Re: [PATCH v6 1/4] block: add task-context bio completion infrastructure

[Christoph Hellwig]

On Wed, Jul 29, 2026 at 10:41:40AM +0200, Jan Kara wrote:
> This looks a bit hacky and at least deserves a comment. Ideally we'd get
> rid of the xfs workqueue and rely on the generic offloading mechanism. But
> as Christoph wrote maybe let's just get something good enough in and we can
> work on improving that later. In particular if Christoph has a plan how to
> clean this up :).

The XFS workqueue can't just go away, as for writes that require
completion time processing (unwritten extent conversion, i_size logging,
zoned block mapping), we want to serialize the processing on a per-inode
basis as otherwise the different helpers would just content on the xfs
inode lock and burn cycles.

We want to get rid of it for read completions, and a patch for that is
in the lazy bounce series I posted a few days ago.