blob: 4a61f98823a6edb581c61d10f015b33aeb1d0599 [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 Czerner367a51a2010-10-27 21:30:11 -040039struct fstrim_range {
Loïc Minier3a3a1af2010-11-24 12:56:53 -080040 __u64 start;
41 __u64 len;
42 __u64 minlen;
Lukas Czerner367a51a2010-10-27 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
61#define MAY_EXEC 1
62#define MAY_WRITE 2
63#define MAY_READ 4
64#define MAY_APPEND 8
Al Viroe6305c42008-07-15 21:03:57 -040065#define MAY_ACCESS 16
Al Viroa1103432008-07-17 09:19:08 -040066#define MAY_OPEN 32
Eric Paris9cfcac82010-07-23 11:43:51 -040067#define MAY_CHDIR 64
Al Viro1fc0f782011-06-20 18:59:02 -040068#define MAY_NOT_BLOCK 128 /* called from RCU mode, don't block */
Linus Torvalds1da177e2005-04-16 15:20:36 -070069
Paul Turner55ec8212009-02-18 14:48:15 -080070/*
71 * flags in file.f_mode. Note that FMODE_READ and FMODE_WRITE must correspond
72 * to O_WRONLY and O_RDWR via the strange trick in __dentry_open()
73 */
74
Christoph Hellwigfc9161e2008-11-05 14:58:46 +010075/* file is open for reading */
Andrew Morton19adf9c2010-03-05 13:42:03 -080076#define FMODE_READ ((__force fmode_t)0x1)
Christoph Hellwigfc9161e2008-11-05 14:58:46 +010077/* file is open for writing */
Andrew Morton19adf9c2010-03-05 13:42:03 -080078#define FMODE_WRITE ((__force fmode_t)0x2)
Christoph Hellwigfc9161e2008-11-05 14:58:46 +010079/* file is seekable */
Andrew Morton19adf9c2010-03-05 13:42:03 -080080#define FMODE_LSEEK ((__force fmode_t)0x4)
Paul Turner55ec8212009-02-18 14:48:15 -080081/* file can be accessed using pread */
Andrew Morton19adf9c2010-03-05 13:42:03 -080082#define FMODE_PREAD ((__force fmode_t)0x8)
Paul Turner55ec8212009-02-18 14:48:15 -080083/* file can be accessed using pwrite */
Andrew Morton19adf9c2010-03-05 13:42:03 -080084#define FMODE_PWRITE ((__force fmode_t)0x10)
Christoph Hellwigfc9161e2008-11-05 14:58:46 +010085/* File is opened for execution with sys_execve / sys_uselib */
Andrew Morton19adf9c2010-03-05 13:42:03 -080086#define FMODE_EXEC ((__force fmode_t)0x20)
Christoph Hellwigfc9161e2008-11-05 14:58:46 +010087/* File is opened with O_NDELAY (only set for block devices) */
Andrew Morton19adf9c2010-03-05 13:42:03 -080088#define FMODE_NDELAY ((__force fmode_t)0x40)
Christoph Hellwigfc9161e2008-11-05 14:58:46 +010089/* File is opened with O_EXCL (only set for block devices) */
Andrew Morton19adf9c2010-03-05 13:42:03 -080090#define FMODE_EXCL ((__force fmode_t)0x80)
Christoph Hellwigfc9161e2008-11-05 14:58:46 +010091/* File is opened using open(.., 3, ..) and is writeable only for ioctls
92 (specialy hack for floppy.c) */
Andrew Morton19adf9c2010-03-05 13:42:03 -080093#define FMODE_WRITE_IOCTL ((__force fmode_t)0x100)
Oleg Drokinb5005312006-03-25 03:07:01 -080094
Christoph Hellwig4d4be482008-12-09 04:47:33 -050095/*
96 * Don't update ctime and mtime.
97 *
98 * Currently a special hack for the XFS open_by_handle ioctl, but we'll
99 * hopefully graduate it to a proper O_CMTIME flag supported by open(2) soon.
100 */
Andrew Morton19adf9c2010-03-05 13:42:03 -0800101#define FMODE_NOCMTIME ((__force fmode_t)0x800)
Christoph Hellwig4d4be482008-12-09 04:47:33 -0500102
Wu Fengguang01414502010-03-05 13:42:03 -0800103/* Expect random access pattern */
Andrew Morton19adf9c2010-03-05 13:42:03 -0800104#define FMODE_RANDOM ((__force fmode_t)0x1000)
Wu Fengguang01414502010-03-05 13:42:03 -0800105
KAMEZAWA Hiroyuki4a3956c2010-10-01 14:20:22 -0700106/* File is huge (eg. /dev/kmem): treat loff_t as unsigned */
107#define FMODE_UNSIGNED_OFFSET ((__force fmode_t)0x2000)
108
Al Viro1abf0c72011-03-13 03:51:11 -0400109/* File is opened with O_PATH; almost nothing can be done with it */
110#define FMODE_PATH ((__force fmode_t)0x4000)
111
Eric Parisecf081d2009-12-17 21:24:25 -0500112/* File was opened by fanotify and shouldn't generate fanotify events */
Andrew Morton13bcbc02010-08-10 18:01:09 -0700113#define FMODE_NONOTIFY ((__force fmode_t)0x1000000)
Eric Parisecf081d2009-12-17 21:24:25 -0500114
Jens Axboe48e70bc2009-04-14 08:19:27 +0200115/*
116 * The below are the various read and write types that we support. Some of
117 * them include behavioral modifiers that send information down to the
118 * block layer and IO scheduler. Terminology:
119 *
120 * The block layer uses device plugging to defer IO a little bit, in
121 * the hope that we will see more IO very shortly. This increases
122 * coalescing of adjacent IO and thus reduces the number of IOs we
123 * have to send to the device. It also allows for better queuing,
124 * if the IO isn't mergeable. If the caller is going to be waiting
125 * for the IO, then he must ensure that the device is unplugged so
126 * that the IO is dispatched to the driver.
127 *
128 * All IO is handled async in Linux. This is fine for background
129 * writes, but for reads or writes that someone waits for completion
130 * on, we want to notify the block layer and IO scheduler so that they
131 * know about it. That allows them to make better scheduling
132 * decisions. So when the below references 'sync' and 'async', it
133 * is referencing this priority hint.
134 *
135 * With that in mind, the available types are:
136 *
137 * READ A normal read operation. Device will be plugged.
138 * READ_SYNC A synchronous read. Device is not plugged, caller can
139 * immediately wait on this read without caring about
140 * unplugging.
141 * READA Used for read-ahead operations. Lower priority, and the
Tejun Heo7cc01582010-08-03 13:14:58 +0200142 * block layer could (in theory) choose to ignore this
Jens Axboe48e70bc2009-04-14 08:19:27 +0200143 * request if it runs into resource problems.
144 * WRITE A normal async write. Device will be plugged.
Jens Axboe721a9602011-03-09 11:56:30 +0100145 * WRITE_SYNC Synchronous write. Identical to WRITE, but passes down
Jens Axboe48e70bc2009-04-14 08:19:27 +0200146 * the hint that someone will be waiting on this IO
Jens Axboe721a9602011-03-09 11:56:30 +0100147 * shortly. The write equivalent of READ_SYNC.
148 * WRITE_ODIRECT Special case write for O_DIRECT only.
Tejun Heo4fed9472010-09-03 11:56:17 +0200149 * WRITE_FLUSH Like WRITE_SYNC but with preceding cache flush.
150 * WRITE_FUA Like WRITE_SYNC but data is guaranteed to be on
151 * non-volatile media on completion.
152 * WRITE_FLUSH_FUA Combination of WRITE_FLUSH and FUA. The IO is preceded
153 * by a cache flush and data is guaranteed to be on
154 * non-volatile media on completion.
Jens Axboe48e70bc2009-04-14 08:19:27 +0200155 *
156 */
Tejun Heo7cc01582010-08-03 13:14:58 +0200157#define RW_MASK REQ_WRITE
158#define RWA_MASK REQ_RAHEAD
Christoph Hellwig7b6d91d2010-08-07 18:20:39 +0200159
160#define READ 0
Tejun Heo7cc01582010-08-03 13:14:58 +0200161#define WRITE RW_MASK
162#define READA RWA_MASK
Christoph Hellwig7b6d91d2010-08-07 18:20:39 +0200163
Jens Axboe721a9602011-03-09 11:56:30 +0100164#define READ_SYNC (READ | REQ_SYNC)
Christoph Hellwig7b6d91d2010-08-07 18:20:39 +0200165#define READ_META (READ | REQ_META)
Jens Axboe721a9602011-03-09 11:56:30 +0100166#define WRITE_SYNC (WRITE | REQ_SYNC | REQ_NOIDLE)
167#define WRITE_ODIRECT (WRITE | REQ_SYNC)
Christoph Hellwig7b6d91d2010-08-07 18:20:39 +0200168#define WRITE_META (WRITE | REQ_META)
Jens Axboe721a9602011-03-09 11:56:30 +0100169#define WRITE_FLUSH (WRITE | REQ_SYNC | REQ_NOIDLE | REQ_FLUSH)
170#define WRITE_FUA (WRITE | REQ_SYNC | REQ_NOIDLE | REQ_FUA)
171#define WRITE_FLUSH_FUA (WRITE | REQ_SYNC | REQ_NOIDLE | REQ_FLUSH | REQ_FUA)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700172
173#define SEL_IN 1
174#define SEL_OUT 2
175#define SEL_EX 4
176
177/* public flags for file_system_type */
178#define FS_REQUIRES_DEV 1
179#define FS_BINARY_MOUNTDATA 2
Miklos Szeredi79c0b2d2007-05-08 00:25:43 -0700180#define FS_HAS_SUBTYPE 4
Linus Torvalds1da177e2005-04-16 15:20:36 -0700181#define FS_REVAL_DOT 16384 /* Check the paths ".", ".." for staleness */
Mark Fasheh349457c2006-09-08 14:22:21 -0700182#define FS_RENAME_DOES_D_MOVE 32768 /* FS will handle d_move()
183 * during rename() internally.
184 */
185
Linus Torvalds1da177e2005-04-16 15:20:36 -0700186/*
187 * These are the fs-independent mount-flags: up to 32 flags are supported
188 */
189#define MS_RDONLY 1 /* Mount read-only */
190#define MS_NOSUID 2 /* Ignore suid and sgid bits */
191#define MS_NODEV 4 /* Disallow access to device special files */
192#define MS_NOEXEC 8 /* Disallow program execution */
193#define MS_SYNCHRONOUS 16 /* Writes are synced at once */
194#define MS_REMOUNT 32 /* Alter flags of a mounted FS */
195#define MS_MANDLOCK 64 /* Allow mandatory locks on an FS */
196#define MS_DIRSYNC 128 /* Directory modifications are synchronous */
197#define MS_NOATIME 1024 /* Do not update access times. */
198#define MS_NODIRATIME 2048 /* Do not update directory access times */
199#define MS_BIND 4096
200#define MS_MOVE 8192
201#define MS_REC 16384
Theodore Ts'o9b04c992006-03-24 03:15:10 -0800202#define MS_VERBOSE 32768 /* War is peace. Verbosity is silence.
203 MS_VERBOSE is deprecated. */
204#define MS_SILENT 32768
Miklos Szeredibf066c72006-01-08 01:03:19 -0800205#define MS_POSIXACL (1<<16) /* VFS does not apply the umask */
Ram Pai9676f0c2005-11-07 17:21:20 -0500206#define MS_UNBINDABLE (1<<17) /* change to unbindable */
Ram Pai07b20882005-11-07 17:19:07 -0500207#define MS_PRIVATE (1<<18) /* change to private */
Ram Paia58b0eb2005-11-07 17:20:48 -0500208#define MS_SLAVE (1<<19) /* change to slave */
Ram Pai03e06e62005-11-07 17:19:33 -0500209#define MS_SHARED (1<<20) /* change to shared */
Valerie Henson47ae32d2006-12-13 00:34:34 -0800210#define MS_RELATIME (1<<21) /* Update atime relative to mtime/ctime. */
Pavel Emelyanov8bf97252007-10-18 23:40:02 -0700211#define MS_KERNMOUNT (1<<22) /* this is a kern_mount call */
Jean Noel Cordenner7a224222008-01-28 23:58:27 -0500212#define MS_I_VERSION (1<<23) /* Update inode I_version field */
Matthew Garrettd0adde52009-03-26 17:49:56 +0000213#define MS_STRICTATIME (1<<24) /* Always perform atime updates */
Al Viro9e1f1de2011-06-03 18:24:58 -0400214#define MS_NOSEC (1<<28)
Al Viro7a4dec52010-08-09 12:05:43 -0400215#define MS_BORN (1<<29)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700216#define MS_ACTIVE (1<<30)
217#define MS_NOUSER (1<<31)
218
219/*
220 * Superblock flags that can be altered by MS_REMOUNT
221 */
Mimi Zohar08b9fe62008-10-13 00:09:50 -0400222#define MS_RMT_MASK (MS_RDONLY|MS_SYNCHRONOUS|MS_MANDLOCK|MS_I_VERSION)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700223
224/*
225 * Old magic mount flag and mask
226 */
227#define MS_MGC_VAL 0xC0ED0000
228#define MS_MGC_MSK 0xffff0000
229
230/* Inode flags - they have nothing to superblock flags now */
231
232#define S_SYNC 1 /* Writes are synced at once */
233#define S_NOATIME 2 /* Do not update access times */
234#define S_APPEND 4 /* Append-only file */
235#define S_IMMUTABLE 8 /* Immutable file */
236#define S_DEAD 16 /* removed, but still open directory */
237#define S_NOQUOTA 32 /* Inode is not counted to quota */
238#define S_DIRSYNC 64 /* Directory modifications are synchronous */
239#define S_NOCMTIME 128 /* Do not update file c/mtime */
240#define S_SWAPFILE 256 /* Do not truncate: swapon got its bmaps */
241#define S_PRIVATE 512 /* Inode is fs-internal */
Eric Paris196f5182010-10-25 14:42:19 -0400242#define S_IMA 1024 /* Inode has an associated IMA struct */
David Howells9875cf82011-01-14 18:45:21 +0000243#define S_AUTOMOUNT 2048 /* Automount/referral quasi-directory */
Andi Kleen69b45732011-05-28 08:25:51 -0700244#define S_NOSEC 4096 /* no suid or xattr security attributes */
Linus Torvalds1da177e2005-04-16 15:20:36 -0700245
246/*
247 * Note that nosuid etc flags are inode-specific: setting some file-system
248 * flags just means all the inodes inherit those flags by default. It might be
249 * possible to override it selectively if you really wanted to with some
250 * ioctl() that is not currently implemented.
251 *
252 * Exception: MS_RDONLY is always applied to the entire file system.
253 *
254 * Unfortunately, it is possible to change a filesystems flags with it mounted
255 * with files in use. This means that all of the inodes will not have their
256 * i_flags updated. Hence, i_flags no longer inherit the superblock mount
257 * flags, so these have to be checked separately. -- rmk@arm.uk.linux.org
258 */
259#define __IS_FLG(inode,flg) ((inode)->i_sb->s_flags & (flg))
260
261#define IS_RDONLY(inode) ((inode)->i_sb->s_flags & MS_RDONLY)
262#define IS_SYNC(inode) (__IS_FLG(inode, MS_SYNCHRONOUS) || \
263 ((inode)->i_flags & S_SYNC))
264#define IS_DIRSYNC(inode) (__IS_FLG(inode, MS_SYNCHRONOUS|MS_DIRSYNC) || \
265 ((inode)->i_flags & (S_SYNC|S_DIRSYNC)))
266#define IS_MANDLOCK(inode) __IS_FLG(inode, MS_MANDLOCK)
Eric Dumazet37756ce2007-02-10 01:44:49 -0800267#define IS_NOATIME(inode) __IS_FLG(inode, MS_RDONLY|MS_NOATIME)
Jean Noel Cordenner7a224222008-01-28 23:58:27 -0500268#define IS_I_VERSION(inode) __IS_FLG(inode, MS_I_VERSION)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700269
270#define IS_NOQUOTA(inode) ((inode)->i_flags & S_NOQUOTA)
271#define IS_APPEND(inode) ((inode)->i_flags & S_APPEND)
272#define IS_IMMUTABLE(inode) ((inode)->i_flags & S_IMMUTABLE)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700273#define IS_POSIXACL(inode) __IS_FLG(inode, MS_POSIXACL)
274
275#define IS_DEADDIR(inode) ((inode)->i_flags & S_DEAD)
276#define IS_NOCMTIME(inode) ((inode)->i_flags & S_NOCMTIME)
277#define IS_SWAPFILE(inode) ((inode)->i_flags & S_SWAPFILE)
278#define IS_PRIVATE(inode) ((inode)->i_flags & S_PRIVATE)
Eric Paris196f5182010-10-25 14:42:19 -0400279#define IS_IMA(inode) ((inode)->i_flags & S_IMA)
David Howells9875cf82011-01-14 18:45:21 +0000280#define IS_AUTOMOUNT(inode) ((inode)->i_flags & S_AUTOMOUNT)
Andi Kleen69b45732011-05-28 08:25:51 -0700281#define IS_NOSEC(inode) ((inode)->i_flags & S_NOSEC)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700282
283/* the read-only stuff doesn't really belong here, but any other place is
284 probably as bad and I don't want to create yet another include file. */
285
286#define BLKROSET _IO(0x12,93) /* set device read-only (0 = read-write) */
287#define BLKROGET _IO(0x12,94) /* get read-only status (0 = read_write) */
288#define BLKRRPART _IO(0x12,95) /* re-read partition table */
289#define BLKGETSIZE _IO(0x12,96) /* return device size /512 (long *arg) */
290#define BLKFLSBUF _IO(0x12,97) /* flush buffer cache */
291#define BLKRASET _IO(0x12,98) /* set read ahead for block device */
292#define BLKRAGET _IO(0x12,99) /* get current read ahead setting */
293#define BLKFRASET _IO(0x12,100)/* set filesystem (mm/filemap.c) read-ahead */
294#define BLKFRAGET _IO(0x12,101)/* get filesystem (mm/filemap.c) read-ahead */
295#define BLKSECTSET _IO(0x12,102)/* set max sectors per request (ll_rw_blk.c) */
296#define BLKSECTGET _IO(0x12,103)/* get max sectors per request (ll_rw_blk.c) */
297#define BLKSSZGET _IO(0x12,104)/* get block device sector size */
298#if 0
299#define BLKPG _IO(0x12,105)/* See blkpg.h */
300
301/* Some people are morons. Do not use sizeof! */
302
303#define BLKELVGET _IOR(0x12,106,size_t)/* elevator get */
304#define BLKELVSET _IOW(0x12,107,size_t)/* elevator set */
305/* This was here just to show that the number is taken -
306 probably all these _IO(0x12,*) ioctls should be moved to blkpg.h. */
307#endif
308/* A jump here: 108-111 have been used for various private purposes. */
309#define BLKBSZGET _IOR(0x12,112,size_t)
310#define BLKBSZSET _IOW(0x12,113,size_t)
311#define BLKGETSIZE64 _IOR(0x12,114,size_t) /* return device size in bytes (u64 *arg) */
Jens Axboe2056a782006-03-23 20:00:26 +0100312#define BLKTRACESETUP _IOWR(0x12,115,struct blk_user_trace_setup)
313#define BLKTRACESTART _IO(0x12,116)
314#define BLKTRACESTOP _IO(0x12,117)
315#define BLKTRACETEARDOWN _IO(0x12,118)
David Woodhoused30a2602008-08-11 15:58:42 +0100316#define BLKDISCARD _IO(0x12,119)
Martin K. Petersenac481c22009-10-03 20:52:01 +0200317#define BLKIOMIN _IO(0x12,120)
318#define BLKIOOPT _IO(0x12,121)
319#define BLKALIGNOFF _IO(0x12,122)
320#define BLKPBSZGET _IO(0x12,123)
Martin K. Petersen98262f22009-12-03 09:24:48 +0100321#define BLKDISCARDZEROES _IO(0x12,124)
Adrian Hunter8d57a982010-08-11 14:17:49 -0700322#define BLKSECDISCARD _IO(0x12,125)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700323
324#define BMAP_IOCTL 1 /* obsolete - kept for compatibility */
325#define FIBMAP _IO(0x00,1) /* bmap access */
326#define FIGETBSZ _IO(0x00,2) /* get the block size used for bmap */
Takashi Satofcccf502009-01-09 16:40:59 -0800327#define FIFREEZE _IOWR('X', 119, int) /* Freeze */
328#define FITHAW _IOWR('X', 120, int) /* Thaw */
Lukas Czerner367a51a2010-10-27 21:30:11 -0400329#define FITRIM _IOWR('X', 121, struct fstrim_range) /* Trim */
Linus Torvalds1da177e2005-04-16 15:20:36 -0700330
David Howells36695672006-08-29 19:06:16 +0100331#define FS_IOC_GETFLAGS _IOR('f', 1, long)
332#define FS_IOC_SETFLAGS _IOW('f', 2, long)
333#define FS_IOC_GETVERSION _IOR('v', 1, long)
334#define FS_IOC_SETVERSION _IOW('v', 2, long)
Mark Fashehc4b929b2008-10-08 19:44:18 -0400335#define FS_IOC_FIEMAP _IOWR('f', 11, struct fiemap)
David Howells36695672006-08-29 19:06:16 +0100336#define FS_IOC32_GETFLAGS _IOR('f', 1, int)
337#define FS_IOC32_SETFLAGS _IOW('f', 2, int)
338#define FS_IOC32_GETVERSION _IOR('v', 1, int)
339#define FS_IOC32_SETVERSION _IOW('v', 2, int)
340
341/*
342 * Inode flags (FS_IOC_GETFLAGS / FS_IOC_SETFLAGS)
343 */
344#define FS_SECRM_FL 0x00000001 /* Secure deletion */
345#define FS_UNRM_FL 0x00000002 /* Undelete */
346#define FS_COMPR_FL 0x00000004 /* Compress file */
347#define FS_SYNC_FL 0x00000008 /* Synchronous updates */
348#define FS_IMMUTABLE_FL 0x00000010 /* Immutable file */
349#define FS_APPEND_FL 0x00000020 /* writes to file may only append */
350#define FS_NODUMP_FL 0x00000040 /* do not dump file */
351#define FS_NOATIME_FL 0x00000080 /* do not update atime */
352/* Reserved for compression usage... */
353#define FS_DIRTY_FL 0x00000100
354#define FS_COMPRBLK_FL 0x00000200 /* One or more compressed clusters */
355#define FS_NOCOMP_FL 0x00000400 /* Don't compress */
356#define FS_ECOMPR_FL 0x00000800 /* Compression error */
357/* End compression flags --- maybe not all used */
358#define FS_BTREE_FL 0x00001000 /* btree format dir */
359#define FS_INDEX_FL 0x00001000 /* hash-indexed directory */
360#define FS_IMAGIC_FL 0x00002000 /* AFS directory */
361#define FS_JOURNAL_DATA_FL 0x00004000 /* Reserved for ext3 */
362#define FS_NOTAIL_FL 0x00008000 /* file tail should not be merged */
363#define FS_DIRSYNC_FL 0x00010000 /* dirsync behaviour (directories only) */
364#define FS_TOPDIR_FL 0x00020000 /* Top of directory hierarchies*/
Steven Whitehouse128e5eb2006-10-02 11:24:43 -0400365#define FS_EXTENT_FL 0x00080000 /* Extents */
366#define FS_DIRECTIO_FL 0x00100000 /* Use direct i/o */
liubo32471f62011-03-21 08:54:27 +0000367#define FS_NOCOW_FL 0x00800000 /* Do not cow file */
David Howells36695672006-08-29 19:06:16 +0100368#define FS_RESERVED_FL 0x80000000 /* reserved for ext2 lib */
369
370#define FS_FL_USER_VISIBLE 0x0003DFFF /* User visible flags */
371#define FS_FL_USER_MODIFIABLE 0x000380FF /* User modifiable flags */
372
373
Andrew Mortone46e4902006-05-22 22:35:24 -0700374#define SYNC_FILE_RANGE_WAIT_BEFORE 1
375#define SYNC_FILE_RANGE_WRITE 2
376#define SYNC_FILE_RANGE_WAIT_AFTER 4
377
Linus Torvalds1da177e2005-04-16 15:20:36 -0700378#ifdef __KERNEL__
379
380#include <linux/linkage.h>
381#include <linux/wait.h>
382#include <linux/types.h>
383#include <linux/kdev_t.h>
384#include <linux/dcache.h>
Al Viro3f8206d2008-07-26 03:46:43 -0400385#include <linux/path.h>
Linus Torvalds1da177e2005-04-16 15:20:36 -0700386#include <linux/stat.h>
387#include <linux/cache.h>
Linus Torvalds1da177e2005-04-16 15:20:36 -0700388#include <linux/list.h>
389#include <linux/radix-tree.h>
390#include <linux/prio_tree.h>
391#include <linux/init.h>
Al Viro914e2632006-10-18 13:55:46 -0400392#include <linux/pid.h>
Jes Sorensen1b1dcc12006-01-09 15:59:24 -0800393#include <linux/mutex.h>
Satyam Sharma3bd858a2007-07-17 15:00:08 +0530394#include <linux/capability.h>
Matthew Wilcox6188e102008-04-18 22:21:05 -0400395#include <linux/semaphore.h>
Mark Fashehc4b929b2008-10-08 19:44:18 -0400396#include <linux/fiemap.h>
Nick Pigginceb5bdc2011-01-07 17:50:05 +1100397#include <linux/rculist_bl.h>
Dave Chinnerb0d40c92011-07-08 14:14:42 +1000398#include <linux/shrinker.h>
Al Viro07b8ce12011-06-20 10:52:57 -0400399#include <linux/atomic.h>
Linus Torvalds1da177e2005-04-16 15:20:36 -0700400
Linus Torvalds1da177e2005-04-16 15:20:36 -0700401#include <asm/byteorder.h>
402
Christoph Hellwiga5694252007-07-17 04:04:28 -0700403struct export_operations;
Christoph Hellwiga885c8c2006-01-08 01:02:50 -0800404struct hd_geometry;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700405struct iovec;
406struct nameidata;
Christoph Hellwig92198f72005-06-23 22:00:59 -0700407struct kiocb;
Alexey Dobriyan57cc7212011-01-10 08:18:25 +0200408struct kobject;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700409struct pipe_inode_info;
410struct poll_table_struct;
411struct kstatfs;
412struct vm_area_struct;
413struct vfsmount;
David Howells745ca242008-11-14 10:39:22 +1100414struct cred;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700415
Denis Cheng74bf17c2007-10-16 23:26:30 -0700416extern void __init inode_init(void);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700417extern void __init inode_init_early(void);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700418extern void __init files_init(unsigned long);
419
Jan Engelhardtdded4f42008-12-01 14:34:50 -0800420extern struct files_stat_struct files_stat;
Eric Dumazet518de9b2010-10-26 14:22:44 -0700421extern unsigned long get_max_files(void);
Jan Engelhardtdded4f42008-12-01 14:34:50 -0800422extern int sysctl_nr_open;
423extern struct inodes_stat_t inodes_stat;
424extern int leases_enable, lease_break_time;
Jan Engelhardtdded4f42008-12-01 14:34:50 -0800425
Linus Torvalds1da177e2005-04-16 15:20:36 -0700426struct buffer_head;
427typedef int (get_block_t)(struct inode *inode, sector_t iblock,
428 struct buffer_head *bh_result, int create);
Christoph Hellwig92198f72005-06-23 22:00:59 -0700429typedef void (dio_iodone_t)(struct kiocb *iocb, loff_t offset,
Christoph Hellwig40e2e972010-07-18 21:17:09 +0000430 ssize_t bytes, void *private, int ret,
431 bool is_async);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700432
433/*
434 * Attribute flags. These should be or-ed together to figure out what
435 * has been changed!
436 */
Miklos Szeredi9767d742008-07-01 15:01:26 +0200437#define ATTR_MODE (1 << 0)
438#define ATTR_UID (1 << 1)
439#define ATTR_GID (1 << 2)
440#define ATTR_SIZE (1 << 3)
441#define ATTR_ATIME (1 << 4)
442#define ATTR_MTIME (1 << 5)
443#define ATTR_CTIME (1 << 6)
444#define ATTR_ATIME_SET (1 << 7)
445#define ATTR_MTIME_SET (1 << 8)
446#define ATTR_FORCE (1 << 9) /* Not a change, but a change it */
447#define ATTR_ATTR_FLAG (1 << 10)
448#define ATTR_KILL_SUID (1 << 11)
449#define ATTR_KILL_SGID (1 << 12)
450#define ATTR_FILE (1 << 13)
451#define ATTR_KILL_PRIV (1 << 14)
452#define ATTR_OPEN (1 << 15) /* Truncating from open(O_TRUNC) */
453#define ATTR_TIMES_SET (1 << 16)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700454
455/*
456 * This is the Inode Attributes structure, used for notify_change(). It
457 * uses the above definitions as flags, to know which values have changed.
458 * Also, in this manner, a Filesystem can look at only the values it cares
459 * about. Basically, these are the attributes that the VFS layer can
460 * request to change from the FS layer.
461 *
462 * Derek Atkins <warlord@MIT.EDU> 94-10-20
463 */
464struct iattr {
465 unsigned int ia_valid;
466 umode_t ia_mode;
467 uid_t ia_uid;
468 gid_t ia_gid;
469 loff_t ia_size;
470 struct timespec ia_atime;
471 struct timespec ia_mtime;
472 struct timespec ia_ctime;
Miklos Szeredicc4e69d2005-11-07 00:59:49 -0800473
474 /*
Lucas De Marchi25985ed2011-03-30 22:57:33 -0300475 * Not an attribute, but an auxiliary info for filesystems wanting to
Miklos Szeredicc4e69d2005-11-07 00:59:49 -0800476 * implement an ftruncate() like method. NOTE: filesystem should
477 * check for (ia_valid & ATTR_FILE), and not for (ia_file != NULL).
478 */
479 struct file *ia_file;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700480};
481
482/*
Linus Torvalds1da177e2005-04-16 15:20:36 -0700483 * Includes for diskquotas.
484 */
485#include <linux/quota.h>
486
Zach Brown994fc28c2005-12-15 14:28:17 -0800487/**
488 * enum positive_aop_returns - aop return codes with specific semantics
489 *
490 * @AOP_WRITEPAGE_ACTIVATE: Informs the caller that page writeback has
491 * completed, that the page is still locked, and
492 * should be considered active. The VM uses this hint
493 * to return the page to the active list -- it won't
494 * be a candidate for writeback again in the near
495 * future. Other callers must be careful to unlock
496 * the page if they get this return. Returned by
497 * writepage();
498 *
499 * @AOP_TRUNCATED_PAGE: The AOP method that was handed a locked page has
500 * unlocked it and the page might have been truncated.
501 * The caller should back up to acquiring a new page and
502 * trying again. The aop will be taking reasonable
503 * precautions not to livelock. If the caller held a page
504 * reference, it should drop it before retrying. Returned
Nick Piggin55144762007-10-16 01:25:26 -0700505 * by readpage().
Zach Brown994fc28c2005-12-15 14:28:17 -0800506 *
507 * address_space_operation functions return these large constants to indicate
508 * special semantics to the caller. These are much larger than the bytes in a
509 * page to allow for functions that return the number of bytes operated on in a
510 * given page.
511 */
512
513enum positive_aop_returns {
514 AOP_WRITEPAGE_ACTIVATE = 0x80000,
515 AOP_TRUNCATED_PAGE = 0x80001,
516};
517
Nick Pigginafddba42007-10-16 01:25:01 -0700518#define AOP_FLAG_UNINTERRUPTIBLE 0x0001 /* will not do a short write */
Nick Piggin89e10782007-10-16 01:25:07 -0700519#define AOP_FLAG_CONT_EXPAND 0x0002 /* called from cont_expand */
Nick Piggin54566b22009-01-04 12:00:53 -0800520#define AOP_FLAG_NOFS 0x0004 /* used by filesystem to direct
521 * helper code (eg buffer layer)
522 * to clear GFP_FS from alloc */
Nick Pigginafddba42007-10-16 01:25:01 -0700523
Linus Torvalds1da177e2005-04-16 15:20:36 -0700524/*
525 * oh the beauties of C type declarations.
526 */
527struct page;
528struct address_space;
529struct writeback_control;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700530
Nick Piggin2f718ff2007-10-16 01:24:59 -0700531struct iov_iter {
532 const struct iovec *iov;
533 unsigned long nr_segs;
534 size_t iov_offset;
535 size_t count;
536};
537
538size_t iov_iter_copy_from_user_atomic(struct page *page,
539 struct iov_iter *i, unsigned long offset, size_t bytes);
540size_t iov_iter_copy_from_user(struct page *page,
541 struct iov_iter *i, unsigned long offset, size_t bytes);
542void iov_iter_advance(struct iov_iter *i, size_t bytes);
Nick Pigginafddba42007-10-16 01:25:01 -0700543int iov_iter_fault_in_readable(struct iov_iter *i, size_t bytes);
Nick Piggin2f718ff2007-10-16 01:24:59 -0700544size_t iov_iter_single_seg_count(struct iov_iter *i);
545
546static inline void iov_iter_init(struct iov_iter *i,
547 const struct iovec *iov, unsigned long nr_segs,
548 size_t count, size_t written)
549{
550 i->iov = iov;
551 i->nr_segs = nr_segs;
552 i->iov_offset = 0;
553 i->count = count + written;
554
555 iov_iter_advance(i, written);
556}
557
558static inline size_t iov_iter_count(struct iov_iter *i)
559{
560 return i->count;
561}
562
Hisashi Hifumi8ab22b92008-07-28 15:46:36 -0700563/*
564 * "descriptor" for what we're up to with a read.
565 * This allows us to use the same read code yet
566 * have multiple different users of the data that
567 * we read from a file.
568 *
569 * The simplest case just copies the data to user
570 * mode.
571 */
572typedef struct {
573 size_t written;
574 size_t count;
575 union {
576 char __user *buf;
577 void *data;
578 } arg;
579 int error;
580} read_descriptor_t;
581
582typedef int (*read_actor_t)(read_descriptor_t *, struct page *,
583 unsigned long, unsigned long);
Nick Piggin2f718ff2007-10-16 01:24:59 -0700584
Linus Torvalds1da177e2005-04-16 15:20:36 -0700585struct address_space_operations {
586 int (*writepage)(struct page *page, struct writeback_control *wbc);
587 int (*readpage)(struct file *, struct page *);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700588
589 /* Write back some dirty pages from this mapping. */
590 int (*writepages)(struct address_space *, struct writeback_control *);
591
Andrew Morton4741c9f2006-03-24 03:18:11 -0800592 /* Set a page dirty. Return true if this dirtied it */
Linus Torvalds1da177e2005-04-16 15:20:36 -0700593 int (*set_page_dirty)(struct page *page);
594
595 int (*readpages)(struct file *filp, struct address_space *mapping,
596 struct list_head *pages, unsigned nr_pages);
597
Nick Pigginafddba42007-10-16 01:25:01 -0700598 int (*write_begin)(struct file *, struct address_space *mapping,
599 loff_t pos, unsigned len, unsigned flags,
600 struct page **pagep, void **fsdata);
601 int (*write_end)(struct file *, struct address_space *mapping,
602 loff_t pos, unsigned len, unsigned copied,
603 struct page *page, void *fsdata);
604
Linus Torvalds1da177e2005-04-16 15:20:36 -0700605 /* Unfortunately this kludge is needed for FIBMAP. Don't use it */
606 sector_t (*bmap)(struct address_space *, sector_t);
NeilBrown2ff28e22006-03-26 01:37:18 -0800607 void (*invalidatepage) (struct page *, unsigned long);
Al Viro27496a82005-10-21 03:20:48 -0400608 int (*releasepage) (struct page *, gfp_t);
Linus Torvalds6072d132010-12-01 13:35:19 -0500609 void (*freepage)(struct page *);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700610 ssize_t (*direct_IO)(int, struct kiocb *, const struct iovec *iov,
611 loff_t offset, unsigned long nr_segs);
Nick Piggin70688e42008-04-28 02:13:02 -0700612 int (*get_xip_mem)(struct address_space *, pgoff_t, int,
613 void **, unsigned long *);
Christoph Lametere965f962006-02-01 03:05:41 -0800614 /* migrate the contents of a page to the specified target */
Christoph Lameter2d1db3b2006-06-23 02:03:33 -0700615 int (*migratepage) (struct address_space *,
616 struct page *, struct page *);
Trond Myklebuste3db7692007-01-10 23:15:39 -0800617 int (*launder_page) (struct page *);
Hisashi Hifumi8ab22b92008-07-28 15:46:36 -0700618 int (*is_partially_uptodate) (struct page *, read_descriptor_t *,
619 unsigned long);
Andi Kleen25718732009-09-16 11:50:13 +0200620 int (*error_remove_page)(struct address_space *, struct page *);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700621};
622
Jens Axboe7dcda1c2011-04-05 23:51:48 +0200623extern const struct address_space_operations empty_aops;
624
Nick Pigginafddba42007-10-16 01:25:01 -0700625/*
626 * pagecache_write_begin/pagecache_write_end must be used by general code
627 * to write into the pagecache.
628 */
629int pagecache_write_begin(struct file *, struct address_space *mapping,
630 loff_t pos, unsigned len, unsigned flags,
631 struct page **pagep, void **fsdata);
632
633int pagecache_write_end(struct file *, struct address_space *mapping,
634 loff_t pos, unsigned len, unsigned copied,
635 struct page *page, void *fsdata);
636
Linus Torvalds1da177e2005-04-16 15:20:36 -0700637struct backing_dev_info;
638struct address_space {
639 struct inode *host; /* owner: inode, block_device */
640 struct radix_tree_root page_tree; /* radix tree of all pages */
Nick Piggin19fd6232008-07-25 19:45:32 -0700641 spinlock_t tree_lock; /* and lock protecting it */
Linus Torvalds1da177e2005-04-16 15:20:36 -0700642 unsigned int i_mmap_writable;/* count VM_SHARED mappings */
643 struct prio_tree_root i_mmap; /* tree of private and shared mappings */
644 struct list_head i_mmap_nonlinear;/*list VM_NONLINEAR mappings */
Peter Zijlstra3d48ae42011-05-24 17:12:06 -0700645 struct mutex i_mmap_mutex; /* protect tree, count, list */
Jan Kara08142572011-06-27 16:18:10 -0700646 /* Protected by tree_lock together with the radix tree */
Linus Torvalds1da177e2005-04-16 15:20:36 -0700647 unsigned long nrpages; /* number of total pages */
648 pgoff_t writeback_index;/* writeback starts here */
Christoph Hellwigf5e54d62006-06-28 04:26:44 -0700649 const struct address_space_operations *a_ops; /* methods */
Linus Torvalds1da177e2005-04-16 15:20:36 -0700650 unsigned long flags; /* error bits/gfp mask */
651 struct backing_dev_info *backing_dev_info; /* device readahead, etc */
652 spinlock_t private_lock; /* for use by the address_space */
653 struct list_head private_list; /* ditto */
654 struct address_space *assoc_mapping; /* ditto */
655} __attribute__((aligned(sizeof(long))));
656 /*
657 * On most architectures that alignment is already the case; but
Lucas De Marchi25985ed2011-03-30 22:57:33 -0300658 * must be enforced here for CRIS, to let the least significant bit
Linus Torvalds1da177e2005-04-16 15:20:36 -0700659 * of struct page's "mapping" pointer be used for PAGE_MAPPING_ANON.
660 */
661
662struct block_device {
663 dev_t bd_dev; /* not a kdev_t - it's a search key */
Richard Kennedyf3ccfcd2011-03-22 16:33:56 -0700664 int bd_openers;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700665 struct inode * bd_inode; /* will die */
Theodore Ts'o87d8fe12009-01-03 09:47:09 -0500666 struct super_block * bd_super;
Arjan van de Venc039e312006-03-23 03:00:28 -0800667 struct mutex bd_mutex; /* open/close mutex */
Linus Torvalds1da177e2005-04-16 15:20:36 -0700668 struct list_head bd_inodes;
Tejun Heo6b4517a2010-04-07 18:53:59 +0900669 void * bd_claiming;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700670 void * bd_holder;
671 int bd_holders;
Tejun Heo77ea8872010-12-08 20:57:37 +0100672 bool bd_write_holder;
Jun'ichi Nomura641dc632006-03-27 01:17:57 -0800673#ifdef CONFIG_SYSFS
Tejun Heo49731ba2011-01-14 18:43:57 +0100674 struct list_head bd_holder_disks;
Jun'ichi Nomura641dc632006-03-27 01:17:57 -0800675#endif
Linus Torvalds1da177e2005-04-16 15:20:36 -0700676 struct block_device * bd_contains;
677 unsigned bd_block_size;
678 struct hd_struct * bd_part;
679 /* number of times partitions within this device have been opened. */
680 unsigned bd_part_count;
681 int bd_invalidated;
682 struct gendisk * bd_disk;
683 struct list_head bd_list;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700684 /*
685 * Private data. You must have bd_claim'ed the block_device
686 * to use this. NOTE: bd_claim allows an owner to claim
687 * the same device multiple times, the owner must take special
688 * care to not mess up bd_private for that case.
689 */
690 unsigned long bd_private;
Takashi Satofcccf502009-01-09 16:40:59 -0800691
692 /* The counter of freeze processes */
693 int bd_fsfreeze_count;
694 /* Mutex for freeze */
695 struct mutex bd_fsfreeze_mutex;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700696};
697
698/*
699 * Radix-tree tags, for tagging dirty and writeback pages within the pagecache
700 * radix trees
701 */
702#define PAGECACHE_TAG_DIRTY 0
703#define PAGECACHE_TAG_WRITEBACK 1
Jan Karaf446daae2010-08-09 17:19:12 -0700704#define PAGECACHE_TAG_TOWRITE 2
Linus Torvalds1da177e2005-04-16 15:20:36 -0700705
706int mapping_tagged(struct address_space *mapping, int tag);
707
708/*
709 * Might pages of this file be mapped into userspace?
710 */
711static inline int mapping_mapped(struct address_space *mapping)
712{
713 return !prio_tree_empty(&mapping->i_mmap) ||
714 !list_empty(&mapping->i_mmap_nonlinear);
715}
716
717/*
718 * Might pages of this file have been modified in userspace?
719 * Note that i_mmap_writable counts all VM_SHARED vmas: do_mmap_pgoff
720 * marks vma as VM_SHARED if it is shared, and the file was opened for
721 * writing i.e. vma may be mprotected writable even if now readonly.
722 */
723static inline int mapping_writably_mapped(struct address_space *mapping)
724{
725 return mapping->i_mmap_writable != 0;
726}
727
728/*
729 * Use sequence counter to get consistent i_size on 32-bit processors.
730 */
731#if BITS_PER_LONG==32 && defined(CONFIG_SMP)
732#include <linux/seqlock.h>
733#define __NEED_I_SIZE_ORDERED
734#define i_size_ordered_init(inode) seqcount_init(&inode->i_size_seqcount)
735#else
736#define i_size_ordered_init(inode) do { } while (0)
737#endif
738
Al Virof19d4a82009-06-08 19:50:45 -0400739struct posix_acl;
740#define ACL_NOT_CACHED ((void *)(-1))
741
Linus Torvalds1da177e2005-04-16 15:20:36 -0700742struct inode {
Nick Piggin44a7d7a2011-01-07 17:49:56 +1100743 /* RCU path lookup touches following: */
744 umode_t i_mode;
745 uid_t i_uid;
746 gid_t i_gid;
747 const struct inode_operations *i_op;
748 struct super_block *i_sb;
749
750 spinlock_t i_lock; /* i_blocks, i_bytes, maybe i_size */
751 unsigned int i_flags;
Linus Torvalds79568f52011-06-20 20:13:49 -0700752 unsigned long i_state;
Linus Torvalds13e12d12011-06-08 15:18:19 -0700753#ifdef CONFIG_SECURITY
754 void *i_security;
755#endif
Nick Piggin44a7d7a2011-01-07 17:49:56 +1100756 struct mutex i_mutex;
757
Linus Torvalds13e12d12011-06-08 15:18:19 -0700758
Nick Piggin44a7d7a2011-01-07 17:49:56 +1100759 unsigned long dirtied_when; /* jiffies of first dirtying */
760
Linus Torvalds1da177e2005-04-16 15:20:36 -0700761 struct hlist_node i_hash;
Nick Piggin7ccf19a2010-10-21 11:49:30 +1100762 struct list_head i_wb_list; /* backing dev IO list */
763 struct list_head i_lru; /* inode LRU list */
Linus Torvalds1da177e2005-04-16 15:20:36 -0700764 struct list_head i_sb_list;
Nick Pigginfa0d7e3d2011-01-07 17:49:49 +1100765 union {
766 struct list_head i_dentry;
767 struct rcu_head i_rcu;
768 };
Linus Torvalds1da177e2005-04-16 15:20:36 -0700769 unsigned long i_ino;
770 atomic_t i_count;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700771 unsigned int i_nlink;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700772 dev_t i_rdev;
Richard Kennedy4e70af52010-02-15 11:16:11 +0000773 unsigned int i_blkbits;
Jean Noel Cordenner7a224222008-01-28 23:58:27 -0500774 u64 i_version;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700775 loff_t i_size;
Eric Dumazet83b7b442006-12-06 20:38:53 -0800776#ifdef __NEED_I_SIZE_ORDERED
777 seqcount_t i_size_seqcount;
778#endif
Linus Torvalds1da177e2005-04-16 15:20:36 -0700779 struct timespec i_atime;
780 struct timespec i_mtime;
781 struct timespec i_ctime;
Takashi Satoa0f62ac2006-03-26 01:37:52 -0800782 blkcnt_t i_blocks;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700783 unsigned short i_bytes;
Christoph Hellwigbd5fe6c2011-06-24 14:29:43 -0400784 atomic_t i_dio_count;
Arjan van de Ven99ac48f2006-03-28 01:56:41 -0800785 const struct file_operations *i_fop; /* former ->i_op->default_file_ops */
Linus Torvalds1da177e2005-04-16 15:20:36 -0700786 struct file_lock *i_flock;
787 struct address_space *i_mapping;
788 struct address_space i_data;
789#ifdef CONFIG_QUOTA
790 struct dquot *i_dquot[MAXQUOTAS];
791#endif
Linus Torvalds1da177e2005-04-16 15:20:36 -0700792 struct list_head i_devices;
Theodore Ts'o4c154162006-09-27 01:50:47 -0700793 union {
794 struct pipe_inode_info *i_pipe;
Theodore Ts'oeaf796e2006-09-27 01:50:48 -0700795 struct block_device *i_bdev;
Theodore Ts'o577c4eb2006-09-27 01:50:49 -0700796 struct cdev *i_cdev;
Theodore Ts'o4c154162006-09-27 01:50:47 -0700797 };
Linus Torvalds1da177e2005-04-16 15:20:36 -0700798
799 __u32 i_generation;
800
Eric Paris3be25f42009-05-21 17:01:26 -0400801#ifdef CONFIG_FSNOTIFY
802 __u32 i_fsnotify_mask; /* all events this inode cares about */
Eric Parise61ce862009-12-17 21:24:24 -0500803 struct hlist_head i_fsnotify_marks;
Robert Love0eeca282005-07-12 17:06:03 -0400804#endif
805
Eric Parisa178d202010-10-25 14:41:59 -0400806#ifdef CONFIG_IMA
Mimi Zohara68a27b2010-11-02 10:10:56 -0400807 atomic_t i_readcount; /* struct files open RO */
Eric Parisa178d202010-10-25 14:41:59 -0400808#endif
Linus Torvalds1da177e2005-04-16 15:20:36 -0700809 atomic_t i_writecount;
Al Virof19d4a82009-06-08 19:50:45 -0400810#ifdef CONFIG_FS_POSIX_ACL
811 struct posix_acl *i_acl;
812 struct posix_acl *i_default_acl;
813#endif
Theodore Ts'o8e18e292006-09-27 01:50:46 -0700814 void *i_private; /* fs or device private pointer */
Linus Torvalds1da177e2005-04-16 15:20:36 -0700815};
816
Al Viro1d3382cb2010-10-23 15:19:20 -0400817static inline int inode_unhashed(struct inode *inode)
818{
819 return hlist_unhashed(&inode->i_hash);
820}
821
Linus Torvalds1da177e2005-04-16 15:20:36 -0700822/*
Ingo Molnarf2eace22006-07-03 00:25:05 -0700823 * inode->i_mutex nesting subclasses for the lock validator:
824 *
825 * 0: the object of the current VFS operation
826 * 1: parent
827 * 2: child/target
828 * 3: quota file
829 *
830 * The locking order between these classes is
Ingo Molnar4df46242006-08-27 01:23:56 -0700831 * parent -> child -> normal -> xattr -> quota
Ingo Molnarf2eace22006-07-03 00:25:05 -0700832 */
833enum inode_i_mutex_lock_class
834{
835 I_MUTEX_NORMAL,
836 I_MUTEX_PARENT,
837 I_MUTEX_CHILD,
Ingo Molnar4df46242006-08-27 01:23:56 -0700838 I_MUTEX_XATTR,
Ingo Molnarf2eace22006-07-03 00:25:05 -0700839 I_MUTEX_QUOTA
840};
841
842/*
Linus Torvalds1da177e2005-04-16 15:20:36 -0700843 * NOTE: in a 32bit arch with a preemptable kernel and
844 * an UP compile the i_size_read/write must be atomic
845 * with respect to the local cpu (unlike with preempt disabled),
846 * but they don't need to be atomic with respect to other cpus like in
847 * true SMP (so they need either to either locally disable irq around
848 * the read or for example on x86 they can be still implemented as a
849 * cmpxchg8b without the need of the lock prefix). For SMP compiles
850 * and 64bit archs it makes no difference if preempt is enabled or not.
851 */
Jan Engelhardt48ed2142006-12-06 20:35:37 -0800852static inline loff_t i_size_read(const struct inode *inode)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700853{
854#if BITS_PER_LONG==32 && defined(CONFIG_SMP)
855 loff_t i_size;
856 unsigned int seq;
857
858 do {
859 seq = read_seqcount_begin(&inode->i_size_seqcount);
860 i_size = inode->i_size;
861 } while (read_seqcount_retry(&inode->i_size_seqcount, seq));
862 return i_size;
863#elif BITS_PER_LONG==32 && defined(CONFIG_PREEMPT)
864 loff_t i_size;
865
866 preempt_disable();
867 i_size = inode->i_size;
868 preempt_enable();
869 return i_size;
870#else
871 return inode->i_size;
872#endif
873}
874
Miklos Szeredi7762f5a2006-10-17 00:10:07 -0700875/*
876 * NOTE: unlike i_size_read(), i_size_write() does need locking around it
877 * (normally i_mutex), otherwise on 32bit/SMP an update of i_size_seqcount
878 * can be lost, resulting in subsequent i_size_read() calls spinning forever.
879 */
Linus Torvalds1da177e2005-04-16 15:20:36 -0700880static inline void i_size_write(struct inode *inode, loff_t i_size)
881{
882#if BITS_PER_LONG==32 && defined(CONFIG_SMP)
883 write_seqcount_begin(&inode->i_size_seqcount);
884 inode->i_size = i_size;
885 write_seqcount_end(&inode->i_size_seqcount);
886#elif BITS_PER_LONG==32 && defined(CONFIG_PREEMPT)
887 preempt_disable();
888 inode->i_size = i_size;
889 preempt_enable();
890#else
891 inode->i_size = i_size;
892#endif
893}
894
Jan Engelhardt48ed2142006-12-06 20:35:37 -0800895static inline unsigned iminor(const struct inode *inode)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700896{
897 return MINOR(inode->i_rdev);
898}
899
Jan Engelhardt48ed2142006-12-06 20:35:37 -0800900static inline unsigned imajor(const struct inode *inode)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700901{
902 return MAJOR(inode->i_rdev);
903}
904
905extern struct block_device *I_BDEV(struct inode *inode);
906
907struct fown_struct {
908 rwlock_t lock; /* protects pid, uid, euid fields */
Eric W. Biederman609d7fa2006-10-02 02:17:15 -0700909 struct pid *pid; /* pid or -pgrp where SIGIO should be sent */
910 enum pid_type pid_type; /* Kind of process group SIGIO should be sent to */
Linus Torvalds1da177e2005-04-16 15:20:36 -0700911 uid_t uid, euid; /* uid/euid of process setting the owner */
Linus Torvalds1da177e2005-04-16 15:20:36 -0700912 int signum; /* posix.1b rt signal to be delivered on IO */
913};
914
915/*
916 * Track a single file's readahead state
917 */
918struct file_ra_state {
Fengguang Wu937085a2007-10-16 01:24:31 -0700919 pgoff_t start; /* where readahead started */
920 unsigned int size; /* # of readahead pages */
921 unsigned int async_size; /* do asynchronous readahead when
Fengguang Wuf9acc8c2007-07-19 01:48:08 -0700922 there are only # of pages ahead */
Fengguang Wu5ce11102007-07-19 01:47:59 -0700923
Fengguang Wu937085a2007-10-16 01:24:31 -0700924 unsigned int ra_pages; /* Maximum readahead window */
Wu Fengguang1ebf26a2009-06-16 15:31:19 -0700925 unsigned int mmap_miss; /* Cache miss stat for mmap accesses */
Fengguang Wuf4e6b492007-10-16 01:24:33 -0700926 loff_t prev_pos; /* Cache last read() position */
Linus Torvalds1da177e2005-04-16 15:20:36 -0700927};
Linus Torvalds1da177e2005-04-16 15:20:36 -0700928
Fengguang Wu5ce11102007-07-19 01:47:59 -0700929/*
Fengguang Wu5ce11102007-07-19 01:47:59 -0700930 * Check if @index falls in the readahead windows.
931 */
932static inline int ra_has_index(struct file_ra_state *ra, pgoff_t index)
933{
Fengguang Wuf9acc8c2007-07-19 01:48:08 -0700934 return (index >= ra->start &&
935 index < ra->start + ra->size);
Fengguang Wu5ce11102007-07-19 01:47:59 -0700936}
937
Dave Hansenad775f52008-02-15 14:38:01 -0800938#define FILE_MNT_WRITE_TAKEN 1
939#define FILE_MNT_WRITE_RELEASED 2
940
Linus Torvalds1da177e2005-04-16 15:20:36 -0700941struct file {
Eric Dumazet2f512012005-10-30 15:02:16 -0800942 /*
943 * fu_list becomes invalid after file_free is called and queued via
944 * fu_rcuhead for RCU freeing
945 */
946 union {
947 struct list_head fu_list;
948 struct rcu_head fu_rcuhead;
949 } f_u;
Josef "Jeff" Sipek0f7fc9e2006-12-08 02:36:35 -0800950 struct path f_path;
951#define f_dentry f_path.dentry
952#define f_vfsmnt f_path.mnt
Arjan van de Ven99ac48f2006-03-28 01:56:41 -0800953 const struct file_operations *f_op;
Jonathan Corbet4a6a4492009-03-27 12:24:31 -0600954 spinlock_t f_lock; /* f_ep_links, f_flags, no IRQ */
Nick Piggin6416ccb2010-08-18 04:37:38 +1000955#ifdef CONFIG_SMP
956 int f_sb_list_cpu;
957#endif
Al Viro516e0cc2008-07-26 00:39:17 -0400958 atomic_long_t f_count;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700959 unsigned int f_flags;
Al Viroaeb5d722008-09-02 15:28:45 -0400960 fmode_t f_mode;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700961 loff_t f_pos;
962 struct fown_struct f_owner;
David Howellsd76b0d92008-11-14 10:39:25 +1100963 const struct cred *f_cred;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700964 struct file_ra_state f_ra;
965
Mathieu Desnoyers2b47c362007-10-16 23:27:21 -0700966 u64 f_version;
Alexey Dobriyan50462062006-09-29 02:00:01 -0700967#ifdef CONFIG_SECURITY
Linus Torvalds1da177e2005-04-16 15:20:36 -0700968 void *f_security;
Alexey Dobriyan50462062006-09-29 02:00:01 -0700969#endif
Linus Torvalds1da177e2005-04-16 15:20:36 -0700970 /* needed for tty driver, and maybe others */
971 void *private_data;
972
973#ifdef CONFIG_EPOLL
974 /* Used by fs/eventpoll.c to link all the hooks to this file */
975 struct list_head f_ep_links;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700976#endif /* #ifdef CONFIG_EPOLL */
977 struct address_space *f_mapping;
Dave Hansenad775f52008-02-15 14:38:01 -0800978#ifdef CONFIG_DEBUG_WRITECOUNT
979 unsigned long f_mnt_write_state;
980#endif
Linus Torvalds1da177e2005-04-16 15:20:36 -0700981};
Linus Torvalds1da177e2005-04-16 15:20:36 -0700982
Aneesh Kumar K.V990d6c22011-01-29 18:43:26 +0530983struct file_handle {
984 __u32 handle_bytes;
985 int handle_type;
986 /* file identifier */
987 unsigned char f_handle[0];
988};
989
Al Viro516e0cc2008-07-26 00:39:17 -0400990#define get_file(x) atomic_long_inc(&(x)->f_count)
Al Virod7065da2010-05-26 15:13:55 -0400991#define fput_atomic(x) atomic_long_add_unless(&(x)->f_count, -1, 1)
Al Viro516e0cc2008-07-26 00:39:17 -0400992#define file_count(x) atomic_long_read(&(x)->f_count)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700993
Dave Hansenad775f52008-02-15 14:38:01 -0800994#ifdef CONFIG_DEBUG_WRITECOUNT
995static inline void file_take_write(struct file *f)
996{
997 WARN_ON(f->f_mnt_write_state != 0);
998 f->f_mnt_write_state = FILE_MNT_WRITE_TAKEN;
999}
1000static inline void file_release_write(struct file *f)
1001{
1002 f->f_mnt_write_state |= FILE_MNT_WRITE_RELEASED;
1003}
1004static inline void file_reset_write(struct file *f)
1005{
1006 f->f_mnt_write_state = 0;
1007}
1008static inline void file_check_state(struct file *f)
1009{
1010 /*
1011 * At this point, either both or neither of these bits
1012 * should be set.
1013 */
1014 WARN_ON(f->f_mnt_write_state == FILE_MNT_WRITE_TAKEN);
1015 WARN_ON(f->f_mnt_write_state == FILE_MNT_WRITE_RELEASED);
1016}
1017static inline int file_check_writeable(struct file *f)
1018{
1019 if (f->f_mnt_write_state == FILE_MNT_WRITE_TAKEN)
1020 return 0;
1021 printk(KERN_WARNING "writeable file with no "
1022 "mnt_want_write()\n");
1023 WARN_ON(1);
1024 return -EINVAL;
1025}
1026#else /* !CONFIG_DEBUG_WRITECOUNT */
1027static inline void file_take_write(struct file *filp) {}
1028static inline void file_release_write(struct file *filp) {}
1029static inline void file_reset_write(struct file *filp) {}
1030static inline void file_check_state(struct file *filp) {}
1031static inline int file_check_writeable(struct file *filp)
1032{
1033 return 0;
1034}
1035#endif /* CONFIG_DEBUG_WRITECOUNT */
1036
Linus Torvalds1da177e2005-04-16 15:20:36 -07001037#define MAX_NON_LFS ((1UL<<31) - 1)
1038
1039/* Page cache limit. The filesystems should put that into their s_maxbytes
1040 limits, otherwise bad things can happen in VM. */
1041#if BITS_PER_LONG==32
1042#define MAX_LFS_FILESIZE (((u64)PAGE_CACHE_SIZE << (BITS_PER_LONG-1))-1)
1043#elif BITS_PER_LONG==64
1044#define MAX_LFS_FILESIZE 0x7fffffffffffffffUL
1045#endif
1046
1047#define FL_POSIX 1
1048#define FL_FLOCK 2
1049#define FL_ACCESS 8 /* not trying to lock, just looking */
Trond Myklebustf475ae92006-06-29 16:38:32 -04001050#define FL_EXISTS 16 /* when unlocking, test for existence */
Linus Torvalds1da177e2005-04-16 15:20:36 -07001051#define FL_LEASE 32 /* lease held on this file */
Miklos Szeredi75e1fcc2006-06-23 02:05:12 -07001052#define FL_CLOSE 64 /* unlock on close */
Linus Torvalds1da177e2005-04-16 15:20:36 -07001053#define FL_SLEEP 128 /* A blocking lock */
1054
1055/*
Miklos Szeredibde74e42008-07-25 01:48:57 -07001056 * Special return value from posix_lock_file() and vfs_lock_file() for
1057 * asynchronous locking.
1058 */
1059#define FILE_LOCK_DEFERRED 1
1060
1061/*
Linus Torvalds1da177e2005-04-16 15:20:36 -07001062 * The POSIX file lock owner is determined by
1063 * the "struct files_struct" in the thread group
1064 * (or NULL for no owner - BSD locks).
1065 *
1066 * Lockd stuffs a "host" pointer into this.
1067 */
1068typedef struct files_struct *fl_owner_t;
1069
1070struct file_lock_operations {
Linus Torvalds1da177e2005-04-16 15:20:36 -07001071 void (*fl_copy_lock)(struct file_lock *, struct file_lock *);
1072 void (*fl_release_private)(struct file_lock *);
1073};
1074
1075struct lock_manager_operations {
1076 int (*fl_compare_owner)(struct file_lock *, struct file_lock *);
1077 void (*fl_notify)(struct file_lock *); /* unblock callback */
Marc Eshel2beb6612006-12-05 23:31:28 -05001078 int (*fl_grant)(struct file_lock *, struct file_lock *, int);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001079 void (*fl_release_private)(struct file_lock *);
1080 void (*fl_break)(struct file_lock *);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001081 int (*fl_change)(struct file_lock **, int);
1082};
1083
J. Bruce Fieldsaf558e32007-09-06 12:34:25 -04001084struct lock_manager {
1085 struct list_head list;
1086};
1087
1088void locks_start_grace(struct lock_manager *);
1089void locks_end_grace(struct lock_manager *);
1090int locks_in_grace(void);
1091
Linus Torvalds1da177e2005-04-16 15:20:36 -07001092/* that will die - we need it for nfs_lock_info */
1093#include <linux/nfs_fs_i.h>
1094
1095struct file_lock {
1096 struct file_lock *fl_next; /* singly linked list for this inode */
1097 struct list_head fl_link; /* doubly linked list of all locks */
1098 struct list_head fl_block; /* circular list of blocked processes */
1099 fl_owner_t fl_owner;
Richard Kennedyafc12462008-07-11 17:20:49 -07001100 unsigned char fl_flags;
1101 unsigned char fl_type;
Linus Torvalds1da177e2005-04-16 15:20:36 -07001102 unsigned int fl_pid;
Vitaliy Gusevab1f1612008-01-17 00:07:08 +00001103 struct pid *fl_nspid;
Linus Torvalds1da177e2005-04-16 15:20:36 -07001104 wait_queue_head_t fl_wait;
1105 struct file *fl_file;
Linus Torvalds1da177e2005-04-16 15:20:36 -07001106 loff_t fl_start;
1107 loff_t fl_end;
1108
1109 struct fasync_struct * fl_fasync; /* for lease break notifications */
1110 unsigned long fl_break_time; /* for nonblocking lease breaks */
1111
Alexey Dobriyan6aed6282009-09-21 17:01:11 -07001112 const struct file_lock_operations *fl_ops; /* Callbacks for filesystems */
Alexey Dobriyan7b021962009-09-21 17:01:12 -07001113 const struct lock_manager_operations *fl_lmops; /* Callbacks for lockmanagers */
Linus Torvalds1da177e2005-04-16 15:20:36 -07001114 union {
1115 struct nfs_lock_info nfs_fl;
Trond Myklebust8d0a8a92005-06-22 17:16:32 +00001116 struct nfs4_lock_info nfs4_fl;
David Howellse8d6c552007-07-15 23:40:12 -07001117 struct {
1118 struct list_head link; /* link in AFS vnode's pending_locks list */
1119 int state; /* state of grant or error if -ve */
1120 } afs;
Linus Torvalds1da177e2005-04-16 15:20:36 -07001121 } fl_u;
1122};
1123
1124/* The following constant reflects the upper bound of the file/locking space */
1125#ifndef OFFSET_MAX
1126#define INT_LIMIT(x) (~((x)1 << (sizeof(x)*8 - 1)))
1127#define OFFSET_MAX INT_LIMIT(loff_t)
1128#define OFFT_OFFSET_MAX INT_LIMIT(off_t)
1129#endif
1130
Linus Torvalds1da177e2005-04-16 15:20:36 -07001131#include <linux/fcntl.h>
1132
Thomas Petazzonibfcd17a2008-08-06 15:12:22 +02001133extern void send_sigio(struct fown_struct *fown, int fd, int band);
1134
Thomas Petazzonibfcd17a2008-08-06 15:12:22 +02001135#ifdef CONFIG_FILE_LOCKING
Linus Torvalds1da177e2005-04-16 15:20:36 -07001136extern int fcntl_getlk(struct file *, struct flock __user *);
Peter Staubachc2936212005-07-27 11:45:09 -07001137extern int fcntl_setlk(unsigned int, struct file *, unsigned int,
1138 struct flock __user *);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001139
1140#if BITS_PER_LONG == 32
1141extern int fcntl_getlk64(struct file *, struct flock64 __user *);
Peter Staubachc2936212005-07-27 11:45:09 -07001142extern int fcntl_setlk64(unsigned int, struct file *, unsigned int,
1143 struct flock64 __user *);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001144#endif
1145
Linus Torvalds1da177e2005-04-16 15:20:36 -07001146extern int fcntl_setlease(unsigned int fd, struct file *filp, long arg);
1147extern int fcntl_getlease(struct file *filp);
1148
1149/* fs/locks.c */
J. Bruce Fields05fa3132010-10-30 17:31:15 -04001150void locks_free_lock(struct file_lock *fl);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001151extern void locks_init_lock(struct file_lock *);
Arnd Bergmannc5b1f0d2010-10-27 15:46:08 +02001152extern struct file_lock * locks_alloc_lock(void);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001153extern void locks_copy_lock(struct file_lock *, struct file_lock *);
J. Bruce Fields1a747ee2008-04-24 10:08:22 -04001154extern void __locks_copy_lock(struct file_lock *, const struct file_lock *);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001155extern void locks_remove_posix(struct file *, fl_owner_t);
1156extern void locks_remove_flock(struct file *);
Felix Blyakhera9e61e22009-03-31 15:12:56 -05001157extern void locks_release_private(struct file_lock *);
J. Bruce Fields6d34ac12007-05-11 16:09:32 -04001158extern void posix_test_lock(struct file *, struct file_lock *);
Marc Eshel150b3932007-01-18 16:15:35 -05001159extern int posix_lock_file(struct file *, struct file_lock *, struct file_lock *);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001160extern int posix_lock_file_wait(struct file *, struct file_lock *);
J. Bruce Fields64a318e2006-01-03 09:55:46 +01001161extern int posix_unblock_lock(struct file *, struct file_lock *);
J. Bruce Fields3ee17ab2007-02-21 00:58:50 -05001162extern int vfs_test_lock(struct file *, struct file_lock *);
Marc Eshel150b3932007-01-18 16:15:35 -05001163extern int vfs_lock_file(struct file *, unsigned int, struct file_lock *, struct file_lock *);
Marc Eshel9b9d2ab2007-01-18 17:52:58 -05001164extern int vfs_cancel_lock(struct file *filp, struct file_lock *fl);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001165extern int flock_lock_file_wait(struct file *filp, struct file_lock *fl);
1166extern int __break_lease(struct inode *inode, unsigned int flags);
1167extern void lease_get_mtime(struct inode *, struct timespec *time);
Christoph Hellwig0af1a452007-07-31 00:39:22 -07001168extern int generic_setlease(struct file *, long, struct file_lock **);
J. Bruce Fieldsa9933ce2007-06-07 17:09:49 -04001169extern int vfs_setlease(struct file *, long, struct file_lock **);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001170extern int lease_modify(struct file_lock **, int);
1171extern int lock_may_read(struct inode *, loff_t start, unsigned long count);
1172extern int lock_may_write(struct inode *, loff_t start, unsigned long count);
Arnd Bergmannb89f4322010-09-18 15:09:31 +02001173extern void lock_flocks(void);
1174extern void unlock_flocks(void);
Thomas Petazzonibfcd17a2008-08-06 15:12:22 +02001175#else /* !CONFIG_FILE_LOCKING */
Steven Whitehousec2aca5e2009-01-20 10:29:45 +00001176static inline int fcntl_getlk(struct file *file, struct flock __user *user)
1177{
1178 return -EINVAL;
1179}
1180
1181static inline int fcntl_setlk(unsigned int fd, struct file *file,
1182 unsigned int cmd, struct flock __user *user)
1183{
1184 return -EACCES;
1185}
1186
Thomas Petazzonibfcd17a2008-08-06 15:12:22 +02001187#if BITS_PER_LONG == 32
Steven Whitehousec2aca5e2009-01-20 10:29:45 +00001188static inline int fcntl_getlk64(struct file *file, struct flock64 __user *user)
1189{
1190 return -EINVAL;
1191}
1192
1193static inline int fcntl_setlk64(unsigned int fd, struct file *file,
1194 unsigned int cmd, struct flock64 __user *user)
1195{
1196 return -EACCES;
1197}
Thomas Petazzonibfcd17a2008-08-06 15:12:22 +02001198#endif
Steven Whitehousec2aca5e2009-01-20 10:29:45 +00001199static inline int fcntl_setlease(unsigned int fd, struct file *filp, long arg)
1200{
1201 return 0;
1202}
1203
1204static inline int fcntl_getlease(struct file *filp)
1205{
1206 return 0;
1207}
1208
1209static inline void locks_init_lock(struct file_lock *fl)
1210{
1211 return;
1212}
1213
1214static inline void __locks_copy_lock(struct file_lock *new, struct file_lock *fl)
1215{
1216 return;
1217}
1218
1219static inline void locks_copy_lock(struct file_lock *new, struct file_lock *fl)
1220{
1221 return;
1222}
1223
1224static inline void locks_remove_posix(struct file *filp, fl_owner_t owner)
1225{
1226 return;
1227}
1228
1229static inline void locks_remove_flock(struct file *filp)
1230{
1231 return;
1232}
1233
1234static inline void posix_test_lock(struct file *filp, struct file_lock *fl)
1235{
1236 return;
1237}
1238
1239static inline int posix_lock_file(struct file *filp, struct file_lock *fl,
1240 struct file_lock *conflock)
1241{
1242 return -ENOLCK;
1243}
1244
1245static inline int posix_lock_file_wait(struct file *filp, struct file_lock *fl)
1246{
1247 return -ENOLCK;
1248}
1249
1250static inline int posix_unblock_lock(struct file *filp,
1251 struct file_lock *waiter)
1252{
1253 return -ENOENT;
1254}
1255
1256static inline int vfs_test_lock(struct file *filp, struct file_lock *fl)
1257{
1258 return 0;
1259}
1260
1261static inline int vfs_lock_file(struct file *filp, unsigned int cmd,
1262 struct file_lock *fl, struct file_lock *conf)
1263{
1264 return -ENOLCK;
1265}
1266
1267static inline int vfs_cancel_lock(struct file *filp, struct file_lock *fl)
1268{
1269 return 0;
1270}
1271
1272static inline int flock_lock_file_wait(struct file *filp,
1273 struct file_lock *request)
1274{
1275 return -ENOLCK;
1276}
1277
1278static inline int __break_lease(struct inode *inode, unsigned int mode)
1279{
1280 return 0;
1281}
1282
1283static inline void lease_get_mtime(struct inode *inode, struct timespec *time)
1284{
1285 return;
1286}
1287
1288static inline int generic_setlease(struct file *filp, long arg,
1289 struct file_lock **flp)
1290{
1291 return -EINVAL;
1292}
1293
1294static inline int vfs_setlease(struct file *filp, long arg,
1295 struct file_lock **lease)
1296{
1297 return -EINVAL;
1298}
1299
1300static inline int lease_modify(struct file_lock **before, int arg)
1301{
1302 return -EINVAL;
1303}
1304
1305static inline int lock_may_read(struct inode *inode, loff_t start,
1306 unsigned long len)
1307{
1308 return 1;
1309}
1310
1311static inline int lock_may_write(struct inode *inode, loff_t start,
1312 unsigned long len)
1313{
1314 return 1;
1315}
1316
Arnd Bergmannb89f4322010-09-18 15:09:31 +02001317static inline void lock_flocks(void)
1318{
1319}
1320
1321static inline void unlock_flocks(void)
1322{
1323}
1324
Thomas Petazzonibfcd17a2008-08-06 15:12:22 +02001325#endif /* !CONFIG_FILE_LOCKING */
1326
Linus Torvalds1da177e2005-04-16 15:20:36 -07001327
1328struct fasync_struct {
Eric Dumazet989a2972010-04-14 09:55:35 +00001329 spinlock_t fa_lock;
1330 int magic;
1331 int fa_fd;
1332 struct fasync_struct *fa_next; /* singly linked list */
1333 struct file *fa_file;
1334 struct rcu_head fa_rcu;
Linus Torvalds1da177e2005-04-16 15:20:36 -07001335};
1336
1337#define FASYNC_MAGIC 0x4601
1338
1339/* SMP safe fasync helpers: */
1340extern int fasync_helper(int, struct file *, int, struct fasync_struct **);
Linus Torvaldsf7347ce2010-10-27 12:38:12 -04001341extern struct fasync_struct *fasync_insert_entry(int, struct file *, struct fasync_struct **, struct fasync_struct *);
1342extern int fasync_remove_entry(struct file *, struct fasync_struct **);
1343extern struct fasync_struct *fasync_alloc(void);
1344extern void fasync_free(struct fasync_struct *);
1345
Linus Torvalds1da177e2005-04-16 15:20:36 -07001346/* can be called from interrupts */
1347extern void kill_fasync(struct fasync_struct **, int, int);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001348
Eric W. Biederman609d7fa2006-10-02 02:17:15 -07001349extern int __f_setown(struct file *filp, struct pid *, enum pid_type, int force);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001350extern int f_setown(struct file *filp, unsigned long arg, int force);
1351extern void f_delown(struct file *filp);
Eric W. Biederman609d7fa2006-10-02 02:17:15 -07001352extern pid_t f_getown(struct file *filp);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001353extern int send_sigurg(struct fown_struct *fown);
1354
1355/*
1356 * Umount options
1357 */
1358
1359#define MNT_FORCE 0x00000001 /* Attempt to forcibily umount */
1360#define MNT_DETACH 0x00000002 /* Just detach from the tree */
1361#define MNT_EXPIRE 0x00000004 /* Mark for expiry */
Miklos Szeredidb1f05b2010-02-10 12:15:53 +01001362#define UMOUNT_NOFOLLOW 0x00000008 /* Don't follow symlink on umount */
1363#define UMOUNT_UNUSED 0x80000000 /* Flag guaranteed to be unused */
Linus Torvalds1da177e2005-04-16 15:20:36 -07001364
1365extern struct list_head super_blocks;
1366extern spinlock_t sb_lock;
1367
Linus Torvalds1da177e2005-04-16 15:20:36 -07001368struct super_block {
1369 struct list_head s_list; /* Keep this first */
1370 dev_t s_dev; /* search index; _not_ kdev_t */
Linus Torvalds1da177e2005-04-16 15:20:36 -07001371 unsigned char s_dirt;
Richard Kennedy270ba5f2010-01-26 14:12:43 +00001372 unsigned char s_blocksize_bits;
1373 unsigned long s_blocksize;
Jeff Layton42cb56a2009-09-18 13:05:53 -07001374 loff_t s_maxbytes; /* Max file size */
Linus Torvalds1da177e2005-04-16 15:20:36 -07001375 struct file_system_type *s_type;
Josef 'Jeff' Sipekee9b6d62007-02-12 00:55:41 -08001376 const struct super_operations *s_op;
Alexey Dobriyan61e225d2009-09-21 17:01:08 -07001377 const struct dquot_operations *dq_op;
Alexey Dobriyan0d54b212009-09-21 17:01:09 -07001378 const struct quotactl_ops *s_qcop;
Christoph Hellwig39655162007-10-21 16:42:17 -07001379 const struct export_operations *s_export_op;
Linus Torvalds1da177e2005-04-16 15:20:36 -07001380 unsigned long s_flags;
1381 unsigned long s_magic;
1382 struct dentry *s_root;
1383 struct rw_semaphore s_umount;
Ingo Molnar7892f2f42006-01-09 15:59:25 -08001384 struct mutex s_lock;
Linus Torvalds1da177e2005-04-16 15:20:36 -07001385 int s_count;
Linus Torvalds1da177e2005-04-16 15:20:36 -07001386 atomic_t s_active;
Alexey Dobriyan50462062006-09-29 02:00:01 -07001387#ifdef CONFIG_SECURITY
Linus Torvalds1da177e2005-04-16 15:20:36 -07001388 void *s_security;
Alexey Dobriyan50462062006-09-29 02:00:01 -07001389#endif
Stephen Hemmingerbb435452010-05-13 17:53:14 -07001390 const struct xattr_handler **s_xattr;
Linus Torvalds1da177e2005-04-16 15:20:36 -07001391
1392 struct list_head s_inodes; /* all inodes */
Nick Pigginceb5bdc2011-01-07 17:50:05 +11001393 struct hlist_bl_head s_anon; /* anonymous dentries for (nfs) exporting */
Nick Piggin6416ccb2010-08-18 04:37:38 +10001394#ifdef CONFIG_SMP
1395 struct list_head __percpu *s_files;
1396#else
Linus Torvalds1da177e2005-04-16 15:20:36 -07001397 struct list_head s_files;
Nick Piggin6416ccb2010-08-18 04:37:38 +10001398#endif
Nick Pigginb5c84bf2011-01-07 17:49:38 +11001399 /* s_dentry_lru, s_nr_dentry_unused protected by dcache.c lru locks */
Kentaro Makitada3bbdd2008-07-23 21:27:13 -07001400 struct list_head s_dentry_lru; /* unused dentry lru */
1401 int s_nr_dentry_unused; /* # of dentry on lru */
Linus Torvalds1da177e2005-04-16 15:20:36 -07001402
Dave Chinner09cc9fc2011-07-08 14:14:40 +10001403 /* s_inode_lru_lock protects s_inode_lru and s_nr_inodes_unused */
1404 spinlock_t s_inode_lru_lock ____cacheline_aligned_in_smp;
Dave Chinner98b745c2011-07-08 14:14:39 +10001405 struct list_head s_inode_lru; /* unused inode lru */
1406 int s_nr_inodes_unused; /* # of inodes on lru */
1407
Linus Torvalds1da177e2005-04-16 15:20:36 -07001408 struct block_device *s_bdev;
Jens Axboe32a88aa2009-09-16 15:02:33 +02001409 struct backing_dev_info *s_bdi;
David Howellsacaebfd2007-05-10 22:51:50 -07001410 struct mtd_info *s_mtd;
Linus Torvalds1da177e2005-04-16 15:20:36 -07001411 struct list_head s_instances;
1412 struct quota_info s_dquot; /* Diskquota specific options */
1413
1414 int s_frozen;
1415 wait_queue_head_t s_wait_unfrozen;
1416
1417 char s_id[32]; /* Informational name */
Linus Torvalds0bba0162011-04-12 15:21:04 -07001418 u8 s_uuid[16]; /* UUID */
Linus Torvalds1da177e2005-04-16 15:20:36 -07001419
1420 void *s_fs_info; /* Filesystem private info */
Al Viro30c40d22008-02-22 19:50:45 -05001421 fmode_t s_mode;
Linus Torvalds1da177e2005-04-16 15:20:36 -07001422
Richard Kennedy270ba5f2010-01-26 14:12:43 +00001423 /* Granularity of c/m/atime in ns.
1424 Cannot be worse than a second */
1425 u32 s_time_gran;
1426
Linus Torvalds1da177e2005-04-16 15:20:36 -07001427 /*
1428 * The next field is for VFS *only*. No filesystems have any business
1429 * even looking at it. You had been warned.
1430 */
Arjan van de Vena11f3a02006-03-23 03:00:33 -08001431 struct mutex s_vfs_rename_mutex; /* Kludge */
Linus Torvalds1da177e2005-04-16 15:20:36 -07001432
Miklos Szeredi79c0b2d2007-05-08 00:25:43 -07001433 /*
1434 * Filesystem subtype. If non-empty the filesystem type field
1435 * in /proc/mounts will be "type.subtype"
1436 */
1437 char *s_subtype;
Miklos Szeredib3b304a2008-02-08 04:21:35 -08001438
1439 /*
1440 * Saved mount options for lazy filesystems using
1441 * generic_show_options()
1442 */
Arnd Bergmann4d2deb42010-02-24 20:01:56 +01001443 char __rcu *s_options;
Al Viroc8aebb02010-12-18 10:22:30 -05001444 const struct dentry_operations *s_d_op; /* default d_op for dentries */
Dan Magenheimer9fdfdcf2011-05-26 10:01:19 -06001445
1446 /*
1447 * Saved pool identifier for cleancache (-1 means none)
1448 */
1449 int cleancache_poolid;
Dave Chinnerb0d40c92011-07-08 14:14:42 +10001450
1451 struct shrinker s_shrink; /* per-sb shrinker handle */
Linus Torvalds1da177e2005-04-16 15:20:36 -07001452};
1453
Dave Chinnerb0d40c92011-07-08 14:14:42 +10001454/* superblock cache pruning functions */
1455extern void prune_icache_sb(struct super_block *sb, int nr_to_scan);
1456extern void prune_dcache_sb(struct super_block *sb, int nr_to_scan);
1457
Linus Torvalds1da177e2005-04-16 15:20:36 -07001458extern struct timespec current_fs_time(struct super_block *sb);
1459
1460/*
1461 * Snapshotting support.
1462 */
1463enum {
1464 SB_UNFROZEN = 0,
1465 SB_FREEZE_WRITE = 1,
1466 SB_FREEZE_TRANS = 2,
1467};
1468
1469#define vfs_check_frozen(sb, level) \
1470 wait_event((sb)->s_wait_unfrozen, ((sb)->s_frozen < (level)))
1471
Al Viro914e2632006-10-18 13:55:46 -04001472#define get_fs_excl() atomic_inc(&current->fs_excl)
1473#define put_fs_excl() atomic_dec(&current->fs_excl)
1474#define has_fs_excl() atomic_read(&current->fs_excl)
Jens Axboe22e2c502005-06-27 10:55:12 +02001475
Serge E. Hallyne795b712011-03-23 16:43:25 -07001476/*
1477 * until VFS tracks user namespaces for inodes, just make all files
1478 * belong to init_user_ns
1479 */
1480extern struct user_namespace init_user_ns;
1481#define inode_userns(inode) (&init_user_ns)
Serge E. Hallyn2e149672011-03-23 16:43:26 -07001482extern bool inode_owner_or_capable(const struct inode *inode);
Satyam Sharma3bd858a2007-07-17 15:00:08 +05301483
Al Viro914e2632006-10-18 13:55:46 -04001484/* not quite ready to be deprecated, but... */
1485extern void lock_super(struct super_block *);
1486extern void unlock_super(struct super_block *);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001487
1488/*
1489 * VFS helper functions..
1490 */
1491extern int vfs_create(struct inode *, struct dentry *, int, struct nameidata *);
1492extern int vfs_mkdir(struct inode *, struct dentry *, int);
1493extern int vfs_mknod(struct inode *, struct dentry *, int, dev_t);
Miklos Szeredidb2e7472008-06-24 16:50:16 +02001494extern int vfs_symlink(struct inode *, struct dentry *, const char *);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001495extern int vfs_link(struct dentry *, struct inode *, struct dentry *);
1496extern int vfs_rmdir(struct inode *, struct dentry *);
1497extern int vfs_unlink(struct inode *, struct dentry *);
1498extern int vfs_rename(struct inode *, struct dentry *, struct inode *, struct dentry *);
1499
1500/*
1501 * VFS dentry helper functions.
1502 */
1503extern void dentry_unhash(struct dentry *dentry);
1504
1505/*
Christoph Hellwig8c744fb2005-11-08 21:35:04 -08001506 * VFS file helper functions.
1507 */
Dmitry Monakhova1bd1202010-03-04 17:29:14 +03001508extern void inode_init_owner(struct inode *inode, const struct inode *dir,
1509 mode_t mode);
Christoph Hellwig8c744fb2005-11-08 21:35:04 -08001510/*
Mark Fashehc4b929b2008-10-08 19:44:18 -04001511 * VFS FS_IOC_FIEMAP helper definitions.
1512 */
1513struct fiemap_extent_info {
1514 unsigned int fi_flags; /* Flags as passed from user */
1515 unsigned int fi_extents_mapped; /* Number of mapped extents */
1516 unsigned int fi_extents_max; /* Size of fiemap_extent array */
Namhyung Kimecf56322011-01-16 23:28:17 +09001517 struct fiemap_extent __user *fi_extents_start; /* Start of
1518 fiemap_extent array */
Mark Fashehc4b929b2008-10-08 19:44:18 -04001519};
1520int fiemap_fill_next_extent(struct fiemap_extent_info *info, u64 logical,
1521 u64 phys, u64 len, u32 flags);
1522int fiemap_check_flags(struct fiemap_extent_info *fieinfo, u32 fs_flags);
1523
1524/*
Linus Torvalds1da177e2005-04-16 15:20:36 -07001525 * File types
1526 *
1527 * NOTE! These match bits 12..15 of stat.st_mode
1528 * (ie "(i_mode >> 12) & 15").
1529 */
1530#define DT_UNKNOWN 0
1531#define DT_FIFO 1
1532#define DT_CHR 2
1533#define DT_DIR 4
1534#define DT_BLK 6
1535#define DT_REG 8
1536#define DT_LNK 10
1537#define DT_SOCK 12
1538#define DT_WHT 14
1539
Linus Torvalds1da177e2005-04-16 15:20:36 -07001540/*
1541 * This is the "filldir" function type, used by readdir() to let
1542 * the kernel specify what kind of dirent layout it wants to have.
1543 * This allows the kernel to read directories into kernel space or
1544 * to have different dirent layouts depending on the binary type.
1545 */
David Howellsafefdbb2006-10-03 01:13:46 -07001546typedef int (*filldir_t)(void *, const char *, int, loff_t, u64, unsigned);
Al Viro08f85852007-10-08 13:26:20 -04001547struct block_device_operations;
Linus Torvalds1da177e2005-04-16 15:20:36 -07001548
Linus Torvalds1da177e2005-04-16 15:20:36 -07001549/* These macros are for out of kernel modules to test that
1550 * the kernel supports the unlocked_ioctl and compat_ioctl
1551 * fields in struct file_operations. */
1552#define HAVE_COMPAT_IOCTL 1
1553#define HAVE_UNLOCKED_IOCTL 1
1554
1555/*
1556 * NOTE:
Arnd Bergmannb19dd422010-07-04 00:15:10 +02001557 * all file operations except setlease can be called without
1558 * the big kernel lock held in all filesystems.
Linus Torvalds1da177e2005-04-16 15:20:36 -07001559 */
1560struct file_operations {
1561 struct module *owner;
1562 loff_t (*llseek) (struct file *, loff_t, int);
1563 ssize_t (*read) (struct file *, char __user *, size_t, loff_t *);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001564 ssize_t (*write) (struct file *, const char __user *, size_t, loff_t *);
Badari Pulavarty027445c2006-09-30 23:28:46 -07001565 ssize_t (*aio_read) (struct kiocb *, const struct iovec *, unsigned long, loff_t);
1566 ssize_t (*aio_write) (struct kiocb *, const struct iovec *, unsigned long, loff_t);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001567 int (*readdir) (struct file *, void *, filldir_t);
1568 unsigned int (*poll) (struct file *, struct poll_table_struct *);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001569 long (*unlocked_ioctl) (struct file *, unsigned int, unsigned long);
1570 long (*compat_ioctl) (struct file *, unsigned int, unsigned long);
1571 int (*mmap) (struct file *, struct vm_area_struct *);
1572 int (*open) (struct inode *, struct file *);
Miklos Szeredi75e1fcc2006-06-23 02:05:12 -07001573 int (*flush) (struct file *, fl_owner_t id);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001574 int (*release) (struct inode *, struct file *);
Christoph Hellwig7ea80852010-05-26 17:53:25 +02001575 int (*fsync) (struct file *, int datasync);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001576 int (*aio_fsync) (struct kiocb *, int datasync);
1577 int (*fasync) (int, struct file *, int);
1578 int (*lock) (struct file *, int, struct file_lock *);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001579 ssize_t (*sendpage) (struct file *, struct page *, int, size_t, loff_t *, int);
1580 unsigned long (*get_unmapped_area)(struct file *, unsigned long, unsigned long, unsigned long, unsigned long);
1581 int (*check_flags)(int);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001582 int (*flock) (struct file *, int, struct file_lock *);
Jens Axboecbb7e572006-04-11 14:57:50 +02001583 ssize_t (*splice_write)(struct pipe_inode_info *, struct file *, loff_t *, size_t, unsigned int);
1584 ssize_t (*splice_read)(struct file *, loff_t *, struct pipe_inode_info *, size_t, unsigned int);
J. Bruce Fieldsf9ffed22006-11-14 15:51:40 -05001585 int (*setlease)(struct file *, long, struct file_lock **);
Christoph Hellwig2fe17c12011-01-14 13:07:43 +01001586 long (*fallocate)(struct file *file, int mode, loff_t offset,
1587 loff_t len);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001588};
1589
1590struct inode_operations {
Linus Torvalds1da177e2005-04-16 15:20:36 -07001591 struct dentry * (*lookup) (struct inode *,struct dentry *, struct nameidata *);
Nick Piggin44a7d7a2011-01-07 17:49:56 +11001592 void * (*follow_link) (struct dentry *, struct nameidata *);
Al Viro10556cb2011-06-20 19:28:19 -04001593 int (*permission) (struct inode *, int);
Al Viro7e401452011-06-20 19:12:17 -04001594 int (*check_acl)(struct inode *, int);
Nick Piggin44a7d7a2011-01-07 17:49:56 +11001595
1596 int (*readlink) (struct dentry *, char __user *,int);
1597 void (*put_link) (struct dentry *, struct nameidata *, void *);
1598
1599 int (*create) (struct inode *,struct dentry *,int, struct nameidata *);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001600 int (*link) (struct dentry *,struct inode *,struct dentry *);
1601 int (*unlink) (struct inode *,struct dentry *);
1602 int (*symlink) (struct inode *,struct dentry *,const char *);
1603 int (*mkdir) (struct inode *,struct dentry *,int);
1604 int (*rmdir) (struct inode *,struct dentry *);
1605 int (*mknod) (struct inode *,struct dentry *,int,dev_t);
1606 int (*rename) (struct inode *, struct dentry *,
1607 struct inode *, struct dentry *);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001608 void (*truncate) (struct inode *);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001609 int (*setattr) (struct dentry *, struct iattr *);
1610 int (*getattr) (struct vfsmount *mnt, struct dentry *, struct kstat *);
1611 int (*setxattr) (struct dentry *, const char *,const void *,size_t,int);
1612 ssize_t (*getxattr) (struct dentry *, const char *, void *, size_t);
1613 ssize_t (*listxattr) (struct dentry *, char *, size_t);
1614 int (*removexattr) (struct dentry *, const char *);
Badari Pulavartyf6b3ec22006-01-06 00:10:38 -08001615 void (*truncate_range)(struct inode *, loff_t, loff_t);
Mark Fashehc4b929b2008-10-08 19:44:18 -04001616 int (*fiemap)(struct inode *, struct fiemap_extent_info *, u64 start,
1617 u64 len);
Nick Piggin44a7d7a2011-01-07 17:49:56 +11001618} ____cacheline_aligned;
Linus Torvalds1da177e2005-04-16 15:20:36 -07001619
1620struct seq_file;
1621
Badari Pulavartyeed4e512006-09-30 23:28:49 -07001622ssize_t rw_copy_check_uvector(int type, const struct iovec __user * uvector,
1623 unsigned long nr_segs, unsigned long fast_segs,
1624 struct iovec *fast_pointer,
1625 struct iovec **ret_pointer);
1626
Linus Torvalds1da177e2005-04-16 15:20:36 -07001627extern ssize_t vfs_read(struct file *, char __user *, size_t, loff_t *);
1628extern ssize_t vfs_write(struct file *, const char __user *, size_t, loff_t *);
1629extern ssize_t vfs_readv(struct file *, const struct iovec __user *,
1630 unsigned long, loff_t *);
1631extern ssize_t vfs_writev(struct file *, const struct iovec __user *,
1632 unsigned long, loff_t *);
1633
Linus Torvalds1da177e2005-04-16 15:20:36 -07001634struct super_operations {
1635 struct inode *(*alloc_inode)(struct super_block *sb);
1636 void (*destroy_inode)(struct inode *);
1637
Christoph Hellwigaa385722011-05-27 06:53:02 -04001638 void (*dirty_inode) (struct inode *, int flags);
Christoph Hellwiga9185b42010-03-05 09:21:37 +01001639 int (*write_inode) (struct inode *, struct writeback_control *wbc);
Al Viro45321ac2010-06-07 13:43:19 -04001640 int (*drop_inode) (struct inode *);
Al Virobe7ce412010-06-04 19:40:39 -04001641 void (*evict_inode) (struct inode *);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001642 void (*put_super) (struct super_block *);
1643 void (*write_super) (struct super_block *);
1644 int (*sync_fs)(struct super_block *sb, int wait);
Takashi Satoc4be0c12009-01-09 16:40:58 -08001645 int (*freeze_fs) (struct super_block *);
1646 int (*unfreeze_fs) (struct super_block *);
David Howells726c3342006-06-23 02:02:58 -07001647 int (*statfs) (struct dentry *, struct kstatfs *);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001648 int (*remount_fs) (struct super_block *, int *, char *);
Al Viro42faad92008-04-24 07:21:56 -04001649 void (*umount_begin) (struct super_block *);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001650
1651 int (*show_options)(struct seq_file *, struct vfsmount *);
Al Viroc7f404b2011-03-16 06:59:40 -04001652 int (*show_devname)(struct seq_file *, struct vfsmount *);
1653 int (*show_path)(struct seq_file *, struct vfsmount *);
Chuck Leverb4629fe2006-03-20 13:44:12 -05001654 int (*show_stats)(struct seq_file *, struct vfsmount *);
Alexey Dobriyan0e51a722006-09-29 01:59:56 -07001655#ifdef CONFIG_QUOTA
Linus Torvalds1da177e2005-04-16 15:20:36 -07001656 ssize_t (*quota_read)(struct super_block *, int, char *, size_t, loff_t);
1657 ssize_t (*quota_write)(struct super_block *, int, const char *, size_t, loff_t);
Alexey Dobriyan0e51a722006-09-29 01:59:56 -07001658#endif
Theodore Ts'o87d8fe12009-01-03 09:47:09 -05001659 int (*bdev_try_to_free_page)(struct super_block*, struct page*, gfp_t);
Dave Chinner0e1fdaf2011-07-08 14:14:44 +10001660 int (*nr_cached_objects)(struct super_block *);
1661 void (*free_cached_objects)(struct super_block *, int);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001662};
1663
Joern Engel1c0eeaf2007-10-16 23:30:44 -07001664/*
Dave Chinner250df6e2011-03-22 22:23:36 +11001665 * Inode state bits. Protected by inode->i_lock
Joern Engel1c0eeaf2007-10-16 23:30:44 -07001666 *
1667 * Three bits determine the dirty state of the inode, I_DIRTY_SYNC,
1668 * I_DIRTY_DATASYNC and I_DIRTY_PAGES.
1669 *
1670 * Four bits define the lifetime of an inode. Initially, inodes are I_NEW,
1671 * until that flag is cleared. I_WILL_FREE, I_FREEING and I_CLEAR are set at
1672 * various stages of removing an inode.
1673 *
Christoph Hellwigeaff8072009-12-17 14:25:01 +01001674 * Two bits are used for locking and completion notification, I_NEW and I_SYNC.
Joern Engel1c0eeaf2007-10-16 23:30:44 -07001675 *
Joern Engele7ca2d42008-02-06 01:36:59 -08001676 * I_DIRTY_SYNC Inode is dirty, but doesn't have to be written on
1677 * fdatasync(). i_atime is the usual cause.
Jan Kara000cb482008-02-14 19:31:32 -08001678 * I_DIRTY_DATASYNC Data-related inode changes pending. We keep track of
1679 * these changes separately from I_DIRTY_SYNC so that we
1680 * don't have to write inode on fdatasync() when only
1681 * mtime has changed in it.
Joern Engel1c0eeaf2007-10-16 23:30:44 -07001682 * I_DIRTY_PAGES Inode has dirty pages. Inode itself may be clean.
Christoph Hellwigeaff8072009-12-17 14:25:01 +01001683 * I_NEW Serves as both a mutex and completion notification.
1684 * New inodes set I_NEW. If two processes both create
1685 * the same inode, one of them will release its inode and
1686 * wait for I_NEW to be released before returning.
1687 * Inodes in I_WILL_FREE, I_FREEING or I_CLEAR state can
1688 * also cause waiting on I_NEW, without I_NEW actually
1689 * being set. find_inode() uses this to prevent returning
1690 * nearly-dead inodes.
Joern Engel1c0eeaf2007-10-16 23:30:44 -07001691 * I_WILL_FREE Must be set when calling write_inode_now() if i_count
1692 * is zero. I_FREEING must be set when I_WILL_FREE is
1693 * cleared.
1694 * I_FREEING Set when inode is about to be freed but still has dirty
1695 * pages or buffers attached or the inode itself is still
1696 * dirty.
Al Virob57922d2010-06-07 14:34:48 -04001697 * I_CLEAR Added by end_writeback(). In this state the inode is clean
Al Viroa4ffdde2010-06-02 17:38:30 -04001698 * and can be destroyed. Inode keeps I_FREEING.
Joern Engel1c0eeaf2007-10-16 23:30:44 -07001699 *
1700 * Inodes that are I_WILL_FREE, I_FREEING or I_CLEAR are
1701 * prohibited for many purposes. iget() must wait for
1702 * the inode to be completely released, then create it
1703 * anew. Other functions will just ignore such inodes,
Christoph Hellwigeaff8072009-12-17 14:25:01 +01001704 * if appropriate. I_NEW is used for waiting.
Joern Engel1c0eeaf2007-10-16 23:30:44 -07001705 *
Christoph Hellwigeaff8072009-12-17 14:25:01 +01001706 * I_SYNC Synchonized write of dirty inode data. The bits is
1707 * set during data writeback, and cleared with a wakeup
1708 * on the bit address once it is done.
Joern Engel1c0eeaf2007-10-16 23:30:44 -07001709 *
Christoph Hellwigbd5fe6c2011-06-24 14:29:43 -04001710 * I_REFERENCED Marks the inode as recently references on the LRU list.
1711 *
1712 * I_DIO_WAKEUP Never set. Only used as a key for wait_on_bit().
1713 *
Joern Engel1c0eeaf2007-10-16 23:30:44 -07001714 * Q: What is the difference between I_WILL_FREE and I_FREEING?
Joern Engel1c0eeaf2007-10-16 23:30:44 -07001715 */
Nick Piggin9e38d862010-10-23 06:55:17 -04001716#define I_DIRTY_SYNC (1 << 0)
1717#define I_DIRTY_DATASYNC (1 << 1)
1718#define I_DIRTY_PAGES (1 << 2)
Christoph Hellwigeaff8072009-12-17 14:25:01 +01001719#define __I_NEW 3
1720#define I_NEW (1 << __I_NEW)
Nick Piggin9e38d862010-10-23 06:55:17 -04001721#define I_WILL_FREE (1 << 4)
1722#define I_FREEING (1 << 5)
1723#define I_CLEAR (1 << 6)
Christoph Hellwigeaff8072009-12-17 14:25:01 +01001724#define __I_SYNC 7
Joern Engel1c0eeaf2007-10-16 23:30:44 -07001725#define I_SYNC (1 << __I_SYNC)
Nick Piggin9e38d862010-10-23 06:55:17 -04001726#define I_REFERENCED (1 << 8)
Christoph Hellwigbd5fe6c2011-06-24 14:29:43 -04001727#define __I_DIO_WAKEUP 9
1728#define I_DIO_WAKEUP (1 << I_DIO_WAKEUP)
Linus Torvalds1da177e2005-04-16 15:20:36 -07001729
1730#define I_DIRTY (I_DIRTY_SYNC | I_DIRTY_DATASYNC | I_DIRTY_PAGES)
1731
1732extern void __mark_inode_dirty(struct inode *, int);
1733static inline void mark_inode_dirty(struct inode *inode)
1734{
1735 __mark_inode_dirty(inode, I_DIRTY);
1736}
1737
1738static inline void mark_inode_dirty_sync(struct inode *inode)
1739{
1740 __mark_inode_dirty(inode, I_DIRTY_SYNC);
1741}
1742
Dave Hansen71c42152007-06-23 17:16:42 -07001743/**
1744 * inc_nlink - directly increment an inode's link count
1745 * @inode: inode
1746 *
1747 * This is a low-level filesystem helper to replace any
1748 * direct filesystem manipulation of i_nlink. Currently,
1749 * it is only here for parity with dec_nlink().
1750 */
Dave Hansend8c76e62006-09-30 23:29:04 -07001751static inline void inc_nlink(struct inode *inode)
Alexey Dobriyan66d21732006-03-23 03:00:51 -08001752{
1753 inode->i_nlink++;
Dave Hansend8c76e62006-09-30 23:29:04 -07001754}
1755
1756static inline void inode_inc_link_count(struct inode *inode)
1757{
1758 inc_nlink(inode);
Alexey Dobriyan66d21732006-03-23 03:00:51 -08001759 mark_inode_dirty(inode);
1760}
1761
Dave Hansen71c42152007-06-23 17:16:42 -07001762/**
1763 * drop_nlink - directly drop an inode's link count
1764 * @inode: inode
1765 *
1766 * This is a low-level filesystem helper to replace any
1767 * direct filesystem manipulation of i_nlink. In cases
1768 * where we are attempting to track writes to the
1769 * filesystem, a decrement to zero means an imminent
1770 * write when the file is truncated and actually unlinked
1771 * on the filesystem.
1772 */
Dave Hansen9a53c3a2006-09-30 23:29:03 -07001773static inline void drop_nlink(struct inode *inode)
Alexey Dobriyan66d21732006-03-23 03:00:51 -08001774{
1775 inode->i_nlink--;
Dave Hansen9a53c3a2006-09-30 23:29:03 -07001776}
1777
Dave Hansen71c42152007-06-23 17:16:42 -07001778/**
1779 * clear_nlink - directly zero an inode's link count
1780 * @inode: inode
1781 *
1782 * This is a low-level filesystem helper to replace any
1783 * direct filesystem manipulation of i_nlink. See
1784 * drop_nlink() for why we care about i_nlink hitting zero.
1785 */
Dave Hansence71ec32006-09-30 23:29:06 -07001786static inline void clear_nlink(struct inode *inode)
1787{
1788 inode->i_nlink = 0;
1789}
1790
Dave Hansen9a53c3a2006-09-30 23:29:03 -07001791static inline void inode_dec_link_count(struct inode *inode)
1792{
1793 drop_nlink(inode);
Alexey Dobriyan66d21732006-03-23 03:00:51 -08001794 mark_inode_dirty(inode);
1795}
1796
Jean Noel Cordenner25ec56b2008-01-28 23:58:27 -05001797/**
1798 * inode_inc_iversion - increments i_version
1799 * @inode: inode that need to be updated
1800 *
1801 * Every time the inode is modified, the i_version field will be incremented.
1802 * The filesystem has to be mounted with i_version flag
1803 */
1804
1805static inline void inode_inc_iversion(struct inode *inode)
1806{
1807 spin_lock(&inode->i_lock);
1808 inode->i_version++;
1809 spin_unlock(&inode->i_lock);
1810}
1811
Christoph Hellwig869243a2006-01-09 20:52:03 -08001812extern void touch_atime(struct vfsmount *mnt, struct dentry *dentry);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001813static inline void file_accessed(struct file *file)
1814{
1815 if (!(file->f_flags & O_NOATIME))
Josef "Jeff" Sipek0f7fc9e2006-12-08 02:36:35 -08001816 touch_atime(file->f_path.mnt, file->f_path.dentry);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001817}
1818
1819int sync_inode(struct inode *inode, struct writeback_control *wbc);
Christoph Hellwigc37650162010-10-06 10:48:20 +02001820int sync_inode_metadata(struct inode *inode, int wait);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001821
Linus Torvalds1da177e2005-04-16 15:20:36 -07001822struct file_system_type {
1823 const char *name;
1824 int fs_flags;
Al Viroc96e41e2010-07-25 00:17:56 +04001825 struct dentry *(*mount) (struct file_system_type *, int,
1826 const char *, void *);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001827 void (*kill_sb) (struct super_block *);
1828 struct module *owner;
1829 struct file_system_type * next;
1830 struct list_head fs_supers;
Peter Zijlstrad475fd42007-10-15 14:51:31 +02001831
Ingo Molnarcf516242006-07-03 00:25:27 -07001832 struct lock_class_key s_lock_key;
Arjan van de Ven897c6ff2006-07-03 00:25:28 -07001833 struct lock_class_key s_umount_key;
Roland Dreier51ee0492010-04-27 14:23:57 -07001834 struct lock_class_key s_vfs_rename_key;
Peter Zijlstrad475fd42007-10-15 14:51:31 +02001835
1836 struct lock_class_key i_lock_key;
1837 struct lock_class_key i_mutex_key;
Peter Zijlstra14358e62007-10-14 01:38:33 +02001838 struct lock_class_key i_mutex_dir_key;
Linus Torvalds1da177e2005-04-16 15:20:36 -07001839};
1840
Al Viroceefda62010-07-26 13:16:50 +04001841extern struct dentry *mount_ns(struct file_system_type *fs_type, int flags,
1842 void *data, int (*fill_super)(struct super_block *, void *, int));
Al Viro152a0832010-07-25 00:46:55 +04001843extern struct dentry *mount_bdev(struct file_system_type *fs_type,
1844 int flags, const char *dev_name, void *data,
1845 int (*fill_super)(struct super_block *, void *, int));
Al Virofc14f2f2010-07-25 01:48:30 +04001846extern struct dentry *mount_single(struct file_system_type *fs_type,
1847 int flags, void *data,
1848 int (*fill_super)(struct super_block *, void *, int));
Al Viro3c26ff62010-07-25 11:46:36 +04001849extern struct dentry *mount_nodev(struct file_system_type *fs_type,
1850 int flags, void *data,
1851 int (*fill_super)(struct super_block *, void *, int));
Linus Torvalds1da177e2005-04-16 15:20:36 -07001852void generic_shutdown_super(struct super_block *sb);
1853void kill_block_super(struct super_block *sb);
1854void kill_anon_super(struct super_block *sb);
1855void kill_litter_super(struct super_block *sb);
1856void deactivate_super(struct super_block *sb);
Al Viro74dbbdd2009-05-06 01:07:50 -04001857void deactivate_locked_super(struct super_block *sb);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001858int set_anon_super(struct super_block *s, void *data);
Al Viro0ee5dc62011-07-07 15:44:25 -04001859int get_anon_bdev(dev_t *);
1860void free_anon_bdev(dev_t);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001861struct super_block *sget(struct file_system_type *type,
1862 int (*test)(struct super_block *,void *),
1863 int (*set)(struct super_block *,void *),
1864 void *data);
Al Viro51139ad2010-07-25 23:47:46 +04001865extern struct dentry *mount_pseudo(struct file_system_type *, char *,
Al Viroc74a1cb2011-01-12 16:59:34 -05001866 const struct super_operations *ops,
1867 const struct dentry_operations *dops,
1868 unsigned long);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001869
Artem Bityutskiy140236b2010-06-10 13:56:33 +03001870static inline void sb_mark_dirty(struct super_block *sb)
1871{
1872 sb->s_dirt = 1;
1873}
1874static inline void sb_mark_clean(struct super_block *sb)
1875{
1876 sb->s_dirt = 0;
1877}
1878static inline int sb_is_dirty(struct super_block *sb)
1879{
1880 return sb->s_dirt;
1881}
1882
Linus Torvalds1da177e2005-04-16 15:20:36 -07001883/* Alas, no aliases. Too much hassle with bringing module.h everywhere */
1884#define fops_get(fops) \
1885 (((fops) && try_module_get((fops)->owner) ? (fops) : NULL))
1886#define fops_put(fops) \
1887 do { if (fops) module_put((fops)->owner); } while(0)
1888
1889extern int register_filesystem(struct file_system_type *);
1890extern int unregister_filesystem(struct file_system_type *);
Pavel Emelyanov8bf97252007-10-18 23:40:02 -07001891extern struct vfsmount *kern_mount_data(struct file_system_type *, void *data);
1892#define kern_mount(type) kern_mount_data(type, NULL)
Linus Torvalds1da177e2005-04-16 15:20:36 -07001893extern int may_umount_tree(struct vfsmount *);
1894extern int may_umount(struct vfsmount *);
1895extern long do_mount(char *, char *, char *, unsigned long, void *);
Al Viro589ff872009-04-18 03:28:19 -04001896extern struct vfsmount *collect_mounts(struct path *);
Al Viro8aec0802007-06-07 12:20:32 -04001897extern void drop_collected_mounts(struct vfsmount *);
Al Viro1f707132010-01-30 22:51:25 -05001898extern int iterate_mounts(int (*)(struct vfsmount *, void *), void *,
1899 struct vfsmount *);
Christoph Hellwigebabe9a2010-07-07 18:53:11 +02001900extern int vfs_statfs(struct path *, struct kstatfs *);
Al Viroc8b91ac2011-03-12 10:41:39 -05001901extern int user_statfs(const char __user *, struct kstatfs *);
1902extern int fd_statfs(int, struct kstatfs *);
Christoph Hellwigebabe9a2010-07-07 18:53:11 +02001903extern int statfs_by_dentry(struct dentry *, struct kstatfs *);
Josef Bacik18e9e512010-03-23 10:34:56 -04001904extern int freeze_super(struct super_block *super);
1905extern int thaw_super(struct super_block *super);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001906
Al Viroce3b0f82009-03-29 19:08:22 -04001907extern int current_umask(void);
1908
Miklos Szeredif87fd4c2006-01-16 22:14:23 -08001909/* /sys/fs */
Greg Kroah-Hartman00d26662007-10-29 14:17:23 -06001910extern struct kobject *fs_kobj;
Miklos Szeredif87fd4c2006-01-16 22:14:23 -08001911
Linus Torvalds435f49a2010-10-29 10:36:49 -07001912#define MAX_RW_COUNT (INT_MAX & PAGE_CACHE_MASK)
Thomas Petazzonibfcd17a2008-08-06 15:12:22 +02001913extern int rw_verify_area(int, struct file *, loff_t *, size_t);
1914
Linus Torvalds1da177e2005-04-16 15:20:36 -07001915#define FLOCK_VERIFY_READ 1
1916#define FLOCK_VERIFY_WRITE 2
1917
Thomas Petazzonibfcd17a2008-08-06 15:12:22 +02001918#ifdef CONFIG_FILE_LOCKING
Linus Torvalds1da177e2005-04-16 15:20:36 -07001919extern int locks_mandatory_locked(struct inode *);
1920extern int locks_mandatory_area(int, struct inode *, struct file *, loff_t, size_t);
1921
1922/*
1923 * Candidates for mandatory locking have the setgid bit set
1924 * but no group execute bit - an otherwise meaningless combination.
1925 */
Pavel Emelyanova16877c2007-10-01 14:41:11 -07001926
1927static inline int __mandatory_lock(struct inode *ino)
1928{
1929 return (ino->i_mode & (S_ISGID | S_IXGRP)) == S_ISGID;
1930}
1931
1932/*
1933 * ... and these candidates should be on MS_MANDLOCK mounted fs,
1934 * otherwise these will be advisory locks
1935 */
1936
1937static inline int mandatory_lock(struct inode *ino)
1938{
1939 return IS_MANDLOCK(ino) && __mandatory_lock(ino);
1940}
Linus Torvalds1da177e2005-04-16 15:20:36 -07001941
1942static inline int locks_verify_locked(struct inode *inode)
1943{
Pavel Emelyanova16877c2007-10-01 14:41:11 -07001944 if (mandatory_lock(inode))
Linus Torvalds1da177e2005-04-16 15:20:36 -07001945 return locks_mandatory_locked(inode);
1946 return 0;
1947}
1948
Linus Torvalds1da177e2005-04-16 15:20:36 -07001949static inline int locks_verify_truncate(struct inode *inode,
1950 struct file *filp,
1951 loff_t size)
1952{
Pavel Emelyanova16877c2007-10-01 14:41:11 -07001953 if (inode->i_flock && mandatory_lock(inode))
Linus Torvalds1da177e2005-04-16 15:20:36 -07001954 return locks_mandatory_area(
1955 FLOCK_VERIFY_WRITE, inode, filp,
1956 size < inode->i_size ? size : inode->i_size,
1957 (size < inode->i_size ? inode->i_size - size
1958 : size - inode->i_size)
1959 );
1960 return 0;
1961}
1962
1963static inline int break_lease(struct inode *inode, unsigned int mode)
1964{
1965 if (inode->i_flock)
1966 return __break_lease(inode, mode);
1967 return 0;
1968}
Thomas Petazzonibfcd17a2008-08-06 15:12:22 +02001969#else /* !CONFIG_FILE_LOCKING */
Steven Whitehouseaf5df562009-01-20 10:29:46 +00001970static inline int locks_mandatory_locked(struct inode *inode)
1971{
1972 return 0;
1973}
1974
1975static inline int locks_mandatory_area(int rw, struct inode *inode,
1976 struct file *filp, loff_t offset,
1977 size_t count)
1978{
1979 return 0;
1980}
1981
1982static inline int __mandatory_lock(struct inode *inode)
1983{
1984 return 0;
1985}
1986
1987static inline int mandatory_lock(struct inode *inode)
1988{
1989 return 0;
1990}
1991
1992static inline int locks_verify_locked(struct inode *inode)
1993{
1994 return 0;
1995}
1996
1997static inline int locks_verify_truncate(struct inode *inode, struct file *filp,
1998 size_t size)
1999{
2000 return 0;
2001}
2002
2003static inline int break_lease(struct inode *inode, unsigned int mode)
2004{
2005 return 0;
2006}
2007
Thomas Petazzonibfcd17a2008-08-06 15:12:22 +02002008#endif /* CONFIG_FILE_LOCKING */
Linus Torvalds1da177e2005-04-16 15:20:36 -07002009
2010/* fs/open.c */
2011
NeilBrown4a301312006-01-08 01:02:39 -08002012extern int do_truncate(struct dentry *, loff_t start, unsigned int time_attrs,
2013 struct file *filp);
Ankit Jain3e63cbb2009-06-19 14:28:07 -04002014extern int do_fallocate(struct file *file, int mode, loff_t offset,
2015 loff_t len);
Jason Uhlenkott8e8a1402007-10-20 03:16:18 +02002016extern long do_sys_open(int dfd, const char __user *filename, int flags,
Ulrich Drepper5590ff02006-01-18 17:43:53 -08002017 int mode);
Linus Torvalds1da177e2005-04-16 15:20:36 -07002018extern struct file *filp_open(const char *, int, int);
Al Viro73d049a2011-03-11 12:08:24 -05002019extern struct file *file_open_root(struct dentry *, struct vfsmount *,
2020 const char *, int);
David Howells745ca242008-11-14 10:39:22 +11002021extern struct file * dentry_open(struct dentry *, struct vfsmount *, int,
2022 const struct cred *);
Linus Torvalds1da177e2005-04-16 15:20:36 -07002023extern int filp_close(struct file *, fl_owner_t id);
2024extern char * getname(const char __user *);
2025
Ankit Jain3e63cbb2009-06-19 14:28:07 -04002026/* fs/ioctl.c */
2027
2028extern int ioctl_preallocate(struct file *filp, void __user *argp);
2029
Linus Torvalds1da177e2005-04-16 15:20:36 -07002030/* fs/dcache.c */
2031extern void __init vfs_caches_init_early(void);
2032extern void __init vfs_caches_init(unsigned long);
2033
Christoph Lameterb86c0892006-12-06 20:32:57 -08002034extern struct kmem_cache *names_cachep;
2035
Vegard Nossum3446a8a2009-05-16 11:22:14 +02002036#define __getname_gfp(gfp) kmem_cache_alloc(names_cachep, (gfp))
2037#define __getname() __getname_gfp(GFP_KERNEL)
2038#define __putname(name) kmem_cache_free(names_cachep, (void *)(name))
Linus Torvalds1da177e2005-04-16 15:20:36 -07002039#ifndef CONFIG_AUDITSYSCALL
2040#define putname(name) __putname(name)
2041#else
2042extern void putname(const char *name);
2043#endif
2044
David Howells93614012006-09-30 20:45:40 +02002045#ifdef CONFIG_BLOCK
Linus Torvalds1da177e2005-04-16 15:20:36 -07002046extern int register_blkdev(unsigned int, const char *);
Akinobu Mitaf4480242007-07-17 04:03:47 -07002047extern void unregister_blkdev(unsigned int, const char *);
Linus Torvalds1da177e2005-04-16 15:20:36 -07002048extern struct block_device *bdget(dev_t);
Alan Jenkinsdddac6a2009-07-29 21:07:55 +02002049extern struct block_device *bdgrab(struct block_device *bdev);
Linus Torvalds1da177e2005-04-16 15:20:36 -07002050extern void bd_set_size(struct block_device *, loff_t size);
2051extern void bd_forget(struct inode *inode);
2052extern void bdput(struct block_device *);
Nick Piggin585d3bc2009-02-25 10:44:19 +01002053extern void invalidate_bdev(struct block_device *);
2054extern int sync_blockdev(struct block_device *bdev);
2055extern struct super_block *freeze_bdev(struct block_device *);
Eric Sandeenc2d75432009-03-31 15:23:46 -07002056extern void emergency_thaw_all(void);
Nick Piggin585d3bc2009-02-25 10:44:19 +01002057extern int thaw_bdev(struct block_device *bdev, struct super_block *sb);
2058extern int fsync_bdev(struct block_device *);
David Howells93614012006-09-30 20:45:40 +02002059#else
2060static inline void bd_forget(struct inode *inode) {}
Al Viro47e44912009-04-01 07:07:16 -04002061static inline int sync_blockdev(struct block_device *bdev) { return 0; }
2062static inline void invalidate_bdev(struct block_device *bdev) {}
2063
2064static inline struct super_block *freeze_bdev(struct block_device *sb)
2065{
2066 return NULL;
2067}
2068
2069static inline int thaw_bdev(struct block_device *bdev, struct super_block *sb)
2070{
2071 return 0;
2072}
David Howells93614012006-09-30 20:45:40 +02002073#endif
Jan Kara60b06802009-04-27 16:43:53 +02002074extern int sync_filesystem(struct super_block *);
David Howells93614012006-09-30 20:45:40 +02002075extern const struct file_operations def_blk_fops;
Arjan van de Ven4b6f5d22006-03-28 01:56:42 -08002076extern const struct file_operations def_chr_fops;
2077extern const struct file_operations bad_sock_fops;
2078extern const struct file_operations def_fifo_fops;
David Howells93614012006-09-30 20:45:40 +02002079#ifdef CONFIG_BLOCK
Linus Torvalds1da177e2005-04-16 15:20:36 -07002080extern int ioctl_by_bdev(struct block_device *, unsigned, unsigned long);
Al Viro56b26ad2008-09-19 03:17:36 -04002081extern int blkdev_ioctl(struct block_device *, fmode_t, unsigned, unsigned long);
Linus Torvalds1da177e2005-04-16 15:20:36 -07002082extern long compat_blkdev_ioctl(struct file *, unsigned, unsigned long);
Tejun Heoe525fd82010-11-13 11:55:17 +01002083extern int blkdev_get(struct block_device *bdev, fmode_t mode, void *holder);
Tejun Heod4d77622010-11-13 11:55:18 +01002084extern struct block_device *blkdev_get_by_path(const char *path, fmode_t mode,
2085 void *holder);
2086extern struct block_device *blkdev_get_by_dev(dev_t dev, fmode_t mode,
2087 void *holder);
Tejun Heoe525fd82010-11-13 11:55:17 +01002088extern int blkdev_put(struct block_device *bdev, fmode_t mode);
Jun'ichi Nomura641dc632006-03-27 01:17:57 -08002089#ifdef CONFIG_SYSFS
Tejun Heoe09b4572010-11-13 11:55:17 +01002090extern int bd_link_disk_holder(struct block_device *bdev, struct gendisk *disk);
Tejun Heo49731ba2011-01-14 18:43:57 +01002091extern void bd_unlink_disk_holder(struct block_device *bdev,
2092 struct gendisk *disk);
Jun'ichi Nomura641dc632006-03-27 01:17:57 -08002093#else
Tejun Heoe09b4572010-11-13 11:55:17 +01002094static inline int bd_link_disk_holder(struct block_device *bdev,
2095 struct gendisk *disk)
2096{
2097 return 0;
2098}
Tejun Heo49731ba2011-01-14 18:43:57 +01002099static inline void bd_unlink_disk_holder(struct block_device *bdev,
2100 struct gendisk *disk)
2101{
2102}
Jun'ichi Nomura641dc632006-03-27 01:17:57 -08002103#endif
David Howells93614012006-09-30 20:45:40 +02002104#endif
Linus Torvalds1da177e2005-04-16 15:20:36 -07002105
2106/* fs/char_dev.c */
Joe Korty68eef3b2006-03-31 02:30:32 -08002107#define CHRDEV_MAJOR_HASH_SIZE 255
Linus Torvalds1da177e2005-04-16 15:20:36 -07002108extern int alloc_chrdev_region(dev_t *, unsigned, unsigned, const char *);
2109extern int register_chrdev_region(dev_t, unsigned, const char *);
Tejun Heo1905b1b2009-08-06 18:13:23 +09002110extern int __register_chrdev(unsigned int major, unsigned int baseminor,
2111 unsigned int count, const char *name,
2112 const struct file_operations *fops);
2113extern void __unregister_chrdev(unsigned int major, unsigned int baseminor,
2114 unsigned int count, const char *name);
Linus Torvalds1da177e2005-04-16 15:20:36 -07002115extern void unregister_chrdev_region(dev_t, unsigned);
Joe Korty68eef3b2006-03-31 02:30:32 -08002116extern void chrdev_show(struct seq_file *,off_t);
Linus Torvalds1da177e2005-04-16 15:20:36 -07002117
Tejun Heo1905b1b2009-08-06 18:13:23 +09002118static inline int register_chrdev(unsigned int major, const char *name,
2119 const struct file_operations *fops)
2120{
2121 return __register_chrdev(major, 0, 256, name, fops);
2122}
2123
2124static inline void unregister_chrdev(unsigned int major, const char *name)
2125{
2126 __unregister_chrdev(major, 0, 256, name);
2127}
2128
Linus Torvalds1da177e2005-04-16 15:20:36 -07002129/* fs/block_dev.c */
2130#define BDEVNAME_SIZE 32 /* Largest string for a blockdev identifier */
Tejun Heo1f014292008-08-25 19:47:23 +09002131#define BDEVT_SIZE 10 /* Largest string for MAJ:MIN for blkdev */
David Howells93614012006-09-30 20:45:40 +02002132
2133#ifdef CONFIG_BLOCK
2134#define BLKDEV_MAJOR_HASH_SIZE 255
Linus Torvalds1da177e2005-04-16 15:20:36 -07002135extern const char *__bdevname(dev_t, char *buffer);
2136extern const char *bdevname(struct block_device *bdev, char *buffer);
2137extern struct block_device *lookup_bdev(const char *);
Joe Korty68eef3b2006-03-31 02:30:32 -08002138extern void blkdev_show(struct seq_file *,off_t);
Al Viro30c40d22008-02-22 19:50:45 -05002139
David Howells93614012006-09-30 20:45:40 +02002140#else
2141#define BLKDEV_MAJOR_HASH_SIZE 0
2142#endif
Linus Torvalds1da177e2005-04-16 15:20:36 -07002143
2144extern void init_special_inode(struct inode *, umode_t, dev_t);
2145
2146/* Invalid inode operations -- fs/bad_inode.c */
2147extern void make_bad_inode(struct inode *);
2148extern int is_bad_inode(struct inode *);
2149
Denys Vlasenkod2d96482008-07-01 14:16:09 +02002150extern const struct file_operations read_pipefifo_fops;
2151extern const struct file_operations write_pipefifo_fops;
2152extern const struct file_operations rdwr_pipefifo_fops;
Linus Torvalds1da177e2005-04-16 15:20:36 -07002153
2154extern int fs_may_remount_ro(struct super_block *);
2155
David Howells93614012006-09-30 20:45:40 +02002156#ifdef CONFIG_BLOCK
Linus Torvalds1da177e2005-04-16 15:20:36 -07002157/*
2158 * return READ, READA, or WRITE
2159 */
2160#define bio_rw(bio) ((bio)->bi_rw & (RW_MASK | RWA_MASK))
2161
2162/*
2163 * return data direction, READ or WRITE
2164 */
2165#define bio_data_dir(bio) ((bio)->bi_rw & 1)
2166
Andrew Pattersonc3279d12008-09-04 14:27:25 -06002167extern void check_disk_size_change(struct gendisk *disk,
2168 struct block_device *bdev);
Andrew Patterson0c002c22008-09-04 14:27:20 -06002169extern int revalidate_disk(struct gendisk *);
Linus Torvalds1da177e2005-04-16 15:20:36 -07002170extern int check_disk_change(struct block_device *);
NeilBrown93b270f2011-02-24 17:25:47 +11002171extern int __invalidate_device(struct block_device *, bool);
Linus Torvalds1da177e2005-04-16 15:20:36 -07002172extern int invalidate_partition(struct gendisk *, int);
David Howells93614012006-09-30 20:45:40 +02002173#endif
Linus Torvalds1da177e2005-04-16 15:20:36 -07002174unsigned long invalidate_mapping_pages(struct address_space *mapping,
2175 pgoff_t start, pgoff_t end);
Anton Altaparmakov54bc4852007-02-10 01:45:38 -08002176
Linus Torvalds1da177e2005-04-16 15:20:36 -07002177static inline void invalidate_remote_inode(struct inode *inode)
2178{
2179 if (S_ISREG(inode->i_mode) || S_ISDIR(inode->i_mode) ||
2180 S_ISLNK(inode->i_mode))
Andrew Mortonfc0ecff2007-02-10 01:45:39 -08002181 invalidate_mapping_pages(inode->i_mapping, 0, -1);
Linus Torvalds1da177e2005-04-16 15:20:36 -07002182}
2183extern int invalidate_inode_pages2(struct address_space *mapping);
2184extern int invalidate_inode_pages2_range(struct address_space *mapping,
2185 pgoff_t start, pgoff_t end);
2186extern int write_inode_now(struct inode *, int);
2187extern int filemap_fdatawrite(struct address_space *);
2188extern int filemap_flush(struct address_space *);
2189extern int filemap_fdatawait(struct address_space *);
Jan Karad3bccb6f2009-08-17 19:30:27 +02002190extern int filemap_fdatawait_range(struct address_space *, loff_t lstart,
2191 loff_t lend);
Linus Torvalds1da177e2005-04-16 15:20:36 -07002192extern int filemap_write_and_wait(struct address_space *mapping);
2193extern int filemap_write_and_wait_range(struct address_space *mapping,
2194 loff_t lstart, loff_t lend);
Andrew Mortonebcf28e2006-03-24 03:18:04 -08002195extern int __filemap_fdatawrite_range(struct address_space *mapping,
2196 loff_t start, loff_t end, int sync_mode);
Jan Karaf4c0a0f2008-07-11 19:27:31 -04002197extern int filemap_fdatawrite_range(struct address_space *mapping,
2198 loff_t start, loff_t end);
Andrew Mortonebcf28e2006-03-24 03:18:04 -08002199
Christoph Hellwig8018ab02010-03-22 17:32:25 +01002200extern int vfs_fsync_range(struct file *file, loff_t start, loff_t end,
2201 int datasync);
2202extern int vfs_fsync(struct file *file, int datasync);
Jan Kara148f9482009-08-17 19:52:36 +02002203extern int generic_write_sync(struct file *file, loff_t pos, loff_t count);
Linus Torvalds1da177e2005-04-16 15:20:36 -07002204extern void sync_supers(void);
Linus Torvalds1da177e2005-04-16 15:20:36 -07002205extern void emergency_sync(void);
2206extern void emergency_remount(void);
David Howells93614012006-09-30 20:45:40 +02002207#ifdef CONFIG_BLOCK
Linus Torvalds1da177e2005-04-16 15:20:36 -07002208extern sector_t bmap(struct inode *, sector_t);
David Howells93614012006-09-30 20:45:40 +02002209#endif
Linus Torvalds1da177e2005-04-16 15:20:36 -07002210extern int notify_change(struct dentry *, struct iattr *);
Al Virof419a2e2008-07-22 00:07:17 -04002211extern int inode_permission(struct inode *, int);
Al Viro2830ba72011-06-20 19:16:29 -04002212extern int generic_permission(struct inode *, int);
Linus Torvalds1da177e2005-04-16 15:20:36 -07002213
Miklos Szeredif696a362008-07-31 13:41:58 +02002214static inline bool execute_ok(struct inode *inode)
2215{
2216 return (inode->i_mode & S_IXUGO) || S_ISDIR(inode->i_mode);
2217}
2218
Al Viro07b8ce12011-06-20 10:52:57 -04002219/*
2220 * get_write_access() gets write permission for a file.
2221 * put_write_access() releases this write permission.
2222 * This is used for regular files.
2223 * We cannot support write (and maybe mmap read-write shared) accesses and
2224 * MAP_DENYWRITE mmappings simultaneously. The i_writecount field of an inode
2225 * can have the following values:
2226 * 0: no writers, no VM_DENYWRITE mappings
2227 * < 0: (-i_writecount) vm_area_structs with VM_DENYWRITE set exist
2228 * > 0: (i_writecount) users are writing to the file.
2229 *
2230 * Normally we operate on that counter with atomic_{inc,dec} and it's safe
2231 * except for the cases where we don't hold i_writecount yet. Then we need to
2232 * use {get,deny}_write_access() - these functions check the sign and refuse
2233 * to do the change if sign is wrong.
2234 */
2235static inline int get_write_access(struct inode *inode)
2236{
2237 return atomic_inc_unless_negative(&inode->i_writecount) ? 0 : -ETXTBSY;
2238}
2239static inline int deny_write_access(struct file *file)
2240{
2241 struct inode *inode = file->f_path.dentry->d_inode;
2242 return atomic_dec_unless_positive(&inode->i_writecount) ? 0 : -ETXTBSY;
2243}
Linus Torvalds1da177e2005-04-16 15:20:36 -07002244static inline void put_write_access(struct inode * inode)
2245{
2246 atomic_dec(&inode->i_writecount);
2247}
2248static inline void allow_write_access(struct file *file)
2249{
2250 if (file)
Josef "Jeff" Sipek0f7fc9e2006-12-08 02:36:35 -08002251 atomic_inc(&file->f_path.dentry->d_inode->i_writecount);
Linus Torvalds1da177e2005-04-16 15:20:36 -07002252}
Mimi Zohara5c96eb2010-11-02 10:11:37 -04002253#ifdef CONFIG_IMA
2254static inline void i_readcount_dec(struct inode *inode)
2255{
2256 BUG_ON(!atomic_read(&inode->i_readcount));
2257 atomic_dec(&inode->i_readcount);
2258}
2259static inline void i_readcount_inc(struct inode *inode)
2260{
2261 atomic_inc(&inode->i_readcount);
2262}
2263#else
2264static inline void i_readcount_dec(struct inode *inode)
2265{
2266 return;
2267}
2268static inline void i_readcount_inc(struct inode *inode)
2269{
2270 return;
2271}
2272#endif
Ulrich Dreppered8cae82008-07-23 21:29:30 -07002273extern int do_pipe_flags(int *, int);
Ulrich Drepperbe61a862008-07-23 21:29:40 -07002274extern struct file *create_read_pipe(struct file *f, int flags);
2275extern struct file *create_write_pipe(int flags);
Andi Kleend6cbd282006-09-30 23:29:26 -07002276extern void free_write_pipe(struct file *);
Linus Torvalds1da177e2005-04-16 15:20:36 -07002277
Mimi Zohar6777d772009-08-21 14:32:48 -04002278extern int kernel_read(struct file *, loff_t, char *, unsigned long);
Linus Torvalds1da177e2005-04-16 15:20:36 -07002279extern struct file * open_exec(const char *);
2280
2281/* fs/dcache.c -- generic fs support functions */
2282extern int is_subdir(struct dentry *, struct dentry *);
Al Viro2096f752010-01-30 13:16:21 -05002283extern int path_is_under(struct path *, struct path *);
Linus Torvalds1da177e2005-04-16 15:20:36 -07002284extern ino_t find_inode_number(struct dentry *, struct qstr *);
2285
2286#include <linux/err.h>
2287
2288/* needed for stackable file system support */
2289extern loff_t default_llseek(struct file *file, loff_t offset, int origin);
2290
2291extern loff_t vfs_llseek(struct file *file, loff_t offset, int origin);
2292
Christoph Hellwig54e34622009-08-07 14:38:25 -03002293extern int inode_init_always(struct super_block *, struct inode *);
Linus Torvalds1da177e2005-04-16 15:20:36 -07002294extern void inode_init_once(struct inode *);
Miklos Szeredi2aa15892011-02-23 13:49:47 +01002295extern void address_space_init_once(struct address_space *mapping);
Al Viro7de9c6ee2010-10-23 11:11:40 -04002296extern void ihold(struct inode * inode);
Linus Torvalds1da177e2005-04-16 15:20:36 -07002297extern void iput(struct inode *);
2298extern struct inode * igrab(struct inode *);
2299extern ino_t iunique(struct super_block *, ino_t);
2300extern int inode_needs_sync(struct inode *inode);
Al Viro45321ac2010-06-07 13:43:19 -04002301extern int generic_delete_inode(struct inode *inode);
2302extern int generic_drop_inode(struct inode *inode);
Linus Torvalds1da177e2005-04-16 15:20:36 -07002303
Anton Altaparmakov88bd5122005-07-13 01:10:44 -07002304extern struct inode *ilookup5_nowait(struct super_block *sb,
2305 unsigned long hashval, int (*test)(struct inode *, void *),
2306 void *data);
Linus Torvalds1da177e2005-04-16 15:20:36 -07002307extern struct inode *ilookup5(struct super_block *sb, unsigned long hashval,
2308 int (*test)(struct inode *, void *), void *data);
2309extern struct inode *ilookup(struct super_block *sb, unsigned long ino);
2310
2311extern struct inode * iget5_locked(struct super_block *, unsigned long, int (*test)(struct inode *, void *), int (*set)(struct inode *, void *), void *);
2312extern struct inode * iget_locked(struct super_block *, unsigned long);
Al Viro261bca82008-12-30 01:48:21 -05002313extern int insert_inode_locked4(struct inode *, unsigned long, int (*test)(struct inode *, void *), void *);
2314extern int insert_inode_locked(struct inode *);
Linus Torvalds1da177e2005-04-16 15:20:36 -07002315extern void unlock_new_inode(struct inode *);
Christoph Hellwig85fe4022010-10-23 11:19:54 -04002316extern unsigned int get_next_ino(void);
Linus Torvalds1da177e2005-04-16 15:20:36 -07002317
Linus Torvalds1da177e2005-04-16 15:20:36 -07002318extern void __iget(struct inode * inode);
David Howellsb46980f2008-02-07 00:15:27 -08002319extern void iget_failed(struct inode *);
Al Virob0683aa2010-06-04 20:55:25 -04002320extern void end_writeback(struct inode *);
Christoph Hellwig2e00c972009-08-07 14:38:29 -03002321extern void __destroy_inode(struct inode *);
Linus Torvalds1da177e2005-04-16 15:20:36 -07002322extern struct inode *new_inode(struct super_block *);
Nick Pigginff0c7d12011-01-07 17:49:50 +11002323extern void free_inode_nonrcu(struct inode *inode);
Jens Axboe01de85e2006-10-17 19:50:36 +02002324extern int should_remove_suid(struct dentry *);
Miklos Szeredi2f1936b2008-06-24 16:50:14 +02002325extern int file_remove_suid(struct file *);
Linus Torvalds1da177e2005-04-16 15:20:36 -07002326
2327extern void __insert_inode_hash(struct inode *, unsigned long hashval);
2328extern void remove_inode_hash(struct inode *);
Christoph Hellwig646ec462010-10-23 07:15:32 -04002329static inline void insert_inode_hash(struct inode *inode)
2330{
Linus Torvalds1da177e2005-04-16 15:20:36 -07002331 __insert_inode_hash(inode, inode->i_ino);
2332}
Christoph Hellwig646ec462010-10-23 07:15:32 -04002333extern void inode_sb_list_add(struct inode *inode);
Linus Torvalds1da177e2005-04-16 15:20:36 -07002334
David Howells93614012006-09-30 20:45:40 +02002335#ifdef CONFIG_BLOCK
Linus Torvalds1da177e2005-04-16 15:20:36 -07002336extern void submit_bio(int, struct bio *);
2337extern int bdev_read_only(struct block_device *);
David Howells93614012006-09-30 20:45:40 +02002338#endif
Linus Torvalds1da177e2005-04-16 15:20:36 -07002339extern int set_blocksize(struct block_device *, int);
2340extern int sb_set_blocksize(struct super_block *, int);
2341extern int sb_min_blocksize(struct super_block *, int);
2342
2343extern int generic_file_mmap(struct file *, struct vm_area_struct *);
2344extern int generic_file_readonly_mmap(struct file *, struct vm_area_struct *);
2345extern int file_read_actor(read_descriptor_t * desc, struct page *page, unsigned long offset, unsigned long size);
Linus Torvalds1da177e2005-04-16 15:20:36 -07002346int generic_write_checks(struct file *file, loff_t *pos, size_t *count, int isblk);
Badari Pulavarty027445c2006-09-30 23:28:46 -07002347extern ssize_t generic_file_aio_read(struct kiocb *, const struct iovec *, unsigned long, loff_t);
Jan Karae4dd9de2009-08-17 18:10:06 +02002348extern ssize_t __generic_file_aio_write(struct kiocb *, const struct iovec *, unsigned long,
2349 loff_t *);
Badari Pulavarty027445c2006-09-30 23:28:46 -07002350extern ssize_t generic_file_aio_write(struct kiocb *, const struct iovec *, unsigned long, loff_t);
Linus Torvalds1da177e2005-04-16 15:20:36 -07002351extern ssize_t generic_file_direct_write(struct kiocb *, const struct iovec *,
2352 unsigned long *, loff_t, loff_t *, size_t, size_t);
2353extern ssize_t generic_file_buffered_write(struct kiocb *, const struct iovec *,
2354 unsigned long, loff_t, loff_t *, size_t, ssize_t);
2355extern ssize_t do_sync_read(struct file *filp, char __user *buf, size_t len, loff_t *ppos);
2356extern 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 -07002357extern int generic_segment_checks(const struct iovec *iov,
2358 unsigned long *nr_segs, size_t *count, int access_flags);
Andrew Morton88e6fae2006-04-11 13:59:36 +02002359
Christoph Hellwigeef99382009-08-20 17:43:41 +02002360/* fs/block_dev.c */
2361extern ssize_t blkdev_aio_write(struct kiocb *iocb, const struct iovec *iov,
2362 unsigned long nr_segs, loff_t pos);
Christoph Hellwig7ea80852010-05-26 17:53:25 +02002363extern int blkdev_fsync(struct file *filp, int datasync);
Christoph Hellwigeef99382009-08-20 17:43:41 +02002364
Andrew Morton88e6fae2006-04-11 13:59:36 +02002365/* fs/splice.c */
Jens Axboecbb7e572006-04-11 14:57:50 +02002366extern ssize_t generic_file_splice_read(struct file *, loff_t *,
Andrew Morton88e6fae2006-04-11 13:59:36 +02002367 struct pipe_inode_info *, size_t, unsigned int);
Miklos Szeredi68181732009-05-07 15:37:36 +02002368extern ssize_t default_file_splice_read(struct file *, loff_t *,
2369 struct pipe_inode_info *, size_t, unsigned int);
Andrew Morton88e6fae2006-04-11 13:59:36 +02002370extern ssize_t generic_file_splice_write(struct pipe_inode_info *,
Jens Axboecbb7e572006-04-11 14:57:50 +02002371 struct file *, loff_t *, size_t, unsigned int);
Andrew Morton88e6fae2006-04-11 13:59:36 +02002372extern ssize_t generic_splice_sendpage(struct pipe_inode_info *pipe,
Jens Axboecbb7e572006-04-11 14:57:50 +02002373 struct file *out, loff_t *, size_t len, unsigned int flags);
2374extern long do_splice_direct(struct file *in, loff_t *ppos, struct file *out,
Andrew Morton88e6fae2006-04-11 13:59:36 +02002375 size_t len, unsigned int flags);
2376
Linus Torvalds1da177e2005-04-16 15:20:36 -07002377extern void
2378file_ra_state_init(struct file_ra_state *ra, struct address_space *mapping);
jan Blunckae6afc32010-05-26 14:44:48 -07002379extern loff_t noop_llseek(struct file *file, loff_t offset, int origin);
Linus Torvalds1da177e2005-04-16 15:20:36 -07002380extern loff_t no_llseek(struct file *file, loff_t offset, int origin);
2381extern loff_t generic_file_llseek(struct file *file, loff_t offset, int origin);
Andi Kleen9465efc2008-06-27 11:05:24 +02002382extern loff_t generic_file_llseek_unlocked(struct file *file, loff_t offset,
2383 int origin);
Linus Torvalds1da177e2005-04-16 15:20:36 -07002384extern int generic_file_open(struct inode * inode, struct file * filp);
2385extern int nonseekable_open(struct inode * inode, struct file * filp);
2386
Carsten Otteceffc072005-06-23 22:05:25 -07002387#ifdef CONFIG_FS_XIP
Carsten Otteeb6fe0c2005-06-23 22:05:28 -07002388extern ssize_t xip_file_read(struct file *filp, char __user *buf, size_t len,
2389 loff_t *ppos);
Carsten Otteceffc072005-06-23 22:05:25 -07002390extern int xip_file_mmap(struct file * file, struct vm_area_struct * vma);
Carsten Otteeb6fe0c2005-06-23 22:05:28 -07002391extern ssize_t xip_file_write(struct file *filp, const char __user *buf,
2392 size_t len, loff_t *ppos);
Carsten Otteceffc072005-06-23 22:05:25 -07002393extern int xip_truncate_page(struct address_space *mapping, loff_t from);
Carsten Otte6d791252005-06-23 22:05:26 -07002394#else
2395static inline int xip_truncate_page(struct address_space *mapping, loff_t from)
2396{
2397 return 0;
2398}
Carsten Otteceffc072005-06-23 22:05:25 -07002399#endif
2400
David Howells93614012006-09-30 20:45:40 +02002401#ifdef CONFIG_BLOCK
Josef Bacikfacd07b2010-05-23 11:00:55 -04002402typedef void (dio_submit_t)(int rw, struct bio *bio, struct inode *inode,
2403 loff_t file_offset);
Linus Torvalds1da177e2005-04-16 15:20:36 -07002404
2405enum {
Christoph Hellwig1e431f52009-11-03 16:44:53 +01002406 /* need locking between buffered and direct access */
2407 DIO_LOCKING = 0x01,
2408
2409 /* filesystem does not support filling holes */
2410 DIO_SKIP_HOLES = 0x02,
Linus Torvalds1da177e2005-04-16 15:20:36 -07002411};
2412
Christoph Hellwigeafdc7d2010-06-04 11:29:53 +02002413void dio_end_io(struct bio *bio, int error);
Christoph Hellwigbd5fe6c2011-06-24 14:29:43 -04002414void inode_dio_wait(struct inode *inode);
2415void inode_dio_done(struct inode *inode);
npiggin@suse.de7bb46a62010-05-27 01:05:33 +10002416
Christoph Hellwigeafdc7d2010-06-04 11:29:53 +02002417ssize_t __blockdev_direct_IO(int rw, struct kiocb *iocb, struct inode *inode,
2418 struct block_device *bdev, const struct iovec *iov, loff_t offset,
2419 unsigned long nr_segs, get_block_t get_block, dio_iodone_t end_io,
2420 dio_submit_t submit_io, int flags);
2421
Linus Torvalds1da177e2005-04-16 15:20:36 -07002422static inline ssize_t blockdev_direct_IO(int rw, struct kiocb *iocb,
Christoph Hellwigaacfc19c2011-06-24 14:29:47 -04002423 struct inode *inode, const struct iovec *iov, loff_t offset,
2424 unsigned long nr_segs, get_block_t get_block)
Linus Torvalds1da177e2005-04-16 15:20:36 -07002425{
Christoph Hellwigaacfc19c2011-06-24 14:29:47 -04002426 return __blockdev_direct_IO(rw, iocb, inode, inode->i_sb->s_bdev, iov,
2427 offset, nr_segs, get_block, NULL, NULL,
Christoph Hellwig1e431f52009-11-03 16:44:53 +01002428 DIO_LOCKING | DIO_SKIP_HOLES);
Linus Torvalds1da177e2005-04-16 15:20:36 -07002429}
David Howells93614012006-09-30 20:45:40 +02002430#endif
Linus Torvalds1da177e2005-04-16 15:20:36 -07002431
Arjan van de Ven4b6f5d22006-03-28 01:56:42 -08002432extern const struct file_operations generic_ro_fops;
Linus Torvalds1da177e2005-04-16 15:20:36 -07002433
2434#define special_file(m) (S_ISCHR(m)||S_ISBLK(m)||S_ISFIFO(m)||S_ISSOCK(m))
2435
2436extern int vfs_readlink(struct dentry *, char __user *, int, const char *);
2437extern int vfs_follow_link(struct nameidata *, const char *);
2438extern int page_readlink(struct dentry *, char __user *, int);
Linus Torvaldscc314ee2005-08-19 18:02:56 -07002439extern void *page_follow_link_light(struct dentry *, struct nameidata *);
2440extern void page_put_link(struct dentry *, struct nameidata *, void *);
Kirill Korotaev0adb25d2006-03-11 03:27:13 -08002441extern int __page_symlink(struct inode *inode, const char *symname, int len,
Nick Piggin54566b22009-01-04 12:00:53 -08002442 int nofs);
Linus Torvalds1da177e2005-04-16 15:20:36 -07002443extern int page_symlink(struct inode *inode, const char *symname, int len);
Arjan van de Venc5ef1c42007-02-12 00:55:40 -08002444extern const struct inode_operations page_symlink_inode_operations;
Linus Torvalds1da177e2005-04-16 15:20:36 -07002445extern int generic_readlink(struct dentry *, char __user *, int);
2446extern void generic_fillattr(struct inode *, struct kstat *);
2447extern int vfs_getattr(struct vfsmount *, struct dentry *, struct kstat *);
Dmitry Monakhovb4627072009-12-14 15:21:12 +03002448void __inode_add_bytes(struct inode *inode, loff_t bytes);
Linus Torvalds1da177e2005-04-16 15:20:36 -07002449void inode_add_bytes(struct inode *inode, loff_t bytes);
2450void inode_sub_bytes(struct inode *inode, loff_t bytes);
2451loff_t inode_get_bytes(struct inode *inode);
2452void inode_set_bytes(struct inode *inode, loff_t bytes);
2453
2454extern int vfs_readdir(struct file *, filldir_t, void *);
2455
David Howellsc7887322010-08-11 11:26:22 +01002456extern int vfs_stat(const char __user *, struct kstat *);
2457extern int vfs_lstat(const char __user *, struct kstat *);
Linus Torvalds1da177e2005-04-16 15:20:36 -07002458extern int vfs_fstat(unsigned int, struct kstat *);
David Howellsc7887322010-08-11 11:26:22 +01002459extern int vfs_fstatat(int , const char __user *, struct kstat *, int);
Linus Torvalds1da177e2005-04-16 15:20:36 -07002460
Erez Zadokdeb21db2008-02-07 00:13:25 -08002461extern int do_vfs_ioctl(struct file *filp, unsigned int fd, unsigned int cmd,
2462 unsigned long arg);
Steven Whitehousee9079cc2008-10-14 14:43:29 +01002463extern int __generic_block_fiemap(struct inode *inode,
Josef Bacik3a3076f2010-04-23 12:17:17 -04002464 struct fiemap_extent_info *fieinfo,
2465 loff_t start, loff_t len,
2466 get_block_t *get_block);
Josef Bacik68c9d702008-10-03 17:32:43 -04002467extern int generic_block_fiemap(struct inode *inode,
2468 struct fiemap_extent_info *fieinfo, u64 start,
2469 u64 len, get_block_t *get_block);
Linus Torvalds1da177e2005-04-16 15:20:36 -07002470
Miklos Szeredic18479f2007-10-18 23:39:11 -07002471extern void get_filesystem(struct file_system_type *fs);
2472extern void put_filesystem(struct file_system_type *fs);
Linus Torvalds1da177e2005-04-16 15:20:36 -07002473extern struct file_system_type *get_fs_type(const char *name);
2474extern struct super_block *get_super(struct block_device *);
Christoph Hellwig45042302009-08-03 23:28:35 +02002475extern struct super_block *get_active_super(struct block_device *bdev);
Linus Torvalds1da177e2005-04-16 15:20:36 -07002476extern struct super_block *user_get_super(dev_t);
2477extern void drop_super(struct super_block *sb);
Al Viro01a05b32010-03-23 06:06:58 -04002478extern void iterate_supers(void (*)(struct super_block *, void *), void *);
Al Viro43e15cd2011-06-03 20:16:57 -04002479extern void iterate_supers_type(struct file_system_type *,
2480 void (*)(struct super_block *, void *), void *);
Linus Torvalds1da177e2005-04-16 15:20:36 -07002481
2482extern int dcache_dir_open(struct inode *, struct file *);
2483extern int dcache_dir_close(struct inode *, struct file *);
2484extern loff_t dcache_dir_lseek(struct file *, loff_t, int);
2485extern int dcache_readdir(struct file *, void *, filldir_t);
npiggin@suse.de7bb46a62010-05-27 01:05:33 +10002486extern int simple_setattr(struct dentry *, struct iattr *);
Linus Torvalds1da177e2005-04-16 15:20:36 -07002487extern int simple_getattr(struct vfsmount *, struct dentry *, struct kstat *);
David Howells726c3342006-06-23 02:02:58 -07002488extern int simple_statfs(struct dentry *, struct kstatfs *);
Linus Torvalds1da177e2005-04-16 15:20:36 -07002489extern int simple_link(struct dentry *, struct inode *, struct dentry *);
2490extern int simple_unlink(struct inode *, struct dentry *);
2491extern int simple_rmdir(struct inode *, struct dentry *);
2492extern int simple_rename(struct inode *, struct dentry *, struct inode *, struct dentry *);
Christoph Hellwig1b061d92010-05-26 17:53:41 +02002493extern int noop_fsync(struct file *, int);
Linus Torvalds1da177e2005-04-16 15:20:36 -07002494extern int simple_empty(struct dentry *);
2495extern int simple_readpage(struct file *file, struct page *page);
Nick Pigginafddba42007-10-16 01:25:01 -07002496extern int simple_write_begin(struct file *file, struct address_space *mapping,
2497 loff_t pos, unsigned len, unsigned flags,
2498 struct page **pagep, void **fsdata);
2499extern int simple_write_end(struct file *file, struct address_space *mapping,
2500 loff_t pos, unsigned len, unsigned copied,
2501 struct page *page, void *fsdata);
Linus Torvalds1da177e2005-04-16 15:20:36 -07002502
2503extern struct dentry *simple_lookup(struct inode *, struct dentry *, struct nameidata *);
2504extern ssize_t generic_read_dir(struct file *, char __user *, size_t, loff_t *);
Arjan van de Ven4b6f5d22006-03-28 01:56:42 -08002505extern const struct file_operations simple_dir_operations;
Arjan van de Venc5ef1c42007-02-12 00:55:40 -08002506extern const struct inode_operations simple_dir_inode_operations;
Arjan van de Ven4b6f5d22006-03-28 01:56:42 -08002507struct tree_descr { char *name; const struct file_operations *ops; int mode; };
Linus Torvalds1da177e2005-04-16 15:20:36 -07002508struct dentry *d_alloc_name(struct dentry *, const char *);
Roberto Sassu7d683a02010-06-03 11:58:28 +02002509extern int simple_fill_super(struct super_block *, unsigned long, struct tree_descr *);
Trond Myklebust1f5ce9e2006-06-09 09:34:16 -04002510extern int simple_pin_fs(struct file_system_type *, struct vfsmount **mount, int *count);
Linus Torvalds1da177e2005-04-16 15:20:36 -07002511extern void simple_release_fs(struct vfsmount **mount, int *count);
2512
Akinobu Mita93b07112008-06-05 22:46:21 -07002513extern ssize_t simple_read_from_buffer(void __user *to, size_t count,
2514 loff_t *ppos, const void *from, size_t available);
Jiri Slaby6a727b42010-05-01 23:51:22 +02002515extern ssize_t simple_write_to_buffer(void *to, size_t available, loff_t *ppos,
2516 const void __user *from, size_t count);
Linus Torvalds1da177e2005-04-16 15:20:36 -07002517
Christoph Hellwig1b061d92010-05-26 17:53:41 +02002518extern int generic_file_fsync(struct file *, int);
Al Virod5aacad2009-06-07 14:56:44 -04002519
Patrick J. LoPresti30ca22c2010-07-22 15:03:41 -07002520extern int generic_check_addressable(unsigned, u64);
2521
Christoph Lametere965f962006-02-01 03:05:41 -08002522#ifdef CONFIG_MIGRATION
Christoph Lameter2d1db3b2006-06-23 02:03:33 -07002523extern int buffer_migrate_page(struct address_space *,
2524 struct page *, struct page *);
Christoph Lametere965f962006-02-01 03:05:41 -08002525#else
2526#define buffer_migrate_page NULL
2527#endif
2528
npiggin@suse.de25d9e2d2009-08-21 02:35:05 +10002529extern int inode_change_ok(const struct inode *, struct iattr *);
2530extern int inode_newsize_ok(const struct inode *, loff_t offset);
Christoph Hellwig6a1a90a2010-06-04 11:30:00 +02002531extern void setattr_copy(struct inode *inode, const struct iattr *attr);
Linus Torvalds1da177e2005-04-16 15:20:36 -07002532
Christoph Hellwig870f4812006-01-09 20:52:01 -08002533extern void file_update_time(struct file *file);
Linus Torvalds1da177e2005-04-16 15:20:36 -07002534
Miklos Szeredib3b304a2008-02-08 04:21:35 -08002535extern int generic_show_options(struct seq_file *m, struct vfsmount *mnt);
2536extern void save_mount_options(struct super_block *sb, char *options);
Al Viro2a32ceb2009-05-08 16:05:57 -04002537extern void replace_mount_options(struct super_block *sb, char *options);
Miklos Szeredib3b304a2008-02-08 04:21:35 -08002538
Linus Torvalds1da177e2005-04-16 15:20:36 -07002539static inline ino_t parent_ino(struct dentry *dentry)
2540{
2541 ino_t res;
2542
Nick Pigginb5c84bf2011-01-07 17:49:38 +11002543 /*
2544 * Don't strictly need d_lock here? If the parent ino could change
2545 * then surely we'd have a deeper race in the caller?
2546 */
Linus Torvalds1da177e2005-04-16 15:20:36 -07002547 spin_lock(&dentry->d_lock);
2548 res = dentry->d_parent->d_inode->i_ino;
2549 spin_unlock(&dentry->d_lock);
2550 return res;
2551}
2552
Linus Torvalds1da177e2005-04-16 15:20:36 -07002553/* Transaction based IO helpers */
2554
2555/*
2556 * An argresp is stored in an allocated page and holds the
2557 * size of the argument or response, along with its content
2558 */
2559struct simple_transaction_argresp {
2560 ssize_t size;
2561 char data[0];
2562};
2563
2564#define SIMPLE_TRANSACTION_LIMIT (PAGE_SIZE - sizeof(struct simple_transaction_argresp))
2565
2566char *simple_transaction_get(struct file *file, const char __user *buf,
2567 size_t size);
2568ssize_t simple_transaction_read(struct file *file, char __user *buf,
2569 size_t size, loff_t *pos);
2570int simple_transaction_release(struct inode *inode, struct file *file);
2571
Ingo Molnar76791ab2009-03-25 16:48:35 +01002572void simple_transaction_set(struct file *file, size_t n);
Linus Torvalds1da177e2005-04-16 15:20:36 -07002573
Arnd Bergmannacaefc22005-05-18 14:40:59 +02002574/*
2575 * simple attribute files
2576 *
2577 * These attributes behave similar to those in sysfs:
2578 *
2579 * Writing to an attribute immediately sets a value, an open file can be
2580 * written to multiple times.
2581 *
2582 * Reading from an attribute creates a buffer from the value that might get
2583 * read with multiple read calls. When the attribute has been read
2584 * completely, no further read calls are possible until the file is opened
2585 * again.
2586 *
2587 * All attributes contain a text representation of a numeric value
2588 * that are accessed with the get() and set() functions.
2589 */
2590#define DEFINE_SIMPLE_ATTRIBUTE(__fops, __get, __set, __fmt) \
2591static int __fops ## _open(struct inode *inode, struct file *file) \
2592{ \
2593 __simple_attr_check_format(__fmt, 0ull); \
2594 return simple_attr_open(inode, file, __get, __set, __fmt); \
2595} \
Alexey Dobriyan828c0952009-10-01 15:43:56 -07002596static const struct file_operations __fops = { \
Arnd Bergmannacaefc22005-05-18 14:40:59 +02002597 .owner = THIS_MODULE, \
2598 .open = __fops ## _open, \
Christoph Hellwig74bedc42008-02-08 04:20:28 -08002599 .release = simple_attr_release, \
Arnd Bergmannacaefc22005-05-18 14:40:59 +02002600 .read = simple_attr_read, \
2601 .write = simple_attr_write, \
Arnd Bergmann1ec55842010-08-15 21:50:52 +02002602 .llseek = generic_file_llseek, \
Arnd Bergmannacaefc22005-05-18 14:40:59 +02002603};
2604
2605static inline void __attribute__((format(printf, 1, 2)))
2606__simple_attr_check_format(const char *fmt, ...)
2607{
2608 /* don't do anything, just let the compiler check the arguments; */
2609}
2610
2611int simple_attr_open(struct inode *inode, struct file *file,
Christoph Hellwig8b88b092008-02-08 04:20:26 -08002612 int (*get)(void *, u64 *), int (*set)(void *, u64),
Arnd Bergmannacaefc22005-05-18 14:40:59 +02002613 const char *fmt);
Christoph Hellwig74bedc42008-02-08 04:20:28 -08002614int simple_attr_release(struct inode *inode, struct file *file);
Arnd Bergmannacaefc22005-05-18 14:40:59 +02002615ssize_t simple_attr_read(struct file *file, char __user *buf,
2616 size_t len, loff_t *ppos);
2617ssize_t simple_attr_write(struct file *file, const char __user *buf,
2618 size_t len, loff_t *ppos);
2619
Alexey Dobriyan4be28542007-10-16 23:26:21 -07002620struct ctl_table;
Alexey Dobriyan8d65af72009-09-23 15:57:19 -07002621int proc_nr_files(struct ctl_table *table, int write,
Adrian Bunk62239ac2007-07-17 04:03:45 -07002622 void __user *buffer, size_t *lenp, loff_t *ppos);
Christoph Hellwig312d3ca2010-10-10 05:36:23 -04002623int proc_nr_dentry(struct ctl_table *table, int write,
2624 void __user *buffer, size_t *lenp, loff_t *ppos);
Dave Chinnercffbc8a2010-10-23 05:03:02 -04002625int proc_nr_inodes(struct ctl_table *table, int write,
2626 void __user *buffer, size_t *lenp, loff_t *ppos);
Tetsuo Handa38e23c92009-04-09 20:17:52 +09002627int __init get_filesystem_list(char *buf);
Adrian Bunk62239ac2007-07-17 04:03:45 -07002628
Namhyung Kim3cd90ea2011-02-01 15:52:46 -08002629#define __FMODE_EXEC ((__force int) FMODE_EXEC)
Namhyung Kim1a44bc82011-02-01 15:52:46 -08002630#define __FMODE_NONOTIFY ((__force int) FMODE_NONOTIFY)
2631
Al Viro6d125522009-12-24 06:58:56 -05002632#define ACC_MODE(x) ("\004\002\006\006"[(x)&O_ACCMODE])
Eric Parisecf081d2009-12-17 21:24:25 -05002633#define OPEN_FMODE(flag) ((__force fmode_t)(((flag + 1) & O_ACCMODE) | \
Namhyung Kim1a44bc82011-02-01 15:52:46 -08002634 (flag & __FMODE_NONOTIFY)))
Al Viro53009902009-12-19 10:15:07 -05002635
Andi Kleen69b45732011-05-28 08:25:51 -07002636static inline int is_sxid(mode_t mode)
2637{
2638 return (mode & S_ISUID) || ((mode & S_ISGID) && (mode & S_IXGRP));
2639}
2640
2641static inline void inode_has_no_xattr(struct inode *inode)
2642{
Al Viro9e1f1de2011-06-03 18:24:58 -04002643 if (!is_sxid(inode->i_mode) && (inode->i_sb->s_flags & MS_NOSEC))
Andi Kleen69b45732011-05-28 08:25:51 -07002644 inode->i_flags |= S_NOSEC;
2645}
2646
Linus Torvalds1da177e2005-04-16 15:20:36 -07002647#endif /* __KERNEL__ */
2648#endif /* _LINUX_FS_H */