Ocfs2 devel - Feb 2023 - [PATCH] ocfs2: fix non-auto defrag path not working issue

This commit fixes three issues on non-auto defrag path (defragfs.ocfs2
doesn't set OCFS2_MOVE_EXT_FL_AUTO_DEFRAG on range.me_flags):
- For ocfs2_find_victim_alloc_group(), old code forgot enlarge bitmap
  range for global_bitmap case. Old code could generate negative
  vict_bit.
- For ocfs2_probe_alloc_group(), old code forgot back off move_len when
  finding enough bitmap space. Old code has possibility to make data
  corruption.
- For ocfs2_ioctl_move_extents(), this func should set me_threshold for
  both auto & non-auto path. Without setting me_threshold,
  ocfs2_move_extent() will make 'move_max_hop = 0', then
  ocfs2_probe_alloc_group() returns 'phys_cpos == 0' under max_hop is 0.

Another info is current defragfs.ocfs2 doesn't have ability to trigger
non-auto defrag path. I modified defragfs.ocfs2 source code
".me_flags"
and hardcode ".me_goal" for running test.

Signed-off-by: Heming Zhao <heming.zhao at suse.com>
---
 fs/ocfs2/move_extents.c | 24 +++++++++++++-----------
 1 file changed, 13 insertions(+), 11 deletions(-)

diff --git a/fs/ocfs2/move_extents.c b/fs/ocfs2/move_extents.c
index 6251748c695b..b1e32ec4a9d4 100644
--- a/fs/ocfs2/move_extents.c
+++ b/fs/ocfs2/move_extents.c
@@ -434,7 +434,7 @@ static int ocfs2_find_victim_alloc_group(struct inode
*inode,
 			bg = (struct ocfs2_group_desc *)gd_bh->b_data;
 
 			if (vict_blkno < (le64_to_cpu(bg->bg_blkno) +
-						le16_to_cpu(bg->bg_bits))) {
+						(le16_to_cpu(bg->bg_bits) << bits_per_unit))) {
 
 				*ret_bh = gd_bh;
 				*vict_bit = (vict_blkno - blkno) >>
@@ -549,6 +549,7 @@ static void ocfs2_probe_alloc_group(struct inode *inode,
struct buffer_head *bh,
 			last_free_bits++;
 
 		if (last_free_bits == move_len) {
+			i -= move_len;
 			*goal_bit = i;
 			*phys_cpos = base_cpos + i;
 			break;
@@ -1020,18 +1021,19 @@ int ocfs2_ioctl_move_extents(struct file *filp, void
__user *argp)
 
 	context->range = &range;
 
+	/*
+	 * ok, the default theshold for the defragmentation
+	 * is 1M, since our maximum clustersize was 1M also.
+	 * any thought?
+	 */
+	if (!range.me_threshold)
+		range.me_threshold = 1024 * 1024;
+
+	if (range.me_threshold > i_size_read(inode))
+		range.me_threshold = i_size_read(inode);
+
 	if (range.me_flags & OCFS2_MOVE_EXT_FL_AUTO_DEFRAG) {
 		context->auto_defrag = 1;
-		/*
-		 * ok, the default theshold for the defragmentation
-		 * is 1M, since our maximum clustersize was 1M also.
-		 * any thought?
-		 */
-		if (!range.me_threshold)
-			range.me_threshold = 1024 * 1024;
-
-		if (range.me_threshold > i_size_read(inode))
-			range.me_threshold = i_size_read(inode);
 
 		if (range.me_flags & OCFS2_MOVE_EXT_FL_PART_DEFRAG)
 			context->partial = 1;
-- 
2.39.0

On 2/17/23 9:20 AM, Heming Zhao wrote:
> This commit fixes three issues on non-auto defrag path (defragfs.ocfs2
> doesn't set OCFS2_MOVE_EXT_FL_AUTO_DEFRAG on range.me_flags):
> - For ocfs2_find_victim_alloc_group(), old code forgot enlarge bitmap
>   range for global_bitmap case. Old code could generate negative
>   vict_bit.
> - For ocfs2_probe_alloc_group(), old code forgot back off move_len when
>   finding enough bitmap space. Old code has possibility to make data
>   corruption.
> - For ocfs2_ioctl_move_extents(), this func should set me_threshold for
>   both auto & non-auto path. Without setting me_threshold,
>   ocfs2_move_extent() will make 'move_max_hop = 0', then
>   ocfs2_probe_alloc_group() returns 'phys_cpos == 0' under max_hop
is 0.
> 
> Another info is current defragfs.ocfs2 doesn't have ability to trigger
> non-auto defrag path. I modified defragfs.ocfs2 source code
".me_flags"
> and hardcode ".me_goal" for running test.
> 
I'm fine with the code changes. But the commit log needs improve.
How about the following?

This fixes three issues on move extents ioctl without auto defrag:

a) In ocfs2_find_victim_alloc_group(), we have to convert bits to block
first in case of global bitmap.
b) In ocfs2_probe_alloc_group(), when finding enough bits in block group
bitmap, we have to back off move_len to start pos as well, otherwise it
may corrupt filesystem.
c) In ocfs2_ioctl_move_extents(), set me_threshold both for non-auto and
auto defrag paths. Otherwise it will set move_max_hop to 0 and finally
cause unexpectedly ENOSPC error.

Currently there is no tools triggering the above issues since
defragfs.ocfs2 enables auto defrag by default. Test with manually
changing defragfs.ocfs2 to run non auto defrag path.

Thanks,
Joseph

> Signed-off-by: Heming Zhao <heming.zhao at suse.com>
> ---
>  fs/ocfs2/move_extents.c | 24 +++++++++++++-----------
>  1 file changed, 13 insertions(+), 11 deletions(-)
> 
> diff --git a/fs/ocfs2/move_extents.c b/fs/ocfs2/move_extents.c
> index 6251748c695b..b1e32ec4a9d4 100644
> --- a/fs/ocfs2/move_extents.c
> +++ b/fs/ocfs2/move_extents.c
> @@ -434,7 +434,7 @@ static int ocfs2_find_victim_alloc_group(struct inode
*inode,
>  			bg = (struct ocfs2_group_desc *)gd_bh->b_data;
>  
>  			if (vict_blkno < (le64_to_cpu(bg->bg_blkno) +
> -						le16_to_cpu(bg->bg_bits))) {
> +						(le16_to_cpu(bg->bg_bits) << bits_per_unit))) {
>  
>  				*ret_bh = gd_bh;
>  				*vict_bit = (vict_blkno - blkno) >>
> @@ -549,6 +549,7 @@ static void ocfs2_probe_alloc_group(struct inode
*inode, struct buffer_head *bh,
>  			last_free_bits++;
>  
>  		if (last_free_bits == move_len) {
> +			i -= move_len;
>  			*goal_bit = i;
>  			*phys_cpos = base_cpos + i;
>  			break;
> @@ -1020,18 +1021,19 @@ int ocfs2_ioctl_move_extents(struct file *filp,
void __user *argp)
>  
>  	context->range = &range;
>  
> +	/*
> +	 * ok, the default theshold for the defragmentation
> +	 * is 1M, since our maximum clustersize was 1M also.
> +	 * any thought?
> +	 */
> +	if (!range.me_threshold)
> +		range.me_threshold = 1024 * 1024;
> +
> +	if (range.me_threshold > i_size_read(inode))
> +		range.me_threshold = i_size_read(inode);
> +
>  	if (range.me_flags & OCFS2_MOVE_EXT_FL_AUTO_DEFRAG) {
>  		context->auto_defrag = 1;
> -		/*
> -		 * ok, the default theshold for the defragmentation
> -		 * is 1M, since our maximum clustersize was 1M also.
> -		 * any thought?
> -		 */
> -		if (!range.me_threshold)
> -			range.me_threshold = 1024 * 1024;
> -
> -		if (range.me_threshold > i_size_read(inode))
> -			range.me_threshold = i_size_read(inode);
>  
>  		if (range.me_flags & OCFS2_MOVE_EXT_FL_PART_DEFRAG)
>  			context->partial = 1;