Btrfs: take i_mutex before generic_write_checks

btrfs_file_write was incorrectly calling generic_write_checks without
taking i_mutex.  This lead to problems with racing around i_size when
doing O_APPEND writes.

The fix here is to move i_mutex higher.

Signed-off-by: Chris Mason <chris.mason@oracle.com>
This commit is contained in:
Chris Mason 2009-10-01 12:29:10 -04:00
parent 35d62a942d
commit ab93dbecfb

View file

@ -920,26 +920,35 @@ static ssize_t btrfs_file_write(struct file *file, const char __user *buf,
start_pos = pos;
vfs_check_frozen(inode->i_sb, SB_FREEZE_WRITE);
current->backing_dev_info = inode->i_mapping->backing_dev_info;
err = generic_write_checks(file, &pos, &count, S_ISBLK(inode->i_mode));
if (err)
goto out_nolock;
if (count == 0)
goto out_nolock;
err = file_remove_suid(file);
if (err)
goto out_nolock;
/* do the reserve before the mutex lock in case we have to do some
* flushing. We wouldn't deadlock, but this is more polite.
*/
err = btrfs_reserve_metadata_for_delalloc(root, inode, 1);
if (err)
goto out_nolock;
mutex_lock(&inode->i_mutex);
current->backing_dev_info = inode->i_mapping->backing_dev_info;
err = generic_write_checks(file, &pos, &count, S_ISBLK(inode->i_mode));
if (err)
goto out;
if (count == 0)
goto out;
err = file_remove_suid(file);
if (err)
goto out;
file_update_time(file);
pages = kmalloc(nrptrs * sizeof(struct page *), GFP_KERNEL);
mutex_lock(&inode->i_mutex);
/* generic_write_checks can change our pos */
start_pos = pos;
BTRFS_I(inode)->sequence++;
first_index = pos >> PAGE_CACHE_SHIFT;
last_index = (pos + count) >> PAGE_CACHE_SHIFT;