mballoc.h 5.4 KB
Newer Older
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19
/*
 *  fs/ext4/mballoc.h
 *
 *  Written by: Alex Tomas <alex@clusterfs.com>
 *
 */
#ifndef _EXT4_MBALLOC_H
#define _EXT4_MBALLOC_H

#include <linux/time.h>
#include <linux/fs.h>
#include <linux/namei.h>
#include <linux/quotaops.h>
#include <linux/buffer_head.h>
#include <linux/module.h>
#include <linux/swap.h>
#include <linux/proc_fs.h>
#include <linux/pagemap.h>
#include <linux/seq_file.h>
20
#include <linux/blkdev.h>
21
#include <linux/mutex.h>
22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38
#include "ext4_jbd2.h"
#include "ext4.h"

/*
 * with AGGRESSIVE_CHECK allocator runs consistency checks over
 * structures. these checks slow things down a lot
 */
#define AGGRESSIVE_CHECK__

/*
 * with DOUBLE_CHECK defined mballoc creates persistent in-core
 * bitmaps, maintains and uses them to check for double allocations
 */
#define DOUBLE_CHECK__

/*
 */
39
#ifdef CONFIG_EXT4_DEBUG
40
extern ushort ext4_mballoc_debug;
41 42 43

#define mb_debug(n, fmt, a...)	                                        \
	do {								\
44
		if ((n) <= ext4_mballoc_debug) {		        \
45 46 47 48 49
			printk(KERN_DEBUG "(%s, %d): %s: ",		\
			       __FILE__, __LINE__, __func__);		\
			printk(fmt, ## a);				\
		}							\
	} while (0)
50
#else
51
#define mb_debug(n, fmt, a...)
52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70
#endif

#define EXT4_MB_HISTORY_ALLOC		1	/* allocation */
#define EXT4_MB_HISTORY_PREALLOC	2	/* preallocated blocks used */

/*
 * How long mballoc can look for a best extent (in found extents)
 */
#define MB_DEFAULT_MAX_TO_SCAN		200

/*
 * How long mballoc must look for a best extent
 */
#define MB_DEFAULT_MIN_TO_SCAN		10

/*
 * with 'ext4_mb_stats' allocator will collect stats that will be
 * shown at umount. The collecting costs though!
 */
71
#define MB_DEFAULT_STATS		0
72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92

/*
 * files smaller than MB_DEFAULT_STREAM_THRESHOLD are served
 * by the stream allocator, which purpose is to pack requests
 * as close each to other as possible to produce smooth I/O traffic
 * We use locality group prealloc space for stream request.
 * We can tune the same via /proc/fs/ext4/<parition>/stream_req
 */
#define MB_DEFAULT_STREAM_THRESHOLD	16	/* 64K */

/*
 * for which requests use 2^N search using buddies
 */
#define MB_DEFAULT_ORDER2_REQS		2

/*
 * default group prealloc size 512 blocks
 */
#define MB_DEFAULT_GROUP_PREALLOC	512


93
struct ext4_free_data {
B
Bobi Jam 已提交
94 95 96 97
	/* MUST be the first member */
	struct ext4_journal_cb_entry	efd_jce;

	/* ext4_free_data private data starts from here */
98

B
Bobi Jam 已提交
99 100
	/* this links the free block information from group_info */
	struct rb_node			efd_node;
101 102

	/* group which free block extent belongs */
B
Bobi Jam 已提交
103
	ext4_group_t			efd_group;
104 105

	/* free block extent */
B
Bobi Jam 已提交
106 107
	ext4_grpblk_t			efd_start_cluster;
	ext4_grpblk_t			efd_count;
108 109

	/* transaction which freed this extent */
B
Bobi Jam 已提交
110
	tid_t				efd_tid;
111 112 113 114 115 116 117 118 119 120 121 122 123 124
};

struct ext4_prealloc_space {
	struct list_head	pa_inode_list;
	struct list_head	pa_group_list;
	union {
		struct list_head pa_tmp_list;
		struct rcu_head	pa_rcu;
	} u;
	spinlock_t		pa_lock;
	atomic_t		pa_count;
	unsigned		pa_deleted;
	ext4_fsblk_t		pa_pstart;	/* phys. block */
	ext4_lblk_t		pa_lstart;	/* log. block */
125 126
	ext4_grpblk_t		pa_len;		/* len of preallocated chunk */
	ext4_grpblk_t		pa_free;	/* how many blocks are free */
127
	unsigned short		pa_type;	/* pa type. inode or group */
128 129 130 131
	spinlock_t		*pa_obj_lock;
	struct inode		*pa_inode;	/* hack, for history only */
};

132 133 134 135
enum {
	MB_INODE_PA = 0,
	MB_GROUP_PA = 1
};
136 137 138

struct ext4_free_extent {
	ext4_lblk_t fe_logical;
139
	ext4_grpblk_t fe_start;	/* In cluster units */
140
	ext4_group_t fe_group;
141
	ext4_grpblk_t fe_len;	/* In cluster units */
142 143 144 145 146 147
};

/*
 * Locality group:
 *   we try to group all related changes together
 *   so that writeback can flush/allocate them together as well
148 149 150
 *   Size of lg_prealloc_list hash is determined by MB_DEFAULT_GROUP_PREALLOC
 *   (512). We store prealloc space into the hash based on the pa_free blocks
 *   order value.ie, fls(pa_free)-1;
151
 */
152
#define PREALLOC_TB_SIZE 10
153 154
struct ext4_locality_group {
	/* for allocator */
155 156 157 158
	/* to serialize allocates */
	struct mutex		lg_mutex;
	/* list of preallocations */
	struct list_head	lg_prealloc_list[PREALLOC_TB_SIZE];
159 160 161 162 163 164 165 166 167 168
	spinlock_t		lg_prealloc_lock;
};

struct ext4_allocation_context {
	struct inode *ac_inode;
	struct super_block *ac_sb;

	/* original request */
	struct ext4_free_extent ac_o_ex;

169
	/* goal request (normalized ac_o_ex) */
170 171 172 173 174
	struct ext4_free_extent ac_g_ex;

	/* the best found extent */
	struct ext4_free_extent ac_b_ex;

175
	/* copy of the best found extent taken before preallocation efforts */
176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210
	struct ext4_free_extent ac_f_ex;

	/* number of iterations done. we have to track to limit searching */
	unsigned long ac_ex_scanned;
	__u16 ac_groups_scanned;
	__u16 ac_found;
	__u16 ac_tail;
	__u16 ac_buddy;
	__u16 ac_flags;		/* allocation hints */
	__u8 ac_status;
	__u8 ac_criteria;
	__u8 ac_2order;		/* if request is to allocate 2^N blocks and
				 * N > 0, the field stores N, otherwise 0 */
	__u8 ac_op;		/* operation, for history only */
	struct page *ac_bitmap_page;
	struct page *ac_buddy_page;
	struct ext4_prealloc_space *ac_pa;
	struct ext4_locality_group *ac_lg;
};

#define AC_STATUS_CONTINUE	1
#define AC_STATUS_FOUND		2
#define AC_STATUS_BREAK		3

struct ext4_buddy {
	struct page *bd_buddy_page;
	void *bd_buddy;
	struct page *bd_bitmap_page;
	void *bd_bitmap;
	struct ext4_group_info *bd_info;
	struct super_block *bd_sb;
	__u16 bd_blkbits;
	ext4_group_t bd_group;
};

211
static inline ext4_fsblk_t ext4_grp_offs_to_block(struct super_block *sb,
212 213
					struct ext4_free_extent *fex)
{
214 215
	return ext4_group_first_block_no(sb, fex->fe_group) +
		(fex->fe_start << EXT4_SB(sb)->s_cluster_bits);
216 217
}
#endif