adulau/aha
1
0
Fork 0
mirror of https://github.com/adulau/aha.git synced 2024-12-27 11:16:11 +00:00
Code Issues Projects Releases Packages Wiki Activity

Btrfs: Mixed back reference (FORWARD ROLLING FORMAT CHANGE)

Browse source 
This commit introduces a new kind of back reference for btrfs metadata.
Once a filesystem has been mounted with this commit, IT WILL NO LONGER
BE MOUNTABLE BY OLDER KERNELS.

When a tree block in subvolume tree is cow'd, the reference counts of all
extents it points to are increased by one.  At transaction commit time,
the old root of the subvolume is recorded in a "dead root" data structure,
and the btree it points to is later walked, dropping reference counts
and freeing any blocks where the reference count goes to 0.

The increments done during cow and decrements done after commit cancel out,
and the walk is a very expensive way to go about freeing the blocks that
are no longer referenced by the new btree root.  This commit reduces the
transaction overhead by avoiding the need for dead root records.

When a non-shared tree block is cow'd, we free the old block at once, and the
new block inherits old block's references. When a tree block with reference
count > 1 is cow'd, we increase the reference counts of all extents
the new block points to by one, and decrease the old block's reference count by
one.

This dead tree avoidance code removes the need to modify the reference
counts of lower level extents when a non-shared tree block is cow'd.
But we still need to update back ref for all pointers in the block.
This is because the location of the block is recorded in the back ref
item.

We can solve this by introducing a new type of back ref. The new
back ref provides information about pointer's key, level and in which
tree the pointer lives. This information allow us to find the pointer
by searching the tree. The shortcoming of the new back ref is that it
only works for pointers in tree blocks referenced by their owner trees.

This is mostly a problem for snapshots, where resolving one of these
fuzzy back references would be O(number_of_snapshots) and quite slow.
The solution used here is to use the fuzzy back references in the common
case where a given tree block is only referenced by one root,
and use the full back references when multiple roots have a reference
on a given block.

This commit adds per subvolume red-black tree to keep trace of cached
inodes. The red-black tree helps the balancing code to find cached
inodes whose inode numbers within a given range.

This commit improves the balancing code by introducing several data
structures to keep the state of balancing. The most important one
is the back ref cache. It caches how the upper level tree blocks are
referenced. This greatly reduce the overhead of checking back ref.

The improved balancing code scales significantly better with a large
number of snapshots.

This is a very large commit and was written in a number of
pieces.  But, they depend heavily on the disk format change and were
squashed together to make sure git bisect didn't end up in a
bad state wrt space balancing or the format change.

Signed-off-by: Yan Zheng <zheng.yan@oracle.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
This commit is contained in:
Yan Zheng 2009-06-10 10:45:14 -04:00 committed by Chris Mason
parent 5c939df56c
commit 5d4f98a28c
20 changed files with 6953 additions and 2068 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

4
fs/btrfs/Makefile
View file

@ -6,5 +6,5 @@ btrfs-y += super.o ctree.o extent-tree.o print-tree.o root-tree.o dir-item.o \
transaction.o inode.o file.o tree-defrag.o \
extent_map.o sysfs.o struct-funcs.o xattr.o ordered-data.o \
extent_io.o volumes.o async-thread.o ioctl.o locking.o orphan.o \
ref-cache.o export.o tree-log.o acl.o free-space-cache.o zlib.o \
compression.o delayed-ref.o
export.o tree-log.o acl.o free-space-cache.o zlib.o \
compression.o delayed-ref.o relocation.o

3
fs/btrfs/btrfs_inode.h
View file

@ -72,6 +72,9 @@ struct btrfs_inode {
*/
struct list_head ordered_operations;
/* node for the red-black tree that links inodes in subvolume root */
struct rb_node rb_node;
/* the space_info for where this inode's data allocations are done */
struct btrfs_space_info *space_info;

685
fs/btrfs/ctree.c
View file

File diff suppressed because it is too large Load diff

308
fs/btrfs/ctree.h
View file

@ -45,6 +45,8 @@ struct btrfs_ordered_sum;
#define BTRFS_MAX_LEVEL 8
#define BTRFS_COMPAT_EXTENT_TREE_V0
/*
* files bigger than this get some pre-flushing when they are added
* to the ordered operations list. That way we limit the total
@ -267,7 +269,18 @@ static inline unsigned long btrfs_chunk_item_size(int num_stripes)
}
#define BTRFS_FSID_SIZE 16
#define BTRFS_HEADER_FLAG_WRITTEN (1 << 0)
#define BTRFS_HEADER_FLAG_WRITTEN (1ULL << 0)
#define BTRFS_HEADER_FLAG_RELOC (1ULL << 1)
#define BTRFS_SUPER_FLAG_SEEDING (1ULL << 32)
#define BTRFS_SUPER_FLAG_METADUMP (1ULL << 33)
#define BTRFS_BACKREF_REV_MAX 256
#define BTRFS_BACKREF_REV_SHIFT 56
#define BTRFS_BACKREF_REV_MASK (((u64)BTRFS_BACKREF_REV_MAX - 1) << \
BTRFS_BACKREF_REV_SHIFT)
#define BTRFS_OLD_BACKREF_REV 0
#define BTRFS_MIXED_BACKREF_REV 1
/*
* every tree block (leaf or node) starts with this header.
@ -296,7 +309,6 @@ struct btrfs_header {
sizeof(struct btrfs_item) - \
sizeof(struct btrfs_file_extent_item))
#define BTRFS_SUPER_FLAG_SEEDING (1ULL << 32)
/*
* this is a very generous portion of the super block, giving us
@ -355,9 +367,12 @@ struct btrfs_super_block {
* Compat flags that we support. If any incompat flags are set other than the
* ones specified below then we will fail to mount
*/
#define BTRFS_FEATURE_COMPAT_SUPP 0x0
#define BTRFS_FEATURE_COMPAT_RO_SUPP 0x0
#define BTRFS_FEATURE_INCOMPAT_SUPP 0x0
#define BTRFS_FEATURE_INCOMPAT_MIXED_BACKREF (1ULL << 0)
#define BTRFS_FEATURE_COMPAT_SUPP 0ULL
#define BTRFS_FEATURE_COMPAT_RO_SUPP 0ULL
#define BTRFS_FEATURE_INCOMPAT_SUPP \
BTRFS_FEATURE_INCOMPAT_MIXED_BACKREF
/*
* A leaf is full of items. offset and size tell us where to find
@ -421,23 +436,65 @@ struct btrfs_path {
unsigned int keep_locks:1;
unsigned int skip_locking:1;
unsigned int leave_spinning:1;
unsigned int search_commit_root:1;
};
/*
* items in the extent btree are used to record the objectid of the
* owner of the block and the number of references
*/
struct btrfs_extent_item {
__le64 refs;
__le64 generation;
__le64 flags;
} __attribute__ ((__packed__));
struct btrfs_extent_item_v0 {
__le32 refs;
} __attribute__ ((__packed__));
struct btrfs_extent_ref {
#define BTRFS_MAX_EXTENT_ITEM_SIZE(r) ((BTRFS_LEAF_DATA_SIZE(r) >> 4) - \
sizeof(struct btrfs_item))
#define BTRFS_EXTENT_FLAG_DATA (1ULL << 0)
#define BTRFS_EXTENT_FLAG_TREE_BLOCK (1ULL << 1)
/* following flags only apply to tree blocks */
/* use full backrefs for extent pointers in the block */
#define BTRFS_BLOCK_FLAG_FULL_BACKREF (1ULL << 8)
struct btrfs_tree_block_info {
struct btrfs_disk_key key;
u8 level;
} __attribute__ ((__packed__));
struct btrfs_extent_data_ref {
__le64 root;
__le64 objectid;
__le64 offset;
__le32 count;
} __attribute__ ((__packed__));
struct btrfs_shared_data_ref {
__le32 count;
} __attribute__ ((__packed__));
struct btrfs_extent_inline_ref {
u8 type;
u64 offset;
} __attribute__ ((__packed__));
/* old style backrefs item */
struct btrfs_extent_ref_v0 {
__le64 root;
__le64 generation;
__le64 objectid;
__le32 num_refs;
__le32 count;
} __attribute__ ((__packed__));
/* dev extents record free space on individual devices. The owner
* field points back to the chunk allocation mapping tree that allocated
* the extent. The chunk tree uuid field is a way to double check the owner
@ -695,12 +752,7 @@ struct btrfs_block_group_cache {
struct list_head cluster_list;
};
struct btrfs_leaf_ref_tree {
struct rb_root root;
struct list_head list;
spinlock_t lock;
};
struct reloc_control;
struct btrfs_device;
struct btrfs_fs_devices;
struct btrfs_fs_info {
@ -831,18 +883,11 @@ struct btrfs_fs_info {
struct task_struct *cleaner_kthread;
int thread_pool_size;
/* tree relocation relocated fields */
struct list_head dead_reloc_roots;
struct btrfs_leaf_ref_tree reloc_ref_tree;
struct btrfs_leaf_ref_tree shared_ref_tree;
struct kobject super_kobj;
struct completion kobj_unregister;
int do_barriers;
int closing;
int log_root_recovering;
atomic_t throttles;
atomic_t throttle_gen;
u64 total_pinned;
@ -861,6 +906,8 @@ struct btrfs_fs_info {
*/
struct list_head space_info;
struct reloc_control *reloc_ctl;
spinlock_t delalloc_lock;
spinlock_t new_trans_lock;
u64 delalloc_bytes;
@ -891,7 +938,6 @@ struct btrfs_fs_info {
* in ram representation of the tree. extent_root is used for all allocations
* and for the extent tree extent_root root.
*/
struct btrfs_dirty_root;
struct btrfs_root {
struct extent_buffer *node;
@ -899,9 +945,6 @@ struct btrfs_root {
spinlock_t node_lock;
struct extent_buffer *commit_root;
struct btrfs_leaf_ref_tree *ref_tree;
struct btrfs_leaf_ref_tree ref_tree_struct;
struct btrfs_dirty_root *dirty_root;
struct btrfs_root *log_root;
struct btrfs_root *reloc_root;
@ -952,10 +995,15 @@ struct btrfs_root {
/* the dirty list is only used by non-reference counted roots */
struct list_head dirty_list;
struct list_head root_list;
spinlock_t list_lock;
struct list_head dead_list;
struct list_head orphan_list;
spinlock_t inode_lock;
/* red-black tree that keeps track of in-memory inodes */
struct rb_root inode_tree;
/*
* right now this just gets used so that a root has its own devid
* for stat. It may be used for more later
@ -1017,7 +1065,16 @@ struct btrfs_root {
* are used, and how many references there are to each block
*/
#define BTRFS_EXTENT_ITEM_KEY 168
#define BTRFS_EXTENT_REF_KEY 180
#define BTRFS_TREE_BLOCK_REF_KEY 176
#define BTRFS_EXTENT_DATA_REF_KEY 178
#define BTRFS_EXTENT_REF_V0_KEY 180
#define BTRFS_SHARED_BLOCK_REF_KEY 182
#define BTRFS_SHARED_DATA_REF_KEY 184
/*
* block groups give us hints into the extent allocation trees. Which
@ -1317,24 +1374,67 @@ static inline u8 *btrfs_dev_extent_chunk_tree_uuid(struct btrfs_dev_extent *dev)
return (u8 *)((unsigned long)dev + ptr);
}
/* struct btrfs_extent_ref */
BTRFS_SETGET_FUNCS(ref_root, struct btrfs_extent_ref, root, 64);
BTRFS_SETGET_FUNCS(ref_generation, struct btrfs_extent_ref, generation, 64);
BTRFS_SETGET_FUNCS(ref_objectid, struct btrfs_extent_ref, objectid, 64);
BTRFS_SETGET_FUNCS(ref_num_refs, struct btrfs_extent_ref, num_refs, 32);
BTRFS_SETGET_FUNCS(extent_refs, struct btrfs_extent_item, refs, 64);
BTRFS_SETGET_FUNCS(extent_generation, struct btrfs_extent_item,
generation, 64);
BTRFS_SETGET_FUNCS(extent_flags, struct btrfs_extent_item, flags, 64);
BTRFS_SETGET_STACK_FUNCS(stack_ref_root, struct btrfs_extent_ref, root, 64);
BTRFS_SETGET_STACK_FUNCS(stack_ref_generation, struct btrfs_extent_ref,
generation, 64);
BTRFS_SETGET_STACK_FUNCS(stack_ref_objectid, struct btrfs_extent_ref,
objectid, 64);
BTRFS_SETGET_STACK_FUNCS(stack_ref_num_refs, struct btrfs_extent_ref,
num_refs, 32);
BTRFS_SETGET_FUNCS(extent_refs_v0, struct btrfs_extent_item_v0, refs, 32);
/* struct btrfs_extent_item */
BTRFS_SETGET_FUNCS(extent_refs, struct btrfs_extent_item, refs, 32);
BTRFS_SETGET_STACK_FUNCS(stack_extent_refs, struct btrfs_extent_item,
refs, 32);
BTRFS_SETGET_FUNCS(tree_block_level, struct btrfs_tree_block_info, level, 8);
static inline void btrfs_tree_block_key(struct extent_buffer *eb,
struct btrfs_tree_block_info *item,
struct btrfs_disk_key *key)
{
read_eb_member(eb, item, struct btrfs_tree_block_info, key, key);
}
static inline void btrfs_set_tree_block_key(struct extent_buffer *eb,
struct btrfs_tree_block_info *item,
struct btrfs_disk_key *key)
{
write_eb_member(eb, item, struct btrfs_tree_block_info, key, key);
}
BTRFS_SETGET_FUNCS(extent_data_ref_root, struct btrfs_extent_data_ref,
root, 64);
BTRFS_SETGET_FUNCS(extent_data_ref_objectid, struct btrfs_extent_data_ref,
objectid, 64);
BTRFS_SETGET_FUNCS(extent_data_ref_offset, struct btrfs_extent_data_ref,
offset, 64);
BTRFS_SETGET_FUNCS(extent_data_ref_count, struct btrfs_extent_data_ref,
count, 32);
BTRFS_SETGET_FUNCS(shared_data_ref_count, struct btrfs_shared_data_ref,
count, 32);
BTRFS_SETGET_FUNCS(extent_inline_ref_type, struct btrfs_extent_inline_ref,
type, 8);
BTRFS_SETGET_FUNCS(extent_inline_ref_offset, struct btrfs_extent_inline_ref,
offset, 64);
static inline u32 btrfs_extent_inline_ref_size(int type)
{
if (type == BTRFS_TREE_BLOCK_REF_KEY ||
type == BTRFS_SHARED_BLOCK_REF_KEY)
return sizeof(struct btrfs_extent_inline_ref);
if (type == BTRFS_SHARED_DATA_REF_KEY)
return sizeof(struct btrfs_shared_data_ref) +
sizeof(struct btrfs_extent_inline_ref);
if (type == BTRFS_EXTENT_DATA_REF_KEY)
return sizeof(struct btrfs_extent_data_ref) +
offsetof(struct btrfs_extent_inline_ref, offset);
BUG();
return 0;
}
BTRFS_SETGET_FUNCS(ref_root_v0, struct btrfs_extent_ref_v0, root, 64);
BTRFS_SETGET_FUNCS(ref_generation_v0, struct btrfs_extent_ref_v0,
generation, 64);
BTRFS_SETGET_FUNCS(ref_objectid_v0, struct btrfs_extent_ref_v0, objectid, 64);
BTRFS_SETGET_FUNCS(ref_count_v0, struct btrfs_extent_ref_v0, count, 32);
/* struct btrfs_node */
BTRFS_SETGET_FUNCS(key_blockptr, struct btrfs_key_ptr, blockptr, 64);
@ -1558,6 +1658,21 @@ static inline int btrfs_clear_header_flag(struct extent_buffer *eb, u64 flag)
return (flags & flag) == flag;
}
static inline int btrfs_header_backref_rev(struct extent_buffer *eb)
{
u64 flags = btrfs_header_flags(eb);
return flags >> BTRFS_BACKREF_REV_SHIFT;
}
static inline void btrfs_set_header_backref_rev(struct extent_buffer *eb,
int rev)
{
u64 flags = btrfs_header_flags(eb);
flags &= ~BTRFS_BACKREF_REV_MASK;
flags |= (u64)rev << BTRFS_BACKREF_REV_SHIFT;
btrfs_set_header_flags(eb, flags);
}
static inline u8 *btrfs_header_fsid(struct extent_buffer *eb)
{
unsigned long ptr = offsetof(struct btrfs_header, fsid);
@ -1790,39 +1905,32 @@ int btrfs_update_pinned_extents(struct btrfs_root *root,
int btrfs_drop_leaf_ref(struct btrfs_trans_handle *trans,
struct btrfs_root *root, struct extent_buffer *leaf);
int btrfs_cross_ref_exist(struct btrfs_trans_handle *trans,
struct btrfs_root *root, u64 objectid, u64 bytenr);
struct btrfs_root *root,
u64 objectid, u64 offset, u64 bytenr);
int btrfs_copy_pinned(struct btrfs_root *root, struct extent_io_tree *copy);
struct btrfs_block_group_cache *btrfs_lookup_block_group(
struct btrfs_fs_info *info,
u64 bytenr);
void btrfs_put_block_group(struct btrfs_block_group_cache *cache);
u64 btrfs_find_block_group(struct btrfs_root *root,
u64 search_start, u64 search_hint, int owner);
struct extent_buffer *btrfs_alloc_free_block(struct btrfs_trans_handle *trans,
struct btrfs_root *root,
u32 blocksize, u64 parent,
u64 root_objectid,
u64 ref_generation,
int level,
u64 hint,
u64 empty_size);
struct btrfs_root *root, u32 blocksize,
u64 parent, u64 root_objectid,
struct btrfs_disk_key *key, int level,
u64 hint, u64 empty_size);
struct extent_buffer *btrfs_init_new_buffer(struct btrfs_trans_handle *trans,
struct btrfs_root *root,
u64 bytenr, u32 blocksize,
int level);
int btrfs_alloc_extent(struct btrfs_trans_handle *trans,
struct btrfs_root *root,
u64 num_bytes, u64 parent, u64 min_bytes,
u64 root_objectid, u64 ref_generation,
u64 owner, u64 empty_size, u64 hint_byte,
u64 search_end, struct btrfs_key *ins, u64 data);
int btrfs_alloc_reserved_extent(struct btrfs_trans_handle *trans,
struct btrfs_root *root, u64 parent,
u64 root_objectid, u64 ref_generation,
u64 owner, struct btrfs_key *ins);
int btrfs_alloc_logged_extent(struct btrfs_trans_handle *trans,
struct btrfs_root *root, u64 parent,
u64 root_objectid, u64 ref_generation,
u64 owner, struct btrfs_key *ins);
int btrfs_alloc_reserved_file_extent(struct btrfs_trans_handle *trans,
struct btrfs_root *root,
u64 root_objectid, u64 owner,
u64 offset, struct btrfs_key *ins);
int btrfs_alloc_logged_file_extent(struct btrfs_trans_handle *trans,
struct btrfs_root *root,
u64 root_objectid, u64 owner, u64 offset,
struct btrfs_key *ins);
int btrfs_reserve_extent(struct btrfs_trans_handle *trans,
struct btrfs_root *root,
u64 num_bytes, u64 min_alloc_size,
@ -1830,18 +1938,18 @@ int btrfs_reserve_extent(struct btrfs_trans_handle *trans,
u64 search_end, struct btrfs_key *ins,
u64 data);
int btrfs_inc_ref(struct btrfs_trans_handle *trans, struct btrfs_root *root,
struct extent_buffer *orig_buf, struct extent_buffer *buf,
u32 *nr_extents);
int btrfs_cache_ref(struct btrfs_trans_handle *trans, struct btrfs_root *root,
struct extent_buffer *buf, u32 nr_extents);
int btrfs_update_ref(struct btrfs_trans_handle *trans,
struct btrfs_root *root, struct extent_buffer *orig_buf,
struct extent_buffer *buf, int start_slot, int nr);
struct extent_buffer *buf, int full_backref);
int btrfs_dec_ref(struct btrfs_trans_handle *trans, struct btrfs_root *root,
struct extent_buffer *buf, int full_backref);
int btrfs_set_disk_extent_flags(struct btrfs_trans_handle *trans,
struct btrfs_root *root,
u64 bytenr, u64 num_bytes, u64 flags,
int is_data);
int btrfs_free_extent(struct btrfs_trans_handle *trans,
struct btrfs_root *root,
u64 bytenr, u64 num_bytes, u64 parent,
u64 root_objectid, u64 ref_generation,
u64 owner_objectid, int pin);
u64 root_objectid, u64 owner, u64 offset);
int btrfs_free_reserved_extent(struct btrfs_root *root, u64 start, u64 len);
int btrfs_finish_extent_commit(struct btrfs_trans_handle *trans,
struct btrfs_root *root,
@ -1849,13 +1957,8 @@ int btrfs_finish_extent_commit(struct btrfs_trans_handle *trans,
int btrfs_inc_extent_ref(struct btrfs_trans_handle *trans,
struct btrfs_root *root,
u64 bytenr, u64 num_bytes, u64 parent,
u64 root_objectid, u64 ref_generation,
u64 owner_objectid);
int btrfs_update_extent_ref(struct btrfs_trans_handle *trans,
struct btrfs_root *root, u64 bytenr, u64 num_bytes,
u64 orig_parent, u64 parent,
u64 root_objectid, u64 ref_generation,
u64 owner_objectid);
u64 root_objectid, u64 owner, u64 offset);
int btrfs_write_dirty_block_groups(struct btrfs_trans_handle *trans,
struct btrfs_root *root);
int btrfs_extent_readonly(struct btrfs_root *root, u64 bytenr);
@ -1867,16 +1970,9 @@ int btrfs_make_block_group(struct btrfs_trans_handle *trans,
u64 size);
int btrfs_remove_block_group(struct btrfs_trans_handle *trans,
struct btrfs_root *root, u64 group_start);
int btrfs_relocate_block_group(struct btrfs_root *root, u64 group_start);
int btrfs_free_reloc_root(struct btrfs_trans_handle *trans,
struct btrfs_root *root);
int btrfs_drop_dead_reloc_roots(struct btrfs_root *root);
int btrfs_reloc_tree_cache_ref(struct btrfs_trans_handle *trans,
struct btrfs_root *root,
struct extent_buffer *buf, u64 orig_start);
int btrfs_add_dead_reloc_root(struct btrfs_root *root);
int btrfs_cleanup_reloc_trees(struct btrfs_root *root);
int btrfs_reloc_clone_csums(struct inode *inode, u64 file_pos, u64 len);
int btrfs_prepare_block_group_relocation(struct btrfs_root *root,
struct btrfs_block_group_cache *group);
u64 btrfs_reduce_alloc_profile(struct btrfs_root *root, u64 flags);
void btrfs_set_inode_space_info(struct btrfs_root *root, struct inode *ionde);
void btrfs_clear_space_info_full(struct btrfs_fs_info *info);
@ -1891,13 +1987,12 @@ void btrfs_delalloc_reserve_space(struct btrfs_root *root, struct inode *inode,
void btrfs_delalloc_free_space(struct btrfs_root *root, struct inode *inode,
u64 bytes);
/* ctree.c */
int btrfs_bin_search(struct extent_buffer *eb, struct btrfs_key *key,
int level, int *slot);
int btrfs_comp_cpu_keys(struct btrfs_key *k1, struct btrfs_key *k2);
int btrfs_previous_item(struct btrfs_root *root,
struct btrfs_path *path, u64 min_objectid,
int type);
int btrfs_merge_path(struct btrfs_trans_handle *trans,
struct btrfs_root *root,
struct btrfs_key *node_keys,
u64 *nodes, int lowest_level);
int btrfs_set_item_key_safe(struct btrfs_trans_handle *trans,
struct btrfs_root *root, struct btrfs_path *path,
struct btrfs_key *new_key);
@ -1918,6 +2013,8 @@ int btrfs_copy_root(struct btrfs_trans_handle *trans,
struct btrfs_root *root,
struct extent_buffer *buf,
struct extent_buffer **cow_ret, u64 new_root_objectid);
int btrfs_block_can_be_shared(struct btrfs_root *root,
struct extent_buffer *buf);
int btrfs_extend_item(struct btrfs_trans_handle *trans, struct btrfs_root
*root, struct btrfs_path *path, u32 data_size);
int btrfs_truncate_item(struct btrfs_trans_handle *trans,
@ -1944,9 +2041,6 @@ void btrfs_unlock_up_safe(struct btrfs_path *p, int level);
int btrfs_del_items(struct btrfs_trans_handle *trans, struct btrfs_root *root,
struct btrfs_path *path, int slot, int nr);
int btrfs_del_leaf(struct btrfs_trans_handle *trans,
struct btrfs_root *root,
struct btrfs_path *path, u64 bytenr);
static inline int btrfs_del_item(struct btrfs_trans_handle *trans,
struct btrfs_root *root,
struct btrfs_path *path)
@ -2005,8 +2099,9 @@ int btrfs_find_last_root(struct btrfs_root *root, u64 objectid, struct
btrfs_root_item *item, struct btrfs_key *key);
int btrfs_search_root(struct btrfs_root *root, u64 search_start,
u64 *found_objectid);
int btrfs_find_dead_roots(struct btrfs_root *root, u64 objectid,
struct btrfs_root *latest_root);
int btrfs_find_dead_roots(struct btrfs_root *root, u64 objectid);
int btrfs_set_root_node(struct btrfs_root_item *item,
struct extent_buffer *node);
/* dir-item.c */
int btrfs_insert_dir_item(struct btrfs_trans_handle *trans,
struct btrfs_root *root, const char *name,
@ -2139,7 +2234,6 @@ int btrfs_page_mkwrite(struct vm_area_struct *vma, struct vm_fault *vmf);
int btrfs_readpage(struct file *file, struct page *page);
void btrfs_delete_inode(struct inode *inode);
void btrfs_put_inode(struct inode *inode);
void btrfs_read_locked_inode(struct inode *inode);
int btrfs_write_inode(struct inode *inode, int wait);
void btrfs_dirty_inode(struct inode *inode);
struct inode *btrfs_alloc_inode(struct super_block *sb);
@ -2147,12 +2241,8 @@ void btrfs_destroy_inode(struct inode *inode);
int btrfs_init_cachep(void);
void btrfs_destroy_cachep(void);
long btrfs_ioctl_trans_end(struct file *file);
struct inode *btrfs_ilookup(struct super_block *s, u64 objectid,
struct btrfs_root *root, int wait);
struct inode *btrfs_iget_locked(struct super_block *s, u64 objectid,
struct btrfs_root *root);
struct inode *btrfs_iget(struct super_block *s, struct btrfs_key *location,
struct btrfs_root *root, int *is_new);
struct btrfs_root *root);
int btrfs_commit_write(struct file *file, struct page *page,
unsigned from, unsigned to);
struct extent_map *btrfs_get_extent(struct inode *inode, struct page *page,
@ -2209,4 +2299,12 @@ int btrfs_check_acl(struct inode *inode, int mask);
int btrfs_init_acl(struct inode *inode, struct inode *dir);
int btrfs_acl_chmod(struct inode *inode);
/* relocation.c */
int btrfs_relocate_block_group(struct btrfs_root *root, u64 group_start);
int btrfs_init_reloc_root(struct btrfs_trans_handle *trans,
struct btrfs_root *root);
int btrfs_update_reloc_root(struct btrfs_trans_handle *trans,
struct btrfs_root *root);
int btrfs_recover_relocation(struct btrfs_root *root);
int btrfs_reloc_clone_csums(struct inode *inode, u64 file_pos, u64 len);
#endif

527
fs/btrfs/delayed-ref.c
View file

@ -29,27 +29,87 @@
* add extents in the middle of btrfs_search_slot, and it allows
* us to buffer up frequently modified backrefs in an rb tree instead
* of hammering updates on the extent allocation tree.
*
* Right now this code is only used for reference counted trees, but
* the long term goal is to get rid of the similar code for delayed
* extent tree modifications.
*/
/*
* entries in the rb tree are ordered by the byte number of the extent
* and by the byte number of the parent block.
* compare two delayed tree backrefs with same bytenr and type
*/
static int comp_entry(struct btrfs_delayed_ref_node *ref,
u64 bytenr, u64 parent)
static int comp_tree_refs(struct btrfs_delayed_tree_ref *ref2,
struct btrfs_delayed_tree_ref *ref1)
{
if (bytenr < ref->bytenr)
if (ref1->node.type == BTRFS_TREE_BLOCK_REF_KEY) {
if (ref1->root < ref2->root)
return -1;
if (ref1->root > ref2->root)
return 1;
} else {
if (ref1->parent < ref2->parent)
return -1;
if (ref1->parent > ref2->parent)
return 1;
}
return 0;
}
/*
* compare two delayed data backrefs with same bytenr and type
*/
static int comp_data_refs(struct btrfs_delayed_data_ref *ref2,
struct btrfs_delayed_data_ref *ref1)
{
if (ref1->node.type == BTRFS_EXTENT_DATA_REF_KEY) {
if (ref1->root < ref2->root)
return -1;
if (ref1->root > ref2->root)
return 1;
if (ref1->objectid < ref2->objectid)
return -1;
if (ref1->objectid > ref2->objectid)
return 1;
if (ref1->offset < ref2->offset)
return -1;
if (ref1->offset > ref2->offset)
return 1;
} else {
if (ref1->parent < ref2->parent)
return -1;
if (ref1->parent > ref2->parent)
return 1;
}
return 0;
}
/*
* entries in the rb tree are ordered by the byte number of the extent,
* type of the delayed backrefs and content of delayed backrefs.
*/
static int comp_entry(struct btrfs_delayed_ref_node *ref2,
struct btrfs_delayed_ref_node *ref1)
{
if (ref1->bytenr < ref2->bytenr)
return -1;
if (bytenr > ref->bytenr)
if (ref1->bytenr > ref2->bytenr)
return 1;
if (parent < ref->parent)
if (ref1->is_head && ref2->is_head)
return 0;
if (ref2->is_head)
return -1;
if (parent > ref->parent)
if (ref1->is_head)
return 1;
if (ref1->type < ref2->type)
return -1;
if (ref1->type > ref2->type)
return 1;
if (ref1->type == BTRFS_TREE_BLOCK_REF_KEY ||
ref1->type == BTRFS_SHARED_BLOCK_REF_KEY) {
return comp_tree_refs(btrfs_delayed_node_to_tree_ref(ref2),
btrfs_delayed_node_to_tree_ref(ref1));
} else if (ref1->type == BTRFS_EXTENT_DATA_REF_KEY ||
ref1->type == BTRFS_SHARED_DATA_REF_KEY) {
return comp_data_refs(btrfs_delayed_node_to_data_ref(ref2),
btrfs_delayed_node_to_data_ref(ref1));
}
BUG();
return 0;
}
@ -59,20 +119,21 @@ static int comp_entry(struct btrfs_delayed_ref_node *ref,
* inserted.
*/
static struct btrfs_delayed_ref_node *tree_insert(struct rb_root *root,
u64 bytenr, u64 parent,
struct rb_node *node)
{
struct rb_node **p = &root->rb_node;
struct rb_node *parent_node = NULL;
struct btrfs_delayed_ref_node *entry;
struct btrfs_delayed_ref_node *ins;
int cmp;
ins = rb_entry(node, struct btrfs_delayed_ref_node, rb_node);
while (*p) {
parent_node = *p;
entry = rb_entry(parent_node, struct btrfs_delayed_ref_node,
rb_node);
cmp = comp_entry(entry, bytenr, parent);
cmp = comp_entry(entry, ins);
if (cmp < 0)
p = &(*p)->rb_left;
else if (cmp > 0)
@ -81,18 +142,17 @@ static struct btrfs_delayed_ref_node *tree_insert(struct rb_root *root,
return entry;
}
entry = rb_entry(node, struct btrfs_delayed_ref_node, rb_node);
rb_link_node(node, parent_node, p);
rb_insert_color(node, root);
return NULL;
}
/*
* find an entry based on (bytenr,parent). This returns the delayed
* ref if it was able to find one, or NULL if nothing was in that spot
* find an head entry based on bytenr. This returns the delayed ref
* head if it was able to find one, or NULL if nothing was in that spot
*/
static struct btrfs_delayed_ref_node *tree_search(struct rb_root *root,
u64 bytenr, u64 parent,
static struct btrfs_delayed_ref_node *find_ref_head(struct rb_root *root,
u64 bytenr,
struct btrfs_delayed_ref_node **last)
{
struct rb_node *n = root->rb_node;
@ -105,7 +165,15 @@ static struct btrfs_delayed_ref_node *tree_search(struct rb_root *root,
if (last)
*last = entry;
cmp = comp_entry(entry, bytenr, parent);
if (bytenr < entry->bytenr)
cmp = -1;
else if (bytenr > entry->bytenr)
cmp = 1;
else if (!btrfs_delayed_ref_is_head(entry))
cmp = 1;
else
cmp = 0;
if (cmp < 0)
n = n->rb_left;
else if (cmp > 0)
@ -154,7 +222,7 @@ int btrfs_find_ref_cluster(struct btrfs_trans_handle *trans,
node = rb_first(&delayed_refs->root);
} else {
ref = NULL;
tree_search(&delayed_refs->root, start, (u64)-1, &ref);
find_ref_head(&delayed_refs->root, start, &ref);
if (ref) {
struct btrfs_delayed_ref_node *tmp;
@ -234,7 +302,7 @@ int btrfs_delayed_ref_pending(struct btrfs_trans_handle *trans, u64 bytenr)
delayed_refs = &trans->transaction->delayed_refs;
spin_lock(&delayed_refs->lock);
ref = tree_search(&delayed_refs->root, bytenr, (u64)-1, NULL);
ref = find_ref_head(&delayed_refs->root, bytenr, NULL);
if (ref) {
prev_node = rb_prev(&ref->rb_node);
if (!prev_node)
@ -250,25 +318,28 @@ out:
}
/*
* helper function to lookup reference count
* helper function to lookup reference count and flags of extent.
*
* the head node for delayed ref is used to store the sum of all the
* reference count modifications queued up in the rbtree. This way you
* can check to see what the reference count would be if all of the
* delayed refs are processed.
* reference count modifications queued up in the rbtree. the head
* node may also store the extent flags to set. This way you can check
* to see what the reference count and extent flags would be if all of
* the delayed refs are not processed.
*/
int btrfs_lookup_extent_ref(struct btrfs_trans_handle *trans,
struct btrfs_root *root, u64 bytenr,
u64 num_bytes, u32 *refs)
int btrfs_lookup_extent_info(struct btrfs_trans_handle *trans,
struct btrfs_root *root, u64 bytenr,
u64 num_bytes, u64 *refs, u64 *flags)
{
struct btrfs_delayed_ref_node *ref;
struct btrfs_delayed_ref_head *head;
struct btrfs_delayed_ref_root *delayed_refs;
struct btrfs_path *path;
struct extent_buffer *leaf;
struct btrfs_extent_item *ei;
struct extent_buffer *leaf;
struct btrfs_key key;
u32 num_refs;
u32 item_size;
u64 num_refs;
u64 extent_flags;
int ret;
path = btrfs_alloc_path();
@ -287,37 +358,60 @@ again:
if (ret == 0) {
leaf = path->nodes[0];
ei = btrfs_item_ptr(leaf, path->slots[0],
struct btrfs_extent_item);
num_refs = btrfs_extent_refs(leaf, ei);
item_size = btrfs_item_size_nr(leaf, path->slots[0]);
if (item_size >= sizeof(*ei)) {
ei = btrfs_item_ptr(leaf, path->slots[0],
struct btrfs_extent_item);
num_refs = btrfs_extent_refs(leaf, ei);
extent_flags = btrfs_extent_flags(leaf, ei);
} else {
#ifdef BTRFS_COMPAT_EXTENT_TREE_V0
struct btrfs_extent_item_v0 *ei0;
BUG_ON(item_size != sizeof(*ei0));
ei0 = btrfs_item_ptr(leaf, path->slots[0],
struct btrfs_extent_item_v0);
num_refs = btrfs_extent_refs_v0(leaf, ei0);
/* FIXME: this isn't correct for data */
extent_flags = BTRFS_BLOCK_FLAG_FULL_BACKREF;
#else
BUG();
#endif
}
BUG_ON(num_refs == 0);
} else {
num_refs = 0;
extent_flags = 0;
ret = 0;
}
spin_lock(&delayed_refs->lock);
ref = tree_search(&delayed_refs->root, bytenr, (u64)-1, NULL);
ref = find_ref_head(&delayed_refs->root, bytenr, NULL);
if (ref) {
head = btrfs_delayed_node_to_head(ref);
if (mutex_trylock(&head->mutex)) {
num_refs += ref->ref_mod;
if (!mutex_trylock(&head->mutex)) {
atomic_inc(&ref->refs);
spin_unlock(&delayed_refs->lock);
btrfs_release_path(root->fs_info->extent_root, path);
mutex_lock(&head->mutex);
mutex_unlock(&head->mutex);
*refs = num_refs;
goto out;
btrfs_put_delayed_ref(ref);
goto again;
}
if (head->extent_op && head->extent_op->update_flags)
extent_flags |= head->extent_op->flags_to_set;
else
BUG_ON(num_refs == 0);
atomic_inc(&ref->refs);
spin_unlock(&delayed_refs->lock);
btrfs_release_path(root->fs_info->extent_root, path);
mutex_lock(&head->mutex);
num_refs += ref->ref_mod;
mutex_unlock(&head->mutex);
btrfs_put_delayed_ref(ref);
goto again;
} else {
*refs = num_refs;
}
WARN_ON(num_refs == 0);
if (refs)
*refs = num_refs;
if (flags)
*flags = extent_flags;
out:
spin_unlock(&delayed_refs->lock);
btrfs_free_path(path);
@ -338,16 +432,7 @@ update_existing_ref(struct btrfs_trans_handle *trans,
struct btrfs_delayed_ref_node *existing,
struct btrfs_delayed_ref_node *update)
{
struct btrfs_delayed_ref *existing_ref;
struct btrfs_delayed_ref *ref;
existing_ref = btrfs_delayed_node_to_ref(existing);
ref = btrfs_delayed_node_to_ref(update);
if (ref->pin)
existing_ref->pin = 1;
if (ref->action != existing_ref->action) {
if (update->action != existing->action) {
/*
* this is effectively undoing either an add or a
* drop. We decrement the ref_mod, and if it goes
@ -363,20 +448,13 @@ update_existing_ref(struct btrfs_trans_handle *trans,
delayed_refs->num_entries--;
if (trans->delayed_ref_updates)
trans->delayed_ref_updates--;
} else {
WARN_ON(existing->type == BTRFS_TREE_BLOCK_REF_KEY ||
existing->type == BTRFS_SHARED_BLOCK_REF_KEY);
}
} else {
if (existing_ref->action == BTRFS_ADD_DELAYED_REF) {
/* if we're adding refs, make sure all the
* details match up. The extent could
* have been totally freed and reallocated
* by a different owner before the delayed
* ref entries were removed.
*/
existing_ref->owner_objectid = ref->owner_objectid;
existing_ref->generation = ref->generation;
existing_ref->root = ref->root;
existing->num_bytes = update->num_bytes;
}
WARN_ON(existing->type == BTRFS_TREE_BLOCK_REF_KEY ||
existing->type == BTRFS_SHARED_BLOCK_REF_KEY);
/*
* the action on the existing ref matches
* the action on the ref we're trying to add.
@ -401,6 +479,7 @@ update_existing_head_ref(struct btrfs_delayed_ref_node *existing,
existing_ref = btrfs_delayed_node_to_head(existing);
ref = btrfs_delayed_node_to_head(update);
BUG_ON(existing_ref->is_data != ref->is_data);
if (ref->must_insert_reserved) {
/* if the extent was freed and then
@ -420,6 +499,24 @@ update_existing_head_ref(struct btrfs_delayed_ref_node *existing,
}
if (ref->extent_op) {
if (!existing_ref->extent_op) {
existing_ref->extent_op = ref->extent_op;
} else {
if (ref->extent_op->update_key) {
memcpy(&existing_ref->extent_op->key,
&ref->extent_op->key,
sizeof(ref->extent_op->key));
existing_ref->extent_op->update_key = 1;
}
if (ref->extent_op->update_flags) {
existing_ref->extent_op->flags_to_set |=
ref->extent_op->flags_to_set;
existing_ref->extent_op->update_flags = 1;
}
kfree(ref->extent_op);
}
}
/*
* update the reference mod on the head to reflect this new operation
*/
@ -427,19 +524,16 @@ update_existing_head_ref(struct btrfs_delayed_ref_node *existing,
}
/*
* helper function to actually insert a delayed ref into the rbtree.
* helper function to actually insert a head node into the rbtree.
* this does all the dirty work in terms of maintaining the correct
* overall modification count in the head node and properly dealing
* with updating existing nodes as new modifications are queued.
* overall modification count.
*/
static noinline int __btrfs_add_delayed_ref(struct btrfs_trans_handle *trans,
struct btrfs_delayed_ref_node *ref,
u64 bytenr, u64 num_bytes, u64 parent, u64 ref_root,
u64 ref_generation, u64 owner_objectid, int action,
int pin)
static noinline int add_delayed_ref_head(struct btrfs_trans_handle *trans,
struct btrfs_delayed_ref_node *ref,
u64 bytenr, u64 num_bytes,
int action, int is_data)
{
struct btrfs_delayed_ref_node *existing;
struct btrfs_delayed_ref *full_ref;
struct btrfs_delayed_ref_head *head_ref = NULL;
struct btrfs_delayed_ref_root *delayed_refs;
int count_mod = 1;
@ -449,12 +543,10 @@ static noinline int __btrfs_add_delayed_ref(struct btrfs_trans_handle *trans,
* the head node stores the sum of all the mods, so dropping a ref
* should drop the sum in the head node by one.
*/
if (parent == (u64)-1) {
if (action == BTRFS_DROP_DELAYED_REF)
count_mod = -1;
else if (action == BTRFS_UPDATE_DELAYED_HEAD)
count_mod = 0;
}
if (action == BTRFS_UPDATE_DELAYED_HEAD)
count_mod = 0;
else if (action == BTRFS_DROP_DELAYED_REF)
count_mod = -1;
/*
* BTRFS_ADD_DELAYED_EXTENT means that we need to update
@ -467,57 +559,42 @@ static noinline int __btrfs_add_delayed_ref(struct btrfs_trans_handle *trans,
* Once we record must_insert_reserved, switch the action to
* BTRFS_ADD_DELAYED_REF because other special casing is not required.
*/
if (action == BTRFS_ADD_DELAYED_EXTENT) {
if (action == BTRFS_ADD_DELAYED_EXTENT)
must_insert_reserved = 1;
action = BTRFS_ADD_DELAYED_REF;
} else {
else
must_insert_reserved = 0;
}
delayed_refs = &trans->transaction->delayed_refs;
/* first set the basic ref node struct up */
atomic_set(&ref->refs, 1);
ref->bytenr = bytenr;
ref->parent = parent;
ref->ref_mod = count_mod;
ref->in_tree = 1;
ref->num_bytes = num_bytes;
ref->ref_mod = count_mod;
ref->type = 0;
ref->action = 0;
ref->is_head = 1;
ref->in_tree = 1;
if (btrfs_delayed_ref_is_head(ref)) {
head_ref = btrfs_delayed_node_to_head(ref);
head_ref->must_insert_reserved = must_insert_reserved;
INIT_LIST_HEAD(&head_ref->cluster);
mutex_init(&head_ref->mutex);
} else {
full_ref = btrfs_delayed_node_to_ref(ref);
full_ref->root = ref_root;
full_ref->generation = ref_generation;
full_ref->owner_objectid = owner_objectid;
full_ref->pin = pin;
full_ref->action = action;
}
head_ref = btrfs_delayed_node_to_head(ref);
head_ref->must_insert_reserved = must_insert_reserved;
head_ref->is_data = is_data;
existing = tree_insert(&delayed_refs->root, bytenr,
parent, &ref->rb_node);
INIT_LIST_HEAD(&head_ref->cluster);
mutex_init(&head_ref->mutex);
existing = tree_insert(&delayed_refs->root, &ref->rb_node);
if (existing) {
if (btrfs_delayed_ref_is_head(ref))
update_existing_head_ref(existing, ref);
else
update_existing_ref(trans, delayed_refs, existing, ref);
update_existing_head_ref(existing, ref);
/*
* we've updated the existing ref, free the newly
* allocated ref
*/
kfree(ref);
} else {
if (btrfs_delayed_ref_is_head(ref)) {
delayed_refs->num_heads++;
delayed_refs->num_heads_ready++;
}
delayed_refs->num_heads++;
delayed_refs->num_heads_ready++;
delayed_refs->num_entries++;
trans->delayed_ref_updates++;
}
@ -525,37 +602,139 @@ static noinline int __btrfs_add_delayed_ref(struct btrfs_trans_handle *trans,
}
/*
* add a delayed ref to the tree. This does all of the accounting required
* helper to insert a delayed tree ref into the rbtree.
*/
static noinline int add_delayed_tree_ref(struct btrfs_trans_handle *trans,
struct btrfs_delayed_ref_node *ref,
u64 bytenr, u64 num_bytes, u64 parent,
u64 ref_root, int level, int action)
{
struct btrfs_delayed_ref_node *existing;
struct btrfs_delayed_tree_ref *full_ref;
struct btrfs_delayed_ref_root *delayed_refs;
if (action == BTRFS_ADD_DELAYED_EXTENT)
action = BTRFS_ADD_DELAYED_REF;
delayed_refs = &trans->transaction->delayed_refs;
/* first set the basic ref node struct up */
atomic_set(&ref->refs, 1);
ref->bytenr = bytenr;
ref->num_bytes = num_bytes;
ref->ref_mod = 1;
ref->action = action;
ref->is_head = 0;
ref->in_tree = 1;
full_ref = btrfs_delayed_node_to_tree_ref(ref);
if (parent) {
full_ref->parent = parent;
ref->type = BTRFS_SHARED_BLOCK_REF_KEY;
} else {
full_ref->root = ref_root;
ref->type = BTRFS_TREE_BLOCK_REF_KEY;
}
full_ref->level = level;
existing = tree_insert(&delayed_refs->root, &ref->rb_node);
if (existing) {
update_existing_ref(trans, delayed_refs, existing, ref);
/*
* we've updated the existing ref, free the newly
* allocated ref
*/
kfree(ref);
} else {
delayed_refs->num_entries++;
trans->delayed_ref_updates++;
}
return 0;
}
/*
* helper to insert a delayed data ref into the rbtree.
*/
static noinline int add_delayed_data_ref(struct btrfs_trans_handle *trans,
struct btrfs_delayed_ref_node *ref,
u64 bytenr, u64 num_bytes, u64 parent,
u64 ref_root, u64 owner, u64 offset,
int action)
{
struct btrfs_delayed_ref_node *existing;
struct btrfs_delayed_data_ref *full_ref;
struct btrfs_delayed_ref_root *delayed_refs;
if (action == BTRFS_ADD_DELAYED_EXTENT)
action = BTRFS_ADD_DELAYED_REF;
delayed_refs = &trans->transaction->delayed_refs;
/* first set the basic ref node struct up */
atomic_set(&ref->refs, 1);
ref->bytenr = bytenr;
ref->num_bytes = num_bytes;
ref->ref_mod = 1;
ref->action = action;
ref->is_head = 0;
ref->in_tree = 1;
full_ref = btrfs_delayed_node_to_data_ref(ref);
if (parent) {
full_ref->parent = parent;
ref->type = BTRFS_SHARED_DATA_REF_KEY;
} else {
full_ref->root = ref_root;
ref->type = BTRFS_EXTENT_DATA_REF_KEY;
}
full_ref->objectid = owner;
full_ref->offset = offset;
existing = tree_insert(&delayed_refs->root, &ref->rb_node);
if (existing) {
update_existing_ref(trans, delayed_refs, existing, ref);
/*
* we've updated the existing ref, free the newly
* allocated ref
*/
kfree(ref);
} else {
delayed_refs->num_entries++;
trans->delayed_ref_updates++;
}
return 0;
}
/*
* add a delayed tree ref. This does all of the accounting required
* to make sure the delayed ref is eventually processed before this
* transaction commits.
*/
int btrfs_add_delayed_ref(struct btrfs_trans_handle *trans,
u64 bytenr, u64 num_bytes, u64 parent, u64 ref_root,
u64 ref_generation, u64 owner_objectid, int action,
int pin)
int btrfs_add_delayed_tree_ref(struct btrfs_trans_handle *trans,
u64 bytenr, u64 num_bytes, u64 parent,
u64 ref_root, int level, int action,
struct btrfs_delayed_extent_op *extent_op)
{
struct btrfs_delayed_ref *ref;
struct btrfs_delayed_tree_ref *ref;
struct btrfs_delayed_ref_head *head_ref;
struct btrfs_delayed_ref_root *delayed_refs;
int ret;
BUG_ON(extent_op && extent_op->is_data);
ref = kmalloc(sizeof(*ref), GFP_NOFS);
if (!ref)
return -ENOMEM;
/*
* the parent = 0 case comes from cases where we don't actually
* know the parent yet. It will get updated later via a add/drop
* pair.
*/
if (parent == 0)
parent = bytenr;
head_ref = kmalloc(sizeof(*head_ref), GFP_NOFS);
if (!head_ref) {
kfree(ref);
return -ENOMEM;
}
head_ref->extent_op = extent_op;
delayed_refs = &trans->transaction->delayed_refs;
spin_lock(&delayed_refs->lock);
@ -563,18 +742,88 @@ int btrfs_add_delayed_ref(struct btrfs_trans_handle *trans,
* insert both the head node and the new ref without dropping
* the spin lock
*/
ret = __btrfs_add_delayed_ref(trans, &head_ref->node, bytenr, num_bytes,
(u64)-1, 0, 0, 0, action, pin);
ret = add_delayed_ref_head(trans, &head_ref->node, bytenr, num_bytes,
action, 0);
BUG_ON(ret);
ret = __btrfs_add_delayed_ref(trans, &ref->node, bytenr, num_bytes,
parent, ref_root, ref_generation,
owner_objectid, action, pin);
ret = add_delayed_tree_ref(trans, &ref->node, bytenr, num_bytes,
parent, ref_root, level, action);
BUG_ON(ret);
spin_unlock(&delayed_refs->lock);
return 0;
}
/*
* add a delayed data ref. it's similar to btrfs_add_delayed_tree_ref.
*/
int btrfs_add_delayed_data_ref(struct btrfs_trans_handle *trans,
u64 bytenr, u64 num_bytes,
u64 parent, u64 ref_root,
u64 owner, u64 offset, int action,
struct btrfs_delayed_extent_op *extent_op)
{
struct btrfs_delayed_data_ref *ref;
struct btrfs_delayed_ref_head *head_ref;
struct btrfs_delayed_ref_root *delayed_refs;
int ret;
BUG_ON(extent_op && !extent_op->is_data);
ref = kmalloc(sizeof(*ref), GFP_NOFS);
if (!ref)
return -ENOMEM;
head_ref = kmalloc(sizeof(*head_ref), GFP_NOFS);
if (!head_ref) {
kfree(ref);
return -ENOMEM;
}
head_ref->extent_op = extent_op;
delayed_refs = &trans->transaction->delayed_refs;
spin_lock(&delayed_refs->lock);
/*
* insert both the head node and the new ref without dropping
* the spin lock
*/
ret = add_delayed_ref_head(trans, &head_ref->node, bytenr, num_bytes,
action, 1);
BUG_ON(ret);
ret = add_delayed_data_ref(trans, &ref->node, bytenr, num_bytes,
parent, ref_root, owner, offset, action);
BUG_ON(ret);
spin_unlock(&delayed_refs->lock);
return 0;
}
int btrfs_add_delayed_extent_op(struct btrfs_trans_handle *trans,
u64 bytenr, u64 num_bytes,
struct btrfs_delayed_extent_op *extent_op)
{
struct btrfs_delayed_ref_head *head_ref;
struct btrfs_delayed_ref_root *delayed_refs;
int ret;
head_ref = kmalloc(sizeof(*head_ref), GFP_NOFS);
if (!head_ref)
return -ENOMEM;
head_ref->extent_op = extent_op;
delayed_refs = &trans->transaction->delayed_refs;
spin_lock(&delayed_refs->lock);
ret = add_delayed_ref_head(trans, &head_ref->node, bytenr,
num_bytes, BTRFS_UPDATE_DELAYED_HEAD,
extent_op->is_data);
BUG_ON(ret);
spin_unlock(&delayed_refs->lock);
return 0;
}
/*
* this does a simple search for the head node for a given extent.
* It must be called with the delayed ref spinlock held, and it returns
@ -587,7 +836,7 @@ btrfs_find_delayed_ref_head(struct btrfs_trans_handle *trans, u64 bytenr)
struct btrfs_delayed_ref_root *delayed_refs;
delayed_refs = &trans->transaction->delayed_refs;
ref = tree_search(&delayed_refs->root, bytenr, (u64)-1, NULL);
ref = find_ref_head(&delayed_refs->root, bytenr, NULL);
if (ref)
return btrfs_delayed_node_to_head(ref);
return NULL;
@ -603,6 +852,7 @@ btrfs_find_delayed_ref_head(struct btrfs_trans_handle *trans, u64 bytenr)
*
* It is the same as doing a ref add and delete in two separate calls.
*/
#if 0
int btrfs_update_delayed_ref(struct btrfs_trans_handle *trans,
u64 bytenr, u64 num_bytes, u64 orig_parent,
u64 parent, u64 orig_ref_root, u64 ref_root,
@ -666,3 +916,4 @@ int btrfs_update_delayed_ref(struct btrfs_trans_handle *trans,
spin_unlock(&delayed_refs->lock);
return 0;
}
#endif

85
fs/btrfs/delayed-ref.h
View file

@ -30,9 +30,6 @@ struct btrfs_delayed_ref_node {
/* the starting bytenr of the extent */
u64 bytenr;
/* the parent our backref will point to */
u64 parent;
/* the size of the extent */
u64 num_bytes;
@ -50,10 +47,21 @@ struct btrfs_delayed_ref_node {
*/
int ref_mod;
unsigned int action:8;
unsigned int type:8;
/* is this node still in the rbtree? */
unsigned int is_head:1;
unsigned int in_tree:1;
};
struct btrfs_delayed_extent_op {
struct btrfs_disk_key key;
u64 flags_to_set;
unsigned int update_key:1;
unsigned int update_flags:1;
unsigned int is_data:1;
};
/*
* the head refs are used to hold a lock on a given extent, which allows us
* to make sure that only one process is running the delayed refs
@ -71,6 +79,7 @@ struct btrfs_delayed_ref_head {
struct list_head cluster;
struct btrfs_delayed_extent_op *extent_op;
/*
* when a new extent is allocated, it is just reserved in memory
* The actual extent isn't inserted into the extent allocation tree
@ -84,27 +93,26 @@ struct btrfs_delayed_ref_head {
* the free has happened.
*/
unsigned int must_insert_reserved:1;
unsigned int is_data:1;
};
struct btrfs_delayed_ref {
struct btrfs_delayed_tree_ref {
struct btrfs_delayed_ref_node node;
union {
u64 root;
u64 parent;
};
int level;
};
/* the root objectid our ref will point to */
u64 root;
/* the generation for the backref */
u64 generation;
/* owner_objectid of the backref */
u64 owner_objectid;
/* operation done by this entry in the rbtree */
u8 action;
/* if pin == 1, when the extent is freed it will be pinned until
* transaction commit
*/
unsigned int pin:1;
struct btrfs_delayed_data_ref {
struct btrfs_delayed_ref_node node;
union {
u64 root;
u64 parent;
};
u64 objectid;
u64 offset;
};
struct btrfs_delayed_ref_root {
@ -143,17 +151,25 @@ static inline void btrfs_put_delayed_ref(struct btrfs_delayed_ref_node *ref)
}
}
int btrfs_add_delayed_ref(struct btrfs_trans_handle *trans,
u64 bytenr, u64 num_bytes, u64 parent, u64 ref_root,
u64 ref_generation, u64 owner_objectid, int action,
int pin);
int btrfs_add_delayed_tree_ref(struct btrfs_trans_handle *trans,
u64 bytenr, u64 num_bytes, u64 parent,
u64 ref_root, int level, int action,
struct btrfs_delayed_extent_op *extent_op);
int btrfs_add_delayed_data_ref(struct btrfs_trans_handle *trans,
u64 bytenr, u64 num_bytes,
u64 parent, u64 ref_root,
u64 owner, u64 offset, int action,
struct btrfs_delayed_extent_op *extent_op);
int btrfs_add_delayed_extent_op(struct btrfs_trans_handle *trans,
u64 bytenr, u64 num_bytes,
struct btrfs_delayed_extent_op *extent_op);
struct btrfs_delayed_ref_head *
btrfs_find_delayed_ref_head(struct btrfs_trans_handle *trans, u64 bytenr);
int btrfs_delayed_ref_pending(struct btrfs_trans_handle *trans, u64 bytenr);
int btrfs_lookup_extent_ref(struct btrfs_trans_handle *trans,
struct btrfs_root *root, u64 bytenr,
u64 num_bytes, u32 *refs);
int btrfs_lookup_extent_info(struct btrfs_trans_handle *trans,
struct btrfs_root *root, u64 bytenr,
u64 num_bytes, u64 *refs, u64 *flags);
int btrfs_update_delayed_ref(struct btrfs_trans_handle *trans,
u64 bytenr, u64 num_bytes, u64 orig_parent,
u64 parent, u64 orig_ref_root, u64 ref_root,
@ -169,18 +185,24 @@ int btrfs_find_ref_cluster(struct btrfs_trans_handle *trans,
*/
static int btrfs_delayed_ref_is_head(struct btrfs_delayed_ref_node *node)
{
return node->parent == (u64)-1;
return node->is_head;
}
/*
* helper functions to cast a node into its container
*/
static inline struct btrfs_delayed_ref *
btrfs_delayed_node_to_ref(struct btrfs_delayed_ref_node *node)
static inline struct btrfs_delayed_tree_ref *
btrfs_delayed_node_to_tree_ref(struct btrfs_delayed_ref_node *node)
{
WARN_ON(btrfs_delayed_ref_is_head(node));
return container_of(node, struct btrfs_delayed_ref, node);
return container_of(node, struct btrfs_delayed_tree_ref, node);
}
static inline struct btrfs_delayed_data_ref *
btrfs_delayed_node_to_data_ref(struct btrfs_delayed_ref_node *node)
{
WARN_ON(btrfs_delayed_ref_is_head(node));
return container_of(node, struct btrfs_delayed_data_ref, node);
}
static inline struct btrfs_delayed_ref_head *
@ -188,6 +210,5 @@ btrfs_delayed_node_to_head(struct btrfs_delayed_ref_node *node)
{
WARN_ON(!btrfs_delayed_ref_is_head(node));
return container_of(node, struct btrfs_delayed_ref_head, node);
}
#endif

95
fs/btrfs/disk-io.c
View file

@ -36,7 +36,6 @@
#include "print-tree.h"
#include "async-thread.h"
#include "locking.h"
#include "ref-cache.h"
#include "tree-log.h"
#include "free-space-cache.h"
@ -884,7 +883,6 @@ static int __setup_root(u32 nodesize, u32 leafsize, u32 sectorsize,
{
root->node = NULL;
root->commit_root = NULL;
root->ref_tree = NULL;
root->sectorsize = sectorsize;
root->nodesize = nodesize;
root->leafsize = leafsize;
@ -899,12 +897,14 @@ static int __setup_root(u32 nodesize, u32 leafsize, u32 sectorsize,
root->last_inode_alloc = 0;
root->name = NULL;
root->in_sysfs = 0;
root->inode_tree.rb_node = NULL;
INIT_LIST_HEAD(&root->dirty_list);
INIT_LIST_HEAD(&root->orphan_list);
INIT_LIST_HEAD(&root->dead_list);
INIT_LIST_HEAD(&root->root_list);
spin_lock_init(&root->node_lock);
spin_lock_init(&root->list_lock);
spin_lock_init(&root->inode_lock);
mutex_init(&root->objectid_mutex);
mutex_init(&root->log_mutex);
init_waitqueue_head(&root->log_writer_wait);
@ -918,9 +918,6 @@ static int __setup_root(u32 nodesize, u32 leafsize, u32 sectorsize,
extent_io_tree_init(&root->dirty_log_pages,
fs_info->btree_inode->i_mapping, GFP_NOFS);
btrfs_leaf_ref_tree_init(&root->ref_tree_struct);
root->ref_tree = &root->ref_tree_struct;
memset(&root->root_key, 0, sizeof(root->root_key));
memset(&root->root_item, 0, sizeof(root->root_item));
memset(&root->defrag_progress, 0, sizeof(root->defrag_progress));
@ -959,6 +956,7 @@ static int find_and_setup_root(struct btrfs_root *tree_root,
blocksize = btrfs_level_size(root, btrfs_root_level(&root->root_item));
root->node = read_tree_block(root, btrfs_root_bytenr(&root->root_item),
blocksize, generation);
root->commit_root = btrfs_root_node(root);
BUG_ON(!root->node);
return 0;
}
@ -1025,20 +1023,19 @@ static struct btrfs_root *alloc_log_tree(struct btrfs_trans_handle *trans,
*/
root->ref_cows = 0;
leaf = btrfs_alloc_free_block(trans, root, root->leafsize,
0, BTRFS_TREE_LOG_OBJECTID,
trans->transid, 0, 0, 0);
leaf = btrfs_alloc_free_block(trans, root, root->leafsize, 0,
BTRFS_TREE_LOG_OBJECTID, NULL, 0, 0, 0);
if (IS_ERR(leaf)) {
kfree(root);
return ERR_CAST(leaf);
}
memset_extent_buffer(leaf, 0, 0, sizeof(struct btrfs_header));
btrfs_set_header_bytenr(leaf, leaf->start);
btrfs_set_header_generation(leaf, trans->transid);
btrfs_set_header_backref_rev(leaf, BTRFS_MIXED_BACKREF_REV);
btrfs_set_header_owner(leaf, BTRFS_TREE_LOG_OBJECTID);
root->node = leaf;
btrfs_set_header_nritems(root->node, 0);
btrfs_set_header_level(root->node, 0);
btrfs_set_header_bytenr(root->node, root->node->start);
btrfs_set_header_generation(root->node, trans->transid);
btrfs_set_header_owner(root->node, BTRFS_TREE_LOG_OBJECTID);
write_extent_buffer(root->node, root->fs_info->fsid,
(unsigned long)btrfs_header_fsid(root->node),
@ -1081,8 +1078,7 @@ int btrfs_add_log_tree(struct btrfs_trans_handle *trans,
inode_item->nbytes = cpu_to_le64(root->leafsize);
inode_item->mode = cpu_to_le32(S_IFDIR | 0755);
btrfs_set_root_bytenr(&log_root->root_item, log_root->node->start);
btrfs_set_root_generation(&log_root->root_item, trans->transid);
btrfs_set_root_node(&log_root->root_item, log_root->node);
WARN_ON(root->log_root);
root->log_root = log_root;
@ -1144,6 +1140,7 @@ out:
blocksize = btrfs_level_size(root, btrfs_root_level(&root->root_item));
root->node = read_tree_block(root, btrfs_root_bytenr(&root->root_item),
blocksize, generation);
root->commit_root = btrfs_root_node(root);
BUG_ON(!root->node);
insert:
if (location->objectid != BTRFS_TREE_LOG_OBJECTID) {
@ -1210,7 +1207,7 @@ struct btrfs_root *btrfs_read_fs_root_no_name(struct btrfs_fs_info *fs_info,
}
if (!(fs_info->sb->s_flags & MS_RDONLY)) {
ret = btrfs_find_dead_roots(fs_info->tree_root,
root->root_key.objectid, root);
root->root_key.objectid);
BUG_ON(ret);
btrfs_orphan_cleanup(root);
}
@ -1569,8 +1566,6 @@ struct btrfs_root *open_ctree(struct super_block *sb,
atomic_set(&fs_info->async_delalloc_pages, 0);
atomic_set(&fs_info->async_submit_draining, 0);
atomic_set(&fs_info->nr_async_bios, 0);
atomic_set(&fs_info->throttles, 0);
atomic_set(&fs_info->throttle_gen, 0);
fs_info->sb = sb;
fs_info->max_extent = (u64)-1;
fs_info->max_inline = 8192 * 1024;
@ -1598,6 +1593,7 @@ struct btrfs_root *open_ctree(struct super_block *sb,
fs_info->btree_inode->i_mapping->a_ops = &btree_aops;
fs_info->btree_inode->i_mapping->backing_dev_info = &fs_info->bdi;
RB_CLEAR_NODE(&BTRFS_I(fs_info->btree_inode)->rb_node);
extent_io_tree_init(&BTRFS_I(fs_info->btree_inode)->io_tree,
fs_info->btree_inode->i_mapping,
GFP_NOFS);
@ -1613,10 +1609,6 @@ struct btrfs_root *open_ctree(struct super_block *sb,
fs_info->btree_inode->i_mapping, GFP_NOFS);
fs_info->do_barriers = 1;
INIT_LIST_HEAD(&fs_info->dead_reloc_roots);
btrfs_leaf_ref_tree_init(&fs_info->reloc_ref_tree);
btrfs_leaf_ref_tree_init(&fs_info->shared_ref_tree);
BTRFS_I(fs_info->btree_inode)->root = tree_root;
memset(&BTRFS_I(fs_info->btree_inode)->location, 0,
sizeof(struct btrfs_key));
@ -1674,6 +1666,12 @@ struct btrfs_root *open_ctree(struct super_block *sb,
goto fail_iput;
}
features = btrfs_super_incompat_flags(disk_super);
if (!(features & BTRFS_FEATURE_INCOMPAT_MIXED_BACKREF)) {
features |= BTRFS_FEATURE_INCOMPAT_MIXED_BACKREF;
btrfs_set_super_incompat_flags(disk_super, features);
}
features = btrfs_super_compat_ro_flags(disk_super) &
~BTRFS_FEATURE_COMPAT_RO_SUPP;
if (!(sb->s_flags & MS_RDONLY) && features) {
@ -1771,7 +1769,7 @@ struct btrfs_root *open_ctree(struct super_block *sb,
if (ret) {
printk(KERN_WARNING "btrfs: failed to read the system "
"array on %s\n", sb->s_id);
goto fail_sys_array;
goto fail_sb_buffer;
}
blocksize = btrfs_level_size(tree_root,
@ -1785,6 +1783,8 @@ struct btrfs_root *open_ctree(struct super_block *sb,
btrfs_super_chunk_root(disk_super),
blocksize, generation);
BUG_ON(!chunk_root->node);
btrfs_set_root_node(&chunk_root->root_item, chunk_root->node);
chunk_root->commit_root = btrfs_root_node(chunk_root);
read_extent_buffer(chunk_root->node, fs_info->chunk_tree_uuid,
(unsigned long)btrfs_header_chunk_tree_uuid(chunk_root->node),
@ -1810,7 +1810,8 @@ struct btrfs_root *open_ctree(struct super_block *sb,
blocksize, generation);
if (!tree_root->node)
goto fail_chunk_root;
btrfs_set_root_node(&tree_root->root_item, tree_root->node);
tree_root->commit_root = btrfs_root_node(tree_root);
ret = find_and_setup_root(tree_root, fs_info,
BTRFS_EXTENT_TREE_OBJECTID, extent_root);
@ -1820,14 +1821,14 @@ struct btrfs_root *open_ctree(struct super_block *sb,
ret = find_and_setup_root(tree_root, fs_info,
BTRFS_DEV_TREE_OBJECTID, dev_root);
dev_root->track_dirty = 1;
if (ret)
goto fail_extent_root;
dev_root->track_dirty = 1;
ret = find_and_setup_root(tree_root, fs_info,
BTRFS_CSUM_TREE_OBJECTID, csum_root);
if (ret)
goto fail_extent_root;
goto fail_dev_root;
csum_root->track_dirty = 1;
@ -1881,7 +1882,7 @@ struct btrfs_root *open_ctree(struct super_block *sb,
}
if (!(sb->s_flags & MS_RDONLY)) {
ret = btrfs_cleanup_reloc_trees(tree_root);
ret = btrfs_recover_relocation(tree_root);
BUG_ON(ret);
}
@ -1908,14 +1909,19 @@ fail_cleaner:
fail_csum_root:
free_extent_buffer(csum_root->node);
free_extent_buffer(csum_root->commit_root);
fail_dev_root:
free_extent_buffer(dev_root->node);
free_extent_buffer(dev_root->commit_root);
fail_extent_root:
free_extent_buffer(extent_root->node);
free_extent_buffer(extent_root->commit_root);
fail_tree_root:
free_extent_buffer(tree_root->node);
free_extent_buffer(tree_root->commit_root);
fail_chunk_root:
free_extent_buffer(chunk_root->node);
fail_sys_array:
free_extent_buffer(dev_root->node);
free_extent_buffer(chunk_root->commit_root);
fail_sb_buffer:
btrfs_stop_workers(&fs_info->fixup_workers);
btrfs_stop_workers(&fs_info->delalloc_workers);
@ -2173,6 +2179,7 @@ int write_ctree_super(struct btrfs_trans_handle *trans,
int btrfs_free_fs_root(struct btrfs_fs_info *fs_info, struct btrfs_root *root)
{
WARN_ON(!RB_EMPTY_ROOT(&root->inode_tree));
radix_tree_delete(&fs_info->fs_roots_radix,
(unsigned long)root->root_key.objectid);
if (root->anon_super.s_dev) {
@ -2219,10 +2226,12 @@ int btrfs_cleanup_fs_roots(struct btrfs_fs_info *fs_info)
ARRAY_SIZE(gang));
if (!ret)
break;
root_objectid = gang[ret - 1]->root_key.objectid + 1;
for (i = 0; i < ret; i++) {
root_objectid = gang[i]->root_key.objectid;
ret = btrfs_find_dead_roots(fs_info->tree_root,
root_objectid, gang[i]);
root_objectid);
BUG_ON(ret);
btrfs_orphan_cleanup(gang[i]);
}
@ -2278,20 +2287,16 @@ int close_ctree(struct btrfs_root *root)
(unsigned long long)fs_info->total_ref_cache_size);
}
if (fs_info->extent_root->node)
free_extent_buffer(fs_info->extent_root->node);
if (fs_info->tree_root->node)
free_extent_buffer(fs_info->tree_root->node);
if (root->fs_info->chunk_root->node)
free_extent_buffer(root->fs_info->chunk_root->node);
if (root->fs_info->dev_root->node)
free_extent_buffer(root->fs_info->dev_root->node);
if (root->fs_info->csum_root->node)
free_extent_buffer(root->fs_info->csum_root->node);
free_extent_buffer(fs_info->extent_root->node);
free_extent_buffer(fs_info->extent_root->commit_root);
free_extent_buffer(fs_info->tree_root->node);
free_extent_buffer(fs_info->tree_root->commit_root);
free_extent_buffer(root->fs_info->chunk_root->node);
free_extent_buffer(root->fs_info->chunk_root->commit_root);
free_extent_buffer(root->fs_info->dev_root->node);
free_extent_buffer(root->fs_info->dev_root->commit_root);
free_extent_buffer(root->fs_info->csum_root->node);
free_extent_buffer(root->fs_info->csum_root->commit_root);
btrfs_free_block_groups(root->fs_info);

4
fs/btrfs/export.c
View file

@ -78,7 +78,7 @@ static struct dentry *btrfs_get_dentry(struct super_block *sb, u64 objectid,
btrfs_set_key_type(&key, BTRFS_INODE_ITEM_KEY);
key.offset = 0;
inode = btrfs_iget(sb, &key, root, NULL);
inode = btrfs_iget(sb, &key, root);
if (IS_ERR(inode))
return (void *)inode;
@ -192,7 +192,7 @@ static struct dentry *btrfs_get_parent(struct dentry *child)
btrfs_set_key_type(&key, BTRFS_INODE_ITEM_KEY);
key.offset = 0;
return d_obtain_alias(btrfs_iget(root->fs_info->sb, &key, root, NULL));
return d_obtain_alias(btrfs_iget(root->fs_info->sb, &key, root));
}
const struct export_operations btrfs_export_ops = {

2650
fs/btrfs/extent-tree.c
View file

File diff suppressed because it is too large Load diff

76
fs/btrfs/file.c
View file

@ -291,16 +291,12 @@ noinline int btrfs_drop_extents(struct btrfs_trans_handle *trans,
{
u64 extent_end = 0;
u64 search_start = start;
u64 leaf_start;
u64 ram_bytes = 0;
u64 orig_parent = 0;
u64 disk_bytenr = 0;
u64 orig_locked_end = locked_end;
u8 compression;
u8 encryption;
u16 other_encoding = 0;
u64 root_gen;
u64 root_owner;
struct extent_buffer *leaf;
struct btrfs_file_extent_item *extent;
struct btrfs_path *path;
@ -340,9 +336,6 @@ next_slot:
bookend = 0;
found_extent = 0;
found_inline = 0;
leaf_start = 0;
root_gen = 0;
root_owner = 0;
compression = 0;
encryption = 0;
extent = NULL;
@ -417,9 +410,6 @@ next_slot:
if (found_extent) {
read_extent_buffer(leaf, &old, (unsigned long)extent,
sizeof(old));
root_gen = btrfs_header_generation(leaf);
root_owner = btrfs_header_owner(leaf);
leaf_start = leaf->start;
}
if (end < extent_end && end >= key.offset) {
@ -443,14 +433,14 @@ next_slot:
}
locked_end = extent_end;
}
orig_parent = path->nodes[0]->start;
disk_bytenr = le64_to_cpu(old.disk_bytenr);
if (disk_bytenr != 0) {
ret = btrfs_inc_extent_ref(trans, root,
disk_bytenr,
le64_to_cpu(old.disk_num_bytes),
orig_parent, root->root_key.objectid,
trans->transid, inode->i_ino);
le64_to_cpu(old.disk_num_bytes), 0,
root->root_key.objectid,
key.objectid, key.offset -
le64_to_cpu(old.offset));
BUG_ON(ret);
}
}
@ -568,17 +558,6 @@ next_slot:
btrfs_mark_buffer_dirty(path->nodes[0]);
btrfs_set_lock_blocking(path->nodes[0]);
if (disk_bytenr != 0) {
ret = btrfs_update_extent_ref(trans, root,
disk_bytenr,
le64_to_cpu(old.disk_num_bytes),
orig_parent,
leaf->start,
root->root_key.objectid,
trans->transid, ins.objectid);
BUG_ON(ret);
}
path->leave_spinning = 0;
btrfs_release_path(root, path);
if (disk_bytenr != 0)
@ -594,8 +573,9 @@ next_slot:
ret = btrfs_free_extent(trans, root,
old_disk_bytenr,
le64_to_cpu(old.disk_num_bytes),
leaf_start, root_owner,
root_gen, key.objectid, 0);
0, root->root_key.objectid,
key.objectid, key.offset -
le64_to_cpu(old.offset));
BUG_ON(ret);
*hint_byte = old_disk_bytenr;
}
@ -664,12 +644,11 @@ int btrfs_mark_extent_written(struct btrfs_trans_handle *trans,
u64 bytenr;
u64 num_bytes;
u64 extent_end;
u64 extent_offset;
u64 orig_offset;
u64 other_start;
u64 other_end;
u64 split = start;
u64 locked_end = end;
u64 orig_parent;
int extent_type;
int split_end = 1;
int ret;
@ -703,7 +682,7 @@ again:
bytenr = btrfs_file_extent_disk_bytenr(leaf, fi);
num_bytes = btrfs_file_extent_disk_num_bytes(leaf, fi);
extent_offset = btrfs_file_extent_offset(leaf, fi);
orig_offset = key.offset - btrfs_file_extent_offset(leaf, fi);
if (key.offset == start)
split = end;
@ -711,8 +690,6 @@ again:
if (key.offset == start && extent_end == end) {
int del_nr = 0;
int del_slot = 0;
u64 leaf_owner = btrfs_header_owner(leaf);
u64 leaf_gen = btrfs_header_generation(leaf);
other_start = end;
other_end = 0;
if (extent_mergeable(leaf, path->slots[0] + 1, inode->i_ino,
@ -721,8 +698,8 @@ again:
del_slot = path->slots[0] + 1;
del_nr++;
ret = btrfs_free_extent(trans, root, bytenr, num_bytes,
leaf->start, leaf_owner,
leaf_gen, inode->i_ino, 0);
0, root->root_key.objectid,
inode->i_ino, orig_offset);
BUG_ON(ret);
}
other_start = 0;
@ -733,8 +710,8 @@ again:
del_slot = path->slots[0];
del_nr++;
ret = btrfs_free_extent(trans, root, bytenr, num_bytes,
leaf->start, leaf_owner,
leaf_gen, inode->i_ino, 0);
0, root->root_key.objectid,
inode->i_ino, orig_offset);
BUG_ON(ret);
}
split_end = 0;
@ -768,13 +745,12 @@ again:
locked_end = extent_end;
}
btrfs_set_file_extent_num_bytes(leaf, fi, split - key.offset);
extent_offset += split - key.offset;
} else {
BUG_ON(key.offset != start);
btrfs_set_file_extent_offset(leaf, fi, extent_offset +
split - key.offset);
btrfs_set_file_extent_num_bytes(leaf, fi, extent_end - split);
key.offset = split;
btrfs_set_file_extent_offset(leaf, fi, key.offset -
orig_offset);
btrfs_set_file_extent_num_bytes(leaf, fi, extent_end - split);
btrfs_set_item_key_safe(trans, root, path, &key);
extent_end = split;
}
@ -793,7 +769,8 @@ again:
struct btrfs_file_extent_item);
key.offset = split;
btrfs_set_item_key_safe(trans, root, path, &key);
btrfs_set_file_extent_offset(leaf, fi, extent_offset);
btrfs_set_file_extent_offset(leaf, fi, key.offset -
orig_offset);
btrfs_set_file_extent_num_bytes(leaf, fi,
other_end - split);
goto done;
@ -815,10 +792,9 @@ again:
btrfs_mark_buffer_dirty(leaf);
orig_parent = leaf->start;
ret = btrfs_inc_extent_ref(trans, root, bytenr, num_bytes,
orig_parent, root->root_key.objectid,
trans->transid, inode->i_ino);
ret = btrfs_inc_extent_ref(trans, root, bytenr, num_bytes, 0,
root->root_key.objectid,
inode->i_ino, orig_offset);
BUG_ON(ret);
btrfs_release_path(root, path);
@ -833,20 +809,12 @@ again:
btrfs_set_file_extent_type(leaf, fi, extent_type);
btrfs_set_file_extent_disk_bytenr(leaf, fi, bytenr);
btrfs_set_file_extent_disk_num_bytes(leaf, fi, num_bytes);
btrfs_set_file_extent_offset(leaf, fi, extent_offset);
btrfs_set_file_extent_offset(leaf, fi, key.offset - orig_offset);
btrfs_set_file_extent_num_bytes(leaf, fi, extent_end - key.offset);
btrfs_set_file_extent_ram_bytes(leaf, fi, num_bytes);
btrfs_set_file_extent_compression(leaf, fi, 0);
btrfs_set_file_extent_encryption(leaf, fi, 0);
btrfs_set_file_extent_other_encoding(leaf, fi, 0);
if (orig_parent != leaf->start) {
ret = btrfs_update_extent_ref(trans, root, bytenr, num_bytes,
orig_parent, leaf->start,
root->root_key.objectid,
trans->transid, inode->i_ino);
BUG_ON(ret);
}
done:
btrfs_mark_buffer_dirty(leaf);

132
fs/btrfs/inode.c
View file

@ -48,7 +48,6 @@
#include "ordered-data.h"
#include "xattr.h"
#include "tree-log.h"
#include "ref-cache.h"
#include "compression.h"
#include "locking.h"
@ -944,6 +943,7 @@ static noinline int run_delalloc_nocow(struct inode *inode,
u64 cow_start;
u64 cur_offset;
u64 extent_end;
u64 extent_offset;
u64 disk_bytenr;
u64 num_bytes;
int extent_type;
@ -1005,6 +1005,7 @@ next_slot:
if (extent_type == BTRFS_FILE_EXTENT_REG ||
extent_type == BTRFS_FILE_EXTENT_PREALLOC) {
disk_bytenr = btrfs_file_extent_disk_bytenr(leaf, fi);
extent_offset = btrfs_file_extent_offset(leaf, fi);
extent_end = found_key.offset +
btrfs_file_extent_num_bytes(leaf, fi);
if (extent_end <= start) {
@ -1022,9 +1023,10 @@ next_slot:
if (btrfs_extent_readonly(root, disk_bytenr))
goto out_check;
if (btrfs_cross_ref_exist(trans, root, inode->i_ino,
disk_bytenr))
found_key.offset -
extent_offset, disk_bytenr))
goto out_check;
disk_bytenr += btrfs_file_extent_offset(leaf, fi);
disk_bytenr += extent_offset;
disk_bytenr += cur_offset - found_key.offset;
num_bytes = min(end + 1, extent_end) - cur_offset;
/*
@ -1489,9 +1491,9 @@ static int insert_reserved_file_extent(struct btrfs_trans_handle *trans,
ins.objectid = disk_bytenr;
ins.offset = disk_num_bytes;
ins.type = BTRFS_EXTENT_ITEM_KEY;
ret = btrfs_alloc_reserved_extent(trans, root, leaf->start,
root->root_key.objectid,
trans->transid, inode->i_ino, &ins);
ret = btrfs_alloc_reserved_file_extent(trans, root,
root->root_key.objectid,
inode->i_ino, file_pos, &ins);
BUG_ON(ret);
btrfs_free_path(path);
@ -1956,23 +1958,13 @@ void btrfs_orphan_cleanup(struct btrfs_root *root)
* crossing root thing. we store the inode number in the
* offset of the orphan item.
*/
inode = btrfs_iget_locked(root->fs_info->sb,
found_key.offset, root);
if (!inode)
found_key.objectid = found_key.offset;
found_key.type = BTRFS_INODE_ITEM_KEY;
found_key.offset = 0;
inode = btrfs_iget(root->fs_info->sb, &found_key, root);
if (IS_ERR(inode))
break;
if (inode->i_state & I_NEW) {
BTRFS_I(inode)->root = root;
/* have to set the location manually */
BTRFS_I(inode)->location.objectid = inode->i_ino;
BTRFS_I(inode)->location.type = BTRFS_INODE_ITEM_KEY;
BTRFS_I(inode)->location.offset = 0;
btrfs_read_locked_inode(inode);
unlock_new_inode(inode);
}
/*
* add this inode to the orphan list so btrfs_orphan_del does
* the proper thing when we hit it
@ -2069,7 +2061,7 @@ static noinline int acls_after_inode_item(struct extent_buffer *leaf,
/*
* read an inode from the btree into the in-memory inode
*/
void btrfs_read_locked_inode(struct inode *inode)
static void btrfs_read_locked_inode(struct inode *inode)
{
struct btrfs_path *path;
struct extent_buffer *leaf;
@ -2599,9 +2591,8 @@ noinline int btrfs_truncate_inode_items(struct btrfs_trans_handle *trans,
struct btrfs_file_extent_item *fi;
u64 extent_start = 0;
u64 extent_num_bytes = 0;
u64 extent_offset = 0;
u64 item_end = 0;
u64 root_gen = 0;
u64 root_owner = 0;
int found_extent;
int del_item;
int pending_del_nr = 0;
@ -2716,6 +2707,9 @@ search_again:
extent_num_bytes =
btrfs_file_extent_disk_num_bytes(leaf,
fi);
extent_offset = found_key.offset -
btrfs_file_extent_offset(leaf, fi);
/* FIXME blocksize != 4096 */
num_dec = btrfs_file_extent_num_bytes(leaf, fi);
if (extent_start != 0) {
@ -2723,8 +2717,6 @@ search_again:
if (root->ref_cows)
inode_sub_bytes(inode, num_dec);
}
root_gen = btrfs_header_generation(leaf);
root_owner = btrfs_header_owner(leaf);
}
} else if (extent_type == BTRFS_FILE_EXTENT_INLINE) {
/*
@ -2768,12 +2760,12 @@ delete:
} else {
break;
}
if (found_extent) {
if (found_extent && root->ref_cows) {
btrfs_set_path_blocking(path);
ret = btrfs_free_extent(trans, root, extent_start,
extent_num_bytes,
leaf->start, root_owner,
root_gen, inode->i_ino, 0);
extent_num_bytes, 0,
btrfs_header_owner(leaf),
inode->i_ino, extent_offset);
BUG_ON(ret);
}
next:
@ -3105,6 +3097,45 @@ static int fixup_tree_root_location(struct btrfs_root *root,
return 0;
}
static void inode_tree_add(struct inode *inode)
{
struct btrfs_root *root = BTRFS_I(inode)->root;
struct btrfs_inode *entry;
struct rb_node **p = &root->inode_tree.rb_node;
struct rb_node *parent = NULL;
spin_lock(&root->inode_lock);
while (*p) {
parent = *p;
entry = rb_entry(parent, struct btrfs_inode, rb_node);
if (inode->i_ino < entry->vfs_inode.i_ino)
p = &(*p)->rb_left;
else if (inode->i_ino > entry->vfs_inode.i_ino)
p = &(*p)->rb_right;
else {
WARN_ON(!(entry->vfs_inode.i_state &
(I_WILL_FREE | I_FREEING | I_CLEAR)));
break;
}
}
rb_link_node(&BTRFS_I(inode)->rb_node, parent, p);
rb_insert_color(&BTRFS_I(inode)->rb_node, &root->inode_tree);
spin_unlock(&root->inode_lock);
}
static void inode_tree_del(struct inode *inode)
{
struct btrfs_root *root = BTRFS_I(inode)->root;
if (!RB_EMPTY_NODE(&BTRFS_I(inode)->rb_node)) {
spin_lock(&root->inode_lock);
rb_erase(&BTRFS_I(inode)->rb_node, &root->inode_tree);
spin_unlock(&root->inode_lock);
RB_CLEAR_NODE(&BTRFS_I(inode)->rb_node);
}
}
static noinline void init_btrfs_i(struct inode *inode)
{
struct btrfs_inode *bi = BTRFS_I(inode);
@ -3130,6 +3161,7 @@ static noinline void init_btrfs_i(struct inode *inode)
inode->i_mapping, GFP_NOFS);
INIT_LIST_HEAD(&BTRFS_I(inode)->delalloc_inodes);
INIT_LIST_HEAD(&BTRFS_I(inode)->ordered_operations);
RB_CLEAR_NODE(&BTRFS_I(inode)->rb_node);
btrfs_ordered_inode_tree_init(&BTRFS_I(inode)->ordered_tree);
mutex_init(&BTRFS_I(inode)->extent_mutex);
mutex_init(&BTRFS_I(inode)->log_mutex);
@ -3152,26 +3184,9 @@ static int btrfs_find_actor(struct inode *inode, void *opaque)
args->root == BTRFS_I(inode)->root;
}
struct inode *btrfs_ilookup(struct super_block *s, u64 objectid,
struct btrfs_root *root, int wait)
{
struct inode *inode;
struct btrfs_iget_args args;
args.ino = objectid;
args.root = root;
if (wait) {
inode = ilookup5(s, objectid, btrfs_find_actor,
(void *)&args);
} else {
inode = ilookup5_nowait(s, objectid, btrfs_find_actor,
(void *)&args);
}
return inode;
}
struct inode *btrfs_iget_locked(struct super_block *s, u64 objectid,
struct btrfs_root *root)
static struct inode *btrfs_iget_locked(struct super_block *s,
u64 objectid,
struct btrfs_root *root)
{
struct inode *inode;
struct btrfs_iget_args args;
@ -3188,24 +3203,21 @@ struct inode *btrfs_iget_locked(struct super_block *s, u64 objectid,
* Returns in *is_new if the inode was read from disk
*/
struct inode *btrfs_iget(struct super_block *s, struct btrfs_key *location,
struct btrfs_root *root, int *is_new)
struct btrfs_root *root)
{
struct inode *inode;
inode = btrfs_iget_locked(s, location->objectid, root);
if (!inode)
return ERR_PTR(-EACCES);
return ERR_PTR(-ENOMEM);
if (inode->i_state & I_NEW) {
BTRFS_I(inode)->root = root;
memcpy(&BTRFS_I(inode)->location, location, sizeof(*location));
btrfs_read_locked_inode(inode);
inode_tree_add(inode);
unlock_new_inode(inode);
if (is_new)
*is_new = 1;
} else {
if (is_new)
*is_new = 0;
}
return inode;
@ -3218,7 +3230,7 @@ struct inode *btrfs_lookup_dentry(struct inode *dir, struct dentry *dentry)
struct btrfs_root *root = bi->root;
struct btrfs_root *sub_root = root;
struct btrfs_key location;
int ret, new;
int ret;
if (dentry->d_name.len > BTRFS_NAME_LEN)
return ERR_PTR(-ENAMETOOLONG);
@ -3236,7 +3248,7 @@ struct inode *btrfs_lookup_dentry(struct inode *dir, struct dentry *dentry)
return ERR_PTR(ret);
if (ret > 0)
return ERR_PTR(-ENOENT);
inode = btrfs_iget(dir->i_sb, &location, sub_root, &new);
inode = btrfs_iget(dir->i_sb, &location, sub_root);
if (IS_ERR(inode))
return ERR_CAST(inode);
}
@ -3631,6 +3643,7 @@ static struct inode *btrfs_new_inode(struct btrfs_trans_handle *trans,
btrfs_set_key_type(location, BTRFS_INODE_ITEM_KEY);
insert_inode_hash(inode);
inode_tree_add(inode);
return inode;
fail:
if (dir)
@ -4683,6 +4696,7 @@ void btrfs_destroy_inode(struct inode *inode)
btrfs_put_ordered_extent(ordered);
}
}
inode_tree_del(inode);
btrfs_drop_extent_cache(inode, 0, (u64)-1, 0);
kmem_cache_free(btrfs_inode_cachep, BTRFS_I(inode));
}

21
fs/btrfs/ioctl.c
View file

@ -82,22 +82,25 @@ static noinline int create_subvol(struct btrfs_root *root,
if (ret)
goto fail;
leaf = btrfs_alloc_free_block(trans, root, root->leafsize, 0,
objectid, trans->transid, 0, 0, 0);
leaf = btrfs_alloc_free_block(trans, root, root->leafsize,
0, objectid, NULL, 0, 0, 0);
if (IS_ERR(leaf)) {
ret = PTR_ERR(leaf);
goto fail;
}
btrfs_set_header_nritems(leaf, 0);
btrfs_set_header_level(leaf, 0);
memset_extent_buffer(leaf, 0, 0, sizeof(struct btrfs_header));
btrfs_set_header_bytenr(leaf, leaf->start);
btrfs_set_header_generation(leaf, trans->transid);
btrfs_set_header_backref_rev(leaf, BTRFS_MIXED_BACKREF_REV);
btrfs_set_header_owner(leaf, objectid);
write_extent_buffer(leaf, root->fs_info->fsid,
(unsigned long)btrfs_header_fsid(leaf),
BTRFS_FSID_SIZE);
write_extent_buffer(leaf, root->fs_info->chunk_tree_uuid,
(unsigned long)btrfs_header_chunk_tree_uuid(leaf),
BTRFS_UUID_SIZE);
btrfs_mark_buffer_dirty(leaf);
inode_item = &root_item.inode;
@ -125,7 +128,7 @@ static noinline int create_subvol(struct btrfs_root *root,
btrfs_set_root_dirid(&root_item, new_dirid);
key.objectid = objectid;
key.offset = 1;
key.offset = 0;
btrfs_set_key_type(&key, BTRFS_ROOT_ITEM_KEY);
ret = btrfs_insert_root(trans, root->fs_info->tree_root, &key,
&root_item);
@ -911,10 +914,10 @@ static long btrfs_ioctl_clone(struct file *file, unsigned long srcfd,
if (disko) {
inode_add_bytes(inode, datal);
ret = btrfs_inc_extent_ref(trans, root,
disko, diskl, leaf->start,
root->root_key.objectid,
trans->transid,
inode->i_ino);
disko, diskl, 0,
root->root_key.objectid,
inode->i_ino,
new_key.offset - datao);
BUG_ON(ret);
}
} else if (type == BTRFS_FILE_EXTENT_INLINE) {

153
fs/btrfs/print-tree.c
View file

@ -45,22 +45,132 @@ static void print_dev_item(struct extent_buffer *eb,
(unsigned long long)btrfs_device_total_bytes(eb, dev_item),
(unsigned long long)btrfs_device_bytes_used(eb, dev_item));
}
static void print_extent_data_ref(struct extent_buffer *eb,
struct btrfs_extent_data_ref *ref)
{
printk(KERN_INFO "\t\textent data backref root %llu "
"objectid %llu offset %llu count %u\n",
(unsigned long long)btrfs_extent_data_ref_root(eb, ref),
(unsigned long long)btrfs_extent_data_ref_objectid(eb, ref),
(unsigned long long)btrfs_extent_data_ref_offset(eb, ref),
btrfs_extent_data_ref_count(eb, ref));
}
static void print_extent_item(struct extent_buffer *eb, int slot)
{
struct btrfs_extent_item *ei;
struct btrfs_extent_inline_ref *iref;
struct btrfs_extent_data_ref *dref;
struct btrfs_shared_data_ref *sref;
struct btrfs_disk_key key;
unsigned long end;
unsigned long ptr;
int type;
u32 item_size = btrfs_item_size_nr(eb, slot);
u64 flags;
u64 offset;
if (item_size < sizeof(*ei)) {
#ifdef BTRFS_COMPAT_EXTENT_TREE_V0
struct btrfs_extent_item_v0 *ei0;
BUG_ON(item_size != sizeof(*ei0));
ei0 = btrfs_item_ptr(eb, slot, struct btrfs_extent_item_v0);
printk(KERN_INFO "\t\textent refs %u\n",
btrfs_extent_refs_v0(eb, ei0));
return;
#else
BUG();
#endif
}
ei = btrfs_item_ptr(eb, slot, struct btrfs_extent_item);
flags = btrfs_extent_flags(eb, ei);
printk(KERN_INFO "\t\textent refs %llu gen %llu flags %llu\n",
(unsigned long long)btrfs_extent_refs(eb, ei),
(unsigned long long)btrfs_extent_generation(eb, ei),
(unsigned long long)flags);
if (flags & BTRFS_EXTENT_FLAG_TREE_BLOCK) {
struct btrfs_tree_block_info *info;
info = (struct btrfs_tree_block_info *)(ei + 1);
btrfs_tree_block_key(eb, info, &key);
printk(KERN_INFO "\t\ttree block key (%llu %x %llu) "
"level %d\n",
(unsigned long long)btrfs_disk_key_objectid(&key),
key.type,
(unsigned long long)btrfs_disk_key_offset(&key),
btrfs_tree_block_level(eb, info));
iref = (struct btrfs_extent_inline_ref *)(info + 1);
} else {
iref = (struct btrfs_extent_inline_ref *)(ei + 1);
}
ptr = (unsigned long)iref;
end = (unsigned long)ei + item_size;
while (ptr < end) {
iref = (struct btrfs_extent_inline_ref *)ptr;
type = btrfs_extent_inline_ref_type(eb, iref);
offset = btrfs_extent_inline_ref_offset(eb, iref);
switch (type) {
case BTRFS_TREE_BLOCK_REF_KEY:
printk(KERN_INFO "\t\ttree block backref "
"root %llu\n", (unsigned long long)offset);
break;
case BTRFS_SHARED_BLOCK_REF_KEY:
printk(KERN_INFO "\t\tshared block backref "
"parent %llu\n", (unsigned long long)offset);
break;
case BTRFS_EXTENT_DATA_REF_KEY:
dref = (struct btrfs_extent_data_ref *)(&iref->offset);
print_extent_data_ref(eb, dref);
break;
case BTRFS_SHARED_DATA_REF_KEY:
sref = (struct btrfs_shared_data_ref *)(iref + 1);
printk(KERN_INFO "\t\tshared data backref "
"parent %llu count %u\n",
(unsigned long long)offset,
btrfs_shared_data_ref_count(eb, sref));
break;
default:
BUG();
}
ptr += btrfs_extent_inline_ref_size(type);
}
WARN_ON(ptr > end);
}
#ifdef BTRFS_COMPAT_EXTENT_TREE_V0
static void print_extent_ref_v0(struct extent_buffer *eb, int slot)
{
struct btrfs_extent_ref_v0 *ref0;
ref0 = btrfs_item_ptr(eb, slot, struct btrfs_extent_ref_v0);
printk("\t\textent back ref root %llu gen %llu "
"owner %llu num_refs %lu\n",
(unsigned long long)btrfs_ref_root_v0(eb, ref0),
(unsigned long long)btrfs_ref_generation_v0(eb, ref0),
(unsigned long long)btrfs_ref_objectid_v0(eb, ref0),
(unsigned long)btrfs_ref_count_v0(eb, ref0));
}
#endif
void btrfs_print_leaf(struct btrfs_root *root, struct extent_buffer *l)
{
int i;
u32 type;
u32 nr = btrfs_header_nritems(l);
struct btrfs_item *item;
struct btrfs_extent_item *ei;
struct btrfs_root_item *ri;
struct btrfs_dir_item *di;
struct btrfs_inode_item *ii;
struct btrfs_block_group_item *bi;
struct btrfs_file_extent_item *fi;
struct btrfs_extent_data_ref *dref;
struct btrfs_shared_data_ref *sref;
struct btrfs_dev_extent *dev_extent;
struct btrfs_key key;
struct btrfs_key found_key;
struct btrfs_extent_ref *ref;
struct btrfs_dev_extent *dev_extent;
u32 type;
printk(KERN_INFO "leaf %llu total ptrs %d free space %d\n",
(unsigned long long)btrfs_header_bytenr(l), nr,
@ -100,20 +210,25 @@ void btrfs_print_leaf(struct btrfs_root *root, struct extent_buffer *l)
btrfs_disk_root_refs(l, ri));
break;
case BTRFS_EXTENT_ITEM_KEY:
ei = btrfs_item_ptr(l, i, struct btrfs_extent_item);
printk(KERN_INFO "\t\textent data refs %u\n",
btrfs_extent_refs(l, ei));
print_extent_item(l, i);
break;
case BTRFS_EXTENT_REF_KEY:
ref = btrfs_item_ptr(l, i, struct btrfs_extent_ref);
printk(KERN_INFO "\t\textent back ref root %llu "
"gen %llu owner %llu num_refs %lu\n",
(unsigned long long)btrfs_ref_root(l, ref),
(unsigned long long)btrfs_ref_generation(l, ref),
(unsigned long long)btrfs_ref_objectid(l, ref),
(unsigned long)btrfs_ref_num_refs(l, ref));
case BTRFS_TREE_BLOCK_REF_KEY:
printk(KERN_INFO "\t\ttree block backref\n");
break;
case BTRFS_SHARED_BLOCK_REF_KEY:
printk(KERN_INFO "\t\tshared block backref\n");
break;
case BTRFS_EXTENT_DATA_REF_KEY:
dref = btrfs_item_ptr(l, i,
struct btrfs_extent_data_ref);
print_extent_data_ref(l, dref);
break;
case BTRFS_SHARED_DATA_REF_KEY:
sref = btrfs_item_ptr(l, i,
struct btrfs_shared_data_ref);
printk(KERN_INFO "\t\tshared data backref count %u\n",
btrfs_shared_data_ref_count(l, sref));
break;
case BTRFS_EXTENT_DATA_KEY:
fi = btrfs_item_ptr(l, i,
struct btrfs_file_extent_item);
@ -139,6 +254,12 @@ void btrfs_print_leaf(struct btrfs_root *root, struct extent_buffer *l)
(unsigned long long)
btrfs_file_extent_ram_bytes(l, fi));
break;
case BTRFS_EXTENT_REF_V0_KEY:
#ifdef BTRFS_COMPAT_EXTENT_TREE_V0
print_extent_ref_v0(l, i);
#else
BUG();
#endif
case BTRFS_BLOCK_GROUP_ITEM_KEY:
bi = btrfs_item_ptr(l, i,
struct btrfs_block_group_item);

3711
fs/btrfs/relocation.c Normal file
View file

File diff suppressed because it is too large Load diff

17
fs/btrfs/root-tree.c
View file

@ -111,6 +111,15 @@ out:
return ret;
}
int btrfs_set_root_node(struct btrfs_root_item *item,
struct extent_buffer *node)
{
btrfs_set_root_bytenr(item, node->start);
btrfs_set_root_level(item, btrfs_header_level(node));
btrfs_set_root_generation(item, btrfs_header_generation(node));
return 0;
}
/*
* copy the data in 'item' into the btree
*/
@ -164,8 +173,7 @@ int btrfs_insert_root(struct btrfs_trans_handle *trans, struct btrfs_root
* offset lower than the latest root. They need to be queued for deletion to
* finish what was happening when we crashed.
*/
int btrfs_find_dead_roots(struct btrfs_root *root, u64 objectid,
struct btrfs_root *latest)
int btrfs_find_dead_roots(struct btrfs_root *root, u64 objectid)
{
struct btrfs_root *dead_root;
struct btrfs_item *item;
@ -227,10 +235,7 @@ again:
goto err;
}
if (objectid == BTRFS_TREE_RELOC_OBJECTID)
ret = btrfs_add_dead_reloc_root(dead_root);
else
ret = btrfs_add_dead_root(dead_root, latest);
ret = btrfs_add_dead_root(dead_root);
if (ret)
goto err;
goto again;

26
fs/btrfs/super.c
View file

@ -52,7 +52,6 @@
#include "export.h"
#include "compression.h"
static struct super_operations btrfs_super_ops;
static void btrfs_put_super(struct super_block *sb)
@ -322,7 +321,7 @@ static int btrfs_fill_super(struct super_block *sb,
struct dentry *root_dentry;
struct btrfs_super_block *disk_super;
struct btrfs_root *tree_root;
struct btrfs_inode *bi;
struct btrfs_key key;
int err;
sb->s_maxbytes = MAX_LFS_FILESIZE;
@ -341,23 +340,15 @@ static int btrfs_fill_super(struct super_block *sb,
}
sb->s_fs_info = tree_root;
disk_super = &tree_root->fs_info->super_copy;
inode = btrfs_iget_locked(sb, BTRFS_FIRST_FREE_OBJECTID,
tree_root->fs_info->fs_root);
bi = BTRFS_I(inode);
bi->location.objectid = inode->i_ino;
bi->location.offset = 0;
bi->root = tree_root->fs_info->fs_root;
btrfs_set_key_type(&bi->location, BTRFS_INODE_ITEM_KEY);
if (!inode) {
err = -ENOMEM;
key.objectid = BTRFS_FIRST_FREE_OBJECTID;
key.type = BTRFS_INODE_ITEM_KEY;
key.offset = 0;
inode = btrfs_iget(sb, &key, tree_root->fs_info->fs_root);
if (IS_ERR(inode)) {
err = PTR_ERR(inode);
goto fail_close;
}
if (inode->i_state & I_NEW) {
btrfs_read_locked_inode(inode);
unlock_new_inode(inode);
}
root_dentry = d_alloc_root(inode);
if (!root_dentry) {
@ -584,7 +575,8 @@ static int btrfs_remount(struct super_block *sb, int *flags, char *data)
if (btrfs_super_log_root(&root->fs_info->super_copy) != 0)
return -EINVAL;
ret = btrfs_cleanup_reloc_trees(root);
/* recover relocation */
ret = btrfs_recover_relocation(root);
WARN_ON(ret);
ret = btrfs_cleanup_fs_roots(root->fs_info);

408
fs/btrfs/transaction.c
View file

@ -25,7 +25,6 @@
#include "disk-io.h"
#include "transaction.h"
#include "locking.h"
#include "ref-cache.h"
#include "tree-log.h"
#define BTRFS_ROOT_TRANS_TAG 0
@ -94,48 +93,40 @@ static noinline int join_transaction(struct btrfs_root *root)
* to make sure the old root from before we joined the transaction is deleted
* when the transaction commits
*/
noinline int btrfs_record_root_in_trans(struct btrfs_root *root)
static noinline int record_root_in_trans(struct btrfs_trans_handle *trans,
struct btrfs_root *root)
{
struct btrfs_dirty_root *dirty;
u64 running_trans_id = root->fs_info->running_transaction->transid;
if (root->ref_cows && root->last_trans < running_trans_id) {
if (root->ref_cows && root->last_trans < trans->transid) {
WARN_ON(root == root->fs_info->extent_root);
if (root->root_item.refs != 0) {
radix_tree_tag_set(&root->fs_info->fs_roots_radix,
(unsigned long)root->root_key.objectid,
BTRFS_ROOT_TRANS_TAG);
WARN_ON(root->root_item.refs == 0);
WARN_ON(root->commit_root != root->node);
dirty = kmalloc(sizeof(*dirty), GFP_NOFS);
BUG_ON(!dirty);
dirty->root = kmalloc(sizeof(*dirty->root), GFP_NOFS);
BUG_ON(!dirty->root);
dirty->latest_root = root;
INIT_LIST_HEAD(&dirty->list);
root->commit_root = btrfs_root_node(root);
memcpy(dirty->root, root, sizeof(*root));
spin_lock_init(&dirty->root->node_lock);
spin_lock_init(&dirty->root->list_lock);
mutex_init(&dirty->root->objectid_mutex);
mutex_init(&dirty->root->log_mutex);
INIT_LIST_HEAD(&dirty->root->dead_list);
dirty->root->node = root->commit_root;
dirty->root->commit_root = NULL;
spin_lock(&root->list_lock);
list_add(&dirty->root->dead_list, &root->dead_list);
spin_unlock(&root->list_lock);
root->dirty_root = dirty;
} else {
WARN_ON(1);
}
root->last_trans = running_trans_id;
radix_tree_tag_set(&root->fs_info->fs_roots_radix,
(unsigned long)root->root_key.objectid,
BTRFS_ROOT_TRANS_TAG);
root->last_trans = trans->transid;
btrfs_init_reloc_root(trans, root);
}
return 0;
}
int btrfs_record_root_in_trans(struct btrfs_trans_handle *trans,
struct btrfs_root *root)
{
if (!root->ref_cows)
return 0;
mutex_lock(&root->fs_info->trans_mutex);
if (root->last_trans == trans->transid) {
mutex_unlock(&root->fs_info->trans_mutex);
return 0;
}
record_root_in_trans(trans, root);
mutex_unlock(&root->fs_info->trans_mutex);
return 0;
}
/* wait for commit against the current transaction to become unblocked
* when this is done, it is safe to start a new transaction, but the current
* transaction might not be fully on disk.
@ -181,7 +172,6 @@ static struct btrfs_trans_handle *start_transaction(struct btrfs_root *root,
ret = join_transaction(root);
BUG_ON(ret);
btrfs_record_root_in_trans(root);
h->transid = root->fs_info->running_transaction->transid;
h->transaction = root->fs_info->running_transaction;
h->blocks_reserved = num_blocks;
@ -192,6 +182,7 @@ static struct btrfs_trans_handle *start_transaction(struct btrfs_root *root,
h->delayed_ref_updates = 0;
root->fs_info->running_transaction->use_count++;
record_root_in_trans(h, root);
mutex_unlock(&root->fs_info->trans_mutex);
return h;
}
@ -233,6 +224,7 @@ static noinline int wait_for_commit(struct btrfs_root *root,
return 0;
}
#if 0
/*
* rate limit against the drop_snapshot code. This helps to slow down new
* operations if the drop_snapshot code isn't able to keep up.
@ -273,6 +265,7 @@ harder:
goto harder;
}
}
#endif
void btrfs_throttle(struct btrfs_root *root)
{
@ -280,7 +273,6 @@ void btrfs_throttle(struct btrfs_root *root)
if (!root->fs_info->open_ioctl_trans)
wait_current_trans(root);
mutex_unlock(&root->fs_info->trans_mutex);
throttle_on_drops(root);
}
static int __btrfs_end_transaction(struct btrfs_trans_handle *trans,
@ -323,9 +315,6 @@ static int __btrfs_end_transaction(struct btrfs_trans_handle *trans,
memset(trans, 0, sizeof(*trans));
kmem_cache_free(btrfs_trans_handle_cachep, trans);
if (throttle)
throttle_on_drops(root);
return 0;
}
@ -462,12 +451,8 @@ static int update_cowonly_root(struct btrfs_trans_handle *trans,
old_root_bytenr = btrfs_root_bytenr(&root->root_item);
if (old_root_bytenr == root->node->start)
break;
btrfs_set_root_bytenr(&root->root_item,
root->node->start);
btrfs_set_root_level(&root->root_item,
btrfs_header_level(root->node));
btrfs_set_root_generation(&root->root_item, trans->transid);
btrfs_set_root_node(&root->root_item, root->node);
ret = btrfs_update_root(trans, tree_root,
&root->root_key,
&root->root_item);
@ -477,14 +462,16 @@ static int update_cowonly_root(struct btrfs_trans_handle *trans,
ret = btrfs_run_delayed_refs(trans, root, (unsigned long)-1);
BUG_ON(ret);
}
free_extent_buffer(root->commit_root);
root->commit_root = btrfs_root_node(root);
return 0;
}
/*
* update all the cowonly tree roots on disk
*/
int btrfs_commit_tree_roots(struct btrfs_trans_handle *trans,
struct btrfs_root *root)
static noinline int commit_cowonly_roots(struct btrfs_trans_handle *trans,
struct btrfs_root *root)
{
struct btrfs_fs_info *fs_info = root->fs_info;
struct list_head *next;
@ -520,118 +507,54 @@ int btrfs_commit_tree_roots(struct btrfs_trans_handle *trans,
* a dirty root struct and adds it into the list of dead roots that need to
* be deleted
*/
int btrfs_add_dead_root(struct btrfs_root *root, struct btrfs_root *latest)
int btrfs_add_dead_root(struct btrfs_root *root)
{
struct btrfs_dirty_root *dirty;
dirty = kmalloc(sizeof(*dirty), GFP_NOFS);
if (!dirty)
return -ENOMEM;
dirty->root = root;
dirty->latest_root = latest;
mutex_lock(&root->fs_info->trans_mutex);
list_add(&dirty->list, &latest->fs_info->dead_roots);
list_add(&root->root_list, &root->fs_info->dead_roots);
mutex_unlock(&root->fs_info->trans_mutex);
return 0;
}
/*
* at transaction commit time we need to schedule the old roots for
* deletion via btrfs_drop_snapshot. This runs through all the
* reference counted roots that were modified in the current
* transaction and puts them into the drop list
* update all the cowonly tree roots on disk
*/
static noinline int add_dirty_roots(struct btrfs_trans_handle *trans,
struct radix_tree_root *radix,
struct list_head *list)
static noinline int commit_fs_roots(struct btrfs_trans_handle *trans,
struct btrfs_root *root)
{
struct btrfs_dirty_root *dirty;
struct btrfs_root *gang[8];
struct btrfs_root *root;
struct btrfs_fs_info *fs_info = root->fs_info;
int i;
int ret;
int err = 0;
u32 refs;
while (1) {
ret = radix_tree_gang_lookup_tag(radix, (void **)gang, 0,
ret = radix_tree_gang_lookup_tag(&fs_info->fs_roots_radix,
(void **)gang, 0,
ARRAY_SIZE(gang),
BTRFS_ROOT_TRANS_TAG);
if (ret == 0)
break;
for (i = 0; i < ret; i++) {
root = gang[i];
radix_tree_tag_clear(radix,
(unsigned long)root->root_key.objectid,
BTRFS_ROOT_TRANS_TAG);
BUG_ON(!root->ref_tree);
dirty = root->dirty_root;
radix_tree_tag_clear(&fs_info->fs_roots_radix,
(unsigned long)root->root_key.objectid,
BTRFS_ROOT_TRANS_TAG);
btrfs_free_log(trans, root);
btrfs_free_reloc_root(trans, root);
btrfs_update_reloc_root(trans, root);
if (root->commit_root == root->node) {
WARN_ON(root->node->start !=
btrfs_root_bytenr(&root->root_item));
free_extent_buffer(root->commit_root);
root->commit_root = NULL;
root->dirty_root = NULL;
spin_lock(&root->list_lock);
list_del_init(&dirty->root->dead_list);
spin_unlock(&root->list_lock);
kfree(dirty->root);
kfree(dirty);
/* make sure to update the root on disk
* so we get any updates to the block used
* counts
*/
err = btrfs_update_root(trans,
root->fs_info->tree_root,
&root->root_key,
&root->root_item);
if (root->commit_root == root->node)
continue;
}
memset(&root->root_item.drop_progress, 0,
sizeof(struct btrfs_disk_key));
root->root_item.drop_level = 0;
root->commit_root = NULL;
root->dirty_root = NULL;
root->root_key.offset = root->fs_info->generation;
btrfs_set_root_bytenr(&root->root_item,
root->node->start);
btrfs_set_root_level(&root->root_item,
btrfs_header_level(root->node));
btrfs_set_root_generation(&root->root_item,
root->root_key.offset);
free_extent_buffer(root->commit_root);
root->commit_root = btrfs_root_node(root);
err = btrfs_insert_root(trans, root->fs_info->tree_root,
btrfs_set_root_node(&root->root_item, root->node);
err = btrfs_update_root(trans, fs_info->tree_root,
&root->root_key,
&root->root_item);
if (err)
break;
refs = btrfs_root_refs(&dirty->root->root_item);
btrfs_set_root_refs(&dirty->root->root_item, refs - 1);
err = btrfs_update_root(trans, root->fs_info->tree_root,
&dirty->root->root_key,
&dirty->root->root_item);
BUG_ON(err);
if (refs == 1) {
list_add(&dirty->list, list);
} else {
WARN_ON(1);
free_extent_buffer(dirty->root->node);
kfree(dirty->root);
kfree(dirty);
}
}
}
return err;
@ -688,12 +611,8 @@ static noinline int wait_transaction_pre_flush(struct btrfs_fs_info *info)
TASK_UNINTERRUPTIBLE);
mutex_unlock(&info->trans_mutex);
atomic_dec(&info->throttles);
wake_up(&info->transaction_throttle);
schedule();
atomic_inc(&info->throttles);
mutex_lock(&info->trans_mutex);
finish_wait(&info->transaction_wait, &wait);
}
@ -705,111 +624,61 @@ static noinline int wait_transaction_pre_flush(struct btrfs_fs_info *info)
* Given a list of roots that need to be deleted, call btrfs_drop_snapshot on
* all of them
*/
static noinline int drop_dirty_roots(struct btrfs_root *tree_root,
struct list_head *list)
int btrfs_drop_dead_root(struct btrfs_root *root)
{
struct btrfs_dirty_root *dirty;
struct btrfs_trans_handle *trans;
struct btrfs_root *tree_root = root->fs_info->tree_root;
unsigned long nr;
u64 num_bytes;
u64 bytes_used;
u64 max_useless;
int ret = 0;
int err;
int ret;
while (!list_empty(list)) {
struct btrfs_root *root;
while (1) {
/*
* we don't want to jump in and create a bunch of
* delayed refs if the transaction is starting to close
*/
wait_transaction_pre_flush(tree_root->fs_info);
trans = btrfs_start_transaction(tree_root, 1);
dirty = list_entry(list->prev, struct btrfs_dirty_root, list);
list_del_init(&dirty->list);
num_bytes = btrfs_root_used(&dirty->root->root_item);
root = dirty->latest_root;
atomic_inc(&root->fs_info->throttles);
while (1) {
/*
* we don't want to jump in and create a bunch of
* delayed refs if the transaction is starting to close
*/
wait_transaction_pre_flush(tree_root->fs_info);
trans = btrfs_start_transaction(tree_root, 1);
/*
* we've joined a transaction, make sure it isn't
* closing right now
*/
if (trans->transaction->delayed_refs.flushing) {
btrfs_end_transaction(trans, tree_root);
continue;
}
mutex_lock(&root->fs_info->drop_mutex);
ret = btrfs_drop_snapshot(trans, dirty->root);
if (ret != -EAGAIN)
break;
mutex_unlock(&root->fs_info->drop_mutex);
err = btrfs_update_root(trans,
tree_root,
&dirty->root->root_key,
&dirty->root->root_item);
if (err)
ret = err;
nr = trans->blocks_used;
ret = btrfs_end_transaction(trans, tree_root);
BUG_ON(ret);
btrfs_btree_balance_dirty(tree_root, nr);
cond_resched();
}
BUG_ON(ret);
atomic_dec(&root->fs_info->throttles);
wake_up(&root->fs_info->transaction_throttle);
num_bytes -= btrfs_root_used(&dirty->root->root_item);
bytes_used = btrfs_root_used(&root->root_item);
if (num_bytes) {
mutex_lock(&root->fs_info->trans_mutex);
btrfs_record_root_in_trans(root);
mutex_unlock(&root->fs_info->trans_mutex);
btrfs_set_root_used(&root->root_item,
bytes_used - num_bytes);
/*
* we've joined a transaction, make sure it isn't
* closing right now
*/
if (trans->transaction->delayed_refs.flushing) {
btrfs_end_transaction(trans, tree_root);
continue;
}
ret = btrfs_del_root(trans, tree_root, &dirty->root->root_key);
if (ret) {
BUG();
ret = btrfs_drop_snapshot(trans, root);
if (ret != -EAGAIN)
break;
}
mutex_unlock(&root->fs_info->drop_mutex);
spin_lock(&root->list_lock);
list_del_init(&dirty->root->dead_list);
if (!list_empty(&root->dead_list)) {
struct btrfs_root *oldest;
oldest = list_entry(root->dead_list.prev,
struct btrfs_root, dead_list);
max_useless = oldest->root_key.offset - 1;
} else {
max_useless = root->root_key.offset - 1;
}
spin_unlock(&root->list_lock);
ret = btrfs_update_root(trans, tree_root,
&root->root_key,
&root->root_item);
if (ret)
break;
nr = trans->blocks_used;
ret = btrfs_end_transaction(trans, tree_root);
BUG_ON(ret);
ret = btrfs_remove_leaf_refs(root, max_useless, 0);
BUG_ON(ret);
free_extent_buffer(dirty->root->node);
kfree(dirty->root);
kfree(dirty);
btrfs_btree_balance_dirty(tree_root, nr);
cond_resched();
}
BUG_ON(ret);
ret = btrfs_del_root(trans, tree_root, &root->root_key);
BUG_ON(ret);
nr = trans->blocks_used;
ret = btrfs_end_transaction(trans, tree_root);
BUG_ON(ret);
free_extent_buffer(root->node);
free_extent_buffer(root->commit_root);
kfree(root);
btrfs_btree_balance_dirty(tree_root, nr);
return ret;
}
@ -839,24 +708,23 @@ static noinline int create_pending_snapshot(struct btrfs_trans_handle *trans,
if (ret)
goto fail;
btrfs_record_root_in_trans(root);
record_root_in_trans(trans, root);
btrfs_set_root_last_snapshot(&root->root_item, trans->transid);
memcpy(new_root_item, &root->root_item, sizeof(*new_root_item));
key.objectid = objectid;
key.offset = trans->transid;
key.offset = 0;
btrfs_set_key_type(&key, BTRFS_ROOT_ITEM_KEY);
old = btrfs_lock_root_node(root);
btrfs_cow_block(trans, root, old, NULL, 0, &old);
btrfs_set_lock_blocking(old);
btrfs_copy_root(trans, root, old, &tmp, objectid);
btrfs_tree_unlock(old);
free_extent_buffer(old);
btrfs_set_root_bytenr(new_root_item, tmp->start);
btrfs_set_root_level(new_root_item, btrfs_header_level(tmp));
btrfs_set_root_generation(new_root_item, trans->transid);
btrfs_set_root_node(new_root_item, tmp);
ret = btrfs_insert_root(trans, root->fs_info->tree_root, &key,
new_root_item);
btrfs_tree_unlock(tmp);
@ -964,6 +832,24 @@ static noinline int finish_pending_snapshots(struct btrfs_trans_handle *trans,
return 0;
}
static void update_super_roots(struct btrfs_root *root)
{
struct btrfs_root_item *root_item;
struct btrfs_super_block *super;
super = &root->fs_info->super_copy;
root_item = &root->fs_info->chunk_root->root_item;
super->chunk_root = root_item->bytenr;
super->chunk_root_generation = root_item->generation;
super->chunk_root_level = root_item->level;
root_item = &root->fs_info->tree_root->root_item;
super->root = root_item->bytenr;
super->generation = root_item->generation;
super->root_level = root_item->level;
}
int btrfs_commit_transaction(struct btrfs_trans_handle *trans,
struct btrfs_root *root)
{
@ -971,8 +857,6 @@ int btrfs_commit_transaction(struct btrfs_trans_handle *trans,
unsigned long timeout = 1;
struct btrfs_transaction *cur_trans;
struct btrfs_transaction *prev_trans = NULL;
struct btrfs_root *chunk_root = root->fs_info->chunk_root;
struct list_head dirty_fs_roots;
struct extent_io_tree *pinned_copy;
DEFINE_WAIT(wait);
int ret;
@ -999,7 +883,6 @@ int btrfs_commit_transaction(struct btrfs_trans_handle *trans,
BUG_ON(ret);
mutex_lock(&root->fs_info->trans_mutex);
INIT_LIST_HEAD(&dirty_fs_roots);
if (cur_trans->in_commit) {
cur_trans->use_count++;
mutex_unlock(&root->fs_info->trans_mutex);
@ -1105,41 +988,36 @@ int btrfs_commit_transaction(struct btrfs_trans_handle *trans,
* with the tree-log code.
*/
mutex_lock(&root->fs_info->tree_log_mutex);
/*
* keep tree reloc code from adding new reloc trees
*/
mutex_lock(&root->fs_info->tree_reloc_mutex);
ret = add_dirty_roots(trans, &root->fs_info->fs_roots_radix,
&dirty_fs_roots);
ret = commit_fs_roots(trans, root);
BUG_ON(ret);
/* add_dirty_roots gets rid of all the tree log roots, it is now
/* commit_fs_roots gets rid of all the tree log roots, it is now
* safe to free the root of tree log roots
*/
btrfs_free_log_root_tree(trans, root->fs_info);
ret = btrfs_commit_tree_roots(trans, root);
ret = commit_cowonly_roots(trans, root);
BUG_ON(ret);
cur_trans = root->fs_info->running_transaction;
spin_lock(&root->fs_info->new_trans_lock);
root->fs_info->running_transaction = NULL;
spin_unlock(&root->fs_info->new_trans_lock);
btrfs_set_super_generation(&root->fs_info->super_copy,
cur_trans->transid);
btrfs_set_super_root(&root->fs_info->super_copy,
root->fs_info->tree_root->node->start);
btrfs_set_super_root_level(&root->fs_info->super_copy,
btrfs_header_level(root->fs_info->tree_root->node));
btrfs_set_super_chunk_root(&root->fs_info->super_copy,
chunk_root->node->start);
btrfs_set_super_chunk_root_level(&root->fs_info->super_copy,
btrfs_header_level(chunk_root->node));
btrfs_set_super_chunk_root_generation(&root->fs_info->super_copy,
btrfs_header_generation(chunk_root->node));
btrfs_set_root_node(&root->fs_info->tree_root->root_item,
root->fs_info->tree_root->node);
free_extent_buffer(root->fs_info->tree_root->commit_root);
root->fs_info->tree_root->commit_root =
btrfs_root_node(root->fs_info->tree_root);
btrfs_set_root_node(&root->fs_info->chunk_root->root_item,
root->fs_info->chunk_root->node);
free_extent_buffer(root->fs_info->chunk_root->commit_root);
root->fs_info->chunk_root->commit_root =
btrfs_root_node(root->fs_info->chunk_root);
update_super_roots(root);
if (!root->fs_info->log_root_recovering) {
btrfs_set_super_log_root(&root->fs_info->super_copy, 0);
@ -1153,7 +1031,6 @@ int btrfs_commit_transaction(struct btrfs_trans_handle *trans,
trans->transaction->blocked = 0;
wake_up(&root->fs_info->transaction_throttle);
wake_up(&root->fs_info->transaction_wait);
mutex_unlock(&root->fs_info->trans_mutex);
@ -1170,9 +1047,6 @@ int btrfs_commit_transaction(struct btrfs_trans_handle *trans,
btrfs_finish_extent_commit(trans, root, pinned_copy);
kfree(pinned_copy);
btrfs_drop_dead_reloc_roots(root);
mutex_unlock(&root->fs_info->tree_reloc_mutex);
/* do the directory inserts of any pending snapshot creations */
finish_pending_snapshots(trans, root->fs_info);
@ -1186,16 +1060,9 @@ int btrfs_commit_transaction(struct btrfs_trans_handle *trans,
put_transaction(cur_trans);
put_transaction(cur_trans);
list_splice_init(&dirty_fs_roots, &root->fs_info->dead_roots);
if (root->fs_info->closing)
list_splice_init(&root->fs_info->dead_roots, &dirty_fs_roots);
mutex_unlock(&root->fs_info->trans_mutex);
kmem_cache_free(btrfs_trans_handle_cachep, trans);
if (root->fs_info->closing)
drop_dirty_roots(root->fs_info->tree_root, &dirty_fs_roots);
return ret;
}
@ -1204,16 +1071,17 @@ int btrfs_commit_transaction(struct btrfs_trans_handle *trans,
*/
int btrfs_clean_old_snapshots(struct btrfs_root *root)
{
struct list_head dirty_roots;
INIT_LIST_HEAD(&dirty_roots);
again:
mutex_lock(&root->fs_info->trans_mutex);
list_splice_init(&root->fs_info->dead_roots, &dirty_roots);
mutex_unlock(&root->fs_info->trans_mutex);
LIST_HEAD(list);
struct btrfs_fs_info *fs_info = root->fs_info;
if (!list_empty(&dirty_roots)) {
drop_dirty_roots(root, &dirty_roots);
goto again;
mutex_lock(&fs_info->trans_mutex);
list_splice_init(&fs_info->dead_roots, &list);
mutex_unlock(&fs_info->trans_mutex);
while (!list_empty(&list)) {
root = list_entry(list.next, struct btrfs_root, root_list);
list_del_init(&root->root_list);
btrfs_drop_dead_root(root);
}
return 0;
}

12
fs/btrfs/transaction.h
View file

@ -62,12 +62,6 @@ struct btrfs_pending_snapshot {
struct list_head list;
};
struct btrfs_dirty_root {
struct list_head list;
struct btrfs_root *root;
struct btrfs_root *latest_root;
};
static inline void btrfs_set_trans_block_group(struct btrfs_trans_handle *trans,
struct inode *inode)
{
@ -100,7 +94,8 @@ int btrfs_write_and_wait_transaction(struct btrfs_trans_handle *trans,
int btrfs_commit_tree_roots(struct btrfs_trans_handle *trans,
struct btrfs_root *root);
int btrfs_add_dead_root(struct btrfs_root *root, struct btrfs_root *latest);
int btrfs_add_dead_root(struct btrfs_root *root);
int btrfs_drop_dead_root(struct btrfs_root *root);
int btrfs_defrag_root(struct btrfs_root *root, int cacheonly);
int btrfs_clean_old_snapshots(struct btrfs_root *root);
int btrfs_commit_transaction(struct btrfs_trans_handle *trans,
@ -108,7 +103,8 @@ int btrfs_commit_transaction(struct btrfs_trans_handle *trans,
int btrfs_end_transaction_throttle(struct btrfs_trans_handle *trans,
struct btrfs_root *root);
void btrfs_throttle(struct btrfs_root *root);
int btrfs_record_root_in_trans(struct btrfs_root *root);
int btrfs_record_root_in_trans(struct btrfs_trans_handle *trans,
struct btrfs_root *root);
int btrfs_write_and_wait_marked_extents(struct btrfs_root *root,
struct extent_io_tree *dirty_pages);
#endif

102
fs/btrfs/tree-log.c
View file

@ -430,18 +430,16 @@ no_copy:
static noinline struct inode *read_one_inode(struct btrfs_root *root,
u64 objectid)
{
struct btrfs_key key;
struct inode *inode;
inode = btrfs_iget_locked(root->fs_info->sb, objectid, root);
if (inode->i_state & I_NEW) {
BTRFS_I(inode)->root = root;
BTRFS_I(inode)->location.objectid = objectid;
BTRFS_I(inode)->location.type = BTRFS_INODE_ITEM_KEY;
BTRFS_I(inode)->location.offset = 0;
btrfs_read_locked_inode(inode);
unlock_new_inode(inode);
}
if (is_bad_inode(inode)) {
key.objectid = objectid;
key.type = BTRFS_INODE_ITEM_KEY;
key.offset = 0;
inode = btrfs_iget(root->fs_info->sb, &key, root);
if (IS_ERR(inode)) {
inode = NULL;
} else if (is_bad_inode(inode)) {
iput(inode);
inode = NULL;
}
@ -541,6 +539,7 @@ static noinline int replay_one_extent(struct btrfs_trans_handle *trans,
if (found_type == BTRFS_FILE_EXTENT_REG ||
found_type == BTRFS_FILE_EXTENT_PREALLOC) {
u64 offset;
unsigned long dest_offset;
struct btrfs_key ins;
@ -555,6 +554,7 @@ static noinline int replay_one_extent(struct btrfs_trans_handle *trans,
ins.objectid = btrfs_file_extent_disk_bytenr(eb, item);
ins.offset = btrfs_file_extent_disk_num_bytes(eb, item);
ins.type = BTRFS_EXTENT_ITEM_KEY;
offset = key->offset - btrfs_file_extent_offset(eb, item);
if (ins.objectid > 0) {
u64 csum_start;
@ -569,19 +569,16 @@ static noinline int replay_one_extent(struct btrfs_trans_handle *trans,
if (ret == 0) {
ret = btrfs_inc_extent_ref(trans, root,
ins.objectid, ins.offset,
path->nodes[0]->start,
root->root_key.objectid,
trans->transid, key->objectid);
0, root->root_key.objectid,
key->objectid, offset);
} else {
/*
* insert the extent pointer in the extent
* allocation tree
*/
ret = btrfs_alloc_logged_extent(trans, root,
path->nodes[0]->start,
root->root_key.objectid,
trans->transid, key->objectid,
&ins);
ret = btrfs_alloc_logged_file_extent(trans,
root, root->root_key.objectid,
key->objectid, offset, &ins);
BUG_ON(ret);
}
btrfs_release_path(root, path);
@ -1706,9 +1703,6 @@ static noinline int walk_down_log_tree(struct btrfs_trans_handle *trans,
btrfs_wait_tree_block_writeback(next);
btrfs_tree_unlock(next);
ret = btrfs_drop_leaf_ref(trans, root, next);
BUG_ON(ret);
WARN_ON(root_owner !=
BTRFS_TREE_LOG_OBJECTID);
ret = btrfs_free_reserved_extent(root,
@ -1753,10 +1747,6 @@ static noinline int walk_down_log_tree(struct btrfs_trans_handle *trans,
btrfs_wait_tree_block_writeback(next);
btrfs_tree_unlock(next);
if (*level == 0) {
ret = btrfs_drop_leaf_ref(trans, root, next);
BUG_ON(ret);
}
WARN_ON(root_owner != BTRFS_TREE_LOG_OBJECTID);
ret = btrfs_free_reserved_extent(root, bytenr, blocksize);
BUG_ON(ret);
@ -1811,12 +1801,6 @@ static noinline int walk_up_log_tree(struct btrfs_trans_handle *trans,
btrfs_wait_tree_block_writeback(next);
btrfs_tree_unlock(next);
if (*level == 0) {
ret = btrfs_drop_leaf_ref(trans, root,
next);
BUG_ON(ret);
}
WARN_ON(root_owner != BTRFS_TREE_LOG_OBJECTID);
ret = btrfs_free_reserved_extent(root,
path->nodes[*level]->start,
@ -1884,11 +1868,6 @@ static int walk_log_tree(struct btrfs_trans_handle *trans,
btrfs_wait_tree_block_writeback(next);
btrfs_tree_unlock(next);
if (orig_level == 0) {
ret = btrfs_drop_leaf_ref(trans, log,
next);
BUG_ON(ret);
}
WARN_ON(log->root_key.objectid !=
BTRFS_TREE_LOG_OBJECTID);
ret = btrfs_free_reserved_extent(log, next->start,
@ -2027,9 +2006,7 @@ int btrfs_sync_log(struct btrfs_trans_handle *trans,
ret = btrfs_write_and_wait_marked_extents(log, &log->dirty_log_pages);
BUG_ON(ret);
btrfs_set_root_bytenr(&log->root_item, log->node->start);
btrfs_set_root_generation(&log->root_item, trans->transid);
btrfs_set_root_level(&log->root_item, btrfs_header_level(log->node));
btrfs_set_root_node(&log->root_item, log->node);
root->log_batch = 0;
root->log_transid++;
@ -2581,7 +2558,7 @@ static noinline int copy_items(struct btrfs_trans_handle *trans,
ins_keys, ins_sizes, nr);
BUG_ON(ret);
for (i = 0; i < nr; i++) {
for (i = 0; i < nr; i++, dst_path->slots[0]++) {
dst_offset = btrfs_item_ptr_offset(dst_path->nodes[0],
dst_path->slots[0]);
@ -2617,36 +2594,31 @@ static noinline int copy_items(struct btrfs_trans_handle *trans,
found_type = btrfs_file_extent_type(src, extent);
if (found_type == BTRFS_FILE_EXTENT_REG ||
found_type == BTRFS_FILE_EXTENT_PREALLOC) {
u64 ds = btrfs_file_extent_disk_bytenr(src,
extent);
u64 dl = btrfs_file_extent_disk_num_bytes(src,
extent);
u64 cs = btrfs_file_extent_offset(src, extent);
u64 cl = btrfs_file_extent_num_bytes(src,
extent);;
u64 ds, dl, cs, cl;
ds = btrfs_file_extent_disk_bytenr(src,
extent);
/* ds == 0 is a hole */
if (ds == 0)
continue;
dl = btrfs_file_extent_disk_num_bytes(src,
extent);
cs = btrfs_file_extent_offset(src, extent);
cl = btrfs_file_extent_num_bytes(src,
extent);;
if (btrfs_file_extent_compression(src,
extent)) {
cs = 0;
cl = dl;
}
/* ds == 0 is a hole */
if (ds != 0) {
ret = btrfs_inc_extent_ref(trans, log,
ds, dl,
dst_path->nodes[0]->start,
BTRFS_TREE_LOG_OBJECTID,
trans->transid,
ins_keys[i].objectid);
BUG_ON(ret);
ret = btrfs_lookup_csums_range(
log->fs_info->csum_root,
ds + cs, ds + cs + cl - 1,
&ordered_sums);
BUG_ON(ret);
}
ret = btrfs_lookup_csums_range(
log->fs_info->csum_root,
ds + cs, ds + cs + cl - 1,
&ordered_sums);
BUG_ON(ret);
}
}
dst_path->slots[0]++;
}
btrfs_mark_buffer_dirty(dst_path->nodes[0]);
@ -3029,9 +3001,7 @@ again:
BUG_ON(!wc.replay_dest);
wc.replay_dest->log_root = log;
mutex_lock(&fs_info->trans_mutex);
btrfs_record_root_in_trans(wc.replay_dest);
mutex_unlock(&fs_info->trans_mutex);
btrfs_record_root_in_trans(trans, wc.replay_dest);
ret = walk_log_tree(trans, log, &wc);
BUG_ON(ret);

2
fs/btrfs/volumes.c
View file

@ -1671,8 +1671,6 @@ static int btrfs_relocate_chunk(struct btrfs_root *root,
int ret;
int i;
printk(KERN_INFO "btrfs relocating chunk %llu\n",
(unsigned long long)chunk_offset);
root = root->fs_info->chunk_root;
extent_root = root->fs_info->extent_root;
em_tree = &root->fs_info->mapping_tree.map_tree;