blob: 7a049fd2aa4cca0860f9dd51f7a317d7fe3b58d5 [file] [log] [blame]
Linus Torvalds1da177e2005-04-16 15:20:36 -07001#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
Linus Torvalds1da177e2005-04-16 15:20:36 -07009#include <linux/limits.h>
10#include <linux/ioctl.h>
Tejun Heo7cc01582010-08-03 13:14:58 +020011#include <linux/blk_types.h>
Ingo Molnarb31d42a2010-10-28 16:39:24 +020012#include <linux/types.h>
Linus Torvalds1da177e2005-04-16 15:20:36 -070013
14/*
15 * It's silly to have NR_OPEN bigger than NR_FILE, but you can change
16 * the file limit at runtime and only root can increase the per-process
17 * nr_file rlimit, so it's safe to set up a ridiculously high absolute
18 * upper limit on files-per-process.
19 *
20 * Some programs (notably those using select()) may have to be
21 * recompiled to take full advantage of the new limits..
22 */
23
24/* Fixed constants first: */
25#undef NR_OPEN
Tim Gardner0ac1ee02011-05-24 17:13:05 -070026#define INR_OPEN_CUR 1024 /* Initial setting for nfile rlimits */
27#define INR_OPEN_MAX 4096 /* Hard limit for nfile rlimits */
Linus Torvalds1da177e2005-04-16 15:20:36 -070028
29#define BLOCK_SIZE_BITS 10
30#define BLOCK_SIZE (1<<BLOCK_SIZE_BITS)
31
David Howells01bf4662006-07-10 04:44:53 -070032#define SEEK_SET 0 /* seek relative to beginning of file */
33#define SEEK_CUR 1 /* seek relative to current file position */
34#define SEEK_END 2 /* seek relative to end of file */
Josef Bacik982d8162011-07-18 13:21:35 -040035#define SEEK_DATA 3 /* seek to the next data */
36#define SEEK_HOLE 4 /* seek to the next hole */
37#define SEEK_MAX SEEK_HOLE
David Howells01bf4662006-07-10 04:44:53 -070038
Lukas Czerner367a51a32010-10-27 21:30:11 -040039struct fstrim_range {
Loïc Minier3a3a1af2010-11-24 12:56:53 -080040 __u64 start;
41 __u64 len;
42 __u64 minlen;
Lukas Czerner367a51a32010-10-27 21:30:11 -040043};
44
Linus Torvalds1da177e2005-04-16 15:20:36 -070045/* And dynamically-tunable limits and defaults: */
46struct files_stat_struct {
Eric Dumazet518de9b2010-10-26 14:22:44 -070047 unsigned long nr_files; /* read only */
48 unsigned long nr_free_files; /* read only */
49 unsigned long max_files; /* tunable */
Linus Torvalds1da177e2005-04-16 15:20:36 -070050};
Linus Torvalds1da177e2005-04-16 15:20:36 -070051
52struct inodes_stat_t {
53 int nr_inodes;
54 int nr_unused;
Stefan Richter9e7bf242007-07-15 23:40:25 -070055 int dummy[5]; /* padding for sysctl ABI compatibility */
Linus Torvalds1da177e2005-04-16 15:20:36 -070056};
Linus Torvalds1da177e2005-04-16 15:20:36 -070057
Linus Torvalds1da177e2005-04-16 15:20:36 -070058
59#define NR_FILE 8192 /* this can well be larger on a larger system */
60
Aneesh Kumar K.V8522ca52011-10-23 23:13:31 +053061#define MAY_EXEC 0x00000001
62#define MAY_WRITE 0x00000002
63#define MAY_READ 0x00000004
64#define MAY_APPEND 0x00000008
65#define MAY_ACCESS 0x00000010
66#define MAY_OPEN 0x00000020
67#define MAY_CHDIR 0x00000040
68/* called from RCU mode, don't block */
69#define MAY_NOT_BLOCK 0x00000080
Linus Torvalds1da177e2005-04-16 15:20:36 -070070
Paul Turner55ec8212009-02-18 14:48:15 -080071/*
72 * flags in file.f_mode. Note that FMODE_READ and FMODE_WRITE must correspond
73 * to O_WRONLY and O_RDWR via the strange trick in __dentry_open()
74 */
75
Christoph Hellwigfc9161e2008-11-05 14:58:46 +010076/* file is open for reading */
Andrew Morton19adf9c2010-03-05 13:42:03 -080077#define FMODE_READ ((__force fmode_t)0x1)
Christoph Hellwigfc9161e2008-11-05 14:58:46 +010078/* file is open for writing */
Andrew Morton19adf9c2010-03-05 13:42:03 -080079#define FMODE_WRITE ((__force fmode_t)0x2)
Christoph Hellwigfc9161e2008-11-05 14:58:46 +010080/* file is seekable */
Andrew Morton19adf9c2010-03-05 13:42:03 -080081#define FMODE_LSEEK ((__force fmode_t)0x4)
Paul Turner55ec8212009-02-18 14:48:15 -080082/* file can be accessed using pread */
Andrew Morton19adf9c2010-03-05 13:42:03 -080083#define FMODE_PREAD ((__force fmode_t)0x8)
Paul Turner55ec8212009-02-18 14:48:15 -080084/* file can be accessed using pwrite */
Andrew Morton19adf9c2010-03-05 13:42:03 -080085#define FMODE_PWRITE ((__force fmode_t)0x10)
Christoph Hellwigfc9161e2008-11-05 14:58:46 +010086/* File is opened for execution with sys_execve / sys_uselib */
Andrew Morton19adf9c2010-03-05 13:42:03 -080087#define FMODE_EXEC ((__force fmode_t)0x20)
Christoph Hellwigfc9161e2008-11-05 14:58:46 +010088/* File is opened with O_NDELAY (only set for block devices) */
Andrew Morton19adf9c2010-03-05 13:42:03 -080089#define FMODE_NDELAY ((__force fmode_t)0x40)
Christoph Hellwigfc9161e2008-11-05 14:58:46 +010090/* File is opened with O_EXCL (only set for block devices) */
Andrew Morton19adf9c2010-03-05 13:42:03 -080091#define FMODE_EXCL ((__force fmode_t)0x80)
Christoph Hellwigfc9161e2008-11-05 14:58:46 +010092/* File is opened using open(.., 3, ..) and is writeable only for ioctls
93 (specialy hack for floppy.c) */
Andrew Morton19adf9c2010-03-05 13:42:03 -080094#define FMODE_WRITE_IOCTL ((__force fmode_t)0x100)
Oleg Drokinb5005312006-03-25 03:07:01 -080095
Christoph Hellwig4d4be482008-12-09 04:47:33 -050096/*
97 * Don't update ctime and mtime.
98 *
99 * Currently a special hack for the XFS open_by_handle ioctl, but we'll
100 * hopefully graduate it to a proper O_CMTIME flag supported by open(2) soon.
101 */
Andrew Morton19adf9c2010-03-05 13:42:03 -0800102#define FMODE_NOCMTIME ((__force fmode_t)0x800)
Christoph Hellwig4d4be482008-12-09 04:47:33 -0500103
Wu Fengguang01414502010-03-05 13:42:03 -0800104/* Expect random access pattern */
Andrew Morton19adf9c2010-03-05 13:42:03 -0800105#define FMODE_RANDOM ((__force fmode_t)0x1000)
Wu Fengguang01414502010-03-05 13:42:03 -0800106
KAMEZAWA Hiroyuki4a3956c2010-10-01 14:20:22 -0700107/* File is huge (eg. /dev/kmem): treat loff_t as unsigned */
108#define FMODE_UNSIGNED_OFFSET ((__force fmode_t)0x2000)
109
Al Viro1abf0c72011-03-13 03:51:11 -0400110/* File is opened with O_PATH; almost nothing can be done with it */
111#define FMODE_PATH ((__force fmode_t)0x4000)
112
Eric Parisecf081d2009-12-17 21:24:25 -0500113/* File was opened by fanotify and shouldn't generate fanotify events */
Andrew Morton13bcbc02010-08-10 18:01:09 -0700114#define FMODE_NONOTIFY ((__force fmode_t)0x1000000)
Eric Parisecf081d2009-12-17 21:24:25 -0500115
Jens Axboe48e70bc2009-04-14 08:19:27 +0200116/*
117 * The below are the various read and write types that we support. Some of
118 * them include behavioral modifiers that send information down to the
119 * block layer and IO scheduler. Terminology:
120 *
121 * The block layer uses device plugging to defer IO a little bit, in
122 * the hope that we will see more IO very shortly. This increases
123 * coalescing of adjacent IO and thus reduces the number of IOs we
124 * have to send to the device. It also allows for better queuing,
125 * if the IO isn't mergeable. If the caller is going to be waiting
126 * for the IO, then he must ensure that the device is unplugged so
127 * that the IO is dispatched to the driver.
128 *
129 * All IO is handled async in Linux. This is fine for background
130 * writes, but for reads or writes that someone waits for completion
131 * on, we want to notify the block layer and IO scheduler so that they
132 * know about it. That allows them to make better scheduling
133 * decisions. So when the below references 'sync' and 'async', it
134 * is referencing this priority hint.
135 *
136 * With that in mind, the available types are:
137 *
138 * READ A normal read operation. Device will be plugged.
139 * READ_SYNC A synchronous read. Device is not plugged, caller can
140 * immediately wait on this read without caring about
141 * unplugging.
142 * READA Used for read-ahead operations. Lower priority, and the
Tejun Heo7cc01582010-08-03 13:14:58 +0200143 * block layer could (in theory) choose to ignore this
Jens Axboe48e70bc2009-04-14 08:19:27 +0200144 * request if it runs into resource problems.
145 * WRITE A normal async write. Device will be plugged.
Jens Axboe721a9602011-03-09 11:56:30 +0100146 * WRITE_SYNC Synchronous write. Identical to WRITE, but passes down
Jens Axboe48e70bc2009-04-14 08:19:27 +0200147 * the hint that someone will be waiting on this IO
Jens Axboe721a9602011-03-09 11:56:30 +0100148 * shortly. The write equivalent of READ_SYNC.
149 * WRITE_ODIRECT Special case write for O_DIRECT only.
Tejun Heo4fed9472010-09-03 11:56:17 +0200150 * WRITE_FLUSH Like WRITE_SYNC but with preceding cache flush.
151 * WRITE_FUA Like WRITE_SYNC but data is guaranteed to be on
152 * non-volatile media on completion.
153 * WRITE_FLUSH_FUA Combination of WRITE_FLUSH and FUA. The IO is preceded
154 * by a cache flush and data is guaranteed to be on
155 * non-volatile media on completion.
Jens Axboe48e70bc2009-04-14 08:19:27 +0200156 *
157 */
Tejun Heo7cc01582010-08-03 13:14:58 +0200158#define RW_MASK REQ_WRITE
159#define RWA_MASK REQ_RAHEAD
Christoph Hellwig7b6d91d2010-08-07 18:20:39 +0200160
161#define READ 0
Tejun Heo7cc01582010-08-03 13:14:58 +0200162#define WRITE RW_MASK
163#define READA RWA_MASK
Christoph Hellwig7b6d91d2010-08-07 18:20:39 +0200164
Jens Axboe721a9602011-03-09 11:56:30 +0100165#define READ_SYNC (READ | REQ_SYNC)
Jens Axboe721a9602011-03-09 11:56:30 +0100166#define WRITE_SYNC (WRITE | REQ_SYNC | REQ_NOIDLE)
167#define WRITE_ODIRECT (WRITE | REQ_SYNC)
Jens Axboe721a9602011-03-09 11:56:30 +0100168#define WRITE_FLUSH (WRITE | REQ_SYNC | REQ_NOIDLE | REQ_FLUSH)
169#define WRITE_FUA (WRITE | REQ_SYNC | REQ_NOIDLE | REQ_FUA)
170#define WRITE_FLUSH_FUA (WRITE | REQ_SYNC | REQ_NOIDLE | REQ_FLUSH | REQ_FUA)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700171
172#define SEL_IN 1
173#define SEL_OUT 2
174#define SEL_EX 4
175
176/* public flags for file_system_type */
177#define FS_REQUIRES_DEV 1
178#define FS_BINARY_MOUNTDATA 2
Miklos Szeredi79c0b2d2007-05-08 00:25:43 -0700179#define FS_HAS_SUBTYPE 4
Linus Torvalds1da177e2005-04-16 15:20:36 -0700180#define FS_REVAL_DOT 16384 /* Check the paths ".", ".." for staleness */
Mark Fasheh349457c2006-09-08 14:22:21 -0700181#define FS_RENAME_DOES_D_MOVE 32768 /* FS will handle d_move()
182 * during rename() internally.
183 */
184
Linus Torvalds1da177e2005-04-16 15:20:36 -0700185/*
186 * These are the fs-independent mount-flags: up to 32 flags are supported
187 */
188#define MS_RDONLY 1 /* Mount read-only */
189#define MS_NOSUID 2 /* Ignore suid and sgid bits */
190#define MS_NODEV 4 /* Disallow access to device special files */
191#define MS_NOEXEC 8 /* Disallow program execution */
192#define MS_SYNCHRONOUS 16 /* Writes are synced at once */
193#define MS_REMOUNT 32 /* Alter flags of a mounted FS */
194#define MS_MANDLOCK 64 /* Allow mandatory locks on an FS */
195#define MS_DIRSYNC 128 /* Directory modifications are synchronous */
196#define MS_NOATIME 1024 /* Do not update access times. */
197#define MS_NODIRATIME 2048 /* Do not update directory access times */
198#define MS_BIND 4096
199#define MS_MOVE 8192
200#define MS_REC 16384
Theodore Ts'o9b04c992006-03-24 03:15:10 -0800201#define MS_VERBOSE 32768 /* War is peace. Verbosity is silence.
202 MS_VERBOSE is deprecated. */
203#define MS_SILENT 32768
Miklos Szeredibf066c72006-01-08 01:03:19 -0800204#define MS_POSIXACL (1<<16) /* VFS does not apply the umask */
Ram Pai9676f0c2005-11-07 17:21:20 -0500205#define MS_UNBINDABLE (1<<17) /* change to unbindable */
Ram Pai07b20882005-11-07 17:19:07 -0500206#define MS_PRIVATE (1<<18) /* change to private */
Ram Paia58b0eb2005-11-07 17:20:48 -0500207#define MS_SLAVE (1<<19) /* change to slave */
Ram Pai03e06e62005-11-07 17:19:33 -0500208#define MS_SHARED (1<<20) /* change to shared */
Valerie Henson47ae32d2006-12-13 00:34:34 -0800209#define MS_RELATIME (1<<21) /* Update atime relative to mtime/ctime. */
Pavel Emelyanov8bf97252007-10-18 23:40:02 -0700210#define MS_KERNMOUNT (1<<22) /* this is a kern_mount call */
Jean Noel Cordenner7a224222008-01-28 23:58:27 -0500211#define MS_I_VERSION (1<<23) /* Update inode I_version field */
Matthew Garrettd0adde52009-03-26 17:49:56 +0000212#define MS_STRICTATIME (1<<24) /* Always perform atime updates */
Al Viro9e1f1de2011-06-03 18:24:58 -0400213#define MS_NOSEC (1<<28)
Al Viro7a4dec52010-08-09 12:05:43 -0400214#define MS_BORN (1<<29)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700215#define MS_ACTIVE (1<<30)
216#define MS_NOUSER (1<<31)
217
218/*
219 * Superblock flags that can be altered by MS_REMOUNT
220 */
Mimi Zohar08b9fe62008-10-13 00:09:50 -0400221#define MS_RMT_MASK (MS_RDONLY|MS_SYNCHRONOUS|MS_MANDLOCK|MS_I_VERSION)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700222
223/*
224 * Old magic mount flag and mask
225 */
226#define MS_MGC_VAL 0xC0ED0000
227#define MS_MGC_MSK 0xffff0000
228
229/* Inode flags - they have nothing to superblock flags now */
230
231#define S_SYNC 1 /* Writes are synced at once */
232#define S_NOATIME 2 /* Do not update access times */
233#define S_APPEND 4 /* Append-only file */
234#define S_IMMUTABLE 8 /* Immutable file */
235#define S_DEAD 16 /* removed, but still open directory */
236#define S_NOQUOTA 32 /* Inode is not counted to quota */
237#define S_DIRSYNC 64 /* Directory modifications are synchronous */
238#define S_NOCMTIME 128 /* Do not update file c/mtime */
239#define S_SWAPFILE 256 /* Do not truncate: swapon got its bmaps */
240#define S_PRIVATE 512 /* Inode is fs-internal */
Eric Paris196f5182010-10-25 14:42:19 -0400241#define S_IMA 1024 /* Inode has an associated IMA struct */
David Howells9875cf82011-01-14 18:45:21 +0000242#define S_AUTOMOUNT 2048 /* Automount/referral quasi-directory */
Andi Kleen69b45732011-05-28 08:25:51 -0700243#define S_NOSEC 4096 /* no suid or xattr security attributes */
Linus Torvalds1da177e2005-04-16 15:20:36 -0700244
245/*
246 * Note that nosuid etc flags are inode-specific: setting some file-system
247 * flags just means all the inodes inherit those flags by default. It might be
248 * possible to override it selectively if you really wanted to with some
249 * ioctl() that is not currently implemented.
250 *
251 * Exception: MS_RDONLY is always applied to the entire file system.
252 *
253 * Unfortunately, it is possible to change a filesystems flags with it mounted
254 * with files in use. This means that all of the inodes will not have their
255 * i_flags updated. Hence, i_flags no longer inherit the superblock mount
256 * flags, so these have to be checked separately. -- rmk@arm.uk.linux.org
257 */
258#define __IS_FLG(inode,flg) ((inode)->i_sb->s_flags & (flg))
259
260#define IS_RDONLY(inode) ((inode)->i_sb->s_flags & MS_RDONLY)
261#define IS_SYNC(inode) (__IS_FLG(inode, MS_SYNCHRONOUS) || \
262 ((inode)->i_flags & S_SYNC))
263#define IS_DIRSYNC(inode) (__IS_FLG(inode, MS_SYNCHRONOUS|MS_DIRSYNC) || \
264 ((inode)->i_flags & (S_SYNC|S_DIRSYNC)))
265#define IS_MANDLOCK(inode) __IS_FLG(inode, MS_MANDLOCK)
Eric Dumazet37756ce2007-02-10 01:44:49 -0800266#define IS_NOATIME(inode) __IS_FLG(inode, MS_RDONLY|MS_NOATIME)
Jean Noel Cordenner7a224222008-01-28 23:58:27 -0500267#define IS_I_VERSION(inode) __IS_FLG(inode, MS_I_VERSION)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700268
269#define IS_NOQUOTA(inode) ((inode)->i_flags & S_NOQUOTA)
270#define IS_APPEND(inode) ((inode)->i_flags & S_APPEND)
271#define IS_IMMUTABLE(inode) ((inode)->i_flags & S_IMMUTABLE)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700272#define IS_POSIXACL(inode) __IS_FLG(inode, MS_POSIXACL)
273
274#define IS_DEADDIR(inode) ((inode)->i_flags & S_DEAD)
275#define IS_NOCMTIME(inode) ((inode)->i_flags & S_NOCMTIME)
276#define IS_SWAPFILE(inode) ((inode)->i_flags & S_SWAPFILE)
277#define IS_PRIVATE(inode) ((inode)->i_flags & S_PRIVATE)
Eric Paris196f5182010-10-25 14:42:19 -0400278#define IS_IMA(inode) ((inode)->i_flags & S_IMA)
David Howells9875cf82011-01-14 18:45:21 +0000279#define IS_AUTOMOUNT(inode) ((inode)->i_flags & S_AUTOMOUNT)
Andi Kleen69b45732011-05-28 08:25:51 -0700280#define IS_NOSEC(inode) ((inode)->i_flags & S_NOSEC)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700281
282/* the read-only stuff doesn't really belong here, but any other place is
283 probably as bad and I don't want to create yet another include file. */
284
285#define BLKROSET _IO(0x12,93) /* set device read-only (0 = read-write) */
286#define BLKROGET _IO(0x12,94) /* get read-only status (0 = read_write) */
287#define BLKRRPART _IO(0x12,95) /* re-read partition table */
288#define BLKGETSIZE _IO(0x12,96) /* return device size /512 (long *arg) */
289#define BLKFLSBUF _IO(0x12,97) /* flush buffer cache */
290#define BLKRASET _IO(0x12,98) /* set read ahead for block device */
291#define BLKRAGET _IO(0x12,99) /* get current read ahead setting */
292#define BLKFRASET _IO(0x12,100)/* set filesystem (mm/filemap.c) read-ahead */
293#define BLKFRAGET _IO(0x12,101)/* get filesystem (mm/filemap.c) read-ahead */
294#define BLKSECTSET _IO(0x12,102)/* set max sectors per request (ll_rw_blk.c) */
295#define BLKSECTGET _IO(0x12,103)/* get max sectors per request (ll_rw_blk.c) */
296#define BLKSSZGET _IO(0x12,104)/* get block device sector size */
297#if 0
298#define BLKPG _IO(0x12,105)/* See blkpg.h */
299
300/* Some people are morons. Do not use sizeof! */
301
302#define BLKELVGET _IOR(0x12,106,size_t)/* elevator get */
303#define BLKELVSET _IOW(0x12,107,size_t)/* elevator set */
304/* This was here just to show that the number is taken -
305 probably all these _IO(0x12,*) ioctls should be moved to blkpg.h. */
306#endif
307/* A jump here: 108-111 have been used for various private purposes. */
308#define BLKBSZGET _IOR(0x12,112,size_t)
309#define BLKBSZSET _IOW(0x12,113,size_t)
310#define BLKGETSIZE64 _IOR(0x12,114,size_t) /* return device size in bytes (u64 *arg) */
Jens Axboe2056a782006-03-23 20:00:26 +0100311#define BLKTRACESETUP _IOWR(0x12,115,struct blk_user_trace_setup)
312#define BLKTRACESTART _IO(0x12,116)
313#define BLKTRACESTOP _IO(0x12,117)
314#define BLKTRACETEARDOWN _IO(0x12,118)
David Woodhoused30a2602008-08-11 15:58:42 +0100315#define BLKDISCARD _IO(0x12,119)
Martin K. Petersenac481c22009-10-03 20:52:01 +0200316#define BLKIOMIN _IO(0x12,120)
317#define BLKIOOPT _IO(0x12,121)
318#define BLKALIGNOFF _IO(0x12,122)
319#define BLKPBSZGET _IO(0x12,123)
Martin K. Petersen98262f22009-12-03 09:24:48 +0100320#define BLKDISCARDZEROES _IO(0x12,124)
Adrian Hunter8d57a982010-08-11 14:17:49 -0700321#define BLKSECDISCARD _IO(0x12,125)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700322
323#define BMAP_IOCTL 1 /* obsolete - kept for compatibility */
324#define FIBMAP _IO(0x00,1) /* bmap access */
325#define FIGETBSZ _IO(0x00,2) /* get the block size used for bmap */
Takashi Satofcccf502009-01-09 16:40:59 -0800326#define FIFREEZE _IOWR('X', 119, int) /* Freeze */
327#define FITHAW _IOWR('X', 120, int) /* Thaw */
Lukas Czerner367a51a32010-10-27 21:30:11 -0400328#define FITRIM _IOWR('X', 121, struct fstrim_range) /* Trim */
Linus Torvalds1da177e2005-04-16 15:20:36 -0700329
David Howells36695672006-08-29 19:06:16 +0100330#define FS_IOC_GETFLAGS _IOR('f', 1, long)
331#define FS_IOC_SETFLAGS _IOW('f', 2, long)
332#define FS_IOC_GETVERSION _IOR('v', 1, long)
333#define FS_IOC_SETVERSION _IOW('v', 2, long)
Mark Fashehc4b929b2008-10-08 19:44:18 -0400334#define FS_IOC_FIEMAP _IOWR('f', 11, struct fiemap)
David Howells36695672006-08-29 19:06:16 +0100335#define FS_IOC32_GETFLAGS _IOR('f', 1, int)
336#define FS_IOC32_SETFLAGS _IOW('f', 2, int)
337#define FS_IOC32_GETVERSION _IOR('v', 1, int)
338#define FS_IOC32_SETVERSION _IOW('v', 2, int)
339
340/*
341 * Inode flags (FS_IOC_GETFLAGS / FS_IOC_SETFLAGS)
342 */
343#define FS_SECRM_FL 0x00000001 /* Secure deletion */
344#define FS_UNRM_FL 0x00000002 /* Undelete */
345#define FS_COMPR_FL 0x00000004 /* Compress file */
346#define FS_SYNC_FL 0x00000008 /* Synchronous updates */
347#define FS_IMMUTABLE_FL 0x00000010 /* Immutable file */
348#define FS_APPEND_FL 0x00000020 /* writes to file may only append */
349#define FS_NODUMP_FL 0x00000040 /* do not dump file */
350#define FS_NOATIME_FL 0x00000080 /* do not update atime */
351/* Reserved for compression usage... */
352#define FS_DIRTY_FL 0x00000100
353#define FS_COMPRBLK_FL 0x00000200 /* One or more compressed clusters */
354#define FS_NOCOMP_FL 0x00000400 /* Don't compress */
355#define FS_ECOMPR_FL 0x00000800 /* Compression error */
356/* End compression flags --- maybe not all used */
357#define FS_BTREE_FL 0x00001000 /* btree format dir */
358#define FS_INDEX_FL 0x00001000 /* hash-indexed directory */
359#define FS_IMAGIC_FL 0x00002000 /* AFS directory */
360#define FS_JOURNAL_DATA_FL 0x00004000 /* Reserved for ext3 */
361#define FS_NOTAIL_FL 0x00008000 /* file tail should not be merged */
362#define FS_DIRSYNC_FL 0x00010000 /* dirsync behaviour (directories only) */
363#define FS_TOPDIR_FL 0x00020000 /* Top of directory hierarchies*/
Steven Whitehouse128e5eb2006-10-02 11:24:43 -0400364#define FS_EXTENT_FL 0x00080000 /* Extents */
365#define FS_DIRECTIO_FL 0x00100000 /* Use direct i/o */
liubo32471f62011-03-21 08:54:27 +0000366#define FS_NOCOW_FL 0x00800000 /* Do not cow file */
David Howells36695672006-08-29 19:06:16 +0100367#define FS_RESERVED_FL 0x80000000 /* reserved for ext2 lib */
368
369#define FS_FL_USER_VISIBLE 0x0003DFFF /* User visible flags */
370#define FS_FL_USER_MODIFIABLE 0x000380FF /* User modifiable flags */
371
372
Andrew Mortone46e4902006-05-22 22:35:24 -0700373#define SYNC_FILE_RANGE_WAIT_BEFORE 1
374#define SYNC_FILE_RANGE_WRITE 2
375#define SYNC_FILE_RANGE_WAIT_AFTER 4
376
Linus Torvalds1da177e2005-04-16 15:20:36 -0700377#ifdef __KERNEL__
378
379#include <linux/linkage.h>
380#include <linux/wait.h>
Linus Torvalds1da177e2005-04-16 15:20:36 -0700381#include <linux/kdev_t.h>
382#include <linux/dcache.h>
Al Viro3f8206d2008-07-26 03:46:43 -0400383#include <linux/path.h>
Linus Torvalds1da177e2005-04-16 15:20:36 -0700384#include <linux/stat.h>
385#include <linux/cache.h>
Linus Torvalds1da177e2005-04-16 15:20:36 -0700386#include <linux/list.h>
387#include <linux/radix-tree.h>
388#include <linux/prio_tree.h>
389#include <linux/init.h>
Al Viro914e2632006-10-18 13:55:46 -0400390#include <linux/pid.h>
Jes Sorensen1b1dcc12006-01-09 15:59:24 -0800391#include <linux/mutex.h>
Satyam Sharma3bd858a2007-07-17 15:00:08 +0530392#include <linux/capability.h>
Matthew Wilcox6188e102008-04-18 22:21:05 -0400393#include <linux/semaphore.h>
Mark Fashehc4b929b2008-10-08 19:44:18 -0400394#include <linux/fiemap.h>
Nick Pigginceb5bdc2011-01-07 17:50:05 +1100395#include <linux/rculist_bl.h>
Dave Chinnerb0d40c92011-07-08 14:14:42 +1000396#include <linux/shrinker.h>
Al Viro07b8ce12011-06-20 10:52:57 -0400397#include <linux/atomic.h>
Linus Torvalds1da177e2005-04-16 15:20:36 -0700398
Linus Torvalds1da177e2005-04-16 15:20:36 -0700399#include <asm/byteorder.h>
400
Christoph Hellwiga5694252007-07-17 04:04:28 -0700401struct export_operations;
Christoph Hellwiga885c8c2006-01-08 01:02:50 -0800402struct hd_geometry;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700403struct iovec;
404struct nameidata;
Christoph Hellwig92198f72005-06-23 22:00:59 -0700405struct kiocb;
Alexey Dobriyan57cc7212011-01-10 08:18:25 +0200406struct kobject;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700407struct pipe_inode_info;
408struct poll_table_struct;
409struct kstatfs;
410struct vm_area_struct;
411struct vfsmount;
David Howells745ca242008-11-14 10:39:22 +1100412struct cred;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700413
Denis Cheng74bf17c2007-10-16 23:26:30 -0700414extern void __init inode_init(void);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700415extern void __init inode_init_early(void);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700416extern void __init files_init(unsigned long);
417
Jan Engelhardtdded4f42008-12-01 14:34:50 -0800418extern struct files_stat_struct files_stat;
Eric Dumazet518de9b2010-10-26 14:22:44 -0700419extern unsigned long get_max_files(void);
Jan Engelhardtdded4f42008-12-01 14:34:50 -0800420extern int sysctl_nr_open;
421extern struct inodes_stat_t inodes_stat;
422extern int leases_enable, lease_break_time;
Jan Engelhardtdded4f42008-12-01 14:34:50 -0800423
Linus Torvalds1da177e2005-04-16 15:20:36 -0700424struct buffer_head;
425typedef int (get_block_t)(struct inode *inode, sector_t iblock,
426 struct buffer_head *bh_result, int create);
Christoph Hellwig92198f72005-06-23 22:00:59 -0700427typedef void (dio_iodone_t)(struct kiocb *iocb, loff_t offset,
Christoph Hellwig40e2e972010-07-18 21:17:09 +0000428 ssize_t bytes, void *private, int ret,
429 bool is_async);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700430
431/*
432 * Attribute flags. These should be or-ed together to figure out what
433 * has been changed!
434 */
Miklos Szeredi9767d742008-07-01 15:01:26 +0200435#define ATTR_MODE (1 << 0)
436#define ATTR_UID (1 << 1)
437#define ATTR_GID (1 << 2)
438#define ATTR_SIZE (1 << 3)
439#define ATTR_ATIME (1 << 4)
440#define ATTR_MTIME (1 << 5)
441#define ATTR_CTIME (1 << 6)
442#define ATTR_ATIME_SET (1 << 7)
443#define ATTR_MTIME_SET (1 << 8)
444#define ATTR_FORCE (1 << 9) /* Not a change, but a change it */
445#define ATTR_ATTR_FLAG (1 << 10)
446#define ATTR_KILL_SUID (1 << 11)
447#define ATTR_KILL_SGID (1 << 12)
448#define ATTR_FILE (1 << 13)
449#define ATTR_KILL_PRIV (1 << 14)
450#define ATTR_OPEN (1 << 15) /* Truncating from open(O_TRUNC) */
451#define ATTR_TIMES_SET (1 << 16)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700452
453/*
454 * This is the Inode Attributes structure, used for notify_change(). It
455 * uses the above definitions as flags, to know which values have changed.
456 * Also, in this manner, a Filesystem can look at only the values it cares
457 * about. Basically, these are the attributes that the VFS layer can
458 * request to change from the FS layer.
459 *
460 * Derek Atkins <warlord@MIT.EDU> 94-10-20
461 */
462struct iattr {
463 unsigned int ia_valid;
464 umode_t ia_mode;
465 uid_t ia_uid;
466 gid_t ia_gid;
467 loff_t ia_size;
468 struct timespec ia_atime;
469 struct timespec ia_mtime;
470 struct timespec ia_ctime;
Miklos Szeredicc4e69d2005-11-07 00:59:49 -0800471
472 /*
Lucas De Marchi25985ed2011-03-30 22:57:33 -0300473 * Not an attribute, but an auxiliary info for filesystems wanting to
Miklos Szeredicc4e69d2005-11-07 00:59:49 -0800474 * implement an ftruncate() like method. NOTE: filesystem should
475 * check for (ia_valid & ATTR_FILE), and not for (ia_file != NULL).
476 */
477 struct file *ia_file;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700478};
479
480/*
Linus Torvalds1da177e2005-04-16 15:20:36 -0700481 * Includes for diskquotas.
482 */
483#include <linux/quota.h>
484
Zach Brown994fc28c2005-12-15 14:28:17 -0800485/**
486 * enum positive_aop_returns - aop return codes with specific semantics
487 *
488 * @AOP_WRITEPAGE_ACTIVATE: Informs the caller that page writeback has
489 * completed, that the page is still locked, and
490 * should be considered active. The VM uses this hint
491 * to return the page to the active list -- it won't
492 * be a candidate for writeback again in the near
493 * future. Other callers must be careful to unlock
494 * the page if they get this return. Returned by
495 * writepage();
496 *
497 * @AOP_TRUNCATED_PAGE: The AOP method that was handed a locked page has
498 * unlocked it and the page might have been truncated.
499 * The caller should back up to acquiring a new page and
500 * trying again. The aop will be taking reasonable
501 * precautions not to livelock. If the caller held a page
502 * reference, it should drop it before retrying. Returned
Nick Piggin55144762007-10-16 01:25:26 -0700503 * by readpage().
Zach Brown994fc28c2005-12-15 14:28:17 -0800504 *
505 * address_space_operation functions return these large constants to indicate
506 * special semantics to the caller. These are much larger than the bytes in a
507 * page to allow for functions that return the number of bytes operated on in a
508 * given page.
509 */
510
511enum positive_aop_returns {
512 AOP_WRITEPAGE_ACTIVATE = 0x80000,
513 AOP_TRUNCATED_PAGE = 0x80001,
514};
515
Nick Pigginafddba42007-10-16 01:25:01 -0700516#define AOP_FLAG_UNINTERRUPTIBLE 0x0001 /* will not do a short write */
Nick Piggin89e10782007-10-16 01:25:07 -0700517#define AOP_FLAG_CONT_EXPAND 0x0002 /* called from cont_expand */
Nick Piggin54566b22009-01-04 12:00:53 -0800518#define AOP_FLAG_NOFS 0x0004 /* used by filesystem to direct
519 * helper code (eg buffer layer)
520 * to clear GFP_FS from alloc */
Nick Pigginafddba42007-10-16 01:25:01 -0700521
Linus Torvalds1da177e2005-04-16 15:20:36 -0700522/*
523 * oh the beauties of C type declarations.
524 */
525struct page;
526struct address_space;
527struct writeback_control;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700528
Nick Piggin2f718ff2007-10-16 01:24:59 -0700529struct iov_iter {
530 const struct iovec *iov;
531 unsigned long nr_segs;
532 size_t iov_offset;
533 size_t count;
534};
535
536size_t iov_iter_copy_from_user_atomic(struct page *page,
537 struct iov_iter *i, unsigned long offset, size_t bytes);
538size_t iov_iter_copy_from_user(struct page *page,
539 struct iov_iter *i, unsigned long offset, size_t bytes);
540void iov_iter_advance(struct iov_iter *i, size_t bytes);
Nick Pigginafddba42007-10-16 01:25:01 -0700541int iov_iter_fault_in_readable(struct iov_iter *i, size_t bytes);
Nick Piggin2f718ff2007-10-16 01:24:59 -0700542size_t iov_iter_single_seg_count(struct iov_iter *i);
543
544static inline void iov_iter_init(struct iov_iter *i,
545 const struct iovec *iov, unsigned long nr_segs,
546 size_t count, size_t written)
547{
548 i->iov = iov;
549 i->nr_segs = nr_segs;
550 i->iov_offset = 0;
551 i->count = count + written;
552
553 iov_iter_advance(i, written);
554}
555
556static inline size_t iov_iter_count(struct iov_iter *i)
557{
558 return i->count;
559}
560
Hisashi Hifumi8ab22b92008-07-28 15:46:36 -0700561/*
562 * "descriptor" for what we're up to with a read.
563 * This allows us to use the same read code yet
564 * have multiple different users of the data that
565 * we read from a file.
566 *
567 * The simplest case just copies the data to user
568 * mode.
569 */
570typedef struct {
571 size_t written;
572 size_t count;
573 union {
574 char __user *buf;
575 void *data;
576 } arg;
577 int error;
578} read_descriptor_t;
579
580typedef int (*read_actor_t)(read_descriptor_t *, struct page *,
581 unsigned long, unsigned long);
Nick Piggin2f718ff2007-10-16 01:24:59 -0700582
Linus Torvalds1da177e2005-04-16 15:20:36 -0700583struct address_space_operations {
584 int (*writepage)(struct page *page, struct writeback_control *wbc);
585 int (*readpage)(struct file *, struct page *);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700586
587 /* Write back some dirty pages from this mapping. */
588 int (*writepages)(struct address_space *, struct writeback_control *);
589
Andrew Morton4741c9f2006-03-24 03:18:11 -0800590 /* Set a page dirty. Return true if this dirtied it */
Linus Torvalds1da177e2005-04-16 15:20:36 -0700591 int (*set_page_dirty)(struct page *page);
592
593 int (*readpages)(struct file *filp, struct address_space *mapping,
594 struct list_head *pages, unsigned nr_pages);
595
Nick Pigginafddba42007-10-16 01:25:01 -0700596 int (*write_begin)(struct file *, struct address_space *mapping,
597 loff_t pos, unsigned len, unsigned flags,
598 struct page **pagep, void **fsdata);
599 int (*write_end)(struct file *, struct address_space *mapping,
600 loff_t pos, unsigned len, unsigned copied,
601 struct page *page, void *fsdata);
602
Linus Torvalds1da177e2005-04-16 15:20:36 -0700603 /* Unfortunately this kludge is needed for FIBMAP. Don't use it */
604 sector_t (*bmap)(struct address_space *, sector_t);
NeilBrown2ff28e22006-03-26 01:37:18 -0800605 void (*invalidatepage) (struct page *, unsigned long);
Al Viro27496a82005-10-21 03:20:48 -0400606 int (*releasepage) (struct page *, gfp_t);
Linus Torvalds6072d132010-12-01 13:35:19 -0500607 void (*freepage)(struct page *);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700608 ssize_t (*direct_IO)(int, struct kiocb *, const struct iovec *iov,
609 loff_t offset, unsigned long nr_segs);
Nick Piggin70688e42008-04-28 02:13:02 -0700610 int (*get_xip_mem)(struct address_space *, pgoff_t, int,
611 void **, unsigned long *);
Christoph Lametere965f962006-02-01 03:05:41 -0800612 /* migrate the contents of a page to the specified target */
Christoph Lameter2d1db3b2006-06-23 02:03:33 -0700613 int (*migratepage) (struct address_space *,
614 struct page *, struct page *);
Trond Myklebuste3db7692007-01-10 23:15:39 -0800615 int (*launder_page) (struct page *);
Hisashi Hifumi8ab22b92008-07-28 15:46:36 -0700616 int (*is_partially_uptodate) (struct page *, read_descriptor_t *,
617 unsigned long);
Andi Kleen25718732009-09-16 11:50:13 +0200618 int (*error_remove_page)(struct address_space *, struct page *);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700619};
620
Jens Axboe7dcda1c2011-04-05 23:51:48 +0200621extern const struct address_space_operations empty_aops;
622
Nick Pigginafddba42007-10-16 01:25:01 -0700623/*
624 * pagecache_write_begin/pagecache_write_end must be used by general code
625 * to write into the pagecache.
626 */
627int pagecache_write_begin(struct file *, struct address_space *mapping,
628 loff_t pos, unsigned len, unsigned flags,
629 struct page **pagep, void **fsdata);
630
631int pagecache_write_end(struct file *, struct address_space *mapping,
632 loff_t pos, unsigned len, unsigned copied,
633 struct page *page, void *fsdata);
634
Linus Torvalds1da177e2005-04-16 15:20:36 -0700635struct backing_dev_info;
636struct address_space {
637 struct inode *host; /* owner: inode, block_device */
638 struct radix_tree_root page_tree; /* radix tree of all pages */
Nick Piggin19fd6232008-07-25 19:45:32 -0700639 spinlock_t tree_lock; /* and lock protecting it */
Linus Torvalds1da177e2005-04-16 15:20:36 -0700640 unsigned int i_mmap_writable;/* count VM_SHARED mappings */
641 struct prio_tree_root i_mmap; /* tree of private and shared mappings */
642 struct list_head i_mmap_nonlinear;/*list VM_NONLINEAR mappings */
Peter Zijlstra3d48ae42011-05-24 17:12:06 -0700643 struct mutex i_mmap_mutex; /* protect tree, count, list */
Jan Kara08142572011-06-27 16:18:10 -0700644 /* Protected by tree_lock together with the radix tree */
Linus Torvalds1da177e2005-04-16 15:20:36 -0700645 unsigned long nrpages; /* number of total pages */
646 pgoff_t writeback_index;/* writeback starts here */
Christoph Hellwigf5e54d62006-06-28 04:26:44 -0700647 const struct address_space_operations *a_ops; /* methods */
Linus Torvalds1da177e2005-04-16 15:20:36 -0700648 unsigned long flags; /* error bits/gfp mask */
649 struct backing_dev_info *backing_dev_info; /* device readahead, etc */
650 spinlock_t private_lock; /* for use by the address_space */
651 struct list_head private_list; /* ditto */
652 struct address_space *assoc_mapping; /* ditto */
653} __attribute__((aligned(sizeof(long))));
654 /*
655 * On most architectures that alignment is already the case; but
Lucas De Marchi25985ed2011-03-30 22:57:33 -0300656 * must be enforced here for CRIS, to let the least significant bit
Linus Torvalds1da177e2005-04-16 15:20:36 -0700657 * of struct page's "mapping" pointer be used for PAGE_MAPPING_ANON.
658 */
659
660struct block_device {
661 dev_t bd_dev; /* not a kdev_t - it's a search key */
Richard Kennedyf3ccfcd2011-03-22 16:33:56 -0700662 int bd_openers;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700663 struct inode * bd_inode; /* will die */
Theodore Ts'o87d8fe12009-01-03 09:47:09 -0500664 struct super_block * bd_super;
Arjan van de Venc039e312006-03-23 03:00:28 -0800665 struct mutex bd_mutex; /* open/close mutex */
Linus Torvalds1da177e2005-04-16 15:20:36 -0700666 struct list_head bd_inodes;
Tejun Heo6b4517a2010-04-07 18:53:59 +0900667 void * bd_claiming;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700668 void * bd_holder;
669 int bd_holders;
Tejun Heo77ea8872010-12-08 20:57:37 +0100670 bool bd_write_holder;
Jun'ichi Nomura641dc632006-03-27 01:17:57 -0800671#ifdef CONFIG_SYSFS
Tejun Heo49731ba2011-01-14 18:43:57 +0100672 struct list_head bd_holder_disks;
Jun'ichi Nomura641dc632006-03-27 01:17:57 -0800673#endif
Linus Torvalds1da177e2005-04-16 15:20:36 -0700674 struct block_device * bd_contains;
675 unsigned bd_block_size;
676 struct hd_struct * bd_part;
677 /* number of times partitions within this device have been opened. */
678 unsigned bd_part_count;
679 int bd_invalidated;
680 struct gendisk * bd_disk;
681 struct list_head bd_list;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700682 /*
683 * Private data. You must have bd_claim'ed the block_device
684 * to use this. NOTE: bd_claim allows an owner to claim
685 * the same device multiple times, the owner must take special
686 * care to not mess up bd_private for that case.
687 */
688 unsigned long bd_private;
Takashi Satofcccf502009-01-09 16:40:59 -0800689
690 /* The counter of freeze processes */
691 int bd_fsfreeze_count;
692 /* Mutex for freeze */
693 struct mutex bd_fsfreeze_mutex;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700694};
695
696/*
697 * Radix-tree tags, for tagging dirty and writeback pages within the pagecache
698 * radix trees
699 */
700#define PAGECACHE_TAG_DIRTY 0
701#define PAGECACHE_TAG_WRITEBACK 1
Jan Karaf446daae2010-08-09 17:19:12 -0700702#define PAGECACHE_TAG_TOWRITE 2
Linus Torvalds1da177e2005-04-16 15:20:36 -0700703
704int mapping_tagged(struct address_space *mapping, int tag);
705
706/*
707 * Might pages of this file be mapped into userspace?
708 */
709static inline int mapping_mapped(struct address_space *mapping)
710{
711 return !prio_tree_empty(&mapping->i_mmap) ||
712 !list_empty(&mapping->i_mmap_nonlinear);
713}
714
715/*
716 * Might pages of this file have been modified in userspace?
717 * Note that i_mmap_writable counts all VM_SHARED vmas: do_mmap_pgoff
718 * marks vma as VM_SHARED if it is shared, and the file was opened for
719 * writing i.e. vma may be mprotected writable even if now readonly.
720 */
721static inline int mapping_writably_mapped(struct address_space *mapping)
722{
723 return mapping->i_mmap_writable != 0;
724}
725
726/*
727 * Use sequence counter to get consistent i_size on 32-bit processors.
728 */
729#if BITS_PER_LONG==32 && defined(CONFIG_SMP)
730#include <linux/seqlock.h>
731#define __NEED_I_SIZE_ORDERED
732#define i_size_ordered_init(inode) seqcount_init(&inode->i_size_seqcount)
733#else
734#define i_size_ordered_init(inode) do { } while (0)
735#endif
736
Al Virof19d4a82009-06-08 19:50:45 -0400737struct posix_acl;
738#define ACL_NOT_CACHED ((void *)(-1))
739
Linus Torvalds3ddcd052011-08-06 22:45:50 -0700740#define IOP_FASTPERM 0x0001
741#define IOP_LOOKUP 0x0002
742#define IOP_NOFOLLOW 0x0004
743
744/*
745 * Keep mostly read-only and often accessed (especially for
746 * the RCU path lookup and 'stat' data) fields at the beginning
747 * of the 'struct inode'
748 */
Linus Torvalds1da177e2005-04-16 15:20:36 -0700749struct inode {
Nick Piggin44a7d7a2011-01-07 17:49:56 +1100750 umode_t i_mode;
Linus Torvalds3ddcd052011-08-06 22:45:50 -0700751 unsigned short i_opflags;
Nick Piggin44a7d7a2011-01-07 17:49:56 +1100752 uid_t i_uid;
753 gid_t i_gid;
Linus Torvalds3ddcd052011-08-06 22:45:50 -0700754 unsigned int i_flags;
755
756#ifdef CONFIG_FS_POSIX_ACL
757 struct posix_acl *i_acl;
758 struct posix_acl *i_default_acl;
759#endif
760
Nick Piggin44a7d7a2011-01-07 17:49:56 +1100761 const struct inode_operations *i_op;
762 struct super_block *i_sb;
Linus Torvalds3ddcd052011-08-06 22:45:50 -0700763 struct address_space *i_mapping;
Nick Piggin44a7d7a2011-01-07 17:49:56 +1100764
Linus Torvalds13e12d12011-06-08 15:18:19 -0700765#ifdef CONFIG_SECURITY
766 void *i_security;
767#endif
Nick Piggin44a7d7a2011-01-07 17:49:56 +1100768
Linus Torvalds3ddcd052011-08-06 22:45:50 -0700769 /* Stat data, not accessed from path walking */
770 unsigned long i_ino;
771 unsigned int i_nlink;
772 dev_t i_rdev;
773 loff_t i_size;
774 struct timespec i_atime;
775 struct timespec i_mtime;
776 struct timespec i_ctime;
777 unsigned int i_blkbits;
778 blkcnt_t i_blocks;
779
780#ifdef __NEED_I_SIZE_ORDERED
781 seqcount_t i_size_seqcount;
782#endif
783
784 /* Misc */
785 unsigned long i_state;
786 spinlock_t i_lock; /* i_blocks, i_bytes, maybe i_size */
787 struct mutex i_mutex;
Linus Torvalds13e12d12011-06-08 15:18:19 -0700788
Nick Piggin44a7d7a2011-01-07 17:49:56 +1100789 unsigned long dirtied_when; /* jiffies of first dirtying */
790
Linus Torvalds1da177e2005-04-16 15:20:36 -0700791 struct hlist_node i_hash;
Nick Piggin7ccf19a2010-10-21 11:49:30 +1100792 struct list_head i_wb_list; /* backing dev IO list */
793 struct list_head i_lru; /* inode LRU list */
Linus Torvalds1da177e2005-04-16 15:20:36 -0700794 struct list_head i_sb_list;
Nick Pigginfa0d7e3d2011-01-07 17:49:49 +1100795 union {
796 struct list_head i_dentry;
797 struct rcu_head i_rcu;
798 };
Linus Torvalds1da177e2005-04-16 15:20:36 -0700799 atomic_t i_count;
Jean Noel Cordenner7a224222008-01-28 23:58:27 -0500800 u64 i_version;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700801 unsigned short i_bytes;
Christoph Hellwigbd5fe6c2011-06-24 14:29:43 -0400802 atomic_t i_dio_count;
Arjan van de Ven99ac48f2006-03-28 01:56:41 -0800803 const struct file_operations *i_fop; /* former ->i_op->default_file_ops */
Linus Torvalds1da177e2005-04-16 15:20:36 -0700804 struct file_lock *i_flock;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700805 struct address_space i_data;
806#ifdef CONFIG_QUOTA
807 struct dquot *i_dquot[MAXQUOTAS];
808#endif
Linus Torvalds1da177e2005-04-16 15:20:36 -0700809 struct list_head i_devices;
Theodore Ts'o4c154162006-09-27 01:50:47 -0700810 union {
811 struct pipe_inode_info *i_pipe;
Theodore Ts'oeaf796e2006-09-27 01:50:48 -0700812 struct block_device *i_bdev;
Theodore Ts'o577c4eb2006-09-27 01:50:49 -0700813 struct cdev *i_cdev;
Theodore Ts'o4c154162006-09-27 01:50:47 -0700814 };
Linus Torvalds1da177e2005-04-16 15:20:36 -0700815
816 __u32 i_generation;
817
Eric Paris3be25f42009-05-21 17:01:26 -0400818#ifdef CONFIG_FSNOTIFY
819 __u32 i_fsnotify_mask; /* all events this inode cares about */
Eric Parise61ce862009-12-17 21:24:24 -0500820 struct hlist_head i_fsnotify_marks;
Robert Love0eeca282005-07-12 17:06:03 -0400821#endif
822
Eric Parisa178d202010-10-25 14:41:59 -0400823#ifdef CONFIG_IMA
Mimi Zohara68a27b2010-11-02 10:10:56 -0400824 atomic_t i_readcount; /* struct files open RO */
Eric Parisa178d202010-10-25 14:41:59 -0400825#endif
Linus Torvalds1da177e2005-04-16 15:20:36 -0700826 atomic_t i_writecount;
Theodore Ts'o8e18e292006-09-27 01:50:46 -0700827 void *i_private; /* fs or device private pointer */
Linus Torvalds1da177e2005-04-16 15:20:36 -0700828};
829
Al Viro1d3382cb2010-10-23 15:19:20 -0400830static inline int inode_unhashed(struct inode *inode)
831{
832 return hlist_unhashed(&inode->i_hash);
833}
834
Linus Torvalds1da177e2005-04-16 15:20:36 -0700835/*
Ingo Molnarf2eace22006-07-03 00:25:05 -0700836 * inode->i_mutex nesting subclasses for the lock validator:
837 *
838 * 0: the object of the current VFS operation
839 * 1: parent
840 * 2: child/target
841 * 3: quota file
842 *
843 * The locking order between these classes is
Ingo Molnar4df46242006-08-27 01:23:56 -0700844 * parent -> child -> normal -> xattr -> quota
Ingo Molnarf2eace22006-07-03 00:25:05 -0700845 */
846enum inode_i_mutex_lock_class
847{
848 I_MUTEX_NORMAL,
849 I_MUTEX_PARENT,
850 I_MUTEX_CHILD,
Ingo Molnar4df46242006-08-27 01:23:56 -0700851 I_MUTEX_XATTR,
Ingo Molnarf2eace22006-07-03 00:25:05 -0700852 I_MUTEX_QUOTA
853};
854
855/*
Linus Torvalds1da177e2005-04-16 15:20:36 -0700856 * NOTE: in a 32bit arch with a preemptable kernel and
857 * an UP compile the i_size_read/write must be atomic
858 * with respect to the local cpu (unlike with preempt disabled),
859 * but they don't need to be atomic with respect to other cpus like in
860 * true SMP (so they need either to either locally disable irq around
861 * the read or for example on x86 they can be still implemented as a
862 * cmpxchg8b without the need of the lock prefix). For SMP compiles
863 * and 64bit archs it makes no difference if preempt is enabled or not.
864 */
Jan Engelhardt48ed2142006-12-06 20:35:37 -0800865static inline loff_t i_size_read(const struct inode *inode)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700866{
867#if BITS_PER_LONG==32 && defined(CONFIG_SMP)
868 loff_t i_size;
869 unsigned int seq;
870
871 do {
872 seq = read_seqcount_begin(&inode->i_size_seqcount);
873 i_size = inode->i_size;
874 } while (read_seqcount_retry(&inode->i_size_seqcount, seq));
875 return i_size;
876#elif BITS_PER_LONG==32 && defined(CONFIG_PREEMPT)
877 loff_t i_size;
878
879 preempt_disable();
880 i_size = inode->i_size;
881 preempt_enable();
882 return i_size;
883#else
884 return inode->i_size;
885#endif
886}
887
Miklos Szeredi7762f5a2006-10-17 00:10:07 -0700888/*
889 * NOTE: unlike i_size_read(), i_size_write() does need locking around it
890 * (normally i_mutex), otherwise on 32bit/SMP an update of i_size_seqcount
891 * can be lost, resulting in subsequent i_size_read() calls spinning forever.
892 */
Linus Torvalds1da177e2005-04-16 15:20:36 -0700893static inline void i_size_write(struct inode *inode, loff_t i_size)
894{
895#if BITS_PER_LONG==32 && defined(CONFIG_SMP)
896 write_seqcount_begin(&inode->i_size_seqcount);
897 inode->i_size = i_size;
898 write_seqcount_end(&inode->i_size_seqcount);
899#elif BITS_PER_LONG==32 && defined(CONFIG_PREEMPT)
900 preempt_disable();
901 inode->i_size = i_size;
902 preempt_enable();
903#else
904 inode->i_size = i_size;
905#endif
906}
907
Jan Engelhardt48ed2142006-12-06 20:35:37 -0800908static inline unsigned iminor(const struct inode *inode)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700909{
910 return MINOR(inode->i_rdev);
911}
912
Jan Engelhardt48ed2142006-12-06 20:35:37 -0800913static inline unsigned imajor(const struct inode *inode)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700914{
915 return MAJOR(inode->i_rdev);
916}
917
918extern struct block_device *I_BDEV(struct inode *inode);
919
920struct fown_struct {
921 rwlock_t lock; /* protects pid, uid, euid fields */
Eric W. Biederman609d7fa2006-10-02 02:17:15 -0700922 struct pid *pid; /* pid or -pgrp where SIGIO should be sent */
923 enum pid_type pid_type; /* Kind of process group SIGIO should be sent to */
Linus Torvalds1da177e2005-04-16 15:20:36 -0700924 uid_t uid, euid; /* uid/euid of process setting the owner */
Linus Torvalds1da177e2005-04-16 15:20:36 -0700925 int signum; /* posix.1b rt signal to be delivered on IO */
926};
927
928/*
929 * Track a single file's readahead state
930 */
931struct file_ra_state {
Fengguang Wu937085a2007-10-16 01:24:31 -0700932 pgoff_t start; /* where readahead started */
933 unsigned int size; /* # of readahead pages */
934 unsigned int async_size; /* do asynchronous readahead when
Fengguang Wuf9acc8c2007-07-19 01:48:08 -0700935 there are only # of pages ahead */
Fengguang Wu5ce11102007-07-19 01:47:59 -0700936
Fengguang Wu937085a2007-10-16 01:24:31 -0700937 unsigned int ra_pages; /* Maximum readahead window */
Wu Fengguang1ebf26a2009-06-16 15:31:19 -0700938 unsigned int mmap_miss; /* Cache miss stat for mmap accesses */
Fengguang Wuf4e6b492007-10-16 01:24:33 -0700939 loff_t prev_pos; /* Cache last read() position */
Linus Torvalds1da177e2005-04-16 15:20:36 -0700940};
Linus Torvalds1da177e2005-04-16 15:20:36 -0700941
Fengguang Wu5ce11102007-07-19 01:47:59 -0700942/*
Fengguang Wu5ce11102007-07-19 01:47:59 -0700943 * Check if @index falls in the readahead windows.
944 */
945static inline int ra_has_index(struct file_ra_state *ra, pgoff_t index)
946{
Fengguang Wuf9acc8c2007-07-19 01:48:08 -0700947 return (index >= ra->start &&
948 index < ra->start + ra->size);
Fengguang Wu5ce11102007-07-19 01:47:59 -0700949}
950
Dave Hansenad775f52008-02-15 14:38:01 -0800951#define FILE_MNT_WRITE_TAKEN 1
952#define FILE_MNT_WRITE_RELEASED 2
953
Linus Torvalds1da177e2005-04-16 15:20:36 -0700954struct file {
Eric Dumazet2f512012005-10-30 15:02:16 -0800955 /*
956 * fu_list becomes invalid after file_free is called and queued via
957 * fu_rcuhead for RCU freeing
958 */
959 union {
960 struct list_head fu_list;
961 struct rcu_head fu_rcuhead;
962 } f_u;
Josef "Jeff" Sipek0f7fc9e2006-12-08 02:36:35 -0800963 struct path f_path;
964#define f_dentry f_path.dentry
965#define f_vfsmnt f_path.mnt
Arjan van de Ven99ac48f2006-03-28 01:56:41 -0800966 const struct file_operations *f_op;
Andi Kleenef3d0fd2011-09-15 16:06:48 -0700967
968 /*
969 * Protects f_ep_links, f_flags, f_pos vs i_size in lseek SEEK_CUR.
970 * Must not be taken from IRQ context.
971 */
972 spinlock_t f_lock;
Nick Piggin6416ccb2010-08-18 04:37:38 +1000973#ifdef CONFIG_SMP
974 int f_sb_list_cpu;
975#endif
Al Viro516e0cc2008-07-26 00:39:17 -0400976 atomic_long_t f_count;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700977 unsigned int f_flags;
Al Viroaeb5d722008-09-02 15:28:45 -0400978 fmode_t f_mode;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700979 loff_t f_pos;
980 struct fown_struct f_owner;
David Howellsd76b0d92008-11-14 10:39:25 +1100981 const struct cred *f_cred;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700982 struct file_ra_state f_ra;
983
Mathieu Desnoyers2b47c362007-10-16 23:27:21 -0700984 u64 f_version;
Alexey Dobriyan50462062006-09-29 02:00:01 -0700985#ifdef CONFIG_SECURITY
Linus Torvalds1da177e2005-04-16 15:20:36 -0700986 void *f_security;
Alexey Dobriyan50462062006-09-29 02:00:01 -0700987#endif
Linus Torvalds1da177e2005-04-16 15:20:36 -0700988 /* needed for tty driver, and maybe others */
989 void *private_data;
990
991#ifdef CONFIG_EPOLL
992 /* Used by fs/eventpoll.c to link all the hooks to this file */
993 struct list_head f_ep_links;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700994#endif /* #ifdef CONFIG_EPOLL */
995 struct address_space *f_mapping;
Dave Hansenad775f52008-02-15 14:38:01 -0800996#ifdef CONFIG_DEBUG_WRITECOUNT
997 unsigned long f_mnt_write_state;
998#endif
Linus Torvalds1da177e2005-04-16 15:20:36 -0700999};
Linus Torvalds1da177e2005-04-16 15:20:36 -07001000
Aneesh Kumar K.V990d6c22011-01-29 18:43:26 +05301001struct file_handle {
1002 __u32 handle_bytes;
1003 int handle_type;
1004 /* file identifier */
1005 unsigned char f_handle[0];
1006};
1007
Al Viro516e0cc2008-07-26 00:39:17 -04001008#define get_file(x) atomic_long_inc(&(x)->f_count)
Al Virod7065da2010-05-26 15:13:55 -04001009#define fput_atomic(x) atomic_long_add_unless(&(x)->f_count, -1, 1)
Al Viro516e0cc2008-07-26 00:39:17 -04001010#define file_count(x) atomic_long_read(&(x)->f_count)
Linus Torvalds1da177e2005-04-16 15:20:36 -07001011
Dave Hansenad775f52008-02-15 14:38:01 -08001012#ifdef CONFIG_DEBUG_WRITECOUNT
1013static inline void file_take_write(struct file *f)
1014{
1015 WARN_ON(f->f_mnt_write_state != 0);
1016 f->f_mnt_write_state = FILE_MNT_WRITE_TAKEN;
1017}
1018static inline void file_release_write(struct file *f)
1019{
1020 f->f_mnt_write_state |= FILE_MNT_WRITE_RELEASED;
1021}
1022static inline void file_reset_write(struct file *f)
1023{
1024 f->f_mnt_write_state = 0;
1025}
1026static inline void file_check_state(struct file *f)
1027{
1028 /*
1029 * At this point, either both or neither of these bits
1030 * should be set.
1031 */
1032 WARN_ON(f->f_mnt_write_state == FILE_MNT_WRITE_TAKEN);
1033 WARN_ON(f->f_mnt_write_state == FILE_MNT_WRITE_RELEASED);
1034}
1035static inline int file_check_writeable(struct file *f)
1036{
1037 if (f->f_mnt_write_state == FILE_MNT_WRITE_TAKEN)
1038 return 0;
1039 printk(KERN_WARNING "writeable file with no "
1040 "mnt_want_write()\n");
1041 WARN_ON(1);
1042 return -EINVAL;
1043}
1044#else /* !CONFIG_DEBUG_WRITECOUNT */
1045static inline void file_take_write(struct file *filp) {}
1046static inline void file_release_write(struct file *filp) {}
1047static inline void file_reset_write(struct file *filp) {}
1048static inline void file_check_state(struct file *filp) {}
1049static inline int file_check_writeable(struct file *filp)
1050{
1051 return 0;
1052}
1053#endif /* CONFIG_DEBUG_WRITECOUNT */
1054
Linus Torvalds1da177e2005-04-16 15:20:36 -07001055#define MAX_NON_LFS ((1UL<<31) - 1)
1056
1057/* Page cache limit. The filesystems should put that into their s_maxbytes
1058 limits, otherwise bad things can happen in VM. */
1059#if BITS_PER_LONG==32
1060#define MAX_LFS_FILESIZE (((u64)PAGE_CACHE_SIZE << (BITS_PER_LONG-1))-1)
1061#elif BITS_PER_LONG==64
1062#define MAX_LFS_FILESIZE 0x7fffffffffffffffUL
1063#endif
1064
1065#define FL_POSIX 1
1066#define FL_FLOCK 2
1067#define FL_ACCESS 8 /* not trying to lock, just looking */
Trond Myklebustf475ae92006-06-29 16:38:32 -04001068#define FL_EXISTS 16 /* when unlocking, test for existence */
Linus Torvalds1da177e2005-04-16 15:20:36 -07001069#define FL_LEASE 32 /* lease held on this file */
Miklos Szeredi75e1fcc2006-06-23 02:05:12 -07001070#define FL_CLOSE 64 /* unlock on close */
Linus Torvalds1da177e2005-04-16 15:20:36 -07001071#define FL_SLEEP 128 /* A blocking lock */
J. Bruce Fields778fc542011-07-26 18:25:49 -04001072#define FL_DOWNGRADE_PENDING 256 /* Lease is being downgraded */
1073#define FL_UNLOCK_PENDING 512 /* Lease is being broken */
Linus Torvalds1da177e2005-04-16 15:20:36 -07001074
1075/*
Miklos Szeredibde74e42008-07-25 01:48:57 -07001076 * Special return value from posix_lock_file() and vfs_lock_file() for
1077 * asynchronous locking.
1078 */
1079#define FILE_LOCK_DEFERRED 1
1080
1081/*
Linus Torvalds1da177e2005-04-16 15:20:36 -07001082 * The POSIX file lock owner is determined by
1083 * the "struct files_struct" in the thread group
1084 * (or NULL for no owner - BSD locks).
1085 *
1086 * Lockd stuffs a "host" pointer into this.
1087 */
1088typedef struct files_struct *fl_owner_t;
1089
1090struct file_lock_operations {
Linus Torvalds1da177e2005-04-16 15:20:36 -07001091 void (*fl_copy_lock)(struct file_lock *, struct file_lock *);
1092 void (*fl_release_private)(struct file_lock *);
1093};
1094
1095struct lock_manager_operations {
J. Bruce Fields8fb47a42011-07-20 20:21:59 -04001096 int (*lm_compare_owner)(struct file_lock *, struct file_lock *);
1097 void (*lm_notify)(struct file_lock *); /* unblock callback */
1098 int (*lm_grant)(struct file_lock *, struct file_lock *, int);
1099 void (*lm_release_private)(struct file_lock *);
1100 void (*lm_break)(struct file_lock *);
1101 int (*lm_change)(struct file_lock **, int);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001102};
1103
J. Bruce Fieldsaf558e32007-09-06 12:34:25 -04001104struct lock_manager {
1105 struct list_head list;
1106};
1107
1108void locks_start_grace(struct lock_manager *);
1109void locks_end_grace(struct lock_manager *);
1110int locks_in_grace(void);
1111
Linus Torvalds1da177e2005-04-16 15:20:36 -07001112/* that will die - we need it for nfs_lock_info */
1113#include <linux/nfs_fs_i.h>
1114
1115struct file_lock {
1116 struct file_lock *fl_next; /* singly linked list for this inode */
1117 struct list_head fl_link; /* doubly linked list of all locks */
1118 struct list_head fl_block; /* circular list of blocked processes */
1119 fl_owner_t fl_owner;
J. Bruce Fields710b7212011-07-26 16:28:29 -04001120 unsigned int fl_flags;
Richard Kennedyafc12462008-07-11 17:20:49 -07001121 unsigned char fl_type;
Linus Torvalds1da177e2005-04-16 15:20:36 -07001122 unsigned int fl_pid;
Vitaliy Gusevab1f1612008-01-17 00:07:08 +00001123 struct pid *fl_nspid;
Linus Torvalds1da177e2005-04-16 15:20:36 -07001124 wait_queue_head_t fl_wait;
1125 struct file *fl_file;
Linus Torvalds1da177e2005-04-16 15:20:36 -07001126 loff_t fl_start;
1127 loff_t fl_end;
1128
1129 struct fasync_struct * fl_fasync; /* for lease break notifications */
J. Bruce Fields778fc542011-07-26 18:25:49 -04001130 /* for lease breaks: */
1131 unsigned long fl_break_time;
1132 unsigned long fl_downgrade_time;
Linus Torvalds1da177e2005-04-16 15:20:36 -07001133
Alexey Dobriyan6aed6282009-09-21 17:01:11 -07001134 const struct file_lock_operations *fl_ops; /* Callbacks for filesystems */
Alexey Dobriyan7b021962009-09-21 17:01:12 -07001135 const struct lock_manager_operations *fl_lmops; /* Callbacks for lockmanagers */
Linus Torvalds1da177e2005-04-16 15:20:36 -07001136 union {
1137 struct nfs_lock_info nfs_fl;
Trond Myklebust8d0a8a92005-06-22 17:16:32 +00001138 struct nfs4_lock_info nfs4_fl;
David Howellse8d6c552007-07-15 23:40:12 -07001139 struct {
1140 struct list_head link; /* link in AFS vnode's pending_locks list */
1141 int state; /* state of grant or error if -ve */
1142 } afs;
Linus Torvalds1da177e2005-04-16 15:20:36 -07001143 } fl_u;
1144};
1145
1146/* The following constant reflects the upper bound of the file/locking space */
1147#ifndef OFFSET_MAX
1148#define INT_LIMIT(x) (~((x)1 << (sizeof(x)*8 - 1)))
1149#define OFFSET_MAX INT_LIMIT(loff_t)
1150#define OFFT_OFFSET_MAX INT_LIMIT(off_t)
1151#endif
1152
Linus Torvalds1da177e2005-04-16 15:20:36 -07001153#include <linux/fcntl.h>
1154
Thomas Petazzonibfcd17a2008-08-06 15:12:22 +02001155extern void send_sigio(struct fown_struct *fown, int fd, int band);
1156
Thomas Petazzonibfcd17a2008-08-06 15:12:22 +02001157#ifdef CONFIG_FILE_LOCKING
Linus Torvalds1da177e2005-04-16 15:20:36 -07001158extern int fcntl_getlk(struct file *, struct flock __user *);
Peter Staubachc2936212005-07-27 11:45:09 -07001159extern int fcntl_setlk(unsigned int, struct file *, unsigned int,
1160 struct flock __user *);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001161
1162#if BITS_PER_LONG == 32
1163extern int fcntl_getlk64(struct file *, struct flock64 __user *);
Peter Staubachc2936212005-07-27 11:45:09 -07001164extern int fcntl_setlk64(unsigned int, struct file *, unsigned int,
1165 struct flock64 __user *);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001166#endif
1167
Linus Torvalds1da177e2005-04-16 15:20:36 -07001168extern int fcntl_setlease(unsigned int fd, struct file *filp, long arg);
1169extern int fcntl_getlease(struct file *filp);
1170
1171/* fs/locks.c */
J. Bruce Fields05fa3132010-10-30 17:31:15 -04001172void locks_free_lock(struct file_lock *fl);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001173extern void locks_init_lock(struct file_lock *);
Arnd Bergmannc5b1f0d2010-10-27 15:46:08 +02001174extern struct file_lock * locks_alloc_lock(void);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001175extern void locks_copy_lock(struct file_lock *, struct file_lock *);
J. Bruce Fields1a747ee2008-04-24 10:08:22 -04001176extern void __locks_copy_lock(struct file_lock *, const struct file_lock *);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001177extern void locks_remove_posix(struct file *, fl_owner_t);
1178extern void locks_remove_flock(struct file *);
Felix Blyakhera9e61e22009-03-31 15:12:56 -05001179extern void locks_release_private(struct file_lock *);
J. Bruce Fields6d34ac12007-05-11 16:09:32 -04001180extern void posix_test_lock(struct file *, struct file_lock *);
Marc Eshel150b3932007-01-18 16:15:35 -05001181extern int posix_lock_file(struct file *, struct file_lock *, struct file_lock *);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001182extern int posix_lock_file_wait(struct file *, struct file_lock *);
J. Bruce Fields64a318e2006-01-03 09:55:46 +01001183extern int posix_unblock_lock(struct file *, struct file_lock *);
J. Bruce Fields3ee17ab2007-02-21 00:58:50 -05001184extern int vfs_test_lock(struct file *, struct file_lock *);
Marc Eshel150b3932007-01-18 16:15:35 -05001185extern int vfs_lock_file(struct file *, unsigned int, struct file_lock *, struct file_lock *);
Marc Eshel9b9d2ab2007-01-18 17:52:58 -05001186extern int vfs_cancel_lock(struct file *filp, struct file_lock *fl);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001187extern int flock_lock_file_wait(struct file *filp, struct file_lock *fl);
1188extern int __break_lease(struct inode *inode, unsigned int flags);
1189extern void lease_get_mtime(struct inode *, struct timespec *time);
Christoph Hellwig0af1a452007-07-31 00:39:22 -07001190extern int generic_setlease(struct file *, long, struct file_lock **);
J. Bruce Fieldsa9933ce2007-06-07 17:09:49 -04001191extern int vfs_setlease(struct file *, long, struct file_lock **);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001192extern int lease_modify(struct file_lock **, int);
1193extern int lock_may_read(struct inode *, loff_t start, unsigned long count);
1194extern int lock_may_write(struct inode *, loff_t start, unsigned long count);
Arnd Bergmannb89f4322010-09-18 15:09:31 +02001195extern void lock_flocks(void);
1196extern void unlock_flocks(void);
Thomas Petazzonibfcd17a2008-08-06 15:12:22 +02001197#else /* !CONFIG_FILE_LOCKING */
Steven Whitehousec2aca5e2009-01-20 10:29:45 +00001198static inline int fcntl_getlk(struct file *file, struct flock __user *user)
1199{
1200 return -EINVAL;
1201}
1202
1203static inline int fcntl_setlk(unsigned int fd, struct file *file,
1204 unsigned int cmd, struct flock __user *user)
1205{
1206 return -EACCES;
1207}
1208
Thomas Petazzonibfcd17a2008-08-06 15:12:22 +02001209#if BITS_PER_LONG == 32
Steven Whitehousec2aca5e2009-01-20 10:29:45 +00001210static inline int fcntl_getlk64(struct file *file, struct flock64 __user *user)
1211{
1212 return -EINVAL;
1213}
1214
1215static inline int fcntl_setlk64(unsigned int fd, struct file *file,
1216 unsigned int cmd, struct flock64 __user *user)
1217{
1218 return -EACCES;
1219}
Thomas Petazzonibfcd17a2008-08-06 15:12:22 +02001220#endif
Steven Whitehousec2aca5e2009-01-20 10:29:45 +00001221static inline int fcntl_setlease(unsigned int fd, struct file *filp, long arg)
1222{
1223 return 0;
1224}
1225
1226static inline int fcntl_getlease(struct file *filp)
1227{
1228 return 0;
1229}
1230
1231static inline void locks_init_lock(struct file_lock *fl)
1232{
1233 return;
1234}
1235
1236static inline void __locks_copy_lock(struct file_lock *new, struct file_lock *fl)
1237{
1238 return;
1239}
1240
1241static inline void locks_copy_lock(struct file_lock *new, struct file_lock *fl)
1242{
1243 return;
1244}
1245
1246static inline void locks_remove_posix(struct file *filp, fl_owner_t owner)
1247{
1248 return;
1249}
1250
1251static inline void locks_remove_flock(struct file *filp)
1252{
1253 return;
1254}
1255
1256static inline void posix_test_lock(struct file *filp, struct file_lock *fl)
1257{
1258 return;
1259}
1260
1261static inline int posix_lock_file(struct file *filp, struct file_lock *fl,
1262 struct file_lock *conflock)
1263{
1264 return -ENOLCK;
1265}
1266
1267static inline int posix_lock_file_wait(struct file *filp, struct file_lock *fl)
1268{
1269 return -ENOLCK;
1270}
1271
1272static inline int posix_unblock_lock(struct file *filp,
1273 struct file_lock *waiter)
1274{
1275 return -ENOENT;
1276}
1277
1278static inline int vfs_test_lock(struct file *filp, struct file_lock *fl)
1279{
1280 return 0;
1281}
1282
1283static inline int vfs_lock_file(struct file *filp, unsigned int cmd,
1284 struct file_lock *fl, struct file_lock *conf)
1285{
1286 return -ENOLCK;
1287}
1288
1289static inline int vfs_cancel_lock(struct file *filp, struct file_lock *fl)
1290{
1291 return 0;
1292}
1293
1294static inline int flock_lock_file_wait(struct file *filp,
1295 struct file_lock *request)
1296{
1297 return -ENOLCK;
1298}
1299
1300static inline int __break_lease(struct inode *inode, unsigned int mode)
1301{
1302 return 0;
1303}
1304
1305static inline void lease_get_mtime(struct inode *inode, struct timespec *time)
1306{
1307 return;
1308}
1309
1310static inline int generic_setlease(struct file *filp, long arg,
1311 struct file_lock **flp)
1312{
1313 return -EINVAL;
1314}
1315
1316static inline int vfs_setlease(struct file *filp, long arg,
1317 struct file_lock **lease)
1318{
1319 return -EINVAL;
1320}
1321
1322static inline int lease_modify(struct file_lock **before, int arg)
1323{
1324 return -EINVAL;
1325}
1326
1327static inline int lock_may_read(struct inode *inode, loff_t start,
1328 unsigned long len)
1329{
1330 return 1;
1331}
1332
1333static inline int lock_may_write(struct inode *inode, loff_t start,
1334 unsigned long len)
1335{
1336 return 1;
1337}
1338
Arnd Bergmannb89f4322010-09-18 15:09:31 +02001339static inline void lock_flocks(void)
1340{
1341}
1342
1343static inline void unlock_flocks(void)
1344{
1345}
1346
Thomas Petazzonibfcd17a2008-08-06 15:12:22 +02001347#endif /* !CONFIG_FILE_LOCKING */
1348
Linus Torvalds1da177e2005-04-16 15:20:36 -07001349
1350struct fasync_struct {
Eric Dumazet989a2972010-04-14 09:55:35 +00001351 spinlock_t fa_lock;
1352 int magic;
1353 int fa_fd;
1354 struct fasync_struct *fa_next; /* singly linked list */
1355 struct file *fa_file;
1356 struct rcu_head fa_rcu;
Linus Torvalds1da177e2005-04-16 15:20:36 -07001357};
1358
1359#define FASYNC_MAGIC 0x4601
1360
1361/* SMP safe fasync helpers: */
1362extern int fasync_helper(int, struct file *, int, struct fasync_struct **);
Linus Torvaldsf7347ce2010-10-27 12:38:12 -04001363extern struct fasync_struct *fasync_insert_entry(int, struct file *, struct fasync_struct **, struct fasync_struct *);
1364extern int fasync_remove_entry(struct file *, struct fasync_struct **);
1365extern struct fasync_struct *fasync_alloc(void);
1366extern void fasync_free(struct fasync_struct *);
1367
Linus Torvalds1da177e2005-04-16 15:20:36 -07001368/* can be called from interrupts */
1369extern void kill_fasync(struct fasync_struct **, int, int);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001370
Eric W. Biederman609d7fa2006-10-02 02:17:15 -07001371extern int __f_setown(struct file *filp, struct pid *, enum pid_type, int force);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001372extern int f_setown(struct file *filp, unsigned long arg, int force);
1373extern void f_delown(struct file *filp);
Eric W. Biederman609d7fa2006-10-02 02:17:15 -07001374extern pid_t f_getown(struct file *filp);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001375extern int send_sigurg(struct fown_struct *fown);
1376
1377/*
1378 * Umount options
1379 */
1380
1381#define MNT_FORCE 0x00000001 /* Attempt to forcibily umount */
1382#define MNT_DETACH 0x00000002 /* Just detach from the tree */
1383#define MNT_EXPIRE 0x00000004 /* Mark for expiry */
Miklos Szeredidb1f05b2010-02-10 12:15:53 +01001384#define UMOUNT_NOFOLLOW 0x00000008 /* Don't follow symlink on umount */
1385#define UMOUNT_UNUSED 0x80000000 /* Flag guaranteed to be unused */
Linus Torvalds1da177e2005-04-16 15:20:36 -07001386
1387extern struct list_head super_blocks;
1388extern spinlock_t sb_lock;
1389
Linus Torvalds1da177e2005-04-16 15:20:36 -07001390struct super_block {
1391 struct list_head s_list; /* Keep this first */
1392 dev_t s_dev; /* search index; _not_ kdev_t */
Linus Torvalds1da177e2005-04-16 15:20:36 -07001393 unsigned char s_dirt;
Richard Kennedy270ba5f2010-01-26 14:12:43 +00001394 unsigned char s_blocksize_bits;
1395 unsigned long s_blocksize;
Jeff Layton42cb56a2009-09-18 13:05:53 -07001396 loff_t s_maxbytes; /* Max file size */
Linus Torvalds1da177e2005-04-16 15:20:36 -07001397 struct file_system_type *s_type;
Josef 'Jeff' Sipekee9b6d62007-02-12 00:55:41 -08001398 const struct super_operations *s_op;
Alexey Dobriyan61e225d2009-09-21 17:01:08 -07001399 const struct dquot_operations *dq_op;
Alexey Dobriyan0d54b212009-09-21 17:01:09 -07001400 const struct quotactl_ops *s_qcop;
Christoph Hellwig39655162007-10-21 16:42:17 -07001401 const struct export_operations *s_export_op;
Linus Torvalds1da177e2005-04-16 15:20:36 -07001402 unsigned long s_flags;
1403 unsigned long s_magic;
1404 struct dentry *s_root;
1405 struct rw_semaphore s_umount;
Ingo Molnar7892f2f2006-01-09 15:59:25 -08001406 struct mutex s_lock;
Linus Torvalds1da177e2005-04-16 15:20:36 -07001407 int s_count;
Linus Torvalds1da177e2005-04-16 15:20:36 -07001408 atomic_t s_active;
Alexey Dobriyan50462062006-09-29 02:00:01 -07001409#ifdef CONFIG_SECURITY
Linus Torvalds1da177e2005-04-16 15:20:36 -07001410 void *s_security;
Alexey Dobriyan50462062006-09-29 02:00:01 -07001411#endif
Stephen Hemmingerbb435452010-05-13 17:53:14 -07001412 const struct xattr_handler **s_xattr;
Linus Torvalds1da177e2005-04-16 15:20:36 -07001413
1414 struct list_head s_inodes; /* all inodes */
Nick Pigginceb5bdc2011-01-07 17:50:05 +11001415 struct hlist_bl_head s_anon; /* anonymous dentries for (nfs) exporting */
Nick Piggin6416ccb2010-08-18 04:37:38 +10001416#ifdef CONFIG_SMP
1417 struct list_head __percpu *s_files;
1418#else
Linus Torvalds1da177e2005-04-16 15:20:36 -07001419 struct list_head s_files;
Nick Piggin6416ccb2010-08-18 04:37:38 +10001420#endif
Nick Pigginb5c84bf2011-01-07 17:49:38 +11001421 /* s_dentry_lru, s_nr_dentry_unused protected by dcache.c lru locks */
Kentaro Makitada3bbdd2008-07-23 21:27:13 -07001422 struct list_head s_dentry_lru; /* unused dentry lru */
1423 int s_nr_dentry_unused; /* # of dentry on lru */
Linus Torvalds1da177e2005-04-16 15:20:36 -07001424
Dave Chinner09cc9fc2011-07-08 14:14:40 +10001425 /* s_inode_lru_lock protects s_inode_lru and s_nr_inodes_unused */
1426 spinlock_t s_inode_lru_lock ____cacheline_aligned_in_smp;
Dave Chinner98b745c2011-07-08 14:14:39 +10001427 struct list_head s_inode_lru; /* unused inode lru */
1428 int s_nr_inodes_unused; /* # of inodes on lru */
1429
Linus Torvalds1da177e2005-04-16 15:20:36 -07001430 struct block_device *s_bdev;
Jens Axboe32a88aa2009-09-16 15:02:33 +02001431 struct backing_dev_info *s_bdi;
David Howellsacaebfd2007-05-10 22:51:50 -07001432 struct mtd_info *s_mtd;
Linus Torvalds1da177e2005-04-16 15:20:36 -07001433 struct list_head s_instances;
1434 struct quota_info s_dquot; /* Diskquota specific options */
1435
1436 int s_frozen;
1437 wait_queue_head_t s_wait_unfrozen;
1438
1439 char s_id[32]; /* Informational name */
Linus Torvalds0bba0162011-04-12 15:21:04 -07001440 u8 s_uuid[16]; /* UUID */
Linus Torvalds1da177e2005-04-16 15:20:36 -07001441
1442 void *s_fs_info; /* Filesystem private info */
Al Viro30c40d22008-02-22 19:50:45 -05001443 fmode_t s_mode;
Linus Torvalds1da177e2005-04-16 15:20:36 -07001444
Richard Kennedy270ba5f2010-01-26 14:12:43 +00001445 /* Granularity of c/m/atime in ns.
1446 Cannot be worse than a second */
1447 u32 s_time_gran;
1448
Linus Torvalds1da177e2005-04-16 15:20:36 -07001449 /*
1450 * The next field is for VFS *only*. No filesystems have any business
1451 * even looking at it. You had been warned.
1452 */
Arjan van de Vena11f3a02006-03-23 03:00:33 -08001453 struct mutex s_vfs_rename_mutex; /* Kludge */
Linus Torvalds1da177e2005-04-16 15:20:36 -07001454
Miklos Szeredi79c0b2d2007-05-08 00:25:43 -07001455 /*
1456 * Filesystem subtype. If non-empty the filesystem type field
1457 * in /proc/mounts will be "type.subtype"
1458 */
1459 char *s_subtype;
Miklos Szeredib3b304a2008-02-08 04:21:35 -08001460
1461 /*
1462 * Saved mount options for lazy filesystems using
1463 * generic_show_options()
1464 */
Arnd Bergmann4d2deb42010-02-24 20:01:56 +01001465 char __rcu *s_options;
Al Viroc8aebb02010-12-18 10:22:30 -05001466 const struct dentry_operations *s_d_op; /* default d_op for dentries */
Dan Magenheimer9fdfdcf2011-05-26 10:01:19 -06001467
1468 /*
1469 * Saved pool identifier for cleancache (-1 means none)
1470 */
1471 int cleancache_poolid;
Dave Chinnerb0d40c92011-07-08 14:14:42 +10001472
1473 struct shrinker s_shrink; /* per-sb shrinker handle */
Linus Torvalds1da177e2005-04-16 15:20:36 -07001474};
1475
Dave Chinnerb0d40c92011-07-08 14:14:42 +10001476/* superblock cache pruning functions */
1477extern void prune_icache_sb(struct super_block *sb, int nr_to_scan);
1478extern void prune_dcache_sb(struct super_block *sb, int nr_to_scan);
1479
Linus Torvalds1da177e2005-04-16 15:20:36 -07001480extern struct timespec current_fs_time(struct super_block *sb);
1481
1482/*
1483 * Snapshotting support.
1484 */
1485enum {
1486 SB_UNFROZEN = 0,
1487 SB_FREEZE_WRITE = 1,
1488 SB_FREEZE_TRANS = 2,
1489};
1490
1491#define vfs_check_frozen(sb, level) \
1492 wait_event((sb)->s_wait_unfrozen, ((sb)->s_frozen < (level)))
1493
Serge E. Hallyne795b712011-03-23 16:43:25 -07001494/*
1495 * until VFS tracks user namespaces for inodes, just make all files
1496 * belong to init_user_ns
1497 */
1498extern struct user_namespace init_user_ns;
1499#define inode_userns(inode) (&init_user_ns)
Serge E. Hallyn2e149672011-03-23 16:43:26 -07001500extern bool inode_owner_or_capable(const struct inode *inode);
Satyam Sharma3bd858a2007-07-17 15:00:08 +05301501
Al Viro914e2632006-10-18 13:55:46 -04001502/* not quite ready to be deprecated, but... */
1503extern void lock_super(struct super_block *);
1504extern void unlock_super(struct super_block *);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001505
1506/*
1507 * VFS helper functions..
1508 */
1509extern int vfs_create(struct inode *, struct dentry *, int, struct nameidata *);
1510extern int vfs_mkdir(struct inode *, struct dentry *, int);
1511extern int vfs_mknod(struct inode *, struct dentry *, int, dev_t);
Miklos Szeredidb2e7472008-06-24 16:50:16 +02001512extern int vfs_symlink(struct inode *, struct dentry *, const char *);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001513extern int vfs_link(struct dentry *, struct inode *, struct dentry *);
1514extern int vfs_rmdir(struct inode *, struct dentry *);
1515extern int vfs_unlink(struct inode *, struct dentry *);
1516extern int vfs_rename(struct inode *, struct dentry *, struct inode *, struct dentry *);
1517
1518/*
1519 * VFS dentry helper functions.
1520 */
1521extern void dentry_unhash(struct dentry *dentry);
1522
1523/*
Christoph Hellwig8c744fb2005-11-08 21:35:04 -08001524 * VFS file helper functions.
1525 */
Dmitry Monakhova1bd1202010-03-04 17:29:14 +03001526extern void inode_init_owner(struct inode *inode, const struct inode *dir,
1527 mode_t mode);
Christoph Hellwig8c744fb2005-11-08 21:35:04 -08001528/*
Mark Fashehc4b929b2008-10-08 19:44:18 -04001529 * VFS FS_IOC_FIEMAP helper definitions.
1530 */
1531struct fiemap_extent_info {
1532 unsigned int fi_flags; /* Flags as passed from user */
1533 unsigned int fi_extents_mapped; /* Number of mapped extents */
1534 unsigned int fi_extents_max; /* Size of fiemap_extent array */
Namhyung Kimecf56322011-01-16 23:28:17 +09001535 struct fiemap_extent __user *fi_extents_start; /* Start of
1536 fiemap_extent array */
Mark Fashehc4b929b2008-10-08 19:44:18 -04001537};
1538int fiemap_fill_next_extent(struct fiemap_extent_info *info, u64 logical,
1539 u64 phys, u64 len, u32 flags);
1540int fiemap_check_flags(struct fiemap_extent_info *fieinfo, u32 fs_flags);
1541
1542/*
Linus Torvalds1da177e2005-04-16 15:20:36 -07001543 * File types
1544 *
1545 * NOTE! These match bits 12..15 of stat.st_mode
1546 * (ie "(i_mode >> 12) & 15").
1547 */
1548#define DT_UNKNOWN 0
1549#define DT_FIFO 1
1550#define DT_CHR 2
1551#define DT_DIR 4
1552#define DT_BLK 6
1553#define DT_REG 8
1554#define DT_LNK 10
1555#define DT_SOCK 12
1556#define DT_WHT 14
1557
Linus Torvalds1da177e2005-04-16 15:20:36 -07001558/*
1559 * This is the "filldir" function type, used by readdir() to let
1560 * the kernel specify what kind of dirent layout it wants to have.
1561 * This allows the kernel to read directories into kernel space or
1562 * to have different dirent layouts depending on the binary type.
1563 */
David Howellsafefdbb2006-10-03 01:13:46 -07001564typedef int (*filldir_t)(void *, const char *, int, loff_t, u64, unsigned);
Al Viro08f85852007-10-08 13:26:20 -04001565struct block_device_operations;
Linus Torvalds1da177e2005-04-16 15:20:36 -07001566
Linus Torvalds1da177e2005-04-16 15:20:36 -07001567/* These macros are for out of kernel modules to test that
1568 * the kernel supports the unlocked_ioctl and compat_ioctl
1569 * fields in struct file_operations. */
1570#define HAVE_COMPAT_IOCTL 1
1571#define HAVE_UNLOCKED_IOCTL 1
1572
Linus Torvalds1da177e2005-04-16 15:20:36 -07001573struct file_operations {
1574 struct module *owner;
1575 loff_t (*llseek) (struct file *, loff_t, int);
1576 ssize_t (*read) (struct file *, char __user *, size_t, loff_t *);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001577 ssize_t (*write) (struct file *, const char __user *, size_t, loff_t *);
Badari Pulavarty027445c2006-09-30 23:28:46 -07001578 ssize_t (*aio_read) (struct kiocb *, const struct iovec *, unsigned long, loff_t);
1579 ssize_t (*aio_write) (struct kiocb *, const struct iovec *, unsigned long, loff_t);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001580 int (*readdir) (struct file *, void *, filldir_t);
1581 unsigned int (*poll) (struct file *, struct poll_table_struct *);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001582 long (*unlocked_ioctl) (struct file *, unsigned int, unsigned long);
1583 long (*compat_ioctl) (struct file *, unsigned int, unsigned long);
1584 int (*mmap) (struct file *, struct vm_area_struct *);
1585 int (*open) (struct inode *, struct file *);
Miklos Szeredi75e1fcc2006-06-23 02:05:12 -07001586 int (*flush) (struct file *, fl_owner_t id);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001587 int (*release) (struct inode *, struct file *);
Josef Bacik02c24a82011-07-16 20:44:56 -04001588 int (*fsync) (struct file *, loff_t, loff_t, int datasync);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001589 int (*aio_fsync) (struct kiocb *, int datasync);
1590 int (*fasync) (int, struct file *, int);
1591 int (*lock) (struct file *, int, struct file_lock *);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001592 ssize_t (*sendpage) (struct file *, struct page *, int, size_t, loff_t *, int);
1593 unsigned long (*get_unmapped_area)(struct file *, unsigned long, unsigned long, unsigned long, unsigned long);
1594 int (*check_flags)(int);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001595 int (*flock) (struct file *, int, struct file_lock *);
Jens Axboecbb7e572006-04-11 14:57:50 +02001596 ssize_t (*splice_write)(struct pipe_inode_info *, struct file *, loff_t *, size_t, unsigned int);
1597 ssize_t (*splice_read)(struct file *, loff_t *, struct pipe_inode_info *, size_t, unsigned int);
J. Bruce Fieldsf9ffed22006-11-14 15:51:40 -05001598 int (*setlease)(struct file *, long, struct file_lock **);
Christoph Hellwig2fe17c12011-01-14 13:07:43 +01001599 long (*fallocate)(struct file *file, int mode, loff_t offset,
1600 loff_t len);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001601};
1602
1603struct inode_operations {
Linus Torvalds1da177e2005-04-16 15:20:36 -07001604 struct dentry * (*lookup) (struct inode *,struct dentry *, struct nameidata *);
Nick Piggin44a7d7a2011-01-07 17:49:56 +11001605 void * (*follow_link) (struct dentry *, struct nameidata *);
Al Viro10556cb2011-06-20 19:28:19 -04001606 int (*permission) (struct inode *, int);
Christoph Hellwig4e34e712011-07-23 17:37:31 +02001607 struct posix_acl * (*get_acl)(struct inode *, int);
Nick Piggin44a7d7a2011-01-07 17:49:56 +11001608
1609 int (*readlink) (struct dentry *, char __user *,int);
1610 void (*put_link) (struct dentry *, struct nameidata *, void *);
1611
1612 int (*create) (struct inode *,struct dentry *,int, struct nameidata *);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001613 int (*link) (struct dentry *,struct inode *,struct dentry *);
1614 int (*unlink) (struct inode *,struct dentry *);
1615 int (*symlink) (struct inode *,struct dentry *,const char *);
1616 int (*mkdir) (struct inode *,struct dentry *,int);
1617 int (*rmdir) (struct inode *,struct dentry *);
1618 int (*mknod) (struct inode *,struct dentry *,int,dev_t);
1619 int (*rename) (struct inode *, struct dentry *,
1620 struct inode *, struct dentry *);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001621 void (*truncate) (struct inode *);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001622 int (*setattr) (struct dentry *, struct iattr *);
1623 int (*getattr) (struct vfsmount *mnt, struct dentry *, struct kstat *);
1624 int (*setxattr) (struct dentry *, const char *,const void *,size_t,int);
1625 ssize_t (*getxattr) (struct dentry *, const char *, void *, size_t);
1626 ssize_t (*listxattr) (struct dentry *, char *, size_t);
1627 int (*removexattr) (struct dentry *, const char *);
Badari Pulavartyf6b3ec22006-01-06 00:10:38 -08001628 void (*truncate_range)(struct inode *, loff_t, loff_t);
Mark Fashehc4b929b2008-10-08 19:44:18 -04001629 int (*fiemap)(struct inode *, struct fiemap_extent_info *, u64 start,
1630 u64 len);
Nick Piggin44a7d7a2011-01-07 17:49:56 +11001631} ____cacheline_aligned;
Linus Torvalds1da177e2005-04-16 15:20:36 -07001632
1633struct seq_file;
1634
Badari Pulavartyeed4e512006-09-30 23:28:49 -07001635ssize_t rw_copy_check_uvector(int type, const struct iovec __user * uvector,
Christopher Yeohfcf63402011-10-31 17:06:39 -07001636 unsigned long nr_segs, unsigned long fast_segs,
1637 struct iovec *fast_pointer,
1638 struct iovec **ret_pointer,
1639 int check_access);
Badari Pulavartyeed4e512006-09-30 23:28:49 -07001640
Linus Torvalds1da177e2005-04-16 15:20:36 -07001641extern ssize_t vfs_read(struct file *, char __user *, size_t, loff_t *);
1642extern ssize_t vfs_write(struct file *, const char __user *, size_t, loff_t *);
1643extern ssize_t vfs_readv(struct file *, const struct iovec __user *,
1644 unsigned long, loff_t *);
1645extern ssize_t vfs_writev(struct file *, const struct iovec __user *,
1646 unsigned long, loff_t *);
1647
Linus Torvalds1da177e2005-04-16 15:20:36 -07001648struct super_operations {
1649 struct inode *(*alloc_inode)(struct super_block *sb);
1650 void (*destroy_inode)(struct inode *);
1651
Christoph Hellwigaa385722011-05-27 06:53:02 -04001652 void (*dirty_inode) (struct inode *, int flags);
Christoph Hellwiga9185b42010-03-05 09:21:37 +01001653 int (*write_inode) (struct inode *, struct writeback_control *wbc);
Al Viro45321ac2010-06-07 13:43:19 -04001654 int (*drop_inode) (struct inode *);
Al Virobe7ce412010-06-04 19:40:39 -04001655 void (*evict_inode) (struct inode *);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001656 void (*put_super) (struct super_block *);
1657 void (*write_super) (struct super_block *);
1658 int (*sync_fs)(struct super_block *sb, int wait);
Takashi Satoc4be0c12009-01-09 16:40:58 -08001659 int (*freeze_fs) (struct super_block *);
1660 int (*unfreeze_fs) (struct super_block *);
David Howells726c3342006-06-23 02:02:58 -07001661 int (*statfs) (struct dentry *, struct kstatfs *);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001662 int (*remount_fs) (struct super_block *, int *, char *);
Al Viro42faad92008-04-24 07:21:56 -04001663 void (*umount_begin) (struct super_block *);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001664
1665 int (*show_options)(struct seq_file *, struct vfsmount *);
Al Viroc7f404b2011-03-16 06:59:40 -04001666 int (*show_devname)(struct seq_file *, struct vfsmount *);
1667 int (*show_path)(struct seq_file *, struct vfsmount *);
Chuck Leverb4629fe2006-03-20 13:44:12 -05001668 int (*show_stats)(struct seq_file *, struct vfsmount *);
Alexey Dobriyan0e51a722006-09-29 01:59:56 -07001669#ifdef CONFIG_QUOTA
Linus Torvalds1da177e2005-04-16 15:20:36 -07001670 ssize_t (*quota_read)(struct super_block *, int, char *, size_t, loff_t);
1671 ssize_t (*quota_write)(struct super_block *, int, const char *, size_t, loff_t);
Alexey Dobriyan0e51a722006-09-29 01:59:56 -07001672#endif
Theodore Ts'o87d8fe12009-01-03 09:47:09 -05001673 int (*bdev_try_to_free_page)(struct super_block*, struct page*, gfp_t);
Dave Chinner0e1fdaf2011-07-08 14:14:44 +10001674 int (*nr_cached_objects)(struct super_block *);
1675 void (*free_cached_objects)(struct super_block *, int);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001676};
1677
Joern Engel1c0eeaf2007-10-16 23:30:44 -07001678/*
Dave Chinner250df6e2011-03-22 22:23:36 +11001679 * Inode state bits. Protected by inode->i_lock
Joern Engel1c0eeaf2007-10-16 23:30:44 -07001680 *
1681 * Three bits determine the dirty state of the inode, I_DIRTY_SYNC,
1682 * I_DIRTY_DATASYNC and I_DIRTY_PAGES.
1683 *
1684 * Four bits define the lifetime of an inode. Initially, inodes are I_NEW,
1685 * until that flag is cleared. I_WILL_FREE, I_FREEING and I_CLEAR are set at
1686 * various stages of removing an inode.
1687 *
Christoph Hellwigeaff8072009-12-17 14:25:01 +01001688 * Two bits are used for locking and completion notification, I_NEW and I_SYNC.
Joern Engel1c0eeaf2007-10-16 23:30:44 -07001689 *
Joern Engele7ca2d42008-02-06 01:36:59 -08001690 * I_DIRTY_SYNC Inode is dirty, but doesn't have to be written on
1691 * fdatasync(). i_atime is the usual cause.
Jan Kara000cb482008-02-14 19:31:32 -08001692 * I_DIRTY_DATASYNC Data-related inode changes pending. We keep track of
1693 * these changes separately from I_DIRTY_SYNC so that we
1694 * don't have to write inode on fdatasync() when only
1695 * mtime has changed in it.
Joern Engel1c0eeaf2007-10-16 23:30:44 -07001696 * I_DIRTY_PAGES Inode has dirty pages. Inode itself may be clean.
Christoph Hellwigeaff8072009-12-17 14:25:01 +01001697 * I_NEW Serves as both a mutex and completion notification.
1698 * New inodes set I_NEW. If two processes both create
1699 * the same inode, one of them will release its inode and
1700 * wait for I_NEW to be released before returning.
1701 * Inodes in I_WILL_FREE, I_FREEING or I_CLEAR state can
1702 * also cause waiting on I_NEW, without I_NEW actually
1703 * being set. find_inode() uses this to prevent returning
1704 * nearly-dead inodes.
Joern Engel1c0eeaf2007-10-16 23:30:44 -07001705 * I_WILL_FREE Must be set when calling write_inode_now() if i_count
1706 * is zero. I_FREEING must be set when I_WILL_FREE is
1707 * cleared.
1708 * I_FREEING Set when inode is about to be freed but still has dirty
1709 * pages or buffers attached or the inode itself is still
1710 * dirty.
Al Virob57922d2010-06-07 14:34:48 -04001711 * I_CLEAR Added by end_writeback(). In this state the inode is clean
Al Viroa4ffdde2010-06-02 17:38:30 -04001712 * and can be destroyed. Inode keeps I_FREEING.
Joern Engel1c0eeaf2007-10-16 23:30:44 -07001713 *
1714 * Inodes that are I_WILL_FREE, I_FREEING or I_CLEAR are
1715 * prohibited for many purposes. iget() must wait for
1716 * the inode to be completely released, then create it
1717 * anew. Other functions will just ignore such inodes,
Christoph Hellwigeaff8072009-12-17 14:25:01 +01001718 * if appropriate. I_NEW is used for waiting.
Joern Engel1c0eeaf2007-10-16 23:30:44 -07001719 *
Christoph Hellwigeaff8072009-12-17 14:25:01 +01001720 * I_SYNC Synchonized write of dirty inode data. The bits is
1721 * set during data writeback, and cleared with a wakeup
1722 * on the bit address once it is done.
Joern Engel1c0eeaf2007-10-16 23:30:44 -07001723 *
Christoph Hellwigbd5fe6c2011-06-24 14:29:43 -04001724 * I_REFERENCED Marks the inode as recently references on the LRU list.
1725 *
1726 * I_DIO_WAKEUP Never set. Only used as a key for wait_on_bit().
1727 *
Joern Engel1c0eeaf2007-10-16 23:30:44 -07001728 * Q: What is the difference between I_WILL_FREE and I_FREEING?
Joern Engel1c0eeaf2007-10-16 23:30:44 -07001729 */
Nick Piggin9e38d862010-10-23 06:55:17 -04001730#define I_DIRTY_SYNC (1 << 0)
1731#define I_DIRTY_DATASYNC (1 << 1)
1732#define I_DIRTY_PAGES (1 << 2)
Christoph Hellwigeaff8072009-12-17 14:25:01 +01001733#define __I_NEW 3
1734#define I_NEW (1 << __I_NEW)
Nick Piggin9e38d862010-10-23 06:55:17 -04001735#define I_WILL_FREE (1 << 4)
1736#define I_FREEING (1 << 5)
1737#define I_CLEAR (1 << 6)
Christoph Hellwigeaff8072009-12-17 14:25:01 +01001738#define __I_SYNC 7
Joern Engel1c0eeaf2007-10-16 23:30:44 -07001739#define I_SYNC (1 << __I_SYNC)
Nick Piggin9e38d862010-10-23 06:55:17 -04001740#define I_REFERENCED (1 << 8)
Christoph Hellwigbd5fe6c2011-06-24 14:29:43 -04001741#define __I_DIO_WAKEUP 9
1742#define I_DIO_WAKEUP (1 << I_DIO_WAKEUP)
Linus Torvalds1da177e2005-04-16 15:20:36 -07001743
1744#define I_DIRTY (I_DIRTY_SYNC | I_DIRTY_DATASYNC | I_DIRTY_PAGES)
1745
1746extern void __mark_inode_dirty(struct inode *, int);
1747static inline void mark_inode_dirty(struct inode *inode)
1748{
1749 __mark_inode_dirty(inode, I_DIRTY);
1750}
1751
1752static inline void mark_inode_dirty_sync(struct inode *inode)
1753{
1754 __mark_inode_dirty(inode, I_DIRTY_SYNC);
1755}
1756
Dave Hansen71c42152007-06-23 17:16:42 -07001757/**
1758 * inc_nlink - directly increment an inode's link count
1759 * @inode: inode
1760 *
1761 * This is a low-level filesystem helper to replace any
1762 * direct filesystem manipulation of i_nlink. Currently,
1763 * it is only here for parity with dec_nlink().
1764 */
Dave Hansend8c76e62006-09-30 23:29:04 -07001765static inline void inc_nlink(struct inode *inode)
Alexey Dobriyan66d21732006-03-23 03:00:51 -08001766{
1767 inode->i_nlink++;
Dave Hansend8c76e62006-09-30 23:29:04 -07001768}
1769
1770static inline void inode_inc_link_count(struct inode *inode)
1771{
1772 inc_nlink(inode);
Alexey Dobriyan66d21732006-03-23 03:00:51 -08001773 mark_inode_dirty(inode);
1774}
1775
Dave Hansen71c42152007-06-23 17:16:42 -07001776/**
1777 * drop_nlink - directly drop an inode's link count
1778 * @inode: inode
1779 *
1780 * This is a low-level filesystem helper to replace any
1781 * direct filesystem manipulation of i_nlink. In cases
1782 * where we are attempting to track writes to the
1783 * filesystem, a decrement to zero means an imminent
1784 * write when the file is truncated and actually unlinked
1785 * on the filesystem.
1786 */
Dave Hansen9a53c3a2006-09-30 23:29:03 -07001787static inline void drop_nlink(struct inode *inode)
Alexey Dobriyan66d21732006-03-23 03:00:51 -08001788{
1789 inode->i_nlink--;
Dave Hansen9a53c3a2006-09-30 23:29:03 -07001790}
1791
Dave Hansen71c42152007-06-23 17:16:42 -07001792/**
1793 * clear_nlink - directly zero an inode's link count
1794 * @inode: inode
1795 *
1796 * This is a low-level filesystem helper to replace any
1797 * direct filesystem manipulation of i_nlink. See
1798 * drop_nlink() for why we care about i_nlink hitting zero.
1799 */
Dave Hansence71ec32006-09-30 23:29:06 -07001800static inline void clear_nlink(struct inode *inode)
1801{
1802 inode->i_nlink = 0;
1803}
1804
Dave Hansen9a53c3a2006-09-30 23:29:03 -07001805static inline void inode_dec_link_count(struct inode *inode)
1806{
1807 drop_nlink(inode);
Alexey Dobriyan66d21732006-03-23 03:00:51 -08001808 mark_inode_dirty(inode);
1809}
1810
Jean Noel Cordenner25ec56b2008-01-28 23:58:27 -05001811/**
1812 * inode_inc_iversion - increments i_version
1813 * @inode: inode that need to be updated
1814 *
1815 * Every time the inode is modified, the i_version field will be incremented.
1816 * The filesystem has to be mounted with i_version flag
1817 */
1818
1819static inline void inode_inc_iversion(struct inode *inode)
1820{
1821 spin_lock(&inode->i_lock);
1822 inode->i_version++;
1823 spin_unlock(&inode->i_lock);
1824}
1825
Christoph Hellwig869243a2006-01-09 20:52:03 -08001826extern void touch_atime(struct vfsmount *mnt, struct dentry *dentry);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001827static inline void file_accessed(struct file *file)
1828{
1829 if (!(file->f_flags & O_NOATIME))
Josef "Jeff" Sipek0f7fc9e2006-12-08 02:36:35 -08001830 touch_atime(file->f_path.mnt, file->f_path.dentry);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001831}
1832
1833int sync_inode(struct inode *inode, struct writeback_control *wbc);
Christoph Hellwigc37650162010-10-06 10:48:20 +02001834int sync_inode_metadata(struct inode *inode, int wait);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001835
Linus Torvalds1da177e2005-04-16 15:20:36 -07001836struct file_system_type {
1837 const char *name;
1838 int fs_flags;
Al Viroc96e41e2010-07-25 00:17:56 +04001839 struct dentry *(*mount) (struct file_system_type *, int,
1840 const char *, void *);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001841 void (*kill_sb) (struct super_block *);
1842 struct module *owner;
1843 struct file_system_type * next;
1844 struct list_head fs_supers;
Peter Zijlstrad475fd42007-10-15 14:51:31 +02001845
Ingo Molnarcf516242006-07-03 00:25:27 -07001846 struct lock_class_key s_lock_key;
Arjan van de Ven897c6ff2006-07-03 00:25:28 -07001847 struct lock_class_key s_umount_key;
Roland Dreier51ee0492010-04-27 14:23:57 -07001848 struct lock_class_key s_vfs_rename_key;
Peter Zijlstrad475fd42007-10-15 14:51:31 +02001849
1850 struct lock_class_key i_lock_key;
1851 struct lock_class_key i_mutex_key;
Peter Zijlstra14358e62007-10-14 01:38:33 +02001852 struct lock_class_key i_mutex_dir_key;
Linus Torvalds1da177e2005-04-16 15:20:36 -07001853};
1854
Al Viroceefda62010-07-26 13:16:50 +04001855extern struct dentry *mount_ns(struct file_system_type *fs_type, int flags,
1856 void *data, int (*fill_super)(struct super_block *, void *, int));
Al Viro152a0832010-07-25 00:46:55 +04001857extern struct dentry *mount_bdev(struct file_system_type *fs_type,
1858 int flags, const char *dev_name, void *data,
1859 int (*fill_super)(struct super_block *, void *, int));
Al Virofc14f2f2010-07-25 01:48:30 +04001860extern struct dentry *mount_single(struct file_system_type *fs_type,
1861 int flags, void *data,
1862 int (*fill_super)(struct super_block *, void *, int));
Al Viro3c26ff62010-07-25 11:46:36 +04001863extern struct dentry *mount_nodev(struct file_system_type *fs_type,
1864 int flags, void *data,
1865 int (*fill_super)(struct super_block *, void *, int));
Linus Torvalds1da177e2005-04-16 15:20:36 -07001866void generic_shutdown_super(struct super_block *sb);
1867void kill_block_super(struct super_block *sb);
1868void kill_anon_super(struct super_block *sb);
1869void kill_litter_super(struct super_block *sb);
1870void deactivate_super(struct super_block *sb);
Al Viro74dbbdd2009-05-06 01:07:50 -04001871void deactivate_locked_super(struct super_block *sb);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001872int set_anon_super(struct super_block *s, void *data);
Al Viro0ee5dc62011-07-07 15:44:25 -04001873int get_anon_bdev(dev_t *);
1874void free_anon_bdev(dev_t);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001875struct super_block *sget(struct file_system_type *type,
1876 int (*test)(struct super_block *,void *),
1877 int (*set)(struct super_block *,void *),
1878 void *data);
Al Viro51139ad2010-07-25 23:47:46 +04001879extern struct dentry *mount_pseudo(struct file_system_type *, char *,
Al Viroc74a1cb2011-01-12 16:59:34 -05001880 const struct super_operations *ops,
1881 const struct dentry_operations *dops,
1882 unsigned long);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001883
Artem Bityutskiy140236b2010-06-10 13:56:33 +03001884static inline void sb_mark_dirty(struct super_block *sb)
1885{
1886 sb->s_dirt = 1;
1887}
1888static inline void sb_mark_clean(struct super_block *sb)
1889{
1890 sb->s_dirt = 0;
1891}
1892static inline int sb_is_dirty(struct super_block *sb)
1893{
1894 return sb->s_dirt;
1895}
1896
Linus Torvalds1da177e2005-04-16 15:20:36 -07001897/* Alas, no aliases. Too much hassle with bringing module.h everywhere */
1898#define fops_get(fops) \
1899 (((fops) && try_module_get((fops)->owner) ? (fops) : NULL))
1900#define fops_put(fops) \
1901 do { if (fops) module_put((fops)->owner); } while(0)
1902
1903extern int register_filesystem(struct file_system_type *);
1904extern int unregister_filesystem(struct file_system_type *);
Pavel Emelyanov8bf97252007-10-18 23:40:02 -07001905extern struct vfsmount *kern_mount_data(struct file_system_type *, void *data);
1906#define kern_mount(type) kern_mount_data(type, NULL)
Tim Chen423e0ab2011-07-19 09:32:38 -07001907extern void kern_unmount(struct vfsmount *mnt);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001908extern int may_umount_tree(struct vfsmount *);
1909extern int may_umount(struct vfsmount *);
1910extern long do_mount(char *, char *, char *, unsigned long, void *);
Al Viro589ff872009-04-18 03:28:19 -04001911extern struct vfsmount *collect_mounts(struct path *);
Al Viro8aec0802007-06-07 12:20:32 -04001912extern void drop_collected_mounts(struct vfsmount *);
Al Viro1f707132010-01-30 22:51:25 -05001913extern int iterate_mounts(int (*)(struct vfsmount *, void *), void *,
1914 struct vfsmount *);
Christoph Hellwigebabe9a2010-07-07 18:53:11 +02001915extern int vfs_statfs(struct path *, struct kstatfs *);
Al Viroc8b91ac2011-03-12 10:41:39 -05001916extern int user_statfs(const char __user *, struct kstatfs *);
1917extern int fd_statfs(int, struct kstatfs *);
Christoph Hellwigebabe9a2010-07-07 18:53:11 +02001918extern int statfs_by_dentry(struct dentry *, struct kstatfs *);
Josef Bacik18e9e512010-03-23 10:34:56 -04001919extern int freeze_super(struct super_block *super);
1920extern int thaw_super(struct super_block *super);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001921
Al Viroce3b0f82009-03-29 19:08:22 -04001922extern int current_umask(void);
1923
Miklos Szeredif87fd4c2006-01-16 22:14:23 -08001924/* /sys/fs */
Greg Kroah-Hartman00d26662007-10-29 14:17:23 -06001925extern struct kobject *fs_kobj;
Miklos Szeredif87fd4c2006-01-16 22:14:23 -08001926
Linus Torvalds435f49a2010-10-29 10:36:49 -07001927#define MAX_RW_COUNT (INT_MAX & PAGE_CACHE_MASK)
Thomas Petazzonibfcd17a2008-08-06 15:12:22 +02001928extern int rw_verify_area(int, struct file *, loff_t *, size_t);
1929
Linus Torvalds1da177e2005-04-16 15:20:36 -07001930#define FLOCK_VERIFY_READ 1
1931#define FLOCK_VERIFY_WRITE 2
1932
Thomas Petazzonibfcd17a2008-08-06 15:12:22 +02001933#ifdef CONFIG_FILE_LOCKING
Linus Torvalds1da177e2005-04-16 15:20:36 -07001934extern int locks_mandatory_locked(struct inode *);
1935extern int locks_mandatory_area(int, struct inode *, struct file *, loff_t, size_t);
1936
1937/*
1938 * Candidates for mandatory locking have the setgid bit set
1939 * but no group execute bit - an otherwise meaningless combination.
1940 */
Pavel Emelyanova16877c2007-10-01 14:41:11 -07001941
1942static inline int __mandatory_lock(struct inode *ino)
1943{
1944 return (ino->i_mode & (S_ISGID | S_IXGRP)) == S_ISGID;
1945}
1946
1947/*
1948 * ... and these candidates should be on MS_MANDLOCK mounted fs,
1949 * otherwise these will be advisory locks
1950 */
1951
1952static inline int mandatory_lock(struct inode *ino)
1953{
1954 return IS_MANDLOCK(ino) && __mandatory_lock(ino);
1955}
Linus Torvalds1da177e2005-04-16 15:20:36 -07001956
1957static inline int locks_verify_locked(struct inode *inode)
1958{
Pavel Emelyanova16877c2007-10-01 14:41:11 -07001959 if (mandatory_lock(inode))
Linus Torvalds1da177e2005-04-16 15:20:36 -07001960 return locks_mandatory_locked(inode);
1961 return 0;
1962}
1963
Linus Torvalds1da177e2005-04-16 15:20:36 -07001964static inline int locks_verify_truncate(struct inode *inode,
1965 struct file *filp,
1966 loff_t size)
1967{
Pavel Emelyanova16877c2007-10-01 14:41:11 -07001968 if (inode->i_flock && mandatory_lock(inode))
Linus Torvalds1da177e2005-04-16 15:20:36 -07001969 return locks_mandatory_area(
1970 FLOCK_VERIFY_WRITE, inode, filp,
1971 size < inode->i_size ? size : inode->i_size,
1972 (size < inode->i_size ? inode->i_size - size
1973 : size - inode->i_size)
1974 );
1975 return 0;
1976}
1977
1978static inline int break_lease(struct inode *inode, unsigned int mode)
1979{
1980 if (inode->i_flock)
1981 return __break_lease(inode, mode);
1982 return 0;
1983}
Thomas Petazzonibfcd17a2008-08-06 15:12:22 +02001984#else /* !CONFIG_FILE_LOCKING */
Steven Whitehouseaf5df562009-01-20 10:29:46 +00001985static inline int locks_mandatory_locked(struct inode *inode)
1986{
1987 return 0;
1988}
1989
1990static inline int locks_mandatory_area(int rw, struct inode *inode,
1991 struct file *filp, loff_t offset,
1992 size_t count)
1993{
1994 return 0;
1995}
1996
1997static inline int __mandatory_lock(struct inode *inode)
1998{
1999 return 0;
2000}
2001
2002static inline int mandatory_lock(struct inode *inode)
2003{
2004 return 0;
2005}
2006
2007static inline int locks_verify_locked(struct inode *inode)
2008{
2009 return 0;
2010}
2011
2012static inline int locks_verify_truncate(struct inode *inode, struct file *filp,
2013 size_t size)
2014{
2015 return 0;
2016}
2017
2018static inline int break_lease(struct inode *inode, unsigned int mode)
2019{
2020 return 0;
2021}
2022
Thomas Petazzonibfcd17a2008-08-06 15:12:22 +02002023#endif /* CONFIG_FILE_LOCKING */
Linus Torvalds1da177e2005-04-16 15:20:36 -07002024
2025/* fs/open.c */
2026
NeilBrown4a301312006-01-08 01:02:39 -08002027extern int do_truncate(struct dentry *, loff_t start, unsigned int time_attrs,
2028 struct file *filp);
Ankit Jain3e63cbb2009-06-19 14:28:07 -04002029extern int do_fallocate(struct file *file, int mode, loff_t offset,
2030 loff_t len);
Jason Uhlenkott8e8a1402007-10-20 03:16:18 +02002031extern long do_sys_open(int dfd, const char __user *filename, int flags,
Ulrich Drepper5590ff02006-01-18 17:43:53 -08002032 int mode);
Linus Torvalds1da177e2005-04-16 15:20:36 -07002033extern struct file *filp_open(const char *, int, int);
Al Viro73d049a2011-03-11 12:08:24 -05002034extern struct file *file_open_root(struct dentry *, struct vfsmount *,
2035 const char *, int);
David Howells745ca242008-11-14 10:39:22 +11002036extern struct file * dentry_open(struct dentry *, struct vfsmount *, int,
2037 const struct cred *);
Linus Torvalds1da177e2005-04-16 15:20:36 -07002038extern int filp_close(struct file *, fl_owner_t id);
2039extern char * getname(const char __user *);
2040
Ankit Jain3e63cbb2009-06-19 14:28:07 -04002041/* fs/ioctl.c */
2042
2043extern int ioctl_preallocate(struct file *filp, void __user *argp);
2044
Linus Torvalds1da177e2005-04-16 15:20:36 -07002045/* fs/dcache.c */
2046extern void __init vfs_caches_init_early(void);
2047extern void __init vfs_caches_init(unsigned long);
2048
Christoph Lameterb86c0892006-12-06 20:32:57 -08002049extern struct kmem_cache *names_cachep;
2050
Vegard Nossum3446a8a2009-05-16 11:22:14 +02002051#define __getname_gfp(gfp) kmem_cache_alloc(names_cachep, (gfp))
2052#define __getname() __getname_gfp(GFP_KERNEL)
2053#define __putname(name) kmem_cache_free(names_cachep, (void *)(name))
Linus Torvalds1da177e2005-04-16 15:20:36 -07002054#ifndef CONFIG_AUDITSYSCALL
2055#define putname(name) __putname(name)
2056#else
2057extern void putname(const char *name);
2058#endif
2059
David Howells93614012006-09-30 20:45:40 +02002060#ifdef CONFIG_BLOCK
Linus Torvalds1da177e2005-04-16 15:20:36 -07002061extern int register_blkdev(unsigned int, const char *);
Akinobu Mitaf4480242007-07-17 04:03:47 -07002062extern void unregister_blkdev(unsigned int, const char *);
Linus Torvalds1da177e2005-04-16 15:20:36 -07002063extern struct block_device *bdget(dev_t);
Alan Jenkinsdddac6a2009-07-29 21:07:55 +02002064extern struct block_device *bdgrab(struct block_device *bdev);
Linus Torvalds1da177e2005-04-16 15:20:36 -07002065extern void bd_set_size(struct block_device *, loff_t size);
2066extern void bd_forget(struct inode *inode);
2067extern void bdput(struct block_device *);
Nick Piggin585d3bc2009-02-25 10:44:19 +01002068extern void invalidate_bdev(struct block_device *);
2069extern int sync_blockdev(struct block_device *bdev);
2070extern struct super_block *freeze_bdev(struct block_device *);
Eric Sandeenc2d75432009-03-31 15:23:46 -07002071extern void emergency_thaw_all(void);
Nick Piggin585d3bc2009-02-25 10:44:19 +01002072extern int thaw_bdev(struct block_device *bdev, struct super_block *sb);
2073extern int fsync_bdev(struct block_device *);
David Howells93614012006-09-30 20:45:40 +02002074#else
2075static inline void bd_forget(struct inode *inode) {}
Al Viro47e44912009-04-01 07:07:16 -04002076static inline int sync_blockdev(struct block_device *bdev) { return 0; }
2077static inline void invalidate_bdev(struct block_device *bdev) {}
2078
2079static inline struct super_block *freeze_bdev(struct block_device *sb)
2080{
2081 return NULL;
2082}
2083
2084static inline int thaw_bdev(struct block_device *bdev, struct super_block *sb)
2085{
2086 return 0;
2087}
David Howells93614012006-09-30 20:45:40 +02002088#endif
Jan Kara60b06802009-04-27 16:43:53 +02002089extern int sync_filesystem(struct super_block *);
David Howells93614012006-09-30 20:45:40 +02002090extern const struct file_operations def_blk_fops;
Arjan van de Ven4b6f5d22006-03-28 01:56:42 -08002091extern const struct file_operations def_chr_fops;
2092extern const struct file_operations bad_sock_fops;
2093extern const struct file_operations def_fifo_fops;
David Howells93614012006-09-30 20:45:40 +02002094#ifdef CONFIG_BLOCK
Linus Torvalds1da177e2005-04-16 15:20:36 -07002095extern int ioctl_by_bdev(struct block_device *, unsigned, unsigned long);
Al Viro56b26ad2008-09-19 03:17:36 -04002096extern int blkdev_ioctl(struct block_device *, fmode_t, unsigned, unsigned long);
Linus Torvalds1da177e2005-04-16 15:20:36 -07002097extern long compat_blkdev_ioctl(struct file *, unsigned, unsigned long);
Tejun Heoe525fd82010-11-13 11:55:17 +01002098extern int blkdev_get(struct block_device *bdev, fmode_t mode, void *holder);
Tejun Heod4d77622010-11-13 11:55:18 +01002099extern struct block_device *blkdev_get_by_path(const char *path, fmode_t mode,
2100 void *holder);
2101extern struct block_device *blkdev_get_by_dev(dev_t dev, fmode_t mode,
2102 void *holder);
Tejun Heoe525fd82010-11-13 11:55:17 +01002103extern int blkdev_put(struct block_device *bdev, fmode_t mode);
Jun'ichi Nomura641dc632006-03-27 01:17:57 -08002104#ifdef CONFIG_SYSFS
Tejun Heoe09b4572010-11-13 11:55:17 +01002105extern int bd_link_disk_holder(struct block_device *bdev, struct gendisk *disk);
Tejun Heo49731ba2011-01-14 18:43:57 +01002106extern void bd_unlink_disk_holder(struct block_device *bdev,
2107 struct gendisk *disk);
Jun'ichi Nomura641dc632006-03-27 01:17:57 -08002108#else
Tejun Heoe09b4572010-11-13 11:55:17 +01002109static inline int bd_link_disk_holder(struct block_device *bdev,
2110 struct gendisk *disk)
2111{
2112 return 0;
2113}
Tejun Heo49731ba2011-01-14 18:43:57 +01002114static inline void bd_unlink_disk_holder(struct block_device *bdev,
2115 struct gendisk *disk)
2116{
2117}
Jun'ichi Nomura641dc632006-03-27 01:17:57 -08002118#endif
David Howells93614012006-09-30 20:45:40 +02002119#endif
Linus Torvalds1da177e2005-04-16 15:20:36 -07002120
2121/* fs/char_dev.c */
Joe Korty68eef3b2006-03-31 02:30:32 -08002122#define CHRDEV_MAJOR_HASH_SIZE 255
Linus Torvalds1da177e2005-04-16 15:20:36 -07002123extern int alloc_chrdev_region(dev_t *, unsigned, unsigned, const char *);
2124extern int register_chrdev_region(dev_t, unsigned, const char *);
Tejun Heo1905b1b2009-08-06 18:13:23 +09002125extern int __register_chrdev(unsigned int major, unsigned int baseminor,
2126 unsigned int count, const char *name,
2127 const struct file_operations *fops);
2128extern void __unregister_chrdev(unsigned int major, unsigned int baseminor,
2129 unsigned int count, const char *name);
Linus Torvalds1da177e2005-04-16 15:20:36 -07002130extern void unregister_chrdev_region(dev_t, unsigned);
Joe Korty68eef3b2006-03-31 02:30:32 -08002131extern void chrdev_show(struct seq_file *,off_t);
Linus Torvalds1da177e2005-04-16 15:20:36 -07002132
Tejun Heo1905b1b2009-08-06 18:13:23 +09002133static inline int register_chrdev(unsigned int major, const char *name,
2134 const struct file_operations *fops)
2135{
2136 return __register_chrdev(major, 0, 256, name, fops);
2137}
2138
2139static inline void unregister_chrdev(unsigned int major, const char *name)
2140{
2141 __unregister_chrdev(major, 0, 256, name);
2142}
2143
Linus Torvalds1da177e2005-04-16 15:20:36 -07002144/* fs/block_dev.c */
2145#define BDEVNAME_SIZE 32 /* Largest string for a blockdev identifier */
Tejun Heo1f014292008-08-25 19:47:23 +09002146#define BDEVT_SIZE 10 /* Largest string for MAJ:MIN for blkdev */
David Howells93614012006-09-30 20:45:40 +02002147
2148#ifdef CONFIG_BLOCK
2149#define BLKDEV_MAJOR_HASH_SIZE 255
Linus Torvalds1da177e2005-04-16 15:20:36 -07002150extern const char *__bdevname(dev_t, char *buffer);
2151extern const char *bdevname(struct block_device *bdev, char *buffer);
2152extern struct block_device *lookup_bdev(const char *);
Joe Korty68eef3b2006-03-31 02:30:32 -08002153extern void blkdev_show(struct seq_file *,off_t);
Al Viro30c40d22008-02-22 19:50:45 -05002154
David Howells93614012006-09-30 20:45:40 +02002155#else
2156#define BLKDEV_MAJOR_HASH_SIZE 0
2157#endif
Linus Torvalds1da177e2005-04-16 15:20:36 -07002158
2159extern void init_special_inode(struct inode *, umode_t, dev_t);
2160
2161/* Invalid inode operations -- fs/bad_inode.c */
2162extern void make_bad_inode(struct inode *);
2163extern int is_bad_inode(struct inode *);
2164
Denys Vlasenkod2d96482008-07-01 14:16:09 +02002165extern const struct file_operations read_pipefifo_fops;
2166extern const struct file_operations write_pipefifo_fops;
2167extern const struct file_operations rdwr_pipefifo_fops;
Linus Torvalds1da177e2005-04-16 15:20:36 -07002168
2169extern int fs_may_remount_ro(struct super_block *);
2170
David Howells93614012006-09-30 20:45:40 +02002171#ifdef CONFIG_BLOCK
Linus Torvalds1da177e2005-04-16 15:20:36 -07002172/*
2173 * return READ, READA, or WRITE
2174 */
2175#define bio_rw(bio) ((bio)->bi_rw & (RW_MASK | RWA_MASK))
2176
2177/*
2178 * return data direction, READ or WRITE
2179 */
2180#define bio_data_dir(bio) ((bio)->bi_rw & 1)
2181
Andrew Pattersonc3279d12008-09-04 14:27:25 -06002182extern void check_disk_size_change(struct gendisk *disk,
2183 struct block_device *bdev);
Andrew Patterson0c002c22008-09-04 14:27:20 -06002184extern int revalidate_disk(struct gendisk *);
Linus Torvalds1da177e2005-04-16 15:20:36 -07002185extern int check_disk_change(struct block_device *);
NeilBrown93b270f2011-02-24 17:25:47 +11002186extern int __invalidate_device(struct block_device *, bool);
Linus Torvalds1da177e2005-04-16 15:20:36 -07002187extern int invalidate_partition(struct gendisk *, int);
David Howells93614012006-09-30 20:45:40 +02002188#endif
Linus Torvalds1da177e2005-04-16 15:20:36 -07002189unsigned long invalidate_mapping_pages(struct address_space *mapping,
2190 pgoff_t start, pgoff_t end);
Anton Altaparmakov54bc4852007-02-10 01:45:38 -08002191
Linus Torvalds1da177e2005-04-16 15:20:36 -07002192static inline void invalidate_remote_inode(struct inode *inode)
2193{
2194 if (S_ISREG(inode->i_mode) || S_ISDIR(inode->i_mode) ||
2195 S_ISLNK(inode->i_mode))
Andrew Mortonfc0ecff2007-02-10 01:45:39 -08002196 invalidate_mapping_pages(inode->i_mapping, 0, -1);
Linus Torvalds1da177e2005-04-16 15:20:36 -07002197}
2198extern int invalidate_inode_pages2(struct address_space *mapping);
2199extern int invalidate_inode_pages2_range(struct address_space *mapping,
2200 pgoff_t start, pgoff_t end);
2201extern int write_inode_now(struct inode *, int);
2202extern int filemap_fdatawrite(struct address_space *);
2203extern int filemap_flush(struct address_space *);
2204extern int filemap_fdatawait(struct address_space *);
Jan Karad3bccb6f2009-08-17 19:30:27 +02002205extern int filemap_fdatawait_range(struct address_space *, loff_t lstart,
2206 loff_t lend);
Linus Torvalds1da177e2005-04-16 15:20:36 -07002207extern int filemap_write_and_wait(struct address_space *mapping);
2208extern int filemap_write_and_wait_range(struct address_space *mapping,
2209 loff_t lstart, loff_t lend);
Andrew Mortonebcf28e2006-03-24 03:18:04 -08002210extern int __filemap_fdatawrite_range(struct address_space *mapping,
2211 loff_t start, loff_t end, int sync_mode);
Jan Karaf4c0a0f2008-07-11 19:27:31 -04002212extern int filemap_fdatawrite_range(struct address_space *mapping,
2213 loff_t start, loff_t end);
Andrew Mortonebcf28e2006-03-24 03:18:04 -08002214
Christoph Hellwig8018ab02010-03-22 17:32:25 +01002215extern int vfs_fsync_range(struct file *file, loff_t start, loff_t end,
2216 int datasync);
2217extern int vfs_fsync(struct file *file, int datasync);
Jan Kara148f9482009-08-17 19:52:36 +02002218extern int generic_write_sync(struct file *file, loff_t pos, loff_t count);
Linus Torvalds1da177e2005-04-16 15:20:36 -07002219extern void sync_supers(void);
Linus Torvalds1da177e2005-04-16 15:20:36 -07002220extern void emergency_sync(void);
2221extern void emergency_remount(void);
David Howells93614012006-09-30 20:45:40 +02002222#ifdef CONFIG_BLOCK
Linus Torvalds1da177e2005-04-16 15:20:36 -07002223extern sector_t bmap(struct inode *, sector_t);
David Howells93614012006-09-30 20:45:40 +02002224#endif
Linus Torvalds1da177e2005-04-16 15:20:36 -07002225extern int notify_change(struct dentry *, struct iattr *);
Al Virof419a2e2008-07-22 00:07:17 -04002226extern int inode_permission(struct inode *, int);
Al Viro2830ba72011-06-20 19:16:29 -04002227extern int generic_permission(struct inode *, int);
Linus Torvalds1da177e2005-04-16 15:20:36 -07002228
Miklos Szeredif696a362008-07-31 13:41:58 +02002229static inline bool execute_ok(struct inode *inode)
2230{
2231 return (inode->i_mode & S_IXUGO) || S_ISDIR(inode->i_mode);
2232}
2233
Al Viro07b8ce12011-06-20 10:52:57 -04002234/*
2235 * get_write_access() gets write permission for a file.
2236 * put_write_access() releases this write permission.
2237 * This is used for regular files.
2238 * We cannot support write (and maybe mmap read-write shared) accesses and
2239 * MAP_DENYWRITE mmappings simultaneously. The i_writecount field of an inode
2240 * can have the following values:
2241 * 0: no writers, no VM_DENYWRITE mappings
2242 * < 0: (-i_writecount) vm_area_structs with VM_DENYWRITE set exist
2243 * > 0: (i_writecount) users are writing to the file.
2244 *
2245 * Normally we operate on that counter with atomic_{inc,dec} and it's safe
2246 * except for the cases where we don't hold i_writecount yet. Then we need to
2247 * use {get,deny}_write_access() - these functions check the sign and refuse
2248 * to do the change if sign is wrong.
2249 */
2250static inline int get_write_access(struct inode *inode)
2251{
2252 return atomic_inc_unless_negative(&inode->i_writecount) ? 0 : -ETXTBSY;
2253}
2254static inline int deny_write_access(struct file *file)
2255{
2256 struct inode *inode = file->f_path.dentry->d_inode;
2257 return atomic_dec_unless_positive(&inode->i_writecount) ? 0 : -ETXTBSY;
2258}
Linus Torvalds1da177e2005-04-16 15:20:36 -07002259static inline void put_write_access(struct inode * inode)
2260{
2261 atomic_dec(&inode->i_writecount);
2262}
2263static inline void allow_write_access(struct file *file)
2264{
2265 if (file)
Josef "Jeff" Sipek0f7fc9e2006-12-08 02:36:35 -08002266 atomic_inc(&file->f_path.dentry->d_inode->i_writecount);
Linus Torvalds1da177e2005-04-16 15:20:36 -07002267}
Mimi Zohara5c96eb2010-11-02 10:11:37 -04002268#ifdef CONFIG_IMA
2269static inline void i_readcount_dec(struct inode *inode)
2270{
2271 BUG_ON(!atomic_read(&inode->i_readcount));
2272 atomic_dec(&inode->i_readcount);
2273}
2274static inline void i_readcount_inc(struct inode *inode)
2275{
2276 atomic_inc(&inode->i_readcount);
2277}
2278#else
2279static inline void i_readcount_dec(struct inode *inode)
2280{
2281 return;
2282}
2283static inline void i_readcount_inc(struct inode *inode)
2284{
2285 return;
2286}
2287#endif
Ulrich Dreppered8cae82008-07-23 21:29:30 -07002288extern int do_pipe_flags(int *, int);
Ulrich Drepperbe61a862008-07-23 21:29:40 -07002289extern struct file *create_read_pipe(struct file *f, int flags);
2290extern struct file *create_write_pipe(int flags);
Andi Kleend6cbd282006-09-30 23:29:26 -07002291extern void free_write_pipe(struct file *);
Linus Torvalds1da177e2005-04-16 15:20:36 -07002292
Mimi Zohar6777d772009-08-21 14:32:48 -04002293extern int kernel_read(struct file *, loff_t, char *, unsigned long);
Linus Torvalds1da177e2005-04-16 15:20:36 -07002294extern struct file * open_exec(const char *);
2295
2296/* fs/dcache.c -- generic fs support functions */
2297extern int is_subdir(struct dentry *, struct dentry *);
Al Viro2096f752010-01-30 13:16:21 -05002298extern int path_is_under(struct path *, struct path *);
Linus Torvalds1da177e2005-04-16 15:20:36 -07002299extern ino_t find_inode_number(struct dentry *, struct qstr *);
2300
2301#include <linux/err.h>
2302
2303/* needed for stackable file system support */
2304extern loff_t default_llseek(struct file *file, loff_t offset, int origin);
2305
2306extern loff_t vfs_llseek(struct file *file, loff_t offset, int origin);
2307
Christoph Hellwig54e34622009-08-07 14:38:25 -03002308extern int inode_init_always(struct super_block *, struct inode *);
Linus Torvalds1da177e2005-04-16 15:20:36 -07002309extern void inode_init_once(struct inode *);
Miklos Szeredi2aa15892011-02-23 13:49:47 +01002310extern void address_space_init_once(struct address_space *mapping);
Al Viro7de9c6ee2010-10-23 11:11:40 -04002311extern void ihold(struct inode * inode);
Linus Torvalds1da177e2005-04-16 15:20:36 -07002312extern void iput(struct inode *);
2313extern struct inode * igrab(struct inode *);
2314extern ino_t iunique(struct super_block *, ino_t);
2315extern int inode_needs_sync(struct inode *inode);
Al Viro45321ac2010-06-07 13:43:19 -04002316extern int generic_delete_inode(struct inode *inode);
2317extern int generic_drop_inode(struct inode *inode);
Linus Torvalds1da177e2005-04-16 15:20:36 -07002318
Anton Altaparmakov88bd5122005-07-13 01:10:44 -07002319extern struct inode *ilookup5_nowait(struct super_block *sb,
2320 unsigned long hashval, int (*test)(struct inode *, void *),
2321 void *data);
Linus Torvalds1da177e2005-04-16 15:20:36 -07002322extern struct inode *ilookup5(struct super_block *sb, unsigned long hashval,
2323 int (*test)(struct inode *, void *), void *data);
2324extern struct inode *ilookup(struct super_block *sb, unsigned long ino);
2325
2326extern struct inode * iget5_locked(struct super_block *, unsigned long, int (*test)(struct inode *, void *), int (*set)(struct inode *, void *), void *);
2327extern struct inode * iget_locked(struct super_block *, unsigned long);
Al Viro261bca82008-12-30 01:48:21 -05002328extern int insert_inode_locked4(struct inode *, unsigned long, int (*test)(struct inode *, void *), void *);
2329extern int insert_inode_locked(struct inode *);
Josh Boyere096d0c2011-08-25 07:48:12 -04002330#ifdef CONFIG_DEBUG_LOCK_ALLOC
2331extern void lockdep_annotate_inode_mutex_key(struct inode *inode);
2332#else
2333static inline void lockdep_annotate_inode_mutex_key(struct inode *inode) { };
2334#endif
Linus Torvalds1da177e2005-04-16 15:20:36 -07002335extern void unlock_new_inode(struct inode *);
Christoph Hellwig85fe4022010-10-23 11:19:54 -04002336extern unsigned int get_next_ino(void);
Linus Torvalds1da177e2005-04-16 15:20:36 -07002337
Linus Torvalds1da177e2005-04-16 15:20:36 -07002338extern void __iget(struct inode * inode);
David Howellsb46980f2008-02-07 00:15:27 -08002339extern void iget_failed(struct inode *);
Al Virob0683aa2010-06-04 20:55:25 -04002340extern void end_writeback(struct inode *);
Christoph Hellwig2e00c972009-08-07 14:38:29 -03002341extern void __destroy_inode(struct inode *);
Eric Dumazeta209dfc2011-07-26 11:36:34 +02002342extern struct inode *new_inode_pseudo(struct super_block *sb);
2343extern struct inode *new_inode(struct super_block *sb);
Nick Pigginff0c7d12011-01-07 17:49:50 +11002344extern void free_inode_nonrcu(struct inode *inode);
Jens Axboe01de85e2006-10-17 19:50:36 +02002345extern int should_remove_suid(struct dentry *);
Miklos Szeredi2f1936b2008-06-24 16:50:14 +02002346extern int file_remove_suid(struct file *);
Linus Torvalds1da177e2005-04-16 15:20:36 -07002347
2348extern void __insert_inode_hash(struct inode *, unsigned long hashval);
Christoph Hellwig646ec462010-10-23 07:15:32 -04002349static inline void insert_inode_hash(struct inode *inode)
2350{
Linus Torvalds1da177e2005-04-16 15:20:36 -07002351 __insert_inode_hash(inode, inode->i_ino);
2352}
Eric Dumazetf2ee7ab2011-07-28 06:41:09 +02002353
2354extern void __remove_inode_hash(struct inode *);
2355static inline void remove_inode_hash(struct inode *inode)
2356{
2357 if (!inode_unhashed(inode))
2358 __remove_inode_hash(inode);
2359}
2360
Christoph Hellwig646ec462010-10-23 07:15:32 -04002361extern void inode_sb_list_add(struct inode *inode);
Linus Torvalds1da177e2005-04-16 15:20:36 -07002362
David Howells93614012006-09-30 20:45:40 +02002363#ifdef CONFIG_BLOCK
Linus Torvalds1da177e2005-04-16 15:20:36 -07002364extern void submit_bio(int, struct bio *);
2365extern int bdev_read_only(struct block_device *);
David Howells93614012006-09-30 20:45:40 +02002366#endif
Linus Torvalds1da177e2005-04-16 15:20:36 -07002367extern int set_blocksize(struct block_device *, int);
2368extern int sb_set_blocksize(struct super_block *, int);
2369extern int sb_min_blocksize(struct super_block *, int);
2370
2371extern int generic_file_mmap(struct file *, struct vm_area_struct *);
2372extern int generic_file_readonly_mmap(struct file *, struct vm_area_struct *);
2373extern int file_read_actor(read_descriptor_t * desc, struct page *page, unsigned long offset, unsigned long size);
Linus Torvalds1da177e2005-04-16 15:20:36 -07002374int generic_write_checks(struct file *file, loff_t *pos, size_t *count, int isblk);
Badari Pulavarty027445c2006-09-30 23:28:46 -07002375extern ssize_t generic_file_aio_read(struct kiocb *, const struct iovec *, unsigned long, loff_t);
Jan Karae4dd9de2009-08-17 18:10:06 +02002376extern ssize_t __generic_file_aio_write(struct kiocb *, const struct iovec *, unsigned long,
2377 loff_t *);
Badari Pulavarty027445c2006-09-30 23:28:46 -07002378extern ssize_t generic_file_aio_write(struct kiocb *, const struct iovec *, unsigned long, loff_t);
Linus Torvalds1da177e2005-04-16 15:20:36 -07002379extern ssize_t generic_file_direct_write(struct kiocb *, const struct iovec *,
2380 unsigned long *, loff_t, loff_t *, size_t, size_t);
2381extern ssize_t generic_file_buffered_write(struct kiocb *, const struct iovec *,
2382 unsigned long, loff_t, loff_t *, size_t, ssize_t);
2383extern ssize_t do_sync_read(struct file *filp, char __user *buf, size_t len, loff_t *ppos);
2384extern ssize_t do_sync_write(struct file *filp, const char __user *buf, size_t len, loff_t *ppos);
Dmitriy Monakhov0ceb3312007-05-08 00:23:02 -07002385extern int generic_segment_checks(const struct iovec *iov,
2386 unsigned long *nr_segs, size_t *count, int access_flags);
Andrew Morton88e6fae2006-04-11 13:59:36 +02002387
Christoph Hellwigeef99382009-08-20 17:43:41 +02002388/* fs/block_dev.c */
2389extern ssize_t blkdev_aio_write(struct kiocb *iocb, const struct iovec *iov,
2390 unsigned long nr_segs, loff_t pos);
Josef Bacik02c24a82011-07-16 20:44:56 -04002391extern int blkdev_fsync(struct file *filp, loff_t start, loff_t end,
2392 int datasync);
Christoph Hellwigeef99382009-08-20 17:43:41 +02002393
Andrew Morton88e6fae2006-04-11 13:59:36 +02002394/* fs/splice.c */
Jens Axboecbb7e572006-04-11 14:57:50 +02002395extern ssize_t generic_file_splice_read(struct file *, loff_t *,
Andrew Morton88e6fae2006-04-11 13:59:36 +02002396 struct pipe_inode_info *, size_t, unsigned int);
Miklos Szeredi68181732009-05-07 15:37:36 +02002397extern ssize_t default_file_splice_read(struct file *, loff_t *,
2398 struct pipe_inode_info *, size_t, unsigned int);
Andrew Morton88e6fae2006-04-11 13:59:36 +02002399extern ssize_t generic_file_splice_write(struct pipe_inode_info *,
Jens Axboecbb7e572006-04-11 14:57:50 +02002400 struct file *, loff_t *, size_t, unsigned int);
Andrew Morton88e6fae2006-04-11 13:59:36 +02002401extern ssize_t generic_splice_sendpage(struct pipe_inode_info *pipe,
Jens Axboecbb7e572006-04-11 14:57:50 +02002402 struct file *out, loff_t *, size_t len, unsigned int flags);
2403extern long do_splice_direct(struct file *in, loff_t *ppos, struct file *out,
Andrew Morton88e6fae2006-04-11 13:59:36 +02002404 size_t len, unsigned int flags);
2405
Linus Torvalds1da177e2005-04-16 15:20:36 -07002406extern void
2407file_ra_state_init(struct file_ra_state *ra, struct address_space *mapping);
jan Blunckae6afc32010-05-26 14:44:48 -07002408extern loff_t noop_llseek(struct file *file, loff_t offset, int origin);
Linus Torvalds1da177e2005-04-16 15:20:36 -07002409extern loff_t no_llseek(struct file *file, loff_t offset, int origin);
2410extern loff_t generic_file_llseek(struct file *file, loff_t offset, int origin);
Andi Kleen57604952011-09-15 16:06:50 -07002411extern loff_t generic_file_llseek_size(struct file *file, loff_t offset,
2412 int origin, loff_t maxsize);
Linus Torvalds1da177e2005-04-16 15:20:36 -07002413extern int generic_file_open(struct inode * inode, struct file * filp);
2414extern int nonseekable_open(struct inode * inode, struct file * filp);
2415
Carsten Otteceffc072005-06-23 22:05:25 -07002416#ifdef CONFIG_FS_XIP
Carsten Otteeb6fe0c2005-06-23 22:05:28 -07002417extern ssize_t xip_file_read(struct file *filp, char __user *buf, size_t len,
2418 loff_t *ppos);
Carsten Otteceffc072005-06-23 22:05:25 -07002419extern int xip_file_mmap(struct file * file, struct vm_area_struct * vma);
Carsten Otteeb6fe0c2005-06-23 22:05:28 -07002420extern ssize_t xip_file_write(struct file *filp, const char __user *buf,
2421 size_t len, loff_t *ppos);
Carsten Otteceffc072005-06-23 22:05:25 -07002422extern int xip_truncate_page(struct address_space *mapping, loff_t from);
Carsten Otte6d791252005-06-23 22:05:26 -07002423#else
2424static inline int xip_truncate_page(struct address_space *mapping, loff_t from)
2425{
2426 return 0;
2427}
Carsten Otteceffc072005-06-23 22:05:25 -07002428#endif
2429
David Howells93614012006-09-30 20:45:40 +02002430#ifdef CONFIG_BLOCK
Josef Bacikfacd07b2010-05-23 11:00:55 -04002431typedef void (dio_submit_t)(int rw, struct bio *bio, struct inode *inode,
2432 loff_t file_offset);
Linus Torvalds1da177e2005-04-16 15:20:36 -07002433
2434enum {
Christoph Hellwig1e431f52009-11-03 16:44:53 +01002435 /* need locking between buffered and direct access */
2436 DIO_LOCKING = 0x01,
2437
2438 /* filesystem does not support filling holes */
2439 DIO_SKIP_HOLES = 0x02,
Linus Torvalds1da177e2005-04-16 15:20:36 -07002440};
2441
Christoph Hellwigeafdc7d2010-06-04 11:29:53 +02002442void dio_end_io(struct bio *bio, int error);
Christoph Hellwigbd5fe6c2011-06-24 14:29:43 -04002443void inode_dio_wait(struct inode *inode);
2444void inode_dio_done(struct inode *inode);
npiggin@suse.de7bb46a62010-05-27 01:05:33 +10002445
Christoph Hellwigeafdc7d2010-06-04 11:29:53 +02002446ssize_t __blockdev_direct_IO(int rw, struct kiocb *iocb, struct inode *inode,
2447 struct block_device *bdev, const struct iovec *iov, loff_t offset,
2448 unsigned long nr_segs, get_block_t get_block, dio_iodone_t end_io,
2449 dio_submit_t submit_io, int flags);
2450
Linus Torvalds1da177e2005-04-16 15:20:36 -07002451static inline ssize_t blockdev_direct_IO(int rw, struct kiocb *iocb,
Christoph Hellwigaacfc19c2011-06-24 14:29:47 -04002452 struct inode *inode, const struct iovec *iov, loff_t offset,
2453 unsigned long nr_segs, get_block_t get_block)
Linus Torvalds1da177e2005-04-16 15:20:36 -07002454{
Christoph Hellwigaacfc19c2011-06-24 14:29:47 -04002455 return __blockdev_direct_IO(rw, iocb, inode, inode->i_sb->s_bdev, iov,
2456 offset, nr_segs, get_block, NULL, NULL,
Christoph Hellwig1e431f52009-11-03 16:44:53 +01002457 DIO_LOCKING | DIO_SKIP_HOLES);
Linus Torvalds1da177e2005-04-16 15:20:36 -07002458}
Randy Dunlaped70afc2011-07-21 13:55:37 -07002459#else
2460static inline void inode_dio_wait(struct inode *inode)
2461{
2462}
David Howells93614012006-09-30 20:45:40 +02002463#endif
Linus Torvalds1da177e2005-04-16 15:20:36 -07002464
Arjan van de Ven4b6f5d22006-03-28 01:56:42 -08002465extern const struct file_operations generic_ro_fops;
Linus Torvalds1da177e2005-04-16 15:20:36 -07002466
2467#define special_file(m) (S_ISCHR(m)||S_ISBLK(m)||S_ISFIFO(m)||S_ISSOCK(m))
2468
2469extern int vfs_readlink(struct dentry *, char __user *, int, const char *);
2470extern int vfs_follow_link(struct nameidata *, const char *);
2471extern int page_readlink(struct dentry *, char __user *, int);
Linus Torvaldscc314ee2005-08-19 18:02:56 -07002472extern void *page_follow_link_light(struct dentry *, struct nameidata *);
2473extern void page_put_link(struct dentry *, struct nameidata *, void *);
Kirill Korotaev0adb25d2006-03-11 03:27:13 -08002474extern int __page_symlink(struct inode *inode, const char *symname, int len,
Nick Piggin54566b22009-01-04 12:00:53 -08002475 int nofs);
Linus Torvalds1da177e2005-04-16 15:20:36 -07002476extern int page_symlink(struct inode *inode, const char *symname, int len);
Arjan van de Venc5ef1c42007-02-12 00:55:40 -08002477extern const struct inode_operations page_symlink_inode_operations;
Linus Torvalds1da177e2005-04-16 15:20:36 -07002478extern int generic_readlink(struct dentry *, char __user *, int);
2479extern void generic_fillattr(struct inode *, struct kstat *);
2480extern int vfs_getattr(struct vfsmount *, struct dentry *, struct kstat *);
Dmitry Monakhovb4627072009-12-14 15:21:12 +03002481void __inode_add_bytes(struct inode *inode, loff_t bytes);
Linus Torvalds1da177e2005-04-16 15:20:36 -07002482void inode_add_bytes(struct inode *inode, loff_t bytes);
2483void inode_sub_bytes(struct inode *inode, loff_t bytes);
2484loff_t inode_get_bytes(struct inode *inode);
2485void inode_set_bytes(struct inode *inode, loff_t bytes);
2486
2487extern int vfs_readdir(struct file *, filldir_t, void *);
2488
David Howellsc7887322010-08-11 11:26:22 +01002489extern int vfs_stat(const char __user *, struct kstat *);
2490extern int vfs_lstat(const char __user *, struct kstat *);
Linus Torvalds1da177e2005-04-16 15:20:36 -07002491extern int vfs_fstat(unsigned int, struct kstat *);
David Howellsc7887322010-08-11 11:26:22 +01002492extern int vfs_fstatat(int , const char __user *, struct kstat *, int);
Linus Torvalds1da177e2005-04-16 15:20:36 -07002493
Erez Zadokdeb21db2008-02-07 00:13:25 -08002494extern int do_vfs_ioctl(struct file *filp, unsigned int fd, unsigned int cmd,
2495 unsigned long arg);
Steven Whitehousee9079cc2008-10-14 14:43:29 +01002496extern int __generic_block_fiemap(struct inode *inode,
Josef Bacik3a3076f2010-04-23 12:17:17 -04002497 struct fiemap_extent_info *fieinfo,
2498 loff_t start, loff_t len,
2499 get_block_t *get_block);
Josef Bacik68c9d702008-10-03 17:32:43 -04002500extern int generic_block_fiemap(struct inode *inode,
2501 struct fiemap_extent_info *fieinfo, u64 start,
2502 u64 len, get_block_t *get_block);
Linus Torvalds1da177e2005-04-16 15:20:36 -07002503
Miklos Szeredic18479f2007-10-18 23:39:11 -07002504extern void get_filesystem(struct file_system_type *fs);
2505extern void put_filesystem(struct file_system_type *fs);
Linus Torvalds1da177e2005-04-16 15:20:36 -07002506extern struct file_system_type *get_fs_type(const char *name);
2507extern struct super_block *get_super(struct block_device *);
Christoph Hellwig45042302009-08-03 23:28:35 +02002508extern struct super_block *get_active_super(struct block_device *bdev);
Linus Torvalds1da177e2005-04-16 15:20:36 -07002509extern struct super_block *user_get_super(dev_t);
2510extern void drop_super(struct super_block *sb);
Al Viro01a05b32010-03-23 06:06:58 -04002511extern void iterate_supers(void (*)(struct super_block *, void *), void *);
Al Viro43e15cd2011-06-03 20:16:57 -04002512extern void iterate_supers_type(struct file_system_type *,
2513 void (*)(struct super_block *, void *), void *);
Linus Torvalds1da177e2005-04-16 15:20:36 -07002514
2515extern int dcache_dir_open(struct inode *, struct file *);
2516extern int dcache_dir_close(struct inode *, struct file *);
2517extern loff_t dcache_dir_lseek(struct file *, loff_t, int);
2518extern int dcache_readdir(struct file *, void *, filldir_t);
npiggin@suse.de7bb46a62010-05-27 01:05:33 +10002519extern int simple_setattr(struct dentry *, struct iattr *);
Linus Torvalds1da177e2005-04-16 15:20:36 -07002520extern int simple_getattr(struct vfsmount *, struct dentry *, struct kstat *);
David Howells726c3342006-06-23 02:02:58 -07002521extern int simple_statfs(struct dentry *, struct kstatfs *);
Linus Torvalds1da177e2005-04-16 15:20:36 -07002522extern int simple_link(struct dentry *, struct inode *, struct dentry *);
2523extern int simple_unlink(struct inode *, struct dentry *);
2524extern int simple_rmdir(struct inode *, struct dentry *);
2525extern int simple_rename(struct inode *, struct dentry *, struct inode *, struct dentry *);
Josef Bacik02c24a82011-07-16 20:44:56 -04002526extern int noop_fsync(struct file *, loff_t, loff_t, int);
Linus Torvalds1da177e2005-04-16 15:20:36 -07002527extern int simple_empty(struct dentry *);
2528extern int simple_readpage(struct file *file, struct page *page);
Nick Pigginafddba42007-10-16 01:25:01 -07002529extern int simple_write_begin(struct file *file, struct address_space *mapping,
2530 loff_t pos, unsigned len, unsigned flags,
2531 struct page **pagep, void **fsdata);
2532extern int simple_write_end(struct file *file, struct address_space *mapping,
2533 loff_t pos, unsigned len, unsigned copied,
2534 struct page *page, void *fsdata);
Linus Torvalds1da177e2005-04-16 15:20:36 -07002535
2536extern struct dentry *simple_lookup(struct inode *, struct dentry *, struct nameidata *);
2537extern ssize_t generic_read_dir(struct file *, char __user *, size_t, loff_t *);
Arjan van de Ven4b6f5d22006-03-28 01:56:42 -08002538extern const struct file_operations simple_dir_operations;
Arjan van de Venc5ef1c42007-02-12 00:55:40 -08002539extern const struct inode_operations simple_dir_inode_operations;
Arjan van de Ven4b6f5d22006-03-28 01:56:42 -08002540struct tree_descr { char *name; const struct file_operations *ops; int mode; };
Linus Torvalds1da177e2005-04-16 15:20:36 -07002541struct dentry *d_alloc_name(struct dentry *, const char *);
Roberto Sassu7d683a02010-06-03 11:58:28 +02002542extern int simple_fill_super(struct super_block *, unsigned long, struct tree_descr *);
Trond Myklebust1f5ce9e2006-06-09 09:34:16 -04002543extern int simple_pin_fs(struct file_system_type *, struct vfsmount **mount, int *count);
Linus Torvalds1da177e2005-04-16 15:20:36 -07002544extern void simple_release_fs(struct vfsmount **mount, int *count);
2545
Akinobu Mita93b07112008-06-05 22:46:21 -07002546extern ssize_t simple_read_from_buffer(void __user *to, size_t count,
2547 loff_t *ppos, const void *from, size_t available);
Jiri Slaby6a727b42010-05-01 23:51:22 +02002548extern ssize_t simple_write_to_buffer(void *to, size_t available, loff_t *ppos,
2549 const void __user *from, size_t count);
Linus Torvalds1da177e2005-04-16 15:20:36 -07002550
Josef Bacik02c24a82011-07-16 20:44:56 -04002551extern int generic_file_fsync(struct file *, loff_t, loff_t, int);
Al Virod5aacad2009-06-07 14:56:44 -04002552
Patrick J. LoPresti30ca22c2010-07-22 15:03:41 -07002553extern int generic_check_addressable(unsigned, u64);
2554
Christoph Lametere965f962006-02-01 03:05:41 -08002555#ifdef CONFIG_MIGRATION
Christoph Lameter2d1db3b2006-06-23 02:03:33 -07002556extern int buffer_migrate_page(struct address_space *,
2557 struct page *, struct page *);
Christoph Lametere965f962006-02-01 03:05:41 -08002558#else
2559#define buffer_migrate_page NULL
2560#endif
2561
npiggin@suse.de25d9e2d2009-08-21 02:35:05 +10002562extern int inode_change_ok(const struct inode *, struct iattr *);
2563extern int inode_newsize_ok(const struct inode *, loff_t offset);
Christoph Hellwig6a1a90a2010-06-04 11:30:00 +02002564extern void setattr_copy(struct inode *inode, const struct iattr *attr);
Linus Torvalds1da177e2005-04-16 15:20:36 -07002565
Christoph Hellwig870f4812006-01-09 20:52:01 -08002566extern void file_update_time(struct file *file);
Linus Torvalds1da177e2005-04-16 15:20:36 -07002567
Miklos Szeredib3b304a2008-02-08 04:21:35 -08002568extern int generic_show_options(struct seq_file *m, struct vfsmount *mnt);
2569extern void save_mount_options(struct super_block *sb, char *options);
Al Viro2a32ceb2009-05-08 16:05:57 -04002570extern void replace_mount_options(struct super_block *sb, char *options);
Miklos Szeredib3b304a2008-02-08 04:21:35 -08002571
Linus Torvalds1da177e2005-04-16 15:20:36 -07002572static inline ino_t parent_ino(struct dentry *dentry)
2573{
2574 ino_t res;
2575
Nick Pigginb5c84bf2011-01-07 17:49:38 +11002576 /*
2577 * Don't strictly need d_lock here? If the parent ino could change
2578 * then surely we'd have a deeper race in the caller?
2579 */
Linus Torvalds1da177e2005-04-16 15:20:36 -07002580 spin_lock(&dentry->d_lock);
2581 res = dentry->d_parent->d_inode->i_ino;
2582 spin_unlock(&dentry->d_lock);
2583 return res;
2584}
2585
Linus Torvalds1da177e2005-04-16 15:20:36 -07002586/* Transaction based IO helpers */
2587
2588/*
2589 * An argresp is stored in an allocated page and holds the
2590 * size of the argument or response, along with its content
2591 */
2592struct simple_transaction_argresp {
2593 ssize_t size;
2594 char data[0];
2595};
2596
2597#define SIMPLE_TRANSACTION_LIMIT (PAGE_SIZE - sizeof(struct simple_transaction_argresp))
2598
2599char *simple_transaction_get(struct file *file, const char __user *buf,
2600 size_t size);
2601ssize_t simple_transaction_read(struct file *file, char __user *buf,
2602 size_t size, loff_t *pos);
2603int simple_transaction_release(struct inode *inode, struct file *file);
2604
Ingo Molnar76791ab2009-03-25 16:48:35 +01002605void simple_transaction_set(struct file *file, size_t n);
Linus Torvalds1da177e2005-04-16 15:20:36 -07002606
Arnd Bergmannacaefc22005-05-18 14:40:59 +02002607/*
2608 * simple attribute files
2609 *
2610 * These attributes behave similar to those in sysfs:
2611 *
2612 * Writing to an attribute immediately sets a value, an open file can be
2613 * written to multiple times.
2614 *
2615 * Reading from an attribute creates a buffer from the value that might get
2616 * read with multiple read calls. When the attribute has been read
2617 * completely, no further read calls are possible until the file is opened
2618 * again.
2619 *
2620 * All attributes contain a text representation of a numeric value
2621 * that are accessed with the get() and set() functions.
2622 */
2623#define DEFINE_SIMPLE_ATTRIBUTE(__fops, __get, __set, __fmt) \
2624static int __fops ## _open(struct inode *inode, struct file *file) \
2625{ \
2626 __simple_attr_check_format(__fmt, 0ull); \
2627 return simple_attr_open(inode, file, __get, __set, __fmt); \
2628} \
Alexey Dobriyan828c0952009-10-01 15:43:56 -07002629static const struct file_operations __fops = { \
Arnd Bergmannacaefc22005-05-18 14:40:59 +02002630 .owner = THIS_MODULE, \
2631 .open = __fops ## _open, \
Christoph Hellwig74bedc42008-02-08 04:20:28 -08002632 .release = simple_attr_release, \
Arnd Bergmannacaefc22005-05-18 14:40:59 +02002633 .read = simple_attr_read, \
2634 .write = simple_attr_write, \
Arnd Bergmann1ec55842010-08-15 21:50:52 +02002635 .llseek = generic_file_llseek, \
Arnd Bergmannacaefc22005-05-18 14:40:59 +02002636};
2637
Joe Perchesb9075fa2011-10-31 17:11:33 -07002638static inline __printf(1, 2)
2639void __simple_attr_check_format(const char *fmt, ...)
Arnd Bergmannacaefc22005-05-18 14:40:59 +02002640{
2641 /* don't do anything, just let the compiler check the arguments; */
2642}
2643
2644int simple_attr_open(struct inode *inode, struct file *file,
Christoph Hellwig8b88b092008-02-08 04:20:26 -08002645 int (*get)(void *, u64 *), int (*set)(void *, u64),
Arnd Bergmannacaefc22005-05-18 14:40:59 +02002646 const char *fmt);
Christoph Hellwig74bedc42008-02-08 04:20:28 -08002647int simple_attr_release(struct inode *inode, struct file *file);
Arnd Bergmannacaefc22005-05-18 14:40:59 +02002648ssize_t simple_attr_read(struct file *file, char __user *buf,
2649 size_t len, loff_t *ppos);
2650ssize_t simple_attr_write(struct file *file, const char __user *buf,
2651 size_t len, loff_t *ppos);
2652
Alexey Dobriyan4be28542007-10-16 23:26:21 -07002653struct ctl_table;
Alexey Dobriyan8d65af72009-09-23 15:57:19 -07002654int proc_nr_files(struct ctl_table *table, int write,
Adrian Bunk62239ac2007-07-17 04:03:45 -07002655 void __user *buffer, size_t *lenp, loff_t *ppos);
Christoph Hellwig312d3ca2010-10-10 05:36:23 -04002656int proc_nr_dentry(struct ctl_table *table, int write,
2657 void __user *buffer, size_t *lenp, loff_t *ppos);
Dave Chinnercffbc8a2010-10-23 05:03:02 -04002658int proc_nr_inodes(struct ctl_table *table, int write,
2659 void __user *buffer, size_t *lenp, loff_t *ppos);
Tetsuo Handa38e23c92009-04-09 20:17:52 +09002660int __init get_filesystem_list(char *buf);
Adrian Bunk62239ac2007-07-17 04:03:45 -07002661
Namhyung Kim3cd90ea2011-02-01 15:52:46 -08002662#define __FMODE_EXEC ((__force int) FMODE_EXEC)
Namhyung Kim1a44bc82011-02-01 15:52:46 -08002663#define __FMODE_NONOTIFY ((__force int) FMODE_NONOTIFY)
2664
Al Viro6d125522009-12-24 06:58:56 -05002665#define ACC_MODE(x) ("\004\002\006\006"[(x)&O_ACCMODE])
Eric Parisecf081d2009-12-17 21:24:25 -05002666#define OPEN_FMODE(flag) ((__force fmode_t)(((flag + 1) & O_ACCMODE) | \
Namhyung Kim1a44bc82011-02-01 15:52:46 -08002667 (flag & __FMODE_NONOTIFY)))
Al Viro53009902009-12-19 10:15:07 -05002668
Andi Kleen69b45732011-05-28 08:25:51 -07002669static inline int is_sxid(mode_t mode)
2670{
2671 return (mode & S_ISUID) || ((mode & S_ISGID) && (mode & S_IXGRP));
2672}
2673
2674static inline void inode_has_no_xattr(struct inode *inode)
2675{
Al Viro9e1f1de2011-06-03 18:24:58 -04002676 if (!is_sxid(inode->i_mode) && (inode->i_sb->s_flags & MS_NOSEC))
Andi Kleen69b45732011-05-28 08:25:51 -07002677 inode->i_flags |= S_NOSEC;
2678}
2679
Linus Torvalds1da177e2005-04-16 15:20:36 -07002680#endif /* __KERNEL__ */
2681#endif /* _LINUX_FS_H */