ext4: Automatically allocate delay allocated blocks on rename
authorTheodore Ts'o <tytso@mit.edu>
Tue, 24 Feb 2009 04:05:27 +0000 (23:05 -0500)
committerTheodore Ts'o <tytso@mit.edu>
Tue, 24 Feb 2009 04:05:27 +0000 (23:05 -0500)
commit8750c6d5fcbd3342b3d908d157f81d345c5325a7
tree323e83b9f81379c696acc861555143a06a84eb7b
parent7d8f9f7d150dded7b68e61ca6403a1f166fb4edf
ext4: Automatically allocate delay allocated blocks on rename

When renaming a file such that a link to another inode is overwritten,
force any delay allocated blocks that to be allocated so that if the
filesystem is mounted with data=ordered, the data blocks will be
pushed out to disk along with the journal commit.  Many application
programs expect this, so we do this to avoid zero length files if the
system crashes unexpectedly.

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
fs/ext4/namei.c