[PATCH] nvme: re-read ANA log on NS CHANGED AEN

[Keith Busch]

On Fri, Dec 04, 2020 at 04:03:10PM +0100, Hannes Reinecke wrote:
> -	error = nvme_parse_ana_log(ctrl, &nr_change_groups,
> -			nvme_update_ana_state);
> -	if (error)
> -		goto out_unlock;
> +	/*
> +	 * Don't update ANA groups if triggered by an NS CHANGED
> +	 * AEN; we'll be rescanning all namespaces anyway afterwards.
> +	 */
> +	if (!test_bit(NVME_AER_NOTICE_NS_CHANGED, &ctrl->events)) {
> +		error = nvme_parse_ana_log(ctrl, &nr_change_groups,
> +					   nvme_update_ana_state);
> +		if (error)
> +			goto out_unlock;
> +	}
>  
>  	/*
>  	 * In theory we should have an ANATT timer per group as they might enter
> @@ -557,6 +563,10 @@ static int nvme_read_ana_log(struct nvme_ctrl *ctrl)
>  		del_timer_sync(&ctrl->anatt_timer);
>  out_unlock:
>  	mutex_unlock(&ctrl->ana_lock);
> +
> +	if (test_bit(NVME_AER_NOTICE_NS_CHANGED, &ctrl->events))
> +		nvme_queue_scan(ctrl);
> +
>  	return error;
>  }

The scan work can be started from other contexts too, so there are some
unfortunate races here if one of those other contexts clears ctrl->events
before the ANA work gets to see it, which then gets you back to having
unusable namespaces when that scan work uses a stale ana log.