diff mbox series

ext4: Enable meta_bg only when new desc blocks are needed

Message ID 20240227131329.2608466-1-srivathsa.d.dara@oracle.com
State Awaiting Upstream
Headers show
Series ext4: Enable meta_bg only when new desc blocks are needed | expand

Commit Message

Srivathsa Dara Feb. 27, 2024, 1:13 p.m. UTC
This patch addresses an issue observed when resize_inode is disabled
and an online extension of a filesysyem is performed. When a filesystem
is expanded to a size that does not require a addition of a new
descriptor block, the meta_bg feature is being enabled even though no
part of the filesystem uses this layout.

This patch ensures that the meta_bg feature is only enabled if
any of the added block groups utilize meta_bg layout.

Signed-off-by: Srivathsa Dara <srivathsa.d.dara@oracle.com>
---
 fs/ext4/resize.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

Comments

Ritesh Harjani (IBM) March 6, 2024, 4:02 a.m. UTC | #1
Srivathsa Dara <srivathsa.d.dara@oracle.com> writes:

> This patch addresses an issue observed when resize_inode is disabled

Other than "meta_bg unwatedly getting enabled when not required", is
there any other issue you observed?

> and an online extension of a filesysyem is performed. When a filesystem
> is expanded to a size that does not require a addition of a new
> descriptor block, the meta_bg feature is being enabled even though no
> part of the filesystem uses this layout.
>
> This patch ensures that the meta_bg feature is only enabled if
> any of the added block groups utilize meta_bg layout.

Make sense to me. 

>
> Signed-off-by: Srivathsa Dara <srivathsa.d.dara@oracle.com>
> ---
>  fs/ext4/resize.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/fs/ext4/resize.c b/fs/ext4/resize.c
> index 928700d57eb6..99b52f26e818 100644
> --- a/fs/ext4/resize.c
> +++ b/fs/ext4/resize.c
> @@ -1996,7 +1996,7 @@ int ext4_resize_fs(struct super_block *sb, ext4_fsblk_t n_blocks_count)
>  		}
>  	}
>  
> -	if ((!resize_inode && !meta_bg) || n_blocks_count == o_blocks_count) {
> +	if ((!resize_inode && !meta_bg && n_desc_blocks > o_desc_blocks) || n_blocks_count == o_blocks_count) {

Beyond 80 chars line. You might want to fix that.

-ritesh

>  		err = ext4_convert_meta_bg(sb, resize_inode);
>  		if (err)
>  			goto out;
> -- 
> 2.39.3
Srivathsa Dara March 6, 2024, 6:55 a.m. UTC | #2
Hi Ritesh,

> -----Original Message-----
> From: Ritesh Harjani <ritesh.list@gmail.com> 
> Sent: Wednesday, March 6, 2024 9:32 AM
> To: Srivathsa Dara <srivathsa.d.dara@oracle.com>; linux-ext4@vger.kernel.org
> Cc: tytso@mit.edu; adilger.kernel@dilger.ca; Rajesh Sivaramasubramaniom <rajesh.sivaramasubramaniom@oracle.com>; Junxiao Bi <junxiao.bi@oracle.com>
> Subject: Re: [PATCH] ext4: Enable meta_bg only when new desc blocks are needed
> 
> Srivathsa Dara <srivathsa.d.dara@oracle.com> writes:
> 
> > This patch addresses an issue observed when resize_inode is disabled
> 
> Other than "meta_bg unwatedly getting enabled when not required", is there any other issue you observed?

Yes, we have seen an issue where a data block is being corrupted by group 
descriptor block because of this. The following patch fixes that issue.

	ext4: fix corruption during on-line resize
	commit: 3a944549dd26ccaf1f898a4be952e75a42bf37dd
	
> 
> > and an online extension of a filesysyem is performed. When a 
> > filesystem is expanded to a size that does not require a addition of a 
> > new descriptor block, the meta_bg feature is being enabled even though 
> > no part of the filesystem uses this layout.
> >
> > This patch ensures that the meta_bg feature is only enabled if any of 
> > the added block groups utilize meta_bg layout.
> 
> Make sense to me. 
> 
> >
> > Signed-off-by: Srivathsa Dara <srivathsa.d.dara@oracle.com>
> > ---
> >  fs/ext4/resize.c | 2 +-
> >  1 file changed, 1 insertion(+), 1 deletion(-)
> >
> > diff --git a/fs/ext4/resize.c b/fs/ext4/resize.c index 
> > 928700d57eb6..99b52f26e818 100644
> > --- a/fs/ext4/resize.c
> > +++ b/fs/ext4/resize.c
> > @@ -1996,7 +1996,7 @@ int ext4_resize_fs(struct super_block *sb, ext4_fsblk_t n_blocks_count)
> >  		}
> >  	}
> >  
> > -	if ((!resize_inode && !meta_bg) || n_blocks_count == o_blocks_count) {
> > +	if ((!resize_inode && !meta_bg && n_desc_blocks > o_desc_blocks) || 
> > +n_blocks_count == o_blocks_count) {
> 
> Beyond 80 chars line. You might want to fix that.

Sorry, I missed that, will send a V2.

> 
> -ritesh
> 
> >  		err = ext4_convert_meta_bg(sb, resize_inode);
> >  		if (err)
> >  			goto out;
> > --
> > 2.39.3

Thanks,
Srivathsa Dara
Theodore Ts'o March 14, 2024, 3:54 a.m. UTC | #3
On Tue, 27 Feb 2024 13:13:29 +0000, Srivathsa Dara wrote:
> This patch addresses an issue observed when resize_inode is disabled
> and an online extension of a filesysyem is performed. When a filesystem
> is expanded to a size that does not require a addition of a new
> descriptor block, the meta_bg feature is being enabled even though no
> part of the filesystem uses this layout.
> 
> This patch ensures that the meta_bg feature is only enabled if
> any of the added block groups utilize meta_bg layout.
> 
> [...]

Applied, thanks!

[1/1] ext4: Enable meta_bg only when new desc blocks are needed
      commit: 07be778c70149321f785611a9c50125b904b0508

Best regards,
diff mbox series

Patch

diff --git a/fs/ext4/resize.c b/fs/ext4/resize.c
index 928700d57eb6..99b52f26e818 100644
--- a/fs/ext4/resize.c
+++ b/fs/ext4/resize.c
@@ -1996,7 +1996,7 @@  int ext4_resize_fs(struct super_block *sb, ext4_fsblk_t n_blocks_count)
 		}
 	}
 
-	if ((!resize_inode && !meta_bg) || n_blocks_count == o_blocks_count) {
+	if ((!resize_inode && !meta_bg && n_desc_blocks > o_desc_blocks) || n_blocks_count == o_blocks_count) {
 		err = ext4_convert_meta_bg(sb, resize_inode);
 		if (err)
 			goto out;