[deliverable/linux.git] / fs / ocfs2 / alloc.h

/* -*- mode: c; c-basic-offset: 8; -*-
 * vim: noexpandtab sw=8 ts=8 sts=0:
 *
 * alloc.h
 *
 * Function prototypes
 *
 * Copyright (C) 2002, 2004 Oracle.  All rights reserved.
 *
 * This program is free software; you can redistribute it and/or
 * modify it under the terms of the GNU General Public
 * License as published by the Free Software Foundation; either
 * version 2 of the License, or (at your option) any later version.
 *
 * This program is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
 * General Public License for more details.
 *
 * You should have received a copy of the GNU General Public
 * License along with this program; if not, write to the
 * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
 * Boston, MA 021110-1307, USA.
 */

#ifndef OCFS2_ALLOC_H
#define OCFS2_ALLOC_H


/*
 * For xattr tree leaf, we limit the leaf byte size to be 64K.
 */
#define OCFS2_MAX_XATTR_TREE_LEAF_SIZE 65536

/*
 * ocfs2_extent_tree and ocfs2_extent_tree_operations are used to abstract
 * the b-tree operations in ocfs2. Now all the b-tree operations are not
 * limited to ocfs2_dinode only. Any data which need to allocate clusters
 * to store can use b-tree. And it only needs to implement its ocfs2_extent_tree
 * and operation.
 *
 * ocfs2_extent_tree becomes the first-class object for extent tree
 * manipulation.  Callers of the alloc.c code need to fill it via one of
 * the ocfs2_init_*_extent_tree() operations below.
 *
 * ocfs2_extent_tree contains info for the root of the b-tree, it must have a
 * root ocfs2_extent_list and a root_bh so that they can be used in the b-tree
 * functions.
 * ocfs2_extent_tree_operations abstract the normal operations we do for
 * the root of extent b-tree.
 */
struct ocfs2_extent_tree_operations;
struct ocfs2_extent_tree {
	struct ocfs2_extent_tree_operations	*et_ops;
	struct buffer_head			*et_root_bh;
	struct ocfs2_extent_list		*et_root_el;
	void					*et_object;
	unsigned int				et_max_leaf_clusters;
};

/*
 * ocfs2_init_*_extent_tree() will fill an ocfs2_extent_tree from the
 * specified object buffer.
 */
void ocfs2_init_dinode_extent_tree(struct ocfs2_extent_tree *et,
				   struct inode *inode,
				   struct buffer_head *bh);
void ocfs2_init_xattr_tree_extent_tree(struct ocfs2_extent_tree *et,
				       struct inode *inode,
				       struct buffer_head *bh);
void ocfs2_init_xattr_value_extent_tree(struct ocfs2_extent_tree *et,
					struct inode *inode,
					struct buffer_head *bh,
					struct ocfs2_xattr_value_root *xv);

struct ocfs2_alloc_context;
int ocfs2_insert_extent(struct ocfs2_super *osb,
			handle_t *handle,
			struct inode *inode,
			struct ocfs2_extent_tree *et,
			u32 cpos,
			u64 start_blk,
			u32 new_clusters,
			u8 flags,
			struct ocfs2_alloc_context *meta_ac);

enum ocfs2_alloc_restarted {
	RESTART_NONE = 0,
	RESTART_TRANS,
	RESTART_META
};
int ocfs2_add_clusters_in_btree(struct ocfs2_super *osb,
				struct inode *inode,
				u32 *logical_offset,
				u32 clusters_to_add,
				int mark_unwritten,
				struct ocfs2_extent_tree *et,
				handle_t *handle,
				struct ocfs2_alloc_context *data_ac,
				struct ocfs2_alloc_context *meta_ac,
				enum ocfs2_alloc_restarted *reason_ret);
struct ocfs2_cached_dealloc_ctxt;
int ocfs2_mark_extent_written(struct inode *inode,
			      struct ocfs2_extent_tree *et,
			      handle_t *handle, u32 cpos, u32 len, u32 phys,
			      struct ocfs2_alloc_context *meta_ac,
			      struct ocfs2_cached_dealloc_ctxt *dealloc);
int ocfs2_remove_extent(struct inode *inode,
			struct ocfs2_extent_tree *et,
			u32 cpos, u32 len, handle_t *handle,
			struct ocfs2_alloc_context *meta_ac,
			struct ocfs2_cached_dealloc_ctxt *dealloc);
int ocfs2_remove_btree_range(struct inode *inode,
			     struct ocfs2_extent_tree *et,
			     u32 cpos, u32 phys_cpos, u32 len,
			     struct ocfs2_cached_dealloc_ctxt *dealloc);

int ocfs2_num_free_extents(struct ocfs2_super *osb,
			   struct inode *inode,
			   struct ocfs2_extent_tree *et);

/*
 * how many new metadata chunks would an allocation need at maximum?
 *
 * Please note that the caller must make sure that root_el is the root
 * of extent tree. So for an inode, it should be &fe->id2.i_list. Otherwise
 * the result may be wrong.
 */
static inline int ocfs2_extend_meta_needed(struct ocfs2_extent_list *root_el)
{
	/*
	 * Rather than do all the work of determining how much we need
	 * (involves a ton of reads and locks), just ask for the
	 * maximal limit.  That's a tree depth shift.  So, one block for
	 * level of the tree (current l_tree_depth), one block for the
	 * new tree_depth==0 extent_block, and one block at the new
	 * top-of-the tree.
	 */
	return le16_to_cpu(root_el->l_tree_depth) + 2;
}

void ocfs2_dinode_new_extent_list(struct inode *inode, struct ocfs2_dinode *di);
void ocfs2_set_inode_data_inline(struct inode *inode, struct ocfs2_dinode *di);
int ocfs2_convert_inline_data_to_extents(struct inode *inode,
					 struct buffer_head *di_bh);

int ocfs2_truncate_log_init(struct ocfs2_super *osb);
void ocfs2_truncate_log_shutdown(struct ocfs2_super *osb);
void ocfs2_schedule_truncate_log_flush(struct ocfs2_super *osb,
				       int cancel);
int ocfs2_flush_truncate_log(struct ocfs2_super *osb);
int ocfs2_begin_truncate_log_recovery(struct ocfs2_super *osb,
				      int slot_num,
				      struct ocfs2_dinode **tl_copy);
int ocfs2_complete_truncate_log_recovery(struct ocfs2_super *osb,
					 struct ocfs2_dinode *tl_copy);
int ocfs2_truncate_log_needs_flush(struct ocfs2_super *osb);
int ocfs2_truncate_log_append(struct ocfs2_super *osb,
			      handle_t *handle,
			      u64 start_blk,
			      unsigned int num_clusters);
int __ocfs2_flush_truncate_log(struct ocfs2_super *osb);

/*
 * Process local structure which describes the block unlinks done
 * during an operation. This is populated via
 * ocfs2_cache_block_dealloc().
 *
 * ocfs2_run_deallocs() should be called after the potentially
 * de-allocating routines. No journal handles should be open, and most
 * locks should have been dropped.
 */
struct ocfs2_cached_dealloc_ctxt {
	struct ocfs2_per_slot_free_list		*c_first_suballocator;
	struct ocfs2_cached_block_free 		*c_global_allocator;
};
static inline void ocfs2_init_dealloc_ctxt(struct ocfs2_cached_dealloc_ctxt *c)
{
	c->c_first_suballocator = NULL;
	c->c_global_allocator = NULL;
}
int ocfs2_cache_cluster_dealloc(struct ocfs2_cached_dealloc_ctxt *ctxt,
				u64 blkno, unsigned int bit);
static inline int ocfs2_dealloc_has_cluster(struct ocfs2_cached_dealloc_ctxt *c)
{
	return c->c_global_allocator != NULL;
}
int ocfs2_run_deallocs(struct ocfs2_super *osb,
		       struct ocfs2_cached_dealloc_ctxt *ctxt);

struct ocfs2_truncate_context {
	struct ocfs2_cached_dealloc_ctxt tc_dealloc;
	int tc_ext_alloc_locked; /* is it cluster locked? */
	/* these get destroyed once it's passed to ocfs2_commit_truncate. */
	struct buffer_head *tc_last_eb_bh;
};

int ocfs2_zero_range_for_truncate(struct inode *inode, handle_t *handle,
				  u64 range_start, u64 range_end);
int ocfs2_prepare_truncate(struct ocfs2_super *osb,
			   struct inode *inode,
			   struct buffer_head *fe_bh,
			   struct ocfs2_truncate_context **tc);
int ocfs2_commit_truncate(struct ocfs2_super *osb,
			  struct inode *inode,
			  struct buffer_head *fe_bh,
			  struct ocfs2_truncate_context *tc);
int ocfs2_truncate_inline(struct inode *inode, struct buffer_head *di_bh,
			  unsigned int start, unsigned int end, int trunc);

int ocfs2_find_leaf(struct inode *inode, struct ocfs2_extent_list *root_el,
		    u32 cpos, struct buffer_head **leaf_bh);
int ocfs2_search_extent_list(struct ocfs2_extent_list *el, u32 v_cluster);

/*
 * Helper function to look at the # of clusters in an extent record.
 */
static inline unsigned int ocfs2_rec_clusters(struct ocfs2_extent_list *el,
					      struct ocfs2_extent_rec *rec)
{
	/*
	 * Cluster count in extent records is slightly different
	 * between interior nodes and leaf nodes. This is to support
	 * unwritten extents which need a flags field in leaf node
	 * records, thus shrinking the available space for a clusters
	 * field.
	 */
	if (el->l_tree_depth)
		return le32_to_cpu(rec->e_int_clusters);
	else
		return le16_to_cpu(rec->e_leaf_clusters);
}

/*
 * This is only valid for leaf nodes, which are the only ones that can
 * have empty extents anyway.
 */
static inline int ocfs2_is_empty_extent(struct ocfs2_extent_rec *rec)
{
	return !rec->e_leaf_clusters;
}

#endif /* OCFS2_ALLOC_H */
Commit	Line	Data
ccd979bd MF	1	/* -- mode: c; c-basic-offset: 8; --
	2	* vim: noexpandtab sw=8 ts=8 sts=0:
	3	*
	4	* alloc.h
	5	*
	6	* Function prototypes
	7	*
	8	* Copyright (C) 2002, 2004 Oracle. All rights reserved.
	9	*
	10	* This program is free software; you can redistribute it and/or
	11	* modify it under the terms of the GNU General Public
	12	* License as published by the Free Software Foundation; either
	13	* version 2 of the License, or (at your option) any later version.
	14	*
	15	* This program is distributed in the hope that it will be useful,
	16	* but WITHOUT ANY WARRANTY; without even the implied warranty of
	17	* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
	18	* General Public License for more details.
	19	*
	20	* You should have received a copy of the GNU General Public
	21	* License along with this program; if not, write to the
	22	* Free Software Foundation, Inc., 59 Temple Place - Suite 330,
	23	* Boston, MA 021110-1307, USA.
	24	*/
	25
	26	#ifndef OCFS2_ALLOC_H
	27	#define OCFS2_ALLOC_H
	28
e7d4cb6b	29
ca12b7c4 TM	30	/*
	31	* For xattr tree leaf, we limit the leaf byte size to be 64K.
	32	*/
	33	#define OCFS2_MAX_XATTR_TREE_LEAF_SIZE 65536
	34
f99b9b7c JB	35	/*
	36	* ocfs2_extent_tree and ocfs2_extent_tree_operations are used to abstract
	37	* the b-tree operations in ocfs2. Now all the b-tree operations are not
	38	* limited to ocfs2_dinode only. Any data which need to allocate clusters
	39	* to store can use b-tree. And it only needs to implement its ocfs2_extent_tree
	40	* and operation.
	41	*
	42	* ocfs2_extent_tree becomes the first-class object for extent tree
	43	* manipulation. Callers of the alloc.c code need to fill it via one of
8d6220d6	44	* the ocfs2_init_*_extent_tree() operations below.
f99b9b7c JB	45	*
	46	* ocfs2_extent_tree contains info for the root of the b-tree, it must have a
	47	* root ocfs2_extent_list and a root_bh so that they can be used in the b-tree
	48	* functions.
	49	* ocfs2_extent_tree_operations abstract the normal operations we do for
	50	* the root of extent b-tree.
	51	*/
	52	struct ocfs2_extent_tree_operations;
	53	struct ocfs2_extent_tree {
	54	struct ocfs2_extent_tree_operations *et_ops;
	55	struct buffer_head *et_root_bh;
	56	struct ocfs2_extent_list *et_root_el;
	57	void *et_object;
	58	unsigned int et_max_leaf_clusters;
	59	};
	60
	61	/*
8d6220d6 JB	62	* ocfs2_init_*_extent_tree() will fill an ocfs2_extent_tree from the
8d6220d6 JB	63	* specified object buffer.
f99b9b7c	64	*/
8d6220d6 JB	65	void ocfs2_init_dinode_extent_tree(struct ocfs2_extent_tree *et,
	66	struct inode *inode,
	67	struct buffer_head *bh);
	68	void ocfs2_init_xattr_tree_extent_tree(struct ocfs2_extent_tree *et,
f99b9b7c	69	struct inode *inode,
8d6220d6 JB	70	struct buffer_head *bh);
	71	void ocfs2_init_xattr_value_extent_tree(struct ocfs2_extent_tree *et,
	72	struct inode *inode,
	73	struct buffer_head *bh,
	74	struct ocfs2_xattr_value_root *xv);
f99b9b7c	75
ccd979bd	76	struct ocfs2_alloc_context;
f99b9b7c JB	77	int ocfs2_insert_extent(struct ocfs2_super *osb,
	78	handle_t *handle,
	79	struct inode *inode,
	80	struct ocfs2_extent_tree *et,
	81	u32 cpos,
	82	u64 start_blk,
	83	u32 new_clusters,
	84	u8 flags,
	85	struct ocfs2_alloc_context *meta_ac);
	86
0eb8d47e TM	87	enum ocfs2_alloc_restarted {
	88	RESTART_NONE = 0,
	89	RESTART_TRANS,
	90	RESTART_META
	91	};
	92	int ocfs2_add_clusters_in_btree(struct ocfs2_super *osb,
	93	struct inode *inode,
	94	u32 *logical_offset,
	95	u32 clusters_to_add,
	96	int mark_unwritten,
f99b9b7c	97	struct ocfs2_extent_tree *et,
0eb8d47e TM	98	handle_t *handle,
	99	struct ocfs2_alloc_context *data_ac,
	100	struct ocfs2_alloc_context *meta_ac,
f99b9b7c	101	enum ocfs2_alloc_restarted *reason_ret);
328d5752	102	struct ocfs2_cached_dealloc_ctxt;
f99b9b7c JB	103	int ocfs2_mark_extent_written(struct inode *inode,
f99b9b7c JB	104	struct ocfs2_extent_tree *et,
328d5752 MF	105	handle_t *handle, u32 cpos, u32 len, u32 phys,
328d5752 MF	106	struct ocfs2_alloc_context *meta_ac,
f99b9b7c JB	107	struct ocfs2_cached_dealloc_ctxt *dealloc);
	108	int ocfs2_remove_extent(struct inode *inode,
	109	struct ocfs2_extent_tree *et,
063c4561 MF	110	u32 cpos, u32 len, handle_t *handle,
063c4561 MF	111	struct ocfs2_alloc_context *meta_ac,
f99b9b7c	112	struct ocfs2_cached_dealloc_ctxt *dealloc);
fecc0112 MF	113	int ocfs2_remove_btree_range(struct inode *inode,
	114	struct ocfs2_extent_tree *et,
	115	u32 cpos, u32 phys_cpos, u32 len,
	116	struct ocfs2_cached_dealloc_ctxt *dealloc);
	117
ccd979bd MF	118	int ocfs2_num_free_extents(struct ocfs2_super *osb,
ccd979bd MF	119	struct inode *inode,
f99b9b7c	120	struct ocfs2_extent_tree *et);
e7d4cb6b	121
811f933d TM	122	/*
	123	* how many new metadata chunks would an allocation need at maximum?
	124	*
	125	* Please note that the caller must make sure that root_el is the root
	126	* of extent tree. So for an inode, it should be &fe->id2.i_list. Otherwise
	127	* the result may be wrong.
	128	*/
	129	static inline int ocfs2_extend_meta_needed(struct ocfs2_extent_list *root_el)
ccd979bd MF	130	{
	131	/*
	132	* Rather than do all the work of determining how much we need
	133	* (involves a ton of reads and locks), just ask for the
	134	* maximal limit. That's a tree depth shift. So, one block for
	135	* level of the tree (current l_tree_depth), one block for the
	136	* new tree_depth==0 extent_block, and one block at the new
	137	* top-of-the tree.
	138	*/
811f933d	139	return le16_to_cpu(root_el->l_tree_depth) + 2;
ccd979bd MF	140	}
ccd979bd MF	141
5b6a3a2b	142	void ocfs2_dinode_new_extent_list(struct inode inode, struct ocfs2_dinode di);
1afc32b9 MF	143	void ocfs2_set_inode_data_inline(struct inode inode, struct ocfs2_dinode di);
	144	int ocfs2_convert_inline_data_to_extents(struct inode *inode,
	145	struct buffer_head *di_bh);
	146
ccd979bd MF	147	int ocfs2_truncate_log_init(struct ocfs2_super *osb);
	148	void ocfs2_truncate_log_shutdown(struct ocfs2_super *osb);
	149	void ocfs2_schedule_truncate_log_flush(struct ocfs2_super *osb,
	150	int cancel);
	151	int ocfs2_flush_truncate_log(struct ocfs2_super *osb);
	152	int ocfs2_begin_truncate_log_recovery(struct ocfs2_super *osb,
	153	int slot_num,
	154	struct ocfs2_dinode **tl_copy);
	155	int ocfs2_complete_truncate_log_recovery(struct ocfs2_super *osb,
	156	struct ocfs2_dinode *tl_copy);
063c4561 MF	157	int ocfs2_truncate_log_needs_flush(struct ocfs2_super *osb);
	158	int ocfs2_truncate_log_append(struct ocfs2_super *osb,
	159	handle_t *handle,
	160	u64 start_blk,
	161	unsigned int num_clusters);
	162	int __ocfs2_flush_truncate_log(struct ocfs2_super *osb);
ccd979bd	163
2b604351 MF	164	/*
	165	* Process local structure which describes the block unlinks done
	166	* during an operation. This is populated via
	167	* ocfs2_cache_block_dealloc().
	168	*
	169	* ocfs2_run_deallocs() should be called after the potentially
	170	* de-allocating routines. No journal handles should be open, and most
	171	* locks should have been dropped.
	172	*/
	173	struct ocfs2_cached_dealloc_ctxt {
	174	struct ocfs2_per_slot_free_list *c_first_suballocator;
2891d290	175	struct ocfs2_cached_block_free *c_global_allocator;
2b604351 MF	176	};
	177	static inline void ocfs2_init_dealloc_ctxt(struct ocfs2_cached_dealloc_ctxt *c)
	178	{
	179	c->c_first_suballocator = NULL;
2891d290	180	c->c_global_allocator = NULL;
2b604351	181	}
2891d290 TM	182	int ocfs2_cache_cluster_dealloc(struct ocfs2_cached_dealloc_ctxt *ctxt,
2891d290 TM	183	u64 blkno, unsigned int bit);
78f30c31 TM	184	static inline int ocfs2_dealloc_has_cluster(struct ocfs2_cached_dealloc_ctxt *c)
	185	{
	186	return c->c_global_allocator != NULL;
	187	}
2b604351 MF	188	int ocfs2_run_deallocs(struct ocfs2_super *osb,
	189	struct ocfs2_cached_dealloc_ctxt *ctxt);
	190
ccd979bd	191	struct ocfs2_truncate_context {
59a5e416	192	struct ocfs2_cached_dealloc_ctxt tc_dealloc;
ccd979bd MF	193	int tc_ext_alloc_locked; /* is it cluster locked? */
	194	/* these get destroyed once it's passed to ocfs2_commit_truncate. */
	195	struct buffer_head *tc_last_eb_bh;
	196	};
	197
35edec1d MF	198	int ocfs2_zero_range_for_truncate(struct inode inode, handle_t handle,
35edec1d MF	199	u64 range_start, u64 range_end);
ccd979bd MF	200	int ocfs2_prepare_truncate(struct ocfs2_super *osb,
	201	struct inode *inode,
	202	struct buffer_head *fe_bh,
	203	struct ocfs2_truncate_context **tc);
	204	int ocfs2_commit_truncate(struct ocfs2_super *osb,
	205	struct inode *inode,
	206	struct buffer_head *fe_bh,
	207	struct ocfs2_truncate_context *tc);
1afc32b9 MF	208	int ocfs2_truncate_inline(struct inode inode, struct buffer_head di_bh,
1afc32b9 MF	209	unsigned int start, unsigned int end, int trunc);
ccd979bd	210
363041a5 MF	211	int ocfs2_find_leaf(struct inode inode, struct ocfs2_extent_list root_el,
363041a5 MF	212	u32 cpos, struct buffer_head **leaf_bh);
328d5752	213	int ocfs2_search_extent_list(struct ocfs2_extent_list *el, u32 v_cluster);
363041a5	214
e48edee2 MF	215	/*
	216	* Helper function to look at the # of clusters in an extent record.
	217	*/
	218	static inline unsigned int ocfs2_rec_clusters(struct ocfs2_extent_list *el,
	219	struct ocfs2_extent_rec *rec)
	220	{
	221	/*
	222	* Cluster count in extent records is slightly different
	223	* between interior nodes and leaf nodes. This is to support
	224	* unwritten extents which need a flags field in leaf node
	225	* records, thus shrinking the available space for a clusters
	226	* field.
	227	*/
	228	if (el->l_tree_depth)
	229	return le32_to_cpu(rec->e_int_clusters);
	230	else
	231	return le16_to_cpu(rec->e_leaf_clusters);
	232	}
	233
00dc417f MF	234	/*
	235	* This is only valid for leaf nodes, which are the only ones that can
	236	* have empty extents anyway.
	237	*/
	238	static inline int ocfs2_is_empty_extent(struct ocfs2_extent_rec *rec)
	239	{
	240	return !rec->e_leaf_clusters;
	241	}
	242
ccd979bd	243	#endif /* OCFS2_ALLOC_H */