Commit | Line | Data |
---|---|---|
1da177e4 LT |
1 | #ifndef _LINUX_FS_H |
2 | #define _LINUX_FS_H | |
3 | ||
4 | /* | |
5 | * This file has definitions for some important file table | |
6 | * structures etc. | |
7 | */ | |
8 | ||
1da177e4 LT |
9 | #include <linux/limits.h> |
10 | #include <linux/ioctl.h> | |
11 | ||
12 | /* | |
13 | * It's silly to have NR_OPEN bigger than NR_FILE, but you can change | |
14 | * the file limit at runtime and only root can increase the per-process | |
15 | * nr_file rlimit, so it's safe to set up a ridiculously high absolute | |
16 | * upper limit on files-per-process. | |
17 | * | |
18 | * Some programs (notably those using select()) may have to be | |
19 | * recompiled to take full advantage of the new limits.. | |
20 | */ | |
21 | ||
22 | /* Fixed constants first: */ | |
23 | #undef NR_OPEN | |
9cfe015a | 24 | extern int sysctl_nr_open; |
1da177e4 LT |
25 | #define INR_OPEN 1024 /* Initial setting for nfile rlimits */ |
26 | ||
27 | #define BLOCK_SIZE_BITS 10 | |
28 | #define BLOCK_SIZE (1<<BLOCK_SIZE_BITS) | |
29 | ||
01bf466e DH |
30 | #define SEEK_SET 0 /* seek relative to beginning of file */ |
31 | #define SEEK_CUR 1 /* seek relative to current file position */ | |
32 | #define SEEK_END 2 /* seek relative to end of file */ | |
1ae7075b | 33 | #define SEEK_MAX SEEK_END |
01bf466e | 34 | |
1da177e4 LT |
35 | /* And dynamically-tunable limits and defaults: */ |
36 | struct files_stat_struct { | |
37 | int nr_files; /* read only */ | |
38 | int nr_free_files; /* read only */ | |
39 | int max_files; /* tunable */ | |
40 | }; | |
41 | extern struct files_stat_struct files_stat; | |
529bf6be | 42 | extern int get_max_files(void); |
1da177e4 LT |
43 | |
44 | struct inodes_stat_t { | |
45 | int nr_inodes; | |
46 | int nr_unused; | |
9e7bf24b | 47 | int dummy[5]; /* padding for sysctl ABI compatibility */ |
1da177e4 LT |
48 | }; |
49 | extern struct inodes_stat_t inodes_stat; | |
50 | ||
51 | extern int leases_enable, lease_break_time; | |
52 | ||
53 | #ifdef CONFIG_DNOTIFY | |
54 | extern int dir_notify_enable; | |
55 | #endif | |
56 | ||
57 | #define NR_FILE 8192 /* this can well be larger on a larger system */ | |
58 | ||
59 | #define MAY_EXEC 1 | |
60 | #define MAY_WRITE 2 | |
61 | #define MAY_READ 4 | |
62 | #define MAY_APPEND 8 | |
63 | ||
64 | #define FMODE_READ 1 | |
65 | #define FMODE_WRITE 2 | |
66 | ||
67 | /* Internal kernel extensions */ | |
68 | #define FMODE_LSEEK 4 | |
69 | #define FMODE_PREAD 8 | |
70 | #define FMODE_PWRITE FMODE_PREAD /* These go hand in hand */ | |
71 | ||
b500531e OD |
72 | /* File is being opened for execution. Primary users of this flag are |
73 | distributed filesystems that can use it to achieve correct ETXTBUSY | |
74 | behavior for cross-node execution/opening_for_writing of files */ | |
75 | #define FMODE_EXEC 16 | |
76 | ||
1da177e4 LT |
77 | #define RW_MASK 1 |
78 | #define RWA_MASK 2 | |
79 | #define READ 0 | |
80 | #define WRITE 1 | |
81 | #define READA 2 /* read-ahead - don't block if no resources */ | |
a7662236 | 82 | #define SWRITE 3 /* for ll_rw_block() - wait for buffer lock */ |
1da177e4 | 83 | #define READ_SYNC (READ | (1 << BIO_RW_SYNC)) |
5404bc7a | 84 | #define READ_META (READ | (1 << BIO_RW_META)) |
1da177e4 | 85 | #define WRITE_SYNC (WRITE | (1 << BIO_RW_SYNC)) |
18ce3751 | 86 | #define SWRITE_SYNC (SWRITE | (1 << BIO_RW_SYNC)) |
1da177e4 LT |
87 | #define WRITE_BARRIER ((1 << BIO_RW) | (1 << BIO_RW_BARRIER)) |
88 | ||
89 | #define SEL_IN 1 | |
90 | #define SEL_OUT 2 | |
91 | #define SEL_EX 4 | |
92 | ||
93 | /* public flags for file_system_type */ | |
94 | #define FS_REQUIRES_DEV 1 | |
95 | #define FS_BINARY_MOUNTDATA 2 | |
79c0b2df | 96 | #define FS_HAS_SUBTYPE 4 |
1da177e4 | 97 | #define FS_REVAL_DOT 16384 /* Check the paths ".", ".." for staleness */ |
349457cc MF |
98 | #define FS_RENAME_DOES_D_MOVE 32768 /* FS will handle d_move() |
99 | * during rename() internally. | |
100 | */ | |
101 | ||
1da177e4 LT |
102 | /* |
103 | * These are the fs-independent mount-flags: up to 32 flags are supported | |
104 | */ | |
105 | #define MS_RDONLY 1 /* Mount read-only */ | |
106 | #define MS_NOSUID 2 /* Ignore suid and sgid bits */ | |
107 | #define MS_NODEV 4 /* Disallow access to device special files */ | |
108 | #define MS_NOEXEC 8 /* Disallow program execution */ | |
109 | #define MS_SYNCHRONOUS 16 /* Writes are synced at once */ | |
110 | #define MS_REMOUNT 32 /* Alter flags of a mounted FS */ | |
111 | #define MS_MANDLOCK 64 /* Allow mandatory locks on an FS */ | |
112 | #define MS_DIRSYNC 128 /* Directory modifications are synchronous */ | |
113 | #define MS_NOATIME 1024 /* Do not update access times. */ | |
114 | #define MS_NODIRATIME 2048 /* Do not update directory access times */ | |
115 | #define MS_BIND 4096 | |
116 | #define MS_MOVE 8192 | |
117 | #define MS_REC 16384 | |
9b04c997 TT |
118 | #define MS_VERBOSE 32768 /* War is peace. Verbosity is silence. |
119 | MS_VERBOSE is deprecated. */ | |
120 | #define MS_SILENT 32768 | |
bf066c7d | 121 | #define MS_POSIXACL (1<<16) /* VFS does not apply the umask */ |
9676f0c6 | 122 | #define MS_UNBINDABLE (1<<17) /* change to unbindable */ |
07b20889 | 123 | #define MS_PRIVATE (1<<18) /* change to private */ |
a58b0eb8 | 124 | #define MS_SLAVE (1<<19) /* change to slave */ |
03e06e68 | 125 | #define MS_SHARED (1<<20) /* change to shared */ |
47ae32d6 | 126 | #define MS_RELATIME (1<<21) /* Update atime relative to mtime/ctime. */ |
8bf9725c | 127 | #define MS_KERNMOUNT (1<<22) /* this is a kern_mount call */ |
7a224228 | 128 | #define MS_I_VERSION (1<<23) /* Update inode I_version field */ |
1da177e4 LT |
129 | #define MS_ACTIVE (1<<30) |
130 | #define MS_NOUSER (1<<31) | |
131 | ||
132 | /* | |
133 | * Superblock flags that can be altered by MS_REMOUNT | |
134 | */ | |
fc33a7bb | 135 | #define MS_RMT_MASK (MS_RDONLY|MS_SYNCHRONOUS|MS_MANDLOCK) |
1da177e4 LT |
136 | |
137 | /* | |
138 | * Old magic mount flag and mask | |
139 | */ | |
140 | #define MS_MGC_VAL 0xC0ED0000 | |
141 | #define MS_MGC_MSK 0xffff0000 | |
142 | ||
143 | /* Inode flags - they have nothing to superblock flags now */ | |
144 | ||
145 | #define S_SYNC 1 /* Writes are synced at once */ | |
146 | #define S_NOATIME 2 /* Do not update access times */ | |
147 | #define S_APPEND 4 /* Append-only file */ | |
148 | #define S_IMMUTABLE 8 /* Immutable file */ | |
149 | #define S_DEAD 16 /* removed, but still open directory */ | |
150 | #define S_NOQUOTA 32 /* Inode is not counted to quota */ | |
151 | #define S_DIRSYNC 64 /* Directory modifications are synchronous */ | |
152 | #define S_NOCMTIME 128 /* Do not update file c/mtime */ | |
153 | #define S_SWAPFILE 256 /* Do not truncate: swapon got its bmaps */ | |
154 | #define S_PRIVATE 512 /* Inode is fs-internal */ | |
155 | ||
156 | /* | |
157 | * Note that nosuid etc flags are inode-specific: setting some file-system | |
158 | * flags just means all the inodes inherit those flags by default. It might be | |
159 | * possible to override it selectively if you really wanted to with some | |
160 | * ioctl() that is not currently implemented. | |
161 | * | |
162 | * Exception: MS_RDONLY is always applied to the entire file system. | |
163 | * | |
164 | * Unfortunately, it is possible to change a filesystems flags with it mounted | |
165 | * with files in use. This means that all of the inodes will not have their | |
166 | * i_flags updated. Hence, i_flags no longer inherit the superblock mount | |
167 | * flags, so these have to be checked separately. -- rmk@arm.uk.linux.org | |
168 | */ | |
169 | #define __IS_FLG(inode,flg) ((inode)->i_sb->s_flags & (flg)) | |
170 | ||
171 | #define IS_RDONLY(inode) ((inode)->i_sb->s_flags & MS_RDONLY) | |
172 | #define IS_SYNC(inode) (__IS_FLG(inode, MS_SYNCHRONOUS) || \ | |
173 | ((inode)->i_flags & S_SYNC)) | |
174 | #define IS_DIRSYNC(inode) (__IS_FLG(inode, MS_SYNCHRONOUS|MS_DIRSYNC) || \ | |
175 | ((inode)->i_flags & (S_SYNC|S_DIRSYNC))) | |
176 | #define IS_MANDLOCK(inode) __IS_FLG(inode, MS_MANDLOCK) | |
37756ced | 177 | #define IS_NOATIME(inode) __IS_FLG(inode, MS_RDONLY|MS_NOATIME) |
7a224228 | 178 | #define IS_I_VERSION(inode) __IS_FLG(inode, MS_I_VERSION) |
1da177e4 LT |
179 | |
180 | #define IS_NOQUOTA(inode) ((inode)->i_flags & S_NOQUOTA) | |
181 | #define IS_APPEND(inode) ((inode)->i_flags & S_APPEND) | |
182 | #define IS_IMMUTABLE(inode) ((inode)->i_flags & S_IMMUTABLE) | |
1da177e4 LT |
183 | #define IS_POSIXACL(inode) __IS_FLG(inode, MS_POSIXACL) |
184 | ||
185 | #define IS_DEADDIR(inode) ((inode)->i_flags & S_DEAD) | |
186 | #define IS_NOCMTIME(inode) ((inode)->i_flags & S_NOCMTIME) | |
187 | #define IS_SWAPFILE(inode) ((inode)->i_flags & S_SWAPFILE) | |
188 | #define IS_PRIVATE(inode) ((inode)->i_flags & S_PRIVATE) | |
189 | ||
190 | /* the read-only stuff doesn't really belong here, but any other place is | |
191 | probably as bad and I don't want to create yet another include file. */ | |
192 | ||
193 | #define BLKROSET _IO(0x12,93) /* set device read-only (0 = read-write) */ | |
194 | #define BLKROGET _IO(0x12,94) /* get read-only status (0 = read_write) */ | |
195 | #define BLKRRPART _IO(0x12,95) /* re-read partition table */ | |
196 | #define BLKGETSIZE _IO(0x12,96) /* return device size /512 (long *arg) */ | |
197 | #define BLKFLSBUF _IO(0x12,97) /* flush buffer cache */ | |
198 | #define BLKRASET _IO(0x12,98) /* set read ahead for block device */ | |
199 | #define BLKRAGET _IO(0x12,99) /* get current read ahead setting */ | |
200 | #define BLKFRASET _IO(0x12,100)/* set filesystem (mm/filemap.c) read-ahead */ | |
201 | #define BLKFRAGET _IO(0x12,101)/* get filesystem (mm/filemap.c) read-ahead */ | |
202 | #define BLKSECTSET _IO(0x12,102)/* set max sectors per request (ll_rw_blk.c) */ | |
203 | #define BLKSECTGET _IO(0x12,103)/* get max sectors per request (ll_rw_blk.c) */ | |
204 | #define BLKSSZGET _IO(0x12,104)/* get block device sector size */ | |
205 | #if 0 | |
206 | #define BLKPG _IO(0x12,105)/* See blkpg.h */ | |
207 | ||
208 | /* Some people are morons. Do not use sizeof! */ | |
209 | ||
210 | #define BLKELVGET _IOR(0x12,106,size_t)/* elevator get */ | |
211 | #define BLKELVSET _IOW(0x12,107,size_t)/* elevator set */ | |
212 | /* This was here just to show that the number is taken - | |
213 | probably all these _IO(0x12,*) ioctls should be moved to blkpg.h. */ | |
214 | #endif | |
215 | /* A jump here: 108-111 have been used for various private purposes. */ | |
216 | #define BLKBSZGET _IOR(0x12,112,size_t) | |
217 | #define BLKBSZSET _IOW(0x12,113,size_t) | |
218 | #define BLKGETSIZE64 _IOR(0x12,114,size_t) /* return device size in bytes (u64 *arg) */ | |
2056a782 JA |
219 | #define BLKTRACESETUP _IOWR(0x12,115,struct blk_user_trace_setup) |
220 | #define BLKTRACESTART _IO(0x12,116) | |
221 | #define BLKTRACESTOP _IO(0x12,117) | |
222 | #define BLKTRACETEARDOWN _IO(0x12,118) | |
1da177e4 LT |
223 | |
224 | #define BMAP_IOCTL 1 /* obsolete - kept for compatibility */ | |
225 | #define FIBMAP _IO(0x00,1) /* bmap access */ | |
226 | #define FIGETBSZ _IO(0x00,2) /* get the block size used for bmap */ | |
227 | ||
36695673 DH |
228 | #define FS_IOC_GETFLAGS _IOR('f', 1, long) |
229 | #define FS_IOC_SETFLAGS _IOW('f', 2, long) | |
230 | #define FS_IOC_GETVERSION _IOR('v', 1, long) | |
231 | #define FS_IOC_SETVERSION _IOW('v', 2, long) | |
232 | #define FS_IOC32_GETFLAGS _IOR('f', 1, int) | |
233 | #define FS_IOC32_SETFLAGS _IOW('f', 2, int) | |
234 | #define FS_IOC32_GETVERSION _IOR('v', 1, int) | |
235 | #define FS_IOC32_SETVERSION _IOW('v', 2, int) | |
236 | ||
237 | /* | |
238 | * Inode flags (FS_IOC_GETFLAGS / FS_IOC_SETFLAGS) | |
239 | */ | |
240 | #define FS_SECRM_FL 0x00000001 /* Secure deletion */ | |
241 | #define FS_UNRM_FL 0x00000002 /* Undelete */ | |
242 | #define FS_COMPR_FL 0x00000004 /* Compress file */ | |
243 | #define FS_SYNC_FL 0x00000008 /* Synchronous updates */ | |
244 | #define FS_IMMUTABLE_FL 0x00000010 /* Immutable file */ | |
245 | #define FS_APPEND_FL 0x00000020 /* writes to file may only append */ | |
246 | #define FS_NODUMP_FL 0x00000040 /* do not dump file */ | |
247 | #define FS_NOATIME_FL 0x00000080 /* do not update atime */ | |
248 | /* Reserved for compression usage... */ | |
249 | #define FS_DIRTY_FL 0x00000100 | |
250 | #define FS_COMPRBLK_FL 0x00000200 /* One or more compressed clusters */ | |
251 | #define FS_NOCOMP_FL 0x00000400 /* Don't compress */ | |
252 | #define FS_ECOMPR_FL 0x00000800 /* Compression error */ | |
253 | /* End compression flags --- maybe not all used */ | |
254 | #define FS_BTREE_FL 0x00001000 /* btree format dir */ | |
255 | #define FS_INDEX_FL 0x00001000 /* hash-indexed directory */ | |
256 | #define FS_IMAGIC_FL 0x00002000 /* AFS directory */ | |
257 | #define FS_JOURNAL_DATA_FL 0x00004000 /* Reserved for ext3 */ | |
258 | #define FS_NOTAIL_FL 0x00008000 /* file tail should not be merged */ | |
259 | #define FS_DIRSYNC_FL 0x00010000 /* dirsync behaviour (directories only) */ | |
260 | #define FS_TOPDIR_FL 0x00020000 /* Top of directory hierarchies*/ | |
128e5eba SW |
261 | #define FS_EXTENT_FL 0x00080000 /* Extents */ |
262 | #define FS_DIRECTIO_FL 0x00100000 /* Use direct i/o */ | |
36695673 DH |
263 | #define FS_RESERVED_FL 0x80000000 /* reserved for ext2 lib */ |
264 | ||
265 | #define FS_FL_USER_VISIBLE 0x0003DFFF /* User visible flags */ | |
266 | #define FS_FL_USER_MODIFIABLE 0x000380FF /* User modifiable flags */ | |
267 | ||
268 | ||
e46e4903 AM |
269 | #define SYNC_FILE_RANGE_WAIT_BEFORE 1 |
270 | #define SYNC_FILE_RANGE_WRITE 2 | |
271 | #define SYNC_FILE_RANGE_WAIT_AFTER 4 | |
272 | ||
1da177e4 LT |
273 | #ifdef __KERNEL__ |
274 | ||
275 | #include <linux/linkage.h> | |
276 | #include <linux/wait.h> | |
277 | #include <linux/types.h> | |
278 | #include <linux/kdev_t.h> | |
279 | #include <linux/dcache.h> | |
0f7fc9e4 | 280 | #include <linux/namei.h> |
1da177e4 LT |
281 | #include <linux/stat.h> |
282 | #include <linux/cache.h> | |
283 | #include <linux/kobject.h> | |
284 | #include <linux/list.h> | |
285 | #include <linux/radix-tree.h> | |
286 | #include <linux/prio_tree.h> | |
287 | #include <linux/init.h> | |
914e2637 | 288 | #include <linux/pid.h> |
1b1dcc1b | 289 | #include <linux/mutex.h> |
3bd858ab | 290 | #include <linux/capability.h> |
6188e10d | 291 | #include <linux/semaphore.h> |
1da177e4 LT |
292 | |
293 | #include <asm/atomic.h> | |
1da177e4 LT |
294 | #include <asm/byteorder.h> |
295 | ||
a5694255 | 296 | struct export_operations; |
a885c8c4 | 297 | struct hd_geometry; |
1da177e4 LT |
298 | struct iovec; |
299 | struct nameidata; | |
92198f7e | 300 | struct kiocb; |
1da177e4 LT |
301 | struct pipe_inode_info; |
302 | struct poll_table_struct; | |
303 | struct kstatfs; | |
304 | struct vm_area_struct; | |
305 | struct vfsmount; | |
306 | ||
74bf17cf | 307 | extern void __init inode_init(void); |
1da177e4 | 308 | extern void __init inode_init_early(void); |
1da177e4 LT |
309 | extern void __init files_init(unsigned long); |
310 | ||
311 | struct buffer_head; | |
312 | typedef int (get_block_t)(struct inode *inode, sector_t iblock, | |
313 | struct buffer_head *bh_result, int create); | |
92198f7e | 314 | typedef void (dio_iodone_t)(struct kiocb *iocb, loff_t offset, |
1da177e4 LT |
315 | ssize_t bytes, void *private); |
316 | ||
317 | /* | |
318 | * Attribute flags. These should be or-ed together to figure out what | |
319 | * has been changed! | |
320 | */ | |
321 | #define ATTR_MODE 1 | |
322 | #define ATTR_UID 2 | |
323 | #define ATTR_GID 4 | |
324 | #define ATTR_SIZE 8 | |
325 | #define ATTR_ATIME 16 | |
326 | #define ATTR_MTIME 32 | |
327 | #define ATTR_CTIME 64 | |
328 | #define ATTR_ATIME_SET 128 | |
329 | #define ATTR_MTIME_SET 256 | |
330 | #define ATTR_FORCE 512 /* Not a change, but a change it */ | |
331 | #define ATTR_ATTR_FLAG 1024 | |
332 | #define ATTR_KILL_SUID 2048 | |
333 | #define ATTR_KILL_SGID 4096 | |
cc4e69de | 334 | #define ATTR_FILE 8192 |
b5376771 | 335 | #define ATTR_KILL_PRIV 16384 |
d139d7ff | 336 | #define ATTR_OPEN 32768 /* Truncating from open(O_TRUNC) */ |
1da177e4 LT |
337 | |
338 | /* | |
339 | * This is the Inode Attributes structure, used for notify_change(). It | |
340 | * uses the above definitions as flags, to know which values have changed. | |
341 | * Also, in this manner, a Filesystem can look at only the values it cares | |
342 | * about. Basically, these are the attributes that the VFS layer can | |
343 | * request to change from the FS layer. | |
344 | * | |
345 | * Derek Atkins <warlord@MIT.EDU> 94-10-20 | |
346 | */ | |
347 | struct iattr { | |
348 | unsigned int ia_valid; | |
349 | umode_t ia_mode; | |
350 | uid_t ia_uid; | |
351 | gid_t ia_gid; | |
352 | loff_t ia_size; | |
353 | struct timespec ia_atime; | |
354 | struct timespec ia_mtime; | |
355 | struct timespec ia_ctime; | |
cc4e69de MS |
356 | |
357 | /* | |
358 | * Not an attribute, but an auxilary info for filesystems wanting to | |
359 | * implement an ftruncate() like method. NOTE: filesystem should | |
360 | * check for (ia_valid & ATTR_FILE), and not for (ia_file != NULL). | |
361 | */ | |
362 | struct file *ia_file; | |
1da177e4 LT |
363 | }; |
364 | ||
1da177e4 LT |
365 | /* |
366 | * Includes for diskquotas. | |
367 | */ | |
368 | #include <linux/quota.h> | |
369 | ||
994fc28c ZB |
370 | /** |
371 | * enum positive_aop_returns - aop return codes with specific semantics | |
372 | * | |
373 | * @AOP_WRITEPAGE_ACTIVATE: Informs the caller that page writeback has | |
374 | * completed, that the page is still locked, and | |
375 | * should be considered active. The VM uses this hint | |
376 | * to return the page to the active list -- it won't | |
377 | * be a candidate for writeback again in the near | |
378 | * future. Other callers must be careful to unlock | |
379 | * the page if they get this return. Returned by | |
380 | * writepage(); | |
381 | * | |
382 | * @AOP_TRUNCATED_PAGE: The AOP method that was handed a locked page has | |
383 | * unlocked it and the page might have been truncated. | |
384 | * The caller should back up to acquiring a new page and | |
385 | * trying again. The aop will be taking reasonable | |
386 | * precautions not to livelock. If the caller held a page | |
387 | * reference, it should drop it before retrying. Returned | |
55144768 | 388 | * by readpage(). |
994fc28c ZB |
389 | * |
390 | * address_space_operation functions return these large constants to indicate | |
391 | * special semantics to the caller. These are much larger than the bytes in a | |
392 | * page to allow for functions that return the number of bytes operated on in a | |
393 | * given page. | |
394 | */ | |
395 | ||
396 | enum positive_aop_returns { | |
397 | AOP_WRITEPAGE_ACTIVATE = 0x80000, | |
398 | AOP_TRUNCATED_PAGE = 0x80001, | |
399 | }; | |
400 | ||
afddba49 | 401 | #define AOP_FLAG_UNINTERRUPTIBLE 0x0001 /* will not do a short write */ |
89e10787 | 402 | #define AOP_FLAG_CONT_EXPAND 0x0002 /* called from cont_expand */ |
afddba49 | 403 | |
1da177e4 LT |
404 | /* |
405 | * oh the beauties of C type declarations. | |
406 | */ | |
407 | struct page; | |
408 | struct address_space; | |
409 | struct writeback_control; | |
1da177e4 | 410 | |
2f718ffc NP |
411 | struct iov_iter { |
412 | const struct iovec *iov; | |
413 | unsigned long nr_segs; | |
414 | size_t iov_offset; | |
415 | size_t count; | |
416 | }; | |
417 | ||
418 | size_t iov_iter_copy_from_user_atomic(struct page *page, | |
419 | struct iov_iter *i, unsigned long offset, size_t bytes); | |
420 | size_t iov_iter_copy_from_user(struct page *page, | |
421 | struct iov_iter *i, unsigned long offset, size_t bytes); | |
422 | void iov_iter_advance(struct iov_iter *i, size_t bytes); | |
afddba49 | 423 | int iov_iter_fault_in_readable(struct iov_iter *i, size_t bytes); |
2f718ffc NP |
424 | size_t iov_iter_single_seg_count(struct iov_iter *i); |
425 | ||
426 | static inline void iov_iter_init(struct iov_iter *i, | |
427 | const struct iovec *iov, unsigned long nr_segs, | |
428 | size_t count, size_t written) | |
429 | { | |
430 | i->iov = iov; | |
431 | i->nr_segs = nr_segs; | |
432 | i->iov_offset = 0; | |
433 | i->count = count + written; | |
434 | ||
435 | iov_iter_advance(i, written); | |
436 | } | |
437 | ||
438 | static inline size_t iov_iter_count(struct iov_iter *i) | |
439 | { | |
440 | return i->count; | |
441 | } | |
442 | ||
443 | ||
1da177e4 LT |
444 | struct address_space_operations { |
445 | int (*writepage)(struct page *page, struct writeback_control *wbc); | |
446 | int (*readpage)(struct file *, struct page *); | |
3978d717 | 447 | void (*sync_page)(struct page *); |
1da177e4 LT |
448 | |
449 | /* Write back some dirty pages from this mapping. */ | |
450 | int (*writepages)(struct address_space *, struct writeback_control *); | |
451 | ||
4741c9fd | 452 | /* Set a page dirty. Return true if this dirtied it */ |
1da177e4 LT |
453 | int (*set_page_dirty)(struct page *page); |
454 | ||
455 | int (*readpages)(struct file *filp, struct address_space *mapping, | |
456 | struct list_head *pages, unsigned nr_pages); | |
457 | ||
458 | /* | |
459 | * ext3 requires that a successful prepare_write() call be followed | |
460 | * by a commit_write() call - they must be balanced | |
461 | */ | |
462 | int (*prepare_write)(struct file *, struct page *, unsigned, unsigned); | |
463 | int (*commit_write)(struct file *, struct page *, unsigned, unsigned); | |
afddba49 NP |
464 | |
465 | int (*write_begin)(struct file *, struct address_space *mapping, | |
466 | loff_t pos, unsigned len, unsigned flags, | |
467 | struct page **pagep, void **fsdata); | |
468 | int (*write_end)(struct file *, struct address_space *mapping, | |
469 | loff_t pos, unsigned len, unsigned copied, | |
470 | struct page *page, void *fsdata); | |
471 | ||
1da177e4 LT |
472 | /* Unfortunately this kludge is needed for FIBMAP. Don't use it */ |
473 | sector_t (*bmap)(struct address_space *, sector_t); | |
2ff28e22 | 474 | void (*invalidatepage) (struct page *, unsigned long); |
27496a8c | 475 | int (*releasepage) (struct page *, gfp_t); |
1da177e4 LT |
476 | ssize_t (*direct_IO)(int, struct kiocb *, const struct iovec *iov, |
477 | loff_t offset, unsigned long nr_segs); | |
70688e4d NP |
478 | int (*get_xip_mem)(struct address_space *, pgoff_t, int, |
479 | void **, unsigned long *); | |
e965f963 | 480 | /* migrate the contents of a page to the specified target */ |
2d1db3b1 CL |
481 | int (*migratepage) (struct address_space *, |
482 | struct page *, struct page *); | |
e3db7691 | 483 | int (*launder_page) (struct page *); |
1da177e4 LT |
484 | }; |
485 | ||
afddba49 NP |
486 | /* |
487 | * pagecache_write_begin/pagecache_write_end must be used by general code | |
488 | * to write into the pagecache. | |
489 | */ | |
490 | int pagecache_write_begin(struct file *, struct address_space *mapping, | |
491 | loff_t pos, unsigned len, unsigned flags, | |
492 | struct page **pagep, void **fsdata); | |
493 | ||
494 | int pagecache_write_end(struct file *, struct address_space *mapping, | |
495 | loff_t pos, unsigned len, unsigned copied, | |
496 | struct page *page, void *fsdata); | |
497 | ||
1da177e4 LT |
498 | struct backing_dev_info; |
499 | struct address_space { | |
500 | struct inode *host; /* owner: inode, block_device */ | |
501 | struct radix_tree_root page_tree; /* radix tree of all pages */ | |
502 | rwlock_t tree_lock; /* and rwlock protecting it */ | |
503 | unsigned int i_mmap_writable;/* count VM_SHARED mappings */ | |
504 | struct prio_tree_root i_mmap; /* tree of private and shared mappings */ | |
505 | struct list_head i_mmap_nonlinear;/*list VM_NONLINEAR mappings */ | |
506 | spinlock_t i_mmap_lock; /* protect tree, count, list */ | |
507 | unsigned int truncate_count; /* Cover race condition with truncate */ | |
508 | unsigned long nrpages; /* number of total pages */ | |
509 | pgoff_t writeback_index;/* writeback starts here */ | |
f5e54d6e | 510 | const struct address_space_operations *a_ops; /* methods */ |
1da177e4 LT |
511 | unsigned long flags; /* error bits/gfp mask */ |
512 | struct backing_dev_info *backing_dev_info; /* device readahead, etc */ | |
513 | spinlock_t private_lock; /* for use by the address_space */ | |
514 | struct list_head private_list; /* ditto */ | |
515 | struct address_space *assoc_mapping; /* ditto */ | |
516 | } __attribute__((aligned(sizeof(long)))); | |
517 | /* | |
518 | * On most architectures that alignment is already the case; but | |
519 | * must be enforced here for CRIS, to let the least signficant bit | |
520 | * of struct page's "mapping" pointer be used for PAGE_MAPPING_ANON. | |
521 | */ | |
522 | ||
523 | struct block_device { | |
524 | dev_t bd_dev; /* not a kdev_t - it's a search key */ | |
525 | struct inode * bd_inode; /* will die */ | |
526 | int bd_openers; | |
c039e313 | 527 | struct mutex bd_mutex; /* open/close mutex */ |
f73ca1b7 | 528 | struct semaphore bd_mount_sem; |
1da177e4 LT |
529 | struct list_head bd_inodes; |
530 | void * bd_holder; | |
531 | int bd_holders; | |
641dc636 JN |
532 | #ifdef CONFIG_SYSFS |
533 | struct list_head bd_holder_list; | |
534 | #endif | |
1da177e4 LT |
535 | struct block_device * bd_contains; |
536 | unsigned bd_block_size; | |
537 | struct hd_struct * bd_part; | |
538 | /* number of times partitions within this device have been opened. */ | |
539 | unsigned bd_part_count; | |
540 | int bd_invalidated; | |
541 | struct gendisk * bd_disk; | |
542 | struct list_head bd_list; | |
543 | struct backing_dev_info *bd_inode_backing_dev_info; | |
544 | /* | |
545 | * Private data. You must have bd_claim'ed the block_device | |
546 | * to use this. NOTE: bd_claim allows an owner to claim | |
547 | * the same device multiple times, the owner must take special | |
548 | * care to not mess up bd_private for that case. | |
549 | */ | |
550 | unsigned long bd_private; | |
551 | }; | |
552 | ||
553 | /* | |
554 | * Radix-tree tags, for tagging dirty and writeback pages within the pagecache | |
555 | * radix trees | |
556 | */ | |
557 | #define PAGECACHE_TAG_DIRTY 0 | |
558 | #define PAGECACHE_TAG_WRITEBACK 1 | |
559 | ||
560 | int mapping_tagged(struct address_space *mapping, int tag); | |
561 | ||
562 | /* | |
563 | * Might pages of this file be mapped into userspace? | |
564 | */ | |
565 | static inline int mapping_mapped(struct address_space *mapping) | |
566 | { | |
567 | return !prio_tree_empty(&mapping->i_mmap) || | |
568 | !list_empty(&mapping->i_mmap_nonlinear); | |
569 | } | |
570 | ||
571 | /* | |
572 | * Might pages of this file have been modified in userspace? | |
573 | * Note that i_mmap_writable counts all VM_SHARED vmas: do_mmap_pgoff | |
574 | * marks vma as VM_SHARED if it is shared, and the file was opened for | |
575 | * writing i.e. vma may be mprotected writable even if now readonly. | |
576 | */ | |
577 | static inline int mapping_writably_mapped(struct address_space *mapping) | |
578 | { | |
579 | return mapping->i_mmap_writable != 0; | |
580 | } | |
581 | ||
582 | /* | |
583 | * Use sequence counter to get consistent i_size on 32-bit processors. | |
584 | */ | |
585 | #if BITS_PER_LONG==32 && defined(CONFIG_SMP) | |
586 | #include <linux/seqlock.h> | |
587 | #define __NEED_I_SIZE_ORDERED | |
588 | #define i_size_ordered_init(inode) seqcount_init(&inode->i_size_seqcount) | |
589 | #else | |
590 | #define i_size_ordered_init(inode) do { } while (0) | |
591 | #endif | |
592 | ||
593 | struct inode { | |
594 | struct hlist_node i_hash; | |
595 | struct list_head i_list; | |
596 | struct list_head i_sb_list; | |
597 | struct list_head i_dentry; | |
598 | unsigned long i_ino; | |
599 | atomic_t i_count; | |
1da177e4 LT |
600 | unsigned int i_nlink; |
601 | uid_t i_uid; | |
602 | gid_t i_gid; | |
603 | dev_t i_rdev; | |
7a224228 | 604 | u64 i_version; |
1da177e4 | 605 | loff_t i_size; |
83b7b44e ED |
606 | #ifdef __NEED_I_SIZE_ORDERED |
607 | seqcount_t i_size_seqcount; | |
608 | #endif | |
1da177e4 LT |
609 | struct timespec i_atime; |
610 | struct timespec i_mtime; | |
611 | struct timespec i_ctime; | |
612 | unsigned int i_blkbits; | |
a0f62ac6 | 613 | blkcnt_t i_blocks; |
1da177e4 | 614 | unsigned short i_bytes; |
12d40e43 | 615 | umode_t i_mode; |
1da177e4 | 616 | spinlock_t i_lock; /* i_blocks, i_bytes, maybe i_size */ |
1b1dcc1b | 617 | struct mutex i_mutex; |
1da177e4 | 618 | struct rw_semaphore i_alloc_sem; |
c5ef1c42 | 619 | const struct inode_operations *i_op; |
99ac48f5 | 620 | const struct file_operations *i_fop; /* former ->i_op->default_file_ops */ |
1da177e4 LT |
621 | struct super_block *i_sb; |
622 | struct file_lock *i_flock; | |
623 | struct address_space *i_mapping; | |
624 | struct address_space i_data; | |
625 | #ifdef CONFIG_QUOTA | |
626 | struct dquot *i_dquot[MAXQUOTAS]; | |
627 | #endif | |
1da177e4 | 628 | struct list_head i_devices; |
4c154168 TT |
629 | union { |
630 | struct pipe_inode_info *i_pipe; | |
eaf796e7 | 631 | struct block_device *i_bdev; |
577c4eb0 | 632 | struct cdev *i_cdev; |
4c154168 | 633 | }; |
1da177e4 LT |
634 | int i_cindex; |
635 | ||
636 | __u32 i_generation; | |
637 | ||
638 | #ifdef CONFIG_DNOTIFY | |
639 | unsigned long i_dnotify_mask; /* Directory notify events */ | |
640 | struct dnotify_struct *i_dnotify; /* for directory notifications */ | |
641 | #endif | |
642 | ||
0eeca283 RL |
643 | #ifdef CONFIG_INOTIFY |
644 | struct list_head inotify_watches; /* watches on this inode */ | |
d4f9af9d | 645 | struct mutex inotify_mutex; /* protects the watches list */ |
0eeca283 RL |
646 | #endif |
647 | ||
1da177e4 LT |
648 | unsigned long i_state; |
649 | unsigned long dirtied_when; /* jiffies of first dirtying */ | |
650 | ||
651 | unsigned int i_flags; | |
652 | ||
653 | atomic_t i_writecount; | |
50462062 | 654 | #ifdef CONFIG_SECURITY |
1da177e4 | 655 | void *i_security; |
50462062 | 656 | #endif |
8e18e294 | 657 | void *i_private; /* fs or device private pointer */ |
1da177e4 LT |
658 | }; |
659 | ||
f2eace23 IM |
660 | /* |
661 | * inode->i_mutex nesting subclasses for the lock validator: | |
662 | * | |
663 | * 0: the object of the current VFS operation | |
664 | * 1: parent | |
665 | * 2: child/target | |
666 | * 3: quota file | |
667 | * | |
668 | * The locking order between these classes is | |
4df46240 | 669 | * parent -> child -> normal -> xattr -> quota |
f2eace23 IM |
670 | */ |
671 | enum inode_i_mutex_lock_class | |
672 | { | |
673 | I_MUTEX_NORMAL, | |
674 | I_MUTEX_PARENT, | |
675 | I_MUTEX_CHILD, | |
4df46240 | 676 | I_MUTEX_XATTR, |
f2eace23 IM |
677 | I_MUTEX_QUOTA |
678 | }; | |
679 | ||
62752ee1 MF |
680 | extern void inode_double_lock(struct inode *inode1, struct inode *inode2); |
681 | extern void inode_double_unlock(struct inode *inode1, struct inode *inode2); | |
682 | ||
1da177e4 LT |
683 | /* |
684 | * NOTE: in a 32bit arch with a preemptable kernel and | |
685 | * an UP compile the i_size_read/write must be atomic | |
686 | * with respect to the local cpu (unlike with preempt disabled), | |
687 | * but they don't need to be atomic with respect to other cpus like in | |
688 | * true SMP (so they need either to either locally disable irq around | |
689 | * the read or for example on x86 they can be still implemented as a | |
690 | * cmpxchg8b without the need of the lock prefix). For SMP compiles | |
691 | * and 64bit archs it makes no difference if preempt is enabled or not. | |
692 | */ | |
48ed214d | 693 | static inline loff_t i_size_read(const struct inode *inode) |
1da177e4 LT |
694 | { |
695 | #if BITS_PER_LONG==32 && defined(CONFIG_SMP) | |
696 | loff_t i_size; | |
697 | unsigned int seq; | |
698 | ||
699 | do { | |
700 | seq = read_seqcount_begin(&inode->i_size_seqcount); | |
701 | i_size = inode->i_size; | |
702 | } while (read_seqcount_retry(&inode->i_size_seqcount, seq)); | |
703 | return i_size; | |
704 | #elif BITS_PER_LONG==32 && defined(CONFIG_PREEMPT) | |
705 | loff_t i_size; | |
706 | ||
707 | preempt_disable(); | |
708 | i_size = inode->i_size; | |
709 | preempt_enable(); | |
710 | return i_size; | |
711 | #else | |
712 | return inode->i_size; | |
713 | #endif | |
714 | } | |
715 | ||
7762f5a0 MS |
716 | /* |
717 | * NOTE: unlike i_size_read(), i_size_write() does need locking around it | |
718 | * (normally i_mutex), otherwise on 32bit/SMP an update of i_size_seqcount | |
719 | * can be lost, resulting in subsequent i_size_read() calls spinning forever. | |
720 | */ | |
1da177e4 LT |
721 | static inline void i_size_write(struct inode *inode, loff_t i_size) |
722 | { | |
723 | #if BITS_PER_LONG==32 && defined(CONFIG_SMP) | |
724 | write_seqcount_begin(&inode->i_size_seqcount); | |
725 | inode->i_size = i_size; | |
726 | write_seqcount_end(&inode->i_size_seqcount); | |
727 | #elif BITS_PER_LONG==32 && defined(CONFIG_PREEMPT) | |
728 | preempt_disable(); | |
729 | inode->i_size = i_size; | |
730 | preempt_enable(); | |
731 | #else | |
732 | inode->i_size = i_size; | |
733 | #endif | |
734 | } | |
735 | ||
48ed214d | 736 | static inline unsigned iminor(const struct inode *inode) |
1da177e4 LT |
737 | { |
738 | return MINOR(inode->i_rdev); | |
739 | } | |
740 | ||
48ed214d | 741 | static inline unsigned imajor(const struct inode *inode) |
1da177e4 LT |
742 | { |
743 | return MAJOR(inode->i_rdev); | |
744 | } | |
745 | ||
746 | extern struct block_device *I_BDEV(struct inode *inode); | |
747 | ||
748 | struct fown_struct { | |
749 | rwlock_t lock; /* protects pid, uid, euid fields */ | |
609d7fa9 EB |
750 | struct pid *pid; /* pid or -pgrp where SIGIO should be sent */ |
751 | enum pid_type pid_type; /* Kind of process group SIGIO should be sent to */ | |
1da177e4 | 752 | uid_t uid, euid; /* uid/euid of process setting the owner */ |
1da177e4 LT |
753 | int signum; /* posix.1b rt signal to be delivered on IO */ |
754 | }; | |
755 | ||
756 | /* | |
757 | * Track a single file's readahead state | |
758 | */ | |
759 | struct file_ra_state { | |
937085aa FW |
760 | pgoff_t start; /* where readahead started */ |
761 | unsigned int size; /* # of readahead pages */ | |
762 | unsigned int async_size; /* do asynchronous readahead when | |
f9acc8c7 | 763 | there are only # of pages ahead */ |
5ce1110b | 764 | |
937085aa | 765 | unsigned int ra_pages; /* Maximum readahead window */ |
0bb7ba6b | 766 | int mmap_miss; /* Cache miss stat for mmap accesses */ |
f4e6b498 | 767 | loff_t prev_pos; /* Cache last read() position */ |
1da177e4 | 768 | }; |
1da177e4 | 769 | |
5ce1110b FW |
770 | /* |
771 | * Check if @index falls in the readahead windows. | |
772 | */ | |
773 | static inline int ra_has_index(struct file_ra_state *ra, pgoff_t index) | |
774 | { | |
f9acc8c7 FW |
775 | return (index >= ra->start && |
776 | index < ra->start + ra->size); | |
5ce1110b FW |
777 | } |
778 | ||
ad775f5a DH |
779 | #define FILE_MNT_WRITE_TAKEN 1 |
780 | #define FILE_MNT_WRITE_RELEASED 2 | |
781 | ||
1da177e4 | 782 | struct file { |
2f512016 ED |
783 | /* |
784 | * fu_list becomes invalid after file_free is called and queued via | |
785 | * fu_rcuhead for RCU freeing | |
786 | */ | |
787 | union { | |
788 | struct list_head fu_list; | |
789 | struct rcu_head fu_rcuhead; | |
790 | } f_u; | |
0f7fc9e4 JJS |
791 | struct path f_path; |
792 | #define f_dentry f_path.dentry | |
793 | #define f_vfsmnt f_path.mnt | |
99ac48f5 | 794 | const struct file_operations *f_op; |
1da177e4 LT |
795 | atomic_t f_count; |
796 | unsigned int f_flags; | |
797 | mode_t f_mode; | |
1da177e4 LT |
798 | loff_t f_pos; |
799 | struct fown_struct f_owner; | |
800 | unsigned int f_uid, f_gid; | |
801 | struct file_ra_state f_ra; | |
802 | ||
2b47c361 | 803 | u64 f_version; |
50462062 | 804 | #ifdef CONFIG_SECURITY |
1da177e4 | 805 | void *f_security; |
50462062 | 806 | #endif |
1da177e4 LT |
807 | /* needed for tty driver, and maybe others */ |
808 | void *private_data; | |
809 | ||
810 | #ifdef CONFIG_EPOLL | |
811 | /* Used by fs/eventpoll.c to link all the hooks to this file */ | |
812 | struct list_head f_ep_links; | |
813 | spinlock_t f_ep_lock; | |
814 | #endif /* #ifdef CONFIG_EPOLL */ | |
815 | struct address_space *f_mapping; | |
ad775f5a DH |
816 | #ifdef CONFIG_DEBUG_WRITECOUNT |
817 | unsigned long f_mnt_write_state; | |
818 | #endif | |
1da177e4 LT |
819 | }; |
820 | extern spinlock_t files_lock; | |
821 | #define file_list_lock() spin_lock(&files_lock); | |
822 | #define file_list_unlock() spin_unlock(&files_lock); | |
823 | ||
095975da | 824 | #define get_file(x) atomic_inc(&(x)->f_count) |
1da177e4 LT |
825 | #define file_count(x) atomic_read(&(x)->f_count) |
826 | ||
ad775f5a DH |
827 | #ifdef CONFIG_DEBUG_WRITECOUNT |
828 | static inline void file_take_write(struct file *f) | |
829 | { | |
830 | WARN_ON(f->f_mnt_write_state != 0); | |
831 | f->f_mnt_write_state = FILE_MNT_WRITE_TAKEN; | |
832 | } | |
833 | static inline void file_release_write(struct file *f) | |
834 | { | |
835 | f->f_mnt_write_state |= FILE_MNT_WRITE_RELEASED; | |
836 | } | |
837 | static inline void file_reset_write(struct file *f) | |
838 | { | |
839 | f->f_mnt_write_state = 0; | |
840 | } | |
841 | static inline void file_check_state(struct file *f) | |
842 | { | |
843 | /* | |
844 | * At this point, either both or neither of these bits | |
845 | * should be set. | |
846 | */ | |
847 | WARN_ON(f->f_mnt_write_state == FILE_MNT_WRITE_TAKEN); | |
848 | WARN_ON(f->f_mnt_write_state == FILE_MNT_WRITE_RELEASED); | |
849 | } | |
850 | static inline int file_check_writeable(struct file *f) | |
851 | { | |
852 | if (f->f_mnt_write_state == FILE_MNT_WRITE_TAKEN) | |
853 | return 0; | |
854 | printk(KERN_WARNING "writeable file with no " | |
855 | "mnt_want_write()\n"); | |
856 | WARN_ON(1); | |
857 | return -EINVAL; | |
858 | } | |
859 | #else /* !CONFIG_DEBUG_WRITECOUNT */ | |
860 | static inline void file_take_write(struct file *filp) {} | |
861 | static inline void file_release_write(struct file *filp) {} | |
862 | static inline void file_reset_write(struct file *filp) {} | |
863 | static inline void file_check_state(struct file *filp) {} | |
864 | static inline int file_check_writeable(struct file *filp) | |
865 | { | |
866 | return 0; | |
867 | } | |
868 | #endif /* CONFIG_DEBUG_WRITECOUNT */ | |
869 | ||
1da177e4 LT |
870 | #define MAX_NON_LFS ((1UL<<31) - 1) |
871 | ||
872 | /* Page cache limit. The filesystems should put that into their s_maxbytes | |
873 | limits, otherwise bad things can happen in VM. */ | |
874 | #if BITS_PER_LONG==32 | |
875 | #define MAX_LFS_FILESIZE (((u64)PAGE_CACHE_SIZE << (BITS_PER_LONG-1))-1) | |
876 | #elif BITS_PER_LONG==64 | |
877 | #define MAX_LFS_FILESIZE 0x7fffffffffffffffUL | |
878 | #endif | |
879 | ||
880 | #define FL_POSIX 1 | |
881 | #define FL_FLOCK 2 | |
882 | #define FL_ACCESS 8 /* not trying to lock, just looking */ | |
f475ae95 | 883 | #define FL_EXISTS 16 /* when unlocking, test for existence */ |
1da177e4 | 884 | #define FL_LEASE 32 /* lease held on this file */ |
75e1fcc0 | 885 | #define FL_CLOSE 64 /* unlock on close */ |
1da177e4 LT |
886 | #define FL_SLEEP 128 /* A blocking lock */ |
887 | ||
888 | /* | |
889 | * The POSIX file lock owner is determined by | |
890 | * the "struct files_struct" in the thread group | |
891 | * (or NULL for no owner - BSD locks). | |
892 | * | |
893 | * Lockd stuffs a "host" pointer into this. | |
894 | */ | |
895 | typedef struct files_struct *fl_owner_t; | |
896 | ||
897 | struct file_lock_operations { | |
1da177e4 LT |
898 | void (*fl_copy_lock)(struct file_lock *, struct file_lock *); |
899 | void (*fl_release_private)(struct file_lock *); | |
900 | }; | |
901 | ||
902 | struct lock_manager_operations { | |
903 | int (*fl_compare_owner)(struct file_lock *, struct file_lock *); | |
904 | void (*fl_notify)(struct file_lock *); /* unblock callback */ | |
2beb6614 | 905 | int (*fl_grant)(struct file_lock *, struct file_lock *, int); |
1da177e4 LT |
906 | void (*fl_copy_lock)(struct file_lock *, struct file_lock *); |
907 | void (*fl_release_private)(struct file_lock *); | |
908 | void (*fl_break)(struct file_lock *); | |
909 | int (*fl_mylease)(struct file_lock *, struct file_lock *); | |
910 | int (*fl_change)(struct file_lock **, int); | |
911 | }; | |
912 | ||
913 | /* that will die - we need it for nfs_lock_info */ | |
914 | #include <linux/nfs_fs_i.h> | |
915 | ||
916 | struct file_lock { | |
917 | struct file_lock *fl_next; /* singly linked list for this inode */ | |
918 | struct list_head fl_link; /* doubly linked list of all locks */ | |
919 | struct list_head fl_block; /* circular list of blocked processes */ | |
920 | fl_owner_t fl_owner; | |
921 | unsigned int fl_pid; | |
ab1f1611 | 922 | struct pid *fl_nspid; |
1da177e4 LT |
923 | wait_queue_head_t fl_wait; |
924 | struct file *fl_file; | |
925 | unsigned char fl_flags; | |
926 | unsigned char fl_type; | |
927 | loff_t fl_start; | |
928 | loff_t fl_end; | |
929 | ||
930 | struct fasync_struct * fl_fasync; /* for lease break notifications */ | |
931 | unsigned long fl_break_time; /* for nonblocking lease breaks */ | |
932 | ||
933 | struct file_lock_operations *fl_ops; /* Callbacks for filesystems */ | |
934 | struct lock_manager_operations *fl_lmops; /* Callbacks for lockmanagers */ | |
935 | union { | |
936 | struct nfs_lock_info nfs_fl; | |
8d0a8a9d | 937 | struct nfs4_lock_info nfs4_fl; |
e8d6c554 DH |
938 | struct { |
939 | struct list_head link; /* link in AFS vnode's pending_locks list */ | |
940 | int state; /* state of grant or error if -ve */ | |
941 | } afs; | |
1da177e4 LT |
942 | } fl_u; |
943 | }; | |
944 | ||
945 | /* The following constant reflects the upper bound of the file/locking space */ | |
946 | #ifndef OFFSET_MAX | |
947 | #define INT_LIMIT(x) (~((x)1 << (sizeof(x)*8 - 1))) | |
948 | #define OFFSET_MAX INT_LIMIT(loff_t) | |
949 | #define OFFT_OFFSET_MAX INT_LIMIT(off_t) | |
950 | #endif | |
951 | ||
1da177e4 LT |
952 | #include <linux/fcntl.h> |
953 | ||
954 | extern int fcntl_getlk(struct file *, struct flock __user *); | |
c293621b PS |
955 | extern int fcntl_setlk(unsigned int, struct file *, unsigned int, |
956 | struct flock __user *); | |
1da177e4 LT |
957 | |
958 | #if BITS_PER_LONG == 32 | |
959 | extern int fcntl_getlk64(struct file *, struct flock64 __user *); | |
c293621b PS |
960 | extern int fcntl_setlk64(unsigned int, struct file *, unsigned int, |
961 | struct flock64 __user *); | |
1da177e4 LT |
962 | #endif |
963 | ||
964 | extern void send_sigio(struct fown_struct *fown, int fd, int band); | |
965 | extern int fcntl_setlease(unsigned int fd, struct file *filp, long arg); | |
966 | extern int fcntl_getlease(struct file *filp); | |
967 | ||
f79e2abb | 968 | /* fs/sync.c */ |
5b04aa3a MF |
969 | extern int do_sync_mapping_range(struct address_space *mapping, loff_t offset, |
970 | loff_t endbyte, unsigned int flags); | |
f79e2abb | 971 | |
1da177e4 LT |
972 | /* fs/locks.c */ |
973 | extern void locks_init_lock(struct file_lock *); | |
974 | extern void locks_copy_lock(struct file_lock *, struct file_lock *); | |
1a747ee0 | 975 | extern void __locks_copy_lock(struct file_lock *, const struct file_lock *); |
1da177e4 LT |
976 | extern void locks_remove_posix(struct file *, fl_owner_t); |
977 | extern void locks_remove_flock(struct file *); | |
6d34ac19 | 978 | extern void posix_test_lock(struct file *, struct file_lock *); |
150b3934 | 979 | extern int posix_lock_file(struct file *, struct file_lock *, struct file_lock *); |
1da177e4 | 980 | extern int posix_lock_file_wait(struct file *, struct file_lock *); |
64a318ee | 981 | extern int posix_unblock_lock(struct file *, struct file_lock *); |
3ee17abd | 982 | extern int vfs_test_lock(struct file *, struct file_lock *); |
150b3934 | 983 | extern int vfs_lock_file(struct file *, unsigned int, struct file_lock *, struct file_lock *); |
9b9d2ab4 | 984 | extern int vfs_cancel_lock(struct file *filp, struct file_lock *fl); |
1da177e4 LT |
985 | extern int flock_lock_file_wait(struct file *filp, struct file_lock *fl); |
986 | extern int __break_lease(struct inode *inode, unsigned int flags); | |
987 | extern void lease_get_mtime(struct inode *, struct timespec *time); | |
0af1a450 | 988 | extern int generic_setlease(struct file *, long, struct file_lock **); |
a9933cea | 989 | extern int vfs_setlease(struct file *, long, struct file_lock **); |
1da177e4 LT |
990 | extern int lease_modify(struct file_lock **, int); |
991 | extern int lock_may_read(struct inode *, loff_t start, unsigned long count); | |
992 | extern int lock_may_write(struct inode *, loff_t start, unsigned long count); | |
7f8ada98 | 993 | extern struct seq_operations locks_seq_operations; |
1da177e4 LT |
994 | |
995 | struct fasync_struct { | |
996 | int magic; | |
997 | int fa_fd; | |
998 | struct fasync_struct *fa_next; /* singly linked list */ | |
999 | struct file *fa_file; | |
1000 | }; | |
1001 | ||
1002 | #define FASYNC_MAGIC 0x4601 | |
1003 | ||
1004 | /* SMP safe fasync helpers: */ | |
1005 | extern int fasync_helper(int, struct file *, int, struct fasync_struct **); | |
1006 | /* can be called from interrupts */ | |
1007 | extern void kill_fasync(struct fasync_struct **, int, int); | |
1008 | /* only for net: no internal synchronization */ | |
1009 | extern void __kill_fasync(struct fasync_struct *, int, int); | |
1010 | ||
609d7fa9 | 1011 | extern int __f_setown(struct file *filp, struct pid *, enum pid_type, int force); |
1da177e4 LT |
1012 | extern int f_setown(struct file *filp, unsigned long arg, int force); |
1013 | extern void f_delown(struct file *filp); | |
609d7fa9 | 1014 | extern pid_t f_getown(struct file *filp); |
1da177e4 LT |
1015 | extern int send_sigurg(struct fown_struct *fown); |
1016 | ||
1017 | /* | |
1018 | * Umount options | |
1019 | */ | |
1020 | ||
1021 | #define MNT_FORCE 0x00000001 /* Attempt to forcibily umount */ | |
1022 | #define MNT_DETACH 0x00000002 /* Just detach from the tree */ | |
1023 | #define MNT_EXPIRE 0x00000004 /* Mark for expiry */ | |
1024 | ||
1025 | extern struct list_head super_blocks; | |
1026 | extern spinlock_t sb_lock; | |
1027 | ||
1da177e4 LT |
1028 | #define S_BIAS (1<<30) |
1029 | struct super_block { | |
1030 | struct list_head s_list; /* Keep this first */ | |
1031 | dev_t s_dev; /* search index; _not_ kdev_t */ | |
1032 | unsigned long s_blocksize; | |
1da177e4 LT |
1033 | unsigned char s_blocksize_bits; |
1034 | unsigned char s_dirt; | |
1035 | unsigned long long s_maxbytes; /* Max file size */ | |
1036 | struct file_system_type *s_type; | |
ee9b6d61 | 1037 | const struct super_operations *s_op; |
1da177e4 LT |
1038 | struct dquot_operations *dq_op; |
1039 | struct quotactl_ops *s_qcop; | |
39655164 | 1040 | const struct export_operations *s_export_op; |
1da177e4 LT |
1041 | unsigned long s_flags; |
1042 | unsigned long s_magic; | |
1043 | struct dentry *s_root; | |
1044 | struct rw_semaphore s_umount; | |
7892f2f4 | 1045 | struct mutex s_lock; |
1da177e4 LT |
1046 | int s_count; |
1047 | int s_syncing; | |
1048 | int s_need_sync_fs; | |
1049 | atomic_t s_active; | |
50462062 | 1050 | #ifdef CONFIG_SECURITY |
1da177e4 | 1051 | void *s_security; |
50462062 | 1052 | #endif |
1da177e4 LT |
1053 | struct xattr_handler **s_xattr; |
1054 | ||
1055 | struct list_head s_inodes; /* all inodes */ | |
1056 | struct list_head s_dirty; /* dirty inodes */ | |
1057 | struct list_head s_io; /* parked for writeback */ | |
0e0f4fc2 | 1058 | struct list_head s_more_io; /* parked for more writeback */ |
1da177e4 LT |
1059 | struct hlist_head s_anon; /* anonymous dentries for (nfs) exporting */ |
1060 | struct list_head s_files; | |
1061 | ||
1062 | struct block_device *s_bdev; | |
acaebfd8 | 1063 | struct mtd_info *s_mtd; |
1da177e4 LT |
1064 | struct list_head s_instances; |
1065 | struct quota_info s_dquot; /* Diskquota specific options */ | |
1066 | ||
1067 | int s_frozen; | |
1068 | wait_queue_head_t s_wait_unfrozen; | |
1069 | ||
1070 | char s_id[32]; /* Informational name */ | |
1071 | ||
1072 | void *s_fs_info; /* Filesystem private info */ | |
1073 | ||
1074 | /* | |
1075 | * The next field is for VFS *only*. No filesystems have any business | |
1076 | * even looking at it. You had been warned. | |
1077 | */ | |
a11f3a05 | 1078 | struct mutex s_vfs_rename_mutex; /* Kludge */ |
1da177e4 | 1079 | |
8ba8e95e | 1080 | /* Granularity of c/m/atime in ns. |
1da177e4 LT |
1081 | Cannot be worse than a second */ |
1082 | u32 s_time_gran; | |
79c0b2df MS |
1083 | |
1084 | /* | |
1085 | * Filesystem subtype. If non-empty the filesystem type field | |
1086 | * in /proc/mounts will be "type.subtype" | |
1087 | */ | |
1088 | char *s_subtype; | |
b3b304a2 MS |
1089 | |
1090 | /* | |
1091 | * Saved mount options for lazy filesystems using | |
1092 | * generic_show_options() | |
1093 | */ | |
1094 | char *s_options; | |
1da177e4 LT |
1095 | }; |
1096 | ||
1097 | extern struct timespec current_fs_time(struct super_block *sb); | |
1098 | ||
1099 | /* | |
1100 | * Snapshotting support. | |
1101 | */ | |
1102 | enum { | |
1103 | SB_UNFROZEN = 0, | |
1104 | SB_FREEZE_WRITE = 1, | |
1105 | SB_FREEZE_TRANS = 2, | |
1106 | }; | |
1107 | ||
1108 | #define vfs_check_frozen(sb, level) \ | |
1109 | wait_event((sb)->s_wait_unfrozen, ((sb)->s_frozen < (level))) | |
1110 | ||
914e2637 AV |
1111 | #define get_fs_excl() atomic_inc(¤t->fs_excl) |
1112 | #define put_fs_excl() atomic_dec(¤t->fs_excl) | |
1113 | #define has_fs_excl() atomic_read(¤t->fs_excl) | |
22e2c507 | 1114 | |
3bd858ab SS |
1115 | #define is_owner_or_cap(inode) \ |
1116 | ((current->fsuid == (inode)->i_uid) || capable(CAP_FOWNER)) | |
1117 | ||
914e2637 AV |
1118 | /* not quite ready to be deprecated, but... */ |
1119 | extern void lock_super(struct super_block *); | |
1120 | extern void unlock_super(struct super_block *); | |
1da177e4 LT |
1121 | |
1122 | /* | |
1123 | * VFS helper functions.. | |
1124 | */ | |
e4543edd | 1125 | extern int vfs_permission(struct nameidata *, int); |
1da177e4 LT |
1126 | extern int vfs_create(struct inode *, struct dentry *, int, struct nameidata *); |
1127 | extern int vfs_mkdir(struct inode *, struct dentry *, int); | |
1128 | extern int vfs_mknod(struct inode *, struct dentry *, int, dev_t); | |
1129 | extern int vfs_symlink(struct inode *, struct dentry *, const char *, int); | |
1130 | extern int vfs_link(struct dentry *, struct inode *, struct dentry *); | |
1131 | extern int vfs_rmdir(struct inode *, struct dentry *); | |
1132 | extern int vfs_unlink(struct inode *, struct dentry *); | |
1133 | extern int vfs_rename(struct inode *, struct dentry *, struct inode *, struct dentry *); | |
1134 | ||
1135 | /* | |
1136 | * VFS dentry helper functions. | |
1137 | */ | |
1138 | extern void dentry_unhash(struct dentry *dentry); | |
1139 | ||
8c744fb8 CH |
1140 | /* |
1141 | * VFS file helper functions. | |
1142 | */ | |
1143 | extern int file_permission(struct file *, int); | |
1144 | ||
1da177e4 LT |
1145 | /* |
1146 | * File types | |
1147 | * | |
1148 | * NOTE! These match bits 12..15 of stat.st_mode | |
1149 | * (ie "(i_mode >> 12) & 15"). | |
1150 | */ | |
1151 | #define DT_UNKNOWN 0 | |
1152 | #define DT_FIFO 1 | |
1153 | #define DT_CHR 2 | |
1154 | #define DT_DIR 4 | |
1155 | #define DT_BLK 6 | |
1156 | #define DT_REG 8 | |
1157 | #define DT_LNK 10 | |
1158 | #define DT_SOCK 12 | |
1159 | #define DT_WHT 14 | |
1160 | ||
1161 | #define OSYNC_METADATA (1<<0) | |
1162 | #define OSYNC_DATA (1<<1) | |
1163 | #define OSYNC_INODE (1<<2) | |
1164 | int generic_osync_inode(struct inode *, struct address_space *, int); | |
1165 | ||
1166 | /* | |
1167 | * This is the "filldir" function type, used by readdir() to let | |
1168 | * the kernel specify what kind of dirent layout it wants to have. | |
1169 | * This allows the kernel to read directories into kernel space or | |
1170 | * to have different dirent layouts depending on the binary type. | |
1171 | */ | |
afefdbb2 | 1172 | typedef int (*filldir_t)(void *, const char *, int, loff_t, u64, unsigned); |
1da177e4 LT |
1173 | |
1174 | struct block_device_operations { | |
1175 | int (*open) (struct inode *, struct file *); | |
1176 | int (*release) (struct inode *, struct file *); | |
1177 | int (*ioctl) (struct inode *, struct file *, unsigned, unsigned long); | |
bb93e3a5 | 1178 | long (*unlocked_ioctl) (struct file *, unsigned, unsigned long); |
1da177e4 | 1179 | long (*compat_ioctl) (struct file *, unsigned, unsigned long); |
30afcb4b JH |
1180 | int (*direct_access) (struct block_device *, sector_t, |
1181 | void **, unsigned long *); | |
1da177e4 LT |
1182 | int (*media_changed) (struct gendisk *); |
1183 | int (*revalidate_disk) (struct gendisk *); | |
a885c8c4 | 1184 | int (*getgeo)(struct block_device *, struct hd_geometry *); |
1da177e4 LT |
1185 | struct module *owner; |
1186 | }; | |
1187 | ||
1188 | /* | |
d96e6e71 | 1189 | * "descriptor" for what we're up to with a read. |
1da177e4 LT |
1190 | * This allows us to use the same read code yet |
1191 | * have multiple different users of the data that | |
1192 | * we read from a file. | |
1193 | * | |
1194 | * The simplest case just copies the data to user | |
1195 | * mode. | |
1196 | */ | |
1197 | typedef struct { | |
1198 | size_t written; | |
1199 | size_t count; | |
1200 | union { | |
1201 | char __user * buf; | |
1202 | void *data; | |
1203 | } arg; | |
1204 | int error; | |
1205 | } read_descriptor_t; | |
1206 | ||
1207 | typedef int (*read_actor_t)(read_descriptor_t *, struct page *, unsigned long, unsigned long); | |
1208 | ||
1209 | /* These macros are for out of kernel modules to test that | |
1210 | * the kernel supports the unlocked_ioctl and compat_ioctl | |
1211 | * fields in struct file_operations. */ | |
1212 | #define HAVE_COMPAT_IOCTL 1 | |
1213 | #define HAVE_UNLOCKED_IOCTL 1 | |
1214 | ||
1215 | /* | |
1216 | * NOTE: | |
1217 | * read, write, poll, fsync, readv, writev, unlocked_ioctl and compat_ioctl | |
1218 | * can be called without the big kernel lock held in all filesystems. | |
1219 | */ | |
1220 | struct file_operations { | |
1221 | struct module *owner; | |
1222 | loff_t (*llseek) (struct file *, loff_t, int); | |
1223 | ssize_t (*read) (struct file *, char __user *, size_t, loff_t *); | |
1da177e4 | 1224 | ssize_t (*write) (struct file *, const char __user *, size_t, loff_t *); |
027445c3 BP |
1225 | ssize_t (*aio_read) (struct kiocb *, const struct iovec *, unsigned long, loff_t); |
1226 | ssize_t (*aio_write) (struct kiocb *, const struct iovec *, unsigned long, loff_t); | |
1da177e4 LT |
1227 | int (*readdir) (struct file *, void *, filldir_t); |
1228 | unsigned int (*poll) (struct file *, struct poll_table_struct *); | |
1229 | int (*ioctl) (struct inode *, struct file *, unsigned int, unsigned long); | |
1230 | long (*unlocked_ioctl) (struct file *, unsigned int, unsigned long); | |
1231 | long (*compat_ioctl) (struct file *, unsigned int, unsigned long); | |
1232 | int (*mmap) (struct file *, struct vm_area_struct *); | |
1233 | int (*open) (struct inode *, struct file *); | |
75e1fcc0 | 1234 | int (*flush) (struct file *, fl_owner_t id); |
1da177e4 LT |
1235 | int (*release) (struct inode *, struct file *); |
1236 | int (*fsync) (struct file *, struct dentry *, int datasync); | |
1237 | int (*aio_fsync) (struct kiocb *, int datasync); | |
1238 | int (*fasync) (int, struct file *, int); | |
1239 | int (*lock) (struct file *, int, struct file_lock *); | |
1da177e4 LT |
1240 | ssize_t (*sendpage) (struct file *, struct page *, int, size_t, loff_t *, int); |
1241 | unsigned long (*get_unmapped_area)(struct file *, unsigned long, unsigned long, unsigned long, unsigned long); | |
1242 | int (*check_flags)(int); | |
1243 | int (*dir_notify)(struct file *filp, unsigned long arg); | |
1244 | int (*flock) (struct file *, int, struct file_lock *); | |
cbb7e577 JA |
1245 | ssize_t (*splice_write)(struct pipe_inode_info *, struct file *, loff_t *, size_t, unsigned int); |
1246 | ssize_t (*splice_read)(struct file *, loff_t *, struct pipe_inode_info *, size_t, unsigned int); | |
f9ffed26 | 1247 | int (*setlease)(struct file *, long, struct file_lock **); |
1da177e4 LT |
1248 | }; |
1249 | ||
1250 | struct inode_operations { | |
1251 | int (*create) (struct inode *,struct dentry *,int, struct nameidata *); | |
1252 | struct dentry * (*lookup) (struct inode *,struct dentry *, struct nameidata *); | |
1253 | int (*link) (struct dentry *,struct inode *,struct dentry *); | |
1254 | int (*unlink) (struct inode *,struct dentry *); | |
1255 | int (*symlink) (struct inode *,struct dentry *,const char *); | |
1256 | int (*mkdir) (struct inode *,struct dentry *,int); | |
1257 | int (*rmdir) (struct inode *,struct dentry *); | |
1258 | int (*mknod) (struct inode *,struct dentry *,int,dev_t); | |
1259 | int (*rename) (struct inode *, struct dentry *, | |
1260 | struct inode *, struct dentry *); | |
1261 | int (*readlink) (struct dentry *, char __user *,int); | |
cc314eef LT |
1262 | void * (*follow_link) (struct dentry *, struct nameidata *); |
1263 | void (*put_link) (struct dentry *, struct nameidata *, void *); | |
1da177e4 LT |
1264 | void (*truncate) (struct inode *); |
1265 | int (*permission) (struct inode *, int, struct nameidata *); | |
1266 | int (*setattr) (struct dentry *, struct iattr *); | |
1267 | int (*getattr) (struct vfsmount *mnt, struct dentry *, struct kstat *); | |
1268 | int (*setxattr) (struct dentry *, const char *,const void *,size_t,int); | |
1269 | ssize_t (*getxattr) (struct dentry *, const char *, void *, size_t); | |
1270 | ssize_t (*listxattr) (struct dentry *, char *, size_t); | |
1271 | int (*removexattr) (struct dentry *, const char *); | |
f6b3ec23 | 1272 | void (*truncate_range)(struct inode *, loff_t, loff_t); |
97ac7350 AA |
1273 | long (*fallocate)(struct inode *inode, int mode, loff_t offset, |
1274 | loff_t len); | |
1da177e4 LT |
1275 | }; |
1276 | ||
1277 | struct seq_file; | |
1278 | ||
eed4e51f BP |
1279 | ssize_t rw_copy_check_uvector(int type, const struct iovec __user * uvector, |
1280 | unsigned long nr_segs, unsigned long fast_segs, | |
1281 | struct iovec *fast_pointer, | |
1282 | struct iovec **ret_pointer); | |
1283 | ||
1da177e4 LT |
1284 | extern ssize_t vfs_read(struct file *, char __user *, size_t, loff_t *); |
1285 | extern ssize_t vfs_write(struct file *, const char __user *, size_t, loff_t *); | |
1286 | extern ssize_t vfs_readv(struct file *, const struct iovec __user *, | |
1287 | unsigned long, loff_t *); | |
1288 | extern ssize_t vfs_writev(struct file *, const struct iovec __user *, | |
1289 | unsigned long, loff_t *); | |
1290 | ||
1da177e4 LT |
1291 | struct super_operations { |
1292 | struct inode *(*alloc_inode)(struct super_block *sb); | |
1293 | void (*destroy_inode)(struct inode *); | |
1294 | ||
1da177e4 LT |
1295 | void (*dirty_inode) (struct inode *); |
1296 | int (*write_inode) (struct inode *, int); | |
1da177e4 LT |
1297 | void (*drop_inode) (struct inode *); |
1298 | void (*delete_inode) (struct inode *); | |
1299 | void (*put_super) (struct super_block *); | |
1300 | void (*write_super) (struct super_block *); | |
1301 | int (*sync_fs)(struct super_block *sb, int wait); | |
1302 | void (*write_super_lockfs) (struct super_block *); | |
1303 | void (*unlockfs) (struct super_block *); | |
726c3342 | 1304 | int (*statfs) (struct dentry *, struct kstatfs *); |
1da177e4 LT |
1305 | int (*remount_fs) (struct super_block *, int *, char *); |
1306 | void (*clear_inode) (struct inode *); | |
42faad99 | 1307 | void (*umount_begin) (struct super_block *); |
1da177e4 LT |
1308 | |
1309 | int (*show_options)(struct seq_file *, struct vfsmount *); | |
b4629fe2 | 1310 | int (*show_stats)(struct seq_file *, struct vfsmount *); |
0e51a720 | 1311 | #ifdef CONFIG_QUOTA |
1da177e4 LT |
1312 | ssize_t (*quota_read)(struct super_block *, int, char *, size_t, loff_t); |
1313 | ssize_t (*quota_write)(struct super_block *, int, const char *, size_t, loff_t); | |
0e51a720 | 1314 | #endif |
1da177e4 LT |
1315 | }; |
1316 | ||
1c0eeaf5 JE |
1317 | /* |
1318 | * Inode state bits. Protected by inode_lock. | |
1319 | * | |
1320 | * Three bits determine the dirty state of the inode, I_DIRTY_SYNC, | |
1321 | * I_DIRTY_DATASYNC and I_DIRTY_PAGES. | |
1322 | * | |
1323 | * Four bits define the lifetime of an inode. Initially, inodes are I_NEW, | |
1324 | * until that flag is cleared. I_WILL_FREE, I_FREEING and I_CLEAR are set at | |
1325 | * various stages of removing an inode. | |
1326 | * | |
1327 | * Two bits are used for locking and completion notification, I_LOCK and I_SYNC. | |
1328 | * | |
e7ca2d41 JE |
1329 | * I_DIRTY_SYNC Inode is dirty, but doesn't have to be written on |
1330 | * fdatasync(). i_atime is the usual cause. | |
000cb48e JK |
1331 | * I_DIRTY_DATASYNC Data-related inode changes pending. We keep track of |
1332 | * these changes separately from I_DIRTY_SYNC so that we | |
1333 | * don't have to write inode on fdatasync() when only | |
1334 | * mtime has changed in it. | |
1c0eeaf5 JE |
1335 | * I_DIRTY_PAGES Inode has dirty pages. Inode itself may be clean. |
1336 | * I_NEW get_new_inode() sets i_state to I_LOCK|I_NEW. Both | |
1337 | * are cleared by unlock_new_inode(), called from iget(). | |
1338 | * I_WILL_FREE Must be set when calling write_inode_now() if i_count | |
1339 | * is zero. I_FREEING must be set when I_WILL_FREE is | |
1340 | * cleared. | |
1341 | * I_FREEING Set when inode is about to be freed but still has dirty | |
1342 | * pages or buffers attached or the inode itself is still | |
1343 | * dirty. | |
1344 | * I_CLEAR Set by clear_inode(). In this state the inode is clean | |
1345 | * and can be destroyed. | |
1346 | * | |
1347 | * Inodes that are I_WILL_FREE, I_FREEING or I_CLEAR are | |
1348 | * prohibited for many purposes. iget() must wait for | |
1349 | * the inode to be completely released, then create it | |
1350 | * anew. Other functions will just ignore such inodes, | |
1351 | * if appropriate. I_LOCK is used for waiting. | |
1352 | * | |
1353 | * I_LOCK Serves as both a mutex and completion notification. | |
1354 | * New inodes set I_LOCK. If two processes both create | |
1355 | * the same inode, one of them will release its inode and | |
1356 | * wait for I_LOCK to be released before returning. | |
1357 | * Inodes in I_WILL_FREE, I_FREEING or I_CLEAR state can | |
1358 | * also cause waiting on I_LOCK, without I_LOCK actually | |
1359 | * being set. find_inode() uses this to prevent returning | |
1360 | * nearly-dead inodes. | |
1361 | * I_SYNC Similar to I_LOCK, but limited in scope to writeback | |
fd3f8984 | 1362 | * of inode dirty data. Having a separate lock for this |
1c0eeaf5 JE |
1363 | * purpose reduces latency and prevents some filesystem- |
1364 | * specific deadlocks. | |
1365 | * | |
1c0eeaf5 JE |
1366 | * Q: What is the difference between I_WILL_FREE and I_FREEING? |
1367 | * Q: igrab() only checks on (I_FREEING|I_WILL_FREE). Should it also check on | |
1368 | * I_CLEAR? If not, why? | |
1369 | */ | |
1370 | #define I_DIRTY_SYNC 1 | |
1371 | #define I_DIRTY_DATASYNC 2 | |
1372 | #define I_DIRTY_PAGES 4 | |
1373 | #define I_NEW 8 | |
1374 | #define I_WILL_FREE 16 | |
1375 | #define I_FREEING 32 | |
1376 | #define I_CLEAR 64 | |
1377 | #define __I_LOCK 7 | |
1da177e4 | 1378 | #define I_LOCK (1 << __I_LOCK) |
1c0eeaf5 JE |
1379 | #define __I_SYNC 8 |
1380 | #define I_SYNC (1 << __I_SYNC) | |
1da177e4 LT |
1381 | |
1382 | #define I_DIRTY (I_DIRTY_SYNC | I_DIRTY_DATASYNC | I_DIRTY_PAGES) | |
1383 | ||
1384 | extern void __mark_inode_dirty(struct inode *, int); | |
1385 | static inline void mark_inode_dirty(struct inode *inode) | |
1386 | { | |
1387 | __mark_inode_dirty(inode, I_DIRTY); | |
1388 | } | |
1389 | ||
1390 | static inline void mark_inode_dirty_sync(struct inode *inode) | |
1391 | { | |
1392 | __mark_inode_dirty(inode, I_DIRTY_SYNC); | |
1393 | } | |
1394 | ||
71c42157 DH |
1395 | /** |
1396 | * inc_nlink - directly increment an inode's link count | |
1397 | * @inode: inode | |
1398 | * | |
1399 | * This is a low-level filesystem helper to replace any | |
1400 | * direct filesystem manipulation of i_nlink. Currently, | |
1401 | * it is only here for parity with dec_nlink(). | |
1402 | */ | |
d8c76e6f | 1403 | static inline void inc_nlink(struct inode *inode) |
66d21730 AD |
1404 | { |
1405 | inode->i_nlink++; | |
d8c76e6f DH |
1406 | } |
1407 | ||
1408 | static inline void inode_inc_link_count(struct inode *inode) | |
1409 | { | |
1410 | inc_nlink(inode); | |
66d21730 AD |
1411 | mark_inode_dirty(inode); |
1412 | } | |
1413 | ||
71c42157 DH |
1414 | /** |
1415 | * drop_nlink - directly drop an inode's link count | |
1416 | * @inode: inode | |
1417 | * | |
1418 | * This is a low-level filesystem helper to replace any | |
1419 | * direct filesystem manipulation of i_nlink. In cases | |
1420 | * where we are attempting to track writes to the | |
1421 | * filesystem, a decrement to zero means an imminent | |
1422 | * write when the file is truncated and actually unlinked | |
1423 | * on the filesystem. | |
1424 | */ | |
9a53c3a7 | 1425 | static inline void drop_nlink(struct inode *inode) |
66d21730 AD |
1426 | { |
1427 | inode->i_nlink--; | |
9a53c3a7 DH |
1428 | } |
1429 | ||
71c42157 DH |
1430 | /** |
1431 | * clear_nlink - directly zero an inode's link count | |
1432 | * @inode: inode | |
1433 | * | |
1434 | * This is a low-level filesystem helper to replace any | |
1435 | * direct filesystem manipulation of i_nlink. See | |
1436 | * drop_nlink() for why we care about i_nlink hitting zero. | |
1437 | */ | |
ce71ec36 DH |
1438 | static inline void clear_nlink(struct inode *inode) |
1439 | { | |
1440 | inode->i_nlink = 0; | |
1441 | } | |
1442 | ||
9a53c3a7 DH |
1443 | static inline void inode_dec_link_count(struct inode *inode) |
1444 | { | |
1445 | drop_nlink(inode); | |
66d21730 AD |
1446 | mark_inode_dirty(inode); |
1447 | } | |
1448 | ||
25ec56b5 JNC |
1449 | /** |
1450 | * inode_inc_iversion - increments i_version | |
1451 | * @inode: inode that need to be updated | |
1452 | * | |
1453 | * Every time the inode is modified, the i_version field will be incremented. | |
1454 | * The filesystem has to be mounted with i_version flag | |
1455 | */ | |
1456 | ||
1457 | static inline void inode_inc_iversion(struct inode *inode) | |
1458 | { | |
1459 | spin_lock(&inode->i_lock); | |
1460 | inode->i_version++; | |
1461 | spin_unlock(&inode->i_lock); | |
1462 | } | |
1463 | ||
869243a0 | 1464 | extern void touch_atime(struct vfsmount *mnt, struct dentry *dentry); |
1da177e4 LT |
1465 | static inline void file_accessed(struct file *file) |
1466 | { | |
1467 | if (!(file->f_flags & O_NOATIME)) | |
0f7fc9e4 | 1468 | touch_atime(file->f_path.mnt, file->f_path.dentry); |
1da177e4 LT |
1469 | } |
1470 | ||
1471 | int sync_inode(struct inode *inode, struct writeback_control *wbc); | |
1472 | ||
1da177e4 LT |
1473 | struct file_system_type { |
1474 | const char *name; | |
1475 | int fs_flags; | |
454e2398 DH |
1476 | int (*get_sb) (struct file_system_type *, int, |
1477 | const char *, void *, struct vfsmount *); | |
1da177e4 LT |
1478 | void (*kill_sb) (struct super_block *); |
1479 | struct module *owner; | |
1480 | struct file_system_type * next; | |
1481 | struct list_head fs_supers; | |
d475fd42 | 1482 | |
cf516249 | 1483 | struct lock_class_key s_lock_key; |
897c6ff9 | 1484 | struct lock_class_key s_umount_key; |
d475fd42 PZ |
1485 | |
1486 | struct lock_class_key i_lock_key; | |
1487 | struct lock_class_key i_mutex_key; | |
14358e6d | 1488 | struct lock_class_key i_mutex_dir_key; |
d475fd42 | 1489 | struct lock_class_key i_alloc_sem_key; |
1da177e4 LT |
1490 | }; |
1491 | ||
454e2398 | 1492 | extern int get_sb_bdev(struct file_system_type *fs_type, |
1da177e4 | 1493 | int flags, const char *dev_name, void *data, |
454e2398 DH |
1494 | int (*fill_super)(struct super_block *, void *, int), |
1495 | struct vfsmount *mnt); | |
1496 | extern int get_sb_single(struct file_system_type *fs_type, | |
1da177e4 | 1497 | int flags, void *data, |
454e2398 DH |
1498 | int (*fill_super)(struct super_block *, void *, int), |
1499 | struct vfsmount *mnt); | |
1500 | extern int get_sb_nodev(struct file_system_type *fs_type, | |
1da177e4 | 1501 | int flags, void *data, |
454e2398 DH |
1502 | int (*fill_super)(struct super_block *, void *, int), |
1503 | struct vfsmount *mnt); | |
1da177e4 LT |
1504 | void generic_shutdown_super(struct super_block *sb); |
1505 | void kill_block_super(struct super_block *sb); | |
1506 | void kill_anon_super(struct super_block *sb); | |
1507 | void kill_litter_super(struct super_block *sb); | |
1508 | void deactivate_super(struct super_block *sb); | |
1509 | int set_anon_super(struct super_block *s, void *data); | |
1510 | struct super_block *sget(struct file_system_type *type, | |
1511 | int (*test)(struct super_block *,void *), | |
1512 | int (*set)(struct super_block *,void *), | |
1513 | void *data); | |
454e2398 | 1514 | extern int get_sb_pseudo(struct file_system_type *, char *, |
ee9b6d61 | 1515 | const struct super_operations *ops, unsigned long, |
454e2398 DH |
1516 | struct vfsmount *mnt); |
1517 | extern int simple_set_mnt(struct vfsmount *mnt, struct super_block *sb); | |
1da177e4 LT |
1518 | int __put_super_and_need_restart(struct super_block *sb); |
1519 | void unnamed_dev_init(void); | |
1520 | ||
1521 | /* Alas, no aliases. Too much hassle with bringing module.h everywhere */ | |
1522 | #define fops_get(fops) \ | |
1523 | (((fops) && try_module_get((fops)->owner) ? (fops) : NULL)) | |
1524 | #define fops_put(fops) \ | |
1525 | do { if (fops) module_put((fops)->owner); } while(0) | |
1526 | ||
1527 | extern int register_filesystem(struct file_system_type *); | |
1528 | extern int unregister_filesystem(struct file_system_type *); | |
8bf9725c PE |
1529 | extern struct vfsmount *kern_mount_data(struct file_system_type *, void *data); |
1530 | #define kern_mount(type) kern_mount_data(type, NULL) | |
1da177e4 LT |
1531 | extern int may_umount_tree(struct vfsmount *); |
1532 | extern int may_umount(struct vfsmount *); | |
1533 | extern long do_mount(char *, char *, char *, unsigned long, void *); | |
8aec0809 AV |
1534 | extern struct vfsmount *collect_mounts(struct vfsmount *, struct dentry *); |
1535 | extern void drop_collected_mounts(struct vfsmount *); | |
1da177e4 | 1536 | |
726c3342 | 1537 | extern int vfs_statfs(struct dentry *, struct kstatfs *); |
1da177e4 | 1538 | |
f87fd4c2 | 1539 | /* /sys/fs */ |
00d26666 | 1540 | extern struct kobject *fs_kobj; |
f87fd4c2 | 1541 | |
1da177e4 LT |
1542 | #define FLOCK_VERIFY_READ 1 |
1543 | #define FLOCK_VERIFY_WRITE 2 | |
1544 | ||
1545 | extern int locks_mandatory_locked(struct inode *); | |
1546 | extern int locks_mandatory_area(int, struct inode *, struct file *, loff_t, size_t); | |
1547 | ||
1548 | /* | |
1549 | * Candidates for mandatory locking have the setgid bit set | |
1550 | * but no group execute bit - an otherwise meaningless combination. | |
1551 | */ | |
a16877ca PE |
1552 | |
1553 | static inline int __mandatory_lock(struct inode *ino) | |
1554 | { | |
1555 | return (ino->i_mode & (S_ISGID | S_IXGRP)) == S_ISGID; | |
1556 | } | |
1557 | ||
1558 | /* | |
1559 | * ... and these candidates should be on MS_MANDLOCK mounted fs, | |
1560 | * otherwise these will be advisory locks | |
1561 | */ | |
1562 | ||
1563 | static inline int mandatory_lock(struct inode *ino) | |
1564 | { | |
1565 | return IS_MANDLOCK(ino) && __mandatory_lock(ino); | |
1566 | } | |
1da177e4 LT |
1567 | |
1568 | static inline int locks_verify_locked(struct inode *inode) | |
1569 | { | |
a16877ca | 1570 | if (mandatory_lock(inode)) |
1da177e4 LT |
1571 | return locks_mandatory_locked(inode); |
1572 | return 0; | |
1573 | } | |
1574 | ||
1575 | extern int rw_verify_area(int, struct file *, loff_t *, size_t); | |
1576 | ||
1577 | static inline int locks_verify_truncate(struct inode *inode, | |
1578 | struct file *filp, | |
1579 | loff_t size) | |
1580 | { | |
a16877ca | 1581 | if (inode->i_flock && mandatory_lock(inode)) |
1da177e4 LT |
1582 | return locks_mandatory_area( |
1583 | FLOCK_VERIFY_WRITE, inode, filp, | |
1584 | size < inode->i_size ? size : inode->i_size, | |
1585 | (size < inode->i_size ? inode->i_size - size | |
1586 | : size - inode->i_size) | |
1587 | ); | |
1588 | return 0; | |
1589 | } | |
1590 | ||
1591 | static inline int break_lease(struct inode *inode, unsigned int mode) | |
1592 | { | |
1593 | if (inode->i_flock) | |
1594 | return __break_lease(inode, mode); | |
1595 | return 0; | |
1596 | } | |
1597 | ||
1598 | /* fs/open.c */ | |
1599 | ||
4a30131e N |
1600 | extern int do_truncate(struct dentry *, loff_t start, unsigned int time_attrs, |
1601 | struct file *filp); | |
8e8a1407 | 1602 | extern long do_sys_open(int dfd, const char __user *filename, int flags, |
5590ff0d | 1603 | int mode); |
1da177e4 LT |
1604 | extern struct file *filp_open(const char *, int, int); |
1605 | extern struct file * dentry_open(struct dentry *, struct vfsmount *, int); | |
1606 | extern int filp_close(struct file *, fl_owner_t id); | |
1607 | extern char * getname(const char __user *); | |
1608 | ||
1609 | /* fs/dcache.c */ | |
1610 | extern void __init vfs_caches_init_early(void); | |
1611 | extern void __init vfs_caches_init(unsigned long); | |
1612 | ||
b86c089b CL |
1613 | extern struct kmem_cache *names_cachep; |
1614 | ||
e94b1766 | 1615 | #define __getname() kmem_cache_alloc(names_cachep, GFP_KERNEL) |
1da177e4 LT |
1616 | #define __putname(name) kmem_cache_free(names_cachep, (void *)(name)) |
1617 | #ifndef CONFIG_AUDITSYSCALL | |
1618 | #define putname(name) __putname(name) | |
1619 | #else | |
1620 | extern void putname(const char *name); | |
1621 | #endif | |
1622 | ||
9361401e | 1623 | #ifdef CONFIG_BLOCK |
1da177e4 | 1624 | extern int register_blkdev(unsigned int, const char *); |
f4480240 | 1625 | extern void unregister_blkdev(unsigned int, const char *); |
1da177e4 LT |
1626 | extern struct block_device *bdget(dev_t); |
1627 | extern void bd_set_size(struct block_device *, loff_t size); | |
1628 | extern void bd_forget(struct inode *inode); | |
1629 | extern void bdput(struct block_device *); | |
1630 | extern struct block_device *open_by_devnum(dev_t, unsigned); | |
9361401e DH |
1631 | #else |
1632 | static inline void bd_forget(struct inode *inode) {} | |
1633 | #endif | |
1634 | extern const struct file_operations def_blk_fops; | |
4b6f5d20 AV |
1635 | extern const struct file_operations def_chr_fops; |
1636 | extern const struct file_operations bad_sock_fops; | |
1637 | extern const struct file_operations def_fifo_fops; | |
9361401e | 1638 | #ifdef CONFIG_BLOCK |
1da177e4 LT |
1639 | extern int ioctl_by_bdev(struct block_device *, unsigned, unsigned long); |
1640 | extern int blkdev_ioctl(struct inode *, struct file *, unsigned, unsigned long); | |
7006f6ec AK |
1641 | extern int blkdev_driver_ioctl(struct inode *inode, struct file *file, |
1642 | struct gendisk *disk, unsigned cmd, | |
1643 | unsigned long arg); | |
1da177e4 LT |
1644 | extern long compat_blkdev_ioctl(struct file *, unsigned, unsigned long); |
1645 | extern int blkdev_get(struct block_device *, mode_t, unsigned); | |
1646 | extern int blkdev_put(struct block_device *); | |
1647 | extern int bd_claim(struct block_device *, void *); | |
1648 | extern void bd_release(struct block_device *); | |
641dc636 JN |
1649 | #ifdef CONFIG_SYSFS |
1650 | extern int bd_claim_by_disk(struct block_device *, void *, struct gendisk *); | |
1651 | extern void bd_release_from_disk(struct block_device *, struct gendisk *); | |
1652 | #else | |
1653 | #define bd_claim_by_disk(bdev, holder, disk) bd_claim(bdev, holder) | |
1654 | #define bd_release_from_disk(bdev, disk) bd_release(bdev) | |
1655 | #endif | |
9361401e | 1656 | #endif |
1da177e4 LT |
1657 | |
1658 | /* fs/char_dev.c */ | |
68eef3b4 | 1659 | #define CHRDEV_MAJOR_HASH_SIZE 255 |
1da177e4 LT |
1660 | extern int alloc_chrdev_region(dev_t *, unsigned, unsigned, const char *); |
1661 | extern int register_chrdev_region(dev_t, unsigned, const char *); | |
1662 | extern int register_chrdev(unsigned int, const char *, | |
99ac48f5 | 1663 | const struct file_operations *); |
e53252d9 | 1664 | extern void unregister_chrdev(unsigned int, const char *); |
1da177e4 | 1665 | extern void unregister_chrdev_region(dev_t, unsigned); |
68eef3b4 | 1666 | extern void chrdev_show(struct seq_file *,off_t); |
1da177e4 LT |
1667 | |
1668 | /* fs/block_dev.c */ | |
1669 | #define BDEVNAME_SIZE 32 /* Largest string for a blockdev identifier */ | |
9361401e DH |
1670 | |
1671 | #ifdef CONFIG_BLOCK | |
1672 | #define BLKDEV_MAJOR_HASH_SIZE 255 | |
1da177e4 LT |
1673 | extern const char *__bdevname(dev_t, char *buffer); |
1674 | extern const char *bdevname(struct block_device *bdev, char *buffer); | |
1675 | extern struct block_device *lookup_bdev(const char *); | |
1676 | extern struct block_device *open_bdev_excl(const char *, int, void *); | |
1677 | extern void close_bdev_excl(struct block_device *); | |
68eef3b4 | 1678 | extern void blkdev_show(struct seq_file *,off_t); |
9361401e DH |
1679 | #else |
1680 | #define BLKDEV_MAJOR_HASH_SIZE 0 | |
1681 | #endif | |
1da177e4 LT |
1682 | |
1683 | extern void init_special_inode(struct inode *, umode_t, dev_t); | |
1684 | ||
1685 | /* Invalid inode operations -- fs/bad_inode.c */ | |
1686 | extern void make_bad_inode(struct inode *); | |
1687 | extern int is_bad_inode(struct inode *); | |
1688 | ||
4b6f5d20 AV |
1689 | extern const struct file_operations read_fifo_fops; |
1690 | extern const struct file_operations write_fifo_fops; | |
1691 | extern const struct file_operations rdwr_fifo_fops; | |
1da177e4 LT |
1692 | |
1693 | extern int fs_may_remount_ro(struct super_block *); | |
1694 | ||
9361401e | 1695 | #ifdef CONFIG_BLOCK |
1da177e4 LT |
1696 | /* |
1697 | * return READ, READA, or WRITE | |
1698 | */ | |
1699 | #define bio_rw(bio) ((bio)->bi_rw & (RW_MASK | RWA_MASK)) | |
1700 | ||
1701 | /* | |
1702 | * return data direction, READ or WRITE | |
1703 | */ | |
1704 | #define bio_data_dir(bio) ((bio)->bi_rw & 1) | |
1705 | ||
1706 | extern int check_disk_change(struct block_device *); | |
2ef41634 | 1707 | extern int __invalidate_device(struct block_device *); |
1da177e4 | 1708 | extern int invalidate_partition(struct gendisk *, int); |
9361401e DH |
1709 | #endif |
1710 | extern int invalidate_inodes(struct super_block *); | |
fc9a07e7 AM |
1711 | unsigned long __invalidate_mapping_pages(struct address_space *mapping, |
1712 | pgoff_t start, pgoff_t end, | |
1713 | bool be_atomic); | |
1da177e4 LT |
1714 | unsigned long invalidate_mapping_pages(struct address_space *mapping, |
1715 | pgoff_t start, pgoff_t end); | |
54bc4855 | 1716 | |
fc0ecff6 | 1717 | static inline unsigned long __deprecated |
54bc4855 AA |
1718 | invalidate_inode_pages(struct address_space *mapping) |
1719 | { | |
1720 | return invalidate_mapping_pages(mapping, 0, ~0UL); | |
1721 | } | |
1722 | ||
1da177e4 LT |
1723 | static inline void invalidate_remote_inode(struct inode *inode) |
1724 | { | |
1725 | if (S_ISREG(inode->i_mode) || S_ISDIR(inode->i_mode) || | |
1726 | S_ISLNK(inode->i_mode)) | |
fc0ecff6 | 1727 | invalidate_mapping_pages(inode->i_mapping, 0, -1); |
1da177e4 LT |
1728 | } |
1729 | extern int invalidate_inode_pages2(struct address_space *mapping); | |
1730 | extern int invalidate_inode_pages2_range(struct address_space *mapping, | |
1731 | pgoff_t start, pgoff_t end); | |
1732 | extern int write_inode_now(struct inode *, int); | |
1733 | extern int filemap_fdatawrite(struct address_space *); | |
1734 | extern int filemap_flush(struct address_space *); | |
1735 | extern int filemap_fdatawait(struct address_space *); | |
1736 | extern int filemap_write_and_wait(struct address_space *mapping); | |
1737 | extern int filemap_write_and_wait_range(struct address_space *mapping, | |
1738 | loff_t lstart, loff_t lend); | |
ebcf28e1 AM |
1739 | extern int wait_on_page_writeback_range(struct address_space *mapping, |
1740 | pgoff_t start, pgoff_t end); | |
1741 | extern int __filemap_fdatawrite_range(struct address_space *mapping, | |
1742 | loff_t start, loff_t end, int sync_mode); | |
1743 | ||
18e79b40 | 1744 | extern long do_fsync(struct file *file, int datasync); |
1da177e4 LT |
1745 | extern void sync_supers(void); |
1746 | extern void sync_filesystems(int wait); | |
cf9a2ae8 | 1747 | extern void __fsync_super(struct super_block *sb); |
1da177e4 LT |
1748 | extern void emergency_sync(void); |
1749 | extern void emergency_remount(void); | |
1750 | extern int do_remount_sb(struct super_block *sb, int flags, | |
1751 | void *data, int force); | |
9361401e | 1752 | #ifdef CONFIG_BLOCK |
1da177e4 | 1753 | extern sector_t bmap(struct inode *, sector_t); |
9361401e | 1754 | #endif |
1da177e4 LT |
1755 | extern int notify_change(struct dentry *, struct iattr *); |
1756 | extern int permission(struct inode *, int, struct nameidata *); | |
1757 | extern int generic_permission(struct inode *, int, | |
1758 | int (*check_acl)(struct inode *, int)); | |
1759 | ||
1760 | extern int get_write_access(struct inode *); | |
1761 | extern int deny_write_access(struct file *); | |
1762 | static inline void put_write_access(struct inode * inode) | |
1763 | { | |
1764 | atomic_dec(&inode->i_writecount); | |
1765 | } | |
1766 | static inline void allow_write_access(struct file *file) | |
1767 | { | |
1768 | if (file) | |
0f7fc9e4 | 1769 | atomic_inc(&file->f_path.dentry->d_inode->i_writecount); |
1da177e4 LT |
1770 | } |
1771 | extern int do_pipe(int *); | |
d6cbd281 AK |
1772 | extern struct file *create_read_pipe(struct file *f); |
1773 | extern struct file *create_write_pipe(void); | |
1774 | extern void free_write_pipe(struct file *); | |
1da177e4 | 1775 | |
a70e65df CH |
1776 | extern struct file *do_filp_open(int dfd, const char *pathname, |
1777 | int open_flag, int mode); | |
1da177e4 LT |
1778 | extern int may_open(struct nameidata *, int, int); |
1779 | ||
1780 | extern int kernel_read(struct file *, unsigned long, char *, unsigned long); | |
1781 | extern struct file * open_exec(const char *); | |
1782 | ||
1783 | /* fs/dcache.c -- generic fs support functions */ | |
1784 | extern int is_subdir(struct dentry *, struct dentry *); | |
1785 | extern ino_t find_inode_number(struct dentry *, struct qstr *); | |
1786 | ||
1787 | #include <linux/err.h> | |
1788 | ||
1789 | /* needed for stackable file system support */ | |
1790 | extern loff_t default_llseek(struct file *file, loff_t offset, int origin); | |
1791 | ||
1792 | extern loff_t vfs_llseek(struct file *file, loff_t offset, int origin); | |
1793 | ||
1794 | extern void inode_init_once(struct inode *); | |
1795 | extern void iput(struct inode *); | |
1796 | extern struct inode * igrab(struct inode *); | |
1797 | extern ino_t iunique(struct super_block *, ino_t); | |
1798 | extern int inode_needs_sync(struct inode *inode); | |
1799 | extern void generic_delete_inode(struct inode *inode); | |
cb2c0233 | 1800 | extern void generic_drop_inode(struct inode *inode); |
1da177e4 | 1801 | |
88bd5121 AA |
1802 | extern struct inode *ilookup5_nowait(struct super_block *sb, |
1803 | unsigned long hashval, int (*test)(struct inode *, void *), | |
1804 | void *data); | |
1da177e4 LT |
1805 | extern struct inode *ilookup5(struct super_block *sb, unsigned long hashval, |
1806 | int (*test)(struct inode *, void *), void *data); | |
1807 | extern struct inode *ilookup(struct super_block *sb, unsigned long ino); | |
1808 | ||
1809 | extern struct inode * iget5_locked(struct super_block *, unsigned long, int (*test)(struct inode *, void *), int (*set)(struct inode *, void *), void *); | |
1810 | extern struct inode * iget_locked(struct super_block *, unsigned long); | |
1811 | extern void unlock_new_inode(struct inode *); | |
1812 | ||
1da177e4 | 1813 | extern void __iget(struct inode * inode); |
b46980fe | 1814 | extern void iget_failed(struct inode *); |
1da177e4 LT |
1815 | extern void clear_inode(struct inode *); |
1816 | extern void destroy_inode(struct inode *); | |
1817 | extern struct inode *new_inode(struct super_block *); | |
01de85e0 | 1818 | extern int should_remove_suid(struct dentry *); |
1da177e4 | 1819 | extern int remove_suid(struct dentry *); |
1da177e4 LT |
1820 | |
1821 | extern void __insert_inode_hash(struct inode *, unsigned long hashval); | |
1822 | extern void remove_inode_hash(struct inode *); | |
1823 | static inline void insert_inode_hash(struct inode *inode) { | |
1824 | __insert_inode_hash(inode, inode->i_ino); | |
1825 | } | |
1826 | ||
1827 | extern struct file * get_empty_filp(void); | |
1828 | extern void file_move(struct file *f, struct list_head *list); | |
1829 | extern void file_kill(struct file *f); | |
9361401e | 1830 | #ifdef CONFIG_BLOCK |
1da177e4 LT |
1831 | struct bio; |
1832 | extern void submit_bio(int, struct bio *); | |
1833 | extern int bdev_read_only(struct block_device *); | |
9361401e | 1834 | #endif |
1da177e4 LT |
1835 | extern int set_blocksize(struct block_device *, int); |
1836 | extern int sb_set_blocksize(struct super_block *, int); | |
1837 | extern int sb_min_blocksize(struct super_block *, int); | |
08d8e974 | 1838 | extern int sb_has_dirty_inodes(struct super_block *); |
1da177e4 LT |
1839 | |
1840 | extern int generic_file_mmap(struct file *, struct vm_area_struct *); | |
1841 | extern int generic_file_readonly_mmap(struct file *, struct vm_area_struct *); | |
1842 | extern int file_read_actor(read_descriptor_t * desc, struct page *page, unsigned long offset, unsigned long size); | |
1da177e4 | 1843 | int generic_write_checks(struct file *file, loff_t *pos, size_t *count, int isblk); |
027445c3 | 1844 | extern ssize_t generic_file_aio_read(struct kiocb *, const struct iovec *, unsigned long, loff_t); |
027445c3 | 1845 | extern ssize_t generic_file_aio_write(struct kiocb *, const struct iovec *, unsigned long, loff_t); |
1da177e4 | 1846 | extern ssize_t generic_file_aio_write_nolock(struct kiocb *, const struct iovec *, |
027445c3 | 1847 | unsigned long, loff_t); |
1da177e4 LT |
1848 | extern ssize_t generic_file_direct_write(struct kiocb *, const struct iovec *, |
1849 | unsigned long *, loff_t, loff_t *, size_t, size_t); | |
1850 | extern ssize_t generic_file_buffered_write(struct kiocb *, const struct iovec *, | |
1851 | unsigned long, loff_t, loff_t *, size_t, ssize_t); | |
1852 | extern ssize_t do_sync_read(struct file *filp, char __user *buf, size_t len, loff_t *ppos); | |
1853 | extern ssize_t do_sync_write(struct file *filp, const char __user *buf, size_t len, loff_t *ppos); | |
0ceb3314 DM |
1854 | extern int generic_segment_checks(const struct iovec *iov, |
1855 | unsigned long *nr_segs, size_t *count, int access_flags); | |
88e6faef AM |
1856 | |
1857 | /* fs/splice.c */ | |
cbb7e577 | 1858 | extern ssize_t generic_file_splice_read(struct file *, loff_t *, |
88e6faef AM |
1859 | struct pipe_inode_info *, size_t, unsigned int); |
1860 | extern ssize_t generic_file_splice_write(struct pipe_inode_info *, | |
cbb7e577 | 1861 | struct file *, loff_t *, size_t, unsigned int); |
6da61809 MF |
1862 | extern ssize_t generic_file_splice_write_nolock(struct pipe_inode_info *, |
1863 | struct file *, loff_t *, size_t, unsigned int); | |
88e6faef | 1864 | extern ssize_t generic_splice_sendpage(struct pipe_inode_info *pipe, |
cbb7e577 JA |
1865 | struct file *out, loff_t *, size_t len, unsigned int flags); |
1866 | extern long do_splice_direct(struct file *in, loff_t *ppos, struct file *out, | |
88e6faef AM |
1867 | size_t len, unsigned int flags); |
1868 | ||
1da177e4 LT |
1869 | extern void |
1870 | file_ra_state_init(struct file_ra_state *ra, struct address_space *mapping); | |
1da177e4 LT |
1871 | extern loff_t no_llseek(struct file *file, loff_t offset, int origin); |
1872 | extern loff_t generic_file_llseek(struct file *file, loff_t offset, int origin); | |
1873 | extern loff_t remote_llseek(struct file *file, loff_t offset, int origin); | |
1874 | extern int generic_file_open(struct inode * inode, struct file * filp); | |
1875 | extern int nonseekable_open(struct inode * inode, struct file * filp); | |
1876 | ||
ceffc078 | 1877 | #ifdef CONFIG_FS_XIP |
eb6fe0c3 CO |
1878 | extern ssize_t xip_file_read(struct file *filp, char __user *buf, size_t len, |
1879 | loff_t *ppos); | |
ceffc078 | 1880 | extern int xip_file_mmap(struct file * file, struct vm_area_struct * vma); |
eb6fe0c3 CO |
1881 | extern ssize_t xip_file_write(struct file *filp, const char __user *buf, |
1882 | size_t len, loff_t *ppos); | |
ceffc078 | 1883 | extern int xip_truncate_page(struct address_space *mapping, loff_t from); |
6d79125b CO |
1884 | #else |
1885 | static inline int xip_truncate_page(struct address_space *mapping, loff_t from) | |
1886 | { | |
1887 | return 0; | |
1888 | } | |
ceffc078 CO |
1889 | #endif |
1890 | ||
9361401e | 1891 | #ifdef CONFIG_BLOCK |
1da177e4 LT |
1892 | ssize_t __blockdev_direct_IO(int rw, struct kiocb *iocb, struct inode *inode, |
1893 | struct block_device *bdev, const struct iovec *iov, loff_t offset, | |
1d8fa7a2 | 1894 | unsigned long nr_segs, get_block_t get_block, dio_iodone_t end_io, |
1da177e4 LT |
1895 | int lock_type); |
1896 | ||
1897 | enum { | |
1898 | DIO_LOCKING = 1, /* need locking between buffered and direct access */ | |
1899 | DIO_NO_LOCKING, /* bdev; no locking at all between buffered/direct */ | |
1900 | DIO_OWN_LOCKING, /* filesystem locks buffered and direct internally */ | |
1901 | }; | |
1902 | ||
1903 | static inline ssize_t blockdev_direct_IO(int rw, struct kiocb *iocb, | |
1904 | struct inode *inode, struct block_device *bdev, const struct iovec *iov, | |
1d8fa7a2 | 1905 | loff_t offset, unsigned long nr_segs, get_block_t get_block, |
1da177e4 LT |
1906 | dio_iodone_t end_io) |
1907 | { | |
1908 | return __blockdev_direct_IO(rw, iocb, inode, bdev, iov, offset, | |
1d8fa7a2 | 1909 | nr_segs, get_block, end_io, DIO_LOCKING); |
1da177e4 LT |
1910 | } |
1911 | ||
1912 | static inline ssize_t blockdev_direct_IO_no_locking(int rw, struct kiocb *iocb, | |
1913 | struct inode *inode, struct block_device *bdev, const struct iovec *iov, | |
1d8fa7a2 | 1914 | loff_t offset, unsigned long nr_segs, get_block_t get_block, |
1da177e4 LT |
1915 | dio_iodone_t end_io) |
1916 | { | |
1917 | return __blockdev_direct_IO(rw, iocb, inode, bdev, iov, offset, | |
1d8fa7a2 | 1918 | nr_segs, get_block, end_io, DIO_NO_LOCKING); |
1da177e4 LT |
1919 | } |
1920 | ||
1921 | static inline ssize_t blockdev_direct_IO_own_locking(int rw, struct kiocb *iocb, | |
1922 | struct inode *inode, struct block_device *bdev, const struct iovec *iov, | |
1d8fa7a2 | 1923 | loff_t offset, unsigned long nr_segs, get_block_t get_block, |
1da177e4 LT |
1924 | dio_iodone_t end_io) |
1925 | { | |
1926 | return __blockdev_direct_IO(rw, iocb, inode, bdev, iov, offset, | |
1d8fa7a2 | 1927 | nr_segs, get_block, end_io, DIO_OWN_LOCKING); |
1da177e4 | 1928 | } |
9361401e | 1929 | #endif |
1da177e4 | 1930 | |
4b6f5d20 | 1931 | extern const struct file_operations generic_ro_fops; |
1da177e4 LT |
1932 | |
1933 | #define special_file(m) (S_ISCHR(m)||S_ISBLK(m)||S_ISFIFO(m)||S_ISSOCK(m)) | |
1934 | ||
1935 | extern int vfs_readlink(struct dentry *, char __user *, int, const char *); | |
1936 | extern int vfs_follow_link(struct nameidata *, const char *); | |
1937 | extern int page_readlink(struct dentry *, char __user *, int); | |
cc314eef LT |
1938 | extern void *page_follow_link_light(struct dentry *, struct nameidata *); |
1939 | extern void page_put_link(struct dentry *, struct nameidata *, void *); | |
0adb25d2 KK |
1940 | extern int __page_symlink(struct inode *inode, const char *symname, int len, |
1941 | gfp_t gfp_mask); | |
1da177e4 | 1942 | extern int page_symlink(struct inode *inode, const char *symname, int len); |
c5ef1c42 | 1943 | extern const struct inode_operations page_symlink_inode_operations; |
1da177e4 LT |
1944 | extern int generic_readlink(struct dentry *, char __user *, int); |
1945 | extern void generic_fillattr(struct inode *, struct kstat *); | |
1946 | extern int vfs_getattr(struct vfsmount *, struct dentry *, struct kstat *); | |
1947 | void inode_add_bytes(struct inode *inode, loff_t bytes); | |
1948 | void inode_sub_bytes(struct inode *inode, loff_t bytes); | |
1949 | loff_t inode_get_bytes(struct inode *inode); | |
1950 | void inode_set_bytes(struct inode *inode, loff_t bytes); | |
1951 | ||
1952 | extern int vfs_readdir(struct file *, filldir_t, void *); | |
1953 | ||
1954 | extern int vfs_stat(char __user *, struct kstat *); | |
1955 | extern int vfs_lstat(char __user *, struct kstat *); | |
5590ff0d UD |
1956 | extern int vfs_stat_fd(int dfd, char __user *, struct kstat *); |
1957 | extern int vfs_lstat_fd(int dfd, char __user *, struct kstat *); | |
1da177e4 LT |
1958 | extern int vfs_fstat(unsigned int, struct kstat *); |
1959 | ||
deb21db7 EZ |
1960 | extern int do_vfs_ioctl(struct file *filp, unsigned int fd, unsigned int cmd, |
1961 | unsigned long arg); | |
1da177e4 | 1962 | |
c18479fe MS |
1963 | extern void get_filesystem(struct file_system_type *fs); |
1964 | extern void put_filesystem(struct file_system_type *fs); | |
1da177e4 LT |
1965 | extern struct file_system_type *get_fs_type(const char *name); |
1966 | extern struct super_block *get_super(struct block_device *); | |
1967 | extern struct super_block *user_get_super(dev_t); | |
1968 | extern void drop_super(struct super_block *sb); | |
1969 | ||
1970 | extern int dcache_dir_open(struct inode *, struct file *); | |
1971 | extern int dcache_dir_close(struct inode *, struct file *); | |
1972 | extern loff_t dcache_dir_lseek(struct file *, loff_t, int); | |
1973 | extern int dcache_readdir(struct file *, void *, filldir_t); | |
1974 | extern int simple_getattr(struct vfsmount *, struct dentry *, struct kstat *); | |
726c3342 | 1975 | extern int simple_statfs(struct dentry *, struct kstatfs *); |
1da177e4 LT |
1976 | extern int simple_link(struct dentry *, struct inode *, struct dentry *); |
1977 | extern int simple_unlink(struct inode *, struct dentry *); | |
1978 | extern int simple_rmdir(struct inode *, struct dentry *); | |
1979 | extern int simple_rename(struct inode *, struct dentry *, struct inode *, struct dentry *); | |
1980 | extern int simple_sync_file(struct file *, struct dentry *, int); | |
1981 | extern int simple_empty(struct dentry *); | |
1982 | extern int simple_readpage(struct file *file, struct page *page); | |
1983 | extern int simple_prepare_write(struct file *file, struct page *page, | |
1984 | unsigned offset, unsigned to); | |
afddba49 NP |
1985 | extern int simple_write_begin(struct file *file, struct address_space *mapping, |
1986 | loff_t pos, unsigned len, unsigned flags, | |
1987 | struct page **pagep, void **fsdata); | |
1988 | extern int simple_write_end(struct file *file, struct address_space *mapping, | |
1989 | loff_t pos, unsigned len, unsigned copied, | |
1990 | struct page *page, void *fsdata); | |
1da177e4 LT |
1991 | |
1992 | extern struct dentry *simple_lookup(struct inode *, struct dentry *, struct nameidata *); | |
1993 | extern ssize_t generic_read_dir(struct file *, char __user *, size_t, loff_t *); | |
4b6f5d20 | 1994 | extern const struct file_operations simple_dir_operations; |
c5ef1c42 | 1995 | extern const struct inode_operations simple_dir_inode_operations; |
4b6f5d20 | 1996 | struct tree_descr { char *name; const struct file_operations *ops; int mode; }; |
1da177e4 LT |
1997 | struct dentry *d_alloc_name(struct dentry *, const char *); |
1998 | extern int simple_fill_super(struct super_block *, int, struct tree_descr *); | |
1f5ce9e9 | 1999 | extern int simple_pin_fs(struct file_system_type *, struct vfsmount **mount, int *count); |
1da177e4 LT |
2000 | extern void simple_release_fs(struct vfsmount **mount, int *count); |
2001 | ||
93b07113 AM |
2002 | extern ssize_t simple_read_from_buffer(void __user *to, size_t count, |
2003 | loff_t *ppos, const void *from, size_t available); | |
2004 | extern ssize_t memory_read_from_buffer(void *to, size_t count, loff_t *ppos, | |
2005 | const void *from, size_t available); | |
1da177e4 | 2006 | |
e965f963 | 2007 | #ifdef CONFIG_MIGRATION |
2d1db3b1 CL |
2008 | extern int buffer_migrate_page(struct address_space *, |
2009 | struct page *, struct page *); | |
e965f963 CL |
2010 | #else |
2011 | #define buffer_migrate_page NULL | |
2012 | #endif | |
2013 | ||
1da177e4 LT |
2014 | extern int inode_change_ok(struct inode *, struct iattr *); |
2015 | extern int __must_check inode_setattr(struct inode *, struct iattr *); | |
2016 | ||
870f4817 | 2017 | extern void file_update_time(struct file *file); |
1da177e4 | 2018 | |
b3b304a2 MS |
2019 | extern int generic_show_options(struct seq_file *m, struct vfsmount *mnt); |
2020 | extern void save_mount_options(struct super_block *sb, char *options); | |
2021 | ||
1da177e4 LT |
2022 | static inline ino_t parent_ino(struct dentry *dentry) |
2023 | { | |
2024 | ino_t res; | |
2025 | ||
2026 | spin_lock(&dentry->d_lock); | |
2027 | res = dentry->d_parent->d_inode->i_ino; | |
2028 | spin_unlock(&dentry->d_lock); | |
2029 | return res; | |
2030 | } | |
2031 | ||
1da177e4 LT |
2032 | /* Transaction based IO helpers */ |
2033 | ||
2034 | /* | |
2035 | * An argresp is stored in an allocated page and holds the | |
2036 | * size of the argument or response, along with its content | |
2037 | */ | |
2038 | struct simple_transaction_argresp { | |
2039 | ssize_t size; | |
2040 | char data[0]; | |
2041 | }; | |
2042 | ||
2043 | #define SIMPLE_TRANSACTION_LIMIT (PAGE_SIZE - sizeof(struct simple_transaction_argresp)) | |
2044 | ||
2045 | char *simple_transaction_get(struct file *file, const char __user *buf, | |
2046 | size_t size); | |
2047 | ssize_t simple_transaction_read(struct file *file, char __user *buf, | |
2048 | size_t size, loff_t *pos); | |
2049 | int simple_transaction_release(struct inode *inode, struct file *file); | |
2050 | ||
2051 | static inline void simple_transaction_set(struct file *file, size_t n) | |
2052 | { | |
2053 | struct simple_transaction_argresp *ar = file->private_data; | |
2054 | ||
2055 | BUG_ON(n > SIMPLE_TRANSACTION_LIMIT); | |
2056 | ||
2057 | /* | |
2058 | * The barrier ensures that ar->size will really remain zero until | |
2059 | * ar->data is ready for reading. | |
2060 | */ | |
2061 | smp_mb(); | |
2062 | ar->size = n; | |
2063 | } | |
2064 | ||
acaefc25 AB |
2065 | /* |
2066 | * simple attribute files | |
2067 | * | |
2068 | * These attributes behave similar to those in sysfs: | |
2069 | * | |
2070 | * Writing to an attribute immediately sets a value, an open file can be | |
2071 | * written to multiple times. | |
2072 | * | |
2073 | * Reading from an attribute creates a buffer from the value that might get | |
2074 | * read with multiple read calls. When the attribute has been read | |
2075 | * completely, no further read calls are possible until the file is opened | |
2076 | * again. | |
2077 | * | |
2078 | * All attributes contain a text representation of a numeric value | |
2079 | * that are accessed with the get() and set() functions. | |
2080 | */ | |
2081 | #define DEFINE_SIMPLE_ATTRIBUTE(__fops, __get, __set, __fmt) \ | |
2082 | static int __fops ## _open(struct inode *inode, struct file *file) \ | |
2083 | { \ | |
2084 | __simple_attr_check_format(__fmt, 0ull); \ | |
2085 | return simple_attr_open(inode, file, __get, __set, __fmt); \ | |
2086 | } \ | |
2087 | static struct file_operations __fops = { \ | |
2088 | .owner = THIS_MODULE, \ | |
2089 | .open = __fops ## _open, \ | |
74bedc4d | 2090 | .release = simple_attr_release, \ |
acaefc25 AB |
2091 | .read = simple_attr_read, \ |
2092 | .write = simple_attr_write, \ | |
2093 | }; | |
2094 | ||
2095 | static inline void __attribute__((format(printf, 1, 2))) | |
2096 | __simple_attr_check_format(const char *fmt, ...) | |
2097 | { | |
2098 | /* don't do anything, just let the compiler check the arguments; */ | |
2099 | } | |
2100 | ||
2101 | int simple_attr_open(struct inode *inode, struct file *file, | |
8b88b099 | 2102 | int (*get)(void *, u64 *), int (*set)(void *, u64), |
acaefc25 | 2103 | const char *fmt); |
74bedc4d | 2104 | int simple_attr_release(struct inode *inode, struct file *file); |
acaefc25 AB |
2105 | ssize_t simple_attr_read(struct file *file, char __user *buf, |
2106 | size_t len, loff_t *ppos); | |
2107 | ssize_t simple_attr_write(struct file *file, const char __user *buf, | |
2108 | size_t len, loff_t *ppos); | |
2109 | ||
2110 | ||
1da177e4 LT |
2111 | #ifdef CONFIG_SECURITY |
2112 | static inline char *alloc_secdata(void) | |
2113 | { | |
2114 | return (char *)get_zeroed_page(GFP_KERNEL); | |
2115 | } | |
2116 | ||
2117 | static inline void free_secdata(void *secdata) | |
2118 | { | |
2119 | free_page((unsigned long)secdata); | |
2120 | } | |
2121 | #else | |
2122 | static inline char *alloc_secdata(void) | |
2123 | { | |
2124 | return (char *)1; | |
2125 | } | |
2126 | ||
2127 | static inline void free_secdata(void *secdata) | |
2128 | { } | |
2129 | #endif /* CONFIG_SECURITY */ | |
2130 | ||
4be28540 AD |
2131 | struct ctl_table; |
2132 | int proc_nr_files(struct ctl_table *table, int write, struct file *filp, | |
62239ac2 AB |
2133 | void __user *buffer, size_t *lenp, loff_t *ppos); |
2134 | ||
011e3fcd | 2135 | int get_filesystem_list(char * buf); |
62239ac2 | 2136 | |
1da177e4 LT |
2137 | #endif /* __KERNEL__ */ |
2138 | #endif /* _LINUX_FS_H */ |