*/
#include <linux/module.h>
-#include <linux/ext4_jbd2.h>
-#include <linux/ext4_fs_extents.h>
+#include "ext4_jbd2.h"
+#include "ext4_extents.h"
/*
* The contiguous blocks details which can be
if (IS_ERR(path)) {
retval = PTR_ERR(path);
+ path = NULL;
goto err_out;
}
* credit. But below we try to not accumalate too much
* of them by restarting the journal.
*/
- needed = ext4_ext_calc_credits_for_insert(inode, path);
+ needed = ext4_ext_calc_credits_for_single_extent(inode,
+ lb->last_block - lb->first_block + 1, path);
/*
* Make sure the credit we accumalated is not really high
*/
- if (needed && handle->h_buffer_credits >= EXT4_RESERVE_TRANS_BLOCKS) {
+ if (needed && ext4_handle_has_enough_credits(handle,
+ EXT4_RESERVE_TRANS_BLOCKS)) {
retval = ext4_journal_restart(handle, needed);
if (retval)
goto err_out;
- }
- if (needed) {
+ } else if (needed) {
retval = ext4_journal_extend(handle, needed);
- if (retval != 0) {
+ if (retval) {
/*
* IF not able to extend the journal restart the journal
*/
goto err_out;
}
}
- retval = ext4_ext_insert_extent(handle, inode, path, &newext);
+ retval = ext4_ext_insert_extent(handle, inode, path, &newext, 0);
err_out:
+ if (path) {
+ ext4_ext_drop_refs(path);
+ kfree(path);
+ }
lb->first_pblock = 0;
return retval;
}
}
+static int extend_credit_for_blkdel(handle_t *handle, struct inode *inode)
+{
+ int retval = 0, needed;
+
+ if (ext4_handle_has_enough_credits(handle, EXT4_RESERVE_TRANS_BLOCKS+1))
+ return 0;
+ /*
+ * We are freeing a blocks. During this we touch
+ * superblock, group descriptor and block bitmap.
+ * So allocate a credit of 3. We may update
+ * quota (user and group).
+ */
+ needed = 3 + EXT4_MAXQUOTAS_TRANS_BLOCKS(inode->i_sb);
+
+ if (ext4_journal_extend(handle, needed) != 0)
+ retval = ext4_journal_restart(handle, needed);
+
+ return retval;
+}
+
static int free_dind_blocks(handle_t *handle,
struct inode *inode, __le32 i_data)
{
tmp_idata = (__le32 *)bh->b_data;
for (i = 0; i < max_entries; i++) {
- if (tmp_idata[i])
- ext4_free_blocks(handle, inode,
- le32_to_cpu(tmp_idata[i]), 1, 1);
+ if (tmp_idata[i]) {
+ extend_credit_for_blkdel(handle, inode);
+ ext4_free_blocks(handle, inode, 0,
+ le32_to_cpu(tmp_idata[i]), 1,
+ EXT4_FREE_BLOCKS_METADATA |
+ EXT4_FREE_BLOCKS_FORGET);
+ }
}
put_bh(bh);
- ext4_free_blocks(handle, inode, le32_to_cpu(i_data), 1, 1);
+ extend_credit_for_blkdel(handle, inode);
+ ext4_free_blocks(handle, inode, 0, le32_to_cpu(i_data), 1,
+ EXT4_FREE_BLOCKS_METADATA |
+ EXT4_FREE_BLOCKS_FORGET);
return 0;
}
}
}
put_bh(bh);
- ext4_free_blocks(handle, inode, le32_to_cpu(i_data), 1, 1);
+ extend_credit_for_blkdel(handle, inode);
+ ext4_free_blocks(handle, inode, 0, le32_to_cpu(i_data), 1,
+ EXT4_FREE_BLOCKS_METADATA |
+ EXT4_FREE_BLOCKS_FORGET);
return 0;
}
-static int free_ind_block(handle_t *handle, struct inode *inode)
+static int free_ind_block(handle_t *handle, struct inode *inode, __le32 *i_data)
{
int retval;
- struct ext4_inode_info *ei = EXT4_I(inode);
- if (ei->i_data[EXT4_IND_BLOCK])
- ext4_free_blocks(handle, inode,
- le32_to_cpu(ei->i_data[EXT4_IND_BLOCK]), 1, 1);
+ /* ei->i_data[EXT4_IND_BLOCK] */
+ if (i_data[0]) {
+ extend_credit_for_blkdel(handle, inode);
+ ext4_free_blocks(handle, inode, 0,
+ le32_to_cpu(i_data[0]), 1,
+ EXT4_FREE_BLOCKS_METADATA |
+ EXT4_FREE_BLOCKS_FORGET);
+ }
- if (ei->i_data[EXT4_DIND_BLOCK]) {
- retval = free_dind_blocks(handle, inode,
- ei->i_data[EXT4_DIND_BLOCK]);
+ /* ei->i_data[EXT4_DIND_BLOCK] */
+ if (i_data[1]) {
+ retval = free_dind_blocks(handle, inode, i_data[1]);
if (retval)
return retval;
}
- if (ei->i_data[EXT4_TIND_BLOCK]) {
- retval = free_tind_blocks(handle, inode,
- ei->i_data[EXT4_TIND_BLOCK]);
+ /* ei->i_data[EXT4_TIND_BLOCK] */
+ if (i_data[2]) {
+ retval = free_tind_blocks(handle, inode, i_data[2]);
if (retval)
return retval;
}
}
static int ext4_ext_swap_inode_data(handle_t *handle, struct inode *inode,
- struct inode *tmp_inode, int retval)
+ struct inode *tmp_inode)
{
+ int retval;
+ __le32 i_data[3];
struct ext4_inode_info *ei = EXT4_I(inode);
struct ext4_inode_info *tmp_ei = EXT4_I(tmp_inode);
- retval = free_ind_block(handle, inode);
- if (retval)
- goto err_out;
-
/*
* One credit accounted for writing the
* i_data field of the original inode
*/
retval = ext4_journal_extend(handle, 1);
- if (retval != 0) {
+ if (retval) {
retval = ext4_journal_restart(handle, 1);
if (retval)
goto err_out;
}
+ i_data[0] = ei->i_data[EXT4_IND_BLOCK];
+ i_data[1] = ei->i_data[EXT4_DIND_BLOCK];
+ i_data[2] = ei->i_data[EXT4_TIND_BLOCK];
+
+ down_write(&EXT4_I(inode)->i_data_sem);
+ /*
+ * if EXT4_STATE_EXT_MIGRATE is cleared a block allocation
+ * happened after we started the migrate. We need to
+ * fail the migrate
+ */
+ if (!(EXT4_I(inode)->i_state & EXT4_STATE_EXT_MIGRATE)) {
+ retval = -EAGAIN;
+ up_write(&EXT4_I(inode)->i_data_sem);
+ goto err_out;
+ } else
+ EXT4_I(inode)->i_state &= ~EXT4_STATE_EXT_MIGRATE;
/*
* We have the extent map build with the tmp inode.
* Now copy the i_data across
spin_lock(&inode->i_lock);
inode->i_blocks += tmp_inode->i_blocks;
spin_unlock(&inode->i_lock);
+ up_write(&EXT4_I(inode)->i_data_sem);
+ /*
+ * We mark the inode dirty after, because we decrement the
+ * i_blocks when freeing the indirect meta-data blocks
+ */
+ retval = free_ind_block(handle, inode, i_data);
ext4_mark_inode_dirty(handle, inode);
+
err_out:
return retval;
}
}
}
put_bh(bh);
- ext4_free_blocks(handle, inode, block, 1, 1);
+ extend_credit_for_blkdel(handle, inode);
+ ext4_free_blocks(handle, inode, 0, block, 1,
+ EXT4_FREE_BLOCKS_METADATA | EXT4_FREE_BLOCKS_FORGET);
return retval;
}
}
-int ext4_ext_migrate(struct inode *inode, struct file *filp,
- unsigned int cmd, unsigned long arg)
+int ext4_ext_migrate(struct inode *inode)
{
handle_t *handle;
int retval = 0, i;
struct inode *tmp_inode = NULL;
struct list_blocks_struct lb;
unsigned long max_entries;
+ __u32 goal;
- if (!test_opt(inode->i_sb, EXTENTS))
- /*
- * if mounted with noextents we don't allow the migrate
- */
- return -EINVAL;
-
- if ((EXT4_I(inode)->i_flags & EXT4_EXTENTS_FL))
+ /*
+ * If the filesystem does not support extents, or the inode
+ * already is extent-based, error out.
+ */
+ if (!EXT4_HAS_INCOMPAT_FEATURE(inode->i_sb,
+ EXT4_FEATURE_INCOMPAT_EXTENTS) ||
+ (EXT4_I(inode)->i_flags & EXT4_EXTENTS_FL))
return -EINVAL;
if (S_ISLNK(inode->i_mode) && inode->i_blocks == 0)
*/
return retval;
- down_write(&EXT4_I(inode)->i_data_sem);
handle = ext4_journal_start(inode,
EXT4_DATA_TRANS_BLOCKS(inode->i_sb) +
EXT4_INDEX_EXTRA_TRANS_BLOCKS + 3 +
- 2 * EXT4_QUOTA_INIT_BLOCKS(inode->i_sb)
+ EXT4_MAXQUOTAS_INIT_BLOCKS(inode->i_sb)
+ 1);
if (IS_ERR(handle)) {
retval = PTR_ERR(handle);
- goto err_out;
+ return retval;
}
- tmp_inode = ext4_new_inode(handle,
- inode->i_sb->s_root->d_inode,
- S_IFREG);
+ goal = (((inode->i_ino - 1) / EXT4_INODES_PER_GROUP(inode->i_sb)) *
+ EXT4_INODES_PER_GROUP(inode->i_sb)) + 1;
+ tmp_inode = ext4_new_inode(handle, inode->i_sb->s_root->d_inode,
+ S_IFREG, 0, goal);
if (IS_ERR(tmp_inode)) {
retval = -ENOMEM;
ext4_journal_stop(handle);
- tmp_inode = NULL;
- goto err_out;
+ return retval;
}
i_size_write(tmp_inode, i_size_read(inode));
/*
ext4_orphan_add(handle, tmp_inode);
ext4_journal_stop(handle);
- ei = EXT4_I(inode);
- i_data = ei->i_data;
- memset(&lb, 0, sizeof(lb));
-
- /* 32 bit block address 4 bytes */
- max_entries = inode->i_sb->s_blocksize >> 2;
-
/*
* start with one credit accounted for
* superblock modification.
* trascation that created the inode. Later as and
* when we add extents we extent the journal
*/
+ /*
+ * Even though we take i_mutex we can still cause block
+ * allocation via mmap write to holes. If we have allocated
+ * new blocks we fail migrate. New block allocation will
+ * clear EXT4_STATE_EXT_MIGRATE flag. The flag is updated
+ * with i_data_sem held to prevent racing with block
+ * allocation.
+ */
+ down_read((&EXT4_I(inode)->i_data_sem));
+ EXT4_I(inode)->i_state |= EXT4_STATE_EXT_MIGRATE;
+ up_read((&EXT4_I(inode)->i_data_sem));
+
handle = ext4_journal_start(inode, 1);
+
+ ei = EXT4_I(inode);
+ i_data = ei->i_data;
+ memset(&lb, 0, sizeof(lb));
+
+ /* 32 bit block address 4 bytes */
+ max_entries = inode->i_sb->s_blocksize >> 2;
for (i = 0; i < EXT4_NDIR_BLOCKS; i++, blk_count++) {
if (i_data[i]) {
retval = update_extent_range(handle, tmp_inode,
*/
retval = finish_range(handle, tmp_inode, &lb);
err_out:
- /*
- * We are either freeing extent information or indirect
- * blocks. During this we touch superblock, group descriptor
- * and block bitmap. Later we mark the tmp_inode dirty
- * via ext4_ext_tree_init. So allocate a credit of 4
- * We may update quota (user and group).
- *
- * FIXME!! we may be touching bitmaps in different block groups.
- */
- if (ext4_journal_extend(handle,
- 4 + 2*EXT4_QUOTA_TRANS_BLOCKS(inode->i_sb)) != 0)
- ext4_journal_restart(handle,
- 4 + 2*EXT4_QUOTA_TRANS_BLOCKS(inode->i_sb));
if (retval)
/*
* Failure case delete the extent information with the
* tmp_inode
*/
free_ext_block(handle, tmp_inode);
- else
- retval = ext4_ext_swap_inode_data(handle, inode,
- tmp_inode, retval);
+ else {
+ retval = ext4_ext_swap_inode_data(handle, inode, tmp_inode);
+ if (retval)
+ /*
+ * if we fail to swap inode data free the extent
+ * details of the tmp inode
+ */
+ free_ext_block(handle, tmp_inode);
+ }
+
+ /* We mark the tmp_inode dirty via ext4_ext_tree_init. */
+ if (ext4_journal_extend(handle, 1) != 0)
+ ext4_journal_restart(handle, 1);
/*
* Mark the tmp_inode as of size zero
tmp_inode->i_nlink = 0;
ext4_journal_stop(handle);
-
- up_write(&EXT4_I(inode)->i_data_sem);
-
- if (tmp_inode)
- iput(tmp_inode);
+ unlock_new_inode(tmp_inode);
+ iput(tmp_inode);
return retval;
}
git://ftp.safe.ca / safe / jmp / linux-2.6 / blobdiff