[PATCH] f2fs: don't call iput() from f2fs_drop_inode()

[PATCH] f2fs: don't call iput() from f2fs_drop_inode()

[Mateusz Guzik]

iput() calls the problematic routine, which does a ->i_count inc/dec
cycle. Undoing it with iput() recurses into the problem.

Note f2fs should not be playing games with the refcount to begin with,
but that will be handled later. Right now solve the immediate
regression.

Fixes: bc986b1d756482a ("fs: stop accessing ->i_count directly in f2fs and gfs2")
Reported-by: kernel test robot <oliver.sang@intel.com>
Closes: https://lore.kernel.org/oe-lkp/202509301450.138b448f-lkp@intel.com
Signed-off-by: Mateusz Guzik <mjguzik@gmail.com>
---
 fs/f2fs/super.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/fs/f2fs/super.c b/fs/f2fs/super.c
index 2619cbbd7d2d..26ec31eb8c80 100644
--- a/fs/f2fs/super.c
+++ b/fs/f2fs/super.c
@@ -1769,7 +1769,7 @@ static int f2fs_drop_inode(struct inode *inode)
 			sb_end_intwrite(inode->i_sb);
 
 			spin_lock(&inode->i_lock);
-			iput(inode);
+			atomic_dec(&inode->i_count);
 		}
 		trace_f2fs_drop_inode(inode, 0);
 		return 0;
-- 
2.43.0

Re: [PATCH] f2fs: don't call iput() from f2fs_drop_inode()

[Jaegeuk Kim]

On 10/01, Mateusz Guzik wrote:
> iput() calls the problematic routine, which does a ->i_count inc/dec
> cycle. Undoing it with iput() recurses into the problem.
> 
> Note f2fs should not be playing games with the refcount to begin with,
> but that will be handled later. Right now solve the immediate
> regression.
> 
> Fixes: bc986b1d756482a ("fs: stop accessing ->i_count directly in f2fs and gfs2")
> Reported-by: kernel test robot <oliver.sang@intel.com>
> Closes: https://lore.kernel.org/oe-lkp/202509301450.138b448f-lkp@intel.com
> Signed-off-by: Mateusz Guzik <mjguzik@gmail.com>
> ---
>  fs/f2fs/super.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/fs/f2fs/super.c b/fs/f2fs/super.c
> index 2619cbbd7d2d..26ec31eb8c80 100644
> --- a/fs/f2fs/super.c
> +++ b/fs/f2fs/super.c
> @@ -1769,7 +1769,7 @@ static int f2fs_drop_inode(struct inode *inode)
>  			sb_end_intwrite(inode->i_sb);
>  
>  			spin_lock(&inode->i_lock);
> -			iput(inode);
> +			atomic_dec(&inode->i_count);

It seems this was applied by Josef [1], added in 6.18-rc1. Let me apply this fix
after my f2fs pull request, since I don't have this issue in my -next tree yet.

[1] https://lore.kernel.org/all/b8e6eb8a3e690ce082828d3580415bf70dfa93aa.1755806649.git.josef@toxicpanda.com/

>  		}
>  		trace_f2fs_drop_inode(inode, 0);
>  		return 0;
> -- 
> 2.43.0
> 

Re: [f2fs-dev] [PATCH] f2fs: don't call iput() from f2fs_drop_inode()

[patchwork-bot+f2fs]

Hello:

This patch was applied to jaegeuk/f2fs.git (dev)
by Jaegeuk Kim <jaegeuk@kernel.org>:

On Wed,  1 Oct 2025 01:29:57 +0200 you wrote:
> iput() calls the problematic routine, which does a ->i_count inc/dec
> cycle. Undoing it with iput() recurses into the problem.
> 
> Note f2fs should not be playing games with the refcount to begin with,
> but that will be handled later. Right now solve the immediate
> regression.
> 
> [...]

Here is the summary with links:
  - [f2fs-dev] f2fs: don't call iput() from f2fs_drop_inode()
    https://git.kernel.org/jaegeuk/f2fs/c/8ec5fc1ff77e

You are awesome, thank you!
-- 
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html

Re: [f2fs-dev] [PATCH] f2fs: don't call iput() from f2fs_drop_inode()

[Chao Yu]

On 10/1/2025 7:29 AM, Mateusz Guzik wrote:
> iput() calls the problematic routine, which does a ->i_count inc/dec
> cycle. Undoing it with iput() recurses into the problem.
> 
> Note f2fs should not be playing games with the refcount to begin with,
> but that will be handled later. Right now solve the immediate
> regression.
> 
> Fixes: bc986b1d756482a ("fs: stop accessing ->i_count directly in f2fs and gfs2")
> Reported-by: kernel test robot <oliver.sang@intel.com>
> Closes: https://lore.kernel.org/oe-lkp/202509301450.138b448f-lkp@intel.com
> Signed-off-by: Mateusz Guzik <mjguzik@gmail.com>

Reviewed-by: Chao Yu <chao@kernel.org>

Thanks,