Btrfs: Add a write ahead tree log to optimize synchronous operations
File syncs and directory syncs are optimized by copying their
items into a special (copy-on-write) log tree. There is one log tree per
subvolume and the btrfs super block points to a tree of log tree roots.
After a crash, items are copied out of the log tree and back into the
subvolume. See tree-log.c for all the details.
Signed-off-by: Chris Mason <chris.mason@oracle.com>
diff --git a/fs/btrfs/Makefile b/fs/btrfs/Makefile
index 3a01065..b7addbf 100644
--- a/fs/btrfs/Makefile
+++ b/fs/btrfs/Makefile
@@ -7,8 +7,7 @@
transaction.o bit-radix.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 acl.o export.o
-
+ ref-cache.o export.o tree-log.o acl.o
else
# Normal Makefile
diff --git a/fs/btrfs/btrfs_inode.h b/fs/btrfs/btrfs_inode.h
index 111f905..fcc8cf2 100644
--- a/fs/btrfs/btrfs_inode.h
+++ b/fs/btrfs/btrfs_inode.h
@@ -33,6 +33,7 @@
struct extent_io_tree io_failure_tree;
struct mutex csum_mutex;
struct mutex extent_mutex;
+ struct mutex log_mutex;
struct inode vfs_inode;
struct btrfs_ordered_inode_tree ordered_tree;
@@ -44,10 +45,17 @@
struct list_head delalloc_inodes;
+ /* full 64 bit generation number */
+ u64 generation;
+
/*
* transid of the trans_handle that last modified this inode
*/
u64 last_trans;
+ /*
+ * transid that last logged this inode
+ */
+ u64 logged_trans;
u64 delalloc_bytes;
u64 disk_i_size;
u32 flags;
diff --git a/fs/btrfs/compat.h b/fs/btrfs/compat.h
index d45fb37..b0ed188 100644
--- a/fs/btrfs/compat.h
+++ b/fs/btrfs/compat.h
@@ -22,6 +22,21 @@
}
#endif
+#if LINUX_VERSION_CODE <= KERNEL_VERSION(2,6,18)
+static inline void btrfs_drop_nlink(struct inode *inode)
+{
+ inode->i_nlink--;
+}
+
+static inline void btrfs_inc_nlink(struct inode *inode)
+{
+ inode->i_nlink++;
+}
+#else
+# define btrfs_drop_nlink(inode) drop_nlink(inode)
+# define btrfs_inc_nlink(inode) inc_nlink(inode)
+#endif
+
/*
* Even if AppArmor isn't enabled, it still has different prototypes.
* Add more distro/version pairs here to declare which has AppArmor applied.
diff --git a/fs/btrfs/ctree.c b/fs/btrfs/ctree.c
index 7114faa..57912404 100644
--- a/fs/btrfs/ctree.c
+++ b/fs/btrfs/ctree.c
@@ -60,7 +60,7 @@
kmem_cache_free(btrfs_path_cachep, p);
}
-void btrfs_release_path(struct btrfs_root *root, struct btrfs_path *p)
+void noinline btrfs_release_path(struct btrfs_root *root, struct btrfs_path *p)
{
int i;
@@ -176,7 +176,7 @@
return 0;
}
-int __btrfs_cow_block(struct btrfs_trans_handle *trans,
+int noinline __btrfs_cow_block(struct btrfs_trans_handle *trans,
struct btrfs_root *root,
struct extent_buffer *buf,
struct extent_buffer *parent, int parent_slot,
@@ -294,7 +294,7 @@
return 0;
}
-int btrfs_cow_block(struct btrfs_trans_handle *trans,
+int noinline btrfs_cow_block(struct btrfs_trans_handle *trans,
struct btrfs_root *root, struct extent_buffer *buf,
struct extent_buffer *parent, int parent_slot,
struct extent_buffer **cow_ret, u64 prealloc_dest)
@@ -677,9 +677,10 @@
*
* slot may point to max if the key is bigger than all of the keys
*/
-static int generic_bin_search(struct extent_buffer *eb, unsigned long p,
- int item_size, struct btrfs_key *key,
- int max, int *slot)
+static noinline int generic_bin_search(struct extent_buffer *eb,
+ unsigned long p,
+ int item_size, struct btrfs_key *key,
+ int max, int *slot)
{
int low = 0;
int high = max;
@@ -765,7 +766,7 @@
return -1;
}
-static struct extent_buffer *read_node_slot(struct btrfs_root *root,
+static noinline struct extent_buffer *read_node_slot(struct btrfs_root *root,
struct extent_buffer *parent, int slot)
{
int level = btrfs_header_level(parent);
@@ -781,7 +782,7 @@
btrfs_node_ptr_generation(parent, slot));
}
-static int balance_level(struct btrfs_trans_handle *trans,
+static noinline int balance_level(struct btrfs_trans_handle *trans,
struct btrfs_root *root,
struct btrfs_path *path, int level)
{
@@ -1128,8 +1129,9 @@
/*
* readahead one full node of leaves
*/
-static void reada_for_search(struct btrfs_root *root, struct btrfs_path *path,
- int level, int slot, u64 objectid)
+static noinline void reada_for_search(struct btrfs_root *root,
+ struct btrfs_path *path,
+ int level, int slot, u64 objectid)
{
struct extent_buffer *node;
struct btrfs_disk_key disk_key;
@@ -1201,7 +1203,8 @@
}
}
-static void unlock_up(struct btrfs_path *path, int level, int lowest_unlock)
+static noinline void unlock_up(struct btrfs_path *path, int level,
+ int lowest_unlock)
{
int i;
int skip_level = level;
@@ -1759,8 +1762,9 @@
*
* returns 0 on success and < 0 on failure
*/
-static int split_node(struct btrfs_trans_handle *trans, struct btrfs_root
- *root, struct btrfs_path *path, int level)
+static noinline int split_node(struct btrfs_trans_handle *trans,
+ struct btrfs_root *root,
+ struct btrfs_path *path, int level)
{
u64 root_gen;
struct extent_buffer *c;
@@ -1874,7 +1878,8 @@
* the start of the leaf data. IOW, how much room
* the leaf has left for both items and data
*/
-int btrfs_leaf_free_space(struct btrfs_root *root, struct extent_buffer *leaf)
+int noinline btrfs_leaf_free_space(struct btrfs_root *root,
+ struct extent_buffer *leaf)
{
int nritems = btrfs_header_nritems(leaf);
int ret;
@@ -2283,9 +2288,11 @@
*
* returns 0 if all went well and < 0 on failure.
*/
-static int split_leaf(struct btrfs_trans_handle *trans, struct btrfs_root
- *root, struct btrfs_key *ins_key,
- struct btrfs_path *path, int data_size, int extend)
+static noinline int split_leaf(struct btrfs_trans_handle *trans,
+ struct btrfs_root *root,
+ struct btrfs_key *ins_key,
+ struct btrfs_path *path, int data_size,
+ int extend)
{
u64 root_gen;
struct extent_buffer *l;
@@ -3079,6 +3086,7 @@
* was nothing in the tree that matched the search criteria.
*/
int btrfs_search_forward(struct btrfs_root *root, struct btrfs_key *min_key,
+ struct btrfs_key *max_key,
struct btrfs_path *path, int cache_only,
u64 min_trans)
{
@@ -3093,6 +3101,7 @@
again:
cur = btrfs_lock_root_node(root);
level = btrfs_header_level(cur);
+ WARN_ON(path->nodes[level]);
path->nodes[level] = cur;
path->locks[level] = 1;
@@ -3107,6 +3116,8 @@
/* at level = 0, we're done, setup the path and exit */
if (level == 0) {
+ if (slot >= nritems)
+ goto find_next_key;
ret = 0;
path->slots[level] = slot;
btrfs_item_key_to_cpu(cur, &found_key, slot);
@@ -3123,6 +3134,8 @@
u64 blockptr;
u64 gen;
struct extent_buffer *tmp;
+ struct btrfs_disk_key disk_key;
+
blockptr = btrfs_node_blockptr(cur, slot);
gen = btrfs_node_ptr_generation(cur, slot);
if (gen < min_trans) {
@@ -3132,6 +3145,14 @@
if (!cache_only)
break;
+ if (max_key) {
+ btrfs_node_key(cur, &disk_key, slot);
+ if (comp_keys(&disk_key, max_key) >= 0) {
+ ret = 1;
+ goto out;
+ }
+ }
+
tmp = btrfs_find_tree_block(root, blockptr,
btrfs_level_size(root, level - 1));
@@ -3143,14 +3164,16 @@
free_extent_buffer(tmp);
slot++;
}
+find_next_key:
/*
* we didn't find a candidate key in this node, walk forward
* and find another one
*/
if (slot >= nritems) {
- ret = btrfs_find_next_key(root, path, min_key, level,
+ path->slots[level] = slot;
+ sret = btrfs_find_next_key(root, path, min_key, level,
cache_only, min_trans);
- if (ret == 0) {
+ if (sret == 0) {
btrfs_release_path(root, path);
goto again;
} else {
@@ -3351,6 +3374,7 @@
{
struct btrfs_key found_key;
struct extent_buffer *leaf;
+ u32 nritems;
int ret;
while(1) {
@@ -3362,9 +3386,20 @@
path->slots[0]--;
}
leaf = path->nodes[0];
+ nritems = btrfs_header_nritems(leaf);
+ if (nritems == 0)
+ return 1;
+ if (path->slots[0] == nritems)
+ path->slots[0]--;
+
btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]);
if (found_key.type == type)
return 0;
+ if (found_key.objectid < min_objectid)
+ break;
+ if (found_key.objectid == min_objectid &&
+ found_key.type < type)
+ break;
}
return 1;
}
diff --git a/fs/btrfs/ctree.h b/fs/btrfs/ctree.h
index b305ae7..6532b60 100644
--- a/fs/btrfs/ctree.h
+++ b/fs/btrfs/ctree.h
@@ -77,6 +77,10 @@
/* orhpan objectid for tracking unlinked/truncated files */
#define BTRFS_ORPHAN_OBJECTID -5ULL
+/* does write ahead logging to speed up fsyncs */
+#define BTRFS_TREE_LOG_OBJECTID -6ULL
+#define BTRFS_TREE_LOG_FIXUP_OBJECTID -7ULL
+
/*
* All files have objectids higher than this.
*/
@@ -276,6 +280,7 @@
__le64 generation;
__le64 root;
__le64 chunk_root;
+ __le64 log_root;
__le64 total_bytes;
__le64 bytes_used;
__le64 root_dir_objectid;
@@ -287,6 +292,7 @@
__le32 sys_chunk_array_size;
u8 root_level;
u8 chunk_root_level;
+ u8 log_root_level;
struct btrfs_dev_item dev_item;
char label[BTRFS_LABEL_SIZE];
u8 sys_chunk_array[BTRFS_SYSTEM_CHUNK_ARRAY_SIZE];
@@ -392,7 +398,10 @@
* make a new item type
*/
struct btrfs_inode_item {
+ /* nfs style generation number */
__le64 generation;
+ /* transid that last touched this inode */
+ __le64 transid;
__le64 size;
__le64 nblocks;
__le64 block_group;
@@ -409,8 +418,13 @@
struct btrfs_timespec otime;
} __attribute__ ((__packed__));
+struct btrfs_dir_log_item {
+ __le64 end;
+} __attribute__ ((__packed__));
+
struct btrfs_dir_item {
struct btrfs_disk_key location;
+ __le64 transid;
__le16 data_len;
__le16 name_len;
u8 type;
@@ -505,6 +519,9 @@
struct btrfs_root *tree_root;
struct btrfs_root *chunk_root;
struct btrfs_root *dev_root;
+
+ /* the log root tree is a directory of all the other log roots */
+ struct btrfs_root *log_root_tree;
struct radix_tree_root fs_roots_radix;
struct extent_io_tree free_space_cache;
@@ -518,6 +535,7 @@
u64 generation;
u64 last_trans_committed;
+ u64 last_trans_new_blockgroup;
u64 open_ioctl_trans;
unsigned long mount_opt;
u64 max_extent;
@@ -527,6 +545,9 @@
wait_queue_head_t transaction_throttle;
wait_queue_head_t transaction_wait;
wait_queue_head_t async_submit_wait;
+
+ wait_queue_head_t tree_log_wait;
+
struct btrfs_super_block super_copy;
struct btrfs_super_block super_for_commit;
struct block_device *__bdev;
@@ -535,6 +556,7 @@
struct backing_dev_info bdi;
spinlock_t hash_lock;
struct mutex trans_mutex;
+ struct mutex tree_log_mutex;
struct mutex transaction_kthread_mutex;
struct mutex cleaner_mutex;
struct mutex alloc_mutex;
@@ -544,8 +566,13 @@
struct list_head trans_list;
struct list_head hashers;
struct list_head dead_roots;
+
atomic_t nr_async_submits;
atomic_t nr_async_bios;
+ atomic_t tree_log_writers;
+ atomic_t tree_log_commit;
+ unsigned long tree_log_batch;
+ u64 tree_log_transid;
/*
* this is used by the balancing code to wait for all the pending
@@ -583,6 +610,7 @@
struct completion kobj_unregister;
int do_barriers;
int closing;
+ int log_root_recovering;
atomic_t throttles;
atomic_t throttle_gen;
@@ -596,6 +624,7 @@
u64 delalloc_bytes;
u64 last_alloc;
u64 last_data_alloc;
+ u64 last_log_alloc;
spinlock_t ref_cache_lock;
u64 total_ref_cache_size;
@@ -632,6 +661,7 @@
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_item root_item;
struct btrfs_key root_key;
@@ -640,6 +670,7 @@
struct kobject root_kobj;
struct completion kobj_unregister;
struct mutex objectid_mutex;
+ struct mutex log_mutex;
u64 objectid;
u64 last_trans;
@@ -692,6 +723,8 @@
* dir items are the name -> inode pointers in a directory. There is one
* for every name in a directory.
*/
+#define BTRFS_DIR_LOG_ITEM_KEY 14
+#define BTRFS_DIR_LOG_INDEX_KEY 15
#define BTRFS_DIR_ITEM_KEY 16
#define BTRFS_DIR_INDEX_KEY 17
/*
@@ -703,7 +736,8 @@
*/
#define BTRFS_CSUM_ITEM_KEY 19
-/* reserve 20-31 for other file stuff */
+
+/* reserve 21-31 for other file/dir stuff */
/*
* root items point to tree roots. There are typically in the root
@@ -938,6 +972,7 @@
/* struct btrfs_inode_item */
BTRFS_SETGET_FUNCS(inode_generation, struct btrfs_inode_item, generation, 64);
+BTRFS_SETGET_FUNCS(inode_transid, struct btrfs_inode_item, transid, 64);
BTRFS_SETGET_FUNCS(inode_size, struct btrfs_inode_item, size, 64);
BTRFS_SETGET_FUNCS(inode_nblocks, struct btrfs_inode_item, nblocks, 64);
BTRFS_SETGET_FUNCS(inode_block_group, struct btrfs_inode_item, block_group, 64);
@@ -1126,10 +1161,13 @@
write_eb_member(eb, item, struct btrfs_item, key, disk_key);
}
+BTRFS_SETGET_FUNCS(dir_log_end, struct btrfs_dir_log_item, end, 64);
+
/* struct btrfs_dir_item */
BTRFS_SETGET_FUNCS(dir_data_len, struct btrfs_dir_item, data_len, 16);
BTRFS_SETGET_FUNCS(dir_type, struct btrfs_dir_item, type, 8);
BTRFS_SETGET_FUNCS(dir_name_len, struct btrfs_dir_item, name_len, 16);
+BTRFS_SETGET_FUNCS(dir_transid, struct btrfs_dir_item, transid, 64);
static inline void btrfs_dir_item_key(struct extent_buffer *eb,
struct btrfs_dir_item *item,
@@ -1301,7 +1339,11 @@
BTRFS_SETGET_STACK_FUNCS(super_chunk_root, struct btrfs_super_block,
chunk_root, 64);
BTRFS_SETGET_STACK_FUNCS(super_chunk_root_level, struct btrfs_super_block,
- chunk_root_level, 64);
+ chunk_root_level, 8);
+BTRFS_SETGET_STACK_FUNCS(super_log_root, struct btrfs_super_block,
+ log_root, 64);
+BTRFS_SETGET_STACK_FUNCS(super_log_root_level, struct btrfs_super_block,
+ log_root_level, 8);
BTRFS_SETGET_STACK_FUNCS(super_total_bytes, struct btrfs_super_block,
total_bytes, 64);
BTRFS_SETGET_STACK_FUNCS(super_bytes_used, struct btrfs_super_block,
@@ -1405,6 +1447,12 @@
}
/* extent-tree.c */
+int btrfs_lookup_extent(struct btrfs_root *root, struct btrfs_path *path,
+ u64 start, u64 len);
+int btrfs_update_pinned_extents(struct btrfs_root *root,
+ u64 bytenr, u64 num, int pin);
+int btrfs_drop_leaf_ref(struct btrfs_trans_handle *trans,
+ struct btrfs_root *root, struct extent_buffer *leaf);
int btrfs_cross_ref_exists(struct btrfs_trans_handle *trans,
struct btrfs_root *root,
struct btrfs_key *key, u64 bytenr);
@@ -1448,6 +1496,11 @@
u64 root_objectid, u64 ref_generation,
u64 owner, u64 owner_offset,
struct btrfs_key *ins);
+int btrfs_alloc_logged_extent(struct btrfs_trans_handle *trans,
+ struct btrfs_root *root,
+ u64 root_objectid, u64 ref_generation,
+ u64 owner, u64 owner_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,
@@ -1488,9 +1541,9 @@
struct btrfs_key *key, int lowest_level,
int cache_only, u64 min_trans);
int btrfs_search_forward(struct btrfs_root *root, struct btrfs_key *min_key,
+ struct btrfs_key *max_key,
struct btrfs_path *path, int cache_only,
u64 min_trans);
-
int btrfs_cow_block(struct btrfs_trans_handle *trans,
struct btrfs_root *root, struct extent_buffer *buf,
struct extent_buffer *parent, int parent_slot,
@@ -1656,6 +1709,18 @@
#define PageChecked PageFsMisc
#endif
+int btrfs_unlink_inode(struct btrfs_trans_handle *trans,
+ struct btrfs_root *root,
+ struct inode *dir, struct inode *inode,
+ const char *name, int name_len);
+int btrfs_add_link(struct btrfs_trans_handle *trans,
+ struct inode *parent_inode, struct inode *inode,
+ const char *name, int name_len, int add_backref, u64 index);
+int btrfs_truncate_inode_items(struct btrfs_trans_handle *trans,
+ struct btrfs_root *root,
+ struct inode *inode, u64 new_size,
+ u32 min_type);
+
int btrfs_start_delalloc_inodes(struct btrfs_root *root);
int btrfs_set_extent_delalloc(struct inode *inode, u64 start, u64 end);
int btrfs_writepages(struct address_space *mapping,
@@ -1715,6 +1780,7 @@
long btrfs_ioctl(struct file *file, unsigned int cmd, unsigned long arg);
/* file.c */
+int btrfs_sync_file(struct file *file, struct dentry *dentry, int datasync);
int btrfs_drop_extent_cache(struct inode *inode, u64 start, u64 end);
int btrfs_check_file(struct btrfs_root *root, struct inode *inode);
extern struct file_operations btrfs_file_operations;
diff --git a/fs/btrfs/dir-item.c b/fs/btrfs/dir-item.c
index 1250946..e4f3009 100644
--- a/fs/btrfs/dir-item.c
+++ b/fs/btrfs/dir-item.c
@@ -96,6 +96,7 @@
btrfs_set_dir_item_key(leaf, dir_item, &disk_key);
btrfs_set_dir_type(leaf, dir_item, BTRFS_FT_XATTR);
btrfs_set_dir_name_len(leaf, dir_item, name_len);
+ btrfs_set_dir_transid(leaf, dir_item, trans->transid);
btrfs_set_dir_data_len(leaf, dir_item, data_len);
name_ptr = (unsigned long)(dir_item + 1);
data_ptr = (unsigned long)((char *)name_ptr + name_len);
@@ -142,6 +143,7 @@
btrfs_set_dir_type(leaf, dir_item, type);
btrfs_set_dir_data_len(leaf, dir_item, 0);
btrfs_set_dir_name_len(leaf, dir_item, name_len);
+ btrfs_set_dir_transid(leaf, dir_item, trans->transid);
name_ptr = (unsigned long)(dir_item + 1);
write_extent_buffer(leaf, name, name_ptr, name_len);
@@ -169,6 +171,7 @@
btrfs_set_dir_type(leaf, dir_item, type);
btrfs_set_dir_data_len(leaf, dir_item, 0);
btrfs_set_dir_name_len(leaf, dir_item, name_len);
+ btrfs_set_dir_transid(leaf, dir_item, trans->transid);
name_ptr = (unsigned long)(dir_item + 1);
write_extent_buffer(leaf, name, name_ptr, name_len);
btrfs_mark_buffer_dirty(leaf);
diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c
index 8e7a938..a4373db 100644
--- a/fs/btrfs/disk-io.c
+++ b/fs/btrfs/disk-io.c
@@ -41,6 +41,7 @@
#include "async-thread.h"
#include "locking.h"
#include "ref-cache.h"
+#include "tree-log.h"
#if 0
static int check_tree_block(struct btrfs_root *root, struct extent_buffer *buf)
@@ -694,6 +695,18 @@
}
+int btrfs_write_tree_block(struct extent_buffer *buf)
+{
+ return btrfs_fdatawrite_range(buf->first_page->mapping, buf->start,
+ buf->start + buf->len - 1, WB_SYNC_NONE);
+}
+
+int btrfs_wait_tree_block_writeback(struct extent_buffer *buf)
+{
+ return btrfs_wait_on_page_writeback_range(buf->first_page->mapping,
+ buf->start, buf->start + buf->len -1);
+}
+
struct extent_buffer *read_tree_block(struct btrfs_root *root, u64 bytenr,
u32 blocksize, u64 parent_transid)
{
@@ -732,15 +745,6 @@
return 0;
}
-int wait_on_tree_block_writeback(struct btrfs_root *root,
- struct extent_buffer *buf)
-{
- struct inode *btree_inode = root->fs_info->btree_inode;
- wait_on_extent_buffer_writeback(&BTRFS_I(btree_inode)->io_tree,
- buf);
- return 0;
-}
-
static int __setup_root(u32 nodesize, u32 leafsize, u32 sectorsize,
u32 stripesize, struct btrfs_root *root,
struct btrfs_fs_info *fs_info,
@@ -771,6 +775,7 @@
spin_lock_init(&root->node_lock);
spin_lock_init(&root->list_lock);
mutex_init(&root->objectid_mutex);
+ mutex_init(&root->log_mutex);
btrfs_leaf_ref_tree_init(&root->ref_tree_struct);
root->ref_tree = &root->ref_tree_struct;
@@ -809,11 +814,74 @@
return 0;
}
-struct btrfs_root *btrfs_read_fs_root_no_radix(struct btrfs_fs_info *fs_info,
- struct btrfs_key *location)
+int btrfs_free_log_root_tree(struct btrfs_trans_handle *trans,
+ struct btrfs_fs_info *fs_info)
+{
+ struct extent_buffer *eb;
+ int ret;
+
+ if (!fs_info->log_root_tree)
+ return 0;
+
+ eb = fs_info->log_root_tree->node;
+
+ WARN_ON(btrfs_header_level(eb) != 0);
+ WARN_ON(btrfs_header_nritems(eb) != 0);
+
+ ret = btrfs_free_extent(trans, fs_info->tree_root,
+ eb->start, eb->len,
+ BTRFS_TREE_LOG_OBJECTID, 0, 0, 0, 1);
+ BUG_ON(ret);
+
+ free_extent_buffer(eb);
+ kfree(fs_info->log_root_tree);
+ fs_info->log_root_tree = NULL;
+ return 0;
+}
+
+int btrfs_init_log_root_tree(struct btrfs_trans_handle *trans,
+ struct btrfs_fs_info *fs_info)
{
struct btrfs_root *root;
struct btrfs_root *tree_root = fs_info->tree_root;
+
+ root = kzalloc(sizeof(*root), GFP_NOFS);
+ if (!root)
+ return -ENOMEM;
+
+ __setup_root(tree_root->nodesize, tree_root->leafsize,
+ tree_root->sectorsize, tree_root->stripesize,
+ root, fs_info, BTRFS_TREE_LOG_OBJECTID);
+
+ root->root_key.objectid = BTRFS_TREE_LOG_OBJECTID;
+ root->root_key.type = BTRFS_ROOT_ITEM_KEY;
+ root->root_key.offset = BTRFS_TREE_LOG_OBJECTID;
+ root->ref_cows = 0;
+
+ root->node = btrfs_alloc_free_block(trans, root, root->leafsize,
+ BTRFS_TREE_LOG_OBJECTID,
+ 0, 0, 0, 0, 0);
+
+ 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),
+ BTRFS_FSID_SIZE);
+ btrfs_mark_buffer_dirty(root->node);
+ btrfs_tree_unlock(root->node);
+ fs_info->log_root_tree = root;
+ return 0;
+}
+
+struct btrfs_root *btrfs_read_fs_root_no_radix(struct btrfs_root *tree_root,
+ struct btrfs_key *location)
+{
+ struct btrfs_root *root;
+ struct btrfs_fs_info *fs_info = tree_root->fs_info;
struct btrfs_path *path;
struct extent_buffer *l;
u64 highest_inode;
@@ -863,11 +931,13 @@
blocksize, 0);
BUG_ON(!root->node);
insert:
- root->ref_cows = 1;
- ret = btrfs_find_highest_inode(root, &highest_inode);
- if (ret == 0) {
- root->highest_inode = highest_inode;
- root->last_inode_alloc = highest_inode;
+ if (location->objectid != BTRFS_TREE_LOG_OBJECTID) {
+ root->ref_cows = 1;
+ ret = btrfs_find_highest_inode(root, &highest_inode);
+ if (ret == 0) {
+ root->highest_inode = highest_inode;
+ root->last_inode_alloc = highest_inode;
+ }
}
return root;
}
@@ -907,7 +977,7 @@
if (root)
return root;
- root = btrfs_read_fs_root_no_radix(fs_info, location);
+ root = btrfs_read_fs_root_no_radix(fs_info->tree_root, location);
if (IS_ERR(root))
return root;
ret = radix_tree_insert(&fs_info->fs_roots_radix,
@@ -1250,16 +1320,18 @@
u32 blocksize;
u32 stripesize;
struct buffer_head *bh;
- struct btrfs_root *extent_root = kmalloc(sizeof(struct btrfs_root),
+ struct btrfs_root *extent_root = kzalloc(sizeof(struct btrfs_root),
GFP_NOFS);
- struct btrfs_root *tree_root = kmalloc(sizeof(struct btrfs_root),
+ struct btrfs_root *tree_root = kzalloc(sizeof(struct btrfs_root),
GFP_NOFS);
struct btrfs_fs_info *fs_info = kzalloc(sizeof(*fs_info),
GFP_NOFS);
- struct btrfs_root *chunk_root = kmalloc(sizeof(struct btrfs_root),
+ struct btrfs_root *chunk_root = kzalloc(sizeof(struct btrfs_root),
GFP_NOFS);
- struct btrfs_root *dev_root = kmalloc(sizeof(struct btrfs_root),
+ struct btrfs_root *dev_root = kzalloc(sizeof(struct btrfs_root),
GFP_NOFS);
+ struct btrfs_root *log_tree_root;
+
int ret;
int err = -EINVAL;
@@ -1343,6 +1415,7 @@
mapping_set_gfp_mask(fs_info->btree_inode->i_mapping, GFP_NOFS);
mutex_init(&fs_info->trans_mutex);
+ mutex_init(&fs_info->tree_log_mutex);
mutex_init(&fs_info->drop_mutex);
mutex_init(&fs_info->alloc_mutex);
mutex_init(&fs_info->chunk_mutex);
@@ -1352,6 +1425,10 @@
init_waitqueue_head(&fs_info->transaction_throttle);
init_waitqueue_head(&fs_info->transaction_wait);
init_waitqueue_head(&fs_info->async_submit_wait);
+ init_waitqueue_head(&fs_info->tree_log_wait);
+ atomic_set(&fs_info->tree_log_commit, 0);
+ atomic_set(&fs_info->tree_log_writers, 0);
+ fs_info->tree_log_transid = 0;
#if 0
ret = add_hasher(fs_info, "crc32c");
@@ -1532,7 +1609,26 @@
if (!fs_info->transaction_kthread)
goto fail_cleaner;
+ if (btrfs_super_log_root(disk_super) != 0) {
+ u32 blocksize;
+ u64 bytenr = btrfs_super_log_root(disk_super);
+ blocksize =
+ btrfs_level_size(tree_root,
+ btrfs_super_log_root_level(disk_super));
+
+ log_tree_root = kzalloc(sizeof(struct btrfs_root),
+ GFP_NOFS);
+
+ __setup_root(nodesize, leafsize, sectorsize, stripesize,
+ log_tree_root, fs_info, BTRFS_TREE_LOG_OBJECTID);
+
+ log_tree_root->node = read_tree_block(tree_root, bytenr,
+ blocksize, 0);
+ ret = btrfs_recover_log_trees(log_tree_root);
+ BUG_ON(ret);
+ }
+ fs_info->last_trans_committed = btrfs_super_generation(disk_super);
return tree_root;
fail_cleaner:
diff --git a/fs/btrfs/disk-io.h b/fs/btrfs/disk-io.h
index 2562a27..6b6fdc6 100644
--- a/fs/btrfs/disk-io.h
+++ b/fs/btrfs/disk-io.h
@@ -45,7 +45,7 @@
struct btrfs_root *btrfs_read_fs_root(struct btrfs_fs_info *fs_info,
struct btrfs_key *location,
const char *name, int namelen);
-struct btrfs_root *btrfs_read_fs_root_no_radix(struct btrfs_fs_info *fs_info,
+struct btrfs_root *btrfs_read_fs_root_no_radix(struct btrfs_root *tree_root,
struct btrfs_key *location);
struct btrfs_root *btrfs_read_fs_root_no_name(struct btrfs_fs_info *fs_info,
struct btrfs_key *location);
@@ -74,4 +74,10 @@
extent_submit_bio_hook_t *submit_bio_hook);
int btrfs_congested_async(struct btrfs_fs_info *info, int iodone);
unsigned long btrfs_async_submit_limit(struct btrfs_fs_info *info);
+int btrfs_write_tree_block(struct extent_buffer *buf);
+int btrfs_wait_tree_block_writeback(struct extent_buffer *buf);
+int btrfs_free_log_root_tree(struct btrfs_trans_handle *trans,
+ struct btrfs_fs_info *fs_info);
+int btrfs_init_log_root_tree(struct btrfs_trans_handle *trans,
+ struct btrfs_fs_info *fs_info);
#endif
diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c
index e63b3b4..646b914 100644
--- a/fs/btrfs/extent-tree.c
+++ b/fs/btrfs/extent-tree.c
@@ -496,6 +496,23 @@
return ret == 0;
}
+/* simple helper to search for an existing extent at a given offset */
+int btrfs_lookup_extent(struct btrfs_root *root, struct btrfs_path *path,
+ u64 start, u64 len)
+{
+ int ret;
+ struct btrfs_key key;
+
+ maybe_lock_mutex(root);
+ key.objectid = start;
+ key.offset = len;
+ btrfs_set_key_type(&key, BTRFS_EXTENT_ITEM_KEY);
+ ret = btrfs_search_slot(NULL, root->fs_info->extent_root, &key, path,
+ 0, 0);
+ maybe_unlock_mutex(root);
+ return ret;
+}
+
static int noinline lookup_extent_backref(struct btrfs_trans_handle *trans,
struct btrfs_root *root,
struct btrfs_path *path, u64 bytenr,
@@ -1409,7 +1426,7 @@
}
-static int update_pinned_extents(struct btrfs_root *root,
+int btrfs_update_pinned_extents(struct btrfs_root *root,
u64 bytenr, u64 num, int pin)
{
u64 len;
@@ -1492,7 +1509,7 @@
EXTENT_DIRTY);
if (ret)
break;
- update_pinned_extents(root, start, end + 1 - start, 0);
+ btrfs_update_pinned_extents(root, start, end + 1 - start, 0);
clear_extent_dirty(unpin, start, end, GFP_NOFS);
set_extent_dirty(free_space_cache, start, end, GFP_NOFS);
if (need_resched()) {
@@ -1538,14 +1555,11 @@
clear_extent_bits(&info->extent_ins, start, end, EXTENT_LOCKED,
GFP_NOFS);
- eb = btrfs_find_tree_block(extent_root, ins.objectid,
+ eb = btrfs_find_create_tree_block(extent_root, ins.objectid,
ins.offset);
- if (!btrfs_buffer_uptodate(eb, trans->transid)) {
- mutex_unlock(&extent_root->fs_info->alloc_mutex);
+ if (!btrfs_buffer_uptodate(eb, trans->transid))
btrfs_read_buffer(eb, trans->transid);
- mutex_lock(&extent_root->fs_info->alloc_mutex);
- }
btrfs_tree_lock(eb);
level = btrfs_header_level(eb);
@@ -1585,13 +1599,20 @@
struct extent_buffer *buf;
buf = btrfs_find_tree_block(root, bytenr, num_bytes);
if (buf) {
+ /* we can reuse a block if it hasn't been written
+ * and it is from this transaction. We can't
+ * reuse anything from the tree log root because
+ * it has tiny sub-transactions.
+ */
if (btrfs_buffer_uptodate(buf, 0) &&
btrfs_try_tree_lock(buf)) {
u64 transid =
root->fs_info->running_transaction->transid;
u64 header_transid =
btrfs_header_generation(buf);
- if (header_transid == transid &&
+ if (btrfs_header_owner(buf) !=
+ BTRFS_TREE_LOG_OBJECTID &&
+ header_transid == transid &&
!btrfs_header_flag(buf,
BTRFS_HEADER_FLAG_WRITTEN)) {
clean_tree_block(NULL, root, buf);
@@ -1603,7 +1624,7 @@
}
free_extent_buffer(buf);
}
- update_pinned_extents(root, bytenr, num_bytes, 1);
+ btrfs_update_pinned_extents(root, bytenr, num_bytes, 1);
} else {
set_extent_bits(&root->fs_info->pending_del,
bytenr, bytenr + num_bytes - 1,
@@ -1801,7 +1822,7 @@
GFP_NOFS);
if (!test_range_bit(&extent_root->fs_info->extent_ins,
start, end, EXTENT_LOCKED, 0)) {
- update_pinned_extents(extent_root, start,
+ btrfs_update_pinned_extents(extent_root, start,
end + 1 - start, 1);
ret = __free_extent(trans, extent_root,
start, end + 1 - start,
@@ -1919,6 +1940,12 @@
if ((data & BTRFS_BLOCK_GROUP_DATA) && btrfs_test_opt(root, SSD)) {
last_ptr = &root->fs_info->last_data_alloc;
}
+ if (root->root_key.objectid == BTRFS_TREE_LOG_OBJECTID) {
+ last_ptr = &root->fs_info->last_log_alloc;
+ if (!last_ptr == 0 && root->fs_info->last_alloc) {
+ *last_ptr = root->fs_info->last_alloc + empty_cluster;
+ }
+ }
if (last_ptr) {
if (*last_ptr)
@@ -2268,6 +2295,35 @@
maybe_unlock_mutex(root);
return ret;
}
+
+/*
+ * this is used by the tree logging recovery code. It records that
+ * an extent has been allocated and makes sure to clear the free
+ * space cache bits as well
+ */
+int btrfs_alloc_logged_extent(struct btrfs_trans_handle *trans,
+ struct btrfs_root *root,
+ u64 root_objectid, u64 ref_generation,
+ u64 owner, u64 owner_offset,
+ struct btrfs_key *ins)
+{
+ int ret;
+ struct btrfs_block_group_cache *block_group;
+
+ maybe_lock_mutex(root);
+ block_group = btrfs_lookup_block_group(root->fs_info, ins->objectid);
+ cache_block_group(root, block_group);
+
+ clear_extent_dirty(&root->fs_info->free_space_cache,
+ ins->objectid, ins->objectid + ins->offset - 1,
+ GFP_NOFS);
+ ret = __btrfs_alloc_reserved_extent(trans, root, root_objectid,
+ ref_generation, owner,
+ owner_offset, ins);
+ maybe_unlock_mutex(root);
+ return ret;
+}
+
/*
* finds a free extent and does all the dirty work required for allocation
* returns the key for the extent through ins, and a tree buffer for
@@ -2350,9 +2406,8 @@
return buf;
}
-static int noinline drop_leaf_ref_no_cache(struct btrfs_trans_handle *trans,
- struct btrfs_root *root,
- struct extent_buffer *leaf)
+int btrfs_drop_leaf_ref(struct btrfs_trans_handle *trans,
+ struct btrfs_root *root, struct extent_buffer *leaf)
{
u64 leaf_owner;
u64 leaf_generation;
@@ -2402,9 +2457,9 @@
return 0;
}
-static int noinline drop_leaf_ref(struct btrfs_trans_handle *trans,
- struct btrfs_root *root,
- struct btrfs_leaf_ref *ref)
+static int noinline cache_drop_leaf_ref(struct btrfs_trans_handle *trans,
+ struct btrfs_root *root,
+ struct btrfs_leaf_ref *ref)
{
int i;
int ret;
@@ -2512,7 +2567,7 @@
btrfs_header_nritems(cur))
break;
if (*level == 0) {
- ret = drop_leaf_ref_no_cache(trans, root, cur);
+ ret = btrfs_drop_leaf_ref(trans, root, cur);
BUG_ON(ret);
break;
}
@@ -2552,7 +2607,7 @@
btrfs_node_key_to_cpu(cur, &key, path->slots[*level]);
ref = btrfs_lookup_leaf_ref(root, bytenr);
if (ref) {
- ret = drop_leaf_ref(trans, root, ref);
+ ret = cache_drop_leaf_ref(trans, root, ref);
BUG_ON(ret);
btrfs_remove_leaf_ref(root, ref);
btrfs_free_leaf_ref(root, ref);
@@ -3628,6 +3683,8 @@
extent_root = root->fs_info->extent_root;
block_group_cache = &root->fs_info->block_group_cache;
+ root->fs_info->last_trans_new_blockgroup = trans->transid;
+
cache = kzalloc(sizeof(*cache), GFP_NOFS);
BUG_ON(!cache);
cache->key.objectid = chunk_offset;
diff --git a/fs/btrfs/file.c b/fs/btrfs/file.c
index e9e86fb..84ecf3a 100644
--- a/fs/btrfs/file.c
+++ b/fs/btrfs/file.c
@@ -36,6 +36,8 @@
#include "btrfs_inode.h"
#include "ioctl.h"
#include "print-tree.h"
+#include "tree-log.h"
+#include "locking.h"
#include "compat.h"
@@ -988,10 +990,27 @@
*ppos = pos;
if (num_written > 0 && ((file->f_flags & O_SYNC) || IS_SYNC(inode))) {
- err = sync_page_range(inode, inode->i_mapping,
- start_pos, num_written);
+ struct btrfs_trans_handle *trans;
+
+ err = btrfs_fdatawrite_range(inode->i_mapping, start_pos,
+ start_pos + num_written -1,
+ WB_SYNC_NONE);
if (err < 0)
num_written = err;
+
+ err = btrfs_wait_on_page_writeback_range(inode->i_mapping,
+ start_pos, start_pos + num_written - 1);
+ if (err < 0)
+ num_written = err;
+
+ trans = btrfs_start_transaction(root, 1);
+ ret = btrfs_log_dentry_safe(trans, root, file->f_dentry);
+ if (ret == 0) {
+ btrfs_sync_log(trans, root);
+ btrfs_end_transaction(trans, root);
+ } else {
+ btrfs_commit_transaction(trans, root);
+ }
} else if (num_written > 0 && (file->f_flags & O_DIRECT)) {
#if LINUX_VERSION_CODE < KERNEL_VERSION(2,6,22)
do_sync_file_range(file, start_pos,
@@ -1019,8 +1038,7 @@
return 0;
}
-static int btrfs_sync_file(struct file *file,
- struct dentry *dentry, int datasync)
+int btrfs_sync_file(struct file *file, struct dentry *dentry, int datasync)
{
struct inode *inode = dentry->d_inode;
struct btrfs_root *root = BTRFS_I(inode)->root;
@@ -1043,6 +1061,8 @@
}
mutex_unlock(&root->fs_info->trans_mutex);
+ filemap_fdatawait(inode->i_mapping);
+
/*
* ok we haven't committed the transaction yet, lets do a commit
*/
@@ -1054,7 +1074,16 @@
ret = -ENOMEM;
goto out;
}
- ret = btrfs_commit_transaction(trans, root);
+
+ ret = btrfs_log_dentry_safe(trans, root, file->f_dentry);
+ if (ret < 0)
+ goto out;
+ if (ret > 0) {
+ ret = btrfs_commit_transaction(trans, root);
+ } else {
+ btrfs_sync_log(trans, root);
+ ret = btrfs_end_transaction(trans, root);
+ }
out:
return ret > 0 ? EIO : ret;
}
diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c
index 43d3f264..65df9d8 100644
--- a/fs/btrfs/inode.c
+++ b/fs/btrfs/inode.c
@@ -46,6 +46,8 @@
#include "volumes.h"
#include "ordered-data.h"
#include "xattr.h"
+#include "compat.h"
+#include "tree-log.h"
struct btrfs_iget_args {
u64 ino;
@@ -586,6 +588,7 @@
&ordered_extent->list);
btrfs_ordered_update_i_size(inode, ordered_extent);
+ btrfs_update_inode(trans, root, inode);
btrfs_remove_ordered_extent(inode, ordered_extent);
/* once for us */
@@ -593,7 +596,6 @@
/* once for the tree */
btrfs_put_ordered_extent(ordered_extent);
- btrfs_update_inode(trans, root, inode);
btrfs_end_transaction(trans, root);
return 0;
}
@@ -1007,7 +1009,8 @@
inode->i_ctime.tv_nsec = btrfs_timespec_nsec(leaf, tspec);
inode->i_blocks = btrfs_inode_nblocks(leaf, inode_item);
- inode->i_generation = btrfs_inode_generation(leaf, inode_item);
+ BTRFS_I(inode)->generation = btrfs_inode_generation(leaf, inode_item);
+ inode->i_generation = BTRFS_I(inode)->generation;
inode->i_rdev = 0;
rdev = btrfs_inode_rdev(leaf, inode_item);
@@ -1056,7 +1059,8 @@
make_bad_inode(inode);
}
-static void fill_inode_item(struct extent_buffer *leaf,
+static void fill_inode_item(struct btrfs_trans_handle *trans,
+ struct extent_buffer *leaf,
struct btrfs_inode_item *item,
struct inode *inode)
{
@@ -1082,7 +1086,8 @@
inode->i_ctime.tv_nsec);
btrfs_set_inode_nblocks(leaf, item, inode->i_blocks);
- btrfs_set_inode_generation(leaf, item, inode->i_generation);
+ btrfs_set_inode_generation(leaf, item, BTRFS_I(inode)->generation);
+ btrfs_set_inode_transid(leaf, item, trans->transid);
btrfs_set_inode_rdev(leaf, item, inode->i_rdev);
btrfs_set_inode_flags(leaf, item, BTRFS_I(inode)->flags);
btrfs_set_inode_block_group(leaf, item,
@@ -1112,7 +1117,7 @@
inode_item = btrfs_item_ptr(leaf, path->slots[0],
struct btrfs_inode_item);
- fill_inode_item(leaf, inode_item, inode);
+ fill_inode_item(trans, leaf, inode_item, inode);
btrfs_mark_buffer_dirty(leaf);
btrfs_set_inode_last_trans(trans, inode);
ret = 0;
@@ -1122,14 +1127,12 @@
}
-static int btrfs_unlink_trans(struct btrfs_trans_handle *trans,
- struct btrfs_root *root,
- struct inode *dir,
- struct dentry *dentry)
+int btrfs_unlink_inode(struct btrfs_trans_handle *trans,
+ struct btrfs_root *root,
+ struct inode *dir, struct inode *inode,
+ const char *name, int name_len)
{
struct btrfs_path *path;
- const char *name = dentry->d_name.name;
- int name_len = dentry->d_name.len;
int ret = 0;
struct extent_buffer *leaf;
struct btrfs_dir_item *di;
@@ -1160,13 +1163,12 @@
btrfs_release_path(root, path);
ret = btrfs_del_inode_ref(trans, root, name, name_len,
- dentry->d_inode->i_ino,
- dentry->d_parent->d_inode->i_ino, &index);
+ inode->i_ino,
+ dir->i_ino, &index);
if (ret) {
printk("failed to delete reference to %.*s, "
"inode %lu parent %lu\n", name_len, name,
- dentry->d_inode->i_ino,
- dentry->d_parent->d_inode->i_ino);
+ inode->i_ino, dir->i_ino);
goto err;
}
@@ -1183,21 +1185,25 @@
ret = btrfs_delete_one_dir_name(trans, root, path, di);
btrfs_release_path(root, path);
- dentry->d_inode->i_ctime = dir->i_ctime;
+ ret = btrfs_del_inode_ref_in_log(trans, root, name, name_len,
+ inode, dir->i_ino);
+ BUG_ON(ret);
+
+ ret = btrfs_del_dir_entries_in_log(trans, root, name, name_len,
+ dir, index);
+ BUG_ON(ret);
err:
btrfs_free_path(path);
- if (!ret) {
- btrfs_i_size_write(dir, dir->i_size - name_len * 2);
- dir->i_mtime = dir->i_ctime = CURRENT_TIME;
- btrfs_update_inode(trans, root, dir);
-#if LINUX_VERSION_CODE <= KERNEL_VERSION(2,6,18)
- dentry->d_inode->i_nlink--;
-#else
- drop_nlink(dentry->d_inode);
-#endif
- ret = btrfs_update_inode(trans, root, dentry->d_inode);
- dir->i_sb->s_dirt = 1;
- }
+ if (ret)
+ goto out;
+
+ btrfs_i_size_write(dir, dir->i_size - name_len * 2);
+ inode->i_ctime = dir->i_mtime = dir->i_ctime = CURRENT_TIME;
+ btrfs_update_inode(trans, root, dir);
+ btrfs_drop_nlink(inode);
+ ret = btrfs_update_inode(trans, root, inode);
+ dir->i_sb->s_dirt = 1;
+out:
return ret;
}
@@ -1218,7 +1224,8 @@
trans = btrfs_start_transaction(root, 1);
btrfs_set_trans_block_group(trans, dir);
- ret = btrfs_unlink_trans(trans, root, dir, dentry);
+ ret = btrfs_unlink_inode(trans, root, dir, dentry->d_inode,
+ dentry->d_name.name, dentry->d_name.len);
if (inode->i_nlink == 0)
ret = btrfs_orphan_add(trans, inode);
@@ -1256,7 +1263,8 @@
goto fail_trans;
/* now the directory is empty */
- err = btrfs_unlink_trans(trans, root, dir, dentry);
+ err = btrfs_unlink_inode(trans, root, dir, dentry->d_inode,
+ dentry->d_name.name, dentry->d_name.len);
if (!err) {
btrfs_i_size_write(inode, 0);
}
@@ -1283,10 +1291,10 @@
* min_type is the minimum key type to truncate down to. If set to 0, this
* will kill all the items on this inode, including the INODE_ITEM_KEY.
*/
-static int btrfs_truncate_in_trans(struct btrfs_trans_handle *trans,
- struct btrfs_root *root,
- struct inode *inode,
- u32 min_type)
+noinline int btrfs_truncate_inode_items(struct btrfs_trans_handle *trans,
+ struct btrfs_root *root,
+ struct inode *inode,
+ u64 new_size, u32 min_type)
{
int ret;
struct btrfs_path *path;
@@ -1307,7 +1315,9 @@
int extent_type = -1;
u64 mask = root->sectorsize - 1;
- btrfs_drop_extent_cache(inode, inode->i_size & (~mask), (u64)-1);
+ if (root->ref_cows)
+ btrfs_drop_extent_cache(inode,
+ new_size & (~mask), (u64)-1);
path = btrfs_alloc_path();
path->reada = -1;
BUG_ON(!path);
@@ -1324,7 +1334,13 @@
goto error;
}
if (ret > 0) {
- BUG_ON(path->slots[0] == 0);
+ /* there are no items in the tree for us to truncate, we're
+ * done
+ */
+ if (path->slots[0] == 0) {
+ ret = 0;
+ goto error;
+ }
path->slots[0]--;
}
@@ -1358,10 +1374,10 @@
}
if (found_type == BTRFS_CSUM_ITEM_KEY) {
ret = btrfs_csum_truncate(trans, root, path,
- inode->i_size);
+ new_size);
BUG_ON(ret);
}
- if (item_end < inode->i_size) {
+ if (item_end < new_size) {
if (found_type == BTRFS_DIR_ITEM_KEY) {
found_type = BTRFS_INODE_ITEM_KEY;
} else if (found_type == BTRFS_EXTENT_ITEM_KEY) {
@@ -1378,7 +1394,7 @@
btrfs_set_key_type(&key, found_type);
goto next;
}
- if (found_key.offset >= inode->i_size)
+ if (found_key.offset >= new_size)
del_item = 1;
else
del_item = 0;
@@ -1394,7 +1410,7 @@
if (!del_item) {
u64 orig_num_bytes =
btrfs_file_extent_num_bytes(leaf, fi);
- extent_num_bytes = inode->i_size -
+ extent_num_bytes = new_size -
found_key.offset + root->sectorsize - 1;
extent_num_bytes = extent_num_bytes &
~((u64)root->sectorsize - 1);
@@ -1402,7 +1418,7 @@
extent_num_bytes);
num_dec = (orig_num_bytes -
extent_num_bytes);
- if (extent_start != 0)
+ if (root->ref_cows && extent_start != 0)
dec_i_blocks(inode, num_dec);
btrfs_mark_buffer_dirty(leaf);
} else {
@@ -1413,22 +1429,29 @@
num_dec = btrfs_file_extent_num_bytes(leaf, fi);
if (extent_start != 0) {
found_extent = 1;
- dec_i_blocks(inode, num_dec);
+ if (root->ref_cows)
+ dec_i_blocks(inode, num_dec);
}
- root_gen = btrfs_header_generation(leaf);
+ if (root->ref_cows) {
+ root_gen =
+ btrfs_header_generation(leaf);
+ }
root_owner = btrfs_header_owner(leaf);
}
} else if (extent_type == BTRFS_FILE_EXTENT_INLINE) {
if (!del_item) {
- u32 newsize = inode->i_size - found_key.offset;
- dec_i_blocks(inode, item_end + 1 -
- found_key.offset - newsize);
- newsize =
- btrfs_file_extent_calc_inline_size(newsize);
+ u32 size = new_size - found_key.offset;
+
+ if (root->ref_cows) {
+ dec_i_blocks(inode, item_end + 1 -
+ found_key.offset - size);
+ }
+ size =
+ btrfs_file_extent_calc_inline_size(size);
ret = btrfs_truncate_item(trans, root, path,
- newsize, 1);
+ size, 1);
BUG_ON(ret);
- } else {
+ } else if (root->ref_cows) {
dec_i_blocks(inode, item_end + 1 -
found_key.offset);
}
@@ -1666,7 +1689,7 @@
trans = btrfs_start_transaction(root, 1);
btrfs_set_trans_block_group(trans, inode);
- ret = btrfs_truncate_in_trans(trans, root, inode, 0);
+ ret = btrfs_truncate_inode_items(trans, root, inode, inode->i_size, 0);
if (ret) {
btrfs_orphan_del(NULL, inode);
goto no_delete_lock;
@@ -1753,15 +1776,20 @@
return 0;
}
-static int btrfs_init_locked_inode(struct inode *inode, void *p)
+static noinline void init_btrfs_i(struct inode *inode)
{
- struct btrfs_iget_args *args = p;
- inode->i_ino = args->ino;
- BTRFS_I(inode)->root = args->root;
- BTRFS_I(inode)->delalloc_bytes = 0;
- inode->i_mapping->writeback_index = 0;
- BTRFS_I(inode)->disk_i_size = 0;
- BTRFS_I(inode)->index_cnt = (u64)-1;
+ struct btrfs_inode *bi = BTRFS_I(inode);
+
+ bi->i_acl = NULL;
+ bi->i_default_acl = NULL;
+
+ bi->generation = 0;
+ bi->last_trans = 0;
+ bi->logged_trans = 0;
+ bi->delalloc_bytes = 0;
+ bi->disk_i_size = 0;
+ bi->flags = 0;
+ bi->index_cnt = (u64)-1;
extent_map_tree_init(&BTRFS_I(inode)->extent_tree, GFP_NOFS);
extent_io_tree_init(&BTRFS_I(inode)->io_tree,
inode->i_mapping, GFP_NOFS);
@@ -1771,6 +1799,15 @@
btrfs_ordered_inode_tree_init(&BTRFS_I(inode)->ordered_tree);
mutex_init(&BTRFS_I(inode)->csum_mutex);
mutex_init(&BTRFS_I(inode)->extent_mutex);
+ mutex_init(&BTRFS_I(inode)->log_mutex);
+}
+
+static int btrfs_init_locked_inode(struct inode *inode, void *p)
+{
+ struct btrfs_iget_args *args = p;
+ inode->i_ino = args->ino;
+ init_btrfs_i(inode);
+ BTRFS_I(inode)->root = args->root;
return 0;
}
@@ -2263,21 +2300,10 @@
* btrfs_get_inode_index_count has an explanation for the magic
* number
*/
+ init_btrfs_i(inode);
BTRFS_I(inode)->index_cnt = 2;
-
- extent_map_tree_init(&BTRFS_I(inode)->extent_tree, GFP_NOFS);
- extent_io_tree_init(&BTRFS_I(inode)->io_tree,
- inode->i_mapping, GFP_NOFS);
- extent_io_tree_init(&BTRFS_I(inode)->io_failure_tree,
- inode->i_mapping, GFP_NOFS);
- btrfs_ordered_inode_tree_init(&BTRFS_I(inode)->ordered_tree);
- INIT_LIST_HEAD(&BTRFS_I(inode)->delalloc_inodes);
- mutex_init(&BTRFS_I(inode)->csum_mutex);
- mutex_init(&BTRFS_I(inode)->extent_mutex);
- BTRFS_I(inode)->delalloc_bytes = 0;
- inode->i_mapping->writeback_index = 0;
- BTRFS_I(inode)->disk_i_size = 0;
BTRFS_I(inode)->root = root;
+ BTRFS_I(inode)->generation = trans->transid;
if (mode & S_IFDIR)
owner = 0;
@@ -2290,7 +2316,6 @@
new_inode_group = group;
}
BTRFS_I(inode)->block_group = new_inode_group;
- BTRFS_I(inode)->flags = 0;
key[0].objectid = objectid;
btrfs_set_key_type(&key[0], BTRFS_INODE_ITEM_KEY);
@@ -2318,7 +2343,7 @@
inode->i_mtime = inode->i_atime = inode->i_ctime = CURRENT_TIME;
inode_item = btrfs_item_ptr(path->nodes[0], path->slots[0],
struct btrfs_inode_item);
- fill_inode_item(path->nodes[0], inode_item, inode);
+ fill_inode_item(trans, path->nodes[0], inode_item, inode);
ref = btrfs_item_ptr(path->nodes[0], path->slots[0] + 1,
struct btrfs_inode_ref);
@@ -2349,38 +2374,34 @@
return btrfs_type_by_mode[(inode->i_mode & S_IFMT) >> S_SHIFT];
}
-static int btrfs_add_link(struct btrfs_trans_handle *trans,
- struct dentry *dentry, struct inode *inode,
- int add_backref, u64 index)
+int btrfs_add_link(struct btrfs_trans_handle *trans,
+ struct inode *parent_inode, struct inode *inode,
+ const char *name, int name_len, int add_backref, u64 index)
{
int ret;
struct btrfs_key key;
- struct btrfs_root *root = BTRFS_I(dentry->d_parent->d_inode)->root;
- struct inode *parent_inode = dentry->d_parent->d_inode;
+ struct btrfs_root *root = BTRFS_I(parent_inode)->root;
key.objectid = inode->i_ino;
btrfs_set_key_type(&key, BTRFS_INODE_ITEM_KEY);
key.offset = 0;
- ret = btrfs_insert_dir_item(trans, root,
- dentry->d_name.name, dentry->d_name.len,
- dentry->d_parent->d_inode->i_ino,
+ ret = btrfs_insert_dir_item(trans, root, name, name_len,
+ parent_inode->i_ino,
&key, btrfs_inode_type(inode),
index);
if (ret == 0) {
if (add_backref) {
ret = btrfs_insert_inode_ref(trans, root,
- dentry->d_name.name,
- dentry->d_name.len,
- inode->i_ino,
- parent_inode->i_ino,
- index);
+ name, name_len,
+ inode->i_ino,
+ parent_inode->i_ino,
+ index);
}
btrfs_i_size_write(parent_inode, parent_inode->i_size +
- dentry->d_name.len * 2);
+ name_len * 2);
parent_inode->i_mtime = parent_inode->i_ctime = CURRENT_TIME;
- ret = btrfs_update_inode(trans, root,
- dentry->d_parent->d_inode);
+ ret = btrfs_update_inode(trans, root, parent_inode);
}
return ret;
}
@@ -2389,7 +2410,9 @@
struct dentry *dentry, struct inode *inode,
int backref, u64 index)
{
- int err = btrfs_add_link(trans, dentry, inode, backref, index);
+ int err = btrfs_add_link(trans, dentry->d_parent->d_inode,
+ inode, dentry->d_name.name,
+ dentry->d_name.len, backref, index);
if (!err) {
d_instantiate(dentry, inode);
return 0;
@@ -2513,19 +2536,7 @@
inode->i_mapping->backing_dev_info = &root->fs_info->bdi;
inode->i_fop = &btrfs_file_operations;
inode->i_op = &btrfs_file_inode_operations;
- extent_map_tree_init(&BTRFS_I(inode)->extent_tree, GFP_NOFS);
- extent_io_tree_init(&BTRFS_I(inode)->io_tree,
- inode->i_mapping, GFP_NOFS);
- extent_io_tree_init(&BTRFS_I(inode)->io_failure_tree,
- inode->i_mapping, GFP_NOFS);
- INIT_LIST_HEAD(&BTRFS_I(inode)->delalloc_inodes);
- mutex_init(&BTRFS_I(inode)->csum_mutex);
- mutex_init(&BTRFS_I(inode)->extent_mutex);
- BTRFS_I(inode)->delalloc_bytes = 0;
- BTRFS_I(inode)->disk_i_size = 0;
- inode->i_mapping->writeback_index = 0;
BTRFS_I(inode)->io_tree.ops = &btrfs_extent_io_ops;
- btrfs_ordered_inode_tree_init(&BTRFS_I(inode)->ordered_tree);
}
dir->i_sb->s_dirt = 1;
btrfs_update_inode_block_group(trans, inode);
@@ -2556,11 +2567,7 @@
if (inode->i_nlink == 0)
return -ENOENT;
-#if LINUX_VERSION_CODE <= KERNEL_VERSION(2,6,18)
- inode->i_nlink++;
-#else
- inc_nlink(inode);
-#endif
+ btrfs_inc_nlink(inode);
err = btrfs_check_free_space(root, 1, 0);
if (err)
goto fail;
@@ -2650,7 +2657,9 @@
if (err)
goto out_fail;
- err = btrfs_add_link(trans, dentry, inode, 0, index);
+ err = btrfs_add_link(trans, dentry->d_parent->d_inode,
+ inode, dentry->d_name.name,
+ dentry->d_name.len, 0, index);
if (err)
goto out_fail;
@@ -3221,7 +3230,7 @@
if (ret)
goto out;
/* FIXME, add redo link to tree so we don't leak on crash */
- ret = btrfs_truncate_in_trans(trans, root, inode,
+ ret = btrfs_truncate_inode_items(trans, root, inode, inode->i_size,
BTRFS_EXTENT_DATA_KEY);
btrfs_update_inode(trans, root, inode);
@@ -3304,6 +3313,7 @@
if (!ei)
return NULL;
ei->last_trans = 0;
+ ei->logged_trans = 0;
btrfs_ordered_inode_tree_init(&ei->ordered_tree);
ei->i_acl = BTRFS_ACL_NOT_CACHED;
ei->i_default_acl = BTRFS_ACL_NOT_CACHED;
@@ -3463,31 +3473,39 @@
btrfs_set_trans_block_group(trans, new_dir);
- old_dentry->d_inode->i_nlink++;
+ btrfs_inc_nlink(old_dentry->d_inode);
old_dir->i_ctime = old_dir->i_mtime = ctime;
new_dir->i_ctime = new_dir->i_mtime = ctime;
old_inode->i_ctime = ctime;
- ret = btrfs_unlink_trans(trans, root, old_dir, old_dentry);
+ ret = btrfs_unlink_inode(trans, root, old_dir, old_dentry->d_inode,
+ old_dentry->d_name.name,
+ old_dentry->d_name.len);
if (ret)
goto out_fail;
if (new_inode) {
new_inode->i_ctime = CURRENT_TIME;
- ret = btrfs_unlink_trans(trans, root, new_dir, new_dentry);
+ ret = btrfs_unlink_inode(trans, root, new_dir,
+ new_dentry->d_inode,
+ new_dentry->d_name.name,
+ new_dentry->d_name.len);
if (ret)
goto out_fail;
if (new_inode->i_nlink == 0) {
- ret = btrfs_orphan_add(trans, new_inode);
+ ret = btrfs_orphan_add(trans, new_dentry->d_inode);
if (ret)
goto out_fail;
}
+
}
ret = btrfs_set_inode_index(new_dir, old_inode, &index);
if (ret)
goto out_fail;
- ret = btrfs_add_link(trans, new_dentry, old_inode, 1, index);
+ ret = btrfs_add_link(trans, new_dentry->d_parent->d_inode,
+ old_inode, new_dentry->d_name.name,
+ new_dentry->d_name.len, 1, index);
if (ret)
goto out_fail;
@@ -3577,19 +3595,7 @@
inode->i_mapping->backing_dev_info = &root->fs_info->bdi;
inode->i_fop = &btrfs_file_operations;
inode->i_op = &btrfs_file_inode_operations;
- extent_map_tree_init(&BTRFS_I(inode)->extent_tree, GFP_NOFS);
- extent_io_tree_init(&BTRFS_I(inode)->io_tree,
- inode->i_mapping, GFP_NOFS);
- extent_io_tree_init(&BTRFS_I(inode)->io_failure_tree,
- inode->i_mapping, GFP_NOFS);
- INIT_LIST_HEAD(&BTRFS_I(inode)->delalloc_inodes);
- mutex_init(&BTRFS_I(inode)->csum_mutex);
- mutex_init(&BTRFS_I(inode)->extent_mutex);
- BTRFS_I(inode)->delalloc_bytes = 0;
- BTRFS_I(inode)->disk_i_size = 0;
- inode->i_mapping->writeback_index = 0;
BTRFS_I(inode)->io_tree.ops = &btrfs_extent_io_ops;
- btrfs_ordered_inode_tree_init(&BTRFS_I(inode)->ordered_tree);
}
dir->i_sb->s_dirt = 1;
btrfs_update_inode_block_group(trans, inode);
@@ -3691,6 +3697,7 @@
.compat_ioctl = btrfs_ioctl,
#endif
.release = btrfs_release_file,
+ .fsync = btrfs_sync_file,
};
static struct extent_io_ops btrfs_extent_io_ops = {
diff --git a/fs/btrfs/root-tree.c b/fs/btrfs/root-tree.c
index 3672669..e3984f9 100644
--- a/fs/btrfs/root-tree.c
+++ b/fs/btrfs/root-tree.c
@@ -202,8 +202,9 @@
memcpy(&found_key, &key, sizeof(key));
key.offset++;
btrfs_release_path(root, path);
- dead_root = btrfs_read_fs_root_no_radix(root->fs_info,
- &found_key);
+ dead_root =
+ btrfs_read_fs_root_no_radix(root->fs_info->tree_root,
+ &found_key);
if (IS_ERR(dead_root)) {
ret = PTR_ERR(dead_root);
goto err;
diff --git a/fs/btrfs/transaction.c b/fs/btrfs/transaction.c
index eff3ad7..49c4f5b4 100644
--- a/fs/btrfs/transaction.c
+++ b/fs/btrfs/transaction.c
@@ -25,6 +25,7 @@
#include "transaction.h"
#include "locking.h"
#include "ref-cache.h"
+#include "tree-log.h"
static int total_trans = 0;
extern struct kmem_cache *btrfs_trans_handle_cachep;
@@ -57,6 +58,7 @@
root->fs_info->generation++;
root->fs_info->last_alloc = 0;
root->fs_info->last_data_alloc = 0;
+ root->fs_info->last_log_alloc = 0;
cur_trans->num_writers = 1;
cur_trans->num_joined = 0;
cur_trans->transid = root->fs_info->generation;
@@ -83,7 +85,7 @@
return 0;
}
-static noinline int record_root_in_trans(struct btrfs_root *root)
+noinline int btrfs_record_root_in_trans(struct btrfs_root *root)
{
struct btrfs_dirty_root *dirty;
u64 running_trans_id = root->fs_info->running_transaction->transid;
@@ -151,7 +153,7 @@
}
}
-struct btrfs_trans_handle *start_transaction(struct btrfs_root *root,
+static struct btrfs_trans_handle *start_transaction(struct btrfs_root *root,
int num_blocks, int wait)
{
struct btrfs_trans_handle *h =
@@ -164,7 +166,7 @@
ret = join_transaction(root);
BUG_ON(ret);
- record_root_in_trans(root);
+ 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;
@@ -456,6 +458,8 @@
BUG_ON(!root->ref_tree);
dirty = root->dirty_root;
+ btrfs_free_log(trans, root);
+
if (root->commit_root == root->node) {
WARN_ON(root->node->start !=
btrfs_root_bytenr(&root->root_item));
@@ -600,7 +604,7 @@
num_bytes -= btrfs_root_used(&dirty->root->root_item);
bytes_used = btrfs_root_used(&root->root_item);
if (num_bytes) {
- record_root_in_trans(root);
+ btrfs_record_root_in_trans(root);
btrfs_set_root_used(&root->root_item,
bytes_used - num_bytes);
}
@@ -745,7 +749,6 @@
int ret;
INIT_LIST_HEAD(&dirty_fs_roots);
-
mutex_lock(&root->fs_info->trans_mutex);
if (trans->transaction->in_commit) {
cur_trans = trans->transaction;
@@ -821,10 +824,30 @@
WARN_ON(cur_trans != trans->transaction);
+ /* btrfs_commit_tree_roots is responsible for getting the
+ * various roots consistent with each other. Every pointer
+ * in the tree of tree roots has to point to the most up to date
+ * root for every subvolume and other tree. So, we have to keep
+ * the tree logging code from jumping in and changing any
+ * of the trees.
+ *
+ * At this point in the commit, there can't be any tree-log
+ * writers, but a little lower down we drop the trans mutex
+ * and let new people in. By holding the tree_log_mutex
+ * from now until after the super is written, we avoid races
+ * with the tree-log code.
+ */
+ mutex_lock(&root->fs_info->tree_log_mutex);
+
ret = add_dirty_roots(trans, &root->fs_info->fs_roots_radix,
&dirty_fs_roots);
BUG_ON(ret);
+ /* add_dirty_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);
BUG_ON(ret);
@@ -843,6 +866,12 @@
chunk_root->node->start);
btrfs_set_super_chunk_root_level(&root->fs_info->super_copy,
btrfs_header_level(chunk_root->node));
+
+ if (!root->fs_info->log_root_recovering) {
+ btrfs_set_super_log_root(&root->fs_info->super_copy, 0);
+ btrfs_set_super_log_root_level(&root->fs_info->super_copy, 0);
+ }
+
memcpy(&root->fs_info->super_for_commit, &root->fs_info->super_copy,
sizeof(root->fs_info->super_copy));
@@ -857,6 +886,12 @@
BUG_ON(ret);
write_ctree_super(trans, root);
+ /*
+ * the super is written, we can safely allow the tree-loggers
+ * to go about their business
+ */
+ mutex_unlock(&root->fs_info->tree_log_mutex);
+
btrfs_finish_extent_commit(trans, root, pinned_copy);
mutex_lock(&root->fs_info->trans_mutex);
diff --git a/fs/btrfs/transaction.h b/fs/btrfs/transaction.h
index 598baa3..cc63650 100644
--- a/fs/btrfs/transaction.h
+++ b/fs/btrfs/transaction.h
@@ -98,4 +98,5 @@
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);
#endif
diff --git a/fs/btrfs/tree-defrag.c b/fs/btrfs/tree-defrag.c
index cc2650b..b3bb5bb 100644
--- a/fs/btrfs/tree-defrag.c
+++ b/fs/btrfs/tree-defrag.c
@@ -81,12 +81,12 @@
memcpy(&key, &root->defrag_progress, sizeof(key));
}
- path->lowest_level = 1;
path->keep_locks = 1;
if (cache_only)
min_trans = root->defrag_trans_start;
- ret = btrfs_search_forward(root, &key, path, cache_only, min_trans);
+ ret = btrfs_search_forward(root, &key, NULL, path,
+ cache_only, min_trans);
if (ret < 0)
goto out;
if (ret > 0) {
diff --git a/fs/btrfs/tree-log.c b/fs/btrfs/tree-log.c
new file mode 100644
index 0000000..d1ce831
--- /dev/null
+++ b/fs/btrfs/tree-log.c
@@ -0,0 +1,2804 @@
+/*
+ * Copyright (C) 2008 Oracle. All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public
+ * License v2 as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ * General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public
+ * License along with this program; if not, write to the
+ * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
+ * Boston, MA 021110-1307, USA.
+ */
+
+#include <linux/sched.h>
+#include "ctree.h"
+#include "transaction.h"
+#include "disk-io.h"
+#include "locking.h"
+#include "print-tree.h"
+#include "compat.h"
+
+/* magic values for the inode_only field in btrfs_log_inode:
+ *
+ * LOG_INODE_ALL means to log everything
+ * LOG_INODE_EXISTS means to log just enough to recreate the inode
+ * during log replay
+ */
+#define LOG_INODE_ALL 0
+#define LOG_INODE_EXISTS 1
+
+/*
+ * stages for the tree walking. The first
+ * stage (0) is to only pin down the blocks we find
+ * the second stage (1) is to make sure that all the inodes
+ * we find in the log are created in the subvolume.
+ *
+ * The last stage is to deal with directories and links and extents
+ * and all the other fun semantics
+ */
+#define LOG_WALK_PIN_ONLY 0
+#define LOG_WALK_REPLAY_INODES 1
+#define LOG_WALK_REPLAY_ALL 2
+
+static int __btrfs_log_inode(struct btrfs_trans_handle *trans,
+ struct btrfs_root *root, struct inode *inode,
+ int inode_only);
+
+/*
+ * tree logging is a special write ahead log used to make sure that
+ * fsyncs and O_SYNCs can happen without doing full tree commits.
+ *
+ * Full tree commits are expensive because they require commonly
+ * modified blocks to be recowed, creating many dirty pages in the
+ * extent tree an 4x-6x higher write load than ext3.
+ *
+ * Instead of doing a tree commit on every fsync, we use the
+ * key ranges and transaction ids to find items for a given file or directory
+ * that have changed in this transaction. Those items are copied into
+ * a special tree (one per subvolume root), that tree is written to disk
+ * and then the fsync is considered complete.
+ *
+ * After a crash, items are copied out of the log-tree back into the
+ * subvolume tree. Any file data extents found are recorded in the extent
+ * allocation tree, and the log-tree freed.
+ *
+ * The log tree is read three times, once to pin down all the extents it is
+ * using in ram and once, once to create all the inodes logged in the tree
+ * and once to do all the other items.
+ */
+
+/*
+ * btrfs_add_log_tree adds a new per-subvolume log tree into the
+ * tree of log tree roots. This must be called with a tree log transaction
+ * running (see start_log_trans).
+ */
+int btrfs_add_log_tree(struct btrfs_trans_handle *trans,
+ struct btrfs_root *root)
+{
+ struct btrfs_key key;
+ struct btrfs_root_item root_item;
+ struct btrfs_inode_item *inode_item;
+ struct extent_buffer *leaf;
+ struct btrfs_root *new_root = root;
+ int ret;
+ u64 objectid = root->root_key.objectid;
+
+ leaf = btrfs_alloc_free_block(trans, root, root->leafsize,
+ BTRFS_TREE_LOG_OBJECTID,
+ 0, 0, 0, 0, 0);
+ if (IS_ERR(leaf)) {
+ ret = PTR_ERR(leaf);
+ return ret;
+ }
+
+ btrfs_set_header_nritems(leaf, 0);
+ btrfs_set_header_level(leaf, 0);
+ btrfs_set_header_bytenr(leaf, leaf->start);
+ btrfs_set_header_generation(leaf, trans->transid);
+ btrfs_set_header_owner(leaf, BTRFS_TREE_LOG_OBJECTID);
+
+ write_extent_buffer(leaf, root->fs_info->fsid,
+ (unsigned long)btrfs_header_fsid(leaf),
+ BTRFS_FSID_SIZE);
+ btrfs_mark_buffer_dirty(leaf);
+
+ inode_item = &root_item.inode;
+ memset(inode_item, 0, sizeof(*inode_item));
+ inode_item->generation = cpu_to_le64(1);
+ inode_item->size = cpu_to_le64(3);
+ inode_item->nlink = cpu_to_le32(1);
+ inode_item->nblocks = cpu_to_le64(1);
+ inode_item->mode = cpu_to_le32(S_IFDIR | 0755);
+
+ btrfs_set_root_bytenr(&root_item, leaf->start);
+ btrfs_set_root_level(&root_item, 0);
+ btrfs_set_root_refs(&root_item, 0);
+ btrfs_set_root_used(&root_item, 0);
+
+ memset(&root_item.drop_progress, 0, sizeof(root_item.drop_progress));
+ root_item.drop_level = 0;
+
+ btrfs_tree_unlock(leaf);
+ free_extent_buffer(leaf);
+ leaf = NULL;
+
+ btrfs_set_root_dirid(&root_item, 0);
+
+ key.objectid = BTRFS_TREE_LOG_OBJECTID;
+ key.offset = objectid;
+ btrfs_set_key_type(&key, BTRFS_ROOT_ITEM_KEY);
+ ret = btrfs_insert_root(trans, root->fs_info->log_root_tree, &key,
+ &root_item);
+ if (ret)
+ goto fail;
+
+ new_root = btrfs_read_fs_root_no_radix(root->fs_info->log_root_tree,
+ &key);
+ BUG_ON(!new_root);
+
+ WARN_ON(root->log_root);
+ root->log_root = new_root;
+
+ /*
+ * log trees do not get reference counted because they go away
+ * before a real commit is actually done. They do store pointers
+ * to file data extents, and those reference counts still get
+ * updated (along with back refs to the log tree).
+ */
+ new_root->ref_cows = 0;
+ new_root->last_trans = trans->transid;
+fail:
+ return ret;
+}
+
+/*
+ * start a sub transaction and setup the log tree
+ * this increments the log tree writer count to make the people
+ * syncing the tree wait for us to finish
+ */
+static int start_log_trans(struct btrfs_trans_handle *trans,
+ struct btrfs_root *root)
+{
+ int ret;
+ mutex_lock(&root->fs_info->tree_log_mutex);
+ if (!root->fs_info->log_root_tree) {
+ ret = btrfs_init_log_root_tree(trans, root->fs_info);
+ BUG_ON(ret);
+ }
+ if (!root->log_root) {
+ ret = btrfs_add_log_tree(trans, root);
+ BUG_ON(ret);
+ }
+ atomic_inc(&root->fs_info->tree_log_writers);
+ root->fs_info->tree_log_batch++;
+ mutex_unlock(&root->fs_info->tree_log_mutex);
+ return 0;
+}
+
+/*
+ * returns 0 if there was a log transaction running and we were able
+ * to join, or returns -ENOENT if there were not transactions
+ * in progress
+ */
+static int join_running_log_trans(struct btrfs_root *root)
+{
+ int ret = -ENOENT;
+
+ smp_mb();
+ if (!root->log_root)
+ return -ENOENT;
+
+ mutex_lock(&root->fs_info->tree_log_mutex);
+ if (root->log_root) {
+ ret = 0;
+ atomic_inc(&root->fs_info->tree_log_writers);
+ root->fs_info->tree_log_batch++;
+ }
+ mutex_unlock(&root->fs_info->tree_log_mutex);
+ return ret;
+}
+
+/*
+ * indicate we're done making changes to the log tree
+ * and wake up anyone waiting to do a sync
+ */
+static int end_log_trans(struct btrfs_root *root)
+{
+ atomic_dec(&root->fs_info->tree_log_writers);
+ smp_mb();
+ if (waitqueue_active(&root->fs_info->tree_log_wait))
+ wake_up(&root->fs_info->tree_log_wait);
+ return 0;
+}
+
+
+/*
+ * the walk control struct is used to pass state down the chain when
+ * processing the log tree. The stage field tells us which part
+ * of the log tree processing we are currently doing. The others
+ * are state fields used for that specific part
+ */
+struct walk_control {
+ /* should we free the extent on disk when done? This is used
+ * at transaction commit time while freeing a log tree
+ */
+ int free;
+
+ /* should we write out the extent buffer? This is used
+ * while flushing the log tree to disk during a sync
+ */
+ int write;
+
+ /* should we wait for the extent buffer io to finish? Also used
+ * while flushing the log tree to disk for a sync
+ */
+ int wait;
+
+ /* pin only walk, we record which extents on disk belong to the
+ * log trees
+ */
+ int pin;
+
+ /* what stage of the replay code we're currently in */
+ int stage;
+
+ /* the root we are currently replaying */
+ struct btrfs_root *replay_dest;
+
+ /* the trans handle for the current replay */
+ struct btrfs_trans_handle *trans;
+
+ /* the function that gets used to process blocks we find in the
+ * tree. Note the extent_buffer might not be up to date when it is
+ * passed in, and it must be checked or read if you need the data
+ * inside it
+ */
+ int (*process_func)(struct btrfs_root *log, struct extent_buffer *eb,
+ struct walk_control *wc, u64 gen);
+};
+
+/*
+ * process_func used to pin down extents, write them or wait on them
+ */
+static int process_one_buffer(struct btrfs_root *log,
+ struct extent_buffer *eb,
+ struct walk_control *wc, u64 gen)
+{
+ if (wc->pin) {
+ mutex_lock(&log->fs_info->alloc_mutex);
+ btrfs_update_pinned_extents(log->fs_info->extent_root,
+ eb->start, eb->len, 1);
+ mutex_unlock(&log->fs_info->alloc_mutex);
+ }
+
+ if (btrfs_buffer_uptodate(eb, gen)) {
+ if (wc->write)
+ btrfs_write_tree_block(eb);
+ if (wc->wait)
+ btrfs_wait_tree_block_writeback(eb);
+ }
+ return 0;
+}
+
+/*
+ * Item overwrite used by replay and tree logging. eb, slot and key all refer
+ * to the src data we are copying out.
+ *
+ * root is the tree we are copying into, and path is a scratch
+ * path for use in this function (it should be released on entry and
+ * will be released on exit).
+ *
+ * If the key is already in the destination tree the existing item is
+ * overwritten. If the existing item isn't big enough, it is extended.
+ * If it is too large, it is truncated.
+ *
+ * If the key isn't in the destination yet, a new item is inserted.
+ */
+static noinline int overwrite_item(struct btrfs_trans_handle *trans,
+ struct btrfs_root *root,
+ struct btrfs_path *path,
+ struct extent_buffer *eb, int slot,
+ struct btrfs_key *key)
+{
+ int ret;
+ u32 item_size;
+ u64 saved_i_size = 0;
+ int save_old_i_size = 0;
+ unsigned long src_ptr;
+ unsigned long dst_ptr;
+ int overwrite_root = 0;
+
+ if (root->root_key.objectid != BTRFS_TREE_LOG_OBJECTID)
+ overwrite_root = 1;
+
+ item_size = btrfs_item_size_nr(eb, slot);
+ src_ptr = btrfs_item_ptr_offset(eb, slot);
+
+ /* look for the key in the destination tree */
+ ret = btrfs_search_slot(NULL, root, key, path, 0, 0);
+ if (ret == 0) {
+ char *src_copy;
+ char *dst_copy;
+ u32 dst_size = btrfs_item_size_nr(path->nodes[0],
+ path->slots[0]);
+ if (dst_size != item_size)
+ goto insert;
+
+ if (item_size == 0) {
+ btrfs_release_path(root, path);
+ return 0;
+ }
+ dst_copy = kmalloc(item_size, GFP_NOFS);
+ src_copy = kmalloc(item_size, GFP_NOFS);
+
+ read_extent_buffer(eb, src_copy, src_ptr, item_size);
+
+ dst_ptr = btrfs_item_ptr_offset(path->nodes[0], path->slots[0]);
+ read_extent_buffer(path->nodes[0], dst_copy, dst_ptr,
+ item_size);
+ ret = memcmp(dst_copy, src_copy, item_size);
+
+ kfree(dst_copy);
+ kfree(src_copy);
+ /*
+ * they have the same contents, just return, this saves
+ * us from cowing blocks in the destination tree and doing
+ * extra writes that may not have been done by a previous
+ * sync
+ */
+ if (ret == 0) {
+ btrfs_release_path(root, path);
+ return 0;
+ }
+
+ }
+insert:
+ btrfs_release_path(root, path);
+ /* try to insert the key into the destination tree */
+ ret = btrfs_insert_empty_item(trans, root, path,
+ key, item_size);
+
+ /* make sure any existing item is the correct size */
+ if (ret == -EEXIST) {
+ u32 found_size;
+ found_size = btrfs_item_size_nr(path->nodes[0],
+ path->slots[0]);
+ if (found_size > item_size) {
+ btrfs_truncate_item(trans, root, path, item_size, 1);
+ } else if (found_size < item_size) {
+ ret = btrfs_del_item(trans, root,
+ path);
+ BUG_ON(ret);
+
+ btrfs_release_path(root, path);
+ ret = btrfs_insert_empty_item(trans,
+ root, path, key, item_size);
+ BUG_ON(ret);
+ }
+ } else if (ret) {
+ BUG();
+ }
+ dst_ptr = btrfs_item_ptr_offset(path->nodes[0],
+ path->slots[0]);
+
+ /* don't overwrite an existing inode if the generation number
+ * was logged as zero. This is done when the tree logging code
+ * is just logging an inode to make sure it exists after recovery.
+ *
+ * Also, don't overwrite i_size on directories during replay.
+ * log replay inserts and removes directory items based on the
+ * state of the tree found in the subvolume, and i_size is modified
+ * as it goes
+ */
+ if (key->type == BTRFS_INODE_ITEM_KEY && ret == -EEXIST) {
+ struct btrfs_inode_item *src_item;
+ struct btrfs_inode_item *dst_item;
+
+ src_item = (struct btrfs_inode_item *)src_ptr;
+ dst_item = (struct btrfs_inode_item *)dst_ptr;
+
+ if (btrfs_inode_generation(eb, src_item) == 0)
+ goto no_copy;
+
+ if (overwrite_root &&
+ S_ISDIR(btrfs_inode_mode(eb, src_item)) &&
+ S_ISDIR(btrfs_inode_mode(path->nodes[0], dst_item))) {
+ save_old_i_size = 1;
+ saved_i_size = btrfs_inode_size(path->nodes[0],
+ dst_item);
+ }
+ }
+
+ copy_extent_buffer(path->nodes[0], eb, dst_ptr,
+ src_ptr, item_size);
+
+ if (save_old_i_size) {
+ struct btrfs_inode_item *dst_item;
+ dst_item = (struct btrfs_inode_item *)dst_ptr;
+ btrfs_set_inode_size(path->nodes[0], dst_item, saved_i_size);
+ }
+
+ /* make sure the generation is filled in */
+ if (key->type == BTRFS_INODE_ITEM_KEY) {
+ struct btrfs_inode_item *dst_item;
+ dst_item = (struct btrfs_inode_item *)dst_ptr;
+ if (btrfs_inode_generation(path->nodes[0], dst_item) == 0) {
+ btrfs_set_inode_generation(path->nodes[0], dst_item,
+ trans->transid);
+ }
+ }
+no_copy:
+ btrfs_mark_buffer_dirty(path->nodes[0]);
+ btrfs_release_path(root, path);
+ return 0;
+}
+
+/*
+ * simple helper to read an inode off the disk from a given root
+ * This can only be called for subvolume roots and not for the log
+ */
+static noinline struct inode *read_one_inode(struct btrfs_root *root,
+ u64 objectid)
+{
+ 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)) {
+ iput(inode);
+ inode = NULL;
+ }
+ return inode;
+}
+
+/* replays a single extent in 'eb' at 'slot' with 'key' into the
+ * subvolume 'root'. path is released on entry and should be released
+ * on exit.
+ *
+ * extents in the log tree have not been allocated out of the extent
+ * tree yet. So, this completes the allocation, taking a reference
+ * as required if the extent already exists or creating a new extent
+ * if it isn't in the extent allocation tree yet.
+ *
+ * The extent is inserted into the file, dropping any existing extents
+ * from the file that overlap the new one.
+ */
+static noinline int replay_one_extent(struct btrfs_trans_handle *trans,
+ struct btrfs_root *root,
+ struct btrfs_path *path,
+ struct extent_buffer *eb, int slot,
+ struct btrfs_key *key)
+{
+ int found_type;
+ u64 mask = root->sectorsize - 1;
+ u64 extent_end;
+ u64 alloc_hint;
+ u64 start = key->offset;
+ struct btrfs_file_extent_item *item;
+ struct inode *inode = NULL;
+ unsigned long size;
+ int ret = 0;
+
+ item = btrfs_item_ptr(eb, slot, struct btrfs_file_extent_item);
+ found_type = btrfs_file_extent_type(eb, item);
+
+ if (found_type == BTRFS_FILE_EXTENT_REG)
+ extent_end = start + btrfs_file_extent_num_bytes(eb, item);
+ else if (found_type == BTRFS_FILE_EXTENT_INLINE) {
+ size = btrfs_file_extent_inline_len(eb,
+ btrfs_item_nr(eb, slot));
+ extent_end = (start + size + mask) & ~mask;
+ } else {
+ ret = 0;
+ goto out;
+ }
+
+ inode = read_one_inode(root, key->objectid);
+ if (!inode) {
+ ret = -EIO;
+ goto out;
+ }
+
+ /*
+ * first check to see if we already have this extent in the
+ * file. This must be done before the btrfs_drop_extents run
+ * so we don't try to drop this extent.
+ */
+ ret = btrfs_lookup_file_extent(trans, root, path, inode->i_ino,
+ start, 0);
+
+ if (ret == 0 && found_type == BTRFS_FILE_EXTENT_REG) {
+ struct btrfs_file_extent_item cmp1;
+ struct btrfs_file_extent_item cmp2;
+ struct btrfs_file_extent_item *existing;
+ struct extent_buffer *leaf;
+
+ leaf = path->nodes[0];
+ existing = btrfs_item_ptr(leaf, path->slots[0],
+ struct btrfs_file_extent_item);
+
+ read_extent_buffer(eb, &cmp1, (unsigned long)item,
+ sizeof(cmp1));
+ read_extent_buffer(leaf, &cmp2, (unsigned long)existing,
+ sizeof(cmp2));
+
+ /*
+ * we already have a pointer to this exact extent,
+ * we don't have to do anything
+ */
+ if (memcmp(&cmp1, &cmp2, sizeof(cmp1)) == 0) {
+ btrfs_release_path(root, path);
+ goto out;
+ }
+ }
+ btrfs_release_path(root, path);
+
+ /* drop any overlapping extents */
+ ret = btrfs_drop_extents(trans, root, inode,
+ start, extent_end, start, &alloc_hint);
+ BUG_ON(ret);
+
+ BUG_ON(ret);
+ if (found_type == BTRFS_FILE_EXTENT_REG) {
+ struct btrfs_key ins;
+
+ 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;
+
+ /* insert the extent pointer in the file */
+ ret = overwrite_item(trans, root, path, eb, slot, key);
+ BUG_ON(ret);
+
+ /*
+ * is this extent already allocated in the extent
+ * allocation tree? If so, just add a reference
+ */
+ ret = btrfs_lookup_extent(root, path, ins.objectid, ins.offset);
+ btrfs_release_path(root, path);
+ if (ret == 0) {
+ ret = btrfs_inc_extent_ref(trans, root,
+ ins.objectid, ins.offset,
+ root->root_key.objectid,
+ trans->transid, key->objectid, start);
+ } else {
+ /*
+ * insert the extent pointer in the extent
+ * allocation tree
+ */
+ ret = btrfs_alloc_logged_extent(trans, root,
+ root->root_key.objectid,
+ trans->transid, key->objectid,
+ start, &ins);
+ BUG_ON(ret);
+ }
+ } else if (found_type == BTRFS_FILE_EXTENT_INLINE) {
+ /* inline extents are easy, we just overwrite them */
+ ret = overwrite_item(trans, root, path, eb, slot, key);
+ BUG_ON(ret);
+ }
+ /* btrfs_drop_extents changes i_blocks, update it here */
+ inode->i_blocks += (extent_end - start) >> 9;
+ btrfs_update_inode(trans, root, inode);
+out:
+ if (inode)
+ iput(inode);
+ return ret;
+}
+
+/*
+ * when cleaning up conflicts between the directory names in the
+ * subvolume, directory names in the log and directory names in the
+ * inode back references, we may have to unlink inodes from directories.
+ *
+ * This is a helper function to do the unlink of a specific directory
+ * item
+ */
+static noinline int drop_one_dir_item(struct btrfs_trans_handle *trans,
+ struct btrfs_root *root,
+ struct btrfs_path *path,
+ struct inode *dir,
+ struct btrfs_dir_item *di)
+{
+ struct inode *inode;
+ char *name;
+ int name_len;
+ struct extent_buffer *leaf;
+ struct btrfs_key location;
+ int ret;
+
+ leaf = path->nodes[0];
+
+ btrfs_dir_item_key_to_cpu(leaf, di, &location);
+ name_len = btrfs_dir_name_len(leaf, di);
+ name = kmalloc(name_len, GFP_NOFS);
+ read_extent_buffer(leaf, name, (unsigned long)(di + 1), name_len);
+ btrfs_release_path(root, path);
+
+ inode = read_one_inode(root, location.objectid);
+ BUG_ON(!inode);
+
+ btrfs_inc_nlink(inode);
+ ret = btrfs_unlink_inode(trans, root, dir, inode, name, name_len);
+ kfree(name);
+
+ iput(inode);
+ return ret;
+}
+
+/*
+ * helper function to see if a given name and sequence number found
+ * in an inode back reference are already in a directory and correctly
+ * point to this inode
+ */
+static noinline int inode_in_dir(struct btrfs_root *root,
+ struct btrfs_path *path,
+ u64 dirid, u64 objectid, u64 index,
+ const char *name, int name_len)
+{
+ struct btrfs_dir_item *di;
+ struct btrfs_key location;
+ int match = 0;
+
+ di = btrfs_lookup_dir_index_item(NULL, root, path, dirid,
+ index, name, name_len, 0);
+ if (di && !IS_ERR(di)) {
+ btrfs_dir_item_key_to_cpu(path->nodes[0], di, &location);
+ if (location.objectid != objectid)
+ goto out;
+ } else
+ goto out;
+ btrfs_release_path(root, path);
+
+ di = btrfs_lookup_dir_item(NULL, root, path, dirid, name, name_len, 0);
+ if (di && !IS_ERR(di)) {
+ btrfs_dir_item_key_to_cpu(path->nodes[0], di, &location);
+ if (location.objectid != objectid)
+ goto out;
+ } else
+ goto out;
+ match = 1;
+out:
+ btrfs_release_path(root, path);
+ return match;
+}
+
+/*
+ * helper function to check a log tree for a named back reference in
+ * an inode. This is used to decide if a back reference that is
+ * found in the subvolume conflicts with what we find in the log.
+ *
+ * inode backreferences may have multiple refs in a single item,
+ * during replay we process one reference at a time, and we don't
+ * want to delete valid links to a file from the subvolume if that
+ * link is also in the log.
+ */
+static noinline int backref_in_log(struct btrfs_root *log,
+ struct btrfs_key *key,
+ char *name, int namelen)
+{
+ struct btrfs_path *path;
+ struct btrfs_inode_ref *ref;
+ unsigned long ptr;
+ unsigned long ptr_end;
+ unsigned long name_ptr;
+ int found_name_len;
+ int item_size;
+ int ret;
+ int match = 0;
+
+ path = btrfs_alloc_path();
+ ret = btrfs_search_slot(NULL, log, key, path, 0, 0);
+ if (ret != 0)
+ goto out;
+
+ item_size = btrfs_item_size_nr(path->nodes[0], path->slots[0]);
+ ptr = btrfs_item_ptr_offset(path->nodes[0], path->slots[0]);
+ ptr_end = ptr + item_size;
+ while (ptr < ptr_end) {
+ ref = (struct btrfs_inode_ref *)ptr;
+ found_name_len = btrfs_inode_ref_name_len(path->nodes[0], ref);
+ if (found_name_len == namelen) {
+ name_ptr = (unsigned long)(ref + 1);
+ ret = memcmp_extent_buffer(path->nodes[0], name,
+ name_ptr, namelen);
+ if (ret == 0) {
+ match = 1;
+ goto out;
+ }
+ }
+ ptr = (unsigned long)(ref + 1) + found_name_len;
+ }
+out:
+ btrfs_free_path(path);
+ return match;
+}
+
+
+/*
+ * replay one inode back reference item found in the log tree.
+ * eb, slot and key refer to the buffer and key found in the log tree.
+ * root is the destination we are replaying into, and path is for temp
+ * use by this function. (it should be released on return).
+ */
+static noinline int add_inode_ref(struct btrfs_trans_handle *trans,
+ struct btrfs_root *root,
+ struct btrfs_root *log,
+ struct btrfs_path *path,
+ struct extent_buffer *eb, int slot,
+ struct btrfs_key *key)
+{
+ struct inode *dir;
+ int ret;
+ struct btrfs_key location;
+ struct btrfs_inode_ref *ref;
+ struct btrfs_dir_item *di;
+ struct inode *inode;
+ char *name;
+ int namelen;
+ unsigned long ref_ptr;
+ unsigned long ref_end;
+
+ location.objectid = key->objectid;
+ location.type = BTRFS_INODE_ITEM_KEY;
+ location.offset = 0;
+
+ /*
+ * it is possible that we didn't log all the parent directories
+ * for a given inode. If we don't find the dir, just don't
+ * copy the back ref in. The link count fixup code will take
+ * care of the rest
+ */
+ dir = read_one_inode(root, key->offset);
+ if (!dir)
+ return -ENOENT;
+
+ inode = read_one_inode(root, key->objectid);
+ BUG_ON(!dir);
+
+ ref_ptr = btrfs_item_ptr_offset(eb, slot);
+ ref_end = ref_ptr + btrfs_item_size_nr(eb, slot);
+
+again:
+ ref = (struct btrfs_inode_ref *)ref_ptr;
+
+ namelen = btrfs_inode_ref_name_len(eb, ref);
+ name = kmalloc(namelen, GFP_NOFS);
+ BUG_ON(!name);
+
+ read_extent_buffer(eb, name, (unsigned long)(ref + 1), namelen);
+
+ /* if we already have a perfect match, we're done */
+ if (inode_in_dir(root, path, dir->i_ino, inode->i_ino,
+ btrfs_inode_ref_index(eb, ref),
+ name, namelen)) {
+ goto out;
+ }
+
+ /*
+ * look for a conflicting back reference in the metadata.
+ * if we find one we have to unlink that name of the file
+ * before we add our new link. Later on, we overwrite any
+ * existing back reference, and we don't want to create
+ * dangling pointers in the directory.
+ */
+conflict_again:
+ ret = btrfs_search_slot(NULL, root, key, path, 0, 0);
+ if (ret == 0) {
+ char *victim_name;
+ int victim_name_len;
+ struct btrfs_inode_ref *victim_ref;
+ unsigned long ptr;
+ unsigned long ptr_end;
+ struct extent_buffer *leaf = path->nodes[0];
+
+ /* are we trying to overwrite a back ref for the root directory
+ * if so, just jump out, we're done
+ */
+ if (key->objectid == key->offset)
+ goto out_nowrite;
+
+ /* check all the names in this back reference to see
+ * if they are in the log. if so, we allow them to stay
+ * otherwise they must be unlinked as a conflict
+ */
+ ptr = btrfs_item_ptr_offset(leaf, path->slots[0]);
+ ptr_end = ptr + btrfs_item_size_nr(leaf, path->slots[0]);
+ while(ptr < ptr_end) {
+ victim_ref = (struct btrfs_inode_ref *)ptr;
+ victim_name_len = btrfs_inode_ref_name_len(leaf,
+ victim_ref);
+ victim_name = kmalloc(victim_name_len, GFP_NOFS);
+ BUG_ON(!victim_name);
+
+ read_extent_buffer(leaf, victim_name,
+ (unsigned long)(victim_ref + 1),
+ victim_name_len);
+
+ if (!backref_in_log(log, key, victim_name,
+ victim_name_len)) {
+ btrfs_inc_nlink(inode);
+ btrfs_release_path(root, path);
+ ret = btrfs_unlink_inode(trans, root, dir,
+ inode, victim_name,
+ victim_name_len);
+ kfree(victim_name);
+ btrfs_release_path(root, path);
+ goto conflict_again;
+ }
+ kfree(victim_name);
+ ptr = (unsigned long)(victim_ref + 1) + victim_name_len;
+ }
+ BUG_ON(ret);
+ }
+ btrfs_release_path(root, path);
+
+ /* look for a conflicting sequence number */
+ di = btrfs_lookup_dir_index_item(trans, root, path, dir->i_ino,
+ btrfs_inode_ref_index(eb, ref),
+ name, namelen, 0);
+ if (di && !IS_ERR(di)) {
+ ret = drop_one_dir_item(trans, root, path, dir, di);
+ BUG_ON(ret);
+ }
+ btrfs_release_path(root, path);
+
+
+ /* look for a conflicting name */
+ di = btrfs_lookup_dir_item(trans, root, path, dir->i_ino,
+ name, namelen, 0);
+ if (di && !IS_ERR(di)) {
+ ret = drop_one_dir_item(trans, root, path, dir, di);
+ BUG_ON(ret);
+ }
+ btrfs_release_path(root, path);
+
+ /* insert our name */
+ ret = btrfs_add_link(trans, dir, inode, name, namelen, 0,
+ btrfs_inode_ref_index(eb, ref));
+ BUG_ON(ret);
+
+ btrfs_update_inode(trans, root, inode);
+
+out:
+ ref_ptr = (unsigned long)(ref + 1) + namelen;
+ kfree(name);
+ if (ref_ptr < ref_end)
+ goto again;
+
+ /* finally write the back reference in the inode */
+ ret = overwrite_item(trans, root, path, eb, slot, key);
+ BUG_ON(ret);
+
+out_nowrite:
+ btrfs_release_path(root, path);
+ iput(dir);
+ iput(inode);
+ return 0;
+}
+
+/*
+ * replay one csum item from the log tree into the subvolume 'root'
+ * eb, slot and key all refer to the log tree
+ * path is for temp use by this function and should be released on return
+ *
+ * This copies the checksums out of the log tree and inserts them into
+ * the subvolume. Any existing checksums for this range in the file
+ * are overwritten, and new items are added where required.
+ *
+ * We keep this simple by reusing the btrfs_ordered_sum code from
+ * the data=ordered mode. This basically means making a copy
+ * of all the checksums in ram, which we have to do anyway for kmap
+ * rules.
+ *
+ * The copy is then sent down to btrfs_csum_file_blocks, which
+ * does all the hard work of finding existing items in the file
+ * or adding new ones.
+ */
+static noinline int replay_one_csum(struct btrfs_trans_handle *trans,
+ struct btrfs_root *root,
+ struct btrfs_path *path,
+ struct extent_buffer *eb, int slot,
+ struct btrfs_key *key)
+{
+ int ret;
+ u32 item_size = btrfs_item_size_nr(eb, slot);
+ u64 cur_offset;
+ unsigned long file_bytes;
+ struct btrfs_ordered_sum *sums;
+ struct btrfs_sector_sum *sector_sum;
+ struct inode *inode;
+ unsigned long ptr;
+
+ file_bytes = (item_size / BTRFS_CRC32_SIZE) * root->sectorsize;
+ inode = read_one_inode(root, key->objectid);
+ if (!inode) {
+ return -EIO;
+ }
+
+ sums = kzalloc(btrfs_ordered_sum_size(root, file_bytes), GFP_NOFS);
+ if (!sums) {
+ iput(inode);
+ return -ENOMEM;
+ }
+
+ INIT_LIST_HEAD(&sums->list);
+ sums->len = file_bytes;
+ sums->file_offset = key->offset;
+
+ /*
+ * copy all the sums into the ordered sum struct
+ */
+ sector_sum = sums->sums;
+ cur_offset = key->offset;
+ ptr = btrfs_item_ptr_offset(eb, slot);
+ while(item_size > 0) {
+ sector_sum->offset = cur_offset;
+ read_extent_buffer(eb, §or_sum->sum, ptr, BTRFS_CRC32_SIZE);
+ sector_sum++;
+ item_size -= BTRFS_CRC32_SIZE;
+ ptr += BTRFS_CRC32_SIZE;
+ cur_offset += root->sectorsize;
+ }
+
+ /* let btrfs_csum_file_blocks add them into the file */
+ ret = btrfs_csum_file_blocks(trans, root, inode, sums);
+ BUG_ON(ret);
+ kfree(sums);
+ iput(inode);
+
+ return 0;
+}
+/*
+ * There are a few corners where the link count of the file can't
+ * be properly maintained during replay. So, instead of adding
+ * lots of complexity to the log code, we just scan the backrefs
+ * for any file that has been through replay.
+ *
+ * The scan will update the link count on the inode to reflect the
+ * number of back refs found. If it goes down to zero, the iput
+ * will free the inode.
+ */
+static noinline int fixup_inode_link_count(struct btrfs_trans_handle *trans,
+ struct btrfs_root *root,
+ struct inode *inode)
+{
+ struct btrfs_path *path;
+ int ret;
+ struct btrfs_key key;
+ u64 nlink = 0;
+ unsigned long ptr;
+ unsigned long ptr_end;
+ int name_len;
+
+ key.objectid = inode->i_ino;
+ key.type = BTRFS_INODE_REF_KEY;
+ key.offset = (u64)-1;
+
+ path = btrfs_alloc_path();
+
+ while(1) {
+ ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
+ if (ret < 0)
+ break;
+ if (ret > 0) {
+ if (path->slots[0] == 0)
+ break;
+ path->slots[0]--;
+ }
+ btrfs_item_key_to_cpu(path->nodes[0], &key,
+ path->slots[0]);
+ if (key.objectid != inode->i_ino ||
+ key.type != BTRFS_INODE_REF_KEY)
+ break;
+ ptr = btrfs_item_ptr_offset(path->nodes[0], path->slots[0]);
+ ptr_end = ptr + btrfs_item_size_nr(path->nodes[0],
+ path->slots[0]);
+ while(ptr < ptr_end) {
+ struct btrfs_inode_ref *ref;
+
+ ref = (struct btrfs_inode_ref *)ptr;
+ name_len = btrfs_inode_ref_name_len(path->nodes[0],
+ ref);
+ ptr = (unsigned long)(ref + 1) + name_len;
+ nlink++;
+ }
+
+ if (key.offset == 0)
+ break;
+ key.offset--;
+ btrfs_release_path(root, path);
+ }
+ btrfs_free_path(path);
+ if (nlink != inode->i_nlink) {
+ inode->i_nlink = nlink;
+ btrfs_update_inode(trans, root, inode);
+ }
+
+ return 0;
+}
+
+static noinline int fixup_inode_link_counts(struct btrfs_trans_handle *trans,
+ struct btrfs_root *root,
+ struct btrfs_path *path)
+{
+ int ret;
+ struct btrfs_key key;
+ struct inode *inode;
+
+ key.objectid = BTRFS_TREE_LOG_FIXUP_OBJECTID;
+ key.type = BTRFS_ORPHAN_ITEM_KEY;
+ key.offset = (u64)-1;
+ while(1) {
+ ret = btrfs_search_slot(trans, root, &key, path, -1, 1);
+ if (ret < 0)
+ break;
+
+ if (ret == 1) {
+ if (path->slots[0] == 0)
+ break;
+ path->slots[0]--;
+ }
+
+ btrfs_item_key_to_cpu(path->nodes[0], &key, path->slots[0]);
+ if (key.objectid != BTRFS_TREE_LOG_FIXUP_OBJECTID ||
+ key.type != BTRFS_ORPHAN_ITEM_KEY)
+ break;
+
+ ret = btrfs_del_item(trans, root, path);
+ BUG_ON(ret);
+
+ btrfs_release_path(root, path);
+ inode = read_one_inode(root, key.offset);
+ BUG_ON(!inode);
+
+ ret = fixup_inode_link_count(trans, root, inode);
+ BUG_ON(ret);
+
+ iput(inode);
+
+ if (key.offset == 0)
+ break;
+ key.offset--;
+ }
+ btrfs_release_path(root, path);
+ return 0;
+}
+
+
+/*
+ * record a given inode in the fixup dir so we can check its link
+ * count when replay is done. The link count is incremented here
+ * so the inode won't go away until we check it
+ */
+static noinline int link_to_fixup_dir(struct btrfs_trans_handle *trans,
+ struct btrfs_root *root,
+ struct btrfs_path *path,
+ u64 objectid)
+{
+ struct btrfs_key key;
+ int ret = 0;
+ struct inode *inode;
+
+ inode = read_one_inode(root, objectid);
+ BUG_ON(!inode);
+
+ key.objectid = BTRFS_TREE_LOG_FIXUP_OBJECTID;
+ btrfs_set_key_type(&key, BTRFS_ORPHAN_ITEM_KEY);
+ key.offset = objectid;
+
+ ret = btrfs_insert_empty_item(trans, root, path, &key, 0);
+
+ btrfs_release_path(root, path);
+ if (ret == 0) {
+ btrfs_inc_nlink(inode);
+ btrfs_update_inode(trans, root, inode);
+ } else if (ret == -EEXIST) {
+ ret = 0;
+ } else {
+ BUG();
+ }
+ iput(inode);
+
+ return ret;
+}
+
+/*
+ * when replaying the log for a directory, we only insert names
+ * for inodes that actually exist. This means an fsync on a directory
+ * does not implicitly fsync all the new files in it
+ */
+static noinline int insert_one_name(struct btrfs_trans_handle *trans,
+ struct btrfs_root *root,
+ struct btrfs_path *path,
+ u64 dirid, u64 index,
+ char *name, int name_len, u8 type,
+ struct btrfs_key *location)
+{
+ struct inode *inode;
+ struct inode *dir;
+ int ret;
+
+ inode = read_one_inode(root, location->objectid);
+ if (!inode)
+ return -ENOENT;
+
+ dir = read_one_inode(root, dirid);
+ if (!dir) {
+ iput(inode);
+ return -EIO;
+ }
+ ret = btrfs_add_link(trans, dir, inode, name, name_len, 1, index);
+
+ /* FIXME, put inode into FIXUP list */
+
+ iput(inode);
+ iput(dir);
+ return ret;
+}
+
+/*
+ * take a single entry in a log directory item and replay it into
+ * the subvolume.
+ *
+ * if a conflicting item exists in the subdirectory already,
+ * the inode it points to is unlinked and put into the link count
+ * fix up tree.
+ *
+ * If a name from the log points to a file or directory that does
+ * not exist in the FS, it is skipped. fsyncs on directories
+ * do not force down inodes inside that directory, just changes to the
+ * names or unlinks in a directory.
+ */
+static noinline int replay_one_name(struct btrfs_trans_handle *trans,
+ struct btrfs_root *root,
+ struct btrfs_path *path,
+ struct extent_buffer *eb,
+ struct btrfs_dir_item *di,
+ struct btrfs_key *key)
+{
+ char *name;
+ int name_len;
+ struct btrfs_dir_item *dst_di;
+ struct btrfs_key found_key;
+ struct btrfs_key log_key;
+ struct inode *dir;
+ struct inode *inode;
+ u8 log_type;
+ int ret;
+
+ dir = read_one_inode(root, key->objectid);
+ BUG_ON(!dir);
+
+ name_len = btrfs_dir_name_len(eb, di);
+ name = kmalloc(name_len, GFP_NOFS);
+ log_type = btrfs_dir_type(eb, di);
+ read_extent_buffer(eb, name, (unsigned long)(di + 1),
+ name_len);
+
+ btrfs_dir_item_key_to_cpu(eb, di, &log_key);
+ if (key->type == BTRFS_DIR_ITEM_KEY) {
+ dst_di = btrfs_lookup_dir_item(trans, root, path, key->objectid,
+ name, name_len, 1);
+ }
+ else if (key->type == BTRFS_DIR_INDEX_KEY) {
+ dst_di = btrfs_lookup_dir_index_item(trans, root, path,
+ key->objectid,
+ key->offset, name,
+ name_len, 1);
+ } else {
+ BUG();
+ }
+ if (!dst_di || IS_ERR(dst_di)) {
+ /* we need a sequence number to insert, so we only
+ * do inserts for the BTRFS_DIR_INDEX_KEY types
+ */
+ if (key->type != BTRFS_DIR_INDEX_KEY)
+ goto out;
+ goto insert;
+ }
+
+ btrfs_dir_item_key_to_cpu(path->nodes[0], dst_di, &found_key);
+ /* the existing item matches the logged item */
+ if (found_key.objectid == log_key.objectid &&
+ found_key.type == log_key.type &&
+ found_key.offset == log_key.offset &&
+ btrfs_dir_type(path->nodes[0], dst_di) == log_type) {
+ goto out;
+ }
+
+ /*
+ * don't drop the conflicting directory entry if the inode
+ * for the new entry doesn't exist
+ */
+ inode = read_one_inode(root, log_key.objectid);
+ if (!inode)
+ goto out;
+
+ iput(inode);
+ ret = drop_one_dir_item(trans, root, path, dir, dst_di);
+ BUG_ON(ret);
+
+ if (key->type == BTRFS_DIR_INDEX_KEY)
+ goto insert;
+out:
+ btrfs_release_path(root, path);
+ kfree(name);
+ iput(dir);
+ return 0;
+
+insert:
+ btrfs_release_path(root, path);
+ ret = insert_one_name(trans, root, path, key->objectid, key->offset,
+ name, name_len, log_type, &log_key);
+
+ if (ret && ret != -ENOENT)
+ BUG();
+ goto out;
+}
+
+/*
+ * find all the names in a directory item and reconcile them into
+ * the subvolume. Only BTRFS_DIR_ITEM_KEY types will have more than
+ * one name in a directory item, but the same code gets used for
+ * both directory index types
+ */
+static noinline int replay_one_dir_item(struct btrfs_trans_handle *trans,
+ struct btrfs_root *root,
+ struct btrfs_path *path,
+ struct extent_buffer *eb, int slot,
+ struct btrfs_key *key)
+{
+ int ret;
+ u32 item_size = btrfs_item_size_nr(eb, slot);
+ struct btrfs_dir_item *di;
+ int name_len;
+ unsigned long ptr;
+ unsigned long ptr_end;
+
+ ptr = btrfs_item_ptr_offset(eb, slot);
+ ptr_end = ptr + item_size;
+ while(ptr < ptr_end) {
+ di = (struct btrfs_dir_item *)ptr;
+ name_len = btrfs_dir_name_len(eb, di);
+ ret = replay_one_name(trans, root, path, eb, di, key);
+ BUG_ON(ret);
+ ptr = (unsigned long)(di + 1);
+ ptr += name_len;
+ }
+ return 0;
+}
+
+/*
+ * directory replay has two parts. There are the standard directory
+ * items in the log copied from the subvolume, and range items
+ * created in the log while the subvolume was logged.
+ *
+ * The range items tell us which parts of the key space the log
+ * is authoritative for. During replay, if a key in the subvolume
+ * directory is in a logged range item, but not actually in the log
+ * that means it was deleted from the directory before the fsync
+ * and should be removed.
+ */
+static noinline int find_dir_range(struct btrfs_root *root,
+ struct btrfs_path *path,
+ u64 dirid, int key_type,
+ u64 *start_ret, u64 *end_ret)
+{
+ struct btrfs_key key;
+ u64 found_end;
+ struct btrfs_dir_log_item *item;
+ int ret;
+ int nritems;
+
+ if (*start_ret == (u64)-1)
+ return 1;
+
+ key.objectid = dirid;
+ key.type = key_type;
+ key.offset = *start_ret;
+
+ ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
+ if (ret < 0)
+ goto out;
+ if (ret > 0) {
+ if (path->slots[0] == 0)
+ goto out;
+ path->slots[0]--;
+ }
+ if (ret != 0)
+ btrfs_item_key_to_cpu(path->nodes[0], &key, path->slots[0]);
+
+ if (key.type != key_type || key.objectid != dirid) {
+ ret = 1;
+ goto next;
+ }
+ item = btrfs_item_ptr(path->nodes[0], path->slots[0],
+ struct btrfs_dir_log_item);
+ found_end = btrfs_dir_log_end(path->nodes[0], item);
+
+ if (*start_ret >= key.offset && *start_ret <= found_end) {
+ ret = 0;
+ *start_ret = key.offset;
+ *end_ret = found_end;
+ goto out;
+ }
+ ret = 1;
+next:
+ /* check the next slot in the tree to see if it is a valid item */
+ nritems = btrfs_header_nritems(path->nodes[0]);
+ if (path->slots[0] >= nritems) {
+ ret = btrfs_next_leaf(root, path);
+ if (ret)
+ goto out;
+ } else {
+ path->slots[0]++;
+ }
+
+ btrfs_item_key_to_cpu(path->nodes[0], &key, path->slots[0]);
+
+ if (key.type != key_type || key.objectid != dirid) {
+ ret = 1;
+ goto out;
+ }
+ item = btrfs_item_ptr(path->nodes[0], path->slots[0],
+ struct btrfs_dir_log_item);
+ found_end = btrfs_dir_log_end(path->nodes[0], item);
+ *start_ret = key.offset;
+ *end_ret = found_end;
+ ret = 0;
+out:
+ btrfs_release_path(root, path);
+ return ret;
+}
+
+/*
+ * this looks for a given directory item in the log. If the directory
+ * item is not in the log, the item is removed and the inode it points
+ * to is unlinked
+ */
+static noinline int check_item_in_log(struct btrfs_trans_handle *trans,
+ struct btrfs_root *root,
+ struct btrfs_root *log,
+ struct btrfs_path *path,
+ struct btrfs_path *log_path,
+ struct inode *dir,
+ struct btrfs_key *dir_key)
+{
+ int ret;
+ struct extent_buffer *eb;
+ int slot;
+ u32 item_size;
+ struct btrfs_dir_item *di;
+ struct btrfs_dir_item *log_di;
+ int name_len;
+ unsigned long ptr;
+ unsigned long ptr_end;
+ char *name;
+ struct inode *inode;
+ struct btrfs_key location;
+
+again:
+ eb = path->nodes[0];
+ slot = path->slots[0];
+ item_size = btrfs_item_size_nr(eb, slot);
+ ptr = btrfs_item_ptr_offset(eb, slot);
+ ptr_end = ptr + item_size;
+ while(ptr < ptr_end) {
+ di = (struct btrfs_dir_item *)ptr;
+ name_len = btrfs_dir_name_len(eb, di);
+ name = kmalloc(name_len, GFP_NOFS);
+ if (!name) {
+ ret = -ENOMEM;
+ goto out;
+ }
+ read_extent_buffer(eb, name, (unsigned long)(di + 1),
+ name_len);
+ log_di = NULL;
+ if (dir_key->type == BTRFS_DIR_ITEM_KEY) {
+ log_di = btrfs_lookup_dir_item(trans, log, log_path,
+ dir_key->objectid,
+ name, name_len, 0);
+ } else if (dir_key->type == BTRFS_DIR_INDEX_KEY) {
+ log_di = btrfs_lookup_dir_index_item(trans, log,
+ log_path,
+ dir_key->objectid,
+ dir_key->offset,
+ name, name_len, 0);
+ }
+ if (!log_di || IS_ERR(log_di)) {
+ btrfs_dir_item_key_to_cpu(eb, di, &location);
+ btrfs_release_path(root, path);
+ btrfs_release_path(log, log_path);
+ inode = read_one_inode(root, location.objectid);
+ BUG_ON(!inode);
+
+ ret = link_to_fixup_dir(trans, root,
+ path, location.objectid);
+ BUG_ON(ret);
+ btrfs_inc_nlink(inode);
+ ret = btrfs_unlink_inode(trans, root, dir, inode,
+ name, name_len);
+ BUG_ON(ret);
+ kfree(name);
+ iput(inode);
+
+ /* there might still be more names under this key
+ * check and repeat if required
+ */
+ ret = btrfs_search_slot(NULL, root, dir_key, path,
+ 0, 0);
+ if (ret == 0)
+ goto again;
+ ret = 0;
+ goto out;
+ }
+ btrfs_release_path(log, log_path);
+ kfree(name);
+
+ ptr = (unsigned long)(di + 1);
+ ptr += name_len;
+ }
+ ret = 0;
+out:
+ btrfs_release_path(root, path);
+ btrfs_release_path(log, log_path);
+ return ret;
+}
+
+/*
+ * deletion replay happens before we copy any new directory items
+ * out of the log or out of backreferences from inodes. It
+ * scans the log to find ranges of keys that log is authoritative for,
+ * and then scans the directory to find items in those ranges that are
+ * not present in the log.
+ *
+ * Anything we don't find in the log is unlinked and removed from the
+ * directory.
+ */
+static noinline int replay_dir_deletes(struct btrfs_trans_handle *trans,
+ struct btrfs_root *root,
+ struct btrfs_root *log,
+ struct btrfs_path *path,
+ u64 dirid)
+{
+ u64 range_start;
+ u64 range_end;
+ int key_type = BTRFS_DIR_LOG_ITEM_KEY;
+ int ret = 0;
+ struct btrfs_key dir_key;
+ struct btrfs_key found_key;
+ struct btrfs_path *log_path;
+ struct inode *dir;
+
+ dir_key.objectid = dirid;
+ dir_key.type = BTRFS_DIR_ITEM_KEY;
+ log_path = btrfs_alloc_path();
+ if (!log_path)
+ return -ENOMEM;
+
+ dir = read_one_inode(root, dirid);
+ /* it isn't an error if the inode isn't there, that can happen
+ * because we replay the deletes before we copy in the inode item
+ * from the log
+ */
+ if (!dir) {
+ btrfs_free_path(log_path);
+ return 0;
+ }
+again:
+ range_start = 0;
+ range_end = 0;
+ while(1) {
+ ret = find_dir_range(log, path, dirid, key_type,
+ &range_start, &range_end);
+ if (ret != 0)
+ break;
+
+ dir_key.offset = range_start;
+ while(1) {
+ int nritems;
+ ret = btrfs_search_slot(NULL, root, &dir_key, path,
+ 0, 0);
+ if (ret < 0)
+ goto out;
+
+ nritems = btrfs_header_nritems(path->nodes[0]);
+ if (path->slots[0] >= nritems) {
+ ret = btrfs_next_leaf(root, path);
+ if (ret)
+ break;
+ }
+ btrfs_item_key_to_cpu(path->nodes[0], &found_key,
+ path->slots[0]);
+ if (found_key.objectid != dirid ||
+ found_key.type != dir_key.type)
+ goto next_type;
+
+ if (found_key.offset > range_end)
+ break;
+
+ ret = check_item_in_log(trans, root, log, path,
+ log_path, dir, &found_key);
+ BUG_ON(ret);
+ if (found_key.offset == (u64)-1)
+ break;
+ dir_key.offset = found_key.offset + 1;
+ }
+ btrfs_release_path(root, path);
+ if (range_end == (u64)-1)
+ break;
+ range_start = range_end + 1;
+ }
+
+next_type:
+ ret = 0;
+ if (key_type == BTRFS_DIR_LOG_ITEM_KEY) {
+ key_type = BTRFS_DIR_LOG_INDEX_KEY;
+ dir_key.type = BTRFS_DIR_INDEX_KEY;
+ btrfs_release_path(root, path);
+ goto again;
+ }
+out:
+ btrfs_release_path(root, path);
+ btrfs_free_path(log_path);
+ iput(dir);
+ return ret;
+}
+
+/*
+ * the process_func used to replay items from the log tree. This
+ * gets called in two different stages. The first stage just looks
+ * for inodes and makes sure they are all copied into the subvolume.
+ *
+ * The second stage copies all the other item types from the log into
+ * the subvolume. The two stage approach is slower, but gets rid of
+ * lots of complexity around inodes referencing other inodes that exist
+ * only in the log (references come from either directory items or inode
+ * back refs).
+ */
+static int replay_one_buffer(struct btrfs_root *log, struct extent_buffer *eb,
+ struct walk_control *wc, u64 gen)
+{
+ int nritems;
+ struct btrfs_path *path;
+ struct btrfs_root *root = wc->replay_dest;
+ struct btrfs_key key;
+ u32 item_size;
+ int level;
+ int i;
+ int ret;
+
+ btrfs_read_buffer(eb, gen);
+
+ level = btrfs_header_level(eb);
+
+ if (level != 0)
+ return 0;
+
+ path = btrfs_alloc_path();
+ BUG_ON(!path);
+
+ nritems = btrfs_header_nritems(eb);
+ for (i = 0; i < nritems; i++) {
+ btrfs_item_key_to_cpu(eb, &key, i);
+ item_size = btrfs_item_size_nr(eb, i);
+
+ /* inode keys are done during the first stage */
+ if (key.type == BTRFS_INODE_ITEM_KEY &&
+ wc->stage == LOG_WALK_REPLAY_INODES) {
+ struct inode *inode;
+ struct btrfs_inode_item *inode_item;
+ u32 mode;
+
+ inode_item = btrfs_item_ptr(eb, i,
+ struct btrfs_inode_item);
+ mode = btrfs_inode_mode(eb, inode_item);
+ if (S_ISDIR(mode)) {
+ ret = replay_dir_deletes(wc->trans,
+ root, log, path, key.objectid);
+ BUG_ON(ret);
+ }
+ ret = overwrite_item(wc->trans, root, path,
+ eb, i, &key);
+ BUG_ON(ret);
+
+ /* for regular files, truncate away
+ * extents past the new EOF
+ */
+ if (S_ISREG(mode)) {
+ inode = read_one_inode(root,
+ key.objectid);
+ BUG_ON(!inode);
+
+ ret = btrfs_truncate_inode_items(wc->trans,
+ root, inode, inode->i_size,
+ BTRFS_EXTENT_DATA_KEY);
+ BUG_ON(ret);
+ iput(inode);
+ }
+ ret = link_to_fixup_dir(wc->trans, root,
+ path, key.objectid);
+ BUG_ON(ret);
+ }
+ if (wc->stage < LOG_WALK_REPLAY_ALL)
+ continue;
+
+ /* these keys are simply copied */
+ if (key.type == BTRFS_XATTR_ITEM_KEY) {
+ ret = overwrite_item(wc->trans, root, path,
+ eb, i, &key);
+ BUG_ON(ret);
+ } else if (key.type == BTRFS_INODE_REF_KEY) {
+ ret = add_inode_ref(wc->trans, root, log, path,
+ eb, i, &key);
+ BUG_ON(ret && ret != -ENOENT);
+ } else if (key.type == BTRFS_EXTENT_DATA_KEY) {
+ ret = replay_one_extent(wc->trans, root, path,
+ eb, i, &key);
+ BUG_ON(ret);
+ } else if (key.type == BTRFS_CSUM_ITEM_KEY) {
+ ret = replay_one_csum(wc->trans, root, path,
+ eb, i, &key);
+ BUG_ON(ret);
+ } else if (key.type == BTRFS_DIR_ITEM_KEY ||
+ key.type == BTRFS_DIR_INDEX_KEY) {
+ ret = replay_one_dir_item(wc->trans, root, path,
+ eb, i, &key);
+ BUG_ON(ret);
+ }
+ }
+ btrfs_free_path(path);
+ return 0;
+}
+
+static int noinline walk_down_log_tree(struct btrfs_trans_handle *trans,
+ struct btrfs_root *root,
+ struct btrfs_path *path, int *level,
+ struct walk_control *wc)
+{
+ u64 root_owner;
+ u64 root_gen;
+ u64 bytenr;
+ u64 ptr_gen;
+ struct extent_buffer *next;
+ struct extent_buffer *cur;
+ struct extent_buffer *parent;
+ u32 blocksize;
+ int ret = 0;
+
+ WARN_ON(*level < 0);
+ WARN_ON(*level >= BTRFS_MAX_LEVEL);
+
+ while(*level > 0) {
+ WARN_ON(*level < 0);
+ WARN_ON(*level >= BTRFS_MAX_LEVEL);
+ cur = path->nodes[*level];
+
+ if (btrfs_header_level(cur) != *level)
+ WARN_ON(1);
+
+ if (path->slots[*level] >=
+ btrfs_header_nritems(cur))
+ break;
+
+ bytenr = btrfs_node_blockptr(cur, path->slots[*level]);
+ ptr_gen = btrfs_node_ptr_generation(cur, path->slots[*level]);
+ blocksize = btrfs_level_size(root, *level - 1);
+
+ parent = path->nodes[*level];
+ root_owner = btrfs_header_owner(parent);
+ root_gen = btrfs_header_generation(parent);
+
+ next = btrfs_find_create_tree_block(root, bytenr, blocksize);
+
+ wc->process_func(root, next, wc, ptr_gen);
+
+ if (*level == 1) {
+ path->slots[*level]++;
+ if (wc->free) {
+ btrfs_read_buffer(next, ptr_gen);
+
+ btrfs_tree_lock(next);
+ clean_tree_block(trans, root, next);
+ 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_extent(trans, root, bytenr,
+ blocksize, root_owner,
+ root_gen, 0, 0, 1);
+ BUG_ON(ret);
+ }
+ free_extent_buffer(next);
+ continue;
+ }
+ btrfs_read_buffer(next, ptr_gen);
+
+ WARN_ON(*level <= 0);
+ if (path->nodes[*level-1])
+ free_extent_buffer(path->nodes[*level-1]);
+ path->nodes[*level-1] = next;
+ *level = btrfs_header_level(next);
+ path->slots[*level] = 0;
+ cond_resched();
+ }
+ WARN_ON(*level < 0);
+ WARN_ON(*level >= BTRFS_MAX_LEVEL);
+
+ if (path->nodes[*level] == root->node) {
+ parent = path->nodes[*level];
+ } else {
+ parent = path->nodes[*level + 1];
+ }
+ bytenr = path->nodes[*level]->start;
+
+ blocksize = btrfs_level_size(root, *level);
+ root_owner = btrfs_header_owner(parent);
+ root_gen = btrfs_header_generation(parent);
+
+ wc->process_func(root, path->nodes[*level], wc,
+ btrfs_header_generation(path->nodes[*level]));
+
+ if (wc->free) {
+ next = path->nodes[*level];
+ btrfs_tree_lock(next);
+ clean_tree_block(trans, root, next);
+ 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_extent(trans, root, bytenr, blocksize,
+ root_owner, root_gen, 0, 0, 1);
+ BUG_ON(ret);
+ }
+ free_extent_buffer(path->nodes[*level]);
+ path->nodes[*level] = NULL;
+ *level += 1;
+
+ cond_resched();
+ return 0;
+}
+
+static int noinline walk_up_log_tree(struct btrfs_trans_handle *trans,
+ struct btrfs_root *root,
+ struct btrfs_path *path, int *level,
+ struct walk_control *wc)
+{
+ u64 root_owner;
+ u64 root_gen;
+ int i;
+ int slot;
+ int ret;
+
+ for(i = *level; i < BTRFS_MAX_LEVEL - 1 && path->nodes[i]; i++) {
+ slot = path->slots[i];
+ if (slot < btrfs_header_nritems(path->nodes[i]) - 1) {
+ struct extent_buffer *node;
+ node = path->nodes[i];
+ path->slots[i]++;
+ *level = i;
+ WARN_ON(*level == 0);
+ return 0;
+ } else {
+ if (path->nodes[*level] == root->node) {
+ root_owner = root->root_key.objectid;
+ root_gen =
+ btrfs_header_generation(path->nodes[*level]);
+ } else {
+ struct extent_buffer *node;
+ node = path->nodes[*level + 1];
+ root_owner = btrfs_header_owner(node);
+ root_gen = btrfs_header_generation(node);
+ }
+ wc->process_func(root, path->nodes[*level], wc,
+ btrfs_header_generation(path->nodes[*level]));
+ if (wc->free) {
+ struct extent_buffer *next;
+
+ next = path->nodes[*level];
+
+ btrfs_tree_lock(next);
+ clean_tree_block(trans, root, next);
+ 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_extent(trans, root,
+ path->nodes[*level]->start,
+ path->nodes[*level]->len,
+ root_owner, root_gen, 0, 0, 1);
+ BUG_ON(ret);
+ }
+ free_extent_buffer(path->nodes[*level]);
+ path->nodes[*level] = NULL;
+ *level = i + 1;
+ }
+ }
+ return 1;
+}
+
+/*
+ * drop the reference count on the tree rooted at 'snap'. This traverses
+ * the tree freeing any blocks that have a ref count of zero after being
+ * decremented.
+ */
+static int walk_log_tree(struct btrfs_trans_handle *trans,
+ struct btrfs_root *log, struct walk_control *wc)
+{
+ int ret = 0;
+ int wret;
+ int level;
+ struct btrfs_path *path;
+ int i;
+ int orig_level;
+
+ path = btrfs_alloc_path();
+ BUG_ON(!path);
+
+ level = btrfs_header_level(log->node);
+ orig_level = level;
+ path->nodes[level] = log->node;
+ extent_buffer_get(log->node);
+ path->slots[level] = 0;
+
+ while(1) {
+ wret = walk_down_log_tree(trans, log, path, &level, wc);
+ if (wret > 0)
+ break;
+ if (wret < 0)
+ ret = wret;
+
+ wret = walk_up_log_tree(trans, log, path, &level, wc);
+ if (wret > 0)
+ break;
+ if (wret < 0)
+ ret = wret;
+ }
+
+ /* was the root node processed? if not, catch it here */
+ if (path->nodes[orig_level]) {
+ wc->process_func(log, path->nodes[orig_level], wc,
+ btrfs_header_generation(path->nodes[orig_level]));
+ if (wc->free) {
+ struct extent_buffer *next;
+
+ next = path->nodes[orig_level];
+
+ btrfs_tree_lock(next);
+ clean_tree_block(trans, log, next);
+ 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_extent(trans, log,
+ next->start, next->len,
+ log->root_key.objectid,
+ btrfs_header_generation(next),
+ 0, 0, 1);
+ BUG_ON(ret);
+ }
+ }
+
+ for (i = 0; i <= orig_level; i++) {
+ if (path->nodes[i]) {
+ free_extent_buffer(path->nodes[i]);
+ path->nodes[i] = NULL;
+ }
+ }
+ btrfs_free_path(path);
+ if (wc->free)
+ free_extent_buffer(log->node);
+ return ret;
+}
+
+int wait_log_commit(struct btrfs_root *log)
+{
+ DEFINE_WAIT(wait);
+ u64 transid = log->fs_info->tree_log_transid;
+
+ do {
+ prepare_to_wait(&log->fs_info->tree_log_wait, &wait,
+ TASK_UNINTERRUPTIBLE);
+ mutex_unlock(&log->fs_info->tree_log_mutex);
+ if (atomic_read(&log->fs_info->tree_log_commit))
+ schedule();
+ finish_wait(&log->fs_info->tree_log_wait, &wait);
+ mutex_lock(&log->fs_info->tree_log_mutex);
+ } while(transid == log->fs_info->tree_log_transid &&
+ atomic_read(&log->fs_info->tree_log_commit));
+ return 0;
+}
+
+/*
+ * btrfs_sync_log does sends a given tree log down to the disk and
+ * updates the super blocks to record it. When this call is done,
+ * you know that any inodes previously logged are safely on disk
+ */
+int btrfs_sync_log(struct btrfs_trans_handle *trans,
+ struct btrfs_root *root)
+{
+ int ret;
+ unsigned long batch;
+ struct btrfs_root *log = root->log_root;
+ struct walk_control wc = {
+ .write = 1,
+ .process_func = process_one_buffer
+ };
+
+ mutex_lock(&log->fs_info->tree_log_mutex);
+ if (atomic_read(&log->fs_info->tree_log_commit)) {
+ wait_log_commit(log);
+ goto out;
+ }
+ atomic_set(&log->fs_info->tree_log_commit, 1);
+
+ while(1) {
+ mutex_unlock(&log->fs_info->tree_log_mutex);
+ schedule_timeout_uninterruptible(1);
+ mutex_lock(&log->fs_info->tree_log_mutex);
+ batch = log->fs_info->tree_log_batch;
+
+ while(atomic_read(&log->fs_info->tree_log_writers)) {
+ DEFINE_WAIT(wait);
+ prepare_to_wait(&log->fs_info->tree_log_wait, &wait,
+ TASK_UNINTERRUPTIBLE);
+ batch = log->fs_info->tree_log_batch;
+ mutex_unlock(&log->fs_info->tree_log_mutex);
+ if (atomic_read(&log->fs_info->tree_log_writers))
+ schedule();
+ mutex_lock(&log->fs_info->tree_log_mutex);
+ finish_wait(&log->fs_info->tree_log_wait, &wait);
+ }
+ if (batch == log->fs_info->tree_log_batch)
+ break;
+ }
+ ret = walk_log_tree(trans, log, &wc);
+ BUG_ON(ret);
+
+ ret = walk_log_tree(trans, log->fs_info->log_root_tree, &wc);
+ BUG_ON(ret);
+
+ wc.wait = 1;
+
+ ret = walk_log_tree(trans, log, &wc);
+ BUG_ON(ret);
+
+ ret = walk_log_tree(trans, log->fs_info->log_root_tree, &wc);
+ BUG_ON(ret);
+
+ btrfs_set_super_log_root(&root->fs_info->super_for_commit,
+ log->fs_info->log_root_tree->node->start);
+ btrfs_set_super_log_root_level(&root->fs_info->super_for_commit,
+ btrfs_header_level(log->fs_info->log_root_tree->node));
+
+ write_ctree_super(trans, log->fs_info->tree_root);
+ log->fs_info->tree_log_transid++;
+ log->fs_info->tree_log_batch = 0;
+ atomic_set(&log->fs_info->tree_log_commit, 0);
+ smp_mb();
+ if (waitqueue_active(&log->fs_info->tree_log_wait))
+ wake_up(&log->fs_info->tree_log_wait);
+out:
+ mutex_unlock(&log->fs_info->tree_log_mutex);
+ return 0;
+
+}
+
+/*
+ * free all the extents used by the tree log. This should be called
+ * at commit time of the full transaction
+ */
+int btrfs_free_log(struct btrfs_trans_handle *trans, struct btrfs_root *root)
+{
+ int ret;
+ struct btrfs_root *log;
+ struct key;
+ struct walk_control wc = {
+ .free = 1,
+ .process_func = process_one_buffer
+ };
+
+ if (!root->log_root)
+ return 0;
+
+ log = root->log_root;
+ ret = walk_log_tree(trans, log, &wc);
+ BUG_ON(ret);
+
+ log = root->log_root;
+ ret = btrfs_del_root(trans, root->fs_info->log_root_tree,
+ &log->root_key);
+ BUG_ON(ret);
+ root->log_root = NULL;
+ kfree(root->log_root);
+ return 0;
+}
+
+/*
+ * helper function to update the item for a given subvolumes log root
+ * in the tree of log roots
+ */
+static int update_log_root(struct btrfs_trans_handle *trans,
+ struct btrfs_root *log)
+{
+ u64 bytenr = btrfs_root_bytenr(&log->root_item);
+ int ret;
+
+ if (log->node->start == bytenr)
+ return 0;
+
+ btrfs_set_root_bytenr(&log->root_item, log->node->start);
+ btrfs_set_root_level(&log->root_item, btrfs_header_level(log->node));
+ ret = btrfs_update_root(trans, log->fs_info->log_root_tree,
+ &log->root_key, &log->root_item);
+ BUG_ON(ret);
+ return ret;
+}
+
+/*
+ * If both a file and directory are logged, and unlinks or renames are
+ * mixed in, we have a few interesting corners:
+ *
+ * create file X in dir Y
+ * link file X to X.link in dir Y
+ * fsync file X
+ * unlink file X but leave X.link
+ * fsync dir Y
+ *
+ * After a crash we would expect only X.link to exist. But file X
+ * didn't get fsync'd again so the log has back refs for X and X.link.
+ *
+ * We solve this by removing directory entries and inode backrefs from the
+ * log when a file that was logged in the current transaction is
+ * unlinked. Any later fsync will include the updated log entries, and
+ * we'll be able to reconstruct the proper directory items from backrefs.
+ *
+ * This optimizations allows us to avoid relogging the entire inode
+ * or the entire directory.
+ */
+int btrfs_del_dir_entries_in_log(struct btrfs_trans_handle *trans,
+ struct btrfs_root *root,
+ const char *name, int name_len,
+ struct inode *dir, u64 index)
+{
+ struct btrfs_root *log;
+ struct btrfs_dir_item *di;
+ struct btrfs_path *path;
+ int ret;
+ int bytes_del = 0;
+
+ ret = join_running_log_trans(root);
+ if (ret)
+ return 0;
+
+ mutex_lock(&BTRFS_I(dir)->log_mutex);
+
+ log = root->log_root;
+ path = btrfs_alloc_path();
+ di = btrfs_lookup_dir_item(trans, log, path, dir->i_ino,
+ name, name_len, -1);
+ if (di && !IS_ERR(di)) {
+ ret = btrfs_delete_one_dir_name(trans, log, path, di);
+ bytes_del += name_len;
+ BUG_ON(ret);
+ }
+ btrfs_release_path(log, path);
+ di = btrfs_lookup_dir_index_item(trans, log, path, dir->i_ino,
+ index, name, name_len, -1);
+ if (di && !IS_ERR(di)) {
+ ret = btrfs_delete_one_dir_name(trans, log, path, di);
+ bytes_del += name_len;
+ BUG_ON(ret);
+ }
+
+ /* update the directory size in the log to reflect the names
+ * we have removed
+ */
+ if (bytes_del) {
+ struct btrfs_key key;
+
+ key.objectid = dir->i_ino;
+ key.offset = 0;
+ key.type = BTRFS_INODE_ITEM_KEY;
+ btrfs_release_path(log, path);
+
+ ret = btrfs_search_slot(trans, log, &key, path, 0, 1);
+ if (ret == 0) {
+ struct btrfs_inode_item *item;
+ u64 i_size;
+
+ item = btrfs_item_ptr(path->nodes[0], path->slots[0],
+ struct btrfs_inode_item);
+ i_size = btrfs_inode_size(path->nodes[0], item);
+ if (i_size > bytes_del)
+ i_size -= bytes_del;
+ else
+ i_size = 0;
+ btrfs_set_inode_size(path->nodes[0], item, i_size);
+ btrfs_mark_buffer_dirty(path->nodes[0]);
+ } else
+ ret = 0;
+ btrfs_release_path(log, path);
+ }
+
+ btrfs_free_path(path);
+ mutex_unlock(&BTRFS_I(dir)->log_mutex);
+ end_log_trans(root);
+
+ return 0;
+}
+
+/* see comments for btrfs_del_dir_entries_in_log */
+int btrfs_del_inode_ref_in_log(struct btrfs_trans_handle *trans,
+ struct btrfs_root *root,
+ const char *name, int name_len,
+ struct inode *inode, u64 dirid)
+{
+ struct btrfs_root *log;
+ u64 index;
+ int ret;
+
+ ret = join_running_log_trans(root);
+ if (ret)
+ return 0;
+ log = root->log_root;
+ mutex_lock(&BTRFS_I(inode)->log_mutex);
+
+ ret = btrfs_del_inode_ref(trans, log, name, name_len, inode->i_ino,
+ dirid, &index);
+ mutex_unlock(&BTRFS_I(inode)->log_mutex);
+ end_log_trans(root);
+
+ if (ret == 0 || ret == -ENOENT)
+ return 0;
+ return ret;
+}
+
+/*
+ * creates a range item in the log for 'dirid'. first_offset and
+ * last_offset tell us which parts of the key space the log should
+ * be considered authoritative for.
+ */
+static noinline int insert_dir_log_key(struct btrfs_trans_handle *trans,
+ struct btrfs_root *log,
+ struct btrfs_path *path,
+ int key_type, u64 dirid,
+ u64 first_offset, u64 last_offset)
+{
+ int ret;
+ struct btrfs_key key;
+ struct btrfs_dir_log_item *item;
+
+ key.objectid = dirid;
+ key.offset = first_offset;
+ if (key_type == BTRFS_DIR_ITEM_KEY)
+ key.type = BTRFS_DIR_LOG_ITEM_KEY;
+ else
+ key.type = BTRFS_DIR_LOG_INDEX_KEY;
+ ret = btrfs_insert_empty_item(trans, log, path, &key, sizeof(*item));
+ BUG_ON(ret);
+
+ item = btrfs_item_ptr(path->nodes[0], path->slots[0],
+ struct btrfs_dir_log_item);
+ btrfs_set_dir_log_end(path->nodes[0], item, last_offset);
+ btrfs_mark_buffer_dirty(path->nodes[0]);
+ btrfs_release_path(log, path);
+ return 0;
+}
+
+/*
+ * log all the items included in the current transaction for a given
+ * directory. This also creates the range items in the log tree required
+ * to replay anything deleted before the fsync
+ */
+static noinline int log_dir_items(struct btrfs_trans_handle *trans,
+ struct btrfs_root *root, struct inode *inode,
+ struct btrfs_path *path,
+ struct btrfs_path *dst_path, int key_type,
+ u64 min_offset, u64 *last_offset_ret)
+{
+ struct btrfs_key min_key;
+ struct btrfs_key max_key;
+ struct btrfs_root *log = root->log_root;
+ struct extent_buffer *src;
+ int ret;
+ int i;
+ int nritems;
+ u64 first_offset = min_offset;
+ u64 last_offset = (u64)-1;
+
+ log = root->log_root;
+ max_key.objectid = inode->i_ino;
+ max_key.offset = (u64)-1;
+ max_key.type = key_type;
+
+ min_key.objectid = inode->i_ino;
+ min_key.type = key_type;
+ min_key.offset = min_offset;
+
+ path->keep_locks = 1;
+
+ ret = btrfs_search_forward(root, &min_key, &max_key,
+ path, 0, trans->transid);
+
+ /*
+ * we didn't find anything from this transaction, see if there
+ * is anything at all
+ */
+ if (ret != 0 || min_key.objectid != inode->i_ino ||
+ min_key.type != key_type) {
+ min_key.objectid = inode->i_ino;
+ min_key.type = key_type;
+ min_key.offset = (u64)-1;
+ btrfs_release_path(root, path);
+ ret = btrfs_search_slot(NULL, root, &min_key, path, 0, 0);
+ if (ret < 0) {
+ btrfs_release_path(root, path);
+ return ret;
+ }
+ ret = btrfs_previous_item(root, path, inode->i_ino, key_type);
+
+ /* if ret == 0 there are items for this type,
+ * create a range to tell us the last key of this type.
+ * otherwise, there are no items in this directory after
+ * *min_offset, and we create a range to indicate that.
+ */
+ if (ret == 0) {
+ struct btrfs_key tmp;
+ btrfs_item_key_to_cpu(path->nodes[0], &tmp,
+ path->slots[0]);
+ if (key_type == tmp.type) {
+ first_offset = max(min_offset, tmp.offset) + 1;
+ }
+ }
+ goto done;
+ }
+
+ /* go backward to find any previous key */
+ ret = btrfs_previous_item(root, path, inode->i_ino, key_type);
+ if (ret == 0) {
+ struct btrfs_key tmp;
+ btrfs_item_key_to_cpu(path->nodes[0], &tmp, path->slots[0]);
+ if (key_type == tmp.type) {
+ first_offset = tmp.offset;
+ ret = overwrite_item(trans, log, dst_path,
+ path->nodes[0], path->slots[0],
+ &tmp);
+ }
+ }
+ btrfs_release_path(root, path);
+
+ /* find the first key from this transaction again */
+ ret = btrfs_search_slot(NULL, root, &min_key, path, 0, 0);
+ if (ret != 0) {
+ WARN_ON(1);
+ goto done;
+ }
+
+ /*
+ * we have a block from this transaction, log every item in it
+ * from our directory
+ */
+ while(1) {
+ struct btrfs_key tmp;
+ src = path->nodes[0];
+ nritems = btrfs_header_nritems(src);
+ for (i = path->slots[0]; i < nritems; i++) {
+ btrfs_item_key_to_cpu(src, &min_key, i);
+
+ if (min_key.objectid != inode->i_ino ||
+ min_key.type != key_type)
+ goto done;
+ ret = overwrite_item(trans, log, dst_path, src, i,
+ &min_key);
+ BUG_ON(ret);
+ }
+ path->slots[0] = nritems;
+
+ /*
+ * look ahead to the next item and see if it is also
+ * from this directory and from this transaction
+ */
+ ret = btrfs_next_leaf(root, path);
+ if (ret == 1) {
+ last_offset = (u64)-1;
+ goto done;
+ }
+ btrfs_item_key_to_cpu(path->nodes[0], &tmp, path->slots[0]);
+ if (tmp.objectid != inode->i_ino || tmp.type != key_type) {
+ last_offset = (u64)-1;
+ goto done;
+ }
+ if (btrfs_header_generation(path->nodes[0]) != trans->transid) {
+ ret = overwrite_item(trans, log, dst_path,
+ path->nodes[0], path->slots[0],
+ &tmp);
+
+ BUG_ON(ret);
+ last_offset = tmp.offset;
+ goto done;
+ }
+ }
+done:
+ *last_offset_ret = last_offset;
+ btrfs_release_path(root, path);
+ btrfs_release_path(log, dst_path);
+
+ /* insert the log range keys to indicate where the log is valid */
+ ret = insert_dir_log_key(trans, log, path, key_type, inode->i_ino,
+ first_offset, last_offset);
+ BUG_ON(ret);
+ return 0;
+}
+
+/*
+ * logging directories is very similar to logging inodes, We find all the items
+ * from the current transaction and write them to the log.
+ *
+ * The recovery code scans the directory in the subvolume, and if it finds a
+ * key in the range logged that is not present in the log tree, then it means
+ * that dir entry was unlinked during the transaction.
+ *
+ * In order for that scan to work, we must include one key smaller than
+ * the smallest logged by this transaction and one key larger than the largest
+ * key logged by this transaction.
+ */
+static noinline int log_directory_changes(struct btrfs_trans_handle *trans,
+ struct btrfs_root *root, struct inode *inode,
+ struct btrfs_path *path,
+ struct btrfs_path *dst_path)
+{
+ u64 min_key;
+ u64 max_key;
+ int ret;
+ int key_type = BTRFS_DIR_ITEM_KEY;
+
+again:
+ min_key = 0;
+ max_key = 0;
+ while(1) {
+ ret = log_dir_items(trans, root, inode, path,
+ dst_path, key_type, min_key,
+ &max_key);
+ BUG_ON(ret);
+ if (max_key == (u64)-1)
+ break;
+ min_key = max_key + 1;
+ }
+
+ if (key_type == BTRFS_DIR_ITEM_KEY) {
+ key_type = BTRFS_DIR_INDEX_KEY;
+ goto again;
+ }
+ return 0;
+}
+
+/*
+ * a helper function to drop items from the log before we relog an
+ * inode. max_key_type indicates the highest item type to remove.
+ * This cannot be run for file data extents because it does not
+ * free the extents they point to.
+ */
+static int drop_objectid_items(struct btrfs_trans_handle *trans,
+ struct btrfs_root *log,
+ struct btrfs_path *path,
+ u64 objectid, int max_key_type)
+{
+ int ret;
+ struct btrfs_key key;
+ struct btrfs_key found_key;
+
+ key.objectid = objectid;
+ key.type = max_key_type;
+ key.offset = (u64)-1;
+
+ while(1) {
+ ret = btrfs_search_slot(trans, log, &key, path, -1, 1);
+
+ if (ret != 1)
+ break;
+
+ if (path->slots[0] == 0)
+ break;
+
+ path->slots[0]--;
+ btrfs_item_key_to_cpu(path->nodes[0], &found_key,
+ path->slots[0]);
+
+ if (found_key.objectid != objectid)
+ break;
+
+ ret = btrfs_del_item(trans, log, path);
+ BUG_ON(ret);
+ btrfs_release_path(log, path);
+ }
+ btrfs_release_path(log, path);
+ return 0;
+}
+
+/* log a single inode in the tree log.
+ * At least one parent directory for this inode must exist in the tree
+ * or be logged already.
+ *
+ * Any items from this inode changed by the current transaction are copied
+ * to the log tree. An extra reference is taken on any extents in this
+ * file, allowing us to avoid a whole pile of corner cases around logging
+ * blocks that have been removed from the tree.
+ *
+ * See LOG_INODE_ALL and related defines for a description of what inode_only
+ * does.
+ *
+ * This handles both files and directories.
+ */
+static int __btrfs_log_inode(struct btrfs_trans_handle *trans,
+ struct btrfs_root *root, struct inode *inode,
+ int inode_only)
+{
+ struct btrfs_path *path;
+ struct btrfs_path *dst_path;
+ struct btrfs_key min_key;
+ struct btrfs_key max_key;
+ struct btrfs_root *log = root->log_root;
+ unsigned long src_offset;
+ unsigned long dst_offset;
+ struct extent_buffer *src;
+ struct btrfs_file_extent_item *extent;
+ struct btrfs_inode_item *inode_item;
+ u32 size;
+ int ret;
+
+ log = root->log_root;
+
+ path = btrfs_alloc_path();
+ dst_path = btrfs_alloc_path();
+
+ min_key.objectid = inode->i_ino;
+ min_key.type = BTRFS_INODE_ITEM_KEY;
+ min_key.offset = 0;
+
+ max_key.objectid = inode->i_ino;
+ if (inode_only == LOG_INODE_EXISTS || S_ISDIR(inode->i_mode))
+ max_key.type = BTRFS_XATTR_ITEM_KEY;
+ else
+ max_key.type = (u8)-1;
+ max_key.offset = (u64)-1;
+
+ /*
+ * if this inode has already been logged and we're in inode_only
+ * mode, we don't want to delete the things that have already
+ * been written to the log.
+ *
+ * But, if the inode has been through an inode_only log,
+ * the logged_trans field is not set. This allows us to catch
+ * any new names for this inode in the backrefs by logging it
+ * again
+ */
+ if (inode_only == LOG_INODE_EXISTS &&
+ BTRFS_I(inode)->logged_trans == trans->transid) {
+ btrfs_free_path(path);
+ btrfs_free_path(dst_path);
+ goto out;
+ }
+ mutex_lock(&BTRFS_I(inode)->log_mutex);
+
+ /*
+ * a brute force approach to making sure we get the most uptodate
+ * copies of everything.
+ */
+ if (S_ISDIR(inode->i_mode)) {
+ int max_key_type = BTRFS_DIR_LOG_INDEX_KEY;
+
+ if (inode_only == LOG_INODE_EXISTS)
+ max_key_type = BTRFS_XATTR_ITEM_KEY;
+ ret = drop_objectid_items(trans, log, path,
+ inode->i_ino, max_key_type);
+ } else {
+ ret = btrfs_truncate_inode_items(trans, log, inode, 0, 0);
+ }
+ BUG_ON(ret);
+ path->keep_locks = 1;
+
+ while(1) {
+ ret = btrfs_search_forward(root, &min_key, &max_key,
+ path, 0, trans->transid);
+ if (ret != 0)
+ break;
+
+ if (min_key.objectid != inode->i_ino)
+ break;
+ if (min_key.type > max_key.type)
+ break;
+
+ src = path->nodes[0];
+ size = btrfs_item_size_nr(src, path->slots[0]);
+ ret = btrfs_insert_empty_item(trans, log, dst_path, &min_key,
+ size);
+ if (ret)
+ BUG();
+
+ dst_offset = btrfs_item_ptr_offset(dst_path->nodes[0],
+ dst_path->slots[0]);
+
+ src_offset = btrfs_item_ptr_offset(src, path->slots[0]);
+
+ copy_extent_buffer(dst_path->nodes[0], src, dst_offset,
+ src_offset, size);
+
+ if (inode_only == LOG_INODE_EXISTS &&
+ min_key.type == BTRFS_INODE_ITEM_KEY) {
+ inode_item = btrfs_item_ptr(dst_path->nodes[0],
+ dst_path->slots[0],
+ struct btrfs_inode_item);
+ btrfs_set_inode_size(dst_path->nodes[0], inode_item, 0);
+
+ /* set the generation to zero so the recover code
+ * can tell the difference between an logging
+ * just to say 'this inode exists' and a logging
+ * to say 'update this inode with these values'
+ */
+ btrfs_set_inode_generation(dst_path->nodes[0],
+ inode_item, 0);
+ }
+ /* take a reference on file data extents so that truncates
+ * or deletes of this inode don't have to relog the inode
+ * again
+ */
+ if (btrfs_key_type(&min_key) == BTRFS_EXTENT_DATA_KEY) {
+ int found_type;
+ extent = btrfs_item_ptr(src, path->slots[0],
+ struct btrfs_file_extent_item);
+
+ found_type = btrfs_file_extent_type(src, extent);
+ if (found_type == BTRFS_FILE_EXTENT_REG) {
+ u64 ds = btrfs_file_extent_disk_bytenr(src,
+ extent);
+ u64 dl = btrfs_file_extent_disk_num_bytes(src,
+ extent);
+ /* ds == 0 is a hole */
+ if (ds != 0) {
+ ret = btrfs_inc_extent_ref(trans, log,
+ ds, dl,
+ log->root_key.objectid,
+ 0,
+ inode->i_ino,
+ min_key.offset);
+ BUG_ON(ret);
+ }
+ }
+ }
+
+ btrfs_mark_buffer_dirty(dst_path->nodes[0]);
+ btrfs_release_path(root, path);
+ btrfs_release_path(log, dst_path);
+
+ if (min_key.offset < (u64)-1)
+ min_key.offset++;
+ else if (min_key.type < (u8)-1)
+ min_key.type++;
+ else if (min_key.objectid < (u64)-1)
+ min_key.objectid++;
+ else
+ break;
+ }
+ if (inode_only == LOG_INODE_ALL && S_ISDIR(inode->i_mode)) {
+ btrfs_release_path(root, path);
+ btrfs_release_path(log, dst_path);
+ ret = log_directory_changes(trans, root, inode, path, dst_path);
+ BUG_ON(ret);
+ }
+ mutex_unlock(&BTRFS_I(inode)->log_mutex);
+
+ btrfs_free_path(path);
+ btrfs_free_path(dst_path);
+
+ mutex_lock(&root->fs_info->tree_log_mutex);
+ ret = update_log_root(trans, log);
+ BUG_ON(ret);
+ mutex_unlock(&root->fs_info->tree_log_mutex);
+out:
+ return 0;
+}
+
+int btrfs_log_inode(struct btrfs_trans_handle *trans,
+ struct btrfs_root *root, struct inode *inode,
+ int inode_only)
+{
+ int ret;
+
+ start_log_trans(trans, root);
+ ret = __btrfs_log_inode(trans, root, inode, inode_only);
+ end_log_trans(root);
+ return ret;
+}
+
+/*
+ * helper function around btrfs_log_inode to make sure newly created
+ * parent directories also end up in the log. A minimal inode and backref
+ * only logging is done of any parent directories that are older than
+ * the last committed transaction
+ */
+int btrfs_log_dentry(struct btrfs_trans_handle *trans,
+ struct btrfs_root *root, struct dentry *dentry)
+{
+ int inode_only = LOG_INODE_ALL;
+ struct super_block *sb;
+ int ret;
+
+ start_log_trans(trans, root);
+ sb = dentry->d_inode->i_sb;
+ while(1) {
+ ret = __btrfs_log_inode(trans, root, dentry->d_inode,
+ inode_only);
+ BUG_ON(ret);
+ inode_only = LOG_INODE_EXISTS;
+
+ dentry = dentry->d_parent;
+ if (!dentry || !dentry->d_inode || sb != dentry->d_inode->i_sb)
+ break;
+
+ if (BTRFS_I(dentry->d_inode)->generation <=
+ root->fs_info->last_trans_committed)
+ break;
+ }
+ end_log_trans(root);
+ return 0;
+}
+
+/*
+ * it is not safe to log dentry if the chunk root has added new
+ * chunks. This returns 0 if the dentry was logged, and 1 otherwise.
+ * If this returns 1, you must commit the transaction to safely get your
+ * data on disk.
+ */
+int btrfs_log_dentry_safe(struct btrfs_trans_handle *trans,
+ struct btrfs_root *root, struct dentry *dentry)
+{
+ u64 gen;
+ gen = root->fs_info->last_trans_new_blockgroup;
+ if (gen > root->fs_info->last_trans_committed)
+ return 1;
+ else
+ return btrfs_log_dentry(trans, root, dentry);
+}
+
+/*
+ * should be called during mount to recover any replay any log trees
+ * from the FS
+ */
+int btrfs_recover_log_trees(struct btrfs_root *log_root_tree)
+{
+ int ret;
+ struct btrfs_path *path;
+ struct btrfs_trans_handle *trans;
+ struct btrfs_key key;
+ struct btrfs_key found_key;
+ struct btrfs_key tmp_key;
+ struct btrfs_root *log;
+ struct btrfs_fs_info *fs_info = log_root_tree->fs_info;
+ struct walk_control wc = {
+ .process_func = process_one_buffer,
+ .stage = 0,
+ };
+
+ fs_info->log_root_recovering = 1;
+ path = btrfs_alloc_path();
+ BUG_ON(!path);
+
+ trans = btrfs_start_transaction(fs_info->tree_root, 1);
+
+ wc.trans = trans;
+ wc.pin = 1;
+
+ walk_log_tree(trans, log_root_tree, &wc);
+
+again:
+ key.objectid = BTRFS_TREE_LOG_OBJECTID;
+ key.offset = (u64)-1;
+ btrfs_set_key_type(&key, BTRFS_ROOT_ITEM_KEY);
+
+ while(1) {
+ ret = btrfs_search_slot(NULL, log_root_tree, &key, path, 0, 0);
+ if (ret < 0)
+ break;
+ if (ret > 0) {
+ if (path->slots[0] == 0)
+ break;
+ path->slots[0]--;
+ }
+ btrfs_item_key_to_cpu(path->nodes[0], &found_key,
+ path->slots[0]);
+ btrfs_release_path(log_root_tree, path);
+ if (found_key.objectid != BTRFS_TREE_LOG_OBJECTID)
+ break;
+
+ log = btrfs_read_fs_root_no_radix(log_root_tree,
+ &found_key);
+ BUG_ON(!log);
+
+
+ tmp_key.objectid = found_key.offset;
+ tmp_key.type = BTRFS_ROOT_ITEM_KEY;
+ tmp_key.offset = (u64)-1;
+
+ wc.replay_dest = btrfs_read_fs_root_no_name(fs_info, &tmp_key);
+
+ BUG_ON(!wc.replay_dest);
+
+ btrfs_record_root_in_trans(wc.replay_dest);
+ ret = walk_log_tree(trans, log, &wc);
+ BUG_ON(ret);
+
+ if (wc.stage == LOG_WALK_REPLAY_ALL) {
+ ret = fixup_inode_link_counts(trans, wc.replay_dest,
+ path);
+ BUG_ON(ret);
+ }
+
+ key.offset = found_key.offset - 1;
+ free_extent_buffer(log->node);
+ kfree(log);
+
+ if (found_key.offset == 0)
+ break;
+ }
+ btrfs_release_path(log_root_tree, path);
+
+ /* step one is to pin it all, step two is to replay just inodes */
+ if (wc.pin) {
+ wc.pin = 0;
+ wc.process_func = replay_one_buffer;
+ wc.stage = LOG_WALK_REPLAY_INODES;
+ goto again;
+ }
+ /* step three is to replay everything */
+ if (wc.stage < LOG_WALK_REPLAY_ALL) {
+ wc.stage++;
+ goto again;
+ }
+
+ btrfs_free_path(path);
+
+ free_extent_buffer(log_root_tree->node);
+ log_root_tree->log_root = NULL;
+ fs_info->log_root_recovering = 0;
+
+ /* step 4: commit the transaction, which also unpins the blocks */
+ btrfs_commit_transaction(trans, fs_info->tree_root);
+
+ kfree(log_root_tree);
+ return 0;
+}
diff --git a/fs/btrfs/tree-log.h b/fs/btrfs/tree-log.h
new file mode 100644
index 0000000..b9409b3
--- /dev/null
+++ b/fs/btrfs/tree-log.h
@@ -0,0 +1,41 @@
+/*
+ * Copyright (C) 2008 Oracle. All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public
+ * License v2 as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ * General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public
+ * License along with this program; if not, write to the
+ * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
+ * Boston, MA 021110-1307, USA.
+ */
+
+#ifndef __TREE_LOG_
+#define __TREE_LOG_
+
+int btrfs_sync_log(struct btrfs_trans_handle *trans,
+ struct btrfs_root *root);
+int btrfs_free_log(struct btrfs_trans_handle *trans, struct btrfs_root *root);
+int btrfs_log_dentry(struct btrfs_trans_handle *trans,
+ struct btrfs_root *root, struct dentry *dentry);
+int btrfs_recover_log_trees(struct btrfs_root *tree_root);
+int btrfs_log_dentry_safe(struct btrfs_trans_handle *trans,
+ struct btrfs_root *root, struct dentry *dentry);
+int btrfs_log_inode(struct btrfs_trans_handle *trans,
+ struct btrfs_root *root, struct inode *inode,
+ int inode_only);
+int btrfs_del_dir_entries_in_log(struct btrfs_trans_handle *trans,
+ struct btrfs_root *root,
+ const char *name, int name_len,
+ struct inode *dir, u64 index);
+int btrfs_del_inode_ref_in_log(struct btrfs_trans_handle *trans,
+ struct btrfs_root *root,
+ const char *name, int name_len,
+ struct inode *inode, u64 dirid);
+#endif
