Re: [RFC v1 4/4] nvmet-discovery: do not use invalid port

[Daniel Wagner]

On Mon, Sep 11, 2023 at 04:44:47PM +0200, Daniel Wagner wrote:
> On Tue, Sep 05, 2023 at 12:40:25PM +0200, Daniel Wagner wrote:
> > > But I'm still confused how we can get here without req->port
> > > set.  Can you try to do a little more analysis as I suspect we have
> > > a deeper problem somewhere.
> 
> The problem is that nvme/005 starts to cleanup all resources and there
> is a race between the cleanup path and the host trying to figure out
> what's going on (get log page).
> 
> We have 3 association:
> 
>  assoc 0: systemd/udev triggered 'nvme connect-all' discovery controller
>  assoc 1: discovery controller from nvme/005
>  assoc 2: i/o controller from nvme/005

Actually, assoc 1 is also a i/o controller for the same hostnqn as assoc
2. This looks more like it assoc 0 and 1 are both from systemd/udev
trigger. But why the heck is the hostnqn for assoc 1 the same as the
hostnqn we use in blktests? Something is really off here.

The rest of my analysis is still valid.