]> git.ipfire.org Git - thirdparty/linux.git/blame - include/linux/fs.h
Merge tag 'io_uring-5.7-2020-05-22' of git://git.kernel.dk/linux-block
[thirdparty/linux.git] / include / linux / fs.h
CommitLineData
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>
d9a08a9e 39#include <linux/ioprio.h>
bbe7449e 40#include <linux/fs_types.h>
f1fffbd4
RV
41#include <linux/build_bug.h>
42#include <linux/stddef.h>
1da177e4 43
1da177e4 44#include <asm/byteorder.h>
607ca46e 45#include <uapi/linux/fs.h>
1da177e4 46
b83ae6d4 47struct backing_dev_info;
52ebea74 48struct bdi_writeback;
2f8b5444 49struct bio;
a5694255 50struct export_operations;
a885c8c4 51struct hd_geometry;
1da177e4 52struct iovec;
92198f7e 53struct kiocb;
57cc7215 54struct kobject;
1da177e4
LT
55struct pipe_inode_info;
56struct poll_table_struct;
57struct kstatfs;
58struct vm_area_struct;
59struct vfsmount;
745ca247 60struct cred;
a509bc1a 61struct swap_info_struct;
55985dd7 62struct seq_file;
7b7a8665 63struct workqueue_struct;
3b93f911 64struct iov_iter;
0b81d077
JK
65struct fscrypt_info;
66struct fscrypt_operations;
5585f2af
EB
67struct fsverity_info;
68struct fsverity_operations;
f3a09c92 69struct fs_context;
d7167b14 70struct fs_parameter_spec;
1da177e4 71
74bf17cf 72extern void __init inode_init(void);
1da177e4 73extern void __init inode_init_early(void);
4248b0da
MG
74extern void __init files_init(void);
75extern void __init files_maxfiles_init(void);
1da177e4 76
dded4f4d 77extern struct files_stat_struct files_stat;
518de9b3 78extern unsigned long get_max_files(void);
9b80a184 79extern unsigned int sysctl_nr_open;
dded4f4d
JE
80extern struct inodes_stat_t inodes_stat;
81extern int leases_enable, lease_break_time;
800179c9
KC
82extern int sysctl_protected_symlinks;
83extern int sysctl_protected_hardlinks;
30aba665
SM
84extern int sysctl_protected_fifos;
85extern int sysctl_protected_regular;
dded4f4d 86
ddef7ed2
CH
87typedef __kernel_rwf_t rwf_t;
88
1da177e4
LT
89struct buffer_head;
90typedef int (get_block_t)(struct inode *inode, sector_t iblock,
91 struct buffer_head *bh_result, int create);
187372a3 92typedef int (dio_iodone_t)(struct kiocb *iocb, loff_t offset,
7b7a8665 93 ssize_t bytes, void *private);
1da177e4 94
bbc1096a
DH
95#define MAY_EXEC 0x00000001
96#define MAY_WRITE 0x00000002
97#define MAY_READ 0x00000004
98#define MAY_APPEND 0x00000008
99#define MAY_ACCESS 0x00000010
100#define MAY_OPEN 0x00000020
101#define MAY_CHDIR 0x00000040
102/* called from RCU mode, don't block */
103#define MAY_NOT_BLOCK 0x00000080
104
105/*
106 * flags in file.f_mode. Note that FMODE_READ and FMODE_WRITE must correspond
75abe329 107 * to O_WRONLY and O_RDWR via the strange trick in do_dentry_open()
bbc1096a
DH
108 */
109
110/* file is open for reading */
111#define FMODE_READ ((__force fmode_t)0x1)
112/* file is open for writing */
113#define FMODE_WRITE ((__force fmode_t)0x2)
114/* file is seekable */
115#define FMODE_LSEEK ((__force fmode_t)0x4)
116/* file can be accessed using pread */
117#define FMODE_PREAD ((__force fmode_t)0x8)
118/* file can be accessed using pwrite */
119#define FMODE_PWRITE ((__force fmode_t)0x10)
120/* File is opened for execution with sys_execve / sys_uselib */
121#define FMODE_EXEC ((__force fmode_t)0x20)
122/* File is opened with O_NDELAY (only set for block devices) */
123#define FMODE_NDELAY ((__force fmode_t)0x40)
124/* File is opened with O_EXCL (only set for block devices) */
125#define FMODE_EXCL ((__force fmode_t)0x80)
126/* File is opened using open(.., 3, ..) and is writeable only for ioctls
127 (specialy hack for floppy.c) */
128#define FMODE_WRITE_IOCTL ((__force fmode_t)0x100)
129/* 32bit hashes as llseek() offset (for directories) */
130#define FMODE_32BITHASH ((__force fmode_t)0x200)
131/* 64bit hashes as llseek() offset (for directories) */
132#define FMODE_64BITHASH ((__force fmode_t)0x400)
133
134/*
135 * Don't update ctime and mtime.
136 *
137 * Currently a special hack for the XFS open_by_handle ioctl, but we'll
138 * hopefully graduate it to a proper O_CMTIME flag supported by open(2) soon.
139 */
140#define FMODE_NOCMTIME ((__force fmode_t)0x800)
141
142/* Expect random access pattern */
143#define FMODE_RANDOM ((__force fmode_t)0x1000)
144
145/* File is huge (eg. /dev/kmem): treat loff_t as unsigned */
146#define FMODE_UNSIGNED_OFFSET ((__force fmode_t)0x2000)
147
148/* File is opened with O_PATH; almost nothing can be done with it */
149#define FMODE_PATH ((__force fmode_t)0x4000)
150
2be7d348
LT
151/* File needs atomic accesses to f_pos */
152#define FMODE_ATOMIC_POS ((__force fmode_t)0x8000)
83f936c7
AV
153/* Write access to underlying fs */
154#define FMODE_WRITER ((__force fmode_t)0x10000)
7f7f25e8
AV
155/* Has read method(s) */
156#define FMODE_CAN_READ ((__force fmode_t)0x20000)
157/* Has write method(s) */
158#define FMODE_CAN_WRITE ((__force fmode_t)0x40000)
9c225f26 159
f5d11409 160#define FMODE_OPENED ((__force fmode_t)0x80000)
73a09dd9 161#define FMODE_CREATED ((__force fmode_t)0x100000)
f5d11409 162
10dce8af
KS
163/* File is stream-like */
164#define FMODE_STREAM ((__force fmode_t)0x200000)
165
bbc1096a 166/* File was opened by fanotify and shouldn't generate fanotify events */
75069f2b 167#define FMODE_NONOTIFY ((__force fmode_t)0x4000000)
bbc1096a 168
91f9943e 169/* File is capable of returning -EAGAIN if I/O will block */
a07b2000
AV
170#define FMODE_NOWAIT ((__force fmode_t)0x8000000)
171
172/* File represents mount that needs unmounting */
173#define FMODE_NEED_UNMOUNT ((__force fmode_t)0x10000000)
b745fafa 174
d3b1084d 175/* File does not contribute to nr_files count */
a07b2000 176#define FMODE_NOACCOUNT ((__force fmode_t)0x20000000)
d3b1084d 177
bbc1096a
DH
178/*
179 * Flag for rw_copy_check_uvector and compat_rw_copy_check_uvector
180 * that indicates that they should check the contents of the iovec are
181 * valid, but not check the memory that the iovec elements
182 * points too.
183 */
184#define CHECK_IOVEC_ONLY -1
185
1da177e4
LT
186/*
187 * Attribute flags. These should be or-ed together to figure out what
188 * has been changed!
189 */
9767d749
MS
190#define ATTR_MODE (1 << 0)
191#define ATTR_UID (1 << 1)
192#define ATTR_GID (1 << 2)
193#define ATTR_SIZE (1 << 3)
194#define ATTR_ATIME (1 << 4)
195#define ATTR_MTIME (1 << 5)
196#define ATTR_CTIME (1 << 6)
197#define ATTR_ATIME_SET (1 << 7)
198#define ATTR_MTIME_SET (1 << 8)
199#define ATTR_FORCE (1 << 9) /* Not a change, but a change it */
9767d749
MS
200#define ATTR_KILL_SUID (1 << 11)
201#define ATTR_KILL_SGID (1 << 12)
202#define ATTR_FILE (1 << 13)
203#define ATTR_KILL_PRIV (1 << 14)
204#define ATTR_OPEN (1 << 15) /* Truncating from open(O_TRUNC) */
205#define ATTR_TIMES_SET (1 << 16)
f2b20f6e 206#define ATTR_TOUCH (1 << 17)
1da177e4 207
787fb6bc
MS
208/*
209 * Whiteout is represented by a char device. The following constants define the
210 * mode and device number to use.
211 */
212#define WHITEOUT_MODE 0
213#define WHITEOUT_DEV 0
214
1da177e4
LT
215/*
216 * This is the Inode Attributes structure, used for notify_change(). It
217 * uses the above definitions as flags, to know which values have changed.
218 * Also, in this manner, a Filesystem can look at only the values it cares
219 * about. Basically, these are the attributes that the VFS layer can
220 * request to change from the FS layer.
221 *
222 * Derek Atkins <warlord@MIT.EDU> 94-10-20
223 */
224struct iattr {
225 unsigned int ia_valid;
226 umode_t ia_mode;
92361636
EB
227 kuid_t ia_uid;
228 kgid_t ia_gid;
1da177e4 229 loff_t ia_size;
95582b00
DD
230 struct timespec64 ia_atime;
231 struct timespec64 ia_mtime;
232 struct timespec64 ia_ctime;
cc4e69de
MS
233
234 /*
25985edc 235 * Not an attribute, but an auxiliary info for filesystems wanting to
cc4e69de
MS
236 * implement an ftruncate() like method. NOTE: filesystem should
237 * check for (ia_valid & ATTR_FILE), and not for (ia_file != NULL).
238 */
239 struct file *ia_file;
1da177e4
LT
240};
241
1da177e4
LT
242/*
243 * Includes for diskquotas.
244 */
245#include <linux/quota.h>
246
69c433ed
MS
247/*
248 * Maximum number of layers of fs stack. Needs to be limited to
249 * prevent kernel stack overflow
250 */
251#define FILESYSTEM_MAX_STACK_DEPTH 2
252
994fc28c
ZB
253/**
254 * enum positive_aop_returns - aop return codes with specific semantics
255 *
256 * @AOP_WRITEPAGE_ACTIVATE: Informs the caller that page writeback has
257 * completed, that the page is still locked, and
258 * should be considered active. The VM uses this hint
259 * to return the page to the active list -- it won't
260 * be a candidate for writeback again in the near
261 * future. Other callers must be careful to unlock
262 * the page if they get this return. Returned by
263 * writepage();
264 *
265 * @AOP_TRUNCATED_PAGE: The AOP method that was handed a locked page has
266 * unlocked it and the page might have been truncated.
267 * The caller should back up to acquiring a new page and
268 * trying again. The aop will be taking reasonable
269 * precautions not to livelock. If the caller held a page
270 * reference, it should drop it before retrying. Returned
55144768 271 * by readpage().
994fc28c
ZB
272 *
273 * address_space_operation functions return these large constants to indicate
274 * special semantics to the caller. These are much larger than the bytes in a
275 * page to allow for functions that return the number of bytes operated on in a
276 * given page.
277 */
278
279enum positive_aop_returns {
280 AOP_WRITEPAGE_ACTIVATE = 0x80000,
281 AOP_TRUNCATED_PAGE = 0x80001,
282};
283
c718a975
TH
284#define AOP_FLAG_CONT_EXPAND 0x0001 /* called from cont_expand */
285#define AOP_FLAG_NOFS 0x0002 /* used by filesystem to direct
54566b2c
NP
286 * helper code (eg buffer layer)
287 * to clear GFP_FS from alloc */
afddba49 288
1da177e4
LT
289/*
290 * oh the beauties of C type declarations.
291 */
292struct page;
293struct address_space;
294struct writeback_control;
1da177e4 295
c75b1d94
JA
296/*
297 * Write life time hint values.
69448867 298 * Stored in struct inode as u8.
c75b1d94
JA
299 */
300enum rw_hint {
301 WRITE_LIFE_NOT_SET = 0,
302 WRITE_LIFE_NONE = RWH_WRITE_LIFE_NONE,
303 WRITE_LIFE_SHORT = RWH_WRITE_LIFE_SHORT,
304 WRITE_LIFE_MEDIUM = RWH_WRITE_LIFE_MEDIUM,
305 WRITE_LIFE_LONG = RWH_WRITE_LIFE_LONG,
306 WRITE_LIFE_EXTREME = RWH_WRITE_LIFE_EXTREME,
307};
308
e2e40f2c 309#define IOCB_EVENTFD (1 << 0)
2ba48ce5
AV
310#define IOCB_APPEND (1 << 1)
311#define IOCB_DIRECT (1 << 2)
97be7ebe 312#define IOCB_HIPRI (1 << 3)
dde0c2e7
CH
313#define IOCB_DSYNC (1 << 4)
314#define IOCB_SYNC (1 << 5)
70fe2f48 315#define IOCB_WRITE (1 << 6)
b745fafa 316#define IOCB_NOWAIT (1 << 7)
e2e40f2c
CH
317
318struct kiocb {
319 struct file *ki_filp;
84c4e1f8
LT
320
321 /* The 'ki_filp' pointer is shared in a union for aio */
322 randomized_struct_fields_start
323
e2e40f2c
CH
324 loff_t ki_pos;
325 void (*ki_complete)(struct kiocb *iocb, long ret, long ret2);
326 void *private;
327 int ki_flags;
fc28724d 328 u16 ki_hint;
d9a08a9e 329 u16 ki_ioprio; /* See linux/ioprio.h */
fb7e1600 330 unsigned int ki_cookie; /* for ->iopoll */
84c4e1f8
LT
331
332 randomized_struct_fields_end
333};
e2e40f2c
CH
334
335static inline bool is_sync_kiocb(struct kiocb *kiocb)
336{
337 return kiocb->ki_complete == NULL;
338}
339
8ab22b9a
HH
340/*
341 * "descriptor" for what we're up to with a read.
342 * This allows us to use the same read code yet
343 * have multiple different users of the data that
344 * we read from a file.
345 *
346 * The simplest case just copies the data to user
347 * mode.
348 */
349typedef struct {
350 size_t written;
351 size_t count;
352 union {
353 char __user *buf;
354 void *data;
355 } arg;
356 int error;
357} read_descriptor_t;
358
359typedef int (*read_actor_t)(read_descriptor_t *, struct page *,
360 unsigned long, unsigned long);
2f718ffc 361
1da177e4
LT
362struct address_space_operations {
363 int (*writepage)(struct page *page, struct writeback_control *wbc);
364 int (*readpage)(struct file *, struct page *);
1da177e4
LT
365
366 /* Write back some dirty pages from this mapping. */
367 int (*writepages)(struct address_space *, struct writeback_control *);
368
4741c9fd 369 /* Set a page dirty. Return true if this dirtied it */
1da177e4
LT
370 int (*set_page_dirty)(struct page *page);
371
74c8164e
JA
372 /*
373 * Reads in the requested pages. Unlike ->readpage(), this is
374 * PURELY used for read-ahead!.
375 */
1da177e4
LT
376 int (*readpages)(struct file *filp, struct address_space *mapping,
377 struct list_head *pages, unsigned nr_pages);
378
afddba49
NP
379 int (*write_begin)(struct file *, struct address_space *mapping,
380 loff_t pos, unsigned len, unsigned flags,
381 struct page **pagep, void **fsdata);
382 int (*write_end)(struct file *, struct address_space *mapping,
383 loff_t pos, unsigned len, unsigned copied,
384 struct page *page, void *fsdata);
385
1da177e4
LT
386 /* Unfortunately this kludge is needed for FIBMAP. Don't use it */
387 sector_t (*bmap)(struct address_space *, sector_t);
d47992f8 388 void (*invalidatepage) (struct page *, unsigned int, unsigned int);
27496a8c 389 int (*releasepage) (struct page *, gfp_t);
6072d13c 390 void (*freepage)(struct page *);
c8b8e32d 391 ssize_t (*direct_IO)(struct kiocb *, struct iov_iter *iter);
b969c4ab 392 /*
ef277c73
TC
393 * migrate the contents of a page to the specified target. If
394 * migrate_mode is MIGRATE_ASYNC, it must not block.
b969c4ab 395 */
2d1db3b1 396 int (*migratepage) (struct address_space *,
a6bc32b8 397 struct page *, struct page *, enum migrate_mode);
bda807d4
MK
398 bool (*isolate_page)(struct page *, isolate_mode_t);
399 void (*putback_page)(struct page *);
e3db7691 400 int (*launder_page) (struct page *);
c186afb4 401 int (*is_partially_uptodate) (struct page *, unsigned long,
8ab22b9a 402 unsigned long);
b4597226 403 void (*is_dirty_writeback) (struct page *, bool *, bool *);
25718736 404 int (*error_remove_page)(struct address_space *, struct page *);
62c230bc
MG
405
406 /* swapfile support */
a509bc1a
MG
407 int (*swap_activate)(struct swap_info_struct *sis, struct file *file,
408 sector_t *span);
409 void (*swap_deactivate)(struct file *file);
1da177e4
LT
410};
411
7dcda1c9
JA
412extern const struct address_space_operations empty_aops;
413
afddba49
NP
414/*
415 * pagecache_write_begin/pagecache_write_end must be used by general code
416 * to write into the pagecache.
417 */
418int pagecache_write_begin(struct file *, struct address_space *mapping,
419 loff_t pos, unsigned len, unsigned flags,
420 struct page **pagep, void **fsdata);
421
422int pagecache_write_end(struct file *, struct address_space *mapping,
423 loff_t pos, unsigned len, unsigned copied,
424 struct page *page, void *fsdata);
425
eb797a8e
MW
426/**
427 * struct address_space - Contents of a cacheable, mappable object.
428 * @host: Owner, either the inode or the block_device.
429 * @i_pages: Cached pages.
430 * @gfp_mask: Memory allocation flags to use for allocating pages.
431 * @i_mmap_writable: Number of VM_SHARED mappings.
09d91cda 432 * @nr_thps: Number of THPs in the pagecache (non-shmem only).
eb797a8e
MW
433 * @i_mmap: Tree of private and shared mappings.
434 * @i_mmap_rwsem: Protects @i_mmap and @i_mmap_writable.
435 * @nrpages: Number of page entries, protected by the i_pages lock.
436 * @nrexceptional: Shadow or DAX entries, protected by the i_pages lock.
437 * @writeback_index: Writeback starts here.
438 * @a_ops: Methods.
439 * @flags: Error bits and flags (AS_*).
440 * @wb_err: The most recent error which has occurred.
441 * @private_lock: For use by the owner of the address_space.
442 * @private_list: For use by the owner of the address_space.
443 * @private_data: For use by the owner of the address_space.
444 */
1da177e4 445struct address_space {
eb797a8e
MW
446 struct inode *host;
447 struct xarray i_pages;
448 gfp_t gfp_mask;
449 atomic_t i_mmap_writable;
09d91cda
SL
450#ifdef CONFIG_READ_ONLY_THP_FOR_FS
451 /* number of thp, only for non-shmem files */
452 atomic_t nr_thps;
453#endif
eb797a8e
MW
454 struct rb_root_cached i_mmap;
455 struct rw_semaphore i_mmap_rwsem;
456 unsigned long nrpages;
f9fe48be 457 unsigned long nrexceptional;
eb797a8e
MW
458 pgoff_t writeback_index;
459 const struct address_space_operations *a_ops;
460 unsigned long flags;
5660e13d 461 errseq_t wb_err;
eb797a8e
MW
462 spinlock_t private_lock;
463 struct list_head private_list;
464 void *private_data;
3859a271 465} __attribute__((aligned(sizeof(long)))) __randomize_layout;
1da177e4
LT
466 /*
467 * On most architectures that alignment is already the case; but
25985edc 468 * must be enforced here for CRIS, to let the least significant bit
1da177e4
LT
469 * of struct page's "mapping" pointer be used for PAGE_MAPPING_ANON.
470 */
87192a2a 471struct request_queue;
1da177e4
LT
472
473struct block_device {
474 dev_t bd_dev; /* not a kdev_t - it's a search key */
f3ccfcda 475 int bd_openers;
1da177e4 476 struct inode * bd_inode; /* will die */
87d8fe1e 477 struct super_block * bd_super;
c039e313 478 struct mutex bd_mutex; /* open/close mutex */
6b4517a7 479 void * bd_claiming;
1da177e4
LT
480 void * bd_holder;
481 int bd_holders;
77ea887e 482 bool bd_write_holder;
641dc636 483#ifdef CONFIG_SYSFS
49731baa 484 struct list_head bd_holder_disks;
641dc636 485#endif
1da177e4
LT
486 struct block_device * bd_contains;
487 unsigned bd_block_size;
c2ee070f 488 u8 bd_partno;
1da177e4
LT
489 struct hd_struct * bd_part;
490 /* number of times partitions within this device have been opened. */
491 unsigned bd_part_count;
492 int bd_invalidated;
493 struct gendisk * bd_disk;
87192a2a 494 struct request_queue * bd_queue;
b1d2dc56 495 struct backing_dev_info *bd_bdi;
1da177e4 496 struct list_head bd_list;
1da177e4
LT
497 /*
498 * Private data. You must have bd_claim'ed the block_device
499 * to use this. NOTE: bd_claim allows an owner to claim
500 * the same device multiple times, the owner must take special
501 * care to not mess up bd_private for that case.
502 */
503 unsigned long bd_private;
fcccf502
TS
504
505 /* The counter of freeze processes */
506 int bd_fsfreeze_count;
507 /* Mutex for freeze */
508 struct mutex bd_fsfreeze_mutex;
3859a271 509} __randomize_layout;
1da177e4 510
ff9c745b
MW
511/* XArray tags, for tagging dirty and writeback pages in the pagecache. */
512#define PAGECACHE_TAG_DIRTY XA_MARK_0
513#define PAGECACHE_TAG_WRITEBACK XA_MARK_1
514#define PAGECACHE_TAG_TOWRITE XA_MARK_2
515
1da177e4 516/*
ff9c745b 517 * Returns true if any of the pages in the mapping are marked with the tag.
1da177e4 518 */
ff9c745b
MW
519static inline bool mapping_tagged(struct address_space *mapping, xa_mark_t tag)
520{
521 return xa_marked(&mapping->i_pages, tag);
522}
1da177e4 523
8b28f621
DB
524static inline void i_mmap_lock_write(struct address_space *mapping)
525{
c8c06efa 526 down_write(&mapping->i_mmap_rwsem);
8b28f621
DB
527}
528
c0d0381a
MK
529static inline int i_mmap_trylock_write(struct address_space *mapping)
530{
531 return down_write_trylock(&mapping->i_mmap_rwsem);
532}
533
8b28f621
DB
534static inline void i_mmap_unlock_write(struct address_space *mapping)
535{
c8c06efa 536 up_write(&mapping->i_mmap_rwsem);
8b28f621
DB
537}
538
3dec0ba0
DB
539static inline void i_mmap_lock_read(struct address_space *mapping)
540{
541 down_read(&mapping->i_mmap_rwsem);
542}
543
544static inline void i_mmap_unlock_read(struct address_space *mapping)
545{
546 up_read(&mapping->i_mmap_rwsem);
547}
548
1da177e4
LT
549/*
550 * Might pages of this file be mapped into userspace?
551 */
552static inline int mapping_mapped(struct address_space *mapping)
553{
f808c13f 554 return !RB_EMPTY_ROOT(&mapping->i_mmap.rb_root);
1da177e4
LT
555}
556
557/*
558 * Might pages of this file have been modified in userspace?
559 * Note that i_mmap_writable counts all VM_SHARED vmas: do_mmap_pgoff
560 * marks vma as VM_SHARED if it is shared, and the file was opened for
561 * writing i.e. vma may be mprotected writable even if now readonly.
4bb5f5d9
DR
562 *
563 * If i_mmap_writable is negative, no new writable mappings are allowed. You
564 * can only deny writable mappings, if none exists right now.
1da177e4
LT
565 */
566static inline int mapping_writably_mapped(struct address_space *mapping)
567{
4bb5f5d9
DR
568 return atomic_read(&mapping->i_mmap_writable) > 0;
569}
570
571static inline int mapping_map_writable(struct address_space *mapping)
572{
573 return atomic_inc_unless_negative(&mapping->i_mmap_writable) ?
574 0 : -EPERM;
575}
576
577static inline void mapping_unmap_writable(struct address_space *mapping)
578{
579 atomic_dec(&mapping->i_mmap_writable);
580}
581
582static inline int mapping_deny_writable(struct address_space *mapping)
583{
584 return atomic_dec_unless_positive(&mapping->i_mmap_writable) ?
585 0 : -EBUSY;
586}
587
588static inline void mapping_allow_writable(struct address_space *mapping)
589{
590 atomic_inc(&mapping->i_mmap_writable);
1da177e4
LT
591}
592
593/*
594 * Use sequence counter to get consistent i_size on 32-bit processors.
595 */
596#if BITS_PER_LONG==32 && defined(CONFIG_SMP)
597#include <linux/seqlock.h>
598#define __NEED_I_SIZE_ORDERED
599#define i_size_ordered_init(inode) seqcount_init(&inode->i_size_seqcount)
600#else
601#define i_size_ordered_init(inode) do { } while (0)
602#endif
603
f19d4a8f
AV
604struct posix_acl;
605#define ACL_NOT_CACHED ((void *)(-1))
2a3a2a3f 606#define ACL_DONT_CACHE ((void *)(-3))
f19d4a8f 607
b8a7a3a6
AG
608static inline struct posix_acl *
609uncached_acl_sentinel(struct task_struct *task)
610{
611 return (void *)task + 1;
612}
613
614static inline bool
615is_uncached_acl(struct posix_acl *acl)
616{
617 return (long)acl & 1;
618}
619
3ddcd056
LT
620#define IOP_FASTPERM 0x0001
621#define IOP_LOOKUP 0x0002
622#define IOP_NOFOLLOW 0x0004
d0a5b995 623#define IOP_XATTR 0x0008
76fca90e 624#define IOP_DEFAULT_READLINK 0x0010
3ddcd056 625
9dd813c1
JK
626struct fsnotify_mark_connector;
627
3ddcd056
LT
628/*
629 * Keep mostly read-only and often accessed (especially for
630 * the RCU path lookup and 'stat' data) fields at the beginning
631 * of the 'struct inode'
632 */
1da177e4 633struct inode {
44a7d7a8 634 umode_t i_mode;
3ddcd056 635 unsigned short i_opflags;
92361636
EB
636 kuid_t i_uid;
637 kgid_t i_gid;
3ddcd056
LT
638 unsigned int i_flags;
639
640#ifdef CONFIG_FS_POSIX_ACL
641 struct posix_acl *i_acl;
642 struct posix_acl *i_default_acl;
643#endif
644
44a7d7a8
NP
645 const struct inode_operations *i_op;
646 struct super_block *i_sb;
3ddcd056 647 struct address_space *i_mapping;
44a7d7a8 648
13e12d14
LT
649#ifdef CONFIG_SECURITY
650 void *i_security;
651#endif
44a7d7a8 652
3ddcd056
LT
653 /* Stat data, not accessed from path walking */
654 unsigned long i_ino;
a78ef704
MS
655 /*
656 * Filesystems may only read i_nlink directly. They shall use the
657 * following functions for modification:
658 *
659 * (set|clear|inc|drop)_nlink
660 * inode_(inc|dec)_link_count
661 */
662 union {
663 const unsigned int i_nlink;
664 unsigned int __i_nlink;
665 };
3ddcd056 666 dev_t i_rdev;
2f9d3df8 667 loff_t i_size;
95582b00
DD
668 struct timespec64 i_atime;
669 struct timespec64 i_mtime;
670 struct timespec64 i_ctime;
6cdbb0ef
TT
671 spinlock_t i_lock; /* i_blocks, i_bytes, maybe i_size */
672 unsigned short i_bytes;
69448867
AG
673 u8 i_blkbits;
674 u8 i_write_hint;
3ddcd056
LT
675 blkcnt_t i_blocks;
676
677#ifdef __NEED_I_SIZE_ORDERED
678 seqcount_t i_size_seqcount;
679#endif
680
681 /* Misc */
682 unsigned long i_state;
9902af79 683 struct rw_semaphore i_rwsem;
13e12d14 684
44a7d7a8 685 unsigned long dirtied_when; /* jiffies of first dirtying */
a2f48706 686 unsigned long dirtied_time_when;
44a7d7a8 687
1da177e4 688 struct hlist_node i_hash;
c7f54084 689 struct list_head i_io_list; /* backing dev IO list */
52ebea74
TH
690#ifdef CONFIG_CGROUP_WRITEBACK
691 struct bdi_writeback *i_wb; /* the associated cgroup wb */
2a814908
TH
692
693 /* foreign inode detection, see wbc_detach_inode() */
694 int i_wb_frn_winner;
695 u16 i_wb_frn_avg_time;
696 u16 i_wb_frn_history;
52ebea74 697#endif
7ccf19a8 698 struct list_head i_lru; /* inode LRU list */
1da177e4 699 struct list_head i_sb_list;
6c60d2b5 700 struct list_head i_wb_list; /* backing dev writeback list */
fa0d7e3d 701 union {
b3d9b7a3 702 struct hlist_head i_dentry;
fa0d7e3d
NP
703 struct rcu_head i_rcu;
704 };
f02a9ad1 705 atomic64_t i_version;
8019ad13 706 atomic64_t i_sequence; /* see futex */
2f9d3df8 707 atomic_t i_count;
bd5fe6c5 708 atomic_t i_dio_count;
6cdbb0ef 709 atomic_t i_writecount;
387e3746 710#if defined(CONFIG_IMA) || defined(CONFIG_FILE_LOCKING)
d984ea60
MZ
711 atomic_t i_readcount; /* struct files open RO */
712#endif
fdb0da89
AV
713 union {
714 const struct file_operations *i_fop; /* former ->i_op->default_file_ops */
715 void (*free_inode)(struct inode *);
716 };
4a075e39 717 struct file_lock_context *i_flctx;
1da177e4 718 struct address_space i_data;
1da177e4 719 struct list_head i_devices;
4c154168
TT
720 union {
721 struct pipe_inode_info *i_pipe;
eaf796e7 722 struct block_device *i_bdev;
577c4eb0 723 struct cdev *i_cdev;
61ba64fc 724 char *i_link;
84e710da 725 unsigned i_dir_seq;
4c154168 726 };
1da177e4
LT
727
728 __u32 i_generation;
729
3be25f49
EP
730#ifdef CONFIG_FSNOTIFY
731 __u32 i_fsnotify_mask; /* all events this inode cares about */
08991e83 732 struct fsnotify_mark_connector __rcu *i_fsnotify_marks;
0eeca283
RL
733#endif
734
643fa961 735#ifdef CONFIG_FS_ENCRYPTION
0b81d077
JK
736 struct fscrypt_info *i_crypt_info;
737#endif
738
5585f2af
EB
739#ifdef CONFIG_FS_VERITY
740 struct fsverity_info *i_verity_info;
741#endif
742
8e18e294 743 void *i_private; /* fs or device private pointer */
3859a271 744} __randomize_layout;
1da177e4 745
50e17c00
DD
746struct timespec64 timestamp_truncate(struct timespec64 t, struct inode *inode);
747
93407472
FF
748static inline unsigned int i_blocksize(const struct inode *node)
749{
750 return (1 << node->i_blkbits);
751}
752
1d3382cb
AV
753static inline int inode_unhashed(struct inode *inode)
754{
755 return hlist_unhashed(&inode->i_hash);
756}
757
5bef9151
AV
758/*
759 * __mark_inode_dirty expects inodes to be hashed. Since we don't
760 * want special inodes in the fileset inode space, we make them
761 * appear hashed, but do not put on any lists. hlist_del()
762 * will work fine and require no locking.
763 */
764static inline void inode_fake_hash(struct inode *inode)
765{
766 hlist_add_fake(&inode->i_hash);
767}
768
f2eace23
IM
769/*
770 * inode->i_mutex nesting subclasses for the lock validator:
771 *
772 * 0: the object of the current VFS operation
773 * 1: parent
774 * 2: child/target
40bd22c9
BF
775 * 3: xattr
776 * 4: second non-directory
d1b72cc6
MS
777 * 5: second parent (when locking independent directories in rename)
778 *
779 * I_MUTEX_NONDIR2 is for certain operations (such as rename) which lock two
40bd22c9 780 * non-directories at once.
f2eace23
IM
781 *
782 * The locking order between these classes is
d1b72cc6 783 * parent[2] -> child -> grandchild -> normal -> xattr -> second non-directory
f2eace23
IM
784 */
785enum inode_i_mutex_lock_class
786{
787 I_MUTEX_NORMAL,
788 I_MUTEX_PARENT,
789 I_MUTEX_CHILD,
4df46240 790 I_MUTEX_XATTR,
d1b72cc6
MS
791 I_MUTEX_NONDIR2,
792 I_MUTEX_PARENT2,
f2eace23
IM
793};
794
5955102c
AV
795static inline void inode_lock(struct inode *inode)
796{
9902af79 797 down_write(&inode->i_rwsem);
5955102c
AV
798}
799
800static inline void inode_unlock(struct inode *inode)
801{
9902af79
AV
802 up_write(&inode->i_rwsem);
803}
804
805static inline void inode_lock_shared(struct inode *inode)
806{
807 down_read(&inode->i_rwsem);
808}
809
810static inline void inode_unlock_shared(struct inode *inode)
811{
812 up_read(&inode->i_rwsem);
5955102c
AV
813}
814
815static inline int inode_trylock(struct inode *inode)
816{
9902af79
AV
817 return down_write_trylock(&inode->i_rwsem);
818}
819
820static inline int inode_trylock_shared(struct inode *inode)
821{
822 return down_read_trylock(&inode->i_rwsem);
5955102c
AV
823}
824
825static inline int inode_is_locked(struct inode *inode)
826{
9902af79 827 return rwsem_is_locked(&inode->i_rwsem);
5955102c
AV
828}
829
830static inline void inode_lock_nested(struct inode *inode, unsigned subclass)
831{
9902af79 832 down_write_nested(&inode->i_rwsem, subclass);
5955102c
AV
833}
834
01c2e13d
DW
835static inline void inode_lock_shared_nested(struct inode *inode, unsigned subclass)
836{
837 down_read_nested(&inode->i_rwsem, subclass);
838}
839
375e289e
BF
840void lock_two_nondirectories(struct inode *, struct inode*);
841void unlock_two_nondirectories(struct inode *, struct inode*);
842
1da177e4
LT
843/*
844 * NOTE: in a 32bit arch with a preemptable kernel and
845 * an UP compile the i_size_read/write must be atomic
846 * with respect to the local cpu (unlike with preempt disabled),
847 * but they don't need to be atomic with respect to other cpus like in
848 * true SMP (so they need either to either locally disable irq around
849 * the read or for example on x86 they can be still implemented as a
850 * cmpxchg8b without the need of the lock prefix). For SMP compiles
851 * and 64bit archs it makes no difference if preempt is enabled or not.
852 */
48ed214d 853static inline loff_t i_size_read(const struct inode *inode)
1da177e4
LT
854{
855#if BITS_PER_LONG==32 && defined(CONFIG_SMP)
856 loff_t i_size;
857 unsigned int seq;
858
859 do {
860 seq = read_seqcount_begin(&inode->i_size_seqcount);
861 i_size = inode->i_size;
862 } while (read_seqcount_retry(&inode->i_size_seqcount, seq));
863 return i_size;
2496396f 864#elif BITS_PER_LONG==32 && defined(CONFIG_PREEMPTION)
1da177e4
LT
865 loff_t i_size;
866
867 preempt_disable();
868 i_size = inode->i_size;
869 preempt_enable();
870 return i_size;
871#else
872 return inode->i_size;
873#endif
874}
875
7762f5a0
MS
876/*
877 * NOTE: unlike i_size_read(), i_size_write() does need locking around it
878 * (normally i_mutex), otherwise on 32bit/SMP an update of i_size_seqcount
879 * can be lost, resulting in subsequent i_size_read() calls spinning forever.
880 */
1da177e4
LT
881static inline void i_size_write(struct inode *inode, loff_t i_size)
882{
883#if BITS_PER_LONG==32 && defined(CONFIG_SMP)
74e3d1e1 884 preempt_disable();
1da177e4
LT
885 write_seqcount_begin(&inode->i_size_seqcount);
886 inode->i_size = i_size;
887 write_seqcount_end(&inode->i_size_seqcount);
74e3d1e1 888 preempt_enable();
2496396f 889#elif BITS_PER_LONG==32 && defined(CONFIG_PREEMPTION)
1da177e4
LT
890 preempt_disable();
891 inode->i_size = i_size;
892 preempt_enable();
893#else
894 inode->i_size = i_size;
895#endif
896}
897
48ed214d 898static inline unsigned iminor(const struct inode *inode)
1da177e4
LT
899{
900 return MINOR(inode->i_rdev);
901}
902
48ed214d 903static inline unsigned imajor(const struct inode *inode)
1da177e4
LT
904{
905 return MAJOR(inode->i_rdev);
906}
907
908extern struct block_device *I_BDEV(struct inode *inode);
909
910struct fown_struct {
911 rwlock_t lock; /* protects pid, uid, euid fields */
609d7fa9
EB
912 struct pid *pid; /* pid or -pgrp where SIGIO should be sent */
913 enum pid_type pid_type; /* Kind of process group SIGIO should be sent to */
92361636 914 kuid_t uid, euid; /* uid/euid of process setting the owner */
1da177e4
LT
915 int signum; /* posix.1b rt signal to be delivered on IO */
916};
917
918/*
919 * Track a single file's readahead state
920 */
921struct file_ra_state {
937085aa
FW
922 pgoff_t start; /* where readahead started */
923 unsigned int size; /* # of readahead pages */
924 unsigned int async_size; /* do asynchronous readahead when
f9acc8c7 925 there are only # of pages ahead */
5ce1110b 926
937085aa 927 unsigned int ra_pages; /* Maximum readahead window */
1ebf26a9 928 unsigned int mmap_miss; /* Cache miss stat for mmap accesses */
f4e6b498 929 loff_t prev_pos; /* Cache last read() position */
1da177e4 930};
1da177e4 931
5ce1110b
FW
932/*
933 * Check if @index falls in the readahead windows.
934 */
935static inline int ra_has_index(struct file_ra_state *ra, pgoff_t index)
936{
f9acc8c7
FW
937 return (index >= ra->start &&
938 index < ra->start + ra->size);
5ce1110b
FW
939}
940
1da177e4 941struct file {
2f512016 942 union {
4f5e65a1 943 struct llist_node fu_llist;
2f512016
ED
944 struct rcu_head fu_rcuhead;
945 } f_u;
0f7fc9e4 946 struct path f_path;
dd37978c 947 struct inode *f_inode; /* cached value */
99ac48f5 948 const struct file_operations *f_op;
ef3d0fd2
AK
949
950 /*
9c225f26 951 * Protects f_ep_links, f_flags.
ef3d0fd2
AK
952 * Must not be taken from IRQ context.
953 */
954 spinlock_t f_lock;
c75b1d94 955 enum rw_hint f_write_hint;
516e0cc5 956 atomic_long_t f_count;
1da177e4 957 unsigned int f_flags;
aeb5d727 958 fmode_t f_mode;
9c225f26 959 struct mutex f_pos_lock;
1da177e4
LT
960 loff_t f_pos;
961 struct fown_struct f_owner;
d76b0d9b 962 const struct cred *f_cred;
1da177e4
LT
963 struct file_ra_state f_ra;
964
2b47c361 965 u64 f_version;
50462062 966#ifdef CONFIG_SECURITY
1da177e4 967 void *f_security;
50462062 968#endif
1da177e4
LT
969 /* needed for tty driver, and maybe others */
970 void *private_data;
971
972#ifdef CONFIG_EPOLL
973 /* Used by fs/eventpoll.c to link all the hooks to this file */
974 struct list_head f_ep_links;
28d82dc1 975 struct list_head f_tfile_llink;
1da177e4
LT
976#endif /* #ifdef CONFIG_EPOLL */
977 struct address_space *f_mapping;
5660e13d 978 errseq_t f_wb_err;
3859a271
KC
979} __randomize_layout
980 __attribute__((aligned(4))); /* lest something weird decides that 2 is OK */
1da177e4 981
990d6c2d
AK
982struct file_handle {
983 __u32 handle_bytes;
984 int handle_type;
985 /* file identifier */
9d82973e 986 unsigned char f_handle[];
990d6c2d
AK
987};
988
cb0942b8
AV
989static inline struct file *get_file(struct file *f)
990{
991 atomic_long_inc(&f->f_count);
992 return f;
993}
091141a4
JA
994#define get_file_rcu_many(x, cnt) \
995 atomic_long_add_unless(&(x)->f_count, (cnt), 0)
996#define get_file_rcu(x) get_file_rcu_many((x), 1)
516e0cc5 997#define file_count(x) atomic_long_read(&(x)->f_count)
1da177e4
LT
998
999#define MAX_NON_LFS ((1UL<<31) - 1)
1000
1001/* Page cache limit. The filesystems should put that into their s_maxbytes
1002 limits, otherwise bad things can happen in VM. */
1003#if BITS_PER_LONG==32
0cc3b0ec 1004#define MAX_LFS_FILESIZE ((loff_t)ULONG_MAX << PAGE_SHIFT)
1da177e4 1005#elif BITS_PER_LONG==64
0cc3b0ec 1006#define MAX_LFS_FILESIZE ((loff_t)LLONG_MAX)
1da177e4
LT
1007#endif
1008
1009#define FL_POSIX 1
1010#define FL_FLOCK 2
617588d5 1011#define FL_DELEG 4 /* NFSv4 delegation */
1da177e4 1012#define FL_ACCESS 8 /* not trying to lock, just looking */
f475ae95 1013#define FL_EXISTS 16 /* when unlocking, test for existence */
1da177e4 1014#define FL_LEASE 32 /* lease held on this file */
75e1fcc0 1015#define FL_CLOSE 64 /* unlock on close */
1da177e4 1016#define FL_SLEEP 128 /* A blocking lock */
778fc546
BF
1017#define FL_DOWNGRADE_PENDING 256 /* Lease is being downgraded */
1018#define FL_UNLOCK_PENDING 512 /* Lease is being broken */
cff2fce5 1019#define FL_OFDLCK 1024 /* lock is "owned" by struct file */
11afe9f7 1020#define FL_LAYOUT 2048 /* outstanding pNFS layout */
1da177e4 1021
50f2112c
BC
1022#define FL_CLOSE_POSIX (FL_POSIX | FL_CLOSE)
1023
bde74e4b
MS
1024/*
1025 * Special return value from posix_lock_file() and vfs_lock_file() for
1026 * asynchronous locking.
1027 */
1028#define FILE_LOCK_DEFERRED 1
1029
7ca76311 1030/* legacy typedef, should eventually be removed */
17fa388d 1031typedef void *fl_owner_t;
1da177e4 1032
a7231a97
JL
1033struct file_lock;
1034
1da177e4 1035struct file_lock_operations {
1da177e4
LT
1036 void (*fl_copy_lock)(struct file_lock *, struct file_lock *);
1037 void (*fl_release_private)(struct file_lock *);
1038};
1039
1040struct lock_manager_operations {
cae80b30
JL
1041 fl_owner_t (*lm_get_owner)(fl_owner_t);
1042 void (*lm_put_owner)(fl_owner_t);
8fb47a4f 1043 void (*lm_notify)(struct file_lock *); /* unblock callback */
d0449b90 1044 int (*lm_grant)(struct file_lock *, int);
4d01b7f5 1045 bool (*lm_break)(struct file_lock *);
7448cc37 1046 int (*lm_change)(struct file_lock *, int, struct list_head *);
1c7dd2ff 1047 void (*lm_setup)(struct file_lock *, void **);
1da177e4
LT
1048};
1049
af558e33
BF
1050struct lock_manager {
1051 struct list_head list;
c87fb4a3
BF
1052 /*
1053 * NFSv4 and up also want opens blocked during the grace period;
1054 * NLM doesn't care:
1055 */
1056 bool block_opens;
af558e33
BF
1057};
1058
5ccb0066
SK
1059struct net;
1060void locks_start_grace(struct net *, struct lock_manager *);
af558e33 1061void locks_end_grace(struct lock_manager *);
003278e4
CL
1062bool locks_in_grace(struct net *);
1063bool opens_in_grace(struct net *);
af558e33 1064
1da177e4
LT
1065/* that will die - we need it for nfs_lock_info */
1066#include <linux/nfs_fs_i.h>
1067
1cb36012
JL
1068/*
1069 * struct file_lock represents a generic "file lock". It's used to represent
1070 * POSIX byte range locks, BSD (flock) locks, and leases. It's important to
1071 * note that the same struct is used to represent both a request for a lock and
1072 * the lock itself, but the same object is never used for both.
1073 *
1074 * FIXME: should we create a separate "struct lock_request" to help distinguish
1075 * these two uses?
1076 *
8116bf4c 1077 * The varous i_flctx lists are ordered by:
1cb36012 1078 *
8116bf4c
JL
1079 * 1) lock owner
1080 * 2) lock range start
1081 * 3) lock range end
1cb36012
JL
1082 *
1083 * Obviously, the last two criteria only matter for POSIX locks.
1084 */
1da177e4 1085struct file_lock {
ada5c1da 1086 struct file_lock *fl_blocker; /* The lock, that is blocking us */
6dee60f6 1087 struct list_head fl_list; /* link into file_lock_context */
139ca04e 1088 struct hlist_node fl_link; /* node in global lists */
ada5c1da
N
1089 struct list_head fl_blocked_requests; /* list of requests with
1090 * ->fl_blocker pointing here
1091 */
1092 struct list_head fl_blocked_member; /* node in
1093 * ->fl_blocker->fl_blocked_requests
1094 */
1da177e4 1095 fl_owner_t fl_owner;
710b7216 1096 unsigned int fl_flags;
afc1246f 1097 unsigned char fl_type;
1da177e4 1098 unsigned int fl_pid;
7012b02a 1099 int fl_link_cpu; /* what cpu's list is this on? */
1da177e4
LT
1100 wait_queue_head_t fl_wait;
1101 struct file *fl_file;
1da177e4
LT
1102 loff_t fl_start;
1103 loff_t fl_end;
1104
1105 struct fasync_struct * fl_fasync; /* for lease break notifications */
778fc546
BF
1106 /* for lease breaks: */
1107 unsigned long fl_break_time;
1108 unsigned long fl_downgrade_time;
1da177e4 1109
6aed6285 1110 const struct file_lock_operations *fl_ops; /* Callbacks for filesystems */
7b021967 1111 const struct lock_manager_operations *fl_lmops; /* Callbacks for lockmanagers */
1da177e4
LT
1112 union {
1113 struct nfs_lock_info nfs_fl;
8d0a8a9d 1114 struct nfs4_lock_info nfs4_fl;
e8d6c554
DH
1115 struct {
1116 struct list_head link; /* link in AFS vnode's pending_locks list */
1117 int state; /* state of grant or error if -ve */
d4696601 1118 unsigned int debug_id;
e8d6c554 1119 } afs;
1da177e4 1120 } fl_u;
3859a271 1121} __randomize_layout;
1da177e4 1122
4a075e39 1123struct file_lock_context {
6109c850 1124 spinlock_t flc_lock;
4a075e39 1125 struct list_head flc_flock;
bd61e0a9 1126 struct list_head flc_posix;
8634b51f 1127 struct list_head flc_lease;
4a075e39
JL
1128};
1129
1da177e4
LT
1130/* The following constant reflects the upper bound of the file/locking space */
1131#ifndef OFFSET_MAX
1132#define INT_LIMIT(x) (~((x)1 << (sizeof(x)*8 - 1)))
1133#define OFFSET_MAX INT_LIMIT(loff_t)
1134#define OFFT_OFFSET_MAX INT_LIMIT(off_t)
1135#endif
1136
bfcd17a6
TP
1137extern void send_sigio(struct fown_struct *fown, int fd, int band);
1138
de2a4a50 1139#define locks_inode(f) file_inode(f)
c568d683 1140
bfcd17a6 1141#ifdef CONFIG_FILE_LOCKING
a75d30c7 1142extern int fcntl_getlk(struct file *, unsigned int, struct flock *);
c293621b 1143extern int fcntl_setlk(unsigned int, struct file *, unsigned int,
a75d30c7 1144 struct flock *);
1da177e4
LT
1145
1146#if BITS_PER_LONG == 32
a75d30c7 1147extern int fcntl_getlk64(struct file *, unsigned int, struct flock64 *);
c293621b 1148extern int fcntl_setlk64(unsigned int, struct file *, unsigned int,
a75d30c7 1149 struct flock64 *);
1da177e4
LT
1150#endif
1151
1da177e4
LT
1152extern int fcntl_setlease(unsigned int fd, struct file *filp, long arg);
1153extern int fcntl_getlease(struct file *filp);
1154
1155/* fs/locks.c */
f27a0fe0 1156void locks_free_lock_context(struct inode *inode);
05fa3135 1157void locks_free_lock(struct file_lock *fl);
1da177e4 1158extern void locks_init_lock(struct file_lock *);
c5b1f0d9 1159extern struct file_lock * locks_alloc_lock(void);
1da177e4 1160extern void locks_copy_lock(struct file_lock *, struct file_lock *);
3fe0fff1 1161extern void locks_copy_conflock(struct file_lock *, struct file_lock *);
1da177e4 1162extern void locks_remove_posix(struct file *, fl_owner_t);
78ed8a13 1163extern void locks_remove_file(struct file *);
a9e61e25 1164extern void locks_release_private(struct file_lock *);
6d34ac19 1165extern void posix_test_lock(struct file *, struct file_lock *);
150b3934 1166extern int posix_lock_file(struct file *, struct file_lock *, struct file_lock *);
cb03f94f 1167extern int locks_delete_block(struct file_lock *);
3ee17abd 1168extern int vfs_test_lock(struct file *, struct file_lock *);
150b3934 1169extern int vfs_lock_file(struct file *, unsigned int, struct file_lock *, struct file_lock *);
9b9d2ab4 1170extern int vfs_cancel_lock(struct file *filp, struct file_lock *fl);
e55c34a6 1171extern int locks_lock_inode_wait(struct inode *inode, struct file_lock *fl);
df4e8d2c 1172extern int __break_lease(struct inode *inode, unsigned int flags, unsigned int type);
95582b00 1173extern void lease_get_mtime(struct inode *, struct timespec64 *time);
e6f5c789
JL
1174extern int generic_setlease(struct file *, long, struct file_lock **, void **priv);
1175extern int vfs_setlease(struct file *, long, struct file_lock **, void **);
7448cc37 1176extern int lease_modify(struct file_lock *, int, struct list_head *);
18f6622e
JL
1177
1178struct notifier_block;
1179extern int lease_register_notifier(struct notifier_block *);
1180extern void lease_unregister_notifier(struct notifier_block *);
1181
6c8c9031
AV
1182struct files_struct;
1183extern void show_fd_locks(struct seq_file *f,
1184 struct file *filp, struct files_struct *files);
bfcd17a6 1185#else /* !CONFIG_FILE_LOCKING */
c1e62b8f
JL
1186static inline int fcntl_getlk(struct file *file, unsigned int cmd,
1187 struct flock __user *user)
c2aca5e5
SW
1188{
1189 return -EINVAL;
1190}
1191
1192static inline int fcntl_setlk(unsigned int fd, struct file *file,
1193 unsigned int cmd, struct flock __user *user)
1194{
1195 return -EACCES;
1196}
1197
bfcd17a6 1198#if BITS_PER_LONG == 32
c1e62b8f
JL
1199static inline int fcntl_getlk64(struct file *file, unsigned int cmd,
1200 struct flock64 __user *user)
c2aca5e5
SW
1201{
1202 return -EINVAL;
1203}
1204
1205static inline int fcntl_setlk64(unsigned int fd, struct file *file,
1206 unsigned int cmd, struct flock64 __user *user)
1207{
1208 return -EACCES;
1209}
bfcd17a6 1210#endif
c2aca5e5
SW
1211static inline int fcntl_setlease(unsigned int fd, struct file *filp, long arg)
1212{
1b2b32dc 1213 return -EINVAL;
c2aca5e5
SW
1214}
1215
1216static inline int fcntl_getlease(struct file *filp)
1217{
1b2b32dc 1218 return F_UNLCK;
c2aca5e5
SW
1219}
1220
4a075e39 1221static inline void
f27a0fe0 1222locks_free_lock_context(struct inode *inode)
4a075e39
JL
1223{
1224}
1225
c2aca5e5
SW
1226static inline void locks_init_lock(struct file_lock *fl)
1227{
1228 return;
1229}
1230
3fe0fff1 1231static inline void locks_copy_conflock(struct file_lock *new, struct file_lock *fl)
c2aca5e5
SW
1232{
1233 return;
1234}
1235
1236static inline void locks_copy_lock(struct file_lock *new, struct file_lock *fl)
1237{
1238 return;
1239}
1240
1241static inline void locks_remove_posix(struct file *filp, fl_owner_t owner)
1242{
1243 return;
1244}
1245
78ed8a13 1246static inline void locks_remove_file(struct file *filp)
c2aca5e5
SW
1247{
1248 return;
1249}
1250
1251static inline void posix_test_lock(struct file *filp, struct file_lock *fl)
1252{
1253 return;
1254}
1255
1256static inline int posix_lock_file(struct file *filp, struct file_lock *fl,
1257 struct file_lock *conflock)
1258{
1259 return -ENOLCK;
1260}
1261
cb03f94f 1262static inline int locks_delete_block(struct file_lock *waiter)
c2aca5e5
SW
1263{
1264 return -ENOENT;
1265}
1266
1267static inline int vfs_test_lock(struct file *filp, struct file_lock *fl)
1268{
1269 return 0;
1270}
1271
1272static inline int vfs_lock_file(struct file *filp, unsigned int cmd,
1273 struct file_lock *fl, struct file_lock *conf)
1274{
1275 return -ENOLCK;
1276}
1277
1278static inline int vfs_cancel_lock(struct file *filp, struct file_lock *fl)
1279{
1280 return 0;
1281}
1282
e55c34a6
BC
1283static inline int locks_lock_inode_wait(struct inode *inode, struct file_lock *fl)
1284{
1285 return -ENOLCK;
1286}
1287
df4e8d2c 1288static inline int __break_lease(struct inode *inode, unsigned int mode, unsigned int type)
c2aca5e5
SW
1289{
1290 return 0;
1291}
1292
95582b00
DD
1293static inline void lease_get_mtime(struct inode *inode,
1294 struct timespec64 *time)
c2aca5e5
SW
1295{
1296 return;
1297}
1298
1299static inline int generic_setlease(struct file *filp, long arg,
e6f5c789 1300 struct file_lock **flp, void **priv)
c2aca5e5
SW
1301{
1302 return -EINVAL;
1303}
1304
1305static inline int vfs_setlease(struct file *filp, long arg,
e6f5c789 1306 struct file_lock **lease, void **priv)
c2aca5e5
SW
1307{
1308 return -EINVAL;
1309}
1310
7448cc37 1311static inline int lease_modify(struct file_lock *fl, int arg,
c45198ed 1312 struct list_head *dispose)
c2aca5e5
SW
1313{
1314 return -EINVAL;
1315}
6c8c9031
AV
1316
1317struct files_struct;
1318static inline void show_fd_locks(struct seq_file *f,
1319 struct file *filp, struct files_struct *files) {}
bfcd17a6
TP
1320#endif /* !CONFIG_FILE_LOCKING */
1321
ee296d7c
JL
1322static inline struct inode *file_inode(const struct file *f)
1323{
1324 return f->f_inode;
1325}
1326
d101a125
MS
1327static inline struct dentry *file_dentry(const struct file *file)
1328{
fb16043b 1329 return d_real(file->f_path.dentry, file_inode(file));
d101a125
MS
1330}
1331
e55c34a6
BC
1332static inline int locks_lock_file_wait(struct file *filp, struct file_lock *fl)
1333{
c568d683 1334 return locks_lock_inode_wait(locks_inode(filp), fl);
e55c34a6
BC
1335}
1336
1da177e4 1337struct fasync_struct {
7a107c0f 1338 rwlock_t fa_lock;
989a2979
ED
1339 int magic;
1340 int fa_fd;
1341 struct fasync_struct *fa_next; /* singly linked list */
1342 struct file *fa_file;
1343 struct rcu_head fa_rcu;
1da177e4
LT
1344};
1345
1346#define FASYNC_MAGIC 0x4601
1347
1348/* SMP safe fasync helpers: */
1349extern int fasync_helper(int, struct file *, int, struct fasync_struct **);
f7347ce4
LT
1350extern struct fasync_struct *fasync_insert_entry(int, struct file *, struct fasync_struct **, struct fasync_struct *);
1351extern int fasync_remove_entry(struct file *, struct fasync_struct **);
1352extern struct fasync_struct *fasync_alloc(void);
1353extern void fasync_free(struct fasync_struct *);
1354
1da177e4
LT
1355/* can be called from interrupts */
1356extern void kill_fasync(struct fasync_struct **, int, int);
1da177e4 1357
e0b93edd 1358extern void __f_setown(struct file *filp, struct pid *, enum pid_type, int force);
393cc3f5 1359extern int f_setown(struct file *filp, unsigned long arg, int force);
1da177e4 1360extern void f_delown(struct file *filp);
609d7fa9 1361extern pid_t f_getown(struct file *filp);
1da177e4
LT
1362extern int send_sigurg(struct fown_struct *fown);
1363
e462ec50
DH
1364/*
1365 * sb->s_flags. Note that these mirror the equivalent MS_* flags where
1366 * represented in both.
1367 */
1368#define SB_RDONLY 1 /* Mount read-only */
1369#define SB_NOSUID 2 /* Ignore suid and sgid bits */
1370#define SB_NODEV 4 /* Disallow access to device special files */
1371#define SB_NOEXEC 8 /* Disallow program execution */
1372#define SB_SYNCHRONOUS 16 /* Writes are synced at once */
1373#define SB_MANDLOCK 64 /* Allow mandatory locks on an FS */
1374#define SB_DIRSYNC 128 /* Directory modifications are synchronous */
1375#define SB_NOATIME 1024 /* Do not update access times. */
1376#define SB_NODIRATIME 2048 /* Do not update directory access times */
1377#define SB_SILENT 32768
1378#define SB_POSIXACL (1<<16) /* VFS does not apply the umask */
1379#define SB_KERNMOUNT (1<<22) /* this is a kern_mount call */
1380#define SB_I_VERSION (1<<23) /* Update inode I_version field */
1381#define SB_LAZYTIME (1<<25) /* Update the on-disk [acm]times lazily */
1382
1383/* These sb flags are internal to the kernel */
1384#define SB_SUBMOUNT (1<<26)
8d0347f6 1385#define SB_FORCE (1<<27)
e462ec50
DH
1386#define SB_NOSEC (1<<28)
1387#define SB_BORN (1<<29)
1388#define SB_ACTIVE (1<<30)
1389#define SB_NOUSER (1<<31)
1390
1da177e4
LT
1391/*
1392 * Umount options
1393 */
1394
1395#define MNT_FORCE 0x00000001 /* Attempt to forcibily umount */
1396#define MNT_DETACH 0x00000002 /* Just detach from the tree */
1397#define MNT_EXPIRE 0x00000004 /* Mark for expiry */
db1f05bb
MS
1398#define UMOUNT_NOFOLLOW 0x00000008 /* Don't follow symlink on umount */
1399#define UMOUNT_UNUSED 0x80000000 /* Flag guaranteed to be unused */
1da177e4 1400
46b15caa
TH
1401/* sb->s_iflags */
1402#define SB_I_CGROUPWB 0x00000001 /* cgroup-aware writeback enabled */
90f8572b 1403#define SB_I_NOEXEC 0x00000002 /* Ignore executables on this fs */
a2982cc9 1404#define SB_I_NODEV 0x00000004 /* Ignore devices on this fs */
95dd7758 1405#define SB_I_MULTIROOT 0x00000008 /* Multiple roots to the dentry tree */
1da177e4 1406
8654df4e
EB
1407/* sb->s_iflags to limit user namespace mounts */
1408#define SB_I_USERNS_VISIBLE 0x00000010 /* fstype already mounted */
57b56ac6
MZ
1409#define SB_I_IMA_UNVERIFIABLE_SIGNATURE 0x00000020
1410#define SB_I_UNTRUSTED_MOUNTER 0x00000040
1da177e4 1411
5accdf82
JK
1412/* Possible states of 'frozen' field */
1413enum {
1414 SB_UNFROZEN = 0, /* FS is unfrozen */
1415 SB_FREEZE_WRITE = 1, /* Writes, dir ops, ioctls frozen */
5accdf82
JK
1416 SB_FREEZE_PAGEFAULT = 2, /* Page faults stopped as well */
1417 SB_FREEZE_FS = 3, /* For internal FS use (e.g. to stop
1418 * internal threads if needed) */
1419 SB_FREEZE_COMPLETE = 4, /* ->freeze_fs finished successfully */
1420};
1421
1422#define SB_FREEZE_LEVELS (SB_FREEZE_COMPLETE - 1)
1423
1424struct sb_writers {
8129ed29
ON
1425 int frozen; /* Is sb frozen? */
1426 wait_queue_head_t wait_unfrozen; /* for get_super_thawed() */
1427 struct percpu_rw_semaphore rw_sem[SB_FREEZE_LEVELS];
5accdf82
JK
1428};
1429
1da177e4
LT
1430struct super_block {
1431 struct list_head s_list; /* Keep this first */
1432 dev_t s_dev; /* search index; _not_ kdev_t */
270ba5f7
RK
1433 unsigned char s_blocksize_bits;
1434 unsigned long s_blocksize;
42cb56ae 1435 loff_t s_maxbytes; /* Max file size */
1da177e4 1436 struct file_system_type *s_type;
ee9b6d61 1437 const struct super_operations *s_op;
61e225dc 1438 const struct dquot_operations *dq_op;
0d54b217 1439 const struct quotactl_ops *s_qcop;
39655164 1440 const struct export_operations *s_export_op;
1da177e4 1441 unsigned long s_flags;
46b15caa 1442 unsigned long s_iflags; /* internal SB_I_* flags */
1da177e4
LT
1443 unsigned long s_magic;
1444 struct dentry *s_root;
1445 struct rw_semaphore s_umount;
1da177e4 1446 int s_count;
1da177e4 1447 atomic_t s_active;
50462062 1448#ifdef CONFIG_SECURITY
1da177e4 1449 void *s_security;
50462062 1450#endif
bb435453 1451 const struct xattr_handler **s_xattr;
643fa961 1452#ifdef CONFIG_FS_ENCRYPTION
0b81d077 1453 const struct fscrypt_operations *s_cop;
22d94f49 1454 struct key *s_master_keys; /* master crypto keys in use */
5585f2af
EB
1455#endif
1456#ifdef CONFIG_FS_VERITY
1457 const struct fsverity_operations *s_vop;
bbbc3fb6 1458#endif
f1ee6162 1459 struct hlist_bl_head s_roots; /* alternate root dentries for NFS */
39f7c4db 1460 struct list_head s_mounts; /* list of mounts; _not_ for fs use */
1da177e4 1461 struct block_device *s_bdev;
32a88aa1 1462 struct backing_dev_info *s_bdi;
acaebfd8 1463 struct mtd_info *s_mtd;
a5166169 1464 struct hlist_node s_instances;
2c5f648a 1465 unsigned int s_quota_types; /* Bitmask of supported quota types */
1da177e4
LT
1466 struct quota_info s_dquot; /* Diskquota specific options */
1467
5accdf82 1468 struct sb_writers s_writers;
1da177e4 1469
99c228a9
AG
1470 /*
1471 * Keep s_fs_info, s_time_gran, s_fsnotify_mask, and
1472 * s_fsnotify_marks together for cache efficiency. They are frequently
1473 * accessed and rarely modified.
1474 */
1475 void *s_fs_info; /* Filesystem private info */
1476
1477 /* Granularity of c/m/atime in ns (cannot be worse than a second) */
1478 u32 s_time_gran;
188d20bc
DD
1479 /* Time limits for c/m/atime in seconds */
1480 time64_t s_time_min;
1481 time64_t s_time_max;
99c228a9
AG
1482#ifdef CONFIG_FSNOTIFY
1483 __u32 s_fsnotify_mask;
1484 struct fsnotify_mark_connector __rcu *s_fsnotify_marks;
1485#endif
1486
85787090
CH
1487 char s_id[32]; /* Informational name */
1488 uuid_t s_uuid; /* UUID */
1da177e4 1489
8de52778 1490 unsigned int s_max_links;
30c40d2c 1491 fmode_t s_mode;
1da177e4
LT
1492
1493 /*
1494 * The next field is for VFS *only*. No filesystems have any business
1495 * even looking at it. You had been warned.
1496 */
a11f3a05 1497 struct mutex s_vfs_rename_mutex; /* Kludge */
1da177e4 1498
79c0b2df
MS
1499 /*
1500 * Filesystem subtype. If non-empty the filesystem type field
1501 * in /proc/mounts will be "type.subtype"
1502 */
a0c9a8b8 1503 const char *s_subtype;
b3b304a2 1504
c8aebb0c 1505 const struct dentry_operations *s_d_op; /* default d_op for dentries */
9fdfdcf1
DM
1506
1507 /*
1508 * Saved pool identifier for cleancache (-1 means none)
1509 */
1510 int cleancache_poolid;
b0d40c92
DC
1511
1512 struct shrinker s_shrink; /* per-sb shrinker handle */
4ed5e82f 1513
7ada4db8
MS
1514 /* Number of inodes with nlink == 0 but still referenced */
1515 atomic_long_t s_remove_count;
1516
721fb6fb
JK
1517 /* Pending fsnotify inode refs */
1518 atomic_long_t s_fsnotify_inode_refs;
1519
4ed5e82f
MS
1520 /* Being remounted read-only */
1521 int s_readonly_remount;
7b7a8665
CH
1522
1523 /* AIO completions deferred from interrupt context */
1524 struct workqueue_struct *s_dio_done_wq;
215752fc 1525 struct hlist_head s_pins;
f6041567 1526
6e4eab57
EB
1527 /*
1528 * Owning user namespace and default context in which to
1529 * interpret filesystem uids, gids, quotas, device nodes,
1530 * xattrs and security labels.
1531 */
1532 struct user_namespace *s_user_ns;
1533
f6041567 1534 /*
7d10f70f
WL
1535 * The list_lru structure is essentially just a pointer to a table
1536 * of per-node lru lists, each of which has its own spinlock.
1537 * There is no need to put them into separate cachelines.
f6041567 1538 */
7d10f70f
WL
1539 struct list_lru s_dentry_lru;
1540 struct list_lru s_inode_lru;
e2fec7c3 1541 struct rcu_head rcu;
853b39a7 1542 struct work_struct destroy_work;
69c433ed 1543
e97fedb9 1544 struct mutex s_sync_lock; /* sync serialisation lock */
69c433ed
MS
1545
1546 /*
1547 * Indicates how deep in a filesystem stack this SB is
1548 */
1549 int s_stack_depth;
74278da9
DC
1550
1551 /* s_inode_list_lock protects s_inodes */
1552 spinlock_t s_inode_list_lock ____cacheline_aligned_in_smp;
1553 struct list_head s_inodes; /* all inodes */
6c60d2b5
DC
1554
1555 spinlock_t s_inode_wblist_lock;
1556 struct list_head s_inodes_wb; /* writeback inodes */
3859a271 1557} __randomize_layout;
1da177e4 1558
81754357
SF
1559/* Helper functions so that in most cases filesystems will
1560 * not need to deal directly with kuid_t and kgid_t and can
1561 * instead deal with the raw numeric values that are stored
1562 * in the filesystem.
1563 */
1564static inline uid_t i_uid_read(const struct inode *inode)
1565{
1566 return from_kuid(inode->i_sb->s_user_ns, inode->i_uid);
1567}
1568
1569static inline gid_t i_gid_read(const struct inode *inode)
1570{
1571 return from_kgid(inode->i_sb->s_user_ns, inode->i_gid);
1572}
1573
1574static inline void i_uid_write(struct inode *inode, uid_t uid)
1575{
1576 inode->i_uid = make_kuid(inode->i_sb->s_user_ns, uid);
1577}
1578
1579static inline void i_gid_write(struct inode *inode, gid_t gid)
1580{
1581 inode->i_gid = make_kgid(inode->i_sb->s_user_ns, gid);
1582}
1583
95582b00 1584extern struct timespec64 current_time(struct inode *inode);
1da177e4
LT
1585
1586/*
1587 * Snapshotting support.
1588 */
1da177e4 1589
5accdf82
JK
1590void __sb_end_write(struct super_block *sb, int level);
1591int __sb_start_write(struct super_block *sb, int level, bool wait);
1592
bee9182d 1593#define __sb_writers_acquired(sb, lev) \
8129ed29 1594 percpu_rwsem_acquire(&(sb)->s_writers.rw_sem[(lev)-1], 1, _THIS_IP_)
bee9182d 1595#define __sb_writers_release(sb, lev) \
8129ed29 1596 percpu_rwsem_release(&(sb)->s_writers.rw_sem[(lev)-1], 1, _THIS_IP_)
bee9182d 1597
5accdf82
JK
1598/**
1599 * sb_end_write - drop write access to a superblock
1600 * @sb: the super we wrote to
1601 *
1602 * Decrement number of writers to the filesystem. Wake up possible waiters
1603 * wanting to freeze the filesystem.
1604 */
1605static inline void sb_end_write(struct super_block *sb)
1606{
1607 __sb_end_write(sb, SB_FREEZE_WRITE);
1608}
1609
1610/**
1611 * sb_end_pagefault - drop write access to a superblock from a page fault
1612 * @sb: the super we wrote to
1613 *
1614 * Decrement number of processes handling write page fault to the filesystem.
1615 * Wake up possible waiters wanting to freeze the filesystem.
1616 */
1617static inline void sb_end_pagefault(struct super_block *sb)
1618{
1619 __sb_end_write(sb, SB_FREEZE_PAGEFAULT);
1620}
1621
1622/**
1623 * sb_end_intwrite - drop write access to a superblock for internal fs purposes
1624 * @sb: the super we wrote to
1625 *
1626 * Decrement fs-internal number of writers to the filesystem. Wake up possible
1627 * waiters wanting to freeze the filesystem.
1628 */
1629static inline void sb_end_intwrite(struct super_block *sb)
1630{
1631 __sb_end_write(sb, SB_FREEZE_FS);
1632}
1633
1634/**
1635 * sb_start_write - get write access to a superblock
1636 * @sb: the super we write to
1637 *
1638 * When a process wants to write data or metadata to a file system (i.e. dirty
1639 * a page or an inode), it should embed the operation in a sb_start_write() -
1640 * sb_end_write() pair to get exclusion against file system freezing. This
1641 * function increments number of writers preventing freezing. If the file
1642 * system is already frozen, the function waits until the file system is
1643 * thawed.
1644 *
1645 * Since freeze protection behaves as a lock, users have to preserve
1646 * ordering of freeze protection and other filesystem locks. Generally,
1647 * freeze protection should be the outermost lock. In particular, we have:
1648 *
1649 * sb_start_write
1650 * -> i_mutex (write path, truncate, directory ops, ...)
1651 * -> s_umount (freeze_super, thaw_super)
1652 */
1653static inline void sb_start_write(struct super_block *sb)
1654{
1655 __sb_start_write(sb, SB_FREEZE_WRITE, true);
1656}
1657
1658static inline int sb_start_write_trylock(struct super_block *sb)
1659{
1660 return __sb_start_write(sb, SB_FREEZE_WRITE, false);
1661}
1662
1663/**
1664 * sb_start_pagefault - get write access to a superblock from a page fault
1665 * @sb: the super we write to
1666 *
1667 * When a process starts handling write page fault, it should embed the
1668 * operation into sb_start_pagefault() - sb_end_pagefault() pair to get
1669 * exclusion against file system freezing. This is needed since the page fault
1670 * is going to dirty a page. This function increments number of running page
1671 * faults preventing freezing. If the file system is already frozen, the
1672 * function waits until the file system is thawed.
1673 *
1674 * Since page fault freeze protection behaves as a lock, users have to preserve
1675 * ordering of freeze protection and other filesystem locks. It is advised to
1676 * put sb_start_pagefault() close to mmap_sem in lock ordering. Page fault
1677 * handling code implies lock dependency:
1678 *
1679 * mmap_sem
1680 * -> sb_start_pagefault
1681 */
1682static inline void sb_start_pagefault(struct super_block *sb)
1683{
1684 __sb_start_write(sb, SB_FREEZE_PAGEFAULT, true);
1685}
1686
1687/*
1688 * sb_start_intwrite - get write access to a superblock for internal fs purposes
1689 * @sb: the super we write to
1690 *
1691 * This is the third level of protection against filesystem freezing. It is
1692 * free for use by a filesystem. The only requirement is that it must rank
1693 * below sb_start_pagefault.
1694 *
1695 * For example filesystem can call sb_start_intwrite() when starting a
1696 * transaction which somewhat eases handling of freezing for internal sources
1697 * of filesystem changes (internal fs threads, discarding preallocation on file
1698 * close, etc.).
1699 */
1700static inline void sb_start_intwrite(struct super_block *sb)
1701{
1702 __sb_start_write(sb, SB_FREEZE_FS, true);
1703}
1da177e4 1704
0c8e3fe3
AG
1705static inline int sb_start_intwrite_trylock(struct super_block *sb)
1706{
1707 return __sb_start_write(sb, SB_FREEZE_FS, false);
1708}
1709
1da177e4 1710
2e149670 1711extern bool inode_owner_or_capable(const struct inode *inode);
3bd858ab 1712
1da177e4
LT
1713/*
1714 * VFS helper functions..
1715 */
312b63fb 1716extern int vfs_create(struct inode *, struct dentry *, umode_t, bool);
18bb1db3 1717extern int vfs_mkdir(struct inode *, struct dentry *, umode_t);
1a67aafb 1718extern int vfs_mknod(struct inode *, struct dentry *, umode_t, dev_t);
db2e747b 1719extern int vfs_symlink(struct inode *, struct dentry *, const char *);
146a8595 1720extern int vfs_link(struct dentry *, struct inode *, struct dentry *, struct inode **);
1da177e4 1721extern int vfs_rmdir(struct inode *, struct dentry *);
b21996e3 1722extern int vfs_unlink(struct inode *, struct dentry *, struct inode **);
520c8b16 1723extern int vfs_rename(struct inode *, struct dentry *, struct inode *, struct dentry *, struct inode **, unsigned int);
787fb6bc 1724extern int vfs_whiteout(struct inode *, struct dentry *);
1da177e4 1725
af7bd4dc
AG
1726extern struct dentry *vfs_tmpfile(struct dentry *dentry, umode_t mode,
1727 int open_flag);
1728
8e6c848e
AV
1729int vfs_mkobj(struct dentry *, umode_t,
1730 int (*f)(struct dentry *, umode_t, void *),
1731 void *);
1732
9df6702a
MS
1733extern long vfs_ioctl(struct file *file, unsigned int cmd, unsigned long arg);
1734
2952db0f
AB
1735#ifdef CONFIG_COMPAT
1736extern long compat_ptr_ioctl(struct file *file, unsigned int cmd,
1737 unsigned long arg);
1738#else
1739#define compat_ptr_ioctl NULL
1740#endif
1741
8c744fb8
CH
1742/*
1743 * VFS file helper functions.
1744 */
a1bd120d 1745extern void inode_init_owner(struct inode *inode, const struct inode *dir,
62bb1091 1746 umode_t mode);
a2982cc9 1747extern bool may_open_dev(const struct path *path);
c4b929b8
MF
1748/*
1749 * VFS FS_IOC_FIEMAP helper definitions.
1750 */
1751struct fiemap_extent_info {
1752 unsigned int fi_flags; /* Flags as passed from user */
1753 unsigned int fi_extents_mapped; /* Number of mapped extents */
1754 unsigned int fi_extents_max; /* Size of fiemap_extent array */
ecf5632d
NK
1755 struct fiemap_extent __user *fi_extents_start; /* Start of
1756 fiemap_extent array */
c4b929b8
MF
1757};
1758int fiemap_fill_next_extent(struct fiemap_extent_info *info, u64 logical,
1759 u64 phys, u64 len, u32 flags);
1760int fiemap_check_flags(struct fiemap_extent_info *fieinfo, u32 fs_flags);
1761
1da177e4
LT
1762/*
1763 * This is the "filldir" function type, used by readdir() to let
1764 * the kernel specify what kind of dirent layout it wants to have.
1765 * This allows the kernel to read directories into kernel space or
1766 * to have different dirent layouts depending on the binary type.
1767 */
ac7576f4
MS
1768struct dir_context;
1769typedef int (*filldir_t)(struct dir_context *, const char *, int, loff_t, u64,
1770 unsigned);
1771
5c0ba4e0 1772struct dir_context {
a09acf4b 1773 filldir_t actor;
bb6f619b 1774 loff_t pos;
5c0ba4e0 1775};
bb6f619b 1776
08f85851 1777struct block_device_operations;
1da177e4 1778
1da177e4
LT
1779/* These macros are for out of kernel modules to test that
1780 * the kernel supports the unlocked_ioctl and compat_ioctl
1781 * fields in struct file_operations. */
1782#define HAVE_COMPAT_IOCTL 1
1783#define HAVE_UNLOCKED_IOCTL 1
1784
b4caecd4
CH
1785/*
1786 * These flags let !MMU mmap() govern direct device mapping vs immediate
1787 * copying more easily for MAP_PRIVATE, especially for ROM filesystems.
1788 *
1789 * NOMMU_MAP_COPY: Copy can be mapped (MAP_PRIVATE)
1790 * NOMMU_MAP_DIRECT: Can be mapped directly (MAP_SHARED)
1791 * NOMMU_MAP_READ: Can be mapped for reading
1792 * NOMMU_MAP_WRITE: Can be mapped for writing
1793 * NOMMU_MAP_EXEC: Can be mapped for execution
1794 */
1795#define NOMMU_MAP_COPY 0x00000001
1796#define NOMMU_MAP_DIRECT 0x00000008
1797#define NOMMU_MAP_READ VM_MAYREAD
1798#define NOMMU_MAP_WRITE VM_MAYWRITE
1799#define NOMMU_MAP_EXEC VM_MAYEXEC
1800
1801#define NOMMU_VMFLAGS \
1802 (NOMMU_MAP_READ | NOMMU_MAP_WRITE | NOMMU_MAP_EXEC)
1803
2e5dfc99
DW
1804/*
1805 * These flags control the behavior of the remap_file_range function pointer.
1806 * If it is called with len == 0 that means "remap to end of source file".
5c437fa2 1807 * See Documentation/filesystems/vfs.rst for more details about this call.
2e5dfc99
DW
1808 *
1809 * REMAP_FILE_DEDUP: only remap if contents identical (i.e. deduplicate)
eca3654e 1810 * REMAP_FILE_CAN_SHORTEN: caller can handle a shortened request
2e5dfc99
DW
1811 */
1812#define REMAP_FILE_DEDUP (1 << 0)
eca3654e 1813#define REMAP_FILE_CAN_SHORTEN (1 << 1)
2e5dfc99
DW
1814
1815/*
1816 * These flags signal that the caller is ok with altering various aspects of
1817 * the behavior of the remap operation. The changes must be made by the
1818 * implementation; the vfs remap helper functions can take advantage of them.
1819 * Flags in this category exist to preserve the quirky behavior of the hoisted
1820 * btrfs clone/dedupe ioctls.
2e5dfc99 1821 */
eca3654e 1822#define REMAP_FILE_ADVISORY (REMAP_FILE_CAN_SHORTEN)
b4caecd4 1823
293bc982
AV
1824struct iov_iter;
1825
1da177e4
LT
1826struct file_operations {
1827 struct module *owner;
1828 loff_t (*llseek) (struct file *, loff_t, int);
1829 ssize_t (*read) (struct file *, char __user *, size_t, loff_t *);
1da177e4 1830 ssize_t (*write) (struct file *, const char __user *, size_t, loff_t *);
293bc982
AV
1831 ssize_t (*read_iter) (struct kiocb *, struct iov_iter *);
1832 ssize_t (*write_iter) (struct kiocb *, struct iov_iter *);
fb7e1600 1833 int (*iopoll)(struct kiocb *kiocb, bool spin);
bb6f619b 1834 int (*iterate) (struct file *, struct dir_context *);
61922694 1835 int (*iterate_shared) (struct file *, struct dir_context *);
a3f8683b 1836 __poll_t (*poll) (struct file *, struct poll_table_struct *);
1da177e4
LT
1837 long (*unlocked_ioctl) (struct file *, unsigned int, unsigned long);
1838 long (*compat_ioctl) (struct file *, unsigned int, unsigned long);
1839 int (*mmap) (struct file *, struct vm_area_struct *);
1c972597 1840 unsigned long mmap_supported_flags;
1da177e4 1841 int (*open) (struct inode *, struct file *);
75e1fcc0 1842 int (*flush) (struct file *, fl_owner_t id);
1da177e4 1843 int (*release) (struct inode *, struct file *);
02c24a82 1844 int (*fsync) (struct file *, loff_t, loff_t, int datasync);
1da177e4
LT
1845 int (*fasync) (int, struct file *, int);
1846 int (*lock) (struct file *, int, struct file_lock *);
1da177e4
LT
1847 ssize_t (*sendpage) (struct file *, struct page *, int, size_t, loff_t *, int);
1848 unsigned long (*get_unmapped_area)(struct file *, unsigned long, unsigned long, unsigned long, unsigned long);
1849 int (*check_flags)(int);
1da177e4 1850 int (*flock) (struct file *, int, struct file_lock *);
cbb7e577
JA
1851 ssize_t (*splice_write)(struct pipe_inode_info *, struct file *, loff_t *, size_t, unsigned int);
1852 ssize_t (*splice_read)(struct file *, loff_t *, struct pipe_inode_info *, size_t, unsigned int);
e6f5c789 1853 int (*setlease)(struct file *, long, struct file_lock **, void **);
2fe17c10
CH
1854 long (*fallocate)(struct file *file, int mode, loff_t offset,
1855 loff_t len);
a3816ab0 1856 void (*show_fdinfo)(struct seq_file *m, struct file *f);
b4caecd4
CH
1857#ifndef CONFIG_MMU
1858 unsigned (*mmap_capabilities)(struct file *);
1859#endif
04b38d60
CH
1860 ssize_t (*copy_file_range)(struct file *, loff_t, struct file *,
1861 loff_t, size_t, unsigned int);
42ec3d4c
DW
1862 loff_t (*remap_file_range)(struct file *file_in, loff_t pos_in,
1863 struct file *file_out, loff_t pos_out,
1864 loff_t len, unsigned int remap_flags);
45cd0faa 1865 int (*fadvise)(struct file *, loff_t, loff_t, int);
3859a271 1866} __randomize_layout;
1da177e4
LT
1867
1868struct inode_operations {
00cd8dd3 1869 struct dentry * (*lookup) (struct inode *,struct dentry *, unsigned int);
fceef393 1870 const char * (*get_link) (struct dentry *, struct inode *, struct delayed_call *);
10556cb2 1871 int (*permission) (struct inode *, int);
4e34e719 1872 struct posix_acl * (*get_acl)(struct inode *, int);
44a7d7a8
NP
1873
1874 int (*readlink) (struct dentry *, char __user *,int);
44a7d7a8 1875
ebfc3b49 1876 int (*create) (struct inode *,struct dentry *, umode_t, bool);
1da177e4
LT
1877 int (*link) (struct dentry *,struct inode *,struct dentry *);
1878 int (*unlink) (struct inode *,struct dentry *);
1879 int (*symlink) (struct inode *,struct dentry *,const char *);
18bb1db3 1880 int (*mkdir) (struct inode *,struct dentry *,umode_t);
1da177e4 1881 int (*rmdir) (struct inode *,struct dentry *);
1a67aafb 1882 int (*mknod) (struct inode *,struct dentry *,umode_t,dev_t);
1da177e4 1883 int (*rename) (struct inode *, struct dentry *,
520c8b16 1884 struct inode *, struct dentry *, unsigned int);
1da177e4 1885 int (*setattr) (struct dentry *, struct iattr *);
a528d35e 1886 int (*getattr) (const struct path *, struct kstat *, u32, unsigned int);
1da177e4 1887 ssize_t (*listxattr) (struct dentry *, char *, size_t);
c4b929b8
MF
1888 int (*fiemap)(struct inode *, struct fiemap_extent_info *, u64 start,
1889 u64 len);
95582b00 1890 int (*update_time)(struct inode *, struct timespec64 *, int);
d9585277 1891 int (*atomic_open)(struct inode *, struct dentry *,
30d90494 1892 struct file *, unsigned open_flag,
44907d79 1893 umode_t create_mode);
60545d0d 1894 int (*tmpfile) (struct inode *, struct dentry *, umode_t);
893d46e4 1895 int (*set_acl)(struct inode *, struct posix_acl *, int);
44a7d7a8 1896} ____cacheline_aligned;
1da177e4 1897
bb7462b6
MS
1898static inline ssize_t call_read_iter(struct file *file, struct kiocb *kio,
1899 struct iov_iter *iter)
1900{
1901 return file->f_op->read_iter(kio, iter);
1902}
1903
1904static inline ssize_t call_write_iter(struct file *file, struct kiocb *kio,
1905 struct iov_iter *iter)
1906{
1907 return file->f_op->write_iter(kio, iter);
1908}
1909
f74ac015
MS
1910static inline int call_mmap(struct file *file, struct vm_area_struct *vma)
1911{
1912 return file->f_op->mmap(file, vma);
1913}
1914
eed4e51f 1915ssize_t rw_copy_check_uvector(int type, const struct iovec __user * uvector,
fcf63409
CY
1916 unsigned long nr_segs, unsigned long fast_segs,
1917 struct iovec *fast_pointer,
ac34ebb3 1918 struct iovec **ret_pointer);
eed4e51f 1919
6fb5032e 1920extern ssize_t __vfs_read(struct file *, char __user *, size_t, loff_t *);
1da177e4
LT
1921extern ssize_t vfs_read(struct file *, char __user *, size_t, loff_t *);
1922extern ssize_t vfs_write(struct file *, const char __user *, size_t, loff_t *);
1923extern ssize_t vfs_readv(struct file *, const struct iovec __user *,
ddef7ed2 1924 unsigned long, loff_t *, rwf_t);
29732938
ZB
1925extern ssize_t vfs_copy_file_range(struct file *, loff_t , struct file *,
1926 loff_t, size_t, unsigned int);
f16acc9d
DC
1927extern ssize_t generic_copy_file_range(struct file *file_in, loff_t pos_in,
1928 struct file *file_out, loff_t pos_out,
1929 size_t len, unsigned int flags);
a83ab01a
DW
1930extern int generic_remap_file_range_prep(struct file *file_in, loff_t pos_in,
1931 struct file *file_out, loff_t pos_out,
42ec3d4c
DW
1932 loff_t *count,
1933 unsigned int remap_flags);
1934extern loff_t do_clone_file_range(struct file *file_in, loff_t pos_in,
1935 struct file *file_out, loff_t pos_out,
452ce659 1936 loff_t len, unsigned int remap_flags);
42ec3d4c
DW
1937extern loff_t vfs_clone_file_range(struct file *file_in, loff_t pos_in,
1938 struct file *file_out, loff_t pos_out,
452ce659 1939 loff_t len, unsigned int remap_flags);
54dbc151
DW
1940extern int vfs_dedupe_file_range(struct file *file,
1941 struct file_dedupe_range *same);
42ec3d4c
DW
1942extern loff_t vfs_dedupe_file_range_one(struct file *src_file, loff_t src_pos,
1943 struct file *dst_file, loff_t dst_pos,
df365836 1944 loff_t len, unsigned int remap_flags);
f1825366 1945
1da177e4 1946
1da177e4
LT
1947struct super_operations {
1948 struct inode *(*alloc_inode)(struct super_block *sb);
1949 void (*destroy_inode)(struct inode *);
fdb0da89 1950 void (*free_inode)(struct inode *);
1da177e4 1951
aa385729 1952 void (*dirty_inode) (struct inode *, int flags);
a9185b41 1953 int (*write_inode) (struct inode *, struct writeback_control *wbc);
45321ac5 1954 int (*drop_inode) (struct inode *);
be7ce416 1955 void (*evict_inode) (struct inode *);
1da177e4 1956 void (*put_super) (struct super_block *);
1da177e4 1957 int (*sync_fs)(struct super_block *sb, int wait);
48b6bca6 1958 int (*freeze_super) (struct super_block *);
c4be0c1d 1959 int (*freeze_fs) (struct super_block *);
48b6bca6 1960 int (*thaw_super) (struct super_block *);
c4be0c1d 1961 int (*unfreeze_fs) (struct super_block *);
726c3342 1962 int (*statfs) (struct dentry *, struct kstatfs *);
1da177e4 1963 int (*remount_fs) (struct super_block *, int *, char *);
42faad99 1964 void (*umount_begin) (struct super_block *);
1da177e4 1965
34c80b1d 1966 int (*show_options)(struct seq_file *, struct dentry *);
d861c630 1967 int (*show_devname)(struct seq_file *, struct dentry *);
a6322de6 1968 int (*show_path)(struct seq_file *, struct dentry *);
64132379 1969 int (*show_stats)(struct seq_file *, struct dentry *);
0e51a720 1970#ifdef CONFIG_QUOTA
1da177e4
LT
1971 ssize_t (*quota_read)(struct super_block *, int, char *, size_t, loff_t);
1972 ssize_t (*quota_write)(struct super_block *, int, const char *, size_t, loff_t);
2d0fa467 1973 struct dquot **(*get_dquots)(struct inode *);
0e51a720 1974#endif
87d8fe1e 1975 int (*bdev_try_to_free_page)(struct super_block*, struct page*, gfp_t);
4101b624
VD
1976 long (*nr_cached_objects)(struct super_block *,
1977 struct shrink_control *);
1978 long (*free_cached_objects)(struct super_block *,
1979 struct shrink_control *);
1da177e4
LT
1980};
1981
bbc1096a
DH
1982/*
1983 * Inode flags - they have no relation to superblock flags now
1984 */
1985#define S_SYNC 1 /* Writes are synced at once */
1986#define S_NOATIME 2 /* Do not update access times */
1987#define S_APPEND 4 /* Append-only file */
1988#define S_IMMUTABLE 8 /* Immutable file */
1989#define S_DEAD 16 /* removed, but still open directory */
1990#define S_NOQUOTA 32 /* Inode is not counted to quota */
1991#define S_DIRSYNC 64 /* Directory modifications are synchronous */
1992#define S_NOCMTIME 128 /* Do not update file c/mtime */
1993#define S_SWAPFILE 256 /* Do not truncate: swapon got its bmaps */
1994#define S_PRIVATE 512 /* Inode is fs-internal */
1995#define S_IMA 1024 /* Inode has an associated IMA struct */
1996#define S_AUTOMOUNT 2048 /* Automount/referral quasi-directory */
1997#define S_NOSEC 4096 /* no suid or xattr security attributes */
6cd176a5 1998#ifdef CONFIG_FS_DAX
fbbbad4b
MW
1999#define S_DAX 8192 /* Direct Access, avoiding the page cache */
2000#else
2001#define S_DAX 0 /* Make all the DAX code disappear */
2002#endif
2ee6a576 2003#define S_ENCRYPTED 16384 /* Encrypted file (using fs/crypto/) */
b886ee3e 2004#define S_CASEFOLD 32768 /* Casefolded file */
5585f2af 2005#define S_VERITY 65536 /* Verity file (using fs/verity/) */
bbc1096a
DH
2006
2007/*
2008 * Note that nosuid etc flags are inode-specific: setting some file-system
2009 * flags just means all the inodes inherit those flags by default. It might be
2010 * possible to override it selectively if you really wanted to with some
2011 * ioctl() that is not currently implemented.
2012 *
e462ec50 2013 * Exception: SB_RDONLY is always applied to the entire file system.
bbc1096a
DH
2014 *
2015 * Unfortunately, it is possible to change a filesystems flags with it mounted
2016 * with files in use. This means that all of the inodes will not have their
2017 * i_flags updated. Hence, i_flags no longer inherit the superblock mount
2018 * flags, so these have to be checked separately. -- rmk@arm.uk.linux.org
2019 */
2020#define __IS_FLG(inode, flg) ((inode)->i_sb->s_flags & (flg))
2021
1751e8a6 2022static inline bool sb_rdonly(const struct super_block *sb) { return sb->s_flags & SB_RDONLY; }
94e92e7a 2023#define IS_RDONLY(inode) sb_rdonly((inode)->i_sb)
e462ec50 2024#define IS_SYNC(inode) (__IS_FLG(inode, SB_SYNCHRONOUS) || \
bbc1096a 2025 ((inode)->i_flags & S_SYNC))
e462ec50 2026#define IS_DIRSYNC(inode) (__IS_FLG(inode, SB_SYNCHRONOUS|SB_DIRSYNC) || \
bbc1096a 2027 ((inode)->i_flags & (S_SYNC|S_DIRSYNC)))
e462ec50
DH
2028#define IS_MANDLOCK(inode) __IS_FLG(inode, SB_MANDLOCK)
2029#define IS_NOATIME(inode) __IS_FLG(inode, SB_RDONLY|SB_NOATIME)
2030#define IS_I_VERSION(inode) __IS_FLG(inode, SB_I_VERSION)
bbc1096a
DH
2031
2032#define IS_NOQUOTA(inode) ((inode)->i_flags & S_NOQUOTA)
2033#define IS_APPEND(inode) ((inode)->i_flags & S_APPEND)
2034#define IS_IMMUTABLE(inode) ((inode)->i_flags & S_IMMUTABLE)
e462ec50 2035#define IS_POSIXACL(inode) __IS_FLG(inode, SB_POSIXACL)
bbc1096a
DH
2036
2037#define IS_DEADDIR(inode) ((inode)->i_flags & S_DEAD)
2038#define IS_NOCMTIME(inode) ((inode)->i_flags & S_NOCMTIME)
2039#define IS_SWAPFILE(inode) ((inode)->i_flags & S_SWAPFILE)
2040#define IS_PRIVATE(inode) ((inode)->i_flags & S_PRIVATE)
2041#define IS_IMA(inode) ((inode)->i_flags & S_IMA)
2042#define IS_AUTOMOUNT(inode) ((inode)->i_flags & S_AUTOMOUNT)
2043#define IS_NOSEC(inode) ((inode)->i_flags & S_NOSEC)
fbbbad4b 2044#define IS_DAX(inode) ((inode)->i_flags & S_DAX)
2ee6a576 2045#define IS_ENCRYPTED(inode) ((inode)->i_flags & S_ENCRYPTED)
b886ee3e 2046#define IS_CASEFOLDED(inode) ((inode)->i_flags & S_CASEFOLD)
5585f2af 2047#define IS_VERITY(inode) ((inode)->i_flags & S_VERITY)
bbc1096a 2048
787fb6bc
MS
2049#define IS_WHITEOUT(inode) (S_ISCHR(inode->i_mode) && \
2050 (inode)->i_rdev == WHITEOUT_DEV)
2051
0bd23d09
EB
2052static inline bool HAS_UNMAPPED_ID(struct inode *inode)
2053{
2054 return !uid_valid(inode->i_uid) || !gid_valid(inode->i_gid);
2055}
2056
c75b1d94
JA
2057static inline enum rw_hint file_write_hint(struct file *file)
2058{
2059 if (file->f_write_hint != WRITE_LIFE_NOT_SET)
2060 return file->f_write_hint;
2061
2062 return file_inode(file)->i_write_hint;
2063}
2064
2065static inline int iocb_flags(struct file *file);
2066
fc28724d
AM
2067static inline u16 ki_hint_validate(enum rw_hint hint)
2068{
2069 typeof(((struct kiocb *)0)->ki_hint) max_hint = -1;
2070
2071 if (hint <= max_hint)
2072 return hint;
2073 return 0;
2074}
2075
c75b1d94
JA
2076static inline void init_sync_kiocb(struct kiocb *kiocb, struct file *filp)
2077{
2078 *kiocb = (struct kiocb) {
2079 .ki_filp = filp,
2080 .ki_flags = iocb_flags(filp),
fc28724d 2081 .ki_hint = ki_hint_validate(file_write_hint(filp)),
20578bdf 2082 .ki_ioprio = get_current_ioprio(),
c75b1d94
JA
2083 };
2084}
2085
5dcdc43e
JX
2086static inline void kiocb_clone(struct kiocb *kiocb, struct kiocb *kiocb_src,
2087 struct file *filp)
2088{
2089 *kiocb = (struct kiocb) {
2090 .ki_filp = filp,
2091 .ki_flags = kiocb_src->ki_flags,
2092 .ki_hint = kiocb_src->ki_hint,
2093 .ki_ioprio = kiocb_src->ki_ioprio,
2094 .ki_pos = kiocb_src->ki_pos,
2095 };
2096}
2097
1c0eeaf5 2098/*
250df6ed 2099 * Inode state bits. Protected by inode->i_lock
1c0eeaf5
JE
2100 *
2101 * Three bits determine the dirty state of the inode, I_DIRTY_SYNC,
2102 * I_DIRTY_DATASYNC and I_DIRTY_PAGES.
2103 *
2104 * Four bits define the lifetime of an inode. Initially, inodes are I_NEW,
2105 * until that flag is cleared. I_WILL_FREE, I_FREEING and I_CLEAR are set at
2106 * various stages of removing an inode.
2107 *
eaff8079 2108 * Two bits are used for locking and completion notification, I_NEW and I_SYNC.
1c0eeaf5 2109 *
e7ca2d41
JE
2110 * I_DIRTY_SYNC Inode is dirty, but doesn't have to be written on
2111 * fdatasync(). i_atime is the usual cause.
000cb48e
JK
2112 * I_DIRTY_DATASYNC Data-related inode changes pending. We keep track of
2113 * these changes separately from I_DIRTY_SYNC so that we
2114 * don't have to write inode on fdatasync() when only
2115 * mtime has changed in it.
1c0eeaf5 2116 * I_DIRTY_PAGES Inode has dirty pages. Inode itself may be clean.
eaff8079
CH
2117 * I_NEW Serves as both a mutex and completion notification.
2118 * New inodes set I_NEW. If two processes both create
2119 * the same inode, one of them will release its inode and
2120 * wait for I_NEW to be released before returning.
2121 * Inodes in I_WILL_FREE, I_FREEING or I_CLEAR state can
2122 * also cause waiting on I_NEW, without I_NEW actually
2123 * being set. find_inode() uses this to prevent returning
2124 * nearly-dead inodes.
1c0eeaf5
JE
2125 * I_WILL_FREE Must be set when calling write_inode_now() if i_count
2126 * is zero. I_FREEING must be set when I_WILL_FREE is
2127 * cleared.
2128 * I_FREEING Set when inode is about to be freed but still has dirty
2129 * pages or buffers attached or the inode itself is still
2130 * dirty.
dbd5768f
JK
2131 * I_CLEAR Added by clear_inode(). In this state the inode is
2132 * clean and can be destroyed. Inode keeps I_FREEING.
1c0eeaf5
JE
2133 *
2134 * Inodes that are I_WILL_FREE, I_FREEING or I_CLEAR are
2135 * prohibited for many purposes. iget() must wait for
2136 * the inode to be completely released, then create it
2137 * anew. Other functions will just ignore such inodes,
eaff8079 2138 * if appropriate. I_NEW is used for waiting.
1c0eeaf5 2139 *
169ebd90
JK
2140 * I_SYNC Writeback of inode is running. The bit is set during
2141 * data writeback, and cleared with a wakeup on the bit
2142 * address once it is done. The bit is also used to pin
2143 * the inode in memory for flusher thread.
1c0eeaf5 2144 *
bd5fe6c5
CH
2145 * I_REFERENCED Marks the inode as recently references on the LRU list.
2146 *
2147 * I_DIO_WAKEUP Never set. Only used as a key for wait_on_bit().
2148 *
682aa8e1
TH
2149 * I_WB_SWITCH Cgroup bdi_writeback switching in progress. Used to
2150 * synchronize competing switching instances and to tell
b93b0163 2151 * wb stat updates to grab the i_pages lock. See
a9519def 2152 * inode_switch_wbs_work_fn() for details.
682aa8e1 2153 *
ad0af710
AG
2154 * I_OVL_INUSE Used by overlayfs to get exclusive ownership on upper
2155 * and work dirs among overlayfs mounts.
2156 *
c2b6d621
AV
2157 * I_CREATING New object's inode in the middle of setting up.
2158 *
1c0eeaf5 2159 * Q: What is the difference between I_WILL_FREE and I_FREEING?
1c0eeaf5 2160 */
9e38d86f
NP
2161#define I_DIRTY_SYNC (1 << 0)
2162#define I_DIRTY_DATASYNC (1 << 1)
2163#define I_DIRTY_PAGES (1 << 2)
eaff8079
CH
2164#define __I_NEW 3
2165#define I_NEW (1 << __I_NEW)
9e38d86f
NP
2166#define I_WILL_FREE (1 << 4)
2167#define I_FREEING (1 << 5)
2168#define I_CLEAR (1 << 6)
eaff8079 2169#define __I_SYNC 7
1c0eeaf5 2170#define I_SYNC (1 << __I_SYNC)
9e38d86f 2171#define I_REFERENCED (1 << 8)
bd5fe6c5 2172#define __I_DIO_WAKEUP 9
ac74d8d6 2173#define I_DIO_WAKEUP (1 << __I_DIO_WAKEUP)
f4e0c30c 2174#define I_LINKABLE (1 << 10)
0ae45f63
TT
2175#define I_DIRTY_TIME (1 << 11)
2176#define __I_DIRTY_TIME_EXPIRED 12
2177#define I_DIRTY_TIME_EXPIRED (1 << __I_DIRTY_TIME_EXPIRED)
682aa8e1 2178#define I_WB_SWITCH (1 << 13)
c2b6d621
AV
2179#define I_OVL_INUSE (1 << 14)
2180#define I_CREATING (1 << 15)
1da177e4 2181
0e11f644
CH
2182#define I_DIRTY_INODE (I_DIRTY_SYNC | I_DIRTY_DATASYNC)
2183#define I_DIRTY (I_DIRTY_INODE | I_DIRTY_PAGES)
0ae45f63 2184#define I_DIRTY_ALL (I_DIRTY | I_DIRTY_TIME)
1da177e4
LT
2185
2186extern void __mark_inode_dirty(struct inode *, int);
2187static inline void mark_inode_dirty(struct inode *inode)
2188{
2189 __mark_inode_dirty(inode, I_DIRTY);
2190}
2191
2192static inline void mark_inode_dirty_sync(struct inode *inode)
2193{
2194 __mark_inode_dirty(inode, I_DIRTY_SYNC);
2195}
2196
7ada4db8
MS
2197extern void inc_nlink(struct inode *inode);
2198extern void drop_nlink(struct inode *inode);
2199extern void clear_nlink(struct inode *inode);
2200extern void set_nlink(struct inode *inode, unsigned int nlink);
d8c76e6f
DH
2201
2202static inline void inode_inc_link_count(struct inode *inode)
2203{
2204 inc_nlink(inode);
66d21730
AD
2205 mark_inode_dirty(inode);
2206}
2207
9a53c3a7
DH
2208static inline void inode_dec_link_count(struct inode *inode)
2209{
2210 drop_nlink(inode);
66d21730
AD
2211 mark_inode_dirty(inode);
2212}
2213
c3b2da31
JB
2214enum file_time_flags {
2215 S_ATIME = 1,
2216 S_MTIME = 2,
2217 S_CTIME = 4,
2218 S_VERSION = 8,
2219};
2220
c6718543 2221extern bool atime_needs_update(const struct path *, struct inode *);
badcf2b7 2222extern void touch_atime(const struct path *);
1da177e4
LT
2223static inline void file_accessed(struct file *file)
2224{
2225 if (!(file->f_flags & O_NOATIME))
68ac1234 2226 touch_atime(&file->f_path);
1da177e4
LT
2227}
2228
e38f7f53
AG
2229extern int file_modified(struct file *file);
2230
1da177e4 2231int sync_inode(struct inode *inode, struct writeback_control *wbc);
c3765016 2232int sync_inode_metadata(struct inode *inode, int wait);
1da177e4 2233
1da177e4
LT
2234struct file_system_type {
2235 const char *name;
2236 int fs_flags;
bbc1096a
DH
2237#define FS_REQUIRES_DEV 1
2238#define FS_BINARY_MOUNTDATA 2
2239#define FS_HAS_SUBTYPE 4
0c55cfc4 2240#define FS_USERNS_MOUNT 8 /* Can be mounted by userns root */
0b3b094a 2241#define FS_DISALLOW_NOTIFY_PERM 16 /* Disable fanotify permission events */
bbc1096a 2242#define FS_RENAME_DOES_D_MOVE 32768 /* FS will handle d_move() during rename() internally. */
f3a09c92 2243 int (*init_fs_context)(struct fs_context *);
d7167b14 2244 const struct fs_parameter_spec *parameters;
c96e41e9
AV
2245 struct dentry *(*mount) (struct file_system_type *, int,
2246 const char *, void *);
1da177e4
LT
2247 void (*kill_sb) (struct super_block *);
2248 struct module *owner;
2249 struct file_system_type * next;
a5166169 2250 struct hlist_head fs_supers;
d475fd42 2251
cf516249 2252 struct lock_class_key s_lock_key;
897c6ff9 2253 struct lock_class_key s_umount_key;
51ee049e 2254 struct lock_class_key s_vfs_rename_key;
5accdf82 2255 struct lock_class_key s_writers_key[SB_FREEZE_LEVELS];
d475fd42
PZ
2256
2257 struct lock_class_key i_lock_key;
2258 struct lock_class_key i_mutex_key;
14358e6d 2259 struct lock_class_key i_mutex_dir_key;
1da177e4
LT
2260};
2261
7f78e035
EB
2262#define MODULE_ALIAS_FS(NAME) MODULE_ALIAS("fs-" NAME)
2263
f175f307 2264#ifdef CONFIG_BLOCK
152a0836
AV
2265extern struct dentry *mount_bdev(struct file_system_type *fs_type,
2266 int flags, const char *dev_name, void *data,
2267 int (*fill_super)(struct super_block *, void *, int));
f175f307
AV
2268#else
2269static inline struct dentry *mount_bdev(struct file_system_type *fs_type,
2270 int flags, const char *dev_name, void *data,
2271 int (*fill_super)(struct super_block *, void *, int))
2272{
2273 return ERR_PTR(-ENODEV);
2274}
2275#endif
fc14f2fe
AV
2276extern struct dentry *mount_single(struct file_system_type *fs_type,
2277 int flags, void *data,
2278 int (*fill_super)(struct super_block *, void *, int));
3c26ff6e
AV
2279extern struct dentry *mount_nodev(struct file_system_type *fs_type,
2280 int flags, void *data,
2281 int (*fill_super)(struct super_block *, void *, int));
ea441d11 2282extern struct dentry *mount_subtree(struct vfsmount *mnt, const char *path);
1da177e4 2283void generic_shutdown_super(struct super_block *sb);
f175f307 2284#ifdef CONFIG_BLOCK
1da177e4 2285void kill_block_super(struct super_block *sb);
f175f307
AV
2286#else
2287static inline void kill_block_super(struct super_block *sb)
2288{
2289 BUG();
2290}
2291#endif
1da177e4
LT
2292void kill_anon_super(struct super_block *sb);
2293void kill_litter_super(struct super_block *sb);
2294void deactivate_super(struct super_block *sb);
74dbbdd7 2295void deactivate_locked_super(struct super_block *sb);
1da177e4 2296int set_anon_super(struct super_block *s, void *data);
cb50b348 2297int set_anon_super_fc(struct super_block *s, struct fs_context *fc);
0ee5dc67
AV
2298int get_anon_bdev(dev_t *);
2299void free_anon_bdev(dev_t);
cb50b348
AV
2300struct super_block *sget_fc(struct fs_context *fc,
2301 int (*test)(struct super_block *, struct fs_context *),
2302 int (*set)(struct super_block *, struct fs_context *));
1da177e4
LT
2303struct super_block *sget(struct file_system_type *type,
2304 int (*test)(struct super_block *,void *),
2305 int (*set)(struct super_block *,void *),
9249e17f 2306 int flags, void *data);
1da177e4
LT
2307
2308/* Alas, no aliases. Too much hassle with bringing module.h everywhere */
2309#define fops_get(fops) \
2310 (((fops) && try_module_get((fops)->owner) ? (fops) : NULL))
2311#define fops_put(fops) \
2312 do { if (fops) module_put((fops)->owner); } while(0)
e84f9e57
AV
2313/*
2314 * This one is to be used *ONLY* from ->open() instances.
2315 * fops must be non-NULL, pinned down *and* module dependencies
2316 * should be sufficient to pin the caller down as well.
2317 */
2318#define replace_fops(f, fops) \
2319 do { \
2320 struct file *__file = (f); \
2321 fops_put(__file->f_op); \
2322 BUG_ON(!(__file->f_op = (fops))); \
2323 } while(0)
1da177e4
LT
2324
2325extern int register_filesystem(struct file_system_type *);
2326extern int unregister_filesystem(struct file_system_type *);
d911b458 2327extern struct vfsmount *kern_mount(struct file_system_type *);
423e0ab0 2328extern void kern_unmount(struct vfsmount *mnt);
1da177e4
LT
2329extern int may_umount_tree(struct vfsmount *);
2330extern int may_umount(struct vfsmount *);
5e6123f3
SL
2331extern long do_mount(const char *, const char __user *,
2332 const char *, unsigned long, void *);
ca71cf71 2333extern struct vfsmount *collect_mounts(const struct path *);
8aec0809 2334extern void drop_collected_mounts(struct vfsmount *);
1f707137
AV
2335extern int iterate_mounts(int (*)(struct vfsmount *, void *), void *,
2336 struct vfsmount *);
f0bb5aaf 2337extern int vfs_statfs(const struct path *, struct kstatfs *);
c8b91acc
AV
2338extern int user_statfs(const char __user *, struct kstatfs *);
2339extern int fd_statfs(int, struct kstatfs *);
18e9e510
JB
2340extern int freeze_super(struct super_block *super);
2341extern int thaw_super(struct super_block *super);
02125a82 2342extern bool our_mnt(struct vfsmount *mnt);
fca39346
JK
2343extern __printf(2, 3)
2344int super_setup_bdi_name(struct super_block *sb, char *fmt, ...);
2345extern int super_setup_bdi(struct super_block *sb);
1da177e4 2346
ce3b0f8d
AV
2347extern int current_umask(void);
2348
5a14696c
BF
2349extern void ihold(struct inode * inode);
2350extern void iput(struct inode *);
95582b00 2351extern int generic_update_time(struct inode *, struct timespec64 *, int);
5a14696c 2352
f87fd4c2 2353/* /sys/fs */
00d26666 2354extern struct kobject *fs_kobj;
f87fd4c2 2355
09cbfeaf 2356#define MAX_RW_COUNT (INT_MAX & PAGE_MASK)
bfcd17a6 2357
9e8925b6 2358#ifdef CONFIG_MANDATORY_FILE_LOCKING
d7a06983 2359extern int locks_mandatory_locked(struct file *);
acc15575 2360extern int locks_mandatory_area(struct inode *, struct file *, loff_t, loff_t, unsigned char);
1da177e4
LT
2361
2362/*
2363 * Candidates for mandatory locking have the setgid bit set
2364 * but no group execute bit - an otherwise meaningless combination.
2365 */
a16877ca
PE
2366
2367static inline int __mandatory_lock(struct inode *ino)
2368{
2369 return (ino->i_mode & (S_ISGID | S_IXGRP)) == S_ISGID;
2370}
2371
2372/*
e462ec50 2373 * ... and these candidates should be on SB_MANDLOCK mounted fs,
a16877ca
PE
2374 * otherwise these will be advisory locks
2375 */
2376
2377static inline int mandatory_lock(struct inode *ino)
2378{
2379 return IS_MANDLOCK(ino) && __mandatory_lock(ino);
2380}
1da177e4 2381
d7a06983 2382static inline int locks_verify_locked(struct file *file)
1da177e4 2383{
c568d683 2384 if (mandatory_lock(locks_inode(file)))
d7a06983 2385 return locks_mandatory_locked(file);
1da177e4
LT
2386 return 0;
2387}
2388
1da177e4 2389static inline int locks_verify_truncate(struct inode *inode,
acc15575 2390 struct file *f,
1da177e4
LT
2391 loff_t size)
2392{
acc15575
CH
2393 if (!inode->i_flctx || !mandatory_lock(inode))
2394 return 0;
2395
2396 if (size < inode->i_size) {
2397 return locks_mandatory_area(inode, f, size, inode->i_size - 1,
2398 F_WRLCK);
2399 } else {
2400 return locks_mandatory_area(inode, f, inode->i_size, size - 1,
2401 F_WRLCK);
2402 }
1da177e4
LT
2403}
2404
9e8925b6
JL
2405#else /* !CONFIG_MANDATORY_FILE_LOCKING */
2406
2407static inline int locks_mandatory_locked(struct file *file)
2408{
2409 return 0;
2410}
2411
fce205e9
LT
2412static inline int locks_mandatory_area(struct inode *inode, struct file *filp,
2413 loff_t start, loff_t end, unsigned char type)
9e8925b6
JL
2414{
2415 return 0;
2416}
2417
2418static inline int __mandatory_lock(struct inode *inode)
2419{
2420 return 0;
2421}
2422
2423static inline int mandatory_lock(struct inode *inode)
2424{
2425 return 0;
2426}
2427
2428static inline int locks_verify_locked(struct file *file)
2429{
2430 return 0;
2431}
2432
2433static inline int locks_verify_truncate(struct inode *inode, struct file *filp,
2434 size_t size)
2435{
2436 return 0;
2437}
2438
2439#endif /* CONFIG_MANDATORY_FILE_LOCKING */
2440
2441
2442#ifdef CONFIG_FILE_LOCKING
1da177e4
LT
2443static inline int break_lease(struct inode *inode, unsigned int mode)
2444{
24cbe784
JL
2445 /*
2446 * Since this check is lockless, we must ensure that any refcounts
8116bf4c
JL
2447 * taken are done before checking i_flctx->flc_lease. Otherwise, we
2448 * could end up racing with tasks trying to set a new lease on this
2449 * file.
24cbe784
JL
2450 */
2451 smp_mb();
8634b51f 2452 if (inode->i_flctx && !list_empty_careful(&inode->i_flctx->flc_lease))
df4e8d2c 2453 return __break_lease(inode, mode, FL_LEASE);
1da177e4
LT
2454 return 0;
2455}
df4e8d2c
BF
2456
2457static inline int break_deleg(struct inode *inode, unsigned int mode)
2458{
962bd40b
JL
2459 /*
2460 * Since this check is lockless, we must ensure that any refcounts
8116bf4c
JL
2461 * taken are done before checking i_flctx->flc_lease. Otherwise, we
2462 * could end up racing with tasks trying to set a new lease on this
2463 * file.
962bd40b
JL
2464 */
2465 smp_mb();
8634b51f 2466 if (inode->i_flctx && !list_empty_careful(&inode->i_flctx->flc_lease))
df4e8d2c 2467 return __break_lease(inode, mode, FL_DELEG);
1da177e4
LT
2468 return 0;
2469}
df4e8d2c 2470
5a14696c
BF
2471static inline int try_break_deleg(struct inode *inode, struct inode **delegated_inode)
2472{
2473 int ret;
2474
2475 ret = break_deleg(inode, O_WRONLY|O_NONBLOCK);
2476 if (ret == -EWOULDBLOCK && delegated_inode) {
2477 *delegated_inode = inode;
2478 ihold(inode);
2479 }
2480 return ret;
2481}
2482
2483static inline int break_deleg_wait(struct inode **delegated_inode)
2484{
2485 int ret;
2486
2487 ret = break_deleg(*delegated_inode, O_WRONLY);
2488 iput(*delegated_inode);
2489 *delegated_inode = NULL;
2490 return ret;
2491}
2492
11afe9f7
CH
2493static inline int break_layout(struct inode *inode, bool wait)
2494{
2495 smp_mb();
2496 if (inode->i_flctx && !list_empty_careful(&inode->i_flctx->flc_lease))
2497 return __break_lease(inode,
2498 wait ? O_WRONLY : O_WRONLY | O_NONBLOCK,
2499 FL_LAYOUT);
2500 return 0;
2501}
2502
bfcd17a6 2503#else /* !CONFIG_FILE_LOCKING */
af5df566
SW
2504static inline int break_lease(struct inode *inode, unsigned int mode)
2505{
2506 return 0;
2507}
2508
df4e8d2c
BF
2509static inline int break_deleg(struct inode *inode, unsigned int mode)
2510{
2511 return 0;
2512}
5a14696c
BF
2513
2514static inline int try_break_deleg(struct inode *inode, struct inode **delegated_inode)
2515{
2516 return 0;
2517}
2518
2519static inline int break_deleg_wait(struct inode **delegated_inode)
2520{
2521 BUG();
2522 return 0;
2523}
2524
11afe9f7
CH
2525static inline int break_layout(struct inode *inode, bool wait)
2526{
2527 return 0;
2528}
2529
bfcd17a6 2530#endif /* CONFIG_FILE_LOCKING */
1da177e4
LT
2531
2532/* fs/open.c */
adb5c247 2533struct audit_names;
91a27b2a 2534struct filename {
adb5c247
JL
2535 const char *name; /* pointer to actual string */
2536 const __user char *uptr; /* original userland pointer */
55422d0b 2537 int refcnt;
1c949843 2538 struct audit_names *aname;
fd2f7cb5 2539 const char iname[];
91a27b2a 2540};
f1fffbd4 2541static_assert(offsetof(struct filename, iname) % sizeof(long) == 0);
1da177e4 2542
7df818b2 2543extern long vfs_truncate(const struct path *, loff_t);
4a30131e
N
2544extern int do_truncate(struct dentry *, loff_t start, unsigned int time_attrs,
2545 struct file *filp);
72c72bdf 2546extern int vfs_fallocate(struct file *file, int mode, loff_t offset,
3e63cbb1 2547 loff_t len);
8e8a1407 2548extern long do_sys_open(int dfd, const char __user *filename, int flags,
a218d0fd 2549 umode_t mode);
669abf4e 2550extern struct file *file_open_name(struct filename *, int, umode_t);
a218d0fd 2551extern struct file *filp_open(const char *, int, umode_t);
73d049a4 2552extern struct file *file_open_root(struct dentry *, struct vfsmount *,
378c6520 2553 const char *, int, umode_t);
765927b2 2554extern struct file * dentry_open(const struct path *, int, const struct cred *);
2abc77af
AV
2555extern struct file * open_with_fake_path(const struct path *, int,
2556 struct inode*, const struct cred *);
19f391eb
AV
2557static inline struct file *file_clone_open(struct file *file)
2558{
2559 return dentry_open(&file->f_path, file->f_flags, file->f_cred);
2560}
1da177e4 2561extern int filp_close(struct file *, fl_owner_t id);
91a27b2a 2562
51f39a1f 2563extern struct filename *getname_flags(const char __user *, int, int *);
91a27b2a 2564extern struct filename *getname(const char __user *);
c4ad8f98 2565extern struct filename *getname_kernel(const char *);
55422d0b 2566extern void putname(struct filename *name);
91a27b2a 2567
30d90494 2568extern int finish_open(struct file *file, struct dentry *dentry,
be12af3e 2569 int (*open)(struct inode *, struct file *));
e45198a6 2570extern int finish_no_open(struct file *file, struct dentry *dentry);
1da177e4
LT
2571
2572/* fs/dcache.c */
2573extern void __init vfs_caches_init_early(void);
4248b0da 2574extern void __init vfs_caches_init(void);
1da177e4 2575
b86c089b
CL
2576extern struct kmem_cache *names_cachep;
2577
a608ca21 2578#define __getname() kmem_cache_alloc(names_cachep, GFP_KERNEL)
3446a8aa 2579#define __putname(name) kmem_cache_free(names_cachep, (void *)(name))
1da177e4 2580
9361401e 2581#ifdef CONFIG_BLOCK
1da177e4 2582extern int register_blkdev(unsigned int, const char *);
f4480240 2583extern void unregister_blkdev(unsigned int, const char *);
f44f1ab5 2584extern void bdev_unhash_inode(dev_t dev);
1da177e4 2585extern struct block_device *bdget(dev_t);
dddac6a7 2586extern struct block_device *bdgrab(struct block_device *bdev);
1da177e4
LT
2587extern void bd_set_size(struct block_device *, loff_t size);
2588extern void bd_forget(struct inode *inode);
2589extern void bdput(struct block_device *);
585d3bc0 2590extern void invalidate_bdev(struct block_device *);
5c0d6b60 2591extern void iterate_bdevs(void (*)(struct block_device *, void *), void *);
585d3bc0 2592extern int sync_blockdev(struct block_device *bdev);
ff01bb48 2593extern void kill_bdev(struct block_device *);
585d3bc0 2594extern struct super_block *freeze_bdev(struct block_device *);
c2d75438 2595extern void emergency_thaw_all(void);
08fdc8a0 2596extern void emergency_thaw_bdev(struct super_block *sb);
585d3bc0
NP
2597extern int thaw_bdev(struct block_device *bdev, struct super_block *sb);
2598extern int fsync_bdev(struct block_device *);
a212b105
TH
2599
2600extern struct super_block *blockdev_superblock;
2601
2602static inline bool sb_is_blkdev_sb(struct super_block *sb)
2603{
2604 return sb == blockdev_superblock;
2605}
9361401e
DH
2606#else
2607static inline void bd_forget(struct inode *inode) {}
47e4491b 2608static inline int sync_blockdev(struct block_device *bdev) { return 0; }
ff01bb48 2609static inline void kill_bdev(struct block_device *bdev) {}
47e4491b
AV
2610static inline void invalidate_bdev(struct block_device *bdev) {}
2611
2612static inline struct super_block *freeze_bdev(struct block_device *sb)
2613{
2614 return NULL;
2615}
2616
2617static inline int thaw_bdev(struct block_device *bdev, struct super_block *sb)
2618{
2619 return 0;
2620}
5c0d6b60 2621
08fdc8a0
MG
2622static inline int emergency_thaw_bdev(struct super_block *sb)
2623{
2624 return 0;
2625}
2626
5c0d6b60
JK
2627static inline void iterate_bdevs(void (*f)(struct block_device *, void *), void *arg)
2628{
2629}
a8855990 2630
3e004eea 2631static inline bool sb_is_blkdev_sb(struct super_block *sb)
a8855990 2632{
3e004eea 2633 return false;
a8855990 2634}
9361401e 2635#endif
60b0680f 2636extern int sync_filesystem(struct super_block *);
9361401e 2637extern const struct file_operations def_blk_fops;
4b6f5d20 2638extern const struct file_operations def_chr_fops;
9361401e 2639#ifdef CONFIG_BLOCK
1da177e4 2640extern int ioctl_by_bdev(struct block_device *, unsigned, unsigned long);
56b26add 2641extern int blkdev_ioctl(struct block_device *, fmode_t, unsigned, unsigned long);
1da177e4 2642extern long compat_blkdev_ioctl(struct file *, unsigned, unsigned long);
e525fd89 2643extern int blkdev_get(struct block_device *bdev, fmode_t mode, void *holder);
d4d77629
TH
2644extern struct block_device *blkdev_get_by_path(const char *path, fmode_t mode,
2645 void *holder);
2646extern struct block_device *blkdev_get_by_dev(dev_t dev, fmode_t mode,
2647 void *holder);
89e524c0
JK
2648extern struct block_device *bd_start_claiming(struct block_device *bdev,
2649 void *holder);
2650extern void bd_finish_claiming(struct block_device *bdev,
2651 struct block_device *whole, void *holder);
2652extern void bd_abort_claiming(struct block_device *bdev,
2653 struct block_device *whole, void *holder);
4385bab1 2654extern void blkdev_put(struct block_device *bdev, fmode_t mode);
be324177 2655
641dc636 2656#ifdef CONFIG_SYSFS
e09b457b 2657extern int bd_link_disk_holder(struct block_device *bdev, struct gendisk *disk);
49731baa
TH
2658extern void bd_unlink_disk_holder(struct block_device *bdev,
2659 struct gendisk *disk);
641dc636 2660#else
e09b457b
TH
2661static inline int bd_link_disk_holder(struct block_device *bdev,
2662 struct gendisk *disk)
2663{
2664 return 0;
2665}
49731baa
TH
2666static inline void bd_unlink_disk_holder(struct block_device *bdev,
2667 struct gendisk *disk)
2668{
2669}
641dc636 2670#endif
9361401e 2671#endif
1da177e4
LT
2672
2673/* fs/char_dev.c */
8a932f73 2674#define CHRDEV_MAJOR_MAX 512
49db08c3
LW
2675/* Marks the bottom of the first segment of free char majors */
2676#define CHRDEV_MAJOR_DYN_END 234
a5d31a3f
LG
2677/* Marks the top and bottom of the second segment of free char majors */
2678#define CHRDEV_MAJOR_DYN_EXT_START 511
2679#define CHRDEV_MAJOR_DYN_EXT_END 384
2680
1da177e4
LT
2681extern int alloc_chrdev_region(dev_t *, unsigned, unsigned, const char *);
2682extern int register_chrdev_region(dev_t, unsigned, const char *);
1905b1bf
TH
2683extern int __register_chrdev(unsigned int major, unsigned int baseminor,
2684 unsigned int count, const char *name,
2685 const struct file_operations *fops);
2686extern void __unregister_chrdev(unsigned int major, unsigned int baseminor,
2687 unsigned int count, const char *name);
1da177e4 2688extern void unregister_chrdev_region(dev_t, unsigned);
68eef3b4 2689extern void chrdev_show(struct seq_file *,off_t);
1da177e4 2690
1905b1bf
TH
2691static inline int register_chrdev(unsigned int major, const char *name,
2692 const struct file_operations *fops)
2693{
2694 return __register_chrdev(major, 0, 256, name, fops);
2695}
2696
2697static inline void unregister_chrdev(unsigned int major, const char *name)
2698{
2699 __unregister_chrdev(major, 0, 256, name);
2700}
2701
1da177e4
LT
2702/* fs/block_dev.c */
2703#define BDEVNAME_SIZE 32 /* Largest string for a blockdev identifier */
1f014290 2704#define BDEVT_SIZE 10 /* Largest string for MAJ:MIN for blkdev */
9361401e
DH
2705
2706#ifdef CONFIG_BLOCK
133d55cd 2707#define BLKDEV_MAJOR_MAX 512
1da177e4
LT
2708extern const char *bdevname(struct block_device *bdev, char *buffer);
2709extern struct block_device *lookup_bdev(const char *);
68eef3b4 2710extern void blkdev_show(struct seq_file *,off_t);
30c40d2c 2711
9361401e 2712#else
133d55cd 2713#define BLKDEV_MAJOR_MAX 0
9361401e 2714#endif
1da177e4
LT
2715
2716extern void init_special_inode(struct inode *, umode_t, dev_t);
2717
2718/* Invalid inode operations -- fs/bad_inode.c */
2719extern void make_bad_inode(struct inode *);
0e3ef1fe 2720extern bool is_bad_inode(struct inode *);
1da177e4 2721
c11f0c0b 2722#ifdef CONFIG_BLOCK
0c002c2f 2723extern int revalidate_disk(struct gendisk *);
1da177e4 2724extern int check_disk_change(struct block_device *);
93b270f7 2725extern int __invalidate_device(struct block_device *, bool);
1da177e4 2726extern int invalidate_partition(struct gendisk *, int);
9361401e 2727#endif
1da177e4
LT
2728unsigned long invalidate_mapping_pages(struct address_space *mapping,
2729 pgoff_t start, pgoff_t end);
54bc4855 2730
1da177e4
LT
2731static inline void invalidate_remote_inode(struct inode *inode)
2732{
2733 if (S_ISREG(inode->i_mode) || S_ISDIR(inode->i_mode) ||
2734 S_ISLNK(inode->i_mode))
fc0ecff6 2735 invalidate_mapping_pages(inode->i_mapping, 0, -1);
1da177e4
LT
2736}
2737extern int invalidate_inode_pages2(struct address_space *mapping);
2738extern int invalidate_inode_pages2_range(struct address_space *mapping,
2739 pgoff_t start, pgoff_t end);
2740extern int write_inode_now(struct inode *, int);
2741extern int filemap_fdatawrite(struct address_space *);
2742extern int filemap_flush(struct address_space *);
76341cab 2743extern int filemap_fdatawait_keep_errors(struct address_space *mapping);
d3bccb6f
JK
2744extern int filemap_fdatawait_range(struct address_space *, loff_t lstart,
2745 loff_t lend);
aa0bfcd9
RZ
2746extern int filemap_fdatawait_range_keep_errors(struct address_space *mapping,
2747 loff_t start_byte, loff_t end_byte);
ffb959bb
JL
2748
2749static inline int filemap_fdatawait(struct address_space *mapping)
2750{
2751 return filemap_fdatawait_range(mapping, 0, LLONG_MAX);
2752}
2753
7fc9e472
GR
2754extern bool filemap_range_has_page(struct address_space *, loff_t lstart,
2755 loff_t lend);
1da177e4
LT
2756extern int filemap_write_and_wait_range(struct address_space *mapping,
2757 loff_t lstart, loff_t lend);
ebcf28e1
AM
2758extern int __filemap_fdatawrite_range(struct address_space *mapping,
2759 loff_t start, loff_t end, int sync_mode);
f4c0a0fd
JK
2760extern int filemap_fdatawrite_range(struct address_space *mapping,
2761 loff_t start, loff_t end);
d72d9e2a 2762extern int filemap_check_errors(struct address_space *mapping);
5660e13d 2763extern void __filemap_set_wb_err(struct address_space *mapping, int err);
a823e458 2764
ddf8f376
IW
2765static inline int filemap_write_and_wait(struct address_space *mapping)
2766{
2767 return filemap_write_and_wait_range(mapping, 0, LLONG_MAX);
2768}
2769
a823e458
JL
2770extern int __must_check file_fdatawait_range(struct file *file, loff_t lstart,
2771 loff_t lend);
5660e13d
JL
2772extern int __must_check file_check_and_advance_wb_err(struct file *file);
2773extern int __must_check file_write_and_wait_range(struct file *file,
2774 loff_t start, loff_t end);
2775
a823e458
JL
2776static inline int file_write_and_wait(struct file *file)
2777{
2778 return file_write_and_wait_range(file, 0, LLONG_MAX);
2779}
2780
5660e13d
JL
2781/**
2782 * filemap_set_wb_err - set a writeback error on an address_space
2783 * @mapping: mapping in which to set writeback error
2784 * @err: error to be set in mapping
2785 *
2786 * When writeback fails in some way, we must record that error so that
2787 * userspace can be informed when fsync and the like are called. We endeavor
2788 * to report errors on any file that was open at the time of the error. Some
2789 * internal callers also need to know when writeback errors have occurred.
2790 *
2791 * When a writeback error occurs, most filesystems will want to call
2792 * filemap_set_wb_err to record the error in the mapping so that it will be
2793 * automatically reported whenever fsync is called on the file.
5660e13d
JL
2794 */
2795static inline void filemap_set_wb_err(struct address_space *mapping, int err)
2796{
2797 /* Fastpath for common case of no error */
2798 if (unlikely(err))
2799 __filemap_set_wb_err(mapping, err);
2800}
2801
2802/**
2803 * filemap_check_wb_error - has an error occurred since the mark was sampled?
2804 * @mapping: mapping to check for writeback errors
2805 * @since: previously-sampled errseq_t
2806 *
2807 * Grab the errseq_t value from the mapping, and see if it has changed "since"
2808 * the given value was sampled.
2809 *
2810 * If it has then report the latest error set, otherwise return 0.
2811 */
2812static inline int filemap_check_wb_err(struct address_space *mapping,
2813 errseq_t since)
2814{
2815 return errseq_check(&mapping->wb_err, since);
2816}
2817
2818/**
2819 * filemap_sample_wb_err - sample the current errseq_t to test for later errors
2820 * @mapping: mapping to be sampled
2821 *
2822 * Writeback errors are always reported relative to a particular sample point
2823 * in the past. This function provides those sample points.
2824 */
2825static inline errseq_t filemap_sample_wb_err(struct address_space *mapping)
2826{
2827 return errseq_sample(&mapping->wb_err);
2828}
2829
09d91cda
SL
2830static inline int filemap_nr_thps(struct address_space *mapping)
2831{
2832#ifdef CONFIG_READ_ONLY_THP_FOR_FS
2833 return atomic_read(&mapping->nr_thps);
2834#else
2835 return 0;
2836#endif
2837}
2838
2839static inline void filemap_nr_thps_inc(struct address_space *mapping)
2840{
2841#ifdef CONFIG_READ_ONLY_THP_FOR_FS
2842 atomic_inc(&mapping->nr_thps);
2843#else
2844 WARN_ON_ONCE(1);
2845#endif
2846}
2847
2848static inline void filemap_nr_thps_dec(struct address_space *mapping)
2849{
2850#ifdef CONFIG_READ_ONLY_THP_FOR_FS
2851 atomic_dec(&mapping->nr_thps);
2852#else
2853 WARN_ON_ONCE(1);
2854#endif
2855}
2856
8018ab05
CH
2857extern int vfs_fsync_range(struct file *file, loff_t start, loff_t end,
2858 int datasync);
2859extern int vfs_fsync(struct file *file, int datasync);
e2592217 2860
22f96b38
JA
2861extern int sync_file_range(struct file *file, loff_t offset, loff_t nbytes,
2862 unsigned int flags);
2863
e2592217
CH
2864/*
2865 * Sync the bytes written if this was a synchronous write. Expect ki_pos
2866 * to already be updated for the write, and will return either the amount
2867 * of bytes passed in, or an error if syncing the file failed.
2868 */
2869static inline ssize_t generic_write_sync(struct kiocb *iocb, ssize_t count)
2870{
2871 if (iocb->ki_flags & IOCB_DSYNC) {
2872 int ret = vfs_fsync_range(iocb->ki_filp,
2873 iocb->ki_pos - count, iocb->ki_pos - 1,
2874 (iocb->ki_flags & IOCB_SYNC) ? 0 : 1);
2875 if (ret)
2876 return ret;
2877 }
2878
2879 return count;
d311d79d 2880}
e2592217 2881
1da177e4
LT
2882extern void emergency_sync(void);
2883extern void emergency_remount(void);
30460e1e 2884
9361401e 2885#ifdef CONFIG_BLOCK
30460e1e
CM
2886extern int bmap(struct inode *inode, sector_t *block);
2887#else
2888static inline int bmap(struct inode *inode, sector_t *block)
2889{
2890 return -EINVAL;
2891}
9361401e 2892#endif
30460e1e 2893
27ac0ffe 2894extern int notify_change(struct dentry *, struct iattr *, struct inode **);
f419a2e3 2895extern int inode_permission(struct inode *, int);
2830ba7f 2896extern int generic_permission(struct inode *, int);
cbdf35bc 2897extern int __check_sticky(struct inode *dir, struct inode *inode);
1da177e4 2898
f696a365
MS
2899static inline bool execute_ok(struct inode *inode)
2900{
2901 return (inode->i_mode & S_IXUGO) || S_ISDIR(inode->i_mode);
2902}
2903
8d71db4f
AV
2904static inline void file_start_write(struct file *file)
2905{
2906 if (!S_ISREG(file_inode(file)->i_mode))
2907 return;
2908 __sb_start_write(file_inode(file)->i_sb, SB_FREEZE_WRITE, true);
2909}
2910
5ae98f15
JK
2911static inline bool file_start_write_trylock(struct file *file)
2912{
2913 if (!S_ISREG(file_inode(file)->i_mode))
2914 return true;
2915 return __sb_start_write(file_inode(file)->i_sb, SB_FREEZE_WRITE, false);
2916}
2917
8d71db4f
AV
2918static inline void file_end_write(struct file *file)
2919{
2920 if (!S_ISREG(file_inode(file)->i_mode))
2921 return;
2922 __sb_end_write(file_inode(file)->i_sb, SB_FREEZE_WRITE);
2923}
bfe219d3 2924
07b8ce1e
AV
2925/*
2926 * get_write_access() gets write permission for a file.
2927 * put_write_access() releases this write permission.
2928 * This is used for regular files.
2929 * We cannot support write (and maybe mmap read-write shared) accesses and
2930 * MAP_DENYWRITE mmappings simultaneously. The i_writecount field of an inode
2931 * can have the following values:
2932 * 0: no writers, no VM_DENYWRITE mappings
2933 * < 0: (-i_writecount) vm_area_structs with VM_DENYWRITE set exist
2934 * > 0: (i_writecount) users are writing to the file.
2935 *
2936 * Normally we operate on that counter with atomic_{inc,dec} and it's safe
2937 * except for the cases where we don't hold i_writecount yet. Then we need to
2938 * use {get,deny}_write_access() - these functions check the sign and refuse
2939 * to do the change if sign is wrong.
2940 */
2941static inline int get_write_access(struct inode *inode)
2942{
2943 return atomic_inc_unless_negative(&inode->i_writecount) ? 0 : -ETXTBSY;
2944}
2945static inline int deny_write_access(struct file *file)
2946{
496ad9aa 2947 struct inode *inode = file_inode(file);
07b8ce1e
AV
2948 return atomic_dec_unless_positive(&inode->i_writecount) ? 0 : -ETXTBSY;
2949}
1da177e4
LT
2950static inline void put_write_access(struct inode * inode)
2951{
2952 atomic_dec(&inode->i_writecount);
2953}
2954static inline void allow_write_access(struct file *file)
2955{
2956 if (file)
496ad9aa 2957 atomic_inc(&file_inode(file)->i_writecount);
1da177e4 2958}
f1fe29b4
DH
2959static inline bool inode_is_open_for_write(const struct inode *inode)
2960{
2961 return atomic_read(&inode->i_writecount) > 0;
2962}
2963
387e3746 2964#if defined(CONFIG_IMA) || defined(CONFIG_FILE_LOCKING)
a5c96ebf
MZ
2965static inline void i_readcount_dec(struct inode *inode)
2966{
2967 BUG_ON(!atomic_read(&inode->i_readcount));
2968 atomic_dec(&inode->i_readcount);
2969}
2970static inline void i_readcount_inc(struct inode *inode)
2971{
2972 atomic_inc(&inode->i_readcount);
2973}
2974#else
2975static inline void i_readcount_dec(struct inode *inode)
2976{
2977 return;
2978}
2979static inline void i_readcount_inc(struct inode *inode)
2980{
2981 return;
2982}
2983#endif
ed8cae8b 2984extern int do_pipe_flags(int *, int);
1da177e4 2985
1284ab5b
MZ
2986#define __kernel_read_file_id(id) \
2987 id(UNKNOWN, unknown) \
2988 id(FIRMWARE, firmware) \
a098ecd2 2989 id(FIRMWARE_PREALLOC_BUFFER, firmware) \
e4c2c0ff 2990 id(FIRMWARE_EFI_EMBEDDED, firmware) \
1284ab5b
MZ
2991 id(MODULE, kernel-module) \
2992 id(KEXEC_IMAGE, kexec-image) \
2993 id(KEXEC_INITRAMFS, kexec-initramfs) \
2994 id(POLICY, security-policy) \
a7d3d039 2995 id(X509_CERTIFICATE, x509-certificate) \
1284ab5b
MZ
2996 id(MAX_ID, )
2997
2998#define __fid_enumify(ENUM, dummy) READING_ ## ENUM,
2999#define __fid_stringify(dummy, str) #str,
3000
bc8ca5b9 3001enum kernel_read_file_id {
1284ab5b
MZ
3002 __kernel_read_file_id(__fid_enumify)
3003};
3004
3005static const char * const kernel_read_file_str[] = {
3006 __kernel_read_file_id(__fid_stringify)
bc8ca5b9
MZ
3007};
3008
da20dfe6 3009static inline const char *kernel_read_file_id_str(enum kernel_read_file_id id)
1284ab5b 3010{
cbfd0c10 3011 if ((unsigned)id >= READING_MAX_ID)
1284ab5b
MZ
3012 return kernel_read_file_str[READING_UNKNOWN];
3013
3014 return kernel_read_file_str[id];
3015}
3016
bc8ca5b9
MZ
3017extern int kernel_read_file(struct file *, void **, loff_t *, loff_t,
3018 enum kernel_read_file_id);
711aab1d 3019extern int kernel_read_file_from_path(const char *, void **, loff_t *, loff_t,
09596b94 3020 enum kernel_read_file_id);
901cff7c
TM
3021extern int kernel_read_file_from_path_initns(const char *, void **, loff_t *, loff_t,
3022 enum kernel_read_file_id);
b844f0ec
MZ
3023extern int kernel_read_file_from_fd(int, void **, loff_t *, loff_t,
3024 enum kernel_read_file_id);
bdd1d2d3 3025extern ssize_t kernel_read(struct file *, void *, size_t, loff_t *);
e13ec939 3026extern ssize_t kernel_write(struct file *, const void *, size_t, loff_t *);
73e18f7c 3027extern ssize_t __kernel_write(struct file *, const void *, size_t, loff_t *);
1da177e4
LT
3028extern struct file * open_exec(const char *);
3029
3030/* fs/dcache.c -- generic fs support functions */
a6e5787f 3031extern bool is_subdir(struct dentry *, struct dentry *);
640eb7e7 3032extern bool path_is_under(const struct path *, const struct path *);
1da177e4 3033
9bf39ab2
MS
3034extern char *file_path(struct file *, char *, int);
3035
1da177e4
LT
3036#include <linux/err.h>
3037
3038/* needed for stackable file system support */
965c8e59 3039extern loff_t default_llseek(struct file *file, loff_t offset, int whence);
1da177e4 3040
965c8e59 3041extern loff_t vfs_llseek(struct file *file, loff_t offset, int whence);
1da177e4 3042
54e34621 3043extern int inode_init_always(struct super_block *, struct inode *);
1da177e4 3044extern void inode_init_once(struct inode *);
2aa15890 3045extern void address_space_init_once(struct address_space *mapping);
1da177e4
LT
3046extern struct inode * igrab(struct inode *);
3047extern ino_t iunique(struct super_block *, ino_t);
3048extern int inode_needs_sync(struct inode *inode);
45321ac5 3049extern int generic_delete_inode(struct inode *inode);
9bcb4b73
AV
3050static inline int generic_drop_inode(struct inode *inode)
3051{
3052 return !inode->i_nlink || inode_unhashed(inode);
3053}
1da177e4 3054
88bd5121
AA
3055extern struct inode *ilookup5_nowait(struct super_block *sb,
3056 unsigned long hashval, int (*test)(struct inode *, void *),
3057 void *data);
1da177e4
LT
3058extern struct inode *ilookup5(struct super_block *sb, unsigned long hashval,
3059 int (*test)(struct inode *, void *), void *data);
3060extern struct inode *ilookup(struct super_block *sb, unsigned long ino);
3061
80ea09a0
MS
3062extern struct inode *inode_insert5(struct inode *inode, unsigned long hashval,
3063 int (*test)(struct inode *, void *),
3064 int (*set)(struct inode *, void *),
3065 void *data);
1da177e4
LT
3066extern struct inode * iget5_locked(struct super_block *, unsigned long, int (*test)(struct inode *, void *), int (*set)(struct inode *, void *), void *);
3067extern struct inode * iget_locked(struct super_block *, unsigned long);
fe032c42
TT
3068extern struct inode *find_inode_nowait(struct super_block *,
3069 unsigned long,
3070 int (*match)(struct inode *,
3071 unsigned long, void *),
3072 void *data);
261bca86
AV
3073extern int insert_inode_locked4(struct inode *, unsigned long, int (*test)(struct inode *, void *), void *);
3074extern int insert_inode_locked(struct inode *);
e096d0c7
JB
3075#ifdef CONFIG_DEBUG_LOCK_ALLOC
3076extern void lockdep_annotate_inode_mutex_key(struct inode *inode);
3077#else
3078static inline void lockdep_annotate_inode_mutex_key(struct inode *inode) { };
3079#endif
1da177e4 3080extern void unlock_new_inode(struct inode *);
c2b6d621 3081extern void discard_new_inode(struct inode *);
85fe4025 3082extern unsigned int get_next_ino(void);
799ea9e9 3083extern void evict_inodes(struct super_block *sb);
1da177e4 3084
1da177e4 3085extern void __iget(struct inode * inode);
b46980fe 3086extern void iget_failed(struct inode *);
dbd5768f 3087extern void clear_inode(struct inode *);
2e00c97e 3088extern void __destroy_inode(struct inode *);
a209dfc7
ED
3089extern struct inode *new_inode_pseudo(struct super_block *sb);
3090extern struct inode *new_inode(struct super_block *sb);
ff0c7d15 3091extern void free_inode_nonrcu(struct inode *inode);
01de85e0 3092extern int should_remove_suid(struct dentry *);
5fa8e0a1 3093extern int file_remove_privs(struct file *);
1da177e4
LT
3094
3095extern void __insert_inode_hash(struct inode *, unsigned long hashval);
646ec461
CH
3096static inline void insert_inode_hash(struct inode *inode)
3097{
1da177e4
LT
3098 __insert_inode_hash(inode, inode->i_ino);
3099}
f2ee7abf
ED
3100
3101extern void __remove_inode_hash(struct inode *);
3102static inline void remove_inode_hash(struct inode *inode)
3103{
cbedaac6 3104 if (!inode_unhashed(inode) && !hlist_fake(&inode->i_hash))
f2ee7abf
ED
3105 __remove_inode_hash(inode);
3106}
3107
646ec461 3108extern void inode_sb_list_add(struct inode *inode);
1da177e4 3109
9361401e 3110#ifdef CONFIG_BLOCK
1da177e4 3111extern int bdev_read_only(struct block_device *);
9361401e 3112#endif
1da177e4
LT
3113extern int set_blocksize(struct block_device *, int);
3114extern int sb_set_blocksize(struct super_block *, int);
3115extern int sb_min_blocksize(struct super_block *, int);
3116
3117extern int generic_file_mmap(struct file *, struct vm_area_struct *);
3118extern int generic_file_readonly_mmap(struct file *, struct vm_area_struct *);
3309dd04 3119extern ssize_t generic_write_checks(struct kiocb *, struct iov_iter *);
1383a7ed
DW
3120extern int generic_remap_checks(struct file *file_in, loff_t pos_in,
3121 struct file *file_out, loff_t pos_out,
42ec3d4c 3122 loff_t *count, unsigned int remap_flags);
a3171351 3123extern int generic_file_rw_checks(struct file *file_in, struct file *file_out);
96e6e8f4
AG
3124extern int generic_copy_file_checks(struct file *file_in, loff_t pos_in,
3125 struct file *file_out, loff_t pos_out,
3126 size_t *count, unsigned int flags);
ed978a81 3127extern ssize_t generic_file_read_iter(struct kiocb *, struct iov_iter *);
8174202b 3128extern ssize_t __generic_file_write_iter(struct kiocb *, struct iov_iter *);
8174202b 3129extern ssize_t generic_file_write_iter(struct kiocb *, struct iov_iter *);
1af5bb49 3130extern ssize_t generic_file_direct_write(struct kiocb *, struct iov_iter *);
3b93f911 3131extern ssize_t generic_perform_write(struct file *, struct iov_iter *, loff_t);
88e6faef 3132
18e9710e 3133ssize_t vfs_iter_read(struct file *file, struct iov_iter *iter, loff_t *ppos,
ddef7ed2 3134 rwf_t flags);
abbb6589 3135ssize_t vfs_iter_write(struct file *file, struct iov_iter *iter, loff_t *ppos,
ddef7ed2 3136 rwf_t flags);
5dcdc43e
JX
3137ssize_t vfs_iocb_iter_read(struct file *file, struct kiocb *iocb,
3138 struct iov_iter *iter);
3139ssize_t vfs_iocb_iter_write(struct file *file, struct kiocb *iocb,
3140 struct iov_iter *iter);
dbe4e192 3141
eef99380 3142/* fs/block_dev.c */
b2de525f 3143extern ssize_t blkdev_read_iter(struct kiocb *iocb, struct iov_iter *to);
1456c0a8 3144extern ssize_t blkdev_write_iter(struct kiocb *iocb, struct iov_iter *from);
02c24a82
JB
3145extern int blkdev_fsync(struct file *filp, loff_t start, loff_t end,
3146 int datasync);
ff01bb48 3147extern void block_sync_page(struct page *page);
eef99380 3148
88e6faef 3149/* fs/splice.c */
cbb7e577 3150extern ssize_t generic_file_splice_read(struct file *, loff_t *,
88e6faef 3151 struct pipe_inode_info *, size_t, unsigned int);
8d020765 3152extern ssize_t iter_file_splice_write(struct pipe_inode_info *,
cbb7e577 3153 struct file *, loff_t *, size_t, unsigned int);
88e6faef 3154extern ssize_t generic_splice_sendpage(struct pipe_inode_info *pipe,
cbb7e577 3155 struct file *out, loff_t *, size_t len, unsigned int flags);
1c118596
MS
3156extern long do_splice_direct(struct file *in, loff_t *ppos, struct file *out,
3157 loff_t *opos, size_t len, unsigned int flags);
3158
88e6faef 3159
1da177e4
LT
3160extern void
3161file_ra_state_init(struct file_ra_state *ra, struct address_space *mapping);
965c8e59
AM
3162extern loff_t noop_llseek(struct file *file, loff_t offset, int whence);
3163extern loff_t no_llseek(struct file *file, loff_t offset, int whence);
46a1c2c7 3164extern loff_t vfs_setpos(struct file *file, loff_t offset, loff_t maxsize);
965c8e59 3165extern loff_t generic_file_llseek(struct file *file, loff_t offset, int whence);
5760495a 3166extern loff_t generic_file_llseek_size(struct file *file, loff_t offset,
965c8e59 3167 int whence, loff_t maxsize, loff_t eof);
1bf9d14d
AV
3168extern loff_t fixed_size_llseek(struct file *file, loff_t offset,
3169 int whence, loff_t size);
b25472f9
AV
3170extern loff_t no_seek_end_llseek_size(struct file *, loff_t, int, loff_t);
3171extern loff_t no_seek_end_llseek(struct file *, loff_t, int);
1da177e4
LT
3172extern int generic_file_open(struct inode * inode, struct file * filp);
3173extern int nonseekable_open(struct inode * inode, struct file * filp);
10dce8af 3174extern int stream_open(struct inode * inode, struct file * filp);
1da177e4 3175
9361401e 3176#ifdef CONFIG_BLOCK
8a4c1e42 3177typedef void (dio_submit_t)(struct bio *bio, struct inode *inode,
facd07b0 3178 loff_t file_offset);
1da177e4
LT
3179
3180enum {
1e431f5c
CH
3181 /* need locking between buffered and direct access */
3182 DIO_LOCKING = 0x01,
3183
3184 /* filesystem does not support filling holes */
3185 DIO_SKIP_HOLES = 0x02,
1da177e4
LT
3186};
3187
4055351c 3188void dio_end_io(struct bio *bio);
eafdc7d1 3189
17f8c842
OS
3190ssize_t __blockdev_direct_IO(struct kiocb *iocb, struct inode *inode,
3191 struct block_device *bdev, struct iov_iter *iter,
c8b8e32d 3192 get_block_t get_block,
17f8c842
OS
3193 dio_iodone_t end_io, dio_submit_t submit_io,
3194 int flags);
7bb46a67 3195
17f8c842
OS
3196static inline ssize_t blockdev_direct_IO(struct kiocb *iocb,
3197 struct inode *inode,
c8b8e32d 3198 struct iov_iter *iter,
17f8c842 3199 get_block_t get_block)
1da177e4 3200{
17f8c842 3201 return __blockdev_direct_IO(iocb, inode, inode->i_sb->s_bdev, iter,
c8b8e32d 3202 get_block, NULL, NULL, DIO_LOCKING | DIO_SKIP_HOLES);
1da177e4 3203}
9361401e 3204#endif
1da177e4 3205
1d59d61f 3206void inode_dio_wait(struct inode *inode);
fe0f07d0
JA
3207
3208/*
3209 * inode_dio_begin - signal start of a direct I/O requests
3210 * @inode: inode the direct I/O happens on
3211 *
3212 * This is called once we've finished processing a direct I/O request,
3213 * and is used to wake up callers waiting for direct I/O to be quiesced.
3214 */
3215static inline void inode_dio_begin(struct inode *inode)
3216{
3217 atomic_inc(&inode->i_dio_count);
3218}
3219
3220/*
3221 * inode_dio_end - signal finish of a direct I/O requests
3222 * @inode: inode the direct I/O happens on
3223 *
3224 * This is called once we've finished processing a direct I/O request,
3225 * and is used to wake up callers waiting for direct I/O to be quiesced.
3226 */
3227static inline void inode_dio_end(struct inode *inode)
3228{
3229 if (atomic_dec_and_test(&inode->i_dio_count))
3230 wake_up_bit(&inode->i_state, __I_DIO_WAKEUP);
3231}
1d59d61f 3232
a92853b6
KK
3233/*
3234 * Warn about a page cache invalidation failure diring a direct I/O write.
3235 */
3236void dio_warn_stale_pagecache(struct file *filp);
3237
5f16f322
TT
3238extern void inode_set_flags(struct inode *inode, unsigned int flags,
3239 unsigned int mask);
3240
4b6f5d20 3241extern const struct file_operations generic_ro_fops;
1da177e4
LT
3242
3243#define special_file(m) (S_ISCHR(m)||S_ISBLK(m)||S_ISFIFO(m)||S_ISSOCK(m))
3244
5d826c84 3245extern int readlink_copy(char __user *, int, const char *);
1da177e4 3246extern int page_readlink(struct dentry *, char __user *, int);
fceef393
AV
3247extern const char *page_get_link(struct dentry *, struct inode *,
3248 struct delayed_call *);
3249extern void page_put_link(void *);
0adb25d2 3250extern int __page_symlink(struct inode *inode, const char *symname, int len,
54566b2c 3251 int nofs);
1da177e4 3252extern int page_symlink(struct inode *inode, const char *symname, int len);
c5ef1c42 3253extern const struct inode_operations page_symlink_inode_operations;
fceef393 3254extern void kfree_link(void *);
1da177e4 3255extern void generic_fillattr(struct inode *, struct kstat *);
a528d35e
DH
3256extern int vfs_getattr_nosec(const struct path *, struct kstat *, u32, unsigned int);
3257extern int vfs_getattr(const struct path *, struct kstat *, u32, unsigned int);
b462707e 3258void __inode_add_bytes(struct inode *inode, loff_t bytes);
1da177e4 3259void inode_add_bytes(struct inode *inode, loff_t bytes);
1c8924eb 3260void __inode_sub_bytes(struct inode *inode, loff_t bytes);
1da177e4 3261void inode_sub_bytes(struct inode *inode, loff_t bytes);
f4a8116a
JK
3262static inline loff_t __inode_get_bytes(struct inode *inode)
3263{
3264 return (((loff_t)inode->i_blocks) << 9) + inode->i_bytes;
3265}
1da177e4
LT
3266loff_t inode_get_bytes(struct inode *inode);
3267void inode_set_bytes(struct inode *inode, loff_t bytes);
fceef393
AV
3268const char *simple_get_link(struct dentry *, struct inode *,
3269 struct delayed_call *);
61ba64fc 3270extern const struct inode_operations simple_symlink_inode_operations;
1da177e4 3271
5c0ba4e0 3272extern int iterate_dir(struct file *, struct dir_context *);
1da177e4 3273
a528d35e
DH
3274extern int vfs_statx(int, const char __user *, int, struct kstat *, u32);
3275extern int vfs_statx_fd(unsigned int, struct kstat *, u32, unsigned int);
3276
3277static inline int vfs_stat(const char __user *filename, struct kstat *stat)
3278{
deccf497
DH
3279 return vfs_statx(AT_FDCWD, filename, AT_NO_AUTOMOUNT,
3280 stat, STATX_BASIC_STATS);
a528d35e
DH
3281}
3282static inline int vfs_lstat(const char __user *name, struct kstat *stat)
3283{
deccf497 3284 return vfs_statx(AT_FDCWD, name, AT_SYMLINK_NOFOLLOW | AT_NO_AUTOMOUNT,
a528d35e
DH
3285 stat, STATX_BASIC_STATS);
3286}
3287static inline int vfs_fstatat(int dfd, const char __user *filename,
3288 struct kstat *stat, int flags)
3289{
5d38f049
IK
3290 return vfs_statx(dfd, filename, flags | AT_NO_AUTOMOUNT,
3291 stat, STATX_BASIC_STATS);
a528d35e
DH
3292}
3293static inline int vfs_fstat(int fd, struct kstat *stat)
3294{
3295 return vfs_statx_fd(fd, stat, STATX_BASIC_STATS, 0);
3296}
3297
3298
d60874cd 3299extern const char *vfs_get_link(struct dentry *, struct delayed_call *);
fd4a0edf 3300extern int vfs_readlink(struct dentry *, char __user *, int);
1da177e4 3301
e9079cce 3302extern int __generic_block_fiemap(struct inode *inode,
3a3076f4
JB
3303 struct fiemap_extent_info *fieinfo,
3304 loff_t start, loff_t len,
3305 get_block_t *get_block);
68c9d702
JB
3306extern int generic_block_fiemap(struct inode *inode,
3307 struct fiemap_extent_info *fieinfo, u64 start,
3308 u64 len, get_block_t *get_block);
1da177e4 3309
ee416bcd 3310extern struct file_system_type *get_filesystem(struct file_system_type *fs);
c18479fe 3311extern void put_filesystem(struct file_system_type *fs);
1da177e4
LT
3312extern struct file_system_type *get_fs_type(const char *name);
3313extern struct super_block *get_super(struct block_device *);
6b6dc836 3314extern struct super_block *get_super_thawed(struct block_device *);
ba6379f7 3315extern struct super_block *get_super_exclusive_thawed(struct block_device *bdev);
4504230a 3316extern struct super_block *get_active_super(struct block_device *bdev);
1da177e4 3317extern void drop_super(struct super_block *sb);
ba6379f7 3318extern void drop_super_exclusive(struct super_block *sb);
01a05b33 3319extern void iterate_supers(void (*)(struct super_block *, void *), void *);
43e15cdb
AV
3320extern void iterate_supers_type(struct file_system_type *,
3321 void (*)(struct super_block *, void *), void *);
1da177e4
LT
3322
3323extern int dcache_dir_open(struct inode *, struct file *);
3324extern int dcache_dir_close(struct inode *, struct file *);
3325extern loff_t dcache_dir_lseek(struct file *, loff_t, int);
5f99f4e7 3326extern int dcache_readdir(struct file *, struct dir_context *);
7bb46a67 3327extern int simple_setattr(struct dentry *, struct iattr *);
a528d35e 3328extern int simple_getattr(const struct path *, struct kstat *, u32, unsigned int);
726c3342 3329extern int simple_statfs(struct dentry *, struct kstatfs *);
20955e89 3330extern int simple_open(struct inode *inode, struct file *file);
1da177e4
LT
3331extern int simple_link(struct dentry *, struct inode *, struct dentry *);
3332extern int simple_unlink(struct inode *, struct dentry *);
3333extern int simple_rmdir(struct inode *, struct dentry *);
e0e0be8a
MS
3334extern int simple_rename(struct inode *, struct dentry *,
3335 struct inode *, struct dentry *, unsigned int);
a3d1e7eb
AV
3336extern void simple_recursive_removal(struct dentry *,
3337 void (*callback)(struct dentry *));
02c24a82 3338extern int noop_fsync(struct file *, loff_t, loff_t, int);
f44c7763
DW
3339extern int noop_set_page_dirty(struct page *page);
3340extern void noop_invalidatepage(struct page *page, unsigned int offset,
3341 unsigned int length);
3342extern ssize_t noop_direct_IO(struct kiocb *iocb, struct iov_iter *iter);
1da177e4
LT
3343extern int simple_empty(struct dentry *);
3344extern int simple_readpage(struct file *file, struct page *page);
afddba49
NP
3345extern int simple_write_begin(struct file *file, struct address_space *mapping,
3346 loff_t pos, unsigned len, unsigned flags,
3347 struct page **pagep, void **fsdata);
3348extern int simple_write_end(struct file *file, struct address_space *mapping,
3349 loff_t pos, unsigned len, unsigned copied,
3350 struct page *page, void *fsdata);
b26d4cd3 3351extern int always_delete_dentry(const struct dentry *);
6987843f 3352extern struct inode *alloc_anon_inode(struct super_block *);
e6f5c789 3353extern int simple_nosetlease(struct file *, long, struct file_lock **, void **);
b26d4cd3 3354extern const struct dentry_operations simple_dentry_operations;
1da177e4 3355
00cd8dd3 3356extern struct dentry *simple_lookup(struct inode *, struct dentry *, unsigned int flags);
1da177e4 3357extern ssize_t generic_read_dir(struct file *, char __user *, size_t, loff_t *);
4b6f5d20 3358extern const struct file_operations simple_dir_operations;
c5ef1c42 3359extern const struct inode_operations simple_dir_inode_operations;
fbabfd0f
EB
3360extern void make_empty_dir_inode(struct inode *inode);
3361extern bool is_empty_dir_inode(struct inode *inode);
cda37124 3362struct tree_descr { const char *name; const struct file_operations *ops; int mode; };
1da177e4 3363struct dentry *d_alloc_name(struct dentry *, const char *);
cda37124
EB
3364extern int simple_fill_super(struct super_block *, unsigned long,
3365 const struct tree_descr *);
1f5ce9e9 3366extern int simple_pin_fs(struct file_system_type *, struct vfsmount **mount, int *count);
1da177e4
LT
3367extern void simple_release_fs(struct vfsmount **mount, int *count);
3368
93b07113
AM
3369extern ssize_t simple_read_from_buffer(void __user *to, size_t count,
3370 loff_t *ppos, const void *from, size_t available);
6a727b43
JS
3371extern ssize_t simple_write_to_buffer(void *to, size_t available, loff_t *ppos,
3372 const void __user *from, size_t count);
1da177e4 3373
ac13a829 3374extern int __generic_file_fsync(struct file *, loff_t, loff_t, int);
02c24a82 3375extern int generic_file_fsync(struct file *, loff_t, loff_t, int);
d5aacad5 3376
30ca22c7
PL
3377extern int generic_check_addressable(unsigned, u64);
3378
e965f963 3379#ifdef CONFIG_MIGRATION
2d1db3b1 3380extern int buffer_migrate_page(struct address_space *,
a6bc32b8
MG
3381 struct page *, struct page *,
3382 enum migrate_mode);
89cb0888
JK
3383extern int buffer_migrate_page_norefs(struct address_space *,
3384 struct page *, struct page *,
3385 enum migrate_mode);
e965f963
CL
3386#else
3387#define buffer_migrate_page NULL
89cb0888 3388#define buffer_migrate_page_norefs NULL
e965f963
CL
3389#endif
3390
31051c85 3391extern int setattr_prepare(struct dentry *, struct iattr *);
25d9e2d1 3392extern int inode_newsize_ok(const struct inode *, loff_t offset);
6a1a90ad 3393extern void setattr_copy(struct inode *inode, const struct iattr *attr);
1da177e4 3394
c3b2da31 3395extern int file_update_time(struct file *file);
1da177e4 3396
d475c634
MW
3397static inline bool io_is_direct(struct file *filp)
3398{
65f87ee7 3399 return (filp->f_flags & O_DIRECT) || IS_DAX(filp->f_mapping->host);
d475c634
MW
3400}
3401
f05a3849 3402static inline bool vma_is_dax(const struct vm_area_struct *vma)
baabda26
DW
3403{
3404 return vma->vm_file && IS_DAX(vma->vm_file->f_mapping->host);
3405}
3406
2bb6d283
DW
3407static inline bool vma_is_fsdax(struct vm_area_struct *vma)
3408{
3409 struct inode *inode;
3410
3411 if (!vma->vm_file)
3412 return false;
3413 if (!vma_is_dax(vma))
3414 return false;
3415 inode = file_inode(vma->vm_file);
230f5a89 3416 if (S_ISCHR(inode->i_mode))
2bb6d283
DW
3417 return false; /* device-dax */
3418 return true;
3419}
3420
2ba48ce5
AV
3421static inline int iocb_flags(struct file *file)
3422{
3423 int res = 0;
3424 if (file->f_flags & O_APPEND)
3425 res |= IOCB_APPEND;
3426 if (io_is_direct(file))
3427 res |= IOCB_DIRECT;
dde0c2e7
CH
3428 if ((file->f_flags & O_DSYNC) || IS_SYNC(file->f_mapping->host))
3429 res |= IOCB_DSYNC;
3430 if (file->f_flags & __O_SYNC)
3431 res |= IOCB_SYNC;
2ba48ce5
AV
3432 return res;
3433}
3434
ddef7ed2 3435static inline int kiocb_set_rw_flags(struct kiocb *ki, rwf_t flags)
fdd2f5b7
GR
3436{
3437 if (unlikely(flags & ~RWF_SUPPORTED))
3438 return -EOPNOTSUPP;
3439
b745fafa 3440 if (flags & RWF_NOWAIT) {
91f9943e 3441 if (!(ki->ki_filp->f_mode & FMODE_NOWAIT))
b745fafa
GR
3442 return -EOPNOTSUPP;
3443 ki->ki_flags |= IOCB_NOWAIT;
3444 }
fdd2f5b7
GR
3445 if (flags & RWF_HIPRI)
3446 ki->ki_flags |= IOCB_HIPRI;
3447 if (flags & RWF_DSYNC)
3448 ki->ki_flags |= IOCB_DSYNC;
3449 if (flags & RWF_SYNC)
3450 ki->ki_flags |= (IOCB_DSYNC | IOCB_SYNC);
e1fc742e
JB
3451 if (flags & RWF_APPEND)
3452 ki->ki_flags |= IOCB_APPEND;
fdd2f5b7
GR
3453 return 0;
3454}
3455
1da177e4
LT
3456static inline ino_t parent_ino(struct dentry *dentry)
3457{
3458 ino_t res;
3459
b5c84bf6
NP
3460 /*
3461 * Don't strictly need d_lock here? If the parent ino could change
3462 * then surely we'd have a deeper race in the caller?
3463 */
1da177e4
LT
3464 spin_lock(&dentry->d_lock);
3465 res = dentry->d_parent->d_inode->i_ino;
3466 spin_unlock(&dentry->d_lock);
3467 return res;
3468}
3469
1da177e4
LT
3470/* Transaction based IO helpers */
3471
3472/*
3473 * An argresp is stored in an allocated page and holds the
3474 * size of the argument or response, along with its content
3475 */
3476struct simple_transaction_argresp {
3477 ssize_t size;
3478 char data[0];
3479};
3480
3481#define SIMPLE_TRANSACTION_LIMIT (PAGE_SIZE - sizeof(struct simple_transaction_argresp))
3482
3483char *simple_transaction_get(struct file *file, const char __user *buf,
3484 size_t size);
3485ssize_t simple_transaction_read(struct file *file, char __user *buf,
3486 size_t size, loff_t *pos);
3487int simple_transaction_release(struct inode *inode, struct file *file);
3488
76791ab2 3489void simple_transaction_set(struct file *file, size_t n);
1da177e4 3490
acaefc25
AB
3491/*
3492 * simple attribute files
3493 *
3494 * These attributes behave similar to those in sysfs:
3495 *
3496 * Writing to an attribute immediately sets a value, an open file can be
3497 * written to multiple times.
3498 *
3499 * Reading from an attribute creates a buffer from the value that might get
3500 * read with multiple read calls. When the attribute has been read
3501 * completely, no further read calls are possible until the file is opened
3502 * again.
3503 *
3504 * All attributes contain a text representation of a numeric value
3505 * that are accessed with the get() and set() functions.
3506 */
3507#define DEFINE_SIMPLE_ATTRIBUTE(__fops, __get, __set, __fmt) \
3508static int __fops ## _open(struct inode *inode, struct file *file) \
3509{ \
3510 __simple_attr_check_format(__fmt, 0ull); \
3511 return simple_attr_open(inode, file, __get, __set, __fmt); \
3512} \
828c0950 3513static const struct file_operations __fops = { \
acaefc25
AB
3514 .owner = THIS_MODULE, \
3515 .open = __fops ## _open, \
74bedc4d 3516 .release = simple_attr_release, \
acaefc25
AB
3517 .read = simple_attr_read, \
3518 .write = simple_attr_write, \
1ec5584e 3519 .llseek = generic_file_llseek, \
68be3029 3520}
acaefc25 3521
b9075fa9
JP
3522static inline __printf(1, 2)
3523void __simple_attr_check_format(const char *fmt, ...)
acaefc25
AB
3524{
3525 /* don't do anything, just let the compiler check the arguments; */
3526}
3527
3528int simple_attr_open(struct inode *inode, struct file *file,
8b88b099 3529 int (*get)(void *, u64 *), int (*set)(void *, u64),
acaefc25 3530 const char *fmt);
74bedc4d 3531int simple_attr_release(struct inode *inode, struct file *file);
acaefc25
AB
3532ssize_t simple_attr_read(struct file *file, char __user *buf,
3533 size_t len, loff_t *ppos);
3534ssize_t simple_attr_write(struct file *file, const char __user *buf,
3535 size_t len, loff_t *ppos);
3536
4be28540 3537struct ctl_table;
8d65af78 3538int proc_nr_files(struct ctl_table *table, int write,
62239ac2 3539 void __user *buffer, size_t *lenp, loff_t *ppos);
312d3ca8
CH
3540int proc_nr_dentry(struct ctl_table *table, int write,
3541 void __user *buffer, size_t *lenp, loff_t *ppos);
cffbc8aa
DC
3542int proc_nr_inodes(struct ctl_table *table, int write,
3543 void __user *buffer, size_t *lenp, loff_t *ppos);
38e23c95 3544int __init get_filesystem_list(char *buf);
62239ac2 3545
3cd90ea4 3546#define __FMODE_EXEC ((__force int) FMODE_EXEC)
1a44bc8c
NK
3547#define __FMODE_NONOTIFY ((__force int) FMODE_NONOTIFY)
3548
6d125529 3549#define ACC_MODE(x) ("\004\002\006\006"[(x)&O_ACCMODE])
ecf081d1 3550#define OPEN_FMODE(flag) ((__force fmode_t)(((flag + 1) & O_ACCMODE) | \
1a44bc8c 3551 (flag & __FMODE_NONOTIFY)))
5300990c 3552
d37177ba 3553static inline bool is_sxid(umode_t mode)
69b45732
AK
3554{
3555 return (mode & S_ISUID) || ((mode & S_ISGID) && (mode & S_IXGRP));
3556}
3557
cbdf35bc
MS
3558static inline int check_sticky(struct inode *dir, struct inode *inode)
3559{
3560 if (!(dir->i_mode & S_ISVTX))
3561 return 0;
3562
3563 return __check_sticky(dir, inode);
3564}
3565
69b45732
AK
3566static inline void inode_has_no_xattr(struct inode *inode)
3567{
e462ec50 3568 if (!is_sxid(inode->i_mode) && (inode->i_sb->s_flags & SB_NOSEC))
69b45732
AK
3569 inode->i_flags |= S_NOSEC;
3570}
3571
a7400222
AV
3572static inline bool is_root_inode(struct inode *inode)
3573{
3574 return inode == inode->i_sb->s_root->d_inode;
3575}
3576
5f99f4e7
AV
3577static inline bool dir_emit(struct dir_context *ctx,
3578 const char *name, int namelen,
3579 u64 ino, unsigned type)
3580{
3581 return ctx->actor(ctx, name, namelen, ctx->pos, ino, type) == 0;
3582}
3583static inline bool dir_emit_dot(struct file *file, struct dir_context *ctx)
3584{
3585 return ctx->actor(ctx, ".", 1, ctx->pos,
3586 file->f_path.dentry->d_inode->i_ino, DT_DIR) == 0;
3587}
3588static inline bool dir_emit_dotdot(struct file *file, struct dir_context *ctx)
3589{
3590 return ctx->actor(ctx, "..", 2, ctx->pos,
3591 parent_ino(file->f_path.dentry), DT_DIR) == 0;
3592}
3593static inline bool dir_emit_dots(struct file *file, struct dir_context *ctx)
3594{
3595 if (ctx->pos == 0) {
3596 if (!dir_emit_dot(file, ctx))
3597 return false;
3598 ctx->pos = 1;
3599 }
3600 if (ctx->pos == 1) {
3601 if (!dir_emit_dotdot(file, ctx))
3602 return false;
3603 ctx->pos = 2;
3604 }
3605 return true;
3606}
5ded75ec
AV
3607static inline bool dir_relax(struct inode *inode)
3608{
5955102c
AV
3609 inode_unlock(inode);
3610 inode_lock(inode);
5ded75ec
AV
3611 return !IS_DEADDIR(inode);
3612}
5f99f4e7 3613
ae05327a
AV
3614static inline bool dir_relax_shared(struct inode *inode)
3615{
3616 inode_unlock_shared(inode);
3617 inode_lock_shared(inode);
3618 return !IS_DEADDIR(inode);
3619}
3620
90f8572b 3621extern bool path_noexec(const struct path *path);
21fc61c7 3622extern void inode_nohighmem(struct inode *inode);
90f8572b 3623
45cd0faa
AG
3624/* mm/fadvise.c */
3625extern int vfs_fadvise(struct file *file, loff_t offset, loff_t len,
3626 int advice);
cf1ea059
JK
3627extern int generic_fadvise(struct file *file, loff_t offset, loff_t len,
3628 int advice);
45cd0faa 3629
2b188cc1
JA
3630#if defined(CONFIG_IO_URING)
3631extern struct sock *io_uring_get_socket(struct file *file);
3632#else
3633static inline struct sock *io_uring_get_socket(struct file *file)
3634{
3635 return NULL;
3636}
3637#endif
3638
5aca2842
DW
3639int vfs_ioc_setflags_prepare(struct inode *inode, unsigned int oldflags,
3640 unsigned int flags);
3641
7b0e492e
DW
3642int vfs_ioc_fssetxattr_check(struct inode *inode, const struct fsxattr *old_fa,
3643 struct fsxattr *fa);
3644
3645static inline void simple_fill_fsxattr(struct fsxattr *fa, __u32 xflags)
3646{
3647 memset(fa, 0, sizeof(*fa));
3648 fa->fsx_xflags = xflags;
3649}
3650
dc617f29
DW
3651/*
3652 * Flush file data before changing attributes. Caller must hold any locks
3653 * required to prevent further writes to this file until we're done setting
3654 * flags.
3655 */
3656static inline int inode_drain_writes(struct inode *inode)
3657{
3658 inode_dio_wait(inode);
3659 return filemap_write_and_wait(inode->i_mapping);
3660}
3661
1da177e4 3662#endif /* _LINUX_FS_H */