-#include <linux/module.h>
+/*
+ * Copyright (C) 2007 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 "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);
-static int push_node_left(struct btrfs_trans_handle *trans, struct btrfs_root
- *root, struct buffer_head *dst, struct buffer_head
- *src);
-static int balance_node_right(struct btrfs_trans_handle *trans, struct
- btrfs_root *root, struct buffer_head *dst_buf,
- struct buffer_head *src_buf);
+ 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, int empty);
+static int balance_node_right(struct btrfs_trans_handle *trans,
+ struct btrfs_root *root,
+ struct extent_buffer *dst_buf,
+ struct extent_buffer *src_buf);
static int del_ptr(struct btrfs_trans_handle *trans, struct btrfs_root *root,
struct btrfs_path *path, int level, int slot);
{
struct btrfs_path *path;
path = kmem_cache_alloc(btrfs_path_cachep, GFP_NOFS);
- if (path)
+ if (path) {
btrfs_init_path(path);
+ path->reada = 1;
+ }
return path;
}
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;
+
for (i = 0; i < BTRFS_MAX_LEVEL; i++) {
+ p->slots[i] = 0;
if (!p->nodes[i])
+ continue;
+ if (p->locks[i]) {
+ btrfs_tree_unlock(p->nodes[i]);
+ p->locks[i] = 0;
+ }
+ free_extent_buffer(p->nodes[i]);
+ p->nodes[i] = NULL;
+ }
+}
+
+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;
- btrfs_block_release(root, p->nodes[i]);
+ }
+ spin_unlock(&root->node_lock);
+
+ btrfs_tree_unlock(eb);
+ free_extent_buffer(eb);
}
- memset(p, 0, sizeof(*p));
+ return eb;
}
-static int btrfs_cow_block(struct btrfs_trans_handle *trans, struct btrfs_root
- *root, struct buffer_head *buf, struct buffer_head
- *parent, int parent_slot, struct buffer_head
- **cow_ret)
+static void add_root_to_dirty_list(struct btrfs_root *root)
{
- struct buffer_head *cow;
- struct btrfs_node *cow_node;
+ if (root->track_dirty && list_empty(&root->dirty_list)) {
+ list_add(&root->dirty_list,
+ &root->fs_info->dirty_cowonly_roots);
+ }
+}
- if (btrfs_header_generation(btrfs_buffer_header(buf)) ==
- trans->transid) {
- *cow_ret = buf;
- return 0;
+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, root);
- cow_node = btrfs_buffer_node(cow);
- if (buf->b_size != root->blocksize || cow->b_size != root->blocksize)
- WARN_ON(1);
- memcpy(cow_node, btrfs_buffer_node(buf), root->blocksize);
- btrfs_set_header_blocknr(&cow_node->header, bh_blocknr(cow));
- btrfs_set_header_generation(&cow_node->header, trans->transid);
- btrfs_inc_ref(trans, root, buf);
+ 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, 0);
+ kfree(new_root);
+
+ if (ret)
+ return ret;
+
+ btrfs_mark_buffer_dirty(cow);
+ *cow_ret = cow;
+ return 0;
+}
+
+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 search_start, u64 empty_size,
+ u64 prealloc_dest)
+{
+ 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;
+ }
+ if (prealloc_dest) {
+ struct btrfs_key ins;
+
+ ins.objectid = prealloc_dest;
+ ins.offset = buf->len;
+ ins.type = BTRFS_EXTENT_ITEM_KEY;
+
+ ret = btrfs_alloc_reserved_extent(trans, root,
+ root->root_key.objectid,
+ root_gen, level,
+ first_key.objectid,
+ &ins);
+ BUG_ON(ret);
+ cow = btrfs_init_new_buffer(trans, root, prealloc_dest,
+ buf->len);
+ } else {
+ 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);
+
+ 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, 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) {
+ different_trans = 1;
+ ret = btrfs_inc_ref(trans, root, buf, 1);
+ if (ret)
+ return ret;
+ } else {
+ clean_tree_block(trans, root, buf);
+ }
+
if (buf == root->node) {
+ WARN_ON(parent && parent != buf);
+ root_gen = btrfs_header_generation(buf);
+
+ spin_lock(&root->node_lock);
root->node = cow;
- get_bh(cow);
+ extent_buffer_get(cow);
+ spin_unlock(&root->node_lock);
+
if (buf != root->commit_root) {
- btrfs_free_extent(trans, root, bh_blocknr(buf), 1, 1);
+ btrfs_free_extent(trans, root, buf->start,
+ buf->len, root->root_key.objectid,
+ root_gen, 0, 0, 1);
}
- btrfs_block_release(root, buf);
+ free_extent_buffer(buf);
+ add_root_to_dirty_list(root);
} else {
- btrfs_set_node_blockptr(btrfs_buffer_node(parent), parent_slot,
- bh_blocknr(cow));
+ 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);
- btrfs_free_extent(trans, root, bh_blocknr(buf), 1, 1);
+ WARN_ON(btrfs_header_generation(parent) != trans->transid);
+ btrfs_free_extent(trans, root, buf->start, buf->len,
+ btrfs_header_owner(parent), root_gen,
+ 0, 0, 1);
}
- btrfs_block_release(root, buf);
- mark_buffer_dirty(cow);
+ if (unlock_orig)
+ btrfs_tree_unlock(buf);
+ free_extent_buffer(buf);
+ btrfs_mark_buffer_dirty(cow);
*cow_ret = cow;
return 0;
}
-/*
- * The leaf data grows from end-to-front in the node.
- * this returns the address of the start of the last item,
- * which is the stop of the leaf data stack
- */
-static inline unsigned int leaf_data_end(struct btrfs_root *root,
- struct btrfs_leaf *leaf)
+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)
{
- u32 nr = btrfs_header_nritems(&leaf->header);
- if (nr == 0)
- return BTRFS_LEAF_DATA_SIZE(root);
- return btrfs_item_offset(leaf->items + nr - 1);
+ 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);
+ WARN_ON(1);
+ }
+ if (trans->transid != root->fs_info->generation) {
+ printk(KERN_CRIT "trans %Lu running %Lu\n", trans->transid,
+ root->fs_info->generation);
+ WARN_ON(1);
+ }
+
+ 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);
+ WARN_ON(prealloc_dest);
+ return 0;
+ }
+ 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,
+ prealloc_dest);
+ return ret;
+}
+
+static int close_blocks(u64 blocknr, u64 other, u32 blocksize)
+{
+ if (blocknr < other && other - (blocknr + blocksize) < 32768)
+ return 1;
+ if (blocknr > other && blocknr - (other + blocksize) < 32768)
+ return 1;
+ return 0;
}
/*
return 1;
if (k1.objectid < k2->objectid)
return -1;
- if (k1.flags > k2->flags)
+ if (k1.type > k2->type)
return 1;
- if (k1.flags < k2->flags)
+ if (k1.type < k2->type)
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 btrfs_key *progress)
+{
+ struct extent_buffer *cur;
+ u64 blocknr;
+ u64 gen;
+ u64 search_start = *last_ret;
+ u64 last_block = 0;
+ u64 other;
+ u32 parent_nritems;
+ int end_slot;
+ int i;
+ int err = 0;
+ int parent_level;
+ int uptodate;
+ u32 blocksize;
+ int progress_passed = 0;
+ struct btrfs_disk_key disk_key;
+
+ 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->running_transaction->transid);
+ WARN_ON(1);
+ }
+ if (trans->transid != root->fs_info->generation) {
+ printk(KERN_CRIT "trans %Lu running %Lu\n", trans->transid,
+ root->fs_info->generation);
+ WARN_ON(1);
+ }
+
+ parent_nritems = btrfs_header_nritems(parent);
+ blocksize = btrfs_level_size(root, parent_level - 1);
+ end_slot = parent_nritems;
+
+ if (parent_nritems == 1)
+ 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 - 2) {
+ other = btrfs_node_blockptr(parent, i + 1);
+ close = close_blocks(blocknr, other, blocksize);
+ }
+ if (close) {
+ 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, gen);
+ else
+ uptodate = 0;
+ if (!cur || !uptodate) {
+ if (cache_only) {
+ free_extent_buffer(cur);
+ continue;
+ }
+ if (!cur) {
+ cur = read_tree_block(root, blocknr,
+ blocksize, gen);
+ } else if (!uptodate) {
+ btrfs_read_buffer(cur, gen);
+ }
+ }
+ if (search_start == 0)
+ search_start = last_block;
+
+ btrfs_tree_lock(cur);
+ err = __btrfs_cow_block(trans, root, cur, parent, i,
+ &cur, search_start,
+ min(16 * blocksize,
+ (end_slot - i) * blocksize), 0);
+ if (err) {
+ btrfs_tree_unlock(cur);
+ free_extent_buffer(cur);
+ break;
+ }
+ search_start = cur->start;
+ last_block = cur->start;
+ *last_ret = search_start;
+ btrfs_tree_unlock(cur);
+ free_extent_buffer(cur);
+ }
+ if (parent->map_token) {
+ unmap_extent_buffer(parent, parent->map_token,
+ KM_USER1);
+ parent->map_token = NULL;
+ }
+ return err;
+}
+
+/*
+ * The leaf data grows from end-to-front in the node.
+ * this returns the address of the start of the last item,
+ * which is the stop of the leaf data stack
+ */
+static inline unsigned int leaf_data_end(struct btrfs_root *root,
+ struct extent_buffer *leaf)
+{
+ u32 nr = btrfs_header_nritems(leaf);
+ if (nr == 0)
+ return BTRFS_LEAF_DATA_SIZE(root);
+ return btrfs_item_offset_nr(leaf, nr - 1);
+}
+
static int check_node(struct btrfs_root *root, struct btrfs_path *path,
int level)
{
- int i;
- struct btrfs_node *parent = NULL;
- struct btrfs_node *node = btrfs_buffer_node(path->nodes[level]);
+ struct extent_buffer *parent = NULL;
+ struct extent_buffer *node = path->nodes[level];
+ struct btrfs_disk_key parent_key;
+ struct btrfs_disk_key node_key;
int parent_slot;
- u32 nritems = btrfs_header_nritems(&node->header);
+ int slot;
+ struct btrfs_key cpukey;
+ u32 nritems = btrfs_header_nritems(node);
if (path->nodes[level + 1])
- parent = btrfs_buffer_node(path->nodes[level + 1]);
- parent_slot = path->slots[level + 1];
+ parent = path->nodes[level + 1];
+
+ slot = path->slots[level];
BUG_ON(nritems == 0);
if (parent) {
- struct btrfs_disk_key *parent_key;
- parent_key = &parent->ptrs[parent_slot].key;
- BUG_ON(memcmp(parent_key, &node->ptrs[0].key,
+ parent_slot = path->slots[level + 1];
+ btrfs_node_key(parent, &parent_key, parent_slot);
+ btrfs_node_key(node, &node_key, 0);
+ BUG_ON(memcmp(&parent_key, &node_key,
sizeof(struct btrfs_disk_key)));
BUG_ON(btrfs_node_blockptr(parent, parent_slot) !=
- btrfs_header_blocknr(&node->header));
+ btrfs_header_bytenr(node));
}
BUG_ON(nritems > BTRFS_NODEPTRS_PER_BLOCK(root));
- for (i = 0; nritems > 1 && i < nritems - 2; i++) {
- struct btrfs_key cpukey;
- btrfs_disk_key_to_cpu(&cpukey, &node->ptrs[i + 1].key);
- BUG_ON(comp_keys(&node->ptrs[i].key, &cpukey) >= 0);
+ if (slot != 0) {
+ btrfs_node_key_to_cpu(node, &cpukey, slot - 1);
+ btrfs_node_key(node, &node_key, slot);
+ BUG_ON(comp_keys(&node_key, &cpukey) <= 0);
+ }
+ if (slot < nritems - 1) {
+ btrfs_node_key_to_cpu(node, &cpukey, slot + 1);
+ btrfs_node_key(node, &node_key, slot);
+ BUG_ON(comp_keys(&node_key, &cpukey) >= 0);
}
return 0;
}
static int check_leaf(struct btrfs_root *root, struct btrfs_path *path,
int level)
{
- int i;
- struct btrfs_leaf *leaf = btrfs_buffer_leaf(path->nodes[level]);
- struct btrfs_node *parent = NULL;
+ struct extent_buffer *leaf = path->nodes[level];
+ struct extent_buffer *parent = NULL;
int parent_slot;
- u32 nritems = btrfs_header_nritems(&leaf->header);
+ struct btrfs_key cpukey;
+ struct btrfs_disk_key parent_key;
+ struct btrfs_disk_key leaf_key;
+ int slot = path->slots[0];
+
+ u32 nritems = btrfs_header_nritems(leaf);
if (path->nodes[level + 1])
- parent = btrfs_buffer_node(path->nodes[level + 1]);
- parent_slot = path->slots[level + 1];
- BUG_ON(btrfs_leaf_free_space(root, leaf) < 0);
+ parent = path->nodes[level + 1];
if (nritems == 0)
return 0;
if (parent) {
- struct btrfs_disk_key *parent_key;
- parent_key = &parent->ptrs[parent_slot].key;
- BUG_ON(memcmp(parent_key, &leaf->items[0].key,
+ parent_slot = path->slots[level + 1];
+ btrfs_node_key(parent, &parent_key, parent_slot);
+ btrfs_item_key(leaf, &leaf_key, 0);
+
+ BUG_ON(memcmp(&parent_key, &leaf_key,
sizeof(struct btrfs_disk_key)));
BUG_ON(btrfs_node_blockptr(parent, parent_slot) !=
- btrfs_header_blocknr(&leaf->header));
+ btrfs_header_bytenr(leaf));
}
+#if 0
for (i = 0; nritems > 1 && i < nritems - 2; i++) {
- struct btrfs_key cpukey;
- btrfs_disk_key_to_cpu(&cpukey, &leaf->items[i + 1].key);
- BUG_ON(comp_keys(&leaf->items[i].key,
- &cpukey) >= 0);
- BUG_ON(btrfs_item_offset(leaf->items + i) !=
- btrfs_item_end(leaf->items + i + 1));
+ btrfs_item_key_to_cpu(leaf, &cpukey, i + 1);
+ btrfs_item_key(leaf, &leaf_key, i);
+ if (comp_keys(&leaf_key, &cpukey) >= 0) {
+ btrfs_print_leaf(root, leaf);
+ printk("slot %d offset bad key\n", i);
+ BUG_ON(1);
+ }
+ if (btrfs_item_offset_nr(leaf, i) !=
+ btrfs_item_end_nr(leaf, i + 1)) {
+ btrfs_print_leaf(root, leaf);
+ printk("slot %d offset bad\n", i);
+ BUG_ON(1);
+ }
if (i == 0) {
- BUG_ON(btrfs_item_offset(leaf->items + i) +
- btrfs_item_size(leaf->items + i) !=
- BTRFS_LEAF_DATA_SIZE(root));
+ if (btrfs_item_offset_nr(leaf, i) +
+ btrfs_item_size_nr(leaf, i) !=
+ BTRFS_LEAF_DATA_SIZE(root)) {
+ btrfs_print_leaf(root, leaf);
+ printk("slot %d first offset bad\n", i);
+ BUG_ON(1);
+ }
+ }
+ }
+ if (nritems > 0) {
+ if (btrfs_item_size_nr(leaf, nritems - 1) > 4096) {
+ btrfs_print_leaf(root, leaf);
+ printk("slot %d bad size \n", nritems - 1);
+ BUG_ON(1);
+ }
+ }
+#endif
+ if (slot != 0 && slot < nritems - 1) {
+ btrfs_item_key(leaf, &leaf_key, slot);
+ btrfs_item_key_to_cpu(leaf, &cpukey, slot - 1);
+ if (comp_keys(&leaf_key, &cpukey) <= 0) {
+ btrfs_print_leaf(root, leaf);
+ printk("slot %d offset bad key\n", slot);
+ BUG_ON(1);
+ }
+ if (btrfs_item_offset_nr(leaf, slot - 1) !=
+ btrfs_item_end_nr(leaf, slot)) {
+ btrfs_print_leaf(root, leaf);
+ printk("slot %d offset bad\n", slot);
+ BUG_ON(1);
+ }
+ }
+ if (slot < nritems - 1) {
+ btrfs_item_key(leaf, &leaf_key, slot);
+ btrfs_item_key_to_cpu(leaf, &cpukey, slot + 1);
+ BUG_ON(comp_keys(&leaf_key, &cpukey) >= 0);
+ if (btrfs_item_offset_nr(leaf, slot) !=
+ btrfs_item_end_nr(leaf, slot + 1)) {
+ btrfs_print_leaf(root, leaf);
+ printk("slot %d offset bad\n", slot);
+ BUG_ON(1);
}
}
+ BUG_ON(btrfs_item_offset_nr(leaf, 0) +
+ btrfs_item_size_nr(leaf, 0) != BTRFS_LEAF_DATA_SIZE(root));
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)
{
- struct btrfs_node *node = btrfs_buffer_node(path->nodes[level]);
- if (memcmp(node->header.fsid, root->fs_info->disk_super->fsid,
- sizeof(node->header.fsid)))
- BUG();
+ 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];
+
+ if (memcmp_extent_buffer(buf, root->fs_info->fsid,
+ (unsigned long)btrfs_header_fsid(buf),
+ BTRFS_FSID_SIZE)) {
+ printk("warning bad block %Lu\n", buf->start);
+ return 1;
+ }
+#endif
if (level == 0)
return check_leaf(root, path, level);
return check_node(root, path, level);
}
/*
- * search for key in the array p. items p are item_size apart
- * and there are 'max' items in p
+ * search for key in the extent_buffer. The items start at offset p,
+ * and they are item_size apart. There are 'max' items in p.
+ *
* the slot in the array is returned via slot, and it points to
* the place where you would insert key if it is not found in
* the array.
*
* slot may point to max if the key is bigger than all of the keys
*/
-static int generic_bin_search(char *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;
int mid;
int ret;
- struct btrfs_disk_key *tmp;
+ struct btrfs_disk_key *tmp = NULL;
+ struct btrfs_disk_key unaligned;
+ unsigned long offset;
+ char *map_token = NULL;
+ char *kaddr = NULL;
+ unsigned long map_start = 0;
+ unsigned long map_len = 0;
+ int err;
while(low < high) {
mid = (low + high) / 2;
- tmp = (struct btrfs_disk_key *)(p + mid * item_size);
+ offset = p + mid * item_size;
+
+ if (!map_token || offset < map_start ||
+ (offset + sizeof(struct btrfs_disk_key)) >
+ map_start + map_len) {
+ if (map_token) {
+ unmap_extent_buffer(eb, map_token, KM_USER0);
+ map_token = NULL;
+ }
+ err = map_extent_buffer(eb, offset,
+ sizeof(struct btrfs_disk_key),
+ &map_token, &kaddr,
+ &map_start, &map_len, KM_USER0);
+
+ if (!err) {
+ tmp = (struct btrfs_disk_key *)(kaddr + offset -
+ map_start);
+ } else {
+ read_extent_buffer(eb, &unaligned,
+ offset, sizeof(unaligned));
+ tmp = &unaligned;
+ }
+
+ } else {
+ tmp = (struct btrfs_disk_key *)(kaddr + offset -
+ map_start);
+ }
ret = comp_keys(tmp, key);
if (ret < 0)
high = mid;
else {
*slot = mid;
+ if (map_token)
+ unmap_extent_buffer(eb, map_token, KM_USER0);
return 0;
}
}
*slot = low;
+ if (map_token)
+ unmap_extent_buffer(eb, map_token, KM_USER0);
return 1;
}
* simple bin_search frontend that does the right thing for
* leaves vs nodes
*/
-static int bin_search(struct btrfs_node *c, struct btrfs_key *key, int *slot)
+static int bin_search(struct extent_buffer *eb, struct btrfs_key *key,
+ int level, int *slot)
{
- if (btrfs_is_leaf(c)) {
- struct btrfs_leaf *l = (struct btrfs_leaf *)c;
- return generic_bin_search((void *)l->items,
+ if (level == 0) {
+ return generic_bin_search(eb,
+ offsetof(struct btrfs_leaf, items),
sizeof(struct btrfs_item),
- key, btrfs_header_nritems(&c->header),
+ key, btrfs_header_nritems(eb),
slot);
} else {
- return generic_bin_search((void *)c->ptrs,
+ return generic_bin_search(eb,
+ offsetof(struct btrfs_node, ptrs),
sizeof(struct btrfs_key_ptr),
- key, btrfs_header_nritems(&c->header),
+ key, btrfs_header_nritems(eb),
slot);
}
return -1;
}
-static struct buffer_head *read_node_slot(struct btrfs_root *root,
- struct buffer_head *parent_buf,
- int slot)
+static noinline struct extent_buffer *read_node_slot(struct btrfs_root *root,
+ struct extent_buffer *parent, int slot)
{
- struct btrfs_node *node = btrfs_buffer_node(parent_buf);
+ int level = btrfs_header_level(parent);
if (slot < 0)
return NULL;
- if (slot >= btrfs_header_nritems(&node->header))
+ if (slot >= btrfs_header_nritems(parent))
return NULL;
- return read_tree_block(root, btrfs_node_blockptr(node, slot));
+
+ BUG_ON(level == 0);
+
+ return read_tree_block(root, btrfs_node_blockptr(parent, slot),
+ 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 noinline int balance_level(struct btrfs_trans_handle *trans,
+ struct btrfs_root *root,
+ struct btrfs_path *path, int level)
{
- struct buffer_head *right_buf;
- struct buffer_head *mid_buf;
- struct buffer_head *left_buf;
- struct buffer_head *parent_buf = NULL;
- struct btrfs_node *right = NULL;
- struct btrfs_node *mid;
- struct btrfs_node *left = NULL;
- struct btrfs_node *parent = NULL;
+ struct extent_buffer *right = NULL;
+ struct extent_buffer *mid;
+ struct extent_buffer *left = NULL;
+ struct extent_buffer *parent = NULL;
int ret = 0;
int wret;
int pslot;
int orig_slot = path->slots[level];
+ int err_on_enospc = 0;
u64 orig_ptr;
if (level == 0)
return 0;
- mid_buf = path->nodes[level];
- mid = btrfs_buffer_node(mid_buf);
+ 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)
- parent_buf = path->nodes[level + 1];
+ parent = path->nodes[level + 1];
pslot = path->slots[level + 1];
/*
* deal with the case where there is only one pointer in the root
* by promoting the node below to a root
*/
- if (!parent_buf) {
- struct buffer_head *child;
- u64 blocknr = bh_blocknr(mid_buf);
+ if (!parent) {
+ struct extent_buffer *child;
- if (btrfs_header_nritems(&mid->header) != 1)
+ if (btrfs_header_nritems(mid) != 1)
return 0;
/* promote the child to a root */
- child = read_node_slot(root, mid_buf, 0);
+ child = read_node_slot(root, mid, 0);
+ btrfs_tree_lock(child);
BUG_ON(!child);
+ ret = btrfs_cow_block(trans, root, child, mid, 0, &child, 0);
+ 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_buf);
- wait_on_buffer(mid_buf);
+ clean_tree_block(trans, root, mid);
+ btrfs_tree_unlock(mid);
/* once for the path */
- btrfs_block_release(root, mid_buf);
+ free_extent_buffer(mid);
+ 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 */
- btrfs_block_release(root, mid_buf);
- return btrfs_free_extent(trans, root, blocknr, 1, 1);
+ free_extent_buffer(mid);
+ return ret;
}
- parent = btrfs_buffer_node(parent_buf);
-
- if (btrfs_header_nritems(&mid->header) >
+ if (btrfs_header_nritems(mid) >
BTRFS_NODEPTRS_PER_BLOCK(root) / 4)
return 0;
- left_buf = read_node_slot(root, parent_buf, pslot - 1);
- right_buf = read_node_slot(root, parent_buf, pslot + 1);
+ if (btrfs_header_nritems(mid) < 2)
+ err_on_enospc = 1;
+
+ 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, 0);
+ if (wret) {
+ ret = wret;
+ goto enospc;
+ }
+ }
+ 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, 0);
+ if (wret) {
+ ret = wret;
+ goto enospc;
+ }
+ }
/* first, try to make some room in the middle buffer */
- if (left_buf) {
- btrfs_cow_block(trans, root, left_buf, parent_buf, pslot - 1,
- &left_buf);
- left = btrfs_buffer_node(left_buf);
- orig_slot += btrfs_header_nritems(&left->header);
- wret = push_node_left(trans, root, left_buf, mid_buf);
+ if (left) {
+ orig_slot += btrfs_header_nritems(left);
+ wret = push_node_left(trans, root, left, mid, 1);
if (wret < 0)
ret = wret;
+ if (btrfs_header_nritems(mid) < 2)
+ err_on_enospc = 1;
}
/*
* then try to empty the right most buffer into the middle
*/
- if (right_buf) {
- btrfs_cow_block(trans, root, right_buf, parent_buf, pslot + 1,
- &right_buf);
- right = btrfs_buffer_node(right_buf);
- wret = push_node_left(trans, root, mid_buf, right_buf);
- if (wret < 0)
+ if (right) {
+ wret = push_node_left(trans, root, mid, right, 1);
+ if (wret < 0 && wret != -ENOSPC)
ret = wret;
- if (btrfs_header_nritems(&right->header) == 0) {
- u64 blocknr = bh_blocknr(right_buf);
- clean_tree_block(trans, root, right_buf);
- wait_on_buffer(right_buf);
- btrfs_block_release(root, right_buf);
- right_buf = NULL;
+ 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);
+ btrfs_tree_unlock(right);
+ free_extent_buffer(right);
right = NULL;
wret = del_ptr(trans, root, path, level + 1, pslot +
1);
if (wret)
ret = wret;
- wret = btrfs_free_extent(trans, root, blocknr, 1, 1);
+ wret = btrfs_free_extent(trans, root, bytenr,
+ blocksize,
+ btrfs_header_owner(parent),
+ generation, 0, 0, 1);
if (wret)
ret = wret;
} else {
- btrfs_memcpy(root, parent,
- &parent->ptrs[pslot + 1].key,
- &right->ptrs[0].key,
- sizeof(struct btrfs_disk_key));
- btrfs_mark_buffer_dirty(parent_buf);
+ struct btrfs_disk_key right_key;
+ btrfs_node_key(right, &right_key, 0);
+ btrfs_set_node_key(parent, &right_key, pslot + 1);
+ btrfs_mark_buffer_dirty(parent);
}
}
- if (btrfs_header_nritems(&mid->header) == 1) {
+ if (btrfs_header_nritems(mid) == 1) {
/*
* we're not allowed to leave a node with one item in the
* tree during a delete. A deletion from lower in the tree
* otherwise we would have pulled some pointers from the
* right
*/
- BUG_ON(!left_buf);
- wret = balance_node_right(trans, root, mid_buf, left_buf);
- if (wret < 0)
+ BUG_ON(!left);
+ wret = balance_node_right(trans, root, mid, left);
+ if (wret < 0) {
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->header) == 0) {
+ if (btrfs_header_nritems(mid) == 0) {
/* we've managed to empty the middle node, drop it */
- u64 blocknr = bh_blocknr(mid_buf);
- clean_tree_block(trans, root, mid_buf);
- wait_on_buffer(mid_buf);
- btrfs_block_release(root, mid_buf);
- mid_buf = NULL;
+ u64 root_gen = btrfs_header_generation(parent);
+ u64 bytenr = mid->start;
+ u32 blocksize = mid->len;
+
+ clean_tree_block(trans, 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, blocknr, 1, 1);
+ wret = btrfs_free_extent(trans, root, bytenr, blocksize,
+ btrfs_header_owner(parent),
+ root_gen, 0, 0, 1);
if (wret)
ret = wret;
} else {
/* update the parent key to reflect our changes */
- btrfs_memcpy(root, parent,
- &parent->ptrs[pslot].key, &mid->ptrs[0].key,
- sizeof(struct btrfs_disk_key));
- btrfs_mark_buffer_dirty(parent_buf);
+ struct btrfs_disk_key mid_key;
+ btrfs_node_key(mid, &mid_key, 0);
+ btrfs_set_node_key(parent, &mid_key, pslot);
+ btrfs_mark_buffer_dirty(parent);
}
/* update the path */
- if (left_buf) {
- if (btrfs_header_nritems(&left->header) > orig_slot) {
- get_bh(left_buf);
- path->nodes[level] = left_buf;
+ 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_buf)
- btrfs_block_release(root, mid_buf);
+ if (mid) {
+ btrfs_tree_unlock(mid);
+ free_extent_buffer(mid);
+ }
} else {
- orig_slot -= btrfs_header_nritems(&left->header);
+ orig_slot -= btrfs_header_nritems(left);
path->slots[level] = orig_slot;
}
}
/* double check we haven't messed things up */
check_block(root, path, level);
if (orig_ptr !=
- btrfs_node_blockptr(btrfs_buffer_node(path->nodes[level]),
- path->slots[level]))
+ btrfs_node_blockptr(path->nodes[level], path->slots[level]))
BUG();
-
- if (right_buf)
- btrfs_block_release(root, right_buf);
- if (left_buf)
- btrfs_block_release(root, left_buf);
+enospc:
+ if (right) {
+ btrfs_tree_unlock(right);
+ free_extent_buffer(right);
+ }
+ 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 buffer_head *right_buf;
- struct buffer_head *mid_buf;
- struct buffer_head *left_buf;
- struct buffer_head *parent_buf = NULL;
- struct btrfs_node *right = NULL;
- struct btrfs_node *mid;
- struct btrfs_node *left = NULL;
- struct btrfs_node *parent = NULL;
+ struct extent_buffer *right = NULL;
+ struct extent_buffer *mid;
+ struct extent_buffer *left = NULL;
+ struct extent_buffer *parent = NULL;
int ret = 0;
int wret;
int pslot;
if (level == 0)
return 1;
- mid_buf = path->nodes[level];
- mid = btrfs_buffer_node(mid_buf);
+ 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)
- parent_buf = path->nodes[level + 1];
+ parent = path->nodes[level + 1];
pslot = path->slots[level + 1];
- if (!parent_buf)
+ if (!parent)
return 1;
- parent = btrfs_buffer_node(parent_buf);
- left_buf = read_node_slot(root, parent_buf, pslot - 1);
+ left = read_node_slot(root, parent, pslot - 1);
/* first, try to make some room in the middle buffer */
- if (left_buf) {
+ if (left) {
u32 left_nr;
- left = btrfs_buffer_node(left_buf);
- left_nr = btrfs_header_nritems(&left->header);
+
+ btrfs_tree_lock(left);
+ left_nr = btrfs_header_nritems(left);
if (left_nr >= BTRFS_NODEPTRS_PER_BLOCK(root) - 1) {
wret = 1;
} else {
- btrfs_cow_block(trans, root, left_buf, parent_buf,
- pslot - 1, &left_buf);
- left = btrfs_buffer_node(left_buf);
- wret = push_node_left(trans, root, left_buf, mid_buf);
+ ret = btrfs_cow_block(trans, root, left, parent,
+ pslot - 1, &left, 0);
+ if (ret)
+ wret = 1;
+ else {
+ wret = push_node_left(trans, root,
+ left, mid, 0);
+ }
}
if (wret < 0)
ret = wret;
if (wret == 0) {
+ struct btrfs_disk_key disk_key;
orig_slot += left_nr;
- btrfs_memcpy(root, parent,
- &parent->ptrs[pslot].key,
- &mid->ptrs[0].key,
- sizeof(struct btrfs_disk_key));
- btrfs_mark_buffer_dirty(parent_buf);
- if (btrfs_header_nritems(&left->header) > orig_slot) {
- path->nodes[level] = left_buf;
+ btrfs_node_key(mid, &disk_key, 0);
+ btrfs_set_node_key(parent, &disk_key, pslot);
+ btrfs_mark_buffer_dirty(parent);
+ if (btrfs_header_nritems(left) > orig_slot) {
+ path->nodes[level] = left;
path->slots[level + 1] -= 1;
path->slots[level] = orig_slot;
- btrfs_block_release(root, mid_buf);
+ btrfs_tree_unlock(mid);
+ free_extent_buffer(mid);
} else {
orig_slot -=
- btrfs_header_nritems(&left->header);
+ btrfs_header_nritems(left);
path->slots[level] = orig_slot;
- btrfs_block_release(root, left_buf);
+ btrfs_tree_unlock(left);
+ free_extent_buffer(left);
}
- check_node(root, path, level);
return 0;
}
- btrfs_block_release(root, left_buf);
+ btrfs_tree_unlock(left);
+ free_extent_buffer(left);
}
- right_buf = read_node_slot(root, parent_buf, pslot + 1);
+ right = read_node_slot(root, parent, pslot + 1);
/*
* then try to empty the right most buffer into the middle
*/
- if (right_buf) {
+ if (right) {
u32 right_nr;
- right = btrfs_buffer_node(right_buf);
- right_nr = btrfs_header_nritems(&right->header);
+ btrfs_tree_lock(right);
+ right_nr = btrfs_header_nritems(right);
if (right_nr >= BTRFS_NODEPTRS_PER_BLOCK(root) - 1) {
wret = 1;
} else {
- btrfs_cow_block(trans, root, right_buf,
- parent_buf, pslot + 1, &right_buf);
- right = btrfs_buffer_node(right_buf);
- wret = balance_node_right(trans, root,
- right_buf, mid_buf);
+ ret = btrfs_cow_block(trans, root, right,
+ parent, pslot + 1,
+ &right, 0);
+ if (ret)
+ wret = 1;
+ else {
+ wret = balance_node_right(trans, root,
+ right, mid);
+ }
}
if (wret < 0)
ret = wret;
if (wret == 0) {
- btrfs_memcpy(root, parent,
- &parent->ptrs[pslot + 1].key,
- &right->ptrs[0].key,
- sizeof(struct btrfs_disk_key));
- btrfs_mark_buffer_dirty(parent_buf);
- if (btrfs_header_nritems(&mid->header) <= orig_slot) {
- path->nodes[level] = right_buf;
+ struct btrfs_disk_key disk_key;
+
+ btrfs_node_key(right, &disk_key, 0);
+ btrfs_set_node_key(parent, &disk_key, pslot + 1);
+ btrfs_mark_buffer_dirty(parent);
+
+ if (btrfs_header_nritems(mid) <= orig_slot) {
+ path->nodes[level] = right;
path->slots[level + 1] += 1;
path->slots[level] = orig_slot -
- btrfs_header_nritems(&mid->header);
- btrfs_block_release(root, mid_buf);
+ btrfs_header_nritems(mid);
+ btrfs_tree_unlock(mid);
+ free_extent_buffer(mid);
} else {
- btrfs_block_release(root, right_buf);
+ btrfs_tree_unlock(right);
+ free_extent_buffer(right);
}
- check_node(root, path, level);
return 0;
}
- btrfs_block_release(root, right_buf);
+ btrfs_tree_unlock(right);
+ free_extent_buffer(right);
}
- check_node(root, path, level);
return 1;
}
+/*
+ * readahead one full node of leaves
+ */
+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;
+ u32 nritems;
+ u64 search;
+ u64 lowest_read;
+ u64 highest_read;
+ u64 nread = 0;
+ int direction = path->reada;
+ struct extent_buffer *eb;
+ u32 nr;
+ u32 blocksize;
+ u32 nscan = 0;
+
+ if (level != 1)
+ return;
+
+ if (!path->nodes[level])
+ return;
+
+ node = path->nodes[level];
+
+ search = btrfs_node_blockptr(node, slot);
+ blocksize = btrfs_level_size(root, level - 1);
+ eb = btrfs_find_tree_block(root, search, blocksize);
+ if (eb) {
+ free_extent_buffer(eb);
+ return;
+ }
+
+ highest_read = search;
+ lowest_read = search;
+
+ nritems = btrfs_header_nritems(node);
+ nr = slot;
+ while(1) {
+ if (direction < 0) {
+ if (nr == 0)
+ break;
+ nr--;
+ } else if (direction > 0) {
+ nr++;
+ 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,
+ btrfs_node_ptr_generation(node, nr));
+ nread += blocksize;
+ }
+ nscan++;
+ if (path->reada < 2 && (nread > (256 * 1024) || nscan > 32))
+ break;
+ if(nread > (1024 * 1024) || nscan > 128)
+ break;
+
+ if (search < lowest_read)
+ lowest_read = search;
+ if (search > highest_read)
+ highest_read = search;
+ }
+}
+
+static noinline void unlock_up(struct btrfs_path *path, int level,
+ int lowest_unlock)
+{
+ int i;
+ int skip_level = level;
+ int no_skips = 0;
+ struct extent_buffer *t;
+
+ for (i = level; i < BTRFS_MAX_LEVEL; i++) {
+ if (!path->nodes[i])
+ break;
+ if (!path->locks[i])
+ break;
+ if (!no_skips && path->slots[i] == 0) {
+ skip_level = i + 1;
+ continue;
+ }
+ if (!no_skips && path->keep_locks) {
+ u32 nritems;
+ t = path->nodes[i];
+ nritems = btrfs_header_nritems(t);
+ if (nritems < 1 || path->slots[i] >= nritems - 1) {
+ skip_level = i + 1;
+ continue;
+ }
+ }
+ if (skip_level < i && i >= lowest_unlock)
+ no_skips = 1;
+
+ t = path->nodes[i];
+ if (i >= lowest_unlock && i > skip_level && path->locks[i]) {
+ 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
*root, struct btrfs_key *key, struct btrfs_path *p, int
ins_len, int cow)
{
- struct buffer_head *b;
- struct buffer_head *cow_buf;
- struct btrfs_node *c;
+ struct extent_buffer *b;
+ struct extent_buffer *tmp;
int slot;
int ret;
int level;
+ int should_reada = p->reada;
+ int lowest_unlock = 1;
+ int blocksize;
+ u8 lowest_level = 0;
+ u64 blocknr;
+ u64 gen;
+ struct btrfs_key prealloc_block;
+ 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(cow && root == root->fs_info->extent_root &&
+ !mutex_is_locked(&root->fs_info->alloc_mutex));
+ if (ins_len < 0)
+ lowest_unlock = 2;
+
+ prealloc_block.objectid = 0;
+
again:
- b = root->node;
- get_bh(b);
+ if (p->skip_locking)
+ b = btrfs_root_node(root);
+ else
+ b = btrfs_lock_root_node(root);
+
while (b) {
- c = btrfs_buffer_node(b);
- level = btrfs_header_level(&c->header);
+ level = btrfs_header_level(b);
+
+ /*
+ * setup the path here so we can release it under lock
+ * contention with the cow code
+ */
+ p->nodes[level] = b;
+ if (!p->skip_locking)
+ p->locks[level] = 1;
+
if (cow) {
int wret;
+
+ /* is a cow on this block not required */
+ spin_lock(&root->fs_info->hash_lock);
+ if (btrfs_header_generation(b) == trans->transid &&
+ !btrfs_header_flag(b, BTRFS_HEADER_FLAG_WRITTEN)) {
+ spin_unlock(&root->fs_info->hash_lock);
+ goto cow_done;
+ }
+ spin_unlock(&root->fs_info->hash_lock);
+
+ /* ok, we have to cow, is our old prealloc the right
+ * size?
+ */
+ if (prealloc_block.objectid &&
+ prealloc_block.offset != b->len) {
+ btrfs_free_reserved_extent(root,
+ prealloc_block.objectid,
+ prealloc_block.offset);
+ prealloc_block.objectid = 0;
+ }
+
+ /*
+ * for higher level blocks, try not to allocate blocks
+ * with the block and the parent locks held.
+ */
+ if (level > 1 && !prealloc_block.objectid &&
+ btrfs_path_lock_waiting(p, level)) {
+ u32 size = b->len;
+ u64 hint = b->start;
+
+ btrfs_release_path(root, p);
+ ret = btrfs_reserve_extent(trans, root,
+ size, size, 0,
+ hint, (u64)-1,
+ &prealloc_block, 0);
+ BUG_ON(ret);
+ goto again;
+ }
+
wret = btrfs_cow_block(trans, root, b,
p->nodes[level + 1],
p->slots[level + 1],
- &cow_buf);
- b = cow_buf;
- c = btrfs_buffer_node(b);
+ &b, prealloc_block.objectid);
+ prealloc_block.objectid = 0;
+ if (wret) {
+ free_extent_buffer(b);
+ ret = wret;
+ goto done;
+ }
}
+cow_done:
BUG_ON(!cow && ins_len);
- if (level != btrfs_header_level(&c->header))
+ if (level != btrfs_header_level(b))
WARN_ON(1);
- level = btrfs_header_level(&c->header);
+ 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(c, key, &slot);
- if (!btrfs_is_leaf(c)) {
+ if (ret) {
+ ret = -1;
+ goto done;
+ }
+
+ 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(&c->header) >=
- BTRFS_NODEPTRS_PER_BLOCK(root) - 1) {
+ if (ins_len > 0 && btrfs_header_nritems(b) >=
+ BTRFS_NODEPTRS_PER_BLOCK(root) - 3) {
int sret = split_node(trans, root, p, level);
BUG_ON(sret > 0);
- if (sret)
- return sret;
+ if (sret) {
+ ret = sret;
+ goto done;
+ }
b = p->nodes[level];
- c = btrfs_buffer_node(b);
slot = p->slots[level];
} else if (ins_len < 0) {
int sret = balance_level(trans, root, p,
level);
- if (sret)
- return sret;
+ if (sret) {
+ ret = sret;
+ goto done;
+ }
b = p->nodes[level];
- if (!b)
+ if (!b) {
+ btrfs_release_path(NULL, p);
goto again;
- c = btrfs_buffer_node(b);
+ }
slot = p->slots[level];
- BUG_ON(btrfs_header_nritems(&c->header) == 1);
+ BUG_ON(btrfs_header_nritems(b) == 1);
}
- b = read_tree_block(root, btrfs_node_blockptr(c, slot));
+ unlock_up(p, level, lowest_unlock);
+
+ /* this is only true while dropping a snapshot */
+ if (level == lowest_level) {
+ break;
+ }
+
+ blocknr = btrfs_node_blockptr(b, slot);
+ gen = btrfs_node_ptr_generation(b, slot);
+ blocksize = btrfs_level_size(root, level - 1);
+
+ tmp = btrfs_find_tree_block(root, blocknr, blocksize);
+ if (tmp && btrfs_buffer_uptodate(tmp, gen)) {
+ b = tmp;
+ } else {
+ /*
+ * reduce lock contention at high levels
+ * of the btree by dropping locks before
+ * we read.
+ */
+ if (level > 1) {
+ btrfs_release_path(NULL, p);
+ if (tmp)
+ free_extent_buffer(tmp);
+ if (should_reada)
+ reada_for_search(root, p,
+ level, slot,
+ key->objectid);
+
+ tmp = read_tree_block(root, blocknr,
+ blocksize, gen);
+ if (tmp)
+ free_extent_buffer(tmp);
+ goto again;
+ } else {
+ if (tmp)
+ free_extent_buffer(tmp);
+ if (should_reada)
+ reada_for_search(root, p,
+ level, slot,
+ key->objectid);
+ b = read_node_slot(root, b, slot);
+ }
+ }
+ if (!p->skip_locking)
+ btrfs_tree_lock(b);
} else {
- struct btrfs_leaf *l = (struct btrfs_leaf *)c;
p->slots[level] = slot;
- if (ins_len > 0 && btrfs_leaf_free_space(root, l) <
+ 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;
+ if (sret) {
+ ret = sret;
+ goto done;
+ }
}
- return ret;
+ unlock_up(p, level, lowest_unlock);
+ goto done;
}
}
- return 1;
+ ret = 1;
+done:
+ if (prealloc_block.objectid) {
+ btrfs_free_reserved_extent(root,
+ prealloc_block.objectid,
+ prealloc_block.offset);
+ }
+
+ return ret;
}
/*
* If this fails to write a tree block, it returns -1, but continues
* fixing up the blocks in ram so the tree is consistent.
*/
-static int fixup_low_keys(struct btrfs_trans_handle *trans, struct btrfs_root
- *root, struct btrfs_path *path, struct btrfs_disk_key
- *key, int level)
+static int fixup_low_keys(struct btrfs_trans_handle *trans,
+ struct btrfs_root *root, struct btrfs_path *path,
+ struct btrfs_disk_key *key, int level)
{
int i;
int ret = 0;
+ struct extent_buffer *t;
+
for (i = level; i < BTRFS_MAX_LEVEL; i++) {
- struct btrfs_node *t;
int tslot = path->slots[i];
if (!path->nodes[i])
break;
- t = btrfs_buffer_node(path->nodes[i]);
- btrfs_memcpy(root, t, &t->ptrs[tslot].key, key, sizeof(*key));
+ t = path->nodes[i];
+ btrfs_set_node_key(t, key, tslot);
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 buffer_head *dst_buf, struct
- buffer_head *src_buf)
+static int push_node_left(struct btrfs_trans_handle *trans,
+ struct btrfs_root *root, struct extent_buffer *dst,
+ struct extent_buffer *src, int empty)
{
- struct btrfs_node *src = btrfs_buffer_node(src_buf);
- struct btrfs_node *dst = btrfs_buffer_node(dst_buf);
int push_items = 0;
int src_nritems;
int dst_nritems;
int ret = 0;
- src_nritems = btrfs_header_nritems(&src->header);
- dst_nritems = btrfs_header_nritems(&dst->header);
+ 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),
+ btrfs_node_key_ptr_offset(0),
+ push_items * sizeof(struct btrfs_key_ptr));
- btrfs_memcpy(root, dst, dst->ptrs + dst_nritems, src->ptrs,
- push_items * sizeof(struct btrfs_key_ptr));
if (push_items < src_nritems) {
- btrfs_memmove(root, src, src->ptrs, src->ptrs + push_items,
- (src_nritems - push_items) *
- sizeof(struct btrfs_key_ptr));
- }
- btrfs_set_header_nritems(&src->header, src_nritems - push_items);
- btrfs_set_header_nritems(&dst->header, dst_nritems + push_items);
- btrfs_mark_buffer_dirty(src_buf);
- btrfs_mark_buffer_dirty(dst_buf);
+ memmove_extent_buffer(src, btrfs_node_key_ptr_offset(0),
+ btrfs_node_key_ptr_offset(push_items),
+ (src_nritems - push_items) *
+ sizeof(struct btrfs_key_ptr));
+ }
+ btrfs_set_header_nritems(src, src_nritems - push_items);
+ btrfs_set_header_nritems(dst, dst_nritems + push_items);
+ btrfs_mark_buffer_dirty(src);
+ btrfs_mark_buffer_dirty(dst);
return ret;
}
*
* this will only push up to 1/2 the contents of the left node over
*/
-static int balance_node_right(struct btrfs_trans_handle *trans, struct
- btrfs_root *root, struct buffer_head *dst_buf,
- struct buffer_head *src_buf)
+static int balance_node_right(struct btrfs_trans_handle *trans,
+ struct btrfs_root *root,
+ struct extent_buffer *dst,
+ struct extent_buffer *src)
{
- struct btrfs_node *src = btrfs_buffer_node(src_buf);
- struct btrfs_node *dst = btrfs_buffer_node(dst_buf);
int push_items = 0;
int max_push;
int src_nritems;
int dst_nritems;
int ret = 0;
- src_nritems = btrfs_header_nritems(&src->header);
- dst_nritems = btrfs_header_nritems(&dst->header);
+ 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) {
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;
- btrfs_memmove(root, dst, dst->ptrs + push_items, dst->ptrs,
- dst_nritems * sizeof(struct btrfs_key_ptr));
+ memmove_extent_buffer(dst, btrfs_node_key_ptr_offset(push_items),
+ btrfs_node_key_ptr_offset(0),
+ (dst_nritems) *
+ sizeof(struct btrfs_key_ptr));
- btrfs_memcpy(root, dst, dst->ptrs,
- src->ptrs + src_nritems - push_items,
- push_items * sizeof(struct btrfs_key_ptr));
+ copy_extent_buffer(dst, src,
+ btrfs_node_key_ptr_offset(0),
+ btrfs_node_key_ptr_offset(src_nritems - push_items),
+ push_items * sizeof(struct btrfs_key_ptr));
- btrfs_set_header_nritems(&src->header, src_nritems - push_items);
- btrfs_set_header_nritems(&dst->header, dst_nritems + push_items);
+ btrfs_set_header_nritems(src, src_nritems - push_items);
+ btrfs_set_header_nritems(dst, dst_nritems + push_items);
- btrfs_mark_buffer_dirty(src_buf);
- btrfs_mark_buffer_dirty(dst_buf);
+ btrfs_mark_buffer_dirty(src);
+ btrfs_mark_buffer_dirty(dst);
return ret;
}
*
* returns zero on success or < 0 on failure.
*/
-static int insert_new_root(struct btrfs_trans_handle *trans, struct btrfs_root
- *root, struct btrfs_path *path, int level)
+static int noinline insert_new_root(struct btrfs_trans_handle *trans,
+ struct btrfs_root *root,
+ struct btrfs_path *path, int level)
{
- struct buffer_head *t;
- struct btrfs_node *lower;
- struct btrfs_node *c;
- struct btrfs_disk_key *lower_key;
+ 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);
- t = btrfs_alloc_free_block(trans, root);
- c = btrfs_buffer_node(t);
- memset(c, 0, root->blocksize);
- btrfs_set_header_nritems(&c->header, 1);
- btrfs_set_header_level(&c->header, level);
- btrfs_set_header_blocknr(&c->header, bh_blocknr(t));
- btrfs_set_header_generation(&c->header, trans->transid);
- lower = btrfs_buffer_node(path->nodes[level-1]);
- memcpy(c->header.fsid, root->fs_info->disk_super->fsid,
- sizeof(c->header.fsid));
- if (btrfs_is_leaf(lower))
- lower_key = &((struct btrfs_leaf *)lower)->items[0].key;
+ if (root->ref_cows)
+ root_gen = trans->transid;
else
- lower_key = &lower->ptrs[0].key;
- btrfs_memcpy(root, c, &c->ptrs[0].key, lower_key,
- sizeof(struct btrfs_disk_key));
- btrfs_set_node_blockptr(c, 0, bh_blocknr(path->nodes[level - 1]));
+ 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);
+
+ write_extent_buffer(c, root->fs_info->fsid,
+ (unsigned long)btrfs_header_fsid(c),
+ BTRFS_FSID_SIZE);
+
+ write_extent_buffer(c, root->fs_info->chunk_tree_uuid,
+ (unsigned long)btrfs_header_chunk_tree_uuid(c),
+ BTRFS_UUID_SIZE);
- btrfs_mark_buffer_dirty(t);
+ 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);
+
+ 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 */
- btrfs_block_release(root, root->node);
- root->node = t;
- get_bh(t);
- path->nodes[level] = t;
+ 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;
}
*/
static int insert_ptr(struct btrfs_trans_handle *trans, struct btrfs_root
*root, struct btrfs_path *path, struct btrfs_disk_key
- *key, u64 blocknr, int slot, int level)
+ *key, u64 bytenr, int slot, int level)
{
- struct btrfs_node *lower;
+ struct extent_buffer *lower;
int nritems;
BUG_ON(!path->nodes[level]);
- lower = btrfs_buffer_node(path->nodes[level]);
- nritems = btrfs_header_nritems(&lower->header);
+ lower = path->nodes[level];
+ nritems = btrfs_header_nritems(lower);
if (slot > nritems)
BUG();
if (nritems == BTRFS_NODEPTRS_PER_BLOCK(root))
BUG();
if (slot != nritems) {
- btrfs_memmove(root, lower, lower->ptrs + slot + 1,
- lower->ptrs + slot,
+ memmove_extent_buffer(lower,
+ btrfs_node_key_ptr_offset(slot + 1),
+ btrfs_node_key_ptr_offset(slot),
(nritems - slot) * sizeof(struct btrfs_key_ptr));
}
- btrfs_memcpy(root, lower, &lower->ptrs[slot].key,
- key, sizeof(struct btrfs_disk_key));
- btrfs_set_node_blockptr(lower, slot, blocknr);
- btrfs_set_header_nritems(&lower->header, nritems + 1);
- btrfs_mark_buffer_dirty(path->nodes[level]);
+ 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;
}
*
* 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)
{
- struct buffer_head *t;
- struct btrfs_node *c;
- struct buffer_head *split_buffer;
- struct btrfs_node *split;
+ u64 root_gen;
+ struct extent_buffer *c;
+ struct extent_buffer *split;
+ struct btrfs_disk_key disk_key;
int mid;
int ret;
int wret;
u32 c_nritems;
- t = path->nodes[level];
- c = btrfs_buffer_node(t);
- if (t == root->node) {
+ 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);
if (ret)
return ret;
} else {
ret = push_nodes_for_insert(trans, root, path, level);
- t = path->nodes[level];
- c = btrfs_buffer_node(t);
- if (!ret &&
- btrfs_header_nritems(&c->header) <
- BTRFS_NODEPTRS_PER_BLOCK(root) - 1)
+ c = path->nodes[level];
+ if (!ret && btrfs_header_nritems(c) <
+ BTRFS_NODEPTRS_PER_BLOCK(root) - 3)
return 0;
+ if (ret < 0)
+ return ret;
}
- c_nritems = btrfs_header_nritems(&c->header);
- split_buffer = btrfs_alloc_free_block(trans, root);
- split = btrfs_buffer_node(split_buffer);
- btrfs_set_header_flags(&split->header, btrfs_header_flags(&c->header));
- btrfs_set_header_level(&split->header, btrfs_header_level(&c->header));
- btrfs_set_header_blocknr(&split->header, bh_blocknr(split_buffer));
- btrfs_set_header_generation(&split->header, trans->transid);
- memcpy(split->header.fsid, root->fs_info->disk_super->fsid,
- sizeof(split->header.fsid));
+ 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,
+ 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_flags(split, btrfs_header_flags(c));
+ btrfs_set_header_level(split, btrfs_header_level(c));
+ 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;
- btrfs_memcpy(root, split, split->ptrs, c->ptrs + mid,
- (c_nritems - mid) * sizeof(struct btrfs_key_ptr));
- btrfs_set_header_nritems(&split->header, c_nritems - mid);
- btrfs_set_header_nritems(&c->header, mid);
+
+ copy_extent_buffer(split, c,
+ btrfs_node_key_ptr_offset(0),
+ btrfs_node_key_ptr_offset(mid),
+ (c_nritems - mid) * sizeof(struct btrfs_key_ptr));
+ btrfs_set_header_nritems(split, c_nritems - mid);
+ btrfs_set_header_nritems(c, mid);
ret = 0;
- btrfs_mark_buffer_dirty(t);
- btrfs_mark_buffer_dirty(split_buffer);
- wret = insert_ptr(trans, root, path, &split->ptrs[0].key,
- bh_blocknr(split_buffer), path->slots[level + 1] + 1,
+ btrfs_mark_buffer_dirty(c);
+ btrfs_mark_buffer_dirty(split);
+
+ btrfs_node_key(split, &disk_key, 0);
+ wret = insert_ptr(trans, root, path, &disk_key, split->start,
+ path->slots[level + 1] + 1,
level + 1);
if (wret)
ret = wret;
if (path->slots[level] >= mid) {
path->slots[level] -= mid;
- btrfs_block_release(root, t);
- path->nodes[level] = split_buffer;
+ btrfs_tree_unlock(c);
+ free_extent_buffer(c);
+ path->nodes[level] = split;
path->slots[level + 1] += 1;
} else {
- btrfs_block_release(root, split_buffer);
+ btrfs_tree_unlock(split);
+ free_extent_buffer(split);
}
return ret;
}
* and nr indicate which items in the leaf to check. This totals up the
* space used both by the item structs and the item data
*/
-static int leaf_space_used(struct btrfs_leaf *l, int start, int nr)
+static int leaf_space_used(struct extent_buffer *l, int start, int nr)
{
int data_len;
- int nritems = btrfs_header_nritems(&l->header);
+ int nritems = btrfs_header_nritems(l);
int end = min(nritems, start + nr) - 1;
if (!nr)
return 0;
- data_len = btrfs_item_end(l->items + start);
- data_len = data_len - btrfs_item_offset(l->items + end);
+ data_len = btrfs_item_end_nr(l, start);
+ data_len = data_len - btrfs_item_offset_nr(l, end);
data_len += sizeof(struct btrfs_item) * nr;
WARN_ON(data_len < 0);
return data_len;
* 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 btrfs_leaf *leaf)
+int noinline btrfs_leaf_free_space(struct btrfs_root *root,
+ struct extent_buffer *leaf)
{
- int nritems = btrfs_header_nritems(&leaf->header);
- return BTRFS_LEAF_DATA_SIZE(root) - leaf_space_used(leaf, 0, nritems);
+ int nritems = btrfs_header_nritems(leaf);
+ int 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, (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 buffer_head *left_buf = path->nodes[0];
- struct btrfs_leaf *left = btrfs_buffer_leaf(left_buf);
- struct btrfs_leaf *right;
- struct buffer_head *right_buf;
- struct buffer_head *upper;
- struct btrfs_node *upper_node;
+ 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;
+ int ret;
slot = path->slots[1];
if (!path->nodes[1]) {
return 1;
}
upper = path->nodes[1];
- upper_node = btrfs_buffer_node(upper);
- if (slot >= btrfs_header_nritems(&upper_node->header) - 1) {
+ if (slot >= btrfs_header_nritems(upper) - 1)
return 1;
- }
- right_buf = read_tree_block(root,
- btrfs_node_blockptr(btrfs_buffer_node(upper), slot + 1));
- right = btrfs_buffer_leaf(right_buf);
+
+ 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)) {
- btrfs_block_release(root, right_buf);
- return 1;
- }
+ if (free_space < data_size + sizeof(struct btrfs_item))
+ goto out_unlock;
+
/* cow and double check */
- btrfs_cow_block(trans, root, right_buf, upper, slot + 1, &right_buf);
- right = btrfs_buffer_leaf(right_buf);
+ ret = btrfs_cow_block(trans, root, right, upper,
+ slot + 1, &right, 0);
+ if (ret)
+ goto out_unlock;
+
free_space = btrfs_leaf_free_space(root, right);
- if (free_space < data_size + sizeof(struct btrfs_item)) {
- btrfs_block_release(root, right_buf);
- return 1;
- }
+ if (free_space < data_size + sizeof(struct btrfs_item))
+ goto out_unlock;
+
+ left_nritems = btrfs_header_nritems(left);
+ if (left_nritems == 0)
+ goto out_unlock;
+
+ if (empty)
+ nr = 0;
+ else
+ nr = 1;
+
+ i = left_nritems - 1;
+ while (i >= nr) {
+ item = btrfs_item_nr(left, i);
- left_nritems = btrfs_header_nritems(&left->header);
- if (left_nritems == 0) {
- btrfs_block_release(root, right_buf);
- return 1;
- }
- for (i = left_nritems - 1; i >= 1; i--) {
- item = left->items + i;
if (path->slots[0] == i)
push_space += data_size + sizeof(*item);
- if (btrfs_item_size(item) + sizeof(*item) + push_space >
- free_space)
+
+ if (!left->map_token) {
+ map_extent_buffer(left, (unsigned long)item,
+ sizeof(struct btrfs_item),
+ &left->map_token, &left->kaddr,
+ &left->map_start, &left->map_len,
+ KM_USER1);
+ }
+
+ this_item_size = btrfs_item_size(left, item);
+ if (this_item_size + sizeof(*item) + push_space > free_space)
break;
push_items++;
- push_space += btrfs_item_size(item) + sizeof(*item);
+ push_space += this_item_size + sizeof(*item);
+ if (i == 0)
+ break;
+ i--;
}
- if (push_items == 0) {
- btrfs_block_release(root, right_buf);
- return 1;
+ if (left->map_token) {
+ unmap_extent_buffer(left, left->map_token, KM_USER1);
+ left->map_token = NULL;
}
- if (push_items == left_nritems)
+
+ if (push_items == 0)
+ goto out_unlock;
+
+ if (!empty && push_items == left_nritems)
WARN_ON(1);
- right_nritems = btrfs_header_nritems(&right->header);
+
/* push left to right */
- push_space = btrfs_item_end(left->items + left_nritems - push_items);
+ right_nritems = btrfs_header_nritems(right);
+
+ push_space = btrfs_item_end_nr(left, left_nritems - push_items);
push_space -= leaf_data_end(root, left);
+
/* make room in the right data area */
- btrfs_memmove(root, right, btrfs_leaf_data(right) +
- leaf_data_end(root, right) - push_space,
- btrfs_leaf_data(right) +
- leaf_data_end(root, right), BTRFS_LEAF_DATA_SIZE(root) -
- leaf_data_end(root, right));
+ data_end = leaf_data_end(root, right);
+ memmove_extent_buffer(right,
+ btrfs_leaf_data(right) + data_end - push_space,
+ btrfs_leaf_data(right) + data_end,
+ BTRFS_LEAF_DATA_SIZE(root) - data_end);
+
/* copy from the left data area */
- btrfs_memcpy(root, right, btrfs_leaf_data(right) +
+ copy_extent_buffer(right, left, btrfs_leaf_data(right) +
BTRFS_LEAF_DATA_SIZE(root) - push_space,
btrfs_leaf_data(left) + leaf_data_end(root, left),
push_space);
- btrfs_memmove(root, right, right->items + push_items, right->items,
- right_nritems * sizeof(struct btrfs_item));
+
+ memmove_extent_buffer(right, btrfs_item_nr_offset(push_items),
+ btrfs_item_nr_offset(0),
+ right_nritems * sizeof(struct btrfs_item));
+
/* copy the items from left to right */
- btrfs_memcpy(root, right, right->items, left->items +
- left_nritems - push_items,
- push_items * sizeof(struct btrfs_item));
+ copy_extent_buffer(right, left, btrfs_item_nr_offset(0),
+ btrfs_item_nr_offset(left_nritems - push_items),
+ push_items * sizeof(struct btrfs_item));
/* update the item pointers */
right_nritems += push_items;
- btrfs_set_header_nritems(&right->header, right_nritems);
+ btrfs_set_header_nritems(right, right_nritems);
push_space = BTRFS_LEAF_DATA_SIZE(root);
for (i = 0; i < right_nritems; i++) {
- btrfs_set_item_offset(right->items + i, push_space -
- btrfs_item_size(right->items + i));
- push_space = btrfs_item_offset(right->items + i);
+ item = btrfs_item_nr(right, i);
+ if (!right->map_token) {
+ map_extent_buffer(right, (unsigned long)item,
+ sizeof(struct btrfs_item),
+ &right->map_token, &right->kaddr,
+ &right->map_start, &right->map_len,
+ KM_USER1);
+ }
+ push_space -= btrfs_item_size(right, item);
+ btrfs_set_item_offset(right, item, push_space);
+ }
+
+ if (right->map_token) {
+ unmap_extent_buffer(right, right->map_token, KM_USER1);
+ right->map_token = NULL;
}
left_nritems -= push_items;
- btrfs_set_header_nritems(&left->header, left_nritems);
+ btrfs_set_header_nritems(left, left_nritems);
- btrfs_mark_buffer_dirty(left_buf);
- btrfs_mark_buffer_dirty(right_buf);
+ if (left_nritems)
+ btrfs_mark_buffer_dirty(left);
+ btrfs_mark_buffer_dirty(right);
- btrfs_memcpy(root, upper_node, &upper_node->ptrs[slot + 1].key,
- &right->items[0].key, sizeof(struct btrfs_disk_key));
+ btrfs_item_key(right, &disk_key, 0);
+ btrfs_set_node_key(upper, &disk_key, slot + 1);
btrfs_mark_buffer_dirty(upper);
/* then fixup the leaf pointer in the path */
if (path->slots[0] >= left_nritems) {
path->slots[0] -= left_nritems;
- btrfs_block_release(root, path->nodes[0]);
- path->nodes[0] = right_buf;
+ 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_block_release(root, right_buf);
+ 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 buffer_head *right_buf = path->nodes[0];
- struct btrfs_leaf *right = btrfs_buffer_leaf(right_buf);
- struct buffer_head *t;
- struct btrfs_leaf *left;
+ struct btrfs_disk_key disk_key;
+ struct extent_buffer *right = path->nodes[0];
+ struct extent_buffer *left;
int slot;
int i;
int free_space;
int push_items = 0;
struct btrfs_item *item;
u32 old_left_nritems;
+ u32 right_nritems;
+ u32 nr;
int ret = 0;
int wret;
+ u32 this_item_size;
+ u32 old_left_item_size;
slot = path->slots[1];
- if (slot == 0) {
+ if (slot == 0)
return 1;
- }
- if (!path->nodes[1]) {
+ if (!path->nodes[1])
+ return 1;
+
+ right_nritems = btrfs_header_nritems(right);
+ if (right_nritems == 0) {
return 1;
}
- t = read_tree_block(root,
- btrfs_node_blockptr(btrfs_buffer_node(path->nodes[1]), slot - 1));
- left = btrfs_buffer_leaf(t);
+
+ 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)) {
- btrfs_block_release(root, t);
- return 1;
+ ret = 1;
+ goto out;
}
/* cow and double check */
- btrfs_cow_block(trans, root, t, path->nodes[1], slot - 1, &t);
- left = btrfs_buffer_leaf(t);
+ ret = btrfs_cow_block(trans, root, left,
+ path->nodes[1], slot - 1, &left, 0);
+ if (ret) {
+ /* we hit -ENOSPC, but it isn't fatal here */
+ ret = 1;
+ goto out;
+ }
+
free_space = btrfs_leaf_free_space(root, left);
if (free_space < data_size + sizeof(struct btrfs_item)) {
- btrfs_block_release(root, t);
- return 1;
+ ret = 1;
+ goto out;
}
- if (btrfs_header_nritems(&right->header) == 0) {
- btrfs_block_release(root, t);
- return 1;
- }
+ if (empty)
+ nr = right_nritems;
+ else
+ nr = right_nritems - 1;
+
+ for (i = 0; i < nr; i++) {
+ item = btrfs_item_nr(right, i);
+ if (!right->map_token) {
+ map_extent_buffer(right, (unsigned long)item,
+ sizeof(struct btrfs_item),
+ &right->map_token, &right->kaddr,
+ &right->map_start, &right->map_len,
+ KM_USER1);
+ }
- for (i = 0; i < btrfs_header_nritems(&right->header) - 1; i++) {
- item = right->items + i;
if (path->slots[0] == i)
push_space += data_size + sizeof(*item);
- if (btrfs_item_size(item) + sizeof(*item) + push_space >
- free_space)
+
+ this_item_size = btrfs_item_size(right, item);
+ if (this_item_size + sizeof(*item) + push_space > free_space)
break;
+
push_items++;
- push_space += btrfs_item_size(item) + sizeof(*item);
+ push_space += this_item_size + sizeof(*item);
+ }
+
+ if (right->map_token) {
+ unmap_extent_buffer(right, right->map_token, KM_USER1);
+ right->map_token = NULL;
}
+
if (push_items == 0) {
- btrfs_block_release(root, t);
- return 1;
+ ret = 1;
+ goto out;
}
- if (push_items == btrfs_header_nritems(&right->header))
+ if (!empty && push_items == btrfs_header_nritems(right))
WARN_ON(1);
+
/* push data from right to left */
- btrfs_memcpy(root, left, left->items +
- btrfs_header_nritems(&left->header),
- right->items, push_items * sizeof(struct btrfs_item));
+ copy_extent_buffer(left, right,
+ btrfs_item_nr_offset(btrfs_header_nritems(left)),
+ btrfs_item_nr_offset(0),
+ push_items * sizeof(struct btrfs_item));
+
push_space = BTRFS_LEAF_DATA_SIZE(root) -
- btrfs_item_offset(right->items + push_items -1);
- btrfs_memcpy(root, left, btrfs_leaf_data(left) +
+ btrfs_item_offset_nr(right, push_items -1);
+
+ copy_extent_buffer(left, right, btrfs_leaf_data(left) +
leaf_data_end(root, left) - push_space,
btrfs_leaf_data(right) +
- btrfs_item_offset(right->items + push_items - 1),
+ btrfs_item_offset_nr(right, push_items - 1),
push_space);
- old_left_nritems = btrfs_header_nritems(&left->header);
+ old_left_nritems = btrfs_header_nritems(left);
BUG_ON(old_left_nritems < 0);
+ old_left_item_size = btrfs_item_offset_nr(left, old_left_nritems - 1);
for (i = old_left_nritems; i < old_left_nritems + push_items; i++) {
- u32 ioff = btrfs_item_offset(left->items + i);
- btrfs_set_item_offset(left->items + i, ioff -
- (BTRFS_LEAF_DATA_SIZE(root) -
- btrfs_item_offset(left->items +
- old_left_nritems - 1)));
+ u32 ioff;
+
+ item = btrfs_item_nr(left, i);
+ if (!left->map_token) {
+ map_extent_buffer(left, (unsigned long)item,
+ sizeof(struct btrfs_item),
+ &left->map_token, &left->kaddr,
+ &left->map_start, &left->map_len,
+ KM_USER1);
+ }
+
+ ioff = btrfs_item_offset(left, item);
+ btrfs_set_item_offset(left, item,
+ ioff - (BTRFS_LEAF_DATA_SIZE(root) - old_left_item_size));
+ }
+ btrfs_set_header_nritems(left, old_left_nritems + push_items);
+ if (left->map_token) {
+ unmap_extent_buffer(left, left->map_token, KM_USER1);
+ left->map_token = NULL;
}
- btrfs_set_header_nritems(&left->header, old_left_nritems + push_items);
/* fixup right node */
- push_space = btrfs_item_offset(right->items + push_items - 1) -
- leaf_data_end(root, right);
- btrfs_memmove(root, right, btrfs_leaf_data(right) +
- BTRFS_LEAF_DATA_SIZE(root) - push_space,
- btrfs_leaf_data(right) +
- leaf_data_end(root, right), push_space);
- btrfs_memmove(root, right, right->items, right->items + push_items,
- (btrfs_header_nritems(&right->header) - push_items) *
- sizeof(struct btrfs_item));
- btrfs_set_header_nritems(&right->header,
- btrfs_header_nritems(&right->header) -
- push_items);
+ 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 -= 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) {
+ map_extent_buffer(right, (unsigned long)item,
+ sizeof(struct btrfs_item),
+ &right->map_token, &right->kaddr,
+ &right->map_start, &right->map_len,
+ KM_USER1);
+ }
- for (i = 0; i < btrfs_header_nritems(&right->header); i++) {
- btrfs_set_item_offset(right->items + i, push_space -
- btrfs_item_size(right->items + i));
- push_space = btrfs_item_offset(right->items + i);
+ push_space = push_space - btrfs_item_size(right, item);
+ btrfs_set_item_offset(right, item, push_space);
+ }
+ if (right->map_token) {
+ unmap_extent_buffer(right, right->map_token, KM_USER1);
+ right->map_token = NULL;
}
- btrfs_mark_buffer_dirty(t);
- btrfs_mark_buffer_dirty(right_buf);
- wret = fixup_low_keys(trans, root, path, &right->items[0].key, 1);
+ btrfs_mark_buffer_dirty(left);
+ 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);
if (wret)
ret = wret;
/* then fixup the leaf pointer in the path */
if (path->slots[0] < push_items) {
path->slots[0] += old_left_nritems;
- btrfs_block_release(root, path->nodes[0]);
- path->nodes[0] = t;
+ 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_block_release(root, t);
+ 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;
}
/*
*
* 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)
+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)
{
- struct buffer_head *l_buf;
- struct btrfs_leaf *l;
+ u64 root_gen;
+ struct extent_buffer *l;
u32 nritems;
int mid;
int slot;
- struct btrfs_leaf *right;
- struct buffer_head *right_buffer;
+ struct extent_buffer *right;
int space_needed = data_size + sizeof(struct btrfs_item);
int data_copy_size;
int rt_data_off;
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_buf = path->nodes[0];
- l = btrfs_buffer_leaf(l_buf);
+ }
+ 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]) {
ret = insert_new_root(trans, root, path, 1);
if (ret)
return ret;
}
+again:
+ double_split = 0;
+ l = path->nodes[0];
slot = path->slots[0];
- nritems = btrfs_header_nritems(&l->header);
+ nritems = btrfs_header_nritems(l);
mid = (nritems + 1)/ 2;
- right_buffer = btrfs_alloc_free_block(trans, root);
- BUG_ON(!right_buffer);
- right = btrfs_buffer_leaf(right_buffer);
- memset(&right->header, 0, sizeof(right->header));
- btrfs_set_header_blocknr(&right->header, bh_blocknr(right_buffer));
- btrfs_set_header_generation(&right->header, trans->transid);
- btrfs_set_header_level(&right->header, 0);
- memcpy(right->header.fsid, root->fs_info->disk_super->fsid,
- sizeof(right->header.fsid));
+
+ btrfs_item_key(l, &disk_key, 0);
+
+ right = btrfs_alloc_free_block(trans, root, root->leafsize,
+ 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);
+ 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);
+
+ 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 >
BTRFS_LEAF_DATA_SIZE(root)) {
if (slot >= nritems) {
btrfs_cpu_key_to_disk(&disk_key, ins_key);
- btrfs_set_header_nritems(&right->header, 0);
+ btrfs_set_header_nritems(right, 0);
wret = insert_ptr(trans, root, path,
- &disk_key,
- bh_blocknr(right_buffer),
+ &disk_key, right->start,
path->slots[1] + 1, 1);
if (wret)
ret = wret;
- btrfs_block_release(root, path->nodes[0]);
- path->nodes[0] = right_buffer;
+
+ 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;
- double_split = 1;
+ if (mid != nritems &&
+ leaf_space_used(l, mid, nritems - mid) +
+ space_needed > BTRFS_LEAF_DATA_SIZE(root)) {
+ double_split = 1;
+ }
}
} 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->header, 0);
+ btrfs_set_header_nritems(right, 0);
wret = insert_ptr(trans, root, path,
&disk_key,
- bh_blocknr(right_buffer),
- path->slots[1] - 1, 1);
+ right->start,
+ path->slots[1], 1);
if (wret)
ret = wret;
- btrfs_block_release(root, path->nodes[0]);
- path->nodes[0] = right_buffer;
+ btrfs_tree_unlock(path->nodes[0]);
+ free_extent_buffer(path->nodes[0]);
+ path->nodes[0] = right;
path->slots[0] = 0;
- path->slots[1] -= 1;
if (path->slots[1] == 0) {
wret = fixup_low_keys(trans, root,
path, &disk_key, 1);
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;
}
}
- btrfs_set_header_nritems(&right->header, nritems - mid);
- data_copy_size = btrfs_item_end(l->items + mid) -
- leaf_data_end(root, l);
- btrfs_memcpy(root, right, right->items, l->items + mid,
- (nritems - mid) * sizeof(struct btrfs_item));
- btrfs_memcpy(root, right,
+ nritems = nritems - mid;
+ btrfs_set_header_nritems(right, nritems);
+ data_copy_size = btrfs_item_end_nr(l, mid) - leaf_data_end(root, l);
+
+ copy_extent_buffer(right, l, btrfs_item_nr_offset(0),
+ btrfs_item_nr_offset(mid),
+ nritems * sizeof(struct btrfs_item));
+
+ copy_extent_buffer(right, l,
btrfs_leaf_data(right) + BTRFS_LEAF_DATA_SIZE(root) -
data_copy_size, btrfs_leaf_data(l) +
leaf_data_end(root, l), data_copy_size);
+
rt_data_off = BTRFS_LEAF_DATA_SIZE(root) -
- btrfs_item_end(l->items + mid);
+ btrfs_item_end_nr(l, mid);
+
+ for (i = 0; i < nritems; i++) {
+ struct btrfs_item *item = btrfs_item_nr(right, i);
+ u32 ioff;
+
+ if (!right->map_token) {
+ map_extent_buffer(right, (unsigned long)item,
+ sizeof(struct btrfs_item),
+ &right->map_token, &right->kaddr,
+ &right->map_start, &right->map_len,
+ KM_USER1);
+ }
+
+ ioff = btrfs_item_offset(right, item);
+ btrfs_set_item_offset(right, item, ioff + rt_data_off);
+ }
- for (i = 0; i < btrfs_header_nritems(&right->header); i++) {
- u32 ioff = btrfs_item_offset(right->items + i);
- btrfs_set_item_offset(right->items + i, ioff + rt_data_off);
+ if (right->map_token) {
+ unmap_extent_buffer(right, right->map_token, KM_USER1);
+ right->map_token = NULL;
}
- btrfs_set_header_nritems(&l->header, mid);
+ btrfs_set_header_nritems(l, mid);
ret = 0;
- wret = insert_ptr(trans, root, path, &right->items[0].key,
- bh_blocknr(right_buffer), path->slots[1] + 1, 1);
+ btrfs_item_key(right, &disk_key, 0);
+ wret = insert_ptr(trans, root, path, &disk_key, right->start,
+ path->slots[1] + 1, 1);
if (wret)
ret = wret;
- btrfs_mark_buffer_dirty(right_buffer);
- btrfs_mark_buffer_dirty(l_buf);
+
+ btrfs_mark_buffer_dirty(right);
+ btrfs_mark_buffer_dirty(l);
BUG_ON(path->slots[0] != slot);
+
if (mid <= slot) {
- btrfs_block_release(root, path->nodes[0]);
- path->nodes[0] = right_buffer;
+ 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
- btrfs_block_release(root, right_buffer);
+ } else {
+ btrfs_tree_unlock(right);
+ free_extent_buffer(right);
+ }
+
BUG_ON(path->slots[0] < 0);
- if (!double_split)
- return ret;
- right_buffer = btrfs_alloc_free_block(trans, root);
- BUG_ON(!right_buffer);
- right = btrfs_buffer_leaf(right_buffer);
- memset(&right->header, 0, sizeof(right->header));
- btrfs_set_header_blocknr(&right->header, bh_blocknr(right_buffer));
- btrfs_set_header_generation(&right->header, trans->transid);
- btrfs_set_header_level(&right->header, 0);
- memcpy(right->header.fsid, root->fs_info->disk_super->fsid,
- sizeof(right->header.fsid));
- btrfs_cpu_key_to_disk(&disk_key, ins_key);
- btrfs_set_header_nritems(&right->header, 0);
- wret = insert_ptr(trans, root, path,
- &disk_key,
- bh_blocknr(right_buffer),
- 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;
}
- btrfs_block_release(root, path->nodes[0]);
- path->nodes[0] = right_buffer;
- path->slots[0] = 0;
- check_node(root, path, 1);
- check_leaf(root, path, 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;
int slot_orig;
- struct btrfs_leaf *leaf;
- struct buffer_head *leaf_buf;
+ struct extent_buffer *leaf;
+ struct btrfs_item *item;
u32 nritems;
unsigned int data_end;
unsigned int old_data_start;
int i;
slot_orig = path->slots[0];
- leaf_buf = path->nodes[0];
- leaf = btrfs_buffer_leaf(leaf_buf);
+ 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->header);
+ nritems = btrfs_header_nritems(leaf);
data_end = leaf_data_end(root, leaf);
- slot = path->slots[0];
- old_data_start = btrfs_item_offset(leaf->items + slot);
- old_size = btrfs_item_size(leaf->items + slot);
- BUG_ON(old_size <= new_size);
+ old_data_start = btrfs_item_offset_nr(leaf, slot);
+
size_diff = old_size - new_size;
BUG_ON(slot < 0);
*/
/* first correct the data pointers */
for (i = slot; i < nritems; i++) {
- u32 ioff = btrfs_item_offset(leaf->items + i);
- btrfs_set_item_offset(leaf->items + i,
- ioff + size_diff);
+ u32 ioff;
+ item = btrfs_item_nr(leaf, i);
+
+ if (!leaf->map_token) {
+ map_extent_buffer(leaf, (unsigned long)item,
+ sizeof(struct btrfs_item),
+ &leaf->map_token, &leaf->kaddr,
+ &leaf->map_start, &leaf->map_len,
+ KM_USER1);
+ }
+
+ ioff = btrfs_item_offset(leaf, item);
+ btrfs_set_item_offset(leaf, item, ioff + size_diff);
}
+
+ if (leaf->map_token) {
+ unmap_extent_buffer(leaf, leaf->map_token, KM_USER1);
+ leaf->map_token = NULL;
+ }
+
/* shift the data */
- btrfs_memmove(root, leaf, btrfs_leaf_data(leaf) +
- data_end + size_diff, btrfs_leaf_data(leaf) +
- data_end, old_data_start + new_size - data_end);
- btrfs_set_item_size(leaf->items + slot, new_size);
- btrfs_mark_buffer_dirty(leaf_buf);
+ 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);
+ btrfs_mark_buffer_dirty(leaf);
ret = 0;
- if (btrfs_leaf_free_space(root, leaf) < 0)
+ if (btrfs_leaf_free_space(root, leaf) < 0) {
+ btrfs_print_leaf(root, leaf);
BUG();
- check_leaf(root, path, 0);
+ }
return ret;
}
-int btrfs_extend_item(struct btrfs_trans_handle *trans, struct btrfs_root
- *root, struct btrfs_path *path, u32 data_size)
+int btrfs_extend_item(struct btrfs_trans_handle *trans,
+ struct btrfs_root *root, struct btrfs_path *path,
+ u32 data_size)
{
int ret = 0;
int slot;
int slot_orig;
- struct btrfs_leaf *leaf;
- struct buffer_head *leaf_buf;
+ struct extent_buffer *leaf;
+ struct btrfs_item *item;
u32 nritems;
unsigned int data_end;
unsigned int old_data;
int i;
slot_orig = path->slots[0];
- leaf_buf = path->nodes[0];
- leaf = btrfs_buffer_leaf(leaf_buf);
+ leaf = path->nodes[0];
- nritems = btrfs_header_nritems(&leaf->header);
+ nritems = btrfs_header_nritems(leaf);
data_end = leaf_data_end(root, leaf);
- if (btrfs_leaf_free_space(root, leaf) < data_size)
+ if (btrfs_leaf_free_space(root, leaf) < data_size) {
+ btrfs_print_leaf(root, leaf);
BUG();
+ }
slot = path->slots[0];
- old_data = btrfs_item_end(leaf->items + slot);
+ old_data = btrfs_item_end_nr(leaf, slot);
BUG_ON(slot < 0);
- BUG_ON(slot >= nritems);
+ if (slot >= nritems) {
+ btrfs_print_leaf(root, leaf);
+ printk("slot %d too large, nritems %d\n", slot, nritems);
+ BUG_ON(1);
+ }
/*
* item0..itemN ... dataN.offset..dataN.size .. data0.size
*/
/* first correct the data pointers */
for (i = slot; i < nritems; i++) {
- u32 ioff = btrfs_item_offset(leaf->items + i);
- btrfs_set_item_offset(leaf->items + i,
- ioff - data_size);
+ u32 ioff;
+ item = btrfs_item_nr(leaf, i);
+
+ if (!leaf->map_token) {
+ map_extent_buffer(leaf, (unsigned long)item,
+ sizeof(struct btrfs_item),
+ &leaf->map_token, &leaf->kaddr,
+ &leaf->map_start, &leaf->map_len,
+ KM_USER1);
+ }
+ ioff = btrfs_item_offset(leaf, item);
+ btrfs_set_item_offset(leaf, item, ioff - data_size);
}
+
+ if (leaf->map_token) {
+ unmap_extent_buffer(leaf, leaf->map_token, KM_USER1);
+ leaf->map_token = NULL;
+ }
+
/* shift the data */
- btrfs_memmove(root, leaf, btrfs_leaf_data(leaf) +
+ memmove_extent_buffer(leaf, btrfs_leaf_data(leaf) +
data_end - data_size, btrfs_leaf_data(leaf) +
data_end, old_data - data_end);
+
data_end = old_data;
- old_size = btrfs_item_size(leaf->items + slot);
- btrfs_set_item_size(leaf->items + slot, old_size + data_size);
- btrfs_mark_buffer_dirty(leaf_buf);
+ old_size = btrfs_item_size_nr(leaf, slot);
+ item = btrfs_item_nr(leaf, slot);
+ btrfs_set_item_size(leaf, item, old_size + data_size);
+ btrfs_mark_buffer_dirty(leaf);
ret = 0;
- if (btrfs_leaf_free_space(root, leaf) < 0)
+ if (btrfs_leaf_free_space(root, leaf) < 0) {
+ btrfs_print_leaf(root, leaf);
BUG();
- check_leaf(root, path, 0);
+ }
return ret;
}
* 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, struct btrfs_root
- *root, struct btrfs_path *path, struct btrfs_key
- *cpu_key, u32 data_size)
+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,
+ int nr)
{
+ struct extent_buffer *leaf;
+ struct btrfs_item *item;
int ret = 0;
int slot;
int slot_orig;
- struct btrfs_leaf *leaf;
- struct buffer_head *leaf_buf;
+ 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);
+ for (i = 0; i < nr; i++) {
+ total_data += data_size[i];
+ }
- /* create a root if there isn't one */
- if (!root->node)
- BUG();
- ret = btrfs_search_slot(trans, root, cpu_key, path, data_size, 1);
+ total_size = total_data + (nr * sizeof(struct btrfs_item));
+ ret = btrfs_search_slot(trans, root, cpu_key, path, total_size, 1);
if (ret == 0) {
return -EEXIST;
}
goto out;
slot_orig = path->slots[0];
- leaf_buf = path->nodes[0];
- leaf = btrfs_buffer_leaf(leaf_buf);
+ leaf = path->nodes[0];
- nritems = btrfs_header_nritems(&leaf->header);
+ nritems = btrfs_header_nritems(leaf);
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",
+ total_size, btrfs_leaf_free_space(root, leaf));
BUG();
}
+
slot = path->slots[0];
BUG_ON(slot < 0);
+
if (slot != nritems) {
int i;
- unsigned int old_data = btrfs_item_end(leaf->items + slot);
+ unsigned int old_data = btrfs_item_end_nr(leaf, slot);
+ if (old_data < data_end) {
+ btrfs_print_leaf(root, leaf);
+ printk("slot %d old_data %d data_end %d\n",
+ slot, old_data, data_end);
+ BUG_ON(1);
+ }
/*
* item0..itemN ... dataN.offset..dataN.size .. data0.size
*/
/* first correct the data pointers */
+ WARN_ON(leaf->map_token);
for (i = slot; i < nritems; i++) {
- u32 ioff = btrfs_item_offset(leaf->items + i);
- btrfs_set_item_offset(leaf->items + i,
- ioff - data_size);
+ u32 ioff;
+
+ item = btrfs_item_nr(leaf, i);
+ if (!leaf->map_token) {
+ map_extent_buffer(leaf, (unsigned long)item,
+ sizeof(struct btrfs_item),
+ &leaf->map_token, &leaf->kaddr,
+ &leaf->map_start, &leaf->map_len,
+ KM_USER1);
+ }
+
+ ioff = btrfs_item_offset(leaf, item);
+ btrfs_set_item_offset(leaf, item, ioff - total_data);
+ }
+ if (leaf->map_token) {
+ unmap_extent_buffer(leaf, leaf->map_token, KM_USER1);
+ leaf->map_token = NULL;
}
/* shift the items */
- btrfs_memmove(root, leaf, leaf->items + slot + 1,
- leaf->items + slot,
+ memmove_extent_buffer(leaf, btrfs_item_nr_offset(slot + nr),
+ btrfs_item_nr_offset(slot),
(nritems - slot) * sizeof(struct btrfs_item));
/* shift the data */
- btrfs_memmove(root, leaf, btrfs_leaf_data(leaf) +
- data_end - data_size, btrfs_leaf_data(leaf) +
+ memmove_extent_buffer(leaf, 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_memcpy(root, leaf, &leaf->items[slot].key, &disk_key,
- sizeof(struct btrfs_disk_key));
- btrfs_set_item_offset(leaf->items + slot, data_end - data_size);
- btrfs_set_item_size(leaf->items + slot, data_size);
- btrfs_set_header_nritems(&leaf->header, nritems + 1);
- btrfs_mark_buffer_dirty(leaf_buf);
+ 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)
+ if (btrfs_leaf_free_space(root, leaf) < 0) {
+ btrfs_print_leaf(root, leaf);
BUG();
- check_leaf(root, path, 0);
+ }
out:
return ret;
}
{
int ret = 0;
struct btrfs_path *path;
- u8 *ptr;
+ struct extent_buffer *leaf;
+ unsigned long ptr;
path = btrfs_alloc_path();
BUG_ON(!path);
- btrfs_init_path(path);
ret = btrfs_insert_empty_item(trans, root, path, cpu_key, data_size);
if (!ret) {
- ptr = btrfs_item_ptr(btrfs_buffer_leaf(path->nodes[0]),
- path->slots[0], u8);
- btrfs_memcpy(root, path->nodes[0]->b_data,
- ptr, data, data_size);
- btrfs_mark_buffer_dirty(path->nodes[0]);
+ leaf = path->nodes[0];
+ ptr = btrfs_item_ptr_offset(leaf, path->slots[0]);
+ write_extent_buffer(leaf, data, ptr, data_size);
+ btrfs_mark_buffer_dirty(leaf);
}
- btrfs_release_path(root, path);
btrfs_free_path(path);
return ret;
}
static int del_ptr(struct btrfs_trans_handle *trans, struct btrfs_root *root,
struct btrfs_path *path, int level, int slot)
{
- struct btrfs_node *node;
- struct buffer_head *parent = path->nodes[level];
+ struct extent_buffer *parent = path->nodes[level];
u32 nritems;
int ret = 0;
int wret;
- node = btrfs_buffer_node(parent);
- nritems = btrfs_header_nritems(&node->header);
+ nritems = btrfs_header_nritems(parent);
if (slot != nritems -1) {
- btrfs_memmove(root, node, node->ptrs + slot,
- node->ptrs + slot + 1,
+ memmove_extent_buffer(parent,
+ btrfs_node_key_ptr_offset(slot),
+ btrfs_node_key_ptr_offset(slot + 1),
sizeof(struct btrfs_key_ptr) *
(nritems - slot - 1));
}
nritems--;
- btrfs_set_header_nritems(&node->header, nritems);
+ btrfs_set_header_nritems(parent, nritems);
if (nritems == 0 && parent == root->node) {
- struct btrfs_header *header = btrfs_buffer_header(root->node);
- BUG_ON(btrfs_header_level(header) != 1);
+ BUG_ON(btrfs_header_level(root->node) != 1);
/* just turn the root into a leaf and break */
- btrfs_set_header_level(header, 0);
+ btrfs_set_header_level(root->node, 0);
} else if (slot == 0) {
- wret = fixup_low_keys(trans, root, path, &node->ptrs[0].key,
- level + 1);
+ struct btrfs_disk_key disk_key;
+
+ btrfs_node_key(parent, &disk_key, 0);
+ wret = fixup_low_keys(trans, root, path, &disk_key, level + 1);
if (wret)
ret = wret;
}
* 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 btrfs_leaf *leaf;
- struct buffer_head *leaf_buf;
- int doff;
- int dsize;
+ struct extent_buffer *leaf;
+ struct btrfs_item *item;
+ int last_off;
+ int dsize = 0;
int ret = 0;
int wret;
+ int i;
u32 nritems;
- leaf_buf = path->nodes[0];
- leaf = btrfs_buffer_leaf(leaf_buf);
- slot = path->slots[0];
- doff = btrfs_item_offset(leaf->items + slot);
- dsize = btrfs_item_size(leaf->items + slot);
- nritems = btrfs_header_nritems(&leaf->header);
+ leaf = path->nodes[0];
+ last_off = btrfs_item_offset_nr(leaf, slot + nr - 1);
+
+ for (i = 0; i < nr; i++)
+ dsize += btrfs_item_size_nr(leaf, slot + i);
- if (slot != nritems - 1) {
+ nritems = btrfs_header_nritems(leaf);
+
+ if (slot + nr != nritems) {
int i;
int data_end = leaf_data_end(root, leaf);
- btrfs_memmove(root, leaf, btrfs_leaf_data(leaf) +
+
+ memmove_extent_buffer(leaf, btrfs_leaf_data(leaf) +
data_end + dsize,
btrfs_leaf_data(leaf) + data_end,
- doff - data_end);
- for (i = slot + 1; i < nritems; i++) {
- u32 ioff = btrfs_item_offset(leaf->items + i);
- btrfs_set_item_offset(leaf->items + i, ioff + dsize);
+ last_off - data_end);
+
+ for (i = slot + nr; i < nritems; i++) {
+ u32 ioff;
+
+ item = btrfs_item_nr(leaf, i);
+ if (!leaf->map_token) {
+ map_extent_buffer(leaf, (unsigned long)item,
+ sizeof(struct btrfs_item),
+ &leaf->map_token, &leaf->kaddr,
+ &leaf->map_start, &leaf->map_len,
+ KM_USER1);
+ }
+ ioff = btrfs_item_offset(leaf, item);
+ btrfs_set_item_offset(leaf, item, ioff + dsize);
+ }
+
+ if (leaf->map_token) {
+ unmap_extent_buffer(leaf, leaf->map_token, KM_USER1);
+ leaf->map_token = NULL;
}
- btrfs_memmove(root, leaf, leaf->items + slot,
- leaf->items + slot + 1,
+
+ memmove_extent_buffer(leaf, btrfs_item_nr_offset(slot),
+ btrfs_item_nr_offset(slot + nr),
sizeof(struct btrfs_item) *
- (nritems - slot - 1));
+ (nritems - slot - nr));
}
- btrfs_set_header_nritems(&leaf->header, 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_buf == root->node) {
- btrfs_set_header_level(&leaf->header, 0);
+ if (leaf == root->node) {
+ btrfs_set_header_level(leaf, 0);
} else {
- clean_tree_block(trans, root, leaf_buf);
- wait_on_buffer(leaf_buf);
+ 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,
- bh_blocknr(leaf_buf), 1, 1);
+ leaf->start, leaf->len,
+ btrfs_header_owner(path->nodes[1]),
+ root_gen, 0, 0, 1);
if (wret)
ret = wret;
}
} else {
int used = leaf_space_used(leaf, 0, nritems);
if (slot == 0) {
+ struct btrfs_disk_key disk_key;
+
+ btrfs_item_key(leaf, &disk_key, 0);
wret = fixup_low_keys(trans, root, path,
- &leaf->items[0].key, 1);
+ &disk_key, 1);
if (wret)
ret = wret;
}
/* delete the leaf if it is mostly empty */
- if (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];
- get_bh(leaf_buf);
- wret = push_leaf_left(trans, root, path, 1);
- if (wret < 0)
+ extent_buffer_get(leaf);
+
+ wret = push_leaf_left(trans, root, path, 1, 1);
+ if (wret < 0 && wret != -ENOSPC)
ret = wret;
- if (path->nodes[0] == leaf_buf &&
- btrfs_header_nritems(&leaf->header)) {
- wret = push_leaf_right(trans, root, path, 1);
- if (wret < 0)
+
+ if (path->nodes[0] == leaf &&
+ btrfs_header_nritems(leaf)) {
+ wret = push_leaf_right(trans, root, path, 1, 1);
+ if (wret < 0 && wret != -ENOSPC)
ret = wret;
}
- if (btrfs_header_nritems(&leaf->header) == 0) {
- u64 blocknr = bh_blocknr(leaf_buf);
- clean_tree_block(trans, root, leaf_buf);
- wait_on_buffer(leaf_buf);
+
+ if (btrfs_header_nritems(leaf) == 0) {
+ u64 root_gen;
+ u64 bytenr = leaf->start;
+ u32 blocksize = leaf->len;
+
+ root_gen = btrfs_header_generation(
+ path->nodes[1]);
+
wret = del_ptr(trans, root, path, 1, slot);
if (wret)
ret = wret;
- btrfs_block_release(root, leaf_buf);
- wret = btrfs_free_extent(trans, root, blocknr,
- 1, 1);
+
+ free_extent_buffer(leaf);
+ wret = btrfs_free_extent(trans, root, bytenr,
+ blocksize,
+ btrfs_header_owner(path->nodes[1]),
+ root_gen, 0, 0, 1);
if (wret)
ret = wret;
} else {
- btrfs_mark_buffer_dirty(leaf_buf);
- btrfs_block_release(root, leaf_buf);
+ /* 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 {
- btrfs_mark_buffer_dirty(leaf_buf);
+ btrfs_mark_buffer_dirty(leaf);
+ }
+ }
+ return ret;
+}
+
+/*
+ * 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;
+}
+
+/*
+ * A helper function to walk down the tree starting at min_key, and looking
+ * for nodes or leaves that are either in cache or have a minimum
+ * transaction id. This is used by the btree defrag code, but could
+ * also be used to search for blocks that have changed since a given
+ * transaction id.
+ *
+ * This does not cow, but it does stuff the starting key it finds back
+ * into min_key, so you can call btrfs_search_slot with cow=1 on the
+ * key and get a writable path.
+ *
+ * This does lock as it descends, and path->keep_locks should be set
+ * to 1 by the caller.
+ *
+ * This honors path->lowest_level to prevent descent past a given level
+ * of the tree.
+ *
+ * returns zero if something useful was found, < 0 on error and 1 if there
+ * 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)
+{
+ struct extent_buffer *cur;
+ struct btrfs_key found_key;
+ int slot;
+ int sret;
+ u32 nritems;
+ int level;
+ int ret = 1;
+
+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;
+
+ if (btrfs_header_generation(cur) < min_trans) {
+ ret = 1;
+ goto out;
+ }
+ while(1) {
+ nritems = btrfs_header_nritems(cur);
+ level = btrfs_header_level(cur);
+ sret = bin_search(cur, min_key, level, &slot);
+
+ /* 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);
+ goto out;
+ }
+ if (sret && slot > 0)
+ slot--;
+ /*
+ * check this node pointer against the cache_only and
+ * min_trans parameters. If it isn't in cache or is too
+ * old, skip to the next one.
+ */
+ while(slot < nritems) {
+ 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) {
+ slot++;
+ continue;
+ }
+ 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));
+
+ if (tmp && btrfs_buffer_uptodate(tmp, gen)) {
+ free_extent_buffer(tmp);
+ break;
+ }
+ if (tmp)
+ 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) {
+ path->slots[level] = slot;
+ sret = btrfs_find_next_key(root, path, min_key, level,
+ cache_only, min_trans);
+ if (sret == 0) {
+ btrfs_release_path(root, path);
+ goto again;
+ } else {
+ goto out;
+ }
+ }
+ /* save our key for returning back */
+ btrfs_node_key_to_cpu(cur, &found_key, slot);
+ path->slots[level] = slot;
+ if (level == path->lowest_level) {
+ ret = 0;
+ unlock_up(path, level, 1);
+ goto out;
+ }
+ cur = read_node_slot(root, cur, slot);
+
+ btrfs_tree_lock(cur);
+ path->locks[level - 1] = 1;
+ path->nodes[level - 1] = cur;
+ unlock_up(path, level, 1);
}
+out:
+ if (ret == 0)
+ memcpy(min_key, &found_key, sizeof(found_key));
return ret;
}
/*
- * walk up the tree as far as required to find the next leaf.
+ * this is similar to btrfs_next_leaf, but does not try to preserve
+ * and fixup the path. It looks for and returns the next key in the
+ * tree based on the current path and the cache_only and min_trans
+ * parameters.
+ *
+ * 0 is returned if another key is found, < 0 if there are any errors
+ * and 1 is returned if there are no higher keys in the tree
+ *
+ * path->keep_locks should be set to 1 on the search made before
+ * calling this function.
+ */
+int btrfs_find_next_key(struct btrfs_root *root, struct btrfs_path *path,
+ struct btrfs_key *key, int lowest_level,
+ int cache_only, u64 min_trans)
+{
+ int level = lowest_level;
+ int slot;
+ struct extent_buffer *c;
+
+ while(level < BTRFS_MAX_LEVEL) {
+ if (!path->nodes[level])
+ return 1;
+
+ slot = path->slots[level] + 1;
+ c = path->nodes[level];
+next:
+ if (slot >= btrfs_header_nritems(c)) {
+ level++;
+ if (level == BTRFS_MAX_LEVEL) {
+ return 1;
+ }
+ continue;
+ }
+ if (level == 0)
+ btrfs_item_key_to_cpu(c, key, slot);
+ else {
+ u64 blockptr = btrfs_node_blockptr(c, slot);
+ u64 gen = btrfs_node_ptr_generation(c, slot);
+
+ if (cache_only) {
+ struct extent_buffer *cur;
+ cur = btrfs_find_tree_block(root, blockptr,
+ btrfs_level_size(root, level - 1));
+ if (!cur || !btrfs_buffer_uptodate(cur, gen)) {
+ slot++;
+ if (cur)
+ free_extent_buffer(cur);
+ goto next;
+ }
+ free_extent_buffer(cur);
+ }
+ if (gen < min_trans) {
+ slot++;
+ goto next;
+ }
+ btrfs_node_key_to_cpu(c, key, slot);
+ }
+ 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 blocknr;
- struct buffer_head *c;
- struct btrfs_node *c_node;
- struct buffer_head *next = NULL;
+ 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]);
+ /*
+ * by releasing the path above we dropped all our locks. A balance
+ * could have added more items next to the key that used to be
+ * at the very end of the block. So, check again here and
+ * advance the path if there are now more items available.
+ */
+ if (nritems > 0 && path->slots[0] < nritems - 1) {
+ path->slots[0]++;
+ goto done;
+ }
while(level < BTRFS_MAX_LEVEL) {
if (!path->nodes[level])
return 1;
+
slot = path->slots[level] + 1;
c = path->nodes[level];
- c_node = btrfs_buffer_node(c);
- if (slot >= btrfs_header_nritems(&c_node->header)) {
+ if (slot >= btrfs_header_nritems(c)) {
level++;
+ if (level == BTRFS_MAX_LEVEL) {
+ return 1;
+ }
continue;
}
- blocknr = btrfs_node_blockptr(c_node, slot);
- if (next)
- btrfs_block_release(root, next);
- next = read_tree_block(root, blocknr);
+
+ if (next) {
+ btrfs_tree_unlock(next);
+ free_extent_buffer(next);
+ }
+
+ if (level == 1 && (path->locks[1] || path->skip_locking) &&
+ path->reada)
+ reada_for_search(root, path, level, slot, 0);
+
+ next = read_node_slot(root, c, slot);
+ if (!path->skip_locking) {
+ WARN_ON(!btrfs_tree_locked(c));
+ btrfs_tree_lock(next);
+ }
break;
}
path->slots[level] = slot;
while(1) {
level--;
c = path->nodes[level];
- btrfs_block_release(root, c);
+ if (path->locks[level])
+ btrfs_tree_unlock(c);
+ free_extent_buffer(c);
path->nodes[level] = next;
path->slots[level] = 0;
+ if (!path->skip_locking)
+ path->locks[level] = 1;
if (!level)
break;
- next = read_tree_block(root,
- btrfs_node_blockptr(btrfs_buffer_node(next), 0));
+ 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;
}
+
+/*
+ * this uses btrfs_prev_leaf to walk backwards in the tree, and keeps
+ * searching until it gets past min_objectid or finds an item of 'type'
+ *
+ * returns 0 if something is found, 1 if nothing was found and < 0 on error
+ */
+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;
+ u32 nritems;
+ 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];
+ 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;
+}
git.openpandora.org / pandora-kernel.git / blobdiff