[SCSI] pm80xx: thermal, sas controller config and error handling update

Modified thermal configuration to happen after interrupt registration
Added SAS controller configuration during initialization
Added error handling logic to handle I_T_Nexus errors and variants

[jejb: fix up tabs and spaces issues]
Signed-off-by: NAnand Kumar S <AnandKumar.Santhanam@pmcs.com>
Acked-by: NJack Wang <jack_wang@usish.com>
Reviewed-by: NHannes Reinecke <hare@suse.de>
Signed-off-by: NJames Bottomley <JBottomley@Parallels.com>

drivers/scsi/pm8001/pm8001_hwi.c

@@ -1670,7 +1670,7 @@ void pm8001_work_fn(struct work_struct *work)
 	}	break;
 	case IO_OPEN_CNX_ERROR_IT_NEXUS_LOSS:
 		dev = pm8001_dev->sas_device;
-		pm8001_I_T_nexus_reset(dev);
+		pm8001_I_T_nexus_event_handler(dev);
 		break;
 	case IO_OPEN_CNX_ERROR_STP_RESOURCES_BUSY:
 		dev = pm8001_dev->sas_device;

drivers/scsi/pm8001/pm8001_init.c

@@ -838,6 +838,8 @@ static int pm8001_pci_probe(struct pci_dev *pdev,
 	if (pm8001_ha->chip_id != chip_8001) {
 		for (i = 1; i < pm8001_ha->number_of_intr; i++)
 			PM8001_CHIP_DISP->interrupt_enable(pm8001_ha, i);
+		/* setup thermal configuration. */
+		pm80xx_set_thermal_config(pm8001_ha);
 	}
 
 	pm8001_init_sas_add(pm8001_ha);

drivers/scsi/pm8001/pm8001_sas.c

@@ -1018,6 +1018,72 @@ int pm8001_I_T_nexus_reset(struct domain_device *dev)
 	return rc;
 }
 
+/*
+* This function handle the IT_NEXUS_XXX event or completion
+* status code for SSP/SATA/SMP I/O request.
+*/
+int pm8001_I_T_nexus_event_handler(struct domain_device *dev)
+{
+	int rc = TMF_RESP_FUNC_FAILED;
+	struct pm8001_device *pm8001_dev;
+	struct pm8001_hba_info *pm8001_ha;
+	struct sas_phy *phy;
+	u32 device_id = 0;
+
+	if (!dev || !dev->lldd_dev)
+		return -1;
+
+	pm8001_dev = dev->lldd_dev;
+	device_id = pm8001_dev->device_id;
+	pm8001_ha = pm8001_find_ha_by_dev(dev);
+
+	PM8001_EH_DBG(pm8001_ha,
+			pm8001_printk("I_T_Nexus handler invoked !!"));
+
+	phy = sas_get_local_phy(dev);
+
+	if (dev_is_sata(dev)) {
+		DECLARE_COMPLETION_ONSTACK(completion_setstate);
+		if (scsi_is_sas_phy_local(phy)) {
+			rc = 0;
+			goto out;
+		}
+		/* send internal ssp/sata/smp abort command to FW */
+		rc = pm8001_exec_internal_task_abort(pm8001_ha, pm8001_dev ,
+							dev, 1, 0);
+		msleep(100);
+
+		/* deregister the target device */
+		pm8001_dev_gone_notify(dev);
+		msleep(200);
+
+		/*send phy reset to hard reset target */
+		rc = sas_phy_reset(phy, 1);
+		msleep(2000);
+		pm8001_dev->setds_completion = &completion_setstate;
+
+		wait_for_completion(&completion_setstate);
+	} else {
+		/* send internal ssp/sata/smp abort command to FW */
+		rc = pm8001_exec_internal_task_abort(pm8001_ha, pm8001_dev ,
+							dev, 1, 0);
+		msleep(100);
+
+		/* deregister the target device */
+		pm8001_dev_gone_notify(dev);
+		msleep(200);
+
+		/*send phy reset to hard reset target */
+		rc = sas_phy_reset(phy, 1);
+		msleep(2000);
+	}
+	PM8001_EH_DBG(pm8001_ha, pm8001_printk(" for device[%x]:rc=%d\n",
+		pm8001_dev->device_id, rc));
+out:
+	sas_put_local_phy(phy);
+
+	return rc;
+}
 /* mandatory SAM-3, the task reset the specified LUN*/
 int pm8001_lu_reset(struct domain_device *dev, u8 *lun)
 {

drivers/scsi/pm8001/pm8001_sas.h

@@ -563,6 +563,7 @@ int pm8001_dev_found(struct domain_device *dev);
 void pm8001_dev_gone(struct domain_device *dev);
 int pm8001_lu_reset(struct domain_device *dev, u8 *lun);
 int pm8001_I_T_nexus_reset(struct domain_device *dev);
+int pm8001_I_T_nexus_event_handler(struct domain_device *dev);
 int pm8001_query_task(struct sas_task *task);
 void pm8001_open_reject_retry(
 	struct pm8001_hba_info *pm8001_ha,
@@ -625,6 +626,7 @@ void pm8001_free_task(struct sas_task *task);
 void pm8001_tag_free(struct pm8001_hba_info *pm8001_ha, u32 tag);
 struct pm8001_device *pm8001_find_dev(struct pm8001_hba_info *pm8001_ha,
 					u32 device_id);
+int pm80xx_set_thermal_config(struct pm8001_hba_info *pm8001_ha);
 
 int pm8001_bar4_shift(struct pm8001_hba_info *pm8001_ha, u32 shiftValue);
 

drivers/scsi/pm8001/pm80xx_hwi.c

@@ -613,7 +613,7 @@ static void init_pci_device_addresses(struct pm8001_hba_info *pm8001_ha)
  * pm80xx_set_thermal_config - support the thermal configuration
  * @pm8001_ha: our hba card information.
  */
-static int
+int
 pm80xx_set_thermal_config(struct pm8001_hba_info *pm8001_ha)
 {
 	struct set_ctrl_cfg_req payload;
@@ -638,6 +638,86 @@ pm80xx_set_thermal_config(struct pm8001_hba_info *pm8001_ha)
 
 }
 
+/**
+* pm80xx_set_sas_protocol_timer_config - support the SAS Protocol
+* Timer configuration page
+* @pm8001_ha: our hba card information.
+*/
+static int
+pm80xx_set_sas_protocol_timer_config(struct pm8001_hba_info *pm8001_ha)
+{
+	struct set_ctrl_cfg_req payload;
+	struct inbound_queue_table *circularQ;
+	SASProtocolTimerConfig_t SASConfigPage;
+	int rc;
+	u32 tag;
+	u32 opc = OPC_INB_SET_CONTROLLER_CONFIG;
+
+	memset(&payload, 0, sizeof(struct set_ctrl_cfg_req));
+	memset(&SASConfigPage, 0, sizeof(SASProtocolTimerConfig_t));
+
+	rc = pm8001_tag_alloc(pm8001_ha, &tag);
+
+	if (rc)
+		return -1;
+
+	circularQ = &pm8001_ha->inbnd_q_tbl[0];
+	payload.tag = cpu_to_le32(tag);
+
+	SASConfigPage.pageCode        =  SAS_PROTOCOL_TIMER_CONFIG_PAGE;
+	SASConfigPage.MST_MSI         =  3 << 15;
+	SASConfigPage.STP_SSP_MCT_TMO =  (STP_MCT_TMO << 16) | SSP_MCT_TMO;
+	SASConfigPage.STP_FRM_TMO     = (SAS_MAX_OPEN_TIME << 24) |
+				(SMP_MAX_CONN_TIMER << 16) | STP_FRM_TIMER;
+	SASConfigPage.STP_IDLE_TMO    =  STP_IDLE_TIME;
+
+	if (SASConfigPage.STP_IDLE_TMO > 0x3FFFFFF)
+		SASConfigPage.STP_IDLE_TMO = 0x3FFFFFF;
+
+
+	SASConfigPage.OPNRJT_RTRY_INTVL =         (SAS_MFD << 16) |
+						SAS_OPNRJT_RTRY_INTVL;
+	SASConfigPage.Data_Cmd_OPNRJT_RTRY_TMO =  (SAS_DOPNRJT_RTRY_TMO << 16)
+						| SAS_COPNRJT_RTRY_TMO;
+	SASConfigPage.Data_Cmd_OPNRJT_RTRY_THR =  (SAS_DOPNRJT_RTRY_THR << 16)
+						| SAS_COPNRJT_RTRY_THR;
+	SASConfigPage.MAX_AIP =  SAS_MAX_AIP;
+
+	PM8001_INIT_DBG(pm8001_ha,
+			pm8001_printk("SASConfigPage.pageCode "
+			"0x%08x\n", SASConfigPage.pageCode));
+	PM8001_INIT_DBG(pm8001_ha,
+			pm8001_printk("SASConfigPage.MST_MSI "
+			" 0x%08x\n", SASConfigPage.MST_MSI));
+	PM8001_INIT_DBG(pm8001_ha,
+			pm8001_printk("SASConfigPage.STP_SSP_MCT_TMO "
+			" 0x%08x\n", SASConfigPage.STP_SSP_MCT_TMO));
+	PM8001_INIT_DBG(pm8001_ha,
+			pm8001_printk("SASConfigPage.STP_FRM_TMO "
+			" 0x%08x\n", SASConfigPage.STP_FRM_TMO));
+	PM8001_INIT_DBG(pm8001_ha,
+			pm8001_printk("SASConfigPage.STP_IDLE_TMO "
+			" 0x%08x\n", SASConfigPage.STP_IDLE_TMO));
+	PM8001_INIT_DBG(pm8001_ha,
+			pm8001_printk("SASConfigPage.OPNRJT_RTRY_INTVL "
+			" 0x%08x\n", SASConfigPage.OPNRJT_RTRY_INTVL));
+	PM8001_INIT_DBG(pm8001_ha,
+			pm8001_printk("SASConfigPage.Data_Cmd_OPNRJT_RTRY_TMO "
+			" 0x%08x\n", SASConfigPage.Data_Cmd_OPNRJT_RTRY_TMO));
+	PM8001_INIT_DBG(pm8001_ha,
+			pm8001_printk("SASConfigPage.Data_Cmd_OPNRJT_RTRY_THR "
+			" 0x%08x\n", SASConfigPage.Data_Cmd_OPNRJT_RTRY_THR));
+	PM8001_INIT_DBG(pm8001_ha, pm8001_printk("SASConfigPage.MAX_AIP "
+			" 0x%08x\n", SASConfigPage.MAX_AIP));
+
+	memcpy(&payload.cfg_pg, &SASConfigPage,
+			 sizeof(SASProtocolTimerConfig_t));
+
+	rc = pm8001_mpi_build_cmd(pm8001_ha, circularQ, opc, &payload, 0);
+
+	return rc;
+}
+
 /**
  * pm80xx_get_encrypt_info - Check for encryption
  * @pm8001_ha: our hba card information.
@@ -800,11 +880,8 @@ static int pm80xx_chip_init(struct pm8001_hba_info *pm8001_ha)
 	} else
 		return -EBUSY;
 
-	/* configure thermal */
-	pm80xx_set_thermal_config(pm8001_ha);
-
-	PM8001_INIT_DBG(pm8001_ha,
-		pm8001_printk("Thermal configuration successful!\n"));
+	/* send SAS protocol timer configuration page to FW */
+	ret = pm80xx_set_sas_protocol_timer_config(pm8001_ha);
 
 	/* Check for encryption */
 	if (pm8001_ha->chip->encrypt) {
@@ -1269,6 +1346,11 @@ mpi_ssp_completion(struct pm8001_hba_info *pm8001_ha , void *piomb)
 		ts->open_rej_reason = SAS_OREJ_RSVD_RETRY;
 		break;
 	case IO_OPEN_CNX_ERROR_IT_NEXUS_LOSS:
+	case IO_XFER_OPEN_RETRY_BACKOFF_THRESHOLD_REACHED:
+	case IO_OPEN_CNX_ERROR_IT_NEXUS_LOSS_OPEN_TMO:
+	case IO_OPEN_CNX_ERROR_IT_NEXUS_LOSS_NO_DEST:
+	case IO_OPEN_CNX_ERROR_IT_NEXUS_LOSS_OPEN_COLLIDE:
+	case IO_OPEN_CNX_ERROR_IT_NEXUS_LOSS_PATHWAY_BLOCKED:
 		PM8001_IO_DBG(pm8001_ha,
 			pm8001_printk("IO_OPEN_CNX_ERROR_IT_NEXUS_LOSS\n"));
 		ts->resp = SAS_TASK_COMPLETE;
@@ -1472,6 +1554,11 @@ static void mpi_ssp_event(struct pm8001_hba_info *pm8001_ha , void *piomb)
 		ts->open_rej_reason = SAS_OREJ_RSVD_RETRY;
 		break;
 	case IO_OPEN_CNX_ERROR_IT_NEXUS_LOSS:
+	case IO_XFER_OPEN_RETRY_BACKOFF_THRESHOLD_REACHED:
+	case IO_OPEN_CNX_ERROR_IT_NEXUS_LOSS_OPEN_TMO:
+	case IO_OPEN_CNX_ERROR_IT_NEXUS_LOSS_NO_DEST:
+	case IO_OPEN_CNX_ERROR_IT_NEXUS_LOSS_OPEN_COLLIDE:
+	case IO_OPEN_CNX_ERROR_IT_NEXUS_LOSS_PATHWAY_BLOCKED:
 		PM8001_IO_DBG(pm8001_ha,
 			pm8001_printk("IO_OPEN_CNX_ERROR_IT_NEXUS_LOSS\n"));
 		ts->resp = SAS_TASK_COMPLETE;
@@ -1557,6 +1644,13 @@ static void mpi_ssp_event(struct pm8001_hba_info *pm8001_ha , void *piomb)
 		ts->resp = SAS_TASK_COMPLETE;
 		ts->stat = SAS_DATA_OVERRUN;
 		break;
+	case IO_XFER_ERROR_INTERNAL_CRC_ERROR:
+		PM8001_IO_DBG(pm8001_ha,
+			pm8001_printk("IO_XFR_ERROR_INTERNAL_CRC_ERROR\n"));
+		/* TBC: used default set values */
+		ts->resp = SAS_TASK_COMPLETE;
+		ts->stat = SAS_DATA_OVERRUN;
+		break;
 	case IO_XFER_CMD_FRAME_ISSUED:
 		PM8001_IO_DBG(pm8001_ha,
 			pm8001_printk("IO_XFER_CMD_FRAME_ISSUED\n"));
@@ -1761,6 +1855,11 @@ mpi_sata_completion(struct pm8001_hba_info *pm8001_ha, void *piomb)
 		ts->open_rej_reason = SAS_OREJ_RSVD_CONT0;
 		break;
 	case IO_OPEN_CNX_ERROR_IT_NEXUS_LOSS:
+	case IO_XFER_OPEN_RETRY_BACKOFF_THRESHOLD_REACHED:
+	case IO_OPEN_CNX_ERROR_IT_NEXUS_LOSS_OPEN_TMO:
+	case IO_OPEN_CNX_ERROR_IT_NEXUS_LOSS_NO_DEST:
+	case IO_OPEN_CNX_ERROR_IT_NEXUS_LOSS_OPEN_COLLIDE:
+	case IO_OPEN_CNX_ERROR_IT_NEXUS_LOSS_PATHWAY_BLOCKED:
 		PM8001_IO_DBG(pm8001_ha,
 			pm8001_printk("IO_OPEN_CNX_ERROR_IT_NEXUS_LOSS\n"));
 		ts->resp = SAS_TASK_COMPLETE;
@@ -2051,7 +2150,12 @@ static void mpi_sata_event(struct pm8001_hba_info *pm8001_ha , void *piomb)
 		ts->open_rej_reason = SAS_OREJ_RSVD_CONT0;
 		break;
 	case IO_OPEN_CNX_ERROR_IT_NEXUS_LOSS:
-		PM8001_IO_DBG(pm8001_ha,
+	case IO_XFER_OPEN_RETRY_BACKOFF_THRESHOLD_REACHED:
+	case IO_OPEN_CNX_ERROR_IT_NEXUS_LOSS_OPEN_TMO:
+	case IO_OPEN_CNX_ERROR_IT_NEXUS_LOSS_NO_DEST:
+	case IO_OPEN_CNX_ERROR_IT_NEXUS_LOSS_OPEN_COLLIDE:
+	case IO_OPEN_CNX_ERROR_IT_NEXUS_LOSS_PATHWAY_BLOCKED:
+		PM8001_FAIL_DBG(pm8001_ha,
 			pm8001_printk("IO_OPEN_CNX_ERROR_IT_NEXUS_LOSS\n"));
 		ts->resp = SAS_TASK_UNDELIVERED;
 		ts->stat = SAS_DEV_NO_RESPONSE;
@@ -2154,6 +2258,20 @@ static void mpi_sata_event(struct pm8001_hba_info *pm8001_ha , void *piomb)
 		ts->resp = SAS_TASK_COMPLETE;
 		ts->stat = SAS_OPEN_TO;
 		break;
+	case IO_XFER_ERROR_INTERNAL_CRC_ERROR:
+		PM8001_FAIL_DBG(pm8001_ha,
+			pm8001_printk("IO_XFR_ERROR_INTERNAL_CRC_ERROR\n"));
+		/* TBC: used default set values */
+		ts->resp = SAS_TASK_COMPLETE;
+		ts->stat = SAS_OPEN_TO;
+		break;
+	case IO_XFER_DMA_ACTIVATE_TIMEOUT:
+		PM8001_FAIL_DBG(pm8001_ha,
+			pm8001_printk("IO_XFR_DMA_ACTIVATE_TIMEOUT\n"));
+		/* TBC: used default set values */
+		ts->resp = SAS_TASK_COMPLETE;
+		ts->stat = SAS_OPEN_TO;
+		break;
 	default:
 		PM8001_IO_DBG(pm8001_ha,
 			pm8001_printk("Unknown status 0x%x\n", event));
@@ -2305,6 +2423,11 @@ mpi_smp_completion(struct pm8001_hba_info *pm8001_ha, void *piomb)
 		ts->open_rej_reason = SAS_OREJ_RSVD_CONT0;
 		break;
 	case IO_OPEN_CNX_ERROR_IT_NEXUS_LOSS:
+	case IO_XFER_OPEN_RETRY_BACKOFF_THRESHOLD_REACHED:
+	case IO_OPEN_CNX_ERROR_IT_NEXUS_LOSS_OPEN_TMO:
+	case IO_OPEN_CNX_ERROR_IT_NEXUS_LOSS_NO_DEST:
+	case IO_OPEN_CNX_ERROR_IT_NEXUS_LOSS_OPEN_COLLIDE:
+	case IO_OPEN_CNX_ERROR_IT_NEXUS_LOSS_PATHWAY_BLOCKED:
 		PM8001_IO_DBG(pm8001_ha,
 			pm8001_printk("IO_OPEN_CNX_ERROR_IT_NEXUS_LOSS\n"));
 		ts->resp = SAS_TASK_COMPLETE;
@@ -2862,6 +2985,9 @@ static int mpi_hw_event(struct pm8001_hba_info *pm8001_ha, void *piomb)
 	case HW_EVENT_PORT_RECOVERY_TIMER_TMO:
 		PM8001_MSG_DBG(pm8001_ha,
 			pm8001_printk("HW_EVENT_PORT_RECOVERY_TIMER_TMO\n"));
+		pm80xx_hw_event_ack_req(pm8001_ha, 0,
+			HW_EVENT_PORT_RECOVERY_TIMER_TMO,
+			port_id, phy_id, 0, 0);
 		sas_phy_disconnected(sas_phy);
 		phy->phy_attached = 0;
 		sas_ha->notify_port_event(sas_phy, PORTE_LINK_RESET_ERR);
@@ -3499,10 +3625,7 @@ static int pm80xx_chip_ssp_io_req(struct pm8001_hba_info *pm8001_ha,
 	ssp_cmd.ssp_iu.efb_prio_attr |= (task->ssp_task.task_prio << 3);
 	ssp_cmd.ssp_iu.efb_prio_attr |= (task->ssp_task.task_attr & 7);
 	memcpy(ssp_cmd.ssp_iu.cdb, task->ssp_task.cdb, 16);
-	circularQ = &pm8001_ha->inbnd_q_tbl[inb++];
-
-	/* rotate the inb queue */
-	inb = inb%PM8001_MAX_SPCV_INB_NUM;
+	circularQ = &pm8001_ha->inbnd_q_tbl[0];
 
 	/* Check if encryption is set */
 	if (pm8001_ha->chip->encrypt &&
@@ -3603,10 +3726,7 @@ static int pm80xx_chip_sata_req(struct pm8001_hba_info *pm8001_ha,
 	unsigned long flags;
 	u32 opc = OPC_INB_SATA_HOST_OPSTART;
 	memset(&sata_cmd, 0, sizeof(sata_cmd));
-	circularQ = &pm8001_ha->inbnd_q_tbl[inb++];
-
-	/* rotate the inb queue */
-	inb = inb%PM8001_MAX_SPCV_INB_NUM;
+	circularQ = &pm8001_ha->inbnd_q_tbl[0];
 
 	if (task->data_dir == PCI_DMA_NONE) {
 		ATAP = 0x04; /* no data*/

drivers/scsi/pm8001/pm80xx_hwi.h

@@ -197,6 +197,30 @@
 #define CIPHER_MODE_XTS				0x00000002
 #define KEK_MGMT_SUBOP_KEYCARDUPDATE		0x4
 
+/* SAS protocol timer configuration page */
+#define SAS_PROTOCOL_TIMER_CONFIG_PAGE  0x04
+#define STP_MCT_TMO                     32
+#define SSP_MCT_TMO                     32
+#define SAS_MAX_OPEN_TIME				5
+#define SMP_MAX_CONN_TIMER              0xFF
+#define STP_FRM_TIMER                   0
+#define STP_IDLE_TIME                   5 /* 5 us; controller default */
+#define SAS_MFD                         0
+#define SAS_OPNRJT_RTRY_INTVL           2
+#define SAS_DOPNRJT_RTRY_TMO            128
+#define SAS_COPNRJT_RTRY_TMO            128
+
+/*
+  Making ORR bigger than IT NEXUS LOSS which is 2000000us = 2 second.
+  Assuming a bigger value 3 second, 3000000/128 = 23437.5 where 128
+  is DOPNRJT_RTRY_TMO
+*/
+#define SAS_DOPNRJT_RTRY_THR            23438
+#define SAS_COPNRJT_RTRY_THR            23438
+#define SAS_MAX_AIP                     0x200000
+#define IT_NEXUS_TIMEOUT       0x7D0
+#define PORT_RECOVERY_TIMEOUT  ((IT_NEXUS_TIMEOUT/100) + 30)
+
 struct mpi_msg_hdr {
 	__le32	header;	/* Bits [11:0] - Message operation code */
 	/* Bits [15:12] - Message Category */
@@ -996,6 +1020,23 @@ struct ssp_coalesced_comp_resp {
 
 /* new outbound structure for spcv - ends */
 
+/* brief data structure for SAS protocol timer configuration page.
+ *
+ */
+struct SASProtocolTimerConfig {
+	__le32 pageCode;			/* 0 */
+	__le32 MST_MSI;				/* 1 */
+	__le32 STP_SSP_MCT_TMO;			/* 2 */
+	__le32 STP_FRM_TMO;			/* 3 */
+	__le32 STP_IDLE_TMO;			/* 4 */
+	__le32 OPNRJT_RTRY_INTVL;		/* 5 */
+	__le32 Data_Cmd_OPNRJT_RTRY_TMO;	/* 6 */
+	__le32 Data_Cmd_OPNRJT_RTRY_THR;	/* 7 */
+	__le32 MAX_AIP;				/* 8 */
+} __attribute__((packed, aligned(4)));
+
+typedef struct SASProtocolTimerConfig SASProtocolTimerConfig_t;
+
 #define NDS_BITS 0x0F
 #define PDS_BITS 0xF0
 
@@ -1122,7 +1163,8 @@ struct ssp_coalesced_comp_resp {
 #define IO_DS_INVALID					0x49
 /* WARNING: the value is not contiguous from here */
 #define IO_XFER_ERR_LAST_PIO_DATAIN_CRC_ERR	0x52
-#define IO_XFR_ERROR_INTERNAL_CRC_ERROR		0x54
+#define IO_XFER_DMA_ACTIVATE_TIMEOUT		0x53
+#define IO_XFER_ERROR_INTERNAL_CRC_ERROR	0x54
 #define MPI_IO_RQE_BUSY_FULL			0x55
 #define IO_XFER_ERR_EOB_DATA_OVERRUN		0x56
 #define IO_XFR_ERROR_INVALID_SSP_RSP_FRAME	0x57