Re: [PATCH 1/2] nvme: only allow entering LIVE from CONNECTING state

[Hannes Reinecke <hare@suse.de>]

On 4/29/25 15:55, Daniel Wagner wrote:
> On Mon, Apr 28, 2025 at 03:21:18PM +0200, Hannes Reinecke wrote:
>> On 4/28/25 14:44, Daniel Wagner wrote:
>>> diff --git a/drivers/nvme/host/core.c b/drivers/nvme/host/core.c
>>> index b502ac07483b..d3c4eacf607f 100644
>>> --- a/drivers/nvme/host/core.c
>>> +++ b/drivers/nvme/host/core.c
>>> @@ -4493,7 +4493,8 @@ static void nvme_fw_act_work(struct work_struct *work)
>>>                   msleep(100);
>>>           }
>>>
>>> -       if (!nvme_change_ctrl_state(ctrl, NVME_CTRL_LIVE))
>>> +       if (!nvme_change_ctrl_state(ctrl, NVME_CTRL_CONNECTING) ||
>>> +           !nvme_change_ctrl_state(ctrl, NVME_CTRL_LIVE))
>>>                   return;
>>>
>>>           nvme_unquiesce_io_queues(ctrl);
>>
>> I would rather have a separate state for firmware activation.
>> (Ab-)using the 'RESETTING' state here has direct implications
>> with the error handler, as for the error handler 'RESETTING'
>> means that the error handler has been scheduled.
>> Which is not true for firmware activation.
> 
> Okay, so something like this here (untested, working on it)?
> 
> 
> diff --git a/drivers/nvme/host/core.c b/drivers/nvme/host/core.c
> index b502ac07483b..32482712d0f2 100644
> --- a/drivers/nvme/host/core.c
> +++ b/drivers/nvme/host/core.c
> @@ -565,6 +565,7 @@ bool nvme_change_ctrl_state(struct nvme_ctrl *ctrl,
>   	case NVME_CTRL_LIVE:
>   		switch (old_state) {
>   		case NVME_CTRL_CONNECTING:
> +		case NVME_CTRL_FW_ACTIVATION:
>   			changed = true;
>   			fallthrough;
>   		default:
> @@ -575,6 +576,7 @@ bool nvme_change_ctrl_state(struct nvme_ctrl *ctrl,
>   		switch (old_state) {
>   		case NVME_CTRL_NEW:
>   		case NVME_CTRL_LIVE:
> +		case NVME_CTRL_FW_ACTIVATION:
>   			changed = true;
>   			fallthrough;
>   		default:
> @@ -596,6 +598,7 @@ bool nvme_change_ctrl_state(struct nvme_ctrl *ctrl,
>   		case NVME_CTRL_LIVE:
>   		case NVME_CTRL_RESETTING:
>   		case NVME_CTRL_CONNECTING:
> +		case NVME_CTRL_FW_ACTIVATION:
>   			changed = true;
>   			fallthrough;
>   		default:
> @@ -621,6 +624,15 @@ bool nvme_change_ctrl_state(struct nvme_ctrl *ctrl,
>   			break;
>   		}
>   		break;
> +	case NVME_CTRL_FW_ACTIVATION:
> +		switch (old_state) {
> +		case NVME_CTRL_LIVE:
> +			changed = true;
> +			fallthrough;
> +		default:
> +			break;
> +		}
> +		break;
>   	default:
>   		break;
>   	}
> @@ -4529,7 +4541,7 @@ static bool nvme_handle_aen_notice(struct nvme_ctrl *ctrl, u32 result)
>   		 * recovery actions from interfering with the controller's
>   		 * firmware activation.
>   		 */
> -		if (nvme_change_ctrl_state(ctrl, NVME_CTRL_RESETTING)) {
> +		if (nvme_change_ctrl_state(ctrl, NVME_CTRL_FW_ACTIVATION)) {
>   			requeue = false;
>   			queue_work(nvme_wq, &ctrl->fw_act_work);
>   		}
> diff --git a/drivers/nvme/host/nvme.h b/drivers/nvme/host/nvme.h
> index 51e078642127..3a383225afed 100644
> --- a/drivers/nvme/host/nvme.h
> +++ b/drivers/nvme/host/nvme.h
> @@ -247,6 +247,7 @@ static inline u16 nvme_req_qid(struct request *req)
>    *				shutdown or removal. In this case we forcibly
>    *				kill all inflight I/O as they have no chance to
>    *				complete
> + * @NVME_CTRL_FW_ACTIVATION:	Controller is in firmware activation state.
>    */
>   enum nvme_ctrl_state {
>   	NVME_CTRL_NEW,
> @@ -256,6 +257,7 @@ enum nvme_ctrl_state {
>   	NVME_CTRL_DELETING,
>   	NVME_CTRL_DELETING_NOIO,
>   	NVME_CTRL_DEAD,
> +	NVME_CTRL_FW_ACTIVATION,
>   };
> 
>   struct nvme_fault_inject {
> 
Indeed, something like it.

Cheers,

Hannes
-- 
Dr. Hannes Reinecke                  Kernel Storage Architect
hare@suse.de                                +49 911 74053 688
SUSE Software Solutions GmbH, Frankenstr. 146, 90461 Nürnberg
HRB 36809 (AG Nürnberg), GF: I. Totev, A. McDonald, W. Knoblich