[PATCH v2 00/11] qla2xxx driver bug fixes

[PATCH v2 00/11] qla2xxx driver bug fixes

[Nilesh Javali]

Martin,

Please apply the attached miscellaneous qla2xxx bug fixes to the
scsi tree at your earliest convenience.

v1->v2:
Add patch, "qla2xxx: Revert: Disable T10-DIF feature with FC-NVMe
during probe", within the patchset.

Thanks,
Nilesh

Arun Easi (2):
  qla2xxx: Allow ql2xextended_error_logging special value 1 to be set
    anytime
  qla2xxx: Fix WARN_ON in qla_nvme_register_hba

Nilesh Javali (1):
  Revert "scsi: qla2xxx: Fix crash on qla2x00_mailbox_command"

Quinn Tran (7):
  qla2xxx: flush all sessions on zone disable
  qla2xxx: flush IO on zone disable
  qla2xxx: Indicate correct supported speeds for Mezz card
  qla2xxx: fix login timeout
  qla2xxx: reduce noisy debug message
  qla2xxx: fix null pointer access while connections disconnect from
    subsystem
  qla2xxx: Revert: Disable T10-DIF feature with FC-NVMe during probe

Saurav Kashyap (1):
  qla2xxx: Check if FW supports MQ before enabling

 drivers/scsi/qla2xxx/qla_dbg.h    |  3 ++
 drivers/scsi/qla2xxx/qla_def.h    |  1 +
 drivers/scsi/qla2xxx/qla_gs.c     | 49 +++++++++++++++++++++++++------
 drivers/scsi/qla2xxx/qla_isr.c    |  4 +--
 drivers/scsi/qla2xxx/qla_mbx.c    |  8 -----
 drivers/scsi/qla2xxx/qla_nvme.c   | 15 +++++++++-
 drivers/scsi/qla2xxx/qla_os.c     |  9 +++---
 drivers/scsi/qla2xxx/qla_target.c |  2 +-
 8 files changed, 66 insertions(+), 25 deletions(-)


base-commit: b12149f2698ce25621ed0413cbb4fc26dd8ab3c1
-- 
2.19.0.rc0

[PATCH v2 01/11] qla2xxx: flush all sessions on zone disable

[Nilesh Javali]

From: Quinn Tran <qutran@marvell.com>

On Zone Disable, certain switch would ignore all
commands. This cause timeout for both switch
scan command and abort of that command. On detection
of this condition, all sessions will be shutdown.

Signed-off-by: Quinn Tran <qutran@marvell.com>
Signed-off-by: Himanshu Madhani <hmadhani@marvell.com>
Signed-off-by: Nilesh Javali <njavali@marvell.com>
---
 drivers/scsi/qla2xxx/qla_gs.c | 12 ++++++++++++
 1 file changed, 12 insertions(+)

diff --git a/drivers/scsi/qla2xxx/qla_gs.c b/drivers/scsi/qla2xxx/qla_gs.c
index 42c3ad27f1cb..c5529da1df59 100644
--- a/drivers/scsi/qla2xxx/qla_gs.c
+++ b/drivers/scsi/qla2xxx/qla_gs.c
@@ -3734,6 +3734,18 @@ static void qla2x00_async_gpnft_gnnft_sp_done(srb_t *sp, int res)
 		unsigned long flags;
 		const char *name = sp->name;
 
+		if (res == QLA_OS_TIMER_EXPIRED) {
+			/* switch is ignoring all commands.
+			 * This might be a zone disable behavior.
+			 * This means we hit 64s timeout.
+			 * 22s GPNFT + 44s Abort = 64s
+			 */
+			ql_dbg(ql_dbg_disc, vha, 0xffff,
+			       "%s: Switch Zone check please .\n",
+			       name);
+			qla2x00_mark_all_devices_lost(vha);
+		}
+
 		/*
 		 * We are in an Interrupt context, queue up this
 		 * sp for GNNFT_DONE work. This will allow all
-- 
2.19.0.rc0

[PATCH v2 02/11] qla2xxx: flush IO on zone disable

[Nilesh Javali]

From: Quinn Tran <qutran@marvell.com>

Perform implicit logout to flush io on zone disable.

Signed-off-by: Quinn Tran <qutran@marvell.com>
Signed-off-by: Himanshu Madhani <hmadhani@marvell.com>
Signed-off-by: Nilesh Javali <njavali@marvell.com>
---
 drivers/scsi/qla2xxx/qla_gs.c | 1 -
 1 file changed, 1 deletion(-)

diff --git a/drivers/scsi/qla2xxx/qla_gs.c b/drivers/scsi/qla2xxx/qla_gs.c
index c5529da1df59..d9ce8d31457a 100644
--- a/drivers/scsi/qla2xxx/qla_gs.c
+++ b/drivers/scsi/qla2xxx/qla_gs.c
@@ -3436,7 +3436,6 @@ void qla24xx_async_gnnft_done(scsi_qla_host_t *vha, srb_t *sp)
 			list_for_each_entry(fcport, &vha->vp_fcports, list) {
 				if ((fcport->flags & FCF_FABRIC_DEVICE) != 0) {
 					fcport->scan_state = QLA_FCPORT_SCAN;
-					fcport->logout_on_delete = 0;
 				}
 			}
 			goto login_logout;
-- 
2.19.0.rc0

[PATCH v2 03/11] qla2xxx: Indicate correct supported speeds for Mezz card

[Nilesh Javali]

From: Quinn Tran <qutran@marvell.com>

Correct the supported speeds for 16G Mezz card.

Signed-off-by: Quinn Tran <qutran@marvell.com>
Signed-off-by: Nilesh Javali <njavali@marvell.com>
---
 drivers/scsi/qla2xxx/qla_gs.c | 17 ++++++++++++-----
 1 file changed, 12 insertions(+), 5 deletions(-)

diff --git a/drivers/scsi/qla2xxx/qla_gs.c b/drivers/scsi/qla2xxx/qla_gs.c
index d9ce8d31457a..8c30d9dbb48c 100644
--- a/drivers/scsi/qla2xxx/qla_gs.c
+++ b/drivers/scsi/qla2xxx/qla_gs.c
@@ -1505,11 +1505,11 @@ qla2x00_prep_ct_fdmi_req(struct ct_sns_pkt *p, uint16_t cmd,
 static uint
 qla25xx_fdmi_port_speed_capability(struct qla_hw_data *ha)
 {
+	uint speeds = 0;
+
 	if (IS_CNA_CAPABLE(ha))
 		return FDMI_PORT_SPEED_10GB;
 	if (IS_QLA28XX(ha) || IS_QLA27XX(ha)) {
-		uint speeds = 0;
-
 		if (ha->max_supported_speed == 2) {
 			if (ha->min_supported_speed <= 6)
 				speeds |= FDMI_PORT_SPEED_64GB;
@@ -1536,9 +1536,16 @@ qla25xx_fdmi_port_speed_capability(struct qla_hw_data *ha)
 		}
 		return speeds;
 	}
-	if (IS_QLA2031(ha))
-		return FDMI_PORT_SPEED_16GB|FDMI_PORT_SPEED_8GB|
-			FDMI_PORT_SPEED_4GB;
+	if (IS_QLA2031(ha)) {
+		if ((ha->pdev->subsystem_vendor == 0x103C) &&
+		    (ha->pdev->subsystem_device == 0x8002)) {
+			speeds = FDMI_PORT_SPEED_16GB;
+		} else {
+			speeds = FDMI_PORT_SPEED_16GB|FDMI_PORT_SPEED_8GB|
+				FDMI_PORT_SPEED_4GB;
+		}
+		return speeds;
+	}
 	if (IS_QLA25XX(ha))
 		return FDMI_PORT_SPEED_8GB|FDMI_PORT_SPEED_4GB|
 			FDMI_PORT_SPEED_2GB|FDMI_PORT_SPEED_1GB;
-- 
2.19.0.rc0

[PATCH v2 04/11] qla2xxx: fix login timeout

[Nilesh Javali]

From: Quinn Tran <qutran@marvell.com>

Multipath errors were seen during failback due to login timeout.
The remote device sent LOGO, the local host teared down the session
and did relogin. The RSCN arrived indicates remote device is going
through failover after which the relogin is in a 20s timeout phase.
At this point the driver is stuck in the relogin process.
Add a fix to delete the session as part of abort/flush the login.

Signed-off-by: Quinn Tran <qutran@marvell.com>
Signed-off-by: Nilesh Javali <njavali@marvell.com>
---
 drivers/scsi/qla2xxx/qla_gs.c     | 19 ++++++++++++++++---
 drivers/scsi/qla2xxx/qla_target.c |  2 +-
 2 files changed, 17 insertions(+), 4 deletions(-)

diff --git a/drivers/scsi/qla2xxx/qla_gs.c b/drivers/scsi/qla2xxx/qla_gs.c
index 8c30d9dbb48c..2d7a47a2873b 100644
--- a/drivers/scsi/qla2xxx/qla_gs.c
+++ b/drivers/scsi/qla2xxx/qla_gs.c
@@ -3536,10 +3536,23 @@ void qla24xx_async_gnnft_done(scsi_qla_host_t *vha, srb_t *sp)
 		}
 
 		if (fcport->scan_state != QLA_FCPORT_FOUND) {
+			bool do_delete = false;
+
+			if (fcport->scan_needed &&
+			    fcport->disc_state == DSC_LOGIN_PEND) {
+				/* his cable just got disconnected after we
+				 * send him a login. Do delete to prevent
+				 * timeout
+				 */
+				fcport->logout_on_delete = 1;
+				do_delete = true;
+			}
+
 			fcport->scan_needed = 0;
-			if ((qla_dual_mode_enabled(vha) ||
-				qla_ini_mode_enabled(vha)) &&
-			    atomic_read(&fcport->state) == FCS_ONLINE) {
+			if (((qla_dual_mode_enabled(vha) ||
+			      qla_ini_mode_enabled(vha)) &&
+			    atomic_read(&fcport->state) == FCS_ONLINE) ||
+				do_delete) {
 				if (fcport->loop_id != FC_NO_LOOP_ID) {
 					if (fcport->flags & FCF_FCP2_DEVICE)
 						fcport->logout_on_delete = 0;
diff --git a/drivers/scsi/qla2xxx/qla_target.c b/drivers/scsi/qla2xxx/qla_target.c
index fbb80a043b4f..90289162dbd4 100644
--- a/drivers/scsi/qla2xxx/qla_target.c
+++ b/drivers/scsi/qla2xxx/qla_target.c
@@ -1270,7 +1270,7 @@ void qlt_schedule_sess_for_deletion(struct fc_port *sess)
 
 	qla24xx_chk_fcp_state(sess);
 
-	ql_dbg(ql_dbg_tgt, sess->vha, 0xe001,
+	ql_dbg(ql_dbg_disc, sess->vha, 0xe001,
 	    "Scheduling sess %p for deletion %8phC\n",
 	    sess, sess->port_name);
 
-- 
2.19.0.rc0

[PATCH v2 05/11] qla2xxx: reduce noisy debug message

[Nilesh Javali]

From: Quinn Tran <qutran@marvell.com>

Update debug level and message for ELS IOCB done.

Signed-off-by: Quinn Tran <qutran@marvell.com>
Signed-off-by: Nilesh Javali <njavali@marvell.com>
---
 drivers/scsi/qla2xxx/qla_isr.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/scsi/qla2xxx/qla_isr.c b/drivers/scsi/qla2xxx/qla_isr.c
index 27bcd346af7c..ab5275dbc338 100644
--- a/drivers/scsi/qla2xxx/qla_isr.c
+++ b/drivers/scsi/qla2xxx/qla_isr.c
@@ -2024,8 +2024,8 @@ qla24xx_els_ct_entry(scsi_qla_host_t *vha, struct req_que *req,
 				res = DID_ERROR << 16;
 			}
 		}
-		ql_dbg(ql_dbg_user, vha, 0x503f,
-		    "ELS IOCB Done -%s error hdl=%x comp_status=0x%x error subcode 1=0x%x error subcode 2=0x%x total_byte=0x%x\n",
+		ql_dbg(ql_dbg_disc, vha, 0x503f,
+		    "ELS IOCB Done -%s hdl=%x comp_status=0x%x error subcode 1=0x%x error subcode 2=0x%x total_byte=0x%x\n",
 		    type, sp->handle, comp_status, fw_status[1], fw_status[2],
 		    le32_to_cpu(ese->total_byte_count));
 		goto els_ct_done;
-- 
2.19.0.rc0

[PATCH v2 06/11] qla2xxx: Allow ql2xextended_error_logging special value 1 to be set anytime

[Nilesh Javali]

From: Arun Easi <aeasi@marvell.com>

ql2xextended_error_logging could now be set anytime to 1 to get the default
mask value, as opposed to load time only.

Signed-off-by: Arun Easi <aeasi@marvell.com>
Signed-off-by: Nilesh Javali <njavali@marvell.com>
---
 drivers/scsi/qla2xxx/qla_dbg.h | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/drivers/scsi/qla2xxx/qla_dbg.h b/drivers/scsi/qla2xxx/qla_dbg.h
index 91eb6901815c..e1d7de63e8f8 100644
--- a/drivers/scsi/qla2xxx/qla_dbg.h
+++ b/drivers/scsi/qla2xxx/qla_dbg.h
@@ -380,5 +380,8 @@ extern int qla24xx_soft_reset(struct qla_hw_data *);
 static inline int
 ql_mask_match(uint level)
 {
+	if (ql2xextended_error_logging == 1)
+		ql2xextended_error_logging = QL_DBG_DEFAULT1_MASK;
+
 	return (level & ql2xextended_error_logging) == level;
 }
-- 
2.19.0.rc0

[PATCH v2 07/11] qla2xxx: Fix WARN_ON in qla_nvme_register_hba

[Nilesh Javali]

From: Arun Easi <aeasi@marvell.com>

qla_nvme_register_hba puts out a warning when there are not enough
queue pairs available for FC-NVME.
Just fail the NVME registration rather than a WARNING + Call Trace.

Signed-off-by: Arun Easi <aeasi@marvell.com>
Signed-off-by: Nilesh Javali <njavali@marvell.com>
---
 drivers/scsi/qla2xxx/qla_def.h  |  1 +
 drivers/scsi/qla2xxx/qla_nvme.c | 10 +++++++++-
 2 files changed, 10 insertions(+), 1 deletion(-)

diff --git a/drivers/scsi/qla2xxx/qla_def.h b/drivers/scsi/qla2xxx/qla_def.h
index 8c92af5e4390..1bc090d8a71b 100644
--- a/drivers/scsi/qla2xxx/qla_def.h
+++ b/drivers/scsi/qla2xxx/qla_def.h
@@ -3880,6 +3880,7 @@ struct qla_hw_data {
 		uint32_t	scm_supported_f:1;
 				/* Enabled in Driver */
 		uint32_t	scm_enabled:1;
+		uint32_t	max_req_queue_warned:1;
 	} flags;
 
 	uint16_t max_exchg;
diff --git a/drivers/scsi/qla2xxx/qla_nvme.c b/drivers/scsi/qla2xxx/qla_nvme.c
index d66d47a0f958..be1d49f5c622 100644
--- a/drivers/scsi/qla2xxx/qla_nvme.c
+++ b/drivers/scsi/qla2xxx/qla_nvme.c
@@ -686,7 +686,15 @@ int qla_nvme_register_hba(struct scsi_qla_host *vha)
 	tmpl = &qla_nvme_fc_transport;
 
 	WARN_ON(vha->nvme_local_port);
-	WARN_ON(ha->max_req_queues < 3);
+
+	if (ha->max_req_queues < 3) {
+		if (!ha->flags.max_req_queue_warned)
+			ql_log(ql_log_info, vha, 0x2120,
+			       "%s: Disabling FC-NVME due to lack of free queue pairs (%d).\n",
+			       __func__, ha->max_req_queues);
+		ha->flags.max_req_queue_warned = 1;
+		return ret;
+	}
 
 	qla_nvme_fc_transport.max_hw_queues =
 	    min((uint8_t)(qla_nvme_fc_transport.max_hw_queues),
-- 
2.19.0.rc0

[PATCH v2 08/11] qla2xxx: Check if FW supports MQ before enabling

[Nilesh Javali]

From: Saurav Kashyap <skashyap@marvell.com>

OS boot during Boot from SAN was stuck at dracut emergency shell
after enabling nvme driver parameter. For non MQ support the driver
was enabling MQ. Add a check to confirm if FW supports MQ.

Signed-off-by: Saurav Kashyap <skashyap@marvell.com>
Signed-off-by: Nilesh Javali <njavali@marvell.com>
---
 drivers/scsi/qla2xxx/qla_os.c | 5 +++++
 1 file changed, 5 insertions(+)

diff --git a/drivers/scsi/qla2xxx/qla_os.c b/drivers/scsi/qla2xxx/qla_os.c
index 9b59f032a569..fda812b9b564 100644
--- a/drivers/scsi/qla2xxx/qla_os.c
+++ b/drivers/scsi/qla2xxx/qla_os.c
@@ -2017,6 +2017,11 @@ qla2x00_iospace_config(struct qla_hw_data *ha)
 	/* Determine queue resources */
 	ha->max_req_queues = ha->max_rsp_queues = 1;
 	ha->msix_count = QLA_BASE_VECTORS;
+
+	/* Check if FW supports MQ or not */
+	if (!(ha->fw_attributes & BIT_6))
+		goto mqiobase_exit;
+
 	if (!ql2xmqsupport || !ql2xnvmeenable ||
 	    (!IS_QLA25XX(ha) && !IS_QLA81XX(ha)))
 		goto mqiobase_exit;
-- 
2.19.0.rc0

[PATCH v2 09/11] qla2xxx: fix null pointer access while connections disconnect from subsystem

[Nilesh Javali]

From: Quinn Tran <qutran@marvell.com>

NVMEAsync command is being submitted to QLA, while the same nvme controller
is in the middle of reset. The reset path has deleted the association and
freed aen_op->fcp_req.private. Add a check for this private pointer
before issuing the command.

...
 6 [ffffb656ca11fce0] page_fault at ffffffff8c00114e
    [exception RIP: qla_nvme_post_cmd+394]
    RIP: ffffffffc0d012ba  RSP: ffffb656ca11fd98  RFLAGS: 00010206
    RAX: ffff8fb039eda228  RBX: ffff8fb039eda200  RCX: 00000000000da161
    RDX: ffffffffc0d4d0f0  RSI: ffffffffc0d26c9b  RDI: ffff8fb039eda220
    RBP: 0000000000000013   R8: ffff8fb47ff6aa80   R9: 0000000000000002
    R10: 0000000000000000  R11: ffffb656ca11fdc8  R12: ffff8fb27d04a3b0
    R13: ffff8fc46dd98a58  R14: 0000000000000000  R15: ffff8fc4540f0000
    ORIG_RAX: ffffffffffffffff  CS: 0010  SS: 0018
 7 [ffffb656ca11fe08] nvme_fc_start_fcp_op at ffffffffc0241568 [nvme_fc]
 8 [ffffb656ca11fe50] nvme_fc_submit_async_event at ffffffffc0241901 [nvme_fc]
 9 [ffffb656ca11fe68] nvme_async_event_work at ffffffffc014543d [nvme_core]
10 [ffffb656ca11fe98] process_one_work at ffffffff8b6cd437
11 [ffffb656ca11fed8] worker_thread at ffffffff8b6cdcef
12 [ffffb656ca11ff10] kthread at ffffffff8b6d3402
13 [ffffb656ca11ff50] ret_from_fork at ffffffff8c000255

--
PID: 37824  TASK: ffff8fb033063d80  CPU: 20  COMMAND: "kworker/u97:451"
 0 [ffffb656ce1abc28] __schedule at ffffffff8be629e3
 1 [ffffb656ce1abcc8] schedule at ffffffff8be62fe8
 2 [ffffb656ce1abcd0] schedule_timeout at ffffffff8be671ed
 3 [ffffb656ce1abd70] wait_for_completion at ffffffff8be639cf
 4 [ffffb656ce1abdd0] flush_work at ffffffff8b6ce2d5
 5 [ffffb656ce1abe70] nvme_stop_ctrl at ffffffffc0144900 [nvme_core]
 6 [ffffb656ce1abe80] nvme_fc_reset_ctrl_work at ffffffffc0243445 [nvme_fc]
 7 [ffffb656ce1abe98] process_one_work at ffffffff8b6cd437
 8 [ffffb656ce1abed8] worker_thread at ffffffff8b6cdb50
 9 [ffffb656ce1abf10] kthread at ffffffff8b6d3402
10 [ffffb656ce1abf50] ret_from_fork at ffffffff8c000255

Signed-off-by: Quinn Tran <qutran@marvell.com>
Signed-off-by: Nilesh Javali <njavali@marvell.com>
---
 drivers/scsi/qla2xxx/qla_nvme.c | 5 +++++
 1 file changed, 5 insertions(+)

diff --git a/drivers/scsi/qla2xxx/qla_nvme.c b/drivers/scsi/qla2xxx/qla_nvme.c
index be1d49f5c622..f451683db75c 100644
--- a/drivers/scsi/qla2xxx/qla_nvme.c
+++ b/drivers/scsi/qla2xxx/qla_nvme.c
@@ -535,6 +535,11 @@ static int qla_nvme_post_cmd(struct nvme_fc_local_port *lport,
 	struct nvme_private *priv = fd->private;
 	struct qla_nvme_rport *qla_rport = rport->private;
 
+	if (!priv) {
+		/* nvme association has been torn down */
+		return rval;
+	}
+
 	fcport = qla_rport->fcport;
 
 	if (!qpair || !fcport || (qpair && !qpair->fw_started) ||
-- 
2.19.0.rc0

[PATCH v2 10/11] Revert "scsi: qla2xxx: Fix crash on qla2x00_mailbox_command"

[Nilesh Javali]

FCoE adapter initialization failed for ISP8021.

This reverts commit 3cb182b3fa8b7a61f05c671525494697cba39c6a.

Signed-off-by: Saurav Kashyap <skashyap@marvell.com>
Signed-off-by: Nilesh Javali <njavali@marvell.com>
---
 drivers/scsi/qla2xxx/qla_mbx.c | 8 --------
 1 file changed, 8 deletions(-)

diff --git a/drivers/scsi/qla2xxx/qla_mbx.c b/drivers/scsi/qla2xxx/qla_mbx.c
index 73883435ab58..93aafef7f21c 100644
--- a/drivers/scsi/qla2xxx/qla_mbx.c
+++ b/drivers/scsi/qla2xxx/qla_mbx.c
@@ -334,14 +334,6 @@ qla2x00_mailbox_command(scsi_qla_host_t *vha, mbx_cmd_t *mcp)
 			if (time_after(jiffies, wait_time))
 				break;
 
-			/*
-			 * Check if it's UNLOADING, cause we cannot poll in
-			 * this case, or else a NULL pointer dereference
-			 * is triggered.
-			 */
-			if (unlikely(test_bit(UNLOADING, &base_vha->dpc_flags)))
-				return QLA_FUNCTION_TIMEOUT;
-
 			/* Check for pending interrupts. */
 			qla2x00_poll(ha->rsp_q_map[0]);
 
-- 
2.19.0.rc0

[PATCH v2 11/11] qla2xxx: Revert: Disable T10-DIF feature with FC-NVMe during probe

[Nilesh Javali]

From: Quinn Tran <qutran@marvell.com>

FCP T10-PI and NVME features are independent of each other. This
patch allows both features to co-exist.

Fixes: 5da05a26b8305a6 ("scsi: qla2xxx: Disable T10-DIF feature with FC-NVMe during probe")
Signed-off-by: Quinn Tran <qutran@marvell.com>
Signed-off-by: Nilesh Javali <njavali@marvell.com>
---
 drivers/scsi/qla2xxx/qla_os.c | 4 ----
 1 file changed, 4 deletions(-)

diff --git a/drivers/scsi/qla2xxx/qla_os.c b/drivers/scsi/qla2xxx/qla_os.c
index fda812b9b564..8da00ba54aec 100644
--- a/drivers/scsi/qla2xxx/qla_os.c
+++ b/drivers/scsi/qla2xxx/qla_os.c
@@ -2834,10 +2834,6 @@ qla2x00_probe_one(struct pci_dev *pdev, const struct pci_device_id *id)
 	/* This may fail but that's ok */
 	pci_enable_pcie_error_reporting(pdev);
 
-	/* Turn off T10-DIF when FC-NVMe is enabled */
-	if (ql2xnvmeenable)
-		ql2xenabledif = 0;
-
 	ha = kzalloc(sizeof(struct qla_hw_data), GFP_KERNEL);
 	if (!ha) {
 		ql_log_pci(ql_log_fatal, pdev, 0x0009,
-- 
2.19.0.rc0

Re: [PATCH v2 10/11] Revert "scsi: qla2xxx: Fix crash on qla2x00_mailbox_command"

[Daniel Wagner]

On Thu, Aug 06, 2020 at 04:10:13AM -0700, Nilesh Javali wrote:
> FCoE adapter initialization failed for ISP8021.
> 
> This reverts commit 3cb182b3fa8b7a61f05c671525494697cba39c6a.

But wouldn't this revert not also bring back the crash from 3cb182b3fa8b
("scsi: qla2xxx: Fix crash on qla2x00_mailbox_command"):

    This patch fixes a crash on qla2x00_mailbox_command caused when the driver
    is on UNLOADING state and tries to call qla2x00_poll, which triggers a
    NULL pointer dereference.

RE: [PATCH v2 10/11] Revert "scsi: qla2xxx: Fix crash on qla2x00_mailbox_command"

[Saurav Kashyap]

Hi Daniel,

> -----Original Message-----
> From: linux-scsi-owner@vger.kernel.org <linux-scsi-owner@vger.kernel.org>
> On Behalf Of Daniel Wagner
> Sent: Friday, August 7, 2020 1:24 PM
> To: Nilesh Javali <njavali@marvell.com>
> Cc: martin.petersen@oracle.com; linux-scsi@vger.kernel.org; GR-QLogic-
> Storage-Upstream <GR-QLogic-Storage-Upstream@marvell.com>
> Subject: Re: [PATCH v2 10/11] Revert "scsi: qla2xxx: Fix crash on
> qla2x00_mailbox_command"
> 
> On Thu, Aug 06, 2020 at 04:10:13AM -0700, Nilesh Javali wrote:
> > FCoE adapter initialization failed for ISP8021.
> >
> > This reverts commit 3cb182b3fa8b7a61f05c671525494697cba39c6a.
> 
> But wouldn't this revert not also bring back the crash from 3cb182b3fa8b
> ("scsi: qla2xxx: Fix crash on qla2x00_mailbox_command"):

This patch was never there in OOT driver, and we never hit an original problem. I have tested this patch myself
and this have gone through test cycles as well. If an original issue is hit again, we will do an analysis and provide
the fix. This revert fixes a load issues with ISP82XX.

Thanks,
~Saurav
> 
>     This patch fixes a crash on qla2x00_mailbox_command caused when the
> driver
>     is on UNLOADING state and tries to call qla2x00_poll, which triggers a
>     NULL pointer dereference.

Re: [PATCH v2 10/11] Revert "scsi: qla2xxx: Fix crash on qla2x00_mailbox_command"

[Himanshu Madhani]

> On Aug 10, 2020, at 4:55 AM, Saurav Kashyap <skashyap@marvell.com> wrote:
> 
> Hi Daniel,
> 
>> -----Original Message-----
>> From: linux-scsi-owner@vger.kernel.org <linux-scsi-owner@vger.kernel.org>
>> On Behalf Of Daniel Wagner
>> Sent: Friday, August 7, 2020 1:24 PM
>> To: Nilesh Javali <njavali@marvell.com>
>> Cc: martin.petersen@oracle.com; linux-scsi@vger.kernel.org; GR-QLogic-
>> Storage-Upstream <GR-QLogic-Storage-Upstream@marvell.com>
>> Subject: Re: [PATCH v2 10/11] Revert "scsi: qla2xxx: Fix crash on
>> qla2x00_mailbox_command"
>> 
>> On Thu, Aug 06, 2020 at 04:10:13AM -0700, Nilesh Javali wrote:
>>> FCoE adapter initialization failed for ISP8021.
>>> 
>>> This reverts commit 3cb182b3fa8b7a61f05c671525494697cba39c6a.
>> 
>> But wouldn't this revert not also bring back the crash from 3cb182b3fa8b
>> ("scsi: qla2xxx: Fix crash on qla2x00_mailbox_command"):
> 
> This patch was never there in OOT driver, and we never hit an original problem. I have tested this patch myself
> and this have gone through test cycles as well. If an original issue is hit again, we will do an analysis and provide
> the fix. This revert fixes a load issues with ISP82XX.
> 
> Thanks,
> ~Saurav
>> 
>>    This patch fixes a crash on qla2x00_mailbox_command caused when the
>> driver
>>    is on UNLOADING state and tries to call qla2x00_poll, which triggers a
>>    NULL pointer dereference.
> 

If my memory serves me right. This crash reported was never reproduced during my testing when this patch was submitted and nor did it cause any issue. So I am fine with this change being reverted. 

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

--
Himanshu Madhani	 Oracle Linux Engineering

Re: [PATCH v2 01/11] qla2xxx: flush all sessions on zone disable

[Himanshu Madhani]

> On Aug 6, 2020, at 6:10 AM, Nilesh Javali <njavali@marvell.com> wrote:
> 
> From: Quinn Tran <qutran@marvell.com>
> 
> On Zone Disable, certain switch would ignore all
> commands. This cause timeout for both switch
> scan command and abort of that command. On detection
> of this condition, all sessions will be shutdown.
> 
> Signed-off-by: Quinn Tran <qutran@marvell.com>
> Signed-off-by: Himanshu Madhani <hmadhani@marvell.com>
> Signed-off-by: Nilesh Javali <njavali@marvell.com>
> ---
> drivers/scsi/qla2xxx/qla_gs.c | 12 ++++++++++++
> 1 file changed, 12 insertions(+)
> 
> diff --git a/drivers/scsi/qla2xxx/qla_gs.c b/drivers/scsi/qla2xxx/qla_gs.c
> index 42c3ad27f1cb..c5529da1df59 100644
> --- a/drivers/scsi/qla2xxx/qla_gs.c
> +++ b/drivers/scsi/qla2xxx/qla_gs.c
> @@ -3734,6 +3734,18 @@ static void qla2x00_async_gpnft_gnnft_sp_done(srb_t *sp, int res)
> 		unsigned long flags;
> 		const char *name = sp->name;
> 
> +		if (res == QLA_OS_TIMER_EXPIRED) {
> +			/* switch is ignoring all commands.
> +			 * This might be a zone disable behavior.
> +			 * This means we hit 64s timeout.
> +			 * 22s GPNFT + 44s Abort = 64s
> +			 */
> +			ql_dbg(ql_dbg_disc, vha, 0xffff,
> +			       "%s: Switch Zone check please .\n",
> +			       name);
> +			qla2x00_mark_all_devices_lost(vha);
> +		}
> +
> 		/*
> 		 * We are in an Interrupt context, queue up this
> 		 * sp for GNNFT_DONE work. This will allow all
> -- 
> 2.19.0.rc0
> 


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

--
Himanshu Madhani	 Oracle Linux Engineering

Re: [PATCH v2 02/11] qla2xxx: flush IO on zone disable

[Himanshu Madhani]

> On Aug 6, 2020, at 6:10 AM, Nilesh Javali <njavali@marvell.com> wrote:
> 
> From: Quinn Tran <qutran@marvell.com>
> 
> Perform implicit logout to flush io on zone disable.
> 
> Signed-off-by: Quinn Tran <qutran@marvell.com>
> Signed-off-by: Himanshu Madhani <hmadhani@marvell.com>
> Signed-off-by: Nilesh Javali <njavali@marvell.com>
> ---
> drivers/scsi/qla2xxx/qla_gs.c | 1 -
> 1 file changed, 1 deletion(-)
> 
> diff --git a/drivers/scsi/qla2xxx/qla_gs.c b/drivers/scsi/qla2xxx/qla_gs.c
> index c5529da1df59..d9ce8d31457a 100644
> --- a/drivers/scsi/qla2xxx/qla_gs.c
> +++ b/drivers/scsi/qla2xxx/qla_gs.c
> @@ -3436,7 +3436,6 @@ void qla24xx_async_gnnft_done(scsi_qla_host_t *vha, srb_t *sp)
> 			list_for_each_entry(fcport, &vha->vp_fcports, list) {
> 				if ((fcport->flags & FCF_FABRIC_DEVICE) != 0) {
> 					fcport->scan_state = QLA_FCPORT_SCAN;
> -					fcport->logout_on_delete = 0;
> 				}
> 			}
> 			goto login_logout;
> -- 
> 2.19.0.rc0
> 

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

--
Himanshu Madhani	 Oracle Linux Engineering

Re: [PATCH v2 03/11] qla2xxx: Indicate correct supported speeds for Mezz card

[Himanshu Madhani]

> On Aug 6, 2020, at 6:10 AM, Nilesh Javali <njavali@marvell.com> wrote:
> 
> From: Quinn Tran <qutran@marvell.com>
> 
> Correct the supported speeds for 16G Mezz card.
> 
> Signed-off-by: Quinn Tran <qutran@marvell.com>
> Signed-off-by: Nilesh Javali <njavali@marvell.com>
> ---
> drivers/scsi/qla2xxx/qla_gs.c | 17 ++++++++++++-----
> 1 file changed, 12 insertions(+), 5 deletions(-)
> 
> diff --git a/drivers/scsi/qla2xxx/qla_gs.c b/drivers/scsi/qla2xxx/qla_gs.c
> index d9ce8d31457a..8c30d9dbb48c 100644
> --- a/drivers/scsi/qla2xxx/qla_gs.c
> +++ b/drivers/scsi/qla2xxx/qla_gs.c
> @@ -1505,11 +1505,11 @@ qla2x00_prep_ct_fdmi_req(struct ct_sns_pkt *p, uint16_t cmd,
> static uint
> qla25xx_fdmi_port_speed_capability(struct qla_hw_data *ha)
> {
> +	uint speeds = 0;
> +
> 	if (IS_CNA_CAPABLE(ha))
> 		return FDMI_PORT_SPEED_10GB;
> 	if (IS_QLA28XX(ha) || IS_QLA27XX(ha)) {
> -		uint speeds = 0;
> -
> 		if (ha->max_supported_speed == 2) {
> 			if (ha->min_supported_speed <= 6)
> 				speeds |= FDMI_PORT_SPEED_64GB;
> @@ -1536,9 +1536,16 @@ qla25xx_fdmi_port_speed_capability(struct qla_hw_data *ha)
> 		}
> 		return speeds;
> 	}
> -	if (IS_QLA2031(ha))
> -		return FDMI_PORT_SPEED_16GB|FDMI_PORT_SPEED_8GB|
> -			FDMI_PORT_SPEED_4GB;
> +	if (IS_QLA2031(ha)) {
> +		if ((ha->pdev->subsystem_vendor == 0x103C) &&
> +		    (ha->pdev->subsystem_device == 0x8002)) {
> +			speeds = FDMI_PORT_SPEED_16GB;
> +		} else {
> +			speeds = FDMI_PORT_SPEED_16GB|FDMI_PORT_SPEED_8GB|
> +				FDMI_PORT_SPEED_4GB;
> +		}
> +		return speeds;
> +	}
> 	if (IS_QLA25XX(ha))
> 		return FDMI_PORT_SPEED_8GB|FDMI_PORT_SPEED_4GB|
> 			FDMI_PORT_SPEED_2GB|FDMI_PORT_SPEED_1GB;
> -- 
> 2.19.0.rc0
> 

Looks fine.

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

--
Himanshu Madhani	 Oracle Linux Engineering

Re: [PATCH v2 04/11] qla2xxx: fix login timeout

[Himanshu Madhani]

> On Aug 6, 2020, at 6:10 AM, Nilesh Javali <njavali@marvell.com> wrote:
> 
> From: Quinn Tran <qutran@marvell.com>
> 
> Multipath errors were seen during failback due to login timeout.
> The remote device sent LOGO, the local host teared down the session
> and did relogin. The RSCN arrived indicates remote device is going
> through failover after which the relogin is in a 20s timeout phase.
> At this point the driver is stuck in the relogin process.
> Add a fix to delete the session as part of abort/flush the login.
> 
> Signed-off-by: Quinn Tran <qutran@marvell.com>
> Signed-off-by: Nilesh Javali <njavali@marvell.com>
> ---
> drivers/scsi/qla2xxx/qla_gs.c     | 19 ++++++++++++++++---
> drivers/scsi/qla2xxx/qla_target.c |  2 +-
> 2 files changed, 17 insertions(+), 4 deletions(-)
> 
> diff --git a/drivers/scsi/qla2xxx/qla_gs.c b/drivers/scsi/qla2xxx/qla_gs.c
> index 8c30d9dbb48c..2d7a47a2873b 100644
> --- a/drivers/scsi/qla2xxx/qla_gs.c
> +++ b/drivers/scsi/qla2xxx/qla_gs.c
> @@ -3536,10 +3536,23 @@ void qla24xx_async_gnnft_done(scsi_qla_host_t *vha, srb_t *sp)
> 		}
> 
> 		if (fcport->scan_state != QLA_FCPORT_FOUND) {
> +			bool do_delete = false;
> +
> +			if (fcport->scan_needed &&
> +			    fcport->disc_state == DSC_LOGIN_PEND) {
> +				/* his cable just got disconnected after we
> +				 * send him a login. Do delete to prevent
> +				 * timeout
> +				 */

Small nit.. the comment should describe who is “his” in this case

> +				fcport->logout_on_delete = 1;
> +				do_delete = true;
> +			}
> +
> 			fcport->scan_needed = 0;
> -			if ((qla_dual_mode_enabled(vha) ||
> -				qla_ini_mode_enabled(vha)) &&
> -			    atomic_read(&fcport->state) == FCS_ONLINE) {
> +			if (((qla_dual_mode_enabled(vha) ||
> +			      qla_ini_mode_enabled(vha)) &&
> +			    atomic_read(&fcport->state) == FCS_ONLINE) ||
> +				do_delete) {
> 				if (fcport->loop_id != FC_NO_LOOP_ID) {
> 					if (fcport->flags & FCF_FCP2_DEVICE)
> 						fcport->logout_on_delete = 0;
> diff --git a/drivers/scsi/qla2xxx/qla_target.c b/drivers/scsi/qla2xxx/qla_target.c
> index fbb80a043b4f..90289162dbd4 100644
> --- a/drivers/scsi/qla2xxx/qla_target.c
> +++ b/drivers/scsi/qla2xxx/qla_target.c
> @@ -1270,7 +1270,7 @@ void qlt_schedule_sess_for_deletion(struct fc_port *sess)
> 
> 	qla24xx_chk_fcp_state(sess);
> 
> -	ql_dbg(ql_dbg_tgt, sess->vha, 0xe001,
> +	ql_dbg(ql_dbg_disc, sess->vha, 0xe001,
> 	    "Scheduling sess %p for deletion %8phC\n",
> 	    sess, sess->port_name);
> 
> -- 
> 2.19.0.rc0
> 

Other than small nit. 

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

--
Himanshu Madhani	 Oracle Linux Engineering

Re: [PATCH v2 06/11] qla2xxx: Allow ql2xextended_error_logging special value 1 to be set anytime

[Himanshu Madhani]

> On Aug 6, 2020, at 6:10 AM, Nilesh Javali <njavali@marvell.com> wrote:
> 
> From: Arun Easi <aeasi@marvell.com>
> 
> ql2xextended_error_logging could now be set anytime to 1 to get the default
> mask value, as opposed to load time only.
> 
> Signed-off-by: Arun Easi <aeasi@marvell.com>
> Signed-off-by: Nilesh Javali <njavali@marvell.com>
> ---
> drivers/scsi/qla2xxx/qla_dbg.h | 3 +++
> 1 file changed, 3 insertions(+)
> 
> diff --git a/drivers/scsi/qla2xxx/qla_dbg.h b/drivers/scsi/qla2xxx/qla_dbg.h
> index 91eb6901815c..e1d7de63e8f8 100644
> --- a/drivers/scsi/qla2xxx/qla_dbg.h
> +++ b/drivers/scsi/qla2xxx/qla_dbg.h
> @@ -380,5 +380,8 @@ extern int qla24xx_soft_reset(struct qla_hw_data *);
> static inline int
> ql_mask_match(uint level)
> {
> +	if (ql2xextended_error_logging == 1)
> +		ql2xextended_error_logging = QL_DBG_DEFAULT1_MASK;
> +
> 	return (level & ql2xextended_error_logging) == level;
> }
> -- 
> 2.19.0.rc0
> 

Looks Good.

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

--
Himanshu Madhani	 Oracle Linux Engineering

Re: [PATCH v2 05/11] qla2xxx: reduce noisy debug message

[Himanshu Madhani]

> On Aug 6, 2020, at 6:10 AM, Nilesh Javali <njavali@marvell.com> wrote:
> 
> From: Quinn Tran <qutran@marvell.com>
> 
> Update debug level and message for ELS IOCB done.
> 
> Signed-off-by: Quinn Tran <qutran@marvell.com>
> Signed-off-by: Nilesh Javali <njavali@marvell.com>
> ---
> drivers/scsi/qla2xxx/qla_isr.c | 4 ++--
> 1 file changed, 2 insertions(+), 2 deletions(-)
> 
> diff --git a/drivers/scsi/qla2xxx/qla_isr.c b/drivers/scsi/qla2xxx/qla_isr.c
> index 27bcd346af7c..ab5275dbc338 100644
> --- a/drivers/scsi/qla2xxx/qla_isr.c
> +++ b/drivers/scsi/qla2xxx/qla_isr.c
> @@ -2024,8 +2024,8 @@ qla24xx_els_ct_entry(scsi_qla_host_t *vha, struct req_que *req,
> 				res = DID_ERROR << 16;
> 			}
> 		}
> -		ql_dbg(ql_dbg_user, vha, 0x503f,
> -		    "ELS IOCB Done -%s error hdl=%x comp_status=0x%x error subcode 1=0x%x error subcode 2=0x%x total_byte=0x%x\n",
> +		ql_dbg(ql_dbg_disc, vha, 0x503f,
> +		    "ELS IOCB Done -%s hdl=%x comp_status=0x%x error subcode 1=0x%x error subcode 2=0x%x total_byte=0x%x\n",
> 		    type, sp->handle, comp_status, fw_status[1], fw_status[2],
> 		    le32_to_cpu(ese->total_byte_count));
> 		goto els_ct_done;
> -- 
> 2.19.0.rc0
> 

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

--
Himanshu Madhani	 Oracle Linux Engineering

Re: [PATCH v2 07/11] qla2xxx: Fix WARN_ON in qla_nvme_register_hba

[Himanshu Madhani]

> On Aug 6, 2020, at 6:10 AM, Nilesh Javali <njavali@marvell.com> wrote:
> 
> From: Arun Easi <aeasi@marvell.com>
> 
> qla_nvme_register_hba puts out a warning when there are not enough
> queue pairs available for FC-NVME.
> Just fail the NVME registration rather than a WARNING + Call Trace.
> 
> Signed-off-by: Arun Easi <aeasi@marvell.com>
> Signed-off-by: Nilesh Javali <njavali@marvell.com>
> ---
> drivers/scsi/qla2xxx/qla_def.h  |  1 +
> drivers/scsi/qla2xxx/qla_nvme.c | 10 +++++++++-
> 2 files changed, 10 insertions(+), 1 deletion(-)
> 
> diff --git a/drivers/scsi/qla2xxx/qla_def.h b/drivers/scsi/qla2xxx/qla_def.h
> index 8c92af5e4390..1bc090d8a71b 100644
> --- a/drivers/scsi/qla2xxx/qla_def.h
> +++ b/drivers/scsi/qla2xxx/qla_def.h
> @@ -3880,6 +3880,7 @@ struct qla_hw_data {
> 		uint32_t	scm_supported_f:1;
> 				/* Enabled in Driver */
> 		uint32_t	scm_enabled:1;
> +		uint32_t	max_req_queue_warned:1;
> 	} flags;
> 
> 	uint16_t max_exchg;
> diff --git a/drivers/scsi/qla2xxx/qla_nvme.c b/drivers/scsi/qla2xxx/qla_nvme.c
> index d66d47a0f958..be1d49f5c622 100644
> --- a/drivers/scsi/qla2xxx/qla_nvme.c
> +++ b/drivers/scsi/qla2xxx/qla_nvme.c
> @@ -686,7 +686,15 @@ int qla_nvme_register_hba(struct scsi_qla_host *vha)
> 	tmpl = &qla_nvme_fc_transport;
> 
> 	WARN_ON(vha->nvme_local_port);
> -	WARN_ON(ha->max_req_queues < 3);
> +
> +	if (ha->max_req_queues < 3) {
> +		if (!ha->flags.max_req_queue_warned)
> +			ql_log(ql_log_info, vha, 0x2120,
> +			       "%s: Disabling FC-NVME due to lack of free queue pairs (%d).\n",
> +			       __func__, ha->max_req_queues);
> +		ha->flags.max_req_queue_warned = 1;
> +		return ret;
> +	}
> 
> 	qla_nvme_fc_transport.max_hw_queues =
> 	    min((uint8_t)(qla_nvme_fc_transport.max_hw_queues),
> -- 
> 2.19.0.rc0
> 

Makes Sense

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

--
Himanshu Madhani	 Oracle Linux Engineering

Re: [PATCH v2 08/11] qla2xxx: Check if FW supports MQ before enabling

[Himanshu Madhani]

> On Aug 6, 2020, at 6:10 AM, Nilesh Javali <njavali@marvell.com> wrote:
> 
> From: Saurav Kashyap <skashyap@marvell.com>
> 
> OS boot during Boot from SAN was stuck at dracut emergency shell
> after enabling nvme driver parameter. For non MQ support the driver
> was enabling MQ. Add a check to confirm if FW supports MQ.
> 
> Signed-off-by: Saurav Kashyap <skashyap@marvell.com>
> Signed-off-by: Nilesh Javali <njavali@marvell.com>
> ---
> drivers/scsi/qla2xxx/qla_os.c | 5 +++++
> 1 file changed, 5 insertions(+)
> 
> diff --git a/drivers/scsi/qla2xxx/qla_os.c b/drivers/scsi/qla2xxx/qla_os.c
> index 9b59f032a569..fda812b9b564 100644
> --- a/drivers/scsi/qla2xxx/qla_os.c
> +++ b/drivers/scsi/qla2xxx/qla_os.c
> @@ -2017,6 +2017,11 @@ qla2x00_iospace_config(struct qla_hw_data *ha)
> 	/* Determine queue resources */
> 	ha->max_req_queues = ha->max_rsp_queues = 1;
> 	ha->msix_count = QLA_BASE_VECTORS;
> +
> +	/* Check if FW supports MQ or not */
> +	if (!(ha->fw_attributes & BIT_6))
> +		goto mqiobase_exit;
> +
> 	if (!ql2xmqsupport || !ql2xnvmeenable ||
> 	    (!IS_QLA25XX(ha) && !IS_QLA81XX(ha)))
> 		goto mqiobase_exit;
> -- 
> 2.19.0.rc0
> 

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

--
Himanshu Madhani	 Oracle Linux Engineering

Re: [PATCH v2 09/11] qla2xxx: fix null pointer access while connections disconnect from subsystem

[Himanshu Madhani]

> On Aug 6, 2020, at 6:10 AM, Nilesh Javali <njavali@marvell.com> wrote:
> 
> From: Quinn Tran <qutran@marvell.com>
> 
> NVMEAsync command is being submitted to QLA, while the same nvme controller
> is in the middle of reset. The reset path has deleted the association and
> freed aen_op->fcp_req.private. Add a check for this private pointer
> before issuing the command.
> 
> ...
> 6 [ffffb656ca11fce0] page_fault at ffffffff8c00114e
>    [exception RIP: qla_nvme_post_cmd+394]
>    RIP: ffffffffc0d012ba  RSP: ffffb656ca11fd98  RFLAGS: 00010206
>    RAX: ffff8fb039eda228  RBX: ffff8fb039eda200  RCX: 00000000000da161
>    RDX: ffffffffc0d4d0f0  RSI: ffffffffc0d26c9b  RDI: ffff8fb039eda220
>    RBP: 0000000000000013   R8: ffff8fb47ff6aa80   R9: 0000000000000002
>    R10: 0000000000000000  R11: ffffb656ca11fdc8  R12: ffff8fb27d04a3b0
>    R13: ffff8fc46dd98a58  R14: 0000000000000000  R15: ffff8fc4540f0000
>    ORIG_RAX: ffffffffffffffff  CS: 0010  SS: 0018
> 7 [ffffb656ca11fe08] nvme_fc_start_fcp_op at ffffffffc0241568 [nvme_fc]
> 8 [ffffb656ca11fe50] nvme_fc_submit_async_event at ffffffffc0241901 [nvme_fc]
> 9 [ffffb656ca11fe68] nvme_async_event_work at ffffffffc014543d [nvme_core]
> 10 [ffffb656ca11fe98] process_one_work at ffffffff8b6cd437
> 11 [ffffb656ca11fed8] worker_thread at ffffffff8b6cdcef
> 12 [ffffb656ca11ff10] kthread at ffffffff8b6d3402
> 13 [ffffb656ca11ff50] ret_from_fork at ffffffff8c000255
> 
> --
> PID: 37824  TASK: ffff8fb033063d80  CPU: 20  COMMAND: "kworker/u97:451"
> 0 [ffffb656ce1abc28] __schedule at ffffffff8be629e3
> 1 [ffffb656ce1abcc8] schedule at ffffffff8be62fe8
> 2 [ffffb656ce1abcd0] schedule_timeout at ffffffff8be671ed
> 3 [ffffb656ce1abd70] wait_for_completion at ffffffff8be639cf
> 4 [ffffb656ce1abdd0] flush_work at ffffffff8b6ce2d5
> 5 [ffffb656ce1abe70] nvme_stop_ctrl at ffffffffc0144900 [nvme_core]
> 6 [ffffb656ce1abe80] nvme_fc_reset_ctrl_work at ffffffffc0243445 [nvme_fc]
> 7 [ffffb656ce1abe98] process_one_work at ffffffff8b6cd437
> 8 [ffffb656ce1abed8] worker_thread at ffffffff8b6cdb50
> 9 [ffffb656ce1abf10] kthread at ffffffff8b6d3402
> 10 [ffffb656ce1abf50] ret_from_fork at ffffffff8c000255
> 
> Signed-off-by: Quinn Tran <qutran@marvell.com>
> Signed-off-by: Nilesh Javali <njavali@marvell.com>
> ---
> drivers/scsi/qla2xxx/qla_nvme.c | 5 +++++
> 1 file changed, 5 insertions(+)
> 
> diff --git a/drivers/scsi/qla2xxx/qla_nvme.c b/drivers/scsi/qla2xxx/qla_nvme.c
> index be1d49f5c622..f451683db75c 100644
> --- a/drivers/scsi/qla2xxx/qla_nvme.c
> +++ b/drivers/scsi/qla2xxx/qla_nvme.c
> @@ -535,6 +535,11 @@ static int qla_nvme_post_cmd(struct nvme_fc_local_port *lport,
> 	struct nvme_private *priv = fd->private;
> 	struct qla_nvme_rport *qla_rport = rport->private;
> 
> +	if (!priv) {
> +		/* nvme association has been torn down */
> +		return rval;
> +	}
> +
> 	fcport = qla_rport->fcport;
> 
> 	if (!qpair || !fcport || (qpair && !qpair->fw_started) ||
> -- 
> 2.19.0.rc0
> 

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

--
Himanshu Madhani	 Oracle Linux Engineering

Re: [PATCH v2 11/11] qla2xxx: Revert: Disable T10-DIF feature with FC-NVMe during probe

[Himanshu Madhani]

> On Aug 6, 2020, at 6:10 AM, Nilesh Javali <njavali@marvell.com> wrote:
> 
> From: Quinn Tran <qutran@marvell.com>
> 
> FCP T10-PI and NVME features are independent of each other. This
> patch allows both features to co-exist.
> 
> Fixes: 5da05a26b8305a6 ("scsi: qla2xxx: Disable T10-DIF feature with FC-NVMe during probe")
> Signed-off-by: Quinn Tran <qutran@marvell.com>
> Signed-off-by: Nilesh Javali <njavali@marvell.com>
> ---
> drivers/scsi/qla2xxx/qla_os.c | 4 ----
> 1 file changed, 4 deletions(-)
> 
> diff --git a/drivers/scsi/qla2xxx/qla_os.c b/drivers/scsi/qla2xxx/qla_os.c
> index fda812b9b564..8da00ba54aec 100644
> --- a/drivers/scsi/qla2xxx/qla_os.c
> +++ b/drivers/scsi/qla2xxx/qla_os.c
> @@ -2834,10 +2834,6 @@ qla2x00_probe_one(struct pci_dev *pdev, const struct pci_device_id *id)
> 	/* This may fail but that's ok */
> 	pci_enable_pcie_error_reporting(pdev);
> 
> -	/* Turn off T10-DIF when FC-NVMe is enabled */
> -	if (ql2xnvmeenable)
> -		ql2xenabledif = 0;
> -
> 	ha = kzalloc(sizeof(struct qla_hw_data), GFP_KERNEL);
> 	if (!ha) {
> 		ql_log_pci(ql_log_fatal, pdev, 0x0009,
> -- 
> 2.19.0.rc0
> 

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

--
Himanshu Madhani	 Oracle Linux Engineering

Re: [PATCH v2 00/11] qla2xxx driver bug fixes

[Martin K. Petersen]

On Thu, 6 Aug 2020 04:10:03 -0700, Nilesh Javali wrote:

> Martin,
> 
> Please apply the attached miscellaneous qla2xxx bug fixes to the
> scsi tree at your earliest convenience.
> 
> v1->v2:
> Add patch, "qla2xxx: Revert: Disable T10-DIF feature with FC-NVMe
> during probe", within the patchset.
> 
> [...]

Applied to 5.9/scsi-fixes, thanks!

[01/11] scsi: qla2xxx: Flush all sessions on zone disable
        https://git.kernel.org/mkp/scsi/c/10ae30ba6648
[02/11] scsi: qla2xxx: Flush I/O on zone disable
        https://git.kernel.org/mkp/scsi/c/a117579d0205
[03/11] scsi: qla2xxx: Indicate correct supported speeds for Mezz card
        https://git.kernel.org/mkp/scsi/c/4709272f6327
[04/11] scsi: qla2xxx: Fix login timeout
        https://git.kernel.org/mkp/scsi/c/abb31aeaa9b2
[05/11] scsi: qla2xxx: Reduce noisy debug message
        https://git.kernel.org/mkp/scsi/c/81b9d1e19d62
[06/11] scsi: qla2xxx: Allow ql2xextended_error_logging special value 1 to be set anytime
        https://git.kernel.org/mkp/scsi/c/49030003a38a
[07/11] scsi: qla2xxx: Fix WARN_ON in qla_nvme_register_hba
        https://git.kernel.org/mkp/scsi/c/897d68eb816b
[08/11] scsi: qla2xxx: Check if FW supports MQ before enabling
        https://git.kernel.org/mkp/scsi/c/dffa11453313
[09/11] scsi: qla2xxx: Fix null pointer access during disconnect from subsystem
        https://git.kernel.org/mkp/scsi/c/83949613fac6
[10/11] Revert "scsi: qla2xxx: Fix crash on qla2x00_mailbox_command"
        https://git.kernel.org/mkp/scsi/c/de7e6194301a
[11/11] Revert "scsi: qla2xxx: Disable T10-DIF feature with FC-NVMe during probe"
        https://git.kernel.org/mkp/scsi/c/dca93232b361

-- 
Martin K. Petersen	Oracle Linux Engineering

Re: [PATCH v2 10/11] Revert "scsi: qla2xxx: Fix crash on qla2x00_mailbox_command"

[Daniel Wagner]

Hi Saurav,

Sorry for the late response, was on holiday.

> On Aug 10, 2020, at 4:55 AM, Saurav Kashyap <skashyap@marvell.com> wrote:
> This patch was never there in OOT driver, and we never hit an original problem. I have tested this patch myself
> and this have gone through test cycles as well. If an original issue is hit again, we will do an analysis and provide
> the fix. This revert fixes a load issues with ISP82XX.

Ok. Maybe this info could be added to the commit message (next time).

Thanks,
Daniel

RE: [EXT] Re: [PATCH v2 10/11] Revert "scsi: qla2xxx: Fix crash on qla2x00_mailbox_command"

[Saurav Kashyap]

> -----Original Message-----
> From: Daniel Wagner <dwagner@suse.de>
> Sent: Monday, August 24, 2020 1:34 PM
> To: Himanshu Madhani <himanshu.madhani@oracle.com>
> Cc: Saurav Kashyap <skashyap@marvell.com>; Nilesh Javali
> <njavali@marvell.com>; martin.petersen@oracle.com; linux-
> scsi@vger.kernel.org; GR-QLogic-Storage-Upstream <GR-QLogic-Storage-
> Upstream@marvell.com>
> Subject: [EXT] Re: [PATCH v2 10/11] Revert "scsi: qla2xxx: Fix crash on
> qla2x00_mailbox_command"
> 
> External Email
> 
> ----------------------------------------------------------------------
> Hi Saurav,
> 
> Sorry for the late response, was on holiday.
> 
> > On Aug 10, 2020, at 4:55 AM, Saurav Kashyap <skashyap@marvell.com>
> wrote:
> > This patch was never there in OOT driver, and we never hit an original
> problem. I have tested this patch myself
> > and this have gone through test cycles as well. If an original issue is hit again,
> we will do an analysis and provide
> > the fix. This revert fixes a load issues with ISP82XX.
> 
> Ok. Maybe this info could be added to the commit message (next time).

Make sense.

Thanks,
~Saurav
> 
> Thanks,
> Daniel
> 
>