17  * FITNESS FOR A PARTICULAR PURPOSE, OR NON-INFRINGEMENT, ARE      *
20  * more details, a copy of which can be found in the file COPYING  *
68 /* Used when mapping IRQ vectors in a driver centric manner */
256 		/* dump mem may return a zero when finished or we got a  in lpfc_config_port_prep()
307  * into a human readable string and store it in OptionROMVersion.
328 	/* Decode the Option rom version word to a readable string */  in lpfc_dump_wakeup_param_cmpl()
356 	 * If the name is empty or there exists a soft name  in lpfc_update_vport_wwn()
730  * structure for use as a delayed link up mechanism with the
751  * structure for use as a delayed link up mechanism with the
1185  * this timer fires, a HBA timeout event shall be posted to the lpfc driver
1219  * this timer fires, a RRQ timeout event shall be posted to the lpfc driver
1315 		 * From there, the idle time stat can be calculated as a  in lpfc_idle_stat_delay_work()
1442  * If a HB mbox is not already in progrees, this routine will allocate
1443  * a LPFC_MBOXQ_t, populate it with a MBX_HEARTBEAT (0x31) command,
1452 	/* Is a Heartbeat mbox already in progress */  in lpfc_issue_hb_mbox()
1479  * flag is set, it will force a MBX_HEARTBEAT mbox command, regardless
1482  * try to issue a MBX_HEARTBEAT mbox command.
1503  * is configured and there is no heart-beat mailbox command outstanding, a
1505  * has been a heart-beat mailbox command outstanding, the HBA shall be put
1553 	/* If there is no heart beat outstanding, issue a heartbeat command */  in lpfc_hb_timeout_handler()
1555 		/* If IOs are completing, no need to issue a MBX_HEARTBEAT */  in lpfc_hb_timeout_handler()
1569 		/* Check if a MBX_HEARTBEAT is already in progress */  in lpfc_hb_timeout_handler()
1573 			 * we need to give the hb mailbox cmd a chance to  in lpfc_hb_timeout_handler()
1606 		/* Check to see if we want to force a MBX_HEARTBEAT */  in lpfc_hb_timeout_handler()
1654  * This routine is called to bring a SLI4 HBA offline when HBA hardware error
1718 	 * There was a firmware error. Take the hba offline and then  in lpfc_handle_deferred_eratt()
1739 	 * This is to ptrotect against a race condition in which  in lpfc_handle_deferred_eratt()
1833 		 * There was a firmware error.  Take the hba offline and then  in lpfc_handle_eratt_s3()
1871 		 * failure is a value other than FFER6. Do not call the offline  in lpfc_handle_eratt_s3()
2183  * This routine is invoked from the worker thread to handle a HBA host
2521 	int oneConnect = 0; /* default is not a oneConnect */  in lpfc_get_hba_model_desc()
2791  * lpfc_sli3_post_buffer - Post IOCB(s) with DMA buffer descriptor(s) to a IOCB ring
2793  * @pring: pointer to a IOCB ring.
2796  * This routine posts a given number of IOCBs with the associated DMA buffer
2949 	uint32_t A, B, C, D, E;  in lpfc_sha_iterate()  local
2959 	A = HashResultPointer[0];  in lpfc_sha_iterate()
2975 		TEMP += S(5, A) + E + HashWorkingPointer[t];  in lpfc_sha_iterate()
2979 		B = A;  in lpfc_sha_iterate()
2980 		A = TEMP;  in lpfc_sha_iterate()
2983 	HashResultPointer[0] += A;  in lpfc_sha_iterate()
3037  * lpfc_cleanup - Performs vport cleanups before deleting a vport
3038  * @vport: pointer to a virtual N_Port data structure.
3079 	/* This is a special case flush to return all  in lpfc_cleanup()
3081 	 * two points in the code where a flush is  in lpfc_cleanup()
3084 	 * (this prevents a crash) and the other is  in lpfc_cleanup()
3086 	 * a crash). Both of these exceptions are  in lpfc_cleanup()
3126  * lpfc_stop_vport_timers - Stop all the timers associated with a vport
3127  * @vport: pointer to a virtual N_Port data structure.
3129  * This routine stops all the timers associated with a @vport. This function
3130  * is invoked before disabling or deleting a @vport. Note that the physical
3244 	/* This is a signal to firmware to sync up CMF BW with link speed */  in lpfc_cmf_signal_init()
3308  * This routine stops all the timers associated with a HBA. This function is
3309  * invoked before either putting a HBA offline or unloading the driver.
3348  * lpfc_block_mgmt_io - Mark a HBA's management interface as blocked
3352  * This routine marks a HBA's management interface as blocked. Once the HBA's
3429 						 "0099 RPI alloc error for "  in lpfc_sli4_node_rpi_restore()
3452  * This routine moves a batch of XRIs from lpfc_io_buf_list_put of HWQ 0
3688  * lpfc_online - Initialize and bring a HBA online
3691  * This routine initializes the HBA and brings a HBA online. During this
3777  * lpfc_unblock_mgmt_io - Mark a HBA's management interface to be not blocked
3780  * This routine marks a HBA's management interface as not blocked. Once the
3798  * lpfc_offline_prep - Prepare a HBA to be brought offline
3802  * This routine is invoked to prepare a HBA to be brought offline. It performs
3884  * lpfc_offline - Bring a HBA offline
3887  * This routine actually brings a HBA offline. It stops all the timers
4312 	 * This is because POST_SGL takes a sequential range of XRIs  in lpfc_io_buf_flush()
4468  * them on a list, it post them to the port by using SGL block post.
4472  *   0 = failure, less than num_to_alloc is a partial failure.
4558 		/* add the nvme buffer to a post list */  in lpfc_new_io_buf()
4676  * @instance: a unique integer ID to this FC port.
4679  * This routine creates a FC port for the upper layer protocol. The FC port
4680  * can be created on top of either a physical port or a virtual port provided
4681  * by the HBA. This routine also allocates a SCSI host data structure (shost)
4732 				/* template is for a no reset SCSI Host */  in lpfc_create_port()
4805 		/* SLI-3 has a limited number of hardware queues (3),  in lpfc_create_port()
4873  * This routine destroys a FC port from the upper layer protocol. All the
4895  * lpfc_get_instance - Get a unique integer ID
4897  * This routine allocates a unique integer ID from lpfc_hba_index pool. It
4901  *   instance - a unique integer ID allocated as the new instance.
4918  * This routine is called by the SCSI layer with a SCSI host to determine
5009  * lpfc_host_attrib_init - Initialize SCSI host attributes on a FC port
5012  * This routine initializes a given SCSI host attributes on a FC port. The
5013  * SCSI host can be either on top of a physical port or a virtual port.
5137  * wait period, a new FCF event shall be added to the FCOE async event
5167  * This routine is invoked when there is no I/O on by a VM for the specified
5179 	/* check if there is a need to issue QFPA */  in lpfc_vmid_poll()
5617 	/* Make sure we have a congestion info buffer */  in lpfc_cgn_update_stat()
5706 	/* Make sure we have a congestion info buffer */  in lpfc_cmf_stats_timer()
5860 		 * Collapse all hours into a day. Rotate days  in lpfc_cmf_stats_timer()
5934 	/* This routine works on a ms granularity so sec and usec are  in lpfc_calc_cmf_latency()
6041 		 * a full LPFC_CMF_INTERVAL.  in lpfc_cmf_timer()
6227  * that the event was received and then issue a read_topology mailbox command so
6549 			/* firmware is reporting a status we don't know about */  in lpfc_sli4_async_sli_evt()
6596 		 * Event Data1 - N.A, Event Data2 - N.A  in lpfc_sli4_async_sli_evt()
6643 		/* May be accompanied by a temperature event */  in lpfc_sli4_async_sli_evt()
6665  * lpfc_sli4_perform_vport_cvl - Perform clear virtual link on a vport
6668  * This routine is to perform Clear Virtual Link (CVL) on a vport in
6669  * response to a CVL event.
6688 		/* Cannot find existing Fabric ndlp, so allocate a new one */  in lpfc_sli4_perform_vport_cvl()
6719  * response to a FCF dead event.
6848 		 * table for a fast recovery from case that the current FCF  in lpfc_sli4_async_fip_evt()
6873 			 * as a link down to FCF registration.  in lpfc_sli4_async_fip_evt()
6880 			 * Handling fast FCF failover to a DEAD FCF event is  in lpfc_sli4_async_fip_evt()
6933 			 * the entire FCF table for a fast recovery  in lpfc_sli4_async_fip_evt()
7003  * This routine is to handle the SLI4 asynchronous grp5 event. A grp5 event
7004  * is an asynchronous notified of a logical link speed change.  The Port
7028  * This routine is to handle the SLI4 asynchronous cmstat event. A cmstat event
7029  * is an asynchronous notification of a request to reset CM stats.
7070  * lpfc_cgn_params_parse - Process a FW cong parm change event
7072  * @p_cgn_param: pointer to a data buffer with the FW cong params.
7215  * This routine issues a read_object mailbox command to
7232 	/* Find out if the FW has a new set of congestion parameters. */  in lpfc_sli4_cgn_params_read()
7240 	/* 0 means no data.  A negative means error.  A positive means  in lpfc_sli4_cgn_params_read()
7268  * lpfc_sli4_cgn_parm_chg_evt - Process a FW congestion param change event
7273  * indicates a change to the FW maintained congestion parameters.
7275  * This routine executes a Read_Object mailbox call to obtain the
7296 	/* If the event is claiming an empty object, it's ok.  A write  in lpfc_sli4_cgn_parm_chg_evt()
7297 	 * could have cleared it.  Only error is a negative return  in lpfc_sli4_cgn_parm_chg_evt()
7478  * lpfc_enable_pci_dev - Enable a generic PCI device.
7524  * lpfc_disable_pci_dev - Disable a generic PCI device.
7548  * lpfc_reset_hba - Reset a hba
7551  * This routine is invoked to reset a hba device. It brings the HBA
7552  * offline, performs a board restart, and then brings the board back
7572 			/* Perform a PCI function reset to start from clean */  in lpfc_reset_hba()
7596  * This function enables the PCI SR-IOV virtual functions to a physical
7618  * lpfc_sli_probe_sriov_nr_virtfn - Enable a number of sr-iov virtual functions
7622  * This function enables the PCI SR-IOV virtual functions to a physical
7801 		 * The scsi_buf for a T10-DIF I/O will hold the FCP cmnd,  in lpfc_sli_driver_resource_setup()
7802 		 * the FCP rsp, and a BDE for each. Sice we have no control  in lpfc_sli_driver_resource_setup()
7820 		 * The scsi_buf for a regular I/O will hold the FCP cmnd,  in lpfc_sli_driver_resource_setup()
7821 		 * the FCP rsp, a BDE for each, and a BDE for up to  in lpfc_sli_driver_resource_setup()
7993 	 * we will associate a new ring, for each EQ/CQ/WQ tuple.  in lpfc_sli4_driver_resource_setup()
8091 		 * firmware will try FLOGI before the driver gets a link up.  in lpfc_sli4_driver_resource_setup()
8092 		 * If it fails, the driver should get a MISCONFIGURED async  in lpfc_sli4_driver_resource_setup()
8155 				phba->nvmet_support = 1; /* a match */  in lpfc_sli4_driver_resource_setup()
8221 		 * The scsi_buf for a T10-DIF I/O holds the FCP cmnd,  in lpfc_sli4_driver_resource_setup()
8222 		 * the FCP rsp, and a SGE. Sice we have no control  in lpfc_sli4_driver_resource_setup()
8247 		 * The scsi_buf for a regular I/O holds the FCP cmnd,  in lpfc_sli4_driver_resource_setup()
8248 		 * the FCP rsp, a SGE for each, and a SGE for up to  in lpfc_sli4_driver_resource_setup()
8705  * lpfc_free_sgl_list - Free a given sgl list.
8709  * This routine is invoked to free a give sgl list and memory.
8803  * This is just a place holder for now.
8841  * posts a PAGE_SIZE memory region to the port to hold up to
8876  * This routine is invoked to allocate a single 4KB memory region to
8882  *   A valid rpi hdr on success.
8883  *   A NULL pointer on any failure.
8908 	 * rpi is normalized to a zero base because the physical rpi is  in lpfc_sli4_create_rpi_hdr()
8920 	 * port expects a 4KB DMA-mapped memory region that is 4K aligned.  in lpfc_sli4_create_rpi_hdr()
8996  * lpfc_hba_alloc - Allocate driver hba data structure for a device.
9037  * lpfc_hba_free - Free driver hba data structure with a device.
9094  * This routine is invoked to create HBA physical port and associate a SCSI
9295 	/* Map HBA SLIM to a kernel virtual address. */  in lpfc_sli_pci_mem_setup()
9303 	/* Map HBA Control Registers to a kernel virtual address. */  in lpfc_sli_pci_mem_setup()
9421 			/* Port has a fatal POST error, break out */  in lpfc_sli4_post_status_check()
9432 	 * If there was a port error during POST, then don't proceed with  in lpfc_sli4_post_status_check()
9765 	 * as two 30-bit addresses.  The other data is a bit marking whether  in lpfc_create_bootstrap_mbox()
10522 		/* Get a ptr to the Hardware Queue associated with this CPU */  in lpfc_sli4_queue_create()
11496  * This routine is invoked to allocate and set up a pool of completion queue
11497  * events. The body of the completion queue event is a completion queue entry
11551  * __lpfc_sli4_cq_event_alloc - Allocate a completion-queue event from free pool
11554  * This routine is the lock free version of the API invoked to allocate a
11571  * lpfc_sli4_cq_event_alloc - Allocate a completion-queue event from free pool
11574  * This routine is the lock version of the API invoked to allocate a
11593  * __lpfc_sli4_cq_event_release - Release a completion-queue event to free pool
11597  * This routine is the lock free version of the API invoked to release a
11608  * lpfc_sli4_cq_event_release - Release a completion-queue event to free pool
11612  * This routine is the lock version of the API invoked to release a
11664  * This routine is invoked to request a PCI function reset. It will destroys
11784 	/* Catch the not-ready port failure after a port reset. */  in lpfc_pci_function_reset()
11855 		 * Map SLI4 PCI Config Space Register base to a kernel virtual  in lpfc_sli4_pci_mem_setup()
11891 			 * Map SLI4 if type 0 HBA Control Register base to a  in lpfc_sli4_pci_mem_setup()
11919 		 * Map SLI4 if type 6 HBA Doorbell Register base to a kernel  in lpfc_sli4_pci_mem_setup()
11940 			 * a kernel virtual address and setup the registers.  in lpfc_sli4_pci_mem_setup()
11969 		 * Map SLI4 if type 6 HBA DPP Register base to a kernel  in lpfc_sli4_pci_mem_setup()
12162  * request_irq() to register MSI vector with a interrupt the handler, which
12473 	 * entries to a IRQ on the same phys_id.  in lpfc_cpu_affinity_check()
12486 			/* If so, find a new_cpup that is on the SAME  in lpfc_cpu_affinity_check()
12503 			/* We found a matching phys_id, so copy the IRQ info */  in lpfc_cpu_affinity_check()
12521 	/* Set any unassigned cpu map entries to a IRQ on any phys_id */  in lpfc_cpu_affinity_check()
12532 			/* If so, find a new_cpup thats on ANY phys_id  in lpfc_cpu_affinity_check()
12576 		/* Only FIRST IRQs get a hdwq index assignment. */  in lpfc_cpu_affinity_check()
12580 		/* 1 to 1, the first LPFC_CPU_FIRST_IRQ cpus to a unique hdwq */  in lpfc_cpu_affinity_check()
12589 	/* Associate a hdwq with each cpu_map entry  in lpfc_cpu_affinity_check()
12617 		/* Not a First CPU and all hdw_queues are used.  Reuse a  in lpfc_cpu_affinity_check()
12634 		 * settle for just a phys_id match.  in lpfc_cpu_affinity_check()
12663 	 * a cpu is hot-added. Perform a simple hdwq round robin assignment.  in lpfc_cpu_affinity_check()
12870 		/* Found a valid CPU */  in lpfc_irq_rebalance()
12960  * allocated without requesting OS affinity mapping.  A vector will be
13047 			/* If found a neighboring online cpu, set affinity */  in lpfc_sli4_enable_msix()
13080 				 * each vector receives a place in cpu_map.  in lpfc_sli4_enable_msix()
13127  * calling the request_irq() to register MSI vector with a interrupt the
13304  * a device with SLI-3 interface spec.
13829 	 * accommodate 512K and 1M IOs in a single nvme buf.  in lpfc_get_sli4_parameters()
13917  * This routine is to be called to attach a device with SLI-3 interface spec
14010 		/* Put device to a known state before enabling interrupt */  in lpfc_pci_probe_one_s3()
14084  * This routine is to be called to disattach a device with SLI-3 interface
14186  * minimum PM requirements to a power-aware driver's PM support for the
14225  * driver implements the minimum PM requirements to a power-aware driver's
14363  * subsystem after a PCI bus error affecting this device has been detected.
14408  * reset to restart the PCI card from scratch, as if from a cold-boot.
14428 	dev_printk(KERN_INFO, &pdev->dev, "recovering from a slot reset.\n");  in lpfc_io_slot_reset_s3()
14570 				"by a system configuration setting on "  in lpfc_log_write_firmware_error()
14589  * lpfc_write_firmware - attempt to write a firmware image to the port
14809 	/* Put device to a known state before enabling interrupt */  in lpfc_pci_probe_one_s4()
14871 	 * FC4 type making a check for nvme_support unnecessary.  in lpfc_pci_probe_one_s4()
15017  * requirements to a power-aware driver's PM support for suspend/resume -- all
15057  * implements the minimum PM requirements to a power-aware driver's PM for
15195  * after a PCI bus error affecting this device has been detected. When this
15249  * restart the PCI card from scratch, as if from a cold-boot. During the
15270 	dev_printk(KERN_INFO, &pdev->dev, "recovering from a slot reset.\n");  in lpfc_io_slot_reset_s4()
15491  * function is called by the PCI subsystem after a PCI bus error affecting
15533  * from scratch, as if from a cold-boot. When this routine is invoked, it