+/*
+ * 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/module.h>
#include <linux/fs.h>
#include <linux/blkdev.h>
#include <linux/scatterlist.h>
#include <linux/swap.h>
#include <linux/radix-tree.h>
+#include <linux/writeback.h>
#include "ctree.h"
#include "disk-io.h"
#include "transaction.h"
#include "btrfs_inode.h"
+u64 bh_blocknr(struct buffer_head *bh)
+{
+ return bh->b_blocknr;
+}
+
static int check_tree_block(struct btrfs_root *root, struct buffer_head *buf)
{
struct btrfs_node *node = btrfs_buffer_node(buf);
- if (buf->b_blocknr != btrfs_header_blocknr(&node->header)) {
- BUG();
+ if (bh_blocknr(buf) != btrfs_header_blocknr(&node->header)) {
+ printk(KERN_CRIT "bh_blocknr(buf) is %llu, header is %llu\n",
+ (unsigned long long)bh_blocknr(buf),
+ (unsigned long long)btrfs_header_blocknr(&node->header));
+ return 1;
}
return 0;
}
head = page_buffers(page);
bh = head;
do {
- if (buffer_mapped(bh) && bh->b_blocknr == blocknr) {
+ if (buffer_mapped(bh) && bh_blocknr(bh) == blocknr) {
ret = bh;
get_bh(bh);
goto out_unlock;
} while (bh != head);
out_unlock:
unlock_page(page);
- if (ret) {
- touch_buffer(ret);
- }
page_cache_release(page);
return ret;
}
+int btrfs_map_bh_to_logical(struct btrfs_root *root, struct buffer_head *bh,
+ u64 logical)
+{
+ if (logical == 0) {
+ bh->b_bdev = NULL;
+ bh->b_blocknr = 0;
+ set_buffer_mapped(bh);
+ } else {
+ map_bh(bh, root->fs_info->sb, logical);
+ }
+ return 0;
+}
+
struct buffer_head *btrfs_find_create_tree_block(struct btrfs_root *root,
u64 blocknr)
{
struct buffer_head *bh;
struct buffer_head *head;
struct buffer_head *ret = NULL;
+ int err;
u64 first_block = index << (PAGE_CACHE_SHIFT - blockbits);
- page = grab_cache_page(mapping, index);
+ page = find_or_create_page(mapping, index, GFP_NOFS);
if (!page)
return NULL;
bh = head;
do {
if (!buffer_mapped(bh)) {
- bh->b_bdev = root->fs_info->sb->s_bdev;
- bh->b_blocknr = first_block;
- set_buffer_mapped(bh);
+ err = btrfs_map_bh_to_logical(root, bh, first_block);
+ BUG_ON(err);
}
- if (bh->b_blocknr == blocknr) {
+ if (bh_blocknr(bh) == blocknr) {
ret = bh;
get_bh(bh);
goto out_unlock;
return ret;
}
-static sector_t max_block(struct block_device *bdev)
-{
- sector_t retval = ~((sector_t)0);
- loff_t sz = i_size_read(bdev->bd_inode);
-
- if (sz) {
- unsigned int size = block_size(bdev);
- unsigned int sizebits = blksize_bits(size);
- retval = (sz >> sizebits);
- }
- return retval;
-}
-
static int btree_get_block(struct inode *inode, sector_t iblock,
struct buffer_head *bh, int create)
{
- if (iblock >= max_block(inode->i_sb->s_bdev)) {
- if (create)
- return -EIO;
-
- /*
- * for reads, we're just trying to fill a partial page.
- * return a hole, they will have to call get_block again
- * before they can fill it, and they will get -EIO at that
- * time
- */
- return 0;
- }
- bh->b_bdev = inode->i_sb->s_bdev;
- bh->b_blocknr = iblock;
- set_buffer_mapped(bh);
- return 0;
+ int err;
+ struct btrfs_root *root = BTRFS_I(bh->b_page->mapping->host)->root;
+ err = btrfs_map_bh_to_logical(root, bh, iblock);
+ return err;
}
int btrfs_csum_data(struct btrfs_root * root, char *data, size_t len,
ret = crypto_hash_digest(&desc, &sg, 1, result);
spin_unlock(&root->fs_info->hash_lock);
if (ret) {
- printk("sha256 digest failed\n");
+ printk("digest failed\n");
}
return ret;
}
static int csum_tree_block(struct btrfs_root *root, struct buffer_head *bh,
int verify)
{
- char result[BTRFS_CSUM_SIZE];
+ char result[BTRFS_CRC32_SIZE];
int ret;
struct btrfs_node *node;
if (ret)
return ret;
if (verify) {
- if (memcmp(bh->b_data, result, BTRFS_CSUM_SIZE)) {
- printk("checksum verify failed on %lu\n",
- bh->b_blocknr);
+ if (memcmp(bh->b_data, result, BTRFS_CRC32_SIZE)) {
+ printk("btrfs: %s checksum verify failed on %llu\n",
+ root->fs_info->sb->s_id,
+ (unsigned long long)bh_blocknr(bh));
return 1;
}
} else {
node = btrfs_buffer_node(bh);
- memcpy(node->header.csum, result, BTRFS_CSUM_SIZE);
+ memcpy(node->header.csum, result, BTRFS_CRC32_SIZE);
}
return 0;
}
.sync_page = block_sync_page,
};
+int readahead_tree_block(struct btrfs_root *root, u64 blocknr)
+{
+ struct buffer_head *bh = NULL;
+ int ret = 0;
+
+ bh = btrfs_find_create_tree_block(root, blocknr);
+ if (!bh)
+ return 0;
+ if (buffer_uptodate(bh)) {
+ ret = 1;
+ goto done;
+ }
+ if (test_set_buffer_locked(bh)) {
+ ret = 1;
+ goto done;
+ }
+ if (!buffer_uptodate(bh)) {
+ get_bh(bh);
+ bh->b_end_io = end_buffer_read_sync;
+ submit_bh(READ, bh);
+ } else {
+ unlock_buffer(bh);
+ ret = 1;
+ }
+done:
+ brelse(bh);
+ return ret;
+}
+
struct buffer_head *read_tree_block(struct btrfs_root *root, u64 blocknr)
{
struct buffer_head *bh = NULL;
wait_on_buffer(bh);
if (!buffer_uptodate(bh))
goto fail;
- csum_tree_block(root, bh, 1);
} else {
unlock_buffer(bh);
}
uptodate:
+ if (!buffer_checked(bh)) {
+ csum_tree_block(root, bh, 1);
+ set_buffer_checked(bh);
+ }
if (check_tree_block(root, bh))
- BUG();
+ goto fail;
return bh;
fail:
brelse(bh);
root->last_inode_alloc = 0;
memset(&root->root_key, 0, sizeof(root->root_key));
memset(&root->root_item, 0, sizeof(root->root_item));
+ root->root_key.objectid = objectid;
return 0;
}
u64 highest_inode;
int ret = 0;
-printk("read_fs_root looking for %Lu %Lu %u\n", location->objectid, location->offset, location->flags);
root = radix_tree_lookup(&fs_info->fs_roots_radix,
(unsigned long)location->objectid);
- if (root) {
-printk("found %p in cache\n", root);
+ if (root)
return root;
- }
root = kmalloc(sizeof(*root), GFP_NOFS);
- if (!root) {
-printk("failed1\n");
+ if (!root)
return ERR_PTR(-ENOMEM);
- }
if (location->offset == (u64)-1) {
ret = find_and_setup_root(fs_info->sb->s_blocksize,
fs_info->tree_root, fs_info,
location->objectid, root);
if (ret) {
-printk("failed2\n");
kfree(root);
return ERR_PTR(ret);
}
BUG_ON(!path);
ret = btrfs_search_slot(NULL, tree_root, location, path, 0, 0);
if (ret != 0) {
-printk("internal search_slot gives us %d\n", ret);
if (ret > 0)
ret = -ENOENT;
goto out;
btrfs_root_blocknr(&root->root_item));
BUG_ON(!root->node);
insert:
-printk("inserting %p\n", root);
root->ref_cows = 1;
ret = radix_tree_insert(&fs_info->fs_roots_radix,
(unsigned long)root->root_key.objectid,
root);
if (ret) {
-printk("radix_tree_insert gives us %d\n", ret);
brelse(root->node);
kfree(root);
return ERR_PTR(ret);
if (ret == 0) {
root->highest_inode = highest_inode;
root->last_inode_alloc = highest_inode;
-printk("highest inode is %Lu\n", highest_inode);
}
-printk("all worked\n");
return root;
}
struct btrfs_fs_info *fs_info = kmalloc(sizeof(*fs_info),
GFP_NOFS);
int ret;
+ int err = -EIO;
struct btrfs_super_block *disk_super;
+ if (!extent_root || !tree_root || !fs_info) {
+ err = -ENOMEM;
+ goto fail;
+ }
init_bit_radix(&fs_info->pinned_radix);
init_bit_radix(&fs_info->pending_del_radix);
+ init_bit_radix(&fs_info->extent_map_radix);
INIT_RADIX_TREE(&fs_info->fs_roots_radix, GFP_NOFS);
+ INIT_RADIX_TREE(&fs_info->block_group_radix, GFP_KERNEL);
+ INIT_RADIX_TREE(&fs_info->block_group_data_radix, GFP_KERNEL);
+ INIT_LIST_HEAD(&fs_info->trans_list);
+ INIT_LIST_HEAD(&fs_info->dead_roots);
sb_set_blocksize(sb, 4096);
fs_info->running_transaction = NULL;
fs_info->tree_root = tree_root;
fs_info->btree_inode->i_nlink = 1;
fs_info->btree_inode->i_size = sb->s_bdev->bd_inode->i_size;
fs_info->btree_inode->i_mapping->a_ops = &btree_aops;
+ fs_info->do_barriers = 1;
+ fs_info->extent_tree_insert_nr = 0;
+ fs_info->extent_tree_prealloc_nr = 0;
+ fs_info->closing = 0;
+
+ INIT_DELAYED_WORK(&fs_info->trans_work, btrfs_transaction_cleaner);
BTRFS_I(fs_info->btree_inode)->root = tree_root;
memset(&BTRFS_I(fs_info->btree_inode)->location, 0,
sizeof(struct btrfs_key));
insert_inode_hash(fs_info->btree_inode);
mapping_set_gfp_mask(fs_info->btree_inode->i_mapping, GFP_NOFS);
- fs_info->hash_tfm = crypto_alloc_hash("sha256", 0, CRYPTO_ALG_ASYNC);
+ fs_info->hash_tfm = crypto_alloc_hash("crc32c", 0, CRYPTO_ALG_ASYNC);
spin_lock_init(&fs_info->hash_lock);
+
if (!fs_info->hash_tfm || IS_ERR(fs_info->hash_tfm)) {
- printk("failed to allocate sha256 hash\n");
- return NULL;
+ printk("btrfs: failed hash setup, modprobe cryptomgr?\n");
+ err = -ENOMEM;
+ goto fail_iput;
}
mutex_init(&fs_info->trans_mutex);
mutex_init(&fs_info->fs_mutex);
- memset(&fs_info->current_insert, 0, sizeof(fs_info->current_insert));
- memset(&fs_info->last_insert, 0, sizeof(fs_info->last_insert));
__setup_root(sb->s_blocksize, tree_root,
fs_info, BTRFS_ROOT_TREE_OBJECTID);
+
fs_info->sb_buffer = read_tree_block(tree_root,
BTRFS_SUPER_INFO_OFFSET /
sb->s_blocksize);
if (!fs_info->sb_buffer)
- return NULL;
+ goto fail_iput;
disk_super = (struct btrfs_super_block *)fs_info->sb_buffer->b_data;
+
if (!btrfs_super_root(disk_super))
- return NULL;
+ goto fail_sb_buffer;
+
+ i_size_write(fs_info->btree_inode,
+ btrfs_super_total_blocks(disk_super) <<
+ fs_info->btree_inode->i_blkbits);
fs_info->disk_super = disk_super;
+
+ if (strncmp((char *)(&disk_super->magic), BTRFS_MAGIC,
+ sizeof(disk_super->magic))) {
+ printk("btrfs: valid FS not found on %s\n", sb->s_id);
+ goto fail_sb_buffer;
+ }
tree_root->node = read_tree_block(tree_root,
btrfs_super_root(disk_super));
- BUG_ON(!tree_root->node);
+ if (!tree_root->node)
+ goto fail_sb_buffer;
mutex_lock(&fs_info->fs_mutex);
ret = find_and_setup_root(sb->s_blocksize, tree_root, fs_info,
BTRFS_EXTENT_TREE_OBJECTID, extent_root);
- BUG_ON(ret);
+ if (ret) {
+ mutex_unlock(&fs_info->fs_mutex);
+ goto fail_tree_root;
+ }
+
+ btrfs_read_block_groups(extent_root);
fs_info->generation = btrfs_super_generation(disk_super) + 1;
- memset(&fs_info->kobj, 0, sizeof(fs_info->kobj));
- kobj_set_kset_s(fs_info, btrfs_subsys);
- kobject_set_name(&fs_info->kobj, "%s", sb->s_id);
- kobject_register(&fs_info->kobj);
mutex_unlock(&fs_info->fs_mutex);
return tree_root;
+
+fail_tree_root:
+ btrfs_block_release(tree_root, tree_root->node);
+fail_sb_buffer:
+ btrfs_block_release(tree_root, fs_info->sb_buffer);
+fail_iput:
+ iput(fs_info->btree_inode);
+fail:
+ kfree(extent_root);
+ kfree(tree_root);
+ kfree(fs_info);
+ return ERR_PTR(err);
}
int write_ctree_super(struct btrfs_trans_handle *trans, struct btrfs_root
*root)
{
+ int ret;
struct buffer_head *bh = root->fs_info->sb_buffer;
btrfs_set_super_root(root->fs_info->disk_super,
- root->fs_info->tree_root->node->b_blocknr);
+ bh_blocknr(root->fs_info->tree_root->node));
lock_buffer(bh);
WARN_ON(atomic_read(&bh->b_count) < 1);
clear_buffer_dirty(bh);
csum_tree_block(root, bh, 0);
bh->b_end_io = end_buffer_write_sync;
get_bh(bh);
- submit_bh(WRITE, bh);
+ if (root->fs_info->do_barriers)
+ ret = submit_bh(WRITE_BARRIER, bh);
+ else
+ ret = submit_bh(WRITE, bh);
+ if (ret == -EOPNOTSUPP) {
+ set_buffer_uptodate(bh);
+ root->fs_info->do_barriers = 0;
+ ret = submit_bh(WRITE, bh);
+ }
wait_on_buffer(bh);
if (!buffer_uptodate(bh)) {
WARN_ON(1);
return 0;
}
-int del_fs_roots(struct btrfs_fs_info *fs_info)
+static int del_fs_roots(struct btrfs_fs_info *fs_info)
{
int ret;
struct btrfs_root *gang[8];
struct btrfs_trans_handle *trans;
struct btrfs_fs_info *fs_info = root->fs_info;
+ fs_info->closing = 1;
+ btrfs_transaction_flush_work(root);
mutex_lock(&fs_info->fs_mutex);
trans = btrfs_start_transaction(root, 1);
btrfs_commit_transaction(trans, root);
crypto_free_hash(fs_info->hash_tfm);
truncate_inode_pages(fs_info->btree_inode->i_mapping, 0);
iput(fs_info->btree_inode);
+
+ btrfs_free_block_groups(root->fs_info);
del_fs_roots(fs_info);
kfree(fs_info->extent_root);
kfree(fs_info->tree_root);
- kobject_unregister(&fs_info->kobj);
return 0;
}
brelse(buf);
}
+void btrfs_btree_balance_dirty(struct btrfs_root *root)
+{
+ balance_dirty_pages_ratelimited(root->fs_info->btree_inode->i_mapping);
+}