Re: [PATCH] dm-crypt: defer async write submission to write thread

[Josef Schlehofer <pepe.schlehofer@gmail.com>]

On Mon, Nov 9, 2026 at 8:51 PM Ignat Korchagin <ignat@linux.win> wrote:
> At the very least I think it may be an unexpected/undesirable setup
> for the user: they enable "no_write_workqueue" expecting a fully
> inline setup, but still get io offloaded. Perhaps in such cases we can
> do a pr_warn_once() to let the user know their setup is impossible?

Now that Damien has confirmed async is fine for zoned devices, I would
like to add this warning as a follow-up patch on top of dm-7.4.

Would a pr_warn_once() in kcryptd_crypt_write_io_submit() be acceptable,
printed when an async completion is sent to the write thread despite
no_write_workqueue? The other option is a warning at table load when
no_write_workqueue is set and the cipher is async. Zoned devices would
not get the warning in either case.

Mikulas, which one would you prefer?

Thanks,
Josef