[SCSI] qla2xxx: Handle outstanding mbx cmds on hung f/w scenarios.

Outstanding mailbox commands, have no way to recover on f/w hung, and we
timeout on waiting for mbx response. This in turn affects the recovery process
as follows:
- We might already be in dpc while waiting for mbx to complete, so recovery for
that pci function will never get invoked. Reset Timeout (10 sec) is far less
than mbx timeout (30 sec).
- Other mbx cmds will get stuck due to serial mbx access.

Solution is to identify fw-hung scenario and handle outstanding mbx commands to
have an early-exit instead of waiting for response.
Other mbx commands waiting for access will also do an early-exit if fw-hung is
still applicable.

Signed-off-by: Giridhar Malavali <giridhar.malavali@qlogic.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>

4 files changed, 78 insertions, 24 deletions

diff --git a/drivers/scsi/qla2xxx/qla_def.h b/drivers/scsi/qla2xxx/qla_def.h
index f0e792a82dde..2bb187e23db0 100644
--- a/drivers/scsi/qla2xxx/qla_def.h
+++ b/drivers/scsi/qla2xxx/qla_def.h
@@ -2413,6 +2413,7 @@ struct qla_hw_data {
 		uint32_t	cpu_affinity_enabled	:1;
 		uint32_t	disable_msix_handshake	:1;
 		uint32_t	fcp_prio_enabled	:1;
+		uint32_t	fw_hung	:1;
 	} flags;
 
 	/* This spinlock is used to protect "io transactions", you must
diff --git a/drivers/scsi/qla2xxx/qla_init.c b/drivers/scsi/qla2xxx/qla_init.c
index 9d969b596b14..4c6caccc6ad5 100644
--- a/drivers/scsi/qla2xxx/qla_init.c
+++ b/drivers/scsi/qla2xxx/qla_init.c
@@ -1972,7 +1972,8 @@ qla2x00_fw_ready(scsi_qla_host_t *vha)
 			}
 		} else {
 			/* Mailbox cmd failed. Timeout on min_wait. */
-			if (time_after_eq(jiffies, mtime))
+			if (time_after_eq(jiffies, mtime) ||
+			    (IS_QLA82XX(ha) && ha->flags.fw_hung))
 				break;
 		}
 
diff --git a/drivers/scsi/qla2xxx/qla_mbx.c b/drivers/scsi/qla2xxx/qla_mbx.c
index 10f4815aec77..2f39e3093939 100644
--- a/drivers/scsi/qla2xxx/qla_mbx.c
+++ b/drivers/scsi/qla2xxx/qla_mbx.c
@@ -37,7 +37,7 @@ qla2x00_mailbox_command(scsi_qla_host_t *vha, mbx_cmd_t *mcp)
 	device_reg_t __iomem *reg;
 	uint8_t		abort_active;
 	uint8_t		io_lock_on;
-	uint16_t	command;
+	uint16_t	command = 0;
 	uint16_t	*iptr;
 	uint16_t __iomem *optr;
 	uint32_t	cnt;
@@ -83,6 +83,13 @@ qla2x00_mailbox_command(scsi_qla_host_t *vha, mbx_cmd_t *mcp)
 		return QLA_FUNCTION_TIMEOUT;
 	}
 
+	if (IS_QLA82XX(ha) && ha->flags.fw_hung) {
+		/* Setting Link-Down error */
+		mcp->mb[0] = MBS_LINK_DOWN_ERROR;
+		rval = QLA_FUNCTION_FAILED;
+		goto premature_exit;
+	}
+
 	ha->flags.mbox_busy = 1;
 	/* Save mailbox command for debug */
 	ha->mcp = mcp;
@@ -151,7 +158,8 @@ qla2x00_mailbox_command(scsi_qla_host_t *vha, mbx_cmd_t *mcp)
 				DEBUG2_3_11(printk(KERN_INFO
 				    "%s(%ld): Pending Mailbox timeout. "
 				    "Exiting.\n", __func__, base_vha->host_no));
-				return QLA_FUNCTION_TIMEOUT;
+				rval = QLA_FUNCTION_TIMEOUT;
+				goto premature_exit;
 			}
 			WRT_REG_DWORD(&reg->isp82.hint, HINT_MBX_INT_PENDING);
 		} else if (IS_FWI2_CAPABLE(ha))
@@ -176,7 +184,8 @@ qla2x00_mailbox_command(scsi_qla_host_t *vha, mbx_cmd_t *mcp)
 				DEBUG2_3_11(printk(KERN_INFO
 				    "%s(%ld): Pending Mailbox timeout. "
 				    "Exiting.\n", __func__, base_vha->host_no));
-				return QLA_FUNCTION_TIMEOUT;
+				rval = QLA_FUNCTION_TIMEOUT;
+				goto premature_exit;
 			}
 			WRT_REG_DWORD(&reg->isp82.hint, HINT_MBX_INT_PENDING);
 		} else if (IS_FWI2_CAPABLE(ha))
@@ -214,6 +223,15 @@ qla2x00_mailbox_command(scsi_qla_host_t *vha, mbx_cmd_t *mcp)
 		ha->flags.mbox_int = 0;
 		clear_bit(MBX_INTERRUPT, &ha->mbx_cmd_flags);
 
+		if (IS_QLA82XX(ha) && ha->flags.fw_hung) {
+			ha->flags.mbox_busy = 0;
+			/* Setting Link-Down error */
+			mcp->mb[0] = MBS_LINK_DOWN_ERROR;
+			ha->mcp = NULL;
+			rval = QLA_FUNCTION_FAILED;
+			goto premature_exit;
+		}
+
 		if (ha->mailbox_out[0] != MBS_COMMAND_COMPLETE)
 			rval = QLA_FUNCTION_FAILED;
 
@@ -279,35 +297,51 @@ qla2x00_mailbox_command(scsi_qla_host_t *vha, mbx_cmd_t *mcp)
 			DEBUG2_3_11(printk("%s(%ld): timeout schedule "
 			"isp_abort_needed.\n", __func__,
 			base_vha->host_no));
-			qla_printk(KERN_WARNING, ha,
-			    "Mailbox command timeout occurred. Scheduling ISP "
-			    "abort. eeh_busy: 0x%x\n", ha->flags.eeh_busy);
-			set_bit(ISP_ABORT_NEEDED, &base_vha->dpc_flags);
-			qla2xxx_wake_dpc(vha);
+
+			if (!test_bit(ISP_ABORT_NEEDED, &vha->dpc_flags) &&
+			    !test_bit(ABORT_ISP_ACTIVE, &vha->dpc_flags) &&
+			    !test_bit(ISP_ABORT_RETRY, &vha->dpc_flags)) {
+
+				qla_printk(KERN_WARNING, ha,
+				    "Mailbox command timeout occured. "
+				    "Scheduling ISP " "abort. eeh_busy: 0x%x\n",
+				    ha->flags.eeh_busy);
+				set_bit(ISP_ABORT_NEEDED, &vha->dpc_flags);
+				qla2xxx_wake_dpc(vha);
+			}
 		} else if (!abort_active) {
 			/* call abort directly since we are in the DPC thread */
 			DEBUG(printk("%s(%ld): timeout calling abort_isp\n",
 			    __func__, base_vha->host_no));
 			DEBUG2_3_11(printk("%s(%ld): timeout calling "
 			    "abort_isp\n", __func__, base_vha->host_no));
-			qla_printk(KERN_WARNING, ha,
-			    "Mailbox command timeout occurred. Issuing ISP "
-			    "abort.\n");
-
-			set_bit(ABORT_ISP_ACTIVE, &base_vha->dpc_flags);
-			clear_bit(ISP_ABORT_NEEDED, &base_vha->dpc_flags);
-			if (ha->isp_ops->abort_isp(base_vha)) {
-				/* Failed. retry later. */
-				set_bit(ISP_ABORT_NEEDED, &base_vha->dpc_flags);
+
+			if (!test_bit(ISP_ABORT_NEEDED, &vha->dpc_flags) &&
+			    !test_bit(ABORT_ISP_ACTIVE, &vha->dpc_flags) &&
+			    !test_bit(ISP_ABORT_RETRY, &vha->dpc_flags)) {
+
+				qla_printk(KERN_WARNING, ha,
+				    "Mailbox command timeout occured. "
+				    "Issuing ISP abort.\n");
+
+				set_bit(ABORT_ISP_ACTIVE, &vha->dpc_flags);
+				clear_bit(ISP_ABORT_NEEDED, &vha->dpc_flags);
+				if (ha->isp_ops->abort_isp(vha)) {
+					/* Failed. retry later. */
+					set_bit(ISP_ABORT_NEEDED,
+					    &vha->dpc_flags);
+				}
+				clear_bit(ABORT_ISP_ACTIVE, &vha->dpc_flags);
+				DEBUG(printk("%s(%ld): finished abort_isp\n",
+				    __func__, vha->host_no));
+				DEBUG2_3_11(printk(
+				    "%s(%ld): finished abort_isp\n",
+				    __func__, vha->host_no));
 			}
-			clear_bit(ABORT_ISP_ACTIVE, &base_vha->dpc_flags);
-			DEBUG(printk("%s(%ld): finished abort_isp\n", __func__,
-			    base_vha->host_no));
-			DEBUG2_3_11(printk("%s(%ld): finished abort_isp\n",
-			    __func__, base_vha->host_no));
 		}
 	}
 
+premature_exit:
 	/* Allow next mbx cmd to come in. */
 	complete(&ha->mbx_cmd_comp);
 
diff --git a/drivers/scsi/qla2xxx/qla_nx.c b/drivers/scsi/qla2xxx/qla_nx.c
index 1a9a7343dffa..512ba8a4ac57 100644
--- a/drivers/scsi/qla2xxx/qla_nx.c
+++ b/drivers/scsi/qla2xxx/qla_nx.c
@@ -3543,6 +3543,14 @@ qla82xx_check_fw_alive(scsi_qla_host_t *vha)
 				set_bit(ISP_ABORT_NEEDED, &vha->dpc_flags);
 			}
 			qla2xxx_wake_dpc(vha);
+			if (ha->flags.mbox_busy) {
+				ha->flags.fw_hung = 1;
+				ha->flags.mbox_int = 1;
+				DEBUG2(qla_printk(KERN_ERR, ha,
+				    "Due to fw hung, doing premature "
+				    "completion of mbx command\n"));
+				complete(&ha->mbx_intr_comp);
+			}
 		}
 	}
 	vha->fw_heartbeat_counter = fw_heartbeat_counter;
@@ -3646,6 +3654,14 @@ void qla82xx_watchdog(scsi_qla_host_t *vha)
 				"%s(): Adapter reset needed!\n", __func__);
 			set_bit(ISP_ABORT_NEEDED, &vha->dpc_flags);
 			qla2xxx_wake_dpc(vha);
+			if (ha->flags.mbox_busy) {
+				ha->flags.fw_hung = 1;
+				ha->flags.mbox_int = 1;
+				DEBUG2(qla_printk(KERN_ERR, ha,
+				    "Need reset, doing premature "
+				    "completion of mbx command\n"));
+				complete(&ha->mbx_intr_comp);
+			}
 		} else {
 			qla82xx_check_fw_alive(vha);
 		}
@@ -3701,8 +3717,10 @@ qla82xx_abort_isp(scsi_qla_host_t *vha)
 	qla82xx_clear_rst_ready(ha);
 	qla82xx_idc_unlock(ha);
 
-	if (rval == QLA_SUCCESS)
+	if (rval == QLA_SUCCESS) {
+		ha->flags.fw_hung = 0;
 		qla82xx_restart_isp(vha);
+	}
 
 	if (rval) {
 		vha->flags.online = 1;