[v3] qla2xxx: synchronize rport dev_loss_tmo setting

From: Hannes Reinecke <hare@suse.de>

Currently, the dev_loss_tmo setting is only ever used for SCSI
devices. This patch reshuffles initialisation such that the SCSI
remote ports are registered before the NVMe ones, allowing the
dev_loss_tmo setting to be synchronized between SCSI and NVMe.

Signed-off-by: Hannes Reinecke <hare@suse.de>
Signed-off-by: Daniel Wagner <dwagner@suse.de>
---
We have this patch in our kernels for a while and it works
fine.

v3:
 - added additional check if fcport pointer is valid in
   qla2x00_set_rport_loss_tmo (crash reported)

v2:
 - https://lore.kernel.org/linux-scsi/20210702092052.93202-1-dwagner@suse.de/
 - fixed build failure for !NVME_FC reported by lkp

v1:
 - https://lore.kernel.org/linux-scsi/20210609094956.11286-1-dwagner@suse.de/
 - initial version
 
 drivers/scsi/qla2xxx/qla_attr.c |  6 ++++++
 drivers/scsi/qla2xxx/qla_init.c | 10 +++-------
 drivers/scsi/qla2xxx/qla_nvme.c |  5 ++++-
 3 files changed, 13 insertions(+), 8 deletions(-)

> On Dec 14, 2021, at 5:11 AM, Daniel Wagner <dwagner@suse.de> wrote:
> 
> From: Hannes Reinecke <hare@suse.de>
> 
> Currently, the dev_loss_tmo setting is only ever used for SCSI
> devices. This patch reshuffles initialisation such that the SCSI
> remote ports are registered before the NVMe ones, allowing the
> dev_loss_tmo setting to be synchronized between SCSI and NVMe.
> 
> Signed-off-by: Hannes Reinecke <hare@suse.de>
> Signed-off-by: Daniel Wagner <dwagner@suse.de>
> ---
> We have this patch in our kernels for a while and it works
> fine.
> 
> v3:
> - added additional check if fcport pointer is valid in
>   qla2x00_set_rport_loss_tmo (crash reported)
> 
> v2:
> - https://lore.kernel.org/linux-scsi/20210702092052.93202-1-dwagner@suse.de/
> - fixed build failure for !NVME_FC reported by lkp
> 
> v1:
> - https://lore.kernel.org/linux-scsi/20210609094956.11286-1-dwagner@suse.de/
> - initial version
> 
> drivers/scsi/qla2xxx/qla_attr.c |  6 ++++++
> drivers/scsi/qla2xxx/qla_init.c | 10 +++-------
> drivers/scsi/qla2xxx/qla_nvme.c |  5 ++++-
> 3 files changed, 13 insertions(+), 8 deletions(-)
> 
> diff --git a/drivers/scsi/qla2xxx/qla_attr.c b/drivers/scsi/qla2xxx/qla_attr.c
> index 032efb294ee5..db55737000ab 100644
> --- a/drivers/scsi/qla2xxx/qla_attr.c
> +++ b/drivers/scsi/qla2xxx/qla_attr.c
> @@ -2700,7 +2700,13 @@ qla2x00_get_starget_port_id(struct scsi_target *starget)
> static inline void
> qla2x00_set_rport_loss_tmo(struct fc_rport *rport, uint32_t timeout)
> {
> +	fc_port_t *fcport = *(fc_port_t **)rport->dd_data;
> +
> 	rport->dev_loss_tmo = timeout ? timeout : 1;
> +
> +	if (IS_ENABLED(CONFIG_NVME_FC) && fcport && fcport->nvme_remote_port)
> +		nvme_fc_set_remoteport_devloss(fcport->nvme_remote_port,
> +					       rport->dev_loss_tmo);
> }
> 
> static void
> diff --git a/drivers/scsi/qla2xxx/qla_init.c b/drivers/scsi/qla2xxx/qla_init.c
> index 070b636802d0..1fe4966fc2f6 100644
> --- a/drivers/scsi/qla2xxx/qla_init.c
> +++ b/drivers/scsi/qla2xxx/qla_init.c
> @@ -5828,13 +5828,6 @@ qla2x00_update_fcport(scsi_qla_host_t *vha, fc_port_t *fcport)
> 
> 	qla2x00_dfs_create_rport(vha, fcport);
> 
> -	if (NVME_TARGET(vha->hw, fcport)) {
> -		qla_nvme_register_remote(vha, fcport);
> -		qla2x00_set_fcport_disc_state(fcport, DSC_LOGIN_COMPLETE);
> -		qla2x00_set_fcport_state(fcport, FCS_ONLINE);
> -		return;
> -	}
> -
> 	qla24xx_update_fcport_fcp_prio(vha, fcport);
> 
> 	switch (vha->host->active_mode) {
> @@ -5856,6 +5849,9 @@ qla2x00_update_fcport(scsi_qla_host_t *vha, fc_port_t *fcport)
> 		break;
> 	}
> 
> +	if (NVME_TARGET(vha->hw, fcport))
> +		qla_nvme_register_remote(vha, fcport);
> +
> 	qla2x00_set_fcport_state(fcport, FCS_ONLINE);
> 
> 	if (IS_IIDMA_CAPABLE(vha->hw) && vha->hw->flags.gpsc_supported) {
> diff --git a/drivers/scsi/qla2xxx/qla_nvme.c b/drivers/scsi/qla2xxx/qla_nvme.c
> index 138ffdb5c92c..e22ec7cb65db 100644
> --- a/drivers/scsi/qla2xxx/qla_nvme.c
> +++ b/drivers/scsi/qla2xxx/qla_nvme.c
> @@ -43,7 +43,7 @@ int qla_nvme_register_remote(struct scsi_qla_host *vha, struct fc_port *fcport)
> 	req.port_name = wwn_to_u64(fcport->port_name);
> 	req.node_name = wwn_to_u64(fcport->node_name);
> 	req.port_role = 0;
> -	req.dev_loss_tmo = 0;
> +	req.dev_loss_tmo = fcport->dev_loss_tmo;
> 
> 	if (fcport->nvme_prli_service_param & NVME_PRLI_SP_INITIATOR)
> 		req.port_role = FC_PORT_ROLE_NVME_INITIATOR;
> @@ -70,6 +70,9 @@ int qla_nvme_register_remote(struct scsi_qla_host *vha, struct fc_port *fcport)
> 		return ret;
> 	}
> 
> +	nvme_fc_set_remoteport_devloss(fcport->nvme_remote_port,
> +				       fcport->dev_loss_tmo);
> +
> 	if (fcport->nvme_prli_service_param & NVME_PRLI_SP_SLER)
> 		ql_log(ql_log_info, vha, 0x212a,
> 		       "PortID:%06x Supports SLER\n", req.port_id);
> -- 
> 2.29.2
> 

Looks Good.

Reviewed-by: Himanshu Madhani <himanshu.madhani@oracle.com>

--
Himanshu Madhani	 Oracle Linux Engineering

Daniel,

> Currently, the dev_loss_tmo setting is only ever used for SCSI
> devices. This patch reshuffles initialisation such that the SCSI
> remote ports are registered before the NVMe ones, allowing the
> dev_loss_tmo setting to be synchronized between SCSI and NVMe.

Applied to 5.17/scsi-staging, thanks!

On Tue, 14 Dec 2021 12:11:39 +0100, Daniel Wagner wrote:

> From: Hannes Reinecke <hare@suse.de>
> 
> Currently, the dev_loss_tmo setting is only ever used for SCSI
> devices. This patch reshuffles initialisation such that the SCSI
> remote ports are registered before the NVMe ones, allowing the
> dev_loss_tmo setting to be synchronized between SCSI and NVMe.
> 
> [...]

Applied to 5.17/scsi-queue, thanks!

[1/1] qla2xxx: synchronize rport dev_loss_tmo setting
      https://git.kernel.org/mkp/scsi/c/baea0e833f76