[PATCH v6 13/18] nvme-fc: perform error recovery directly from ioerr_work

[Daniel Wagner]

On Sun, Sep 20, 2026 at 11:28:11AM -0700, Mohamed Khalfella wrote:
> +static void
> +nvme_fc_error_recovery(struct nvme_fc_ctrl *ctrl)
> +{
> +	nvme_stop_keep_alive(&ctrl->ctrl);
> +	flush_work(&ctrl->ctrl.async_event_work);
> +
> +	/* will block while waiting for io to terminate */
> +	nvme_fc_delete_association(ctrl);
> +	nvme_stop_ctrl(&ctrl->ctrl);

Not happy that we have duplicated this part, but after reading on
tcp/rdma it's the 'common' pattern. I think it make sense to move the fc
towards how tcp/rdma is doing it so that we can make common code out of
it eventually.

So from this, looks good.

Reviewed-by: Daniel Wagner <dwagner at suse.de>