#include <linux/xattr.h>
#include <linux/posix_acl.h>
#include <linux/falloc.h>
+#include <linux/slab.h>
#include "compat.h"
#include "ctree.h"
#include "disk-io.h"
inline_len, compressed_size,
compressed_pages);
BUG_ON(ret);
+ btrfs_delalloc_release_metadata(inode, end + 1 - start);
btrfs_drop_extent_cache(inode, start, aligned_end - 1, 0);
return 0;
}
* change at any time if we discover bad compression ratios.
*/
if (!(BTRFS_I(inode)->flags & BTRFS_INODE_NOCOMPRESS) &&
- btrfs_test_opt(root, COMPRESS)) {
+ (btrfs_test_opt(root, COMPRESS) ||
+ (BTRFS_I(inode)->force_compress))) {
WARN_ON(pages);
pages = kzalloc(sizeof(struct page *) * nr_pages, GFP_NOFS);
trans = btrfs_join_transaction(root, 1);
BUG_ON(!trans);
btrfs_set_trans_block_group(trans, inode);
+ trans->block_rsv = &root->fs_info->delalloc_block_rsv;
/* lets try to make an inline extent */
if (ret || total_in < (actual_end - start)) {
start, end, NULL,
EXTENT_CLEAR_UNLOCK_PAGE | EXTENT_CLEAR_DIRTY |
EXTENT_CLEAR_DELALLOC |
- EXTENT_CLEAR_ACCOUNTING |
EXTENT_SET_WRITEBACK | EXTENT_END_WRITEBACK);
btrfs_end_transaction(trans, root);
nr_pages_ret = 0;
/* flag the file so we don't compress in the future */
- BTRFS_I(inode)->flags |= BTRFS_INODE_NOCOMPRESS;
+ if (!btrfs_test_opt(root, FORCE_COMPRESS) &&
+ !(BTRFS_I(inode)->force_compress)) {
+ BTRFS_I(inode)->flags |= BTRFS_INODE_NOCOMPRESS;
+ }
}
if (will_compress) {
*num_added += 1;
unsigned long nr_written = 0;
lock_extent(io_tree, async_extent->start,
- async_extent->start +
- async_extent->ram_size - 1, GFP_NOFS);
+ async_extent->start +
+ async_extent->ram_size - 1, GFP_NOFS);
/* allocate blocks */
ret = cow_file_range(inode, async_cow->locked_page,
trans = btrfs_join_transaction(root, 1);
BUG_ON(!trans);
btrfs_set_trans_block_group(trans, inode);
+ trans->block_rsv = &root->fs_info->delalloc_block_rsv;
actual_end = min_t(u64, isize, end + 1);
EXTENT_CLEAR_UNLOCK_PAGE |
EXTENT_CLEAR_UNLOCK |
EXTENT_CLEAR_DELALLOC |
- EXTENT_CLEAR_ACCOUNTING |
EXTENT_CLEAR_DIRTY |
EXTENT_SET_WRITEBACK |
EXTENT_END_WRITEBACK);
while (disk_num_bytes > 0) {
unsigned long op;
- cur_alloc_size = min(disk_num_bytes, root->fs_info->max_extent);
+ cur_alloc_size = disk_num_bytes;
ret = btrfs_reserve_extent(trans, root, cur_alloc_size,
root->sectorsize, 0, alloc_hint,
(u64)-1, &ins, 1);
num_bytes, num_bytes, type);
BUG_ON(ret);
+ if (root->root_key.objectid ==
+ BTRFS_DATA_RELOC_TREE_OBJECTID) {
+ ret = btrfs_reloc_clone_csums(inode, cur_offset,
+ num_bytes);
+ BUG_ON(ret);
+ }
+
extent_clear_unlock_delalloc(inode, &BTRFS_I(inode)->io_tree,
cur_offset, cur_offset + num_bytes - 1,
locked_page, EXTENT_CLEAR_UNLOCK_PAGE |
else if (BTRFS_I(inode)->flags & BTRFS_INODE_PREALLOC)
ret = run_delalloc_nocow(inode, locked_page, start, end,
page_started, 0, nr_written);
- else if (!btrfs_test_opt(root, COMPRESS))
+ else if (!btrfs_test_opt(root, COMPRESS) &&
+ !(BTRFS_I(inode)->force_compress))
ret = cow_file_range(inode, locked_page, start, end,
page_started, nr_written, 1);
else
}
static int btrfs_split_extent_hook(struct inode *inode,
- struct extent_state *orig, u64 split)
+ struct extent_state *orig, u64 split)
{
- struct btrfs_root *root = BTRFS_I(inode)->root;
- u64 size;
-
+ /* not delalloc, ignore it */
if (!(orig->state & EXTENT_DELALLOC))
return 0;
- size = orig->end - orig->start + 1;
- if (size > root->fs_info->max_extent) {
- u64 num_extents;
- u64 new_size;
-
- new_size = orig->end - split + 1;
- num_extents = div64_u64(size + root->fs_info->max_extent - 1,
- root->fs_info->max_extent);
-
- /*
- * if we break a large extent up then leave oustanding_extents
- * be, since we've already accounted for the large extent.
- */
- if (div64_u64(new_size + root->fs_info->max_extent - 1,
- root->fs_info->max_extent) < num_extents)
- return 0;
- }
-
- spin_lock(&BTRFS_I(inode)->accounting_lock);
- BTRFS_I(inode)->outstanding_extents++;
- spin_unlock(&BTRFS_I(inode)->accounting_lock);
-
+ atomic_inc(&BTRFS_I(inode)->outstanding_extents);
return 0;
}
struct extent_state *new,
struct extent_state *other)
{
- struct btrfs_root *root = BTRFS_I(inode)->root;
- u64 new_size, old_size;
- u64 num_extents;
-
/* not delalloc, ignore it */
if (!(other->state & EXTENT_DELALLOC))
return 0;
- old_size = other->end - other->start + 1;
- if (new->start < other->start)
- new_size = other->end - new->start + 1;
- else
- new_size = new->end - other->start + 1;
-
- /* we're not bigger than the max, unreserve the space and go */
- if (new_size <= root->fs_info->max_extent) {
- spin_lock(&BTRFS_I(inode)->accounting_lock);
- BTRFS_I(inode)->outstanding_extents--;
- spin_unlock(&BTRFS_I(inode)->accounting_lock);
- return 0;
- }
-
- /*
- * If we grew by another max_extent, just return, we want to keep that
- * reserved amount.
- */
- num_extents = div64_u64(old_size + root->fs_info->max_extent - 1,
- root->fs_info->max_extent);
- if (div64_u64(new_size + root->fs_info->max_extent - 1,
- root->fs_info->max_extent) > num_extents)
- return 0;
-
- spin_lock(&BTRFS_I(inode)->accounting_lock);
- BTRFS_I(inode)->outstanding_extents--;
- spin_unlock(&BTRFS_I(inode)->accounting_lock);
-
+ atomic_dec(&BTRFS_I(inode)->outstanding_extents);
return 0;
}
* bytes in this file, and to maintain the list of inodes that
* have pending delalloc work to be done.
*/
-static int btrfs_set_bit_hook(struct inode *inode, u64 start, u64 end,
- unsigned long old, unsigned long bits)
+static int btrfs_set_bit_hook(struct inode *inode,
+ struct extent_state *state, int *bits)
{
/*
* but in this case, we are only testeing for the DELALLOC
* bit, which is only set or cleared with irqs on
*/
- if (!(old & EXTENT_DELALLOC) && (bits & EXTENT_DELALLOC)) {
+ if (!(state->state & EXTENT_DELALLOC) && (*bits & EXTENT_DELALLOC)) {
struct btrfs_root *root = BTRFS_I(inode)->root;
+ u64 len = state->end + 1 - state->start;
+
+ if (*bits & EXTENT_FIRST_DELALLOC)
+ *bits &= ~EXTENT_FIRST_DELALLOC;
+ else
+ atomic_inc(&BTRFS_I(inode)->outstanding_extents);
- spin_lock(&BTRFS_I(inode)->accounting_lock);
- BTRFS_I(inode)->outstanding_extents++;
- spin_unlock(&BTRFS_I(inode)->accounting_lock);
- btrfs_delalloc_reserve_space(root, inode, end - start + 1);
spin_lock(&root->fs_info->delalloc_lock);
- BTRFS_I(inode)->delalloc_bytes += end - start + 1;
- root->fs_info->delalloc_bytes += end - start + 1;
+ BTRFS_I(inode)->delalloc_bytes += len;
+ root->fs_info->delalloc_bytes += len;
if (list_empty(&BTRFS_I(inode)->delalloc_inodes)) {
list_add_tail(&BTRFS_I(inode)->delalloc_inodes,
&root->fs_info->delalloc_inodes);
* extent_io.c clear_bit_hook, see set_bit_hook for why
*/
static int btrfs_clear_bit_hook(struct inode *inode,
- struct extent_state *state, unsigned long bits)
+ struct extent_state *state, int *bits)
{
/*
* set_bit and clear bit hooks normally require _irqsave/restore
* but in this case, we are only testeing for the DELALLOC
* bit, which is only set or cleared with irqs on
*/
- if ((state->state & EXTENT_DELALLOC) && (bits & EXTENT_DELALLOC)) {
+ if ((state->state & EXTENT_DELALLOC) && (*bits & EXTENT_DELALLOC)) {
struct btrfs_root *root = BTRFS_I(inode)->root;
+ u64 len = state->end + 1 - state->start;
- if (bits & EXTENT_DO_ACCOUNTING) {
- spin_lock(&BTRFS_I(inode)->accounting_lock);
- BTRFS_I(inode)->outstanding_extents--;
- spin_unlock(&BTRFS_I(inode)->accounting_lock);
- btrfs_unreserve_metadata_for_delalloc(root, inode, 1);
- }
+ if (*bits & EXTENT_FIRST_DELALLOC)
+ *bits &= ~EXTENT_FIRST_DELALLOC;
+ else if (!(*bits & EXTENT_DO_ACCOUNTING))
+ atomic_dec(&BTRFS_I(inode)->outstanding_extents);
+
+ if (*bits & EXTENT_DO_ACCOUNTING)
+ btrfs_delalloc_release_metadata(inode, len);
+
+ if (root->root_key.objectid != BTRFS_DATA_RELOC_TREE_OBJECTID)
+ btrfs_free_reserved_data_space(inode, len);
spin_lock(&root->fs_info->delalloc_lock);
- if (state->end - state->start + 1 >
- root->fs_info->delalloc_bytes) {
- printk(KERN_INFO "btrfs warning: delalloc account "
- "%llu %llu\n",
- (unsigned long long)
- state->end - state->start + 1,
- (unsigned long long)
- root->fs_info->delalloc_bytes);
- btrfs_delalloc_free_space(root, inode, (u64)-1);
- root->fs_info->delalloc_bytes = 0;
- BTRFS_I(inode)->delalloc_bytes = 0;
- } else {
- btrfs_delalloc_free_space(root, inode,
- state->end -
- state->start + 1);
- root->fs_info->delalloc_bytes -= state->end -
- state->start + 1;
- BTRFS_I(inode)->delalloc_bytes -= state->end -
- state->start + 1;
- }
+ root->fs_info->delalloc_bytes -= len;
+ BTRFS_I(inode)->delalloc_bytes -= len;
+
if (BTRFS_I(inode)->delalloc_bytes == 0 &&
!list_empty(&BTRFS_I(inode)->delalloc_inodes)) {
list_del_init(&BTRFS_I(inode)->delalloc_inodes);
return 0;
}
-int btrfs_set_extent_delalloc(struct inode *inode, u64 start, u64 end)
+int btrfs_set_extent_delalloc(struct inode *inode, u64 start, u64 end,
+ struct extent_state **cached_state)
{
if ((end & (PAGE_CACHE_SIZE - 1)) == 0)
WARN_ON(1);
return set_extent_delalloc(&BTRFS_I(inode)->io_tree, start, end,
- GFP_NOFS);
+ cached_state, GFP_NOFS);
}
/* see btrfs_writepage_start_hook for details on why this is required */
{
struct btrfs_writepage_fixup *fixup;
struct btrfs_ordered_extent *ordered;
+ struct extent_state *cached_state = NULL;
struct page *page;
struct inode *inode;
u64 page_start;
page_start = page_offset(page);
page_end = page_offset(page) + PAGE_CACHE_SIZE - 1;
- lock_extent(&BTRFS_I(inode)->io_tree, page_start, page_end, GFP_NOFS);
+ lock_extent_bits(&BTRFS_I(inode)->io_tree, page_start, page_end, 0,
+ &cached_state, GFP_NOFS);
/* already ordered? We're done */
if (PagePrivate2(page))
ordered = btrfs_lookup_ordered_extent(inode, page_start);
if (ordered) {
- unlock_extent(&BTRFS_I(inode)->io_tree, page_start,
- page_end, GFP_NOFS);
+ unlock_extent_cached(&BTRFS_I(inode)->io_tree, page_start,
+ page_end, &cached_state, GFP_NOFS);
unlock_page(page);
btrfs_start_ordered_extent(inode, ordered, 1);
goto again;
}
- btrfs_set_extent_delalloc(inode, page_start, page_end);
+ BUG();
+ btrfs_set_extent_delalloc(inode, page_start, page_end, &cached_state);
ClearPageChecked(page);
out:
- unlock_extent(&BTRFS_I(inode)->io_tree, page_start, page_end, GFP_NOFS);
+ unlock_extent_cached(&BTRFS_I(inode)->io_tree, page_start, page_end,
+ &cached_state, GFP_NOFS);
out_page:
unlock_page(page);
page_cache_release(page);
* before we start the transaction. It limits the amount of btree
* reads required while inside the transaction.
*/
-static noinline void reada_csum(struct btrfs_root *root,
- struct btrfs_path *path,
- struct btrfs_ordered_extent *ordered_extent)
-{
- struct btrfs_ordered_sum *sum;
- u64 bytenr;
-
- sum = list_entry(ordered_extent->list.next, struct btrfs_ordered_sum,
- list);
- bytenr = sum->sums[0].bytenr;
-
- /*
- * we don't care about the results, the point of this search is
- * just to get the btree leaves into ram
- */
- btrfs_lookup_csum(NULL, root->fs_info->csum_root, path, bytenr, 0);
-}
-
/* as ordered data IO finishes, this gets called so we can finish
* an ordered extent if the range of bytes in the file it covers are
* fully written.
static int btrfs_finish_ordered_io(struct inode *inode, u64 start, u64 end)
{
struct btrfs_root *root = BTRFS_I(inode)->root;
- struct btrfs_trans_handle *trans;
+ struct btrfs_trans_handle *trans = NULL;
struct btrfs_ordered_extent *ordered_extent = NULL;
struct extent_io_tree *io_tree = &BTRFS_I(inode)->io_tree;
- struct btrfs_path *path;
+ struct extent_state *cached_state = NULL;
int compressed = 0;
int ret;
- ret = btrfs_dec_test_ordered_pending(inode, start, end - start + 1);
+ ret = btrfs_dec_test_ordered_pending(inode, &ordered_extent, start,
+ end - start + 1);
if (!ret)
return 0;
-
- /*
- * before we join the transaction, try to do some of our IO.
- * This will limit the amount of IO that we have to do with
- * the transaction running. We're unlikely to need to do any
- * IO if the file extents are new, the disk_i_size checks
- * covers the most common case.
- */
- if (start < BTRFS_I(inode)->disk_i_size) {
- path = btrfs_alloc_path();
- if (path) {
- ret = btrfs_lookup_file_extent(NULL, root, path,
- inode->i_ino,
- start, 0);
- ordered_extent = btrfs_lookup_ordered_extent(inode,
- start);
- if (!list_empty(&ordered_extent->list)) {
- btrfs_release_path(root, path);
- reada_csum(root, path, ordered_extent);
- }
- btrfs_free_path(path);
- }
- }
-
- if (!ordered_extent)
- ordered_extent = btrfs_lookup_ordered_extent(inode, start);
BUG_ON(!ordered_extent);
+
if (test_bit(BTRFS_ORDERED_NOCOW, &ordered_extent->flags)) {
BUG_ON(!list_empty(&ordered_extent->list));
ret = btrfs_ordered_update_i_size(inode, 0, ordered_extent);
if (!ret) {
trans = btrfs_join_transaction(root, 1);
+ btrfs_set_trans_block_group(trans, inode);
+ trans->block_rsv = &root->fs_info->delalloc_block_rsv;
ret = btrfs_update_inode(trans, root, inode);
BUG_ON(ret);
- btrfs_end_transaction(trans, root);
}
goto out;
}
- lock_extent(io_tree, ordered_extent->file_offset,
- ordered_extent->file_offset + ordered_extent->len - 1,
- GFP_NOFS);
+ lock_extent_bits(io_tree, ordered_extent->file_offset,
+ ordered_extent->file_offset + ordered_extent->len - 1,
+ 0, &cached_state, GFP_NOFS);
trans = btrfs_join_transaction(root, 1);
+ btrfs_set_trans_block_group(trans, inode);
+ trans->block_rsv = &root->fs_info->delalloc_block_rsv;
if (test_bit(BTRFS_ORDERED_COMPRESSED, &ordered_extent->flags))
compressed = 1;
ordered_extent->len);
BUG_ON(ret);
}
- unlock_extent(io_tree, ordered_extent->file_offset,
- ordered_extent->file_offset + ordered_extent->len - 1,
- GFP_NOFS);
+ unlock_extent_cached(io_tree, ordered_extent->file_offset,
+ ordered_extent->file_offset +
+ ordered_extent->len - 1, &cached_state, GFP_NOFS);
+
add_pending_csums(trans, inode, ordered_extent->file_offset,
&ordered_extent->list);
- /* this also removes the ordered extent from the tree */
btrfs_ordered_update_i_size(inode, 0, ordered_extent);
ret = btrfs_update_inode(trans, root, inode);
BUG_ON(ret);
- btrfs_end_transaction(trans, root);
out:
+ btrfs_delalloc_release_metadata(inode, ordered_extent->len);
+ if (trans)
+ btrfs_end_transaction(trans, root);
/* once for us */
btrfs_put_ordered_extent(ordered_extent);
/* once for the tree */
}
/*
+ * calculate extra metadata reservation when snapshotting a subvolume
+ * contains orphan files.
+ */
+void btrfs_orphan_pre_snapshot(struct btrfs_trans_handle *trans,
+ struct btrfs_pending_snapshot *pending,
+ u64 *bytes_to_reserve)
+{
+ struct btrfs_root *root;
+ struct btrfs_block_rsv *block_rsv;
+ u64 num_bytes;
+ int index;
+
+ root = pending->root;
+ if (!root->orphan_block_rsv || list_empty(&root->orphan_list))
+ return;
+
+ block_rsv = root->orphan_block_rsv;
+
+ /* orphan block reservation for the snapshot */
+ num_bytes = block_rsv->size;
+
+ /*
+ * after the snapshot is created, COWing tree blocks may use more
+ * space than it frees. So we should make sure there is enough
+ * reserved space.
+ */
+ index = trans->transid & 0x1;
+ if (block_rsv->reserved + block_rsv->freed[index] < block_rsv->size) {
+ num_bytes += block_rsv->size -
+ (block_rsv->reserved + block_rsv->freed[index]);
+ }
+
+ *bytes_to_reserve += num_bytes;
+}
+
+void btrfs_orphan_post_snapshot(struct btrfs_trans_handle *trans,
+ struct btrfs_pending_snapshot *pending)
+{
+ struct btrfs_root *root = pending->root;
+ struct btrfs_root *snap = pending->snap;
+ struct btrfs_block_rsv *block_rsv;
+ u64 num_bytes;
+ int index;
+ int ret;
+
+ if (!root->orphan_block_rsv || list_empty(&root->orphan_list))
+ return;
+
+ /* refill source subvolume's orphan block reservation */
+ block_rsv = root->orphan_block_rsv;
+ index = trans->transid & 0x1;
+ if (block_rsv->reserved + block_rsv->freed[index] < block_rsv->size) {
+ num_bytes = block_rsv->size -
+ (block_rsv->reserved + block_rsv->freed[index]);
+ ret = btrfs_block_rsv_migrate(&pending->block_rsv,
+ root->orphan_block_rsv,
+ num_bytes);
+ BUG_ON(ret);
+ }
+
+ /* setup orphan block reservation for the snapshot */
+ block_rsv = btrfs_alloc_block_rsv(snap);
+ BUG_ON(!block_rsv);
+
+ btrfs_add_durable_block_rsv(root->fs_info, block_rsv);
+ snap->orphan_block_rsv = block_rsv;
+
+ num_bytes = root->orphan_block_rsv->size;
+ ret = btrfs_block_rsv_migrate(&pending->block_rsv,
+ block_rsv, num_bytes);
+ BUG_ON(ret);
+
+#if 0
+ /* insert orphan item for the snapshot */
+ WARN_ON(!root->orphan_item_inserted);
+ ret = btrfs_insert_orphan_item(trans, root->fs_info->tree_root,
+ snap->root_key.objectid);
+ BUG_ON(ret);
+ snap->orphan_item_inserted = 1;
+#endif
+}
+
+enum btrfs_orphan_cleanup_state {
+ ORPHAN_CLEANUP_STARTED = 1,
+ ORPHAN_CLEANUP_DONE = 2,
+};
+
+/*
+ * This is called in transaction commmit time. If there are no orphan
+ * files in the subvolume, it removes orphan item and frees block_rsv
+ * structure.
+ */
+void btrfs_orphan_commit_root(struct btrfs_trans_handle *trans,
+ struct btrfs_root *root)
+{
+ int ret;
+
+ if (!list_empty(&root->orphan_list) ||
+ root->orphan_cleanup_state != ORPHAN_CLEANUP_DONE)
+ return;
+
+ if (root->orphan_item_inserted &&
+ btrfs_root_refs(&root->root_item) > 0) {
+ ret = btrfs_del_orphan_item(trans, root->fs_info->tree_root,
+ root->root_key.objectid);
+ BUG_ON(ret);
+ root->orphan_item_inserted = 0;
+ }
+
+ if (root->orphan_block_rsv) {
+ WARN_ON(root->orphan_block_rsv->size > 0);
+ btrfs_free_block_rsv(root, root->orphan_block_rsv);
+ root->orphan_block_rsv = NULL;
+ }
+}
+
+/*
* This creates an orphan entry for the given inode in case something goes
* wrong in the middle of an unlink/truncate.
+ *
+ * NOTE: caller of this function should reserve 5 units of metadata for
+ * this function.
*/
int btrfs_orphan_add(struct btrfs_trans_handle *trans, struct inode *inode)
{
struct btrfs_root *root = BTRFS_I(inode)->root;
- int ret = 0;
+ struct btrfs_block_rsv *block_rsv = NULL;
+ int reserve = 0;
+ int insert = 0;
+ int ret;
- spin_lock(&root->list_lock);
+ if (!root->orphan_block_rsv) {
+ block_rsv = btrfs_alloc_block_rsv(root);
+ BUG_ON(!block_rsv);
+ }
- /* already on the orphan list, we're good */
- if (!list_empty(&BTRFS_I(inode)->i_orphan)) {
- spin_unlock(&root->list_lock);
- return 0;
+ spin_lock(&root->orphan_lock);
+ if (!root->orphan_block_rsv) {
+ root->orphan_block_rsv = block_rsv;
+ } else if (block_rsv) {
+ btrfs_free_block_rsv(root, block_rsv);
+ block_rsv = NULL;
}
- list_add(&BTRFS_I(inode)->i_orphan, &root->orphan_list);
+ if (list_empty(&BTRFS_I(inode)->i_orphan)) {
+ list_add(&BTRFS_I(inode)->i_orphan, &root->orphan_list);
+#if 0
+ /*
+ * For proper ENOSPC handling, we should do orphan
+ * cleanup when mounting. But this introduces backward
+ * compatibility issue.
+ */
+ if (!xchg(&root->orphan_item_inserted, 1))
+ insert = 2;
+ else
+ insert = 1;
+#endif
+ insert = 1;
+ } else {
+ WARN_ON(!BTRFS_I(inode)->orphan_meta_reserved);
+ }
- spin_unlock(&root->list_lock);
+ if (!BTRFS_I(inode)->orphan_meta_reserved) {
+ BTRFS_I(inode)->orphan_meta_reserved = 1;
+ reserve = 1;
+ }
+ spin_unlock(&root->orphan_lock);
- /*
- * insert an orphan item to track this unlinked/truncated file
- */
- ret = btrfs_insert_orphan_item(trans, root, inode->i_ino);
+ if (block_rsv)
+ btrfs_add_durable_block_rsv(root->fs_info, block_rsv);
- return ret;
+ /* grab metadata reservation from transaction handle */
+ if (reserve) {
+ ret = btrfs_orphan_reserve_metadata(trans, inode);
+ BUG_ON(ret);
+ }
+
+ /* insert an orphan item to track this unlinked/truncated file */
+ if (insert >= 1) {
+ ret = btrfs_insert_orphan_item(trans, root, inode->i_ino);
+ BUG_ON(ret);
+ }
+
+ /* insert an orphan item to track subvolume contains orphan files */
+ if (insert >= 2) {
+ ret = btrfs_insert_orphan_item(trans, root->fs_info->tree_root,
+ root->root_key.objectid);
+ BUG_ON(ret);
+ }
+ return 0;
}
/*
int btrfs_orphan_del(struct btrfs_trans_handle *trans, struct inode *inode)
{
struct btrfs_root *root = BTRFS_I(inode)->root;
+ int delete_item = 0;
+ int release_rsv = 0;
int ret = 0;
- spin_lock(&root->list_lock);
-
- if (list_empty(&BTRFS_I(inode)->i_orphan)) {
- spin_unlock(&root->list_lock);
- return 0;
+ spin_lock(&root->orphan_lock);
+ if (!list_empty(&BTRFS_I(inode)->i_orphan)) {
+ list_del_init(&BTRFS_I(inode)->i_orphan);
+ delete_item = 1;
}
- list_del_init(&BTRFS_I(inode)->i_orphan);
- if (!trans) {
- spin_unlock(&root->list_lock);
- return 0;
+ if (BTRFS_I(inode)->orphan_meta_reserved) {
+ BTRFS_I(inode)->orphan_meta_reserved = 0;
+ release_rsv = 1;
}
+ spin_unlock(&root->orphan_lock);
- spin_unlock(&root->list_lock);
+ if (trans && delete_item) {
+ ret = btrfs_del_orphan_item(trans, root, inode->i_ino);
+ BUG_ON(ret);
+ }
- ret = btrfs_del_orphan_item(trans, root, inode->i_ino);
+ if (release_rsv)
+ btrfs_orphan_release_metadata(inode);
- return ret;
+ return 0;
}
/*
struct inode *inode;
int ret = 0, nr_unlink = 0, nr_truncate = 0;
- if (!xchg(&root->clean_orphans, 0))
+ if (cmpxchg(&root->orphan_cleanup_state, 0, ORPHAN_CLEANUP_STARTED))
return;
path = btrfs_alloc_path();
found_key.objectid = found_key.offset;
found_key.type = BTRFS_INODE_ITEM_KEY;
found_key.offset = 0;
- inode = btrfs_iget(root->fs_info->sb, &found_key, root);
- if (IS_ERR(inode))
- break;
+ inode = btrfs_iget(root->fs_info->sb, &found_key, root, NULL);
+ BUG_ON(IS_ERR(inode));
/*
* add this inode to the orphan list so btrfs_orphan_del does
* the proper thing when we hit it
*/
- spin_lock(&root->list_lock);
+ spin_lock(&root->orphan_lock);
list_add(&BTRFS_I(inode)->i_orphan, &root->orphan_list);
- spin_unlock(&root->list_lock);
+ spin_unlock(&root->orphan_lock);
/*
* if this is a bad inode, means we actually succeeded in
* do a destroy_inode
*/
if (is_bad_inode(inode)) {
- trans = btrfs_start_transaction(root, 1);
+ trans = btrfs_start_transaction(root, 0);
btrfs_orphan_del(trans, inode);
btrfs_end_transaction(trans, root);
iput(inode);
/* this will do delete_inode and everything for us */
iput(inode);
}
+ btrfs_free_path(path);
+
+ root->orphan_cleanup_state = ORPHAN_CLEANUP_DONE;
+
+ if (root->orphan_block_rsv)
+ btrfs_block_rsv_release(root, root->orphan_block_rsv,
+ (u64)-1);
+
+ if (root->orphan_block_rsv || root->orphan_item_inserted) {
+ trans = btrfs_join_transaction(root, 1);
+ btrfs_end_transaction(trans, root);
+ }
if (nr_unlink)
printk(KERN_INFO "btrfs: unlinked %d orphans\n", nr_unlink);
if (nr_truncate)
printk(KERN_INFO "btrfs: truncated %d orphans\n", nr_truncate);
-
- btrfs_free_path(path);
}
/*
return ret;
}
-static int btrfs_unlink(struct inode *dir, struct dentry *dentry)
+/* helper to check if there is any shared block in the path */
+static int check_path_shared(struct btrfs_root *root,
+ struct btrfs_path *path)
+{
+ struct extent_buffer *eb;
+ int level;
+ int ret;
+ u64 refs;
+
+ for (level = 0; level < BTRFS_MAX_LEVEL; level++) {
+ if (!path->nodes[level])
+ break;
+ eb = path->nodes[level];
+ if (!btrfs_block_can_be_shared(root, eb))
+ continue;
+ ret = btrfs_lookup_extent_info(NULL, root, eb->start, eb->len,
+ &refs, NULL);
+ if (refs > 1)
+ return 1;
+ }
+ return 0;
+}
+
+/*
+ * helper to start transaction for unlink and rmdir.
+ *
+ * unlink and rmdir are special in btrfs, they do not always free space.
+ * so in enospc case, we should make sure they will free space before
+ * allowing them to use the global metadata reservation.
+ */
+static struct btrfs_trans_handle *__unlink_start_trans(struct inode *dir,
+ struct dentry *dentry)
{
- struct btrfs_root *root;
struct btrfs_trans_handle *trans;
+ struct btrfs_root *root = BTRFS_I(dir)->root;
+ struct btrfs_path *path;
+ struct btrfs_inode_ref *ref;
+ struct btrfs_dir_item *di;
struct inode *inode = dentry->d_inode;
+ u64 index;
+ int check_link = 1;
+ int err = -ENOSPC;
int ret;
- unsigned long nr = 0;
- root = BTRFS_I(dir)->root;
+ trans = btrfs_start_transaction(root, 10);
+ if (!IS_ERR(trans) || PTR_ERR(trans) != -ENOSPC)
+ return trans;
- /*
- * 5 items for unlink inode
- * 1 for orphan
- */
- ret = btrfs_reserve_metadata_space(root, 6);
- if (ret)
- return ret;
+ if (inode->i_ino == BTRFS_EMPTY_SUBVOL_DIR_OBJECTID)
+ return ERR_PTR(-ENOSPC);
+
+ /* check if there is someone else holds reference */
+ if (S_ISDIR(inode->i_mode) && atomic_read(&inode->i_count) > 1)
+ return ERR_PTR(-ENOSPC);
+
+ if (atomic_read(&inode->i_count) > 2)
+ return ERR_PTR(-ENOSPC);
+
+ if (xchg(&root->fs_info->enospc_unlink, 1))
+ return ERR_PTR(-ENOSPC);
+
+ path = btrfs_alloc_path();
+ if (!path) {
+ root->fs_info->enospc_unlink = 0;
+ return ERR_PTR(-ENOMEM);
+ }
- trans = btrfs_start_transaction(root, 1);
+ trans = btrfs_start_transaction(root, 0);
if (IS_ERR(trans)) {
- btrfs_unreserve_metadata_space(root, 6);
- return PTR_ERR(trans);
+ btrfs_free_path(path);
+ root->fs_info->enospc_unlink = 0;
+ return trans;
}
+ path->skip_locking = 1;
+ path->search_commit_root = 1;
+
+ ret = btrfs_lookup_inode(trans, root, path,
+ &BTRFS_I(dir)->location, 0);
+ if (ret < 0) {
+ err = ret;
+ goto out;
+ }
+ if (ret == 0) {
+ if (check_path_shared(root, path))
+ goto out;
+ } else {
+ check_link = 0;
+ }
+ btrfs_release_path(root, path);
+
+ ret = btrfs_lookup_inode(trans, root, path,
+ &BTRFS_I(inode)->location, 0);
+ if (ret < 0) {
+ err = ret;
+ goto out;
+ }
+ if (ret == 0) {
+ if (check_path_shared(root, path))
+ goto out;
+ } else {
+ check_link = 0;
+ }
+ btrfs_release_path(root, path);
+
+ if (ret == 0 && S_ISREG(inode->i_mode)) {
+ ret = btrfs_lookup_file_extent(trans, root, path,
+ inode->i_ino, (u64)-1, 0);
+ if (ret < 0) {
+ err = ret;
+ goto out;
+ }
+ BUG_ON(ret == 0);
+ if (check_path_shared(root, path))
+ goto out;
+ btrfs_release_path(root, path);
+ }
+
+ if (!check_link) {
+ err = 0;
+ goto out;
+ }
+
+ di = btrfs_lookup_dir_item(trans, root, path, dir->i_ino,
+ dentry->d_name.name, dentry->d_name.len, 0);
+ if (IS_ERR(di)) {
+ err = PTR_ERR(di);
+ goto out;
+ }
+ if (di) {
+ if (check_path_shared(root, path))
+ goto out;
+ } else {
+ err = 0;
+ goto out;
+ }
+ btrfs_release_path(root, path);
+
+ ref = btrfs_lookup_inode_ref(trans, root, path,
+ dentry->d_name.name, dentry->d_name.len,
+ inode->i_ino, dir->i_ino, 0);
+ if (IS_ERR(ref)) {
+ err = PTR_ERR(ref);
+ goto out;
+ }
+ BUG_ON(!ref);
+ if (check_path_shared(root, path))
+ goto out;
+ index = btrfs_inode_ref_index(path->nodes[0], ref);
+ btrfs_release_path(root, path);
+
+ di = btrfs_lookup_dir_index_item(trans, root, path, dir->i_ino, index,
+ dentry->d_name.name, dentry->d_name.len, 0);
+ if (IS_ERR(di)) {
+ err = PTR_ERR(di);
+ goto out;
+ }
+ BUG_ON(ret == -ENOENT);
+ if (check_path_shared(root, path))
+ goto out;
+
+ err = 0;
+out:
+ btrfs_free_path(path);
+ if (err) {
+ btrfs_end_transaction(trans, root);
+ root->fs_info->enospc_unlink = 0;
+ return ERR_PTR(err);
+ }
+
+ trans->block_rsv = &root->fs_info->global_block_rsv;
+ return trans;
+}
+
+static void __unlink_end_trans(struct btrfs_trans_handle *trans,
+ struct btrfs_root *root)
+{
+ if (trans->block_rsv == &root->fs_info->global_block_rsv) {
+ BUG_ON(!root->fs_info->enospc_unlink);
+ root->fs_info->enospc_unlink = 0;
+ }
+ btrfs_end_transaction_throttle(trans, root);
+}
+
+static int btrfs_unlink(struct inode *dir, struct dentry *dentry)
+{
+ struct btrfs_root *root = BTRFS_I(dir)->root;
+ struct btrfs_trans_handle *trans;
+ struct inode *inode = dentry->d_inode;
+ int ret;
+ unsigned long nr = 0;
+
+ trans = __unlink_start_trans(dir, dentry);
+ if (IS_ERR(trans))
+ return PTR_ERR(trans);
+
btrfs_set_trans_block_group(trans, dir);
btrfs_record_unlink_dir(trans, dir, dentry->d_inode, 0);
ret = btrfs_unlink_inode(trans, root, dir, dentry->d_inode,
dentry->d_name.name, dentry->d_name.len);
+ BUG_ON(ret);
- if (inode->i_nlink == 0)
+ if (inode->i_nlink == 0) {
ret = btrfs_orphan_add(trans, inode);
+ BUG_ON(ret);
+ }
nr = trans->blocks_used;
-
- btrfs_end_transaction_throttle(trans, root);
- btrfs_unreserve_metadata_space(root, 6);
+ __unlink_end_trans(trans, root);
btrfs_btree_balance_dirty(root, nr);
return ret;
}
{
struct inode *inode = dentry->d_inode;
int err = 0;
- int ret;
struct btrfs_root *root = BTRFS_I(dir)->root;
struct btrfs_trans_handle *trans;
unsigned long nr = 0;
inode->i_ino == BTRFS_FIRST_FREE_OBJECTID)
return -ENOTEMPTY;
- ret = btrfs_reserve_metadata_space(root, 5);
- if (ret)
- return ret;
-
- trans = btrfs_start_transaction(root, 1);
- if (IS_ERR(trans)) {
- btrfs_unreserve_metadata_space(root, 5);
+ trans = __unlink_start_trans(dir, dentry);
+ if (IS_ERR(trans))
return PTR_ERR(trans);
- }
btrfs_set_trans_block_group(trans, dir);
btrfs_i_size_write(inode, 0);
out:
nr = trans->blocks_used;
- ret = btrfs_end_transaction_throttle(trans, root);
- btrfs_unreserve_metadata_space(root, 5);
+ __unlink_end_trans(trans, root);
btrfs_btree_balance_dirty(root, nr);
- if (ret && !err)
- err = ret;
return err;
}
if (pending_del_nr) {
ret = btrfs_del_items(trans, root, path, pending_del_slot,
pending_del_nr);
+ BUG_ON(ret);
}
btrfs_free_path(path);
return err;
struct btrfs_root *root = BTRFS_I(inode)->root;
struct extent_io_tree *io_tree = &BTRFS_I(inode)->io_tree;
struct btrfs_ordered_extent *ordered;
+ struct extent_state *cached_state = NULL;
char *kaddr;
u32 blocksize = root->sectorsize;
pgoff_t index = from >> PAGE_CACHE_SHIFT;
if ((offset & (blocksize - 1)) == 0)
goto out;
- ret = btrfs_check_data_free_space(root, inode, PAGE_CACHE_SIZE);
- if (ret)
- goto out;
-
- ret = btrfs_reserve_metadata_for_delalloc(root, inode, 1);
+ ret = btrfs_delalloc_reserve_space(inode, PAGE_CACHE_SIZE);
if (ret)
goto out;
again:
page = grab_cache_page(mapping, index);
if (!page) {
- btrfs_free_reserved_data_space(root, inode, PAGE_CACHE_SIZE);
- btrfs_unreserve_metadata_for_delalloc(root, inode, 1);
+ btrfs_delalloc_release_space(inode, PAGE_CACHE_SIZE);
goto out;
}
}
wait_on_page_writeback(page);
- lock_extent(io_tree, page_start, page_end, GFP_NOFS);
+ lock_extent_bits(io_tree, page_start, page_end, 0, &cached_state,
+ GFP_NOFS);
set_page_extent_mapped(page);
ordered = btrfs_lookup_ordered_extent(inode, page_start);
if (ordered) {
- unlock_extent(io_tree, page_start, page_end, GFP_NOFS);
+ unlock_extent_cached(io_tree, page_start, page_end,
+ &cached_state, GFP_NOFS);
unlock_page(page);
page_cache_release(page);
btrfs_start_ordered_extent(inode, ordered, 1);
goto again;
}
- clear_extent_bits(&BTRFS_I(inode)->io_tree, page_start, page_end,
+ clear_extent_bit(&BTRFS_I(inode)->io_tree, page_start, page_end,
EXTENT_DIRTY | EXTENT_DELALLOC | EXTENT_DO_ACCOUNTING,
- GFP_NOFS);
+ 0, 0, &cached_state, GFP_NOFS);
- ret = btrfs_set_extent_delalloc(inode, page_start, page_end);
+ ret = btrfs_set_extent_delalloc(inode, page_start, page_end,
+ &cached_state);
if (ret) {
- unlock_extent(io_tree, page_start, page_end, GFP_NOFS);
+ unlock_extent_cached(io_tree, page_start, page_end,
+ &cached_state, GFP_NOFS);
goto out_unlock;
}
}
ClearPageChecked(page);
set_page_dirty(page);
- unlock_extent(io_tree, page_start, page_end, GFP_NOFS);
+ unlock_extent_cached(io_tree, page_start, page_end, &cached_state,
+ GFP_NOFS);
out_unlock:
if (ret)
- btrfs_free_reserved_data_space(root, inode, PAGE_CACHE_SIZE);
- btrfs_unreserve_metadata_for_delalloc(root, inode, 1);
+ btrfs_delalloc_release_space(inode, PAGE_CACHE_SIZE);
unlock_page(page);
page_cache_release(page);
out:
struct btrfs_trans_handle *trans;
struct btrfs_root *root = BTRFS_I(inode)->root;
struct extent_io_tree *io_tree = &BTRFS_I(inode)->io_tree;
- struct extent_map *em;
+ struct extent_map *em = NULL;
+ struct extent_state *cached_state = NULL;
u64 mask = root->sectorsize - 1;
u64 hole_start = (inode->i_size + mask) & ~mask;
u64 block_end = (size + mask) & ~mask;
struct btrfs_ordered_extent *ordered;
btrfs_wait_ordered_range(inode, hole_start,
block_end - hole_start);
- lock_extent(io_tree, hole_start, block_end - 1, GFP_NOFS);
+ lock_extent_bits(io_tree, hole_start, block_end - 1, 0,
+ &cached_state, GFP_NOFS);
ordered = btrfs_lookup_ordered_extent(inode, hole_start);
if (!ordered)
break;
- unlock_extent(io_tree, hole_start, block_end - 1, GFP_NOFS);
+ unlock_extent_cached(io_tree, hole_start, block_end - 1,
+ &cached_state, GFP_NOFS);
btrfs_put_ordered_extent(ordered);
}
u64 hint_byte = 0;
hole_size = last_byte - cur_offset;
- err = btrfs_reserve_metadata_space(root, 2);
- if (err)
+ trans = btrfs_start_transaction(root, 2);
+ if (IS_ERR(trans)) {
+ err = PTR_ERR(trans);
break;
-
- trans = btrfs_start_transaction(root, 1);
+ }
btrfs_set_trans_block_group(trans, inode);
err = btrfs_drop_extents(trans, inode, cur_offset,
last_byte - 1, 0);
btrfs_end_transaction(trans, root);
- btrfs_unreserve_metadata_space(root, 2);
}
free_extent_map(em);
+ em = NULL;
cur_offset = last_byte;
if (cur_offset >= block_end)
break;
}
- unlock_extent(io_tree, hole_start, block_end - 1, GFP_NOFS);
+ free_extent_map(em);
+ unlock_extent_cached(io_tree, hole_start, block_end - 1, &cached_state,
+ GFP_NOFS);
return err;
}
}
}
- ret = btrfs_reserve_metadata_space(root, 1);
- if (ret)
- return ret;
+ trans = btrfs_start_transaction(root, 5);
+ if (IS_ERR(trans))
+ return PTR_ERR(trans);
- trans = btrfs_start_transaction(root, 1);
btrfs_set_trans_block_group(trans, inode);
ret = btrfs_orphan_add(trans, inode);
nr = trans->blocks_used;
btrfs_end_transaction(trans, root);
- btrfs_unreserve_metadata_space(root, 1);
btrfs_btree_balance_dirty(root, nr);
if (attr->ia_size > inode->i_size) {
i_size_write(inode, attr->ia_size);
btrfs_ordered_update_i_size(inode, inode->i_size, NULL);
- trans = btrfs_start_transaction(root, 1);
+ trans = btrfs_start_transaction(root, 0);
+ BUG_ON(IS_ERR(trans));
btrfs_set_trans_block_group(trans, inode);
+ trans->block_rsv = root->orphan_block_rsv;
+ BUG_ON(!trans->block_rsv);
ret = btrfs_update_inode(trans, root, inode);
BUG_ON(ret);
btrfs_i_size_write(inode, 0);
while (1) {
- trans = btrfs_start_transaction(root, 1);
+ trans = btrfs_start_transaction(root, 0);
+ BUG_ON(IS_ERR(trans));
btrfs_set_trans_block_group(trans, inode);
- ret = btrfs_truncate_inode_items(trans, root, inode, 0, 0);
+ trans->block_rsv = root->orphan_block_rsv;
+
+ ret = btrfs_block_rsv_check(trans, root,
+ root->orphan_block_rsv, 0, 5);
+ if (ret) {
+ BUG_ON(ret != -EAGAIN);
+ ret = btrfs_commit_transaction(trans, root);
+ BUG_ON(ret);
+ continue;
+ }
+ ret = btrfs_truncate_inode_items(trans, root, inode, 0, 0);
if (ret != -EAGAIN)
break;
btrfs_end_transaction(trans, root);
trans = NULL;
btrfs_btree_balance_dirty(root, nr);
+
}
if (ret == 0) {
return 0;
}
-static noinline void init_btrfs_i(struct inode *inode)
-{
- struct btrfs_inode *bi = BTRFS_I(inode);
-
- bi->generation = 0;
- bi->sequence = 0;
- bi->last_trans = 0;
- bi->last_sub_trans = 0;
- bi->logged_trans = 0;
- bi->delalloc_bytes = 0;
- bi->reserved_bytes = 0;
- bi->disk_i_size = 0;
- bi->flags = 0;
- bi->index_cnt = (u64)-1;
- bi->last_unlink_trans = 0;
- bi->ordered_data_close = 0;
- extent_map_tree_init(&BTRFS_I(inode)->extent_tree, GFP_NOFS);
- extent_io_tree_init(&BTRFS_I(inode)->io_tree,
- inode->i_mapping, GFP_NOFS);
- extent_io_tree_init(&BTRFS_I(inode)->io_failure_tree,
- inode->i_mapping, GFP_NOFS);
- INIT_LIST_HEAD(&BTRFS_I(inode)->delalloc_inodes);
- INIT_LIST_HEAD(&BTRFS_I(inode)->ordered_operations);
- RB_CLEAR_NODE(&BTRFS_I(inode)->rb_node);
- btrfs_ordered_inode_tree_init(&BTRFS_I(inode)->ordered_tree);
- mutex_init(&BTRFS_I(inode)->log_mutex);
-}
-
static int btrfs_init_locked_inode(struct inode *inode, void *p)
{
struct btrfs_iget_args *args = p;
inode->i_ino = args->ino;
- init_btrfs_i(inode);
BTRFS_I(inode)->root = args->root;
btrfs_set_inode_space_info(args->root, inode);
return 0;
* Returns in *is_new if the inode was read from disk
*/
struct inode *btrfs_iget(struct super_block *s, struct btrfs_key *location,
- struct btrfs_root *root)
+ struct btrfs_root *root, int *new)
{
struct inode *inode;
inode_tree_add(inode);
unlock_new_inode(inode);
+ if (new)
+ *new = 1;
}
return inode;
if (!inode)
return ERR_PTR(-ENOMEM);
- init_btrfs_i(inode);
-
BTRFS_I(inode)->root = root;
memcpy(&BTRFS_I(inode)->location, key, sizeof(*key));
BTRFS_I(inode)->dummy_inode = 1;
return NULL;
if (location.type == BTRFS_INODE_ITEM_KEY) {
- inode = btrfs_iget(dir->i_sb, &location, root);
- if (unlikely(root->clean_orphans) &&
- !(inode->i_sb->s_flags & MS_RDONLY)) {
- down_read(&root->fs_info->cleanup_work_sem);
- btrfs_orphan_cleanup(root);
- up_read(&root->fs_info->cleanup_work_sem);
- }
+ inode = btrfs_iget(dir->i_sb, &location, root, NULL);
return inode;
}
else
inode = new_simple_dir(dir->i_sb, &location, sub_root);
} else {
- inode = btrfs_iget(dir->i_sb, &location, sub_root);
+ inode = btrfs_iget(dir->i_sb, &location, sub_root, NULL);
}
srcu_read_unlock(&root->fs_info->subvol_srcu, index);
return ret;
}
-int btrfs_write_inode(struct inode *inode, int wait)
+int btrfs_write_inode(struct inode *inode, struct writeback_control *wbc)
{
struct btrfs_root *root = BTRFS_I(inode)->root;
struct btrfs_trans_handle *trans;
int ret = 0;
- if (root->fs_info->btree_inode == inode)
+ if (BTRFS_I(inode)->dummy_inode)
return 0;
- if (wait) {
+ if (wbc->sync_mode == WB_SYNC_ALL) {
trans = btrfs_join_transaction(root, 1);
btrfs_set_trans_block_group(trans, inode);
ret = btrfs_commit_transaction(trans, root);
{
struct btrfs_root *root = BTRFS_I(inode)->root;
struct btrfs_trans_handle *trans;
+ int ret;
+
+ if (BTRFS_I(inode)->dummy_inode)
+ return;
trans = btrfs_join_transaction(root, 1);
btrfs_set_trans_block_group(trans, inode);
- btrfs_update_inode(trans, root, inode);
+
+ ret = btrfs_update_inode(trans, root, inode);
+ if (ret)
+ printk(KERN_ERR"btrfs: fail to dirty inode %lu error %d\n",
+ inode->i_ino, ret);
+
btrfs_end_transaction(trans, root);
}
* btrfs_get_inode_index_count has an explanation for the magic
* number
*/
- init_btrfs_i(inode);
BTRFS_I(inode)->index_cnt = 2;
BTRFS_I(inode)->root = root;
BTRFS_I(inode)->generation = trans->transid;
if (!new_valid_dev(rdev))
return -EINVAL;
+ err = btrfs_find_free_objectid(NULL, root, dir->i_ino, &objectid);
+ if (err)
+ return err;
+
/*
* 2 for inode item and ref
* 2 for dir items
* 1 for xattr if selinux is on
*/
- err = btrfs_reserve_metadata_space(root, 5);
- if (err)
- return err;
+ trans = btrfs_start_transaction(root, 5);
+ if (IS_ERR(trans))
+ return PTR_ERR(trans);
- trans = btrfs_start_transaction(root, 1);
- if (!trans)
- goto fail;
btrfs_set_trans_block_group(trans, dir);
- err = btrfs_find_free_objectid(trans, root, dir->i_ino, &objectid);
- if (err) {
- err = -ENOSPC;
- goto out_unlock;
- }
-
inode = btrfs_new_inode(trans, root, dir, dentry->d_name.name,
dentry->d_name.len,
dentry->d_parent->d_inode->i_ino, objectid,
out_unlock:
nr = trans->blocks_used;
btrfs_end_transaction_throttle(trans, root);
-fail:
- btrfs_unreserve_metadata_space(root, 5);
+ btrfs_btree_balance_dirty(root, nr);
if (drop_inode) {
inode_dec_link_count(inode);
iput(inode);
}
- btrfs_btree_balance_dirty(root, nr);
return err;
}
struct btrfs_trans_handle *trans;
struct btrfs_root *root = BTRFS_I(dir)->root;
struct inode *inode = NULL;
- int err;
int drop_inode = 0;
+ int err;
unsigned long nr = 0;
u64 objectid;
u64 index = 0;
+ err = btrfs_find_free_objectid(NULL, root, dir->i_ino, &objectid);
+ if (err)
+ return err;
/*
* 2 for inode item and ref
* 2 for dir items
* 1 for xattr if selinux is on
*/
- err = btrfs_reserve_metadata_space(root, 5);
- if (err)
- return err;
+ trans = btrfs_start_transaction(root, 5);
+ if (IS_ERR(trans))
+ return PTR_ERR(trans);
- trans = btrfs_start_transaction(root, 1);
- if (!trans)
- goto fail;
btrfs_set_trans_block_group(trans, dir);
- err = btrfs_find_free_objectid(trans, root, dir->i_ino, &objectid);
- if (err) {
- err = -ENOSPC;
- goto out_unlock;
- }
-
inode = btrfs_new_inode(trans, root, dir, dentry->d_name.name,
dentry->d_name.len,
dentry->d_parent->d_inode->i_ino,
out_unlock:
nr = trans->blocks_used;
btrfs_end_transaction_throttle(trans, root);
-fail:
- btrfs_unreserve_metadata_space(root, 5);
if (drop_inode) {
inode_dec_link_count(inode);
iput(inode);
if (root->objectid != BTRFS_I(inode)->root->objectid)
return -EPERM;
- /*
- * 1 item for inode ref
- * 2 items for dir items
- */
- err = btrfs_reserve_metadata_space(root, 3);
- if (err)
- return err;
-
btrfs_inc_nlink(inode);
err = btrfs_set_inode_index(dir, &index);
if (err)
goto fail;
- trans = btrfs_start_transaction(root, 1);
+ /*
+ * 1 item for inode ref
+ * 2 items for dir items
+ */
+ trans = btrfs_start_transaction(root, 3);
+ if (IS_ERR(trans)) {
+ err = PTR_ERR(trans);
+ goto fail;
+ }
btrfs_set_trans_block_group(trans, dir);
atomic_inc(&inode->i_count);
nr = trans->blocks_used;
btrfs_end_transaction_throttle(trans, root);
fail:
- btrfs_unreserve_metadata_space(root, 3);
if (drop_inode) {
inode_dec_link_count(inode);
iput(inode);
u64 index = 0;
unsigned long nr = 1;
+ err = btrfs_find_free_objectid(NULL, root, dir->i_ino, &objectid);
+ if (err)
+ return err;
+
/*
* 2 items for inode and ref
* 2 items for dir items
* 1 for xattr if selinux is on
*/
- err = btrfs_reserve_metadata_space(root, 5);
- if (err)
- return err;
-
- trans = btrfs_start_transaction(root, 1);
- if (!trans) {
- err = -ENOMEM;
- goto out_unlock;
- }
+ trans = btrfs_start_transaction(root, 5);
+ if (IS_ERR(trans))
+ return PTR_ERR(trans);
btrfs_set_trans_block_group(trans, dir);
- err = btrfs_find_free_objectid(trans, root, dir->i_ino, &objectid);
- if (err) {
- err = -ENOSPC;
- goto out_unlock;
- }
-
inode = btrfs_new_inode(trans, root, dir, dentry->d_name.name,
dentry->d_name.len,
dentry->d_parent->d_inode->i_ino, objectid,
out_fail:
nr = trans->blocks_used;
btrfs_end_transaction_throttle(trans, root);
-
-out_unlock:
- btrfs_unreserve_metadata_space(root, 5);
if (drop_on_err)
iput(inode);
btrfs_btree_balance_dirty(root, nr);
}
flush_dcache_page(page);
} else if (create && PageUptodate(page)) {
+ WARN_ON(1);
if (!trans) {
kunmap(page);
free_extent_map(em);
{
struct extent_io_tree *tree;
struct btrfs_ordered_extent *ordered;
+ struct extent_state *cached_state = NULL;
u64 page_start = page_offset(page);
u64 page_end = page_start + PAGE_CACHE_SIZE - 1;
btrfs_releasepage(page, GFP_NOFS);
return;
}
- lock_extent(tree, page_start, page_end, GFP_NOFS);
+ lock_extent_bits(tree, page_start, page_end, 0, &cached_state,
+ GFP_NOFS);
ordered = btrfs_lookup_ordered_extent(page->mapping->host,
page_offset(page));
if (ordered) {
clear_extent_bit(tree, page_start, page_end,
EXTENT_DIRTY | EXTENT_DELALLOC |
EXTENT_LOCKED | EXTENT_DO_ACCOUNTING, 1, 0,
- NULL, GFP_NOFS);
+ &cached_state, GFP_NOFS);
/*
* whoever cleared the private bit is responsible
* for the finish_ordered_io
page_start, page_end);
}
btrfs_put_ordered_extent(ordered);
- lock_extent(tree, page_start, page_end, GFP_NOFS);
+ cached_state = NULL;
+ lock_extent_bits(tree, page_start, page_end, 0, &cached_state,
+ GFP_NOFS);
}
clear_extent_bit(tree, page_start, page_end,
EXTENT_LOCKED | EXTENT_DIRTY | EXTENT_DELALLOC |
- EXTENT_DO_ACCOUNTING, 1, 1, NULL, GFP_NOFS);
+ EXTENT_DO_ACCOUNTING, 1, 1, &cached_state, GFP_NOFS);
__btrfs_releasepage(page, GFP_NOFS);
ClearPageChecked(page);
struct btrfs_root *root = BTRFS_I(inode)->root;
struct extent_io_tree *io_tree = &BTRFS_I(inode)->io_tree;
struct btrfs_ordered_extent *ordered;
+ struct extent_state *cached_state = NULL;
char *kaddr;
unsigned long zero_start;
loff_t size;
u64 page_start;
u64 page_end;
- ret = btrfs_check_data_free_space(root, inode, PAGE_CACHE_SIZE);
+ ret = btrfs_delalloc_reserve_space(inode, PAGE_CACHE_SIZE);
if (ret) {
if (ret == -ENOMEM)
ret = VM_FAULT_OOM;
goto out;
}
- ret = btrfs_reserve_metadata_for_delalloc(root, inode, 1);
- if (ret) {
- btrfs_free_reserved_data_space(root, inode, PAGE_CACHE_SIZE);
- ret = VM_FAULT_SIGBUS;
- goto out;
- }
-
ret = VM_FAULT_NOPAGE; /* make the VM retry the fault */
again:
lock_page(page);
if ((page->mapping != inode->i_mapping) ||
(page_start >= size)) {
- btrfs_free_reserved_data_space(root, inode, PAGE_CACHE_SIZE);
/* page got truncated out from underneath us */
goto out_unlock;
}
wait_on_page_writeback(page);
- lock_extent(io_tree, page_start, page_end, GFP_NOFS);
+ lock_extent_bits(io_tree, page_start, page_end, 0, &cached_state,
+ GFP_NOFS);
set_page_extent_mapped(page);
/*
*/
ordered = btrfs_lookup_ordered_extent(inode, page_start);
if (ordered) {
- unlock_extent(io_tree, page_start, page_end, GFP_NOFS);
+ unlock_extent_cached(io_tree, page_start, page_end,
+ &cached_state, GFP_NOFS);
unlock_page(page);
btrfs_start_ordered_extent(inode, ordered, 1);
btrfs_put_ordered_extent(ordered);
* is probably a better way to do this, but for now keep consistent with
* prepare_pages in the normal write path.
*/
- clear_extent_bits(&BTRFS_I(inode)->io_tree, page_start, page_end,
+ clear_extent_bit(&BTRFS_I(inode)->io_tree, page_start, page_end,
EXTENT_DIRTY | EXTENT_DELALLOC | EXTENT_DO_ACCOUNTING,
- GFP_NOFS);
+ 0, 0, &cached_state, GFP_NOFS);
- ret = btrfs_set_extent_delalloc(inode, page_start, page_end);
+ ret = btrfs_set_extent_delalloc(inode, page_start, page_end,
+ &cached_state);
if (ret) {
- unlock_extent(io_tree, page_start, page_end, GFP_NOFS);
+ unlock_extent_cached(io_tree, page_start, page_end,
+ &cached_state, GFP_NOFS);
ret = VM_FAULT_SIGBUS;
- btrfs_free_reserved_data_space(root, inode, PAGE_CACHE_SIZE);
goto out_unlock;
}
ret = 0;
BTRFS_I(inode)->last_trans = root->fs_info->generation;
BTRFS_I(inode)->last_sub_trans = BTRFS_I(inode)->root->log_transid;
- unlock_extent(io_tree, page_start, page_end, GFP_NOFS);
+ unlock_extent_cached(io_tree, page_start, page_end, &cached_state, GFP_NOFS);
out_unlock:
- btrfs_unreserve_metadata_for_delalloc(root, inode, 1);
if (!ret)
return VM_FAULT_LOCKED;
unlock_page(page);
+ btrfs_delalloc_release_space(inode, PAGE_CACHE_SIZE);
out:
return ret;
}
btrfs_wait_ordered_range(inode, inode->i_size & (~mask), (u64)-1);
btrfs_ordered_update_i_size(inode, inode->i_size, NULL);
- trans = btrfs_start_transaction(root, 1);
+ trans = btrfs_start_transaction(root, 0);
+ BUG_ON(IS_ERR(trans));
btrfs_set_trans_block_group(trans, inode);
+ trans->block_rsv = root->orphan_block_rsv;
/*
* setattr is responsible for setting the ordered_data_close flag,
btrfs_add_ordered_operation(trans, root, inode);
while (1) {
+ if (!trans) {
+ trans = btrfs_start_transaction(root, 0);
+ BUG_ON(IS_ERR(trans));
+ btrfs_set_trans_block_group(trans, inode);
+ trans->block_rsv = root->orphan_block_rsv;
+ }
+
+ ret = btrfs_block_rsv_check(trans, root,
+ root->orphan_block_rsv, 0, 5);
+ if (ret) {
+ BUG_ON(ret != -EAGAIN);
+ ret = btrfs_commit_transaction(trans, root);
+ BUG_ON(ret);
+ trans = NULL;
+ continue;
+ }
+
ret = btrfs_truncate_inode_items(trans, root, inode,
inode->i_size,
BTRFS_EXTENT_DATA_KEY);
nr = trans->blocks_used;
btrfs_end_transaction(trans, root);
+ trans = NULL;
btrfs_btree_balance_dirty(root, nr);
-
- trans = btrfs_start_transaction(root, 1);
- btrfs_set_trans_block_group(trans, inode);
}
if (ret == 0 && inode->i_nlink > 0) {
struct inode *btrfs_alloc_inode(struct super_block *sb)
{
struct btrfs_inode *ei;
+ struct inode *inode;
ei = kmem_cache_alloc(btrfs_inode_cachep, GFP_NOFS);
if (!ei)
return NULL;
+
+ ei->root = NULL;
+ ei->space_info = NULL;
+ ei->generation = 0;
+ ei->sequence = 0;
ei->last_trans = 0;
ei->last_sub_trans = 0;
ei->logged_trans = 0;
- ei->outstanding_extents = 0;
- ei->reserved_extents = 0;
- ei->root = NULL;
+ ei->delalloc_bytes = 0;
+ ei->reserved_bytes = 0;
+ ei->disk_i_size = 0;
+ ei->flags = 0;
+ ei->index_cnt = (u64)-1;
+ ei->last_unlink_trans = 0;
+
spin_lock_init(&ei->accounting_lock);
+ atomic_set(&ei->outstanding_extents, 0);
+ ei->reserved_extents = 0;
+
+ ei->ordered_data_close = 0;
+ ei->orphan_meta_reserved = 0;
+ ei->dummy_inode = 0;
+ ei->force_compress = 0;
+
+ inode = &ei->vfs_inode;
+ extent_map_tree_init(&ei->extent_tree, GFP_NOFS);
+ extent_io_tree_init(&ei->io_tree, &inode->i_data, GFP_NOFS);
+ extent_io_tree_init(&ei->io_failure_tree, &inode->i_data, GFP_NOFS);
+ mutex_init(&ei->log_mutex);
btrfs_ordered_inode_tree_init(&ei->ordered_tree);
INIT_LIST_HEAD(&ei->i_orphan);
+ INIT_LIST_HEAD(&ei->delalloc_inodes);
INIT_LIST_HEAD(&ei->ordered_operations);
- return &ei->vfs_inode;
+ RB_CLEAR_NODE(&ei->rb_node);
+
+ return inode;
}
void btrfs_destroy_inode(struct inode *inode)
WARN_ON(!list_empty(&inode->i_dentry));
WARN_ON(inode->i_data.nrpages);
+ WARN_ON(atomic_read(&BTRFS_I(inode)->outstanding_extents));
+ WARN_ON(BTRFS_I(inode)->reserved_extents);
/*
* This can happen where we create an inode, but somebody else also
spin_unlock(&root->fs_info->ordered_extent_lock);
}
- spin_lock(&root->list_lock);
+ spin_lock(&root->orphan_lock);
if (!list_empty(&BTRFS_I(inode)->i_orphan)) {
printk(KERN_INFO "BTRFS: inode %lu still on the orphan list\n",
inode->i_ino);
list_del_init(&BTRFS_I(inode)->i_orphan);
}
- spin_unlock(&root->list_lock);
+ spin_unlock(&root->orphan_lock);
while (1) {
ordered = btrfs_lookup_first_ordered_extent(inode, (u64)-1);
void btrfs_drop_inode(struct inode *inode)
{
struct btrfs_root *root = BTRFS_I(inode)->root;
-
if (inode->i_nlink > 0 && btrfs_root_refs(&root->root_item) == 0)
generic_delete_inode(inode);
else
if (S_ISDIR(old_inode->i_mode) && new_inode &&
new_inode->i_size > BTRFS_EMPTY_DIR_SIZE)
return -ENOTEMPTY;
-
- /*
- * We want to reserve the absolute worst case amount of items. So if
- * both inodes are subvols and we need to unlink them then that would
- * require 4 item modifications, but if they are both normal inodes it
- * would require 5 item modifications, so we'll assume their normal
- * inodes. So 5 * 2 is 10, plus 1 for the new link, so 11 total items
- * should cover the worst case number of items we'll modify.
- */
- ret = btrfs_reserve_metadata_space(root, 11);
- if (ret)
- return ret;
-
/*
* we're using rename to replace one file with another.
* and the replacement file is large. Start IO on it now so
/* close the racy window with snapshot create/destroy ioctl */
if (old_inode->i_ino == BTRFS_FIRST_FREE_OBJECTID)
down_read(&root->fs_info->subvol_sem);
+ /*
+ * We want to reserve the absolute worst case amount of items. So if
+ * both inodes are subvols and we need to unlink them then that would
+ * require 4 item modifications, but if they are both normal inodes it
+ * would require 5 item modifications, so we'll assume their normal
+ * inodes. So 5 * 2 is 10, plus 1 for the new link, so 11 total items
+ * should cover the worst case number of items we'll modify.
+ */
+ trans = btrfs_start_transaction(root, 20);
+ if (IS_ERR(trans))
+ return PTR_ERR(trans);
- trans = btrfs_start_transaction(root, 1);
btrfs_set_trans_block_group(trans, new_dir);
if (dest != root)
if (old_inode->i_ino == BTRFS_FIRST_FREE_OBJECTID)
up_read(&root->fs_info->subvol_sem);
- btrfs_unreserve_metadata_space(root, 11);
return ret;
}
return 0;
}
+int btrfs_start_one_delalloc_inode(struct btrfs_root *root, int delay_iput)
+{
+ struct btrfs_inode *binode;
+ struct inode *inode = NULL;
+
+ spin_lock(&root->fs_info->delalloc_lock);
+ while (!list_empty(&root->fs_info->delalloc_inodes)) {
+ binode = list_entry(root->fs_info->delalloc_inodes.next,
+ struct btrfs_inode, delalloc_inodes);
+ inode = igrab(&binode->vfs_inode);
+ if (inode) {
+ list_move_tail(&binode->delalloc_inodes,
+ &root->fs_info->delalloc_inodes);
+ break;
+ }
+
+ list_del_init(&binode->delalloc_inodes);
+ cond_resched_lock(&root->fs_info->delalloc_lock);
+ }
+ spin_unlock(&root->fs_info->delalloc_lock);
+
+ if (inode) {
+ write_inode_now(inode, 0);
+ if (delay_iput)
+ btrfs_add_delayed_iput(inode);
+ else
+ iput(inode);
+ return 1;
+ }
+ return 0;
+}
+
static int btrfs_symlink(struct inode *dir, struct dentry *dentry,
const char *symname)
{
if (name_len > BTRFS_MAX_INLINE_DATA_SIZE(root))
return -ENAMETOOLONG;
+ err = btrfs_find_free_objectid(NULL, root, dir->i_ino, &objectid);
+ if (err)
+ return err;
/*
* 2 items for inode item and ref
* 2 items for dir items
* 1 item for xattr if selinux is on
*/
- err = btrfs_reserve_metadata_space(root, 5);
- if (err)
- return err;
+ trans = btrfs_start_transaction(root, 5);
+ if (IS_ERR(trans))
+ return PTR_ERR(trans);
- trans = btrfs_start_transaction(root, 1);
- if (!trans)
- goto out_fail;
btrfs_set_trans_block_group(trans, dir);
- err = btrfs_find_free_objectid(trans, root, dir->i_ino, &objectid);
- if (err) {
- err = -ENOSPC;
- goto out_unlock;
- }
-
inode = btrfs_new_inode(trans, root, dir, dentry->d_name.name,
dentry->d_name.len,
dentry->d_parent->d_inode->i_ino, objectid,
out_unlock:
nr = trans->blocks_used;
btrfs_end_transaction_throttle(trans, root);
-out_fail:
- btrfs_unreserve_metadata_space(root, 5);
if (drop_inode) {
inode_dec_link_count(inode);
iput(inode);
return err;
}
-static int prealloc_file_range(struct inode *inode, u64 start, u64 end,
- u64 alloc_hint, int mode)
+int btrfs_prealloc_file_range(struct inode *inode, int mode,
+ u64 start, u64 num_bytes, u64 min_size,
+ loff_t actual_len, u64 *alloc_hint)
{
struct btrfs_trans_handle *trans;
struct btrfs_root *root = BTRFS_I(inode)->root;
struct btrfs_key ins;
- u64 alloc_size;
u64 cur_offset = start;
- u64 num_bytes = end - start;
int ret = 0;
while (num_bytes > 0) {
- alloc_size = min(num_bytes, root->fs_info->max_extent);
-
- trans = btrfs_start_transaction(root, 1);
-
- ret = btrfs_reserve_extent(trans, root, alloc_size,
- root->sectorsize, 0, alloc_hint,
- (u64)-1, &ins, 1);
- if (ret) {
- WARN_ON(1);
- goto stop_trans;
+ trans = btrfs_start_transaction(root, 3);
+ if (IS_ERR(trans)) {
+ ret = PTR_ERR(trans);
+ break;
}
- ret = btrfs_reserve_metadata_space(root, 3);
+ ret = btrfs_reserve_extent(trans, root, num_bytes, min_size,
+ 0, *alloc_hint, (u64)-1, &ins, 1);
if (ret) {
- btrfs_free_reserved_extent(root, ins.objectid,
- ins.offset);
- goto stop_trans;
+ btrfs_end_transaction(trans, root);
+ break;
}
ret = insert_reserved_file_extent(trans, inode,
num_bytes -= ins.offset;
cur_offset += ins.offset;
- alloc_hint = ins.objectid + ins.offset;
+ *alloc_hint = ins.objectid + ins.offset;
inode->i_ctime = CURRENT_TIME;
BTRFS_I(inode)->flags |= BTRFS_INODE_PREALLOC;
if (!(mode & FALLOC_FL_KEEP_SIZE) &&
- cur_offset > inode->i_size) {
+ (actual_len > inode->i_size) &&
+ (cur_offset > inode->i_size)) {
+ if (cur_offset > actual_len)
+ i_size_write(inode, actual_len);
+ else
+ i_size_write(inode, cur_offset);
i_size_write(inode, cur_offset);
btrfs_ordered_update_i_size(inode, cur_offset, NULL);
}
BUG_ON(ret);
btrfs_end_transaction(trans, root);
- btrfs_unreserve_metadata_space(root, 3);
}
return ret;
-
-stop_trans:
- btrfs_end_transaction(trans, root);
- return ret;
-
}
static long btrfs_fallocate(struct inode *inode, int mode,
loff_t offset, loff_t len)
{
+ struct extent_state *cached_state = NULL;
u64 cur_offset;
u64 last_byte;
u64 alloc_start;
goto out;
}
- ret = btrfs_check_data_free_space(BTRFS_I(inode)->root, inode,
- alloc_end - alloc_start);
+ ret = btrfs_check_data_free_space(inode, alloc_end - alloc_start);
if (ret)
goto out;
/* the extent lock is ordered inside the running
* transaction
*/
- lock_extent(&BTRFS_I(inode)->io_tree, alloc_start, locked_end,
- GFP_NOFS);
+ lock_extent_bits(&BTRFS_I(inode)->io_tree, alloc_start,
+ locked_end, 0, &cached_state, GFP_NOFS);
ordered = btrfs_lookup_first_ordered_extent(inode,
alloc_end - 1);
if (ordered &&
ordered->file_offset + ordered->len > alloc_start &&
ordered->file_offset < alloc_end) {
btrfs_put_ordered_extent(ordered);
- unlock_extent(&BTRFS_I(inode)->io_tree,
- alloc_start, locked_end, GFP_NOFS);
+ unlock_extent_cached(&BTRFS_I(inode)->io_tree,
+ alloc_start, locked_end,
+ &cached_state, GFP_NOFS);
/*
* we can't wait on the range with the transaction
* running or with the extent lock held
if (em->block_start == EXTENT_MAP_HOLE ||
(cur_offset >= inode->i_size &&
!test_bit(EXTENT_FLAG_PREALLOC, &em->flags))) {
- ret = prealloc_file_range(inode,
- cur_offset, last_byte,
- alloc_hint, mode);
+ ret = btrfs_prealloc_file_range(inode, 0, cur_offset,
+ last_byte - cur_offset,
+ 1 << inode->i_blkbits,
+ offset + len,
+ &alloc_hint);
if (ret < 0) {
free_extent_map(em);
break;
}
}
- if (em->block_start <= EXTENT_MAP_LAST_BYTE)
- alloc_hint = em->block_start;
free_extent_map(em);
cur_offset = last_byte;
break;
}
}
- unlock_extent(&BTRFS_I(inode)->io_tree, alloc_start, locked_end,
- GFP_NOFS);
+ unlock_extent_cached(&BTRFS_I(inode)->io_tree, alloc_start, locked_end,
+ &cached_state, GFP_NOFS);
- btrfs_free_reserved_data_space(BTRFS_I(inode)->root, inode,
- alloc_end - alloc_start);
+ btrfs_free_reserved_data_space(inode, alloc_end - alloc_start);
out:
mutex_unlock(&inode->i_mutex);
return ret;
git://ftp.safe.ca / safe / jmp / linux-2.6 / blobdiff