#include <linux/bit_spinlock.h>
#include <linux/version.h>
#include <linux/xattr.h>
+#include <linux/posix_acl.h>
#include "ctree.h"
#include "disk-io.h"
#include "transaction.h"
#include "ioctl.h"
#include "print-tree.h"
#include "volumes.h"
+#include "ordered-data.h"
+#include "xattr.h"
+#include "compat.h"
+#include "tree-log.h"
struct btrfs_iget_args {
u64 ino;
[S_IFLNK >> S_SHIFT] = BTRFS_FT_SYMLINK,
};
+static void btrfs_truncate(struct inode *inode);
+
int btrfs_check_free_space(struct btrfs_root *root, u64 num_required,
int for_del)
{
u64 num_bytes;
u64 cur_alloc_size;
u64 blocksize = root->sectorsize;
- u64 orig_start = start;
u64 orig_num_bytes;
struct btrfs_key ins;
- int ret;
+ struct extent_map *em;
+ struct extent_map_tree *em_tree = &BTRFS_I(inode)->extent_tree;
+ int ret = 0;
- trans = btrfs_start_transaction(root, 1);
+ trans = btrfs_join_transaction(root, 1);
BUG_ON(!trans);
btrfs_set_trans_block_group(trans, inode);
num_bytes = (end - start + blocksize) & ~(blocksize - 1);
num_bytes = max(blocksize, num_bytes);
- ret = btrfs_drop_extents(trans, root, inode,
- start, start + num_bytes, start, &alloc_hint);
orig_num_bytes = num_bytes;
if (alloc_hint == EXTENT_MAP_INLINE)
goto out;
BUG_ON(num_bytes > btrfs_super_total_bytes(&root->fs_info->super_copy));
+ mutex_lock(&BTRFS_I(inode)->extent_mutex);
+ btrfs_drop_extent_cache(inode, start, start + num_bytes - 1);
+ mutex_unlock(&BTRFS_I(inode)->extent_mutex);
while(num_bytes > 0) {
cur_alloc_size = min(num_bytes, root->fs_info->max_extent);
- ret = btrfs_alloc_extent(trans, root, cur_alloc_size,
- root->sectorsize,
- root->root_key.objectid,
- trans->transid,
- inode->i_ino, start, 0,
- alloc_hint, (u64)-1, &ins, 1);
+ ret = btrfs_reserve_extent(trans, root, cur_alloc_size,
+ root->sectorsize, 0, alloc_hint,
+ (u64)-1, &ins, 1);
if (ret) {
WARN_ON(1);
goto out;
}
+ em = alloc_extent_map(GFP_NOFS);
+ em->start = start;
+ em->len = ins.offset;
+ em->block_start = ins.objectid;
+ em->bdev = root->fs_info->fs_devices->latest_bdev;
+ mutex_lock(&BTRFS_I(inode)->extent_mutex);
+ set_bit(EXTENT_FLAG_PINNED, &em->flags);
+ while(1) {
+ spin_lock(&em_tree->lock);
+ ret = add_extent_mapping(em_tree, em);
+ spin_unlock(&em_tree->lock);
+ if (ret != -EEXIST) {
+ free_extent_map(em);
+ break;
+ }
+ btrfs_drop_extent_cache(inode, start,
+ start + ins.offset - 1);
+ }
+ mutex_unlock(&BTRFS_I(inode)->extent_mutex);
+
cur_alloc_size = ins.offset;
- ret = btrfs_insert_file_extent(trans, root, inode->i_ino,
- start, ins.objectid, ins.offset,
+ ret = btrfs_add_ordered_extent(inode, start, ins.objectid,
ins.offset, 0);
- inode->i_blocks += ins.offset >> 9;
- btrfs_check_file(root, inode);
+ BUG_ON(ret);
if (num_bytes < cur_alloc_size) {
printk("num_bytes %Lu cur_alloc %Lu\n", num_bytes,
cur_alloc_size);
alloc_hint = ins.objectid + ins.offset;
start += cur_alloc_size;
}
- btrfs_drop_extent_cache(inode, orig_start,
- orig_start + orig_num_bytes - 1);
- btrfs_add_ordered_inode(inode);
- btrfs_update_inode(trans, root, inode);
out:
btrfs_end_transaction(trans, root);
return ret;
u64 extent_start;
u64 extent_end;
u64 bytenr;
- u64 cow_end;
u64 loops = 0;
u64 total_fs_bytes;
struct btrfs_root *root = BTRFS_I(inode)->root;
struct btrfs_block_group_cache *block_group;
+ struct btrfs_trans_handle *trans;
struct extent_buffer *leaf;
int found_type;
struct btrfs_path *path;
struct btrfs_file_extent_item *item;
int ret;
- int err;
+ int err = 0;
struct btrfs_key found_key;
total_fs_bytes = btrfs_super_total_bytes(&root->fs_info->super_copy);
path = btrfs_alloc_path();
BUG_ON(!path);
+ trans = btrfs_join_transaction(root, 1);
+ BUG_ON(!trans);
again:
ret = btrfs_lookup_file_extent(NULL, root, path,
inode->i_ino, start, 0);
if (ret < 0) {
- btrfs_free_path(path);
- return ret;
+ err = ret;
+ goto out;
}
- cow_end = end;
if (ret != 0) {
if (path->slots[0] == 0)
goto not_found;
if (start < extent_start || start >= extent_end)
goto not_found;
- cow_end = min(end, extent_end - 1);
bytenr = btrfs_file_extent_disk_bytenr(leaf, item);
if (bytenr == 0)
goto not_found;
- if (btrfs_count_snapshots_in_path(root, path, inode->i_ino,
- bytenr) != 1) {
+ if (btrfs_cross_ref_exists(trans, root, &found_key, bytenr))
goto not_found;
- }
-
/*
* we may be called by the resizer, make sure we're inside
* the limits of the FS
if (!block_group || block_group->ro)
goto not_found;
+ bytenr += btrfs_file_extent_offset(leaf, item);
+ extent_num_bytes = min(end + 1, extent_end) - start;
+ ret = btrfs_add_ordered_extent(inode, start, bytenr,
+ extent_num_bytes, 1);
+ if (ret) {
+ err = ret;
+ goto out;
+ }
+
+ btrfs_release_path(root, path);
start = extent_end;
+ if (start <= end) {
+ loops++;
+ goto again;
+ }
} else {
- goto not_found;
- }
-loop:
- if (start > end) {
+not_found:
+ btrfs_end_transaction(trans, root);
btrfs_free_path(path);
- return 0;
+ return cow_file_range(inode, start, end);
}
- btrfs_release_path(root, path);
- loops++;
- goto again;
-
-not_found:
- cow_file_range(inode, start, end);
- start = end + 1;
- goto loop;
+out:
+ WARN_ON(err);
+ btrfs_end_transaction(trans, root);
+ btrfs_free_path(path);
+ return err;
}
static int run_delalloc_range(struct inode *inode, u64 start, u64 end)
spin_lock_irqsave(&root->fs_info->delalloc_lock, flags);
BTRFS_I(inode)->delalloc_bytes += end - start + 1;
root->fs_info->delalloc_bytes += end - start + 1;
+ if (list_empty(&BTRFS_I(inode)->delalloc_inodes)) {
+ list_add_tail(&BTRFS_I(inode)->delalloc_inodes,
+ &root->fs_info->delalloc_inodes);
+ }
spin_unlock_irqrestore(&root->fs_info->delalloc_lock, flags);
}
return 0;
root->fs_info->delalloc_bytes -= end - start + 1;
BTRFS_I(inode)->delalloc_bytes -= end - start + 1;
}
+ if (BTRFS_I(inode)->delalloc_bytes == 0 &&
+ !list_empty(&BTRFS_I(inode)->delalloc_inodes)) {
+ list_del_init(&BTRFS_I(inode)->delalloc_inodes);
+ }
spin_unlock_irqrestore(&root->fs_info->delalloc_lock, flags);
}
return 0;
int mirror_num)
{
struct btrfs_root *root = BTRFS_I(inode)->root;
- struct btrfs_trans_handle *trans;
int ret = 0;
- char *sums = NULL;
-
- ret = btrfs_csum_one_bio(root, bio, &sums);
- BUG_ON(ret);
-
- trans = btrfs_start_transaction(root, 1);
-
- btrfs_set_trans_block_group(trans, inode);
- mutex_lock(&BTRFS_I(inode)->csum_mutex);
- btrfs_csum_file_blocks(trans, root, inode, bio, sums);
- mutex_unlock(&BTRFS_I(inode)->csum_mutex);
- ret = btrfs_end_transaction(trans, root);
+ ret = btrfs_csum_one_bio(root, inode, bio);
BUG_ON(ret);
- kfree(sums);
-
return btrfs_map_bio(root, rw, bio, mirror_num, 1);
}
struct btrfs_root *root = BTRFS_I(inode)->root;
int ret = 0;
- if (!(rw & (1 << BIO_RW))) {
- ret = btrfs_bio_wq_end_io(root->fs_info, bio, 0);
- BUG_ON(ret);
- goto mapit;
- }
+ ret = btrfs_bio_wq_end_io(root->fs_info, bio, 0);
+ BUG_ON(ret);
if (btrfs_test_opt(root, NODATASUM) ||
btrfs_test_flag(inode, NODATASUM)) {
goto mapit;
}
+ if (!(rw & (1 << BIO_RW))) {
+ btrfs_lookup_bio_sums(root, inode, bio);
+ goto mapit;
+ }
return btrfs_wq_submit_bio(BTRFS_I(inode)->root->fs_info,
inode, rw, bio, mirror_num,
__btrfs_submit_bio_hook);
return btrfs_map_bio(root, rw, bio, mirror_num, 0);
}
-int btrfs_readpage_io_hook(struct page *page, u64 start, u64 end)
+static noinline int add_pending_csums(struct btrfs_trans_handle *trans,
+ struct inode *inode, u64 file_offset,
+ struct list_head *list)
+{
+ struct list_head *cur;
+ struct btrfs_ordered_sum *sum;
+
+ btrfs_set_trans_block_group(trans, inode);
+ list_for_each(cur, list) {
+ sum = list_entry(cur, struct btrfs_ordered_sum, list);
+ btrfs_csum_file_blocks(trans, BTRFS_I(inode)->root,
+ inode, sum);
+ }
+ return 0;
+}
+
+int btrfs_set_extent_delalloc(struct inode *inode, u64 start, u64 end)
+{
+ return set_extent_delalloc(&BTRFS_I(inode)->io_tree, start, end,
+ GFP_NOFS);
+}
+
+struct btrfs_writepage_fixup {
+ struct page *page;
+ struct btrfs_work work;
+};
+
+/* see btrfs_writepage_start_hook for details on why this is required */
+void btrfs_writepage_fixup_worker(struct btrfs_work *work)
+{
+ struct btrfs_writepage_fixup *fixup;
+ struct btrfs_ordered_extent *ordered;
+ struct page *page;
+ struct inode *inode;
+ u64 page_start;
+ u64 page_end;
+
+ fixup = container_of(work, struct btrfs_writepage_fixup, work);
+ page = fixup->page;
+again:
+ lock_page(page);
+ if (!page->mapping || !PageDirty(page) || !PageChecked(page)) {
+ ClearPageChecked(page);
+ goto out_page;
+ }
+
+ inode = page->mapping->host;
+ 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);
+
+ /* already ordered? We're done */
+ if (test_range_bit(&BTRFS_I(inode)->io_tree, page_start, page_end,
+ EXTENT_ORDERED, 0)) {
+ goto out;
+ }
+
+ ordered = btrfs_lookup_ordered_extent(inode, page_start);
+ if (ordered) {
+ unlock_extent(&BTRFS_I(inode)->io_tree, page_start,
+ page_end, GFP_NOFS);
+ unlock_page(page);
+ btrfs_start_ordered_extent(inode, ordered, 1);
+ goto again;
+ }
+
+ btrfs_set_extent_delalloc(inode, page_start, page_end);
+ ClearPageChecked(page);
+out:
+ unlock_extent(&BTRFS_I(inode)->io_tree, page_start, page_end, GFP_NOFS);
+out_page:
+ unlock_page(page);
+ page_cache_release(page);
+}
+
+/*
+ * There are a few paths in the higher layers of the kernel that directly
+ * set the page dirty bit without asking the filesystem if it is a
+ * good idea. This causes problems because we want to make sure COW
+ * properly happens and the data=ordered rules are followed.
+ *
+ * In our case any range that doesn't have the EXTENT_ORDERED bit set
+ * hasn't been properly setup for IO. We kick off an async process
+ * to fix it up. The async helper will wait for ordered extents, set
+ * the delalloc bit and make it safe to write the page.
+ */
+int btrfs_writepage_start_hook(struct page *page, u64 start, u64 end)
{
- int ret = 0;
struct inode *inode = page->mapping->host;
+ struct btrfs_writepage_fixup *fixup;
+ struct btrfs_root *root = BTRFS_I(inode)->root;
+ int ret;
+
+ ret = test_range_bit(&BTRFS_I(inode)->io_tree, start, end,
+ EXTENT_ORDERED, 0);
+ if (ret)
+ return 0;
+
+ if (PageChecked(page))
+ return -EAGAIN;
+
+ fixup = kzalloc(sizeof(*fixup), GFP_NOFS);
+ if (!fixup)
+ return -EAGAIN;
+
+ SetPageChecked(page);
+ page_cache_get(page);
+ fixup->work.func = btrfs_writepage_fixup_worker;
+ fixup->page = page;
+ btrfs_queue_worker(&root->fs_info->fixup_workers, &fixup->work);
+ return -EAGAIN;
+}
+
+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_ordered_extent *ordered_extent;
struct extent_io_tree *io_tree = &BTRFS_I(inode)->io_tree;
- struct btrfs_csum_item *item;
+ struct btrfs_file_extent_item *extent_item;
struct btrfs_path *path = NULL;
- u32 csum;
+ struct extent_buffer *leaf;
+ u64 alloc_hint = 0;
+ struct list_head list;
+ struct btrfs_key ins;
+ int ret;
- if (btrfs_test_opt(root, NODATASUM) ||
- btrfs_test_flag(inode, NODATASUM))
+ ret = btrfs_dec_test_ordered_pending(inode, start, end - start + 1);
+ if (!ret)
return 0;
+ trans = btrfs_join_transaction(root, 1);
+
+ ordered_extent = btrfs_lookup_ordered_extent(inode, start);
+ BUG_ON(!ordered_extent);
+ if (test_bit(BTRFS_ORDERED_NOCOW, &ordered_extent->flags))
+ goto nocow;
+
path = btrfs_alloc_path();
- item = btrfs_lookup_csum(NULL, root, path, inode->i_ino, start, 0);
- if (IS_ERR(item)) {
- ret = PTR_ERR(item);
- /* a csum that isn't present is a preallocated region. */
- if (ret == -ENOENT || ret == -EFBIG)
- ret = 0;
- csum = 0;
- printk("no csum found for inode %lu start %Lu\n", inode->i_ino, start);
- goto out;
- }
- read_extent_buffer(path->nodes[0], &csum, (unsigned long)item,
- BTRFS_CRC32_SIZE);
- set_state_private(io_tree, start, csum);
-out:
+ BUG_ON(!path);
+
+ lock_extent(io_tree, ordered_extent->file_offset,
+ ordered_extent->file_offset + ordered_extent->len - 1,
+ GFP_NOFS);
+
+ INIT_LIST_HEAD(&list);
+
+ mutex_lock(&BTRFS_I(inode)->extent_mutex);
+
+ ret = btrfs_drop_extents(trans, root, inode,
+ ordered_extent->file_offset,
+ ordered_extent->file_offset +
+ ordered_extent->len,
+ ordered_extent->file_offset, &alloc_hint);
+ BUG_ON(ret);
+
+ ins.objectid = inode->i_ino;
+ ins.offset = ordered_extent->file_offset;
+ ins.type = BTRFS_EXTENT_DATA_KEY;
+ ret = btrfs_insert_empty_item(trans, root, path, &ins,
+ sizeof(*extent_item));
+ BUG_ON(ret);
+ leaf = path->nodes[0];
+ extent_item = btrfs_item_ptr(leaf, path->slots[0],
+ struct btrfs_file_extent_item);
+ btrfs_set_file_extent_generation(leaf, extent_item, trans->transid);
+ btrfs_set_file_extent_type(leaf, extent_item, BTRFS_FILE_EXTENT_REG);
+ btrfs_set_file_extent_disk_bytenr(leaf, extent_item,
+ ordered_extent->start);
+ btrfs_set_file_extent_disk_num_bytes(leaf, extent_item,
+ ordered_extent->len);
+ btrfs_set_file_extent_offset(leaf, extent_item, 0);
+ btrfs_set_file_extent_num_bytes(leaf, extent_item,
+ ordered_extent->len);
+ btrfs_mark_buffer_dirty(leaf);
+
+ btrfs_drop_extent_cache(inode, ordered_extent->file_offset,
+ ordered_extent->file_offset +
+ ordered_extent->len - 1);
+ mutex_unlock(&BTRFS_I(inode)->extent_mutex);
+
+ ins.objectid = ordered_extent->start;
+ ins.offset = ordered_extent->len;
+ ins.type = BTRFS_EXTENT_ITEM_KEY;
+ ret = btrfs_alloc_reserved_extent(trans, root, leaf->start,
+ root->root_key.objectid,
+ trans->transid, inode->i_ino,
+ ordered_extent->file_offset, &ins);
+ BUG_ON(ret);
+ btrfs_release_path(root, path);
+
+ inode->i_blocks += ordered_extent->len >> 9;
+ unlock_extent(io_tree, ordered_extent->file_offset,
+ ordered_extent->file_offset + ordered_extent->len - 1,
+ GFP_NOFS);
+nocow:
+ add_pending_csums(trans, inode, ordered_extent->file_offset,
+ &ordered_extent->list);
+
+ mutex_lock(&BTRFS_I(inode)->extent_mutex);
+ btrfs_ordered_update_i_size(inode, ordered_extent);
+ btrfs_update_inode(trans, root, inode);
+ btrfs_remove_ordered_extent(inode, ordered_extent);
+ mutex_unlock(&BTRFS_I(inode)->extent_mutex);
+
+ /* once for us */
+ btrfs_put_ordered_extent(ordered_extent);
+ /* once for the tree */
+ btrfs_put_ordered_extent(ordered_extent);
+
+ btrfs_end_transaction(trans, root);
if (path)
btrfs_free_path(path);
- return ret;
+ return 0;
+}
+
+int btrfs_writepage_end_io_hook(struct page *page, u64 start, u64 end,
+ struct extent_state *state, int uptodate)
+{
+ return btrfs_finish_ordered_io(page->mapping->host, start, end);
}
struct io_failure_record {
return -EIO;
}
+/*
+ * This creates an orphan entry for the given inode in case something goes
+ * wrong in the middle of an unlink/truncate.
+ */
+int btrfs_orphan_add(struct btrfs_trans_handle *trans, struct inode *inode)
+{
+ struct btrfs_root *root = BTRFS_I(inode)->root;
+ int ret = 0;
+
+ spin_lock(&root->list_lock);
+
+ /* already on the orphan list, we're good */
+ if (!list_empty(&BTRFS_I(inode)->i_orphan)) {
+ spin_unlock(&root->list_lock);
+ return 0;
+ }
+
+ list_add(&BTRFS_I(inode)->i_orphan, &root->orphan_list);
+
+ spin_unlock(&root->list_lock);
+
+ /*
+ * insert an orphan item to track this unlinked/truncated file
+ */
+ ret = btrfs_insert_orphan_item(trans, root, inode->i_ino);
+
+ return ret;
+}
+
+/*
+ * We have done the truncate/delete so we can go ahead and remove the orphan
+ * item for this particular inode.
+ */
+int btrfs_orphan_del(struct btrfs_trans_handle *trans, struct inode *inode)
+{
+ struct btrfs_root *root = BTRFS_I(inode)->root;
+ int ret = 0;
+
+ spin_lock(&root->list_lock);
+
+ if (list_empty(&BTRFS_I(inode)->i_orphan)) {
+ spin_unlock(&root->list_lock);
+ return 0;
+ }
+
+ list_del_init(&BTRFS_I(inode)->i_orphan);
+ if (!trans) {
+ spin_unlock(&root->list_lock);
+ return 0;
+ }
+
+ spin_unlock(&root->list_lock);
+
+ ret = btrfs_del_orphan_item(trans, root, inode->i_ino);
+
+ return ret;
+}
+
+/*
+ * this cleans up any orphans that may be left on the list from the last use
+ * of this root.
+ */
+void btrfs_orphan_cleanup(struct btrfs_root *root)
+{
+ struct btrfs_path *path;
+ struct extent_buffer *leaf;
+ struct btrfs_item *item;
+ struct btrfs_key key, found_key;
+ struct btrfs_trans_handle *trans;
+ struct inode *inode;
+ int ret = 0, nr_unlink = 0, nr_truncate = 0;
+
+ /* don't do orphan cleanup if the fs is readonly. */
+ if (root->inode->i_sb->s_flags & MS_RDONLY)
+ return;
+
+ path = btrfs_alloc_path();
+ if (!path)
+ return;
+ path->reada = -1;
+
+ key.objectid = BTRFS_ORPHAN_OBJECTID;
+ btrfs_set_key_type(&key, BTRFS_ORPHAN_ITEM_KEY);
+ key.offset = (u64)-1;
+
+ trans = btrfs_start_transaction(root, 1);
+ btrfs_set_trans_block_group(trans, root->inode);
+
+ while (1) {
+ ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
+ if (ret < 0) {
+ printk(KERN_ERR "Error searching slot for orphan: %d"
+ "\n", ret);
+ break;
+ }
+
+ /*
+ * if ret == 0 means we found what we were searching for, which
+ * is weird, but possible, so only screw with path if we didnt
+ * find the key and see if we have stuff that matches
+ */
+ if (ret > 0) {
+ if (path->slots[0] == 0)
+ break;
+ path->slots[0]--;
+ }
+
+ /* pull out the item */
+ leaf = path->nodes[0];
+ item = btrfs_item_nr(leaf, path->slots[0]);
+ btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]);
+
+ /* make sure the item matches what we want */
+ if (found_key.objectid != BTRFS_ORPHAN_OBJECTID)
+ break;
+ if (btrfs_key_type(&found_key) != BTRFS_ORPHAN_ITEM_KEY)
+ break;
+
+ /* release the path since we're done with it */
+ btrfs_release_path(root, path);
+
+ /*
+ * this is where we are basically btrfs_lookup, without the
+ * crossing root thing. we store the inode number in the
+ * offset of the orphan item.
+ */
+ inode = btrfs_iget_locked(root->inode->i_sb,
+ found_key.offset, root);
+ if (!inode)
+ break;
+
+ if (inode->i_state & I_NEW) {
+ BTRFS_I(inode)->root = root;
+
+ /* have to set the location manually */
+ BTRFS_I(inode)->location.objectid = inode->i_ino;
+ BTRFS_I(inode)->location.type = BTRFS_INODE_ITEM_KEY;
+ BTRFS_I(inode)->location.offset = 0;
+
+ btrfs_read_locked_inode(inode);
+ unlock_new_inode(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);
+ list_add(&BTRFS_I(inode)->i_orphan, &root->orphan_list);
+ spin_unlock(&root->list_lock);
+
+ /*
+ * if this is a bad inode, means we actually succeeded in
+ * removing the inode, but not the orphan record, which means
+ * we need to manually delete the orphan since iput will just
+ * do a destroy_inode
+ */
+ if (is_bad_inode(inode)) {
+ btrfs_orphan_del(trans, inode);
+ iput(inode);
+ continue;
+ }
+
+ /* if we have links, this was a truncate, lets do that */
+ if (inode->i_nlink) {
+ nr_truncate++;
+ btrfs_truncate(inode);
+ } else {
+ nr_unlink++;
+ }
+
+ /* this will do delete_inode and everything for us */
+ iput(inode);
+ }
+
+ 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);
+ btrfs_end_transaction(trans, root);
+}
+
void btrfs_read_locked_inode(struct inode *inode)
{
struct btrfs_path *path;
inode->i_nlink = btrfs_inode_nlink(leaf, inode_item);
inode->i_uid = btrfs_inode_uid(leaf, inode_item);
inode->i_gid = btrfs_inode_gid(leaf, inode_item);
- inode->i_size = btrfs_inode_size(leaf, inode_item);
+ btrfs_i_size_write(inode, btrfs_inode_size(leaf, inode_item));
tspec = btrfs_inode_atime(inode_item);
inode->i_atime.tv_sec = btrfs_timespec_sec(leaf, tspec);
inode->i_ctime.tv_nsec = btrfs_timespec_nsec(leaf, tspec);
inode->i_blocks = btrfs_inode_nblocks(leaf, inode_item);
- inode->i_generation = btrfs_inode_generation(leaf, inode_item);
+ BTRFS_I(inode)->generation = btrfs_inode_generation(leaf, inode_item);
+ inode->i_generation = BTRFS_I(inode)->generation;
inode->i_rdev = 0;
rdev = btrfs_inode_rdev(leaf, inode_item);
+ BTRFS_I(inode)->index_cnt = (u64)-1;
+
alloc_group_block = btrfs_inode_block_group(leaf, inode_item);
BTRFS_I(inode)->block_group = btrfs_lookup_block_group(root->fs_info,
alloc_group_block);
make_bad_inode(inode);
}
-static void fill_inode_item(struct extent_buffer *leaf,
+static void fill_inode_item(struct btrfs_trans_handle *trans,
+ struct extent_buffer *leaf,
struct btrfs_inode_item *item,
struct inode *inode)
{
btrfs_set_inode_uid(leaf, item, inode->i_uid);
btrfs_set_inode_gid(leaf, item, inode->i_gid);
- btrfs_set_inode_size(leaf, item, inode->i_size);
+ btrfs_set_inode_size(leaf, item, BTRFS_I(inode)->disk_i_size);
btrfs_set_inode_mode(leaf, item, inode->i_mode);
btrfs_set_inode_nlink(leaf, item, inode->i_nlink);
inode->i_ctime.tv_nsec);
btrfs_set_inode_nblocks(leaf, item, inode->i_blocks);
- btrfs_set_inode_generation(leaf, item, inode->i_generation);
+ btrfs_set_inode_generation(leaf, item, BTRFS_I(inode)->generation);
+ btrfs_set_inode_transid(leaf, item, trans->transid);
btrfs_set_inode_rdev(leaf, item, inode->i_rdev);
btrfs_set_inode_flags(leaf, item, BTRFS_I(inode)->flags);
btrfs_set_inode_block_group(leaf, item,
BTRFS_I(inode)->block_group->key.objectid);
}
-int btrfs_update_inode(struct btrfs_trans_handle *trans,
+int noinline btrfs_update_inode(struct btrfs_trans_handle *trans,
struct btrfs_root *root,
struct inode *inode)
{
inode_item = btrfs_item_ptr(leaf, path->slots[0],
struct btrfs_inode_item);
- fill_inode_item(leaf, inode_item, inode);
+ fill_inode_item(trans, leaf, inode_item, inode);
btrfs_mark_buffer_dirty(leaf);
btrfs_set_inode_last_trans(trans, inode);
ret = 0;
}
-static int btrfs_unlink_trans(struct btrfs_trans_handle *trans,
- struct btrfs_root *root,
- struct inode *dir,
- struct dentry *dentry)
+int btrfs_unlink_inode(struct btrfs_trans_handle *trans,
+ struct btrfs_root *root,
+ struct inode *dir, struct inode *inode,
+ const char *name, int name_len)
{
struct btrfs_path *path;
- const char *name = dentry->d_name.name;
- int name_len = dentry->d_name.len;
int ret = 0;
struct extent_buffer *leaf;
struct btrfs_dir_item *di;
struct btrfs_key key;
+ u64 index;
path = btrfs_alloc_path();
if (!path) {
goto err;
btrfs_release_path(root, path);
+ ret = btrfs_del_inode_ref(trans, root, name, name_len,
+ inode->i_ino,
+ dir->i_ino, &index);
+ if (ret) {
+ printk("failed to delete reference to %.*s, "
+ "inode %lu parent %lu\n", name_len, name,
+ inode->i_ino, dir->i_ino);
+ goto err;
+ }
+
di = btrfs_lookup_dir_index_item(trans, root, path, dir->i_ino,
- key.objectid, name, name_len, -1);
+ index, name, name_len, -1);
if (IS_ERR(di)) {
ret = PTR_ERR(di);
goto err;
ret = btrfs_delete_one_dir_name(trans, root, path, di);
btrfs_release_path(root, path);
- dentry->d_inode->i_ctime = dir->i_ctime;
- ret = btrfs_del_inode_ref(trans, root, name, name_len,
- dentry->d_inode->i_ino,
- dentry->d_parent->d_inode->i_ino);
- if (ret) {
- printk("failed to delete reference to %.*s, "
- "inode %lu parent %lu\n", name_len, name,
- dentry->d_inode->i_ino,
- dentry->d_parent->d_inode->i_ino);
- }
+ ret = btrfs_del_inode_ref_in_log(trans, root, name, name_len,
+ inode, dir->i_ino);
+ BUG_ON(ret != 0 && ret != -ENOENT);
+ if (ret != -ENOENT)
+ BTRFS_I(dir)->log_dirty_trans = trans->transid;
+
+ ret = btrfs_del_dir_entries_in_log(trans, root, name, name_len,
+ dir, index);
+ BUG_ON(ret);
err:
btrfs_free_path(path);
- if (!ret) {
- dir->i_size -= name_len * 2;
- dir->i_mtime = dir->i_ctime = CURRENT_TIME;
- btrfs_update_inode(trans, root, dir);
-#if LINUX_VERSION_CODE <= KERNEL_VERSION(2,6,18)
- dentry->d_inode->i_nlink--;
-#else
- drop_nlink(dentry->d_inode);
-#endif
- ret = btrfs_update_inode(trans, root, dentry->d_inode);
- dir->i_sb->s_dirt = 1;
- }
+ if (ret)
+ goto out;
+
+ btrfs_i_size_write(dir, dir->i_size - name_len * 2);
+ inode->i_ctime = dir->i_mtime = dir->i_ctime = CURRENT_TIME;
+ btrfs_update_inode(trans, root, dir);
+ btrfs_drop_nlink(inode);
+ ret = btrfs_update_inode(trans, root, inode);
+ dir->i_sb->s_dirt = 1;
+out:
return ret;
}
trans = btrfs_start_transaction(root, 1);
btrfs_set_trans_block_group(trans, dir);
- ret = btrfs_unlink_trans(trans, root, dir, dentry);
- nr = trans->blocks_used;
+ ret = btrfs_unlink_inode(trans, root, dir, dentry->d_inode,
+ dentry->d_name.name, dentry->d_name.len);
- if (inode->i_nlink == 0) {
- /* if the inode isn't linked anywhere,
- * we don't need to worry about
- * data=ordered
- */
- btrfs_del_ordered_inode(inode, 1);
- }
+ if (inode->i_nlink == 0)
+ ret = btrfs_orphan_add(trans, inode);
+
+ nr = trans->blocks_used;
btrfs_end_transaction_throttle(trans, root);
fail:
trans = btrfs_start_transaction(root, 1);
btrfs_set_trans_block_group(trans, dir);
+ err = btrfs_orphan_add(trans, inode);
+ if (err)
+ goto fail_trans;
+
/* now the directory is empty */
- err = btrfs_unlink_trans(trans, root, dir, dentry);
+ err = btrfs_unlink_inode(trans, root, dir, dentry->d_inode,
+ dentry->d_name.name, dentry->d_name.len);
if (!err) {
- inode->i_size = 0;
+ btrfs_i_size_write(inode, 0);
}
+fail_trans:
nr = trans->blocks_used;
ret = btrfs_end_transaction_throttle(trans, root);
fail:
*
* csum items that cross the new i_size are truncated to the new size
* as well.
+ *
+ * min_type is the minimum key type to truncate down to. If set to 0, this
+ * will kill all the items on this inode, including the INODE_ITEM_KEY.
*/
-static int btrfs_truncate_in_trans(struct btrfs_trans_handle *trans,
- struct btrfs_root *root,
- struct inode *inode,
- u32 min_type)
+noinline int btrfs_truncate_inode_items(struct btrfs_trans_handle *trans,
+ struct btrfs_root *root,
+ struct inode *inode,
+ u64 new_size, u32 min_type)
{
int ret;
struct btrfs_path *path;
int extent_type = -1;
u64 mask = root->sectorsize - 1;
- btrfs_drop_extent_cache(inode, inode->i_size & (~mask), (u64)-1);
+ if (root->ref_cows)
+ btrfs_drop_extent_cache(inode,
+ new_size & (~mask), (u64)-1);
path = btrfs_alloc_path();
path->reada = -1;
BUG_ON(!path);
goto error;
}
if (ret > 0) {
- BUG_ON(path->slots[0] == 0);
+ /* there are no items in the tree for us to truncate, we're
+ * done
+ */
+ if (path->slots[0] == 0) {
+ ret = 0;
+ goto error;
+ }
path->slots[0]--;
}
}
if (found_type == BTRFS_CSUM_ITEM_KEY) {
ret = btrfs_csum_truncate(trans, root, path,
- inode->i_size);
+ new_size);
BUG_ON(ret);
}
- if (item_end < inode->i_size) {
+ if (item_end < new_size) {
if (found_type == BTRFS_DIR_ITEM_KEY) {
found_type = BTRFS_INODE_ITEM_KEY;
} else if (found_type == BTRFS_EXTENT_ITEM_KEY) {
btrfs_set_key_type(&key, found_type);
goto next;
}
- if (found_key.offset >= inode->i_size)
+ if (found_key.offset >= new_size)
del_item = 1;
else
del_item = 0;
if (!del_item) {
u64 orig_num_bytes =
btrfs_file_extent_num_bytes(leaf, fi);
- extent_num_bytes = inode->i_size -
+ extent_num_bytes = new_size -
found_key.offset + root->sectorsize - 1;
extent_num_bytes = extent_num_bytes &
~((u64)root->sectorsize - 1);
extent_num_bytes);
num_dec = (orig_num_bytes -
extent_num_bytes);
- if (extent_start != 0)
+ if (root->ref_cows && extent_start != 0)
dec_i_blocks(inode, num_dec);
btrfs_mark_buffer_dirty(leaf);
} else {
num_dec = btrfs_file_extent_num_bytes(leaf, fi);
if (extent_start != 0) {
found_extent = 1;
- dec_i_blocks(inode, num_dec);
+ if (root->ref_cows)
+ dec_i_blocks(inode, num_dec);
}
root_gen = btrfs_header_generation(leaf);
root_owner = btrfs_header_owner(leaf);
}
} else if (extent_type == BTRFS_FILE_EXTENT_INLINE) {
if (!del_item) {
- u32 newsize = inode->i_size - found_key.offset;
- dec_i_blocks(inode, item_end + 1 -
- found_key.offset - newsize);
- newsize =
- btrfs_file_extent_calc_inline_size(newsize);
+ u32 size = new_size - found_key.offset;
+
+ if (root->ref_cows) {
+ dec_i_blocks(inode, item_end + 1 -
+ found_key.offset - size);
+ }
+ size =
+ btrfs_file_extent_calc_inline_size(size);
ret = btrfs_truncate_item(trans, root, path,
- newsize, 1);
+ size, 1);
BUG_ON(ret);
- } else {
+ } else if (root->ref_cows) {
dec_i_blocks(inode, item_end + 1 -
found_key.offset);
}
if (found_extent) {
ret = btrfs_free_extent(trans, root, extent_start,
extent_num_bytes,
- root_owner,
+ leaf->start, root_owner,
root_gen, inode->i_ino,
found_key.offset, 0);
BUG_ON(ret);
return ret;
}
-static int btrfs_cow_one_page(struct inode *inode, struct page *page,
- size_t zero_start)
-{
- char *kaddr;
- struct extent_io_tree *io_tree = &BTRFS_I(inode)->io_tree;
- u64 page_start = (u64)page->index << PAGE_CACHE_SHIFT;
- u64 page_end = page_start + PAGE_CACHE_SIZE - 1;
- int ret = 0;
-
- WARN_ON(!PageLocked(page));
- set_page_extent_mapped(page);
-
- lock_extent(io_tree, page_start, page_end, GFP_NOFS);
- set_extent_delalloc(&BTRFS_I(inode)->io_tree, page_start,
- page_end, GFP_NOFS);
-
- if (zero_start != PAGE_CACHE_SIZE) {
- kaddr = kmap(page);
- memset(kaddr + zero_start, 0, PAGE_CACHE_SIZE - zero_start);
- flush_dcache_page(page);
- kunmap(page);
- }
- set_page_dirty(page);
- unlock_extent(io_tree, page_start, page_end, GFP_NOFS);
-
- return ret;
-}
-
/*
* taken from block_truncate_page, but does cow as it zeros out
* any bytes left in the last page in the file.
{
struct inode *inode = mapping->host;
struct btrfs_root *root = BTRFS_I(inode)->root;
+ struct extent_io_tree *io_tree = &BTRFS_I(inode)->io_tree;
+ struct btrfs_ordered_extent *ordered;
+ char *kaddr;
u32 blocksize = root->sectorsize;
pgoff_t index = from >> PAGE_CACHE_SHIFT;
unsigned offset = from & (PAGE_CACHE_SIZE-1);
struct page *page;
int ret = 0;
u64 page_start;
+ u64 page_end;
if ((offset & (blocksize - 1)) == 0)
goto out;
page = grab_cache_page(mapping, index);
if (!page)
goto out;
+
+ page_start = page_offset(page);
+ page_end = page_start + PAGE_CACHE_SIZE - 1;
+
if (!PageUptodate(page)) {
ret = btrfs_readpage(NULL, page);
lock_page(page);
}
if (!PageUptodate(page)) {
ret = -EIO;
- goto out;
+ goto out_unlock;
}
}
-
- page_start = (u64)page->index << PAGE_CACHE_SHIFT;
wait_on_page_writeback(page);
- ret = btrfs_cow_one_page(inode, page, offset);
+ lock_extent(io_tree, page_start, page_end, 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_page(page);
+ page_cache_release(page);
+ btrfs_start_ordered_extent(inode, ordered, 1);
+ btrfs_put_ordered_extent(ordered);
+ goto again;
+ }
+
+ btrfs_set_extent_delalloc(inode, page_start, page_end);
+ ret = 0;
+ if (offset != PAGE_CACHE_SIZE) {
+ kaddr = kmap(page);
+ memset(kaddr + offset, 0, PAGE_CACHE_SIZE - offset);
+ flush_dcache_page(page);
+ kunmap(page);
+ }
+ ClearPageChecked(page);
+ set_page_dirty(page);
+ unlock_extent(io_tree, page_start, page_end, GFP_NOFS);
+
+out_unlock:
unlock_page(page);
page_cache_release(page);
out:
btrfs_truncate_page(inode->i_mapping, inode->i_size);
- lock_extent(io_tree, hole_start, block_end - 1, GFP_NOFS);
hole_size = block_end - hole_start;
+ while(1) {
+ struct btrfs_ordered_extent *ordered;
+ btrfs_wait_ordered_range(inode, hole_start, hole_size);
+
+ lock_extent(io_tree, hole_start, block_end - 1, GFP_NOFS);
+ ordered = btrfs_lookup_ordered_extent(inode, hole_start);
+ if (ordered) {
+ unlock_extent(io_tree, hole_start,
+ block_end - 1, GFP_NOFS);
+ btrfs_put_ordered_extent(ordered);
+ } else {
+ break;
+ }
+ }
trans = btrfs_start_transaction(root, 1);
btrfs_set_trans_block_group(trans, inode);
+ mutex_lock(&BTRFS_I(inode)->extent_mutex);
err = btrfs_drop_extents(trans, root, inode,
hole_start, block_end, hole_start,
&alloc_hint);
(u64)-1);
btrfs_check_file(root, inode);
}
+ mutex_unlock(&BTRFS_I(inode)->extent_mutex);
btrfs_end_transaction(trans, root);
unlock_extent(io_tree, hole_start, block_end - 1, GFP_NOFS);
if (err)
}
out:
err = inode_setattr(inode, attr);
+
+ if (!err && ((attr->ia_valid & ATTR_MODE)))
+ err = btrfs_acl_chmod(inode);
fail:
return err;
}
truncate_inode_pages(&inode->i_data, 0);
if (is_bad_inode(inode)) {
+ btrfs_orphan_del(NULL, inode);
goto no_delete;
}
+ btrfs_wait_ordered_range(inode, 0, (u64)-1);
- inode->i_size = 0;
+ btrfs_i_size_write(inode, 0);
trans = btrfs_start_transaction(root, 1);
btrfs_set_trans_block_group(trans, inode);
- ret = btrfs_truncate_in_trans(trans, root, inode, 0);
- if (ret)
+ ret = btrfs_truncate_inode_items(trans, root, inode, inode->i_size, 0);
+ if (ret) {
+ btrfs_orphan_del(NULL, inode);
goto no_delete_lock;
+ }
+
+ btrfs_orphan_del(trans, inode);
nr = trans->blocks_used;
clear_inode(inode);
struct btrfs_root *root = BTRFS_I(dir)->root;
int ret = 0;
- if (namelen == 1 && strcmp(name, ".") == 0) {
- location->objectid = dir->i_ino;
- location->type = BTRFS_INODE_ITEM_KEY;
- location->offset = 0;
- return 0;
- }
path = btrfs_alloc_path();
BUG_ON(!path);
- if (namelen == 2 && strcmp(name, "..") == 0) {
- struct btrfs_key key;
- struct extent_buffer *leaf;
- u32 nritems;
- int slot;
-
- key.objectid = dir->i_ino;
- btrfs_set_key_type(&key, BTRFS_INODE_REF_KEY);
- key.offset = 0;
- ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
- BUG_ON(ret == 0);
- ret = 0;
-
- leaf = path->nodes[0];
- slot = path->slots[0];
- nritems = btrfs_header_nritems(leaf);
- if (slot >= nritems)
- goto out_err;
-
- btrfs_item_key_to_cpu(leaf, &key, slot);
- if (key.objectid != dir->i_ino ||
- key.type != BTRFS_INODE_REF_KEY) {
- goto out_err;
- }
- location->objectid = key.offset;
- location->type = BTRFS_INODE_ITEM_KEY;
- location->offset = 0;
- goto out;
- }
-
di = btrfs_lookup_dir_item(NULL, root, path, dir->i_ino, name,
namelen, 0);
if (IS_ERR(di))
struct btrfs_root **sub_root,
struct dentry *dentry)
{
- struct btrfs_path *path;
struct btrfs_root_item *ri;
if (btrfs_key_type(location) != BTRFS_ROOT_ITEM_KEY)
if (location->objectid == BTRFS_ROOT_TREE_OBJECTID)
return 0;
- path = btrfs_alloc_path();
- BUG_ON(!path);
-
*sub_root = btrfs_read_fs_root(root->fs_info, location,
dentry->d_name.name,
dentry->d_name.len);
btrfs_set_key_type(location, BTRFS_INODE_ITEM_KEY);
location->offset = 0;
- btrfs_free_path(path);
return 0;
}
-static int btrfs_init_locked_inode(struct inode *inode, void *p)
+static noinline void init_btrfs_i(struct inode *inode)
{
- struct btrfs_iget_args *args = p;
- inode->i_ino = args->ino;
- BTRFS_I(inode)->root = args->root;
- BTRFS_I(inode)->delalloc_bytes = 0;
+ struct btrfs_inode *bi = BTRFS_I(inode);
+
+ bi->i_acl = NULL;
+ bi->i_default_acl = NULL;
+
+ bi->generation = 0;
+ bi->last_trans = 0;
+ bi->logged_trans = 0;
+ bi->delalloc_bytes = 0;
+ bi->disk_i_size = 0;
+ bi->flags = 0;
+ bi->index_cnt = (u64)-1;
+ bi->log_dirty_trans = 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);
+ btrfs_ordered_inode_tree_init(&BTRFS_I(inode)->ordered_tree);
mutex_init(&BTRFS_I(inode)->csum_mutex);
- atomic_set(&BTRFS_I(inode)->ordered_writeback, 0);
+ mutex_init(&BTRFS_I(inode)->extent_mutex);
+ 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;
return 0;
}
args->root == BTRFS_I(inode)->root);
}
-struct inode *btrfs_ilookup(struct super_block *s, u64 objectid,
- u64 root_objectid)
-{
- struct btrfs_iget_args args;
- args.ino = objectid;
- args.root = btrfs_lookup_fs_root(btrfs_sb(s)->fs_info, root_objectid);
-
- if (!args.root)
- return NULL;
-
- return ilookup5(s, objectid, btrfs_find_actor, (void *)&args);
-}
-
struct inode *btrfs_iget_locked(struct super_block *s, u64 objectid,
struct btrfs_root *root)
{
return inode;
}
+/* Get an inode object given its location and corresponding root.
+ * 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, int *is_new)
+{
+ struct inode *inode;
+
+ inode = btrfs_iget_locked(s, location->objectid, root);
+ if (!inode)
+ return ERR_PTR(-EACCES);
+
+ if (inode->i_state & I_NEW) {
+ BTRFS_I(inode)->root = root;
+ memcpy(&BTRFS_I(inode)->location, location, sizeof(*location));
+ btrfs_read_locked_inode(inode);
+ unlock_new_inode(inode);
+ if (is_new)
+ *is_new = 1;
+ } else {
+ if (is_new)
+ *is_new = 0;
+ }
+
+ return inode;
+}
+
static struct dentry *btrfs_lookup(struct inode *dir, struct dentry *dentry,
struct nameidata *nd)
{
struct btrfs_root *root = bi->root;
struct btrfs_root *sub_root = root;
struct btrfs_key location;
- int ret;
+ int ret, new, do_orphan = 0;
if (dentry->d_name.len > BTRFS_NAME_LEN)
return ERR_PTR(-ENAMETOOLONG);
return ERR_PTR(ret);
if (ret > 0)
return ERR_PTR(-ENOENT);
- inode = btrfs_iget_locked(dir->i_sb, location.objectid,
- sub_root);
- if (!inode)
- return ERR_PTR(-EACCES);
- if (inode->i_state & I_NEW) {
- /* the inode and parent dir are two different roots */
- if (sub_root != root) {
- igrab(inode);
- sub_root->inode = inode;
- }
- BTRFS_I(inode)->root = sub_root;
- memcpy(&BTRFS_I(inode)->location, &location,
- sizeof(location));
- btrfs_read_locked_inode(inode);
- unlock_new_inode(inode);
+ inode = btrfs_iget(dir->i_sb, &location, sub_root, &new);
+ if (IS_ERR(inode))
+ return ERR_CAST(inode);
+
+ /* the inode and parent dir are two different roots */
+ if (new && root != sub_root) {
+ igrab(inode);
+ sub_root->inode = inode;
+ do_orphan = 1;
}
}
+
+ if (unlikely(do_orphan))
+ btrfs_orphan_cleanup(sub_root);
+
return d_splice_alias(inode, dentry);
}
DT_UNKNOWN, DT_REG, DT_DIR, DT_CHR, DT_BLK, DT_FIFO, DT_SOCK, DT_LNK
};
-static int btrfs_readdir(struct file *filp, void *dirent, filldir_t filldir)
+static int btrfs_real_readdir(struct file *filp, void *dirent,
+ filldir_t filldir)
{
struct inode *inode = filp->f_dentry->d_inode;
struct btrfs_root *root = BTRFS_I(inode)->root;
if (filp->f_pos == 0) {
over = filldir(dirent, ".", 1,
1, inode->i_ino,
- DT_DIR);
- if (over)
- return 0;
- filp->f_pos = 1;
- }
-
- key.objectid = inode->i_ino;
- path = btrfs_alloc_path();
- path->reada = 2;
-
- /* special case for .., just use the back ref */
- if (filp->f_pos == 1) {
- btrfs_set_key_type(&key, BTRFS_INODE_REF_KEY);
- key.offset = 0;
- ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
- BUG_ON(ret == 0);
- leaf = path->nodes[0];
- slot = path->slots[0];
- nritems = btrfs_header_nritems(leaf);
- if (slot >= nritems) {
- btrfs_release_path(root, path);
- goto read_dir_items;
- }
- btrfs_item_key_to_cpu(leaf, &found_key, slot);
- btrfs_release_path(root, path);
- if (found_key.objectid != key.objectid ||
- found_key.type != BTRFS_INODE_REF_KEY)
- goto read_dir_items;
+ DT_DIR);
+ if (over)
+ return 0;
+ filp->f_pos = 1;
+ }
+ /* special case for .., just use the back ref */
+ if (filp->f_pos == 1) {
+ u64 pino = parent_ino(filp->f_path.dentry);
over = filldir(dirent, "..", 2,
- 2, found_key.offset, DT_DIR);
+ 2, pino, DT_DIR);
if (over)
- goto nopos;
+ return 0;
filp->f_pos = 2;
}
-read_dir_items:
+ path = btrfs_alloc_path();
+ path->reada = 2;
+
btrfs_set_key_type(&key, key_type);
key.offset = filp->f_pos;
+ key.objectid = inode->i_ino;
ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
if (ret < 0)
goto err;
advance = 0;
- while(1) {
+
+ while (1) {
leaf = path->nodes[0];
nritems = btrfs_header_nritems(leaf);
slot = path->slots[0];
if (advance || slot >= nritems) {
- if (slot >= nritems -1) {
+ if (slot >= nritems - 1) {
ret = btrfs_next_leaf(root, path);
if (ret)
break;
continue;
filp->f_pos = found_key.offset;
- advance = 1;
+
di = btrfs_item_ptr(leaf, slot, struct btrfs_dir_item);
di_cur = 0;
di_total = btrfs_item_size(leaf, item);
- while(di_cur < di_total) {
+
+ while (di_cur < di_total) {
struct btrfs_key location;
name_len = btrfs_dir_name_len(leaf, di);
- if (name_len < 32) {
+ if (name_len <= sizeof(tmp_name)) {
name_ptr = tmp_name;
} else {
name_ptr = kmalloc(name_len, GFP_NOFS);
- BUG_ON(!name_ptr);
+ if (!name_ptr) {
+ ret = -ENOMEM;
+ goto err;
+ }
}
read_extent_buffer(leaf, name_ptr,
(unsigned long)(di + 1), name_len);
d_type = btrfs_filetype_table[btrfs_dir_type(leaf, di)];
btrfs_dir_item_key_to_cpu(leaf, di, &location);
over = filldir(dirent, name_ptr, name_len,
- found_key.offset,
- location.objectid,
+ found_key.offset, location.objectid,
d_type);
if (name_ptr != tmp_name)
if (over)
goto nopos;
+
di_len = btrfs_dir_name_len(leaf, di) +
- btrfs_dir_data_len(leaf, di) +sizeof(*di);
+ btrfs_dir_data_len(leaf, di) + sizeof(*di);
di_cur += di_len;
di = (struct btrfs_dir_item *)((char *)di + di_len);
}
}
+
+ /* Reached end of directory/root. Bump pos past the last item. */
if (key_type == BTRFS_DIR_INDEX_KEY)
filp->f_pos = INT_LIMIT(typeof(filp->f_pos));
else
return ret;
}
+/* Kernels earlier than 2.6.28 still have the NFS deadlock where nfsd
+ will call the file system's ->lookup() method from within its
+ filldir callback, which in turn was called from the file system's
+ ->readdir() method. And will deadlock for many file systems. */
+#if LINUX_VERSION_CODE < KERNEL_VERSION(2,6,28)
+
+struct nfshack_dirent {
+ u64 ino;
+ loff_t offset;
+ int namlen;
+ unsigned int d_type;
+ char name[];
+};
+
+struct nfshack_readdir {
+ char *dirent;
+ size_t used;
+ int full;
+};
+
+
+
+static int btrfs_nfshack_filldir(void *__buf, const char *name, int namlen,
+ loff_t offset, u64 ino, unsigned int d_type)
+{
+ struct nfshack_readdir *buf = __buf;
+ struct nfshack_dirent *de = (void *)(buf->dirent + buf->used);
+ unsigned int reclen;
+
+ reclen = ALIGN(sizeof(struct nfshack_dirent) + namlen, sizeof(u64));
+ if (buf->used + reclen > PAGE_SIZE) {
+ buf->full = 1;
+ return -EINVAL;
+ }
+
+ de->namlen = namlen;
+ de->offset = offset;
+ de->ino = ino;
+ de->d_type = d_type;
+ memcpy(de->name, name, namlen);
+ buf->used += reclen;
+
+ return 0;
+}
+
+static int btrfs_nfshack_readdir(struct file *file, void *dirent,
+ filldir_t filldir)
+{
+ struct nfshack_readdir buf;
+ struct nfshack_dirent *de;
+ int err;
+ int size;
+ loff_t offset;
+
+ buf.dirent = (void *)__get_free_page(GFP_KERNEL);
+ if (!buf.dirent)
+ return -ENOMEM;
+
+ offset = file->f_pos;
+
+ do {
+ unsigned int reclen;
+
+ buf.used = 0;
+ buf.full = 0;
+ err = btrfs_real_readdir(file, &buf, btrfs_nfshack_filldir);
+ if (err)
+ break;
+
+ size = buf.used;
+
+ if (!size)
+ break;
+
+ de = (struct nfshack_dirent *)buf.dirent;
+ while (size > 0) {
+ offset = de->offset;
+
+ if (filldir(dirent, de->name, de->namlen, de->offset,
+ de->ino, de->d_type))
+ goto done;
+ offset = file->f_pos;
+
+ reclen = ALIGN(sizeof(*de) + de->namlen,
+ sizeof(u64));
+ size -= reclen;
+ de = (struct nfshack_dirent *)((char *)de + reclen);
+ }
+ } while (buf.full);
+
+ done:
+ free_page((unsigned long)buf.dirent);
+ file->f_pos = offset;
+
+ return err;
+}
+#endif
+
int btrfs_write_inode(struct inode *inode, int wait)
{
struct btrfs_root *root = BTRFS_I(inode)->root;
struct btrfs_trans_handle *trans;
int ret = 0;
+ if (root->fs_info->closing > 1)
+ return 0;
+
if (wait) {
- trans = btrfs_start_transaction(root, 1);
+ 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;
- trans = btrfs_start_transaction(root, 1);
+ trans = btrfs_join_transaction(root, 1);
btrfs_set_trans_block_group(trans, inode);
btrfs_update_inode(trans, root, inode);
btrfs_end_transaction(trans, root);
}
+static int btrfs_set_inode_index_count(struct inode *inode)
+{
+ struct btrfs_root *root = BTRFS_I(inode)->root;
+ struct btrfs_key key, found_key;
+ struct btrfs_path *path;
+ struct extent_buffer *leaf;
+ int ret;
+
+ key.objectid = inode->i_ino;
+ btrfs_set_key_type(&key, BTRFS_DIR_INDEX_KEY);
+ key.offset = (u64)-1;
+
+ path = btrfs_alloc_path();
+ if (!path)
+ return -ENOMEM;
+
+ ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
+ if (ret < 0)
+ goto out;
+ /* FIXME: we should be able to handle this */
+ if (ret == 0)
+ goto out;
+ ret = 0;
+
+ /*
+ * MAGIC NUMBER EXPLANATION:
+ * since we search a directory based on f_pos we have to start at 2
+ * since '.' and '..' have f_pos of 0 and 1 respectively, so everybody
+ * else has to start at 2
+ */
+ if (path->slots[0] == 0) {
+ BTRFS_I(inode)->index_cnt = 2;
+ goto out;
+ }
+
+ path->slots[0]--;
+
+ leaf = path->nodes[0];
+ btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]);
+
+ if (found_key.objectid != inode->i_ino ||
+ btrfs_key_type(&found_key) != BTRFS_DIR_INDEX_KEY) {
+ BTRFS_I(inode)->index_cnt = 2;
+ goto out;
+ }
+
+ BTRFS_I(inode)->index_cnt = found_key.offset + 1;
+out:
+ btrfs_free_path(path);
+ return ret;
+}
+
+static int btrfs_set_inode_index(struct inode *dir, struct inode *inode,
+ u64 *index)
+{
+ int ret = 0;
+
+ if (BTRFS_I(dir)->index_cnt == (u64)-1) {
+ ret = btrfs_set_inode_index_count(dir);
+ if (ret) {
+ return ret;
+ }
+ }
+
+ *index = BTRFS_I(dir)->index_cnt;
+ BTRFS_I(dir)->index_cnt++;
+
+ return ret;
+}
+
static struct inode *btrfs_new_inode(struct btrfs_trans_handle *trans,
struct btrfs_root *root,
+ struct inode *dir,
const char *name, int name_len,
u64 ref_objectid,
u64 objectid,
struct btrfs_block_group_cache *group,
- int mode)
+ int mode, u64 *index)
{
struct inode *inode;
struct btrfs_inode_item *inode_item;
if (!inode)
return ERR_PTR(-ENOMEM);
- 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);
- mutex_init(&BTRFS_I(inode)->csum_mutex);
- atomic_set(&BTRFS_I(inode)->ordered_writeback, 0);
- BTRFS_I(inode)->delalloc_bytes = 0;
+ if (dir) {
+ ret = btrfs_set_inode_index(dir, inode, index);
+ if (ret)
+ return ERR_PTR(ret);
+ }
+ /*
+ * index_cnt is ignored for everything but a dir,
+ * 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 (mode & S_IFDIR)
owner = 0;
new_inode_group = group;
}
BTRFS_I(inode)->block_group = new_inode_group;
- BTRFS_I(inode)->flags = 0;
key[0].objectid = objectid;
btrfs_set_key_type(&key[0], BTRFS_INODE_ITEM_KEY);
inode->i_mtime = inode->i_atime = inode->i_ctime = CURRENT_TIME;
inode_item = btrfs_item_ptr(path->nodes[0], path->slots[0],
struct btrfs_inode_item);
- fill_inode_item(path->nodes[0], inode_item, inode);
+ fill_inode_item(trans, path->nodes[0], inode_item, inode);
ref = btrfs_item_ptr(path->nodes[0], path->slots[0] + 1,
struct btrfs_inode_ref);
btrfs_set_inode_ref_name_len(path->nodes[0], ref, name_len);
+ btrfs_set_inode_ref_index(path->nodes[0], ref, *index);
ptr = (unsigned long)(ref + 1);
write_extent_buffer(path->nodes[0], name, ptr, name_len);
insert_inode_hash(inode);
return inode;
fail:
+ if (dir)
+ BTRFS_I(dir)->index_cnt--;
btrfs_free_path(path);
return ERR_PTR(ret);
}
return btrfs_type_by_mode[(inode->i_mode & S_IFMT) >> S_SHIFT];
}
-static int btrfs_add_link(struct btrfs_trans_handle *trans,
- struct dentry *dentry, struct inode *inode,
- int add_backref)
+int btrfs_add_link(struct btrfs_trans_handle *trans,
+ struct inode *parent_inode, struct inode *inode,
+ const char *name, int name_len, int add_backref, u64 index)
{
int ret;
struct btrfs_key key;
- struct btrfs_root *root = BTRFS_I(dentry->d_parent->d_inode)->root;
- struct inode *parent_inode;
+ struct btrfs_root *root = BTRFS_I(parent_inode)->root;
key.objectid = inode->i_ino;
btrfs_set_key_type(&key, BTRFS_INODE_ITEM_KEY);
key.offset = 0;
- ret = btrfs_insert_dir_item(trans, root,
- dentry->d_name.name, dentry->d_name.len,
- dentry->d_parent->d_inode->i_ino,
- &key, btrfs_inode_type(inode));
+ ret = btrfs_insert_dir_item(trans, root, name, name_len,
+ parent_inode->i_ino,
+ &key, btrfs_inode_type(inode),
+ index);
if (ret == 0) {
if (add_backref) {
ret = btrfs_insert_inode_ref(trans, root,
- dentry->d_name.name,
- dentry->d_name.len,
- inode->i_ino,
- dentry->d_parent->d_inode->i_ino);
+ name, name_len,
+ inode->i_ino,
+ parent_inode->i_ino,
+ index);
}
- parent_inode = dentry->d_parent->d_inode;
- parent_inode->i_size += dentry->d_name.len * 2;
+ btrfs_i_size_write(parent_inode, parent_inode->i_size +
+ name_len * 2);
parent_inode->i_mtime = parent_inode->i_ctime = CURRENT_TIME;
- ret = btrfs_update_inode(trans, root,
- dentry->d_parent->d_inode);
+ ret = btrfs_update_inode(trans, root, parent_inode);
}
return ret;
}
static int btrfs_add_nondir(struct btrfs_trans_handle *trans,
struct dentry *dentry, struct inode *inode,
- int backref)
+ int backref, u64 index)
{
- int err = btrfs_add_link(trans, dentry, inode, backref);
+ int err = btrfs_add_link(trans, dentry->d_parent->d_inode,
+ inode, dentry->d_name.name,
+ dentry->d_name.len, backref, index);
if (!err) {
d_instantiate(dentry, inode);
return 0;
int drop_inode = 0;
u64 objectid;
unsigned long nr = 0;
+ u64 index = 0;
if (!new_valid_dev(rdev))
return -EINVAL;
goto out_unlock;
}
- inode = btrfs_new_inode(trans, root, dentry->d_name.name,
+ inode = btrfs_new_inode(trans, root, dir, dentry->d_name.name,
dentry->d_name.len,
dentry->d_parent->d_inode->i_ino, objectid,
- BTRFS_I(dir)->block_group, mode);
+ BTRFS_I(dir)->block_group, mode, &index);
err = PTR_ERR(inode);
if (IS_ERR(inode))
goto out_unlock;
+ err = btrfs_init_acl(inode, dir);
+ if (err) {
+ drop_inode = 1;
+ goto out_unlock;
+ }
+
btrfs_set_trans_block_group(trans, inode);
- err = btrfs_add_nondir(trans, dentry, inode, 0);
+ err = btrfs_add_nondir(trans, dentry, inode, 0, index);
if (err)
drop_inode = 1;
else {
int drop_inode = 0;
unsigned long nr = 0;
u64 objectid;
+ u64 index = 0;
err = btrfs_check_free_space(root, 1, 0);
if (err)
goto out_unlock;
}
- inode = btrfs_new_inode(trans, root, dentry->d_name.name,
+ inode = btrfs_new_inode(trans, root, dir, dentry->d_name.name,
dentry->d_name.len,
dentry->d_parent->d_inode->i_ino,
- objectid, BTRFS_I(dir)->block_group, mode);
+ objectid, BTRFS_I(dir)->block_group, mode,
+ &index);
err = PTR_ERR(inode);
if (IS_ERR(inode))
goto out_unlock;
+ err = btrfs_init_acl(inode, dir);
+ if (err) {
+ drop_inode = 1;
+ goto out_unlock;
+ }
+
btrfs_set_trans_block_group(trans, inode);
- err = btrfs_add_nondir(trans, dentry, inode, 0);
+ err = btrfs_add_nondir(trans, dentry, inode, 0, index);
if (err)
drop_inode = 1;
else {
inode->i_mapping->backing_dev_info = &root->fs_info->bdi;
inode->i_fop = &btrfs_file_operations;
inode->i_op = &btrfs_file_inode_operations;
- 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);
- mutex_init(&BTRFS_I(inode)->csum_mutex);
- BTRFS_I(inode)->delalloc_bytes = 0;
- atomic_set(&BTRFS_I(inode)->ordered_writeback, 0);
BTRFS_I(inode)->io_tree.ops = &btrfs_extent_io_ops;
}
dir->i_sb->s_dirt = 1;
struct btrfs_trans_handle *trans;
struct btrfs_root *root = BTRFS_I(dir)->root;
struct inode *inode = old_dentry->d_inode;
+ u64 index;
unsigned long nr = 0;
int err;
int drop_inode = 0;
if (inode->i_nlink == 0)
return -ENOENT;
-#if LINUX_VERSION_CODE <= KERNEL_VERSION(2,6,18)
- inode->i_nlink++;
-#else
- inc_nlink(inode);
-#endif
+ btrfs_inc_nlink(inode);
err = btrfs_check_free_space(root, 1, 0);
if (err)
goto fail;
+ err = btrfs_set_inode_index(dir, inode, &index);
+ if (err)
+ goto fail;
+
trans = btrfs_start_transaction(root, 1);
btrfs_set_trans_block_group(trans, dir);
atomic_inc(&inode->i_count);
- err = btrfs_add_nondir(trans, dentry, inode, 1);
+
+ err = btrfs_add_nondir(trans, dentry, inode, 1, index);
if (err)
drop_inode = 1;
int err = 0;
int drop_on_err = 0;
u64 objectid = 0;
+ u64 index = 0;
unsigned long nr = 1;
err = btrfs_check_free_space(root, 1, 0);
goto out_unlock;
}
- inode = btrfs_new_inode(trans, root, dentry->d_name.name,
+ inode = btrfs_new_inode(trans, root, dir, dentry->d_name.name,
dentry->d_name.len,
dentry->d_parent->d_inode->i_ino, objectid,
- BTRFS_I(dir)->block_group, S_IFDIR | mode);
+ BTRFS_I(dir)->block_group, S_IFDIR | mode,
+ &index);
if (IS_ERR(inode)) {
err = PTR_ERR(inode);
goto out_fail;
}
drop_on_err = 1;
+
+ err = btrfs_init_acl(inode, dir);
+ if (err)
+ goto out_fail;
+
inode->i_op = &btrfs_dir_inode_operations;
inode->i_fop = &btrfs_dir_file_operations;
btrfs_set_trans_block_group(trans, inode);
- inode->i_size = 0;
+ btrfs_i_size_write(inode, 0);
err = btrfs_update_inode(trans, root, inode);
if (err)
goto out_fail;
- err = btrfs_add_link(trans, dentry, inode, 0);
+ err = btrfs_add_link(trans, dentry->d_parent->d_inode,
+ inode, dentry->d_name.name,
+ dentry->d_name.len, 0, index);
if (err)
goto out_fail;
static int merge_extent_mapping(struct extent_map_tree *em_tree,
struct extent_map *existing,
- struct extent_map *em)
+ struct extent_map *em,
+ u64 map_start, u64 map_len)
{
u64 start_diff;
- u64 new_end;
- int ret = 0;
- int real_blocks = existing->block_start < EXTENT_MAP_LAST_BYTE;
-
- if (real_blocks && em->block_start >= EXTENT_MAP_LAST_BYTE)
- goto invalid;
-
- if (!real_blocks && em->block_start != existing->block_start)
- goto invalid;
-
- new_end = max(existing->start + existing->len, em->start + em->len);
-
- if (existing->start >= em->start) {
- if (em->start + em->len < existing->start)
- goto invalid;
-
- start_diff = existing->start - em->start;
- if (real_blocks && em->block_start + start_diff !=
- existing->block_start)
- goto invalid;
-
- em->len = new_end - em->start;
- remove_extent_mapping(em_tree, existing);
- /* free for the tree */
- free_extent_map(existing);
- ret = add_extent_mapping(em_tree, em);
-
- } else if (em->start > existing->start) {
-
- if (existing->start + existing->len < em->start)
- goto invalid;
-
- start_diff = em->start - existing->start;
- if (real_blocks && existing->block_start + start_diff !=
- em->block_start)
- goto invalid;
-
- remove_extent_mapping(em_tree, existing);
- em->block_start = existing->block_start;
- em->start = existing->start;
- em->len = new_end - existing->start;
- free_extent_map(existing);
-
- ret = add_extent_mapping(em_tree, em);
- } else {
- goto invalid;
- }
- return ret;
-
-invalid:
- printk("invalid extent map merge [%Lu %Lu %Lu] [%Lu %Lu %Lu]\n",
- existing->start, existing->len, existing->block_start,
- em->start, em->len, em->block_start);
- return -EIO;
+ BUG_ON(map_start < em->start || map_start >= extent_map_end(em));
+ start_diff = map_start - em->start;
+ em->start = map_start;
+ em->len = map_len;
+ if (em->block_start < EXTENT_MAP_LAST_BYTE)
+ em->block_start += start_diff;
+ return add_extent_mapping(em_tree, em);
}
struct extent_map *btrfs_get_extent(struct inode *inode, struct page *page,
u64 extent_end = 0;
u64 objectid = inode->i_ino;
u32 found_type;
- struct btrfs_path *path;
+ struct btrfs_path *path = NULL;
struct btrfs_root *root = BTRFS_I(inode)->root;
struct btrfs_file_extent_item *item;
struct extent_buffer *leaf;
struct extent_io_tree *io_tree = &BTRFS_I(inode)->io_tree;
struct btrfs_trans_handle *trans = NULL;
- path = btrfs_alloc_path();
- BUG_ON(!path);
-
again:
spin_lock(&em_tree->lock);
em = lookup_extent_mapping(em_tree, start, len);
err = -ENOMEM;
goto out;
}
-
+ em->bdev = root->fs_info->fs_devices->latest_bdev;
em->start = EXTENT_MAP_HOLE;
em->len = (u64)-1;
- em->bdev = root->fs_info->fs_devices->latest_bdev;
+
+ if (!path) {
+ path = btrfs_alloc_path();
+ BUG_ON(!path);
+ }
+
ret = btrfs_lookup_file_extent(trans, root, path,
objectid, start, trans != NULL);
if (ret < 0) {
free_extent_map(em);
em = NULL;
btrfs_release_path(root, path);
- trans = btrfs_start_transaction(root, 1);
+ trans = btrfs_join_transaction(root, 1);
goto again;
}
write_extent_buffer(leaf, map + pg_offset, ptr,
*/
if (ret == -EEXIST) {
struct extent_map *existing;
+
+ ret = 0;
+
existing = lookup_extent_mapping(em_tree, start, len);
if (existing && (existing->start > start ||
existing->start + existing->len <= start)) {
em->len);
if (existing) {
err = merge_extent_mapping(em_tree, existing,
- em);
+ em, start,
+ root->sectorsize);
free_extent_map(existing);
if (err) {
free_extent_map(em);
} else {
free_extent_map(em);
em = existing;
+ err = 0;
}
}
spin_unlock(&em_tree->lock);
out:
- btrfs_free_path(path);
+ if (path)
+ btrfs_free_path(path);
if (trans) {
ret = btrfs_end_transaction(trans, root);
- if (!err)
+ if (!err) {
err = ret;
+ }
}
if (err) {
free_extent_map(em);
return extent_write_full_page(tree, page, btrfs_get_extent, wbc);
}
-static int btrfs_writepages(struct address_space *mapping,
- struct writeback_control *wbc)
+int btrfs_writepages(struct address_space *mapping,
+ struct writeback_control *wbc)
{
struct extent_io_tree *tree;
tree = &BTRFS_I(mapping->host)->io_tree;
return extent_readpages(tree, mapping, pages, nr_pages,
btrfs_get_extent);
}
-
-static int btrfs_releasepage(struct page *page, gfp_t gfp_flags)
+static int __btrfs_releasepage(struct page *page, gfp_t gfp_flags)
{
struct extent_io_tree *tree;
struct extent_map_tree *map;
map = &BTRFS_I(page->mapping->host)->extent_tree;
ret = try_release_extent_mapping(map, tree, page, gfp_flags);
if (ret == 1) {
- invalidate_extent_lru(tree, page_offset(page), PAGE_CACHE_SIZE);
ClearPagePrivate(page);
set_page_private(page, 0);
page_cache_release(page);
return ret;
}
+static int btrfs_releasepage(struct page *page, gfp_t gfp_flags)
+{
+ if (PageWriteback(page) || PageDirty(page))
+ return 0;
+ return __btrfs_releasepage(page, gfp_flags);
+}
+
static void btrfs_invalidatepage(struct page *page, unsigned long offset)
{
struct extent_io_tree *tree;
+ struct btrfs_ordered_extent *ordered;
+ u64 page_start = page_offset(page);
+ u64 page_end = page_start + PAGE_CACHE_SIZE - 1;
+ wait_on_page_writeback(page);
tree = &BTRFS_I(page->mapping->host)->io_tree;
- extent_invalidatepage(tree, page, offset);
- btrfs_releasepage(page, GFP_NOFS);
+ if (offset) {
+ btrfs_releasepage(page, GFP_NOFS);
+ return;
+ }
+
+ lock_extent(tree, page_start, page_end, GFP_NOFS);
+ ordered = btrfs_lookup_ordered_extent(page->mapping->host,
+ page_offset(page));
+ if (ordered) {
+ /*
+ * IO on this page will never be started, so we need
+ * to account for any ordered extents now
+ */
+ clear_extent_bit(tree, page_start, page_end,
+ EXTENT_DIRTY | EXTENT_DELALLOC |
+ EXTENT_LOCKED, 1, 0, GFP_NOFS);
+ btrfs_finish_ordered_io(page->mapping->host,
+ page_start, page_end);
+ btrfs_put_ordered_extent(ordered);
+ lock_extent(tree, page_start, page_end, GFP_NOFS);
+ }
+ clear_extent_bit(tree, page_start, page_end,
+ EXTENT_LOCKED | EXTENT_DIRTY | EXTENT_DELALLOC |
+ EXTENT_ORDERED,
+ 1, 1, GFP_NOFS);
+ __btrfs_releasepage(page, GFP_NOFS);
+
+ ClearPageChecked(page);
if (PagePrivate(page)) {
- invalidate_extent_lru(tree, page_offset(page), PAGE_CACHE_SIZE);
ClearPagePrivate(page);
set_page_private(page, 0);
page_cache_release(page);
{
struct inode *inode = fdentry(vma->vm_file)->d_inode;
struct btrfs_root *root = BTRFS_I(inode)->root;
- unsigned long end;
+ struct extent_io_tree *io_tree = &BTRFS_I(inode)->io_tree;
+ struct btrfs_ordered_extent *ordered;
+ char *kaddr;
+ unsigned long zero_start;
loff_t size;
int ret;
u64 page_start;
+ u64 page_end;
ret = btrfs_check_free_space(root, PAGE_CACHE_SIZE, 0);
if (ret)
goto out;
ret = -EINVAL;
-
+again:
lock_page(page);
- wait_on_page_writeback(page);
size = i_size_read(inode);
- page_start = (u64)page->index << PAGE_CACHE_SHIFT;
+ page_start = page_offset(page);
+ page_end = page_start + PAGE_CACHE_SIZE - 1;
if ((page->mapping != inode->i_mapping) ||
- (page_start > size)) {
+ (page_start >= 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);
+ set_page_extent_mapped(page);
+
+ /*
+ * we can't set the delalloc bits if there are pending ordered
+ * extents. Drop our locks and wait for them to finish
+ */
+ ordered = btrfs_lookup_ordered_extent(inode, page_start);
+ if (ordered) {
+ unlock_extent(io_tree, page_start, page_end, GFP_NOFS);
+ unlock_page(page);
+ btrfs_start_ordered_extent(inode, ordered, 1);
+ btrfs_put_ordered_extent(ordered);
+ goto again;
+ }
+
+ btrfs_set_extent_delalloc(inode, page_start, page_end);
+ ret = 0;
/* page is wholly or partially inside EOF */
if (page_start + PAGE_CACHE_SIZE > size)
- end = size & ~PAGE_CACHE_MASK;
+ zero_start = size & ~PAGE_CACHE_MASK;
else
- end = PAGE_CACHE_SIZE;
+ zero_start = PAGE_CACHE_SIZE;
- ret = btrfs_cow_one_page(inode, page, end);
+ if (zero_start != PAGE_CACHE_SIZE) {
+ kaddr = kmap(page);
+ memset(kaddr + zero_start, 0, PAGE_CACHE_SIZE - zero_start);
+ flush_dcache_page(page);
+ kunmap(page);
+ }
+ ClearPageChecked(page);
+ set_page_dirty(page);
+ unlock_extent(io_tree, page_start, page_end, GFP_NOFS);
out_unlock:
unlock_page(page);
int ret;
struct btrfs_trans_handle *trans;
unsigned long nr;
+ u64 mask = root->sectorsize - 1;
if (!S_ISREG(inode->i_mode))
return;
return;
btrfs_truncate_page(inode->i_mapping, inode->i_size);
+ btrfs_wait_ordered_range(inode, inode->i_size & (~mask), (u64)-1);
trans = btrfs_start_transaction(root, 1);
btrfs_set_trans_block_group(trans, inode);
+ btrfs_i_size_write(inode, inode->i_size);
+ ret = btrfs_orphan_add(trans, inode);
+ if (ret)
+ goto out;
/* FIXME, add redo link to tree so we don't leak on crash */
- ret = btrfs_truncate_in_trans(trans, root, inode,
+ ret = btrfs_truncate_inode_items(trans, root, inode, inode->i_size,
BTRFS_EXTENT_DATA_KEY);
btrfs_update_inode(trans, root, inode);
- nr = trans->blocks_used;
+ ret = btrfs_orphan_del(trans, inode);
+ BUG_ON(ret);
+
+out:
+ nr = trans->blocks_used;
ret = btrfs_end_transaction_throttle(trans, root);
BUG_ON(ret);
btrfs_btree_balance_dirty(root, nr);
struct btrfs_block_group_cache *block_group)
{
struct inode *inode;
- int ret;
+ u64 index = 0;
- inode = btrfs_new_inode(trans, new_root, "..", 2, new_dirid,
- new_dirid, block_group, S_IFDIR | 0700);
+ inode = btrfs_new_inode(trans, new_root, NULL, "..", 2, new_dirid,
+ new_dirid, block_group, S_IFDIR | 0700, &index);
if (IS_ERR(inode))
return PTR_ERR(inode);
inode->i_op = &btrfs_dir_inode_operations;
inode->i_fop = &btrfs_dir_file_operations;
new_root->inode = inode;
- ret = btrfs_insert_inode_ref(trans, new_root, "..", 2, new_dirid,
- new_dirid);
inode->i_nlink = 1;
- inode->i_size = 0;
+ btrfs_i_size_write(inode, 0);
return btrfs_update_inode(trans, new_root, inode);
}
if (!ei)
return NULL;
ei->last_trans = 0;
- ei->ordered_trans = 0;
+ ei->logged_trans = 0;
+ btrfs_ordered_inode_tree_init(&ei->ordered_tree);
+ ei->i_acl = BTRFS_ACL_NOT_CACHED;
+ ei->i_default_acl = BTRFS_ACL_NOT_CACHED;
+ INIT_LIST_HEAD(&ei->i_orphan);
return &ei->vfs_inode;
}
void btrfs_destroy_inode(struct inode *inode)
{
+ struct btrfs_ordered_extent *ordered;
WARN_ON(!list_empty(&inode->i_dentry));
WARN_ON(inode->i_data.nrpages);
+ if (BTRFS_I(inode)->i_acl &&
+ BTRFS_I(inode)->i_acl != BTRFS_ACL_NOT_CACHED)
+ posix_acl_release(BTRFS_I(inode)->i_acl);
+ if (BTRFS_I(inode)->i_default_acl &&
+ BTRFS_I(inode)->i_default_acl != BTRFS_ACL_NOT_CACHED)
+ posix_acl_release(BTRFS_I(inode)->i_default_acl);
+
+ spin_lock(&BTRFS_I(inode)->root->list_lock);
+ if (!list_empty(&BTRFS_I(inode)->i_orphan)) {
+ printk(KERN_ERR "BTRFS: inode %lu: inode still on the orphan"
+ " list\n", inode->i_ino);
+ dump_stack();
+ }
+ spin_unlock(&BTRFS_I(inode)->root->list_lock);
+
+ while(1) {
+ ordered = btrfs_lookup_first_ordered_extent(inode, (u64)-1);
+ if (!ordered)
+ break;
+ else {
+ printk("found ordered extent %Lu %Lu\n",
+ ordered->file_offset, ordered->len);
+ btrfs_remove_ordered_extent(inode, ordered);
+ btrfs_put_ordered_extent(ordered);
+ btrfs_put_ordered_extent(ordered);
+ }
+ }
btrfs_drop_extent_cache(inode, 0, (u64)-1);
kmem_cache_free(btrfs_inode_cachep, BTRFS_I(inode));
}
-#if LINUX_VERSION_CODE > KERNEL_VERSION(2,6,23)
+#if LINUX_VERSION_CODE > KERNEL_VERSION(2,6,26)
+static void init_once(void *foo)
+#elif LINUX_VERSION_CODE > KERNEL_VERSION(2,6,23)
static void init_once(struct kmem_cache * cachep, void *foo)
#else
static void init_once(void * foo, struct kmem_cache * cachep,
struct kmem_cache *btrfs_cache_create(const char *name, size_t size,
unsigned long extra_flags,
-#if LINUX_VERSION_CODE > KERNEL_VERSION(2,6,23)
+#if LINUX_VERSION_CODE > KERNEL_VERSION(2,6,26)
+ void (*ctor)(void *)
+#elif LINUX_VERSION_CODE > KERNEL_VERSION(2,6,23)
void (*ctor)(struct kmem_cache *, void *)
#else
void (*ctor)(void *, struct kmem_cache *,
struct inode *new_inode = new_dentry->d_inode;
struct inode *old_inode = old_dentry->d_inode;
struct timespec ctime = CURRENT_TIME;
+ u64 index = 0;
int ret;
if (S_ISDIR(old_inode->i_mode) && new_inode &&
btrfs_set_trans_block_group(trans, new_dir);
- old_dentry->d_inode->i_nlink++;
+ btrfs_inc_nlink(old_dentry->d_inode);
old_dir->i_ctime = old_dir->i_mtime = ctime;
new_dir->i_ctime = new_dir->i_mtime = ctime;
old_inode->i_ctime = ctime;
- ret = btrfs_unlink_trans(trans, root, old_dir, old_dentry);
+ ret = btrfs_unlink_inode(trans, root, old_dir, old_dentry->d_inode,
+ old_dentry->d_name.name,
+ old_dentry->d_name.len);
if (ret)
goto out_fail;
if (new_inode) {
new_inode->i_ctime = CURRENT_TIME;
- ret = btrfs_unlink_trans(trans, root, new_dir, new_dentry);
+ ret = btrfs_unlink_inode(trans, root, new_dir,
+ new_dentry->d_inode,
+ new_dentry->d_name.name,
+ new_dentry->d_name.len);
if (ret)
goto out_fail;
+ if (new_inode->i_nlink == 0) {
+ ret = btrfs_orphan_add(trans, new_dentry->d_inode);
+ if (ret)
+ goto out_fail;
+ }
+
}
- ret = btrfs_add_link(trans, new_dentry, old_inode, 1);
+ ret = btrfs_set_inode_index(new_dir, old_inode, &index);
+ if (ret)
+ goto out_fail;
+
+ ret = btrfs_add_link(trans, new_dentry->d_parent->d_inode,
+ old_inode, new_dentry->d_name.name,
+ new_dentry->d_name.len, 1, index);
if (ret)
goto out_fail;
out_fail:
- btrfs_end_transaction(trans, root);
+ btrfs_end_transaction_throttle(trans, root);
out_unlock:
return ret;
}
+int btrfs_start_delalloc_inodes(struct btrfs_root *root)
+{
+ struct list_head *head = &root->fs_info->delalloc_inodes;
+ struct btrfs_inode *binode;
+ unsigned long flags;
+
+ spin_lock_irqsave(&root->fs_info->delalloc_lock, flags);
+ while(!list_empty(head)) {
+ binode = list_entry(head->next, struct btrfs_inode,
+ delalloc_inodes);
+ atomic_inc(&binode->vfs_inode.i_count);
+ spin_unlock_irqrestore(&root->fs_info->delalloc_lock, flags);
+ filemap_write_and_wait(binode->vfs_inode.i_mapping);
+ iput(&binode->vfs_inode);
+ spin_lock_irqsave(&root->fs_info->delalloc_lock, flags);
+ }
+ spin_unlock_irqrestore(&root->fs_info->delalloc_lock, flags);
+ return 0;
+}
+
static int btrfs_symlink(struct inode *dir, struct dentry *dentry,
const char *symname)
{
int err;
int drop_inode = 0;
u64 objectid;
+ u64 index = 0 ;
int name_len;
int datasize;
unsigned long ptr;
goto out_unlock;
}
- inode = btrfs_new_inode(trans, root, dentry->d_name.name,
+ inode = btrfs_new_inode(trans, root, dir, dentry->d_name.name,
dentry->d_name.len,
dentry->d_parent->d_inode->i_ino, objectid,
- BTRFS_I(dir)->block_group, S_IFLNK|S_IRWXUGO);
+ BTRFS_I(dir)->block_group, S_IFLNK|S_IRWXUGO,
+ &index);
err = PTR_ERR(inode);
if (IS_ERR(inode))
goto out_unlock;
+ err = btrfs_init_acl(inode, dir);
+ if (err) {
+ drop_inode = 1;
+ goto out_unlock;
+ }
+
btrfs_set_trans_block_group(trans, inode);
- err = btrfs_add_nondir(trans, dentry, inode, 0);
+ err = btrfs_add_nondir(trans, dentry, inode, 0, index);
if (err)
drop_inode = 1;
else {
inode->i_mapping->backing_dev_info = &root->fs_info->bdi;
inode->i_fop = &btrfs_file_operations;
inode->i_op = &btrfs_file_inode_operations;
- 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);
- mutex_init(&BTRFS_I(inode)->csum_mutex);
- BTRFS_I(inode)->delalloc_bytes = 0;
- atomic_set(&BTRFS_I(inode)->ordered_writeback, 0);
BTRFS_I(inode)->io_tree.ops = &btrfs_extent_io_ops;
}
dir->i_sb->s_dirt = 1;
inode->i_op = &btrfs_symlink_inode_operations;
inode->i_mapping->a_ops = &btrfs_symlink_aops;
inode->i_mapping->backing_dev_info = &root->fs_info->bdi;
- inode->i_size = name_len - 1;
+ btrfs_i_size_write(inode, name_len - 1);
err = btrfs_update_inode(trans, root, inode);
if (err)
drop_inode = 1;
return err;
}
+static int btrfs_set_page_dirty(struct page *page)
+{
+ return __set_page_dirty_nobuffers(page);
+}
+
+#if LINUX_VERSION_CODE > KERNEL_VERSION(2,6,26)
+static int btrfs_permission(struct inode *inode, int mask)
+#else
static int btrfs_permission(struct inode *inode, int mask,
struct nameidata *nd)
+#endif
{
if (btrfs_test_flag(inode, READONLY) && (mask & MAY_WRITE))
return -EACCES;
- return generic_permission(inode, mask, NULL);
+ return generic_permission(inode, mask, btrfs_check_acl);
}
static struct inode_operations btrfs_dir_inode_operations = {
.symlink = btrfs_symlink,
.setattr = btrfs_setattr,
.mknod = btrfs_mknod,
- .setxattr = generic_setxattr,
- .getxattr = generic_getxattr,
+ .setxattr = btrfs_setxattr,
+ .getxattr = btrfs_getxattr,
.listxattr = btrfs_listxattr,
- .removexattr = generic_removexattr,
+ .removexattr = btrfs_removexattr,
.permission = btrfs_permission,
};
static struct inode_operations btrfs_dir_ro_inode_operations = {
static struct file_operations btrfs_dir_file_operations = {
.llseek = generic_file_llseek,
.read = generic_read_dir,
- .readdir = btrfs_readdir,
+#if LINUX_VERSION_CODE < KERNEL_VERSION(2,6,28)
+ .readdir = btrfs_nfshack_readdir,
+#else /* NFSd readdir/lookup deadlock is fixed */
+ .readdir = btrfs_real_readdir,
+#endif
.unlocked_ioctl = btrfs_ioctl,
#ifdef CONFIG_COMPAT
.compat_ioctl = btrfs_ioctl,
#endif
.release = btrfs_release_file,
+ .fsync = btrfs_sync_file,
};
static struct extent_io_ops btrfs_extent_io_ops = {
.fill_delalloc = run_delalloc_range,
.submit_bio_hook = btrfs_submit_bio_hook,
.merge_bio_hook = btrfs_merge_bio_hook,
- .readpage_io_hook = btrfs_readpage_io_hook,
.readpage_end_io_hook = btrfs_readpage_end_io_hook,
+ .writepage_end_io_hook = btrfs_writepage_end_io_hook,
+ .writepage_start_hook = btrfs_writepage_start_hook,
.readpage_io_failed_hook = btrfs_io_failed_hook,
.set_bit_hook = btrfs_set_bit_hook,
.clear_bit_hook = btrfs_clear_bit_hook,
.direct_IO = btrfs_direct_IO,
.invalidatepage = btrfs_invalidatepage,
.releasepage = btrfs_releasepage,
- .set_page_dirty = __set_page_dirty_nobuffers,
+ .set_page_dirty = btrfs_set_page_dirty,
};
static struct address_space_operations btrfs_symlink_aops = {
.truncate = btrfs_truncate,
.getattr = btrfs_getattr,
.setattr = btrfs_setattr,
- .setxattr = generic_setxattr,
- .getxattr = generic_getxattr,
+ .setxattr = btrfs_setxattr,
+ .getxattr = btrfs_getxattr,
.listxattr = btrfs_listxattr,
- .removexattr = generic_removexattr,
+ .removexattr = btrfs_removexattr,
.permission = btrfs_permission,
};
static struct inode_operations btrfs_special_inode_operations = {
.getattr = btrfs_getattr,
.setattr = btrfs_setattr,
.permission = btrfs_permission,
+ .setxattr = btrfs_setxattr,
+ .getxattr = btrfs_getxattr,
+ .listxattr = btrfs_listxattr,
+ .removexattr = btrfs_removexattr,
};
static struct inode_operations btrfs_symlink_inode_operations = {
.readlink = generic_readlink,
git.openpandora.org / pandora-kernel.git / blobdiff