[PATCH v2 00/13] qla2xxx driver fixes

[PATCH v2 00/13] qla2xxx driver fixes

[Nilesh Javali]

Martin,

Please apply the qla2xxx driver misc bug fixes to the scsi tree
at your earliest convenience.

v2:
- Incorporate all minor review comments
- Add Fixes tag to 11/13
- Add Reviewed-by tag

Thanks,
Nilesh

Arun Easi (3):
  qla2xxx: Fix loss of NVME namespaces after driver reload test
  qla2xxx: Fix missed DMA unmap for NVME ls requests
  qla2xxx: Fix crash during module load unload test

Manish Rangankar (1):
  qla2xxx: Use correct feature type field during rffid processing

Nilesh Javali (1):
  qla2xxx: Update version to 10.02.07.400-k

Quinn Tran (7):
  qla2xxx: Fix incorrect reporting of task management failure
  qla2xxx: Fix disk failure to rediscover
  qla2xxx: fix n2n inconsistent plogi
  qla2xxx: Fix hang due to session stuck
  qla2xxx: Fix laggy FC remote port session recovery
  qla2xxx: reduce false trigger to login
  qla2xxx: Fix stuck session of prli reject

Shreyas Deodhar (1):
  qla2xxx: Increase max limit of ql2xnvme_queues

 drivers/scsi/qla2xxx/qla_def.h     |  5 +++
 drivers/scsi/qla2xxx/qla_gs.c      |  5 +--
 drivers/scsi/qla2xxx/qla_init.c    | 36 ++++++++++------
 drivers/scsi/qla2xxx/qla_iocb.c    |  8 ++--
 drivers/scsi/qla2xxx/qla_isr.c     |  1 +
 drivers/scsi/qla2xxx/qla_nvme.c    | 67 +++++++++++++++++++++++-------
 drivers/scsi/qla2xxx/qla_nvme.h    |  1 -
 drivers/scsi/qla2xxx/qla_os.c      | 23 ++++++++--
 drivers/scsi/qla2xxx/qla_version.h |  4 +-
 9 files changed, 107 insertions(+), 43 deletions(-)


base-commit: ac2beb4e3bd75b0049068516b9d42201bda0ded3
-- 
2.19.0.rc0

[PATCH v2 01/13] qla2xxx: Fix incorrect reporting of task management failure

[Nilesh Javali]

From: Quinn Tran <qutran@marvell.com>

User experienced no task management error while target device
is responding with error. The RSP_CODE field in the status
iocb is in little endian. Driver assumes it's big endian,
where it picked up erroneous data.

Convert the data back to big endian as is on the wire,
where current code will pick up correct status.

Cc: stable@vger.kernel.org
Fixes: faef62d13463 ("[SCSI] qla2xxx: Fix Task Management command asynchronous handling")
Signed-off-by: Quinn Tran <qutran@marvell.com>
Signed-off-by: Nilesh Javali <njavali@marvell.com>
Reviewed-by: Himanshu Madhani <himanshu.madhani@oracle.com>
---
 drivers/scsi/qla2xxx/qla_isr.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/drivers/scsi/qla2xxx/qla_isr.c b/drivers/scsi/qla2xxx/qla_isr.c
index aaf6504570fd..198b782d7790 100644
--- a/drivers/scsi/qla2xxx/qla_isr.c
+++ b/drivers/scsi/qla2xxx/qla_isr.c
@@ -2498,6 +2498,7 @@ qla24xx_tm_iocb_entry(scsi_qla_host_t *vha, struct req_que *req, void *tsk)
 		iocb->u.tmf.data = QLA_FUNCTION_FAILED;
 	} else if ((le16_to_cpu(sts->scsi_status) &
 	    SS_RESPONSE_INFO_LEN_VALID)) {
+		host_to_fcp_swap(sts->data, sizeof(sts->data));
 		if (le32_to_cpu(sts->rsp_data_len) < 4) {
 			ql_log(ql_log_warn, fcport->vha, 0x503b,
 			    "Async-%s error - hdl=%x not enough response(%d).\n",
-- 
2.19.0.rc0

[PATCH v2 02/13] qla2xxx: Fix disk failure to rediscover

[Nilesh Javali]

From: Quinn Tran <qutran@marvell.com>

User experience some of the LUN failed to rediscovered after
long cable pull test. The issue is triggered by a race
condition between driver setting session online state vs
starting the LUN scan process at the same time. Current code
set the online state after notifying the session is
available. In this case, trigger to start the LUN scan process happened
before driver could set the session in online state.
LUN scan ends up with failure due to the session
online check was failing.

Set the online state before reporting of the availability of the session.

Cc: stable@vger.kernel.org
Fixes: aecf043443d3 ("scsi: qla2xxx: Fix Remote port registration")
Signed-off-by: Quinn Tran <qutran@marvell.com>
Signed-off-by: Nilesh Javali <njavali@marvell.com>
Reviewed-by: Himanshu Madhani <himanshu.madhani@oracle.com>
---
 drivers/scsi/qla2xxx/qla_init.c | 5 +++--
 drivers/scsi/qla2xxx/qla_nvme.c | 5 +++++
 2 files changed, 8 insertions(+), 2 deletions(-)

diff --git a/drivers/scsi/qla2xxx/qla_init.c b/drivers/scsi/qla2xxx/qla_init.c
index 835ed4179887..6ffe44b805b6 100644
--- a/drivers/scsi/qla2xxx/qla_init.c
+++ b/drivers/scsi/qla2xxx/qla_init.c
@@ -5758,6 +5758,8 @@ qla2x00_reg_remote_port(scsi_qla_host_t *vha, fc_port_t *fcport)
 	if (atomic_read(&fcport->state) == FCS_ONLINE)
 		return;
 
+	qla2x00_set_fcport_state(fcport, FCS_ONLINE);
+
 	rport_ids.node_name = wwn_to_u64(fcport->node_name);
 	rport_ids.port_name = wwn_to_u64(fcport->port_name);
 	rport_ids.port_id = fcport->d_id.b.domain << 16 |
@@ -5858,6 +5860,7 @@ qla2x00_update_fcport(scsi_qla_host_t *vha, fc_port_t *fcport)
 		qla2x00_reg_remote_port(vha, fcport);
 		break;
 	case MODE_TARGET:
+		qla2x00_set_fcport_state(fcport, FCS_ONLINE);
 		if (!vha->vha_tgt.qla_tgt->tgt_stop &&
 			!vha->vha_tgt.qla_tgt->tgt_stopped)
 			qlt_fc_port_added(vha, fcport);
@@ -5875,8 +5878,6 @@ qla2x00_update_fcport(scsi_qla_host_t *vha, fc_port_t *fcport)
 	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) {
 		if (fcport->id_changed) {
 			fcport->id_changed = 0;
diff --git a/drivers/scsi/qla2xxx/qla_nvme.c b/drivers/scsi/qla2xxx/qla_nvme.c
index 718c761ff5f8..5723082d94d6 100644
--- a/drivers/scsi/qla2xxx/qla_nvme.c
+++ b/drivers/scsi/qla2xxx/qla_nvme.c
@@ -37,6 +37,11 @@ int qla_nvme_register_remote(struct scsi_qla_host *vha, struct fc_port *fcport)
 		(fcport->nvme_flag & NVME_FLAG_REGISTERED))
 		return 0;
 
+	if (atomic_read(&fcport->state) == FCS_ONLINE)
+		return 0;
+
+	qla2x00_set_fcport_state(fcport, FCS_ONLINE);
+
 	fcport->nvme_flag &= ~NVME_FLAG_RESETTING;
 
 	memset(&req, 0, sizeof(struct nvme_fc_port_info));
-- 
2.19.0.rc0

[PATCH v2 03/13] qla2xxx: Fix loss of NVME namespaces after driver reload test

[Nilesh Javali]

From: Arun Easi <aeasi@marvell.com>

Driver registration of localport can race when it
happens at the remote port discovery time. Fix
this by calling the registration under a mutex.

Reported-by: Marco Patalano <mpatalan@redhat.com>
Tested-by: Marco Patalano <mpatalan@redhat.com>
Cc: stable@vger.kernel.org
Fixes: e84067d74301 ("scsi: qla2xxx: Add FC-NVMe F/W initialization and transport registration")
Signed-off-by: Arun Easi <aeasi@marvell.com>
Signed-off-by: Nilesh Javali <njavali@marvell.com>
Reviewed-by: Himanshu Madhani <himanshu.madhani@oracle.com>
---
 drivers/scsi/qla2xxx/qla_nvme.c | 30 ++++++++++++++++++++----------
 1 file changed, 20 insertions(+), 10 deletions(-)

diff --git a/drivers/scsi/qla2xxx/qla_nvme.c b/drivers/scsi/qla2xxx/qla_nvme.c
index 5723082d94d6..3bf5cbd754a7 100644
--- a/drivers/scsi/qla2xxx/qla_nvme.c
+++ b/drivers/scsi/qla2xxx/qla_nvme.c
@@ -782,8 +782,6 @@ int qla_nvme_register_hba(struct scsi_qla_host *vha)
 	ha = vha->hw;
 	tmpl = &qla_nvme_fc_transport;
 
-	WARN_ON(vha->nvme_local_port);
-
 	if (ql2xnvme_queues < MIN_NVME_HW_QUEUES || ql2xnvme_queues > MAX_NVME_HW_QUEUES) {
 		ql_log(ql_log_warn, vha, 0xfffd,
 		    "ql2xnvme_queues=%d is out of range(MIN:%d - MAX:%d). Resetting ql2xnvme_queues to:%d\n",
@@ -797,7 +795,7 @@ int qla_nvme_register_hba(struct scsi_qla_host *vha)
 		(uint8_t)(ha->max_qpairs ? ha->max_qpairs : 1));
 
 	ql_log(ql_log_info, vha, 0xfffb,
-	    "Number of NVME queues used for this port: %d\n",
+	       "Number of NVME queues used for this port: %d\n",
 	    qla_nvme_fc_transport.max_hw_queues);
 
 	pinfo.node_name = wwn_to_u64(vha->node_name);
@@ -805,13 +803,25 @@ int qla_nvme_register_hba(struct scsi_qla_host *vha)
 	pinfo.port_role = FC_PORT_ROLE_NVME_INITIATOR;
 	pinfo.port_id = vha->d_id.b24;
 
-	ql_log(ql_log_info, vha, 0xffff,
-	    "register_localport: host-traddr=nn-0x%llx:pn-0x%llx on portID:%x\n",
-	    pinfo.node_name, pinfo.port_name, pinfo.port_id);
-	qla_nvme_fc_transport.dma_boundary = vha->host->dma_boundary;
-
-	ret = nvme_fc_register_localport(&pinfo, tmpl,
-	    get_device(&ha->pdev->dev), &vha->nvme_local_port);
+	mutex_lock(&ha->vport_lock);
+	/*
+	 * Check again for nvme_local_port to see if any other thread raced
+	 * with this one and finished registration.
+	 */
+	if (!vha->nvme_local_port) {
+		ql_log(ql_log_info, vha, 0xffff,
+		    "register_localport: host-traddr=nn-0x%llx:pn-0x%llx on portID:%x\n",
+		    pinfo.node_name, pinfo.port_name, pinfo.port_id);
+		qla_nvme_fc_transport.dma_boundary = vha->host->dma_boundary;
+
+		ret = nvme_fc_register_localport(&pinfo, tmpl,
+						 get_device(&ha->pdev->dev),
+						 &vha->nvme_local_port);
+		mutex_unlock(&ha->vport_lock);
+	} else {
+		mutex_unlock(&ha->vport_lock);
+		return 0;
+	}
 	if (ret) {
 		ql_log(ql_log_warn, vha, 0xffff,
 		    "register_localport failed: ret=%x\n", ret);
-- 
2.19.0.rc0

[PATCH v2 04/13] qla2xxx: Fix missed DMA unmap for NVME ls requests

[Nilesh Javali]

From: Arun Easi <aeasi@marvell.com>

At NVME ELS request time, request structure is DMA mapped
and never unmapped. Fix this by calling the unmap on
ELS completion.

Cc: stable@vger.kernel.org
Fixes: e84067d74301 ("scsi: qla2xxx: Add FC-NVMe F/W initialization and transport registration")
Signed-off-by: Arun Easi <aeasi@marvell.com>
Signed-off-by: Nilesh Javali <njavali@marvell.com>
Reviewed-by: Himanshu Madhani <himanshu.madhani@oracle.com>
---
 drivers/scsi/qla2xxx/qla_nvme.c | 17 +++++++++++++++++
 1 file changed, 17 insertions(+)

diff --git a/drivers/scsi/qla2xxx/qla_nvme.c b/drivers/scsi/qla2xxx/qla_nvme.c
index 3bf5cbd754a7..794a95b2e3b4 100644
--- a/drivers/scsi/qla2xxx/qla_nvme.c
+++ b/drivers/scsi/qla2xxx/qla_nvme.c
@@ -175,6 +175,18 @@ static void qla_nvme_release_fcp_cmd_kref(struct kref *kref)
 	qla2xxx_rel_qpair_sp(sp->qpair, sp);
 }
 
+static void qla_nvme_ls_unmap(struct srb *sp, struct nvmefc_ls_req *fd)
+{
+	if (sp->flags & SRB_DMA_VALID) {
+		struct srb_iocb *nvme = &sp->u.iocb_cmd;
+		struct qla_hw_data *ha = sp->fcport->vha->hw;
+
+		dma_unmap_single(&ha->pdev->dev, nvme->u.nvme.cmd_dma,
+				 fd->rqstlen, DMA_TO_DEVICE);
+		sp->flags &= ~SRB_DMA_VALID;
+	}
+}
+
 static void qla_nvme_release_ls_cmd_kref(struct kref *kref)
 {
 	struct srb *sp = container_of(kref, struct srb, cmd_kref);
@@ -191,6 +203,8 @@ static void qla_nvme_release_ls_cmd_kref(struct kref *kref)
 	spin_unlock_irqrestore(&priv->cmd_lock, flags);
 
 	fd = priv->fd;
+
+	qla_nvme_ls_unmap(sp, fd);
 	fd->done(fd, priv->comp_status);
 out:
 	qla2x00_rel_sp(sp);
@@ -361,6 +375,8 @@ static int qla_nvme_ls_req(struct nvme_fc_local_port *lport,
 	dma_sync_single_for_device(&ha->pdev->dev, nvme->u.nvme.cmd_dma,
 	    fd->rqstlen, DMA_TO_DEVICE);
 
+	sp->flags |= SRB_DMA_VALID;
+
 	rval = qla2x00_start_sp(sp);
 	if (rval != QLA_SUCCESS) {
 		ql_log(ql_log_warn, vha, 0x700e,
@@ -368,6 +384,7 @@ static int qla_nvme_ls_req(struct nvme_fc_local_port *lport,
 		wake_up(&sp->nvme_ls_waitq);
 		sp->priv = NULL;
 		priv->sp = NULL;
+		qla_nvme_ls_unmap(sp, fd);
 		qla2x00_rel_sp(sp);
 		return rval;
 	}
-- 
2.19.0.rc0

[PATCH v2 05/13] qla2xxx: Fix crash during module load unload test

[Nilesh Javali]

From: Arun Easi <aeasi@marvell.com>

During purex packet handling driver was incorrectly
freeing a pre-allocated structure. Fix this by
skipping that entry.

System crashed with the following stack during a
module unload test.

Call Trace:
          sbitmap_init_node+0x7f/0x1e0
          sbitmap_queue_init_node+0x24/0x150
          blk_mq_init_bitmaps+0x3d/0xa0
          blk_mq_init_tags+0x68/0x90
          blk_mq_alloc_map_and_rqs+0x44/0x120
          blk_mq_alloc_set_map_and_rqs+0x63/0x150
          blk_mq_alloc_tag_set+0x11b/0x230
          scsi_add_host_with_dma.cold+0x3f/0x245
          qla2x00_probe_one+0xd5a/0x1b80 [qla2xxx]

Call Trace with slub_debug and debug kernel:
        kasan_report_invalid_free+0x50/0x80
        __kasan_slab_free+0x137/0x150
        slab_free_freelist_hook+0xc6/0x190
        kfree+0xe8/0x2e0
        qla2x00_free_device+0x3bb/0x5d0 [qla2xxx]
        qla2x00_remove_one+0x668/0xcf0 [qla2xxx]

Reported-by: Marco Patalano <mpatalan@redhat.com>
Tested-by: Marco Patalano <mpatalan@redhat.com>
Cc: stable@vger.kernel.org
Fixes: 62e9dd177732 ("scsi: qla2xxx: Change in PUREX to handle FPIN ELS requests")
Signed-off-by: Arun Easi <aeasi@marvell.com>
Signed-off-by: Nilesh Javali <njavali@marvell.com>
Reviewed-by: Himanshu Madhani <himanshu.madhani@oracle.com>
---
 drivers/scsi/qla2xxx/qla_os.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/drivers/scsi/qla2xxx/qla_os.c b/drivers/scsi/qla2xxx/qla_os.c
index a4546346c18b..d572a76d0fa0 100644
--- a/drivers/scsi/qla2xxx/qla_os.c
+++ b/drivers/scsi/qla2xxx/qla_os.c
@@ -3904,6 +3904,8 @@ qla24xx_free_purex_list(struct purex_list *list)
 	spin_lock_irqsave(&list->lock, flags);
 	list_for_each_entry_safe(item, next, &list->head, list) {
 		list_del(&item->list);
+		if (item == &item->vha->default_item)
+			continue;
 		kfree(item);
 	}
 	spin_unlock_irqrestore(&list->lock, flags);
-- 
2.19.0.rc0

[PATCH v2 06/13] qla2xxx: fix n2n inconsistent plogi

[Nilesh Javali]

From: Quinn Tran <qutran@marvell.com>

For N2N topology, ELS Passthrough is used to send PLOGI.
On failure of ELS pass through PLOGI, driver flipped over
to using LLIOCB PLOGI for N2N. This is not consistent.
This patch would delete the session to restart the
connection, where ELS pass through PLOGI would be used consistently.

Cc: stable@vger.kernel.org
Fixes: c76ae845ea83 ("scsi: qla2xxx: Add error handling for PLOGI ELS passthrough")
Signed-off-by: Quinn Tran <qutran@marvell.com>
Signed-off-by: Nilesh Javali <njavali@marvell.com>
Reviewed-by: Himanshu Madhani <himanshu.madhani@oracle.com>
---
 drivers/scsi/qla2xxx/qla_iocb.c | 8 +++-----
 1 file changed, 3 insertions(+), 5 deletions(-)

diff --git a/drivers/scsi/qla2xxx/qla_iocb.c b/drivers/scsi/qla2xxx/qla_iocb.c
index 5e3ee1f7b43c..e0fe9ddb4bd2 100644
--- a/drivers/scsi/qla2xxx/qla_iocb.c
+++ b/drivers/scsi/qla2xxx/qla_iocb.c
@@ -2958,6 +2958,7 @@ static void qla2x00_els_dcmd2_sp_done(srb_t *sp, int res)
 					set_bit(ISP_ABORT_NEEDED,
 					    &vha->dpc_flags);
 					qla2xxx_wake_dpc(vha);
+					break;
 				}
 				fallthrough;
 			default:
@@ -2967,9 +2968,7 @@ static void qla2x00_els_dcmd2_sp_done(srb_t *sp, int res)
 				    fw_status[0], fw_status[1], fw_status[2]);
 
 				fcport->flags &= ~FCF_ASYNC_SENT;
-				qla2x00_set_fcport_disc_state(fcport,
-				    DSC_LOGIN_FAILED);
-				set_bit(RELOGIN_NEEDED, &vha->dpc_flags);
+				qlt_schedule_sess_for_deletion(fcport);
 				break;
 			}
 			break;
@@ -2981,8 +2980,7 @@ static void qla2x00_els_dcmd2_sp_done(srb_t *sp, int res)
 			    fw_status[0], fw_status[1], fw_status[2]);
 
 			sp->fcport->flags &= ~FCF_ASYNC_SENT;
-			qla2x00_set_fcport_disc_state(fcport, DSC_LOGIN_FAILED);
-			set_bit(RELOGIN_NEEDED, &vha->dpc_flags);
+			qlt_schedule_sess_for_deletion(fcport);
 			break;
 		}
 
-- 
2.19.0.rc0

[PATCH v2 07/13] qla2xxx: Fix hang due to session stuck

[Nilesh Javali]

From: Quinn Tran <qutran@marvell.com>

User experience device lost. The log shows Get port data base
command was queued up, failed, and requeued again. Every time
it is requeued, it set the FCF_ASYNC_ACTIVE. This prevents any
recovery code from occurring because driver thinks a recovery is in
progress for this session. In essence, this session is hung.
The reason it gets into this place is the session deletion got
in front of this call due to link perturbation.

Break the requeue cycle and exit.
The session deletion code will trigger a session relogin.

Cc: stable@vger.kernel.org
Fixes: 726b85487067 ("qla2xxx: Add framework for async fabric discovery")
Signed-off-by: Quinn Tran <qutran@marvell.com>
Signed-off-by: Nilesh Javali <njavali@marvell.com>
Reviewed-by: Himanshu Madhani <himanshu.madhani@oracle.com>
---
 drivers/scsi/qla2xxx/qla_def.h  |  4 ++++
 drivers/scsi/qla2xxx/qla_init.c | 19 +++++++++++++++++--
 2 files changed, 21 insertions(+), 2 deletions(-)

diff --git a/drivers/scsi/qla2xxx/qla_def.h b/drivers/scsi/qla2xxx/qla_def.h
index 47d7fa1c7ae8..25b884701267 100644
--- a/drivers/scsi/qla2xxx/qla_def.h
+++ b/drivers/scsi/qla2xxx/qla_def.h
@@ -5437,4 +5437,8 @@ struct ql_vnd_tgt_stats_resp {
 #include "qla_gbl.h"
 #include "qla_dbg.h"
 #include "qla_inline.h"
+
+#define IS_SESSION_DELETED(_fcport) (_fcport->disc_state == DSC_DELETE_PEND || \
+				      _fcport->disc_state == DSC_DELETED)
+
 #endif
diff --git a/drivers/scsi/qla2xxx/qla_init.c b/drivers/scsi/qla2xxx/qla_init.c
index 6ffe44b805b6..3f05e87ac2d6 100644
--- a/drivers/scsi/qla2xxx/qla_init.c
+++ b/drivers/scsi/qla2xxx/qla_init.c
@@ -575,6 +575,14 @@ qla2x00_async_adisc(struct scsi_qla_host *vha, fc_port_t *fcport,
 	struct srb_iocb *lio;
 	int rval = QLA_FUNCTION_FAILED;
 
+	if (IS_SESSION_DELETED(fcport)) {
+		ql_log(ql_log_warn, vha, 0xffff,
+		       "%s: %8phC is being delete - not sending command.\n",
+		       __func__, fcport->port_name);
+		fcport->flags &= ~FCF_ASYNC_ACTIVE;
+		return rval;
+	}
+
 	if (!vha->flags.online || (fcport->flags & FCF_ASYNC_SENT))
 		return rval;
 
@@ -1338,8 +1346,15 @@ int qla24xx_async_gpdb(struct scsi_qla_host *vha, fc_port_t *fcport, u8 opt)
 	struct port_database_24xx *pd;
 	struct qla_hw_data *ha = vha->hw;
 
-	if (!vha->flags.online || (fcport->flags & FCF_ASYNC_SENT) ||
-	    fcport->loop_id == FC_NO_LOOP_ID) {
+	if (IS_SESSION_DELETED(fcport)) {
+		ql_log(ql_log_warn, vha, 0xffff,
+		       "%s: %8phC is being delete - not sending command.\n",
+		       __func__, fcport->port_name);
+		fcport->flags &= ~FCF_ASYNC_ACTIVE;
+		return rval;
+	}
+
+	if (!vha->flags.online || fcport->flags & FCF_ASYNC_SENT) {
 		ql_log(ql_log_warn, vha, 0xffff,
 		    "%s: %8phC online %d flags %x - not sending command.\n",
 		    __func__, fcport->port_name, vha->flags.online, fcport->flags);
-- 
2.19.0.rc0

[PATCH v2 08/13] qla2xxx: Fix laggy FC remote port session recovery

[Nilesh Javali]

From: Quinn Tran <qutran@marvell.com>

For session recovery, driver relies on the dpc thread to
initiate certain operation. The dpc thread runs exclusively
without the Mailbox interface being occupied. Recent code change
for heartbeat check via mailbox cmd 0 is causing the dpc thread
from carrying out its operation. This patch allows the higher
priority error recovery to run first before running the lower priority
heartbeat check.

Cc: stable@vger.kernel.org
Fixes: d94d8158e184 ("scsi: qla2xxx: Add heartbeat check")
Signed-off-by: Quinn Tran <qutran@marvell.com>
Signed-off-by: Nilesh Javali <njavali@marvell.com>
Reviewed-by: Himanshu Madhani <himanshu.madhani@oracle.com>
---
 drivers/scsi/qla2xxx/qla_def.h |  1 +
 drivers/scsi/qla2xxx/qla_os.c  | 20 +++++++++++++++++---
 2 files changed, 18 insertions(+), 3 deletions(-)

diff --git a/drivers/scsi/qla2xxx/qla_def.h b/drivers/scsi/qla2xxx/qla_def.h
index 25b884701267..2de5920126ca 100644
--- a/drivers/scsi/qla2xxx/qla_def.h
+++ b/drivers/scsi/qla2xxx/qla_def.h
@@ -4621,6 +4621,7 @@ struct qla_hw_data {
 	struct workqueue_struct *wq;
 	struct work_struct heartbeat_work;
 	struct qlfc_fw fw_buf;
+	unsigned long last_heartbeat_run_jiffies;
 
 	/* FCP_CMND priority support */
 	struct qla_fcp_prio_cfg *fcp_prio_cfg;
diff --git a/drivers/scsi/qla2xxx/qla_os.c b/drivers/scsi/qla2xxx/qla_os.c
index d572a76d0fa0..89c7ac36a41a 100644
--- a/drivers/scsi/qla2xxx/qla_os.c
+++ b/drivers/scsi/qla2xxx/qla_os.c
@@ -7218,7 +7218,7 @@ static bool qla_do_heartbeat(struct scsi_qla_host *vha)
 	return do_heartbeat;
 }
 
-static void qla_heart_beat(struct scsi_qla_host *vha)
+static void qla_heart_beat(struct scsi_qla_host *vha, u16 dpc_started)
 {
 	struct qla_hw_data *ha = vha->hw;
 
@@ -7228,8 +7228,19 @@ static void qla_heart_beat(struct scsi_qla_host *vha)
 	if (vha->hw->flags.eeh_busy || qla2x00_chip_is_down(vha))
 		return;
 
-	if (qla_do_heartbeat(vha))
+	/*
+	 * dpc thread cannot run if heartbeat is running at the same time.
+	 * We also do not want to starve heartbeat task. Therefore, do
+	 * heartbeat task at least once every 5 seconds.
+	 */
+	if (dpc_started &&
+	    time_before(jiffies, ha->last_heartbeat_run_jiffies + 5 * HZ))
+		return;
+
+	if (qla_do_heartbeat(vha)) {
+		ha->last_heartbeat_run_jiffies = jiffies;
 		queue_work(ha->wq, &ha->heartbeat_work);
+	}
 }
 
 /**************************************************************************
@@ -7420,6 +7431,8 @@ qla2x00_timer(struct timer_list *t)
 		start_dpc++;
 	}
 
+	/* borrowing w to signify dpc will run */
+	w = 0;
 	/* Schedule the DPC routine if needed */
 	if ((test_bit(ISP_ABORT_NEEDED, &vha->dpc_flags) ||
 	    test_bit(LOOP_RESYNC_NEEDED, &vha->dpc_flags) ||
@@ -7452,9 +7465,10 @@ qla2x00_timer(struct timer_list *t)
 		    test_bit(RELOGIN_NEEDED, &vha->dpc_flags),
 		    test_bit(PROCESS_PUREX_IOCB, &vha->dpc_flags));
 		qla2xxx_wake_dpc(vha);
+		w = 1;
 	}
 
-	qla_heart_beat(vha);
+	qla_heart_beat(vha, w);
 
 	qla2x00_restart_timer(vha, WATCH_INTERVAL);
 }
-- 
2.19.0.rc0

[PATCH v2 09/13] qla2xxx: reduce false trigger to login

[Nilesh Javali]

From: Quinn Tran <qutran@marvell.com>

While a session is in the middle of a relogin,
a late RSCN can be delivered from switch. RSCN trigger fabric
scan where the scan logic can trigger another session login
while a login is in progress.
This patch reduce the extra trigger to prevent multiple login
to the same session.

Cc: stable@vger.kernel.org
Fixes: bee8b84686c4 ("scsi: qla2xxx: Reduce redundant ADISC command for RSCNs")
Signed-off-by: Quinn Tran <qutran@marvell.com>
Signed-off-by: Nilesh Javali <njavali@marvell.com>
Reviewed-by: Himanshu Madhani <himanshu.madhani@oracle.com>
---
 drivers/scsi/qla2xxx/qla_init.c | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)

diff --git a/drivers/scsi/qla2xxx/qla_init.c b/drivers/scsi/qla2xxx/qla_init.c
index 3f05e87ac2d6..a8c27cccf4d6 100644
--- a/drivers/scsi/qla2xxx/qla_init.c
+++ b/drivers/scsi/qla2xxx/qla_init.c
@@ -1644,7 +1644,8 @@ int qla24xx_fcport_handle_login(struct scsi_qla_host *vha, fc_port_t *fcport)
 	    fcport->login_gen, fcport->loop_id, fcport->scan_state,
 	    fcport->fc4_type);
 
-	if (fcport->scan_state != QLA_FCPORT_FOUND)
+	if (fcport->scan_state != QLA_FCPORT_FOUND ||
+	    fcport->disc_state == DSC_DELETE_PEND)
 		return 0;
 
 	if ((fcport->loop_id != FC_NO_LOOP_ID) &&
@@ -1665,7 +1666,7 @@ int qla24xx_fcport_handle_login(struct scsi_qla_host *vha, fc_port_t *fcport)
 	if (vha->host->active_mode == MODE_TARGET && !N2N_TOPO(vha->hw))
 		return 0;
 
-	if (fcport->flags & FCF_ASYNC_SENT) {
+	if (fcport->flags & (FCF_ASYNC_SENT | FCF_ASYNC_ACTIVE)) {
 		set_bit(RELOGIN_NEEDED, &vha->dpc_flags);
 		return 0;
 	}
-- 
2.19.0.rc0

[PATCH v2 10/13] qla2xxx: Fix stuck session of prli reject

[Nilesh Javali]

From: Quinn Tran <qutran@marvell.com>

Remove stale recovery code that prevents normal path recovery.

Cc: stable@vger.kernel.org
Fixes: 1cbc0efcd9be ("scsi: qla2xxx: Fix retry for PRLI RJT with reason of BUSY")
Signed-off-by: Quinn Tran <qutran@marvell.com>
Signed-off-by: Nilesh Javali <njavali@marvell.com>
Reviewed-by: Himanshu Madhani <himanshu.madhani@oracle.com>
---
 drivers/scsi/qla2xxx/qla_init.c | 7 -------
 1 file changed, 7 deletions(-)

diff --git a/drivers/scsi/qla2xxx/qla_init.c b/drivers/scsi/qla2xxx/qla_init.c
index a8c27cccf4d6..7f81525c4fb3 100644
--- a/drivers/scsi/qla2xxx/qla_init.c
+++ b/drivers/scsi/qla2xxx/qla_init.c
@@ -2105,13 +2105,6 @@ qla24xx_handle_prli_done_event(struct scsi_qla_host *vha, struct event_arg *ea)
 		qla24xx_post_gpdb_work(vha, ea->fcport, 0);
 		break;
 	default:
-		if ((ea->iop[0] == LSC_SCODE_ELS_REJECT) &&
-		    (ea->iop[1] == 0x50000)) {   /* reson 5=busy expl:0x0 */
-			set_bit(RELOGIN_NEEDED, &vha->dpc_flags);
-			ea->fcport->fw_login_state = DSC_LS_PLOGI_COMP;
-			break;
-		}
-
 		sp = ea->sp;
 		ql_dbg(ql_dbg_disc, vha, 0x2118,
 		       "%s %d %8phC priority %s, fc4type %x prev try %s\n",
-- 
2.19.0.rc0

[PATCH v2 11/13] qla2xxx: Use correct feature type field during rffid processing

[Nilesh Javali]

From: Manish Rangankar <mrangankar@marvell.com>

During SNS Register FC-4 Features (RFF_ID) for initiator driver was
sending incorrect type field for nvme supported device. Use correct
feature type field.

Cc: stable@vger.kernel.org
Fixes: e374f9f59281 ("scsi: qla2xxx: Migrate switch registration commands away from mailbox interface")
Signed-off-by: Manish Rangankar <mrangankar@marvell.com>
Signed-off-by: Nilesh Javali <njavali@marvell.com>
Reviewed-by: Himanshu Madhani <himanshu.madhani@oracle.com>
---
 drivers/scsi/qla2xxx/qla_gs.c | 5 ++---
 1 file changed, 2 insertions(+), 3 deletions(-)

diff --git a/drivers/scsi/qla2xxx/qla_gs.c b/drivers/scsi/qla2xxx/qla_gs.c
index a812f4a45232..6b67bd561810 100644
--- a/drivers/scsi/qla2xxx/qla_gs.c
+++ b/drivers/scsi/qla2xxx/qla_gs.c
@@ -676,8 +676,7 @@ qla2x00_rff_id(scsi_qla_host_t *vha, u8 type)
 		return (QLA_SUCCESS);
 	}
 
-	return qla_async_rffid(vha, &vha->d_id, qlt_rff_id(vha),
-	    FC4_TYPE_FCP_SCSI);
+	return qla_async_rffid(vha, &vha->d_id, qlt_rff_id(vha), type);
 }
 
 static int qla_async_rffid(scsi_qla_host_t *vha, port_id_t *d_id,
@@ -729,7 +728,7 @@ static int qla_async_rffid(scsi_qla_host_t *vha, port_id_t *d_id,
 	/* Prepare CT arguments -- port_id, FC-4 feature, FC-4 type */
 	ct_req->req.rff_id.port_id = port_id_to_be_id(*d_id);
 	ct_req->req.rff_id.fc4_feature = fc4feature;
-	ct_req->req.rff_id.fc4_type = fc4type;		/* SCSI - FCP */
+	ct_req->req.rff_id.fc4_type = fc4type;		/* SCSI-FCP or FC-NVMe */
 
 	sp->u.iocb_cmd.u.ctarg.req_size = RFF_ID_REQ_SIZE;
 	sp->u.iocb_cmd.u.ctarg.rsp_size = RFF_ID_RSP_SIZE;
-- 
2.19.0.rc0

[PATCH v2 12/13] qla2xxx: Increase max limit of ql2xnvme_queues

[Nilesh Javali]

From: Shreyas Deodhar <sdeodhar@marvell.com>

Increase max limit of ql2xnvme_queues to (max_qpair - 1).

Cc: stable@vger.kernel.org
Fixes: 65120de26a547 ("scsi: qla2xxx: Add ql2xnvme_queues module param to configure number of NVMe queues")
Signed-off-by: Shreyas Deodhar <sdeodhar@marvell.com>
Signed-off-by: Nilesh Javali <njavali@marvell.com>
Reviewed-by: Himanshu Madhani <himanshu.madhani@oracle.com>
---
 drivers/scsi/qla2xxx/qla_nvme.c | 15 ++++++++++-----
 drivers/scsi/qla2xxx/qla_nvme.h |  1 -
 drivers/scsi/qla2xxx/qla_os.c   |  1 -
 3 files changed, 10 insertions(+), 7 deletions(-)

diff --git a/drivers/scsi/qla2xxx/qla_nvme.c b/drivers/scsi/qla2xxx/qla_nvme.c
index 794a95b2e3b4..87c9404aa401 100644
--- a/drivers/scsi/qla2xxx/qla_nvme.c
+++ b/drivers/scsi/qla2xxx/qla_nvme.c
@@ -799,17 +799,22 @@ int qla_nvme_register_hba(struct scsi_qla_host *vha)
 	ha = vha->hw;
 	tmpl = &qla_nvme_fc_transport;
 
-	if (ql2xnvme_queues < MIN_NVME_HW_QUEUES || ql2xnvme_queues > MAX_NVME_HW_QUEUES) {
+	if (ql2xnvme_queues < MIN_NVME_HW_QUEUES) {
 		ql_log(ql_log_warn, vha, 0xfffd,
-		    "ql2xnvme_queues=%d is out of range(MIN:%d - MAX:%d). Resetting ql2xnvme_queues to:%d\n",
-		    ql2xnvme_queues, MIN_NVME_HW_QUEUES, MAX_NVME_HW_QUEUES,
-		    DEF_NVME_HW_QUEUES);
+		    "ql2xnvme_queues=%d is lower than minimum queues: %d. Resetting ql2xnvme_queues to:%d\n",
+		    ql2xnvme_queues, MIN_NVME_HW_QUEUES, DEF_NVME_HW_QUEUES);
 		ql2xnvme_queues = DEF_NVME_HW_QUEUES;
+	} else if (ql2xnvme_queues > (ha->max_qpairs - 1)) {
+		ql_log(ql_log_warn, vha, 0xfffd,
+		       "ql2xnvme_queues=%d is greater than available IRQs: %d. Resetting ql2xnvme_queues to: %d\n",
+		       ql2xnvme_queues, (ha->max_qpairs - 1),
+		       (ha->max_qpairs - 1));
+		ql2xnvme_queues = ((ha->max_qpairs - 1));
 	}
 
 	qla_nvme_fc_transport.max_hw_queues =
 	    min((uint8_t)(ql2xnvme_queues),
-		(uint8_t)(ha->max_qpairs ? ha->max_qpairs : 1));
+		(uint8_t)((ha->max_qpairs - 1) ? (ha->max_qpairs - 1) : 1));
 
 	ql_log(ql_log_info, vha, 0xfffb,
 	       "Number of NVME queues used for this port: %d\n",
diff --git a/drivers/scsi/qla2xxx/qla_nvme.h b/drivers/scsi/qla2xxx/qla_nvme.h
index d0e3c0e07baa..d299478371b2 100644
--- a/drivers/scsi/qla2xxx/qla_nvme.h
+++ b/drivers/scsi/qla2xxx/qla_nvme.h
@@ -14,7 +14,6 @@
 #include "qla_dsd.h"
 
 #define MIN_NVME_HW_QUEUES 1
-#define MAX_NVME_HW_QUEUES 128
 #define DEF_NVME_HW_QUEUES 8
 
 #define NVME_ATIO_CMD_OFF 32
diff --git a/drivers/scsi/qla2xxx/qla_os.c b/drivers/scsi/qla2xxx/qla_os.c
index 89c7ac36a41a..8f47dd421e33 100644
--- a/drivers/scsi/qla2xxx/qla_os.c
+++ b/drivers/scsi/qla2xxx/qla_os.c
@@ -344,7 +344,6 @@ MODULE_PARM_DESC(ql2xnvme_queues,
 	"Number of NVMe Queues that can be configured.\n"
 	"Final value will be min(ql2xnvme_queues, num_cpus,num_chip_queues)\n"
 	"1 - Minimum number of queues supported\n"
-	"128 - Maximum number of queues supported\n"
 	"8 - Default value");
 
 static struct scsi_transport_template *qla2xxx_transport_template = NULL;
-- 
2.19.0.rc0

[PATCH v2 13/13] qla2xxx: Update version to 10.02.07.400-k

[Nilesh Javali]

Signed-off-by: Nilesh Javali <njavali@marvell.com>
Reviewed-by: Himanshu Madhani <himanshu.madhani@oracle.com>
---
 drivers/scsi/qla2xxx/qla_version.h | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/scsi/qla2xxx/qla_version.h b/drivers/scsi/qla2xxx/qla_version.h
index 913d454f4949..b09d7d2080c0 100644
--- a/drivers/scsi/qla2xxx/qla_version.h
+++ b/drivers/scsi/qla2xxx/qla_version.h
@@ -6,9 +6,9 @@
 /*
  * Driver version
  */
-#define QLA2XXX_VERSION      "10.02.07.300-k"
+#define QLA2XXX_VERSION      "10.02.07.400-k"
 
 #define QLA_DRIVER_MAJOR_VER	10
 #define QLA_DRIVER_MINOR_VER	2
 #define QLA_DRIVER_PATCH_VER	7
-#define QLA_DRIVER_BETA_VER	300
+#define QLA_DRIVER_BETA_VER	400
-- 
2.19.0.rc0

Re: [PATCH v2 00/13] qla2xxx driver fixes

[Martin K. Petersen]

Nilesh,

> Please apply the qla2xxx driver misc bug fixes to the scsi tree at
> your earliest convenience.

Applied to 5.18/scsi-staging, thanks!

-- 
Martin K. Petersen	Oracle Linux Engineering

Re: [PATCH v2 00/13] qla2xxx driver fixes

[Martin K. Petersen]

On Thu, 10 Mar 2022 01:25:51 -0800, Nilesh Javali wrote:

> Martin,
> 
> Please apply the qla2xxx driver misc bug fixes to the scsi tree
> at your earliest convenience.
> 
> v2:
> - Incorporate all minor review comments
> - Add Fixes tag to 11/13
> - Add Reviewed-by tag
> 
> [...]

Applied to 5.18/scsi-queue, thanks!

[01/13] qla2xxx: Fix incorrect reporting of task management failure
        https://git.kernel.org/mkp/scsi/c/58ca5999e036
[02/13] qla2xxx: Fix disk failure to rediscover
        https://git.kernel.org/mkp/scsi/c/6a45c8e137d4
[03/13] qla2xxx: Fix loss of NVME namespaces after driver reload test
        https://git.kernel.org/mkp/scsi/c/db212f2eb3fb
[04/13] qla2xxx: Fix missed DMA unmap for NVME ls requests
        https://git.kernel.org/mkp/scsi/c/c85ab7d9e27a
[05/13] qla2xxx: Fix crash during module load unload test
        https://git.kernel.org/mkp/scsi/c/0972252450f9
[06/13] qla2xxx: fix n2n inconsistent plogi
        https://git.kernel.org/mkp/scsi/c/c13ce47c64ea
[07/13] qla2xxx: Fix hang due to session stuck
        https://git.kernel.org/mkp/scsi/c/c02aada06d19
[08/13] qla2xxx: Fix laggy FC remote port session recovery
        https://git.kernel.org/mkp/scsi/c/713b415726f1
[09/13] qla2xxx: reduce false trigger to login
        https://git.kernel.org/mkp/scsi/c/d2646eed7b19
[10/13] qla2xxx: Fix stuck session of prli reject
        https://git.kernel.org/mkp/scsi/c/f3502e2e98a9
[11/13] qla2xxx: Use correct feature type field during rffid processing
        https://git.kernel.org/mkp/scsi/c/a7e05f7a1bcb
[12/13] qla2xxx: Increase max limit of ql2xnvme_queues
        https://git.kernel.org/mkp/scsi/c/3648bcf1c137
[13/13] qla2xxx: Update version to 10.02.07.400-k
        https://git.kernel.org/mkp/scsi/c/811655d005b2

-- 
Martin K. Petersen	Oracle Linux Engineering