return 1;
}
- ret = btrfs_drop_extents(trans, inode, start, aligned_end,
+ ret = btrfs_drop_extents(trans, root, inode, start, aligned_end,
&hint_byte, 1);
if (ret)
return ret;
* If this code finds it can't get good compression, it puts an
* entry onto the work queue to write the uncompressed bytes. This
* makes sure that both compressed inodes and uncompressed inodes
- * are written in the same order that pdflush sent them down.
+ * are written in the same order that the flusher thread sent them
+ * down.
*/
static noinline int compress_file_range(struct inode *inode,
struct page *locked_page,
em->block_start = disk_bytenr;
em->bdev = root->fs_info->fs_devices->latest_bdev;
set_bit(EXTENT_FLAG_PINNED, &em->flags);
+ set_bit(EXTENT_FLAG_PREALLOC, &em->flags);
while (1) {
write_lock(&em_tree->lock);
ret = add_extent_mapping(em_tree, em);
* the caller is expected to unpin it and allow it to be merged
* with the others.
*/
- ret = btrfs_drop_extents(trans, inode, file_pos, file_pos + num_bytes,
+ ret = btrfs_drop_extents(trans, root, inode, file_pos,
+ file_pos + num_bytes,
&hint, 0);
if (ret)
goto out;
trans = btrfs_join_transaction_nolock(root);
else
trans = btrfs_join_transaction(root);
- if (IS_ERR(trans))
- return PTR_ERR(trans);
+ if (IS_ERR(trans)) {
+ ret = PTR_ERR(trans);
+ trans = NULL;
+ goto out;
+ }
trans->block_rsv = &root->fs_info->delalloc_block_rsv;
ret = btrfs_update_inode_fallback(trans, root, inode);
if (ret) /* -ENOMEM or corruption */
ordered_extent->len,
compress_type, 0, 0,
BTRFS_FILE_EXTENT_REG);
- unpin_extent_cache(&BTRFS_I(inode)->extent_tree,
- ordered_extent->file_offset,
- ordered_extent->len);
}
-
+ unpin_extent_cache(&BTRFS_I(inode)->extent_tree,
+ ordered_extent->file_offset, ordered_extent->len,
+ trans->transid);
if (ret < 0) {
btrfs_abort_transaction(trans, root, ret);
goto out_unlock;
btrfs_abort_transaction(trans, root, ret);
goto out_unlock;
}
+ } else {
+ btrfs_set_inode_last_trans(trans, inode);
}
ret = 0;
out_unlock:
insert = 1;
#endif
insert = 1;
- atomic_dec(&root->orphan_inodes);
+ atomic_inc(&root->orphan_inodes);
}
if (!test_and_set_bit(BTRFS_INODE_ORPHAN_META_RESERVED,
inode_set_bytes(inode, btrfs_inode_nbytes(leaf, inode_item));
BTRFS_I(inode)->generation = btrfs_inode_generation(leaf, inode_item);
+ BTRFS_I(inode)->last_trans = btrfs_inode_transid(leaf, inode_item);
+
+ /*
+ * If we were modified in the current generation and evicted from memory
+ * and then re-read we need to do a full sync since we don't have any
+ * idea about which extents were modified before we were evicted from
+ * cache.
+ */
+ if (BTRFS_I(inode)->last_trans == root->fs_info->generation)
+ set_bit(BTRFS_INODE_NEEDS_FULL_SYNC,
+ &BTRFS_I(inode)->runtime_flags);
+
inode->i_version = btrfs_inode_sequence(leaf, inode_item);
inode->i_generation = BTRFS_I(inode)->generation;
inode->i_rdev = 0;
return -ENOMEM;
path->reada = -1;
+ /*
+ * We want to drop from the next block forward in case this new size is
+ * not block aligned since we will be keeping the last block of the
+ * extent just the way it is.
+ */
if (root->ref_cows || root == root->fs_info->tree_root)
- btrfs_drop_extent_cache(inode, new_size & (~mask), (u64)-1, 0);
+ btrfs_drop_extent_cache(inode, (new_size + mask) & (~mask), (u64)-1, 0);
/*
* This function is also used to drop the items in the log tree before
if (path->slots[0] == 0 ||
path->slots[0] != pending_del_slot) {
- if (root->ref_cows &&
- BTRFS_I(inode)->location.objectid !=
- BTRFS_FREE_INO_OBJECTID) {
- err = -EAGAIN;
- goto out;
- }
if (pending_del_nr) {
ret = btrfs_del_items(trans, root, path,
pending_del_slot,
struct extent_io_tree *io_tree = &BTRFS_I(inode)->io_tree;
struct extent_map *em = NULL;
struct extent_state *cached_state = NULL;
+ struct extent_map_tree *em_tree = &BTRFS_I(inode)->extent_tree;
u64 mask = root->sectorsize - 1;
u64 hole_start = (oldsize + mask) & ~mask;
u64 block_end = (size + mask) & ~mask;
last_byte = min(extent_map_end(em), block_end);
last_byte = (last_byte + mask) & ~mask;
if (!test_bit(EXTENT_FLAG_PREALLOC, &em->flags)) {
+ struct extent_map *hole_em;
u64 hint_byte = 0;
hole_size = last_byte - cur_offset;
break;
}
- err = btrfs_drop_extents(trans, inode, cur_offset,
+ err = btrfs_drop_extents(trans, root, inode,
+ cur_offset,
cur_offset + hole_size,
&hint_byte, 1);
if (err) {
break;
}
- btrfs_drop_extent_cache(inode, hole_start,
- last_byte - 1, 0);
+ btrfs_drop_extent_cache(inode, cur_offset,
+ cur_offset + hole_size - 1, 0);
+ hole_em = alloc_extent_map();
+ if (!hole_em) {
+ set_bit(BTRFS_INODE_NEEDS_FULL_SYNC,
+ &BTRFS_I(inode)->runtime_flags);
+ goto next;
+ }
+ hole_em->start = cur_offset;
+ hole_em->len = hole_size;
+ hole_em->orig_start = cur_offset;
+ hole_em->block_start = EXTENT_MAP_HOLE;
+ hole_em->block_len = 0;
+ hole_em->bdev = root->fs_info->fs_devices->latest_bdev;
+ hole_em->compress_type = BTRFS_COMPRESS_NONE;
+ hole_em->generation = trans->transid;
+
+ while (1) {
+ write_lock(&em_tree->lock);
+ err = add_extent_mapping(em_tree, hole_em);
+ if (!err)
+ list_move(&hole_em->list,
+ &em_tree->modified_extents);
+ write_unlock(&em_tree->lock);
+ if (err != -EEXIST)
+ break;
+ btrfs_drop_extent_cache(inode, cur_offset,
+ cur_offset +
+ hole_size - 1, 0);
+ }
+ free_extent_map(hole_em);
+next:
btrfs_update_inode(trans, root, inode);
btrfs_end_transaction(trans, root);
}
goto no_delete;
}
rsv->size = min_size;
+ rsv->failfast = 1;
global_rsv = &root->fs_info->global_block_rsv;
btrfs_i_size_write(inode, 0);
trans->block_rsv = rsv;
ret = btrfs_truncate_inode_items(trans, root, inode, 0, 0);
- if (ret != -EAGAIN)
+ if (ret != -ENOSPC)
break;
nr = trans->blocks_used;
}
static struct dentry *btrfs_lookup(struct inode *dir, struct dentry *dentry,
- struct nameidata *nd)
+ unsigned int flags)
{
struct dentry *ret;
BTRFS_I(inode)->generation = trans->transid;
inode->i_generation = BTRFS_I(inode)->generation;
+ /*
+ * We could have gotten an inode number from somebody who was fsynced
+ * and then removed in this same transaction, so let's just set full
+ * sync since it will be a full sync anyway and this will blow away the
+ * old info in the log.
+ */
+ set_bit(BTRFS_INODE_NEEDS_FULL_SYNC, &BTRFS_I(inode)->runtime_flags);
+
if (S_ISDIR(mode))
owner = 0;
else
}
static int btrfs_create(struct inode *dir, struct dentry *dentry,
- umode_t mode, struct nameidata *nd)
+ umode_t mode, bool excl)
{
struct btrfs_trans_handle *trans;
struct btrfs_root *root = BTRFS_I(dir)->root;
u64 page_start;
u64 page_end;
+ sb_start_pagefault(inode->i_sb);
ret = btrfs_delalloc_reserve_space(inode, PAGE_CACHE_SIZE);
if (!ret) {
ret = file_update_time(vma->vm_file);
BTRFS_I(inode)->last_trans = root->fs_info->generation;
BTRFS_I(inode)->last_sub_trans = BTRFS_I(inode)->root->log_transid;
+ BTRFS_I(inode)->last_log_commit = BTRFS_I(inode)->root->last_log_commit;
unlock_extent_cached(io_tree, page_start, page_end, &cached_state, GFP_NOFS);
out_unlock:
- if (!ret)
+ if (!ret) {
+ sb_end_pagefault(inode->i_sb);
return VM_FAULT_LOCKED;
+ }
unlock_page(page);
out:
btrfs_delalloc_release_space(inode, PAGE_CACHE_SIZE);
out_noreserve:
+ sb_end_pagefault(inode->i_sb);
return ret;
}
if (!rsv)
return -ENOMEM;
rsv->size = min_size;
+ rsv->failfast = 1;
/*
* 1 for the truncate slack space
&BTRFS_I(inode)->runtime_flags))
btrfs_add_ordered_operation(trans, root, inode);
- while (1) {
- ret = btrfs_block_rsv_refill(root, rsv, min_size);
- if (ret) {
- /*
- * This can only happen with the original transaction we
- * started above, every other time we shouldn't have a
- * transaction started yet.
- */
- if (ret == -EAGAIN)
- goto end_trans;
- err = ret;
- break;
- }
-
- if (!trans) {
- /* Just need the 1 for updating the inode */
- trans = btrfs_start_transaction(root, 1);
- if (IS_ERR(trans)) {
- ret = err = PTR_ERR(trans);
- trans = NULL;
- break;
- }
- }
-
- trans->block_rsv = rsv;
+ /*
+ * So if we truncate and then write and fsync we normally would just
+ * write the extents that changed, which is a problem if we need to
+ * first truncate that entire inode. So set this flag so we write out
+ * all of the extents in the inode to the sync log so we're completely
+ * safe.
+ */
+ set_bit(BTRFS_INODE_NEEDS_FULL_SYNC, &BTRFS_I(inode)->runtime_flags);
+ trans->block_rsv = rsv;
+ while (1) {
ret = btrfs_truncate_inode_items(trans, root, inode,
inode->i_size,
BTRFS_EXTENT_DATA_KEY);
- if (ret != -EAGAIN) {
+ if (ret != -ENOSPC) {
err = ret;
break;
}
err = ret;
break;
}
-end_trans:
+
nr = trans->blocks_used;
btrfs_end_transaction(trans, root);
- trans = NULL;
btrfs_btree_balance_dirty(root, nr);
+
+ trans = btrfs_start_transaction(root, 2);
+ if (IS_ERR(trans)) {
+ ret = err = PTR_ERR(trans);
+ trans = NULL;
+ break;
+ }
+
+ ret = btrfs_block_rsv_migrate(&root->fs_info->trans_block_rsv,
+ rsv, min_size);
+ BUG_ON(ret); /* shouldn't happen */
+ trans->block_rsv = rsv;
}
if (ret == 0 && inode->i_nlink > 0) {
ei->csum_bytes = 0;
ei->index_cnt = (u64)-1;
ei->last_unlink_trans = 0;
+ ei->last_log_commit = 0;
spin_lock_init(&ei->lock);
ei->outstanding_extents = 0;
struct btrfs_ordered_extent *ordered;
struct btrfs_root *root = BTRFS_I(inode)->root;
- WARN_ON(!list_empty(&inode->i_dentry));
+ WARN_ON(!hlist_empty(&inode->i_dentry));
WARN_ON(inode->i_data.nrpages);
WARN_ON(BTRFS_I(inode)->outstanding_extents);
WARN_ON(BTRFS_I(inode)->reserved_extents);
loff_t actual_len, u64 *alloc_hint,
struct btrfs_trans_handle *trans)
{
+ struct extent_map_tree *em_tree = &BTRFS_I(inode)->extent_tree;
+ struct extent_map *em;
struct btrfs_root *root = BTRFS_I(inode)->root;
struct btrfs_key ins;
u64 cur_offset = start;
btrfs_drop_extent_cache(inode, cur_offset,
cur_offset + ins.offset -1, 0);
+ em = alloc_extent_map();
+ if (!em) {
+ set_bit(BTRFS_INODE_NEEDS_FULL_SYNC,
+ &BTRFS_I(inode)->runtime_flags);
+ goto next;
+ }
+
+ em->start = cur_offset;
+ em->orig_start = cur_offset;
+ em->len = ins.offset;
+ em->block_start = ins.objectid;
+ em->block_len = ins.offset;
+ em->bdev = root->fs_info->fs_devices->latest_bdev;
+ set_bit(EXTENT_FLAG_PREALLOC, &em->flags);
+ em->generation = trans->transid;
+
+ while (1) {
+ write_lock(&em_tree->lock);
+ ret = add_extent_mapping(em_tree, em);
+ if (!ret)
+ list_move(&em->list,
+ &em_tree->modified_extents);
+ write_unlock(&em_tree->lock);
+ if (ret != -EEXIST)
+ break;
+ btrfs_drop_extent_cache(inode, cur_offset,
+ cur_offset + ins.offset - 1,
+ 0);
+ }
+ free_extent_map(em);
+next:
num_bytes -= ins.offset;
cur_offset += ins.offset;
*alloc_hint = ins.objectid + ins.offset;
projects / cascardo / linux.git / blobdiff