Re: [PATCH 14/17] btrfs: move btrfs_path_cachep out of ctree.h

[Qu Wenruo <wqu@suse.com>]

On 2022/9/14 23:06, Josef Bacik wrote:
> This is local to the ctree code, remove it from ctree.h and inode.c,
> create new init/exit functions for the cachep, and move it locally to
> ctree.c.
> 
> Signed-off-by: Josef Bacik <josef@toxicpanda.com>

Reviewed-by: Qu Wenruo <wqu@suse.com>

But one thing mentioned below.

[...]
> @@ -2813,6 +2817,8 @@ static int __init init_btrfs_fs(void)
>   	extent_buffer_free_cachep();
>   free_extent_cachep:
>   	extent_state_free_cachep();
> +free_ctree:
> +	btrfs_ctree_exit();

The tags will grow larger and larger.

Can we make a new cleanup to make all those _exit() function to be 
unconditional? (e.g. check the cachep and only free them if initialized)

That already happened for open_ctree() which has some incorrect error 
handling.

Thus a new small series cleaning up those tags would be a good idea.

Thanks,
Qu

>   free_transaction:
>   	btrfs_transaction_exit();
>   free_cachep:
> @@ -2826,6 +2832,7 @@ static int __init init_btrfs_fs(void)
>   
>   static void __exit exit_btrfs_fs(void)
>   {
> +	btrfs_ctree_exit();
>   	btrfs_transaction_exit();
>   	btrfs_destroy_cachep();
>   	btrfs_delayed_ref_exit();