2 * Copyright (c) 2000-2003,2005 Silicon Graphics, Inc.
5 * This program is free software; you can redistribute it and/or
6 * modify it under the terms of the GNU General Public License as
7 * published by the Free Software Foundation.
9 * This program is distributed in the hope that it would be useful,
10 * but WITHOUT ANY WARRANTY; without even the implied warranty of
11 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
12 * GNU General Public License for more details.
14 * You should have received a copy of the GNU General Public License
15 * along with this program; if not, write the Free Software Foundation,
16 * Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA
18 #ifndef __XFS_INODE_H__
19 #define __XFS_INODE_H__
24 #define XFS_DATA_FORK 0
25 #define XFS_ATTR_FORK 1
28 * The following xfs_ext_irec_t struct introduces a second (top) level
29 * to the in-core extent allocation scheme. These structs are allocated
30 * in a contiguous block, creating an indirection array where each entry
31 * (irec) contains a pointer to a buffer of in-core extent records which
32 * it manages. Each extent buffer is 4k in size, since 4k is the system
33 * page size on Linux i386 and systems with larger page sizes don't seem
34 * to gain much, if anything, by using their native page size as the
35 * extent buffer size. Also, using 4k extent buffers everywhere provides
36 * a consistent interface for CXFS across different platforms.
38 * There is currently no limit on the number of irec's (extent lists)
39 * allowed, so heavily fragmented files may require an indirection array
40 * which spans multiple system pages of memory. The number of extents
41 * which would require this amount of contiguous memory is very large
42 * and should not cause problems in the foreseeable future. However,
43 * if the memory needed for the contiguous array ever becomes a problem,
44 * it is possible that a third level of indirection may be required.
46 typedef struct xfs_ext_irec {
47 xfs_bmbt_rec_t *er_extbuf; /* block of extent records */
48 xfs_extnum_t er_extoff; /* extent offset in file */
49 xfs_extnum_t er_extcount; /* number of extents in page/block */
53 * File incore extent information, present for each of data & attr forks.
55 #define XFS_IEXT_BUFSZ 4096
56 #define XFS_LINEAR_EXTS (XFS_IEXT_BUFSZ / (uint)sizeof(xfs_bmbt_rec_t))
57 #define XFS_INLINE_EXTS 2
58 #define XFS_INLINE_DATA 32
59 typedef struct xfs_ifork {
60 int if_bytes; /* bytes in if_u1 */
61 int if_real_bytes; /* bytes allocated in if_u1 */
62 xfs_bmbt_block_t *if_broot; /* file's incore btree root */
63 short if_broot_bytes; /* bytes allocated for root */
64 unsigned char if_flags; /* per-fork flags */
65 unsigned char if_ext_max; /* max # of extent records */
66 xfs_extnum_t if_lastex; /* last if_extents used */
68 xfs_bmbt_rec_t *if_extents; /* linear map file exts */
69 xfs_ext_irec_t *if_ext_irec; /* irec map file exts */
70 char *if_data; /* inline file data */
73 xfs_bmbt_rec_t if_inline_ext[XFS_INLINE_EXTS];
74 /* very small file extents */
75 char if_inline_data[XFS_INLINE_DATA];
76 /* very small file data */
77 xfs_dev_t if_rdev; /* dev number if special */
78 uuid_t if_uuid; /* mount point value */
83 * Flags for xfs_ichgtime().
85 #define XFS_ICHGTIME_MOD 0x1 /* data fork modification timestamp */
86 #define XFS_ICHGTIME_ACC 0x2 /* data fork access timestamp */
87 #define XFS_ICHGTIME_CHG 0x4 /* inode field change timestamp */
90 * Per-fork incore inode flags.
92 #define XFS_IFINLINE 0x01 /* Inline data is read in */
93 #define XFS_IFEXTENTS 0x02 /* All extent pointers are read in */
94 #define XFS_IFBROOT 0x04 /* i_broot points to the bmap b-tree root */
95 #define XFS_IFEXTIREC 0x08 /* Indirection array of extent blocks */
98 * Flags for xfs_itobp(), xfs_imap() and xfs_dilocate().
100 #define XFS_IMAP_LOOKUP 0x1
101 #define XFS_IMAP_BULKSTAT 0x2
109 struct xfs_bmap_free;
110 struct xfs_bmbt_irec;
111 struct xfs_bmbt_block;
113 struct xfs_inode_log_item;
118 #if defined(XFS_ILOCK_TRACE)
119 #define XFS_ILOCK_KTRACE_SIZE 32
120 extern ktrace_t *xfs_ilock_trace_buf;
121 extern void xfs_ilock_trace(struct xfs_inode *, int, unsigned int, inst_t *);
123 #define xfs_ilock_trace(i,n,f,ra)
126 typedef struct dm_attrs_s {
127 __uint32_t da_dmevmask; /* DMIG event mask */
128 __uint16_t da_dmstate; /* DMIG state info */
129 __uint16_t da_pad; /* DMIG extra padding */
132 typedef struct xfs_iocore {
133 void *io_obj; /* pointer to container
134 * inode or dcxvn structure */
135 struct xfs_mount *io_mount; /* fs mount struct ptr */
137 mrlock_t *io_lock; /* inode IO lock */
138 mrlock_t *io_iolock; /* inode IO lock */
142 xfs_fsize_t io_new_size; /* sz when write completes */
144 /* Miscellaneous state. */
145 unsigned int io_flags; /* IO related flags */
148 dm_attrs_t io_dmattrs;
152 #define io_dmevmask io_dmattrs.da_dmevmask
153 #define io_dmstate io_dmattrs.da_dmstate
155 #define XFS_IO_INODE(io) ((xfs_inode_t *) ((io)->io_obj))
156 #define XFS_IO_DCXVN(io) ((dcxvn_t *) ((io)->io_obj))
159 * Flags in the flags field
162 #define XFS_IOCORE_RT 0x1
165 * xfs_iocore prototypes
168 extern void xfs_iocore_inode_init(struct xfs_inode *);
169 extern void xfs_iocore_inode_reinit(struct xfs_inode *);
173 * This is the type used in the xfs inode hash table.
174 * An array of these is allocated for each mounted
175 * file system to hash the inodes for that file system.
177 typedef struct xfs_ihash {
178 struct xfs_inode *ih_next;
183 #define XFS_IHASH(mp,ino) ((mp)->m_ihash + (((uint)(ino)) % (mp)->m_ihsize))
186 * This is the xfs inode cluster hash. This hash is used by xfs_iflush to
187 * find inodes that share a cluster and can be flushed to disk at the same
190 typedef struct xfs_chashlist {
191 struct xfs_chashlist *chl_next;
192 struct xfs_chashlist *chl_prev;
193 struct xfs_inode *chl_ip;
194 xfs_daddr_t chl_blkno; /* starting block number of
196 struct xfs_buf *chl_buf; /* the inode buffer */
199 typedef struct xfs_chash {
200 xfs_chashlist_t *ch_list;
204 #define XFS_CHASH(mp,blk) ((mp)->m_chash + (((uint)blk) % (mp)->m_chsize))
208 * This is the xfs in-core inode structure.
209 * Most of the on-disk inode is embedded in the i_d field.
211 * The extent pointers/inline file space, however, are managed
212 * separately. The memory for this information is pointed to by
213 * the if_u1 unions depending on the type of the data.
214 * This is used to linearize the array of extents for fast in-core
215 * access. This is used until the file's number of extents
216 * surpasses XFS_MAX_INCORE_EXTENTS, at which point all extent pointers
217 * are accessed through the buffer cache.
219 * Other state kept in the in-core inode is used for identification,
220 * locking, transactional updating, etc of the inode.
222 * Generally, we do not want to hold the i_rlock while holding the
223 * i_ilock. Hierarchy is i_iolock followed by i_rlock.
225 * xfs_iptr_t contains all the inode fields upto and including the
226 * i_mnext and i_mprev fields, it is used as a marker in the inode
227 * chain off the mount structure by xfs_sync calls.
231 struct xfs_ihash *ip_hash; /* pointer to hash header */
232 struct xfs_inode *ip_next; /* inode hash link forw */
233 struct xfs_inode *ip_mnext; /* next inode in mount list */
234 struct xfs_inode *ip_mprev; /* ptr to prev inode */
235 struct xfs_inode **ip_prevp; /* ptr to prev i_next */
236 struct xfs_mount *ip_mount; /* fs mount struct ptr */
239 typedef struct xfs_inode {
240 /* Inode linking and identification information. */
241 struct xfs_ihash *i_hash; /* pointer to hash header */
242 struct xfs_inode *i_next; /* inode hash link forw */
243 struct xfs_inode *i_mnext; /* next inode in mount list */
244 struct xfs_inode *i_mprev; /* ptr to prev inode */
245 struct xfs_inode **i_prevp; /* ptr to prev i_next */
246 struct xfs_mount *i_mount; /* fs mount struct ptr */
247 struct list_head i_reclaim; /* reclaim list */
248 struct bhv_desc i_bhv_desc; /* inode behavior descriptor*/
249 struct xfs_dquot *i_udquot; /* user dquot */
250 struct xfs_dquot *i_gdquot; /* group dquot */
252 /* Inode location stuff */
253 xfs_ino_t i_ino; /* inode number (agno/agino)*/
254 xfs_daddr_t i_blkno; /* blkno of inode buffer */
255 ushort i_len; /* len of inode buffer */
256 ushort i_boffset; /* off of inode in buffer */
258 /* Extent information. */
259 xfs_ifork_t *i_afp; /* attribute fork pointer */
260 xfs_ifork_t i_df; /* data fork */
262 /* Transaction and locking information. */
263 struct xfs_trans *i_transp; /* ptr to owning transaction*/
264 struct xfs_inode_log_item *i_itemp; /* logging information */
265 mrlock_t i_lock; /* inode lock */
266 mrlock_t i_iolock; /* inode IO lock */
267 sema_t i_flock; /* inode flush lock */
268 atomic_t i_pincount; /* inode pin count */
269 wait_queue_head_t i_ipin_wait; /* inode pinning wait queue */
271 struct xfs_inode **i_refcache; /* ptr to entry in ref cache */
272 struct xfs_inode *i_release; /* inode to unref */
275 xfs_iocore_t i_iocore; /* I/O core */
277 /* Miscellaneous state. */
278 unsigned short i_flags; /* see defined flags below */
279 unsigned char i_update_core; /* timestamps/size is dirty */
280 unsigned char i_update_size; /* di_size field is dirty */
281 unsigned int i_gen; /* generation count */
282 unsigned int i_delayed_blks; /* count of delay alloc blks */
284 xfs_dinode_core_t i_d; /* most of ondisk inode */
285 xfs_chashlist_t *i_chash; /* cluster hash list header */
286 struct xfs_inode *i_cnext; /* cluster hash link forward */
287 struct xfs_inode *i_cprev; /* cluster hash link backward */
289 /* Trace buffers per inode. */
290 #ifdef XFS_BMAP_TRACE
291 struct ktrace *i_xtrace; /* inode extent list trace */
293 #ifdef XFS_BMBT_TRACE
294 struct ktrace *i_btrace; /* inode bmap btree trace */
297 struct ktrace *i_rwtrace; /* inode read/write trace */
299 #ifdef XFS_ILOCK_TRACE
300 struct ktrace *i_lock_trace; /* inode lock/unlock trace */
302 #ifdef XFS_DIR2_TRACE
303 struct ktrace *i_dir_trace; /* inode directory trace */
307 #endif /* __KERNEL__ */
313 #define XFS_IFORK_PTR(ip,w) \
314 ((w) == XFS_DATA_FORK ? &(ip)->i_df : (ip)->i_afp)
315 #define XFS_IFORK_Q(ip) XFS_CFORK_Q(&(ip)->i_d)
316 #define XFS_IFORK_DSIZE(ip) XFS_CFORK_DSIZE(&ip->i_d, ip->i_mount)
317 #define XFS_IFORK_ASIZE(ip) XFS_CFORK_ASIZE(&ip->i_d, ip->i_mount)
318 #define XFS_IFORK_SIZE(ip,w) XFS_CFORK_SIZE(&ip->i_d, ip->i_mount, w)
319 #define XFS_IFORK_FORMAT(ip,w) XFS_CFORK_FORMAT(&ip->i_d, w)
320 #define XFS_IFORK_FMT_SET(ip,w,n) XFS_CFORK_FMT_SET(&ip->i_d, w, n)
321 #define XFS_IFORK_NEXTENTS(ip,w) XFS_CFORK_NEXTENTS(&ip->i_d, w)
322 #define XFS_IFORK_NEXT_SET(ip,w,n) XFS_CFORK_NEXT_SET(&ip->i_d, w, n)
328 * In-core inode flags.
330 #define XFS_IGRIO 0x0001 /* inode used for guaranteed rate i/o */
331 #define XFS_IUIOSZ 0x0002 /* inode i/o sizes have been explicitly set */
332 #define XFS_IQUIESCE 0x0004 /* we have started quiescing for this inode */
333 #define XFS_IRECLAIM 0x0008 /* we have started reclaiming this inode */
334 #define XFS_ISTALE 0x0010 /* inode has been staled */
335 #define XFS_IRECLAIMABLE 0x0020 /* inode can be reclaimed */
336 #define XFS_INEW 0x0040
339 * Flags for inode locking.
341 #define XFS_IOLOCK_EXCL 0x001
342 #define XFS_IOLOCK_SHARED 0x002
343 #define XFS_ILOCK_EXCL 0x004
344 #define XFS_ILOCK_SHARED 0x008
345 #define XFS_IUNLOCK_NONOTIFY 0x010
346 /* XFS_IOLOCK_NESTED 0x020 */
347 #define XFS_EXTENT_TOKEN_RD 0x040
348 #define XFS_SIZE_TOKEN_RD 0x080
349 #define XFS_EXTSIZE_RD (XFS_EXTENT_TOKEN_RD|XFS_SIZE_TOKEN_RD)
350 #define XFS_WILLLEND 0x100 /* Always acquire tokens for lending */
351 #define XFS_EXTENT_TOKEN_WR (XFS_EXTENT_TOKEN_RD | XFS_WILLLEND)
352 #define XFS_SIZE_TOKEN_WR (XFS_SIZE_TOKEN_RD | XFS_WILLLEND)
353 #define XFS_EXTSIZE_WR (XFS_EXTSIZE_RD | XFS_WILLLEND)
354 /* XFS_SIZE_TOKEN_WANT 0x200 */
356 #define XFS_LOCK_MASK \
357 (XFS_IOLOCK_EXCL | XFS_IOLOCK_SHARED | XFS_ILOCK_EXCL | \
358 XFS_ILOCK_SHARED | XFS_EXTENT_TOKEN_RD | XFS_SIZE_TOKEN_RD | \
362 * Flags for xfs_iflush()
364 #define XFS_IFLUSH_DELWRI_ELSE_SYNC 1
365 #define XFS_IFLUSH_DELWRI_ELSE_ASYNC 2
366 #define XFS_IFLUSH_SYNC 3
367 #define XFS_IFLUSH_ASYNC 4
368 #define XFS_IFLUSH_DELWRI 5
371 * Flags for xfs_itruncate_start().
373 #define XFS_ITRUNC_DEFINITE 0x1
374 #define XFS_ITRUNC_MAYBE 0x2
376 #define XFS_ITOV(ip) BHV_TO_VNODE(XFS_ITOBHV(ip))
377 #define XFS_ITOV_NULL(ip) BHV_TO_VNODE_NULL(XFS_ITOBHV(ip))
378 #define XFS_ITOBHV(ip) ((struct bhv_desc *)(&((ip)->i_bhv_desc)))
379 #define XFS_BHVTOI(bhvp) ((xfs_inode_t *)((char *)(bhvp) - \
380 (char *)&(((xfs_inode_t *)0)->i_bhv_desc)))
381 #define BHV_IS_XFS(bdp) (BHV_OPS(bdp) == &xfs_vnodeops)
384 * For multiple groups support: if S_ISGID bit is set in the parent
385 * directory, group of new file is set to that of the parent, and
386 * new subdirectory gets S_ISGID bit from parent.
388 #define XFS_INHERIT_GID(pip, vfsp) \
389 (((vfsp)->vfs_flag & VFS_GRPID) || ((pip)->i_d.di_mode & S_ISGID))
392 * xfs_iget.c prototypes.
395 #define IGET_CREATE 1
397 void xfs_ihash_init(struct xfs_mount *);
398 void xfs_ihash_free(struct xfs_mount *);
399 void xfs_chash_init(struct xfs_mount *);
400 void xfs_chash_free(struct xfs_mount *);
401 xfs_inode_t *xfs_inode_incore(struct xfs_mount *, xfs_ino_t,
403 void xfs_inode_lock_init(xfs_inode_t *, struct vnode *);
404 int xfs_iget(struct xfs_mount *, struct xfs_trans *, xfs_ino_t,
405 uint, uint, xfs_inode_t **, xfs_daddr_t);
406 void xfs_iput(xfs_inode_t *, uint);
407 void xfs_iput_new(xfs_inode_t *, uint);
408 void xfs_ilock(xfs_inode_t *, uint);
409 int xfs_ilock_nowait(xfs_inode_t *, uint);
410 void xfs_iunlock(xfs_inode_t *, uint);
411 void xfs_ilock_demote(xfs_inode_t *, uint);
412 void xfs_iflock(xfs_inode_t *);
413 int xfs_iflock_nowait(xfs_inode_t *);
414 uint xfs_ilock_map_shared(xfs_inode_t *);
415 void xfs_iunlock_map_shared(xfs_inode_t *, uint);
416 void xfs_ifunlock(xfs_inode_t *);
417 void xfs_ireclaim(xfs_inode_t *);
418 int xfs_finish_reclaim(xfs_inode_t *, int, int);
419 int xfs_finish_reclaim_all(struct xfs_mount *, int);
422 * xfs_inode.c prototypes.
424 int xfs_itobp(struct xfs_mount *, struct xfs_trans *,
425 xfs_inode_t *, xfs_dinode_t **, struct xfs_buf **,
427 int xfs_iread(struct xfs_mount *, struct xfs_trans *, xfs_ino_t,
428 xfs_inode_t **, xfs_daddr_t);
429 int xfs_iread_extents(struct xfs_trans *, xfs_inode_t *, int);
430 int xfs_ialloc(struct xfs_trans *, xfs_inode_t *, mode_t,
431 xfs_nlink_t, xfs_dev_t, struct cred *, xfs_prid_t,
432 int, struct xfs_buf **, boolean_t *, xfs_inode_t **);
433 void xfs_xlate_dinode_core(xfs_caddr_t, struct xfs_dinode_core *,
435 uint xfs_ip2xflags(struct xfs_inode *);
436 uint xfs_dic2xflags(struct xfs_dinode_core *);
437 int xfs_ifree(struct xfs_trans *, xfs_inode_t *,
438 struct xfs_bmap_free *);
439 void xfs_itruncate_start(xfs_inode_t *, uint, xfs_fsize_t);
440 int xfs_itruncate_finish(struct xfs_trans **, xfs_inode_t *,
441 xfs_fsize_t, int, int);
442 int xfs_iunlink(struct xfs_trans *, xfs_inode_t *);
443 int xfs_igrow_start(xfs_inode_t *, xfs_fsize_t, struct cred *);
444 void xfs_igrow_finish(struct xfs_trans *, xfs_inode_t *,
447 void xfs_idestroy_fork(xfs_inode_t *, int);
448 void xfs_idestroy(xfs_inode_t *);
449 void xfs_idata_realloc(xfs_inode_t *, int, int);
450 void xfs_iextract(xfs_inode_t *);
451 void xfs_iext_realloc(xfs_inode_t *, int, int);
452 void xfs_iroot_realloc(xfs_inode_t *, int, int);
453 void xfs_ipin(xfs_inode_t *);
454 void xfs_iunpin(xfs_inode_t *);
455 int xfs_iextents_copy(xfs_inode_t *, xfs_bmbt_rec_t *, int);
456 int xfs_iflush(xfs_inode_t *, uint);
457 void xfs_iflush_all(struct xfs_mount *);
458 int xfs_iaccess(xfs_inode_t *, mode_t, cred_t *);
459 uint xfs_iroundup(uint);
460 void xfs_ichgtime(xfs_inode_t *, int);
461 xfs_fsize_t xfs_file_last_byte(xfs_inode_t *);
462 void xfs_lock_inodes(xfs_inode_t **, int, int, uint);
464 xfs_inode_t *xfs_vtoi(struct vnode *vp);
466 void xfs_synchronize_atime(xfs_inode_t *);
468 xfs_bmbt_rec_t *xfs_iext_get_ext(xfs_ifork_t *, xfs_extnum_t);
469 void xfs_iext_insert(xfs_ifork_t *, xfs_extnum_t, xfs_extnum_t,
471 void xfs_iext_add(xfs_ifork_t *, xfs_extnum_t, int);
472 void xfs_iext_add_indirect_multi(xfs_ifork_t *, int, xfs_extnum_t, int);
473 void xfs_iext_remove(xfs_ifork_t *, xfs_extnum_t, int);
474 void xfs_iext_remove_inline(xfs_ifork_t *, xfs_extnum_t, int);
475 void xfs_iext_remove_direct(xfs_ifork_t *, xfs_extnum_t, int);
476 void xfs_iext_remove_indirect(xfs_ifork_t *, xfs_extnum_t, int);
477 void xfs_iext_realloc_direct(xfs_ifork_t *, int);
478 void xfs_iext_realloc_indirect(xfs_ifork_t *, int);
479 void xfs_iext_indirect_to_direct(xfs_ifork_t *);
480 void xfs_iext_direct_to_inline(xfs_ifork_t *, xfs_extnum_t);
481 void xfs_iext_inline_to_direct(xfs_ifork_t *, int);
482 void xfs_iext_destroy(xfs_ifork_t *);
483 xfs_bmbt_rec_t *xfs_iext_bno_to_ext(xfs_ifork_t *, xfs_fileoff_t, int *);
484 xfs_ext_irec_t *xfs_iext_bno_to_irec(xfs_ifork_t *, xfs_fileoff_t, int *);
485 xfs_ext_irec_t *xfs_iext_idx_to_irec(xfs_ifork_t *, xfs_extnum_t *, int *, int);
486 void xfs_iext_irec_init(xfs_ifork_t *);
487 xfs_ext_irec_t *xfs_iext_irec_new(xfs_ifork_t *, int);
488 void xfs_iext_irec_remove(xfs_ifork_t *, int);
489 void xfs_iext_irec_compact(xfs_ifork_t *);
490 void xfs_iext_irec_compact_pages(xfs_ifork_t *);
491 void xfs_iext_irec_compact_full(xfs_ifork_t *);
492 void xfs_iext_irec_update_extoffs(xfs_ifork_t *, int, int);
494 #define xfs_ipincount(ip) ((unsigned int) atomic_read(&ip->i_pincount))
497 void xfs_isize_check(struct xfs_mount *, xfs_inode_t *, xfs_fsize_t);
499 #define xfs_isize_check(mp, ip, isize)
503 void xfs_inobp_check(struct xfs_mount *, struct xfs_buf *);
505 #define xfs_inobp_check(mp, bp)
508 extern struct kmem_zone *xfs_chashlist_zone;
509 extern struct kmem_zone *xfs_ifork_zone;
510 extern struct kmem_zone *xfs_inode_zone;
511 extern struct kmem_zone *xfs_ili_zone;
512 extern struct vnodeops xfs_vnodeops;
514 #endif /* __KERNEL__ */
516 #endif /* __XFS_INODE_H__ */