Loading...
1/* SPDX-License-Identifier: GPL-2.0 */
2/*
3 * Copyright (C) 2007 Oracle. All rights reserved.
4 */
5
6#ifndef BTRFS_ORDERED_DATA_H
7#define BTRFS_ORDERED_DATA_H
8
9#include <linux/types.h>
10#include <linux/list.h>
11#include <linux/refcount.h>
12#include <linux/completion.h>
13#include <linux/rbtree.h>
14#include <linux/wait.h>
15#include "async-thread.h"
16
17struct inode;
18struct page;
19struct extent_state;
20struct btrfs_inode;
21struct btrfs_root;
22struct btrfs_fs_info;
23
24struct btrfs_ordered_sum {
25 /*
26 * Logical start address and length for of the blocks covered by
27 * the sums array.
28 */
29 u64 logical;
30 u32 len;
31
32 struct list_head list;
33 /* last field is a variable length array of csums */
34 u8 sums[];
35};
36
37/*
38 * Bits for btrfs_ordered_extent::flags.
39 *
40 * BTRFS_ORDERED_IO_DONE is set when all of the blocks are written.
41 * It is used to make sure metadata is inserted into the tree only once
42 * per extent.
43 *
44 * BTRFS_ORDERED_COMPLETE is set when the extent is removed from the
45 * rbtree, just before waking any waiters. It is used to indicate the
46 * IO is done and any metadata is inserted into the tree.
47 */
48enum {
49 /*
50 * Different types for ordered extents, one and only one of the 4 types
51 * need to be set when creating ordered extent.
52 *
53 * REGULAR: For regular non-compressed COW write
54 * NOCOW: For NOCOW write into existing non-hole extent
55 * PREALLOC: For NOCOW write into preallocated extent
56 * COMPRESSED: For compressed COW write
57 */
58 BTRFS_ORDERED_REGULAR,
59 BTRFS_ORDERED_NOCOW,
60 BTRFS_ORDERED_PREALLOC,
61 BTRFS_ORDERED_COMPRESSED,
62
63 /*
64 * Extra bit for direct io, can only be set for
65 * REGULAR/NOCOW/PREALLOC. No direct io for compressed extent.
66 */
67 BTRFS_ORDERED_DIRECT,
68
69 /* Extra status bits for ordered extents */
70
71 /* set when all the pages are written */
72 BTRFS_ORDERED_IO_DONE,
73 /* set when removed from the tree */
74 BTRFS_ORDERED_COMPLETE,
75 /* We had an io error when writing this out */
76 BTRFS_ORDERED_IOERR,
77 /* Set when we have to truncate an extent */
78 BTRFS_ORDERED_TRUNCATED,
79 /* Used during fsync to track already logged extents */
80 BTRFS_ORDERED_LOGGED,
81 /* We have already logged all the csums of the ordered extent */
82 BTRFS_ORDERED_LOGGED_CSUM,
83 /* We wait for this extent to complete in the current transaction */
84 BTRFS_ORDERED_PENDING,
85 /* BTRFS_IOC_ENCODED_WRITE */
86 BTRFS_ORDERED_ENCODED,
87};
88
89/* BTRFS_ORDERED_* flags that specify the type of the extent. */
90#define BTRFS_ORDERED_TYPE_FLAGS ((1UL << BTRFS_ORDERED_REGULAR) | \
91 (1UL << BTRFS_ORDERED_NOCOW) | \
92 (1UL << BTRFS_ORDERED_PREALLOC) | \
93 (1UL << BTRFS_ORDERED_COMPRESSED) | \
94 (1UL << BTRFS_ORDERED_DIRECT) | \
95 (1UL << BTRFS_ORDERED_ENCODED))
96
97struct btrfs_ordered_extent {
98 /* logical offset in the file */
99 u64 file_offset;
100
101 /*
102 * These fields directly correspond to the same fields in
103 * btrfs_file_extent_item.
104 */
105 u64 num_bytes;
106 u64 ram_bytes;
107 u64 disk_bytenr;
108 u64 disk_num_bytes;
109 u64 offset;
110
111 /* number of bytes that still need writing */
112 u64 bytes_left;
113
114 /*
115 * If we get truncated we need to adjust the file extent we enter for
116 * this ordered extent so that we do not expose stale data.
117 */
118 u64 truncated_len;
119
120 /* flags (described above) */
121 unsigned long flags;
122
123 /* compression algorithm */
124 int compress_type;
125
126 /* Qgroup reserved space */
127 int qgroup_rsv;
128
129 /* reference count */
130 refcount_t refs;
131
132 /* the inode we belong to */
133 struct btrfs_inode *inode;
134
135 /* list of checksums for insertion when the extent io is done */
136 struct list_head list;
137
138 /* used for fast fsyncs */
139 struct list_head log_list;
140
141 /* used to wait for the BTRFS_ORDERED_COMPLETE bit */
142 wait_queue_head_t wait;
143
144 /* our friendly rbtree entry */
145 struct rb_node rb_node;
146
147 /* a per root list of all the pending ordered extents */
148 struct list_head root_extent_list;
149
150 struct btrfs_work work;
151
152 struct completion completion;
153 struct btrfs_work flush_work;
154 struct list_head work_list;
155
156 struct list_head bioc_list;
157};
158
159int btrfs_finish_one_ordered(struct btrfs_ordered_extent *ordered_extent);
160int btrfs_finish_ordered_io(struct btrfs_ordered_extent *ordered_extent);
161
162void btrfs_put_ordered_extent(struct btrfs_ordered_extent *entry);
163void btrfs_remove_ordered_extent(struct btrfs_inode *btrfs_inode,
164 struct btrfs_ordered_extent *entry);
165void btrfs_finish_ordered_extent(struct btrfs_ordered_extent *ordered,
166 struct folio *folio, u64 file_offset, u64 len,
167 bool uptodate);
168void btrfs_mark_ordered_io_finished(struct btrfs_inode *inode,
169 struct folio *folio, u64 file_offset,
170 u64 num_bytes, bool uptodate);
171bool btrfs_dec_test_ordered_pending(struct btrfs_inode *inode,
172 struct btrfs_ordered_extent **cached,
173 u64 file_offset, u64 io_size);
174
175/*
176 * This represents details about the target file extent item of a write operation.
177 */
178struct btrfs_file_extent {
179 u64 disk_bytenr;
180 u64 disk_num_bytes;
181 u64 num_bytes;
182 u64 ram_bytes;
183 u64 offset;
184 u8 compression;
185};
186
187struct btrfs_ordered_extent *btrfs_alloc_ordered_extent(
188 struct btrfs_inode *inode, u64 file_offset,
189 const struct btrfs_file_extent *file_extent, unsigned long flags);
190void btrfs_add_ordered_sum(struct btrfs_ordered_extent *entry,
191 struct btrfs_ordered_sum *sum);
192struct btrfs_ordered_extent *btrfs_lookup_ordered_extent(struct btrfs_inode *inode,
193 u64 file_offset);
194void btrfs_start_ordered_extent(struct btrfs_ordered_extent *entry);
195int btrfs_wait_ordered_range(struct btrfs_inode *inode, u64 start, u64 len);
196struct btrfs_ordered_extent *
197btrfs_lookup_first_ordered_extent(struct btrfs_inode *inode, u64 file_offset);
198struct btrfs_ordered_extent *btrfs_lookup_first_ordered_range(
199 struct btrfs_inode *inode, u64 file_offset, u64 len);
200struct btrfs_ordered_extent *btrfs_lookup_ordered_range(
201 struct btrfs_inode *inode,
202 u64 file_offset,
203 u64 len);
204void btrfs_get_ordered_extents_for_logging(struct btrfs_inode *inode,
205 struct list_head *list);
206u64 btrfs_wait_ordered_extents(struct btrfs_root *root, u64 nr,
207 const struct btrfs_block_group *bg);
208void btrfs_wait_ordered_roots(struct btrfs_fs_info *fs_info, u64 nr,
209 const struct btrfs_block_group *bg);
210void btrfs_lock_and_flush_ordered_range(struct btrfs_inode *inode, u64 start,
211 u64 end,
212 struct extent_state **cached_state);
213bool btrfs_try_lock_ordered_range(struct btrfs_inode *inode, u64 start, u64 end,
214 struct extent_state **cached_state);
215struct btrfs_ordered_extent *btrfs_split_ordered_extent(
216 struct btrfs_ordered_extent *ordered, u64 len);
217void btrfs_mark_ordered_extent_error(struct btrfs_ordered_extent *ordered);
218int __init ordered_data_init(void);
219void __cold ordered_data_exit(void);
220
221#endif
1/*
2 * Copyright (C) 2007 Oracle. All rights reserved.
3 *
4 * This program is free software; you can redistribute it and/or
5 * modify it under the terms of the GNU General Public
6 * License v2 as published by the Free Software Foundation.
7 *
8 * This program is distributed in the hope that it will be useful,
9 * but WITHOUT ANY WARRANTY; without even the implied warranty of
10 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
11 * General Public License for more details.
12 *
13 * You should have received a copy of the GNU General Public
14 * License along with this program; if not, write to the
15 * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
16 * Boston, MA 021110-1307, USA.
17 */
18
19#ifndef __BTRFS_ORDERED_DATA__
20#define __BTRFS_ORDERED_DATA__
21
22/* one of these per inode */
23struct btrfs_ordered_inode_tree {
24 spinlock_t lock;
25 struct rb_root tree;
26 struct rb_node *last;
27};
28
29struct btrfs_ordered_sum {
30 /* bytenr is the start of this extent on disk */
31 u64 bytenr;
32
33 /*
34 * this is the length in bytes covered by the sums array below.
35 */
36 int len;
37 struct list_head list;
38 /* last field is a variable length array of csums */
39 u32 sums[];
40};
41
42/*
43 * bits for the flags field:
44 *
45 * BTRFS_ORDERED_IO_DONE is set when all of the blocks are written.
46 * It is used to make sure metadata is inserted into the tree only once
47 * per extent.
48 *
49 * BTRFS_ORDERED_COMPLETE is set when the extent is removed from the
50 * rbtree, just before waking any waiters. It is used to indicate the
51 * IO is done and any metadata is inserted into the tree.
52 */
53#define BTRFS_ORDERED_IO_DONE 0 /* set when all the pages are written */
54
55#define BTRFS_ORDERED_COMPLETE 1 /* set when removed from the tree */
56
57#define BTRFS_ORDERED_NOCOW 2 /* set when we want to write in place */
58
59#define BTRFS_ORDERED_COMPRESSED 3 /* writing a zlib compressed extent */
60
61#define BTRFS_ORDERED_PREALLOC 4 /* set when writing to prealloced extent */
62
63#define BTRFS_ORDERED_DIRECT 5 /* set when we're doing DIO with this extent */
64
65#define BTRFS_ORDERED_IOERR 6 /* We had an io error when writing this out */
66
67#define BTRFS_ORDERED_UPDATED_ISIZE 7 /* indicates whether this ordered extent
68 * has done its due diligence in updating
69 * the isize. */
70#define BTRFS_ORDERED_LOGGED_CSUM 8 /* We've logged the csums on this ordered
71 ordered extent */
72#define BTRFS_ORDERED_TRUNCATED 9 /* Set when we have to truncate an extent */
73
74struct btrfs_ordered_extent {
75 /* logical offset in the file */
76 u64 file_offset;
77
78 /* disk byte number */
79 u64 start;
80
81 /* ram length of the extent in bytes */
82 u64 len;
83
84 /* extent length on disk */
85 u64 disk_len;
86
87 /* number of bytes that still need writing */
88 u64 bytes_left;
89
90 /* number of bytes that still need csumming */
91 u64 csum_bytes_left;
92
93 /*
94 * the end of the ordered extent which is behind it but
95 * didn't update disk_i_size. Please see the comment of
96 * btrfs_ordered_update_i_size();
97 */
98 u64 outstanding_isize;
99
100 /*
101 * If we get truncated we need to adjust the file extent we enter for
102 * this ordered extent so that we do not expose stale data.
103 */
104 u64 truncated_len;
105
106 /* flags (described above) */
107 unsigned long flags;
108
109 /* compression algorithm */
110 int compress_type;
111
112 /* reference count */
113 atomic_t refs;
114
115 /* the inode we belong to */
116 struct inode *inode;
117
118 /* list of checksums for insertion when the extent io is done */
119 struct list_head list;
120
121 /* If we need to wait on this to be done */
122 struct list_head log_list;
123
124 /* used to wait for the BTRFS_ORDERED_COMPLETE bit */
125 wait_queue_head_t wait;
126
127 /* our friendly rbtree entry */
128 struct rb_node rb_node;
129
130 /* a per root list of all the pending ordered extents */
131 struct list_head root_extent_list;
132
133 struct btrfs_work work;
134
135 struct completion completion;
136 struct btrfs_work flush_work;
137 struct list_head work_list;
138};
139
140/*
141 * calculates the total size you need to allocate for an ordered sum
142 * structure spanning 'bytes' in the file
143 */
144static inline int btrfs_ordered_sum_size(struct btrfs_root *root,
145 unsigned long bytes)
146{
147 int num_sectors = (int)DIV_ROUND_UP(bytes, root->sectorsize);
148 return sizeof(struct btrfs_ordered_sum) + num_sectors * sizeof(u32);
149}
150
151static inline void
152btrfs_ordered_inode_tree_init(struct btrfs_ordered_inode_tree *t)
153{
154 spin_lock_init(&t->lock);
155 t->tree = RB_ROOT;
156 t->last = NULL;
157}
158
159void btrfs_put_ordered_extent(struct btrfs_ordered_extent *entry);
160void btrfs_remove_ordered_extent(struct inode *inode,
161 struct btrfs_ordered_extent *entry);
162int btrfs_dec_test_ordered_pending(struct inode *inode,
163 struct btrfs_ordered_extent **cached,
164 u64 file_offset, u64 io_size, int uptodate);
165int btrfs_dec_test_first_ordered_pending(struct inode *inode,
166 struct btrfs_ordered_extent **cached,
167 u64 *file_offset, u64 io_size,
168 int uptodate);
169int btrfs_add_ordered_extent(struct inode *inode, u64 file_offset,
170 u64 start, u64 len, u64 disk_len, int type);
171int btrfs_add_ordered_extent_dio(struct inode *inode, u64 file_offset,
172 u64 start, u64 len, u64 disk_len, int type);
173int btrfs_add_ordered_extent_compress(struct inode *inode, u64 file_offset,
174 u64 start, u64 len, u64 disk_len,
175 int type, int compress_type);
176void btrfs_add_ordered_sum(struct inode *inode,
177 struct btrfs_ordered_extent *entry,
178 struct btrfs_ordered_sum *sum);
179struct btrfs_ordered_extent *btrfs_lookup_ordered_extent(struct inode *inode,
180 u64 file_offset);
181void btrfs_start_ordered_extent(struct inode *inode,
182 struct btrfs_ordered_extent *entry, int wait);
183int btrfs_wait_ordered_range(struct inode *inode, u64 start, u64 len);
184struct btrfs_ordered_extent *
185btrfs_lookup_first_ordered_extent(struct inode * inode, u64 file_offset);
186struct btrfs_ordered_extent *btrfs_lookup_ordered_range(struct inode *inode,
187 u64 file_offset,
188 u64 len);
189int btrfs_ordered_update_i_size(struct inode *inode, u64 offset,
190 struct btrfs_ordered_extent *ordered);
191int btrfs_find_ordered_sum(struct inode *inode, u64 offset, u64 disk_bytenr,
192 u32 *sum, int len);
193int btrfs_run_ordered_operations(struct btrfs_trans_handle *trans,
194 struct btrfs_root *root, int wait);
195void btrfs_add_ordered_operation(struct btrfs_trans_handle *trans,
196 struct btrfs_root *root,
197 struct inode *inode);
198int btrfs_wait_ordered_extents(struct btrfs_root *root, int nr);
199void btrfs_wait_ordered_roots(struct btrfs_fs_info *fs_info, int nr);
200void btrfs_get_logged_extents(struct inode *inode,
201 struct list_head *logged_list);
202void btrfs_put_logged_extents(struct list_head *logged_list);
203void btrfs_submit_logged_extents(struct list_head *logged_list,
204 struct btrfs_root *log);
205void btrfs_wait_logged_extents(struct btrfs_root *log, u64 transid);
206void btrfs_free_logged_extents(struct btrfs_root *log, u64 transid);
207int __init ordered_data_init(void);
208void ordered_data_exit(void);
209#endif