xref: /kernel/linux/linux-5.10/fs/gfs2/incore.h (revision 8c2ecf20)
1/* SPDX-License-Identifier: GPL-2.0-only */
2/*
3 * Copyright (C) Sistina Software, Inc.  1997-2003 All rights reserved.
4 * Copyright (C) 2004-2008 Red Hat, Inc.  All rights reserved.
5 */
6
7#ifndef __INCORE_DOT_H__
8#define __INCORE_DOT_H__
9
10#include <linux/fs.h>
11#include <linux/kobject.h>
12#include <linux/workqueue.h>
13#include <linux/dlm.h>
14#include <linux/buffer_head.h>
15#include <linux/rcupdate.h>
16#include <linux/rculist_bl.h>
17#include <linux/completion.h>
18#include <linux/rbtree.h>
19#include <linux/ktime.h>
20#include <linux/percpu.h>
21#include <linux/lockref.h>
22#include <linux/rhashtable.h>
23
24#define DIO_WAIT	0x00000010
25#define DIO_METADATA	0x00000020
26
27struct gfs2_log_operations;
28struct gfs2_bufdata;
29struct gfs2_holder;
30struct gfs2_glock;
31struct gfs2_quota_data;
32struct gfs2_trans;
33struct gfs2_jdesc;
34struct gfs2_sbd;
35struct lm_lockops;
36
37typedef void (*gfs2_glop_bh_t) (struct gfs2_glock *gl, unsigned int ret);
38
39struct gfs2_log_header_host {
40	u64 lh_sequence;	/* Sequence number of this transaction */
41	u32 lh_flags;		/* GFS2_LOG_HEAD_... */
42	u32 lh_tail;		/* Block number of log tail */
43	u32 lh_blkno;
44
45	s64 lh_local_total;
46	s64 lh_local_free;
47	s64 lh_local_dinodes;
48};
49
50/*
51 * Structure of operations that are associated with each
52 * type of element in the log.
53 */
54
55struct gfs2_log_operations {
56	void (*lo_before_commit) (struct gfs2_sbd *sdp, struct gfs2_trans *tr);
57	void (*lo_after_commit) (struct gfs2_sbd *sdp, struct gfs2_trans *tr);
58	void (*lo_before_scan) (struct gfs2_jdesc *jd,
59				struct gfs2_log_header_host *head, int pass);
60	int (*lo_scan_elements) (struct gfs2_jdesc *jd, unsigned int start,
61				 struct gfs2_log_descriptor *ld, __be64 *ptr,
62				 int pass);
63	void (*lo_after_scan) (struct gfs2_jdesc *jd, int error, int pass);
64	const char *lo_name;
65};
66
67#define GBF_FULL 1
68
69/**
70 * Clone bitmaps (bi_clone):
71 *
72 * - When a block is freed, we remember the previous state of the block in the
73 *   clone bitmap, and only mark the block as free in the real bitmap.
74 *
75 * - When looking for a block to allocate, we check for a free block in the
76 *   clone bitmap, and if no clone bitmap exists, in the real bitmap.
77 *
78 * - For allocating a block, we mark it as allocated in the real bitmap, and if
79 *   a clone bitmap exists, also in the clone bitmap.
80 *
81 * - At the end of a log_flush, we copy the real bitmap into the clone bitmap
82 *   to make the clone bitmap reflect the current allocation state.
83 *   (Alternatively, we could remove the clone bitmap.)
84 *
85 * The clone bitmaps are in-core only, and is never written to disk.
86 *
87 * These steps ensure that blocks which have been freed in a transaction cannot
88 * be reallocated in that same transaction.
89 */
90struct gfs2_bitmap {
91	struct buffer_head *bi_bh;
92	char *bi_clone;
93	unsigned long bi_flags;
94	u32 bi_offset;
95	u32 bi_start;
96	u32 bi_bytes;
97	u32 bi_blocks;
98};
99
100struct gfs2_rgrpd {
101	struct rb_node rd_node;		/* Link with superblock */
102	struct gfs2_glock *rd_gl;	/* Glock for this rgrp */
103	u64 rd_addr;			/* grp block disk address */
104	u64 rd_data0;			/* first data location */
105	u32 rd_length;			/* length of rgrp header in fs blocks */
106	u32 rd_data;			/* num of data blocks in rgrp */
107	u32 rd_bitbytes;		/* number of bytes in data bitmaps */
108	u32 rd_free;
109	u32 rd_reserved;                /* number of blocks reserved */
110	u32 rd_free_clone;
111	u32 rd_dinodes;
112	u64 rd_igeneration;
113	struct gfs2_bitmap *rd_bits;
114	struct gfs2_sbd *rd_sbd;
115	struct gfs2_rgrp_lvb *rd_rgl;
116	u32 rd_last_alloc;
117	u32 rd_flags;
118	u32 rd_extfail_pt;		/* extent failure point */
119#define GFS2_RDF_CHECK		0x10000000 /* check for unlinked inodes */
120#define GFS2_RDF_UPTODATE	0x20000000 /* rg is up to date */
121#define GFS2_RDF_ERROR		0x40000000 /* error in rg */
122#define GFS2_RDF_PREFERRED	0x80000000 /* This rgrp is preferred */
123#define GFS2_RDF_MASK		0xf0000000 /* mask for internal flags */
124	spinlock_t rd_rsspin;           /* protects reservation related vars */
125	struct rb_root rd_rstree;       /* multi-block reservation tree */
126};
127
128struct gfs2_rbm {
129	struct gfs2_rgrpd *rgd;
130	u32 offset;		/* The offset is bitmap relative */
131	int bii;		/* Bitmap index */
132};
133
134static inline struct gfs2_bitmap *rbm_bi(const struct gfs2_rbm *rbm)
135{
136	return rbm->rgd->rd_bits + rbm->bii;
137}
138
139static inline u64 gfs2_rbm_to_block(const struct gfs2_rbm *rbm)
140{
141	BUG_ON(rbm->offset >= rbm->rgd->rd_data);
142	return rbm->rgd->rd_data0 + (rbm_bi(rbm)->bi_start * GFS2_NBBY) +
143		rbm->offset;
144}
145
146static inline bool gfs2_rbm_eq(const struct gfs2_rbm *rbm1,
147			       const struct gfs2_rbm *rbm2)
148{
149	return (rbm1->rgd == rbm2->rgd) && (rbm1->bii == rbm2->bii) &&
150	       (rbm1->offset == rbm2->offset);
151}
152
153enum gfs2_state_bits {
154	BH_Pinned = BH_PrivateStart,
155	BH_Escaped = BH_PrivateStart + 1,
156};
157
158BUFFER_FNS(Pinned, pinned)
159TAS_BUFFER_FNS(Pinned, pinned)
160BUFFER_FNS(Escaped, escaped)
161TAS_BUFFER_FNS(Escaped, escaped)
162
163struct gfs2_bufdata {
164	struct buffer_head *bd_bh;
165	struct gfs2_glock *bd_gl;
166	u64 bd_blkno;
167
168	struct list_head bd_list;
169
170	struct gfs2_trans *bd_tr;
171	struct list_head bd_ail_st_list;
172	struct list_head bd_ail_gl_list;
173};
174
175/*
176 * Internally, we prefix things with gdlm_ and GDLM_ (for gfs-dlm) since a
177 * prefix of lock_dlm_ gets awkward.
178 */
179
180#define GDLM_STRNAME_BYTES	25
181#define GDLM_LVB_SIZE		32
182
183/*
184 * ls_recover_flags:
185 *
186 * DFL_BLOCK_LOCKS: dlm is in recovery and will grant locks that had been
187 * held by failed nodes whose journals need recovery.  Those locks should
188 * only be used for journal recovery until the journal recovery is done.
189 * This is set by the dlm recover_prep callback and cleared by the
190 * gfs2_control thread when journal recovery is complete.  To avoid
191 * races between recover_prep setting and gfs2_control clearing, recover_spin
192 * is held while changing this bit and reading/writing recover_block
193 * and recover_start.
194 *
195 * DFL_NO_DLM_OPS: dlm lockspace ops/callbacks are not being used.
196 *
197 * DFL_FIRST_MOUNT: this node is the first to mount this fs and is doing
198 * recovery of all journals before allowing other nodes to mount the fs.
199 * This is cleared when FIRST_MOUNT_DONE is set.
200 *
201 * DFL_FIRST_MOUNT_DONE: this node was the first mounter, and has finished
202 * recovery of all journals, and now allows other nodes to mount the fs.
203 *
204 * DFL_MOUNT_DONE: gdlm_mount has completed successfully and cleared
205 * BLOCK_LOCKS for the first time.  The gfs2_control thread should now
206 * control clearing BLOCK_LOCKS for further recoveries.
207 *
208 * DFL_UNMOUNT: gdlm_unmount sets to keep sdp off gfs2_control_wq.
209 *
210 * DFL_DLM_RECOVERY: set while dlm is in recovery, between recover_prep()
211 * and recover_done(), i.e. set while recover_block == recover_start.
212 */
213
214enum {
215	DFL_BLOCK_LOCKS		= 0,
216	DFL_NO_DLM_OPS		= 1,
217	DFL_FIRST_MOUNT		= 2,
218	DFL_FIRST_MOUNT_DONE	= 3,
219	DFL_MOUNT_DONE		= 4,
220	DFL_UNMOUNT		= 5,
221	DFL_DLM_RECOVERY	= 6,
222};
223
224/*
225 * We are using struct lm_lockname as an rhashtable key.  Avoid holes within
226 * the struct; padding at the end is fine.
227 */
228struct lm_lockname {
229	u64 ln_number;
230	struct gfs2_sbd *ln_sbd;
231	unsigned int ln_type;
232};
233
234#define lm_name_equal(name1, name2) \
235        (((name1)->ln_number == (name2)->ln_number) &&	\
236	 ((name1)->ln_type == (name2)->ln_type) &&	\
237	 ((name1)->ln_sbd == (name2)->ln_sbd))
238
239
240struct gfs2_glock_operations {
241	int (*go_sync) (struct gfs2_glock *gl);
242	int (*go_xmote_bh) (struct gfs2_glock *gl, struct gfs2_holder *gh);
243	void (*go_inval) (struct gfs2_glock *gl, int flags);
244	int (*go_demote_ok) (const struct gfs2_glock *gl);
245	int (*go_lock) (struct gfs2_holder *gh);
246	void (*go_dump)(struct seq_file *seq, struct gfs2_glock *gl,
247			const char *fs_id_buf);
248	void (*go_callback)(struct gfs2_glock *gl, bool remote);
249	void (*go_free)(struct gfs2_glock *gl);
250	const int go_subclass;
251	const int go_type;
252	const unsigned long go_flags;
253#define GLOF_ASPACE 1 /* address space attached */
254#define GLOF_LVB    2 /* Lock Value Block attached */
255#define GLOF_LRU    4 /* LRU managed */
256#define GLOF_NONDISK   8 /* not I/O related */
257};
258
259enum {
260	GFS2_LKS_SRTT = 0,	/* Non blocking smoothed round trip time */
261	GFS2_LKS_SRTTVAR = 1,	/* Non blocking smoothed variance */
262	GFS2_LKS_SRTTB = 2,	/* Blocking smoothed round trip time */
263	GFS2_LKS_SRTTVARB = 3,	/* Blocking smoothed variance */
264	GFS2_LKS_SIRT = 4,	/* Smoothed Inter-request time */
265	GFS2_LKS_SIRTVAR = 5,	/* Smoothed Inter-request variance */
266	GFS2_LKS_DCOUNT = 6,	/* Count of dlm requests */
267	GFS2_LKS_QCOUNT = 7,	/* Count of gfs2_holder queues */
268	GFS2_NR_LKSTATS
269};
270
271struct gfs2_lkstats {
272	u64 stats[GFS2_NR_LKSTATS];
273};
274
275enum {
276	/* States */
277	HIF_HOLDER		= 6,  /* Set for gh that "holds" the glock */
278	HIF_FIRST		= 7,
279	HIF_WAIT		= 10,
280};
281
282struct gfs2_holder {
283	struct list_head gh_list;
284
285	struct gfs2_glock *gh_gl;
286	struct pid *gh_owner_pid;
287	u16 gh_flags;
288	u16 gh_state;
289
290	int gh_error;
291	unsigned long gh_iflags; /* HIF_... */
292	unsigned long gh_ip;
293};
294
295/* Number of quota types we support */
296#define GFS2_MAXQUOTAS 2
297
298struct gfs2_qadata { /* quota allocation data */
299	/* Quota stuff */
300	struct gfs2_quota_data *qa_qd[2 * GFS2_MAXQUOTAS];
301	struct gfs2_holder qa_qd_ghs[2 * GFS2_MAXQUOTAS];
302	unsigned int qa_qd_num;
303	int qa_ref;
304};
305
306/* Resource group multi-block reservation, in order of appearance:
307
308   Step 1. Function prepares to write, allocates a mb, sets the size hint.
309   Step 2. User calls inplace_reserve to target an rgrp, sets the rgrp info
310   Step 3. Function get_local_rgrp locks the rgrp, determines which bits to use
311   Step 4. Bits are assigned from the rgrp based on either the reservation
312           or wherever it can.
313*/
314
315struct gfs2_blkreserv {
316	struct rb_node rs_node;       /* link to other block reservations */
317	struct gfs2_rbm rs_rbm;       /* Start of reservation */
318	u32 rs_free;                  /* how many blocks are still free */
319};
320
321/*
322 * Allocation parameters
323 * @target: The number of blocks we'd ideally like to allocate
324 * @aflags: The flags (e.g. Orlov flag)
325 *
326 * The intent is to gradually expand this structure over time in
327 * order to give more information, e.g. alignment, min extent size
328 * to the allocation code.
329 */
330struct gfs2_alloc_parms {
331	u64 target;
332	u32 min_target;
333	u32 aflags;
334	u64 allowed;
335};
336
337enum {
338	GLF_LOCK			= 1,
339	GLF_DEMOTE			= 3,
340	GLF_PENDING_DEMOTE		= 4,
341	GLF_DEMOTE_IN_PROGRESS		= 5,
342	GLF_DIRTY			= 6,
343	GLF_LFLUSH			= 7,
344	GLF_INVALIDATE_IN_PROGRESS	= 8,
345	GLF_REPLY_PENDING		= 9,
346	GLF_INITIAL			= 10,
347	GLF_FROZEN			= 11,
348	GLF_LRU				= 13,
349	GLF_OBJECT			= 14, /* Used only for tracing */
350	GLF_BLOCKING			= 15,
351	GLF_INODE_CREATING		= 16, /* Inode creation occurring */
352	GLF_PENDING_DELETE		= 17,
353	GLF_FREEING			= 18, /* Wait for glock to be freed */
354};
355
356struct gfs2_glock {
357	unsigned long gl_flags;		/* GLF_... */
358	struct lm_lockname gl_name;
359
360	struct lockref gl_lockref;
361
362	/* State fields protected by gl_lockref.lock */
363	unsigned int gl_state:2,	/* Current state */
364		     gl_target:2,	/* Target state */
365		     gl_demote_state:2,	/* State requested by remote node */
366		     gl_req:2,		/* State in last dlm request */
367		     gl_reply:8;	/* Last reply from the dlm */
368
369	unsigned long gl_demote_time; /* time of first demote request */
370	long gl_hold_time;
371	struct list_head gl_holders;
372
373	const struct gfs2_glock_operations *gl_ops;
374	ktime_t gl_dstamp;
375	struct gfs2_lkstats gl_stats;
376	struct dlm_lksb gl_lksb;
377	unsigned long gl_tchange;
378	void *gl_object;
379
380	struct list_head gl_lru;
381	struct list_head gl_ail_list;
382	atomic_t gl_ail_count;
383	atomic_t gl_revokes;
384	struct delayed_work gl_work;
385	/* For iopen glocks only */
386	struct {
387		struct delayed_work gl_delete;
388		u64 gl_no_formal_ino;
389	};
390	struct rcu_head gl_rcu;
391	struct rhash_head gl_node;
392};
393
394enum {
395	GIF_INVALID		= 0,
396	GIF_QD_LOCKED		= 1,
397	GIF_ALLOC_FAILED	= 2,
398	GIF_SW_PAGED		= 3,
399	GIF_FREE_VFS_INODE      = 5,
400	GIF_GLOP_PENDING	= 6,
401	GIF_DEFERRED_DELETE	= 7,
402};
403
404struct gfs2_inode {
405	struct inode i_inode;
406	u64 i_no_addr;
407	u64 i_no_formal_ino;
408	u64 i_generation;
409	u64 i_eattr;
410	unsigned long i_flags;		/* GIF_... */
411	struct gfs2_glock *i_gl; /* Move into i_gh? */
412	struct gfs2_holder i_iopen_gh;
413	struct gfs2_holder i_gh; /* for prepare/commit_write only */
414	struct gfs2_qadata *i_qadata; /* quota allocation data */
415	struct gfs2_holder i_rgd_gh;
416	struct gfs2_blkreserv i_res; /* rgrp multi-block reservation */
417	u64 i_goal;	/* goal block for allocations */
418	atomic_t i_sizehint;  /* hint of the write size */
419	struct rw_semaphore i_rw_mutex;
420	struct list_head i_ordered;
421	struct list_head i_trunc_list;
422	__be64 *i_hash_cache;
423	u32 i_entries;
424	u32 i_diskflags;
425	u8 i_height;
426	u8 i_depth;
427	u16 i_rahead;
428};
429
430/*
431 * Since i_inode is the first element of struct gfs2_inode,
432 * this is effectively a cast.
433 */
434static inline struct gfs2_inode *GFS2_I(struct inode *inode)
435{
436	return container_of(inode, struct gfs2_inode, i_inode);
437}
438
439static inline struct gfs2_sbd *GFS2_SB(const struct inode *inode)
440{
441	return inode->i_sb->s_fs_info;
442}
443
444struct gfs2_file {
445	struct mutex f_fl_mutex;
446	struct gfs2_holder f_fl_gh;
447};
448
449struct gfs2_revoke_replay {
450	struct list_head rr_list;
451	u64 rr_blkno;
452	unsigned int rr_where;
453};
454
455enum {
456	QDF_CHANGE		= 1,
457	QDF_LOCKED		= 2,
458	QDF_REFRESH		= 3,
459	QDF_QMSG_QUIET          = 4,
460};
461
462struct gfs2_quota_data {
463	struct hlist_bl_node qd_hlist;
464	struct list_head qd_list;
465	struct kqid qd_id;
466	struct gfs2_sbd *qd_sbd;
467	struct lockref qd_lockref;
468	struct list_head qd_lru;
469	unsigned qd_hash;
470
471	unsigned long qd_flags;		/* QDF_... */
472
473	s64 qd_change;
474	s64 qd_change_sync;
475
476	unsigned int qd_slot;
477	unsigned int qd_slot_count;
478
479	struct buffer_head *qd_bh;
480	struct gfs2_quota_change *qd_bh_qc;
481	unsigned int qd_bh_count;
482
483	struct gfs2_glock *qd_gl;
484	struct gfs2_quota_lvb qd_qb;
485
486	u64 qd_sync_gen;
487	unsigned long qd_last_warn;
488	struct rcu_head qd_rcu;
489};
490
491enum {
492	TR_TOUCHED = 1,
493	TR_ATTACHED = 2,
494	TR_ALLOCED = 3,
495};
496
497struct gfs2_trans {
498	unsigned long tr_ip;
499
500	unsigned int tr_blocks;
501	unsigned int tr_revokes;
502	unsigned int tr_reserved;
503	unsigned long tr_flags;
504
505	unsigned int tr_num_buf_new;
506	unsigned int tr_num_databuf_new;
507	unsigned int tr_num_buf_rm;
508	unsigned int tr_num_databuf_rm;
509	unsigned int tr_num_revoke;
510	unsigned int tr_num_revoke_rm;
511
512	struct list_head tr_list;
513	struct list_head tr_databuf;
514	struct list_head tr_buf;
515
516	unsigned int tr_first;
517	struct list_head tr_ail1_list;
518	struct list_head tr_ail2_list;
519};
520
521struct gfs2_journal_extent {
522	struct list_head list;
523
524	unsigned int lblock; /* First logical block */
525	u64 dblock; /* First disk block */
526	u64 blocks;
527};
528
529struct gfs2_jdesc {
530	struct list_head jd_list;
531	struct list_head extent_list;
532	unsigned int nr_extents;
533	struct work_struct jd_work;
534	struct inode *jd_inode;
535	unsigned long jd_flags;
536#define JDF_RECOVERY 1
537	unsigned int jd_jid;
538	u32 jd_blocks;
539	int jd_recover_error;
540	/* Replay stuff */
541
542	unsigned int jd_found_blocks;
543	unsigned int jd_found_revokes;
544	unsigned int jd_replayed_blocks;
545
546	struct list_head jd_revoke_list;
547	unsigned int jd_replay_tail;
548
549	u64 jd_no_addr;
550};
551
552struct gfs2_statfs_change_host {
553	s64 sc_total;
554	s64 sc_free;
555	s64 sc_dinodes;
556};
557
558#define GFS2_QUOTA_DEFAULT	GFS2_QUOTA_OFF
559#define GFS2_QUOTA_OFF		0
560#define GFS2_QUOTA_ACCOUNT	1
561#define GFS2_QUOTA_ON		2
562
563#define GFS2_DATA_DEFAULT	GFS2_DATA_ORDERED
564#define GFS2_DATA_WRITEBACK	1
565#define GFS2_DATA_ORDERED	2
566
567#define GFS2_ERRORS_DEFAULT     GFS2_ERRORS_WITHDRAW
568#define GFS2_ERRORS_WITHDRAW    0
569#define GFS2_ERRORS_CONTINUE    1 /* place holder for future feature */
570#define GFS2_ERRORS_RO          2 /* place holder for future feature */
571#define GFS2_ERRORS_PANIC       3
572
573struct gfs2_args {
574	char ar_lockproto[GFS2_LOCKNAME_LEN];	/* Name of the Lock Protocol */
575	char ar_locktable[GFS2_LOCKNAME_LEN];	/* Name of the Lock Table */
576	char ar_hostdata[GFS2_LOCKNAME_LEN];	/* Host specific data */
577	unsigned int ar_spectator:1;		/* Don't get a journal */
578	unsigned int ar_localflocks:1;		/* Let the VFS do flock|fcntl */
579	unsigned int ar_debug:1;		/* Oops on errors */
580	unsigned int ar_posix_acl:1;		/* Enable posix acls */
581	unsigned int ar_quota:2;		/* off/account/on */
582	unsigned int ar_suiddir:1;		/* suiddir support */
583	unsigned int ar_data:2;			/* ordered/writeback */
584	unsigned int ar_meta:1;			/* mount metafs */
585	unsigned int ar_discard:1;		/* discard requests */
586	unsigned int ar_errors:2;               /* errors=withdraw | panic */
587	unsigned int ar_nobarrier:1;            /* do not send barriers */
588	unsigned int ar_rgrplvb:1;		/* use lvbs for rgrp info */
589	unsigned int ar_loccookie:1;		/* use location based readdir
590						   cookies */
591	s32 ar_commit;				/* Commit interval */
592	s32 ar_statfs_quantum;			/* The fast statfs interval */
593	s32 ar_quota_quantum;			/* The quota interval */
594	s32 ar_statfs_percent;			/* The % change to force sync */
595};
596
597struct gfs2_tune {
598	spinlock_t gt_spin;
599
600	unsigned int gt_logd_secs;
601
602	unsigned int gt_quota_warn_period; /* Secs between quota warn msgs */
603	unsigned int gt_quota_scale_num; /* Numerator */
604	unsigned int gt_quota_scale_den; /* Denominator */
605	unsigned int gt_quota_quantum; /* Secs between syncs to quota file */
606	unsigned int gt_new_files_jdata;
607	unsigned int gt_max_readahead; /* Max bytes to read-ahead from disk */
608	unsigned int gt_complain_secs;
609	unsigned int gt_statfs_quantum;
610	unsigned int gt_statfs_slow;
611};
612
613enum {
614	SDF_JOURNAL_CHECKED	= 0,
615	SDF_JOURNAL_LIVE	= 1,
616	SDF_WITHDRAWN		= 2,
617	SDF_NOBARRIERS		= 3,
618	SDF_NORECOVERY		= 4,
619	SDF_DEMOTE		= 5,
620	SDF_NOJOURNALID		= 6,
621	SDF_RORECOVERY		= 7, /* read only recovery */
622	SDF_SKIP_DLM_UNLOCK	= 8,
623	SDF_FORCE_AIL_FLUSH     = 9,
624	SDF_FS_FROZEN           = 10,
625	SDF_WITHDRAWING		= 11, /* Will withdraw eventually */
626	SDF_WITHDRAW_IN_PROG	= 12, /* Withdraw is in progress */
627	SDF_REMOTE_WITHDRAW	= 13, /* Performing remote recovery */
628	SDF_WITHDRAW_RECOVERY	= 14, /* Wait for journal recovery when we are
629					 withdrawing */
630};
631
632enum gfs2_freeze_state {
633	SFS_UNFROZEN		= 0,
634	SFS_STARTING_FREEZE	= 1,
635	SFS_FROZEN		= 2,
636};
637
638#define GFS2_FSNAME_LEN		256
639
640struct gfs2_inum_host {
641	u64 no_formal_ino;
642	u64 no_addr;
643};
644
645struct gfs2_sb_host {
646	u32 sb_magic;
647	u32 sb_type;
648	u32 sb_format;
649
650	u32 sb_fs_format;
651	u32 sb_multihost_format;
652	u32 sb_bsize;
653	u32 sb_bsize_shift;
654
655	struct gfs2_inum_host sb_master_dir;
656	struct gfs2_inum_host sb_root_dir;
657
658	char sb_lockproto[GFS2_LOCKNAME_LEN];
659	char sb_locktable[GFS2_LOCKNAME_LEN];
660};
661
662/*
663 * lm_mount() return values
664 *
665 * ls_jid - the journal ID this node should use
666 * ls_first - this node is the first to mount the file system
667 * ls_lockspace - lock module's context for this file system
668 * ls_ops - lock module's functions
669 */
670
671struct lm_lockstruct {
672	int ls_jid;
673	unsigned int ls_first;
674	const struct lm_lockops *ls_ops;
675	dlm_lockspace_t *ls_dlm;
676
677	int ls_recover_jid_done;   /* These two are deprecated, */
678	int ls_recover_jid_status; /* used previously by gfs_controld */
679
680	struct dlm_lksb ls_mounted_lksb; /* mounted_lock */
681	struct dlm_lksb ls_control_lksb; /* control_lock */
682	char ls_control_lvb[GDLM_LVB_SIZE]; /* control_lock lvb */
683	struct completion ls_sync_wait; /* {control,mounted}_{lock,unlock} */
684	char *ls_lvb_bits;
685
686	spinlock_t ls_recover_spin; /* protects following fields */
687	unsigned long ls_recover_flags; /* DFL_ */
688	uint32_t ls_recover_mount; /* gen in first recover_done cb */
689	uint32_t ls_recover_start; /* gen in last recover_done cb */
690	uint32_t ls_recover_block; /* copy recover_start in last recover_prep */
691	uint32_t ls_recover_size; /* size of recover_submit, recover_result */
692	uint32_t *ls_recover_submit; /* gen in last recover_slot cb per jid */
693	uint32_t *ls_recover_result; /* result of last jid recovery */
694};
695
696struct gfs2_pcpu_lkstats {
697	/* One struct for each glock type */
698	struct gfs2_lkstats lkstats[10];
699};
700
701/* List of local (per node) statfs inodes */
702struct local_statfs_inode {
703	struct list_head si_list;
704	struct inode *si_sc_inode;
705	unsigned int si_jid; /* journal id this statfs inode corresponds to */
706};
707
708struct gfs2_sbd {
709	struct super_block *sd_vfs;
710	struct gfs2_pcpu_lkstats __percpu *sd_lkstats;
711	struct kobject sd_kobj;
712	struct completion sd_kobj_unregister;
713	unsigned long sd_flags;	/* SDF_... */
714	struct gfs2_sb_host sd_sb;
715
716	/* Constants computed on mount */
717
718	u32 sd_fsb2bb;
719	u32 sd_fsb2bb_shift;
720	u32 sd_diptrs;	/* Number of pointers in a dinode */
721	u32 sd_inptrs;	/* Number of pointers in a indirect block */
722	u32 sd_ldptrs;  /* Number of pointers in a log descriptor block */
723	u32 sd_jbsize;	/* Size of a journaled data block */
724	u32 sd_hash_bsize;	/* sizeof(exhash block) */
725	u32 sd_hash_bsize_shift;
726	u32 sd_hash_ptrs;	/* Number of pointers in a hash block */
727	u32 sd_qc_per_block;
728	u32 sd_blocks_per_bitmap;
729	u32 sd_max_dirres;	/* Max blocks needed to add a directory entry */
730	u32 sd_max_height;	/* Max height of a file's metadata tree */
731	u64 sd_heightsize[GFS2_MAX_META_HEIGHT + 1];
732	u32 sd_max_dents_per_leaf; /* Max number of dirents in a leaf block */
733
734	struct gfs2_args sd_args;	/* Mount arguments */
735	struct gfs2_tune sd_tune;	/* Filesystem tuning structure */
736
737	/* Lock Stuff */
738
739	struct lm_lockstruct sd_lockstruct;
740	struct gfs2_holder sd_live_gh;
741	struct gfs2_glock *sd_rename_gl;
742	struct gfs2_glock *sd_freeze_gl;
743	struct work_struct sd_freeze_work;
744	wait_queue_head_t sd_glock_wait;
745	wait_queue_head_t sd_async_glock_wait;
746	atomic_t sd_glock_disposal;
747	struct completion sd_locking_init;
748	struct completion sd_wdack;
749	struct delayed_work sd_control_work;
750
751	/* Inode Stuff */
752
753	struct dentry *sd_master_dir;
754	struct dentry *sd_root_dir;
755
756	struct inode *sd_jindex;
757	struct inode *sd_statfs_inode;
758	struct inode *sd_sc_inode;
759	struct list_head sd_sc_inodes_list;
760	struct inode *sd_qc_inode;
761	struct inode *sd_rindex;
762	struct inode *sd_quota_inode;
763
764	/* StatFS stuff */
765
766	spinlock_t sd_statfs_spin;
767	struct gfs2_statfs_change_host sd_statfs_master;
768	struct gfs2_statfs_change_host sd_statfs_local;
769	int sd_statfs_force_sync;
770
771	/* Resource group stuff */
772
773	int sd_rindex_uptodate;
774	spinlock_t sd_rindex_spin;
775	struct rb_root sd_rindex_tree;
776	unsigned int sd_rgrps;
777	unsigned int sd_max_rg_data;
778
779	/* Journal index stuff */
780
781	struct list_head sd_jindex_list;
782	spinlock_t sd_jindex_spin;
783	struct mutex sd_jindex_mutex;
784	unsigned int sd_journals;
785
786	struct gfs2_jdesc *sd_jdesc;
787	struct gfs2_holder sd_journal_gh;
788	struct gfs2_holder sd_jinode_gh;
789	struct gfs2_glock *sd_jinode_gl;
790
791	struct gfs2_holder sd_sc_gh;
792	struct gfs2_holder sd_qc_gh;
793
794	struct completion sd_journal_ready;
795
796	/* Daemon stuff */
797
798	struct task_struct *sd_logd_process;
799	struct task_struct *sd_quotad_process;
800
801	/* Quota stuff */
802
803	struct list_head sd_quota_list;
804	atomic_t sd_quota_count;
805	struct mutex sd_quota_mutex;
806	struct mutex sd_quota_sync_mutex;
807	wait_queue_head_t sd_quota_wait;
808	struct list_head sd_trunc_list;
809	spinlock_t sd_trunc_lock;
810
811	unsigned int sd_quota_slots;
812	unsigned long *sd_quota_bitmap;
813	spinlock_t sd_bitmap_lock;
814
815	u64 sd_quota_sync_gen;
816
817	/* Log stuff */
818
819	struct address_space sd_aspace;
820
821	spinlock_t sd_log_lock;
822
823	struct gfs2_trans *sd_log_tr;
824	unsigned int sd_log_blks_reserved;
825	int sd_log_committed_revoke;
826
827	atomic_t sd_log_pinned;
828	unsigned int sd_log_num_revoke;
829
830	struct list_head sd_log_revokes;
831	struct list_head sd_log_ordered;
832	spinlock_t sd_ordered_lock;
833
834	atomic_t sd_log_thresh1;
835	atomic_t sd_log_thresh2;
836	atomic_t sd_log_blks_free;
837	atomic_t sd_log_blks_needed;
838	wait_queue_head_t sd_log_waitq;
839	wait_queue_head_t sd_logd_waitq;
840
841	u64 sd_log_sequence;
842	unsigned int sd_log_head;
843	unsigned int sd_log_tail;
844	int sd_log_idle;
845
846	struct rw_semaphore sd_log_flush_lock;
847	atomic_t sd_log_in_flight;
848	struct bio *sd_log_bio;
849	wait_queue_head_t sd_log_flush_wait;
850	int sd_log_error; /* First log error */
851	wait_queue_head_t sd_withdraw_wait;
852
853	atomic_t sd_reserving_log;
854	wait_queue_head_t sd_reserving_log_wait;
855
856	unsigned int sd_log_flush_head;
857
858	spinlock_t sd_ail_lock;
859	struct list_head sd_ail1_list;
860	struct list_head sd_ail2_list;
861
862	/* For quiescing the filesystem */
863	struct gfs2_holder sd_freeze_gh;
864	atomic_t sd_freeze_state;
865	struct mutex sd_freeze_mutex;
866
867	char sd_fsname[GFS2_FSNAME_LEN + 3 * sizeof(int) + 2];
868	char sd_table_name[GFS2_FSNAME_LEN];
869	char sd_proto_name[GFS2_FSNAME_LEN];
870
871	/* Debugging crud */
872
873	unsigned long sd_last_warning;
874	struct dentry *debugfs_dir;    /* debugfs directory */
875	unsigned long sd_glock_dqs_held;
876};
877
878static inline void gfs2_glstats_inc(struct gfs2_glock *gl, int which)
879{
880	gl->gl_stats.stats[which]++;
881}
882
883static inline void gfs2_sbstats_inc(const struct gfs2_glock *gl, int which)
884{
885	const struct gfs2_sbd *sdp = gl->gl_name.ln_sbd;
886	preempt_disable();
887	this_cpu_ptr(sdp->sd_lkstats)->lkstats[gl->gl_name.ln_type].stats[which]++;
888	preempt_enable();
889}
890
891extern struct gfs2_rgrpd *gfs2_glock2rgrp(struct gfs2_glock *gl);
892
893static inline unsigned gfs2_max_stuffed_size(const struct gfs2_inode *ip)
894{
895	return GFS2_SB(&ip->i_inode)->sd_sb.sb_bsize - sizeof(struct gfs2_dinode);
896}
897
898#endif /* __INCORE_DOT_H__ */
899
900