On Wed, Feb 14, 2024 at 10:25:57AM -0700, Keith Busch wrote:
> On Wed, Feb 14, 2024 at 10:54:57AM +0100, Christoph Hellwig wrote:
> > @@ -2036,11 +1813,15 @@ static int null_validate_conf(struct nullb_device *dev)
> > pr_err("legacy IO path is no longer available\n");
> > return -EINVAL;
> > }
> > + if (dev->queue_mode == NULL_Q_BIO) {
> > + pr_err("BIO-based IO path is no longer available, using blk-mq instead.\n");
> > + dev->queue_mode = NULL_Q_MQ;
> > + }
>
> Seems pointless to keep dev->queue_mode around if only one value is
> valid.
>
> Instead of checking the param here once per device, could we do it just
> once for the module in null_set_queue_mode()?
Maybe. Note that the code would have to be quite a bit more complex
than this - not only due to configfs as noted by Damien, but also
because NULL_Q_BIO isn't rejected but simply ignored to not break
blktests. I can look at this at parsing time, but my gut feeling
is that it's not worth the effort.
