Re: [PATCH v6 11/16] ntfs: update runlist handling and cluster allocator

[Christoph Hellwig <hch@lst.de>]

Suggested commit message:

Updates runlist handling and cluster allocation to support
contiguous allocations and filesystem trimming.

Improve the runlist API to handle allocation failures and introduces
discard support.

> +		if (is_dealloc == true)
> +			ntfs_release_dirty_clusters(vol, rl->length);
>  		up_write(&vol->lcnbmp_lock);
> +		memalloc_nofs_restore(memalloc_flags);
>  		ntfs_debug("Done.");
> +		return rl == NULL ? ERR_PTR(-EIO) : rl;

In general you want the memalloc_nofs_restore to be after goto
labels in a single place, as otherwise debugging is really hard.
In doubt a separate wrapper doing it my be even better.