xfs_mount.h 15.1 KB
Newer Older
D
Dave Chinner 已提交
1
// SPDX-License-Identifier: GPL-2.0
L
Linus Torvalds 已提交
2
/*
3 4
 * Copyright (c) 2000-2005 Silicon Graphics, Inc.
 * All Rights Reserved.
L
Linus Torvalds 已提交
5 6 7 8
 */
#ifndef __XFS_MOUNT_H__
#define	__XFS_MOUNT_H__

9
struct xlog;
L
Linus Torvalds 已提交
10
struct xfs_inode;
11
struct xfs_mru_cache;
D
David Chinner 已提交
12
struct xfs_ail;
C
Christoph Hellwig 已提交
13
struct xfs_quotainfo;
14
struct xfs_da_geometry;
C
Christoph Hellwig 已提交
15

16 17 18 19 20 21 22 23 24 25
/* dynamic preallocation free space thresholds, 5% down to 1% */
enum {
	XFS_LOWSP_1_PCNT = 0,
	XFS_LOWSP_2_PCNT,
	XFS_LOWSP_3_PCNT,
	XFS_LOWSP_4_PCNT,
	XFS_LOWSP_5_PCNT,
	XFS_LOWSP_MAX,
};

26 27 28 29 30 31 32
/*
 * Error Configuration
 *
 * Error classes define the subsystem the configuration belongs to.
 * Error numbers define the errors that are configurable.
 */
enum {
33
	XFS_ERR_METADATA,
34 35 36
	XFS_ERR_CLASS_MAX,
};
enum {
37
	XFS_ERR_DEFAULT,
38 39 40
	XFS_ERR_EIO,
	XFS_ERR_ENOSPC,
	XFS_ERR_ENODEV,
41 42 43
	XFS_ERR_ERRNO_MAX,
};

44 45
#define XFS_ERR_RETRY_FOREVER	-1

46 47 48 49 50 51
/*
 * Although retry_timeout is in jiffies which is normally an unsigned long,
 * we limit the retry timeout to 86400 seconds, or one day.  So even a
 * signed 32-bit long is sufficient for a HZ value up to 24855.  Making it
 * signed lets us store the special "-1" value, meaning retry forever.
 */
52 53 54
struct xfs_error_cfg {
	struct xfs_kobj	kobj;
	int		max_retries;
55
	long		retry_timeout;	/* in jiffies, -1 = infinite */
56 57
};

L
Linus Torvalds 已提交
58
typedef struct xfs_mount {
C
Christoph Hellwig 已提交
59
	struct super_block	*m_super;
60 61 62 63 64 65 66 67 68 69 70 71 72 73

	/*
	 * Bitsets of per-fs metadata that have been checked and/or are sick.
	 * Callers must hold m_sb_lock to access these two fields.
	 */
	uint8_t			m_fs_checked;
	uint8_t			m_fs_sick;
	/*
	 * Bitsets of rt metadata that have been checked and/or are sick.
	 * Callers must hold m_sb_lock to access this field.
	 */
	uint8_t			m_rt_checked;
	uint8_t			m_rt_sick;

D
David Chinner 已提交
74
	struct xfs_ail		*m_ail;		/* fs active log item list */
75 76

	struct xfs_sb		m_sb;		/* copy of fs superblock */
E
Eric Sandeen 已提交
77
	spinlock_t		m_sb_lock;	/* sb counter lock */
78
	struct percpu_counter	m_icount;	/* allocated inodes counter */
79
	struct percpu_counter	m_ifree;	/* free inodes counter */
80
	struct percpu_counter	m_fdblocks;	/* free block counter */
81 82 83 84 85 86
	/*
	 * Count of data device blocks reserved for delayed allocations,
	 * including indlen blocks.  Does not include allocated CoW staging
	 * extents or anything related to the rt device.
	 */
	struct percpu_counter	m_delalloc_blks;
87

L
Linus Torvalds 已提交
88
	struct xfs_buf		*m_sb_bp;	/* buffer for superblock */
89 90
	char			*m_rtname;	/* realtime device name */
	char			*m_logname;	/* external log device name */
L
Linus Torvalds 已提交
91 92 93
	int			m_bsize;	/* fs logical block size */
	xfs_agnumber_t		m_agfrotor;	/* last ag where space found */
	xfs_agnumber_t		m_agirotor;	/* last ag dir inode alloced */
E
Eric Sandeen 已提交
94
	spinlock_t		m_agirotor_lock;/* .. and lock protecting it */
L
Linus Torvalds 已提交
95
	xfs_agnumber_t		m_maxagi;	/* highest inode alloc group */
96 97
	uint			m_allocsize_log;/* min write size log bytes */
	uint			m_allocsize_blocks; /* min write size blocks */
98 99
	struct xfs_da_geometry	*m_dir_geo;	/* directory block geometry */
	struct xfs_da_geometry	*m_attr_geo;	/* attribute block geometry */
100
	struct xlog		*m_log;		/* log specific stuff */
D
Darrick J. Wong 已提交
101
	struct xfs_ino_geometry	m_ino_geo;	/* inode geometry */
L
Linus Torvalds 已提交
102 103 104 105
	int			m_logbufs;	/* number of log buffers */
	int			m_logbsize;	/* size of each log buffer */
	uint			m_rsumlevels;	/* rt summary levels */
	uint			m_rsumsize;	/* size of rt summary, bytes */
106 107 108 109 110 111 112
	/*
	 * Optional cache of rt summary level per bitmap block with the
	 * invariant that m_rsum_cache[bbno] <= the minimum i for which
	 * rsum[i][bbno] != 0. Reads and writes are serialized by the rsumip
	 * inode lock.
	 */
	uint8_t			*m_rsum_cache;
L
Linus Torvalds 已提交
113 114 115 116 117 118 119
	struct xfs_inode	*m_rbmip;	/* pointer to bitmap inode */
	struct xfs_inode	*m_rsumip;	/* pointer to summary inode */
	struct xfs_inode	*m_rootip;	/* pointer to root directory */
	struct xfs_quotainfo	*m_quotainfo;	/* disk quota information */
	xfs_buftarg_t		*m_ddev_targp;	/* saves taking the address */
	xfs_buftarg_t		*m_logdev_targp;/* ptr to log device */
	xfs_buftarg_t		*m_rtdev_targp;	/* ptr to rt device */
120 121 122
	uint8_t			m_blkbit_log;	/* blocklog + NBBY */
	uint8_t			m_blkbb_log;	/* blocklog - BBSHIFT */
	uint8_t			m_agno_log;	/* log #ag's */
L
Linus Torvalds 已提交
123 124 125
	uint			m_blockmask;	/* sb_blocksize-1 */
	uint			m_blockwsize;	/* sb_blocksize in words */
	uint			m_blockwmask;	/* blockwsize-1 */
126 127 128 129
	uint			m_alloc_mxr[2];	/* max alloc btree records */
	uint			m_alloc_mnr[2];	/* min alloc btree records */
	uint			m_bmap_dmxr[2];	/* max bmap btree records */
	uint			m_bmap_dmnr[2];	/* min bmap btree records */
130 131
	uint			m_rmap_mxr[2];	/* max rmap btree records */
	uint			m_rmap_mnr[2];	/* min rmap btree records */
132 133
	uint			m_refc_mxr[2];	/* max refc btree records */
	uint			m_refc_mnr[2];	/* min refc btree records */
L
Linus Torvalds 已提交
134 135
	uint			m_ag_maxlevels;	/* XFS_AG_MAXLEVELS */
	uint			m_bm_maxlevels[2]; /* XFS_BM_MAXLEVELS */
136
	uint			m_rmap_maxlevels; /* max rmap btree levels */
137
	uint			m_refc_maxlevels; /* max refcount btree level */
138
	xfs_extlen_t		m_ag_prealloc_blocks; /* reserved ag blocks */
139 140
	uint			m_alloc_set_aside; /* space we can't use */
	uint			m_ag_max_usable; /* max space per AG */
141 142
	struct radix_tree_root	m_perag_tree;	/* per-ag accounting info */
	spinlock_t		m_perag_lock;	/* lock for m_perag_tree */
143
	struct mutex		m_growlock;	/* growfs mutex */
L
Linus Torvalds 已提交
144
	int			m_fixedfsid[2];	/* unchanged for life of FS */
145
	uint64_t		m_flags;	/* global mount flags */
146
	bool			m_finobt_nores; /* no per-AG finobt resv. */
L
Linus Torvalds 已提交
147
	uint			m_qflags;	/* quota status flags */
148
	struct xfs_trans_resv	m_resv;		/* precomputed res values */
149 150 151
	uint64_t		m_resblks;	/* total reserved blocks */
	uint64_t		m_resblks_avail;/* available reserved blocks */
	uint64_t		m_resblks_save;	/* reserved blks @ remount,ro */
L
Linus Torvalds 已提交
152 153
	int			m_dalign;	/* stripe unit */
	int			m_swidth;	/* stripe width */
154
	uint8_t			m_sectbb_log;	/* sectlog - BBSHIFT */
L
Linus Torvalds 已提交
155
	atomic_t		m_active_trans;	/* number trans frozen */
156
	struct xfs_mru_cache	*m_filestream;  /* per-mount filestream data */
157
	struct delayed_work	m_reclaim_work;	/* background inode reclaim */
158 159
	struct delayed_work	m_eofblocks_work; /* background eof blocks
						     trimming */
160 161
	struct delayed_work	m_cowblocks_work; /* background cow blocks
						     trimming */
162
	bool			m_update_sb;	/* sb needs update in mount */
163 164
	int64_t			m_low_space[XFS_LOWSP_MAX];
						/* low free space thresholds */
B
Brian Foster 已提交
165
	struct xfs_kobj		m_kobj;
166
	struct xfs_kobj		m_error_kobj;
167
	struct xfs_kobj		m_error_meta_kobj;
168
	struct xfs_error_cfg	m_error_cfg[XFS_ERR_CLASS_MAX][XFS_ERR_ERRNO_MAX];
169
	struct xstats		m_stats;	/* per-fs stats */
170
	struct ratelimit_state	m_flush_inodes_ratelimit;
171

172
	struct workqueue_struct *m_buf_workqueue;
173
	struct workqueue_struct	*m_unwritten_workqueue;
174
	struct workqueue_struct	*m_cil_workqueue;
D
Dave Chinner 已提交
175
	struct workqueue_struct	*m_reclaim_workqueue;
176
	struct workqueue_struct *m_eofblocks_workqueue;
177
	struct workqueue_struct	*m_sync_workqueue;
178 179 180 181 182 183 184 185 186 187

	/*
	 * Generation of the filesysyem layout.  This is incremented by each
	 * growfs, and used by the pNFS server to ensure the client updates
	 * its view of the block device once it gets a layout that might
	 * reference the newly added blocks.  Does not need to be persistent
	 * as long as we only allow file system size increments, but if we
	 * ever support shrinks it would have to be persisted in addition
	 * to various other kinds of pain inflicted on the pNFS server.
	 */
188
	uint32_t		m_generation;
189

190
	bool			m_always_cow;
191
	bool			m_fail_unmount;
192
#ifdef DEBUG
193 194 195 196 197 198
	/*
	 * Frequency with which errors are injected.  Replaces xfs_etest; the
	 * value stored in here is the inverse of the frequency with which the
	 * error triggers.  1 = always, 2 = half the time, etc.
	 */
	unsigned int		*m_errortag;
199
	struct xfs_kobj		m_errortag_kobj;
200
#endif
L
Linus Torvalds 已提交
201 202
} xfs_mount_t;

D
Darrick J. Wong 已提交
203 204
#define M_IGEO(mp)		(&(mp)->m_ino_geo)

L
Linus Torvalds 已提交
205 206 207
/*
 * Flags for m_flags.
 */
D
David Chinner 已提交
208
#define XFS_MOUNT_WSYNC		(1ULL << 0)	/* for nfs - all metadata ops
L
Linus Torvalds 已提交
209 210
						   must be synchronous except
						   for space allocations */
211
#define XFS_MOUNT_UNMOUNTING	(1ULL << 1)	/* filesystem is unmounting */
D
David Chinner 已提交
212
#define XFS_MOUNT_WAS_CLEAN	(1ULL << 3)
213
#define XFS_MOUNT_FS_SHUTDOWN	(1ULL << 4)	/* atomic stop of all filesystem
L
Linus Torvalds 已提交
214 215
						   operations, typically for
						   disk errors in metadata */
216
#define XFS_MOUNT_DISCARD	(1ULL << 5)	/* discard unused blocks */
217
#define XFS_MOUNT_NOALIGN	(1ULL << 7)	/* turn off stripe alignment
L
Linus Torvalds 已提交
218
						   allocations */
219
#define XFS_MOUNT_ATTR2		(1ULL << 8)	/* allow use of attr2 format */
220
#define XFS_MOUNT_GRPID		(1ULL << 9)	/* group-ID assigned from directory */
221
#define XFS_MOUNT_NORECOVERY	(1ULL << 10)	/* no recovery - dirty fs */
222
#define XFS_MOUNT_ALLOCSIZE	(1ULL << 12)	/* specified allocation size */
223 224
#define XFS_MOUNT_SMALL_INUMS	(1ULL << 14)	/* user wants 32bit inodes */
#define XFS_MOUNT_32BITINODES	(1ULL << 15)	/* inode32 allocator active */
225
#define XFS_MOUNT_NOUUID	(1ULL << 16)	/* ignore uuid during mount */
226
#define XFS_MOUNT_IKEEP		(1ULL << 18)	/* keep empty inode clusters*/
227
#define XFS_MOUNT_SWALLOC	(1ULL << 19)	/* turn on stripe width
L
Linus Torvalds 已提交
228
						 * allocation */
229
#define XFS_MOUNT_RDONLY	(1ULL << 20)	/* read-only fs */
230
#define XFS_MOUNT_DIRSYNC	(1ULL << 21)	/* synchronous directory ops */
231
#define XFS_MOUNT_LARGEIO	(1ULL << 22)	/* report large preferred
232
						 * I/O size in stat() */
233 234
#define XFS_MOUNT_FILESTREAMS	(1ULL << 24)	/* enable the filestreams
						   allocator */
235
#define XFS_MOUNT_NOATTR2	(1ULL << 25)	/* disable use of attr2 format */
236

D
Dave Chinner 已提交
237 238
#define XFS_MOUNT_DAX		(1ULL << 62)	/* TEST ONLY! */

L
Linus Torvalds 已提交
239
/*
240 241
 * Max and min values for mount-option defined I/O
 * preallocation sizes.
L
Linus Torvalds 已提交
242
 */
243
#define XFS_MAX_IO_LOG		30	/* 1G */
L
Linus Torvalds 已提交
244 245
#define XFS_MIN_IO_LOG		PAGE_SHIFT

D
David Chinner 已提交
246 247
#define XFS_LAST_UNMOUNT_WAS_CLEAN(mp)	\
				((mp)->m_flags & XFS_MOUNT_WAS_CLEAN)
L
Linus Torvalds 已提交
248
#define XFS_FORCED_SHUTDOWN(mp)	((mp)->m_flags & XFS_MOUNT_FS_SHUTDOWN)
249 250
void xfs_do_force_shutdown(struct xfs_mount *mp, int flags, char *fname,
		int lnnum);
L
Linus Torvalds 已提交
251
#define xfs_force_shutdown(m,f)	\
252
	xfs_do_force_shutdown(m, f, __FILE__, __LINE__)
L
Linus Torvalds 已提交
253

C
Christoph Hellwig 已提交
254 255 256 257 258 259 260
#define SHUTDOWN_META_IO_ERROR	0x0001	/* write attempt to metadata failed */
#define SHUTDOWN_LOG_IO_ERROR	0x0002	/* write attempt to the log failed */
#define SHUTDOWN_FORCE_UMOUNT	0x0004	/* shutdown from a forced unmount */
#define SHUTDOWN_CORRUPT_INCORE	0x0008	/* corrupt in-memory data structures */
#define SHUTDOWN_REMOTE_REQ	0x0010	/* shutdown came from remote cell */
#define SHUTDOWN_DEVICE_REQ	0x0020	/* failed all paths to the device */

L
Linus Torvalds 已提交
261 262 263
/*
 * Flags for xfs_mountfs
 */
264
#define XFS_MFSI_QUIET		0x40	/* Be silent if mount errors found */
L
Linus Torvalds 已提交
265

266 267
static inline xfs_agnumber_t
xfs_daddr_to_agno(struct xfs_mount *mp, xfs_daddr_t d)
L
Linus Torvalds 已提交
268
{
E
Eric Sandeen 已提交
269
	xfs_rfsblock_t ld = XFS_BB_TO_FSBT(mp, d);
270 271
	do_div(ld, mp->m_sb.sb_agblocks);
	return (xfs_agnumber_t) ld;
L
Linus Torvalds 已提交
272 273
}

274 275
static inline xfs_agblock_t
xfs_daddr_to_agbno(struct xfs_mount *mp, xfs_daddr_t d)
L
Linus Torvalds 已提交
276
{
E
Eric Sandeen 已提交
277
	xfs_rfsblock_t ld = XFS_BB_TO_FSBT(mp, d);
278
	return (xfs_agblock_t) do_div(ld, mp->m_sb.sb_agblocks);
L
Linus Torvalds 已提交
279 280
}

281 282 283 284 285 286 287 288 289 290
/* per-AG block reservation data structures*/
struct xfs_ag_resv {
	/* number of blocks originally reserved here */
	xfs_extlen_t			ar_orig_reserved;
	/* number of blocks reserved here */
	xfs_extlen_t			ar_reserved;
	/* number of blocks originally asked for */
	xfs_extlen_t			ar_asked;
};

291 292
/*
 * Per-ag incore structure, copies of information in agf and agi, to improve the
293
 * performance of allocation group selection.
294 295 296 297 298 299 300 301 302
 */
typedef struct xfs_perag {
	struct xfs_mount *pag_mount;	/* owner filesystem */
	xfs_agnumber_t	pag_agno;	/* AG this structure belongs to */
	atomic_t	pag_ref;	/* perag reference count */
	char		pagf_init;	/* this agf's entry is initialized */
	char		pagi_init;	/* this agi's entry is initialized */
	char		pagf_metadata;	/* the agf is preferred to be metadata */
	char		pagi_inodeok;	/* The agi is ok for inodes */
303
	uint8_t		pagf_levels[XFS_BTNUM_AGF];
304
					/* # of levels in bno & cnt btree */
305
	bool		pagf_agflreset; /* agfl requires reset before use */
306
	uint32_t	pagf_flcount;	/* count of blocks in freelist */
307 308
	xfs_extlen_t	pagf_freeblks;	/* total free blocks */
	xfs_extlen_t	pagf_longest;	/* longest free space */
309
	uint32_t	pagf_btreeblks;	/* # of blocks held in AGF btrees */
310 311 312 313 314 315 316 317 318 319 320
	xfs_agino_t	pagi_freecount;	/* number of free inodes */
	xfs_agino_t	pagi_count;	/* number of allocated inodes */

	/*
	 * Inode allocation search lookup optimisation.
	 * If the pagino matches, the search for new inodes
	 * doesn't need to search the near ones again straight away
	 */
	xfs_agino_t	pagl_pagino;
	xfs_agino_t	pagl_leftrec;
	xfs_agino_t	pagl_rightrec;
321 322 323 324 325 326 327 328 329

	/*
	 * Bitsets of per-ag metadata that have been checked and/or are sick.
	 * Callers should hold pag_state_lock before accessing this field.
	 */
	uint16_t	pag_checked;
	uint16_t	pag_sick;
	spinlock_t	pag_state_lock;

330 331
	spinlock_t	pagb_lock;	/* lock for pagb_tree */
	struct rb_root	pagb_tree;	/* ordered tree of busy extents */
332 333
	unsigned int	pagb_gen;	/* generation count for pagb_tree */
	wait_queue_head_t pagb_wait;	/* woken when pagb_gen changes */
334 335 336 337 338 339 340 341 342 343

	atomic_t        pagf_fstrms;    /* # of filestreams active in this AG */

	spinlock_t	pag_ici_lock;	/* incore inode cache lock */
	struct radix_tree_root pag_ici_root;	/* incore inode cache root */
	int		pag_ici_reclaimable;	/* reclaimable inodes */
	struct mutex	pag_ici_reclaim_lock;	/* serialisation point */
	unsigned long	pag_ici_reclaim_cursor;	/* reclaim restart point */

	/* buffer cache index */
344 345
	spinlock_t	pag_buf_lock;	/* lock for pag_buf_hash */
	struct rhashtable pag_buf_hash;
346 347 348 349

	/* for rcu-safe freeing */
	struct rcu_head	rcu_head;
	int		pagb_count;	/* pagb slots in use */
350 351 352

	/* Blocks reserved for all kinds of metadata. */
	struct xfs_ag_resv	pag_meta_resv;
353 354
	/* Blocks reserved for the reverse mapping btree. */
	struct xfs_ag_resv	pag_rmapbt_resv;
355 356

	/* reference count */
357
	uint8_t			pagf_refcount_level;
358 359 360 361 362 363 364

	/*
	 * Unlinked inode information.  This incore information reflects
	 * data stored in the AGI, so callers must hold the AGI buffer lock
	 * or have some other means to control concurrency.
	 */
	struct rhashtable	pagi_unlinked_hash;
365 366
} xfs_perag_t;

367 368 369 370 371 372 373 374
static inline struct xfs_ag_resv *
xfs_perag_resv(
	struct xfs_perag	*pag,
	enum xfs_ag_resv_type	type)
{
	switch (type) {
	case XFS_AG_RESV_METADATA:
		return &pag->pag_meta_resv;
375 376
	case XFS_AG_RESV_RMAPBT:
		return &pag->pag_rmapbt_resv;
377 378 379 380 381
	default:
		return NULL;
	}
}

382 383 384
int xfs_buf_hash_init(xfs_perag_t *pag);
void xfs_buf_hash_destroy(xfs_perag_t *pag);

385
extern void	xfs_uuid_table_free(void);
386
extern int	xfs_log_sbcount(xfs_mount_t *);
387
extern uint64_t xfs_default_resblks(xfs_mount_t *mp);
C
Christoph Hellwig 已提交
388
extern int	xfs_mountfs(xfs_mount_t *mp);
389 390
extern int	xfs_initialize_perag(xfs_mount_t *mp, xfs_agnumber_t agcount,
				     xfs_agnumber_t *maxagi);
391
extern void	xfs_unmountfs(xfs_mount_t *);
D
Dave Chinner 已提交
392

393
extern int	xfs_mod_icount(struct xfs_mount *mp, int64_t delta);
394
extern int	xfs_mod_ifree(struct xfs_mount *mp, int64_t delta);
395 396
extern int	xfs_mod_fdblocks(struct xfs_mount *mp, int64_t delta,
				 bool reserved);
D
Dave Chinner 已提交
397 398
extern int	xfs_mod_frextents(struct xfs_mount *mp, int64_t delta);

399
extern struct xfs_buf *xfs_getsb(xfs_mount_t *);
400
extern int	xfs_readsb(xfs_mount_t *, int);
L
Linus Torvalds 已提交
401
extern void	xfs_freesb(xfs_mount_t *);
402
extern bool	xfs_fs_writable(struct xfs_mount *mp, int level);
403
extern int	xfs_sb_validate_fsb_count(struct xfs_sb *, uint64_t);
L
Linus Torvalds 已提交
404

C
Christoph Hellwig 已提交
405 406
extern int	xfs_dev_is_read_only(struct xfs_mount *, char *);

407 408
extern void	xfs_set_low_space_thresholds(struct xfs_mount *);

409 410 411
int	xfs_zero_extent(struct xfs_inode *ip, xfs_fsblock_t start_fsb,
			xfs_off_t count_fsb);

412 413
struct xfs_error_cfg * xfs_error_get_cfg(struct xfs_mount *mp,
		int error_class, int error);
414
void xfs_force_summary_recalc(struct xfs_mount *mp);
415
void xfs_mod_delalloc(struct xfs_mount *mp, int64_t delta);
416

L
Linus Torvalds 已提交
417
#endif	/* __XFS_MOUNT_H__ */