Re: [PATCH v3 0/6] AG aware parallel writeback for XFS

[Christoph Hellwig <hch@lst.de>]

On Fri, Feb 06, 2026 at 03:37:38PM +0530, Kundan Kumar wrote:
> When you say "coarse-grained sharding", do you mean tracking a single
> "home AG" per inode (no per-folio tagging) and using it as a best-effort
> hint for writeback routing?

Yes, but..

> If so, we can align with that and keep the threading model generic by
> relying on bdi writeback contexts. Concretely, XFS would set up a 
> bounded number of bdi wb contexts at mount time, and route each inode to 
> its home shard. Does this align with what you have in mind?

Yes.

> We had implemented a similar approach in earlier versions of this
> series[1]. But the feedback[2] that we got was that mapping high level
> writeback to eventual AG allocation can be sometimes inaccurate (aged
> filesystems, inode spanning accross multiple AGs, etc.), so we moved to
> per folio tagging to make the routing decision closer to the actual IO.
> That said, I agree that the per folio approach is complex.

It is inaccurate, not to say often wrong.  And that's the point I'm
trying to make here:  start with simplifying the allocation policies
so that they make sense in this kind of environment, and then align
that sharding to that.