[PATCH 1/5] nvme-fc: don't hold rport lock when putting ctrl

[Daniel Wagner]

Hi Justin,

On Tue, Oct 28, 2025 at 05:33:17PM -0700, Justin Tee wrote:
> > nvme_fc_ctrl_put can acquire the rport lock when freeing the
> > ctrl object:
> >
> > nvme_fc_ctrl_put
> >   nvme_fc_ctrl_free
> >     spin_lock_irqsave(rport->lock)
> >
> > Thus we can't hold the rport lock when calling nvme_fc_ctrl_put.
> 
> While I agree that we can’t hold the rport lock when calling
> nvme_fc_ctrl_put, nvme_fc_ctrl_free also does a nvme_fc_rport_put,
> which could also trigger nvme_fc_free_rport, making rport invalid.
> Should we also add kref get on the rport before entering the
> list_for_each_entry loop?
> 
> Also, because nvme_fc_ctrl_free removes itself from the
> rport->ctrl_list, should we also start using list_for_each_entry_safe?
> 
> So, something like this?

Yes, this makes sense. Just wondering why I didn't see any KASAN
reports. 

Should I add your change to my patch (obviously mentioning it), or do
you want to send a patch yourself?

In the meantime, I am giving this patch a spin in my test setup.

Thanks,
Daniel