#include "disk-io.h"
#include "transaction.h"
#include "btrfs_inode.h"
+#include "volumes.h"
+#include "print-tree.h"
#if 0
static int check_tree_block(struct btrfs_root *root, struct extent_buffer *buf)
}
#endif
-struct extent_buffer *btrfs_find_tree_block(struct btrfs_root *root,
- u64 blocknr)
-{
- struct inode *btree_inode = root->fs_info->btree_inode;
- struct extent_buffer *eb;
- eb = find_extent_buffer(&BTRFS_I(btree_inode)->extent_tree,
- blocknr * root->sectorsize,
- root->sectorsize, GFP_NOFS);
- if (eb)
- eb->alloc_addr = (unsigned long)__builtin_return_address(0);
- return eb;
-}
-
-struct extent_buffer *btrfs_find_create_tree_block(struct btrfs_root *root,
- u64 blocknr)
-{
- struct inode *btree_inode = root->fs_info->btree_inode;
- struct extent_buffer *eb;
- eb = alloc_extent_buffer(&BTRFS_I(btree_inode)->extent_tree,
- blocknr * root->sectorsize,
- root->sectorsize, GFP_NOFS);
- eb->alloc_addr = (unsigned long)__builtin_return_address(0);
- return eb;
-}
+static struct extent_io_ops btree_extent_io_ops;
struct extent_map *btree_get_extent(struct inode *inode, struct page *page,
- size_t page_offset, u64 start, u64 end,
+ size_t page_offset, u64 start, u64 len,
int create)
{
struct extent_map_tree *em_tree = &BTRFS_I(inode)->extent_tree;
int ret;
again:
- em = lookup_extent_mapping(em_tree, start, end);
+ spin_lock(&em_tree->lock);
+ em = lookup_extent_mapping(em_tree, start, len);
+ spin_unlock(&em_tree->lock);
if (em) {
goto out;
}
goto out;
}
em->start = 0;
- em->end = (i_size_read(inode) & ~((u64)PAGE_CACHE_SIZE -1)) - 1;
+ em->len = i_size_read(inode);
em->block_start = 0;
- em->block_end = em->end;
em->bdev = inode->i_sb->s_bdev;
+
+ spin_lock(&em_tree->lock);
ret = add_extent_mapping(em_tree, em);
+ spin_unlock(&em_tree->lock);
+
if (ret == -EEXIST) {
free_extent_map(em);
em = NULL;
return em;
}
+u32 btrfs_csum_data(struct btrfs_root *root, char *data, u32 seed, size_t len)
+{
+ return crc32c(seed, data, len);
+}
+
+void btrfs_csum_final(u32 crc, char *result)
+{
+ *(__le32 *)result = ~cpu_to_le32(crc);
+}
+
+static int csum_tree_block(struct btrfs_root *root, struct extent_buffer *buf,
+ int verify)
+{
+ char result[BTRFS_CRC32_SIZE];
+ unsigned long len;
+ unsigned long cur_len;
+ unsigned long offset = BTRFS_CSUM_SIZE;
+ char *map_token = NULL;
+ char *kaddr;
+ unsigned long map_start;
+ unsigned long map_len;
+ int err;
+ u32 crc = ~(u32)0;
+
+ len = buf->len - offset;
+ while(len > 0) {
+ err = map_private_extent_buffer(buf, offset, 32,
+ &map_token, &kaddr,
+ &map_start, &map_len, KM_USER0);
+ if (err) {
+ printk("failed to map extent buffer! %lu\n",
+ offset);
+ return 1;
+ }
+ cur_len = min(len, map_len - (offset - map_start));
+ crc = btrfs_csum_data(root, kaddr + offset - map_start,
+ crc, cur_len);
+ len -= cur_len;
+ offset += cur_len;
+ unmap_extent_buffer(buf, map_token, KM_USER0);
+ }
+ btrfs_csum_final(crc, result);
+
+ if (verify) {
+ int from_this_trans = 0;
+
+ if (root->fs_info->running_transaction &&
+ btrfs_header_generation(buf) ==
+ root->fs_info->running_transaction->transid)
+ from_this_trans = 1;
+
+ /* FIXME, this is not good */
+ if (memcmp_extent_buffer(buf, result, 0, BTRFS_CRC32_SIZE)) {
+ u32 val;
+ u32 found = 0;
+ memcpy(&found, result, BTRFS_CRC32_SIZE);
+
+ read_extent_buffer(buf, &val, 0, BTRFS_CRC32_SIZE);
+ WARN_ON(1);
+ printk("btrfs: %s checksum verify failed on %llu "
+ "wanted %X found %X from_this_trans %d "
+ "level %d\n",
+ root->fs_info->sb->s_id,
+ buf->start, val, found, from_this_trans,
+ btrfs_header_level(buf));
+ return 1;
+ }
+ } else {
+ write_extent_buffer(buf, result, 0, BTRFS_CRC32_SIZE);
+ }
+ return 0;
+}
+
+
+int csum_dirty_buffer(struct btrfs_root *root, struct page *page)
+{
+ struct extent_io_tree *tree;
+ u64 start = (u64)page->index << PAGE_CACHE_SHIFT;
+ u64 found_start;
+ int found_level;
+ unsigned long len;
+ struct extent_buffer *eb;
+ tree = &BTRFS_I(page->mapping->host)->io_tree;
+
+ if (page->private == EXTENT_PAGE_PRIVATE)
+ goto out;
+ if (!page->private)
+ goto out;
+ len = page->private >> 2;
+ if (len == 0) {
+ WARN_ON(1);
+ }
+ eb = alloc_extent_buffer(tree, start, len, page, GFP_NOFS);
+ read_extent_buffer_pages(tree, eb, start + PAGE_CACHE_SIZE, 1,
+ btree_get_extent);
+ btrfs_clear_buffer_defrag(eb);
+ found_start = btrfs_header_bytenr(eb);
+ if (found_start != start) {
+ printk("warning: eb start incorrect %Lu buffer %Lu len %lu\n",
+ start, found_start, len);
+ WARN_ON(1);
+ goto err;
+ }
+ if (eb->first_page != page) {
+ printk("bad first page %lu %lu\n", eb->first_page->index,
+ page->index);
+ WARN_ON(1);
+ goto err;
+ }
+ if (!PageUptodate(page)) {
+ printk("csum not up to date page %lu\n", page->index);
+ WARN_ON(1);
+ goto err;
+ }
+ found_level = btrfs_header_level(eb);
+ spin_lock(&root->fs_info->hash_lock);
+ btrfs_set_header_flag(eb, BTRFS_HEADER_FLAG_WRITTEN);
+ spin_unlock(&root->fs_info->hash_lock);
+ csum_tree_block(root, eb, 0);
+err:
+ free_extent_buffer(eb);
+out:
+ return 0;
+}
+
+static int btree_writepage_io_hook(struct page *page, u64 start, u64 end)
+{
+ struct btrfs_root *root = BTRFS_I(page->mapping->host)->root;
+
+ csum_dirty_buffer(root, page);
+ return 0;
+}
+
+static int btree_submit_bio_hook(struct inode *inode, int rw, struct bio *bio)
+{
+ struct btrfs_root *root = BTRFS_I(inode)->root;
+ u64 offset;
+ offset = bio->bi_sector << 9;
+ if (offset == BTRFS_SUPER_INFO_OFFSET) {
+ bio->bi_bdev = root->fs_info->sb->s_bdev;
+ submit_bio(rw, bio);
+ return 0;
+ }
+ return btrfs_map_bio(BTRFS_I(inode)->root, rw, bio);
+}
+
static int btree_writepage(struct page *page, struct writeback_control *wbc)
{
- struct extent_map_tree *tree;
- tree = &BTRFS_I(page->mapping->host)->extent_tree;
+ struct extent_io_tree *tree;
+ tree = &BTRFS_I(page->mapping->host)->io_tree;
return extent_write_full_page(tree, page, btree_get_extent, wbc);
}
+
+static int btree_writepages(struct address_space *mapping,
+ struct writeback_control *wbc)
+{
+ struct extent_io_tree *tree;
+ tree = &BTRFS_I(mapping->host)->io_tree;
+ if (wbc->sync_mode == WB_SYNC_NONE) {
+ u64 num_dirty;
+ u64 start = 0;
+ unsigned long thresh = 96 * 1024 * 1024;
+
+ if (wbc->for_kupdate)
+ return 0;
+
+ if (current_is_pdflush()) {
+ thresh = 96 * 1024 * 1024;
+ } else {
+ thresh = 8 * 1024 * 1024;
+ }
+ num_dirty = count_range_bits(tree, &start, (u64)-1,
+ thresh, EXTENT_DIRTY);
+ if (num_dirty < thresh) {
+ return 0;
+ }
+ }
+ return extent_writepages(tree, mapping, btree_get_extent, wbc);
+}
+
int btree_readpage(struct file *file, struct page *page)
{
- struct extent_map_tree *tree;
- tree = &BTRFS_I(page->mapping->host)->extent_tree;
+ struct extent_io_tree *tree;
+ tree = &BTRFS_I(page->mapping->host)->io_tree;
return extent_read_full_page(tree, page, btree_get_extent);
}
-static int btree_releasepage(struct page *page, gfp_t unused_gfp_flags)
+static int btree_releasepage(struct page *page, gfp_t gfp_flags)
{
- struct extent_map_tree *tree;
+ struct extent_io_tree *tree;
+ struct extent_map_tree *map;
int ret;
- BUG_ON(page->private != 1);
- tree = &BTRFS_I(page->mapping->host)->extent_tree;
- ret = try_release_extent_mapping(tree, page);
+ tree = &BTRFS_I(page->mapping->host)->io_tree;
+ map = &BTRFS_I(page->mapping->host)->extent_tree;
+ ret = try_release_extent_mapping(map, tree, page, gfp_flags);
if (ret == 1) {
ClearPagePrivate(page);
set_page_private(page, 0);
static void btree_invalidatepage(struct page *page, unsigned long offset)
{
- struct extent_map_tree *tree;
- tree = &BTRFS_I(page->mapping->host)->extent_tree;
+ struct extent_io_tree *tree;
+ tree = &BTRFS_I(page->mapping->host)->io_tree;
extent_invalidatepage(tree, page, offset);
btree_releasepage(page, GFP_NOFS);
}
-int btrfs_csum_data(struct btrfs_root * root, char *data, size_t len,
- char *result)
-{
- return 0;
-#if 0
- u32 crc;
- crc = crc32c(0, data, len);
- memcpy(result, &crc, BTRFS_CRC32_SIZE);
- return 0;
-#endif
-}
-
-#if 0
-static int csum_tree_block(struct btrfs_root *root, struct extent_buffer *buf,
- int verify)
-{
- return 0;
- char result[BTRFS_CRC32_SIZE];
- int ret;
- struct btrfs_node *node;
-
- ret = btrfs_csum_data(root, bh->b_data + BTRFS_CSUM_SIZE,
- bh->b_size - BTRFS_CSUM_SIZE, result);
- if (ret)
- return ret;
- if (verify) {
- 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_CRC32_SIZE);
- }
- return 0;
-}
-#endif
-
#if 0
static int btree_writepage(struct page *page, struct writeback_control *wbc)
{
static struct address_space_operations btree_aops = {
.readpage = btree_readpage,
.writepage = btree_writepage,
+ .writepages = btree_writepages,
.releasepage = btree_releasepage,
.invalidatepage = btree_invalidatepage,
.sync_page = block_sync_page,
};
-int readahead_tree_block(struct btrfs_root *root, u64 blocknr)
+int readahead_tree_block(struct btrfs_root *root, u64 bytenr, u32 blocksize)
{
struct extent_buffer *buf = NULL;
struct inode *btree_inode = root->fs_info->btree_inode;
int ret = 0;
- buf = btrfs_find_create_tree_block(root, blocknr);
+ buf = btrfs_find_create_tree_block(root, bytenr, blocksize);
if (!buf)
return 0;
- read_extent_buffer_pages(&BTRFS_I(btree_inode)->extent_tree,
- buf, 0);
+ read_extent_buffer_pages(&BTRFS_I(btree_inode)->io_tree,
+ buf, 0, 0, btree_get_extent);
free_extent_buffer(buf);
return ret;
}
-struct extent_buffer *read_tree_block(struct btrfs_root *root, u64 blocknr)
+static int close_all_devices(struct btrfs_fs_info *fs_info)
+{
+ struct list_head *list;
+ struct list_head *next;
+ struct btrfs_device *device;
+
+ list = &fs_info->fs_devices->devices;
+ list_for_each(next, list) {
+ device = list_entry(next, struct btrfs_device, dev_list);
+ if (device->bdev && device->bdev != fs_info->sb->s_bdev)
+ close_bdev_excl(device->bdev);
+ device->bdev = NULL;
+ }
+ return 0;
+}
+
+int btrfs_verify_block_csum(struct btrfs_root *root,
+ struct extent_buffer *buf)
+{
+ struct extent_io_tree *io_tree;
+ u64 end;
+ int ret;
+
+ io_tree = &BTRFS_I(root->fs_info->btree_inode)->io_tree;
+ if (buf->flags & EXTENT_CSUM)
+ return 0;
+
+ end = min_t(u64, buf->len, PAGE_CACHE_SIZE);
+ end = buf->start + end - 1;
+ if (test_range_bit(io_tree, buf->start, end, EXTENT_CSUM, 1)) {
+ buf->flags |= EXTENT_CSUM;
+ return 0;
+ }
+
+ lock_extent(io_tree, buf->start, end, GFP_NOFS);
+
+ if (test_range_bit(io_tree, buf->start, end, EXTENT_CSUM, 1)) {
+ buf->flags |= EXTENT_CSUM;
+ ret = 0;
+ goto out_unlock;
+ }
+
+ ret = csum_tree_block(root, buf, 1);
+ set_extent_bits(io_tree, buf->start, end, EXTENT_CSUM, GFP_NOFS);
+ buf->flags |= EXTENT_CSUM;
+
+out_unlock:
+ unlock_extent(io_tree, buf->start, end, GFP_NOFS);
+ return ret;
+}
+
+struct extent_buffer *btrfs_find_tree_block(struct btrfs_root *root,
+ u64 bytenr, u32 blocksize)
+{
+ struct inode *btree_inode = root->fs_info->btree_inode;
+ struct extent_buffer *eb;
+ eb = find_extent_buffer(&BTRFS_I(btree_inode)->io_tree,
+ bytenr, blocksize, GFP_NOFS);
+ return eb;
+}
+
+struct extent_buffer *btrfs_find_create_tree_block(struct btrfs_root *root,
+ u64 bytenr, u32 blocksize)
+{
+ struct inode *btree_inode = root->fs_info->btree_inode;
+ struct extent_buffer *eb;
+
+ eb = alloc_extent_buffer(&BTRFS_I(btree_inode)->io_tree,
+ bytenr, blocksize, NULL, GFP_NOFS);
+ return eb;
+}
+
+
+struct extent_buffer *read_tree_block(struct btrfs_root *root, u64 bytenr,
+ u32 blocksize)
{
struct extent_buffer *buf = NULL;
struct inode *btree_inode = root->fs_info->btree_inode;
+ struct extent_io_tree *io_tree;
+ int ret;
- buf = btrfs_find_create_tree_block(root, blocknr);
+ io_tree = &BTRFS_I(btree_inode)->io_tree;
+
+ buf = btrfs_find_create_tree_block(root, bytenr, blocksize);
if (!buf)
return NULL;
- read_extent_buffer_pages(&BTRFS_I(btree_inode)->extent_tree,
- buf, 1);
- buf->alloc_addr = (unsigned long)__builtin_return_address(0);
+ read_extent_buffer_pages(&BTRFS_I(btree_inode)->io_tree, buf, 0, 1,
+ btree_get_extent);
+
+ ret = btrfs_verify_block_csum(root, buf);
return buf;
}
struct extent_buffer *buf)
{
struct inode *btree_inode = root->fs_info->btree_inode;
- clear_extent_buffer_dirty(&BTRFS_I(btree_inode)->extent_tree, buf);
+ if (btrfs_header_generation(buf) ==
+ root->fs_info->running_transaction->transid)
+ clear_extent_buffer_dirty(&BTRFS_I(btree_inode)->io_tree,
+ buf);
return 0;
}
struct extent_buffer *buf)
{
struct inode *btree_inode = root->fs_info->btree_inode;
- wait_on_extent_buffer_writeback(&BTRFS_I(btree_inode)->extent_tree,
+ wait_on_extent_buffer_writeback(&BTRFS_I(btree_inode)->io_tree,
buf);
return 0;
}
-int set_tree_block_dirty(struct btrfs_root *root, struct extent_buffer *buf)
-{
- struct inode *btree_inode = root->fs_info->btree_inode;
- set_extent_buffer_dirty(&BTRFS_I(btree_inode)->extent_tree, buf);
- return 0;
-}
-
-static int __setup_root(int blocksize,
- struct btrfs_root *root,
+static int __setup_root(u32 nodesize, u32 leafsize, u32 sectorsize,
+ u32 stripesize, struct btrfs_root *root,
struct btrfs_fs_info *fs_info,
u64 objectid)
{
root->node = NULL;
root->inode = NULL;
root->commit_root = NULL;
- root->sectorsize = blocksize;
- root->nodesize = blocksize;
- root->leafsize = blocksize;
+ root->sectorsize = sectorsize;
+ root->nodesize = nodesize;
+ root->leafsize = leafsize;
+ root->stripesize = stripesize;
root->ref_cows = 0;
+ root->track_dirty = 0;
+
root->fs_info = fs_info;
root->objectid = objectid;
root->last_trans = 0;
root->highest_inode = 0;
root->last_inode_alloc = 0;
root->name = NULL;
+ root->in_sysfs = 0;
+
+ INIT_LIST_HEAD(&root->dirty_list);
memset(&root->root_key, 0, sizeof(root->root_key));
memset(&root->root_item, 0, sizeof(root->root_item));
memset(&root->defrag_progress, 0, sizeof(root->defrag_progress));
memset(&root->root_kobj, 0, sizeof(root->root_kobj));
init_completion(&root->kobj_unregister);
- init_rwsem(&root->snap_sem);
root->defrag_running = 0;
root->defrag_level = 0;
root->root_key.objectid = objectid;
return 0;
}
-static int find_and_setup_root(int blocksize,
- struct btrfs_root *tree_root,
+static int find_and_setup_root(struct btrfs_root *tree_root,
struct btrfs_fs_info *fs_info,
u64 objectid,
struct btrfs_root *root)
{
int ret;
+ u32 blocksize;
- __setup_root(blocksize, root, fs_info, objectid);
+ __setup_root(tree_root->nodesize, tree_root->leafsize,
+ tree_root->sectorsize, tree_root->stripesize,
+ root, fs_info, objectid);
ret = btrfs_find_last_root(tree_root, objectid,
&root->root_item, &root->root_key);
BUG_ON(ret);
- root->node = read_tree_block(root,
- btrfs_root_blocknr(&root->root_item));
+ blocksize = btrfs_level_size(root, btrfs_root_level(&root->root_item));
+ root->node = read_tree_block(root, btrfs_root_bytenr(&root->root_item),
+ blocksize);
BUG_ON(!root->node);
return 0;
}
struct btrfs_path *path;
struct extent_buffer *l;
u64 highest_inode;
+ u32 blocksize;
int ret = 0;
root = kzalloc(sizeof(*root), GFP_NOFS);
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,
+ ret = find_and_setup_root(tree_root, fs_info,
location->objectid, root);
if (ret) {
kfree(root);
goto insert;
}
- __setup_root(fs_info->sb->s_blocksize, root, fs_info,
- location->objectid);
+ __setup_root(tree_root->nodesize, tree_root->leafsize,
+ tree_root->sectorsize, tree_root->stripesize,
+ root, fs_info, location->objectid);
path = btrfs_alloc_path();
BUG_ON(!path);
read_extent_buffer(l, &root->root_item,
btrfs_item_ptr_offset(l, path->slots[0]),
sizeof(root->root_item));
+ memcpy(&root->root_key, location, sizeof(*location));
ret = 0;
out:
btrfs_release_path(root, path);
kfree(root);
return ERR_PTR(ret);
}
- root->node = read_tree_block(root,
- btrfs_root_blocknr(&root->root_item));
+ blocksize = btrfs_level_size(root, btrfs_root_level(&root->root_item));
+ root->node = read_tree_block(root, btrfs_root_bytenr(&root->root_item),
+ blocksize);
BUG_ON(!root->node);
insert:
root->ref_cows = 1;
return root;
}
-struct btrfs_root *btrfs_read_fs_root(struct btrfs_fs_info *fs_info,
- struct btrfs_key *location,
- const char *name, int namelen)
+struct btrfs_root *btrfs_lookup_fs_root(struct btrfs_fs_info *fs_info,
+ u64 root_objectid)
+{
+ struct btrfs_root *root;
+
+ if (root_objectid == BTRFS_ROOT_TREE_OBJECTID)
+ return fs_info->tree_root;
+ if (root_objectid == BTRFS_EXTENT_TREE_OBJECTID)
+ return fs_info->extent_root;
+
+ root = radix_tree_lookup(&fs_info->fs_roots_radix,
+ (unsigned long)root_objectid);
+ return root;
+}
+
+struct btrfs_root *btrfs_read_fs_root_no_name(struct btrfs_fs_info *fs_info,
+ struct btrfs_key *location)
{
struct btrfs_root *root;
int ret;
+ if (location->objectid == BTRFS_ROOT_TREE_OBJECTID)
+ return fs_info->tree_root;
+ if (location->objectid == BTRFS_EXTENT_TREE_OBJECTID)
+ return fs_info->extent_root;
+
root = radix_tree_lookup(&fs_info->fs_roots_radix,
(unsigned long)location->objectid);
if (root)
kfree(root);
return ERR_PTR(ret);
}
+ ret = btrfs_find_dead_roots(fs_info->tree_root,
+ root->root_key.objectid, root);
+ BUG_ON(ret);
+
+ return root;
+}
+
+struct btrfs_root *btrfs_read_fs_root(struct btrfs_fs_info *fs_info,
+ struct btrfs_key *location,
+ const char *name, int namelen)
+{
+ struct btrfs_root *root;
+ int ret;
+
+ root = btrfs_read_fs_root_no_name(fs_info, location);
+ if (!root)
+ return NULL;
+
+ if (root->in_sysfs)
+ return root;
ret = btrfs_set_root_name(root, name, namelen);
if (ret) {
kfree(root);
return ERR_PTR(ret);
}
+ root->in_sysfs = 1;
+ return root;
+}
+#if 0
+static int add_hasher(struct btrfs_fs_info *info, char *type) {
+ struct btrfs_hasher *hasher;
+
+ hasher = kmalloc(sizeof(*hasher), GFP_NOFS);
+ if (!hasher)
+ return -ENOMEM;
+ hasher->hash_tfm = crypto_alloc_hash(type, 0, CRYPTO_ALG_ASYNC);
+ if (!hasher->hash_tfm) {
+ kfree(hasher);
+ return -EINVAL;
+ }
+ spin_lock(&info->hash_lock);
+ list_add(&hasher->list, &info->hashers);
+ spin_unlock(&info->hash_lock);
+ return 0;
+}
+#endif
- ret = btrfs_find_dead_roots(fs_info->tree_root,
- root->root_key.objectid, root);
- BUG_ON(ret);
+static int btrfs_congested_fn(void *congested_data, int bdi_bits)
+{
+ struct btrfs_fs_info *info = (struct btrfs_fs_info *)congested_data;
+ int ret = 0;
+ struct list_head *cur;
+ struct btrfs_device *device;
+ struct backing_dev_info *bdi;
+
+ list_for_each(cur, &info->fs_devices->devices) {
+ device = list_entry(cur, struct btrfs_device, dev_list);
+ bdi = blk_get_backing_dev_info(device->bdev);
+ if (bdi && bdi_congested(bdi, bdi_bits)) {
+ ret = 1;
+ break;
+ }
+ }
+ return ret;
+}
- return root;
+void btrfs_unplug_io_fn(struct backing_dev_info *bdi, struct page *page)
+{
+ struct list_head *cur;
+ struct btrfs_device *device;
+ struct btrfs_fs_info *info;
+
+ info = (struct btrfs_fs_info *)bdi->unplug_io_data;
+ list_for_each(cur, &info->fs_devices->devices) {
+ device = list_entry(cur, struct btrfs_device, dev_list);
+ bdi = blk_get_backing_dev_info(device->bdev);
+ if (bdi->unplug_io_fn) {
+ bdi->unplug_io_fn(bdi, page);
+ }
+ }
+}
+
+static int setup_bdi(struct btrfs_fs_info *info, struct backing_dev_info *bdi)
+{
+ bdi_init(bdi);
+ bdi->ra_pages = default_backing_dev_info.ra_pages * 4;
+ bdi->state = 0;
+ bdi->capabilities = default_backing_dev_info.capabilities;
+ bdi->unplug_io_fn = btrfs_unplug_io_fn;
+ bdi->unplug_io_data = info;
+ bdi->congested_fn = btrfs_congested_fn;
+ bdi->congested_data = info;
+ return 0;
}
-struct btrfs_root *open_ctree(struct super_block *sb)
+struct btrfs_root *open_ctree(struct super_block *sb,
+ struct btrfs_fs_devices *fs_devices)
{
+ u32 sectorsize;
+ u32 nodesize;
+ u32 leafsize;
+ u32 blocksize;
+ u32 stripesize;
struct btrfs_root *extent_root = kmalloc(sizeof(struct btrfs_root),
GFP_NOFS);
struct btrfs_root *tree_root = kmalloc(sizeof(struct btrfs_root),
GFP_NOFS);
struct btrfs_fs_info *fs_info = kmalloc(sizeof(*fs_info),
GFP_NOFS);
+ struct btrfs_root *chunk_root = kmalloc(sizeof(struct btrfs_root),
+ GFP_NOFS);
+ struct btrfs_root *dev_root = kmalloc(sizeof(struct btrfs_root),
+ GFP_NOFS);
int ret;
- int err = -EIO;
+ int err = -EINVAL;
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_ins_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);
+ INIT_LIST_HEAD(&fs_info->hashers);
+ spin_lock_init(&fs_info->hash_lock);
+ spin_lock_init(&fs_info->delalloc_lock);
+ spin_lock_init(&fs_info->new_trans_lock);
+
memset(&fs_info->super_kobj, 0, sizeof(fs_info->super_kobj));
init_completion(&fs_info->kobj_unregister);
sb_set_blocksize(sb, 4096);
fs_info->last_trans_committed = 0;
fs_info->tree_root = tree_root;
fs_info->extent_root = extent_root;
+ fs_info->chunk_root = chunk_root;
+ fs_info->dev_root = dev_root;
+ fs_info->fs_devices = fs_devices;
+ INIT_LIST_HEAD(&fs_info->dirty_cowonly_roots);
+ INIT_LIST_HEAD(&fs_info->space_info);
+ btrfs_mapping_init(&fs_info->mapping_tree);
fs_info->sb = sb;
+ fs_info->throttles = 0;
+ fs_info->mount_opt = 0;
+ fs_info->max_extent = (u64)-1;
+ fs_info->max_inline = 8192 * 1024;
+ fs_info->delalloc_bytes = 0;
+ setup_bdi(fs_info, &fs_info->bdi);
fs_info->btree_inode = new_inode(sb);
fs_info->btree_inode->i_ino = 1;
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;
- extent_map_tree_init(&BTRFS_I(fs_info->btree_inode)->extent_tree,
+ fs_info->btree_inode->i_mapping->backing_dev_info = &fs_info->bdi;
+
+ extent_io_tree_init(&BTRFS_I(fs_info->btree_inode)->io_tree,
fs_info->btree_inode->i_mapping,
GFP_NOFS);
- extent_map_tree_init(&fs_info->free_space_cache,
+ extent_map_tree_init(&BTRFS_I(fs_info->btree_inode)->extent_tree,
+ GFP_NOFS);
+
+ BTRFS_I(fs_info->btree_inode)->io_tree.ops = &btree_extent_io_ops;
+
+ extent_io_tree_init(&fs_info->free_space_cache,
+ fs_info->btree_inode->i_mapping, GFP_NOFS);
+ extent_io_tree_init(&fs_info->block_group_cache,
+ fs_info->btree_inode->i_mapping, GFP_NOFS);
+ extent_io_tree_init(&fs_info->pinned_extents,
+ fs_info->btree_inode->i_mapping, GFP_NOFS);
+ extent_io_tree_init(&fs_info->pending_del,
+ fs_info->btree_inode->i_mapping, GFP_NOFS);
+ extent_io_tree_init(&fs_info->extent_ins,
fs_info->btree_inode->i_mapping, GFP_NOFS);
fs_info->do_barriers = 1;
fs_info->closing = 0;
-
+ fs_info->total_pinned = 0;
+ fs_info->last_alloc = 0;
+ fs_info->last_data_alloc = 0;
+ fs_info->extra_alloc_bits = 0;
+ fs_info->extra_data_alloc_bits = 0;
+
+#if LINUX_VERSION_CODE <= KERNEL_VERSION(2,6,18)
+ INIT_WORK(&fs_info->trans_work, btrfs_transaction_cleaner, fs_info);
+#else
INIT_DELAYED_WORK(&fs_info->trans_work, btrfs_transaction_cleaner);
+#endif
BTRFS_I(fs_info->btree_inode)->root = tree_root;
memset(&BTRFS_I(fs_info->btree_inode)->location, 0,
sizeof(struct btrfs_key));
mutex_init(&fs_info->trans_mutex);
mutex_init(&fs_info->fs_mutex);
- __setup_root(sb->s_blocksize, tree_root,
+#if 0
+ ret = add_hasher(fs_info, "crc32c");
+ if (ret) {
+ printk("btrfs: failed hash setup, modprobe cryptomgr?\n");
+ err = -ENOMEM;
+ goto fail_iput;
+ }
+#endif
+ __setup_root(4096, 4096, 4096, 4096, tree_root,
fs_info, BTRFS_ROOT_TREE_OBJECTID);
fs_info->sb_buffer = read_tree_block(tree_root,
- BTRFS_SUPER_INFO_OFFSET /
- sb->s_blocksize);
+ BTRFS_SUPER_INFO_OFFSET,
+ 4096);
if (!fs_info->sb_buffer)
goto fail_iput;
read_extent_buffer(fs_info->sb_buffer, fs_info->fsid,
(unsigned long)btrfs_super_fsid(fs_info->sb_buffer),
BTRFS_FSID_SIZE);
+
disk_super = &fs_info->super_copy;
if (!btrfs_super_root(disk_super))
goto fail_sb_buffer;
- i_size_write(fs_info->btree_inode,
- btrfs_super_total_blocks(disk_super) <<
- fs_info->btree_inode->i_blkbits);
+ if (btrfs_super_num_devices(disk_super) != fs_devices->num_devices) {
+ printk("Btrfs: wanted %llu devices, but found %llu\n",
+ (unsigned long long)btrfs_super_num_devices(disk_super),
+ (unsigned long long)fs_devices->num_devices);
+ goto fail_sb_buffer;
+ }
+ nodesize = btrfs_super_nodesize(disk_super);
+ leafsize = btrfs_super_leafsize(disk_super);
+ sectorsize = btrfs_super_sectorsize(disk_super);
+ stripesize = btrfs_super_stripesize(disk_super);
+ tree_root->nodesize = nodesize;
+ tree_root->leafsize = leafsize;
+ tree_root->sectorsize = sectorsize;
+ tree_root->stripesize = stripesize;
+ sb_set_blocksize(sb, sectorsize);
+ i_size_write(fs_info->btree_inode,
+ btrfs_super_total_bytes(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;
}
+
+ mutex_lock(&fs_info->fs_mutex);
+
+ ret = btrfs_read_sys_array(tree_root);
+ BUG_ON(ret);
+
+ blocksize = btrfs_level_size(tree_root,
+ btrfs_super_chunk_root_level(disk_super));
+
+ __setup_root(nodesize, leafsize, sectorsize, stripesize,
+ chunk_root, fs_info, BTRFS_CHUNK_TREE_OBJECTID);
+
+ chunk_root->node = read_tree_block(chunk_root,
+ btrfs_super_chunk_root(disk_super),
+ blocksize);
+ BUG_ON(!chunk_root->node);
+
+ ret = btrfs_read_chunk_tree(chunk_root);
+ BUG_ON(ret);
+
+ blocksize = btrfs_level_size(tree_root,
+ btrfs_super_root_level(disk_super));
+
+
tree_root->node = read_tree_block(tree_root,
- btrfs_super_root(disk_super));
+ btrfs_super_root(disk_super),
+ blocksize);
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,
+
+ ret = find_and_setup_root(tree_root, fs_info,
BTRFS_EXTENT_TREE_OBJECTID, extent_root);
- if (ret) {
- mutex_unlock(&fs_info->fs_mutex);
+ if (ret)
goto fail_tree_root;
- }
+ extent_root->track_dirty = 1;
+
+ ret = find_and_setup_root(tree_root, fs_info,
+ BTRFS_DEV_TREE_OBJECTID, dev_root);
+ dev_root->track_dirty = 1;
+
+ if (ret)
+ goto fail_extent_root;
btrfs_read_block_groups(extent_root);
mutex_unlock(&fs_info->fs_mutex);
return tree_root;
+fail_extent_root:
+ free_extent_buffer(extent_root->node);
fail_tree_root:
+ mutex_unlock(&fs_info->fs_mutex);
free_extent_buffer(tree_root->node);
fail_sb_buffer:
free_extent_buffer(fs_info->sb_buffer);
+ extent_io_tree_empty_lru(&BTRFS_I(fs_info->btree_inode)->io_tree);
fail_iput:
iput(fs_info->btree_inode);
fail:
+ close_all_devices(fs_info);
kfree(extent_root);
kfree(tree_root);
+ bdi_destroy(&fs_info->bdi);
kfree(fs_info);
return ERR_PTR(err);
}
int ret;
struct extent_buffer *super = root->fs_info->sb_buffer;
struct inode *btree_inode = root->fs_info->btree_inode;
+ struct super_block *sb = root->fs_info->sb;
- set_extent_buffer_dirty(&BTRFS_I(btree_inode)->extent_tree, super);
+ if (!btrfs_test_opt(root, NOBARRIER))
+ blkdev_issue_flush(sb->s_bdev, NULL);
+ set_extent_buffer_dirty(&BTRFS_I(btree_inode)->io_tree, super);
ret = sync_page_range_nolock(btree_inode, btree_inode->i_mapping,
super->start, super->len);
+ if (!btrfs_test_opt(root, NOBARRIER))
+ blkdev_issue_flush(sb->s_bdev, NULL);
return ret;
}
{
radix_tree_delete(&fs_info->fs_roots_radix,
(unsigned long)root->root_key.objectid);
- btrfs_sysfs_del_root(root);
+ if (root->in_sysfs)
+ btrfs_sysfs_del_root(root);
if (root->inode)
iput(root->inode);
if (root->node)
write_ctree_super(NULL, root);
mutex_unlock(&fs_info->fs_mutex);
+ if (fs_info->delalloc_bytes) {
+ printk("btrfs: at unmount delalloc count %Lu\n",
+ fs_info->delalloc_bytes);
+ }
if (fs_info->extent_root->node)
free_extent_buffer(fs_info->extent_root->node);
if (fs_info->tree_root->node)
free_extent_buffer(fs_info->tree_root->node);
+ if (root->fs_info->chunk_root->node);
+ free_extent_buffer(root->fs_info->chunk_root->node);
+
+ if (root->fs_info->dev_root->node);
+ free_extent_buffer(root->fs_info->dev_root->node);
+
free_extent_buffer(fs_info->sb_buffer);
- 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);
+
+ filemap_write_and_wait(fs_info->btree_inode->i_mapping);
+
+ extent_io_tree_empty_lru(&fs_info->free_space_cache);
+ extent_io_tree_empty_lru(&fs_info->block_group_cache);
+ extent_io_tree_empty_lru(&fs_info->pinned_extents);
+ extent_io_tree_empty_lru(&fs_info->pending_del);
+ extent_io_tree_empty_lru(&fs_info->extent_ins);
+ extent_io_tree_empty_lru(&BTRFS_I(fs_info->btree_inode)->io_tree);
+
+ truncate_inode_pages(fs_info->btree_inode->i_mapping, 0);
+
+ iput(fs_info->btree_inode);
+#if 0
+ while(!list_empty(&fs_info->hashers)) {
+ struct btrfs_hasher *hasher;
+ hasher = list_entry(fs_info->hashers.next, struct btrfs_hasher,
+ hashers);
+ list_del(&hasher->hashers);
+ crypto_free_hash(&fs_info->hash_tfm);
+ kfree(hasher);
+ }
+#endif
+ close_all_devices(fs_info);
+ btrfs_mapping_tree_free(&fs_info->mapping_tree);
+ bdi_destroy(&fs_info->bdi);
+
kfree(fs_info->extent_root);
kfree(fs_info->tree_root);
+ kfree(fs_info->chunk_root);
+ kfree(fs_info->dev_root);
return 0;
}
int btrfs_buffer_uptodate(struct extent_buffer *buf)
{
struct inode *btree_inode = buf->first_page->mapping->host;
- return extent_buffer_uptodate(&BTRFS_I(btree_inode)->extent_tree, buf);
+ return extent_buffer_uptodate(&BTRFS_I(btree_inode)->io_tree, buf);
}
int btrfs_set_buffer_uptodate(struct extent_buffer *buf)
{
struct inode *btree_inode = buf->first_page->mapping->host;
- return set_extent_buffer_uptodate(&BTRFS_I(btree_inode)->extent_tree,
+ return set_extent_buffer_uptodate(&BTRFS_I(btree_inode)->io_tree,
buf);
}
if (transid != root->fs_info->generation) {
printk(KERN_CRIT "transid mismatch buffer %llu, found %Lu running %Lu\n",
- (unsigned long long)extent_buffer_blocknr(buf),
+ (unsigned long long)buf->start,
transid, root->fs_info->generation);
WARN_ON(1);
}
- set_extent_buffer_dirty(&BTRFS_I(btree_inode)->extent_tree, buf);
+ set_extent_buffer_dirty(&BTRFS_I(btree_inode)->io_tree, buf);
+}
+
+void btrfs_throttle(struct btrfs_root *root)
+{
+ struct backing_dev_info *bdi;
+
+ bdi = root->fs_info->sb->s_bdev->bd_inode->i_mapping->backing_dev_info;
+ if (root->fs_info->throttles && bdi_write_congested(bdi)) {
+#if LINUX_VERSION_CODE > KERNEL_VERSION(2,6,18)
+ congestion_wait(WRITE, HZ/20);
+#else
+ blk_congestion_wait(WRITE, HZ/20);
+#endif
+ }
}
void btrfs_btree_balance_dirty(struct btrfs_root *root, unsigned long nr)
{
balance_dirty_pages_ratelimited_nr(
- root->fs_info->btree_inode->i_mapping, nr);
+ root->fs_info->btree_inode->i_mapping, 1);
+}
+
+void btrfs_set_buffer_defrag(struct extent_buffer *buf)
+{
+ struct btrfs_root *root = BTRFS_I(buf->first_page->mapping->host)->root;
+ struct inode *btree_inode = root->fs_info->btree_inode;
+ set_extent_bits(&BTRFS_I(btree_inode)->io_tree, buf->start,
+ buf->start + buf->len - 1, EXTENT_DEFRAG, GFP_NOFS);
+}
+
+void btrfs_set_buffer_defrag_done(struct extent_buffer *buf)
+{
+ struct btrfs_root *root = BTRFS_I(buf->first_page->mapping->host)->root;
+ struct inode *btree_inode = root->fs_info->btree_inode;
+ set_extent_bits(&BTRFS_I(btree_inode)->io_tree, buf->start,
+ buf->start + buf->len - 1, EXTENT_DEFRAG_DONE,
+ GFP_NOFS);
+}
+
+int btrfs_buffer_defrag(struct extent_buffer *buf)
+{
+ struct btrfs_root *root = BTRFS_I(buf->first_page->mapping->host)->root;
+ struct inode *btree_inode = root->fs_info->btree_inode;
+ return test_range_bit(&BTRFS_I(btree_inode)->io_tree,
+ buf->start, buf->start + buf->len - 1, EXTENT_DEFRAG, 0);
+}
+
+int btrfs_buffer_defrag_done(struct extent_buffer *buf)
+{
+ struct btrfs_root *root = BTRFS_I(buf->first_page->mapping->host)->root;
+ struct inode *btree_inode = root->fs_info->btree_inode;
+ return test_range_bit(&BTRFS_I(btree_inode)->io_tree,
+ buf->start, buf->start + buf->len - 1,
+ EXTENT_DEFRAG_DONE, 0);
+}
+
+int btrfs_clear_buffer_defrag_done(struct extent_buffer *buf)
+{
+ struct btrfs_root *root = BTRFS_I(buf->first_page->mapping->host)->root;
+ struct inode *btree_inode = root->fs_info->btree_inode;
+ return clear_extent_bits(&BTRFS_I(btree_inode)->io_tree,
+ buf->start, buf->start + buf->len - 1,
+ EXTENT_DEFRAG_DONE, GFP_NOFS);
}
+
+int btrfs_clear_buffer_defrag(struct extent_buffer *buf)
+{
+ struct btrfs_root *root = BTRFS_I(buf->first_page->mapping->host)->root;
+ struct inode *btree_inode = root->fs_info->btree_inode;
+ return clear_extent_bits(&BTRFS_I(btree_inode)->io_tree,
+ buf->start, buf->start + buf->len - 1,
+ EXTENT_DEFRAG, GFP_NOFS);
+}
+
+int btrfs_read_buffer(struct extent_buffer *buf)
+{
+ struct btrfs_root *root = BTRFS_I(buf->first_page->mapping->host)->root;
+ struct inode *btree_inode = root->fs_info->btree_inode;
+ return read_extent_buffer_pages(&BTRFS_I(btree_inode)->io_tree,
+ buf, 0, 1, btree_get_extent);
+}
+
+static struct extent_io_ops btree_extent_io_ops = {
+ .writepage_io_hook = btree_writepage_io_hook,
+ .submit_bio_hook = btree_submit_bio_hook,
+ /* note we're sharing with inode.c for the merge bio hook */
+ .merge_bio_hook = btrfs_merge_bio_hook,
+};
projects / linux-2.6-block.git / blobdiff