[PATCH 1/2] ext4: teach resize report old blocks count correctly

[PATCH 1/2] ext4: teach resize report old blocks count correctly

[Yongqiang Yang]

o_blocks_count should not be changed, becasuse it is used later when
reporting old blocks count in debug mode.

Signed-off-by: Yongqiang Yang <xiaoqiangnk@gmail.com>
---
 fs/ext4/resize.c |    3 +--
 1 files changed, 1 insertions(+), 2 deletions(-)

diff --git a/fs/ext4/resize.c b/fs/ext4/resize.c
index 59fa0be..601fea1 100644
--- a/fs/ext4/resize.c
+++ b/fs/ext4/resize.c
@@ -1654,8 +1654,7 @@ int ext4_resize_fs(struct super_block *sb, ext4_fsblk_t n_blocks_count)
 	    es->s_log_groups_per_flex)
 		flexbg_size = 1 << es->s_log_groups_per_flex;
 
-	o_blocks_count = ext4_blocks_count(es);
-	if (o_blocks_count == n_blocks_count)
+	if (ext4_blocks_count(es) == n_blocks_count)
 		goto out;
 
 	flex_gd = alloc_flex_gd(flexbg_size);
-- 
1.7.5.1

[PATCH 2/2] ext4: ignore last group without enough space when resizing

[Yongqiang Yang]

If last gtoup has no enough space for group tables, then we
just ignore it.

Reported-by: Daniel Drake <dsd@laptop.org>
Signed-off-by: Yongqiang Yang <xiaoqiangnk@gmail.com>
---
 fs/ext4/resize.c |   12 +++++++++---
 1 files changed, 9 insertions(+), 3 deletions(-)

diff --git a/fs/ext4/resize.c b/fs/ext4/resize.c
index 601fea1..7a961cc 100644
--- a/fs/ext4/resize.c
+++ b/fs/ext4/resize.c
@@ -199,7 +199,7 @@ static void free_flex_gd(struct ext4_new_flex_group_data *flex_gd)
  *
  * @sb: super block of fs to which the groups belongs
  */
-static void ext4_alloc_group_tables(struct super_block *sb,
+static int ext4_alloc_group_tables(struct super_block *sb,
 				struct ext4_new_flex_group_data *flex_gd,
 				int flexbg_size)
 {
@@ -224,6 +224,8 @@ static void ext4_alloc_group_tables(struct super_block *sb,
 	       (last_group & ~(flexbg_size - 1))));
 next_group:
 	group = group_data[0].group;
+	if (src_group >= group_data[0].group + flex_gd->count)
+		return -EINVAL;
 	start_blk = ext4_group_first_block_no(sb, src_group);
 	last_blk = start_blk + group_data[src_group - group].blocks_count;
 
@@ -233,7 +235,6 @@ next_group:
 
 	start_blk += overhead;
 
-	BUG_ON(src_group >= group_data[0].group + flex_gd->count);
 	/* We collect contiguous blocks as much as possible. */
 	src_group++;
 	for (; src_group <= last_group; src_group++)
@@ -298,6 +299,7 @@ next_group:
 			       group_data[i].free_blocks_count);
 		}
 	}
+	return 0;
 }
 
 static struct buffer_head *bclean(handle_t *handle, struct super_block *sb,
@@ -1668,7 +1670,11 @@ int ext4_resize_fs(struct super_block *sb, ext4_fsblk_t n_blocks_count)
 	 */
 	while (ext4_setup_next_flex_gd(sb, flex_gd, n_blocks_count,
 					      flexbg_size)) {
-		ext4_alloc_group_tables(sb, flex_gd, flexbg_size);
+		err = ext4_alloc_group_tables(sb, flex_gd, flexbg_size);
+		if (err == -EINVAL) {
+			err = 0;
+			break;
+		}
 		err = ext4_flex_group_add(sb, resize_inode, flex_gd);
 		if (unlikely(err))
 			break;
-- 
1.7.5.1

Re: [PATCH 2/2] ext4: ignore last group without enough space when resizing

[Daniel Drake]

Hi,

On Tue, Jan 31, 2012 at 9:01 PM, Yongqiang Yang <xiaoqiangnk@gmail.com> wrote:
> If last gtoup has no enough space for group tables, then we
> just ignore it.
>
> Reported-by: Daniel Drake <dsd@laptop.org>
> Signed-off-by: Yongqiang Yang <xiaoqiangnk@gmail.com>

Thanks, this fixes the crash.

Am I right in thinking that the companion patch "[PATCH] resize2fs:
let online resizing report new blocks count right" is only needed for
cosmetic reasons; the resize operation will be successful even without
that patch applied to resize2fs?

Thanks,
Daniel

Re: [PATCH 2/2] ext4: ignore last group without enough space when resizing

[Yongqiang Yang]

On Mon, Apr 2, 2012 at 11:04 PM, Daniel Drake <dsd@laptop.org> wrote:
> Hi,
>
> On Tue, Jan 31, 2012 at 9:01 PM, Yongqiang Yang <xiaoqiangnk@gmail.com> wrote:
>> If last gtoup has no enough space for group tables, then we
>> just ignore it.
>>
>> Reported-by: Daniel Drake <dsd@laptop.org>
>> Signed-off-by: Yongqiang Yang <xiaoqiangnk@gmail.com>
>
> Thanks, this fixes the crash.
>
> Am I right in thinking that the companion patch "[PATCH] resize2fs:
> let online resizing report new blocks count right" is only needed for
> cosmetic reasons; the resize operation will be successful even without
> that patch applied to resize2fs?
Yep,  that patch just lets resize2fs report right size of resized filesytems.

Thanks,
Yongqiang.
>
> Thanks,
> Daniel



-- 
Best Wishes
Yongqiang Yang

Re: [PATCH 1/2] ext4: teach resize report old blocks count correctly

[Yongqiang Yang]

ping?

The 2nd patch of the series fixes a bug which I meet again when
testing 64bit feature.


Yongqiang.

On Wed, Feb 1, 2012 at 11:01 AM, Yongqiang Yang <xiaoqiangnk@gmail.com> wrote:
> o_blocks_count should not be changed, becasuse it is used later when
> reporting old blocks count in debug mode.
>
> Signed-off-by: Yongqiang Yang <xiaoqiangnk@gmail.com>
> ---
>  fs/ext4/resize.c |    3 +--
>  1 files changed, 1 insertions(+), 2 deletions(-)
>
> diff --git a/fs/ext4/resize.c b/fs/ext4/resize.c
> index 59fa0be..601fea1 100644
> --- a/fs/ext4/resize.c
> +++ b/fs/ext4/resize.c
> @@ -1654,8 +1654,7 @@ int ext4_resize_fs(struct super_block *sb, ext4_fsblk_t n_blocks_count)
>            es->s_log_groups_per_flex)
>                flexbg_size = 1 << es->s_log_groups_per_flex;
>
> -       o_blocks_count = ext4_blocks_count(es);
> -       if (o_blocks_count == n_blocks_count)
> +       if (ext4_blocks_count(es) == n_blocks_count)
>                goto out;
>
>        flex_gd = alloc_flex_gd(flexbg_size);
> --
> 1.7.5.1
>



-- 
Best Wishes
Yongqiang Yang
--
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