aboutsummaryrefslogtreecommitdiffstats
path: root/fs
diff options
context:
space:
mode:
authorLinus Torvalds <torvalds@linux-foundation.org>2012-03-31 16:42:57 -0400
committerLinus Torvalds <torvalds@linux-foundation.org>2012-03-31 16:42:57 -0400
commit8bb1f229527dee95644e0f8496980bb767c6f620 (patch)
tree511551e9772f11f855bd5b759b6d449da47e8820 /fs
parentf22e08a79f3765fecf060b225a46931c94fb0a92 (diff)
parentc0d0259481cc6ec2a38cad810055e455de35c733 (diff)
Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs
Pull second try at vfs part d#2 from Al Viro: "Miklos' first series (with do_lookup() rewrite split into edible chunks) + assorted bits and pieces. The 'untangling of do_lookup()' series is is a splitup of what used to be a monolithic patch from Miklos, so this series is basically "how do I convince myself that his patch is correct (or find a hole in it)". No holes found and I like the resulting cleanup, so in it went..." Changes from try 1: Fix a boot problem with selinux, and commit messages prettied up a bit. * 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs: (24 commits) vfs: fix out-of-date dentry_unhash() comment vfs: split __lookup_hash untangling do_lookup() - take __lookup_hash()-calling case out of line. untangling do_lookup() - switch to calling __lookup_hash() untangling do_lookup() - merge d_alloc_and_lookup() callers untangling do_lookup() - merge failure exits in !dentry case untangling do_lookup() - massage !dentry case towards __lookup_hash() untangling do_lookup() - get rid of need_reval in !dentry case untangling do_lookup() - eliminate a loop. untangling do_lookup() - expand the area under ->i_mutex untangling do_lookup() - isolate !dentry stuff from the rest of it. vfs: move MAY_EXEC check from __lookup_hash() vfs: don't revalidate just looked up dentry vfs: fix d_need_lookup/d_revalidate order in do_lookup ext3: move headers to fs/ext3/ migrate ext2_fs.h guts to fs/ext2/ext2.h new helper: ext2_image_size() get rid of pointless includes of ext2_fs.h ext2: No longer export ext2_fs.h to user space mtdchar: kill persistently held vfsmount ...
Diffstat (limited to 'fs')
-rw-r--r--fs/aio.c32
-rw-r--r--fs/ext2/ext2.h631
-rw-r--r--fs/ext2/xattr_security.c5
-rw-r--r--fs/ext2/xattr_trusted.c5
-rw-r--r--fs/ext2/xip.c2
-rw-r--r--fs/ext3/acl.c8
-rw-r--r--fs/ext3/balloc.c10
-rw-r--r--fs/ext3/bitmap.c4
-rw-r--r--fs/ext3/dir.c7
-rw-r--r--fs/ext3/ext3.h1322
-rw-r--r--fs/ext3/ext3_jbd.c2
-rw-r--r--fs/ext3/file.c6
-rw-r--r--fs/ext3/fsync.c8
-rw-r--r--fs/ext3/hash.c4
-rw-r--r--fs/ext3/ialloc.c13
-rw-r--r--fs/ext3/inode.c12
-rw-r--r--fs/ext3/ioctl.c7
-rw-r--r--fs/ext3/namei.c14
-rw-r--r--fs/ext3/resize.c5
-rw-r--r--fs/ext3/super.c18
-rw-r--r--fs/ext3/symlink.c4
-rw-r--r--fs/ext3/xattr.c7
-rw-r--r--fs/ext3/xattr_security.c6
-rw-r--r--fs/ext3/xattr_trusted.c6
-rw-r--r--fs/ext3/xattr_user.c5
-rw-r--r--fs/gfs2/file.c1
-rw-r--r--fs/namei.c195
-rw-r--r--fs/ocfs2/ioctl.c2
-rw-r--r--fs/pstore/inode.c26
29 files changed, 2074 insertions, 293 deletions
diff --git a/fs/aio.c b/fs/aio.c
index 4f71627264fd..da887604dfc5 100644
--- a/fs/aio.c
+++ b/fs/aio.c
@@ -305,15 +305,18 @@ out_freectx:
305 return ERR_PTR(err); 305 return ERR_PTR(err);
306} 306}
307 307
308/* aio_cancel_all 308/* kill_ctx
309 * Cancels all outstanding aio requests on an aio context. Used 309 * Cancels all outstanding aio requests on an aio context. Used
310 * when the processes owning a context have all exited to encourage 310 * when the processes owning a context have all exited to encourage
311 * the rapid destruction of the kioctx. 311 * the rapid destruction of the kioctx.
312 */ 312 */
313static void aio_cancel_all(struct kioctx *ctx) 313static void kill_ctx(struct kioctx *ctx)
314{ 314{
315 int (*cancel)(struct kiocb *, struct io_event *); 315 int (*cancel)(struct kiocb *, struct io_event *);
316 struct task_struct *tsk = current;
317 DECLARE_WAITQUEUE(wait, tsk);
316 struct io_event res; 318 struct io_event res;
319
317 spin_lock_irq(&ctx->ctx_lock); 320 spin_lock_irq(&ctx->ctx_lock);
318 ctx->dead = 1; 321 ctx->dead = 1;
319 while (!list_empty(&ctx->active_reqs)) { 322 while (!list_empty(&ctx->active_reqs)) {
@@ -329,15 +332,7 @@ static void aio_cancel_all(struct kioctx *ctx)
329 spin_lock_irq(&ctx->ctx_lock); 332 spin_lock_irq(&ctx->ctx_lock);
330 } 333 }
331 } 334 }
332 spin_unlock_irq(&ctx->ctx_lock);
333}
334
335static void wait_for_all_aios(struct kioctx *ctx)
336{
337 struct task_struct *tsk = current;
338 DECLARE_WAITQUEUE(wait, tsk);
339 335
340 spin_lock_irq(&ctx->ctx_lock);
341 if (!ctx->reqs_active) 336 if (!ctx->reqs_active)
342 goto out; 337 goto out;
343 338
@@ -387,9 +382,7 @@ void exit_aio(struct mm_struct *mm)
387 ctx = hlist_entry(mm->ioctx_list.first, struct kioctx, list); 382 ctx = hlist_entry(mm->ioctx_list.first, struct kioctx, list);
388 hlist_del_rcu(&ctx->list); 383 hlist_del_rcu(&ctx->list);
389 384
390 aio_cancel_all(ctx); 385 kill_ctx(ctx);
391
392 wait_for_all_aios(ctx);
393 386
394 if (1 != atomic_read(&ctx->users)) 387 if (1 != atomic_read(&ctx->users))
395 printk(KERN_DEBUG 388 printk(KERN_DEBUG
@@ -1269,8 +1262,7 @@ static void io_destroy(struct kioctx *ioctx)
1269 if (likely(!was_dead)) 1262 if (likely(!was_dead))
1270 put_ioctx(ioctx); /* twice for the list */ 1263 put_ioctx(ioctx); /* twice for the list */
1271 1264
1272 aio_cancel_all(ioctx); 1265 kill_ctx(ioctx);
1273 wait_for_all_aios(ioctx);
1274 1266
1275 /* 1267 /*
1276 * Wake up any waiters. The setting of ctx->dead must be seen 1268 * Wake up any waiters. The setting of ctx->dead must be seen
@@ -1278,7 +1270,6 @@ static void io_destroy(struct kioctx *ioctx)
1278 * locking done by the above calls to ensure this consistency. 1270 * locking done by the above calls to ensure this consistency.
1279 */ 1271 */
1280 wake_up_all(&ioctx->wait); 1272 wake_up_all(&ioctx->wait);
1281 put_ioctx(ioctx); /* once for the lookup */
1282} 1273}
1283 1274
1284/* sys_io_setup: 1275/* sys_io_setup:
@@ -1315,11 +1306,9 @@ SYSCALL_DEFINE2(io_setup, unsigned, nr_events, aio_context_t __user *, ctxp)
1315 ret = PTR_ERR(ioctx); 1306 ret = PTR_ERR(ioctx);
1316 if (!IS_ERR(ioctx)) { 1307 if (!IS_ERR(ioctx)) {
1317 ret = put_user(ioctx->user_id, ctxp); 1308 ret = put_user(ioctx->user_id, ctxp);
1318 if (!ret) { 1309 if (ret)
1319 put_ioctx(ioctx); 1310 io_destroy(ioctx);
1320 return 0; 1311 put_ioctx(ioctx);
1321 }
1322 io_destroy(ioctx);
1323 } 1312 }
1324 1313
1325out: 1314out:
@@ -1337,6 +1326,7 @@ SYSCALL_DEFINE1(io_destroy, aio_context_t, ctx)
1337 struct kioctx *ioctx = lookup_ioctx(ctx); 1326 struct kioctx *ioctx = lookup_ioctx(ctx);
1338 if (likely(NULL != ioctx)) { 1327 if (likely(NULL != ioctx)) {
1339 io_destroy(ioctx); 1328 io_destroy(ioctx);
1329 put_ioctx(ioctx);
1340 return 0; 1330 return 0;
1341 } 1331 }
1342 pr_debug("EINVAL: io_destroy: invalid context id\n"); 1332 pr_debug("EINVAL: io_destroy: invalid context id\n");
diff --git a/fs/ext2/ext2.h b/fs/ext2/ext2.h
index 75ad433c6691..0b2b4db5bdcd 100644
--- a/fs/ext2/ext2.h
+++ b/fs/ext2/ext2.h
@@ -1,5 +1,636 @@
1/*
2 * Copyright (C) 1992, 1993, 1994, 1995
3 * Remy Card (card@masi.ibp.fr)
4 * Laboratoire MASI - Institut Blaise Pascal
5 * Universite Pierre et Marie Curie (Paris VI)
6 *
7 * from
8 *
9 * linux/include/linux/minix_fs.h
10 *
11 * Copyright (C) 1991, 1992 Linus Torvalds
12 */
1#include <linux/fs.h> 13#include <linux/fs.h>
2#include <linux/ext2_fs.h> 14#include <linux/ext2_fs.h>
15#include <linux/blockgroup_lock.h>
16#include <linux/percpu_counter.h>
17#include <linux/rbtree.h>
18
19/* XXX Here for now... not interested in restructing headers JUST now */
20
21/* data type for block offset of block group */
22typedef int ext2_grpblk_t;
23
24/* data type for filesystem-wide blocks number */
25typedef unsigned long ext2_fsblk_t;
26
27#define E2FSBLK "%lu"
28
29struct ext2_reserve_window {
30 ext2_fsblk_t _rsv_start; /* First byte reserved */
31 ext2_fsblk_t _rsv_end; /* Last byte reserved or 0 */
32};
33
34struct ext2_reserve_window_node {
35 struct rb_node rsv_node;
36 __u32 rsv_goal_size;
37 __u32 rsv_alloc_hit;
38 struct ext2_reserve_window rsv_window;
39};
40
41struct ext2_block_alloc_info {
42 /* information about reservation window */
43 struct ext2_reserve_window_node rsv_window_node;
44 /*
45 * was i_next_alloc_block in ext2_inode_info
46 * is the logical (file-relative) number of the
47 * most-recently-allocated block in this file.
48 * We use this for detecting linearly ascending allocation requests.
49 */
50 __u32 last_alloc_logical_block;
51 /*
52 * Was i_next_alloc_goal in ext2_inode_info
53 * is the *physical* companion to i_next_alloc_block.
54 * it the the physical block number of the block which was most-recentl
55 * allocated to this file. This give us the goal (target) for the next
56 * allocation when we detect linearly ascending requests.
57 */
58 ext2_fsblk_t last_alloc_physical_block;
59};
60
61#define rsv_start rsv_window._rsv_start
62#define rsv_end rsv_window._rsv_end
63
64/*
65 * second extended-fs super-block data in memory
66 */
67struct ext2_sb_info {
68 unsigned long s_frag_size; /* Size of a fragment in bytes */
69 unsigned long s_frags_per_block;/* Number of fragments per block */
70 unsigned long s_inodes_per_block;/* Number of inodes per block */
71 unsigned long s_frags_per_group;/* Number of fragments in a group */
72 unsigned long s_blocks_per_group;/* Number of blocks in a group */
73 unsigned long s_inodes_per_group;/* Number of inodes in a group */
74 unsigned long s_itb_per_group; /* Number of inode table blocks per group */
75 unsigned long s_gdb_count; /* Number of group descriptor blocks */
76 unsigned long s_desc_per_block; /* Number of group descriptors per block */
77 unsigned long s_groups_count; /* Number of groups in the fs */
78 unsigned long s_overhead_last; /* Last calculated overhead */
79 unsigned long s_blocks_last; /* Last seen block count */
80 struct buffer_head * s_sbh; /* Buffer containing the super block */
81 struct ext2_super_block * s_es; /* Pointer to the super block in the buffer */
82 struct buffer_head ** s_group_desc;
83 unsigned long s_mount_opt;
84 unsigned long s_sb_block;
85 uid_t s_resuid;
86 gid_t s_resgid;
87 unsigned short s_mount_state;
88 unsigned short s_pad;
89 int s_addr_per_block_bits;
90 int s_desc_per_block_bits;
91 int s_inode_size;
92 int s_first_ino;
93 spinlock_t s_next_gen_lock;
94 u32 s_next_generation;
95 unsigned long s_dir_count;
96 u8 *s_debts;
97 struct percpu_counter s_freeblocks_counter;
98 struct percpu_counter s_freeinodes_counter;
99 struct percpu_counter s_dirs_counter;
100 struct blockgroup_lock *s_blockgroup_lock;
101 /* root of the per fs reservation window tree */
102 spinlock_t s_rsv_window_lock;
103 struct rb_root s_rsv_window_root;
104 struct ext2_reserve_window_node s_rsv_window_head;
105 /*
106 * s_lock protects against concurrent modifications of s_mount_state,
107 * s_blocks_last, s_overhead_last and the content of superblock's
108 * buffer pointed to by sbi->s_es.
109 *
110 * Note: It is used in ext2_show_options() to provide a consistent view
111 * of the mount options.
112 */
113 spinlock_t s_lock;
114};
115
116static inline spinlock_t *
117sb_bgl_lock(struct ext2_sb_info *sbi, unsigned int block_group)
118{
119 return bgl_lock_ptr(sbi->s_blockgroup_lock, block_group);
120}
121
122/*
123 * Define EXT2FS_DEBUG to produce debug messages
124 */
125#undef EXT2FS_DEBUG
126
127/*
128 * Define EXT2_RESERVATION to reserve data blocks for expanding files
129 */
130#define EXT2_DEFAULT_RESERVE_BLOCKS 8
131/*max window size: 1024(direct blocks) + 3([t,d]indirect blocks) */
132#define EXT2_MAX_RESERVE_BLOCKS 1027
133#define EXT2_RESERVE_WINDOW_NOT_ALLOCATED 0
134/*
135 * The second extended file system version
136 */
137#define EXT2FS_DATE "95/08/09"
138#define EXT2FS_VERSION "0.5b"
139
140/*
141 * Debug code
142 */
143#ifdef EXT2FS_DEBUG
144# define ext2_debug(f, a...) { \
145 printk ("EXT2-fs DEBUG (%s, %d): %s:", \
146 __FILE__, __LINE__, __func__); \
147 printk (f, ## a); \
148 }
149#else
150# define ext2_debug(f, a...) /**/
151#endif
152
153/*
154 * Special inode numbers
155 */
156#define EXT2_BAD_INO 1 /* Bad blocks inode */
157#define EXT2_ROOT_INO 2 /* Root inode */
158#define EXT2_BOOT_LOADER_INO 5 /* Boot loader inode */
159#define EXT2_UNDEL_DIR_INO 6 /* Undelete directory inode */
160
161/* First non-reserved inode for old ext2 filesystems */
162#define EXT2_GOOD_OLD_FIRST_INO 11
163
164static inline struct ext2_sb_info *EXT2_SB(struct super_block *sb)
165{
166 return sb->s_fs_info;
167}
168
169/*
170 * Macro-instructions used to manage several block sizes
171 */
172#define EXT2_MIN_BLOCK_SIZE 1024
173#define EXT2_MAX_BLOCK_SIZE 4096
174#define EXT2_MIN_BLOCK_LOG_SIZE 10
175#define EXT2_BLOCK_SIZE(s) ((s)->s_blocksize)
176#define EXT2_ADDR_PER_BLOCK(s) (EXT2_BLOCK_SIZE(s) / sizeof (__u32))
177#define EXT2_BLOCK_SIZE_BITS(s) ((s)->s_blocksize_bits)
178#define EXT2_ADDR_PER_BLOCK_BITS(s) (EXT2_SB(s)->s_addr_per_block_bits)
179#define EXT2_INODE_SIZE(s) (EXT2_SB(s)->s_inode_size)
180#define EXT2_FIRST_INO(s) (EXT2_SB(s)->s_first_ino)
181
182/*
183 * Macro-instructions used to manage fragments
184 */
185#define EXT2_MIN_FRAG_SIZE 1024
186#define EXT2_MAX_FRAG_SIZE 4096
187#define EXT2_MIN_FRAG_LOG_SIZE 10
188#define EXT2_FRAG_SIZE(s) (EXT2_SB(s)->s_frag_size)
189#define EXT2_FRAGS_PER_BLOCK(s) (EXT2_SB(s)->s_frags_per_block)
190
191/*
192 * Structure of a blocks group descriptor
193 */
194struct ext2_group_desc
195{
196 __le32 bg_block_bitmap; /* Blocks bitmap block */
197 __le32 bg_inode_bitmap; /* Inodes bitmap block */
198 __le32 bg_inode_table; /* Inodes table block */
199 __le16 bg_free_blocks_count; /* Free blocks count */
200 __le16 bg_free_inodes_count; /* Free inodes count */
201 __le16 bg_used_dirs_count; /* Directories count */
202 __le16 bg_pad;
203 __le32 bg_reserved[3];
204};
205
206/*
207 * Macro-instructions used to manage group descriptors
208 */
209#define EXT2_BLOCKS_PER_GROUP(s) (EXT2_SB(s)->s_blocks_per_group)
210#define EXT2_DESC_PER_BLOCK(s) (EXT2_SB(s)->s_desc_per_block)
211#define EXT2_INODES_PER_GROUP(s) (EXT2_SB(s)->s_inodes_per_group)
212#define EXT2_DESC_PER_BLOCK_BITS(s) (EXT2_SB(s)->s_desc_per_block_bits)
213
214/*
215 * Constants relative to the data blocks
216 */
217#define EXT2_NDIR_BLOCKS 12
218#define EXT2_IND_BLOCK EXT2_NDIR_BLOCKS
219#define EXT2_DIND_BLOCK (EXT2_IND_BLOCK + 1)
220#define EXT2_TIND_BLOCK (EXT2_DIND_BLOCK + 1)
221#define EXT2_N_BLOCKS (EXT2_TIND_BLOCK + 1)
222
223/*
224 * Inode flags (GETFLAGS/SETFLAGS)
225 */
226#define EXT2_SECRM_FL FS_SECRM_FL /* Secure deletion */
227#define EXT2_UNRM_FL FS_UNRM_FL /* Undelete */
228#define EXT2_COMPR_FL FS_COMPR_FL /* Compress file */
229#define EXT2_SYNC_FL FS_SYNC_FL /* Synchronous updates */
230#define EXT2_IMMUTABLE_FL FS_IMMUTABLE_FL /* Immutable file */
231#define EXT2_APPEND_FL FS_APPEND_FL /* writes to file may only append */
232#define EXT2_NODUMP_FL FS_NODUMP_FL /* do not dump file */
233#define EXT2_NOATIME_FL FS_NOATIME_FL /* do not update atime */
234/* Reserved for compression usage... */
235#define EXT2_DIRTY_FL FS_DIRTY_FL
236#define EXT2_COMPRBLK_FL FS_COMPRBLK_FL /* One or more compressed clusters */
237#define EXT2_NOCOMP_FL FS_NOCOMP_FL /* Don't compress */
238#define EXT2_ECOMPR_FL FS_ECOMPR_FL /* Compression error */
239/* End compression flags --- maybe not all used */
240#define EXT2_BTREE_FL FS_BTREE_FL /* btree format dir */
241#define EXT2_INDEX_FL FS_INDEX_FL /* hash-indexed directory */
242#define EXT2_IMAGIC_FL FS_IMAGIC_FL /* AFS directory */
243#define EXT2_JOURNAL_DATA_FL FS_JOURNAL_DATA_FL /* Reserved for ext3 */
244#define EXT2_NOTAIL_FL FS_NOTAIL_FL /* file tail should not be merged */
245#define EXT2_DIRSYNC_FL FS_DIRSYNC_FL /* dirsync behaviour (directories only) */
246#define EXT2_TOPDIR_FL FS_TOPDIR_FL /* Top of directory hierarchies*/
247#define EXT2_RESERVED_FL FS_RESERVED_FL /* reserved for ext2 lib */
248
249#define EXT2_FL_USER_VISIBLE FS_FL_USER_VISIBLE /* User visible flags */
250#define EXT2_FL_USER_MODIFIABLE FS_FL_USER_MODIFIABLE /* User modifiable flags */
251
252/* Flags that should be inherited by new inodes from their parent. */
253#define EXT2_FL_INHERITED (EXT2_SECRM_FL | EXT2_UNRM_FL | EXT2_COMPR_FL |\
254 EXT2_SYNC_FL | EXT2_NODUMP_FL |\
255 EXT2_NOATIME_FL | EXT2_COMPRBLK_FL |\
256 EXT2_NOCOMP_FL | EXT2_JOURNAL_DATA_FL |\
257 EXT2_NOTAIL_FL | EXT2_DIRSYNC_FL)
258
259/* Flags that are appropriate for regular files (all but dir-specific ones). */
260#define EXT2_REG_FLMASK (~(EXT2_DIRSYNC_FL | EXT2_TOPDIR_FL))
261
262/* Flags that are appropriate for non-directories/regular files. */
263#define EXT2_OTHER_FLMASK (EXT2_NODUMP_FL | EXT2_NOATIME_FL)
264
265/* Mask out flags that are inappropriate for the given type of inode. */
266static inline __u32 ext2_mask_flags(umode_t mode, __u32 flags)
267{
268 if (S_ISDIR(mode))
269 return flags;
270 else if (S_ISREG(mode))
271 return flags & EXT2_REG_FLMASK;
272 else
273 return flags & EXT2_OTHER_FLMASK;
274}
275
276/*
277 * ioctl commands
278 */
279#define EXT2_IOC_GETFLAGS FS_IOC_GETFLAGS
280#define EXT2_IOC_SETFLAGS FS_IOC_SETFLAGS
281#define EXT2_IOC_GETVERSION FS_IOC_GETVERSION
282#define EXT2_IOC_SETVERSION FS_IOC_SETVERSION
283#define EXT2_IOC_GETRSVSZ _IOR('f', 5, long)
284#define EXT2_IOC_SETRSVSZ _IOW('f', 6, long)
285
286/*
287 * ioctl commands in 32 bit emulation
288 */
289#define EXT2_IOC32_GETFLAGS FS_IOC32_GETFLAGS
290#define EXT2_IOC32_SETFLAGS FS_IOC32_SETFLAGS
291#define EXT2_IOC32_GETVERSION FS_IOC32_GETVERSION
292#define EXT2_IOC32_SETVERSION FS_IOC32_SETVERSION
293
294/*
295 * Structure of an inode on the disk
296 */
297struct ext2_inode {
298 __le16 i_mode; /* File mode */
299 __le16 i_uid; /* Low 16 bits of Owner Uid */
300 __le32 i_size; /* Size in bytes */
301 __le32 i_atime; /* Access time */
302 __le32 i_ctime; /* Creation time */
303 __le32 i_mtime; /* Modification time */
304 __le32 i_dtime; /* Deletion Time */
305 __le16 i_gid; /* Low 16 bits of Group Id */
306 __le16 i_links_count; /* Links count */
307 __le32 i_blocks; /* Blocks count */
308 __le32 i_flags; /* File flags */
309 union {
310 struct {
311 __le32 l_i_reserved1;
312 } linux1;
313 struct {
314 __le32 h_i_translator;
315 } hurd1;
316 struct {
317 __le32 m_i_reserved1;
318 } masix1;
319 } osd1; /* OS dependent 1 */
320 __le32 i_block[EXT2_N_BLOCKS];/* Pointers to blocks */
321 __le32 i_generation; /* File version (for NFS) */
322 __le32 i_file_acl; /* File ACL */
323 __le32 i_dir_acl; /* Directory ACL */
324 __le32 i_faddr; /* Fragment address */
325 union {
326 struct {
327 __u8 l_i_frag; /* Fragment number */
328 __u8 l_i_fsize; /* Fragment size */
329 __u16 i_pad1;
330 __le16 l_i_uid_high; /* these 2 fields */
331 __le16 l_i_gid_high; /* were reserved2[0] */
332 __u32 l_i_reserved2;
333 } linux2;
334 struct {
335 __u8 h_i_frag; /* Fragment number */
336 __u8 h_i_fsize; /* Fragment size */
337 __le16 h_i_mode_high;
338 __le16 h_i_uid_high;
339 __le16 h_i_gid_high;
340 __le32 h_i_author;
341 } hurd2;
342 struct {
343 __u8 m_i_frag; /* Fragment number */
344 __u8 m_i_fsize; /* Fragment size */
345 __u16 m_pad1;
346 __u32 m_i_reserved2[2];
347 } masix2;
348 } osd2; /* OS dependent 2 */
349};
350
351#define i_size_high i_dir_acl
352
353#define i_reserved1 osd1.linux1.l_i_reserved1
354#define i_frag osd2.linux2.l_i_frag
355#define i_fsize osd2.linux2.l_i_fsize
356#define i_uid_low i_uid
357#define i_gid_low i_gid
358#define i_uid_high osd2.linux2.l_i_uid_high
359#define i_gid_high osd2.linux2.l_i_gid_high
360#define i_reserved2 osd2.linux2.l_i_reserved2
361
362/*
363 * File system states
364 */
365#define EXT2_VALID_FS 0x0001 /* Unmounted cleanly */
366#define EXT2_ERROR_FS 0x0002 /* Errors detected */
367
368/*
369 * Mount flags
370 */
371#define EXT2_MOUNT_CHECK 0x000001 /* Do mount-time checks */
372#define EXT2_MOUNT_OLDALLOC 0x000002 /* Don't use the new Orlov allocator */
373#define EXT2_MOUNT_GRPID 0x000004 /* Create files with directory's group */
374#define EXT2_MOUNT_DEBUG 0x000008 /* Some debugging messages */
375#define EXT2_MOUNT_ERRORS_CONT 0x000010 /* Continue on errors */
376#define EXT2_MOUNT_ERRORS_RO 0x000020 /* Remount fs ro on errors */
377#define EXT2_MOUNT_ERRORS_PANIC 0x000040 /* Panic on errors */
378#define EXT2_MOUNT_MINIX_DF 0x000080 /* Mimics the Minix statfs */
379#define EXT2_MOUNT_NOBH 0x000100 /* No buffer_heads */
380#define EXT2_MOUNT_NO_UID32 0x000200 /* Disable 32-bit UIDs */
381#define EXT2_MOUNT_XATTR_USER 0x004000 /* Extended user attributes */
382#define EXT2_MOUNT_POSIX_ACL 0x008000 /* POSIX Access Control Lists */
383#define EXT2_MOUNT_XIP 0x010000 /* Execute in place */
384#define EXT2_MOUNT_USRQUOTA 0x020000 /* user quota */
385#define EXT2_MOUNT_GRPQUOTA 0x040000 /* group quota */
386#define EXT2_MOUNT_RESERVATION 0x080000 /* Preallocation */
387
388
389#define clear_opt(o, opt) o &= ~EXT2_MOUNT_##opt
390#define set_opt(o, opt) o |= EXT2_MOUNT_##opt
391#define test_opt(sb, opt) (EXT2_SB(sb)->s_mount_opt & \
392 EXT2_MOUNT_##opt)
393/*
394 * Maximal mount counts between two filesystem checks
395 */
396#define EXT2_DFL_MAX_MNT_COUNT 20 /* Allow 20 mounts */
397#define EXT2_DFL_CHECKINTERVAL 0 /* Don't use interval check */
398
399/*
400 * Behaviour when detecting errors
401 */
402#define EXT2_ERRORS_CONTINUE 1 /* Continue execution */
403#define EXT2_ERRORS_RO 2 /* Remount fs read-only */
404#define EXT2_ERRORS_PANIC 3 /* Panic */
405#define EXT2_ERRORS_DEFAULT EXT2_ERRORS_CONTINUE
406
407/*
408 * Structure of the super block
409 */
410struct ext2_super_block {
411 __le32 s_inodes_count; /* Inodes count */
412 __le32 s_blocks_count; /* Blocks count */
413 __le32 s_r_blocks_count; /* Reserved blocks count */
414 __le32 s_free_blocks_count; /* Free blocks count */
415 __le32 s_free_inodes_count; /* Free inodes count */
416 __le32 s_first_data_block; /* First Data Block */
417 __le32 s_log_block_size; /* Block size */
418 __le32 s_log_frag_size; /* Fragment size */
419 __le32 s_blocks_per_group; /* # Blocks per group */
420 __le32 s_frags_per_group; /* # Fragments per group */
421 __le32 s_inodes_per_group; /* # Inodes per group */
422 __le32 s_mtime; /* Mount time */
423 __le32 s_wtime; /* Write time */
424 __le16 s_mnt_count; /* Mount count */
425 __le16 s_max_mnt_count; /* Maximal mount count */
426 __le16 s_magic; /* Magic signature */
427 __le16 s_state; /* File system state */
428 __le16 s_errors; /* Behaviour when detecting errors */
429 __le16 s_minor_rev_level; /* minor revision level */
430 __le32 s_lastcheck; /* time of last check */
431 __le32 s_checkinterval; /* max. time between checks */
432 __le32 s_creator_os; /* OS */
433 __le32 s_rev_level; /* Revision level */
434 __le16 s_def_resuid; /* Default uid for reserved blocks */
435 __le16 s_def_resgid; /* Default gid for reserved blocks */
436 /*
437 * These fields are for EXT2_DYNAMIC_REV superblocks only.
438 *
439 * Note: the difference between the compatible feature set and
440 * the incompatible feature set is that if there is a bit set
441 * in the incompatible feature set that the kernel doesn't
442 * know about, it should refuse to mount the filesystem.
443 *
444 * e2fsck's requirements are more strict; if it doesn't know
445 * about a feature in either the compatible or incompatible
446 * feature set, it must abort and not try to meddle with
447 * things it doesn't understand...
448 */
449 __le32 s_first_ino; /* First non-reserved inode */
450 __le16 s_inode_size; /* size of inode structure */
451 __le16 s_block_group_nr; /* block group # of this superblock */
452 __le32 s_feature_compat; /* compatible feature set */
453 __le32 s_feature_incompat; /* incompatible feature set */
454 __le32 s_feature_ro_compat; /* readonly-compatible feature set */
455 __u8 s_uuid[16]; /* 128-bit uuid for volume */
456 char s_volume_name[16]; /* volume name */
457 char s_last_mounted[64]; /* directory where last mounted */
458 __le32 s_algorithm_usage_bitmap; /* For compression */
459 /*
460 * Performance hints. Directory preallocation should only
461 * happen if the EXT2_COMPAT_PREALLOC flag is on.
462 */
463 __u8 s_prealloc_blocks; /* Nr of blocks to try to preallocate*/
464 __u8 s_prealloc_dir_blocks; /* Nr to preallocate for dirs */
465 __u16 s_padding1;
466 /*
467 * Journaling support valid if EXT3_FEATURE_COMPAT_HAS_JOURNAL set.
468 */
469 __u8 s_journal_uuid[16]; /* uuid of journal superblock */
470 __u32 s_journal_inum; /* inode number of journal file */
471 __u32 s_journal_dev; /* device number of journal file */
472 __u32 s_last_orphan; /* start of list of inodes to delete */
473 __u32 s_hash_seed[4]; /* HTREE hash seed */
474 __u8 s_def_hash_version; /* Default hash version to use */
475 __u8 s_reserved_char_pad;
476 __u16 s_reserved_word_pad;
477 __le32 s_default_mount_opts;
478 __le32 s_first_meta_bg; /* First metablock block group */
479 __u32 s_reserved[190]; /* Padding to the end of the block */
480};
481
482/*
483 * Codes for operating systems
484 */
485#define EXT2_OS_LINUX 0
486#define EXT2_OS_HURD 1
487#define EXT2_OS_MASIX 2
488#define EXT2_OS_FREEBSD 3
489#define EXT2_OS_LITES 4
490
491/*
492 * Revision levels
493 */
494#define EXT2_GOOD_OLD_REV 0 /* The good old (original) format */
495#define EXT2_DYNAMIC_REV 1 /* V2 format w/ dynamic inode sizes */
496
497#define EXT2_CURRENT_REV EXT2_GOOD_OLD_REV
498#define EXT2_MAX_SUPP_REV EXT2_DYNAMIC_REV
499
500#define EXT2_GOOD_OLD_INODE_SIZE 128
501
502/*
503 * Feature set definitions
504 */
505
506#define EXT2_HAS_COMPAT_FEATURE(sb,mask) \
507 ( EXT2_SB(sb)->s_es->s_feature_compat & cpu_to_le32(mask) )
508#define EXT2_HAS_RO_COMPAT_FEATURE(sb,mask) \
509 ( EXT2_SB(sb)->s_es->s_feature_ro_compat & cpu_to_le32(mask) )
510#define EXT2_HAS_INCOMPAT_FEATURE(sb,mask) \
511 ( EXT2_SB(sb)->s_es->s_feature_incompat & cpu_to_le32(mask) )
512#define EXT2_SET_COMPAT_FEATURE(sb,mask) \
513 EXT2_SB(sb)->s_es->s_feature_compat |= cpu_to_le32(mask)
514#define EXT2_SET_RO_COMPAT_FEATURE(sb,mask) \
515 EXT2_SB(sb)->s_es->s_feature_ro_compat |= cpu_to_le32(mask)
516#define EXT2_SET_INCOMPAT_FEATURE(sb,mask) \
517 EXT2_SB(sb)->s_es->s_feature_incompat |= cpu_to_le32(mask)
518#define EXT2_CLEAR_COMPAT_FEATURE(sb,mask) \
519 EXT2_SB(sb)->s_es->s_feature_compat &= ~cpu_to_le32(mask)
520#define EXT2_CLEAR_RO_COMPAT_FEATURE(sb,mask) \
521 EXT2_SB(sb)->s_es->s_feature_ro_compat &= ~cpu_to_le32(mask)
522#define EXT2_CLEAR_INCOMPAT_FEATURE(sb,mask) \
523 EXT2_SB(sb)->s_es->s_feature_incompat &= ~cpu_to_le32(mask)
524
525#define EXT2_FEATURE_COMPAT_DIR_PREALLOC 0x0001
526#define EXT2_FEATURE_COMPAT_IMAGIC_INODES 0x0002
527#define EXT3_FEATURE_COMPAT_HAS_JOURNAL 0x0004
528#define EXT2_FEATURE_COMPAT_EXT_ATTR 0x0008
529#define EXT2_FEATURE_COMPAT_RESIZE_INO 0x0010
530#define EXT2_FEATURE_COMPAT_DIR_INDEX 0x0020
531#define EXT2_FEATURE_COMPAT_ANY 0xffffffff
532
533#define EXT2_FEATURE_RO_COMPAT_SPARSE_SUPER 0x0001
534#define EXT2_FEATURE_RO_COMPAT_LARGE_FILE 0x0002
535#define EXT2_FEATURE_RO_COMPAT_BTREE_DIR 0x0004
536#define EXT2_FEATURE_RO_COMPAT_ANY 0xffffffff
537
538#define EXT2_FEATURE_INCOMPAT_COMPRESSION 0x0001
539#define EXT2_FEATURE_INCOMPAT_FILETYPE 0x0002
540#define EXT3_FEATURE_INCOMPAT_RECOVER 0x0004
541#define EXT3_FEATURE_INCOMPAT_JOURNAL_DEV 0x0008
542#define EXT2_FEATURE_INCOMPAT_META_BG 0x0010
543#define EXT2_FEATURE_INCOMPAT_ANY 0xffffffff
544
545#define EXT2_FEATURE_COMPAT_SUPP EXT2_FEATURE_COMPAT_EXT_ATTR
546#define EXT2_FEATURE_INCOMPAT_SUPP (EXT2_FEATURE_INCOMPAT_FILETYPE| \
547 EXT2_FEATURE_INCOMPAT_META_BG)
548#define EXT2_FEATURE_RO_COMPAT_SUPP (EXT2_FEATURE_RO_COMPAT_SPARSE_SUPER| \
549 EXT2_FEATURE_RO_COMPAT_LARGE_FILE| \
550 EXT2_FEATURE_RO_COMPAT_BTREE_DIR)
551#define EXT2_FEATURE_RO_COMPAT_UNSUPPORTED ~EXT2_FEATURE_RO_COMPAT_SUPP
552#define EXT2_FEATURE_INCOMPAT_UNSUPPORTED ~EXT2_FEATURE_INCOMPAT_SUPP
553
554/*
555 * Default values for user and/or group using reserved blocks
556 */
557#define EXT2_DEF_RESUID 0
558#define EXT2_DEF_RESGID 0
559
560/*
561 * Default mount options
562 */
563#define EXT2_DEFM_DEBUG 0x0001
564#define EXT2_DEFM_BSDGROUPS 0x0002
565#define EXT2_DEFM_XATTR_USER 0x0004
566#define EXT2_DEFM_ACL 0x0008
567#define EXT2_DEFM_UID16 0x0010
568 /* Not used by ext2, but reserved for use by ext3 */
569#define EXT3_DEFM_JMODE 0x0060
570#define EXT3_DEFM_JMODE_DATA 0x0020
571#define EXT3_DEFM_JMODE_ORDERED 0x0040
572#define EXT3_DEFM_JMODE_WBACK 0x0060
573
574/*
575 * Structure of a directory entry
576 */
577
578struct ext2_dir_entry {
579 __le32 inode; /* Inode number */
580 __le16 rec_len; /* Directory entry length */
581 __le16 name_len; /* Name length */
582 char name[]; /* File name, up to EXT2_NAME_LEN */
583};
584
585/*
586 * The new version of the directory entry. Since EXT2 structures are
587 * stored in intel byte order, and the name_len field could never be
588 * bigger than 255 chars, it's safe to reclaim the extra byte for the
589 * file_type field.
590 */
591struct ext2_dir_entry_2 {
592 __le32 inode; /* Inode number */
593 __le16 rec_len; /* Directory entry length */
594 __u8 name_len; /* Name length */
595 __u8 file_type;
596 char name[]; /* File name, up to EXT2_NAME_LEN */
597};
598
599/*
600 * Ext2 directory file types. Only the low 3 bits are used. The
601 * other bits are reserved for now.
602 */
603enum {
604 EXT2_FT_UNKNOWN = 0,
605 EXT2_FT_REG_FILE = 1,
606 EXT2_FT_DIR = 2,
607 EXT2_FT_CHRDEV = 3,
608 EXT2_FT_BLKDEV = 4,
609 EXT2_FT_FIFO = 5,
610 EXT2_FT_SOCK = 6,
611 EXT2_FT_SYMLINK = 7,
612 EXT2_FT_MAX
613};
614
615/*
616 * EXT2_DIR_PAD defines the directory entries boundaries
617 *
618 * NOTE: It must be a multiple of 4
619 */
620#define EXT2_DIR_PAD 4
621#define EXT2_DIR_ROUND (EXT2_DIR_PAD - 1)
622#define EXT2_DIR_REC_LEN(name_len) (((name_len) + 8 + EXT2_DIR_ROUND) & \
623 ~EXT2_DIR_ROUND)
624#define EXT2_MAX_REC_LEN ((1<<16)-1)
625
626static inline void verify_offsets(void)
627{
628#define A(x,y) BUILD_BUG_ON(x != offsetof(struct ext2_super_block, y));
629 A(EXT2_SB_MAGIC_OFFSET, s_magic);
630 A(EXT2_SB_BLOCKS_OFFSET, s_blocks_count);
631 A(EXT2_SB_BSIZE_OFFSET, s_log_block_size);
632#undef A
633}
3 634
4/* 635/*
5 * ext2 mount options 636 * ext2 mount options
diff --git a/fs/ext2/xattr_security.c b/fs/ext2/xattr_security.c
index be7a8d02c9a7..cfedb2cb0d8c 100644
--- a/fs/ext2/xattr_security.c
+++ b/fs/ext2/xattr_security.c
@@ -3,10 +3,7 @@
3 * Handler for storing security labels as extended attributes. 3 * Handler for storing security labels as extended attributes.
4 */ 4 */
5 5
6#include <linux/slab.h> 6#include "ext2.h"
7#include <linux/string.h>
8#include <linux/fs.h>
9#include <linux/ext2_fs.h>
10#include <linux/security.h> 7#include <linux/security.h>
11#include "xattr.h" 8#include "xattr.h"
12 9
diff --git a/fs/ext2/xattr_trusted.c b/fs/ext2/xattr_trusted.c
index 2989467d3595..7e192574c001 100644
--- a/fs/ext2/xattr_trusted.c
+++ b/fs/ext2/xattr_trusted.c
@@ -5,10 +5,7 @@
5 * Copyright (C) 2003 by Andreas Gruenbacher, <a.gruenbacher@computer.org> 5 * Copyright (C) 2003 by Andreas Gruenbacher, <a.gruenbacher@computer.org>
6 */ 6 */
7 7
8#include <linux/string.h> 8#include "ext2.h"
9#include <linux/capability.h>
10#include <linux/fs.h>
11#include <linux/ext2_fs.h>
12#include "xattr.h" 9#include "xattr.h"
13 10
14static size_t 11static size_t
diff --git a/fs/ext2/xip.c b/fs/ext2/xip.c
index 322a56b2dfb1..1c3312858fcf 100644
--- a/fs/ext2/xip.c
+++ b/fs/ext2/xip.c
@@ -9,8 +9,6 @@
9#include <linux/fs.h> 9#include <linux/fs.h>
10#include <linux/genhd.h> 10#include <linux/genhd.h>
11#include <linux/buffer_head.h> 11#include <linux/buffer_head.h>
12#include <linux/ext2_fs_sb.h>
13#include <linux/ext2_fs.h>
14#include <linux/blkdev.h> 12#include <linux/blkdev.h>
15#include "ext2.h" 13#include "ext2.h"
16#include "xip.h" 14#include "xip.h"
diff --git a/fs/ext3/acl.c b/fs/ext3/acl.c
index 3091f62e55b6..c76832c8d192 100644
--- a/fs/ext3/acl.c
+++ b/fs/ext3/acl.c
@@ -4,13 +4,7 @@
4 * Copyright (C) 2001-2003 Andreas Gruenbacher, <agruen@suse.de> 4 * Copyright (C) 2001-2003 Andreas Gruenbacher, <agruen@suse.de>
5 */ 5 */
6 6
7#include <linux/init.h> 7#include "ext3.h"
8#include <linux/sched.h>
9#include <linux/slab.h>
10#include <linux/capability.h>
11#include <linux/fs.h>
12#include <linux/ext3_jbd.h>
13#include <linux/ext3_fs.h>
14#include "xattr.h" 8#include "xattr.h"
15#include "acl.h" 9#include "acl.h"
16 10
diff --git a/fs/ext3/balloc.c b/fs/ext3/balloc.c
index 1e036b79384c..baac1b129fba 100644
--- a/fs/ext3/balloc.c
+++ b/fs/ext3/balloc.c
@@ -11,17 +11,9 @@
11 * David S. Miller (davem@caip.rutgers.edu), 1995 11 * David S. Miller (davem@caip.rutgers.edu), 1995
12 */ 12 */
13 13
14#include <linux/time.h>
15#include <linux/capability.h>
16#include <linux/fs.h>
17#include <linux/slab.h>
18#include <linux/jbd.h>
19#include <linux/ext3_fs.h>
20#include <linux/ext3_jbd.h>
21#include <linux/quotaops.h> 14#include <linux/quotaops.h>
22#include <linux/buffer_head.h>
23#include <linux/blkdev.h> 15#include <linux/blkdev.h>
24#include <trace/events/ext3.h> 16#include "ext3.h"
25 17
26/* 18/*
27 * balloc.c contains the blocks allocation and deallocation routines 19 * balloc.c contains the blocks allocation and deallocation routines
diff --git a/fs/ext3/bitmap.c b/fs/ext3/bitmap.c
index 6afc39d80253..909d13e26560 100644
--- a/fs/ext3/bitmap.c
+++ b/fs/ext3/bitmap.c
@@ -7,9 +7,7 @@
7 * Universite Pierre et Marie Curie (Paris VI) 7 * Universite Pierre et Marie Curie (Paris VI)
8 */ 8 */
9 9
10#include <linux/buffer_head.h> 10#include "ext3.h"
11#include <linux/jbd.h>
12#include <linux/ext3_fs.h>
13 11
14#ifdef EXT3FS_DEBUG 12#ifdef EXT3FS_DEBUG
15 13
diff --git a/fs/ext3/dir.c b/fs/ext3/dir.c
index 34f0a072b935..cc761ad8fa57 100644
--- a/fs/ext3/dir.c
+++ b/fs/ext3/dir.c
@@ -21,12 +21,7 @@
21 * 21 *
22 */ 22 */
23 23
24#include <linux/fs.h> 24#include "ext3.h"
25#include <linux/jbd.h>
26#include <linux/ext3_fs.h>
27#include <linux/buffer_head.h>
28#include <linux/slab.h>
29#include <linux/rbtree.h>
30 25
31static unsigned char ext3_filetype_table[] = { 26static unsigned char ext3_filetype_table[] = {
32 DT_UNKNOWN, DT_REG, DT_DIR, DT_CHR, DT_BLK, DT_FIFO, DT_SOCK, DT_LNK 27 DT_UNKNOWN, DT_REG, DT_DIR, DT_CHR, DT_BLK, DT_FIFO, DT_SOCK, DT_LNK
diff --git a/fs/ext3/ext3.h b/fs/ext3/ext3.h
new file mode 100644
index 000000000000..b6515fd7e56c
--- /dev/null
+++ b/fs/ext3/ext3.h
@@ -0,0 +1,1322 @@
1/*
2 * Written by Stephen C. Tweedie <sct@redhat.com>, 1999
3 *
4 * Copyright 1998--1999 Red Hat corp --- All Rights Reserved
5 *
6 * This file is part of the Linux kernel and is made available under
7 * the terms of the GNU General Public License, version 2, or at your
8 * option, any later version, incorporated herein by reference.
9 *
10 * Copyright (C) 1992, 1993, 1994, 1995
11 * Remy Card (card@masi.ibp.fr)
12 * Laboratoire MASI - Institut Blaise Pascal
13 * Universite Pierre et Marie Curie (Paris VI)
14 *
15 * from
16 *
17 * linux/include/linux/minix_fs.h
18 *
19 * Copyright (C) 1991, 1992 Linus Torvalds
20 */
21
22#include <linux/fs.h>
23#include <linux/jbd.h>
24#include <linux/magic.h>
25#include <linux/bug.h>
26#include <linux/blockgroup_lock.h>
27
28/*
29 * The second extended filesystem constants/structures
30 */
31
32/*
33 * Define EXT3FS_DEBUG to produce debug messages
34 */
35#undef EXT3FS_DEBUG
36
37/*
38 * Define EXT3_RESERVATION to reserve data blocks for expanding files
39 */
40#define EXT3_DEFAULT_RESERVE_BLOCKS 8
41/*max window size: 1024(direct blocks) + 3([t,d]indirect blocks) */
42#define EXT3_MAX_RESERVE_BLOCKS 1027
43#define EXT3_RESERVE_WINDOW_NOT_ALLOCATED 0
44
45/*
46 * Debug code
47 */
48#ifdef EXT3FS_DEBUG
49#define ext3_debug(f, a...) \
50 do { \
51 printk (KERN_DEBUG "EXT3-fs DEBUG (%s, %d): %s:", \
52 __FILE__, __LINE__, __func__); \
53 printk (KERN_DEBUG f, ## a); \
54 } while (0)
55#else
56#define ext3_debug(f, a...) do {} while (0)
57#endif
58
59/*
60 * Special inodes numbers
61 */
62#define EXT3_BAD_INO 1 /* Bad blocks inode */
63#define EXT3_ROOT_INO 2 /* Root inode */
64#define EXT3_BOOT_LOADER_INO 5 /* Boot loader inode */
65#define EXT3_UNDEL_DIR_INO 6 /* Undelete directory inode */
66#define EXT3_RESIZE_INO 7 /* Reserved group descriptors inode */
67#define EXT3_JOURNAL_INO 8 /* Journal inode */
68
69/* First non-reserved inode for old ext3 filesystems */
70#define EXT3_GOOD_OLD_FIRST_INO 11
71
72/*
73 * Maximal count of links to a file
74 */
75#define EXT3_LINK_MAX 32000
76
77/*
78 * Macro-instructions used to manage several block sizes
79 */
80#define EXT3_MIN_BLOCK_SIZE 1024
81#define EXT3_MAX_BLOCK_SIZE 65536
82#define EXT3_MIN_BLOCK_LOG_SIZE 10
83#define EXT3_BLOCK_SIZE(s) ((s)->s_blocksize)
84#define EXT3_ADDR_PER_BLOCK(s) (EXT3_BLOCK_SIZE(s) / sizeof (__u32))
85#define EXT3_BLOCK_SIZE_BITS(s) ((s)->s_blocksize_bits)
86#define EXT3_ADDR_PER_BLOCK_BITS(s) (EXT3_SB(s)->s_addr_per_block_bits)
87#define EXT3_INODE_SIZE(s) (EXT3_SB(s)->s_inode_size)
88#define EXT3_FIRST_INO(s) (EXT3_SB(s)->s_first_ino)
89
90/*
91 * Macro-instructions used to manage fragments
92 */
93#define EXT3_MIN_FRAG_SIZE 1024
94#define EXT3_MAX_FRAG_SIZE 4096
95#define EXT3_MIN_FRAG_LOG_SIZE 10
96#define EXT3_FRAG_SIZE(s) (EXT3_SB(s)->s_frag_size)
97#define EXT3_FRAGS_PER_BLOCK(s) (EXT3_SB(s)->s_frags_per_block)
98
99/*
100 * Structure of a blocks group descriptor
101 */
102struct ext3_group_desc
103{
104 __le32 bg_block_bitmap; /* Blocks bitmap block */
105 __le32 bg_inode_bitmap; /* Inodes bitmap block */
106 __le32 bg_inode_table; /* Inodes table block */
107 __le16 bg_free_blocks_count; /* Free blocks count */
108 __le16 bg_free_inodes_count; /* Free inodes count */
109 __le16 bg_used_dirs_count; /* Directories count */
110 __u16 bg_pad;
111 __le32 bg_reserved[3];
112};
113
114/*
115 * Macro-instructions used to manage group descriptors
116 */
117#define EXT3_BLOCKS_PER_GROUP(s) (EXT3_SB(s)->s_blocks_per_group)
118#define EXT3_DESC_PER_BLOCK(s) (EXT3_SB(s)->s_desc_per_block)
119#define EXT3_INODES_PER_GROUP(s) (EXT3_SB(s)->s_inodes_per_group)
120#define EXT3_DESC_PER_BLOCK_BITS(s) (EXT3_SB(s)->s_desc_per_block_bits)
121
122/*
123 * Constants relative to the data blocks
124 */
125#define EXT3_NDIR_BLOCKS 12
126#define EXT3_IND_BLOCK EXT3_NDIR_BLOCKS
127#define EXT3_DIND_BLOCK (EXT3_IND_BLOCK + 1)
128#define EXT3_TIND_BLOCK (EXT3_DIND_BLOCK + 1)
129#define EXT3_N_BLOCKS (EXT3_TIND_BLOCK + 1)
130
131/*
132 * Inode flags
133 */
134#define EXT3_SECRM_FL 0x00000001 /* Secure deletion */
135#define EXT3_UNRM_FL 0x00000002 /* Undelete */
136#define EXT3_COMPR_FL 0x00000004 /* Compress file */
137#define EXT3_SYNC_FL 0x00000008 /* Synchronous updates */
138#define EXT3_IMMUTABLE_FL 0x00000010 /* Immutable file */
139#define EXT3_APPEND_FL 0x00000020 /* writes to file may only append */
140#define EXT3_NODUMP_FL 0x00000040 /* do not dump file */
141#define EXT3_NOATIME_FL 0x00000080 /* do not update atime */
142/* Reserved for compression usage... */
143#define EXT3_DIRTY_FL 0x00000100
144#define EXT3_COMPRBLK_FL 0x00000200 /* One or more compressed clusters */
145#define EXT3_NOCOMPR_FL 0x00000400 /* Don't compress */
146#define EXT3_ECOMPR_FL 0x00000800 /* Compression error */
147/* End compression flags --- maybe not all used */
148#define EXT3_INDEX_FL 0x00001000 /* hash-indexed directory */
149#define EXT3_IMAGIC_FL 0x00002000 /* AFS directory */
150#define EXT3_JOURNAL_DATA_FL 0x00004000 /* file data should be journaled */
151#define EXT3_NOTAIL_FL 0x00008000 /* file tail should not be merged */
152#define EXT3_DIRSYNC_FL 0x00010000 /* dirsync behaviour (directories only) */
153#define EXT3_TOPDIR_FL 0x00020000 /* Top of directory hierarchies*/
154#define EXT3_RESERVED_FL 0x80000000 /* reserved for ext3 lib */
155
156#define EXT3_FL_USER_VISIBLE 0x0003DFFF /* User visible flags */
157#define EXT3_FL_USER_MODIFIABLE 0x000380FF /* User modifiable flags */
158
159/* Flags that should be inherited by new inodes from their parent. */
160#define EXT3_FL_INHERITED (EXT3_SECRM_FL | EXT3_UNRM_FL | EXT3_COMPR_FL |\
161 EXT3_SYNC_FL | EXT3_NODUMP_FL |\
162 EXT3_NOATIME_FL | EXT3_COMPRBLK_FL |\
163 EXT3_NOCOMPR_FL | EXT3_JOURNAL_DATA_FL |\
164 EXT3_NOTAIL_FL | EXT3_DIRSYNC_FL)
165
166/* Flags that are appropriate for regular files (all but dir-specific ones). */
167#define EXT3_REG_FLMASK (~(EXT3_DIRSYNC_FL | EXT3_TOPDIR_FL))
168
169/* Flags that are appropriate for non-directories/regular files. */
170#define EXT3_OTHER_FLMASK (EXT3_NODUMP_FL | EXT3_NOATIME_FL)
171
172/* Mask out flags that are inappropriate for the given type of inode. */
173static inline __u32 ext3_mask_flags(umode_t mode, __u32 flags)
174{
175 if (S_ISDIR(mode))
176 return flags;
177 else if (S_ISREG(mode))
178 return flags & EXT3_REG_FLMASK;
179 else
180 return flags & EXT3_OTHER_FLMASK;
181}
182
183/* Used to pass group descriptor data when online resize is done */
184struct ext3_new_group_input {
185 __u32 group; /* Group number for this data */
186 __u32 block_bitmap; /* Absolute block number of block bitmap */
187 __u32 inode_bitmap; /* Absolute block number of inode bitmap */
188 __u32 inode_table; /* Absolute block number of inode table start */
189 __u32 blocks_count; /* Total number of blocks in this group */
190 __u16 reserved_blocks; /* Number of reserved blocks in this group */
191 __u16 unused;
192};
193
194/* The struct ext3_new_group_input in kernel space, with free_blocks_count */
195struct ext3_new_group_data {
196 __u32 group;
197 __u32 block_bitmap;
198 __u32 inode_bitmap;
199 __u32 inode_table;
200 __u32 blocks_count;
201 __u16 reserved_blocks;
202 __u16 unused;
203 __u32 free_blocks_count;
204};
205
206
207/*
208 * ioctl commands
209 */
210#define EXT3_IOC_GETFLAGS FS_IOC_GETFLAGS
211#define EXT3_IOC_SETFLAGS FS_IOC_SETFLAGS
212#define EXT3_IOC_GETVERSION _IOR('f', 3, long)
213#define EXT3_IOC_SETVERSION _IOW('f', 4, long)
214#define EXT3_IOC_GROUP_EXTEND _IOW('f', 7, unsigned long)
215#define EXT3_IOC_GROUP_ADD _IOW('f', 8,struct ext3_new_group_input)
216#define EXT3_IOC_GETVERSION_OLD FS_IOC_GETVERSION
217#define EXT3_IOC_SETVERSION_OLD FS_IOC_SETVERSION
218#ifdef CONFIG_JBD_DEBUG
219#define EXT3_IOC_WAIT_FOR_READONLY _IOR('f', 99, long)
220#endif
221#define EXT3_IOC_GETRSVSZ _IOR('f', 5, long)
222#define EXT3_IOC_SETRSVSZ _IOW('f', 6, long)
223
224/*
225 * ioctl commands in 32 bit emulation
226 */
227#define EXT3_IOC32_GETFLAGS FS_IOC32_GETFLAGS
228#define EXT3_IOC32_SETFLAGS FS_IOC32_SETFLAGS
229#define EXT3_IOC32_GETVERSION _IOR('f', 3, int)
230#define EXT3_IOC32_SETVERSION _IOW('f', 4, int)
231#define EXT3_IOC32_GETRSVSZ _IOR('f', 5, int)
232#define EXT3_IOC32_SETRSVSZ _IOW('f', 6, int)
233#define EXT3_IOC32_GROUP_EXTEND _IOW('f', 7, unsigned int)
234#ifdef CONFIG_JBD_DEBUG
235#define EXT3_IOC32_WAIT_FOR_READONLY _IOR('f', 99, int)
236#endif
237#define EXT3_IOC32_GETVERSION_OLD FS_IOC32_GETVERSION
238#define EXT3_IOC32_SETVERSION_OLD FS_IOC32_SETVERSION
239
240
241/*
242 * Mount options
243 */
244struct ext3_mount_options {
245 unsigned long s_mount_opt;
246 uid_t s_resuid;
247 gid_t s_resgid;
248 unsigned long s_commit_interval;
249#ifdef CONFIG_QUOTA
250 int s_jquota_fmt;
251 char *s_qf_names[MAXQUOTAS];
252#endif
253};
254
255/*
256 * Structure of an inode on the disk
257 */
258struct ext3_inode {
259 __le16 i_mode; /* File mode */
260 __le16 i_uid; /* Low 16 bits of Owner Uid */
261 __le32 i_size; /* Size in bytes */
262 __le32 i_atime; /* Access time */
263 __le32 i_ctime; /* Creation time */
264 __le32 i_mtime; /* Modification time */
265 __le32 i_dtime; /* Deletion Time */
266 __le16 i_gid; /* Low 16 bits of Group Id */
267 __le16 i_links_count; /* Links count */
268 __le32 i_blocks; /* Blocks count */
269 __le32 i_flags; /* File flags */
270 union {
271 struct {
272 __u32 l_i_reserved1;
273 } linux1;
274 struct {
275 __u32 h_i_translator;
276 } hurd1;
277 struct {
278 __u32 m_i_reserved1;
279 } masix1;
280 } osd1; /* OS dependent 1 */
281 __le32 i_block[EXT3_N_BLOCKS];/* Pointers to blocks */
282 __le32 i_generation; /* File version (for NFS) */
283 __le32 i_file_acl; /* File ACL */
284 __le32 i_dir_acl; /* Directory ACL */
285 __le32 i_faddr; /* Fragment address */
286 union {
287 struct {
288 __u8 l_i_frag; /* Fragment number */
289 __u8 l_i_fsize; /* Fragment size */
290 __u16 i_pad1;
291 __le16 l_i_uid_high; /* these 2 fields */
292 __le16 l_i_gid_high; /* were reserved2[0] */
293 __u32 l_i_reserved2;
294 } linux2;
295 struct {
296 __u8 h_i_frag; /* Fragment number */
297 __u8 h_i_fsize; /* Fragment size */
298 __u16 h_i_mode_high;
299 __u16 h_i_uid_high;
300 __u16 h_i_gid_high;
301 __u32 h_i_author;
302 } hurd2;
303 struct {
304 __u8 m_i_frag; /* Fragment number */
305 __u8 m_i_fsize; /* Fragment size */
306 __u16 m_pad1;
307 __u32 m_i_reserved2[2];
308 } masix2;
309 } osd2; /* OS dependent 2 */
310 __le16 i_extra_isize;
311 __le16 i_pad1;
312};
313
314#define i_size_high i_dir_acl
315
316#define i_reserved1 osd1.linux1.l_i_reserved1
317#define i_frag osd2.linux2.l_i_frag
318#define i_fsize osd2.linux2.l_i_fsize
319#define i_uid_low i_uid
320#define i_gid_low i_gid
321#define i_uid_high osd2.linux2.l_i_uid_high
322#define i_gid_high osd2.linux2.l_i_gid_high
323#define i_reserved2 osd2.linux2.l_i_reserved2
324
325/*
326 * File system states
327 */
328#define EXT3_VALID_FS 0x0001 /* Unmounted cleanly */
329#define EXT3_ERROR_FS 0x0002 /* Errors detected */
330#define EXT3_ORPHAN_FS 0x0004 /* Orphans being recovered */
331
332/*
333 * Misc. filesystem flags
334 */
335#define EXT2_FLAGS_SIGNED_HASH 0x0001 /* Signed dirhash in use */
336#define EXT2_FLAGS_UNSIGNED_HASH 0x0002 /* Unsigned dirhash in use */
337#define EXT2_FLAGS_TEST_FILESYS 0x0004 /* to test development code */
338
339/*
340 * Mount flags
341 */
342#define EXT3_MOUNT_CHECK 0x00001 /* Do mount-time checks */
343/* EXT3_MOUNT_OLDALLOC was there */
344#define EXT3_MOUNT_GRPID 0x00004 /* Create files with directory's group */
345#define EXT3_MOUNT_DEBUG 0x00008 /* Some debugging messages */
346#define EXT3_MOUNT_ERRORS_CONT 0x00010 /* Continue on errors */
347#define EXT3_MOUNT_ERRORS_RO 0x00020 /* Remount fs ro on errors */
348#define EXT3_MOUNT_ERRORS_PANIC 0x00040 /* Panic on errors */
349#define EXT3_MOUNT_MINIX_DF 0x00080 /* Mimics the Minix statfs */
350#define EXT3_MOUNT_NOLOAD 0x00100 /* Don't use existing journal*/
351#define EXT3_MOUNT_ABORT 0x00200 /* Fatal error detected */
352#define EXT3_MOUNT_DATA_FLAGS 0x00C00 /* Mode for data writes: */
353#define EXT3_MOUNT_JOURNAL_DATA 0x00400 /* Write data to journal */
354#define EXT3_MOUNT_ORDERED_DATA 0x00800 /* Flush data before commit */
355#define EXT3_MOUNT_WRITEBACK_DATA 0x00C00 /* No data ordering */
356#define EXT3_MOUNT_UPDATE_JOURNAL 0x01000 /* Update the journal format */
357#define EXT3_MOUNT_NO_UID32 0x02000 /* Disable 32-bit UIDs */
358#define EXT3_MOUNT_XATTR_USER 0x04000 /* Extended user attributes */
359#define EXT3_MOUNT_POSIX_ACL 0x08000 /* POSIX Access Control Lists */
360#define EXT3_MOUNT_RESERVATION 0x10000 /* Preallocation */
361#define EXT3_MOUNT_BARRIER 0x20000 /* Use block barriers */
362#define EXT3_MOUNT_QUOTA 0x80000 /* Some quota option set */
363#define EXT3_MOUNT_USRQUOTA 0x100000 /* "old" user quota */
364#define EXT3_MOUNT_GRPQUOTA 0x200000 /* "old" group quota */
365#define EXT3_MOUNT_DATA_ERR_ABORT 0x400000 /* Abort on file data write
366 * error in ordered mode */
367
368/* Compatibility, for having both ext2_fs.h and ext3_fs.h included at once */
369#ifndef _LINUX_EXT2_FS_H
370#define clear_opt(o, opt) o &= ~EXT3_MOUNT_##opt
371#define set_opt(o, opt) o |= EXT3_MOUNT_##opt
372#define test_opt(sb, opt) (EXT3_SB(sb)->s_mount_opt & \
373 EXT3_MOUNT_##opt)
374#else
375#define EXT2_MOUNT_NOLOAD EXT3_MOUNT_NOLOAD
376#define EXT2_MOUNT_ABORT EXT3_MOUNT_ABORT
377#define EXT2_MOUNT_DATA_FLAGS EXT3_MOUNT_DATA_FLAGS
378#endif
379
380#define ext3_set_bit __set_bit_le
381#define ext3_set_bit_atomic ext2_set_bit_atomic
382#define ext3_clear_bit __clear_bit_le
383#define ext3_clear_bit_atomic ext2_clear_bit_atomic
384#define ext3_test_bit test_bit_le
385#define ext3_find_next_zero_bit find_next_zero_bit_le
386
387/*
388 * Maximal mount counts between two filesystem checks
389 */
390#define EXT3_DFL_MAX_MNT_COUNT 20 /* Allow 20 mounts */
391#define EXT3_DFL_CHECKINTERVAL 0 /* Don't use interval check */
392
393/*
394 * Behaviour when detecting errors
395 */
396#define EXT3_ERRORS_CONTINUE 1 /* Continue execution */
397#define EXT3_ERRORS_RO 2 /* Remount fs read-only */
398#define EXT3_ERRORS_PANIC 3 /* Panic */
399#define EXT3_ERRORS_DEFAULT EXT3_ERRORS_CONTINUE
400
401/*
402 * Structure of the super block
403 */
404struct ext3_super_block {
405/*00*/ __le32 s_inodes_count; /* Inodes count */
406 __le32 s_blocks_count; /* Blocks count */
407 __le32 s_r_blocks_count; /* Reserved blocks count */
408 __le32 s_free_blocks_count; /* Free blocks count */
409/*10*/ __le32 s_free_inodes_count; /* Free inodes count */
410 __le32 s_first_data_block; /* First Data Block */
411 __le32 s_log_block_size; /* Block size */
412 __le32 s_log_frag_size; /* Fragment size */
413/*20*/ __le32 s_blocks_per_group; /* # Blocks per group */
414 __le32 s_frags_per_group; /* # Fragments per group */
415 __le32 s_inodes_per_group; /* # Inodes per group */
416 __le32 s_mtime; /* Mount time */
417/*30*/ __le32 s_wtime; /* Write time */
418 __le16 s_mnt_count; /* Mount count */
419 __le16 s_max_mnt_count; /* Maximal mount count */
420 __le16 s_magic; /* Magic signature */
421 __le16 s_state; /* File system state */
422 __le16 s_errors; /* Behaviour when detecting errors */
423 __le16 s_minor_rev_level; /* minor revision level */
424/*40*/ __le32 s_lastcheck; /* time of last check */
425 __le32 s_checkinterval; /* max. time between checks */
426 __le32 s_creator_os; /* OS */
427 __le32 s_rev_level; /* Revision level */
428/*50*/ __le16 s_def_resuid; /* Default uid for reserved blocks */
429 __le16 s_def_resgid; /* Default gid for reserved blocks */
430 /*
431 * These fields are for EXT3_DYNAMIC_REV superblocks only.
432 *
433 * Note: the difference between the compatible feature set and
434 * the incompatible feature set is that if there is a bit set
435 * in the incompatible feature set that the kernel doesn't
436 * know about, it should refuse to mount the filesystem.
437 *
438 * e2fsck's requirements are more strict; if it doesn't know
439 * about a feature in either the compatible or incompatible
440 * feature set, it must abort and not try to meddle with
441 * things it doesn't understand...
442 */
443 __le32 s_first_ino; /* First non-reserved inode */
444 __le16 s_inode_size; /* size of inode structure */
445 __le16 s_block_group_nr; /* block group # of this superblock */
446 __le32 s_feature_compat; /* compatible feature set */
447/*60*/ __le32 s_feature_incompat; /* incompatible feature set */
448 __le32 s_feature_ro_compat; /* readonly-compatible feature set */
449/*68*/ __u8 s_uuid[16]; /* 128-bit uuid for volume */
450/*78*/ char s_volume_name[16]; /* volume name */
451/*88*/ char s_last_mounted[64]; /* directory where last mounted */
452/*C8*/ __le32 s_algorithm_usage_bitmap; /* For compression */
453 /*
454 * Performance hints. Directory preallocation should only
455 * happen if the EXT3_FEATURE_COMPAT_DIR_PREALLOC flag is on.
456 */
457 __u8 s_prealloc_blocks; /* Nr of blocks to try to preallocate*/
458 __u8 s_prealloc_dir_blocks; /* Nr to preallocate for dirs */
459 __le16 s_reserved_gdt_blocks; /* Per group desc for online growth */
460 /*
461 * Journaling support valid if EXT3_FEATURE_COMPAT_HAS_JOURNAL set.
462 */
463/*D0*/ __u8 s_journal_uuid[16]; /* uuid of journal superblock */
464/*E0*/ __le32 s_journal_inum; /* inode number of journal file */
465 __le32 s_journal_dev; /* device number of journal file */
466 __le32 s_last_orphan; /* start of list of inodes to delete */
467 __le32 s_hash_seed[4]; /* HTREE hash seed */
468 __u8 s_def_hash_version; /* Default hash version to use */
469 __u8 s_reserved_char_pad;
470 __u16 s_reserved_word_pad;
471 __le32 s_default_mount_opts;
472 __le32 s_first_meta_bg; /* First metablock block group */
473 __le32 s_mkfs_time; /* When the filesystem was created */
474 __le32 s_jnl_blocks[17]; /* Backup of the journal inode */
475 /* 64bit support valid if EXT4_FEATURE_COMPAT_64BIT */
476/*150*/ __le32 s_blocks_count_hi; /* Blocks count */
477 __le32 s_r_blocks_count_hi; /* Reserved blocks count */
478 __le32 s_free_blocks_count_hi; /* Free blocks count */
479 __le16 s_min_extra_isize; /* All inodes have at least # bytes */
480 __le16 s_want_extra_isize; /* New inodes should reserve # bytes */
481 __le32 s_flags; /* Miscellaneous flags */
482 __le16 s_raid_stride; /* RAID stride */
483 __le16 s_mmp_interval; /* # seconds to wait in MMP checking */
484 __le64 s_mmp_block; /* Block for multi-mount protection */
485 __le32 s_raid_stripe_width; /* blocks on all data disks (N*stride)*/
486 __u8 s_log_groups_per_flex; /* FLEX_BG group size */
487 __u8 s_reserved_char_pad2;
488 __le16 s_reserved_pad;
489 __u32 s_reserved[162]; /* Padding to the end of the block */
490};
491
492/* data type for block offset of block group */
493typedef int ext3_grpblk_t;
494
495/* data type for filesystem-wide blocks number */
496typedef unsigned long ext3_fsblk_t;
497
498#define E3FSBLK "%lu"
499
500struct ext3_reserve_window {
501 ext3_fsblk_t _rsv_start; /* First byte reserved */
502 ext3_fsblk_t _rsv_end; /* Last byte reserved or 0 */
503};
504
505struct ext3_reserve_window_node {
506 struct rb_node rsv_node;
507 __u32 rsv_goal_size;
508 __u32 rsv_alloc_hit;
509 struct ext3_reserve_window rsv_window;
510};
511
512struct ext3_block_alloc_info {
513 /* information about reservation window */
514 struct ext3_reserve_window_node rsv_window_node;
515 /*
516 * was i_next_alloc_block in ext3_inode_info
517 * is the logical (file-relative) number of the
518 * most-recently-allocated block in this file.
519 * We use this for detecting linearly ascending allocation requests.
520 */
521 __u32 last_alloc_logical_block;
522 /*
523 * Was i_next_alloc_goal in ext3_inode_info
524 * is the *physical* companion to i_next_alloc_block.
525 * it the physical block number of the block which was most-recentl
526 * allocated to this file. This give us the goal (target) for the next
527 * allocation when we detect linearly ascending requests.
528 */
529 ext3_fsblk_t last_alloc_physical_block;
530};
531
532#define rsv_start rsv_window._rsv_start
533#define rsv_end rsv_window._rsv_end
534
535/*
536 * third extended file system inode data in memory
537 */
538struct ext3_inode_info {
539 __le32 i_data[15]; /* unconverted */
540 __u32 i_flags;
541#ifdef EXT3_FRAGMENTS
542 __u32 i_faddr;
543 __u8 i_frag_no;
544 __u8 i_frag_size;
545#endif
546 ext3_fsblk_t i_file_acl;
547 __u32 i_dir_acl;
548 __u32 i_dtime;
549
550 /*
551 * i_block_group is the number of the block group which contains
552 * this file's inode. Constant across the lifetime of the inode,
553 * it is ued for making block allocation decisions - we try to
554 * place a file's data blocks near its inode block, and new inodes
555 * near to their parent directory's inode.
556 */
557 __u32 i_block_group;
558 unsigned long i_state_flags; /* Dynamic state flags for ext3 */
559
560 /* block reservation info */
561 struct ext3_block_alloc_info *i_block_alloc_info;
562
563 __u32 i_dir_start_lookup;
564#ifdef CONFIG_EXT3_FS_XATTR
565 /*
566 * Extended attributes can be read independently of the main file
567 * data. Taking i_mutex even when reading would cause contention
568 * between readers of EAs and writers of regular file data, so
569 * instead we synchronize on xattr_sem when reading or changing
570 * EAs.
571 */
572 struct rw_semaphore xattr_sem;
573#endif
574
575 struct list_head i_orphan; /* unlinked but open inodes */
576
577 /*
578 * i_disksize keeps track of what the inode size is ON DISK, not
579 * in memory. During truncate, i_size is set to the new size by
580 * the VFS prior to calling ext3_truncate(), but the filesystem won't
581 * set i_disksize to 0 until the truncate is actually under way.
582 *
583 * The intent is that i_disksize always represents the blocks which
584 * are used by this file. This allows recovery to restart truncate
585 * on orphans if we crash during truncate. We actually write i_disksize
586 * into the on-disk inode when writing inodes out, instead of i_size.
587 *
588 * The only time when i_disksize and i_size may be different is when
589 * a truncate is in progress. The only things which change i_disksize
590 * are ext3_get_block (growth) and ext3_truncate (shrinkth).
591 */
592 loff_t i_disksize;
593
594 /* on-disk additional length */
595 __u16 i_extra_isize;
596
597 /*
598 * truncate_mutex is for serialising ext3_truncate() against
599 * ext3_getblock(). In the 2.4 ext2 design, great chunks of inode's
600 * data tree are chopped off during truncate. We can't do that in
601 * ext3 because whenever we perform intermediate commits during
602 * truncate, the inode and all the metadata blocks *must* be in a
603 * consistent state which allows truncation of the orphans to restart
604 * during recovery. Hence we must fix the get_block-vs-truncate race
605 * by other means, so we have truncate_mutex.
606 */
607 struct mutex truncate_mutex;
608
609 /*
610 * Transactions that contain inode's metadata needed to complete
611 * fsync and fdatasync, respectively.
612 */
613 atomic_t i_sync_tid;
614 atomic_t i_datasync_tid;
615
616 struct inode vfs_inode;
617};
618
619/*
620 * third extended-fs super-block data in memory
621 */
622struct ext3_sb_info {
623 unsigned long s_frag_size; /* Size of a fragment in bytes */
624 unsigned long s_frags_per_block;/* Number of fragments per block */
625 unsigned long s_inodes_per_block;/* Number of inodes per block */
626 unsigned long s_frags_per_group;/* Number of fragments in a group */
627 unsigned long s_blocks_per_group;/* Number of blocks in a group */
628 unsigned long s_inodes_per_group;/* Number of inodes in a group */
629 unsigned long s_itb_per_group; /* Number of inode table blocks per group */
630 unsigned long s_gdb_count; /* Number of group descriptor blocks */
631 unsigned long s_desc_per_block; /* Number of group descriptors per block */
632 unsigned long s_groups_count; /* Number of groups in the fs */
633 unsigned long s_overhead_last; /* Last calculated overhead */
634 unsigned long s_blocks_last; /* Last seen block count */
635 struct buffer_head * s_sbh; /* Buffer containing the super block */
636 struct ext3_super_block * s_es; /* Pointer to the super block in the buffer */
637 struct buffer_head ** s_group_desc;
638 unsigned long s_mount_opt;
639 ext3_fsblk_t s_sb_block;
640 uid_t s_resuid;
641 gid_t s_resgid;
642 unsigned short s_mount_state;
643 unsigned short s_pad;
644 int s_addr_per_block_bits;
645 int s_desc_per_block_bits;
646 int s_inode_size;
647 int s_first_ino;
648 spinlock_t s_next_gen_lock;
649 u32 s_next_generation;
650 u32 s_hash_seed[4];
651 int s_def_hash_version;
652 int s_hash_unsigned; /* 3 if hash should be signed, 0 if not */
653 struct percpu_counter s_freeblocks_counter;
654 struct percpu_counter s_freeinodes_counter;
655 struct percpu_counter s_dirs_counter;
656 struct blockgroup_lock *s_blockgroup_lock;
657
658 /* root of the per fs reservation window tree */
659 spinlock_t s_rsv_window_lock;
660 struct rb_root s_rsv_window_root;
661 struct ext3_reserve_window_node s_rsv_window_head;
662
663 /* Journaling */
664 struct inode * s_journal_inode;
665 struct journal_s * s_journal;
666 struct list_head s_orphan;
667 struct mutex s_orphan_lock;
668 struct mutex s_resize_lock;
669 unsigned long s_commit_interval;
670 struct block_device *journal_bdev;
671#ifdef CONFIG_QUOTA
672 char *s_qf_names[MAXQUOTAS]; /* Names of quota files with journalled quota */
673 int s_jquota_fmt; /* Format of quota to use */
674#endif
675};
676
677static inline spinlock_t *
678sb_bgl_lock(struct ext3_sb_info *sbi, unsigned int block_group)
679{
680 return bgl_lock_ptr(sbi->s_blockgroup_lock, block_group);
681}
682
683static inline struct ext3_sb_info * EXT3_SB(struct super_block *sb)
684{
685 return sb->s_fs_info;
686}
687static inline struct ext3_inode_info *EXT3_I(struct inode *inode)
688{
689 return container_of(inode, struct ext3_inode_info, vfs_inode);
690}
691
692static inline int ext3_valid_inum(struct super_block *sb, unsigned long ino)
693{
694 return ino == EXT3_ROOT_INO ||
695 ino == EXT3_JOURNAL_INO ||
696 ino == EXT3_RESIZE_INO ||
697 (ino >= EXT3_FIRST_INO(sb) &&
698 ino <= le32_to_cpu(EXT3_SB(sb)->s_es->s_inodes_count));
699}
700
701/*
702 * Inode dynamic state flags
703 */
704enum {
705 EXT3_STATE_JDATA, /* journaled data exists */
706 EXT3_STATE_NEW, /* inode is newly created */
707 EXT3_STATE_XATTR, /* has in-inode xattrs */
708 EXT3_STATE_FLUSH_ON_CLOSE, /* flush dirty pages on close */
709};
710
711static inline int ext3_test_inode_state(struct inode *inode, int bit)
712{
713 return test_bit(bit, &EXT3_I(inode)->i_state_flags);
714}
715
716static inline void ext3_set_inode_state(struct inode *inode, int bit)
717{
718 set_bit(bit, &EXT3_I(inode)->i_state_flags);
719}
720
721static inline void ext3_clear_inode_state(struct inode *inode, int bit)
722{
723 clear_bit(bit, &EXT3_I(inode)->i_state_flags);
724}
725
726#define NEXT_ORPHAN(inode) EXT3_I(inode)->i_dtime
727
728/*
729 * Codes for operating systems
730 */
731#define EXT3_OS_LINUX 0
732#define EXT3_OS_HURD 1
733#define EXT3_OS_MASIX 2
734#define EXT3_OS_FREEBSD 3
735#define EXT3_OS_LITES 4
736
737/*
738 * Revision levels
739 */
740#define EXT3_GOOD_OLD_REV 0 /* The good old (original) format */
741#define EXT3_DYNAMIC_REV 1 /* V2 format w/ dynamic inode sizes */
742
743#define EXT3_CURRENT_REV EXT3_GOOD_OLD_REV
744#define EXT3_MAX_SUPP_REV EXT3_DYNAMIC_REV
745
746#define EXT3_GOOD_OLD_INODE_SIZE 128
747
748/*
749 * Feature set definitions
750 */
751
752#define EXT3_HAS_COMPAT_FEATURE(sb,mask) \
753 ( EXT3_SB(sb)->s_es->s_feature_compat & cpu_to_le32(mask) )
754#define EXT3_HAS_RO_COMPAT_FEATURE(sb,mask) \
755 ( EXT3_SB(sb)->s_es->s_feature_ro_compat & cpu_to_le32(mask) )
756#define EXT3_HAS_INCOMPAT_FEATURE(sb,mask) \
757 ( EXT3_SB(sb)->s_es->s_feature_incompat & cpu_to_le32(mask) )
758#define EXT3_SET_COMPAT_FEATURE(sb,mask) \
759 EXT3_SB(sb)->s_es->s_feature_compat |= cpu_to_le32(mask)
760#define EXT3_SET_RO_COMPAT_FEATURE(sb,mask) \
761 EXT3_SB(sb)->s_es->s_feature_ro_compat |= cpu_to_le32(mask)
762#define EXT3_SET_INCOMPAT_FEATURE(sb,mask) \
763 EXT3_SB(sb)->s_es->s_feature_incompat |= cpu_to_le32(mask)
764#define EXT3_CLEAR_COMPAT_FEATURE(sb,mask) \
765 EXT3_SB(sb)->s_es->s_feature_compat &= ~cpu_to_le32(mask)
766#define EXT3_CLEAR_RO_COMPAT_FEATURE(sb,mask) \
767 EXT3_SB(sb)->s_es->s_feature_ro_compat &= ~cpu_to_le32(mask)
768#define EXT3_CLEAR_INCOMPAT_FEATURE(sb,mask) \
769 EXT3_SB(sb)->s_es->s_feature_incompat &= ~cpu_to_le32(mask)
770
771#define EXT3_FEATURE_COMPAT_DIR_PREALLOC 0x0001
772#define EXT3_FEATURE_COMPAT_IMAGIC_INODES 0x0002
773#define EXT3_FEATURE_COMPAT_HAS_JOURNAL 0x0004
774#define EXT3_FEATURE_COMPAT_EXT_ATTR 0x0008
775#define EXT3_FEATURE_COMPAT_RESIZE_INODE 0x0010
776#define EXT3_FEATURE_COMPAT_DIR_INDEX 0x0020
777
778#define EXT3_FEATURE_RO_COMPAT_SPARSE_SUPER 0x0001
779#define EXT3_FEATURE_RO_COMPAT_LARGE_FILE 0x0002
780#define EXT3_FEATURE_RO_COMPAT_BTREE_DIR 0x0004
781
782#define EXT3_FEATURE_INCOMPAT_COMPRESSION 0x0001
783#define EXT3_FEATURE_INCOMPAT_FILETYPE 0x0002
784#define EXT3_FEATURE_INCOMPAT_RECOVER 0x0004 /* Needs recovery */
785#define EXT3_FEATURE_INCOMPAT_JOURNAL_DEV 0x0008 /* Journal device */
786#define EXT3_FEATURE_INCOMPAT_META_BG 0x0010
787
788#define EXT3_FEATURE_COMPAT_SUPP EXT2_FEATURE_COMPAT_EXT_ATTR
789#define EXT3_FEATURE_INCOMPAT_SUPP (EXT3_FEATURE_INCOMPAT_FILETYPE| \
790 EXT3_FEATURE_INCOMPAT_RECOVER| \
791 EXT3_FEATURE_INCOMPAT_META_BG)
792#define EXT3_FEATURE_RO_COMPAT_SUPP (EXT3_FEATURE_RO_COMPAT_SPARSE_SUPER| \
793 EXT3_FEATURE_RO_COMPAT_LARGE_FILE| \
794 EXT3_FEATURE_RO_COMPAT_BTREE_DIR)
795
796/*
797 * Default values for user and/or group using reserved blocks
798 */
799#define EXT3_DEF_RESUID 0
800#define EXT3_DEF_RESGID 0
801
802/*
803 * Default mount options
804 */
805#define EXT3_DEFM_DEBUG 0x0001
806#define EXT3_DEFM_BSDGROUPS 0x0002
807#define EXT3_DEFM_XATTR_USER 0x0004
808#define EXT3_DEFM_ACL 0x0008
809#define EXT3_DEFM_UID16 0x0010
810#define EXT3_DEFM_JMODE 0x0060
811#define EXT3_DEFM_JMODE_DATA 0x0020
812#define EXT3_DEFM_JMODE_ORDERED 0x0040
813#define EXT3_DEFM_JMODE_WBACK 0x0060
814
815/*
816 * Structure of a directory entry
817 */
818#define EXT3_NAME_LEN 255
819
820struct ext3_dir_entry {
821 __le32 inode; /* Inode number */
822 __le16 rec_len; /* Directory entry length */
823 __le16 name_len; /* Name length */
824 char name[EXT3_NAME_LEN]; /* File name */
825};
826
827/*
828 * The new version of the directory entry. Since EXT3 structures are
829 * stored in intel byte order, and the name_len field could never be
830 * bigger than 255 chars, it's safe to reclaim the extra byte for the
831 * file_type field.
832 */
833struct ext3_dir_entry_2 {
834 __le32 inode; /* Inode number */
835 __le16 rec_len; /* Directory entry length */
836 __u8 name_len; /* Name length */
837 __u8 file_type;
838 char name[EXT3_NAME_LEN]; /* File name */
839};
840
841/*
842 * Ext3 directory file types. Only the low 3 bits are used. The
843 * other bits are reserved for now.
844 */
845#define EXT3_FT_UNKNOWN 0
846#define EXT3_FT_REG_FILE 1
847#define EXT3_FT_DIR 2
848#define EXT3_FT_CHRDEV 3
849#define EXT3_FT_BLKDEV 4
850#define EXT3_FT_FIFO 5
851#define EXT3_FT_SOCK 6
852#define EXT3_FT_SYMLINK 7
853
854#define EXT3_FT_MAX 8
855
856/*
857 * EXT3_DIR_PAD defines the directory entries boundaries
858 *
859 * NOTE: It must be a multiple of 4
860 */
861#define EXT3_DIR_PAD 4
862#define EXT3_DIR_ROUND (EXT3_DIR_PAD - 1)
863#define EXT3_DIR_REC_LEN(name_len) (((name_len) + 8 + EXT3_DIR_ROUND) & \
864 ~EXT3_DIR_ROUND)
865#define EXT3_MAX_REC_LEN ((1<<16)-1)
866
867/*
868 * Tests against MAX_REC_LEN etc were put in place for 64k block
869 * sizes; if that is not possible on this arch, we can skip
870 * those tests and speed things up.
871 */
872static inline unsigned ext3_rec_len_from_disk(__le16 dlen)
873{
874 unsigned len = le16_to_cpu(dlen);
875
876#if (PAGE_CACHE_SIZE >= 65536)
877 if (len == EXT3_MAX_REC_LEN)
878 return 1 << 16;
879#endif
880 return len;
881}
882
883static inline __le16 ext3_rec_len_to_disk(unsigned len)
884{
885#if (PAGE_CACHE_SIZE >= 65536)
886 if (len == (1 << 16))
887 return cpu_to_le16(EXT3_MAX_REC_LEN);
888 else if (len > (1 << 16))
889 BUG();
890#endif
891 return cpu_to_le16(len);
892}
893
894/*
895 * Hash Tree Directory indexing
896 * (c) Daniel Phillips, 2001
897 */
898
899#define is_dx(dir) (EXT3_HAS_COMPAT_FEATURE(dir->i_sb, \
900 EXT3_FEATURE_COMPAT_DIR_INDEX) && \
901 (EXT3_I(dir)->i_flags & EXT3_INDEX_FL))
902#define EXT3_DIR_LINK_MAX(dir) (!is_dx(dir) && (dir)->i_nlink >= EXT3_LINK_MAX)
903#define EXT3_DIR_LINK_EMPTY(dir) ((dir)->i_nlink == 2 || (dir)->i_nlink == 1)
904
905/* Legal values for the dx_root hash_version field: */
906
907#define DX_HASH_LEGACY 0
908#define DX_HASH_HALF_MD4 1
909#define DX_HASH_TEA 2
910#define DX_HASH_LEGACY_UNSIGNED 3
911#define DX_HASH_HALF_MD4_UNSIGNED 4
912#define DX_HASH_TEA_UNSIGNED 5
913
914/* hash info structure used by the directory hash */
915struct dx_hash_info
916{
917 u32 hash;
918 u32 minor_hash;
919 int hash_version;
920 u32 *seed;
921};
922
923#define EXT3_HTREE_EOF 0x7fffffff
924
925/*
926 * Control parameters used by ext3_htree_next_block
927 */
928#define HASH_NB_ALWAYS 1
929
930
931/*
932 * Describe an inode's exact location on disk and in memory
933 */
934struct ext3_iloc
935{
936 struct buffer_head *bh;
937 unsigned long offset;
938 unsigned long block_group;
939};
940
941static inline struct ext3_inode *ext3_raw_inode(struct ext3_iloc *iloc)
942{
943 return (struct ext3_inode *) (iloc->bh->b_data + iloc->offset);
944}
945
946/*
947 * This structure is stuffed into the struct file's private_data field
948 * for directories. It is where we put information so that we can do
949 * readdir operations in hash tree order.
950 */
951struct dir_private_info {
952 struct rb_root root;
953 struct rb_node *curr_node;
954 struct fname *extra_fname;
955 loff_t last_pos;
956 __u32 curr_hash;
957 __u32 curr_minor_hash;
958 __u32 next_hash;
959};
960
961/* calculate the first block number of the group */
962static inline ext3_fsblk_t
963ext3_group_first_block_no(struct super_block *sb, unsigned long group_no)
964{
965 return group_no * (ext3_fsblk_t)EXT3_BLOCKS_PER_GROUP(sb) +
966 le32_to_cpu(EXT3_SB(sb)->s_es->s_first_data_block);
967}
968
969/*
970 * Special error return code only used by dx_probe() and its callers.
971 */
972#define ERR_BAD_DX_DIR -75000
973
974/*
975 * Function prototypes
976 */
977
978/*
979 * Ok, these declarations are also in <linux/kernel.h> but none of the
980 * ext3 source programs needs to include it so they are duplicated here.
981 */
982# define NORET_TYPE /**/
983# define ATTRIB_NORET __attribute__((noreturn))
984# define NORET_AND noreturn,
985
986/* balloc.c */
987extern int ext3_bg_has_super(struct super_block *sb, int group);
988extern unsigned long ext3_bg_num_gdb(struct super_block *sb, int group);
989extern ext3_fsblk_t ext3_new_block (handle_t *handle, struct inode *inode,
990 ext3_fsblk_t goal, int *errp);
991extern ext3_fsblk_t ext3_new_blocks (handle_t *handle, struct inode *inode,
992 ext3_fsblk_t goal, unsigned long *count, int *errp);
993extern void ext3_free_blocks (handle_t *handle, struct inode *inode,
994 ext3_fsblk_t block, unsigned long count);
995extern void ext3_free_blocks_sb (handle_t *handle, struct super_block *sb,
996 ext3_fsblk_t block, unsigned long count,
997 unsigned long *pdquot_freed_blocks);
998extern ext3_fsblk_t ext3_count_free_blocks (struct super_block *);
999extern void ext3_check_blocks_bitmap (struct super_block *);
1000extern struct ext3_group_desc * ext3_get_group_desc(struct super_block * sb,
1001 unsigned int block_group,
1002 struct buffer_head ** bh);
1003extern int ext3_should_retry_alloc(struct super_block *sb, int *retries);
1004extern void ext3_init_block_alloc_info(struct inode *);
1005extern void ext3_rsv_window_add(struct super_block *sb, struct ext3_reserve_window_node *rsv);
1006extern int ext3_trim_fs(struct super_block *sb, struct fstrim_range *range);
1007
1008/* dir.c */
1009extern int ext3_check_dir_entry(const char *, struct inode *,
1010 struct ext3_dir_entry_2 *,
1011 struct buffer_head *, unsigned long);
1012extern int ext3_htree_store_dirent(struct file *dir_file, __u32 hash,
1013 __u32 minor_hash,
1014 struct ext3_dir_entry_2 *dirent);
1015extern void ext3_htree_free_dir_info(struct dir_private_info *p);
1016
1017/* fsync.c */
1018extern int ext3_sync_file(struct file *, loff_t, loff_t, int);
1019
1020/* hash.c */
1021extern int ext3fs_dirhash(const char *name, int len, struct
1022 dx_hash_info *hinfo);
1023
1024/* ialloc.c */
1025extern struct inode * ext3_new_inode (handle_t *, struct inode *,
1026 const struct qstr *, umode_t);
1027extern void ext3_free_inode (handle_t *, struct inode *);
1028extern struct inode * ext3_orphan_get (struct super_block *, unsigned long);
1029extern unsigned long ext3_count_free_inodes (struct super_block *);
1030extern unsigned long ext3_count_dirs (struct super_block *);
1031extern void ext3_check_inodes_bitmap (struct super_block *);
1032extern unsigned long ext3_count_free (struct buffer_head *, unsigned);
1033
1034
1035/* inode.c */
1036int ext3_forget(handle_t *handle, int is_metadata, struct inode *inode,
1037 struct buffer_head *bh, ext3_fsblk_t blocknr);
1038struct buffer_head * ext3_getblk (handle_t *, struct inode *, long, int, int *);
1039struct buffer_head * ext3_bread (handle_t *, struct inode *, int, int, int *);
1040int ext3_get_blocks_handle(handle_t *handle, struct inode *inode,
1041 sector_t iblock, unsigned long maxblocks, struct buffer_head *bh_result,
1042 int create);
1043
1044extern struct inode *ext3_iget(struct super_block *, unsigned long);
1045extern int ext3_write_inode (struct inode *, struct writeback_control *);
1046extern int ext3_setattr (struct dentry *, struct iattr *);
1047extern void ext3_evict_inode (struct inode *);
1048extern int ext3_sync_inode (handle_t *, struct inode *);
1049extern void ext3_discard_reservation (struct inode *);
1050extern void ext3_dirty_inode(struct inode *, int);
1051extern int ext3_change_inode_journal_flag(struct inode *, int);
1052extern int ext3_get_inode_loc(struct inode *, struct ext3_iloc *);
1053extern int ext3_can_truncate(struct inode *inode);
1054extern void ext3_truncate(struct inode *inode);
1055extern void ext3_set_inode_flags(struct inode *);
1056extern void ext3_get_inode_flags(struct ext3_inode_info *);
1057extern void ext3_set_aops(struct inode *inode);
1058extern int ext3_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo,
1059 u64 start, u64 len);
1060
1061/* ioctl.c */
1062extern long ext3_ioctl(struct file *, unsigned int, unsigned long);
1063extern long ext3_compat_ioctl(struct file *, unsigned int, unsigned long);
1064
1065/* namei.c */
1066extern int ext3_orphan_add(handle_t *, struct inode *);
1067extern int ext3_orphan_del(handle_t *, struct inode *);
1068extern int ext3_htree_fill_tree(struct file *dir_file, __u32 start_hash,
1069 __u32 start_minor_hash, __u32 *next_hash);
1070
1071/* resize.c */
1072extern int ext3_group_add(struct super_block *sb,
1073 struct ext3_new_group_data *input);
1074extern int ext3_group_extend(struct super_block *sb,
1075 struct ext3_super_block *es,
1076 ext3_fsblk_t n_blocks_count);
1077
1078/* super.c */
1079extern __printf(3, 4)
1080void ext3_error(struct super_block *, const char *, const char *, ...);
1081extern void __ext3_std_error (struct super_block *, const char *, int);
1082extern __printf(3, 4)
1083void ext3_abort(struct super_block *, const char *, const char *, ...);
1084extern __printf(3, 4)
1085void ext3_warning(struct super_block *, const char *, const char *, ...);
1086extern __printf(3, 4)
1087void ext3_msg(struct super_block *, const char *, const char *, ...);
1088extern void ext3_update_dynamic_rev (struct super_block *sb);
1089
1090#define ext3_std_error(sb, errno) \
1091do { \
1092 if ((errno)) \
1093 __ext3_std_error((sb), __func__, (errno)); \
1094} while (0)
1095
1096/*
1097 * Inodes and files operations
1098 */
1099
1100/* dir.c */
1101extern const struct file_operations ext3_dir_operations;
1102
1103/* file.c */
1104extern const struct inode_operations ext3_file_inode_operations;
1105extern const struct file_operations ext3_file_operations;
1106
1107/* namei.c */
1108extern const struct inode_operations ext3_dir_inode_operations;
1109extern const struct inode_operations ext3_special_inode_operations;
1110
1111/* symlink.c */
1112extern const struct inode_operations ext3_symlink_inode_operations;
1113extern const struct inode_operations ext3_fast_symlink_inode_operations;
1114
1115#define EXT3_JOURNAL(inode) (EXT3_SB((inode)->i_sb)->s_journal)
1116
1117/* Define the number of blocks we need to account to a transaction to
1118 * modify one block of data.
1119 *
1120 * We may have to touch one inode, one bitmap buffer, up to three
1121 * indirection blocks, the group and superblock summaries, and the data
1122 * block to complete the transaction. */
1123
1124#define EXT3_SINGLEDATA_TRANS_BLOCKS 8U
1125
1126/* Extended attribute operations touch at most two data buffers,
1127 * two bitmap buffers, and two group summaries, in addition to the inode
1128 * and the superblock, which are already accounted for. */
1129
1130#define EXT3_XATTR_TRANS_BLOCKS 6U
1131
1132/* Define the minimum size for a transaction which modifies data. This
1133 * needs to take into account the fact that we may end up modifying two
1134 * quota files too (one for the group, one for the user quota). The
1135 * superblock only gets updated once, of course, so don't bother
1136 * counting that again for the quota updates. */
1137
1138#define EXT3_DATA_TRANS_BLOCKS(sb) (EXT3_SINGLEDATA_TRANS_BLOCKS + \
1139 EXT3_XATTR_TRANS_BLOCKS - 2 + \
1140 EXT3_MAXQUOTAS_TRANS_BLOCKS(sb))
1141
1142/* Delete operations potentially hit one directory's namespace plus an
1143 * entire inode, plus arbitrary amounts of bitmap/indirection data. Be
1144 * generous. We can grow the delete transaction later if necessary. */
1145
1146#define EXT3_DELETE_TRANS_BLOCKS(sb) (EXT3_MAXQUOTAS_TRANS_BLOCKS(sb) + 64)
1147
1148/* Define an arbitrary limit for the amount of data we will anticipate
1149 * writing to any given transaction. For unbounded transactions such as
1150 * write(2) and truncate(2) we can write more than this, but we always
1151 * start off at the maximum transaction size and grow the transaction
1152 * optimistically as we go. */
1153
1154#define EXT3_MAX_TRANS_DATA 64U
1155
1156/* We break up a large truncate or write transaction once the handle's
1157 * buffer credits gets this low, we need either to extend the
1158 * transaction or to start a new one. Reserve enough space here for
1159 * inode, bitmap, superblock, group and indirection updates for at least
1160 * one block, plus two quota updates. Quota allocations are not
1161 * needed. */
1162
1163#define EXT3_RESERVE_TRANS_BLOCKS 12U
1164
1165#define EXT3_INDEX_EXTRA_TRANS_BLOCKS 8
1166
1167#ifdef CONFIG_QUOTA
1168/* Amount of blocks needed for quota update - we know that the structure was
1169 * allocated so we need to update only inode+data */
1170#define EXT3_QUOTA_TRANS_BLOCKS(sb) (test_opt(sb, QUOTA) ? 2 : 0)
1171/* Amount of blocks needed for quota insert/delete - we do some block writes
1172 * but inode, sb and group updates are done only once */
1173#define EXT3_QUOTA_INIT_BLOCKS(sb) (test_opt(sb, QUOTA) ? (DQUOT_INIT_ALLOC*\
1174 (EXT3_SINGLEDATA_TRANS_BLOCKS-3)+3+DQUOT_INIT_REWRITE) : 0)
1175#define EXT3_QUOTA_DEL_BLOCKS(sb) (test_opt(sb, QUOTA) ? (DQUOT_DEL_ALLOC*\
1176 (EXT3_SINGLEDATA_TRANS_BLOCKS-3)+3+DQUOT_DEL_REWRITE) : 0)
1177#else
1178#define EXT3_QUOTA_TRANS_BLOCKS(sb) 0
1179#define EXT3_QUOTA_INIT_BLOCKS(sb) 0
1180#define EXT3_QUOTA_DEL_BLOCKS(sb) 0
1181#endif
1182#define EXT3_MAXQUOTAS_TRANS_BLOCKS(sb) (MAXQUOTAS*EXT3_QUOTA_TRANS_BLOCKS(sb))
1183#define EXT3_MAXQUOTAS_INIT_BLOCKS(sb) (MAXQUOTAS*EXT3_QUOTA_INIT_BLOCKS(sb))
1184#define EXT3_MAXQUOTAS_DEL_BLOCKS(sb) (MAXQUOTAS*EXT3_QUOTA_DEL_BLOCKS(sb))
1185
1186int
1187ext3_mark_iloc_dirty(handle_t *handle,
1188 struct inode *inode,
1189 struct ext3_iloc *iloc);
1190
1191/*
1192 * On success, We end up with an outstanding reference count against
1193 * iloc->bh. This _must_ be cleaned up later.
1194 */
1195
1196int ext3_reserve_inode_write(handle_t *handle, struct inode *inode,
1197 struct ext3_iloc *iloc);
1198
1199int ext3_mark_inode_dirty(handle_t *handle, struct inode *inode);
1200
1201/*
1202 * Wrapper functions with which ext3 calls into JBD. The intent here is
1203 * to allow these to be turned into appropriate stubs so ext3 can control
1204 * ext2 filesystems, so ext2+ext3 systems only nee one fs. This work hasn't
1205 * been done yet.
1206 */
1207
1208static inline void ext3_journal_release_buffer(handle_t *handle,
1209 struct buffer_head *bh)
1210{
1211 journal_release_buffer(handle, bh);
1212}
1213
1214void ext3_journal_abort_handle(const char *caller, const char *err_fn,
1215 struct buffer_head *bh, handle_t *handle, int err);
1216
1217int __ext3_journal_get_undo_access(const char *where, handle_t *handle,
1218 struct buffer_head *bh);
1219
1220int __ext3_journal_get_write_access(const char *where, handle_t *handle,
1221 struct buffer_head *bh);
1222
1223int __ext3_journal_forget(const char *where, handle_t *handle,
1224 struct buffer_head *bh);
1225
1226int __ext3_journal_revoke(const char *where, handle_t *handle,
1227 unsigned long blocknr, struct buffer_head *bh);
1228
1229int __ext3_journal_get_create_access(const char *where,
1230 handle_t *handle, struct buffer_head *bh);
1231
1232int __ext3_journal_dirty_metadata(const char *where,
1233 handle_t *handle, struct buffer_head *bh);
1234
1235#define ext3_journal_get_undo_access(handle, bh) \
1236 __ext3_journal_get_undo_access(__func__, (handle), (bh))
1237#define ext3_journal_get_write_access(handle, bh) \
1238 __ext3_journal_get_write_access(__func__, (handle), (bh))
1239#define ext3_journal_revoke(handle, blocknr, bh) \
1240 __ext3_journal_revoke(__func__, (handle), (blocknr), (bh))
1241#define ext3_journal_get_create_access(handle, bh) \
1242 __ext3_journal_get_create_access(__func__, (handle), (bh))
1243#define ext3_journal_dirty_metadata(handle, bh) \
1244 __ext3_journal_dirty_metadata(__func__, (handle), (bh))
1245#define ext3_journal_forget(handle, bh) \
1246 __ext3_journal_forget(__func__, (handle), (bh))
1247
1248int ext3_journal_dirty_data(handle_t *handle, struct buffer_head *bh);
1249
1250handle_t *ext3_journal_start_sb(struct super_block *sb, int nblocks);
1251int __ext3_journal_stop(const char *where, handle_t *handle);
1252
1253static inline handle_t *ext3_journal_start(struct inode *inode, int nblocks)
1254{
1255 return ext3_journal_start_sb(inode->i_sb, nblocks);
1256}
1257
1258#define ext3_journal_stop(handle) \
1259 __ext3_journal_stop(__func__, (handle))
1260
1261static inline handle_t *ext3_journal_current_handle(void)
1262{
1263 return journal_current_handle();
1264}
1265
1266static inline int ext3_journal_extend(handle_t *handle, int nblocks)
1267{
1268 return journal_extend(handle, nblocks);
1269}
1270
1271static inline int ext3_journal_restart(handle_t *handle, int nblocks)
1272{
1273 return journal_restart(handle, nblocks);
1274}
1275
1276static inline int ext3_journal_blocks_per_page(struct inode *inode)
1277{
1278 return journal_blocks_per_page(inode);
1279}
1280
1281static inline int ext3_journal_force_commit(journal_t *journal)
1282{
1283 return journal_force_commit(journal);
1284}
1285
1286/* super.c */
1287int ext3_force_commit(struct super_block *sb);
1288
1289static inline int ext3_should_journal_data(struct inode *inode)
1290{
1291 if (!S_ISREG(inode->i_mode))
1292 return 1;
1293 if (test_opt(inode->i_sb, DATA_FLAGS) == EXT3_MOUNT_JOURNAL_DATA)
1294 return 1;
1295 if (EXT3_I(inode)->i_flags & EXT3_JOURNAL_DATA_FL)
1296 return 1;
1297 return 0;
1298}
1299
1300static inline int ext3_should_order_data(struct inode *inode)
1301{
1302 if (!S_ISREG(inode->i_mode))
1303 return 0;
1304 if (EXT3_I(inode)->i_flags & EXT3_JOURNAL_DATA_FL)
1305 return 0;
1306 if (test_opt(inode->i_sb, DATA_FLAGS) == EXT3_MOUNT_ORDERED_DATA)
1307 return 1;
1308 return 0;
1309}
1310
1311static inline int ext3_should_writeback_data(struct inode *inode)
1312{
1313 if (!S_ISREG(inode->i_mode))
1314 return 0;
1315 if (EXT3_I(inode)->i_flags & EXT3_JOURNAL_DATA_FL)
1316 return 0;
1317 if (test_opt(inode->i_sb, DATA_FLAGS) == EXT3_MOUNT_WRITEBACK_DATA)
1318 return 1;
1319 return 0;
1320}
1321
1322#include <trace/events/ext3.h>
diff --git a/fs/ext3/ext3_jbd.c b/fs/ext3/ext3_jbd.c
index d401f148d74d..785a3261a26c 100644
--- a/fs/ext3/ext3_jbd.c
+++ b/fs/ext3/ext3_jbd.c
@@ -2,7 +2,7 @@
2 * Interface between ext3 and JBD 2 * Interface between ext3 and JBD
3 */ 3 */
4 4
5#include <linux/ext3_jbd.h> 5#include "ext3.h"
6 6
7int __ext3_journal_get_undo_access(const char *where, handle_t *handle, 7int __ext3_journal_get_undo_access(const char *where, handle_t *handle,
8 struct buffer_head *bh) 8 struct buffer_head *bh)
diff --git a/fs/ext3/file.c b/fs/ext3/file.c
index 724df69847dc..25cb413277e9 100644
--- a/fs/ext3/file.c
+++ b/fs/ext3/file.c
@@ -18,12 +18,8 @@
18 * (jj@sunsite.ms.mff.cuni.cz) 18 * (jj@sunsite.ms.mff.cuni.cz)
19 */ 19 */
20 20
21#include <linux/time.h>
22#include <linux/fs.h>
23#include <linux/jbd.h>
24#include <linux/quotaops.h> 21#include <linux/quotaops.h>
25#include <linux/ext3_fs.h> 22#include "ext3.h"
26#include <linux/ext3_jbd.h>
27#include "xattr.h" 23#include "xattr.h"
28#include "acl.h" 24#include "acl.h"
29 25
diff --git a/fs/ext3/fsync.c b/fs/ext3/fsync.c
index 1860ed356323..d4dff278cbd8 100644
--- a/fs/ext3/fsync.c
+++ b/fs/ext3/fsync.c
@@ -22,15 +22,9 @@
22 * we can depend on generic_block_fdatasync() to sync the data blocks. 22 * we can depend on generic_block_fdatasync() to sync the data blocks.
23 */ 23 */
24 24
25#include <linux/time.h>
26#include <linux/blkdev.h> 25#include <linux/blkdev.h>
27#include <linux/fs.h>
28#include <linux/sched.h>
29#include <linux/writeback.h> 26#include <linux/writeback.h>
30#include <linux/jbd.h> 27#include "ext3.h"
31#include <linux/ext3_fs.h>
32#include <linux/ext3_jbd.h>
33#include <trace/events/ext3.h>
34 28
35/* 29/*
36 * akpm: A new design for ext3_sync_file(). 30 * akpm: A new design for ext3_sync_file().
diff --git a/fs/ext3/hash.c b/fs/ext3/hash.c
index 7d215b4d4f2e..d10231ddcf8a 100644
--- a/fs/ext3/hash.c
+++ b/fs/ext3/hash.c
@@ -9,9 +9,7 @@
9 * License. 9 * License.
10 */ 10 */
11 11
12#include <linux/fs.h> 12#include "ext3.h"
13#include <linux/jbd.h>
14#include <linux/ext3_fs.h>
15#include <linux/cryptohash.h> 13#include <linux/cryptohash.h>
16 14
17#define DELTA 0x9E3779B9 15#define DELTA 0x9E3779B9
diff --git a/fs/ext3/ialloc.c b/fs/ext3/ialloc.c
index 1cde28438014..e3c39e4cec19 100644
--- a/fs/ext3/ialloc.c
+++ b/fs/ext3/ialloc.c
@@ -12,21 +12,10 @@
12 * David S. Miller (davem@caip.rutgers.edu), 1995 12 * David S. Miller (davem@caip.rutgers.edu), 1995
13 */ 13 */
14 14
15#include <linux/time.h>
16#include <linux/fs.h>
17#include <linux/jbd.h>
18#include <linux/ext3_fs.h>
19#include <linux/ext3_jbd.h>
20#include <linux/stat.h>
21#include <linux/string.h>
22#include <linux/quotaops.h> 15#include <linux/quotaops.h>
23#include <linux/buffer_head.h>
24#include <linux/random.h> 16#include <linux/random.h>
25#include <linux/bitops.h>
26#include <trace/events/ext3.h>
27
28#include <asm/byteorder.h>
29 17
18#include "ext3.h"
30#include "xattr.h" 19#include "xattr.h"
31#include "acl.h" 20#include "acl.h"
32 21
diff --git a/fs/ext3/inode.c b/fs/ext3/inode.c
index 6d3418662b54..10d7812f6021 100644
--- a/fs/ext3/inode.c
+++ b/fs/ext3/inode.c
@@ -22,22 +22,12 @@
22 * Assorted race fixes, rewrite of ext3_get_block() by Al Viro, 2000 22 * Assorted race fixes, rewrite of ext3_get_block() by Al Viro, 2000
23 */ 23 */
24 24
25#include <linux/fs.h>
26#include <linux/time.h>
27#include <linux/ext3_jbd.h>
28#include <linux/jbd.h>
29#include <linux/highuid.h> 25#include <linux/highuid.h>
30#include <linux/pagemap.h>
31#include <linux/quotaops.h> 26#include <linux/quotaops.h>
32#include <linux/string.h>
33#include <linux/buffer_head.h>
34#include <linux/writeback.h> 27#include <linux/writeback.h>
35#include <linux/mpage.h> 28#include <linux/mpage.h>
36#include <linux/uio.h>
37#include <linux/bio.h>
38#include <linux/fiemap.h>
39#include <linux/namei.h> 29#include <linux/namei.h>
40#include <trace/events/ext3.h> 30#include "ext3.h"
41#include "xattr.h" 31#include "xattr.h"
42#include "acl.h" 32#include "acl.h"
43 33
diff --git a/fs/ext3/ioctl.c b/fs/ext3/ioctl.c
index 4af574ce4a46..677a5c27dc69 100644
--- a/fs/ext3/ioctl.c
+++ b/fs/ext3/ioctl.c
@@ -7,15 +7,10 @@
7 * Universite Pierre et Marie Curie (Paris VI) 7 * Universite Pierre et Marie Curie (Paris VI)
8 */ 8 */
9 9
10#include <linux/fs.h>
11#include <linux/jbd.h>
12#include <linux/capability.h>
13#include <linux/ext3_fs.h>
14#include <linux/ext3_jbd.h>
15#include <linux/mount.h> 10#include <linux/mount.h>
16#include <linux/time.h>
17#include <linux/compat.h> 11#include <linux/compat.h>
18#include <asm/uaccess.h> 12#include <asm/uaccess.h>
13#include "ext3.h"
19 14
20long ext3_ioctl(struct file *filp, unsigned int cmd, unsigned long arg) 15long ext3_ioctl(struct file *filp, unsigned int cmd, unsigned long arg)
21{ 16{
diff --git a/fs/ext3/namei.c b/fs/ext3/namei.c
index e8e211795e9f..d7940b24cf68 100644
--- a/fs/ext3/namei.c
+++ b/fs/ext3/namei.c
@@ -24,20 +24,8 @@
24 * Theodore Ts'o, 2002 24 * Theodore Ts'o, 2002
25 */ 25 */
26 26
27#include <linux/fs.h>
28#include <linux/pagemap.h>
29#include <linux/jbd.h>
30#include <linux/time.h>
31#include <linux/ext3_fs.h>
32#include <linux/ext3_jbd.h>
33#include <linux/fcntl.h>
34#include <linux/stat.h>
35#include <linux/string.h>
36#include <linux/quotaops.h> 27#include <linux/quotaops.h>
37#include <linux/buffer_head.h> 28#include "ext3.h"
38#include <linux/bio.h>
39#include <trace/events/ext3.h>
40
41#include "namei.h" 29#include "namei.h"
42#include "xattr.h" 30#include "xattr.h"
43#include "acl.h" 31#include "acl.h"
diff --git a/fs/ext3/resize.c b/fs/ext3/resize.c
index 7916e4ce166a..0f814f3450de 100644
--- a/fs/ext3/resize.c
+++ b/fs/ext3/resize.c
@@ -11,10 +11,7 @@
11 11
12#define EXT3FS_DEBUG 12#define EXT3FS_DEBUG
13 13
14#include <linux/ext3_jbd.h> 14#include "ext3.h"
15
16#include <linux/errno.h>
17#include <linux/slab.h>
18 15
19 16
20#define outside(b, first, last) ((b) < (first) || (b) >= (last)) 17#define outside(b, first, last) ((b) < (first) || (b) >= (last))
diff --git a/fs/ext3/super.c b/fs/ext3/super.c
index e0b45b93327b..cf0b5921cf0f 100644
--- a/fs/ext3/super.c
+++ b/fs/ext3/super.c
@@ -17,22 +17,12 @@
17 */ 17 */
18 18
19#include <linux/module.h> 19#include <linux/module.h>
20#include <linux/string.h>
21#include <linux/fs.h>
22#include <linux/time.h>
23#include <linux/jbd.h>
24#include <linux/ext3_fs.h>
25#include <linux/ext3_jbd.h>
26#include <linux/slab.h>
27#include <linux/init.h>
28#include <linux/blkdev.h> 20#include <linux/blkdev.h>
29#include <linux/parser.h> 21#include <linux/parser.h>
30#include <linux/buffer_head.h>
31#include <linux/exportfs.h> 22#include <linux/exportfs.h>
32#include <linux/vfs.h> 23#include <linux/statfs.h>
33#include <linux/random.h> 24#include <linux/random.h>
34#include <linux/mount.h> 25#include <linux/mount.h>
35#include <linux/namei.h>
36#include <linux/quotaops.h> 26#include <linux/quotaops.h>
37#include <linux/seq_file.h> 27#include <linux/seq_file.h>
38#include <linux/log2.h> 28#include <linux/log2.h>
@@ -40,13 +30,13 @@
40 30
41#include <asm/uaccess.h> 31#include <asm/uaccess.h>
42 32
33#define CREATE_TRACE_POINTS
34
35#include "ext3.h"
43#include "xattr.h" 36#include "xattr.h"
44#include "acl.h" 37#include "acl.h"
45#include "namei.h" 38#include "namei.h"
46 39
47#define CREATE_TRACE_POINTS
48#include <trace/events/ext3.h>
49
50#ifdef CONFIG_EXT3_DEFAULTS_TO_ORDERED 40#ifdef CONFIG_EXT3_DEFAULTS_TO_ORDERED
51 #define EXT3_MOUNT_DEFAULT_DATA_MODE EXT3_MOUNT_ORDERED_DATA 41 #define EXT3_MOUNT_DEFAULT_DATA_MODE EXT3_MOUNT_ORDERED_DATA
52#else 42#else
diff --git a/fs/ext3/symlink.c b/fs/ext3/symlink.c
index 7c4898207776..6b01c3eab1f3 100644
--- a/fs/ext3/symlink.c
+++ b/fs/ext3/symlink.c
@@ -17,10 +17,8 @@
17 * ext3 symlink handling code 17 * ext3 symlink handling code
18 */ 18 */
19 19
20#include <linux/fs.h>
21#include <linux/jbd.h>
22#include <linux/ext3_fs.h>
23#include <linux/namei.h> 20#include <linux/namei.h>
21#include "ext3.h"
24#include "xattr.h" 22#include "xattr.h"
25 23
26static void * ext3_follow_link(struct dentry *dentry, struct nameidata *nd) 24static void * ext3_follow_link(struct dentry *dentry, struct nameidata *nd)
diff --git a/fs/ext3/xattr.c b/fs/ext3/xattr.c
index d565759d82ee..d22ebb7a4f55 100644
--- a/fs/ext3/xattr.c
+++ b/fs/ext3/xattr.c
@@ -50,14 +50,9 @@
50 * by the buffer lock. 50 * by the buffer lock.
51 */ 51 */
52 52
53#include <linux/init.h> 53#include "ext3.h"
54#include <linux/fs.h>
55#include <linux/slab.h>
56#include <linux/ext3_jbd.h>
57#include <linux/ext3_fs.h>
58#include <linux/mbcache.h> 54#include <linux/mbcache.h>
59#include <linux/quotaops.h> 55#include <linux/quotaops.h>
60#include <linux/rwsem.h>
61#include "xattr.h" 56#include "xattr.h"
62#include "acl.h" 57#include "acl.h"
63 58
diff --git a/fs/ext3/xattr_security.c b/fs/ext3/xattr_security.c
index ea26f2acab94..3387664ad70e 100644
--- a/fs/ext3/xattr_security.c
+++ b/fs/ext3/xattr_security.c
@@ -3,12 +3,8 @@
3 * Handler for storing security labels as extended attributes. 3 * Handler for storing security labels as extended attributes.
4 */ 4 */
5 5
6#include <linux/slab.h>
7#include <linux/string.h>
8#include <linux/fs.h>
9#include <linux/ext3_jbd.h>
10#include <linux/ext3_fs.h>
11#include <linux/security.h> 6#include <linux/security.h>
7#include "ext3.h"
12#include "xattr.h" 8#include "xattr.h"
13 9
14static size_t 10static size_t
diff --git a/fs/ext3/xattr_trusted.c b/fs/ext3/xattr_trusted.c
index 2526a8829de8..d75727cc67fa 100644
--- a/fs/ext3/xattr_trusted.c
+++ b/fs/ext3/xattr_trusted.c
@@ -5,11 +5,7 @@
5 * Copyright (C) 2003 by Andreas Gruenbacher, <a.gruenbacher@computer.org> 5 * Copyright (C) 2003 by Andreas Gruenbacher, <a.gruenbacher@computer.org>
6 */ 6 */
7 7
8#include <linux/string.h> 8#include "ext3.h"
9#include <linux/capability.h>
10#include <linux/fs.h>
11#include <linux/ext3_jbd.h>
12#include <linux/ext3_fs.h>
13#include "xattr.h" 9#include "xattr.h"
14 10
15static size_t 11static size_t
diff --git a/fs/ext3/xattr_user.c b/fs/ext3/xattr_user.c
index b32e473a1e33..5612af3567e0 100644
--- a/fs/ext3/xattr_user.c
+++ b/fs/ext3/xattr_user.c
@@ -5,10 +5,7 @@
5 * Copyright (C) 2001 by Andreas Gruenbacher, <a.gruenbacher@computer.org> 5 * Copyright (C) 2001 by Andreas Gruenbacher, <a.gruenbacher@computer.org>
6 */ 6 */
7 7
8#include <linux/string.h> 8#include "ext3.h"
9#include <linux/fs.h>
10#include <linux/ext3_jbd.h>
11#include <linux/ext3_fs.h>
12#include "xattr.h" 9#include "xattr.h"
13 10
14static size_t 11static size_t
diff --git a/fs/gfs2/file.c b/fs/gfs2/file.c
index 76834587a8a4..a3d2c9ee8d66 100644
--- a/fs/gfs2/file.c
+++ b/fs/gfs2/file.c
@@ -18,7 +18,6 @@
18#include <linux/mount.h> 18#include <linux/mount.h>
19#include <linux/fs.h> 19#include <linux/fs.h>
20#include <linux/gfs2_ondisk.h> 20#include <linux/gfs2_ondisk.h>
21#include <linux/ext2_fs.h>
22#include <linux/falloc.h> 21#include <linux/falloc.h>
23#include <linux/swap.h> 22#include <linux/swap.h>
24#include <linux/crc32.h> 23#include <linux/crc32.h>
diff --git a/fs/namei.c b/fs/namei.c
index e615ff37e27d..1898198abc3d 100644
--- a/fs/namei.c
+++ b/fs/namei.c
@@ -1054,53 +1054,65 @@ static void follow_dotdot(struct nameidata *nd)
1054} 1054}
1055 1055
1056/* 1056/*
1057 * Allocate a dentry with name and parent, and perform a parent 1057 * This looks up the name in dcache, possibly revalidates the old dentry and
1058 * directory ->lookup on it. Returns the new dentry, or ERR_PTR 1058 * allocates a new one if not found or not valid. In the need_lookup argument
1059 * on error. parent->d_inode->i_mutex must be held. d_lookup must 1059 * returns whether i_op->lookup is necessary.
1060 * have verified that no child exists while under i_mutex. 1060 *
1061 * dir->d_inode->i_mutex must be held
1061 */ 1062 */
1062static struct dentry *d_alloc_and_lookup(struct dentry *parent, 1063static struct dentry *lookup_dcache(struct qstr *name, struct dentry *dir,
1063 struct qstr *name, struct nameidata *nd) 1064 struct nameidata *nd, bool *need_lookup)
1064{ 1065{
1065 struct inode *inode = parent->d_inode;
1066 struct dentry *dentry; 1066 struct dentry *dentry;
1067 struct dentry *old; 1067 int error;
1068 1068
1069 /* Don't create child dentry for a dead directory. */ 1069 *need_lookup = false;
1070 if (unlikely(IS_DEADDIR(inode))) 1070 dentry = d_lookup(dir, name);
1071 return ERR_PTR(-ENOENT); 1071 if (dentry) {
1072 if (d_need_lookup(dentry)) {
1073 *need_lookup = true;
1074 } else if (dentry->d_flags & DCACHE_OP_REVALIDATE) {
1075 error = d_revalidate(dentry, nd);
1076 if (unlikely(error <= 0)) {
1077 if (error < 0) {
1078 dput(dentry);
1079 return ERR_PTR(error);
1080 } else if (!d_invalidate(dentry)) {
1081 dput(dentry);
1082 dentry = NULL;
1083 }
1084 }
1085 }
1086 }
1072 1087
1073 dentry = d_alloc(parent, name); 1088 if (!dentry) {
1074 if (unlikely(!dentry)) 1089 dentry = d_alloc(dir, name);
1075 return ERR_PTR(-ENOMEM); 1090 if (unlikely(!dentry))
1091 return ERR_PTR(-ENOMEM);
1076 1092
1077 old = inode->i_op->lookup(inode, dentry, nd); 1093 *need_lookup = true;
1078 if (unlikely(old)) {
1079 dput(dentry);
1080 dentry = old;
1081 } 1094 }
1082 return dentry; 1095 return dentry;
1083} 1096}
1084 1097
1085/* 1098/*
1086 * We already have a dentry, but require a lookup to be performed on the parent 1099 * Call i_op->lookup on the dentry. The dentry must be negative but may be
1087 * directory to fill in d_inode. Returns the new dentry, or ERR_PTR on error. 1100 * hashed if it was pouplated with DCACHE_NEED_LOOKUP.
1088 * parent->d_inode->i_mutex must be held. d_lookup must have verified that no 1101 *
1089 * child exists while under i_mutex. 1102 * dir->d_inode->i_mutex must be held
1090 */ 1103 */
1091static struct dentry *d_inode_lookup(struct dentry *parent, struct dentry *dentry, 1104static struct dentry *lookup_real(struct inode *dir, struct dentry *dentry,
1092 struct nameidata *nd) 1105 struct nameidata *nd)
1093{ 1106{
1094 struct inode *inode = parent->d_inode;
1095 struct dentry *old; 1107 struct dentry *old;
1096 1108
1097 /* Don't create child dentry for a dead directory. */ 1109 /* Don't create child dentry for a dead directory. */
1098 if (unlikely(IS_DEADDIR(inode))) { 1110 if (unlikely(IS_DEADDIR(dir))) {
1099 dput(dentry); 1111 dput(dentry);
1100 return ERR_PTR(-ENOENT); 1112 return ERR_PTR(-ENOENT);
1101 } 1113 }
1102 1114
1103 old = inode->i_op->lookup(inode, dentry, nd); 1115 old = dir->i_op->lookup(dir, dentry, nd);
1104 if (unlikely(old)) { 1116 if (unlikely(old)) {
1105 dput(dentry); 1117 dput(dentry);
1106 dentry = old; 1118 dentry = old;
@@ -1108,6 +1120,19 @@ static struct dentry *d_inode_lookup(struct dentry *parent, struct dentry *dentr
1108 return dentry; 1120 return dentry;
1109} 1121}
1110 1122
1123static struct dentry *__lookup_hash(struct qstr *name,
1124 struct dentry *base, struct nameidata *nd)
1125{
1126 bool need_lookup;
1127 struct dentry *dentry;
1128
1129 dentry = lookup_dcache(name, base, nd, &need_lookup);
1130 if (!need_lookup)
1131 return dentry;
1132
1133 return lookup_real(base->d_inode, dentry, nd);
1134}
1135
1111/* 1136/*
1112 * It's more convoluted than I'd like it to be, but... it's still fairly 1137 * It's more convoluted than I'd like it to be, but... it's still fairly
1113 * small and for now I'd prefer to have fast path as straight as possible. 1138 * small and for now I'd prefer to have fast path as straight as possible.
@@ -1139,6 +1164,8 @@ static int do_lookup(struct nameidata *nd, struct qstr *name,
1139 return -ECHILD; 1164 return -ECHILD;
1140 nd->seq = seq; 1165 nd->seq = seq;
1141 1166
1167 if (unlikely(d_need_lookup(dentry)))
1168 goto unlazy;
1142 if (unlikely(dentry->d_flags & DCACHE_OP_REVALIDATE)) { 1169 if (unlikely(dentry->d_flags & DCACHE_OP_REVALIDATE)) {
1143 status = d_revalidate(dentry, nd); 1170 status = d_revalidate(dentry, nd);
1144 if (unlikely(status <= 0)) { 1171 if (unlikely(status <= 0)) {
@@ -1147,8 +1174,6 @@ static int do_lookup(struct nameidata *nd, struct qstr *name,
1147 goto unlazy; 1174 goto unlazy;
1148 } 1175 }
1149 } 1176 }
1150 if (unlikely(d_need_lookup(dentry)))
1151 goto unlazy;
1152 path->mnt = mnt; 1177 path->mnt = mnt;
1153 path->dentry = dentry; 1178 path->dentry = dentry;
1154 if (unlikely(!__follow_mount_rcu(nd, path, inode))) 1179 if (unlikely(!__follow_mount_rcu(nd, path, inode)))
@@ -1163,38 +1188,14 @@ unlazy:
1163 dentry = __d_lookup(parent, name); 1188 dentry = __d_lookup(parent, name);
1164 } 1189 }
1165 1190
1166 if (dentry && unlikely(d_need_lookup(dentry))) { 1191 if (unlikely(!dentry))
1192 goto need_lookup;
1193
1194 if (unlikely(d_need_lookup(dentry))) {
1167 dput(dentry); 1195 dput(dentry);
1168 dentry = NULL; 1196 goto need_lookup;
1169 }
1170retry:
1171 if (unlikely(!dentry)) {
1172 struct inode *dir = parent->d_inode;
1173 BUG_ON(nd->inode != dir);
1174
1175 mutex_lock(&dir->i_mutex);
1176 dentry = d_lookup(parent, name);
1177 if (likely(!dentry)) {
1178 dentry = d_alloc_and_lookup(parent, name, nd);
1179 if (IS_ERR(dentry)) {
1180 mutex_unlock(&dir->i_mutex);
1181 return PTR_ERR(dentry);
1182 }
1183 /* known good */
1184 need_reval = 0;
1185 status = 1;
1186 } else if (unlikely(d_need_lookup(dentry))) {
1187 dentry = d_inode_lookup(parent, dentry, nd);
1188 if (IS_ERR(dentry)) {
1189 mutex_unlock(&dir->i_mutex);
1190 return PTR_ERR(dentry);
1191 }
1192 /* known good */
1193 need_reval = 0;
1194 status = 1;
1195 }
1196 mutex_unlock(&dir->i_mutex);
1197 } 1197 }
1198
1198 if (unlikely(dentry->d_flags & DCACHE_OP_REVALIDATE) && need_reval) 1199 if (unlikely(dentry->d_flags & DCACHE_OP_REVALIDATE) && need_reval)
1199 status = d_revalidate(dentry, nd); 1200 status = d_revalidate(dentry, nd);
1200 if (unlikely(status <= 0)) { 1201 if (unlikely(status <= 0)) {
@@ -1204,12 +1205,10 @@ retry:
1204 } 1205 }
1205 if (!d_invalidate(dentry)) { 1206 if (!d_invalidate(dentry)) {
1206 dput(dentry); 1207 dput(dentry);
1207 dentry = NULL; 1208 goto need_lookup;
1208 need_reval = 1;
1209 goto retry;
1210 } 1209 }
1211 } 1210 }
1212 1211done:
1213 path->mnt = mnt; 1212 path->mnt = mnt;
1214 path->dentry = dentry; 1213 path->dentry = dentry;
1215 err = follow_managed(path, nd->flags); 1214 err = follow_managed(path, nd->flags);
@@ -1221,6 +1220,16 @@ retry:
1221 nd->flags |= LOOKUP_JUMPED; 1220 nd->flags |= LOOKUP_JUMPED;
1222 *inode = path->dentry->d_inode; 1221 *inode = path->dentry->d_inode;
1223 return 0; 1222 return 0;
1223
1224need_lookup:
1225 BUG_ON(nd->inode != parent->d_inode);
1226
1227 mutex_lock(&parent->d_inode->i_mutex);
1228 dentry = __lookup_hash(name, parent, nd);
1229 mutex_unlock(&parent->d_inode->i_mutex);
1230 if (IS_ERR(dentry))
1231 return PTR_ERR(dentry);
1232 goto done;
1224} 1233}
1225 1234
1226static inline int may_lookup(struct nameidata *nd) 1235static inline int may_lookup(struct nameidata *nd)
@@ -1846,59 +1855,6 @@ int vfs_path_lookup(struct dentry *dentry, struct vfsmount *mnt,
1846 return err; 1855 return err;
1847} 1856}
1848 1857
1849static struct dentry *__lookup_hash(struct qstr *name,
1850 struct dentry *base, struct nameidata *nd)
1851{
1852 struct inode *inode = base->d_inode;
1853 struct dentry *dentry;
1854 int err;
1855
1856 err = inode_permission(inode, MAY_EXEC);
1857 if (err)
1858 return ERR_PTR(err);
1859
1860 /*
1861 * Don't bother with __d_lookup: callers are for creat as
1862 * well as unlink, so a lot of the time it would cost
1863 * a double lookup.
1864 */
1865 dentry = d_lookup(base, name);
1866
1867 if (dentry && d_need_lookup(dentry)) {
1868 /*
1869 * __lookup_hash is called with the parent dir's i_mutex already
1870 * held, so we are good to go here.
1871 */
1872 dentry = d_inode_lookup(base, dentry, nd);
1873 if (IS_ERR(dentry))
1874 return dentry;
1875 }
1876
1877 if (dentry && (dentry->d_flags & DCACHE_OP_REVALIDATE)) {
1878 int status = d_revalidate(dentry, nd);
1879 if (unlikely(status <= 0)) {
1880 /*
1881 * The dentry failed validation.
1882 * If d_revalidate returned 0 attempt to invalidate
1883 * the dentry otherwise d_revalidate is asking us
1884 * to return a fail status.
1885 */
1886 if (status < 0) {
1887 dput(dentry);
1888 return ERR_PTR(status);
1889 } else if (!d_invalidate(dentry)) {
1890 dput(dentry);
1891 dentry = NULL;
1892 }
1893 }
1894 }
1895
1896 if (!dentry)
1897 dentry = d_alloc_and_lookup(base, name, nd);
1898
1899 return dentry;
1900}
1901
1902/* 1858/*
1903 * Restricted form of lookup. Doesn't follow links, single-component only, 1859 * Restricted form of lookup. Doesn't follow links, single-component only,
1904 * needs parent already locked. Doesn't follow mounts. 1860 * needs parent already locked. Doesn't follow mounts.
@@ -1924,6 +1880,7 @@ struct dentry *lookup_one_len(const char *name, struct dentry *base, int len)
1924{ 1880{
1925 struct qstr this; 1881 struct qstr this;
1926 unsigned int c; 1882 unsigned int c;
1883 int err;
1927 1884
1928 WARN_ON_ONCE(!mutex_is_locked(&base->d_inode->i_mutex)); 1885 WARN_ON_ONCE(!mutex_is_locked(&base->d_inode->i_mutex));
1929 1886
@@ -1948,6 +1905,10 @@ struct dentry *lookup_one_len(const char *name, struct dentry *base, int len)
1948 return ERR_PTR(err); 1905 return ERR_PTR(err);
1949 } 1906 }
1950 1907
1908 err = inode_permission(base->d_inode, MAY_EXEC);
1909 if (err)
1910 return ERR_PTR(err);
1911
1951 return __lookup_hash(&this, base, NULL); 1912 return __lookup_hash(&this, base, NULL);
1952} 1913}
1953 1914
@@ -2749,7 +2710,7 @@ SYSCALL_DEFINE2(mkdir, const char __user *, pathname, umode_t, mode)
2749 2710
2750/* 2711/*
2751 * The dentry_unhash() helper will try to drop the dentry early: we 2712 * The dentry_unhash() helper will try to drop the dentry early: we
2752 * should have a usage count of 2 if we're the only user of this 2713 * should have a usage count of 1 if we're the only user of this
2753 * dentry, and if that is true (possibly after pruning the dcache), 2714 * dentry, and if that is true (possibly after pruning the dcache),
2754 * then we drop the dentry now. 2715 * then we drop the dentry now.
2755 * 2716 *
diff --git a/fs/ocfs2/ioctl.c b/fs/ocfs2/ioctl.c
index a6fda3c188aa..a1a1bfd652c9 100644
--- a/fs/ocfs2/ioctl.c
+++ b/fs/ocfs2/ioctl.c
@@ -28,8 +28,6 @@
28#include "suballoc.h" 28#include "suballoc.h"
29#include "move_extents.h" 29#include "move_extents.h"
30 30
31#include <linux/ext2_fs.h>
32
33#define o2info_from_user(a, b) \ 31#define o2info_from_user(a, b) \
34 copy_from_user(&(a), (b), sizeof(a)) 32 copy_from_user(&(a), (b), sizeof(a))
35#define o2info_to_user(a, b) \ 33#define o2info_to_user(a, b) \
diff --git a/fs/pstore/inode.c b/fs/pstore/inode.c
index f37c32b94525..50952c9bd06c 100644
--- a/fs/pstore/inode.c
+++ b/fs/pstore/inode.c
@@ -105,26 +105,12 @@ static const struct inode_operations pstore_dir_inode_operations = {
105 .unlink = pstore_unlink, 105 .unlink = pstore_unlink,
106}; 106};
107 107
108static struct inode *pstore_get_inode(struct super_block *sb, 108static struct inode *pstore_get_inode(struct super_block *sb)
109 const struct inode *dir, int mode, dev_t dev)
110{ 109{
111 struct inode *inode = new_inode(sb); 110 struct inode *inode = new_inode(sb);
112
113 if (inode) { 111 if (inode) {
114 inode->i_ino = get_next_ino(); 112 inode->i_ino = get_next_ino();
115 inode->i_uid = inode->i_gid = 0;
116 inode->i_mode = mode;
117 inode->i_atime = inode->i_mtime = inode->i_ctime = CURRENT_TIME; 113 inode->i_atime = inode->i_mtime = inode->i_ctime = CURRENT_TIME;
118 switch (mode & S_IFMT) {
119 case S_IFREG:
120 inode->i_fop = &pstore_file_operations;
121 break;
122 case S_IFDIR:
123 inode->i_op = &pstore_dir_inode_operations;
124 inode->i_fop = &simple_dir_operations;
125 inc_nlink(inode);
126 break;
127 }
128 } 114 }
129 return inode; 115 return inode;
130} 116}
@@ -216,9 +202,11 @@ int pstore_mkfile(enum pstore_type_id type, char *psname, u64 id,
216 return rc; 202 return rc;
217 203
218 rc = -ENOMEM; 204 rc = -ENOMEM;
219 inode = pstore_get_inode(pstore_sb, root->d_inode, S_IFREG | 0444, 0); 205 inode = pstore_get_inode(pstore_sb);
220 if (!inode) 206 if (!inode)
221 goto fail; 207 goto fail;
208 inode->i_mode = S_IFREG | 0444;
209 inode->i_fop = &pstore_file_operations;
222 private = kmalloc(sizeof *private + size, GFP_KERNEL); 210 private = kmalloc(sizeof *private + size, GFP_KERNEL);
223 if (!private) 211 if (!private)
224 goto fail_alloc; 212 goto fail_alloc;
@@ -293,10 +281,12 @@ int pstore_fill_super(struct super_block *sb, void *data, int silent)
293 281
294 parse_options(data); 282 parse_options(data);
295 283
296 inode = pstore_get_inode(sb, NULL, S_IFDIR | 0755, 0); 284 inode = pstore_get_inode(sb);
297 if (inode) { 285 if (inode) {
298 /* override ramfs "dir" options so we catch unlink(2) */ 286 inode->i_mode = S_IFDIR | 0755;
299 inode->i_op = &pstore_dir_inode_operations; 287 inode->i_op = &pstore_dir_inode_operations;
288 inode->i_fop = &simple_dir_operations;
289 inc_nlink(inode);
300 } 290 }
301 sb->s_root = d_make_root(inode); 291 sb->s_root = d_make_root(inode);
302 if (!sb->s_root) 292 if (!sb->s_root)
generated by cgit v1.2.2 (git 2.25.0) at 2025-02-07 17:40:05 -0500