Re: [PATCH 2/3] Btrfs: move btrfs_free_qgroup_config() out of spin_lock and fix comments

[Arne Jansen <sensille@gmx.net>]

On 07.08.2013 07:12, Wang Shilong wrote:
> btrfs_free_qgroup_config() is not only called by open/close_ctree(),but
> also btrfs_disable_quota().And for btrfs_disable_quota(),we have set
> 'quota_root' to be null before calling btrfs_free_qgroup_config(),so it
> is safe to cleanup in-memory structures without lock held.

Why do you do this? Did the spinlock gave you any trouble?

> 
> Signed-off-by: Wang Shilong <wangsl.fnst@cn.fujitsu.com>
> Reviewed-by: Miao Xie <miaox@cn.fujitsu.com>
> ---
>  fs/btrfs/qgroup.c | 9 ++++++---
>  1 file changed, 6 insertions(+), 3 deletions(-)
> 
> diff --git a/fs/btrfs/qgroup.c b/fs/btrfs/qgroup.c
> index 3b103e2..b809616 100644
> --- a/fs/btrfs/qgroup.c
> +++ b/fs/btrfs/qgroup.c
> @@ -432,8 +432,10 @@ out:
>  }
>  
>  /*
> - * This is only called from close_ctree() or open_ctree(), both in single-
> - * treaded paths. Clean up the in-memory structures. No locking needed.
> + * This is called from close_ctree() or open_ctree() or btrfs_quota_disable(),
> + * first two are in single-treaded paths.And for the third one, we have set

You could use the opportunity and correct the typo in 'single-treaded'. Also,
there's a space missing after the period.

-Arne

> + * quota_root to be null with qgroup_lock held before, so it is safe to clean
> + * up the in-memory structures without qgroup_lock held.
>   */
>  void btrfs_free_qgroup_config(struct btrfs_fs_info *fs_info)
>  {
> @@ -937,9 +939,10 @@ int btrfs_quota_disable(struct btrfs_trans_handle *trans,
>  	fs_info->pending_quota_state = 0;
>  	quota_root = fs_info->quota_root;
>  	fs_info->quota_root = NULL;
> -	btrfs_free_qgroup_config(fs_info);
>  	spin_unlock(&fs_info->qgroup_lock);
>  
> +	btrfs_free_qgroup_config(fs_info);
> +
>  	if (!quota_root) {
>  		ret = -EINVAL;
>  		goto out;