#include "disk-io.h"
#include "transaction.h"
#include "print-tree.h"
+#include "locking.h"
static int split_node(struct btrfs_trans_handle *trans, struct btrfs_root
*root, struct btrfs_path *path, int level);
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);
+ struct btrfs_path *path, int data_size, int extend);
static int push_node_left(struct btrfs_trans_handle *trans,
struct btrfs_root *root, struct extent_buffer *dst,
- struct extent_buffer *src);
+ struct extent_buffer *src, int empty);
static int balance_node_right(struct btrfs_trans_handle *trans,
struct btrfs_root *root,
struct extent_buffer *dst_buf,
void btrfs_release_path(struct btrfs_root *root, struct btrfs_path *p)
{
int i;
+ int skip = p->skip_locking;
+ int keep = p->keep_locks;
+
for (i = 0; i < BTRFS_MAX_LEVEL; i++) {
if (!p->nodes[i])
- break;
+ continue;
+ if (p->locks[i]) {
+ btrfs_tree_unlock(p->nodes[i]);
+ p->locks[i] = 0;
+ }
free_extent_buffer(p->nodes[i]);
}
memset(p, 0, sizeof(*p));
+ p->skip_locking = skip;
+ p->keep_locks = keep;
+}
+
+struct extent_buffer *btrfs_root_node(struct btrfs_root *root)
+{
+ struct extent_buffer *eb;
+ spin_lock(&root->node_lock);
+ eb = root->node;
+ extent_buffer_get(eb);
+ spin_unlock(&root->node_lock);
+ return eb;
+}
+
+struct extent_buffer *btrfs_lock_root_node(struct btrfs_root *root)
+{
+ struct extent_buffer *eb;
+
+ while(1) {
+ eb = btrfs_root_node(root);
+ btrfs_tree_lock(eb);
+
+ spin_lock(&root->node_lock);
+ if (eb == root->node) {
+ spin_unlock(&root->node_lock);
+ break;
+ }
+ spin_unlock(&root->node_lock);
+
+ btrfs_tree_unlock(eb);
+ free_extent_buffer(eb);
+ }
+ return eb;
+}
+
+static void add_root_to_dirty_list(struct btrfs_root *root)
+{
+ if (root->track_dirty && list_empty(&root->dirty_list)) {
+ list_add(&root->dirty_list,
+ &root->fs_info->dirty_cowonly_roots);
+ }
+}
+
+int btrfs_copy_root(struct btrfs_trans_handle *trans,
+ struct btrfs_root *root,
+ struct extent_buffer *buf,
+ struct extent_buffer **cow_ret, u64 new_root_objectid)
+{
+ struct extent_buffer *cow;
+ u32 nritems;
+ int ret = 0;
+ int level;
+ struct btrfs_key first_key;
+ struct btrfs_root *new_root;
+
+ new_root = kmalloc(sizeof(*new_root), GFP_NOFS);
+ if (!new_root)
+ return -ENOMEM;
+
+ memcpy(new_root, root, sizeof(*new_root));
+ new_root->root_key.objectid = new_root_objectid;
+
+ WARN_ON(root->ref_cows && trans->transid !=
+ root->fs_info->running_transaction->transid);
+ WARN_ON(root->ref_cows && trans->transid != root->last_trans);
+
+ level = btrfs_header_level(buf);
+ nritems = btrfs_header_nritems(buf);
+ if (nritems) {
+ if (level == 0)
+ btrfs_item_key_to_cpu(buf, &first_key, 0);
+ else
+ btrfs_node_key_to_cpu(buf, &first_key, 0);
+ } else {
+ first_key.objectid = 0;
+ }
+ cow = btrfs_alloc_free_block(trans, new_root, buf->len,
+ new_root_objectid,
+ trans->transid, first_key.objectid,
+ level, buf->start, 0);
+ if (IS_ERR(cow)) {
+ kfree(new_root);
+ return PTR_ERR(cow);
+ }
+
+ copy_extent_buffer(cow, buf, 0, 0, cow->len);
+ btrfs_set_header_bytenr(cow, cow->start);
+ btrfs_set_header_generation(cow, trans->transid);
+ btrfs_set_header_owner(cow, new_root_objectid);
+ btrfs_clear_header_flag(cow, BTRFS_HEADER_FLAG_WRITTEN);
+
+ WARN_ON(btrfs_header_generation(buf) > trans->transid);
+ ret = btrfs_inc_ref(trans, new_root, buf);
+ kfree(new_root);
+
+ if (ret)
+ return ret;
+
+ btrfs_mark_buffer_dirty(cow);
+ *cow_ret = cow;
+ return 0;
}
-static int __btrfs_cow_block(struct btrfs_trans_handle *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,
struct extent_buffer **cow_ret,
u64 search_start, u64 empty_size)
{
+ u64 root_gen;
struct extent_buffer *cow;
+ u32 nritems;
int ret = 0;
int different_trans = 0;
+ int level;
+ int unlock_orig = 0;
+ struct btrfs_key first_key;
+ if (*cow_ret == buf)
+ unlock_orig = 1;
+
+ WARN_ON(!btrfs_tree_locked(buf));
+
+ if (root->ref_cows) {
+ root_gen = trans->transid;
+ } else {
+ root_gen = 0;
+ }
+ WARN_ON(root->ref_cows && trans->transid !=
+ root->fs_info->running_transaction->transid);
WARN_ON(root->ref_cows && trans->transid != root->last_trans);
+ level = btrfs_header_level(buf);
+ nritems = btrfs_header_nritems(buf);
+ if (nritems) {
+ if (level == 0)
+ btrfs_item_key_to_cpu(buf, &first_key, 0);
+ else
+ btrfs_node_key_to_cpu(buf, &first_key, 0);
+ } else {
+ first_key.objectid = 0;
+ }
cow = btrfs_alloc_free_block(trans, root, buf->len,
+ root->root_key.objectid,
+ root_gen, first_key.objectid, level,
search_start, empty_size);
if (IS_ERR(cow))
return PTR_ERR(cow);
btrfs_set_header_bytenr(cow, cow->start);
btrfs_set_header_generation(cow, trans->transid);
btrfs_set_header_owner(cow, root->root_key.objectid);
+ btrfs_clear_header_flag(cow, BTRFS_HEADER_FLAG_WRITTEN);
WARN_ON(btrfs_header_generation(buf) > trans->transid);
if (btrfs_header_generation(buf) != trans->transid) {
}
if (buf == root->node) {
+ WARN_ON(parent && parent != buf);
+ root_gen = btrfs_header_generation(buf);
+
+ spin_lock(&root->node_lock);
root->node = cow;
extent_buffer_get(cow);
+ spin_unlock(&root->node_lock);
+
if (buf != root->commit_root) {
btrfs_free_extent(trans, root, buf->start,
- buf->len, 1);
+ buf->len, root->root_key.objectid,
+ root_gen, 0, 0, 1);
}
free_extent_buffer(buf);
+ add_root_to_dirty_list(root);
} else {
+ root_gen = btrfs_header_generation(parent);
btrfs_set_node_blockptr(parent, parent_slot,
cow->start);
+ WARN_ON(trans->transid == 0);
+ btrfs_set_node_ptr_generation(parent, parent_slot,
+ trans->transid);
btrfs_mark_buffer_dirty(parent);
WARN_ON(btrfs_header_generation(parent) != trans->transid);
- btrfs_free_extent(trans, root, buf->start, buf->len, 1);
+ btrfs_free_extent(trans, root, buf->start, buf->len,
+ btrfs_header_owner(parent), root_gen,
+ 0, 0, 1);
}
+ if (unlock_orig)
+ btrfs_tree_unlock(buf);
free_extent_buffer(buf);
btrfs_mark_buffer_dirty(cow);
*cow_ret = cow;
struct extent_buffer **cow_ret)
{
u64 search_start;
+ u64 header_trans;
int ret;
+
if (trans->transaction != root->fs_info->running_transaction) {
printk(KERN_CRIT "trans %Lu running %Lu\n", trans->transid,
root->fs_info->running_transaction->transid);
root->fs_info->generation);
WARN_ON(1);
}
- if (btrfs_header_generation(buf) == trans->transid) {
+
+ header_trans = btrfs_header_generation(buf);
+ spin_lock(&root->fs_info->hash_lock);
+ if (header_trans == trans->transid &&
+ !btrfs_header_flag(buf, BTRFS_HEADER_FLAG_WRITTEN)) {
*cow_ret = buf;
+ spin_unlock(&root->fs_info->hash_lock);
return 0;
}
-
- search_start = buf->start & ~((u64)BTRFS_BLOCK_GROUP_SIZE - 1);
+ spin_unlock(&root->fs_info->hash_lock);
+ search_start = buf->start & ~((u64)(1024 * 1024 * 1024) - 1);
ret = __btrfs_cow_block(trans, root, buf, parent,
parent_slot, cow_ret, search_start, 0);
return ret;
return 0;
}
-static int should_defrag_leaf(struct extent_buffer *leaf)
+/*
+ * compare two keys in a memcmp fashion
+ */
+static int comp_keys(struct btrfs_disk_key *disk, struct btrfs_key *k2)
{
- struct btrfs_key key;
- u32 nritems;
-
- if (btrfs_buffer_defrag(leaf))
- return 1;
+ struct btrfs_key k1;
- nritems = btrfs_header_nritems(leaf);
- if (nritems == 0)
- return 0;
+ btrfs_disk_key_to_cpu(&k1, disk);
- btrfs_item_key_to_cpu(leaf, &key, 0);
- if (key.type == BTRFS_DIR_ITEM_KEY)
+ if (k1.objectid > k2->objectid)
return 1;
-
-
- btrfs_item_key_to_cpu(leaf, &key, nritems - 1);
- if (key.type == BTRFS_DIR_ITEM_KEY)
+ if (k1.objectid < k2->objectid)
+ return -1;
+ if (k1.type > k2->type)
return 1;
- if (nritems > 4) {
- btrfs_item_key_to_cpu(leaf, &key, nritems / 2);
- if (key.type == BTRFS_DIR_ITEM_KEY)
- return 1;
- }
+ if (k1.type < k2->type)
+ return -1;
+ if (k1.offset > k2->offset)
+ return 1;
+ if (k1.offset < k2->offset)
+ return -1;
return 0;
}
+
int btrfs_realloc_node(struct btrfs_trans_handle *trans,
struct btrfs_root *root, struct extent_buffer *parent,
int start_slot, int cache_only, u64 *last_ret,
struct extent_buffer *cur;
struct extent_buffer *tmp;
u64 blocknr;
+ u64 gen;
u64 search_start = *last_ret;
u64 last_block = 0;
u64 other;
int parent_level;
int uptodate;
u32 blocksize;
+ int progress_passed = 0;
+ struct btrfs_disk_key disk_key;
+
+ /* FIXME this code needs locking */
+ return 0;
+
+ parent_level = btrfs_header_level(parent);
+ if (cache_only && parent_level != 1)
+ return 0;
if (trans->transaction != root->fs_info->running_transaction) {
printk(KERN_CRIT "trans %Lu running %Lu\n", trans->transid,
root->fs_info->generation);
WARN_ON(1);
}
- parent_level = btrfs_header_level(parent);
parent_nritems = btrfs_header_nritems(parent);
blocksize = btrfs_level_size(root, parent_level - 1);
if (parent_nritems == 1)
return 0;
- if (root != root->fs_info->extent_root) {
- struct btrfs_key first_key;
- struct btrfs_key last_key;
-
- btrfs_node_key_to_cpu(parent, &first_key, 0);
- btrfs_node_key_to_cpu(parent, &last_key, parent_nritems - 1);
- if (first_key.objectid != last_key.objectid)
- return 0;
- }
-
for (i = start_slot; i < end_slot; i++) {
int close = 1;
+ if (!parent->map_token) {
+ map_extent_buffer(parent,
+ btrfs_node_key_ptr_offset(i),
+ sizeof(struct btrfs_key_ptr),
+ &parent->map_token, &parent->kaddr,
+ &parent->map_start, &parent->map_len,
+ KM_USER1);
+ }
+ btrfs_node_key(parent, &disk_key, i);
+ if (!progress_passed && comp_keys(&disk_key, progress) < 0)
+ continue;
+
+ progress_passed = 1;
blocknr = btrfs_node_blockptr(parent, i);
+ gen = btrfs_node_ptr_generation(parent, i);
if (last_block == 0)
last_block = blocknr;
+
if (i > 0) {
other = btrfs_node_blockptr(parent, i - 1);
close = close_blocks(blocknr, other, blocksize);
}
- if (close && i < end_slot - 1) {
+ if (!close && i < end_slot - 2) {
other = btrfs_node_blockptr(parent, i + 1);
close = close_blocks(blocknr, other, blocksize);
}
last_block = blocknr;
continue;
}
+ if (parent->map_token) {
+ unmap_extent_buffer(parent, parent->map_token,
+ KM_USER1);
+ parent->map_token = NULL;
+ }
cur = btrfs_find_tree_block(root, blocknr, blocksize);
if (cur)
- uptodate = btrfs_buffer_uptodate(cur);
+ uptodate = btrfs_buffer_uptodate(cur, gen);
else
uptodate = 0;
- if (!cur || !uptodate ||
- (parent_level != 1 && !btrfs_buffer_defrag(cur)) ||
- (parent_level == 1 && !should_defrag_leaf(cur))) {
+ if (!cur || !uptodate) {
if (cache_only) {
free_extent_buffer(cur);
continue;
}
if (!cur) {
cur = read_tree_block(root, blocknr,
- blocksize);
+ blocksize, gen);
} else if (!uptodate) {
- btrfs_read_buffer(cur);
+ btrfs_read_buffer(cur, gen);
}
}
if (search_start == 0)
break;
}
search_start = tmp->start;
+ last_block = tmp->start;
*last_ret = search_start;
if (parent_level == 1)
btrfs_clear_buffer_defrag(tmp);
free_extent_buffer(tmp);
}
+ if (parent->map_token) {
+ unmap_extent_buffer(parent, parent->map_token,
+ KM_USER1);
+ parent->map_token = NULL;
+ }
return err;
}
return btrfs_item_offset_nr(leaf, nr - 1);
}
-/*
- * compare two keys in a memcmp fashion
- */
-static int comp_keys(struct btrfs_disk_key *disk, struct btrfs_key *k2)
-{
- struct btrfs_key k1;
-
- btrfs_disk_key_to_cpu(&k1, disk);
-
- if (k1.objectid > k2->objectid)
- return 1;
- if (k1.objectid < k2->objectid)
- return -1;
- if (k1.type > k2->type)
- return 1;
- if (k1.type < k2->type)
- return -1;
- if (k1.offset > k2->offset)
- return 1;
- if (k1.offset < k2->offset)
- return -1;
- return 0;
-}
-
static int check_node(struct btrfs_root *root, struct btrfs_path *path,
int level)
{
return 0;
}
-static int check_block(struct btrfs_root *root, struct btrfs_path *path,
- int level)
+static int noinline check_block(struct btrfs_root *root,
+ struct btrfs_path *path, int level)
{
+ u64 found_start;
return 0;
+ if (btrfs_header_level(path->nodes[level]) != level)
+ printk("warning: bad level %Lu wanted %d found %d\n",
+ path->nodes[level]->start, level,
+ btrfs_header_level(path->nodes[level]));
+ found_start = btrfs_header_bytenr(path->nodes[level]);
+ if (found_start != path->nodes[level]->start) {
+ printk("warning: bad bytentr %Lu found %Lu\n",
+ path->nodes[level]->start, found_start);
+ }
#if 0
struct extent_buffer *buf = path->nodes[level];
static struct extent_buffer *read_node_slot(struct btrfs_root *root,
struct extent_buffer *parent, int slot)
{
+ int level = btrfs_header_level(parent);
if (slot < 0)
return NULL;
if (slot >= btrfs_header_nritems(parent))
return NULL;
+
+ BUG_ON(level == 0);
+
return read_tree_block(root, btrfs_node_blockptr(parent, slot),
- btrfs_level_size(root, btrfs_header_level(parent) - 1));
+ btrfs_level_size(root, level - 1),
+ btrfs_node_ptr_generation(parent, slot));
}
-static int balance_level(struct btrfs_trans_handle *trans, struct btrfs_root
- *root, struct btrfs_path *path, int level)
+static int balance_level(struct btrfs_trans_handle *trans,
+ struct btrfs_root *root,
+ struct btrfs_path *path, int level)
{
struct extent_buffer *right = NULL;
struct extent_buffer *mid;
return 0;
mid = path->nodes[level];
+ WARN_ON(!path->locks[level]);
+ WARN_ON(btrfs_header_generation(mid) != trans->transid);
+
orig_ptr = btrfs_node_blockptr(mid, orig_slot);
if (level < BTRFS_MAX_LEVEL - 1)
/* promote the child to a root */
child = read_node_slot(root, mid, 0);
+ btrfs_tree_lock(child);
BUG_ON(!child);
+ ret = btrfs_cow_block(trans, root, child, mid, 0, &child);
+ BUG_ON(ret);
+
+ spin_lock(&root->node_lock);
root->node = child;
+ spin_unlock(&root->node_lock);
+
+ add_root_to_dirty_list(root);
+ btrfs_tree_unlock(child);
+ path->locks[level] = 0;
path->nodes[level] = NULL;
clean_tree_block(trans, root, mid);
- wait_on_tree_block_writeback(root, mid);
+ btrfs_tree_unlock(mid);
/* once for the path */
free_extent_buffer(mid);
- ret = btrfs_free_extent(trans, root, mid->start, mid->len, 1);
+ ret = btrfs_free_extent(trans, root, mid->start, mid->len,
+ root->root_key.objectid,
+ btrfs_header_generation(mid), 0, 0, 1);
/* once for the root ptr */
free_extent_buffer(mid);
return ret;
left = read_node_slot(root, parent, pslot - 1);
if (left) {
+ btrfs_tree_lock(left);
wret = btrfs_cow_block(trans, root, left,
parent, pslot - 1, &left);
if (wret) {
}
right = read_node_slot(root, parent, pslot + 1);
if (right) {
+ btrfs_tree_lock(right);
wret = btrfs_cow_block(trans, root, right,
parent, pslot + 1, &right);
if (wret) {
/* first, try to make some room in the middle buffer */
if (left) {
orig_slot += btrfs_header_nritems(left);
- wret = push_node_left(trans, root, left, mid);
+ wret = push_node_left(trans, root, left, mid, 1);
if (wret < 0)
ret = wret;
if (btrfs_header_nritems(mid) < 2)
* then try to empty the right most buffer into the middle
*/
if (right) {
- wret = push_node_left(trans, root, mid, right);
+ wret = push_node_left(trans, root, mid, right, 1);
if (wret < 0 && wret != -ENOSPC)
ret = wret;
if (btrfs_header_nritems(right) == 0) {
u64 bytenr = right->start;
+ u64 generation = btrfs_header_generation(parent);
u32 blocksize = right->len;
clean_tree_block(trans, root, right);
- wait_on_tree_block_writeback(root, right);
+ btrfs_tree_unlock(right);
free_extent_buffer(right);
right = NULL;
wret = del_ptr(trans, root, path, level + 1, pslot +
if (wret)
ret = wret;
wret = btrfs_free_extent(trans, root, bytenr,
- blocksize, 1);
+ blocksize,
+ btrfs_header_owner(parent),
+ generation, 0, 0, 1);
if (wret)
ret = wret;
} else {
ret = wret;
goto enospc;
}
+ if (wret == 1) {
+ wret = push_node_left(trans, root, left, mid, 1);
+ if (wret < 0)
+ ret = wret;
+ }
BUG_ON(wret == 1);
}
if (btrfs_header_nritems(mid) == 0) {
/* we've managed to empty the middle node, drop it */
+ u64 root_gen = btrfs_header_generation(parent);
u64 bytenr = mid->start;
u32 blocksize = mid->len;
+
clean_tree_block(trans, root, mid);
- wait_on_tree_block_writeback(root, mid);
+ btrfs_tree_unlock(mid);
free_extent_buffer(mid);
mid = NULL;
wret = del_ptr(trans, root, path, level + 1, pslot);
if (wret)
ret = wret;
- wret = btrfs_free_extent(trans, root, bytenr, blocksize, 1);
+ wret = btrfs_free_extent(trans, root, bytenr, blocksize,
+ btrfs_header_owner(parent),
+ root_gen, 0, 0, 1);
if (wret)
ret = wret;
} else {
if (left) {
if (btrfs_header_nritems(left) > orig_slot) {
extent_buffer_get(left);
+ /* left was locked after cow */
path->nodes[level] = left;
path->slots[level + 1] -= 1;
path->slots[level] = orig_slot;
- if (mid)
+ if (mid) {
+ btrfs_tree_unlock(mid);
free_extent_buffer(mid);
+ }
} else {
orig_slot -= btrfs_header_nritems(left);
path->slots[level] = orig_slot;
btrfs_node_blockptr(path->nodes[level], path->slots[level]))
BUG();
enospc:
- if (right)
+ if (right) {
+ btrfs_tree_unlock(right);
free_extent_buffer(right);
- if (left)
+ }
+ if (left) {
+ if (path->nodes[level] != left)
+ btrfs_tree_unlock(left);
free_extent_buffer(left);
+ }
return ret;
}
/* returns zero if the push worked, non-zero otherwise */
-static int push_nodes_for_insert(struct btrfs_trans_handle *trans,
- struct btrfs_root *root,
- struct btrfs_path *path, int level)
+static int noinline push_nodes_for_insert(struct btrfs_trans_handle *trans,
+ struct btrfs_root *root,
+ struct btrfs_path *path, int level)
{
struct extent_buffer *right = NULL;
struct extent_buffer *mid;
return 1;
mid = path->nodes[level];
+ WARN_ON(btrfs_header_generation(mid) != trans->transid);
orig_ptr = btrfs_node_blockptr(mid, orig_slot);
if (level < BTRFS_MAX_LEVEL - 1)
/* first, try to make some room in the middle buffer */
if (left) {
u32 left_nr;
+
+ btrfs_tree_lock(left);
left_nr = btrfs_header_nritems(left);
if (left_nr >= BTRFS_NODEPTRS_PER_BLOCK(root) - 1) {
wret = 1;
wret = 1;
else {
wret = push_node_left(trans, root,
- left, mid);
+ left, mid, 0);
}
}
if (wret < 0)
path->nodes[level] = left;
path->slots[level + 1] -= 1;
path->slots[level] = orig_slot;
+ btrfs_tree_unlock(mid);
free_extent_buffer(mid);
} else {
orig_slot -=
btrfs_header_nritems(left);
path->slots[level] = orig_slot;
+ btrfs_tree_unlock(left);
free_extent_buffer(left);
}
return 0;
}
+ btrfs_tree_unlock(left);
free_extent_buffer(left);
}
- right= read_node_slot(root, parent, pslot + 1);
+ right = read_node_slot(root, parent, pslot + 1);
/*
* then try to empty the right most buffer into the middle
*/
if (right) {
u32 right_nr;
+ btrfs_tree_lock(right);
right_nr = btrfs_header_nritems(right);
if (right_nr >= BTRFS_NODEPTRS_PER_BLOCK(root) - 1) {
wret = 1;
path->slots[level + 1] += 1;
path->slots[level] = orig_slot -
btrfs_header_nritems(mid);
+ btrfs_tree_unlock(mid);
free_extent_buffer(mid);
} else {
+ btrfs_tree_unlock(right);
free_extent_buffer(right);
}
return 0;
}
+ btrfs_tree_unlock(right);
free_extent_buffer(right);
}
return 1;
* readahead one full node of leaves
*/
static void reada_for_search(struct btrfs_root *root, struct btrfs_path *path,
- int level, int slot)
+ int level, int slot, u64 objectid)
{
struct extent_buffer *node;
+ struct btrfs_disk_key disk_key;
u32 nritems;
u64 search;
u64 lowest_read;
return;
node = path->nodes[level];
+ WARN_ON(!path->skip_locking && !btrfs_tree_locked(node));
+
search = btrfs_node_blockptr(node, slot);
blocksize = btrfs_level_size(root, level - 1);
eb = btrfs_find_tree_block(root, search, blocksize);
if (nr >= nritems)
break;
}
+ if (path->reada < 0 && objectid) {
+ btrfs_node_key(node, &disk_key, nr);
+ if (btrfs_disk_key_objectid(&disk_key) != objectid)
+ break;
+ }
search = btrfs_node_blockptr(node, nr);
if ((search >= lowest_read && search <= highest_read) ||
(search < lowest_read && lowest_read - search <= 32768) ||
(search > highest_read && search - highest_read <= 32768)) {
- readahead_tree_block(root, search, blocksize);
+ readahead_tree_block(root, search, blocksize,
+ btrfs_node_ptr_generation(node, nr));
nread += blocksize;
}
nscan++;
highest_read = search;
}
}
+
+static void unlock_up(struct btrfs_path *path, int level, int lowest_unlock)
+{
+ int i;
+ int skip_level = level;
+ struct extent_buffer *t;
+
+ for (i = level; i < BTRFS_MAX_LEVEL; i++) {
+ if (!path->nodes[i])
+ break;
+ if (!path->locks[i])
+ break;
+ if (path->slots[i] == 0) {
+ skip_level = i + 1;
+ continue;
+ }
+ if (path->keep_locks) {
+ u32 nritems;
+ t = path->nodes[i];
+ nritems = btrfs_header_nritems(t);
+ if (nritems < 2 || path->slots[i] >= nritems - 2) {
+if (path->keep_locks) {
+//printk("path %p skip level now %d\n", path, skip_level);
+}
+ skip_level = i + 1;
+ continue;
+ }
+ }
+ t = path->nodes[i];
+ if (i >= lowest_unlock && i > skip_level && path->locks[i]) {
+if (path->keep_locks) {
+//printk("path %p unlocking level %d slot %d nritems %d skip_level %d\n", path, i, path->slots[i], btrfs_header_nritems(t), skip_level);
+}
+ btrfs_tree_unlock(t);
+ path->locks[i] = 0;
+ }
+ }
+}
+
/*
* look for key in the tree. path is filled in with nodes along the way
* if key is found, we return zero and you can find the item in the leaf
ins_len, int cow)
{
struct extent_buffer *b;
- u64 bytenr;
int slot;
int ret;
int level;
int should_reada = p->reada;
+ int lowest_unlock = 1;
u8 lowest_level = 0;
lowest_level = p->lowest_level;
WARN_ON(lowest_level && ins_len);
WARN_ON(p->nodes[0] != NULL);
- WARN_ON(!mutex_is_locked(&root->fs_info->fs_mutex));
+ WARN_ON(root == root->fs_info->extent_root &&
+ !mutex_is_locked(&root->fs_info->alloc_mutex));
+ WARN_ON(root == root->fs_info->chunk_root &&
+ !mutex_is_locked(&root->fs_info->chunk_mutex));
+ WARN_ON(root == root->fs_info->dev_root &&
+ !mutex_is_locked(&root->fs_info->chunk_mutex));
+ if (ins_len < 0)
+ lowest_unlock = 2;
again:
- b = root->node;
- extent_buffer_get(b);
+ if (!p->skip_locking)
+ b = btrfs_lock_root_node(root);
+ else
+ b = btrfs_root_node(root);
+
while (b) {
level = btrfs_header_level(b);
if (cow) {
WARN_ON(1);
level = btrfs_header_level(b);
p->nodes[level] = b;
+ if (!p->skip_locking)
+ p->locks[level] = 1;
ret = check_block(root, p, level);
if (ret)
return -1;
+
ret = bin_search(b, key, level, &slot);
if (level != 0) {
if (ret && slot > 0)
slot -= 1;
p->slots[level] = slot;
if (ins_len > 0 && btrfs_header_nritems(b) >=
- BTRFS_NODEPTRS_PER_BLOCK(root) - 1) {
+ BTRFS_NODEPTRS_PER_BLOCK(root) - 3) {
int sret = split_node(trans, root, p, level);
BUG_ON(sret > 0);
if (sret)
BUG_ON(btrfs_header_nritems(b) == 1);
}
/* this is only true while dropping a snapshot */
- if (level == lowest_level)
+ if (level == lowest_level) {
+ unlock_up(p, level, lowest_unlock);
break;
- bytenr = btrfs_node_blockptr(b, slot);
+ }
+
if (should_reada)
- reada_for_search(root, p, level, slot);
- b = read_tree_block(root, bytenr,
- btrfs_level_size(root, level - 1));
+ reada_for_search(root, p, level, slot,
+ key->objectid);
+
+ b = read_node_slot(root, b, slot);
+ if (!p->skip_locking)
+ btrfs_tree_lock(b);
+ unlock_up(p, level + 1, lowest_unlock);
} else {
p->slots[level] = slot;
if (ins_len > 0 && btrfs_leaf_free_space(root, b) <
sizeof(struct btrfs_item) + ins_len) {
int sret = split_leaf(trans, root, key,
- p, ins_len);
+ p, ins_len, ret == 0);
BUG_ON(sret > 0);
if (sret)
return sret;
}
+ unlock_up(p, level, lowest_unlock);
return ret;
}
}
break;
t = path->nodes[i];
btrfs_set_node_key(t, key, tslot);
+ if (!btrfs_tree_locked(path->nodes[i])) {
+ int ii;
+printk("fixup without lock on level %d\n", btrfs_header_level(path->nodes[i]));
+ for (ii = 0; ii < BTRFS_MAX_LEVEL; ii++) {
+printk("level %d slot %d\n", ii, path->slots[ii]);
+ }
+ }
btrfs_mark_buffer_dirty(path->nodes[i]);
if (tslot != 0)
break;
* returns 0 if some ptrs were pushed left, < 0 if there was some horrible
* error, and > 0 if there was no room in the left hand block.
*/
-static int push_node_left(struct btrfs_trans_handle *trans, struct btrfs_root
- *root, struct extent_buffer *dst,
- struct extent_buffer *src)
+static int push_node_left(struct btrfs_trans_handle *trans,
+ struct btrfs_root *root, struct extent_buffer *dst,
+ struct extent_buffer *src, int empty)
{
int push_items = 0;
int src_nritems;
src_nritems = btrfs_header_nritems(src);
dst_nritems = btrfs_header_nritems(dst);
push_items = BTRFS_NODEPTRS_PER_BLOCK(root) - dst_nritems;
+ WARN_ON(btrfs_header_generation(src) != trans->transid);
+ WARN_ON(btrfs_header_generation(dst) != trans->transid);
+
+ if (!empty && src_nritems <= 8)
+ return 1;
if (push_items <= 0) {
return 1;
}
- if (src_nritems < push_items)
- push_items = src_nritems;
+ if (empty) {
+ push_items = min(src_nritems, push_items);
+ if (push_items < src_nritems) {
+ /* leave at least 8 pointers in the node if
+ * we aren't going to empty it
+ */
+ if (src_nritems - push_items < 8) {
+ if (push_items <= 8)
+ return 1;
+ push_items -= 8;
+ }
+ }
+ } else
+ push_items = min(src_nritems - 8, push_items);
copy_extent_buffer(dst, src,
btrfs_node_key_ptr_offset(dst_nritems),
int dst_nritems;
int ret = 0;
+ WARN_ON(btrfs_header_generation(src) != trans->transid);
+ WARN_ON(btrfs_header_generation(dst) != trans->transid);
+
src_nritems = btrfs_header_nritems(src);
dst_nritems = btrfs_header_nritems(dst);
push_items = BTRFS_NODEPTRS_PER_BLOCK(root) - dst_nritems;
- if (push_items <= 0)
+ if (push_items <= 0) {
+ return 1;
+ }
+
+ if (src_nritems < 4) {
return 1;
+ }
max_push = src_nritems / 2 + 1;
/* don't try to empty the node */
- if (max_push >= src_nritems)
+ if (max_push >= src_nritems) {
return 1;
+ }
if (max_push < push_items)
push_items = max_push;
*
* returns zero on success or < 0 on failure.
*/
-static int insert_new_root(struct btrfs_trans_handle *trans,
+static int noinline insert_new_root(struct btrfs_trans_handle *trans,
struct btrfs_root *root,
struct btrfs_path *path, int level)
{
+ u64 root_gen;
+ u64 lower_gen;
struct extent_buffer *lower;
struct extent_buffer *c;
+ struct extent_buffer *old;
struct btrfs_disk_key lower_key;
BUG_ON(path->nodes[level]);
BUG_ON(path->nodes[level-1] != root->node);
+ if (root->ref_cows)
+ root_gen = trans->transid;
+ else
+ root_gen = 0;
+
+ lower = path->nodes[level-1];
+ if (level == 1)
+ btrfs_item_key(lower, &lower_key, 0);
+ else
+ btrfs_node_key(lower, &lower_key, 0);
+
c = btrfs_alloc_free_block(trans, root, root->nodesize,
+ root->root_key.objectid,
+ root_gen, lower_key.objectid, level,
root->node->start, 0);
if (IS_ERR(c))
return PTR_ERR(c);
+
memset_extent_buffer(c, 0, 0, root->nodesize);
btrfs_set_header_nritems(c, 1);
btrfs_set_header_level(c, level);
btrfs_set_header_bytenr(c, c->start);
btrfs_set_header_generation(c, trans->transid);
btrfs_set_header_owner(c, root->root_key.objectid);
- lower = path->nodes[level-1];
write_extent_buffer(c, root->fs_info->fsid,
(unsigned long)btrfs_header_fsid(c),
BTRFS_FSID_SIZE);
- if (level == 1)
- btrfs_item_key(lower, &lower_key, 0);
- else
- btrfs_node_key(lower, &lower_key, 0);
+
+ write_extent_buffer(c, root->fs_info->chunk_tree_uuid,
+ (unsigned long)btrfs_header_chunk_tree_uuid(c),
+ BTRFS_UUID_SIZE);
+
btrfs_set_node_key(c, &lower_key, 0);
btrfs_set_node_blockptr(c, 0, lower->start);
+ lower_gen = btrfs_header_generation(lower);
+ WARN_ON(lower_gen == 0);
+
+ btrfs_set_node_ptr_generation(c, 0, lower_gen);
btrfs_mark_buffer_dirty(c);
- /* the super has an extra ref to root->node */
- free_extent_buffer(root->node);
+ spin_lock(&root->node_lock);
+ old = root->node;
root->node = c;
+ spin_unlock(&root->node_lock);
+
+ /* the super has an extra ref to root->node */
+ free_extent_buffer(old);
+
+ add_root_to_dirty_list(root);
extent_buffer_get(c);
path->nodes[level] = c;
+ path->locks[level] = 1;
path->slots[level] = 0;
+
+ if (root->ref_cows && lower_gen != trans->transid) {
+ struct btrfs_path *back_path = btrfs_alloc_path();
+ int ret;
+ mutex_lock(&root->fs_info->alloc_mutex);
+ ret = btrfs_insert_extent_backref(trans,
+ root->fs_info->extent_root,
+ path, lower->start,
+ root->root_key.objectid,
+ trans->transid, 0, 0);
+ BUG_ON(ret);
+ mutex_unlock(&root->fs_info->alloc_mutex);
+ btrfs_free_path(back_path);
+ }
return 0;
}
}
btrfs_set_node_key(lower, key, slot);
btrfs_set_node_blockptr(lower, slot, bytenr);
+ WARN_ON(trans->transid == 0);
+ btrfs_set_node_ptr_generation(lower, slot, trans->transid);
btrfs_set_header_nritems(lower, nritems + 1);
btrfs_mark_buffer_dirty(lower);
return 0;
static 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;
struct extent_buffer *split;
struct btrfs_disk_key disk_key;
u32 c_nritems;
c = path->nodes[level];
+ WARN_ON(btrfs_header_generation(c) != trans->transid);
if (c == root->node) {
/* trying to split the root, lets make a new one */
ret = insert_new_root(trans, root, path, level + 1);
ret = push_nodes_for_insert(trans, root, path, level);
c = path->nodes[level];
if (!ret && btrfs_header_nritems(c) <
- BTRFS_NODEPTRS_PER_BLOCK(root) - 1)
+ BTRFS_NODEPTRS_PER_BLOCK(root) - 3)
return 0;
if (ret < 0)
return ret;
}
c_nritems = btrfs_header_nritems(c);
+ if (root->ref_cows)
+ root_gen = trans->transid;
+ else
+ root_gen = 0;
+
+ btrfs_node_key(c, &disk_key, 0);
split = btrfs_alloc_free_block(trans, root, root->nodesize,
- c->start, 0);
+ root->root_key.objectid,
+ root_gen,
+ btrfs_disk_key_objectid(&disk_key),
+ level, c->start, 0);
if (IS_ERR(split))
return PTR_ERR(split);
btrfs_set_header_bytenr(split, split->start);
btrfs_set_header_generation(split, trans->transid);
btrfs_set_header_owner(split, root->root_key.objectid);
+ btrfs_set_header_flags(split, 0);
write_extent_buffer(split, root->fs_info->fsid,
(unsigned long)btrfs_header_fsid(split),
BTRFS_FSID_SIZE);
+ write_extent_buffer(split, root->fs_info->chunk_tree_uuid,
+ (unsigned long)btrfs_header_chunk_tree_uuid(split),
+ BTRFS_UUID_SIZE);
mid = (c_nritems + 1) / 2;
if (path->slots[level] >= mid) {
path->slots[level] -= mid;
+ btrfs_tree_unlock(c);
free_extent_buffer(c);
path->nodes[level] = split;
path->slots[level + 1] += 1;
} else {
+ btrfs_tree_unlock(split);
free_extent_buffer(split);
}
return ret;
ret = BTRFS_LEAF_DATA_SIZE(root) - leaf_space_used(leaf, 0, nritems);
if (ret < 0) {
printk("leaf free space ret %d, leaf data size %lu, used %d nritems %d\n",
- ret, BTRFS_LEAF_DATA_SIZE(root),
+ ret, (unsigned long) BTRFS_LEAF_DATA_SIZE(root),
leaf_space_used(leaf, 0, nritems), nritems);
}
return ret;
* room, 0 if everything worked out and < 0 if there were major errors.
*/
static int push_leaf_right(struct btrfs_trans_handle *trans, struct btrfs_root
- *root, struct btrfs_path *path, int data_size)
+ *root, struct btrfs_path *path, int data_size,
+ int empty)
{
struct extent_buffer *left = path->nodes[0];
struct extent_buffer *right;
struct extent_buffer *upper;
struct btrfs_disk_key disk_key;
int slot;
- int i;
+ u32 i;
int free_space;
int push_space = 0;
int push_items = 0;
struct btrfs_item *item;
u32 left_nritems;
+ u32 nr;
u32 right_nritems;
u32 data_end;
u32 this_item_size;
if (slot >= btrfs_header_nritems(upper) - 1)
return 1;
- right = read_tree_block(root, btrfs_node_blockptr(upper, slot + 1),
- root->leafsize);
+ WARN_ON(!btrfs_tree_locked(path->nodes[1]));
+
+ right = read_node_slot(root, upper, slot + 1);
+ btrfs_tree_lock(right);
free_space = btrfs_leaf_free_space(root, right);
- if (free_space < data_size + sizeof(struct btrfs_item)) {
- free_extent_buffer(right);
- return 1;
- }
+ if (free_space < data_size + sizeof(struct btrfs_item))
+ goto out_unlock;
/* cow and double check */
ret = btrfs_cow_block(trans, root, right, upper,
slot + 1, &right);
- if (ret) {
- free_extent_buffer(right);
- return 1;
- }
+ if (ret)
+ goto out_unlock;
+
free_space = btrfs_leaf_free_space(root, right);
- if (free_space < data_size + sizeof(struct btrfs_item)) {
- free_extent_buffer(right);
- return 1;
- }
+ if (free_space < data_size + sizeof(struct btrfs_item))
+ goto out_unlock;
left_nritems = btrfs_header_nritems(left);
- if (left_nritems == 0) {
- free_extent_buffer(right);
- return 1;
- }
+ if (left_nritems == 0)
+ goto out_unlock;
- for (i = left_nritems - 1; i >= 1; i--) {
+ if (empty)
+ nr = 0;
+ else
+ nr = 1;
+
+ i = left_nritems - 1;
+ while (i >= nr) {
item = btrfs_item_nr(left, i);
if (path->slots[0] == i)
break;
push_items++;
push_space += this_item_size + sizeof(*item);
+ if (i == 0)
+ break;
+ i--;
}
if (left->map_token) {
unmap_extent_buffer(left, left->map_token, KM_USER1);
left->map_token = NULL;
}
- if (push_items == 0) {
- free_extent_buffer(right);
- return 1;
- }
+ if (push_items == 0)
+ goto out_unlock;
- if (push_items == left_nritems)
+ if (!empty && push_items == left_nritems)
WARN_ON(1);
/* push left to right */
right_nritems = btrfs_header_nritems(right);
+
push_space = btrfs_item_end_nr(left, left_nritems - push_items);
push_space -= leaf_data_end(root, left);
right_nritems += push_items;
btrfs_set_header_nritems(right, right_nritems);
push_space = BTRFS_LEAF_DATA_SIZE(root);
-
for (i = 0; i < right_nritems; i++) {
item = btrfs_item_nr(right, i);
if (!right->map_token) {
left_nritems -= push_items;
btrfs_set_header_nritems(left, left_nritems);
- btrfs_mark_buffer_dirty(left);
+ if (left_nritems)
+ btrfs_mark_buffer_dirty(left);
btrfs_mark_buffer_dirty(right);
btrfs_item_key(right, &disk_key, 0);
/* then fixup the leaf pointer in the path */
if (path->slots[0] >= left_nritems) {
path->slots[0] -= left_nritems;
+ if (btrfs_header_nritems(path->nodes[0]) == 0)
+ clean_tree_block(trans, root, path->nodes[0]);
+ btrfs_tree_unlock(path->nodes[0]);
free_extent_buffer(path->nodes[0]);
path->nodes[0] = right;
path->slots[1] += 1;
} else {
+ btrfs_tree_unlock(right);
free_extent_buffer(right);
}
return 0;
+
+out_unlock:
+ btrfs_tree_unlock(right);
+ free_extent_buffer(right);
+ return 1;
}
+
/*
* push some data in the path leaf to the left, trying to free up at
* least data_size bytes. returns zero if the push worked, nonzero otherwise
*/
static int push_leaf_left(struct btrfs_trans_handle *trans, struct btrfs_root
- *root, struct btrfs_path *path, int data_size)
+ *root, struct btrfs_path *path, int data_size,
+ int empty)
{
struct btrfs_disk_key disk_key;
struct extent_buffer *right = path->nodes[0];
struct btrfs_item *item;
u32 old_left_nritems;
u32 right_nritems;
+ u32 nr;
int ret = 0;
int wret;
u32 this_item_size;
if (!path->nodes[1])
return 1;
- left = read_tree_block(root, btrfs_node_blockptr(path->nodes[1],
- slot - 1), root->leafsize);
+ right_nritems = btrfs_header_nritems(right);
+ if (right_nritems == 0) {
+ return 1;
+ }
+
+ WARN_ON(!btrfs_tree_locked(path->nodes[1]));
+
+ left = read_node_slot(root, path->nodes[1], slot - 1);
+ btrfs_tree_lock(left);
free_space = btrfs_leaf_free_space(root, left);
if (free_space < data_size + sizeof(struct btrfs_item)) {
- free_extent_buffer(left);
- return 1;
+ ret = 1;
+ goto out;
}
/* cow and double check */
path->nodes[1], slot - 1, &left);
if (ret) {
/* we hit -ENOSPC, but it isn't fatal here */
- free_extent_buffer(left);
- return 1;
+ ret = 1;
+ goto out;
}
+
free_space = btrfs_leaf_free_space(root, left);
if (free_space < data_size + sizeof(struct btrfs_item)) {
- free_extent_buffer(left);
- return 1;
+ ret = 1;
+ goto out;
}
- right_nritems = btrfs_header_nritems(right);
- if (right_nritems == 0) {
- free_extent_buffer(left);
- return 1;
- }
+ if (empty)
+ nr = right_nritems;
+ else
+ nr = right_nritems - 1;
- for (i = 0; i < right_nritems - 1; i++) {
+ for (i = 0; i < nr; i++) {
item = btrfs_item_nr(right, i);
if (!right->map_token) {
map_extent_buffer(right, (unsigned long)item,
}
if (push_items == 0) {
- free_extent_buffer(left);
- return 1;
+ ret = 1;
+ goto out;
}
- if (push_items == btrfs_header_nritems(right))
+ if (!empty && push_items == btrfs_header_nritems(right))
WARN_ON(1);
/* push data from right to left */
}
/* fixup right node */
- push_space = btrfs_item_offset_nr(right, push_items - 1) -
- leaf_data_end(root, right);
- memmove_extent_buffer(right, btrfs_leaf_data(right) +
- BTRFS_LEAF_DATA_SIZE(root) - push_space,
- btrfs_leaf_data(right) +
- leaf_data_end(root, right), push_space);
-
- memmove_extent_buffer(right, btrfs_item_nr_offset(0),
+ if (push_items > right_nritems) {
+ printk("push items %d nr %u\n", push_items, right_nritems);
+ WARN_ON(1);
+ }
+
+ if (push_items < right_nritems) {
+ push_space = btrfs_item_offset_nr(right, push_items - 1) -
+ leaf_data_end(root, right);
+ memmove_extent_buffer(right, btrfs_leaf_data(right) +
+ BTRFS_LEAF_DATA_SIZE(root) - push_space,
+ btrfs_leaf_data(right) +
+ leaf_data_end(root, right), push_space);
+
+ memmove_extent_buffer(right, btrfs_item_nr_offset(0),
btrfs_item_nr_offset(push_items),
(btrfs_header_nritems(right) - push_items) *
sizeof(struct btrfs_item));
-
- right_nritems = btrfs_header_nritems(right) - push_items;
+ }
+ right_nritems -= push_items;
btrfs_set_header_nritems(right, right_nritems);
push_space = BTRFS_LEAF_DATA_SIZE(root);
-
for (i = 0; i < right_nritems; i++) {
item = btrfs_item_nr(right, i);
}
btrfs_mark_buffer_dirty(left);
- btrfs_mark_buffer_dirty(right);
+ if (right_nritems)
+ btrfs_mark_buffer_dirty(right);
btrfs_item_key(right, &disk_key, 0);
wret = fixup_low_keys(trans, root, path, &disk_key, 1);
/* then fixup the leaf pointer in the path */
if (path->slots[0] < push_items) {
path->slots[0] += old_left_nritems;
+ if (btrfs_header_nritems(path->nodes[0]) == 0)
+ clean_tree_block(trans, root, path->nodes[0]);
+ btrfs_tree_unlock(path->nodes[0]);
free_extent_buffer(path->nodes[0]);
path->nodes[0] = left;
path->slots[1] -= 1;
} else {
+ btrfs_tree_unlock(left);
free_extent_buffer(left);
path->slots[0] -= push_items;
}
BUG_ON(path->slots[0] < 0);
return ret;
+out:
+ btrfs_tree_unlock(left);
+ free_extent_buffer(left);
+ return ret;
}
/*
*/
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)
+ struct btrfs_path *path, int data_size, int extend)
{
+ u64 root_gen;
struct extent_buffer *l;
u32 nritems;
int mid;
int i;
int ret = 0;
int wret;
- int double_split = 0;
+ int double_split;
+ int num_doubles = 0;
struct btrfs_disk_key disk_key;
+ if (extend)
+ space_needed = data_size;
+
+ if (root->ref_cows)
+ root_gen = trans->transid;
+ else
+ root_gen = 0;
+
/* first try to make some room by pushing left and right */
- wret = push_leaf_left(trans, root, path, data_size);
- if (wret < 0) {
- return wret;
- }
- if (wret) {
- wret = push_leaf_right(trans, root, path, data_size);
- if (wret < 0)
+ if (ins_key->type != BTRFS_DIR_ITEM_KEY) {
+ wret = push_leaf_right(trans, root, path, data_size, 0);
+ if (wret < 0) {
return wret;
- }
- l = path->nodes[0];
+ }
+ if (wret) {
+ wret = push_leaf_left(trans, root, path, data_size, 0);
+ if (wret < 0)
+ return wret;
+ }
+ l = path->nodes[0];
- /* did the pushes work? */
- if (btrfs_leaf_free_space(root, l) >=
- sizeof(struct btrfs_item) + data_size) {
- return 0;
+ /* did the pushes work? */
+ if (btrfs_leaf_free_space(root, l) >= space_needed)
+ return 0;
}
if (!path->nodes[1]) {
if (ret)
return ret;
}
+again:
+ double_split = 0;
+ l = path->nodes[0];
slot = path->slots[0];
nritems = btrfs_header_nritems(l);
mid = (nritems + 1)/ 2;
+ btrfs_item_key(l, &disk_key, 0);
+
right = btrfs_alloc_free_block(trans, root, root->leafsize,
- l->start, 0);
- if (IS_ERR(right))
+ root->root_key.objectid,
+ root_gen, disk_key.objectid, 0,
+ l->start, 0);
+ if (IS_ERR(right)) {
+ BUG_ON(1);
return PTR_ERR(right);
+ }
memset_extent_buffer(right, 0, 0, sizeof(struct btrfs_header));
btrfs_set_header_bytenr(right, right->start);
(unsigned long)btrfs_header_fsid(right),
BTRFS_FSID_SIZE);
+ write_extent_buffer(right, root->fs_info->chunk_tree_uuid,
+ (unsigned long)btrfs_header_chunk_tree_uuid(right),
+ BTRFS_UUID_SIZE);
if (mid <= slot) {
if (nritems == 1 ||
leaf_space_used(l, mid, nritems - mid) + space_needed >
path->slots[1] + 1, 1);
if (wret)
ret = wret;
+
+ btrfs_tree_unlock(path->nodes[0]);
free_extent_buffer(path->nodes[0]);
path->nodes[0] = right;
path->slots[0] = 0;
path->slots[1] += 1;
+ btrfs_mark_buffer_dirty(right);
return ret;
}
mid = slot;
} else {
if (leaf_space_used(l, 0, mid + 1) + space_needed >
BTRFS_LEAF_DATA_SIZE(root)) {
- if (slot == 0) {
+ if (!extend && slot == 0) {
btrfs_cpu_key_to_disk(&disk_key, ins_key);
btrfs_set_header_nritems(right, 0);
wret = insert_ptr(trans, root, path,
path->slots[1], 1);
if (wret)
ret = wret;
+ btrfs_tree_unlock(path->nodes[0]);
free_extent_buffer(path->nodes[0]);
path->nodes[0] = right;
path->slots[0] = 0;
if (wret)
ret = wret;
}
+ btrfs_mark_buffer_dirty(right);
return ret;
+ } else if (extend && slot == 0) {
+ mid = 1;
+ } else {
+ mid = slot;
+ if (mid != nritems &&
+ leaf_space_used(l, mid, nritems - mid) +
+ space_needed > BTRFS_LEAF_DATA_SIZE(root)) {
+ double_split = 1;
+ }
}
- mid = slot;
- double_split = 1;
}
}
nritems = nritems - mid;
BUG_ON(path->slots[0] != slot);
if (mid <= slot) {
+ btrfs_tree_unlock(path->nodes[0]);
free_extent_buffer(path->nodes[0]);
path->nodes[0] = right;
path->slots[0] -= mid;
path->slots[1] += 1;
- } else
+ } else {
+ btrfs_tree_unlock(right);
free_extent_buffer(right);
-
- BUG_ON(path->slots[0] < 0);
-
- if (!double_split) {
- return ret;
}
- right = btrfs_alloc_free_block(trans, root, root->leafsize,
- l->start, 0);
- if (IS_ERR(right))
- return PTR_ERR(right);
-
- memset_extent_buffer(right, 0, 0, sizeof(struct btrfs_header));
- btrfs_set_header_bytenr(right, right->start);
- btrfs_set_header_generation(right, trans->transid);
- btrfs_set_header_owner(right, root->root_key.objectid);
- btrfs_set_header_level(right, 0);
- write_extent_buffer(right, root->fs_info->fsid,
- (unsigned long)btrfs_header_fsid(right),
- BTRFS_FSID_SIZE);
+ BUG_ON(path->slots[0] < 0);
- btrfs_cpu_key_to_disk(&disk_key, ins_key);
- btrfs_set_header_nritems(right, 0);
- wret = insert_ptr(trans, root, path,
- &disk_key, right->start,
- path->slots[1], 1);
- if (wret)
- ret = wret;
- if (path->slots[1] == 0) {
- wret = fixup_low_keys(trans, root, path, &disk_key, 1);
- if (wret)
- ret = wret;
+ if (double_split) {
+ BUG_ON(num_doubles != 0);
+ num_doubles++;
+ goto again;
}
- free_extent_buffer(path->nodes[0]);
- path->nodes[0] = right;
- path->slots[0] = 0;
return ret;
}
int btrfs_truncate_item(struct btrfs_trans_handle *trans,
struct btrfs_root *root,
struct btrfs_path *path,
- u32 new_size)
+ u32 new_size, int from_end)
{
int ret = 0;
int slot;
slot_orig = path->slots[0];
leaf = path->nodes[0];
+ slot = path->slots[0];
+
+ old_size = btrfs_item_size_nr(leaf, slot);
+ if (old_size == new_size)
+ return 0;
nritems = btrfs_header_nritems(leaf);
data_end = leaf_data_end(root, leaf);
- slot = path->slots[0];
old_data_start = btrfs_item_offset_nr(leaf, slot);
- old_size = btrfs_item_size_nr(leaf, slot);
- BUG_ON(old_size <= new_size);
+
size_diff = old_size - new_size;
BUG_ON(slot < 0);
}
/* shift the data */
- memmove_extent_buffer(leaf, btrfs_leaf_data(leaf) +
- data_end + size_diff, btrfs_leaf_data(leaf) +
- data_end, old_data_start + new_size - data_end);
+ if (from_end) {
+ memmove_extent_buffer(leaf, btrfs_leaf_data(leaf) +
+ data_end + size_diff, btrfs_leaf_data(leaf) +
+ data_end, old_data_start + new_size - data_end);
+ } else {
+ struct btrfs_disk_key disk_key;
+ u64 offset;
+
+ btrfs_item_key(leaf, &disk_key, slot);
+
+ if (btrfs_disk_key_type(&disk_key) == BTRFS_EXTENT_DATA_KEY) {
+ unsigned long ptr;
+ struct btrfs_file_extent_item *fi;
+
+ fi = btrfs_item_ptr(leaf, slot,
+ struct btrfs_file_extent_item);
+ fi = (struct btrfs_file_extent_item *)(
+ (unsigned long)fi - size_diff);
+
+ if (btrfs_file_extent_type(leaf, fi) ==
+ BTRFS_FILE_EXTENT_INLINE) {
+ ptr = btrfs_item_ptr_offset(leaf, slot);
+ memmove_extent_buffer(leaf, ptr,
+ (unsigned long)fi,
+ offsetof(struct btrfs_file_extent_item,
+ disk_bytenr));
+ }
+ }
+
+ memmove_extent_buffer(leaf, btrfs_leaf_data(leaf) +
+ data_end + size_diff, btrfs_leaf_data(leaf) +
+ data_end, old_data_start - data_end);
+
+ offset = btrfs_disk_key_offset(&disk_key);
+ btrfs_set_disk_key_offset(&disk_key, offset + size_diff);
+ btrfs_set_item_key(leaf, &disk_key, slot);
+ if (slot == 0)
+ fixup_low_keys(trans, root, path, &disk_key, 1);
+ }
item = btrfs_item_nr(leaf, slot);
btrfs_set_item_size(leaf, item, new_size);
* Given a key and some data, insert an item into the tree.
* This does all the path init required, making room in the tree if needed.
*/
-int btrfs_insert_empty_item(struct btrfs_trans_handle *trans,
+int btrfs_insert_empty_items(struct btrfs_trans_handle *trans,
struct btrfs_root *root,
struct btrfs_path *path,
- struct btrfs_key *cpu_key, u32 data_size)
+ struct btrfs_key *cpu_key, u32 *data_size,
+ int nr)
{
struct extent_buffer *leaf;
struct btrfs_item *item;
int ret = 0;
int slot;
int slot_orig;
+ int i;
u32 nritems;
+ u32 total_size = 0;
+ u32 total_data = 0;
unsigned int data_end;
struct btrfs_disk_key disk_key;
- btrfs_cpu_key_to_disk(&disk_key, cpu_key);
-
- /* create a root if there isn't one */
- if (!root->node)
- BUG();
+ for (i = 0; i < nr; i++) {
+ total_data += data_size[i];
+ }
- ret = btrfs_search_slot(trans, root, cpu_key, path, data_size, 1);
+ total_size = total_data + (nr - 1) * sizeof(struct btrfs_item);
+ ret = btrfs_search_slot(trans, root, cpu_key, path, total_size, 1);
if (ret == 0) {
return -EEXIST;
}
data_end = leaf_data_end(root, leaf);
if (btrfs_leaf_free_space(root, leaf) <
- sizeof(struct btrfs_item) + data_size) {
+ sizeof(struct btrfs_item) + total_size) {
btrfs_print_leaf(root, leaf);
printk("not enough freespace need %u have %d\n",
- data_size, btrfs_leaf_free_space(root, leaf));
+ total_size, btrfs_leaf_free_space(root, leaf));
BUG();
}
}
ioff = btrfs_item_offset(leaf, item);
- btrfs_set_item_offset(leaf, item, ioff - data_size);
+ btrfs_set_item_offset(leaf, item, ioff - total_data);
}
if (leaf->map_token) {
unmap_extent_buffer(leaf, leaf->map_token, KM_USER1);
}
/* shift the items */
- memmove_extent_buffer(leaf, btrfs_item_nr_offset(slot + 1),
+ memmove_extent_buffer(leaf, btrfs_item_nr_offset(slot + nr),
btrfs_item_nr_offset(slot),
(nritems - slot) * sizeof(struct btrfs_item));
/* shift the data */
memmove_extent_buffer(leaf, btrfs_leaf_data(leaf) +
- data_end - data_size, btrfs_leaf_data(leaf) +
+ data_end - total_data, btrfs_leaf_data(leaf) +
data_end, old_data - data_end);
data_end = old_data;
}
/* setup the item for the new data */
- btrfs_set_item_key(leaf, &disk_key, slot);
- item = btrfs_item_nr(leaf, slot);
- btrfs_set_item_offset(leaf, item, data_end - data_size);
- btrfs_set_item_size(leaf, item, data_size);
- btrfs_set_header_nritems(leaf, nritems + 1);
+ for (i = 0; i < nr; i++) {
+ btrfs_cpu_key_to_disk(&disk_key, cpu_key + i);
+ btrfs_set_item_key(leaf, &disk_key, slot + i);
+ item = btrfs_item_nr(leaf, slot + i);
+ btrfs_set_item_offset(leaf, item, data_end - data_size[i]);
+ data_end -= data_size[i];
+ btrfs_set_item_size(leaf, item, data_size[i]);
+ }
+ btrfs_set_header_nritems(leaf, nritems + nr);
btrfs_mark_buffer_dirty(leaf);
ret = 0;
- if (slot == 0)
+ if (slot == 0) {
+ btrfs_cpu_key_to_disk(&disk_key, cpu_key);
ret = fixup_low_keys(trans, root, path, &disk_key, 1);
+ }
if (btrfs_leaf_free_space(root, leaf) < 0) {
btrfs_print_leaf(root, leaf);
* delete the item at the leaf level in path. If that empties
* the leaf, remove it from the tree
*/
-int btrfs_del_item(struct btrfs_trans_handle *trans, struct btrfs_root *root,
- struct btrfs_path *path)
+int btrfs_del_items(struct btrfs_trans_handle *trans, struct btrfs_root *root,
+ struct btrfs_path *path, int slot, int nr)
{
- int slot;
struct extent_buffer *leaf;
struct btrfs_item *item;
- int doff;
- int dsize;
+ int last_off;
+ int dsize = 0;
int ret = 0;
int wret;
+ int i;
u32 nritems;
leaf = path->nodes[0];
- slot = path->slots[0];
- doff = btrfs_item_offset_nr(leaf, slot);
- dsize = btrfs_item_size_nr(leaf, slot);
+ last_off = btrfs_item_offset_nr(leaf, slot + nr - 1);
+
+ for (i = 0; i < nr; i++)
+ dsize += btrfs_item_size_nr(leaf, slot + i);
+
nritems = btrfs_header_nritems(leaf);
- if (slot != nritems - 1) {
+ if (slot + nr != nritems) {
int i;
int data_end = leaf_data_end(root, leaf);
memmove_extent_buffer(leaf, btrfs_leaf_data(leaf) +
data_end + dsize,
btrfs_leaf_data(leaf) + data_end,
- doff - data_end);
+ last_off - data_end);
- for (i = slot + 1; i < nritems; i++) {
+ for (i = slot + nr; i < nritems; i++) {
u32 ioff;
item = btrfs_item_nr(leaf, i);
}
memmove_extent_buffer(leaf, btrfs_item_nr_offset(slot),
- btrfs_item_nr_offset(slot + 1),
+ btrfs_item_nr_offset(slot + nr),
sizeof(struct btrfs_item) *
- (nritems - slot - 1));
+ (nritems - slot - nr));
}
- btrfs_set_header_nritems(leaf, nritems - 1);
- nritems--;
+ btrfs_set_header_nritems(leaf, nritems - nr);
+ nritems -= nr;
/* delete the leaf if we've emptied it */
if (nritems == 0) {
if (leaf == root->node) {
btrfs_set_header_level(leaf, 0);
} else {
- clean_tree_block(trans, root, leaf);
- wait_on_tree_block_writeback(root, leaf);
+ u64 root_gen = btrfs_header_generation(path->nodes[1]);
wret = del_ptr(trans, root, path, 1, path->slots[1]);
if (wret)
ret = wret;
wret = btrfs_free_extent(trans, root,
- leaf->start, leaf->len, 1);
+ leaf->start, leaf->len,
+ btrfs_header_owner(path->nodes[1]),
+ root_gen, 0, 0, 1);
if (wret)
ret = wret;
}
}
/* delete the leaf if it is mostly empty */
- if (0 && used < BTRFS_LEAF_DATA_SIZE(root) / 3) {
+ if (used < BTRFS_LEAF_DATA_SIZE(root) / 4) {
/* push_leaf_left fixes the path.
* make sure the path still points to our leaf
* for possible call to del_ptr below
slot = path->slots[1];
extent_buffer_get(leaf);
- wret = push_leaf_left(trans, root, path, 1);
+ wret = push_leaf_left(trans, root, path, 1, 1);
if (wret < 0 && wret != -ENOSPC)
ret = wret;
if (path->nodes[0] == leaf &&
btrfs_header_nritems(leaf)) {
- wret = push_leaf_right(trans, root, path, 1);
+ wret = push_leaf_right(trans, root, path, 1, 1);
if (wret < 0 && wret != -ENOSPC)
ret = wret;
}
if (btrfs_header_nritems(leaf) == 0) {
+ u64 root_gen;
u64 bytenr = leaf->start;
u32 blocksize = leaf->len;
- clean_tree_block(trans, root, leaf);
- wait_on_tree_block_writeback(root, leaf);
+ root_gen = btrfs_header_generation(
+ path->nodes[1]);
wret = del_ptr(trans, root, path, 1, slot);
if (wret)
free_extent_buffer(leaf);
wret = btrfs_free_extent(trans, root, bytenr,
- blocksize, 1);
+ blocksize,
+ btrfs_header_owner(path->nodes[1]),
+ root_gen, 0, 0, 1);
if (wret)
ret = wret;
} else {
- btrfs_mark_buffer_dirty(leaf);
+ /* if we're still in the path, make sure
+ * we're dirty. Otherwise, one of the
+ * push_leaf functions must have already
+ * dirtied this buffer
+ */
+ if (path->nodes[0] == leaf)
+ btrfs_mark_buffer_dirty(leaf);
free_extent_buffer(leaf);
}
} else {
}
/*
- * walk up the tree as far as required to find the next leaf.
+ * search the tree again to find a leaf with lesser keys
+ * returns 0 if it found something or 1 if there are no lesser leaves.
+ * returns < 0 on io errors.
+ */
+int btrfs_prev_leaf(struct btrfs_root *root, struct btrfs_path *path)
+{
+ struct btrfs_key key;
+ struct btrfs_disk_key found_key;
+ int ret;
+
+ btrfs_item_key_to_cpu(path->nodes[0], &key, 0);
+
+ if (key.offset > 0)
+ key.offset--;
+ else if (key.type > 0)
+ key.type--;
+ else if (key.objectid > 0)
+ key.objectid--;
+ else
+ return 1;
+
+ btrfs_release_path(root, path);
+ ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
+ if (ret < 0)
+ return ret;
+ btrfs_item_key(path->nodes[0], &found_key, 0);
+ ret = comp_keys(&found_key, &key);
+ if (ret < 0)
+ return 0;
+ return 1;
+}
+
+/*
+ * search the tree again to find a leaf with greater keys
* returns 0 if it found something or 1 if there are no greater leaves.
* returns < 0 on io errors.
*/
{
int slot;
int level = 1;
- u64 bytenr;
struct extent_buffer *c;
struct extent_buffer *next = NULL;
+ struct btrfs_key key;
+ u32 nritems;
+ int ret;
+
+ nritems = btrfs_header_nritems(path->nodes[0]);
+ if (nritems == 0) {
+ return 1;
+ }
+
+ btrfs_item_key_to_cpu(path->nodes[0], &key, nritems - 1);
+
+ btrfs_release_path(root, path);
+ path->keep_locks = 1;
+ ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
+ path->keep_locks = 0;
+
+ if (ret < 0)
+ return ret;
+
+ nritems = btrfs_header_nritems(path->nodes[0]);
+ if (nritems > 0 && path->slots[0] < nritems - 1) {
+ goto done;
+ }
while(level < BTRFS_MAX_LEVEL) {
if (!path->nodes[level])
c = path->nodes[level];
if (slot >= btrfs_header_nritems(c)) {
level++;
+ if (level == BTRFS_MAX_LEVEL) {
+ return 1;
+ }
continue;
}
- bytenr = btrfs_node_blockptr(c, slot);
- if (next)
+ if (next) {
+ btrfs_tree_unlock(next);
free_extent_buffer(next);
+ }
- if (path->reada)
- reada_for_search(root, path, level, slot);
-
- next = read_tree_block(root, bytenr,
- btrfs_level_size(root, level -1));
+ if (level == 1 && path->locks[1] && path->reada)
+ reada_for_search(root, path, level, slot, 0);
+
+ next = read_node_slot(root, c, slot);
+ if (!path->skip_locking) {
+ if (!btrfs_tree_locked(c)) {
+ int i;
+ WARN_ON(1);
+printk("path %p no lock on level %d\n", path, level);
+for (i = 0; i < BTRFS_MAX_LEVEL; i++) {
+printk("path %p level %d slot %d nritems %d\n", path, i, path->slots[i], btrfs_header_nritems(path->nodes[i]));
+}
+ }
+ btrfs_tree_lock(next);
+ }
break;
}
path->slots[level] = slot;
while(1) {
level--;
c = path->nodes[level];
+ if (path->locks[level])
+ btrfs_tree_unlock(c);
free_extent_buffer(c);
path->nodes[level] = next;
path->slots[level] = 0;
+ path->locks[level] = 1;
if (!level)
break;
- if (path->reada)
- reada_for_search(root, path, level, 0);
- next = read_tree_block(root, btrfs_node_blockptr(next, 0),
- btrfs_level_size(root, level - 1));
+ if (level == 1 && path->locks[1] && path->reada)
+ reada_for_search(root, path, level, slot, 0);
+ next = read_node_slot(root, next, 0);
+ if (!path->skip_locking) {
+ WARN_ON(!btrfs_tree_locked(path->nodes[level]));
+ btrfs_tree_lock(next);
+ }
}
+done:
+ unlock_up(path, 0, 1);
return 0;
}
+
+int btrfs_previous_item(struct btrfs_root *root,
+ struct btrfs_path *path, u64 min_objectid,
+ int type)
+{
+ struct btrfs_key found_key;
+ struct extent_buffer *leaf;
+ int ret;
+
+ while(1) {
+ if (path->slots[0] == 0) {
+ ret = btrfs_prev_leaf(root, path);
+ if (ret != 0)
+ return ret;
+ } else {
+ path->slots[0]--;
+ }
+ leaf = path->nodes[0];
+ btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]);
+ if (found_key.type == type)
+ return 0;
+ }
+ return 1;
+}
+
projects / linux-2.6-block.git / blobdiff