2006-10-11 12:20:50 +04:00
/*
2008-04-30 02:13:32 +04:00
* ext4 . h
2006-10-11 12:20:50 +04:00
*
* Copyright ( C ) 1992 , 1993 , 1994 , 1995
* Remy Card ( card @ masi . ibp . fr )
* Laboratoire MASI - Institut Blaise Pascal
* Universite Pierre et Marie Curie ( Paris VI )
*
* from
*
* linux / include / linux / minix_fs . h
*
* Copyright ( C ) 1991 , 1992 Linus Torvalds
*/
2008-04-30 02:13:32 +04:00
# ifndef _EXT4_H
# define _EXT4_H
2006-10-11 12:20:50 +04:00
# include <linux/types.h>
2006-10-11 12:21:05 +04:00
# include <linux/blkdev.h>
2006-10-11 12:20:50 +04:00
# include <linux/magic.h>
2009-01-06 05:49:55 +03:00
# include <linux/jbd2.h>
2009-01-22 20:13:05 +03:00
# include <linux/quota.h>
2009-05-01 21:44:33 +04:00
# include <linux/rwsem.h>
# include <linux/rbtree.h>
# include <linux/seqlock.h>
# include <linux/mutex.h>
2009-05-04 00:33:44 +04:00
# include <linux/timer.h>
# include <linux/wait.h>
# include <linux/blockgroup_lock.h>
# include <linux/percpu_counter.h>
2013-10-18 05:11:01 +04:00
# include <linux/ratelimit.h>
2012-04-30 02:27:10 +04:00
# include <crypto/hash.h>
2014-02-22 15:18:17 +04:00
# include <linux/falloc.h>
2010-05-17 14:00:00 +04:00
# ifdef __KERNEL__
# include <linux/compat.h>
# endif
2008-01-29 08:19:52 +03:00
2006-10-11 12:20:50 +04:00
/*
2008-07-12 03:27:31 +04:00
* The fourth extended filesystem constants / structures
2006-10-11 12:20:50 +04:00
*/
/*
2006-10-11 12:20:53 +04:00
* Define EXT4FS_DEBUG to produce debug messages
2006-10-11 12:20:50 +04:00
*/
2006-10-11 12:20:53 +04:00
# undef EXT4FS_DEBUG
2006-10-11 12:20:50 +04:00
/*
* Debug code
*/
2006-10-11 12:20:53 +04:00
# ifdef EXT4FS_DEBUG
# define ext4_debug(f, a...) \
2006-10-11 12:20:50 +04:00
do { \
2008-09-09 06:25:24 +04:00
printk ( KERN_DEBUG " EXT4-fs DEBUG (%s, %d): %s: " , \
2008-07-14 05:03:29 +04:00
__FILE__ , __LINE__ , __func__ ) ; \
2008-09-09 06:25:24 +04:00
printk ( KERN_DEBUG f , # # a ) ; \
2006-10-11 12:20:50 +04:00
} while ( 0 )
# else
2012-03-20 07:11:43 +04:00
# define ext4_debug(fmt, ...) no_printk(fmt, ##__VA_ARGS__)
2006-10-11 12:20:50 +04:00
# endif
2012-11-28 22:03:30 +04:00
/*
* Turn on EXT_DEBUG to get lots of info about extents operations .
*/
# define EXT_DEBUG__
# ifdef EXT_DEBUG
# define ext_debug(fmt, ...) printk(fmt, ##__VA_ARGS__)
# else
# define ext_debug(fmt, ...) no_printk(fmt, ##__VA_ARGS__)
# endif
2010-03-02 19:46:09 +03:00
# define EXT4_ERROR_INODE(inode, fmt, a...) \
2010-07-27 19:56:40 +04:00
ext4_error_inode ( ( inode ) , __func__ , __LINE__ , 0 , ( fmt ) , # # a )
# define EXT4_ERROR_INODE_BLOCK(inode, block, fmt, a...) \
ext4_error_inode ( ( inode ) , __func__ , __LINE__ , ( block ) , ( fmt ) , # # a )
2010-03-02 19:46:09 +03:00
2011-01-10 20:10:55 +03:00
# define EXT4_ERROR_FILE(file, block, fmt, a...) \
ext4_error_file ( ( file ) , __func__ , __LINE__ , ( block ) , ( fmt ) , # # a )
2010-03-02 19:46:09 +03:00
2009-05-01 21:44:33 +04:00
/* data type for block offset of block group */
typedef int ext4_grpblk_t ;
/* data type for filesystem-wide blocks number */
typedef unsigned long long ext4_fsblk_t ;
/* data type for file logical block number */
typedef __u32 ext4_lblk_t ;
/* data type for block group number */
typedef unsigned int ext4_group_t ;
2009-09-30 08:32:42 +04:00
/*
2010-05-17 15:00:00 +04:00
* Flags used in mballoc ' s allocation_context flags field .
2009-09-30 08:32:42 +04:00
*
* Also used to show what ' s going on for debugging purposes when the
* flag field is exported via the traceport interface
*/
2009-05-01 21:44:33 +04:00
2008-01-29 08:19:52 +03:00
/* prefer goal again. length */
2009-08-10 00:46:13 +04:00
# define EXT4_MB_HINT_MERGE 0x0001
2008-01-29 08:19:52 +03:00
/* blocks already reserved */
2009-08-10 00:46:13 +04:00
# define EXT4_MB_HINT_RESERVED 0x0002
2008-01-29 08:19:52 +03:00
/* metadata is being allocated */
2009-08-10 00:46:13 +04:00
# define EXT4_MB_HINT_METADATA 0x0004
2008-01-29 08:19:52 +03:00
/* first blocks in the file */
2009-08-10 00:46:13 +04:00
# define EXT4_MB_HINT_FIRST 0x0008
2008-01-29 08:19:52 +03:00
/* search for the best chunk */
2009-08-10 00:46:13 +04:00
# define EXT4_MB_HINT_BEST 0x0010
2008-01-29 08:19:52 +03:00
/* data is being allocated */
2009-08-10 00:46:13 +04:00
# define EXT4_MB_HINT_DATA 0x0020
2008-01-29 08:19:52 +03:00
/* don't preallocate (for tails) */
2009-08-10 00:46:13 +04:00
# define EXT4_MB_HINT_NOPREALLOC 0x0040
2008-01-29 08:19:52 +03:00
/* allocate for locality group */
2009-08-10 00:46:13 +04:00
# define EXT4_MB_HINT_GROUP_ALLOC 0x0080
2008-01-29 08:19:52 +03:00
/* allocate goal blocks or none */
2009-08-10 00:46:13 +04:00
# define EXT4_MB_HINT_GOAL_ONLY 0x0100
2008-01-29 08:19:52 +03:00
/* goal is meaningful */
2009-08-10 00:46:13 +04:00
# define EXT4_MB_HINT_TRY_GOAL 0x0200
2008-07-15 01:52:37 +04:00
/* blocks already pre-reserved by delayed allocation */
2009-08-10 00:46:13 +04:00
# define EXT4_MB_DELALLOC_RESERVED 0x0400
2009-08-10 06:01:13 +04:00
/* We are doing stream allocation */
# define EXT4_MB_STREAM_ALLOC 0x0800
2011-05-25 15:41:26 +04:00
/* Use reserved root blocks if needed */
# define EXT4_MB_USE_ROOT_BLOCKS 0x1000
2013-04-10 06:11:22 +04:00
/* Use blocks from reserved pool */
# define EXT4_MB_USE_RESERVED 0x2000
2008-01-29 08:19:52 +03:00
struct ext4_allocation_request {
/* target inode for block we're allocating */
struct inode * inode ;
2009-07-13 18:24:17 +04:00
/* how many blocks we want to allocate */
unsigned int len ;
2008-01-29 08:19:52 +03:00
/* logical block in target inode */
ext4_lblk_t logical ;
/* the closest logical allocated block to the left */
ext4_lblk_t lleft ;
/* the closest logical allocated block to the right */
ext4_lblk_t lright ;
2009-07-13 18:24:17 +04:00
/* phys. target (a hint) */
ext4_fsblk_t goal ;
/* phys. block for the closest logical allocated block to the left */
ext4_fsblk_t pleft ;
/* phys. block for the closest logical allocated block to the right */
2008-01-29 08:19:52 +03:00
ext4_fsblk_t pright ;
/* flags. see above EXT4_MB_HINT_* */
2008-11-05 08:14:04 +03:00
unsigned int flags ;
2008-01-29 08:19:52 +03:00
} ;
2010-05-17 03:00:00 +04:00
/*
* Logical to physical block mapping , used by ext4_map_blocks ( )
*
* This structure is used to pass requests into ext4_map_blocks ( ) as
* well as to store the information returned by ext4_map_blocks ( ) . It
* takes less room on the stack than a struct buffer_head .
*/
# define EXT4_MAP_NEW (1 << BH_New)
# define EXT4_MAP_MAPPED (1 << BH_Mapped)
# define EXT4_MAP_UNWRITTEN (1 << BH_Unwritten)
# define EXT4_MAP_BOUNDARY (1 << BH_Boundary)
2011-09-10 03:04:51 +04:00
/* Sometimes (in the bigalloc case, from ext4_da_get_block_prep) the caller of
* ext4_map_blocks wants to know whether or not the underlying cluster has
* already been accounted for . EXT4_MAP_FROM_CLUSTER conveys to the caller that
* the requested mapping was from previously mapped ( or delayed allocated )
* cluster . We use BH_AllocFromCluster only for this flag . BH_AllocFromCluster
* should never appear on buffer_head ' s state flags .
*/
# define EXT4_MAP_FROM_CLUSTER (1 << BH_AllocFromCluster)
2010-05-17 03:00:00 +04:00
# define EXT4_MAP_FLAGS (EXT4_MAP_NEW | EXT4_MAP_MAPPED |\
EXT4_MAP_UNWRITTEN | EXT4_MAP_BOUNDARY | \
2014-04-21 07:44:47 +04:00
EXT4_MAP_FROM_CLUSTER )
2010-05-17 03:00:00 +04:00
struct ext4_map_blocks {
ext4_fsblk_t m_pblk ;
ext4_lblk_t m_lblk ;
unsigned int m_len ;
unsigned int m_flags ;
} ;
2010-10-28 05:30:10 +04:00
/*
* Flags for ext4_io_end - > flags
*/
# define EXT4_IO_END_UNWRITTEN 0x0001
2012-03-05 19:40:22 +04:00
/*
2014-04-21 07:45:47 +04:00
* For converting unwritten extents on a work queue . ' handle ' is used for
2013-06-04 21:21:11 +04:00
* buffered writeback .
2012-03-05 19:40:22 +04:00
*/
2009-09-28 23:49:08 +04:00
typedef struct ext4_io_end {
2010-07-27 19:56:06 +04:00
struct list_head list ; /* per-file finished IO list */
2013-06-04 21:21:11 +04:00
handle_t * handle ; /* handle reserved for extent
* conversion */
2009-09-28 23:49:08 +04:00
struct inode * inode ; /* file being written to */
2013-06-04 22:23:41 +04:00
struct bio * bio ; /* Linked list of completed
* bios covering the extent */
2009-09-28 23:48:29 +04:00
unsigned int flag ; /* unwritten or not */
2010-02-05 07:58:38 +03:00
loff_t offset ; /* offset in the file */
ssize_t size ; /* size of the extent */
2013-06-04 19:58:58 +04:00
atomic_t count ; /* reference counter */
2009-09-28 23:49:08 +04:00
} ext4_io_end_t ;
2010-10-28 05:30:10 +04:00
struct ext4_io_submit {
int io_op ;
struct bio * io_bio ;
ext4_io_end_t * io_end ;
sector_t io_next_block ;
} ;
2006-10-11 12:20:50 +04:00
/*
* Special inodes numbers
*/
2006-10-11 12:20:53 +04:00
# define EXT4_BAD_INO 1 /* Bad blocks inode */
# define EXT4_ROOT_INO 2 /* Root inode */
2011-05-25 03:00:39 +04:00
# define EXT4_USR_QUOTA_INO 3 /* User quota inode */
# define EXT4_GRP_QUOTA_INO 4 /* Group quota inode */
2006-10-11 12:20:53 +04:00
# define EXT4_BOOT_LOADER_INO 5 /* Boot loader inode */
# define EXT4_UNDEL_DIR_INO 6 /* Undelete directory inode */
# define EXT4_RESIZE_INO 7 /* Reserved group descriptors inode */
# define EXT4_JOURNAL_INO 8 /* Journal inode */
2006-10-11 12:20:50 +04:00
2006-10-11 12:20:53 +04:00
/* First non-reserved inode for old ext4 filesystems */
# define EXT4_GOOD_OLD_FIRST_INO 11
2006-10-11 12:20:50 +04:00
/*
* Maximal count of links to a file
*/
2007-07-18 16:38:01 +04:00
# define EXT4_LINK_MAX 65000
2006-10-11 12:20:50 +04:00
/*
* Macro - instructions used to manage several block sizes
*/
2006-10-11 12:20:53 +04:00
# define EXT4_MIN_BLOCK_SIZE 1024
2008-01-29 07:58:27 +03:00
# define EXT4_MAX_BLOCK_SIZE 65536
# define EXT4_MIN_BLOCK_LOG_SIZE 10
2010-10-28 05:29:12 +04:00
# define EXT4_MAX_BLOCK_LOG_SIZE 16
2006-10-11 12:20:50 +04:00
# ifdef __KERNEL__
2006-10-11 12:20:53 +04:00
# define EXT4_BLOCK_SIZE(s) ((s)->s_blocksize)
2006-10-11 12:20:50 +04:00
# else
2006-10-11 12:20:53 +04:00
# define EXT4_BLOCK_SIZE(s) (EXT4_MIN_BLOCK_SIZE << (s)->s_log_block_size)
2006-10-11 12:20:50 +04:00
# endif
2008-09-09 06:25:24 +04:00
# define EXT4_ADDR_PER_BLOCK(s) (EXT4_BLOCK_SIZE(s) / sizeof(__u32))
2011-09-10 02:34:51 +04:00
# define EXT4_CLUSTER_SIZE(s) (EXT4_BLOCK_SIZE(s) << \
EXT4_SB ( s ) - > s_cluster_bits )
2006-10-11 12:20:50 +04:00
# ifdef __KERNEL__
2006-10-11 12:20:53 +04:00
# define EXT4_BLOCK_SIZE_BITS(s) ((s)->s_blocksize_bits)
2011-09-10 02:34:51 +04:00
# define EXT4_CLUSTER_BITS(s) (EXT4_SB(s)->s_cluster_bits)
2006-10-11 12:20:50 +04:00
# else
2006-10-11 12:20:53 +04:00
# define EXT4_BLOCK_SIZE_BITS(s) ((s)->s_log_block_size + 10)
2006-10-11 12:20:50 +04:00
# endif
# ifdef __KERNEL__
2006-10-11 12:20:53 +04:00
# define EXT4_ADDR_PER_BLOCK_BITS(s) (EXT4_SB(s)->s_addr_per_block_bits)
# define EXT4_INODE_SIZE(s) (EXT4_SB(s)->s_inode_size)
# define EXT4_FIRST_INO(s) (EXT4_SB(s)->s_first_ino)
2006-10-11 12:20:50 +04:00
# else
2006-10-11 12:20:53 +04:00
# define EXT4_INODE_SIZE(s) (((s)->s_rev_level == EXT4_GOOD_OLD_REV) ? \
EXT4_GOOD_OLD_INODE_SIZE : \
2006-10-11 12:20:50 +04:00
( s ) - > s_inode_size )
2006-10-11 12:20:53 +04:00
# define EXT4_FIRST_INO(s) (((s)->s_rev_level == EXT4_GOOD_OLD_REV) ? \
EXT4_GOOD_OLD_FIRST_INO : \
2006-10-11 12:20:50 +04:00
( s ) - > s_first_ino )
# endif
2007-07-18 05:42:41 +04:00
# define EXT4_BLOCK_ALIGN(size, blkbits) ALIGN((size), (1 << (blkbits)))
2006-10-11 12:20:50 +04:00
2011-09-10 02:44:51 +04:00
/* Translate a block number to a cluster number */
# define EXT4_B2C(sbi, blk) ((blk) >> (sbi)->s_cluster_bits)
/* Translate a cluster number to a block number */
# define EXT4_C2B(sbi, cluster) ((cluster) << (sbi)->s_cluster_bits)
/* Translate # of blks to # of clusters */
# define EXT4_NUM_B2C(sbi, blks) (((blks) + (sbi)->s_cluster_ratio - 1) >> \
( sbi ) - > s_cluster_bits )
2013-12-20 18:29:35 +04:00
/* Mask out the low bits to get the starting block of the cluster */
# define EXT4_PBLK_CMASK(s, pblk) ((pblk) & \
~ ( ( ext4_fsblk_t ) ( s ) - > s_cluster_ratio - 1 ) )
# define EXT4_LBLK_CMASK(s, lblk) ((lblk) & \
~ ( ( ext4_lblk_t ) ( s ) - > s_cluster_ratio - 1 ) )
/* Get the cluster offset */
# define EXT4_PBLK_COFF(s, pblk) ((pblk) & \
( ( ext4_fsblk_t ) ( s ) - > s_cluster_ratio - 1 ) )
# define EXT4_LBLK_COFF(s, lblk) ((lblk) & \
( ( ext4_lblk_t ) ( s ) - > s_cluster_ratio - 1 ) )
2011-09-10 02:44:51 +04:00
2006-10-11 12:20:50 +04:00
/*
* Structure of a blocks group descriptor
*/
2006-10-11 12:20:53 +04:00
struct ext4_group_desc
2006-10-11 12:20:50 +04:00
{
2007-10-17 02:38:25 +04:00
__le32 bg_block_bitmap_lo ; /* Blocks bitmap block */
2007-10-17 02:38:25 +04:00
__le32 bg_inode_bitmap_lo ; /* Inodes bitmap block */
__le32 bg_inode_table_lo ; /* Inodes table block */
2009-01-06 06:20:24 +03:00
__le16 bg_free_blocks_count_lo ; /* Free blocks count */
__le16 bg_free_inodes_count_lo ; /* Free inodes count */
__le16 bg_used_dirs_count_lo ; /* Directories count */
Ext4: Uninitialized Block Groups
In pass1 of e2fsck, every inode table in the fileystem is scanned and checked,
regardless of whether it is in use. This is this the most time consuming part
of the filesystem check. The unintialized block group feature can greatly
reduce e2fsck time by eliminating checking of uninitialized inodes.
With this feature, there is a a high water mark of used inodes for each block
group. Block and inode bitmaps can be uninitialized on disk via a flag in the
group descriptor to avoid reading or scanning them at e2fsck time. A checksum
of each group descriptor is used to ensure that corruption in the group
descriptor's bit flags does not cause incorrect operation.
The feature is enabled through a mkfs option
mke2fs /dev/ -O uninit_groups
A patch adding support for uninitialized block groups to e2fsprogs tools has
been posted to the linux-ext4 mailing list.
The patches have been stress tested with fsstress and fsx. In performance
tests testing e2fsck time, we have seen that e2fsck time on ext3 grows
linearly with the total number of inodes in the filesytem. In ext4 with the
uninitialized block groups feature, the e2fsck time is constant, based
solely on the number of used inodes rather than the total inode count.
Since typical ext4 filesystems only use 1-10% of their inodes, this feature can
greatly reduce e2fsck time for users. With performance improvement of 2-20
times, depending on how full the filesystem is.
The attached graph shows the major improvements in e2fsck times in filesystems
with a large total inode count, but few inodes in use.
In each group descriptor if we have
EXT4_BG_INODE_UNINIT set in bg_flags:
Inode table is not initialized/used in this group. So we can skip
the consistency check during fsck.
EXT4_BG_BLOCK_UNINIT set in bg_flags:
No block in the group is used. So we can skip the block bitmap
verification for this group.
We also add two new fields to group descriptor as a part of
uninitialized group patch.
__le16 bg_itable_unused; /* Unused inodes count */
__le16 bg_checksum; /* crc16(sb_uuid+group+desc) */
bg_itable_unused:
If we have EXT4_BG_INODE_UNINIT not set in bg_flags
then bg_itable_unused will give the offset within
the inode table till the inodes are used. This can be
used by fsck to skip list of inodes that are marked unused.
bg_checksum:
Now that we depend on bg_flags and bg_itable_unused to determine
the block and inode usage, we need to make sure group descriptor
is not corrupt. We add checksum to group descriptor to
detect corruption. If the descriptor is found to be corrupt, we
mark all the blocks and inodes in the group used.
Signed-off-by: Avantika Mathur <mathur@us.ibm.com>
Signed-off-by: Andreas Dilger <adilger@clusterfs.com>
Signed-off-by: Mingming Cao <cmm@us.ibm.com>
Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>
2007-10-17 02:38:25 +04:00
__le16 bg_flags ; /* EXT4_BG_flags (INODE_UNINIT, etc) */
2012-04-30 02:23:10 +04:00
__le32 bg_exclude_bitmap_lo ; /* Exclude bitmap for snapshots */
__le16 bg_block_bitmap_csum_lo ; /* crc32c(s_uuid+grp_num+bbitmap) LE */
__le16 bg_inode_bitmap_csum_lo ; /* crc32c(s_uuid+grp_num+ibitmap) LE */
2009-01-06 06:20:24 +03:00
__le16 bg_itable_unused_lo ; /* Unused inodes count */
Ext4: Uninitialized Block Groups
In pass1 of e2fsck, every inode table in the fileystem is scanned and checked,
regardless of whether it is in use. This is this the most time consuming part
of the filesystem check. The unintialized block group feature can greatly
reduce e2fsck time by eliminating checking of uninitialized inodes.
With this feature, there is a a high water mark of used inodes for each block
group. Block and inode bitmaps can be uninitialized on disk via a flag in the
group descriptor to avoid reading or scanning them at e2fsck time. A checksum
of each group descriptor is used to ensure that corruption in the group
descriptor's bit flags does not cause incorrect operation.
The feature is enabled through a mkfs option
mke2fs /dev/ -O uninit_groups
A patch adding support for uninitialized block groups to e2fsprogs tools has
been posted to the linux-ext4 mailing list.
The patches have been stress tested with fsstress and fsx. In performance
tests testing e2fsck time, we have seen that e2fsck time on ext3 grows
linearly with the total number of inodes in the filesytem. In ext4 with the
uninitialized block groups feature, the e2fsck time is constant, based
solely on the number of used inodes rather than the total inode count.
Since typical ext4 filesystems only use 1-10% of their inodes, this feature can
greatly reduce e2fsck time for users. With performance improvement of 2-20
times, depending on how full the filesystem is.
The attached graph shows the major improvements in e2fsck times in filesystems
with a large total inode count, but few inodes in use.
In each group descriptor if we have
EXT4_BG_INODE_UNINIT set in bg_flags:
Inode table is not initialized/used in this group. So we can skip
the consistency check during fsck.
EXT4_BG_BLOCK_UNINIT set in bg_flags:
No block in the group is used. So we can skip the block bitmap
verification for this group.
We also add two new fields to group descriptor as a part of
uninitialized group patch.
__le16 bg_itable_unused; /* Unused inodes count */
__le16 bg_checksum; /* crc16(sb_uuid+group+desc) */
bg_itable_unused:
If we have EXT4_BG_INODE_UNINIT not set in bg_flags
then bg_itable_unused will give the offset within
the inode table till the inodes are used. This can be
used by fsck to skip list of inodes that are marked unused.
bg_checksum:
Now that we depend on bg_flags and bg_itable_unused to determine
the block and inode usage, we need to make sure group descriptor
is not corrupt. We add checksum to group descriptor to
detect corruption. If the descriptor is found to be corrupt, we
mark all the blocks and inodes in the group used.
Signed-off-by: Avantika Mathur <mathur@us.ibm.com>
Signed-off-by: Andreas Dilger <adilger@clusterfs.com>
Signed-off-by: Mingming Cao <cmm@us.ibm.com>
Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>
2007-10-17 02:38:25 +04:00
__le16 bg_checksum ; /* crc16(sb_uuid+group+desc) */
2006-10-11 12:21:15 +04:00
__le32 bg_block_bitmap_hi ; /* Blocks bitmap block MSB */
__le32 bg_inode_bitmap_hi ; /* Inodes bitmap block MSB */
__le32 bg_inode_table_hi ; /* Inodes table block MSB */
2008-01-29 07:58:27 +03:00
__le16 bg_free_blocks_count_hi ; /* Free blocks count MSB */
__le16 bg_free_inodes_count_hi ; /* Free inodes count MSB */
__le16 bg_used_dirs_count_hi ; /* Directories count MSB */
2009-01-06 06:20:24 +03:00
__le16 bg_itable_unused_hi ; /* Unused inodes count MSB */
2012-04-30 02:23:10 +04:00
__le32 bg_exclude_bitmap_hi ; /* Exclude bitmap block MSB */
__le16 bg_block_bitmap_csum_hi ; /* crc32c(s_uuid+grp_num+bbitmap) BE */
__le16 bg_inode_bitmap_csum_hi ; /* crc32c(s_uuid+grp_num+ibitmap) BE */
__u32 bg_reserved ;
2006-10-11 12:20:50 +04:00
} ;
2012-04-30 02:33:10 +04:00
# define EXT4_BG_INODE_BITMAP_CSUM_HI_END \
( offsetof ( struct ext4_group_desc , bg_inode_bitmap_csum_hi ) + \
sizeof ( __le16 ) )
# define EXT4_BG_BLOCK_BITMAP_CSUM_HI_END \
( offsetof ( struct ext4_group_desc , bg_block_bitmap_csum_hi ) + \
sizeof ( __le16 ) )
2008-07-12 03:27:31 +04:00
/*
* Structure of a flex block group info
*/
struct flex_groups {
2013-03-12 07:39:59 +04:00
atomic64_t free_clusters ;
atomic_t free_inodes ;
atomic_t used_dirs ;
2008-07-12 03:27:31 +04:00
} ;
Ext4: Uninitialized Block Groups
In pass1 of e2fsck, every inode table in the fileystem is scanned and checked,
regardless of whether it is in use. This is this the most time consuming part
of the filesystem check. The unintialized block group feature can greatly
reduce e2fsck time by eliminating checking of uninitialized inodes.
With this feature, there is a a high water mark of used inodes for each block
group. Block and inode bitmaps can be uninitialized on disk via a flag in the
group descriptor to avoid reading or scanning them at e2fsck time. A checksum
of each group descriptor is used to ensure that corruption in the group
descriptor's bit flags does not cause incorrect operation.
The feature is enabled through a mkfs option
mke2fs /dev/ -O uninit_groups
A patch adding support for uninitialized block groups to e2fsprogs tools has
been posted to the linux-ext4 mailing list.
The patches have been stress tested with fsstress and fsx. In performance
tests testing e2fsck time, we have seen that e2fsck time on ext3 grows
linearly with the total number of inodes in the filesytem. In ext4 with the
uninitialized block groups feature, the e2fsck time is constant, based
solely on the number of used inodes rather than the total inode count.
Since typical ext4 filesystems only use 1-10% of their inodes, this feature can
greatly reduce e2fsck time for users. With performance improvement of 2-20
times, depending on how full the filesystem is.
The attached graph shows the major improvements in e2fsck times in filesystems
with a large total inode count, but few inodes in use.
In each group descriptor if we have
EXT4_BG_INODE_UNINIT set in bg_flags:
Inode table is not initialized/used in this group. So we can skip
the consistency check during fsck.
EXT4_BG_BLOCK_UNINIT set in bg_flags:
No block in the group is used. So we can skip the block bitmap
verification for this group.
We also add two new fields to group descriptor as a part of
uninitialized group patch.
__le16 bg_itable_unused; /* Unused inodes count */
__le16 bg_checksum; /* crc16(sb_uuid+group+desc) */
bg_itable_unused:
If we have EXT4_BG_INODE_UNINIT not set in bg_flags
then bg_itable_unused will give the offset within
the inode table till the inodes are used. This can be
used by fsck to skip list of inodes that are marked unused.
bg_checksum:
Now that we depend on bg_flags and bg_itable_unused to determine
the block and inode usage, we need to make sure group descriptor
is not corrupt. We add checksum to group descriptor to
detect corruption. If the descriptor is found to be corrupt, we
mark all the blocks and inodes in the group used.
Signed-off-by: Avantika Mathur <mathur@us.ibm.com>
Signed-off-by: Andreas Dilger <adilger@clusterfs.com>
Signed-off-by: Mingming Cao <cmm@us.ibm.com>
Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>
2007-10-17 02:38:25 +04:00
# define EXT4_BG_INODE_UNINIT 0x0001 /* Inode table/bitmap not in use */
# define EXT4_BG_BLOCK_UNINIT 0x0002 /* Block bitmap not in use */
# define EXT4_BG_INODE_ZEROED 0x0004 /* On-disk itable initialized to zero */
2006-10-11 12:20:50 +04:00
/*
* Macro - instructions used to manage group descriptors
*/
2006-10-11 12:21:14 +04:00
# define EXT4_MIN_DESC_SIZE 32
2006-10-11 12:21:15 +04:00
# define EXT4_MIN_DESC_SIZE_64BIT 64
2006-10-11 12:21:14 +04:00
# define EXT4_MAX_DESC_SIZE EXT4_MIN_BLOCK_SIZE
# define EXT4_DESC_SIZE(s) (EXT4_SB(s)->s_desc_size)
2006-10-11 12:20:50 +04:00
# ifdef __KERNEL__
2006-10-11 12:20:53 +04:00
# define EXT4_BLOCKS_PER_GROUP(s) (EXT4_SB(s)->s_blocks_per_group)
2011-09-10 02:34:51 +04:00
# define EXT4_CLUSTERS_PER_GROUP(s) (EXT4_SB(s)->s_clusters_per_group)
2006-10-11 12:20:53 +04:00
# define EXT4_DESC_PER_BLOCK(s) (EXT4_SB(s)->s_desc_per_block)
# define EXT4_INODES_PER_GROUP(s) (EXT4_SB(s)->s_inodes_per_group)
# define EXT4_DESC_PER_BLOCK_BITS(s) (EXT4_SB(s)->s_desc_per_block_bits)
2006-10-11 12:20:50 +04:00
# else
2006-10-11 12:20:53 +04:00
# define EXT4_BLOCKS_PER_GROUP(s) ((s)->s_blocks_per_group)
2006-10-11 12:21:14 +04:00
# define EXT4_DESC_PER_BLOCK(s) (EXT4_BLOCK_SIZE(s) / EXT4_DESC_SIZE(s))
2006-10-11 12:20:53 +04:00
# define EXT4_INODES_PER_GROUP(s) ((s)->s_inodes_per_group)
2006-10-11 12:20:50 +04:00
# endif
/*
* Constants relative to the data blocks
*/
2006-10-11 12:20:53 +04:00
# define EXT4_NDIR_BLOCKS 12
# define EXT4_IND_BLOCK EXT4_NDIR_BLOCKS
# define EXT4_DIND_BLOCK (EXT4_IND_BLOCK + 1)
# define EXT4_TIND_BLOCK (EXT4_DIND_BLOCK + 1)
# define EXT4_N_BLOCKS (EXT4_TIND_BLOCK + 1)
2006-10-11 12:20:50 +04:00
/*
* Inode flags
*/
2006-10-11 12:20:53 +04:00
# define EXT4_SECRM_FL 0x00000001 /* Secure deletion */
# define EXT4_UNRM_FL 0x00000002 /* Undelete */
# define EXT4_COMPR_FL 0x00000004 /* Compress file */
# define EXT4_SYNC_FL 0x00000008 /* Synchronous updates */
# define EXT4_IMMUTABLE_FL 0x00000010 /* Immutable file */
# define EXT4_APPEND_FL 0x00000020 /* writes to file may only append */
# define EXT4_NODUMP_FL 0x00000040 /* do not dump file */
# define EXT4_NOATIME_FL 0x00000080 /* do not update atime */
2006-10-11 12:20:50 +04:00
/* Reserved for compression usage... */
2006-10-11 12:20:53 +04:00
# define EXT4_DIRTY_FL 0x00000100
# define EXT4_COMPRBLK_FL 0x00000200 /* One or more compressed clusters */
# define EXT4_NOCOMPR_FL 0x00000400 /* Don't compress */
# define EXT4_ECOMPR_FL 0x00000800 /* Compression error */
2006-10-11 12:20:50 +04:00
/* End compression flags --- maybe not all used */
2006-10-11 12:20:53 +04:00
# define EXT4_INDEX_FL 0x00001000 /* hash-indexed directory */
# define EXT4_IMAGIC_FL 0x00002000 /* AFS directory */
# define EXT4_JOURNAL_DATA_FL 0x00004000 /* file data should be journaled */
# define EXT4_NOTAIL_FL 0x00008000 /* file tail should not be merged */
# define EXT4_DIRSYNC_FL 0x00010000 /* dirsync behaviour (directories only) */
# define EXT4_TOPDIR_FL 0x00020000 /* Top of directory hierarchies*/
2008-01-29 07:58:27 +03:00
# define EXT4_HUGE_FILE_FL 0x00040000 /* Set to each huge file */
2006-10-11 12:21:03 +04:00
# define EXT4_EXTENTS_FL 0x00080000 /* Inode uses extents */
2010-01-25 11:31:32 +03:00
# define EXT4_EA_INODE_FL 0x00200000 /* Inode used for large EA */
2010-02-24 17:52:53 +03:00
# define EXT4_EOFBLOCKS_FL 0x00400000 /* Blocks allocated beyond EOF */
2012-12-10 23:04:46 +04:00
# define EXT4_INLINE_DATA_FL 0x10000000 /* Inode has inline data. */
2008-01-29 07:58:27 +03:00
# define EXT4_RESERVED_FL 0x80000000 /* reserved for ext4 lib */
2006-10-11 12:20:50 +04:00
2010-02-24 17:52:53 +03:00
# define EXT4_FL_USER_VISIBLE 0x004BDFFF /* User visible flags */
2013-04-04 06:04:52 +04:00
# define EXT4_FL_USER_MODIFIABLE 0x004380FF /* User modifiable flags */
2006-10-11 12:20:50 +04:00
2009-02-16 02:57:26 +03:00
/* Flags that should be inherited by new inodes from their parent. */
# define EXT4_FL_INHERITED (EXT4_SECRM_FL | EXT4_UNRM_FL | EXT4_COMPR_FL |\
2011-08-31 19:54:51 +04:00
EXT4_SYNC_FL | EXT4_NODUMP_FL | EXT4_NOATIME_FL | \
2009-02-16 02:57:26 +03:00
EXT4_NOCOMPR_FL | EXT4_JOURNAL_DATA_FL | \
EXT4_NOTAIL_FL | EXT4_DIRSYNC_FL )
2009-02-16 02:09:20 +03:00
/* Flags that are appropriate for regular files (all but dir-specific ones). */
# define EXT4_REG_FLMASK (~(EXT4_DIRSYNC_FL | EXT4_TOPDIR_FL))
/* Flags that are appropriate for non-directories/regular files. */
# define EXT4_OTHER_FLMASK (EXT4_NODUMP_FL | EXT4_NOATIME_FL)
/* Mask out flags that are inappropriate for the given type of inode. */
static inline __u32 ext4_mask_flags ( umode_t mode , __u32 flags )
{
if ( S_ISDIR ( mode ) )
return flags ;
else if ( S_ISREG ( mode ) )
return flags & EXT4_REG_FLMASK ;
else
return flags & EXT4_OTHER_FLMASK ;
}
2010-05-17 06:00:00 +04:00
/*
* Inode flags used for atomic set / get
*/
enum {
EXT4_INODE_SECRM = 0 , /* Secure deletion */
EXT4_INODE_UNRM = 1 , /* Undelete */
EXT4_INODE_COMPR = 2 , /* Compress file */
EXT4_INODE_SYNC = 3 , /* Synchronous updates */
EXT4_INODE_IMMUTABLE = 4 , /* Immutable file */
EXT4_INODE_APPEND = 5 , /* writes to file may only append */
EXT4_INODE_NODUMP = 6 , /* do not dump file */
EXT4_INODE_NOATIME = 7 , /* do not update atime */
/* Reserved for compression usage... */
EXT4_INODE_DIRTY = 8 ,
EXT4_INODE_COMPRBLK = 9 , /* One or more compressed clusters */
EXT4_INODE_NOCOMPR = 10 , /* Don't compress */
EXT4_INODE_ECOMPR = 11 , /* Compression error */
/* End compression flags --- maybe not all used */
EXT4_INODE_INDEX = 12 , /* hash-indexed directory */
EXT4_INODE_IMAGIC = 13 , /* AFS directory */
EXT4_INODE_JOURNAL_DATA = 14 , /* file data should be journaled */
EXT4_INODE_NOTAIL = 15 , /* file tail should not be merged */
EXT4_INODE_DIRSYNC = 16 , /* dirsync behaviour (directories only) */
EXT4_INODE_TOPDIR = 17 , /* Top of directory hierarchies*/
EXT4_INODE_HUGE_FILE = 18 , /* Set to each huge file */
EXT4_INODE_EXTENTS = 19 , /* Inode uses extents */
EXT4_INODE_EA_INODE = 21 , /* Inode used for large EA */
EXT4_INODE_EOFBLOCKS = 22 , /* Blocks allocated beyond EOF */
2012-12-10 23:04:46 +04:00
EXT4_INODE_INLINE_DATA = 28 , /* Data in inode. */
2010-05-17 06:00:00 +04:00
EXT4_INODE_RESERVED = 31 , /* reserved for ext4 lib */
} ;
/*
ext4: ensure Inode flags consistency are checked at build time
Flags being used by atomic operations in inode flags (e.g.
ext4_test_inode_flag(), should be consistent with that actually stored
in inodes, i.e.: EXT4_XXX_FL.
It ensures that this consistency is checked at build-time, not at
run-time.
Currently, the flags consistency are being checked at run-time, but,
there is no real reason to not do a build-time check instead of a
run-time check. The code is comparing macro defined values with enum
type variables, where both are constants, so, there is no problem in
comparing constants at build-time.
enum variables are treated as constants by the C compiler, according
to the C99 specs (see www.open-std.org/jtc1/sc22/wg14/www/docs/n1124.pdf
sec. 6.2.5, item 16), so, there is no real problem in comparing an
enumeration type at build time
Signed-off-by: Carlos Maiolino <cmaiolino@redhat.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2012-12-11 01:30:45 +04:00
* Since it ' s pretty easy to mix up bit numbers and hex values , we use a
* build - time check to make sure that EXT4_XXX_FL is consistent with respect to
* EXT4_INODE_XXX . If all is well , the macros will be dropped , so , it won ' t cost
* any extra space in the compiled kernel image , otherwise , the build will fail .
* It ' s important that these values are the same , since we are using
* EXT4_INODE_XXX to test for flag values , but EXT4_XXX_FL must be consistent
* with the values of FS_XXX_FL defined in include / linux / fs . h and the on - disk
* values found in ext2 , ext3 and ext4 filesystems , and of course the values
* defined in e2fsprogs .
2010-05-17 06:00:00 +04:00
*
* It ' s not paranoia if the Murphy ' s Law really * is * out to get you . : - )
*/
ext4: ensure Inode flags consistency are checked at build time
Flags being used by atomic operations in inode flags (e.g.
ext4_test_inode_flag(), should be consistent with that actually stored
in inodes, i.e.: EXT4_XXX_FL.
It ensures that this consistency is checked at build-time, not at
run-time.
Currently, the flags consistency are being checked at run-time, but,
there is no real reason to not do a build-time check instead of a
run-time check. The code is comparing macro defined values with enum
type variables, where both are constants, so, there is no problem in
comparing constants at build-time.
enum variables are treated as constants by the C compiler, according
to the C99 specs (see www.open-std.org/jtc1/sc22/wg14/www/docs/n1124.pdf
sec. 6.2.5, item 16), so, there is no real problem in comparing an
enumeration type at build time
Signed-off-by: Carlos Maiolino <cmaiolino@redhat.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2012-12-11 01:30:45 +04:00
# define TEST_FLAG_VALUE(FLAG) (EXT4_##FLAG##_FL == (1 << EXT4_INODE_##FLAG))
# define CHECK_FLAG_VALUE(FLAG) BUILD_BUG_ON(!TEST_FLAG_VALUE(FLAG))
2010-05-17 06:00:00 +04:00
static inline void ext4_check_flag_values ( void )
{
CHECK_FLAG_VALUE ( SECRM ) ;
CHECK_FLAG_VALUE ( UNRM ) ;
CHECK_FLAG_VALUE ( COMPR ) ;
CHECK_FLAG_VALUE ( SYNC ) ;
CHECK_FLAG_VALUE ( IMMUTABLE ) ;
CHECK_FLAG_VALUE ( APPEND ) ;
CHECK_FLAG_VALUE ( NODUMP ) ;
CHECK_FLAG_VALUE ( NOATIME ) ;
CHECK_FLAG_VALUE ( DIRTY ) ;
CHECK_FLAG_VALUE ( COMPRBLK ) ;
CHECK_FLAG_VALUE ( NOCOMPR ) ;
CHECK_FLAG_VALUE ( ECOMPR ) ;
CHECK_FLAG_VALUE ( INDEX ) ;
CHECK_FLAG_VALUE ( IMAGIC ) ;
CHECK_FLAG_VALUE ( JOURNAL_DATA ) ;
CHECK_FLAG_VALUE ( NOTAIL ) ;
CHECK_FLAG_VALUE ( DIRSYNC ) ;
CHECK_FLAG_VALUE ( TOPDIR ) ;
CHECK_FLAG_VALUE ( HUGE_FILE ) ;
CHECK_FLAG_VALUE ( EXTENTS ) ;
CHECK_FLAG_VALUE ( EA_INODE ) ;
CHECK_FLAG_VALUE ( EOFBLOCKS ) ;
2012-12-10 23:04:46 +04:00
CHECK_FLAG_VALUE ( INLINE_DATA ) ;
2010-05-17 06:00:00 +04:00
CHECK_FLAG_VALUE ( RESERVED ) ;
}
2006-10-11 12:20:50 +04:00
/* Used to pass group descriptor data when online resize is done */
2006-10-11 12:20:53 +04:00
struct ext4_new_group_input {
2007-05-24 21:04:54 +04:00
__u32 group ; /* Group number for this data */
__u64 block_bitmap ; /* Absolute block number of block bitmap */
__u64 inode_bitmap ; /* Absolute block number of inode bitmap */
__u64 inode_table ; /* Absolute block number of inode table start */
__u32 blocks_count ; /* Total number of blocks in this group */
__u16 reserved_blocks ; /* Number of reserved blocks in this group */
2006-10-11 12:20:50 +04:00
__u16 unused ;
} ;
2010-05-17 14:00:00 +04:00
# if defined(__KERNEL__) && defined(CONFIG_COMPAT)
struct compat_ext4_new_group_input {
u32 group ;
compat_u64 block_bitmap ;
compat_u64 inode_bitmap ;
compat_u64 inode_table ;
u32 blocks_count ;
u16 reserved_blocks ;
u16 unused ;
} ;
# endif
2006-10-11 12:20:53 +04:00
/* The struct ext4_new_group_input in kernel space, with free_blocks_count */
struct ext4_new_group_data {
2006-10-11 12:20:50 +04:00
__u32 group ;
2006-10-11 12:21:10 +04:00
__u64 block_bitmap ;
__u64 inode_bitmap ;
__u64 inode_table ;
2006-10-11 12:20:50 +04:00
__u32 blocks_count ;
__u16 reserved_blocks ;
__u16 unused ;
__u32 free_blocks_count ;
} ;
2012-01-04 08:32:52 +04:00
/* Indexes used to index group tables in ext4_new_group_data */
enum {
BLOCK_BITMAP = 0 , /* block bitmap */
INODE_BITMAP , /* inode bitmap */
INODE_TABLE , /* inode tables */
GROUP_TABLE_COUNT ,
} ;
2007-07-18 05:42:41 +04:00
/*
2010-07-27 19:56:07 +04:00
* Flags used by ext4_map_blocks ( )
2007-07-18 05:42:41 +04:00
*/
2014-04-21 07:45:47 +04:00
/* Allocate any needed blocks and/or convert an unwritten
2009-05-14 08:58:52 +04:00
extent to be an initialized ext4 */
2009-05-14 21:57:08 +04:00
# define EXT4_GET_BLOCKS_CREATE 0x0001
2014-04-21 07:45:47 +04:00
/* Request the creation of an unwritten extent */
# define EXT4_GET_BLOCKS_UNWRIT_EXT 0x0002
# define EXT4_GET_BLOCKS_CREATE_UNWRIT_EXT (EXT4_GET_BLOCKS_UNWRIT_EXT|\
2009-05-14 08:58:52 +04:00
EXT4_GET_BLOCKS_CREATE )
2013-04-10 06:11:22 +04:00
/* Caller is from the delayed allocation writeout path
* finally doing the actual allocation of delayed blocks */
2009-06-09 08:17:05 +04:00
# define EXT4_GET_BLOCKS_DELALLOC_RESERVE 0x0004
2009-09-28 23:49:08 +04:00
/* caller is from the direct IO path, request to creation of an
2014-04-21 07:45:47 +04:00
unwritten extents if not allocated , split the unwritten
2009-09-28 23:49:08 +04:00
extent if blocks has been preallocated already */
2010-03-02 21:28:44 +03:00
# define EXT4_GET_BLOCKS_PRE_IO 0x0008
2010-01-15 09:27:59 +03:00
# define EXT4_GET_BLOCKS_CONVERT 0x0010
2010-03-02 21:28:44 +03:00
# define EXT4_GET_BLOCKS_IO_CREATE_EXT (EXT4_GET_BLOCKS_PRE_IO|\
2014-04-21 07:45:47 +04:00
EXT4_GET_BLOCKS_CREATE_UNWRIT_EXT )
2010-03-02 21:28:44 +03:00
/* Convert extent to initialized after IO complete */
# define EXT4_GET_BLOCKS_IO_CONVERT_EXT (EXT4_GET_BLOCKS_CONVERT|\
2014-04-21 07:45:47 +04:00
EXT4_GET_BLOCKS_CREATE_UNWRIT_EXT )
2013-04-10 06:11:22 +04:00
/* Eventual metadata allocation (due to growing extent tree)
* should not fail , so try to use reserved blocks for that . */
# define EXT4_GET_BLOCKS_METADATA_NOFAIL 0x0020
2011-05-25 15:41:54 +04:00
/* Don't normalize allocation size (used for fallocate) */
# define EXT4_GET_BLOCKS_NO_NORMALIZE 0x0040
2011-10-25 16:15:12 +04:00
/* Request will not result in inode size update (user for fallocate) */
# define EXT4_GET_BLOCKS_KEEP_SIZE 0x0080
2012-07-10 00:29:29 +04:00
/* Do not take i_data_sem locking in ext4_map_blocks */
# define EXT4_GET_BLOCKS_NO_LOCK 0x0100
2013-02-18 09:29:59 +04:00
/* Do not put hole in extent cache */
# define EXT4_GET_BLOCKS_NO_PUT_HOLE 0x0200
2014-03-19 02:05:35 +04:00
/* Convert written extents to unwritten */
# define EXT4_GET_BLOCKS_CONVERT_UNWRITTEN 0x0400
2006-10-11 12:20:50 +04:00
2013-08-17 05:23:41 +04:00
/*
2013-08-17 06:05:14 +04:00
* The bit position of these flags must not overlap with any of the
* EXT4_GET_BLOCKS_ * . They are used by ext4_ext_find_extent ( ) ,
2013-08-17 05:23:41 +04:00
* read_extent_tree_block ( ) , ext4_split_extent_at ( ) ,
2013-08-17 06:05:14 +04:00
* ext4_ext_insert_extent ( ) , and ext4_ext_create_new_leaf ( ) .
* EXT4_EX_NOCACHE is used to indicate that the we shouldn ' t be
* caching the extents when reading from the extent tree while a
* truncate or punch hole operation is in progress .
2013-08-17 05:23:41 +04:00
*/
# define EXT4_EX_NOCACHE 0x0400
2013-08-17 06:05:14 +04:00
# define EXT4_EX_FORCE_CACHE 0x0800
2013-08-17 05:23:41 +04:00
2009-11-23 15:17:05 +03:00
/*
* Flags used by ext4_free_blocks
*/
# define EXT4_FREE_BLOCKS_METADATA 0x0001
# define EXT4_FREE_BLOCKS_FORGET 0x0002
ext4: Add block validity check when truncating indirect block mapped inodes
Add checks to ext4_free_branches() to make sure a block number found
in an indirect block are valid before trying to free it. If a bad
block number is found, stop freeing the indirect block immediately,
since the file system is corrupt and we will need to run fsck anyway.
This also avoids spamming the logs, and specifically avoids
driver-level "attempt to access beyond end of device" errors obscure
what is really going on.
If you get *really*, *really*, *really* unlucky, without this patch, a
supposed indirect block containing garbage might contain a reference
to a primary block group descriptor, in which case
ext4_free_branches() could end up zero'ing out a block group
descriptor block, and if then one of the block bitmaps for a block
group described by that bg descriptor block is not in memory, and is
read in by ext4_read_block_bitmap(). This function calls
ext4_valid_block_bitmap(), which assumes that bg_inode_table() was
validated at mount time and hasn't been modified since. Since this
assumption is no longer valid, it's possible for the value
(ext4_inode_table(sb, desc) - group_first_block) to go negative, which
will cause ext4_find_next_zero_bit() to trigger a kernel GPF.
Addresses-Google-Bug: #2220436
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2010-01-23 01:40:42 +03:00
# define EXT4_FREE_BLOCKS_VALIDATED 0x0004
ext4: fix i_blocks/quota accounting when extent insertion fails
The current implementation of ext4_free_blocks() always calls
dquot_free_block This looks quite sensible in the most cases: blocks
to be freed are associated with inode and were accounted in quota and
i_blocks some time ago.
However, there is a case when blocks to free were not accounted by the
time calling ext4_free_blocks() yet:
1. delalloc is on, write_begin pre-allocated some space in quota
2. write-back happens, ext4 allocates some blocks in ext4_ext_map_blocks()
3. then ext4_ext_map_blocks() gets an error (e.g. ENOSPC) from
ext4_ext_insert_extent() and calls ext4_free_blocks().
In this scenario, ext4_free_blocks() calls dquot_free_block() who, in
turn, decrements i_blocks for blocks which were not accounted yet (due
to delalloc) After clean umount, e2fsck reports something like:
> Inode 21, i_blocks is 5080, should be 5128. Fix<y>?
because i_blocks was erroneously decremented as explained above.
The patch fixes the problem by passing the new flag
EXT4_FREE_BLOCKS_NO_QUOT_UPDATE to ext4_free_blocks(), to request
that the dquot_free_block() call be skipped.
Signed-off-by: Maxim Patlasov <maxim.patlasov@gmail.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Cc: stable@kernel.org
2011-07-11 03:37:48 +04:00
# define EXT4_FREE_BLOCKS_NO_QUOT_UPDATE 0x0008
2011-09-10 02:50:51 +04:00
# define EXT4_FREE_BLOCKS_NOFREE_FIRST_CLUSTER 0x0010
# define EXT4_FREE_BLOCKS_NOFREE_LAST_CLUSTER 0x0020
2013-08-17 17:36:54 +04:00
# define EXT4_FREE_BLOCKS_RESERVE 0x0040
2009-11-23 15:17:05 +03:00
2006-10-11 12:20:50 +04:00
/*
* ioctl commands
*/
2006-10-11 12:20:53 +04:00
# define EXT4_IOC_GETFLAGS FS_IOC_GETFLAGS
# define EXT4_IOC_SETFLAGS FS_IOC_SETFLAGS
# define EXT4_IOC_GETVERSION _IOR('f', 3, long)
# define EXT4_IOC_SETVERSION _IOW('f', 4, long)
# define EXT4_IOC_GETVERSION_OLD FS_IOC_GETVERSION
# define EXT4_IOC_SETVERSION_OLD FS_IOC_SETVERSION
# define EXT4_IOC_GETRSVSZ _IOR('f', 5, long)
# define EXT4_IOC_SETRSVSZ _IOW('f', 6, long)
2008-09-14 03:54:35 +04:00
# define EXT4_IOC_GROUP_EXTEND _IOW('f', 7, unsigned long)
# define EXT4_IOC_GROUP_ADD _IOW('f', 8, struct ext4_new_group_input)
# define EXT4_IOC_MIGRATE _IO('f', 9)
2009-02-26 09:04:07 +03:00
/* note ioctl 10 reserved for an early version of the FIEMAP ioctl */
2008-09-14 03:54:35 +04:00
/* note ioctl 11 reserved for filesystem-independent FIEMAP ioctl */
2009-02-26 09:04:07 +03:00
# define EXT4_IOC_ALLOC_DA_BLKS _IO('f', 12)
2009-06-18 03:24:03 +04:00
# define EXT4_IOC_MOVE_EXT _IOWR('f', 15, struct move_extent)
2012-01-05 02:09:44 +04:00
# define EXT4_IOC_RESIZE_FS _IOW('f', 16, __u64)
2013-04-08 20:54:05 +04:00
# define EXT4_IOC_SWAP_BOOT _IO('f', 17)
2013-08-17 06:05:14 +04:00
# define EXT4_IOC_PRECACHE_EXTENTS _IO('f', 18)
2006-10-11 12:20:50 +04:00
2010-05-17 13:00:00 +04:00
# if defined(__KERNEL__) && defined(CONFIG_COMPAT)
2006-10-11 12:20:50 +04:00
/*
* ioctl commands in 32 bit emulation
*/
2006-10-11 12:20:53 +04:00
# define EXT4_IOC32_GETFLAGS FS_IOC32_GETFLAGS
# define EXT4_IOC32_SETFLAGS FS_IOC32_SETFLAGS
# define EXT4_IOC32_GETVERSION _IOR('f', 3, int)
# define EXT4_IOC32_SETVERSION _IOW('f', 4, int)
# define EXT4_IOC32_GETRSVSZ _IOR('f', 5, int)
# define EXT4_IOC32_SETRSVSZ _IOW('f', 6, int)
# define EXT4_IOC32_GROUP_EXTEND _IOW('f', 7, unsigned int)
2010-05-17 14:00:00 +04:00
# define EXT4_IOC32_GROUP_ADD _IOW('f', 8, struct compat_ext4_new_group_input)
2006-10-11 12:20:53 +04:00
# define EXT4_IOC32_GETVERSION_OLD FS_IOC32_GETVERSION
# define EXT4_IOC32_SETVERSION_OLD FS_IOC32_SETVERSION
2010-05-17 13:00:00 +04:00
# endif
2006-10-11 12:20:50 +04:00
tree-wide: fix comment/printk typos
"gadget", "through", "command", "maintain", "maintain", "controller", "address",
"between", "initiali[zs]e", "instead", "function", "select", "already",
"equal", "access", "management", "hierarchy", "registration", "interest",
"relative", "memory", "offset", "already",
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
2010-11-01 22:38:34 +03:00
/* Max physical block we can address w/o extents */
ext4: limit block allocations for indirect-block files to < 2^32
Today, the ext4 allocator will happily allocate blocks past
2^32 for indirect-block files, which results in the block
numbers getting truncated, and corruption ensues.
This patch limits such allocations to < 2^32, and adds
BUG_ONs if we do get blocks larger than that.
This should address RH Bug 519471, ext4 bitmap allocator
must limit blocks to < 2^32
* ext4_find_goal() is modified to choose a goal < UINT_MAX,
so that our starting point is in an acceptable range.
* ext4_xattr_block_set() is modified such that the goal block
is < UINT_MAX, as above.
* ext4_mb_regular_allocator() is modified so that the group
search does not continue into groups which are too high
* ext4_mb_use_preallocated() has a check that we don't use
preallocated space which is too far out
* ext4_alloc_blocks() and ext4_xattr_block_set() add some BUG_ONs
No attempt has been made to limit inode locations to < 2^32,
so we may wind up with blocks far from their inodes. Doing
this much already will lead to some odd ENOSPC issues when the
"lower 32" gets full, and further restricting inodes could
make that even weirder.
For high inodes, choosing a goal of the original, % UINT_MAX,
may be a bit odd, but then we're in an odd situation anyway,
and I don't know of a better heuristic.
Signed-off-by: Eric Sandeen <sandeen@redhat.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2009-09-16 22:45:10 +04:00
# define EXT4_MAX_BLOCK_FILE_PHYS 0xFFFFFFFF
2006-10-11 12:20:50 +04:00
/*
* Structure of an inode on the disk
*/
2006-10-11 12:20:53 +04:00
struct ext4_inode {
2006-10-11 12:20:50 +04:00
__le16 i_mode ; /* File mode */
__le16 i_uid ; /* Low 16 bits of Owner Uid */
2008-01-29 07:58:27 +03:00
__le32 i_size_lo ; /* Size in bytes */
2006-10-11 12:20:50 +04:00
__le32 i_atime ; /* Access time */
2007-07-18 17:15:20 +04:00
__le32 i_ctime ; /* Inode Change time */
2006-10-11 12:20:50 +04:00
__le32 i_mtime ; /* Modification time */
__le32 i_dtime ; /* Deletion Time */
__le16 i_gid ; /* Low 16 bits of Group Id */
__le16 i_links_count ; /* Links count */
2008-01-29 07:58:26 +03:00
__le32 i_blocks_lo ; /* Blocks count */
2006-10-11 12:20:50 +04:00
__le32 i_flags ; /* File flags */
union {
struct {
2008-01-29 07:58:27 +03:00
__le32 l_i_version ;
2006-10-11 12:20:50 +04:00
} linux1 ;
struct {
__u32 h_i_translator ;
} hurd1 ;
struct {
__u32 m_i_reserved1 ;
} masix1 ;
} osd1 ; /* OS dependent 1 */
2006-10-11 12:20:53 +04:00
__le32 i_block [ EXT4_N_BLOCKS ] ; /* Pointers to blocks */
2006-10-11 12:20:50 +04:00
__le32 i_generation ; /* File version (for NFS) */
2008-01-29 07:58:27 +03:00
__le32 i_file_acl_lo ; /* File ACL */
2008-01-29 07:58:27 +03:00
__le32 i_size_high ;
2007-10-17 02:38:25 +04:00
__le32 i_obso_faddr ; /* Obsoleted fragment address */
2006-10-11 12:20:50 +04:00
union {
struct {
2008-01-29 07:58:26 +03:00
__le16 l_i_blocks_high ; /* were l_i_reserved1 */
2006-10-11 12:21:09 +04:00
__le16 l_i_file_acl_high ;
2007-05-24 21:04:54 +04:00
__le16 l_i_uid_high ; /* these 2 fields */
2006-10-11 12:20:50 +04:00
__le16 l_i_gid_high ; /* were reserved2[0] */
2012-04-30 02:23:10 +04:00
__le16 l_i_checksum_lo ; /* crc32c(uuid+inum+inode) LE */
__le16 l_i_reserved ;
2006-10-11 12:20:50 +04:00
} linux2 ;
struct {
2007-10-17 02:38:25 +04:00
__le16 h_i_reserved1 ; /* Obsoleted fragment number/size which are removed in ext4 */
2006-10-11 12:20:50 +04:00
__u16 h_i_mode_high ;
__u16 h_i_uid_high ;
__u16 h_i_gid_high ;
__u32 h_i_author ;
} hurd2 ;
struct {
2007-10-17 02:38:25 +04:00
__le16 h_i_reserved1 ; /* Obsoleted fragment number/size which are removed in ext4 */
2006-10-11 12:21:09 +04:00
__le16 m_i_file_acl_high ;
2006-10-11 12:20:50 +04:00
__u32 m_i_reserved2 [ 2 ] ;
} masix2 ;
} osd2 ; /* OS dependent 2 */
__le16 i_extra_isize ;
2012-04-30 02:23:10 +04:00
__le16 i_checksum_hi ; /* crc32c(uuid+inum+inode) BE */
2007-07-18 17:15:20 +04:00
__le32 i_ctime_extra ; /* extra Change time (nsec << 2 | epoch) */
__le32 i_mtime_extra ; /* extra Modification time(nsec << 2 | epoch) */
__le32 i_atime_extra ; /* extra Access time (nsec << 2 | epoch) */
__le32 i_crtime ; /* File Creation time */
__le32 i_crtime_extra ; /* extra FileCreationtime (nsec << 2 | epoch) */
2008-01-29 07:58:27 +03:00
__le32 i_version_hi ; /* high 32 bits for 64-bit version */
2006-10-11 12:20:50 +04:00
} ;
2009-06-18 03:24:03 +04:00
struct move_extent {
__u32 reserved ; /* should be zero */
__u32 donor_fd ; /* donor file descriptor */
__u64 orig_start ; /* logical start offset in block for orig */
__u64 donor_start ; /* logical start offset in block for donor */
__u64 len ; /* block length to be moved */
__u64 moved_len ; /* moved block length */
} ;
2006-10-11 12:20:50 +04:00
2007-07-18 17:15:20 +04:00
# define EXT4_EPOCH_BITS 2
# define EXT4_EPOCH_MASK ((1 << EXT4_EPOCH_BITS) - 1)
# define EXT4_NSEC_MASK (~0UL << EXT4_EPOCH_BITS)
/*
* Extended fields will fit into an inode if the filesystem was formatted
* with large inodes ( - I 256 or larger ) and there are not currently any EAs
* consuming all of the available space . For new inodes we always reserve
* enough space for the kernel ' s known extended fields , but for inodes
* created with an old kernel this might not have been the case . None of
* the extended inode fields is critical for correct filesystem operation .
* This macro checks if a certain field fits in the inode . Note that
* inode - size = GOOD_OLD_INODE_SIZE + i_extra_isize
*/
# define EXT4_FITS_IN_INODE(ext4_inode, einode, field) \
( ( offsetof ( typeof ( * ext4_inode ) , field ) + \
sizeof ( ( ext4_inode ) - > field ) ) \
< = ( EXT4_GOOD_OLD_INODE_SIZE + \
( einode ) - > i_extra_isize ) ) \
static inline __le32 ext4_encode_extra_time ( struct timespec * time )
{
return cpu_to_le32 ( ( sizeof ( time - > tv_sec ) > 4 ?
2009-09-30 09:13:55 +04:00
( time - > tv_sec > > 32 ) & EXT4_EPOCH_MASK : 0 ) |
( ( time - > tv_nsec < < EXT4_EPOCH_BITS ) & EXT4_NSEC_MASK ) ) ;
2007-07-18 17:15:20 +04:00
}
static inline void ext4_decode_extra_time ( struct timespec * time , __le32 extra )
{
if ( sizeof ( time - > tv_sec ) > 4 )
time - > tv_sec | = ( __u64 ) ( le32_to_cpu ( extra ) & EXT4_EPOCH_MASK )
< < 32 ;
2009-09-30 09:13:55 +04:00
time - > tv_nsec = ( le32_to_cpu ( extra ) & EXT4_NSEC_MASK ) > > EXT4_EPOCH_BITS ;
2007-07-18 17:15:20 +04:00
}
# define EXT4_INODE_SET_XTIME(xtime, inode, raw_inode) \
do { \
( raw_inode ) - > xtime = cpu_to_le32 ( ( inode ) - > xtime . tv_sec ) ; \
if ( EXT4_FITS_IN_INODE ( raw_inode , EXT4_I ( inode ) , xtime # # _extra ) ) \
( raw_inode ) - > xtime # # _extra = \
ext4_encode_extra_time ( & ( inode ) - > xtime ) ; \
} while ( 0 )
# define EXT4_EINODE_SET_XTIME(xtime, einode, raw_inode) \
do { \
if ( EXT4_FITS_IN_INODE ( raw_inode , einode , xtime ) ) \
( raw_inode ) - > xtime = cpu_to_le32 ( ( einode ) - > xtime . tv_sec ) ; \
if ( EXT4_FITS_IN_INODE ( raw_inode , einode , xtime # # _extra ) ) \
( raw_inode ) - > xtime # # _extra = \
ext4_encode_extra_time ( & ( einode ) - > xtime ) ; \
} while ( 0 )
# define EXT4_INODE_GET_XTIME(xtime, inode, raw_inode) \
do { \
( inode ) - > xtime . tv_sec = ( signed ) le32_to_cpu ( ( raw_inode ) - > xtime ) ; \
if ( EXT4_FITS_IN_INODE ( raw_inode , EXT4_I ( inode ) , xtime # # _extra ) ) \
ext4_decode_extra_time ( & ( inode ) - > xtime , \
raw_inode - > xtime # # _extra ) ; \
2010-12-20 06:10:31 +03:00
else \
( inode ) - > xtime . tv_nsec = 0 ; \
2007-07-18 17:15:20 +04:00
} while ( 0 )
# define EXT4_EINODE_GET_XTIME(xtime, einode, raw_inode) \
do { \
if ( EXT4_FITS_IN_INODE ( raw_inode , einode , xtime ) ) \
( einode ) - > xtime . tv_sec = \
( signed ) le32_to_cpu ( ( raw_inode ) - > xtime ) ; \
2014-02-17 04:29:32 +04:00
else \
( einode ) - > xtime . tv_sec = 0 ; \
2007-07-18 17:15:20 +04:00
if ( EXT4_FITS_IN_INODE ( raw_inode , einode , xtime # # _extra ) ) \
ext4_decode_extra_time ( & ( einode ) - > xtime , \
raw_inode - > xtime # # _extra ) ; \
2010-12-20 06:10:31 +03:00
else \
( einode ) - > xtime . tv_nsec = 0 ; \
2007-07-18 17:15:20 +04:00
} while ( 0 )
2008-01-29 07:58:27 +03:00
# define i_disk_version osd1.linux1.l_i_version
2006-10-11 12:20:50 +04:00
# if defined(__KERNEL__) || defined(__linux__)
# define i_reserved1 osd1.linux1.l_i_reserved1
2006-10-11 12:21:09 +04:00
# define i_file_acl_high osd2.linux2.l_i_file_acl_high
2008-01-29 07:58:26 +03:00
# define i_blocks_high osd2.linux2.l_i_blocks_high
2006-10-11 12:20:50 +04:00
# define i_uid_low i_uid
# define i_gid_low i_gid
# define i_uid_high osd2.linux2.l_i_uid_high
# define i_gid_high osd2.linux2.l_i_gid_high
2012-04-30 02:23:10 +04:00
# define i_checksum_lo osd2.linux2.l_i_checksum_lo
2006-10-11 12:20:50 +04:00
# elif defined(__GNU__)
# define i_translator osd1.hurd1.h_i_translator
# define i_uid_high osd2.hurd2.h_i_uid_high
# define i_gid_high osd2.hurd2.h_i_gid_high
# define i_author osd2.hurd2.h_i_author
# elif defined(__masix__)
# define i_reserved1 osd1.masix1.m_i_reserved1
2006-10-11 12:21:09 +04:00
# define i_file_acl_high osd2.masix2.m_i_file_acl_high
2006-10-11 12:20:50 +04:00
# define i_reserved2 osd2.masix2.m_i_reserved2
# endif /* defined(__KERNEL__) || defined(__linux__) */
2012-11-09 00:18:54 +04:00
# include "extents_status.h"
2009-05-01 21:44:33 +04:00
/*
* fourth extended file system inode data in memory
*/
struct ext4_inode_info {
__le32 i_data [ 15 ] ; /* unconverted */
__u32 i_dtime ;
2010-05-17 06:00:00 +04:00
ext4_fsblk_t i_file_acl ;
2009-05-01 21:44:33 +04:00
/*
* i_block_group is the number of the block group which contains
* this file ' s inode . Constant across the lifetime of the inode ,
* it is ued for making block allocation decisions - we try to
* place a file ' s data blocks near its inode block , and new inodes
* near to their parent directory ' s inode .
*/
ext4_group_t i_block_group ;
2011-01-10 20:13:42 +03:00
ext4_lblk_t i_dir_start_lookup ;
2011-01-10 20:18:25 +03:00
# if (BITS_PER_LONG < 64)
2010-01-24 22:34:07 +03:00
unsigned long i_state_flags ; /* Dynamic state flags */
2011-01-10 20:18:25 +03:00
# endif
2010-05-17 06:00:00 +04:00
unsigned long i_flags ;
2009-05-01 21:44:33 +04:00
/*
* Extended attributes can be read independently of the main file
* data . Taking i_mutex even when reading would cause contention
* between readers of EAs and writers of regular file data , so
* instead we synchronize on xattr_sem when reading or changing
* EAs .
*/
struct rw_semaphore xattr_sem ;
struct list_head i_orphan ; /* unlinked but open inodes */
/*
* i_disksize keeps track of what the inode size is ON DISK , not
* in memory . During truncate , i_size is set to the new size by
* the VFS prior to calling ext4_truncate ( ) , but the filesystem won ' t
* set i_disksize to 0 until the truncate is actually under way .
*
* The intent is that i_disksize always represents the blocks which
* are used by this file . This allows recovery to restart truncate
* on orphans if we crash during truncate . We actually write i_disksize
* into the on - disk inode when writing inodes out , instead of i_size .
*
* The only time when i_disksize and i_size may be different is when
* a truncate is in progress . The only things which change i_disksize
* are ext4_get_block ( growth ) and ext4_truncate ( shrinkth ) .
*/
loff_t i_disksize ;
/*
* i_data_sem is for serialising ext4_truncate ( ) against
* ext4_getblock ( ) . In the 2.4 ext2 design , great chunks of inode ' s
* data tree are chopped off during truncate . We can ' t do that in
* ext4 because whenever we perform intermediate commits during
* truncate , the inode and all the metadata blocks * must * be in a
* consistent state which allows truncation of the orphans to restart
* during recovery . Hence we must fix the get_block - vs - truncate race
* by other means , so we have i_data_sem .
*/
struct rw_semaphore i_data_sem ;
struct inode vfs_inode ;
2011-01-10 20:29:43 +03:00
struct jbd2_inode * jinode ;
2009-05-01 21:44:33 +04:00
2014-04-21 22:37:55 +04:00
spinlock_t i_raw_lock ; /* protects updates to the raw inode */
2009-05-01 21:44:33 +04:00
/*
* File creation time . Its function is same as that of
* struct timespec i_ { a , c , m } time in the generic inode .
*/
struct timespec i_crtime ;
/* mballoc */
struct list_head i_prealloc_list ;
spinlock_t i_prealloc_lock ;
2012-11-09 00:18:54 +04:00
/* extents status tree */
struct ext4_es_tree i_es_tree ;
rwlock_t i_es_lock ;
2013-02-18 09:32:55 +04:00
struct list_head i_es_lru ;
unsigned int i_es_lru_nr ; /* protected by i_es_lock */
2013-07-01 16:12:37 +04:00
unsigned long i_touch_when ; /* jiffies of last accessing */
2012-11-09 00:18:54 +04:00
2009-05-01 21:44:33 +04:00
/* ialloc */
ext4_group_t i_last_alloc_group ;
/* allocation reservation info for delalloc */
2011-09-10 02:56:51 +04:00
/* In case of bigalloc, these refer to clusters rather than blocks */
2009-05-01 21:44:33 +04:00
unsigned int i_reserved_data_blocks ;
unsigned int i_reserved_meta_blocks ;
unsigned int i_allocated_meta_blocks ;
2011-01-10 20:13:03 +03:00
ext4_lblk_t i_da_metadata_calc_last_lblock ;
2010-01-01 10:41:30 +03:00
int i_da_metadata_calc_len ;
2009-05-01 21:44:33 +04:00
/* on-disk additional length */
__u16 i_extra_isize ;
2012-12-10 23:04:46 +04:00
/* Indicate the inline data space. */
u16 i_inline_off ;
u16 i_inline_size ;
2009-12-14 15:21:14 +03:00
# ifdef CONFIG_QUOTA
/* quota space reservation, managed internally by quota code */
qsize_t i_reserved_quota ;
# endif
2009-09-28 23:48:29 +04:00
2013-06-04 22:21:02 +04:00
/* Lock protecting lists below */
2010-03-05 00:14:02 +03:00
spinlock_t i_completed_io_lock ;
2013-06-04 22:21:02 +04:00
/*
* Completed IOs that need unwritten extents handling and have
* transaction reserved
*/
struct list_head i_rsv_conversion_list ;
/*
* Completed IOs that need unwritten extents handling and don ' t have
* transaction reserved
*/
2011-01-10 20:13:42 +03:00
atomic_t i_ioend_count ; /* Number of outstanding io_end structs */
2012-09-29 07:24:52 +04:00
atomic_t i_unwritten ; /* Nr. of inflight conversions pending */
2013-06-04 22:21:02 +04:00
struct work_struct i_rsv_conversion_work ;
2011-01-10 20:13:42 +03:00
spinlock_t i_block_reservation_lock ;
2009-12-09 07:51:10 +03:00
/*
* Transactions that contain inode ' s metadata needed to complete
* fsync and fdatasync , respectively .
*/
tid_t i_sync_tid ;
tid_t i_datasync_tid ;
2012-04-30 02:31:10 +04:00
/* Precomputed uuid+inum+igen checksum for seeding inode checksums */
__u32 i_csum_seed ;
2009-05-01 21:44:33 +04:00
} ;
2006-10-11 12:20:50 +04:00
/*
* File system states
*/
2006-10-11 12:20:53 +04:00
# define EXT4_VALID_FS 0x0001 /* Unmounted cleanly */
# define EXT4_ERROR_FS 0x0002 /* Errors detected */
# define EXT4_ORPHAN_FS 0x0004 /* Orphans being recovered */
2006-10-11 12:20:50 +04:00
2008-02-10 09:11:44 +03:00
/*
* Misc . filesystem flags
*/
# define EXT2_FLAGS_SIGNED_HASH 0x0001 /* Signed dirhash in use */
# define EXT2_FLAGS_UNSIGNED_HASH 0x0002 /* Unsigned dirhash in use */
# define EXT2_FLAGS_TEST_FILESYS 0x0004 /* to test development code */
2006-10-11 12:20:50 +04:00
/*
2013-04-04 06:12:52 +04:00
* Mount flags set via mount options or defaults
2006-10-11 12:20:50 +04:00
*/
2006-10-11 12:20:53 +04:00
# define EXT4_MOUNT_GRPID 0x00004 /* Create files with directory's group */
# define EXT4_MOUNT_DEBUG 0x00008 /* Some debugging messages */
# define EXT4_MOUNT_ERRORS_CONT 0x00010 /* Continue on errors */
# define EXT4_MOUNT_ERRORS_RO 0x00020 /* Remount fs ro on errors */
# define EXT4_MOUNT_ERRORS_PANIC 0x00040 /* Panic on errors */
2012-03-04 02:56:23 +04:00
# define EXT4_MOUNT_ERRORS_MASK 0x00070
2006-10-11 12:20:53 +04:00
# define EXT4_MOUNT_MINIX_DF 0x00080 /* Mimics the Minix statfs */
# define EXT4_MOUNT_NOLOAD 0x00100 /* Don't use existing journal*/
# define EXT4_MOUNT_DATA_FLAGS 0x00C00 /* Mode for data writes: */
# define EXT4_MOUNT_JOURNAL_DATA 0x00400 /* Write data to journal */
# define EXT4_MOUNT_ORDERED_DATA 0x00800 /* Flush data before commit */
# define EXT4_MOUNT_WRITEBACK_DATA 0x00C00 /* No data ordering */
# define EXT4_MOUNT_UPDATE_JOURNAL 0x01000 /* Update the journal format */
# define EXT4_MOUNT_NO_UID32 0x02000 /* Disable 32-bit UIDs */
# define EXT4_MOUNT_XATTR_USER 0x04000 /* Extended user attributes */
# define EXT4_MOUNT_POSIX_ACL 0x08000 /* POSIX Access Control Lists */
2009-03-17 06:12:23 +03:00
# define EXT4_MOUNT_NO_AUTO_DA_ALLOC 0x10000 /* No auto delalloc mapping */
2006-10-11 12:20:53 +04:00
# define EXT4_MOUNT_BARRIER 0x20000 /* Use block barriers */
# define EXT4_MOUNT_QUOTA 0x80000 /* Some quota option set */
# define EXT4_MOUNT_USRQUOTA 0x100000 /* "old" user quota */
# define EXT4_MOUNT_GRPQUOTA 0x200000 /* "old" group quota */
2010-03-05 00:14:02 +03:00
# define EXT4_MOUNT_DIOREAD_NOLOCK 0x400000 /* Enable support for dio read nolocking */
2009-11-02 21:15:27 +03:00
# define EXT4_MOUNT_JOURNAL_CHECKSUM 0x800000 /* Journal checksums */
2008-01-29 07:58:27 +03:00
# define EXT4_MOUNT_JOURNAL_ASYNC_COMMIT 0x1000000 /* Journal Async Commit */
2008-07-12 03:27:31 +04:00
# define EXT4_MOUNT_DELALLOC 0x8000000 /* Delalloc support */
2008-10-11 06:12:43 +04:00
# define EXT4_MOUNT_DATA_ERR_ABORT 0x10000000 /* Abort on file data write */
2009-05-17 23:38:01 +04:00
# define EXT4_MOUNT_BLOCK_VALIDITY 0x20000000 /* Block validity checking */
2009-11-19 22:25:42 +03:00
# define EXT4_MOUNT_DISCARD 0x40000000 /* Issue DISCARD requests */
ext4: add support for lazy inode table initialization
When the lazy_itable_init extended option is passed to mke2fs, it
considerably speeds up filesystem creation because inode tables are
not zeroed out. The fact that parts of the inode table are
uninitialized is not a problem so long as the block group descriptors,
which contain information regarding how much of the inode table has
been initialized, has not been corrupted However, if the block group
checksums are not valid, e2fsck must scan the entire inode table, and
the the old, uninitialized data could potentially cause e2fsck to
report false problems.
Hence, it is important for the inode tables to be initialized as soon
as possble. This commit adds this feature so that mke2fs can safely
use the lazy inode table initialization feature to speed up formatting
file systems.
This is done via a new new kernel thread called ext4lazyinit, which is
created on demand and destroyed, when it is no longer needed. There
is only one thread for all ext4 filesystems in the system. When the
first filesystem with inititable mount option is mounted, ext4lazyinit
thread is created, then the filesystem can register its request in the
request list.
This thread then walks through the list of requests picking up
scheduled requests and invoking ext4_init_inode_table(). Next schedule
time for the request is computed by multiplying the time it took to
zero out last inode table with wait multiplier, which can be set with
the (init_itable=n) mount option (default is 10). We are doing
this so we do not take the whole I/O bandwidth. When the thread is no
longer necessary (request list is empty) it frees the appropriate
structures and exits (and can be created later later by another
filesystem).
We do not disturb regular inode allocations in any way, it just do not
care whether the inode table is, or is not zeroed. But when zeroing, we
have to skip used inodes, obviously. Also we should prevent new inode
allocations from the group, while zeroing is on the way. For that we
take write alloc_sem lock in ext4_init_inode_table() and read alloc_sem
in the ext4_claim_inode, so when we are unlucky and allocator hits the
group which is currently being zeroed, it just has to wait.
This can be suppresed using the mount option no_init_itable.
Signed-off-by: Lukas Czerner <lczerner@redhat.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2010-10-28 05:30:05 +04:00
# define EXT4_MOUNT_INIT_INODE_TABLE 0x80000000 /* Initialize uninitialized itables */
2008-10-11 06:12:43 +04:00
2013-04-04 06:12:52 +04:00
/*
* Mount flags set either automatically ( could not be set by mount option )
* based on per file system feature or property or in special cases such as
* distinguishing between explicit mount option definition and default .
*/
2011-09-04 02:22:38 +04:00
# define EXT4_MOUNT2_EXPLICIT_DELALLOC 0x00000001 / * User explicitly
specified delalloc */
2013-04-04 06:12:52 +04:00
# define EXT4_MOUNT2_STD_GROUP_SIZE 0x00000002 / * We have standard group
size of blocksize * 8
blocks */
2014-03-24 22:09:06 +04:00
# define EXT4_MOUNT2_HURD_COMPAT 0x00000004 / * Support HURD-castrated
file systems */
2011-09-04 02:22:38 +04:00
2010-12-16 04:26:48 +03:00
# define clear_opt(sb, opt) EXT4_SB(sb)->s_mount_opt &= \
~ EXT4_MOUNT_ # # opt
# define set_opt(sb, opt) EXT4_SB(sb)->s_mount_opt |= \
EXT4_MOUNT_ # # opt
2006-10-11 12:20:53 +04:00
# define test_opt(sb, opt) (EXT4_SB(sb)->s_mount_opt & \
EXT4_MOUNT_ # # opt )
2006-10-11 12:20:50 +04:00
2010-12-16 04:30:48 +03:00
# define clear_opt2(sb, opt) EXT4_SB(sb)->s_mount_opt2 &= \
~ EXT4_MOUNT2_ # # opt
# define set_opt2(sb, opt) EXT4_SB(sb)->s_mount_opt2 |= \
EXT4_MOUNT2_ # # opt
# define test_opt2(sb, opt) (EXT4_SB(sb)->s_mount_opt2 & \
EXT4_MOUNT2_ # # opt )
ext4: use proper little-endian bitops
ext4_{set,clear}_bit() is defined as __test_and_{set,clear}_bit_le() for
ext4. Only two ext4_{set,clear}_bit() calls check the return value. The
rest of calls ignore the return value and they can be replaced with
__{set,clear}_bit_le().
This changes ext4_{set,clear}_bit() from __test_and_{set,clear}_bit_le()
to __{set,clear}_bit_le() and introduces ext4_test_and_{set,clear}_bit()
for the two places where old bit needs to be returned.
This ext4_{set,clear}_bit() change is considered safe, because if someone
uses these macros without noticing the change, new ext4_{set,clear}_bit
don't have return value and causes compiler errors where the return value
is used.
This also removes unused ext4_find_first_zero_bit().
Signed-off-by: Akinobu Mita <akinobu.mita@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2011-12-29 05:32:07 +04:00
# define ext4_test_and_set_bit __test_and_set_bit_le
# define ext4_set_bit __set_bit_le
2006-10-11 12:20:53 +04:00
# define ext4_set_bit_atomic ext2_set_bit_atomic
ext4: use proper little-endian bitops
ext4_{set,clear}_bit() is defined as __test_and_{set,clear}_bit_le() for
ext4. Only two ext4_{set,clear}_bit() calls check the return value. The
rest of calls ignore the return value and they can be replaced with
__{set,clear}_bit_le().
This changes ext4_{set,clear}_bit() from __test_and_{set,clear}_bit_le()
to __{set,clear}_bit_le() and introduces ext4_test_and_{set,clear}_bit()
for the two places where old bit needs to be returned.
This ext4_{set,clear}_bit() change is considered safe, because if someone
uses these macros without noticing the change, new ext4_{set,clear}_bit
don't have return value and causes compiler errors where the return value
is used.
This also removes unused ext4_find_first_zero_bit().
Signed-off-by: Akinobu Mita <akinobu.mita@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2011-12-29 05:32:07 +04:00
# define ext4_test_and_clear_bit __test_and_clear_bit_le
# define ext4_clear_bit __clear_bit_le
2006-10-11 12:20:53 +04:00
# define ext4_clear_bit_atomic ext2_clear_bit_atomic
2011-03-24 02:42:07 +03:00
# define ext4_test_bit test_bit_le
# define ext4_find_next_zero_bit find_next_zero_bit_le
# define ext4_find_next_bit find_next_bit_le
2006-10-11 12:20:50 +04:00
2011-07-27 06:05:53 +04:00
extern void ext4_set_bits ( void * bm , int cur , int len ) ;
2006-10-11 12:20:50 +04:00
/*
* Maximal mount counts between two filesystem checks
*/
2006-10-11 12:20:53 +04:00
# define EXT4_DFL_MAX_MNT_COUNT 20 /* Allow 20 mounts */
# define EXT4_DFL_CHECKINTERVAL 0 /* Don't use interval check */
2006-10-11 12:20:50 +04:00
/*
* Behaviour when detecting errors
*/
2006-10-11 12:20:53 +04:00
# define EXT4_ERRORS_CONTINUE 1 /* Continue execution */
# define EXT4_ERRORS_RO 2 /* Remount fs read-only */
# define EXT4_ERRORS_PANIC 3 /* Panic */
# define EXT4_ERRORS_DEFAULT EXT4_ERRORS_CONTINUE
2006-10-11 12:20:50 +04:00
2012-04-30 02:23:10 +04:00
/* Metadata checksum algorithm codes */
# define EXT4_CRC32C_CHKSUM 1
2006-10-11 12:20:50 +04:00
/*
* Structure of the super block
*/
2006-10-11 12:20:53 +04:00
struct ext4_super_block {
2006-10-11 12:20:50 +04:00
/*00*/ __le32 s_inodes_count ; /* Inodes count */
2007-10-17 02:38:25 +04:00
__le32 s_blocks_count_lo ; /* Blocks count */
2007-10-17 02:38:25 +04:00
__le32 s_r_blocks_count_lo ; /* Reserved blocks count */
__le32 s_free_blocks_count_lo ; /* Free blocks count */
2006-10-11 12:20:50 +04:00
/*10*/ __le32 s_free_inodes_count ; /* Free inodes count */
__le32 s_first_data_block ; /* First Data Block */
__le32 s_log_block_size ; /* Block size */
2011-09-10 02:34:51 +04:00
__le32 s_log_cluster_size ; /* Allocation cluster size */
2006-10-11 12:20:50 +04:00
/*20*/ __le32 s_blocks_per_group ; /* # Blocks per group */
2011-09-10 02:34:51 +04:00
__le32 s_clusters_per_group ; /* # Clusters per group */
2006-10-11 12:20:50 +04:00
__le32 s_inodes_per_group ; /* # Inodes per group */
__le32 s_mtime ; /* Mount time */
/*30*/ __le32 s_wtime ; /* Write time */
__le16 s_mnt_count ; /* Mount count */
__le16 s_max_mnt_count ; /* Maximal mount count */
__le16 s_magic ; /* Magic signature */
__le16 s_state ; /* File system state */
__le16 s_errors ; /* Behaviour when detecting errors */
__le16 s_minor_rev_level ; /* minor revision level */
/*40*/ __le32 s_lastcheck ; /* time of last check */
__le32 s_checkinterval ; /* max. time between checks */
__le32 s_creator_os ; /* OS */
__le32 s_rev_level ; /* Revision level */
/*50*/ __le16 s_def_resuid ; /* Default uid for reserved blocks */
__le16 s_def_resgid ; /* Default gid for reserved blocks */
/*
2006-10-11 12:20:53 +04:00
* These fields are for EXT4_DYNAMIC_REV superblocks only .
2006-10-11 12:20:50 +04:00
*
* Note : the difference between the compatible feature set and
* the incompatible feature set is that if there is a bit set
* in the incompatible feature set that the kernel doesn ' t
* know about , it should refuse to mount the filesystem .
*
* e2fsck ' s requirements are more strict ; if it doesn ' t know
* about a feature in either the compatible or incompatible
* feature set , it must abort and not try to meddle with
* things it doesn ' t understand . . .
*/
__le32 s_first_ino ; /* First non-reserved inode */
2006-10-11 12:21:14 +04:00
__le16 s_inode_size ; /* size of inode structure */
2006-10-11 12:20:50 +04:00
__le16 s_block_group_nr ; /* block group # of this superblock */
__le32 s_feature_compat ; /* compatible feature set */
/*60*/ __le32 s_feature_incompat ; /* incompatible feature set */
__le32 s_feature_ro_compat ; /* readonly-compatible feature set */
/*68*/ __u8 s_uuid [ 16 ] ; /* 128-bit uuid for volume */
/*78*/ char s_volume_name [ 16 ] ; /* volume name */
/*88*/ char s_last_mounted [ 64 ] ; /* directory where last mounted */
/*C8*/ __le32 s_algorithm_usage_bitmap ; /* For compression */
/*
* Performance hints . Directory preallocation should only
2006-10-11 12:20:53 +04:00
* happen if the EXT4_FEATURE_COMPAT_DIR_PREALLOC flag is on .
2006-10-11 12:20:50 +04:00
*/
__u8 s_prealloc_blocks ; /* Nr of blocks to try to preallocate*/
__u8 s_prealloc_dir_blocks ; /* Nr to preallocate for dirs */
__le16 s_reserved_gdt_blocks ; /* Per group desc for online growth */
/*
2006-10-11 12:20:53 +04:00
* Journaling support valid if EXT4_FEATURE_COMPAT_HAS_JOURNAL set .
2006-10-11 12:20:50 +04:00
*/
/*D0*/ __u8 s_journal_uuid [ 16 ] ; /* uuid of journal superblock */
/*E0*/ __le32 s_journal_inum ; /* inode number of journal file */
__le32 s_journal_dev ; /* device number of journal file */
__le32 s_last_orphan ; /* start of list of inodes to delete */
__le32 s_hash_seed [ 4 ] ; /* HTREE hash seed */
__u8 s_def_hash_version ; /* Default hash version to use */
2010-07-27 19:56:04 +04:00
__u8 s_jnl_backup_type ;
2006-10-11 12:21:14 +04:00
__le16 s_desc_size ; /* size of group descriptor */
2006-10-11 12:21:10 +04:00
/*100*/ __le32 s_default_mount_opts ;
2006-10-11 12:20:50 +04:00
__le32 s_first_meta_bg ; /* First metablock block group */
2006-10-11 12:21:10 +04:00
__le32 s_mkfs_time ; /* When the filesystem was created */
__le32 s_jnl_blocks [ 17 ] ; /* Backup of the journal inode */
/* 64bit support valid if EXT4_FEATURE_COMPAT_64BIT */
/*150*/ __le32 s_blocks_count_hi ; /* Blocks count */
__le32 s_r_blocks_count_hi ; /* Reserved blocks count */
__le32 s_free_blocks_count_hi ; /* Free blocks count */
2007-10-17 02:38:25 +04:00
__le16 s_min_extra_isize ; /* All inodes have at least # bytes */
__le16 s_want_extra_isize ; /* New inodes should reserve # bytes */
__le32 s_flags ; /* Miscellaneous flags */
__le16 s_raid_stride ; /* RAID stride */
2011-05-25 02:31:25 +04:00
__le16 s_mmp_update_interval ; /* # seconds to wait in MMP checking */
2007-10-17 02:38:25 +04:00
__le64 s_mmp_block ; /* Block for multi-mount protection */
__le32 s_raid_stripe_width ; /* blocks on all data disks (N*stride)*/
2008-07-12 03:27:31 +04:00
__u8 s_log_groups_per_flex ; /* FLEX_BG group size */
2012-04-30 02:23:10 +04:00
__u8 s_checksum_type ; /* metadata checksum algorithm used */
2008-07-12 03:27:31 +04:00
__le16 s_reserved_pad ;
2009-03-01 03:39:58 +03:00
__le64 s_kbytes_written ; /* nr of lifetime kilobytes written */
2010-06-29 19:00:23 +04:00
__le32 s_snapshot_inum ; /* Inode number of active snapshot */
__le32 s_snapshot_id ; /* sequential ID of active snapshot */
__le64 s_snapshot_r_blocks_count ; /* reserved blocks for active
snapshot ' s future use */
__le32 s_snapshot_list ; /* inode number of the head of the
on - disk snapshot list */
2010-07-27 19:56:03 +04:00
# define EXT4_S_ERR_START offsetof(struct ext4_super_block, s_error_count)
__le32 s_error_count ; /* number of fs errors */
__le32 s_first_error_time ; /* first time an error happened */
__le32 s_first_error_ino ; /* inode involved in first error */
__le64 s_first_error_block ; /* block involved of first error */
__u8 s_first_error_func [ 32 ] ; /* function where the error happened */
__le32 s_first_error_line ; /* line number where error happened */
__le32 s_last_error_time ; /* most recent time of an error */
__le32 s_last_error_ino ; /* inode involved in last error */
__le32 s_last_error_line ; /* line number where error happened */
__le64 s_last_error_block ; /* block involved of last error */
__u8 s_last_error_func [ 32 ] ; /* function where the error happened */
2010-08-02 07:14:20 +04:00
# define EXT4_S_ERR_END offsetof(struct ext4_super_block, s_mount_opts)
__u8 s_mount_opts [ 64 ] ;
2011-09-10 02:34:51 +04:00
__le32 s_usr_quota_inum ; /* inode for tracking user quota */
__le32 s_grp_quota_inum ; /* inode for tracking group quota */
__le32 s_overhead_clusters ; /* overhead blocks/clusters in fs */
2012-04-30 02:23:10 +04:00
__le32 s_reserved [ 108 ] ; /* Padding to the end of the block */
__le32 s_checksum ; /* crc32c(superblock) */
2006-10-11 12:20:50 +04:00
} ;
2010-07-27 19:56:03 +04:00
# define EXT4_S_ERR_LEN (EXT4_S_ERR_END - EXT4_S_ERR_START)
2006-10-11 12:20:50 +04:00
# ifdef __KERNEL__
2009-06-13 18:09:48 +04:00
/*
2009-06-13 18:09:36 +04:00
* run - time mount flags
2009-06-13 18:09:48 +04:00
*/
# define EXT4_MF_MNTDIR_SAMPLED 0x0001
2009-06-13 18:09:36 +04:00
# define EXT4_MF_FS_ABORTED 0x0002 /* Fatal error detected */
2009-06-13 18:09:48 +04:00
2009-05-04 00:33:44 +04:00
/*
* fourth extended - fs super - block data in memory
*/
struct ext4_sb_info {
unsigned long s_desc_size ; /* Size of a group descriptor in bytes */
unsigned long s_inodes_per_block ; /* Number of inodes per block */
unsigned long s_blocks_per_group ; /* Number of blocks in a group */
2011-09-10 02:34:51 +04:00
unsigned long s_clusters_per_group ; /* Number of clusters in a group */
2009-05-04 00:33:44 +04:00
unsigned long s_inodes_per_group ; /* Number of inodes in a group */
unsigned long s_itb_per_group ; /* Number of inode table blocks per group */
unsigned long s_gdb_count ; /* Number of group descriptor blocks */
unsigned long s_desc_per_block ; /* Number of group descriptors per block */
ext4_group_t s_groups_count ; /* Number of groups in the fs */
ext4: limit block allocations for indirect-block files to < 2^32
Today, the ext4 allocator will happily allocate blocks past
2^32 for indirect-block files, which results in the block
numbers getting truncated, and corruption ensues.
This patch limits such allocations to < 2^32, and adds
BUG_ONs if we do get blocks larger than that.
This should address RH Bug 519471, ext4 bitmap allocator
must limit blocks to < 2^32
* ext4_find_goal() is modified to choose a goal < UINT_MAX,
so that our starting point is in an acceptable range.
* ext4_xattr_block_set() is modified such that the goal block
is < UINT_MAX, as above.
* ext4_mb_regular_allocator() is modified so that the group
search does not continue into groups which are too high
* ext4_mb_use_preallocated() has a check that we don't use
preallocated space which is too far out
* ext4_alloc_blocks() and ext4_xattr_block_set() add some BUG_ONs
No attempt has been made to limit inode locations to < 2^32,
so we may wind up with blocks far from their inodes. Doing
this much already will lead to some odd ENOSPC issues when the
"lower 32" gets full, and further restricting inodes could
make that even weirder.
For high inodes, choosing a goal of the original, % UINT_MAX,
may be a bit odd, but then we're in an odd situation anyway,
and I don't know of a better heuristic.
Signed-off-by: Eric Sandeen <sandeen@redhat.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2009-09-16 22:45:10 +04:00
ext4_group_t s_blockfile_groups ; /* Groups acceptable for non-extent files */
2012-07-10 00:27:05 +04:00
unsigned long s_overhead ; /* # of fs overhead clusters */
2011-09-10 02:34:51 +04:00
unsigned int s_cluster_ratio ; /* Number of blocks per cluster */
unsigned int s_cluster_bits ; /* log2 of s_cluster_ratio */
2009-05-04 00:33:44 +04:00
loff_t s_bitmap_maxbytes ; /* max bytes for bitmap files */
struct buffer_head * s_sbh ; /* Buffer containing the super block */
struct ext4_super_block * s_es ; /* Pointer to the super block in the buffer */
struct buffer_head * * s_group_desc ;
2009-06-13 18:09:41 +04:00
unsigned int s_mount_opt ;
2010-12-16 04:30:48 +03:00
unsigned int s_mount_opt2 ;
2009-06-13 18:09:48 +04:00
unsigned int s_mount_flags ;
2012-03-05 04:27:31 +04:00
unsigned int s_def_mount_opt ;
2009-05-04 00:33:44 +04:00
ext4_fsblk_t s_sb_block ;
2013-04-10 06:11:22 +04:00
atomic64_t s_resv_clusters ;
2012-02-08 03:41:49 +04:00
kuid_t s_resuid ;
kgid_t s_resgid ;
2009-05-04 00:33:44 +04:00
unsigned short s_mount_state ;
unsigned short s_pad ;
int s_addr_per_block_bits ;
int s_desc_per_block_bits ;
int s_inode_size ;
int s_first_ino ;
unsigned int s_inode_readahead_blks ;
2009-06-13 19:45:35 +04:00
unsigned int s_inode_goal ;
2009-05-04 00:33:44 +04:00
spinlock_t s_next_gen_lock ;
u32 s_next_generation ;
u32 s_hash_seed [ 4 ] ;
int s_def_hash_version ;
int s_hash_unsigned ; /* 3 if hash should be signed, 0 if not */
2011-09-10 02:56:51 +04:00
struct percpu_counter s_freeclusters_counter ;
2009-05-04 00:33:44 +04:00
struct percpu_counter s_freeinodes_counter ;
struct percpu_counter s_dirs_counter ;
2011-09-10 02:56:51 +04:00
struct percpu_counter s_dirtyclusters_counter ;
2009-05-04 00:33:44 +04:00
struct blockgroup_lock * s_blockgroup_lock ;
struct proc_dir_entry * s_proc ;
struct kobject s_kobj ;
struct completion s_kobj_unregister ;
2012-05-31 06:56:46 +04:00
struct super_block * s_sb ;
2009-05-04 00:33:44 +04:00
/* Journaling */
struct journal_s * s_journal ;
struct list_head s_orphan ;
struct mutex s_orphan_lock ;
2011-07-27 05:35:44 +04:00
unsigned long s_resize_flags ; /* Flags indicating if there
is a resizer */
2009-05-04 00:33:44 +04:00
unsigned long s_commit_interval ;
u32 s_max_batch_time ;
u32 s_min_batch_time ;
struct block_device * journal_bdev ;
# ifdef CONFIG_QUOTA
char * s_qf_names [ MAXQUOTAS ] ; /* Names of quota files with journalled quota */
int s_jquota_fmt ; /* Format of quota to use */
# endif
unsigned int s_want_extra_isize ; /* New inodes should reserve # bytes */
2009-05-17 23:38:01 +04:00
struct rb_root system_blks ;
2009-05-04 00:33:44 +04:00
# ifdef EXTENTS_STATS
/* ext4 extents stats */
unsigned long s_ext_min ;
unsigned long s_ext_max ;
unsigned long s_depth_max ;
spinlock_t s_ext_stats_lock ;
unsigned long s_ext_blocks ;
unsigned long s_ext_extents ;
# endif
/* for buddy allocator */
struct ext4_group_info * * * s_group_info ;
struct inode * s_buddy_cache ;
spinlock_t s_md_lock ;
unsigned short * s_mb_offsets ;
unsigned int * s_mb_maxs ;
2012-09-05 09:31:50 +04:00
unsigned int s_group_info_size ;
2009-05-04 00:33:44 +04:00
/* tunables */
unsigned long s_stripe ;
unsigned int s_mb_stream_request ;
unsigned int s_mb_max_to_scan ;
unsigned int s_mb_min_to_scan ;
unsigned int s_mb_stats ;
unsigned int s_mb_order2_reqs ;
unsigned int s_mb_group_prealloc ;
2012-08-17 17:48:17 +04:00
unsigned int s_max_dir_size_kb ;
2009-05-04 00:33:44 +04:00
/* where last allocation was done - for stream allocation */
unsigned long s_mb_last_group ;
unsigned long s_mb_last_start ;
/* stats for buddy allocator */
atomic_t s_bal_reqs ; /* number of reqs with len > 1 */
atomic_t s_bal_success ; /* we found long enough chunks */
atomic_t s_bal_allocated ; /* in blocks */
atomic_t s_bal_ex_scanned ; /* total extents scanned */
atomic_t s_bal_goals ; /* goal hits */
atomic_t s_bal_breaks ; /* too long searches */
atomic_t s_bal_2orders ; /* 2^order hits */
spinlock_t s_bal_lock ;
unsigned long s_mb_buddies_generated ;
unsigned long long s_mb_generation_time ;
atomic_t s_mb_lost_chunks ;
atomic_t s_mb_preallocated ;
atomic_t s_mb_discarded ;
ext4: Avoid group preallocation for closed files
Currently the group preallocation code tries to find a large (512)
free block from which to do per-cpu group allocation for small files.
The problem with this scheme is that it leaves the filesystem horribly
fragmented. In the worst case, if the filesystem is unmounted and
remounted (after a system shutdown, for example) we forget the fact
that wee were using a particular (now-partially filled) 512 block
extent. So the next time we try to allocate space for a small file,
we will find *another* completely free 512 block chunk to allocate
small files. Given that there are 32,768 blocks in a block group,
after 64 iterations of "mount, write one 4k file in a directory,
unmount", the block group will have 64 files, each separated by 511
blocks, and the block group will no longer have any free 512
completely free chunks of blocks for group preallocation space.
So if we try to allocate blocks for a file that has been closed, such
that we know the final size of the file, and the filesystem is not
busy, avoid using group preallocation.
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2009-09-18 21:34:02 +04:00
atomic_t s_lock_busy ;
2009-05-04 00:33:44 +04:00
/* locality groups */
2010-02-02 08:39:01 +03:00
struct ext4_locality_group __percpu * s_locality_groups ;
2009-05-04 00:33:44 +04:00
/* for write statistics */
unsigned long s_sectors_written_start ;
u64 s_kbytes_written ;
2012-08-17 17:54:17 +04:00
/* the size of zero-out chunk */
unsigned int s_extent_max_zeroout_kb ;
2009-05-04 00:33:44 +04:00
unsigned int s_log_groups_per_flex ;
struct flex_groups * s_flex_groups ;
2012-09-05 09:29:50 +04:00
ext4_group_t s_flex_groups_allocated ;
2009-09-28 23:48:41 +04:00
2013-06-04 22:21:02 +04:00
/* workqueue for reserved extent conversions (buffered io) */
struct workqueue_struct * rsv_conversion_wq ;
2010-07-27 19:56:04 +04:00
/* timer for periodic error stats printing */
struct timer_list s_err_report ;
ext4: add support for lazy inode table initialization
When the lazy_itable_init extended option is passed to mke2fs, it
considerably speeds up filesystem creation because inode tables are
not zeroed out. The fact that parts of the inode table are
uninitialized is not a problem so long as the block group descriptors,
which contain information regarding how much of the inode table has
been initialized, has not been corrupted However, if the block group
checksums are not valid, e2fsck must scan the entire inode table, and
the the old, uninitialized data could potentially cause e2fsck to
report false problems.
Hence, it is important for the inode tables to be initialized as soon
as possble. This commit adds this feature so that mke2fs can safely
use the lazy inode table initialization feature to speed up formatting
file systems.
This is done via a new new kernel thread called ext4lazyinit, which is
created on demand and destroyed, when it is no longer needed. There
is only one thread for all ext4 filesystems in the system. When the
first filesystem with inititable mount option is mounted, ext4lazyinit
thread is created, then the filesystem can register its request in the
request list.
This thread then walks through the list of requests picking up
scheduled requests and invoking ext4_init_inode_table(). Next schedule
time for the request is computed by multiplying the time it took to
zero out last inode table with wait multiplier, which can be set with
the (init_itable=n) mount option (default is 10). We are doing
this so we do not take the whole I/O bandwidth. When the thread is no
longer necessary (request list is empty) it frees the appropriate
structures and exits (and can be created later later by another
filesystem).
We do not disturb regular inode allocations in any way, it just do not
care whether the inode table is, or is not zeroed. But when zeroing, we
have to skip used inodes, obviously. Also we should prevent new inode
allocations from the group, while zeroing is on the way. For that we
take write alloc_sem lock in ext4_init_inode_table() and read alloc_sem
in the ext4_claim_inode, so when we are unlucky and allocator hits the
group which is currently being zeroed, it just has to wait.
This can be suppresed using the mount option no_init_itable.
Signed-off-by: Lukas Czerner <lczerner@redhat.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2010-10-28 05:30:05 +04:00
/* Lazy inode table initialization info */
struct ext4_li_request * s_li_request ;
/* Wait multiplier for lazy initialization thread */
unsigned int s_li_wait_mult ;
2011-05-25 02:31:25 +04:00
/* Kernel thread for multiple mount protection */
struct task_struct * s_mmp_tsk ;
2011-07-11 08:03:38 +04:00
/* record the last minlen when FITRIM is called. */
atomic_t s_last_trim_minblks ;
2012-04-30 02:27:10 +04:00
/* Reference to checksum algorithm driver via cryptoapi */
struct crypto_shash * s_chksum_driver ;
2012-04-30 02:29:10 +04:00
/* Precomputed FS UUID checksum for seeding other checksums */
__u32 s_csum_seed ;
2013-02-18 09:32:55 +04:00
/* Reclaim extents from extent status tree */
struct shrinker s_es_shrinker ;
struct list_head s_es_lru ;
2013-07-01 16:12:37 +04:00
unsigned long s_es_last_sorted ;
2013-03-02 19:27:46 +04:00
struct percpu_counter s_extent_cache_cnt ;
2014-03-19 03:24:49 +04:00
struct mb_cache * s_mb_cache ;
2013-02-18 09:32:55 +04:00
spinlock_t s_es_lru_lock ____cacheline_aligned_in_smp ;
2013-10-18 05:11:01 +04:00
/* Ratelimit ext4 messages. */
struct ratelimit_state s_err_ratelimit_state ;
struct ratelimit_state s_warning_ratelimit_state ;
struct ratelimit_state s_msg_ratelimit_state ;
2009-05-04 00:33:44 +04:00
} ;
2008-09-09 06:25:24 +04:00
static inline struct ext4_sb_info * EXT4_SB ( struct super_block * sb )
2006-10-11 12:20:50 +04:00
{
return sb - > s_fs_info ;
}
2006-10-11 12:20:53 +04:00
static inline struct ext4_inode_info * EXT4_I ( struct inode * inode )
2006-10-11 12:20:50 +04:00
{
2006-10-11 12:20:53 +04:00
return container_of ( inode , struct ext4_inode_info , vfs_inode ) ;
2006-10-11 12:20:50 +04:00
}
2007-07-18 17:15:20 +04:00
static inline struct timespec ext4_current_time ( struct inode * inode )
{
return ( inode - > i_sb - > s_time_gran < NSEC_PER_SEC ) ?
current_fs_time ( inode - > i_sb ) : CURRENT_TIME_SEC ;
}
2006-10-11 12:20:53 +04:00
static inline int ext4_valid_inum ( struct super_block * sb , unsigned long ino )
2006-10-11 12:20:50 +04:00
{
2006-10-11 12:20:53 +04:00
return ino = = EXT4_ROOT_INO | |
ext4: make quota as first class supported feature
This patch adds support for quotas as a first class feature in ext4;
which is to say, the quota files are stored in hidden inodes as file
system metadata, instead of as separate files visible in the file system
directory hierarchy.
It is based on the proposal at:
https://ext4.wiki.kernel.org/index.php/Design_For_1st_Class_Quota_in_Ext4
This patch introduces a new feature - EXT4_FEATURE_RO_COMPAT_QUOTA
which, when turned on, enables quota accounting at mount time
iteself. Also, the quota inodes are stored in two additional superblock
fields. Some changes introduced by this patch that should be pointed
out are:
1) Two new ext4-superblock fields - s_usr_quota_inum and
s_grp_quota_inum for storing the quota inodes in use.
2) Default quota inodes are: inode#3 for tracking userquota and inode#4
for tracking group quota. The superblock fields can be set to use
other inodes as well.
3) If the QUOTA feature and corresponding quota inodes are set in
superblock, the quota usage tracking is turned on at mount time. On
'quotaon' ioctl, the quota limits enforcement is turned
on. 'quotaoff' ioctl turns off only the limits enforcement in this
case.
4) When QUOTA feature is in use, the quota mount options 'quota',
'usrquota', 'grpquota' are ignored by the kernel.
5) mke2fs or tune2fs can be used to set the QUOTA feature and initialize
quota inodes. The default reserved inodes will not be visible to user
as regular files.
6) The quota-tools will need to be modified to support hidden quota
files on ext4. E2fsprogs will also include support for creating and
fixing quota files.
7) Support is only for the new V2 quota file format.
Tested-by: Jan Kara <jack@suse.cz>
Reviewed-by: Jan Kara <jack@suse.cz>
Reviewed-by: Johann Lombardi <johann@whamcloud.com>
Signed-off-by: Aditya Kali <adityakali@google.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2012-07-23 04:21:31 +04:00
ino = = EXT4_USR_QUOTA_INO | |
ino = = EXT4_GRP_QUOTA_INO | |
2013-04-08 20:54:05 +04:00
ino = = EXT4_BOOT_LOADER_INO | |
2006-10-11 12:20:53 +04:00
ino = = EXT4_JOURNAL_INO | |
ino = = EXT4_RESIZE_INO | |
( ino > = EXT4_FIRST_INO ( sb ) & &
ino < = le32_to_cpu ( EXT4_SB ( sb ) - > s_es - > s_inodes_count ) ) ;
2006-10-11 12:20:50 +04:00
}
2010-01-24 22:34:07 +03:00
2011-11-01 01:30:44 +04:00
static inline void ext4_set_io_unwritten_flag ( struct inode * inode ,
struct ext4_io_end * io_end )
{
if ( ! ( io_end - > flag & EXT4_IO_END_UNWRITTEN ) ) {
io_end - > flag | = EXT4_IO_END_UNWRITTEN ;
2012-09-29 07:24:52 +04:00
atomic_inc ( & EXT4_I ( inode ) - > i_unwritten ) ;
2011-11-01 01:30:44 +04:00
}
}
2012-09-29 07:21:09 +04:00
static inline ext4_io_end_t * ext4_inode_aio ( struct inode * inode )
{
return inode - > i_private ;
}
static inline void ext4_inode_aio_set ( struct inode * inode , ext4_io_end_t * io )
{
inode - > i_private = io ;
}
2010-01-24 22:34:07 +03:00
/*
* Inode dynamic state flags
*/
enum {
EXT4_STATE_JDATA , /* journaled data exists */
EXT4_STATE_NEW , /* inode is newly created */
EXT4_STATE_XATTR , /* has in-inode xattrs */
EXT4_STATE_NO_EXPAND , /* No space for expansion */
EXT4_STATE_DA_ALLOC_CLOSE , /* Alloc DA blks on close */
EXT4_STATE_EXT_MIGRATE , /* Inode is migrating */
EXT4_STATE_DIO_UNWRITTEN , /* need convert on dio done*/
2010-05-17 16:00:00 +04:00
EXT4_STATE_NEWENTRY , /* File just added to dir */
2011-01-10 20:12:36 +03:00
EXT4_STATE_DELALLOC_RESERVED , /* blks already reserved for delalloc */
2012-09-29 08:41:21 +04:00
EXT4_STATE_DIOREAD_LOCK , /* Disable support for dio read
nolocking */
2012-12-10 23:04:46 +04:00
EXT4_STATE_MAY_INLINE_DATA , /* may have in-inode data */
2013-04-03 20:39:17 +04:00
EXT4_STATE_ORDERED_MODE , /* data=ordered mode */
2013-08-17 06:05:14 +04:00
EXT4_STATE_EXT_PRECACHED , /* extents have been precached */
2010-01-24 22:34:07 +03:00
} ;
2011-01-10 20:18:25 +03:00
# define EXT4_INODE_BIT_FNS(name, field, offset) \
2010-05-17 06:00:00 +04:00
static inline int ext4_test_inode_ # # name ( struct inode * inode , int bit ) \
{ \
2011-01-10 20:18:25 +03:00
return test_bit ( bit + ( offset ) , & EXT4_I ( inode ) - > i_ # # field ) ; \
2010-05-17 06:00:00 +04:00
} \
static inline void ext4_set_inode_ # # name ( struct inode * inode , int bit ) \
{ \
2011-01-10 20:18:25 +03:00
set_bit ( bit + ( offset ) , & EXT4_I ( inode ) - > i_ # # field ) ; \
2010-05-17 06:00:00 +04:00
} \
static inline void ext4_clear_inode_ # # name ( struct inode * inode , int bit ) \
{ \
2011-01-10 20:18:25 +03:00
clear_bit ( bit + ( offset ) , & EXT4_I ( inode ) - > i_ # # field ) ; \
2010-01-24 22:34:07 +03:00
}
2013-11-12 07:40:40 +04:00
/* Add these declarations here only so that these functions can be
* found by name . Otherwise , they are very hard to locate . */
static inline int ext4_test_inode_flag ( struct inode * inode , int bit ) ;
static inline void ext4_set_inode_flag ( struct inode * inode , int bit ) ;
static inline void ext4_clear_inode_flag ( struct inode * inode , int bit ) ;
2011-01-10 20:18:25 +03:00
EXT4_INODE_BIT_FNS ( flag , flags , 0 )
2013-11-12 07:40:40 +04:00
/* Add these declarations here only so that these functions can be
* found by name . Otherwise , they are very hard to locate . */
static inline int ext4_test_inode_state ( struct inode * inode , int bit ) ;
static inline void ext4_set_inode_state ( struct inode * inode , int bit ) ;
static inline void ext4_clear_inode_state ( struct inode * inode , int bit ) ;
2011-01-10 20:18:25 +03:00
# if (BITS_PER_LONG < 64)
EXT4_INODE_BIT_FNS ( state , state_flags , 0 )
static inline void ext4_clear_state_flags ( struct ext4_inode_info * ei )
{
( ei ) - > i_state_flags = 0 ;
}
# else
EXT4_INODE_BIT_FNS ( state , flags , 32 )
static inline void ext4_clear_state_flags ( struct ext4_inode_info * ei )
{
/* We depend on the fact that callers will set i_flags */
}
# endif
2006-10-11 12:20:50 +04:00
# else
2006-10-11 12:20:53 +04:00
/* Assume that user mode programs are passing in an ext4fs superblock, not
2006-10-11 12:20:50 +04:00
* a kernel struct super_block . This will allow us to call the feature - test
* macros from user land . */
2006-10-11 12:20:53 +04:00
# define EXT4_SB(sb) (sb)
2006-10-11 12:20:50 +04:00
# endif
2006-10-11 12:20:53 +04:00
# define NEXT_ORPHAN(inode) EXT4_I(inode)->i_dtime
2006-10-11 12:20:50 +04:00
/*
* Codes for operating systems
*/
2006-10-11 12:20:53 +04:00
# define EXT4_OS_LINUX 0
# define EXT4_OS_HURD 1
# define EXT4_OS_MASIX 2
# define EXT4_OS_FREEBSD 3
# define EXT4_OS_LITES 4
2006-10-11 12:20:50 +04:00
/*
* Revision levels
*/
2006-10-11 12:20:53 +04:00
# define EXT4_GOOD_OLD_REV 0 /* The good old (original) format */
# define EXT4_DYNAMIC_REV 1 /* V2 format w/ dynamic inode sizes */
2006-10-11 12:20:50 +04:00
2006-10-11 12:20:53 +04:00
# define EXT4_CURRENT_REV EXT4_GOOD_OLD_REV
# define EXT4_MAX_SUPP_REV EXT4_DYNAMIC_REV
2006-10-11 12:20:50 +04:00
2006-10-11 12:20:53 +04:00
# define EXT4_GOOD_OLD_INODE_SIZE 128
2006-10-11 12:20:50 +04:00
/*
* Feature set definitions
*/
2006-10-11 12:20:53 +04:00
# define EXT4_HAS_COMPAT_FEATURE(sb,mask) \
2008-11-22 23:04:59 +03:00
( ( EXT4_SB ( sb ) - > s_es - > s_feature_compat & cpu_to_le32 ( mask ) ) ! = 0 )
2006-10-11 12:20:53 +04:00
# define EXT4_HAS_RO_COMPAT_FEATURE(sb,mask) \
2008-11-22 23:04:59 +03:00
( ( EXT4_SB ( sb ) - > s_es - > s_feature_ro_compat & cpu_to_le32 ( mask ) ) ! = 0 )
2006-10-11 12:20:53 +04:00
# define EXT4_HAS_INCOMPAT_FEATURE(sb,mask) \
2008-11-22 23:04:59 +03:00
( ( EXT4_SB ( sb ) - > s_es - > s_feature_incompat & cpu_to_le32 ( mask ) ) ! = 0 )
2006-10-11 12:20:53 +04:00
# define EXT4_SET_COMPAT_FEATURE(sb,mask) \
EXT4_SB ( sb ) - > s_es - > s_feature_compat | = cpu_to_le32 ( mask )
# define EXT4_SET_RO_COMPAT_FEATURE(sb,mask) \
EXT4_SB ( sb ) - > s_es - > s_feature_ro_compat | = cpu_to_le32 ( mask )
# define EXT4_SET_INCOMPAT_FEATURE(sb,mask) \
EXT4_SB ( sb ) - > s_es - > s_feature_incompat | = cpu_to_le32 ( mask )
# define EXT4_CLEAR_COMPAT_FEATURE(sb,mask) \
EXT4_SB ( sb ) - > s_es - > s_feature_compat & = ~ cpu_to_le32 ( mask )
# define EXT4_CLEAR_RO_COMPAT_FEATURE(sb,mask) \
EXT4_SB ( sb ) - > s_es - > s_feature_ro_compat & = ~ cpu_to_le32 ( mask )
# define EXT4_CLEAR_INCOMPAT_FEATURE(sb,mask) \
EXT4_SB ( sb ) - > s_es - > s_feature_incompat & = ~ cpu_to_le32 ( mask )
# define EXT4_FEATURE_COMPAT_DIR_PREALLOC 0x0001
# define EXT4_FEATURE_COMPAT_IMAGIC_INODES 0x0002
# define EXT4_FEATURE_COMPAT_HAS_JOURNAL 0x0004
# define EXT4_FEATURE_COMPAT_EXT_ATTR 0x0008
# define EXT4_FEATURE_COMPAT_RESIZE_INODE 0x0010
# define EXT4_FEATURE_COMPAT_DIR_INDEX 0x0020
# define EXT4_FEATURE_RO_COMPAT_SPARSE_SUPER 0x0001
# define EXT4_FEATURE_RO_COMPAT_LARGE_FILE 0x0002
# define EXT4_FEATURE_RO_COMPAT_BTREE_DIR 0x0004
2008-01-29 07:58:26 +03:00
# define EXT4_FEATURE_RO_COMPAT_HUGE_FILE 0x0008
Ext4: Uninitialized Block Groups
In pass1 of e2fsck, every inode table in the fileystem is scanned and checked,
regardless of whether it is in use. This is this the most time consuming part
of the filesystem check. The unintialized block group feature can greatly
reduce e2fsck time by eliminating checking of uninitialized inodes.
With this feature, there is a a high water mark of used inodes for each block
group. Block and inode bitmaps can be uninitialized on disk via a flag in the
group descriptor to avoid reading or scanning them at e2fsck time. A checksum
of each group descriptor is used to ensure that corruption in the group
descriptor's bit flags does not cause incorrect operation.
The feature is enabled through a mkfs option
mke2fs /dev/ -O uninit_groups
A patch adding support for uninitialized block groups to e2fsprogs tools has
been posted to the linux-ext4 mailing list.
The patches have been stress tested with fsstress and fsx. In performance
tests testing e2fsck time, we have seen that e2fsck time on ext3 grows
linearly with the total number of inodes in the filesytem. In ext4 with the
uninitialized block groups feature, the e2fsck time is constant, based
solely on the number of used inodes rather than the total inode count.
Since typical ext4 filesystems only use 1-10% of their inodes, this feature can
greatly reduce e2fsck time for users. With performance improvement of 2-20
times, depending on how full the filesystem is.
The attached graph shows the major improvements in e2fsck times in filesystems
with a large total inode count, but few inodes in use.
In each group descriptor if we have
EXT4_BG_INODE_UNINIT set in bg_flags:
Inode table is not initialized/used in this group. So we can skip
the consistency check during fsck.
EXT4_BG_BLOCK_UNINIT set in bg_flags:
No block in the group is used. So we can skip the block bitmap
verification for this group.
We also add two new fields to group descriptor as a part of
uninitialized group patch.
__le16 bg_itable_unused; /* Unused inodes count */
__le16 bg_checksum; /* crc16(sb_uuid+group+desc) */
bg_itable_unused:
If we have EXT4_BG_INODE_UNINIT not set in bg_flags
then bg_itable_unused will give the offset within
the inode table till the inodes are used. This can be
used by fsck to skip list of inodes that are marked unused.
bg_checksum:
Now that we depend on bg_flags and bg_itable_unused to determine
the block and inode usage, we need to make sure group descriptor
is not corrupt. We add checksum to group descriptor to
detect corruption. If the descriptor is found to be corrupt, we
mark all the blocks and inodes in the group used.
Signed-off-by: Avantika Mathur <mathur@us.ibm.com>
Signed-off-by: Andreas Dilger <adilger@clusterfs.com>
Signed-off-by: Mingming Cao <cmm@us.ibm.com>
Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>
2007-10-17 02:38:25 +04:00
# define EXT4_FEATURE_RO_COMPAT_GDT_CSUM 0x0010
2007-07-18 16:38:01 +04:00
# define EXT4_FEATURE_RO_COMPAT_DIR_NLINK 0x0020
2007-07-18 17:15:20 +04:00
# define EXT4_FEATURE_RO_COMPAT_EXTRA_ISIZE 0x0040
2011-05-25 03:00:39 +04:00
# define EXT4_FEATURE_RO_COMPAT_QUOTA 0x0100
2011-09-10 02:34:51 +04:00
# define EXT4_FEATURE_RO_COMPAT_BIGALLOC 0x0200
2012-04-30 02:23:10 +04:00
/*
* METADATA_CSUM also enables group descriptor checksums ( GDT_CSUM ) . When
* METADATA_CSUM is set , group descriptor checksums use the same algorithm as
* all other data structures ' checksums . However , the METADATA_CSUM and
* GDT_CSUM bits are mutually exclusive .
*/
2012-01-05 07:01:53 +04:00
# define EXT4_FEATURE_RO_COMPAT_METADATA_CSUM 0x0400
2006-10-11 12:20:53 +04:00
# define EXT4_FEATURE_INCOMPAT_COMPRESSION 0x0001
# define EXT4_FEATURE_INCOMPAT_FILETYPE 0x0002
# define EXT4_FEATURE_INCOMPAT_RECOVER 0x0004 /* Needs recovery */
# define EXT4_FEATURE_INCOMPAT_JOURNAL_DEV 0x0008 /* Journal device */
# define EXT4_FEATURE_INCOMPAT_META_BG 0x0010
2006-10-11 12:21:03 +04:00
# define EXT4_FEATURE_INCOMPAT_EXTENTS 0x0040 /* extents support */
2006-10-11 12:21:10 +04:00
# define EXT4_FEATURE_INCOMPAT_64BIT 0x0080
2008-01-29 07:58:26 +03:00
# define EXT4_FEATURE_INCOMPAT_MMP 0x0100
2007-10-17 02:38:25 +04:00
# define EXT4_FEATURE_INCOMPAT_FLEX_BG 0x0200
2010-01-25 11:31:32 +03:00
# define EXT4_FEATURE_INCOMPAT_EA_INODE 0x0400 /* EA in inode */
# define EXT4_FEATURE_INCOMPAT_DIRDATA 0x1000 /* data in dirent */
2012-02-21 02:53:01 +04:00
# define EXT4_FEATURE_INCOMPAT_BG_USE_META_CSUM 0x2000 /* use crc32c for bg */
2012-01-05 07:01:53 +04:00
# define EXT4_FEATURE_INCOMPAT_LARGEDIR 0x4000 /* >2GB or 3-lvl htree */
2012-12-10 23:04:46 +04:00
# define EXT4_FEATURE_INCOMPAT_INLINE_DATA 0x8000 /* data in inode */
2006-10-11 12:20:53 +04:00
2011-04-19 01:29:14 +04:00
# define EXT2_FEATURE_COMPAT_SUPP EXT4_FEATURE_COMPAT_EXT_ATTR
# define EXT2_FEATURE_INCOMPAT_SUPP (EXT4_FEATURE_INCOMPAT_FILETYPE| \
EXT4_FEATURE_INCOMPAT_META_BG )
# define EXT2_FEATURE_RO_COMPAT_SUPP (EXT4_FEATURE_RO_COMPAT_SPARSE_SUPER| \
EXT4_FEATURE_RO_COMPAT_LARGE_FILE | \
EXT4_FEATURE_RO_COMPAT_BTREE_DIR )
# define EXT3_FEATURE_COMPAT_SUPP EXT4_FEATURE_COMPAT_EXT_ATTR
# define EXT3_FEATURE_INCOMPAT_SUPP (EXT4_FEATURE_INCOMPAT_FILETYPE| \
EXT4_FEATURE_INCOMPAT_RECOVER | \
EXT4_FEATURE_INCOMPAT_META_BG )
# define EXT3_FEATURE_RO_COMPAT_SUPP (EXT4_FEATURE_RO_COMPAT_SPARSE_SUPER| \
EXT4_FEATURE_RO_COMPAT_LARGE_FILE | \
EXT4_FEATURE_RO_COMPAT_BTREE_DIR )
2006-10-11 12:20:53 +04:00
# define EXT4_FEATURE_COMPAT_SUPP EXT2_FEATURE_COMPAT_EXT_ATTR
# define EXT4_FEATURE_INCOMPAT_SUPP (EXT4_FEATURE_INCOMPAT_FILETYPE| \
EXT4_FEATURE_INCOMPAT_RECOVER | \
2006-10-11 12:21:03 +04:00
EXT4_FEATURE_INCOMPAT_META_BG | \
2006-10-11 12:21:10 +04:00
EXT4_FEATURE_INCOMPAT_EXTENTS | \
2007-10-17 02:38:25 +04:00
EXT4_FEATURE_INCOMPAT_64BIT | \
2011-05-25 02:31:25 +04:00
EXT4_FEATURE_INCOMPAT_FLEX_BG | \
2012-12-10 23:06:03 +04:00
EXT4_FEATURE_INCOMPAT_MMP | \
EXT4_FEATURE_INCOMPAT_INLINE_DATA )
2006-10-11 12:20:53 +04:00
# define EXT4_FEATURE_RO_COMPAT_SUPP (EXT4_FEATURE_RO_COMPAT_SPARSE_SUPER| \
EXT4_FEATURE_RO_COMPAT_LARGE_FILE | \
Ext4: Uninitialized Block Groups
In pass1 of e2fsck, every inode table in the fileystem is scanned and checked,
regardless of whether it is in use. This is this the most time consuming part
of the filesystem check. The unintialized block group feature can greatly
reduce e2fsck time by eliminating checking of uninitialized inodes.
With this feature, there is a a high water mark of used inodes for each block
group. Block and inode bitmaps can be uninitialized on disk via a flag in the
group descriptor to avoid reading or scanning them at e2fsck time. A checksum
of each group descriptor is used to ensure that corruption in the group
descriptor's bit flags does not cause incorrect operation.
The feature is enabled through a mkfs option
mke2fs /dev/ -O uninit_groups
A patch adding support for uninitialized block groups to e2fsprogs tools has
been posted to the linux-ext4 mailing list.
The patches have been stress tested with fsstress and fsx. In performance
tests testing e2fsck time, we have seen that e2fsck time on ext3 grows
linearly with the total number of inodes in the filesytem. In ext4 with the
uninitialized block groups feature, the e2fsck time is constant, based
solely on the number of used inodes rather than the total inode count.
Since typical ext4 filesystems only use 1-10% of their inodes, this feature can
greatly reduce e2fsck time for users. With performance improvement of 2-20
times, depending on how full the filesystem is.
The attached graph shows the major improvements in e2fsck times in filesystems
with a large total inode count, but few inodes in use.
In each group descriptor if we have
EXT4_BG_INODE_UNINIT set in bg_flags:
Inode table is not initialized/used in this group. So we can skip
the consistency check during fsck.
EXT4_BG_BLOCK_UNINIT set in bg_flags:
No block in the group is used. So we can skip the block bitmap
verification for this group.
We also add two new fields to group descriptor as a part of
uninitialized group patch.
__le16 bg_itable_unused; /* Unused inodes count */
__le16 bg_checksum; /* crc16(sb_uuid+group+desc) */
bg_itable_unused:
If we have EXT4_BG_INODE_UNINIT not set in bg_flags
then bg_itable_unused will give the offset within
the inode table till the inodes are used. This can be
used by fsck to skip list of inodes that are marked unused.
bg_checksum:
Now that we depend on bg_flags and bg_itable_unused to determine
the block and inode usage, we need to make sure group descriptor
is not corrupt. We add checksum to group descriptor to
detect corruption. If the descriptor is found to be corrupt, we
mark all the blocks and inodes in the group used.
Signed-off-by: Avantika Mathur <mathur@us.ibm.com>
Signed-off-by: Andreas Dilger <adilger@clusterfs.com>
Signed-off-by: Mingming Cao <cmm@us.ibm.com>
Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>
2007-10-17 02:38:25 +04:00
EXT4_FEATURE_RO_COMPAT_GDT_CSUM | \
2007-07-18 16:38:01 +04:00
EXT4_FEATURE_RO_COMPAT_DIR_NLINK | \
2007-07-18 17:15:20 +04:00
EXT4_FEATURE_RO_COMPAT_EXTRA_ISIZE | \
2008-01-29 07:58:26 +03:00
EXT4_FEATURE_RO_COMPAT_BTREE_DIR | \
2011-09-10 03:06:51 +04:00
EXT4_FEATURE_RO_COMPAT_HUGE_FILE | \
2012-05-27 16:12:42 +04:00
EXT4_FEATURE_RO_COMPAT_BIGALLOC | \
ext4: make quota as first class supported feature
This patch adds support for quotas as a first class feature in ext4;
which is to say, the quota files are stored in hidden inodes as file
system metadata, instead of as separate files visible in the file system
directory hierarchy.
It is based on the proposal at:
https://ext4.wiki.kernel.org/index.php/Design_For_1st_Class_Quota_in_Ext4
This patch introduces a new feature - EXT4_FEATURE_RO_COMPAT_QUOTA
which, when turned on, enables quota accounting at mount time
iteself. Also, the quota inodes are stored in two additional superblock
fields. Some changes introduced by this patch that should be pointed
out are:
1) Two new ext4-superblock fields - s_usr_quota_inum and
s_grp_quota_inum for storing the quota inodes in use.
2) Default quota inodes are: inode#3 for tracking userquota and inode#4
for tracking group quota. The superblock fields can be set to use
other inodes as well.
3) If the QUOTA feature and corresponding quota inodes are set in
superblock, the quota usage tracking is turned on at mount time. On
'quotaon' ioctl, the quota limits enforcement is turned
on. 'quotaoff' ioctl turns off only the limits enforcement in this
case.
4) When QUOTA feature is in use, the quota mount options 'quota',
'usrquota', 'grpquota' are ignored by the kernel.
5) mke2fs or tune2fs can be used to set the QUOTA feature and initialize
quota inodes. The default reserved inodes will not be visible to user
as regular files.
6) The quota-tools will need to be modified to support hidden quota
files on ext4. E2fsprogs will also include support for creating and
fixing quota files.
7) Support is only for the new V2 quota file format.
Tested-by: Jan Kara <jack@suse.cz>
Reviewed-by: Jan Kara <jack@suse.cz>
Reviewed-by: Johann Lombardi <johann@whamcloud.com>
Signed-off-by: Aditya Kali <adityakali@google.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2012-07-23 04:21:31 +04:00
EXT4_FEATURE_RO_COMPAT_METADATA_CSUM | \
EXT4_FEATURE_RO_COMPAT_QUOTA )
2006-10-11 12:20:50 +04:00
/*
* Default values for user and / or group using reserved blocks
*/
2006-10-11 12:20:53 +04:00
# define EXT4_DEF_RESUID 0
# define EXT4_DEF_RESGID 0
2006-10-11 12:20:50 +04:00
2008-10-10 07:53:47 +04:00
# define EXT4_DEF_INODE_READAHEAD_BLKS 32
2006-10-11 12:20:50 +04:00
/*
* Default mount options
*/
2006-10-11 12:20:53 +04:00
# define EXT4_DEFM_DEBUG 0x0001
# define EXT4_DEFM_BSDGROUPS 0x0002
# define EXT4_DEFM_XATTR_USER 0x0004
# define EXT4_DEFM_ACL 0x0008
# define EXT4_DEFM_UID16 0x0010
# define EXT4_DEFM_JMODE 0x0060
# define EXT4_DEFM_JMODE_DATA 0x0020
# define EXT4_DEFM_JMODE_ORDERED 0x0040
# define EXT4_DEFM_JMODE_WBACK 0x0060
2010-08-02 07:14:20 +04:00
# define EXT4_DEFM_NOBARRIER 0x0100
# define EXT4_DEFM_BLOCK_VALIDITY 0x0200
# define EXT4_DEFM_DISCARD 0x0400
# define EXT4_DEFM_NODELALLOC 0x0800
2006-10-11 12:20:50 +04:00
2009-01-04 04:27:38 +03:00
/*
* Default journal batch times
*/
# define EXT4_DEF_MIN_BATCH_TIME 0
# define EXT4_DEF_MAX_BATCH_TIME 15000 /* 15ms */
2009-03-12 19:18:34 +03:00
/*
* Minimum number of groups in a flexgroup before we separate out
* directories into the first block group of a flexgroup
*/
# define EXT4_FLEX_SIZE_DIR_ALLOC_SCHEME 4
2006-10-11 12:20:50 +04:00
/*
* Structure of a directory entry
*/
2006-10-11 12:20:53 +04:00
# define EXT4_NAME_LEN 255
2006-10-11 12:20:50 +04:00
2006-10-11 12:20:53 +04:00
struct ext4_dir_entry {
2006-10-11 12:20:50 +04:00
__le32 inode ; /* Inode number */
__le16 rec_len ; /* Directory entry length */
__le16 name_len ; /* Name length */
2006-10-11 12:20:53 +04:00
char name [ EXT4_NAME_LEN ] ; /* File name */
2006-10-11 12:20:50 +04:00
} ;
/*
2006-10-11 12:20:53 +04:00
* The new version of the directory entry . Since EXT4 structures are
2006-10-11 12:20:50 +04:00
* stored in intel byte order , and the name_len field could never be
* bigger than 255 chars , it ' s safe to reclaim the extra byte for the
* file_type field .
*/
2006-10-11 12:20:53 +04:00
struct ext4_dir_entry_2 {
2006-10-11 12:20:50 +04:00
__le32 inode ; /* Inode number */
__le16 rec_len ; /* Directory entry length */
__u8 name_len ; /* Name length */
__u8 file_type ;
2006-10-11 12:20:53 +04:00
char name [ EXT4_NAME_LEN ] ; /* File name */
2006-10-11 12:20:50 +04:00
} ;
2012-04-30 02:23:10 +04:00
/*
* This is a bogus directory entry at the end of each leaf block that
* records checksums .
*/
struct ext4_dir_entry_tail {
__le32 det_reserved_zero1 ; /* Pretend to be unused */
__le16 det_rec_len ; /* 12 */
__u8 det_reserved_zero2 ; /* Zero name length */
__u8 det_reserved_ft ; /* 0xDE, fake file type */
__le32 det_checksum ; /* crc32c(uuid+inum+dirblock) */
} ;
2012-12-10 23:05:59 +04:00
# define EXT4_DIRENT_TAIL(block, blocksize) \
( ( struct ext4_dir_entry_tail * ) ( ( ( void * ) ( block ) ) + \
( ( blocksize ) - \
sizeof ( struct ext4_dir_entry_tail ) ) ) )
2006-10-11 12:20:50 +04:00
/*
2006-10-11 12:20:53 +04:00
* Ext4 directory file types . Only the low 3 bits are used . The
2006-10-11 12:20:50 +04:00
* other bits are reserved for now .
*/
2006-10-11 12:20:53 +04:00
# define EXT4_FT_UNKNOWN 0
# define EXT4_FT_REG_FILE 1
# define EXT4_FT_DIR 2
# define EXT4_FT_CHRDEV 3
# define EXT4_FT_BLKDEV 4
# define EXT4_FT_FIFO 5
# define EXT4_FT_SOCK 6
# define EXT4_FT_SYMLINK 7
2006-10-11 12:20:50 +04:00
2006-10-11 12:20:53 +04:00
# define EXT4_FT_MAX 8
2006-10-11 12:20:50 +04:00
2012-04-30 02:23:10 +04:00
# define EXT4_FT_DIR_CSUM 0xDE
2006-10-11 12:20:50 +04:00
/*
2006-10-11 12:20:53 +04:00
* EXT4_DIR_PAD defines the directory entries boundaries
2006-10-11 12:20:50 +04:00
*
* NOTE : It must be a multiple of 4
*/
2006-10-11 12:20:53 +04:00
# define EXT4_DIR_PAD 4
# define EXT4_DIR_ROUND (EXT4_DIR_PAD - 1)
# define EXT4_DIR_REC_LEN(name_len) (((name_len) + 8 + EXT4_DIR_ROUND) & \
~ EXT4_DIR_ROUND )
2008-01-29 07:58:27 +03:00
# define EXT4_MAX_REC_LEN ((1<<16)-1)
2010-08-05 09:46:37 +04:00
/*
* If we ever get support for fs block sizes > page_size , we ' ll need
* to remove the # if statements in the next two functions . . .
*/
static inline unsigned int
ext4_rec_len_from_disk ( __le16 dlen , unsigned blocksize )
{
unsigned len = le16_to_cpu ( dlen ) ;
# if (PAGE_CACHE_SIZE >= 65536)
if ( len = = EXT4_MAX_REC_LEN | | len = = 0 )
return blocksize ;
return ( len & 65532 ) | ( ( len & 3 ) < < 16 ) ;
# else
return len ;
# endif
}
static inline __le16 ext4_rec_len_to_disk ( unsigned len , unsigned blocksize )
{
if ( ( len > blocksize ) | | ( blocksize > ( 1 < < 18 ) ) | | ( len & 3 ) )
BUG ( ) ;
# if (PAGE_CACHE_SIZE >= 65536)
if ( len < 65536 )
return cpu_to_le16 ( len ) ;
if ( len = = blocksize ) {
if ( blocksize = = 65536 )
return cpu_to_le16 ( EXT4_MAX_REC_LEN ) ;
else
return cpu_to_le16 ( 0 ) ;
}
return cpu_to_le16 ( ( len & 65532 ) | ( ( len > > 16 ) & 3 ) ) ;
# else
return cpu_to_le16 ( len ) ;
# endif
}
2006-10-11 12:20:50 +04:00
/*
* Hash Tree Directory indexing
* ( c ) Daniel Phillips , 2001
*/
2007-10-17 02:38:25 +04:00
# define is_dx(dir) (EXT4_HAS_COMPAT_FEATURE(dir->i_sb, \
EXT4_FEATURE_COMPAT_DIR_INDEX ) & & \
2010-05-17 06:00:00 +04:00
ext4_test_inode_flag ( ( dir ) , EXT4_INODE_INDEX ) )
2006-10-11 12:20:53 +04:00
# define EXT4_DIR_LINK_MAX(dir) (!is_dx(dir) && (dir)->i_nlink >= EXT4_LINK_MAX)
# define EXT4_DIR_LINK_EMPTY(dir) ((dir)->i_nlink == 2 || (dir)->i_nlink == 1)
2006-10-11 12:20:50 +04:00
/* Legal values for the dx_root hash_version field: */
# define DX_HASH_LEGACY 0
# define DX_HASH_HALF_MD4 1
# define DX_HASH_TEA 2
2008-10-28 20:21:44 +03:00
# define DX_HASH_LEGACY_UNSIGNED 3
# define DX_HASH_HALF_MD4_UNSIGNED 4
# define DX_HASH_TEA_UNSIGNED 5
2006-10-11 12:20:50 +04:00
2012-04-30 02:27:10 +04:00
static inline u32 ext4_chksum ( struct ext4_sb_info * sbi , u32 crc ,
const void * address , unsigned int length )
{
struct {
struct shash_desc shash ;
2012-07-23 04:25:31 +04:00
char ctx [ 4 ] ;
2012-04-30 02:27:10 +04:00
} desc ;
int err ;
2012-07-23 04:25:31 +04:00
BUG_ON ( crypto_shash_descsize ( sbi - > s_chksum_driver ) ! = sizeof ( desc . ctx ) ) ;
2012-04-30 02:27:10 +04:00
desc . shash . tfm = sbi - > s_chksum_driver ;
desc . shash . flags = 0 ;
* ( u32 * ) desc . ctx = crc ;
err = crypto_shash_update ( & desc . shash , address , length ) ;
BUG_ON ( err ) ;
return * ( u32 * ) desc . ctx ;
}
2006-10-11 12:20:50 +04:00
# ifdef __KERNEL__
/* hash info structure used by the directory hash */
struct dx_hash_info
{
u32 hash ;
u32 minor_hash ;
int hash_version ;
u32 * seed ;
} ;
2012-03-19 06:44:40 +04:00
/* 32 and 64 bit signed EOF for dx directories */
# define EXT4_HTREE_EOF_32BIT ((1UL << (32 - 1)) - 1)
# define EXT4_HTREE_EOF_64BIT ((1ULL << (64 - 1)) - 1)
2006-10-11 12:20:50 +04:00
/*
2006-10-11 12:20:53 +04:00
* Control parameters used by ext4_htree_next_block
2006-10-11 12:20:50 +04:00
*/
# define HASH_NB_ALWAYS 1
/*
* Describe an inode ' s exact location on disk and in memory
*/
2006-10-11 12:20:53 +04:00
struct ext4_iloc
2006-10-11 12:20:50 +04:00
{
struct buffer_head * bh ;
unsigned long offset ;
2008-01-29 07:58:27 +03:00
ext4_group_t block_group ;
2006-10-11 12:20:50 +04:00
} ;
2006-10-11 12:20:53 +04:00
static inline struct ext4_inode * ext4_raw_inode ( struct ext4_iloc * iloc )
2006-10-11 12:20:50 +04:00
{
2006-10-11 12:20:53 +04:00
return ( struct ext4_inode * ) ( iloc - > bh - > b_data + iloc - > offset ) ;
2006-10-11 12:20:50 +04:00
}
/*
* This structure is stuffed into the struct file ' s private_data field
* for directories . It is where we put information so that we can do
* readdir operations in hash tree order .
*/
struct dir_private_info {
struct rb_root root ;
struct rb_node * curr_node ;
struct fname * extra_fname ;
loff_t last_pos ;
__u32 curr_hash ;
__u32 curr_minor_hash ;
__u32 next_hash ;
} ;
/* calculate the first block number of the group */
2006-10-11 12:20:53 +04:00
static inline ext4_fsblk_t
2008-01-29 07:58:27 +03:00
ext4_group_first_block_no ( struct super_block * sb , ext4_group_t group_no )
2006-10-11 12:20:50 +04:00
{
2006-10-11 12:20:53 +04:00
return group_no * ( ext4_fsblk_t ) EXT4_BLOCKS_PER_GROUP ( sb ) +
le32_to_cpu ( EXT4_SB ( sb ) - > s_es - > s_first_data_block ) ;
2006-10-11 12:20:50 +04:00
}
/*
* Special error return code only used by dx_probe ( ) and its callers .
*/
# define ERR_BAD_DX_DIR -75000
ext4: add support for lazy inode table initialization
When the lazy_itable_init extended option is passed to mke2fs, it
considerably speeds up filesystem creation because inode tables are
not zeroed out. The fact that parts of the inode table are
uninitialized is not a problem so long as the block group descriptors,
which contain information regarding how much of the inode table has
been initialized, has not been corrupted However, if the block group
checksums are not valid, e2fsck must scan the entire inode table, and
the the old, uninitialized data could potentially cause e2fsck to
report false problems.
Hence, it is important for the inode tables to be initialized as soon
as possble. This commit adds this feature so that mke2fs can safely
use the lazy inode table initialization feature to speed up formatting
file systems.
This is done via a new new kernel thread called ext4lazyinit, which is
created on demand and destroyed, when it is no longer needed. There
is only one thread for all ext4 filesystems in the system. When the
first filesystem with inititable mount option is mounted, ext4lazyinit
thread is created, then the filesystem can register its request in the
request list.
This thread then walks through the list of requests picking up
scheduled requests and invoking ext4_init_inode_table(). Next schedule
time for the request is computed by multiplying the time it took to
zero out last inode table with wait multiplier, which can be set with
the (init_itable=n) mount option (default is 10). We are doing
this so we do not take the whole I/O bandwidth. When the thread is no
longer necessary (request list is empty) it frees the appropriate
structures and exits (and can be created later later by another
filesystem).
We do not disturb regular inode allocations in any way, it just do not
care whether the inode table is, or is not zeroed. But when zeroing, we
have to skip used inodes, obviously. Also we should prevent new inode
allocations from the group, while zeroing is on the way. For that we
take write alloc_sem lock in ext4_init_inode_table() and read alloc_sem
in the ext4_claim_inode, so when we are unlucky and allocator hits the
group which is currently being zeroed, it just has to wait.
This can be suppresed using the mount option no_init_itable.
Signed-off-by: Lukas Czerner <lczerner@redhat.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2010-10-28 05:30:05 +04:00
/*
* Timeout and state flag for lazy initialization inode thread .
*/
# define EXT4_DEF_LI_WAIT_MULT 10
# define EXT4_DEF_LI_MAX_START_DELAY 5
# define EXT4_LAZYINIT_QUIT 0x0001
# define EXT4_LAZYINIT_RUNNING 0x0002
/*
* Lazy inode table initialization info
*/
struct ext4_lazy_init {
unsigned long li_state ;
struct list_head li_request_list ;
struct mutex li_list_mtx ;
} ;
struct ext4_li_request {
struct super_block * lr_super ;
struct ext4_sb_info * lr_sbi ;
ext4_group_t lr_next_group ;
struct list_head lr_request ;
unsigned long lr_next_sched ;
unsigned long lr_timeout ;
} ;
2010-10-28 05:30:05 +04:00
struct ext4_features {
struct kobject f_kobj ;
struct completion f_kobj_unregister ;
} ;
2011-05-25 02:31:25 +04:00
/*
* This structure will be used for multiple mount protection . It will be
* written into the block number saved in the s_mmp_block field in the
* superblock . Programs that check MMP should assume that if
* SEQ_FSCK ( or any unknown code above SEQ_MAX ) is present then it is NOT safe
* to use the filesystem , regardless of how old the timestamp is .
*/
# define EXT4_MMP_MAGIC 0x004D4D50U /* ASCII for MMP */
# define EXT4_MMP_SEQ_CLEAN 0xFF4D4D50U /* mmp_seq value for clean unmount */
# define EXT4_MMP_SEQ_FSCK 0xE24D4D50U /* mmp_seq value when being fscked */
# define EXT4_MMP_SEQ_MAX 0xE24D4D4FU /* maximum valid mmp_seq value */
struct mmp_struct {
__le32 mmp_magic ; /* Magic number for MMP */
__le32 mmp_seq ; /* Sequence no. updated periodically */
/*
* mmp_time , mmp_nodename & mmp_bdevname are only used for information
* purposes and do not affect the correctness of the algorithm
*/
__le64 mmp_time ; /* Time last updated */
char mmp_nodename [ 64 ] ; /* Node which last updated MMP block */
char mmp_bdevname [ 32 ] ; /* Bdev which last updated MMP block */
/*
* mmp_check_interval is used to verify if the MMP block has been
* updated on the block device . The value is updated based on the
* maximum time to write the MMP block during an update cycle .
*/
__le16 mmp_check_interval ;
__le16 mmp_pad1 ;
2012-04-30 02:23:10 +04:00
__le32 mmp_pad2 [ 226 ] ;
__le32 mmp_checksum ; /* crc32c(uuid+mmp_block) */
2011-05-25 02:31:25 +04:00
} ;
/* arguments passed to the mmp thread */
struct mmpd_data {
struct buffer_head * bh ; /* bh from initial read_mmp_block() */
struct super_block * sb ; /* super block of the fs */
} ;
/*
* Check interval multiplier
* The MMP block is written every update interval and initially checked every
* update interval x the multiplier ( the value is then adapted based on the
* write latency ) . The reason is that writes can be delayed under load and we
* don ' t want readers to incorrectly assume that the filesystem is no longer
* in use .
*/
# define EXT4_MMP_CHECK_MULT 2UL
/*
* Minimum interval for MMP checking in seconds .
*/
# define EXT4_MMP_MIN_CHECK_INTERVAL 5UL
/*
* Maximum interval for MMP checking in seconds .
*/
# define EXT4_MMP_MAX_CHECK_INTERVAL 300UL
2006-10-11 12:20:50 +04:00
/*
* Function prototypes
*/
/*
* Ok , these declarations are also in < linux / kernel . h > but none of the
2006-10-11 12:20:53 +04:00
* ext4 source programs needs to include it so they are duplicated here .
2006-10-11 12:20:50 +04:00
*/
2007-05-24 21:04:54 +04:00
# define NORET_TYPE /**/
# define ATTRIB_NORET __attribute__((noreturn))
# define NORET_AND noreturn,
2006-10-11 12:20:50 +04:00
2008-11-05 08:14:04 +03:00
/* bitmap.c */
2012-07-01 03:14:57 +04:00
extern unsigned int ext4_count_free ( char * bitmap , unsigned numchars ) ;
2012-04-30 02:33:10 +04:00
void ext4_inode_bitmap_csum_set ( struct super_block * sb , ext4_group_t group ,
struct ext4_group_desc * gdp ,
struct buffer_head * bh , int sz ) ;
int ext4_inode_bitmap_csum_verify ( struct super_block * sb , ext4_group_t group ,
struct ext4_group_desc * gdp ,
struct buffer_head * bh , int sz ) ;
2012-04-30 02:35:10 +04:00
void ext4_block_bitmap_csum_set ( struct super_block * sb , ext4_group_t group ,
struct ext4_group_desc * gdp ,
2012-10-22 08:34:32 +04:00
struct buffer_head * bh ) ;
2012-04-30 02:35:10 +04:00
int ext4_block_bitmap_csum_verify ( struct super_block * sb , ext4_group_t group ,
struct ext4_group_desc * gdp ,
2012-10-22 08:34:32 +04:00
struct buffer_head * bh ) ;
2008-11-05 08:14:04 +03:00
2006-10-11 12:20:50 +04:00
/* balloc.c */
2013-04-04 07:32:34 +04:00
extern void ext4_get_group_no_and_offset ( struct super_block * sb ,
ext4_fsblk_t blocknr ,
ext4_group_t * blockgrpp ,
ext4_grpblk_t * offsetp ) ;
extern ext4_group_t ext4_get_group_number ( struct super_block * sb ,
ext4_fsblk_t block ) ;
2012-04-30 02:35:10 +04:00
extern void ext4_validate_block_bitmap ( struct super_block * sb ,
struct ext4_group_desc * desc ,
2013-08-28 23:59:51 +04:00
ext4_group_t block_group ,
2012-04-30 02:35:10 +04:00
struct buffer_head * bh ) ;
2006-10-11 12:21:05 +04:00
extern unsigned int ext4_block_group ( struct super_block * sb ,
ext4_fsblk_t blocknr ) ;
extern ext4_grpblk_t ext4_block_group_offset ( struct super_block * sb ,
ext4_fsblk_t blocknr ) ;
2008-01-29 07:58:27 +03:00
extern int ext4_bg_has_super ( struct super_block * sb , ext4_group_t group ) ;
extern unsigned long ext4_bg_num_gdb ( struct super_block * sb ,
ext4_group_t group ) ;
2008-07-12 03:27:31 +04:00
extern ext4_fsblk_t ext4_new_meta_blocks ( handle_t * handle , struct inode * inode ,
2011-05-25 15:41:26 +04:00
ext4_fsblk_t goal ,
unsigned int flags ,
unsigned long * count ,
int * errp ) ;
2011-09-10 03:14:51 +04:00
extern int ext4_claim_free_clusters ( struct ext4_sb_info * sbi ,
s64 nclusters , unsigned int flags ) ;
2011-09-10 03:10:51 +04:00
extern ext4_fsblk_t ext4_count_free_clusters ( struct super_block * ) ;
2008-09-09 06:25:24 +04:00
extern void ext4_check_blocks_bitmap ( struct super_block * ) ;
2006-10-11 12:20:53 +04:00
extern struct ext4_group_desc * ext4_get_group_desc ( struct super_block * sb ,
2008-01-29 07:58:27 +03:00
ext4_group_t block_group ,
2006-10-11 12:20:50 +04:00
struct buffer_head * * bh ) ;
2006-10-11 12:20:53 +04:00
extern int ext4_should_retry_alloc ( struct super_block * sb , int * retries ) ;
2012-02-21 02:52:46 +04:00
extern struct buffer_head * ext4_read_block_bitmap_nowait ( struct super_block * sb ,
ext4_group_t block_group ) ;
extern int ext4_wait_block_bitmap ( struct super_block * sb ,
ext4_group_t block_group ,
struct buffer_head * bh ) ;
extern struct buffer_head * ext4_read_block_bitmap ( struct super_block * sb ,
ext4_group_t block_group ) ;
2011-09-10 02:42:51 +04:00
extern void ext4_init_block_bitmap ( struct super_block * sb ,
struct buffer_head * bh ,
ext4_group_t group ,
struct ext4_group_desc * desc ) ;
2011-09-10 03:12:51 +04:00
extern unsigned ext4_free_clusters_after_init ( struct super_block * sb ,
ext4_group_t block_group ,
struct ext4_group_desc * gdp ) ;
2011-09-10 02:44:51 +04:00
extern unsigned ext4_num_overhead_clusters ( struct super_block * sb ,
ext4_group_t block_group ,
struct ext4_group_desc * gdp ) ;
2011-06-28 18:01:31 +04:00
ext4_fsblk_t ext4_inode_to_goal_block ( struct inode * ) ;
2006-10-11 12:20:50 +04:00
/* dir.c */
2010-07-27 19:54:40 +04:00
extern int __ext4_check_dir_entry ( const char * , unsigned int , struct inode * ,
2011-01-10 20:10:55 +03:00
struct file * ,
2010-07-27 19:54:40 +04:00
struct ext4_dir_entry_2 * ,
2012-12-10 23:05:58 +04:00
struct buffer_head * , char * , int ,
unsigned int ) ;
# define ext4_check_dir_entry(dir, filp, de, bh, buf, size, offset) \
2011-01-10 20:10:55 +03:00
unlikely ( __ext4_check_dir_entry ( __func__ , __LINE__ , ( dir ) , ( filp ) , \
2012-12-10 23:05:58 +04:00
( de ) , ( bh ) , ( buf ) , ( size ) , ( offset ) ) )
2006-10-11 12:20:53 +04:00
extern int ext4_htree_store_dirent ( struct file * dir_file , __u32 hash ,
2006-10-11 12:20:50 +04:00
__u32 minor_hash ,
2006-10-11 12:20:53 +04:00
struct ext4_dir_entry_2 * dirent ) ;
extern void ext4_htree_free_dir_info ( struct dir_private_info * p ) ;
2012-12-10 23:05:58 +04:00
extern int ext4_find_dest_de ( struct inode * dir , struct inode * inode ,
struct buffer_head * bh ,
void * buf , int buf_size ,
const char * name , int namelen ,
struct ext4_dir_entry_2 * * dest_de ) ;
void ext4_insert_dentry ( struct inode * inode ,
struct ext4_dir_entry_2 * de ,
int buf_size ,
const char * name , int namelen ) ;
static inline void ext4_update_dx_flag ( struct inode * inode )
{
if ( ! EXT4_HAS_COMPAT_FEATURE ( inode - > i_sb ,
EXT4_FEATURE_COMPAT_DIR_INDEX ) )
ext4_clear_inode_flag ( inode , EXT4_INODE_INDEX ) ;
}
2012-12-10 23:05:59 +04:00
static unsigned char ext4_filetype_table [ ] = {
DT_UNKNOWN , DT_REG , DT_DIR , DT_CHR , DT_BLK , DT_FIFO , DT_SOCK , DT_LNK
} ;
static inline unsigned char get_dtype ( struct super_block * sb , int filetype )
{
if ( ! EXT4_HAS_INCOMPAT_FEATURE ( sb , EXT4_FEATURE_INCOMPAT_FILETYPE ) | |
( filetype > = EXT4_FT_MAX ) )
return DT_UNKNOWN ;
return ext4_filetype_table [ filetype ] ;
}
2006-10-11 12:20:50 +04:00
/* fsync.c */
2011-07-17 04:44:56 +04:00
extern int ext4_sync_file ( struct file * , loff_t , loff_t , int ) ;
2006-10-11 12:20:50 +04:00
/* hash.c */
2006-10-11 12:20:53 +04:00
extern int ext4fs_dirhash ( const char * name , int len , struct
2006-10-11 12:20:50 +04:00
dx_hash_info * hinfo ) ;
/* ialloc.c */
2013-02-10 01:27:09 +04:00
extern struct inode * __ext4_new_inode ( handle_t * , struct inode * , umode_t ,
const struct qstr * qstr , __u32 goal ,
uid_t * owner , int handle_type ,
unsigned int line_no , int nblocks ) ;
# define ext4_new_inode(handle, dir, mode, qstr, goal, owner) \
__ext4_new_inode ( ( handle ) , ( dir ) , ( mode ) , ( qstr ) , ( goal ) , ( owner ) , \
0 , 0 , 0 )
# define ext4_new_inode_start_handle(dir, mode, qstr, goal, owner, \
type , nblocks ) \
__ext4_new_inode ( NULL , ( dir ) , ( mode ) , ( qstr ) , ( goal ) , ( owner ) , \
( type ) , __LINE__ , ( nblocks ) )
2008-09-09 06:25:24 +04:00
extern void ext4_free_inode ( handle_t * , struct inode * ) ;
extern struct inode * ext4_orphan_get ( struct super_block * , unsigned long ) ;
extern unsigned long ext4_count_free_inodes ( struct super_block * ) ;
extern unsigned long ext4_count_dirs ( struct super_block * ) ;
extern void ext4_check_inodes_bitmap ( struct super_block * ) ;
2010-10-28 05:30:15 +04:00
extern void ext4_mark_bitmap_end ( int start_bit , int end_bit , char * bitmap ) ;
ext4: add support for lazy inode table initialization
When the lazy_itable_init extended option is passed to mke2fs, it
considerably speeds up filesystem creation because inode tables are
not zeroed out. The fact that parts of the inode table are
uninitialized is not a problem so long as the block group descriptors,
which contain information regarding how much of the inode table has
been initialized, has not been corrupted However, if the block group
checksums are not valid, e2fsck must scan the entire inode table, and
the the old, uninitialized data could potentially cause e2fsck to
report false problems.
Hence, it is important for the inode tables to be initialized as soon
as possble. This commit adds this feature so that mke2fs can safely
use the lazy inode table initialization feature to speed up formatting
file systems.
This is done via a new new kernel thread called ext4lazyinit, which is
created on demand and destroyed, when it is no longer needed. There
is only one thread for all ext4 filesystems in the system. When the
first filesystem with inititable mount option is mounted, ext4lazyinit
thread is created, then the filesystem can register its request in the
request list.
This thread then walks through the list of requests picking up
scheduled requests and invoking ext4_init_inode_table(). Next schedule
time for the request is computed by multiplying the time it took to
zero out last inode table with wait multiplier, which can be set with
the (init_itable=n) mount option (default is 10). We are doing
this so we do not take the whole I/O bandwidth. When the thread is no
longer necessary (request list is empty) it frees the appropriate
structures and exits (and can be created later later by another
filesystem).
We do not disturb regular inode allocations in any way, it just do not
care whether the inode table is, or is not zeroed. But when zeroing, we
have to skip used inodes, obviously. Also we should prevent new inode
allocations from the group, while zeroing is on the way. For that we
take write alloc_sem lock in ext4_init_inode_table() and read alloc_sem
in the ext4_claim_inode, so when we are unlucky and allocator hits the
group which is currently being zeroed, it just has to wait.
This can be suppresed using the mount option no_init_itable.
Signed-off-by: Lukas Czerner <lczerner@redhat.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2010-10-28 05:30:05 +04:00
extern int ext4_init_inode_table ( struct super_block * sb ,
ext4_group_t group , int barrier ) ;
2012-02-21 02:52:46 +04:00
extern void ext4_end_bitmap_read ( struct buffer_head * bh , int uptodate ) ;
2006-10-11 12:20:50 +04:00
2008-01-29 08:19:52 +03:00
/* mballoc.c */
extern long ext4_mb_stats ;
extern long ext4_mb_max_to_scan ;
2012-05-28 22:19:25 +04:00
extern int ext4_mb_init ( struct super_block * ) ;
2008-01-29 08:19:52 +03:00
extern int ext4_mb_release ( struct super_block * ) ;
extern ext4_fsblk_t ext4_mb_new_blocks ( handle_t * ,
struct ext4_allocation_request * , int * ) ;
extern int ext4_mb_reserve_blocks ( struct super_block * , int ) ;
2008-10-10 17:40:52 +04:00
extern void ext4_discard_preallocations ( struct inode * ) ;
2010-10-28 05:30:14 +04:00
extern int __init ext4_init_mballoc ( void ) ;
extern void ext4_exit_mballoc ( void ) ;
2009-11-22 15:44:56 +03:00
extern void ext4_free_blocks ( handle_t * handle , struct inode * inode ,
2009-11-23 15:17:05 +03:00
struct buffer_head * bh , ext4_fsblk_t block ,
unsigned long count , int flags ) ;
2012-09-05 09:31:50 +04:00
extern int ext4_mb_alloc_groupinfo ( struct super_block * sb ,
ext4_group_t ngroups ) ;
2009-01-06 05:36:19 +03:00
extern int ext4_mb_add_groupinfo ( struct super_block * sb ,
2008-07-12 03:27:31 +04:00
ext4_group_t i , struct ext4_group_desc * desc ) ;
2011-07-27 05:46:07 +04:00
extern int ext4_group_add_blocks ( handle_t * handle , struct super_block * sb ,
2011-05-09 18:46:41 +04:00
ext4_fsblk_t block , unsigned long count ) ;
2010-10-28 05:30:12 +04:00
extern int ext4_trim_fs ( struct super_block * , struct fstrim_range * ) ;
2006-10-11 12:20:50 +04:00
/* inode.c */
2008-01-29 07:58:27 +03:00
struct buffer_head * ext4_getblk ( handle_t * , struct inode * ,
ext4_lblk_t , int , int * ) ;
struct buffer_head * ext4_bread ( handle_t * , struct inode * ,
ext4_lblk_t , int , int * ) ;
2012-12-10 23:05:51 +04:00
int ext4_get_block_write ( struct inode * inode , sector_t iblock ,
struct buffer_head * bh_result , int create ) ;
2008-10-07 08:46:36 +04:00
int ext4_get_block ( struct inode * inode , sector_t iblock ,
struct buffer_head * bh_result , int create ) ;
2012-12-10 23:05:57 +04:00
int ext4_da_get_block_prep ( struct inode * inode , sector_t iblock ,
struct buffer_head * bh , int create ) ;
2012-12-10 23:05:51 +04:00
int ext4_walk_page_buffers ( handle_t * handle ,
struct buffer_head * head ,
unsigned from ,
unsigned to ,
int * partial ,
int ( * fn ) ( handle_t * handle ,
struct buffer_head * bh ) ) ;
int do_journal_get_write_access ( handle_t * handle ,
struct buffer_head * bh ) ;
2012-12-10 23:05:57 +04:00
# define FALL_BACK_TO_NONDELALLOC 1
# define CONVERT_INLINE_DATA 2
2006-10-11 12:20:50 +04:00
2008-02-07 11:15:37 +03:00
extern struct inode * ext4_iget ( struct super_block * , unsigned long ) ;
2010-03-05 11:21:37 +03:00
extern int ext4_write_inode ( struct inode * , struct writeback_control * ) ;
2008-09-09 06:25:24 +04:00
extern int ext4_setattr ( struct dentry * , struct iattr * ) ;
2008-07-12 03:27:31 +04:00
extern int ext4_getattr ( struct vfsmount * mnt , struct dentry * dentry ,
struct kstat * stat ) ;
2010-06-07 21:16:22 +04:00
extern void ext4_evict_inode ( struct inode * ) ;
extern void ext4_clear_inode ( struct inode * ) ;
2008-09-09 06:25:24 +04:00
extern int ext4_sync_inode ( handle_t * , struct inode * ) ;
2011-05-27 14:53:02 +04:00
extern void ext4_dirty_inode ( struct inode * , int ) ;
2006-10-11 12:20:53 +04:00
extern int ext4_change_inode_journal_flag ( struct inode * , int ) ;
extern int ext4_get_inode_loc ( struct inode * , struct ext4_iloc * ) ;
2013-08-17 05:19:41 +04:00
extern int ext4_inode_attach_jinode ( struct inode * inode ) ;
2008-07-12 03:27:31 +04:00
extern int ext4_can_truncate ( struct inode * inode ) ;
2008-09-09 06:25:24 +04:00
extern void ext4_truncate ( struct inode * ) ;
2013-07-01 16:12:38 +04:00
extern int ext4_punch_hole ( struct inode * inode , loff_t offset , loff_t length ) ;
2009-08-18 06:17:20 +04:00
extern int ext4_truncate_restart_trans ( handle_t * , struct inode * , int nblocks ) ;
2006-10-11 12:20:53 +04:00
extern void ext4_set_inode_flags ( struct inode * ) ;
2007-07-18 17:24:20 +04:00
extern void ext4_get_inode_flags ( struct ext4_inode_info * ) ;
2009-02-26 09:04:07 +03:00
extern int ext4_alloc_da_blocks ( struct inode * inode ) ;
2006-10-11 12:20:53 +04:00
extern void ext4_set_aops ( struct inode * inode ) ;
2006-10-11 12:21:03 +04:00
extern int ext4_writepage_trans_blocks ( struct inode * ) ;
2008-08-20 06:16:03 +04:00
extern int ext4_chunk_trans_blocks ( struct inode * , int nrblocks ) ;
2013-05-28 07:32:35 +04:00
extern int ext4_block_truncate_page ( handle_t * handle ,
struct address_space * mapping , loff_t from ) ;
2013-05-28 07:32:35 +04:00
extern int ext4_zero_partial_blocks ( handle_t * handle , struct inode * inode ,
loff_t lstart , loff_t lend ) ;
2009-04-01 02:23:21 +04:00
extern int ext4_page_mkwrite ( struct vm_area_struct * vma , struct vm_fault * vmf ) ;
2009-12-14 15:21:14 +03:00
extern qsize_t * ext4_get_reserved_space ( struct inode * inode ) ;
2010-01-25 12:00:31 +03:00
extern void ext4_da_update_reserve_space ( struct inode * inode ,
int used , int quota_claim ) ;
2011-06-28 03:40:50 +04:00
/* indirect.c */
extern int ext4_ind_map_blocks ( handle_t * handle , struct inode * inode ,
struct ext4_map_blocks * map , int flags ) ;
extern ssize_t ext4_ind_direct_IO ( int rw , struct kiocb * iocb ,
const struct iovec * iov , loff_t offset ,
unsigned long nr_segs ) ;
extern int ext4_ind_calc_metadata_amount ( struct inode * inode , sector_t lblock ) ;
2013-06-04 20:56:55 +04:00
extern int ext4_ind_trans_blocks ( struct inode * inode , int nrblocks ) ;
2013-04-03 20:47:17 +04:00
extern void ext4_ind_truncate ( handle_t * , struct inode * inode ) ;
2013-04-03 20:45:17 +04:00
extern int ext4_free_hole_blocks ( handle_t * handle , struct inode * inode ,
ext4_lblk_t first , ext4_lblk_t stop ) ;
2011-06-28 00:36:31 +04:00
2006-10-11 12:20:50 +04:00
/* ioctl.c */
2008-04-30 06:03:54 +04:00
extern long ext4_ioctl ( struct file * , unsigned int , unsigned long ) ;
2008-09-09 06:25:24 +04:00
extern long ext4_compat_ioctl ( struct file * , unsigned int , unsigned long ) ;
2006-10-11 12:20:50 +04:00
2008-01-29 07:58:26 +03:00
/* migrate.c */
2008-09-13 20:52:26 +04:00
extern int ext4_ext_migrate ( struct inode * ) ;
2013-04-11 07:32:52 +04:00
extern int ext4_ind_migrate ( struct inode * inode ) ;
2009-02-15 07:01:36 +03:00
2006-10-11 12:20:50 +04:00
/* namei.c */
2012-04-30 02:41:10 +04:00
extern int ext4_dirent_csum_verify ( struct inode * inode ,
struct ext4_dir_entry * dirent ) ;
2006-10-11 12:20:53 +04:00
extern int ext4_orphan_add ( handle_t * , struct inode * ) ;
extern int ext4_orphan_del ( handle_t * , struct inode * ) ;
extern int ext4_htree_fill_tree ( struct file * dir_file , __u32 start_hash ,
2006-10-11 12:20:50 +04:00
__u32 start_minor_hash , __u32 * next_hash ) ;
2012-12-10 23:05:59 +04:00
extern int search_dir ( struct buffer_head * bh ,
char * search_buf ,
int buf_size ,
struct inode * dir ,
const struct qstr * d_name ,
unsigned int offset ,
struct ext4_dir_entry_2 * * res_dir ) ;
2012-12-10 23:06:00 +04:00
extern int ext4_generic_delete_entry ( handle_t * handle ,
struct inode * dir ,
struct ext4_dir_entry_2 * de_del ,
struct buffer_head * bh ,
void * entry_buf ,
int buf_size ,
int csum_size ) ;
2006-10-11 12:20:50 +04:00
/* resize.c */
2006-10-11 12:20:53 +04:00
extern int ext4_group_add ( struct super_block * sb ,
struct ext4_new_group_data * input ) ;
extern int ext4_group_extend ( struct super_block * sb ,
struct ext4_super_block * es ,
ext4_fsblk_t n_blocks_count ) ;
2012-01-05 02:09:44 +04:00
extern int ext4_resize_fs ( struct super_block * sb , ext4_fsblk_t n_blocks_count ) ;
2006-10-11 12:20:50 +04:00
/* super.c */
2012-07-10 00:27:05 +04:00
extern int ext4_calculate_overhead ( struct super_block * sb ) ;
2012-04-30 02:29:10 +04:00
extern int ext4_superblock_csum_verify ( struct super_block * sb ,
struct ext4_super_block * es ) ;
2012-10-10 09:06:58 +04:00
extern void ext4_superblock_csum_set ( struct super_block * sb ) ;
2011-08-01 16:45:02 +04:00
extern void * ext4_kvmalloc ( size_t size , gfp_t flags ) ;
extern void * ext4_kvzalloc ( size_t size , gfp_t flags ) ;
extern void ext4_kvfree ( void * ptr ) ;
2012-09-05 09:29:50 +04:00
extern int ext4_alloc_flex_bg_array ( struct super_block * sb ,
ext4_group_t ngroup ) ;
2013-02-08 22:00:31 +04:00
extern const char * ext4_decode_error ( struct super_block * sb , int errno ,
char nbuf [ 16 ] ) ;
2013-07-01 16:12:37 +04:00
2011-11-01 04:11:33 +04:00
extern __printf ( 4 , 5 )
void __ext4_error ( struct super_block * , const char * , unsigned int ,
const char * , . . . ) ;
extern __printf ( 5 , 6 )
2013-07-01 16:12:37 +04:00
void __ext4_error_inode ( struct inode * , const char * , unsigned int , ext4_fsblk_t ,
2011-11-01 04:11:33 +04:00
const char * , . . . ) ;
extern __printf ( 5 , 6 )
2013-07-01 16:12:37 +04:00
void __ext4_error_file ( struct file * , const char * , unsigned int , ext4_fsblk_t ,
2011-11-01 04:11:33 +04:00
const char * , . . . ) ;
2010-07-27 19:56:40 +04:00
extern void __ext4_std_error ( struct super_block * , const char * ,
unsigned int , int ) ;
2011-11-01 04:11:33 +04:00
extern __printf ( 4 , 5 )
void __ext4_abort ( struct super_block * , const char * , unsigned int ,
const char * , . . . ) ;
extern __printf ( 4 , 5 )
void __ext4_warning ( struct super_block * , const char * , unsigned int ,
const char * , . . . ) ;
extern __printf ( 3 , 4 )
2013-07-01 16:12:37 +04:00
void __ext4_msg ( struct super_block * , const char * , const char * , . . . ) ;
2011-05-25 02:31:25 +04:00
extern void __dump_mmp_msg ( struct super_block * , struct mmp_struct * mmp ,
const char * , unsigned int , const char * ) ;
2011-11-01 04:11:33 +04:00
extern __printf ( 7 , 8 )
void __ext4_grp_locked_error ( const char * , unsigned int ,
struct super_block * , ext4_group_t ,
unsigned long , ext4_fsblk_t ,
const char * , . . . ) ;
2013-07-01 16:12:37 +04:00
# ifdef CONFIG_PRINTK
# define ext4_error_inode(inode, func, line, block, fmt, ...) \
__ext4_error_inode ( inode , func , line , block , fmt , # # __VA_ARGS__ )
# define ext4_error_file(file, func, line, block, fmt, ...) \
__ext4_error_file ( file , func , line , block , fmt , # # __VA_ARGS__ )
# define ext4_error(sb, fmt, ...) \
__ext4_error ( sb , __func__ , __LINE__ , fmt , # # __VA_ARGS__ )
# define ext4_abort(sb, fmt, ...) \
__ext4_abort ( sb , __func__ , __LINE__ , fmt , # # __VA_ARGS__ )
# define ext4_warning(sb, fmt, ...) \
__ext4_warning ( sb , __func__ , __LINE__ , fmt , # # __VA_ARGS__ )
# define ext4_msg(sb, level, fmt, ...) \
__ext4_msg ( sb , level , fmt , # # __VA_ARGS__ )
# define dump_mmp_msg(sb, mmp, msg) \
__dump_mmp_msg ( sb , mmp , __func__ , __LINE__ , msg )
# define ext4_grp_locked_error(sb, grp, ino, block, fmt, ...) \
__ext4_grp_locked_error ( __func__ , __LINE__ , sb , grp , ino , block , \
fmt , # # __VA_ARGS__ )
# else
# define ext4_error_inode(inode, func, line, block, fmt, ...) \
do { \
no_printk ( fmt , # # __VA_ARGS__ ) ; \
__ext4_error_inode ( inode , " " , 0 , block , " " ) ; \
} while ( 0 )
# define ext4_error_file(file, func, line, block, fmt, ...) \
do { \
no_printk ( fmt , # # __VA_ARGS__ ) ; \
__ext4_error_file ( file , " " , 0 , block , " " ) ; \
} while ( 0 )
# define ext4_error(sb, fmt, ...) \
do { \
no_printk ( fmt , # # __VA_ARGS__ ) ; \
__ext4_error ( sb , " " , 0 , " " ) ; \
} while ( 0 )
# define ext4_abort(sb, fmt, ...) \
do { \
no_printk ( fmt , # # __VA_ARGS__ ) ; \
__ext4_abort ( sb , " " , 0 , " " ) ; \
} while ( 0 )
# define ext4_warning(sb, fmt, ...) \
do { \
no_printk ( fmt , # # __VA_ARGS__ ) ; \
__ext4_warning ( sb , " " , 0 , " " ) ; \
} while ( 0 )
# define ext4_msg(sb, level, fmt, ...) \
do { \
no_printk ( fmt , # # __VA_ARGS__ ) ; \
__ext4_msg ( sb , " " , " " ) ; \
} while ( 0 )
# define dump_mmp_msg(sb, mmp, msg) \
__dump_mmp_msg ( sb , mmp , " " , 0 , " " )
# define ext4_grp_locked_error(sb, grp, ino, block, fmt, ...) \
do { \
no_printk ( fmt , # # __VA_ARGS__ ) ; \
__ext4_grp_locked_error ( " " , 0 , sb , grp , ino , block , " " ) ; \
} while ( 0 )
# endif
2008-09-09 06:25:24 +04:00
extern void ext4_update_dynamic_rev ( struct super_block * sb ) ;
2008-01-29 07:58:27 +03:00
extern int ext4_update_compat_feature ( handle_t * handle , struct super_block * sb ,
__u32 compat ) ;
extern int ext4_update_rocompat_feature ( handle_t * handle ,
struct super_block * sb , __u32 rocompat ) ;
extern int ext4_update_incompat_feature ( handle_t * handle ,
struct super_block * sb , __u32 incompat ) ;
2006-10-11 12:21:15 +04:00
extern ext4_fsblk_t ext4_block_bitmap ( struct super_block * sb ,
struct ext4_group_desc * bg ) ;
extern ext4_fsblk_t ext4_inode_bitmap ( struct super_block * sb ,
struct ext4_group_desc * bg ) ;
extern ext4_fsblk_t ext4_inode_table ( struct super_block * sb ,
struct ext4_group_desc * bg ) ;
2011-09-10 03:08:51 +04:00
extern __u32 ext4_free_group_clusters ( struct super_block * sb ,
struct ext4_group_desc * bg ) ;
2009-01-06 06:20:24 +03:00
extern __u32 ext4_free_inodes_count ( struct super_block * sb ,
struct ext4_group_desc * bg ) ;
extern __u32 ext4_used_dirs_count ( struct super_block * sb ,
struct ext4_group_desc * bg ) ;
extern __u32 ext4_itable_unused_count ( struct super_block * sb ,
struct ext4_group_desc * bg ) ;
2006-10-11 12:21:15 +04:00
extern void ext4_block_bitmap_set ( struct super_block * sb ,
struct ext4_group_desc * bg , ext4_fsblk_t blk ) ;
extern void ext4_inode_bitmap_set ( struct super_block * sb ,
struct ext4_group_desc * bg , ext4_fsblk_t blk ) ;
extern void ext4_inode_table_set ( struct super_block * sb ,
struct ext4_group_desc * bg , ext4_fsblk_t blk ) ;
2011-09-10 03:08:51 +04:00
extern void ext4_free_group_clusters_set ( struct super_block * sb ,
struct ext4_group_desc * bg ,
__u32 count ) ;
2009-01-06 06:20:24 +03:00
extern void ext4_free_inodes_set ( struct super_block * sb ,
struct ext4_group_desc * bg , __u32 count ) ;
extern void ext4_used_dirs_set ( struct super_block * sb ,
struct ext4_group_desc * bg , __u32 count ) ;
extern void ext4_itable_unused_set ( struct super_block * sb ,
struct ext4_group_desc * bg , __u32 count ) ;
2012-04-30 02:45:10 +04:00
extern int ext4_group_desc_csum_verify ( struct super_block * sb , __u32 group ,
2009-05-02 03:44:44 +04:00
struct ext4_group_desc * gdp ) ;
2012-04-30 02:45:10 +04:00
extern void ext4_group_desc_csum_set ( struct super_block * sb , __u32 group ,
struct ext4_group_desc * gdp ) ;
2013-01-13 17:41:45 +04:00
extern int ext4_register_li_request ( struct super_block * sb ,
ext4_group_t first_not_zeroed ) ;
2006-10-11 12:21:10 +04:00
2012-04-30 02:33:10 +04:00
static inline int ext4_has_group_desc_csum ( struct super_block * sb )
{
return EXT4_HAS_RO_COMPAT_FEATURE ( sb ,
EXT4_FEATURE_RO_COMPAT_GDT_CSUM |
EXT4_FEATURE_RO_COMPAT_METADATA_CSUM ) ;
}
2006-10-11 12:21:10 +04:00
static inline ext4_fsblk_t ext4_blocks_count ( struct ext4_super_block * es )
{
return ( ( ext4_fsblk_t ) le32_to_cpu ( es - > s_blocks_count_hi ) < < 32 ) |
2007-10-17 02:38:25 +04:00
le32_to_cpu ( es - > s_blocks_count_lo ) ;
2006-10-11 12:21:10 +04:00
}
static inline ext4_fsblk_t ext4_r_blocks_count ( struct ext4_super_block * es )
{
return ( ( ext4_fsblk_t ) le32_to_cpu ( es - > s_r_blocks_count_hi ) < < 32 ) |
2007-10-17 02:38:25 +04:00
le32_to_cpu ( es - > s_r_blocks_count_lo ) ;
2006-10-11 12:21:10 +04:00
}
static inline ext4_fsblk_t ext4_free_blocks_count ( struct ext4_super_block * es )
{
return ( ( ext4_fsblk_t ) le32_to_cpu ( es - > s_free_blocks_count_hi ) < < 32 ) |
2007-10-17 02:38:25 +04:00
le32_to_cpu ( es - > s_free_blocks_count_lo ) ;
2006-10-11 12:21:10 +04:00
}
static inline void ext4_blocks_count_set ( struct ext4_super_block * es ,
ext4_fsblk_t blk )
{
2007-10-17 02:38:25 +04:00
es - > s_blocks_count_lo = cpu_to_le32 ( ( u32 ) blk ) ;
2006-10-11 12:21:10 +04:00
es - > s_blocks_count_hi = cpu_to_le32 ( blk > > 32 ) ;
}
static inline void ext4_free_blocks_count_set ( struct ext4_super_block * es ,
ext4_fsblk_t blk )
{
2007-10-17 02:38:25 +04:00
es - > s_free_blocks_count_lo = cpu_to_le32 ( ( u32 ) blk ) ;
2006-10-11 12:21:10 +04:00
es - > s_free_blocks_count_hi = cpu_to_le32 ( blk > > 32 ) ;
}
static inline void ext4_r_blocks_count_set ( struct ext4_super_block * es ,
ext4_fsblk_t blk )
{
2007-10-17 02:38:25 +04:00
es - > s_r_blocks_count_lo = cpu_to_le32 ( ( u32 ) blk ) ;
2006-10-11 12:21:10 +04:00
es - > s_r_blocks_count_hi = cpu_to_le32 ( blk > > 32 ) ;
}
2008-01-29 07:58:27 +03:00
static inline loff_t ext4_isize ( struct ext4_inode * raw_inode )
{
2009-01-18 02:41:37 +03:00
if ( S_ISREG ( le16_to_cpu ( raw_inode - > i_mode ) ) )
return ( ( loff_t ) le32_to_cpu ( raw_inode - > i_size_high ) < < 32 ) |
le32_to_cpu ( raw_inode - > i_size_lo ) ;
else
return ( loff_t ) le32_to_cpu ( raw_inode - > i_size_lo ) ;
2008-01-29 07:58:27 +03:00
}
2006-10-11 12:21:10 +04:00
2008-01-29 07:58:27 +03:00
static inline void ext4_isize_set ( struct ext4_inode * raw_inode , loff_t i_size )
{
raw_inode - > i_size_lo = cpu_to_le32 ( i_size ) ;
raw_inode - > i_size_high = cpu_to_le32 ( i_size > > 32 ) ;
}
2006-10-11 12:20:50 +04:00
2008-01-29 08:19:52 +03:00
static inline
struct ext4_group_info * ext4_get_group_info ( struct super_block * sb ,
2008-09-09 06:25:24 +04:00
ext4_group_t group )
2008-01-29 08:19:52 +03:00
{
struct ext4_group_info * * * grp_info ;
long indexv , indexh ;
2013-06-06 19:16:43 +04:00
BUG_ON ( group > = EXT4_SB ( sb ) - > s_groups_count ) ;
2008-01-29 08:19:52 +03:00
grp_info = EXT4_SB ( sb ) - > s_group_info ;
indexv = group > > ( EXT4_DESC_PER_BLOCK_BITS ( sb ) ) ;
indexh = group & ( ( EXT4_DESC_PER_BLOCK ( sb ) ) - 1 ) ;
return grp_info [ indexv ] [ indexh ] ;
}
2009-05-01 16:50:38 +04:00
/*
* Reading s_groups_count requires using smp_rmb ( ) afterwards . See
* the locking protocol documented in the comments of ext4_group_add ( )
* in resize . c
*/
static inline ext4_group_t ext4_get_groups_count ( struct super_block * sb )
{
ext4_group_t ngroups = EXT4_SB ( sb ) - > s_groups_count ;
smp_rmb ( ) ;
return ngroups ;
}
2008-01-29 08:19:52 +03:00
2008-07-12 03:27:31 +04:00
static inline ext4_group_t ext4_flex_group ( struct ext4_sb_info * sbi ,
ext4_group_t block_group )
{
return block_group > > sbi - > s_log_groups_per_flex ;
}
static inline unsigned int ext4_flex_bg_size ( struct ext4_sb_info * sbi )
{
return 1 < < sbi - > s_log_groups_per_flex ;
}
2006-10-11 12:20:53 +04:00
# define ext4_std_error(sb, errno) \
2006-10-11 12:20:50 +04:00
do { \
if ( ( errno ) ) \
2010-07-27 19:56:40 +04:00
__ext4_std_error ( ( sb ) , __func__ , __LINE__ , ( errno ) ) ; \
2006-10-11 12:20:50 +04:00
} while ( 0 )
2008-10-09 18:56:23 +04:00
# ifdef CONFIG_SMP
2011-09-10 03:16:51 +04:00
/* Each CPU can accumulate percpu_counter_batch clusters in their local
* counters . So we need to make sure we have free clusters more
2009-01-07 01:41:04 +03:00
* than percpu_counter_batch * nr_cpu_ids . Also add a window of 4 times .
2008-10-09 18:56:23 +04:00
*/
2011-09-10 03:16:51 +04:00
# define EXT4_FREECLUSTERS_WATERMARK (4 * (percpu_counter_batch * nr_cpu_ids))
2008-10-09 18:56:23 +04:00
# else
2011-09-10 03:16:51 +04:00
# define EXT4_FREECLUSTERS_WATERMARK 0
2008-10-09 18:56:23 +04:00
# endif
2013-08-17 18:09:31 +04:00
/* Update i_disksize. Requires i_mutex to avoid races with truncate */
2008-09-13 21:06:18 +04:00
static inline void ext4_update_i_disksize ( struct inode * inode , loff_t newsize )
{
2013-08-17 18:09:31 +04:00
WARN_ON_ONCE ( S_ISREG ( inode - > i_mode ) & &
! mutex_is_locked ( & inode - > i_mutex ) ) ;
down_write ( & EXT4_I ( inode ) - > i_data_sem ) ;
if ( newsize > EXT4_I ( inode ) - > i_disksize )
EXT4_I ( inode ) - > i_disksize = newsize ;
up_write ( & EXT4_I ( inode ) - > i_data_sem ) ;
}
2009-01-06 06:19:52 +03:00
struct ext4_group_info {
unsigned long bb_state ;
struct rb_root bb_free_root ;
2009-08-26 06:36:45 +04:00
ext4_grpblk_t bb_first_free ; /* first free block */
ext4_grpblk_t bb_free ; /* total free blocks */
ext4_grpblk_t bb_fragments ; /* nr of freespace fragments */
2010-05-16 23:00:00 +04:00
ext4_grpblk_t bb_largest_free_order ; /* order of largest frag in BG */
2009-01-06 06:19:52 +03:00
struct list_head bb_prealloc_list ;
# ifdef DOUBLE_CHECK
void * bb_bitmap ;
# endif
struct rw_semaphore alloc_sem ;
2009-08-26 06:36:45 +04:00
ext4_grpblk_t bb_counters [ ] ; /* Nr of free power-of-two-block
* regions , index is order .
* bb_counters [ 3 ] = 5 means
* 5 free 8 - block regions . */
2009-01-06 06:19:52 +03:00
} ;
2011-07-11 08:03:38 +04:00
# define EXT4_GROUP_INFO_NEED_INIT_BIT 0
# define EXT4_GROUP_INFO_WAS_TRIMMED_BIT 1
ext4: mark block group as corrupt on block bitmap error
When we notice a block-bitmap corruption (because of device failure or
something else), we should mark this group as corrupt and prevent
further block allocations/deallocations from it. Currently, we end up
generating one error message for every block in the bitmap. This
potentially could make the system unstable as noticed in some
bugs. With this patch, the error will be printed only the first time
and mark the entire block group as corrupted. This prevents future
access allocations/deallocations from it.
Also tested by corrupting the block
bitmap and forcefully introducing the mb_free_blocks error:
(1) create a largefile (2Gb)
$ dd if=/dev/zero of=largefile oflag=direct bs=10485760 count=200
(2) umount filesystem. use dumpe2fs to see which block-bitmaps
are in use by largefile and note their block numbers
(3) use dd to zero-out the used block bitmaps
$ dd if=/dev/zero of=/dev/hdc4 bs=4096 seek=14 count=8 oflag=direct
(4) mount the FS and delete the largefile.
(5) recreate the largefile. verify that the new largefile does not
get any blocks from the groups marked as bad.
Without the patch, we will see mb_free_blocks error for each bit in
each zero'ed out bitmap at (4). With the patch, we only see the error
once per blockgroup:
[ 309.706803] EXT4-fs error (device sdb4): ext4_mb_generate_buddy:735: group 15: 32768 clusters in bitmap, 0 in gd. blk grp corrupted.
[ 309.720824] EXT4-fs error (device sdb4): ext4_mb_generate_buddy:735: group 14: 32768 clusters in bitmap, 0 in gd. blk grp corrupted.
[ 309.732858] EXT4-fs error (device sdb4) in ext4_free_blocks:4802: IO failure
[ 309.748321] EXT4-fs error (device sdb4): ext4_mb_generate_buddy:735: group 13: 32768 clusters in bitmap, 0 in gd. blk grp corrupted.
[ 309.760331] EXT4-fs error (device sdb4) in ext4_free_blocks:4802: IO failure
[ 309.769695] EXT4-fs error (device sdb4): ext4_mb_generate_buddy:735: group 12: 32768 clusters in bitmap, 0 in gd. blk grp corrupted.
[ 309.781721] EXT4-fs error (device sdb4) in ext4_free_blocks:4802: IO failure
[ 309.798166] EXT4-fs error (device sdb4): ext4_mb_generate_buddy:735: group 11: 32768 clusters in bitmap, 0 in gd. blk grp corrupted.
[ 309.810184] EXT4-fs error (device sdb4) in ext4_free_blocks:4802: IO failure
[ 309.819532] EXT4-fs error (device sdb4): ext4_mb_generate_buddy:735: group 10: 32768 clusters in bitmap, 0 in gd. blk grp corrupted.
Google-Bug-Id: 7258357
[darrick.wong@oracle.com]
Further modifications (by Darrick) to make more obvious that this corruption
bit applies to blocks only. Set the corruption flag if the block group bitmap
verification fails.
Original-author: Aditya Kali <adityakali@google.com>
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2013-08-29 01:35:51 +04:00
# define EXT4_GROUP_INFO_BBITMAP_CORRUPT_BIT 2
2013-08-29 02:32:58 +04:00
# define EXT4_GROUP_INFO_IBITMAP_CORRUPT_BIT 3
2009-01-06 06:19:52 +03:00
# define EXT4_MB_GRP_NEED_INIT(grp) \
( test_bit ( EXT4_GROUP_INFO_NEED_INIT_BIT , & ( ( grp ) - > bb_state ) ) )
ext4: mark block group as corrupt on block bitmap error
When we notice a block-bitmap corruption (because of device failure or
something else), we should mark this group as corrupt and prevent
further block allocations/deallocations from it. Currently, we end up
generating one error message for every block in the bitmap. This
potentially could make the system unstable as noticed in some
bugs. With this patch, the error will be printed only the first time
and mark the entire block group as corrupted. This prevents future
access allocations/deallocations from it.
Also tested by corrupting the block
bitmap and forcefully introducing the mb_free_blocks error:
(1) create a largefile (2Gb)
$ dd if=/dev/zero of=largefile oflag=direct bs=10485760 count=200
(2) umount filesystem. use dumpe2fs to see which block-bitmaps
are in use by largefile and note their block numbers
(3) use dd to zero-out the used block bitmaps
$ dd if=/dev/zero of=/dev/hdc4 bs=4096 seek=14 count=8 oflag=direct
(4) mount the FS and delete the largefile.
(5) recreate the largefile. verify that the new largefile does not
get any blocks from the groups marked as bad.
Without the patch, we will see mb_free_blocks error for each bit in
each zero'ed out bitmap at (4). With the patch, we only see the error
once per blockgroup:
[ 309.706803] EXT4-fs error (device sdb4): ext4_mb_generate_buddy:735: group 15: 32768 clusters in bitmap, 0 in gd. blk grp corrupted.
[ 309.720824] EXT4-fs error (device sdb4): ext4_mb_generate_buddy:735: group 14: 32768 clusters in bitmap, 0 in gd. blk grp corrupted.
[ 309.732858] EXT4-fs error (device sdb4) in ext4_free_blocks:4802: IO failure
[ 309.748321] EXT4-fs error (device sdb4): ext4_mb_generate_buddy:735: group 13: 32768 clusters in bitmap, 0 in gd. blk grp corrupted.
[ 309.760331] EXT4-fs error (device sdb4) in ext4_free_blocks:4802: IO failure
[ 309.769695] EXT4-fs error (device sdb4): ext4_mb_generate_buddy:735: group 12: 32768 clusters in bitmap, 0 in gd. blk grp corrupted.
[ 309.781721] EXT4-fs error (device sdb4) in ext4_free_blocks:4802: IO failure
[ 309.798166] EXT4-fs error (device sdb4): ext4_mb_generate_buddy:735: group 11: 32768 clusters in bitmap, 0 in gd. blk grp corrupted.
[ 309.810184] EXT4-fs error (device sdb4) in ext4_free_blocks:4802: IO failure
[ 309.819532] EXT4-fs error (device sdb4): ext4_mb_generate_buddy:735: group 10: 32768 clusters in bitmap, 0 in gd. blk grp corrupted.
Google-Bug-Id: 7258357
[darrick.wong@oracle.com]
Further modifications (by Darrick) to make more obvious that this corruption
bit applies to blocks only. Set the corruption flag if the block group bitmap
verification fails.
Original-author: Aditya Kali <adityakali@google.com>
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2013-08-29 01:35:51 +04:00
# define EXT4_MB_GRP_BBITMAP_CORRUPT(grp) \
( test_bit ( EXT4_GROUP_INFO_BBITMAP_CORRUPT_BIT , & ( ( grp ) - > bb_state ) ) )
2013-08-29 02:32:58 +04:00
# define EXT4_MB_GRP_IBITMAP_CORRUPT(grp) \
( test_bit ( EXT4_GROUP_INFO_IBITMAP_CORRUPT_BIT , & ( ( grp ) - > bb_state ) ) )
2009-01-06 06:19:52 +03:00
2011-07-11 08:03:38 +04:00
# define EXT4_MB_GRP_WAS_TRIMMED(grp) \
( test_bit ( EXT4_GROUP_INFO_WAS_TRIMMED_BIT , & ( ( grp ) - > bb_state ) ) )
# define EXT4_MB_GRP_SET_TRIMMED(grp) \
( set_bit ( EXT4_GROUP_INFO_WAS_TRIMMED_BIT , & ( ( grp ) - > bb_state ) ) )
# define EXT4_MB_GRP_CLEAR_TRIMMED(grp) \
( clear_bit ( EXT4_GROUP_INFO_WAS_TRIMMED_BIT , & ( ( grp ) - > bb_state ) ) )
ext4: Avoid group preallocation for closed files
Currently the group preallocation code tries to find a large (512)
free block from which to do per-cpu group allocation for small files.
The problem with this scheme is that it leaves the filesystem horribly
fragmented. In the worst case, if the filesystem is unmounted and
remounted (after a system shutdown, for example) we forget the fact
that wee were using a particular (now-partially filled) 512 block
extent. So the next time we try to allocate space for a small file,
we will find *another* completely free 512 block chunk to allocate
small files. Given that there are 32,768 blocks in a block group,
after 64 iterations of "mount, write one 4k file in a directory,
unmount", the block group will have 64 files, each separated by 511
blocks, and the block group will no longer have any free 512
completely free chunks of blocks for group preallocation space.
So if we try to allocate blocks for a file that has been closed, such
that we know the final size of the file, and the filesystem is not
busy, avoid using group preallocation.
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2009-09-18 21:34:02 +04:00
# define EXT4_MAX_CONTENTION 8
# define EXT4_CONTENTION_THRESHOLD 2
2009-05-03 04:35:09 +04:00
static inline spinlock_t * ext4_group_lock_ptr ( struct super_block * sb ,
ext4_group_t group )
2009-01-06 06:19:52 +03:00
{
2009-05-03 04:35:09 +04:00
return bgl_lock_ptr ( EXT4_SB ( sb ) - > s_blockgroup_lock , group ) ;
}
2009-01-06 06:19:52 +03:00
ext4: Avoid group preallocation for closed files
Currently the group preallocation code tries to find a large (512)
free block from which to do per-cpu group allocation for small files.
The problem with this scheme is that it leaves the filesystem horribly
fragmented. In the worst case, if the filesystem is unmounted and
remounted (after a system shutdown, for example) we forget the fact
that wee were using a particular (now-partially filled) 512 block
extent. So the next time we try to allocate space for a small file,
we will find *another* completely free 512 block chunk to allocate
small files. Given that there are 32,768 blocks in a block group,
after 64 iterations of "mount, write one 4k file in a directory,
unmount", the block group will have 64 files, each separated by 511
blocks, and the block group will no longer have any free 512
completely free chunks of blocks for group preallocation space.
So if we try to allocate blocks for a file that has been closed, such
that we know the final size of the file, and the filesystem is not
busy, avoid using group preallocation.
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2009-09-18 21:34:02 +04:00
/*
* Returns true if the filesystem is busy enough that attempts to
* access the block group locks has run into contention .
*/
static inline int ext4_fs_is_busy ( struct ext4_sb_info * sbi )
{
return ( atomic_read ( & sbi - > s_lock_busy ) > EXT4_CONTENTION_THRESHOLD ) ;
}
2009-05-03 04:35:09 +04:00
static inline void ext4_lock_group ( struct super_block * sb , ext4_group_t group )
{
ext4: Avoid group preallocation for closed files
Currently the group preallocation code tries to find a large (512)
free block from which to do per-cpu group allocation for small files.
The problem with this scheme is that it leaves the filesystem horribly
fragmented. In the worst case, if the filesystem is unmounted and
remounted (after a system shutdown, for example) we forget the fact
that wee were using a particular (now-partially filled) 512 block
extent. So the next time we try to allocate space for a small file,
we will find *another* completely free 512 block chunk to allocate
small files. Given that there are 32,768 blocks in a block group,
after 64 iterations of "mount, write one 4k file in a directory,
unmount", the block group will have 64 files, each separated by 511
blocks, and the block group will no longer have any free 512
completely free chunks of blocks for group preallocation space.
So if we try to allocate blocks for a file that has been closed, such
that we know the final size of the file, and the filesystem is not
busy, avoid using group preallocation.
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2009-09-18 21:34:02 +04:00
spinlock_t * lock = ext4_group_lock_ptr ( sb , group ) ;
if ( spin_trylock ( lock ) )
/*
* We ' re able to grab the lock right away , so drop the
* lock contention counter .
*/
atomic_add_unless ( & EXT4_SB ( sb ) - > s_lock_busy , - 1 , 0 ) ;
else {
/*
* The lock is busy , so bump the contention counter ,
* and then wait on the spin lock .
*/
atomic_add_unless ( & EXT4_SB ( sb ) - > s_lock_busy , 1 ,
EXT4_MAX_CONTENTION ) ;
spin_lock ( lock ) ;
}
2009-01-06 06:19:52 +03:00
}
static inline void ext4_unlock_group ( struct super_block * sb ,
ext4_group_t group )
{
2009-05-03 04:35:09 +04:00
spin_unlock ( ext4_group_lock_ptr ( sb , group ) ) ;
2009-01-06 06:19:52 +03:00
}
2011-06-28 03:16:02 +04:00
/*
* Block validity checking
*/
# define ext4_check_indirect_blockref(inode, bh) \
ext4_check_blockref ( __func__ , __LINE__ , inode , \
( __le32 * ) ( bh ) - > b_data , \
EXT4_ADDR_PER_BLOCK ( ( inode ) - > i_sb ) )
# define ext4_ind_check_inode(inode) \
ext4_check_blockref ( __func__ , __LINE__ , inode , \
EXT4_I ( inode ) - > i_data , \
EXT4_NDIR_BLOCKS )
2006-10-11 12:20:50 +04:00
/*
* Inodes and files operations
*/
/* dir.c */
2006-10-11 12:20:53 +04:00
extern const struct file_operations ext4_dir_operations ;
2006-10-11 12:20:50 +04:00
/* file.c */
2007-02-12 11:55:40 +03:00
extern const struct inode_operations ext4_file_inode_operations ;
2006-10-11 12:20:53 +04:00
extern const struct file_operations ext4_file_operations ;
2010-10-28 05:30:06 +04:00
extern loff_t ext4_llseek ( struct file * file , loff_t offset , int origin ) ;
2012-10-05 19:31:55 +04:00
extern void ext4_unwritten_wait ( struct inode * inode ) ;
2006-10-11 12:20:50 +04:00
2013-02-10 00:23:03 +04:00
/* inline.c */
extern int ext4_has_inline_data ( struct inode * inode ) ;
extern int ext4_get_inline_size ( struct inode * inode ) ;
extern int ext4_get_max_inline_size ( struct inode * inode ) ;
extern int ext4_find_inline_data_nolock ( struct inode * inode ) ;
extern void ext4_write_inline_data ( struct inode * inode ,
struct ext4_iloc * iloc ,
void * buffer , loff_t pos ,
unsigned int len ) ;
extern int ext4_prepare_inline_data ( handle_t * handle , struct inode * inode ,
unsigned int len ) ;
extern int ext4_init_inline_data ( handle_t * handle , struct inode * inode ,
unsigned int len ) ;
extern int ext4_destroy_inline_data ( handle_t * handle , struct inode * inode ) ;
extern int ext4_readpage_inline ( struct inode * inode , struct page * page ) ;
extern int ext4_try_to_write_inline_data ( struct address_space * mapping ,
struct inode * inode ,
loff_t pos , unsigned len ,
unsigned flags ,
struct page * * pagep ) ;
extern int ext4_write_inline_data_end ( struct inode * inode ,
loff_t pos , unsigned len ,
unsigned copied ,
struct page * page ) ;
extern struct buffer_head *
ext4_journalled_write_inline_data ( struct inode * inode ,
unsigned len ,
struct page * page ) ;
extern int ext4_da_write_inline_data_begin ( struct address_space * mapping ,
struct inode * inode ,
loff_t pos , unsigned len ,
unsigned flags ,
struct page * * pagep ,
void * * fsdata ) ;
extern int ext4_da_write_inline_data_end ( struct inode * inode , loff_t pos ,
unsigned len , unsigned copied ,
struct page * page ) ;
extern int ext4_try_add_inline_entry ( handle_t * handle , struct dentry * dentry ,
struct inode * inode ) ;
extern int ext4_try_create_inline_dir ( handle_t * handle ,
struct inode * parent ,
struct inode * inode ) ;
extern int ext4_read_inline_dir ( struct file * filp ,
2013-05-18 00:08:53 +04:00
struct dir_context * ctx ,
2013-02-10 00:23:03 +04:00
int * has_inline_data ) ;
2013-04-20 01:53:09 +04:00
extern int htree_inlinedir_to_tree ( struct file * dir_file ,
struct inode * dir , ext4_lblk_t block ,
struct dx_hash_info * hinfo ,
__u32 start_hash , __u32 start_minor_hash ,
int * has_inline_data ) ;
2013-02-10 00:23:03 +04:00
extern struct buffer_head * ext4_find_inline_entry ( struct inode * dir ,
const struct qstr * d_name ,
struct ext4_dir_entry_2 * * res_dir ,
int * has_inline_data ) ;
extern int ext4_delete_inline_entry ( handle_t * handle ,
struct inode * dir ,
struct ext4_dir_entry_2 * de_del ,
struct buffer_head * bh ,
int * has_inline_data ) ;
extern int empty_inline_dir ( struct inode * dir , int * has_inline_data ) ;
extern struct buffer_head * ext4_get_first_inline_block ( struct inode * inode ,
struct ext4_dir_entry_2 * * parent_de ,
int * retval ) ;
extern int ext4_inline_data_fiemap ( struct inode * inode ,
struct fiemap_extent_info * fieinfo ,
int * has_inline ) ;
extern int ext4_try_to_evict_inline_data ( handle_t * handle ,
struct inode * inode ,
int needed ) ;
extern void ext4_inline_data_truncate ( struct inode * inode , int * has_inline ) ;
extern int ext4_convert_inline_data ( struct inode * inode ) ;
2006-10-11 12:20:50 +04:00
/* namei.c */
2007-02-12 11:55:40 +03:00
extern const struct inode_operations ext4_dir_inode_operations ;
extern const struct inode_operations ext4_special_inode_operations ;
2009-05-01 21:49:15 +04:00
extern struct dentry * ext4_get_parent ( struct dentry * child ) ;
2012-12-10 23:05:57 +04:00
extern struct ext4_dir_entry_2 * ext4_init_dot_dotdot ( struct inode * inode ,
struct ext4_dir_entry_2 * de ,
int blocksize , int csum_size ,
unsigned int parent_ino , int dotdot_real_len ) ;
2012-12-10 23:05:59 +04:00
extern void initialize_dirent_tail ( struct ext4_dir_entry_tail * t ,
unsigned int blocksize ) ;
extern int ext4_handle_dirty_dirent_node ( handle_t * handle ,
struct inode * inode ,
struct buffer_head * bh ) ;
2013-04-20 01:53:09 +04:00
# define S_SHIFT 12
static unsigned char ext4_type_by_mode [ S_IFMT > > S_SHIFT ] = {
[ S_IFREG > > S_SHIFT ] = EXT4_FT_REG_FILE ,
[ S_IFDIR > > S_SHIFT ] = EXT4_FT_DIR ,
[ S_IFCHR > > S_SHIFT ] = EXT4_FT_CHRDEV ,
[ S_IFBLK > > S_SHIFT ] = EXT4_FT_BLKDEV ,
[ S_IFIFO > > S_SHIFT ] = EXT4_FT_FIFO ,
[ S_IFSOCK > > S_SHIFT ] = EXT4_FT_SOCK ,
[ S_IFLNK > > S_SHIFT ] = EXT4_FT_SYMLINK ,
} ;
static inline void ext4_set_de_type ( struct super_block * sb ,
struct ext4_dir_entry_2 * de ,
umode_t mode ) {
if ( EXT4_HAS_INCOMPAT_FEATURE ( sb , EXT4_FEATURE_INCOMPAT_FILETYPE ) )
de - > file_type = ext4_type_by_mode [ ( mode & S_IFMT ) > > S_SHIFT ] ;
}
2006-10-11 12:20:50 +04:00
/* symlink.c */
2007-02-12 11:55:40 +03:00
extern const struct inode_operations ext4_symlink_inode_operations ;
extern const struct inode_operations ext4_fast_symlink_inode_operations ;
2006-10-11 12:20:50 +04:00
2009-05-17 23:38:01 +04:00
/* block_validity */
extern void ext4_release_system_zone ( struct super_block * sb ) ;
extern int ext4_setup_system_zone ( struct super_block * sb ) ;
2010-10-28 05:30:14 +04:00
extern int __init ext4_init_system_zone ( void ) ;
extern void ext4_exit_system_zone ( void ) ;
2009-05-17 23:38:01 +04:00
extern int ext4_data_block_valid ( struct ext4_sb_info * sbi ,
ext4_fsblk_t start_blk ,
unsigned int count ) ;
2011-06-28 03:16:02 +04:00
extern int ext4_check_blockref ( const char * , unsigned int ,
struct inode * , __le32 * , unsigned int ) ;
2009-05-17 23:38:01 +04:00
2006-10-11 12:21:03 +04:00
/* extents.c */
2012-11-28 22:03:30 +04:00
struct ext4_ext_path ;
struct ext4_extent ;
2013-08-28 22:47:06 +04:00
/*
* Maximum number of logical blocks in a file ; ext4_extent ' s ee_block is
* __le32 .
*/
# define EXT_MAX_BLOCKS 0xffffffff
2006-10-11 12:21:03 +04:00
extern int ext4_ext_tree_init ( handle_t * handle , struct inode * ) ;
extern int ext4_ext_writepage_trans_blocks ( struct inode * , int ) ;
2013-06-04 21:01:11 +04:00
extern int ext4_ext_index_trans_blocks ( struct inode * inode , int extents ) ;
2010-05-17 03:00:00 +04:00
extern int ext4_ext_map_blocks ( handle_t * handle , struct inode * inode ,
struct ext4_map_blocks * map , int flags ) ;
2013-04-03 20:47:17 +04:00
extern void ext4_ext_truncate ( handle_t * , struct inode * ) ;
2013-04-03 20:45:17 +04:00
extern int ext4_ext_remove_space ( struct inode * inode , ext4_lblk_t start ,
ext4_lblk_t end ) ;
2006-10-11 12:21:03 +04:00
extern void ext4_ext_init ( struct super_block * ) ;
extern void ext4_ext_release ( struct super_block * ) ;
2011-01-14 15:07:43 +03:00
extern long ext4_fallocate ( struct file * file , int mode , loff_t offset ,
2007-07-18 05:42:41 +04:00
loff_t len ) ;
2013-06-04 21:21:11 +04:00
extern int ext4_convert_unwritten_extents ( handle_t * handle , struct inode * inode ,
loff_t offset , ssize_t len ) ;
2010-05-17 03:00:00 +04:00
extern int ext4_map_blocks ( handle_t * handle , struct inode * inode ,
struct ext4_map_blocks * map , int flags ) ;
2012-11-28 22:03:30 +04:00
extern int ext4_ext_calc_metadata_amount ( struct inode * inode ,
ext4_lblk_t lblocks ) ;
extern int ext4_extent_tree_init ( handle_t * , struct inode * ) ;
extern int ext4_ext_calc_credits_for_single_extent ( struct inode * inode ,
int num ,
struct ext4_ext_path * path ) ;
extern int ext4_can_extents_be_merged ( struct inode * inode ,
struct ext4_extent * ex1 ,
struct ext4_extent * ex2 ) ;
extern int ext4_ext_insert_extent ( handle_t * , struct inode * ,
struct ext4_ext_path * ,
struct ext4_extent * , int ) ;
extern struct ext4_ext_path * ext4_ext_find_extent ( struct inode * , ext4_lblk_t ,
2013-08-17 05:23:41 +04:00
struct ext4_ext_path * ,
int flags ) ;
2012-11-28 22:03:30 +04:00
extern void ext4_ext_drop_refs ( struct ext4_ext_path * ) ;
extern int ext4_ext_check_inode ( struct inode * inode ) ;
2013-02-18 09:28:47 +04:00
extern int ext4_find_delalloc_range ( struct inode * inode ,
ext4_lblk_t lblk_start ,
ext4_lblk_t lblk_end ) ;
2012-11-28 22:03:30 +04:00
extern int ext4_find_delalloc_cluster ( struct inode * inode , ext4_lblk_t lblk ) ;
2008-11-22 23:04:59 +03:00
extern int ext4_fiemap ( struct inode * inode , struct fiemap_extent_info * fieinfo ,
__u64 start , __u64 len ) ;
2013-08-17 06:05:14 +04:00
extern int ext4_ext_precache ( struct inode * inode ) ;
2014-02-24 00:18:59 +04:00
extern int ext4_collapse_range ( struct inode * inode , loff_t offset , loff_t len ) ;
2012-11-28 22:03:30 +04:00
2009-06-18 03:24:03 +04:00
/* move_extent.c */
2013-04-08 20:54:05 +04:00
extern void ext4_double_down_write_data_sem ( struct inode * first ,
struct inode * second ) ;
extern void ext4_double_up_write_data_sem ( struct inode * orig_inode ,
struct inode * donor_inode ) ;
2009-06-18 03:24:03 +04:00
extern int ext4_move_extents ( struct file * o_filp , struct file * d_filp ,
__u64 start_orig , __u64 start_donor ,
__u64 len , __u64 * moved_len ) ;
2014-02-24 00:18:59 +04:00
extern int mext_next_extent ( struct inode * inode , struct ext4_ext_path * path ,
struct ext4_extent * * extent ) ;
2009-06-18 03:24:03 +04:00
2010-10-28 05:30:10 +04:00
/* page-io.c */
2010-10-28 05:30:14 +04:00
extern int __init ext4_init_pageio ( void ) ;
extern void ext4_exit_pageio ( void ) ;
2010-10-28 05:30:10 +04:00
extern ext4_io_end_t * ext4_init_io_end ( struct inode * inode , gfp_t flags ) ;
2013-06-04 19:58:58 +04:00
extern ext4_io_end_t * ext4_get_io_end ( ext4_io_end_t * io_end ) ;
extern int ext4_put_io_end ( ext4_io_end_t * io_end ) ;
extern void ext4_put_io_end_defer ( ext4_io_end_t * io_end ) ;
extern void ext4_io_submit_init ( struct ext4_io_submit * io ,
struct writeback_control * wbc ) ;
2013-06-04 22:21:02 +04:00
extern void ext4_end_io_rsv_work ( struct work_struct * work ) ;
2010-10-28 05:30:10 +04:00
extern void ext4_io_submit ( struct ext4_io_submit * io ) ;
extern int ext4_bio_write_page ( struct ext4_io_submit * io ,
struct page * page ,
int len ,
2014-05-12 16:12:25 +04:00
struct writeback_control * wbc ,
bool keep_towrite ) ;
2008-11-22 23:04:59 +03:00
2011-05-25 02:31:25 +04:00
/* mmp.c */
extern int ext4_multi_mount_protect ( struct super_block * , ext4_fsblk_t ) ;
2012-04-30 02:47:10 +04:00
extern void ext4_mmp_csum_set ( struct super_block * sb , struct mmp_struct * mmp ) ;
extern int ext4_mmp_csum_verify ( struct super_block * sb ,
struct mmp_struct * mmp ) ;
2011-05-25 02:31:25 +04:00
2013-06-04 21:19:34 +04:00
/*
* Note that these flags will never ever appear in a buffer_head ' s state flag .
* See EXT4_MAP_ . . . to see where this is used .
*/
2010-03-05 00:14:02 +03:00
enum ext4_state_bits {
2014-04-21 07:44:47 +04:00
BH_AllocFromCluster /* allocated blocks were part of already
2013-06-04 21:19:34 +04:00
* allocated cluster . */
2014-04-21 07:44:47 +04:00
= BH_JBDPrivateStart
2010-03-05 00:14:02 +03:00
} ;
2009-01-06 05:49:55 +03:00
/*
2012-09-20 05:48:00 +04:00
* Add new method to test whether block and inode bitmaps are properly
2009-01-06 05:49:55 +03:00
* initialized . With uninit_bg reading the block from disk is not enough
* to mark the bitmap uptodate . We need to also zero - out the bitmap
*/
# define BH_BITMAP_UPTODATE BH_JBDPrivateStart
static inline int bitmap_uptodate ( struct buffer_head * bh )
{
return ( buffer_uptodate ( bh ) & &
test_bit ( BH_BITMAP_UPTODATE , & ( bh ) - > b_state ) ) ;
}
static inline void set_bitmap_uptodate ( struct buffer_head * bh )
{
set_bit ( BH_BITMAP_UPTODATE , & ( bh ) - > b_state ) ;
}
2012-09-29 08:41:21 +04:00
/*
* Disable DIO read nolock optimization , so new dioreaders will be forced
* to grab i_mutex
*/
static inline void ext4_inode_block_unlocked_dio ( struct inode * inode )
{
ext4_set_inode_state ( inode , EXT4_STATE_DIOREAD_LOCK ) ;
smp_mb ( ) ;
}
static inline void ext4_inode_resume_unlocked_dio ( struct inode * inode )
{
smp_mb ( ) ;
ext4_clear_inode_state ( inode , EXT4_STATE_DIOREAD_LOCK ) ;
}
2010-03-04 07:55:01 +03:00
# define in_range(b, first, len) ((b) >= (first) && (b) <= (first) + (len) - 1)
ext4: serialize unaligned asynchronous DIO
ext4 has a data corruption case when doing non-block-aligned
asynchronous direct IO into a sparse file, as demonstrated
by xfstest 240.
The root cause is that while ext4 preallocates space in the
hole, mappings of that space still look "new" and
dio_zero_block() will zero out the unwritten portions. When
more than one AIO thread is going, they both find this "new"
block and race to zero out their portion; this is uncoordinated
and causes data corruption.
Dave Chinner fixed this for xfs by simply serializing all
unaligned asynchronous direct IO. I've done the same here.
The difference is that we only wait on conversions, not all IO.
This is a very big hammer, and I'm not very pleased with
stuffing this into ext4_file_write(). But since ext4 is
DIO_LOCKING, we need to serialize it at this high level.
I tried to move this into ext4_ext_direct_IO, but by then
we have the i_mutex already, and we will wait on the
work queue to do conversions - which must also take the
i_mutex. So that won't work.
This was originally exposed by qemu-kvm installing to
a raw disk image with a normal sector-63 alignment. I've
tested a backport of this patch with qemu, and it does
avoid the corruption. It is also quite a lot slower
(14 min for package installs, vs. 8 min for well-aligned)
but I'll take slow correctness over fast corruption any day.
Mingming suggested that we can track outstanding
conversions, and wait on those so that non-sparse
files won't be affected, and I've implemented that here;
unaligned AIO to nonsparse files won't take a perf hit.
[tytso@mit.edu: Keep the mutex as a hashed array instead
of bloating the ext4 inode]
[tytso@mit.edu: Fix up namespace issues so that global
variables are protected with an "ext4_" prefix.]
Signed-off-by: Eric Sandeen <sandeen@redhat.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2011-02-12 16:17:34 +03:00
/* For ioend & aio unwritten conversion wait queues */
# define EXT4_WQ_HASH_SZ 37
# define ext4_ioend_wq(v) (&ext4__ioend_wq[((unsigned long)(v)) %\
EXT4_WQ_HASH_SZ ] )
# define ext4_aio_mutex(v) (&ext4__aio_mutex[((unsigned long)(v)) %\
EXT4_WQ_HASH_SZ ] )
extern wait_queue_head_t ext4__ioend_wq [ EXT4_WQ_HASH_SZ ] ;
extern struct mutex ext4__aio_mutex [ EXT4_WQ_HASH_SZ ] ;
2011-07-27 05:35:44 +04:00
# define EXT4_RESIZING 0
extern int ext4_resize_begin ( struct super_block * sb ) ;
extern void ext4_resize_end ( struct super_block * sb ) ;
2006-10-11 12:20:50 +04:00
# endif /* __KERNEL__ */
2008-04-30 02:13:32 +04:00
# endif /* _EXT4_H */