | 3156 | extern const struct inode_operations simple_dir_inode_operations; |
fbabfd0f EB |
3157 | extern void make_empty_dir_inode(struct inode *inode); |
3158 | extern bool is_empty_dir_inode(struct inode *inode); | |
cda37124 | 3159 | struct tree_descr { const char *name; const struct file_operations *ops; int mode; }; |
1da177e4 | 3160 | struct dentry *d_alloc_name(struct dentry *, const char *); |
cda37124 EB |
3161 | extern int simple_fill_super(struct super_block *, unsigned long, |
3162 | const struct tree_descr *); | |
1f5ce9e9 | 3163 | extern int simple_pin_fs(struct file_system_type *, struct vfsmount **mount, int *count); |
1da177e4 LT |
3164 | extern void simple_release_fs(struct vfsmount **mount, int *count); |
3165 | ||
93b07113 AM |
3166 | extern ssize_t simple_read_from_buffer(void __user *to, size_t count, |
3167 | loff_t *ppos, const void *from, size_t available); | |
6a727b43 JS |
3168 | extern ssize_t simple_write_to_buffer(void *to, size_t available, loff_t *ppos, |
3169 | const void __user *from, size_t count); | |
1da177e4 | 3170 | |
ac13a829 | 3171 | extern int __generic_file_fsync(struct file *, loff_t, loff_t, int); |
02c24a82 | 3172 | extern int generic_file_fsync(struct file *, loff_t, loff_t, int); |
d5aacad5 | 3173 | |
30ca22c7 PL |
3174 | extern int generic_check_addressable(unsigned, u64); |
3175 | ||
e965f963 | 3176 | #ifdef CONFIG_MIGRATION |
2d1db3b1 | 3177 | extern int buffer_migrate_page(struct address_space *, |
a6bc32b8 MG |
3178 | struct page *, struct page *, |
3179 | enum migrate_mode); | |
e965f963 CL |
3180 | #else |
3181 | #define buffer_migrate_page NULL | |
3182 | #endif | |
3183 | ||
31051c85 | 3184 | extern int setattr_prepare(struct dentry *, struct iattr *); |
25d9e2d1 | 3185 | extern int inode_newsize_ok(const struct inode *, loff_t offset); |
6a1a90ad | 3186 | extern void setattr_copy(struct inode *inode, const struct iattr *attr); |
1da177e4 | 3187 | |
c3b2da31 | 3188 | extern int file_update_time(struct file *file); |
1da177e4 | 3189 | |
d475c634 MW |
3190 | static inline bool io_is_direct(struct file *filp) |
3191 | { | |
65f87ee7 | 3192 | return (filp->f_flags & O_DIRECT) || IS_DAX(filp->f_mapping->host); |
d475c634 MW |
3193 | } |
3194 | ||
baabda26 DW |
3195 | static inline bool vma_is_dax(struct vm_area_struct *vma) |
3196 | { | |
3197 | return vma->vm_file && IS_DAX(vma->vm_file->f_mapping->host); | |
3198 | } | |
3199 | ||
2bb6d283 DW |
3200 | static inline bool vma_is_fsdax(struct vm_area_struct *vma) |
3201 | { | |
3202 | struct inode *inode; | |
3203 | ||
3204 | if (!vma->vm_file) | |
3205 | return false; | |
3206 | if (!vma_is_dax(vma)) | |
3207 | return false; | |
3208 | inode = file_inode(vma->vm_file); | |
230f5a89 | 3209 | if (S_ISCHR(inode->i_mode)) |
2bb6d283 DW |
3210 | return false; /* device-dax */ |
3211 | return true; | |
3212 | } | |
3213 | ||
2ba48ce5 AV |
3214 | static inline int iocb_flags(struct file *file) |
3215 | { | |
3216 | int res = 0; | |
3217 | if (file->f_flags & O_APPEND) | |
3218 | res |= IOCB_APPEND; | |
3219 | if (io_is_direct(file)) | |
3220 | res |= IOCB_DIRECT; | |
dde0c2e7 CH |
3221 | if ((file->f_flags & O_DSYNC) || IS_SYNC(file->f_mapping->host)) |
3222 | res |= IOCB_DSYNC; | |
3223 | if (file->f_flags & __O_SYNC) | |
3224 | res |= IOCB_SYNC; | |
2ba48ce5 AV |
3225 | return res; |
3226 | } | |
3227 | ||
ddef7ed2 | 3228 | static inline int kiocb_set_rw_flags(struct kiocb *ki, rwf_t flags) |
fdd2f5b7 GR |
3229 | { |
3230 | if (unlikely(flags & ~RWF_SUPPORTED)) | |
3231 | return -EOPNOTSUPP; | |
3232 | ||
b745fafa | 3233 | if (flags & RWF_NOWAIT) { |
91f9943e | 3234 | if (!(ki->ki_filp->f_mode & FMODE_NOWAIT)) |
b745fafa GR |
3235 | return -EOPNOTSUPP; |
3236 | ki->ki_flags |= IOCB_NOWAIT; | |
3237 | } | |
fdd2f5b7 GR |
3238 | if (flags & RWF_HIPRI) |
3239 | ki->ki_flags |= IOCB_HIPRI; | |
3240 | if (flags & RWF_DSYNC) | |
3241 | ki->ki_flags |= IOCB_DSYNC; | |
3242 | if (flags & RWF_SYNC) | |
3243 | ki->ki_flags |= (IOCB_DSYNC | IOCB_SYNC); | |
e1fc742e JB |
3244 | if (flags & RWF_APPEND) |
3245 | ki->ki_flags |= IOCB_APPEND; | |
fdd2f5b7 GR |
3246 | return 0; |
3247 | } | |
3248 | ||
1da177e4 LT |
3249 | static inline ino_t parent_ino(struct dentry *dentry) |
3250 | { | |
3251 | ino_t res; | |
3252 | ||
b5c84bf6 NP |
3253 | /* |
3254 | * Don't strictly need d_lock here? If the parent ino could change | |
3255 | * then surely we'd have a deeper race in the caller? | |
3256 | */ | |
1da177e4 LT |
3257 | spin_lock(&dentry->d_lock); |
3258 | res = dentry->d_parent->d_inode->i_ino; | |
3259 | spin_unlock(&dentry->d_lock); | |
3260 | return res; | |
3261 | } | |
3262 | ||
1da177e4 LT |
3263 | /* Transaction based IO helpers */ |
3264 | ||
3265 | /* | |
3266 | * An argresp is stored in an allocated page and holds the | |
3267 | * size of the argument or response, along with its content | |
3268 | */ | |
3269 | struct simple_transaction_argresp { | |
3270 | ssize_t size; | |
3271 | char data[0]; | |
3272 | }; | |
3273 | ||
3274 | #define SIMPLE_TRANSACTION_LIMIT (PAGE_SIZE - sizeof(struct simple_transaction_argresp)) | |
3275 | ||
3276 | char *simple_transaction_get(struct file *file, const char __user *buf, | |
3277 | size_t size); | |
3278 | ssize_t simple_transaction_read(struct file *file, char __user *buf, | |
3279 | size_t size, loff_t *pos); | |
3280 | int simple_transaction_release(struct inode *inode, struct file *file); | |
3281 | ||
76791ab2 | 3282 | void simple_transaction_set(struct file *file, size_t n); |
1da177e4 | 3283 | |
acaefc25 AB |
3284 | /* |
3285 | * simple attribute files | |
3286 | * | |
3287 | * These attributes behave similar to those in sysfs: | |
3288 | * | |
3289 | * Writing to an attribute immediately sets a value, an open file can be | |
3290 | * written to multiple times. | |
3291 | * | |
3292 | * Reading from an attribute creates a buffer from the value that might get | |
3293 | * read with multiple read calls. When the attribute has been read | |
3294 | * completely, no further read calls are possible until the file is opened | |
3295 | * again. | |
3296 | * | |
3297 | * All attributes contain a text representation of a numeric value | |
3298 | * that are accessed with the get() and set() functions. | |
3299 | */ | |
3300 | #define DEFINE_SIMPLE_ATTRIBUTE(__fops, __get, __set, __fmt) \ | |
3301 | static int __fops ## _open(struct inode *inode, struct file *file) \ | |
3302 | { \ | |
3303 | __simple_attr_check_format(__fmt, 0ull); \ | |
3304 | return simple_attr_open(inode, file, __get, __set, __fmt); \ | |
3305 | } \ | |
828c0950 | 3306 | static const struct file_operations __fops = { \ |
acaefc25 AB |
3307 | .owner = THIS_MODULE, \ |
3308 | .open = __fops ## _open, \ | |
74bedc4d | 3309 | .release = simple_attr_release, \ |
acaefc25 AB |
3310 | .read = simple_attr_read, \ |
3311 | .write = simple_attr_write, \ | |
1ec5584e | 3312 | .llseek = generic_file_llseek, \ |
68be3029 | 3313 | } |
acaefc25 | 3314 | |
b9075fa9 JP |
3315 | static inline __printf(1, 2) |
3316 | void __simple_attr_check_format(const char *fmt, ...) | |
acaefc25 AB |
3317 | { |
3318 | /* don't do anything, just let the compiler check the arguments; */ | |
3319 | } | |
3320 | ||
3321 | int simple_attr_open(struct inode *inode, struct file *file, | |
8b88b099 | 3322 | int (*get)(void *, u64 *), int (*set)(void *, u64), |
acaefc25 | 3323 | const char *fmt); |
74bedc4d | 3324 | int simple_attr_release(struct inode *inode, struct file *file); |
acaefc25 AB |
3325 | ssize_t simple_attr_read(struct file *file, char __user *buf, |
3326 | size_t len, loff_t *ppos); | |
3327 | ssize_t simple_attr_write(struct file *file, const char __user *buf, | |
3328 | size_t len, loff_t *ppos); | |
3329 | ||
4be28540 | 3330 | struct ctl_table; |
8d65af78 | 3331 | int proc_nr_files(struct ctl_table *table, int write, |
62239ac2 | 3332 | void __user *buffer, size_t *lenp, loff_t *ppos); |
312d3ca8 CH |
3333 | int proc_nr_dentry(struct ctl_table *table, int write, |
3334 | void __user *buffer, size_t *lenp, loff_t *ppos); | |
cffbc8aa DC |
3335 | int proc_nr_inodes(struct ctl_table *table, int write, |
3336 | void __user *buffer, size_t *lenp, loff_t *ppos); | |
38e23c95 | 3337 | int __init get_filesystem_list(char *buf); |
62239ac2 | 3338 | |
3cd90ea4 | 3339 | #define __FMODE_EXEC ((__force int) FMODE_EXEC) |
1a44bc8c NK |
3340 | #define __FMODE_NONOTIFY ((__force int) FMODE_NONOTIFY) |
3341 | ||
6d125529 | 3342 | #define ACC_MODE(x) ("\004\002\006\006"[(x)&O_ACCMODE]) |
ecf081d1 | 3343 | #define OPEN_FMODE(flag) ((__force fmode_t)(((flag + 1) & O_ACCMODE) | \ |
1a44bc8c | 3344 | (flag & __FMODE_NONOTIFY))) |
5300990c | 3345 | |
d37177ba | 3346 | static inline bool is_sxid(umode_t mode) |
69b45732 AK |
3347 | { |
3348 | return (mode & S_ISUID) || ((mode & S_ISGID) && (mode & S_IXGRP)); | |
3349 | } | |
3350 | ||
cbdf35bc MS |
3351 | static inline int check_sticky(struct inode *dir, struct inode *inode) |
3352 | { | |
3353 | if (!(dir->i_mode & S_ISVTX)) | |
3354 | return 0; | |
3355 | ||
3356 | return __check_sticky(dir, inode); | |
3357 | } | |
3358 | ||
69b45732 AK |
3359 | static inline void inode_has_no_xattr(struct inode *inode) |
3360 | { | |
e462ec50 | 3361 | if (!is_sxid(inode->i_mode) && (inode->i_sb->s_flags & SB_NOSEC)) |
69b45732 AK |
3362 | inode->i_flags |= S_NOSEC; |
3363 | } | |
3364 | ||
a7400222 AV |
3365 | static inline bool is_root_inode(struct inode *inode) |
3366 | { | |
3367 | return inode == inode->i_sb->s_root->d_inode; | |
3368 | } | |
3369 | ||
5f99f4e7 AV |
3370 | static inline bool dir_emit(struct dir_context *ctx, |
3371 | const char *name, int namelen, | |
3372 | u64 ino, unsigned type) | |
3373 | { | |
3374 | return ctx->actor(ctx, name, namelen, ctx->pos, ino, type) == 0; | |
3375 | } | |
3376 | static inline bool dir_emit_dot(struct file *file, struct dir_context *ctx) | |
3377 | { | |
3378 | return ctx->actor(ctx, ".", 1, ctx->pos, | |
3379 | file->f_path.dentry->d_inode->i_ino, DT_DIR) == 0; | |
3380 | } | |
3381 | static inline bool dir_emit_dotdot(struct file *file, struct dir_context *ctx) | |
3382 | { | |
3383 | return ctx->actor(ctx, "..", 2, ctx->pos, | |
3384 | parent_ino(file->f_path.dentry), DT_DIR) == 0; | |
3385 | } | |
3386 | static inline bool dir_emit_dots(struct file *file, struct dir_context *ctx) | |
3387 | { | |
3388 | if (ctx->pos == 0) { | |
3389 | if (!dir_emit_dot(file, ctx)) | |
3390 | return false; | |
3391 | ctx->pos = 1; | |
3392 | } | |
3393 | if (ctx->pos == 1) { | |
3394 | if (!dir_emit_dotdot(file, ctx)) | |
3395 | return false; | |
3396 | ctx->pos = 2; | |
3397 | } | |
3398 | return true; | |
3399 | } | |
5ded75ec AV |
3400 | static inline bool dir_relax(struct inode *inode) |
3401 | { | |
5955102c AV |
3402 | inode_unlock(inode); |
3403 | inode_lock(inode); | |
5ded75ec AV |
3404 | return !IS_DEADDIR(inode); |
3405 | } | |
5f99f4e7 | 3406 | |
ae05327a AV |
3407 | static inline bool dir_relax_shared(struct inode *inode) |
3408 | { | |
3409 | inode_unlock_shared(inode); | |
3410 | inode_lock_shared(inode); | |
3411 | return !IS_DEADDIR(inode); | |
3412 | } | |
3413 | ||
90f8572b | 3414 | extern bool path_noexec(const struct path *path); |
21fc61c7 | 3415 | extern void inode_nohighmem(struct inode *inode); |
90f8572b | 3416 | |
1da177e4 | 3417 | #endif /* _LINUX_FS_H */ |