[PATCH v0 5/6] nvme-fc: redesign locking and refcounting

[Daniel Wagner]

On Fri, Feb 16, 2024 at 12:09:20PM +0100, Hannes Reinecke wrote:
> Hmm. I'm a bit unsure about this; essentially you change the rport
> refcounting (and not just the controller refcounting).
> And the problem here is that rport refcounting is actually tied to
> the driver-internal rports, which have a different lifetime
> (dev_loss_tmo and all that).
> 
> Would it be possible to break this in two, with one patch changing the
> controller/options refcounting and the other one changing the rport
> refcounting?

Yeah, I see. I completely forgot about the dev_loss_tmo thing. I'll try
to split this patch.