[PATCH 1/3] ext4: add i_data_sem sanity check

[PATCH 1/3] ext4: add i_data_sem sanity check

[Dmitry Monakhov]

Each caller of ext4_ext_dirty must hold i_data_sem,
The only exception is migration code, let's make it convenient.

Signed-off-by: Dmitry Monakhov <dmonakhov@openvz.org>
---
 fs/ext4/extents.c |    2 ++
 fs/ext4/migrate.c |    7 +++++++
 2 files changed, 9 insertions(+), 0 deletions(-)

diff --git a/fs/ext4/extents.c b/fs/ext4/extents.c
index b30172d..ee93f82 100644
--- a/fs/ext4/extents.c
+++ b/fs/ext4/extents.c
@@ -161,6 +161,8 @@ int __ext4_ext_dirty(const char *where, unsigned int line, handle_t *handle,
 		     struct inode *inode, struct ext4_ext_path *path)
 {
 	int err;
+
+	WARN_ON(!rwsem_is_locked(&EXT4_I(inode)->i_data_sem));
 	if (path->p_bh) {
 		ext4_extent_block_csum_set(inode, ext_block_hdr(path->p_bh));
 		/* path points to block */
diff --git a/fs/ext4/migrate.c b/fs/ext4/migrate.c
index ec09243..d3567f2 100644
--- a/fs/ext4/migrate.c
+++ b/fs/ext4/migrate.c
@@ -39,6 +39,8 @@ static int finish_range(handle_t *handle, struct inode *inode,
 	newext.ee_block = cpu_to_le32(lb->first_block);
 	newext.ee_len   = cpu_to_le16(lb->last_block - lb->first_block + 1);
 	ext4_ext_store_pblock(&newext, lb->first_pblock);
+	/* Locking only for convinience since we are operating on temp inode */
+	down_write(&EXT4_I(inode)->i_data_sem);
 	path = ext4_ext_find_extent(inode, lb->first_block, NULL, 0);
 
 	if (IS_ERR(path)) {
@@ -61,7 +63,9 @@ static int finish_range(handle_t *handle, struct inode *inode,
 	 */
 	if (needed && ext4_handle_has_enough_credits(handle,
 						EXT4_RESERVE_TRANS_BLOCKS)) {
+		up_write((&EXT4_I(inode)->i_data_sem));
 		retval = ext4_journal_restart(handle, needed);
+		down_write((&EXT4_I(inode)->i_data_sem));
 		if (retval)
 			goto err_out;
 	} else if (needed) {
@@ -70,13 +74,16 @@ static int finish_range(handle_t *handle, struct inode *inode,
 			/*
 			 * IF not able to extend the journal restart the journal
 			 */
+			up_write((&EXT4_I(inode)->i_data_sem));
 			retval = ext4_journal_restart(handle, needed);
+			down_write((&EXT4_I(inode)->i_data_sem));
 			if (retval)
 				goto err_out;
 		}
 	}
 	retval = ext4_ext_insert_extent(handle, inode, path, &newext, 0);
 err_out:
+	up_write((&EXT4_I(inode)->i_data_sem));
 	if (path) {
 		ext4_ext_drop_refs(path);
 		kfree(path);
-- 
1.7.1

[PATCH 2/3] ext4: use correct depth value

[Dmitry Monakhov]

Inode's depth can be changed from here:
ext4_ext_try_to_merge() ->ext4_ext_try_to_merge_up()
We must use correct value.

Signed-off-by: Dmitry Monakhov <dmonakhov@openvz.org>
---
 fs/ext4/extents.c |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)

diff --git a/fs/ext4/extents.c b/fs/ext4/extents.c
index ee93f82..6950bab 100644
--- a/fs/ext4/extents.c
+++ b/fs/ext4/extents.c
@@ -3254,7 +3254,7 @@ out:
 
 fix_extent_len:
 	ex->ee_len = orig_ex.ee_len;
-	ext4_ext_dirty(handle, inode, path + depth);
+	ext4_ext_dirty(handle, inode, path + path->p_depth);
 	return err;
 }
 
-- 
1.7.1

Re: [PATCH 2/3] ext4: use correct depth value

[Jan Kara]

On Mon 21-07-14 19:04:48, Dmitry Monakhov wrote:
> Inode's depth can be changed from here:
> ext4_ext_try_to_merge() ->ext4_ext_try_to_merge_up()
> We must use correct value.
  Looks good. You can add:
Reviewed-by: Jan Kara <jack@suse.cz>

								Honza

> 
> Signed-off-by: Dmitry Monakhov <dmonakhov@openvz.org>
> ---
>  fs/ext4/extents.c |    2 +-
>  1 files changed, 1 insertions(+), 1 deletions(-)
> 
> diff --git a/fs/ext4/extents.c b/fs/ext4/extents.c
> index ee93f82..6950bab 100644
> --- a/fs/ext4/extents.c
> +++ b/fs/ext4/extents.c
> @@ -3254,7 +3254,7 @@ out:
>  
>  fix_extent_len:
>  	ex->ee_len = orig_ex.ee_len;
> -	ext4_ext_dirty(handle, inode, path + depth);
> +	ext4_ext_dirty(handle, inode, path + path->p_depth);
>  	return err;
>  }
>  
> -- 
> 1.7.1
> 
> --
> To unsubscribe from this list: send the line "unsubscribe linux-ext4" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
-- 
Jan Kara <jack@suse.cz>
SUSE Labs, CR

Re: [PATCH 2/3] ext4: use correct depth value

[Theodore Ts'o]

On Mon, Jul 21, 2014 at 09:48:37PM +0200, Jan Kara wrote:
> On Mon 21-07-14 19:04:48, Dmitry Monakhov wrote:
> > Inode's depth can be changed from here:
> > ext4_ext_try_to_merge() ->ext4_ext_try_to_merge_up()
> > We must use correct value.
>   Looks good. You can add:
> Reviewed-by: Jan Kara <jack@suse.cz>

Thanks, applied.

						- Ted

[PATCH 3/3] ext4: fix incorrect locking in move_extent_per_page

[Dmitry Monakhov]

If we have to copy data we must drop i_data_sem because of
get_blocks() will be called inside mext_page_mkuptodate(), but later we must
reacquire it again because we are about to change extent's tree

Signed-off-by: Dmitry Monakhov <dmonakhov@openvz.org>
---
 fs/ext4/move_extent.c |    3 ++-
 1 files changed, 2 insertions(+), 1 deletions(-)

diff --git a/fs/ext4/move_extent.c b/fs/ext4/move_extent.c
index 2484c7e..671a74b 100644
--- a/fs/ext4/move_extent.c
+++ b/fs/ext4/move_extent.c
@@ -1013,10 +1013,11 @@ data_copy:
 		*err = -EBUSY;
 		goto unlock_pages;
 	}

Re: [PATCH 3/3] ext4: fix incorrect locking in move_extent_per_page

[Jan Kara]

On Mon 21-07-14 19:04:49, Dmitry Monakhov wrote:
> If we have to copy data we must drop i_data_sem because of
> get_blocks() will be called inside mext_page_mkuptodate(), but later we must
> reacquire it again because we are about to change extent's tree
> 
> Signed-off-by: Dmitry Monakhov <dmonakhov@openvz.org>
  Looks good. You can add:
Reviewed-by: Jan Kara <jack@suse.cz>

								Honza

> ---
>  fs/ext4/move_extent.c |    3 ++-
>  1 files changed, 2 insertions(+), 1 deletions(-)
> 
> diff --git a/fs/ext4/move_extent.c b/fs/ext4/move_extent.c
> index 2484c7e..671a74b 100644
> --- a/fs/ext4/move_extent.c
> +++ b/fs/ext4/move_extent.c
> @@ -1013,10 +1013,11 @@ data_copy:
>  		*err = -EBUSY;
>  		goto unlock_pages;
>  	}
> -
> +	ext4_double_down_write_data_sem(orig_inode, donor_inode);
>  	replaced_count = mext_replace_branches(handle, orig_inode, donor_inode,
>  					       orig_blk_offset,
>  					       block_len_in_page, err);
> +	ext4_double_up_write_data_sem(orig_inode, donor_inode);
>  	if (*err) {
>  		if (replaced_count) {
>  			block_len_in_page = replaced_count;
> -- 
> 1.7.1
> 
> --
> To unsubscribe from this list: send the line "unsubscribe linux-ext4" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
-- 
Jan Kara <jack@suse.cz>
SUSE Labs, CR

Re: [PATCH 3/3] ext4: fix incorrect locking in move_extent_per_page

[Theodore Ts'o]

On Mon, Jul 21, 2014 at 09:51:02PM +0200, Jan Kara wrote:
> On Mon 21-07-14 19:04:49, Dmitry Monakhov wrote:
> > If we have to copy data we must drop i_data_sem because of
> > get_blocks() will be called inside mext_page_mkuptodate(), but later we must
> > reacquire it again because we are about to change extent's tree
> > 
> > Signed-off-by: Dmitry Monakhov <dmonakhov@openvz.org>
>   Looks good. You can add:
> Reviewed-by: Jan Kara <jack@suse.cz>

Thanks, applied.

					- Ted

Re: [PATCH 1/3] ext4: add i_data_sem sanity check

[Jan Kara]

On Mon 21-07-14 19:04:47, Dmitry Monakhov wrote:
> Each caller of ext4_ext_dirty must hold i_data_sem,
> The only exception is migration code, let's make it convenient.
> 
> Signed-off-by: Dmitry Monakhov <dmonakhov@openvz.org>
  Looks good. You can add:
Reviewed-by: Jan Kara <jack@suse.cz>

								Honza

> ---
>  fs/ext4/extents.c |    2 ++
>  fs/ext4/migrate.c |    7 +++++++
>  2 files changed, 9 insertions(+), 0 deletions(-)
> 
> diff --git a/fs/ext4/extents.c b/fs/ext4/extents.c
> index b30172d..ee93f82 100644
> --- a/fs/ext4/extents.c
> +++ b/fs/ext4/extents.c
> @@ -161,6 +161,8 @@ int __ext4_ext_dirty(const char *where, unsigned int line, handle_t *handle,
>  		     struct inode *inode, struct ext4_ext_path *path)
>  {
>  	int err;
> +
> +	WARN_ON(!rwsem_is_locked(&EXT4_I(inode)->i_data_sem));
>  	if (path->p_bh) {
>  		ext4_extent_block_csum_set(inode, ext_block_hdr(path->p_bh));
>  		/* path points to block */
> diff --git a/fs/ext4/migrate.c b/fs/ext4/migrate.c
> index ec09243..d3567f2 100644
> --- a/fs/ext4/migrate.c
> +++ b/fs/ext4/migrate.c
> @@ -39,6 +39,8 @@ static int finish_range(handle_t *handle, struct inode *inode,
>  	newext.ee_block = cpu_to_le32(lb->first_block);
>  	newext.ee_len   = cpu_to_le16(lb->last_block - lb->first_block + 1);
>  	ext4_ext_store_pblock(&newext, lb->first_pblock);
> +	/* Locking only for convinience since we are operating on temp inode */
> +	down_write(&EXT4_I(inode)->i_data_sem);
>  	path = ext4_ext_find_extent(inode, lb->first_block, NULL, 0);
>  
>  	if (IS_ERR(path)) {
> @@ -61,7 +63,9 @@ static int finish_range(handle_t *handle, struct inode *inode,
>  	 */
>  	if (needed && ext4_handle_has_enough_credits(handle,
>  						EXT4_RESERVE_TRANS_BLOCKS)) {
> +		up_write((&EXT4_I(inode)->i_data_sem));
>  		retval = ext4_journal_restart(handle, needed);
> +		down_write((&EXT4_I(inode)->i_data_sem));
>  		if (retval)
>  			goto err_out;
>  	} else if (needed) {
> @@ -70,13 +74,16 @@ static int finish_range(handle_t *handle, struct inode *inode,
>  			/*
>  			 * IF not able to extend the journal restart the journal
>  			 */
> +			up_write((&EXT4_I(inode)->i_data_sem));
>  			retval = ext4_journal_restart(handle, needed);
> +			down_write((&EXT4_I(inode)->i_data_sem));
>  			if (retval)
>  				goto err_out;
>  		}
>  	}
>  	retval = ext4_ext_insert_extent(handle, inode, path, &newext, 0);
>  err_out:
> +	up_write((&EXT4_I(inode)->i_data_sem));
>  	if (path) {
>  		ext4_ext_drop_refs(path);
>  		kfree(path);
> -- 
> 1.7.1
> 
> --
> To unsubscribe from this list: send the line "unsubscribe linux-ext4" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
-- 
Jan Kara <jack@suse.cz>
SUSE Labs, CR

Re: [PATCH 1/3] ext4: add i_data_sem sanity check

[Theodore Ts'o]

On Mon, Jul 21, 2014 at 09:35:10PM +0200, Jan Kara wrote:
> On Mon 21-07-14 19:04:47, Dmitry Monakhov wrote:
> > Each caller of ext4_ext_dirty must hold i_data_sem,
> > The only exception is migration code, let's make it convenient.
> > 
> > Signed-off-by: Dmitry Monakhov <dmonakhov@openvz.org>
>   Looks good. You can add:
> Reviewed-by: Jan Kara <jack@suse.cz>

Thanks, applied.

						- Ted