2009-06-10 12:56:03

by Aneesh Kumar K.V

Subject: [PATCH] ext4: Properly mark the extent uninitialized during truncate

We should make sure we don't mark the non uninit extent as
uninit during truncate

Signed-off-by: Aneesh Kumar K.V <[email protected]>
---
fs/ext4/extents.c | 8 ++++++--
1 files changed, 6 insertions(+), 2 deletions(-)

diff --git a/fs/ext4/extents.c b/fs/ext4/extents.c
index f2a2591..764c394 100644
--- a/fs/ext4/extents.c
+++ b/fs/ext4/extents.c
@@ -2083,12 +2083,16 @@ ext4_ext_rm_leaf(handle_t *handle, struct inode *inode,
ex = EXT_LAST_EXTENT(eh);

ex_ee_block = le32_to_cpu(ex->ee_block);
- if (ext4_ext_is_uninitialized(ex))
- uninitialized = 1;
ex_ee_len = ext4_ext_get_actual_len(ex);

while (ex >= EXT_FIRST_EXTENT(eh) &&
ex_ee_block + ex_ee_len > start) {
+
+ if (ext4_ext_is_uninitialized(ex))
+ uninitialized = 1;
+ else
+ uninitialized = 0;
+
ext_debug("remove ext %lu:%u\n", ex_ee_block, ex_ee_len);
path[depth].p_ext = ex;

--
1.6.3.2.214.gf4f78

2009-06-10 19:43:30

by Aneesh Kumar K.V

Subject: Re: [PATCH] ext4: Properly mark the extent uninitialized during truncate

On Wed, Jun 10, 2009 at 06:25:57PM +0530, Aneesh Kumar K.V wrote:
> We should make sure we don't mark the non uninit extent as
> uninit during truncate
>
> Signed-off-by: Aneesh Kumar K.V <[email protected]>
> ---
> fs/ext4/extents.c | 8 ++++++--
> 1 files changed, 6 insertions(+), 2 deletions(-)
>
> diff --git a/fs/ext4/extents.c b/fs/ext4/extents.c
> index f2a2591..764c394 100644
> --- a/fs/ext4/extents.c
> +++ b/fs/ext4/extents.c
> @@ -2083,12 +2083,16 @@ ext4_ext_rm_leaf(handle_t *handle, struct inode *inode,
> ex = EXT_LAST_EXTENT(eh);
>
> ex_ee_block = le32_to_cpu(ex->ee_block);
> - if (ext4_ext_is_uninitialized(ex))
> - uninitialized = 1;
> ex_ee_len = ext4_ext_get_actual_len(ex);
>
> while (ex >= EXT_FIRST_EXTENT(eh) &&
> ex_ee_block + ex_ee_len > start) {
> +
> + if (ext4_ext_is_uninitialized(ex))
> + uninitialized = 1;
> + else
> + uninitialized = 0;
> +
> ext_debug("remove ext %lu:%u\n", ex_ee_block, ex_ee_len);
> path[depth].p_ext = ex;
>

BTW this fix is really important. It fix a data corruption when we write
to an fallocate space and later truncate the same file. So mostly a
stable candidate.

-aneesh

2009-06-10 21:11:32

by Theodore Ts'o

Subject: Re: [PATCH] ext4: Properly mark the extent uninitialized during truncate

On Wed, Jun 10, 2009 at 06:25:57PM +0530, Aneesh Kumar K.V wrote:
> We should make sure we don't mark the non uninit extent as
> uninit during truncate

Good catch! I've applied it to the ext4 patch queue.

Thanks,

- Ted