/******************************************************************* * This file is part of the Emulex Linux Device Driver for * * Fibre Channel Host Bus Adapters. * * Copyright (C) 2004-2005 Emulex. All rights reserved. * * EMULEX and SLI are trademarks of Emulex. * * www.emulex.com * * Portions Copyright (C) 2004-2005 Christoph Hellwig * * * * This program is free software; you can redistribute it and/or * * modify it under the terms of version 2 of the GNU General * * Public License as published by the Free Software Foundation. * * This program is distributed in the hope that it will be useful. * * ALL EXPRESS OR IMPLIED CONDITIONS, REPRESENTATIONS AND * * WARRANTIES, INCLUDING ANY IMPLIED WARRANTY OF MERCHANTABILITY, * * FITNESS FOR A PARTICULAR PURPOSE, OR NON-INFRINGEMENT, ARE * * DISCLAIMED, EXCEPT TO THE EXTENT THAT SUCH DISCLAIMERS ARE HELD * * TO BE LEGALLY INVALID. See the GNU General Public License for * * more details, a copy of which can be found in the file COPYING * * included with this package. * *******************************************************************/ #include <linux/blkdev.h> #include <linux/pci.h> #include <linux/kthread.h> #include <linux/interrupt.h> #include <scsi/scsi.h> #include <scsi/scsi_device.h> #include <scsi/scsi_host.h> #include <scsi/scsi_transport_fc.h> #include "lpfc_hw.h" #include "lpfc_disc.h" #include "lpfc_sli.h" #include "lpfc_scsi.h" #include "lpfc.h" #include "lpfc_logmsg.h" #include "lpfc_crtn.h" /* AlpaArray for assignment of scsid for scan-down and bind_method */ static uint8_t lpfcAlpaArray[] = { 0xEF, 0xE8, 0xE4, 0xE2, 0xE1, 0xE0, 0xDC, 0xDA, 0xD9, 0xD6, 0xD5, 0xD4, 0xD3, 0xD2, 0xD1, 0xCE, 0xCD, 0xCC, 0xCB, 0xCA, 0xC9, 0xC7, 0xC6, 0xC5, 0xC3, 0xBC, 0xBA, 0xB9, 0xB6, 0xB5, 0xB4, 0xB3, 0xB2, 0xB1, 0xAE, 0xAD, 0xAC, 0xAB, 0xAA, 0xA9, 0xA7, 0xA6, 0xA5, 0xA3, 0x9F, 0x9E, 0x9D, 0x9B, 0x98, 0x97, 0x90, 0x8F, 0x88, 0x84, 0x82, 0x81, 0x80, 0x7C, 0x7A, 0x79, 0x76, 0x75, 0x74, 0x73, 0x72, 0x71, 0x6E, 0x6D, 0x6C, 0x6B, 0x6A, 0x69, 0x67, 0x66, 0x65, 0x63, 0x5C, 0x5A, 0x59, 0x56, 0x55, 0x54, 0x53, 0x52, 0x51, 0x4E, 0x4D, 0x4C, 0x4B, 0x4A, 0x49, 0x47, 0x46, 0x45, 0x43, 0x3C, 0x3A, 0x39, 0x36, 0x35, 0x34, 0x33, 0x32, 0x31, 0x2E, 0x2D, 0x2C, 0x2B, 0x2A, 0x29, 0x27, 0x26, 0x25, 0x23, 0x1F, 0x1E, 0x1D, 0x1B, 0x18, 0x17, 0x10, 0x0F, 0x08, 0x04, 0x02, 0x01 }; static void lpfc_disc_timeout_handler(struct lpfc_hba *); static void lpfc_process_nodev_timeout(struct lpfc_hba *phba, struct lpfc_nodelist *ndlp) { int warn_on = 0; spin_lock_irq(phba->host->host_lock); if (!(ndlp->nlp_flag & NLP_NODEV_TMO)) { spin_unlock_irq(phba->host->host_lock); return; } ndlp->nlp_flag &= ~NLP_NODEV_TMO; if (ndlp->nlp_sid != NLP_NO_SID) { warn_on = 1; /* flush the target */ lpfc_sli_abort_iocb(phba, &phba->sli.ring[phba->sli.fcp_ring], ndlp->nlp_sid, 0, 0, LPFC_CTX_TGT); } spin_unlock_irq(phba->host->host_lock); if (warn_on) { lpfc_printf_log(phba, KERN_ERR, LOG_DISCOVERY, "%d:0203 Nodev timeout on NPort x%x " "Data: x%x x%x x%x\n", phba->brd_no, ndlp->nlp_DID, ndlp->nlp_flag, ndlp->nlp_state, ndlp->nlp_rpi); } else { lpfc_printf_log(phba, KERN_INFO, LOG_DISCOVERY, "%d:0204 Nodev timeout on NPort x%x " "Data: x%x x%x x%x\n", phba->brd_no, ndlp->nlp_DID, ndlp->nlp_flag, ndlp->nlp_state, ndlp->nlp_rpi); } lpfc_disc_state_machine(phba, ndlp, NULL, NLP_EVT_DEVICE_RM); return; } static void lpfc_work_list_done(struct lpfc_hba * phba) { struct lpfc_work_evt *evtp = NULL; struct lpfc_nodelist *ndlp; int free_evt; spin_lock_irq(phba->host->host_lock); while(!list_empty(&phba->work_list)) { list_remove_head((&phba->work_list), evtp, typeof(*evtp), evt_listp); spin_unlock_irq(phba->host->host_lock); free_evt = 1; switch(evtp->evt) { case LPFC_EVT_NODEV_TMO: ndlp = (struct lpfc_nodelist *)(evtp->evt_arg1); lpfc_process_nodev_timeout(phba, ndlp); free_evt = 0; break; case LPFC_EVT_ELS_RETRY: ndlp = (struct lpfc_nodelist *)(evtp->evt_arg1); lpfc_els_retry_delay_handler(ndlp); free_evt = 0; break; case LPFC_EVT_ONLINE: *(int *)(evtp->evt_arg1) = lpfc_online(phba); complete((struct completion *)(evtp->evt_arg2)); break; case LPFC_EVT_OFFLINE: *(int *)(evtp->evt_arg1) = lpfc_offline(phba); complete((struct completion *)(evtp->evt_arg2)); break; } if (free_evt) kfree(evtp); spin_lock_irq(phba->host->host_lock); } spin_unlock_irq(phba->host->host_lock); } static void lpfc_work_done(struct lpfc_hba * phba) { struct lpfc_sli_ring *pring; int i; uint32_t ha_copy; uint32_t control; uint32_t work_hba_events; spin_lock_irq(phba->host->host_lock); ha_copy = phba->work_ha; phba->work_ha = 0; work_hba_events=phba->work_hba_events; spin_unlock_irq(phba->host->host_lock); if(ha_copy & HA_ERATT) lpfc_handle_eratt(phba); if(ha_copy & HA_MBATT) lpfc_sli_handle_mb_event(phba); if(ha_copy & HA_LATT) lpfc_handle_latt(phba); if (work_hba_events & WORKER_DISC_TMO) lpfc_disc_timeout_handler(phba); if (work_hba_events & WORKER_ELS_TMO) lpfc_els_timeout_handler(phba); if (work_hba_events & WORKER_MBOX_TMO) lpfc_mbox_timeout_handler(phba); if (work_hba_events & WORKER_FDMI_TMO) lpfc_fdmi_tmo_handler(phba); spin_lock_irq(phba->host->host_lock); phba->work_hba_events &= ~work_hba_events; spin_unlock_irq(phba->host->host_lock); for (i = 0; i < phba->sli.num_rings; i++, ha_copy >>= 4) { pring = &phba->sli.ring[i]; if ((ha_copy & HA_RXATT) || (pring->flag & LPFC_DEFERRED_RING_EVENT)) { if (pring->flag & LPFC_STOP_IOCB_MASK) { pring->flag |= LPFC_DEFERRED_RING_EVENT; } else { lpfc_sli_handle_slow_ring_event(phba, pring, (ha_copy & HA_RXMASK)); pring->flag &= ~LPFC_DEFERRED_RING_EVENT; } /* * Turn on Ring interrupts */ spin_lock_irq(phba->host->host_lock); control = readl(phba->HCregaddr); control |= (HC_R0INT_ENA << i); writel(control, phba->HCregaddr); readl(phba->HCregaddr); /* flush */ spin_unlock_irq(phba->host->host_lock); } } lpfc_work_list_done (phba); } static int check_work_wait_done(struct lpfc_hba *phba) { spin_lock_irq(phba->host->host_lock); if (phba->work_ha || phba->work_hba_events || (!list_empty(&phba->work_list)) || kthread_should_stop()) { spin_unlock_irq(phba->host->host_lock); return 1; } else { spin_unlock_irq(phba->host->host_lock); return 0; } } int lpfc_do_work(void *p) { struct lpfc_hba *phba = p; int rc; DECLARE_WAIT_QUEUE_HEAD(work_waitq); set_user_nice(current, -20); phba->work_wait = &work_waitq; while (1) { rc = wait_event_interruptible(work_waitq, check_work_wait_done(phba)); BUG_ON(rc); if (kthread_should_stop()) break; lpfc_work_done(phba); } phba->work_wait = NULL; return 0; } /* * This is only called to handle FC worker events. Since this a rare * occurance, we allocate a struct lpfc_work_evt structure here instead of * embedding it in the IOCB. */ int lpfc_workq_post_event(struct lpfc_hba * phba, void *arg1, void *arg2, uint32_t evt) { struct lpfc_work_evt *evtp; /* * All Mailbox completions and LPFC_ELS_RING rcv ring IOCB events will * be queued to worker thread for processing */ evtp = kmalloc(sizeof(struct lpfc_work_evt), GFP_KERNEL); if (!evtp) return 0; evtp->evt_arg1 = arg1; evtp->evt_arg2 = arg2; evtp->evt = evt; list_add_tail(&evtp->evt_listp, &phba->work_list); spin_lock_irq(phba->host->host_lock); if (phba->work_wait) wake_up(phba->work_wait); spin_unlock_irq(phba->host->host_lock); return 1; } int lpfc_linkdown(struct lpfc_hba * phba) { struct lpfc_sli *psli; struct lpfc_nodelist *ndlp, *next_ndlp; struct list_head *listp; struct list_head *node_list[7]; LPFC_MBOXQ_t *mb; int rc, i; psli = &phba->sli; spin_lock_irq(phba->host->host_lock); phba->hba_state = LPFC_LINK_DOWN; spin_unlock_irq(phba->host->host_lock); /* Clean up any firmware default rpi's */ if ((mb = mempool_alloc(phba->mbox_mem_pool, GFP_KERNEL))) { lpfc_unreg_did(phba, 0xffffffff, mb); mb->mbox_cmpl=lpfc_sli_def_mbox_cmpl; if (lpfc_sli_issue_mbox(phba, mb, (MBX_NOWAIT | MBX_STOP_IOCB)) == MBX_NOT_FINISHED) { mempool_free( mb, phba->mbox_mem_pool); } } /* Cleanup any outstanding RSCN activity */ lpfc_els_flush_rscn(phba); /* Cleanup any outstanding ELS commands */ lpfc_els_flush_cmd(phba); /* Issue a LINK DOWN event to all nodes */ node_list[0] = &phba->fc_npr_list; /* MUST do this list first */ node_list[1] = &phba->fc_nlpmap_list; node_list[2] = &phba->fc_nlpunmap_list; node_list[3] = &phba->fc_prli_list; node_list[4] = &phba->fc_reglogin_list; node_list[5] = &phba->fc_adisc_list; node_list[6] = &phba->fc_plogi_list; for (i = 0; i < 7; i++) { listp = node_list[i]; if (list_empty(listp)) continue; list_for_each_entry_safe(ndlp, next_ndlp, listp, nlp_listp) { /* Fabric nodes are not handled thru state machine for link down */ if (ndlp->nlp_type & NLP_FABRIC) { /* Remove ALL Fabric nodes except Fabric_DID */ if (ndlp->nlp_DID != Fabric_DID) { /* Take it off current list and free */ lpfc_nlp_list(phba, ndlp, NLP_NO_LIST); } } else { rc = lpfc_disc_state_machine(phba, ndlp, NULL, NLP_EVT_DEVICE_RECOVERY); /* Check config parameter use-adisc or FCP-2 */ if ((rc != NLP_STE_FREED_NODE) && (phba->cfg_use_adisc == 0) && !(ndlp->nlp_fcp_info & NLP_FCP_2_DEVICE)) { /* We know we will have to relogin, so * unreglogin the rpi right now to fail * any outstanding I/Os quickly. */ lpfc_unreg_rpi(phba, ndlp); } } } } /* free any ndlp's on unused list */ list_for_each_entry_safe(ndlp, next_ndlp, &phba->fc_unused_list, nlp_listp) { lpfc_nlp_list(phba, ndlp, NLP_NO_LIST); } /* Setup myDID for link up if we are in pt2pt mode */ if (phba->fc_flag & FC_PT2PT) { phba->fc_myDID = 0; if ((mb = mempool_alloc(phba->mbox_mem_pool, GFP_KERNEL))) { lpfc_config_link(phba, mb); mb->mbox_cmpl=lpfc_sli_def_mbox_cmpl; if (lpfc_sli_issue_mbox (phba, mb, (MBX_NOWAIT | MBX_STOP_IOCB)) == MBX_NOT_FINISHED) { mempool_free( mb, phba->mbox_mem_pool); } } spin_lock_irq(phba->host->host_lock); phba->fc_flag &= ~(FC_PT2PT | FC_PT2PT_PLOGI); spin_unlock_irq(phba->host->host_lock); } spin_lock_irq(phba->host->host_lock); phba->fc_flag &= ~FC_LBIT; spin_unlock_irq(phba->host->host_lock); /* Turn off discovery timer if its running */ lpfc_can_disctmo(phba); /* Must process IOCBs on all rings to handle ABORTed I/Os */ return (0); } static int lpfc_linkup(struct lpfc_hba * phba) { struct lpfc_nodelist *ndlp, *next_ndlp; spin_lock_irq(phba->host->host_lock); phba->hba_state = LPFC_LINK_UP; phba->fc_flag &= ~(FC_PT2PT | FC_PT2PT_PLOGI | FC_ABORT_DISCOVERY | FC_RSCN_MODE | FC_NLP_MORE | FC_RSCN_DISCOVERY); phba->fc_flag |= FC_NDISC_ACTIVE; phba->fc_ns_retry = 0; spin_unlock_irq(phba->host->host_lock); /* * Clean up old Fabric NLP_FABRIC logins. */ list_for_each_entry_safe(ndlp, next_ndlp, &phba->fc_nlpunmap_list, nlp_listp) { if (ndlp->nlp_DID == Fabric_DID) { /* Take it off current list and free */ lpfc_nlp_list(phba, ndlp, NLP_NO_LIST); } } /* free any ndlp's on unused list */ list_for_each_entry_safe(ndlp, next_ndlp, &phba->fc_unused_list, nlp_listp) { lpfc_nlp_list(phba, ndlp, NLP_NO_LIST); } return 0; } /* * This routine handles processing a CLEAR_LA mailbox * command upon completion. It is setup in the LPFC_MBOXQ * as the completion routine when the command is * handed off to the SLI layer. */ void lpfc_mbx_cmpl_clear_la(struct lpfc_hba * phba, LPFC_MBOXQ_t * pmb) { struct lpfc_sli *psli; MAILBOX_t *mb; uint32_t control; psli = &phba->sli; mb = &pmb->mb; /* Since we don't do discovery right now, turn these off here */ psli->ring[psli->ip_ring].flag &= ~LPFC_STOP_IOCB_EVENT; psli->ring[psli->fcp_ring].flag &= ~LPFC_STOP_IOCB_EVENT; psli->ring[psli->next_ring].flag &= ~LPFC_STOP_IOCB_EVENT; /* Check for error */ if ((mb->mbxStatus) && (mb->mbxStatus != 0x1601)) { /* CLEAR_LA mbox error <mbxStatus> state <hba_state> */ lpfc_printf_log(phba, KERN_ERR, LOG_MBOX, "%d:0320 CLEAR_LA mbxStatus error x%x hba " "state x%x\n", phba->brd_no, mb->mbxStatus, phba->hba_state); phba->hba_state = LPFC_HBA_ERROR; goto out; } if (phba->fc_flag & FC_ABORT_DISCOVERY) goto out; phba->num_disc_nodes = 0; /* go thru NPR list and issue ELS PLOGIs */ if (phba->fc_npr_cnt) { lpfc_els_disc_plogi(phba); } if(!phba->num_disc_nodes) { spin_lock_irq(phba->host->host_lock); phba->fc_flag &= ~FC_NDISC_ACTIVE; spin_unlock_irq(phba->host->host_lock); } phba->hba_state = LPFC_HBA_READY; out: /* Device Discovery completes */ lpfc_printf_log(phba, KERN_INFO, LOG_DISCOVERY, "%d:0225 Device Discovery completes\n", phba->brd_no); mempool_free( pmb, phba->mbox_mem_pool); spin_lock_irq(phba->host->host_lock); phba->fc_flag &= ~FC_ABORT_DISCOVERY; if (phba->fc_flag & FC_ESTABLISH_LINK) { phba->fc_flag &= ~FC_ESTABLISH_LINK; } spin_unlock_irq(phba->host->host_lock); del_timer_sync(&phba->fc_estabtmo); lpfc_can_disctmo(phba); /* turn on Link Attention interrupts */ spin_lock_irq(phba->host->host_lock); psli->sli_flag |= LPFC_PROCESS_LA; control = readl(phba->HCregaddr); control |= HC_LAINT_ENA; writel(control, phba->HCregaddr); readl(phba->HCregaddr); /* flush */ spin_unlock_irq(phba->host->host_lock); return; } static void lpfc_mbx_cmpl_config_link(struct lpfc_hba * phba, LPFC_MBOXQ_t * pmb) { struct lpfc_sli *psli; MAILBOX_t *mb; psli = &phba->sli; mb = &pmb->mb; /* Check for error */ if (mb->mbxStatus) { /* CONFIG_LINK mbox error <mbxStatus> state <hba_state> */ lpfc_printf_log(phba, KERN_ERR, LOG_MBOX, "%d:0306 CONFIG_LINK mbxStatus error x%x " "HBA state x%x\n", phba->brd_no, mb->mbxStatus, phba->hba_state); lpfc_linkdown(phba); phba->hba_state = LPFC_HBA_ERROR; goto out; } if (phba->hba_state == LPFC_LOCAL_CFG_LINK) { if (phba->fc_topology == TOPOLOGY_LOOP) { /* If we are public loop and L bit was set */ if ((phba->fc_flag & FC_PUBLIC_LOOP) && !(phba->fc_flag & FC_LBIT)) { /* Need to wait for FAN - use discovery timer * for timeout. hba_state is identically * LPFC_LOCAL_CFG_LINK while waiting for FAN */ lpfc_set_disctmo(phba); mempool_free( pmb, phba->mbox_mem_pool); return; } } /* Start discovery by sending a FLOGI hba_state is identically * LPFC_FLOGI while waiting for FLOGI cmpl */ phba->hba_state = LPFC_FLOGI; lpfc_set_disctmo(phba); lpfc_initial_flogi(phba); mempool_free( pmb, phba->mbox_mem_pool); return; } if (phba->hba_state == LPFC_FABRIC_CFG_LINK) { mempool_free( pmb, phba->mbox_mem_pool); return; } out: /* CONFIG_LINK bad hba state <hba_state> */ lpfc_printf_log(phba, KERN_ERR, LOG_DISCOVERY, "%d:0200 CONFIG_LINK bad hba state x%x\n", phba->brd_no, phba->hba_state); if (phba->hba_state != LPFC_CLEAR_LA) { lpfc_clear_la(phba, pmb); pmb->mbox_cmpl = lpfc_mbx_cmpl_clear_la; if (lpfc_sli_issue_mbox(phba, pmb, (MBX_NOWAIT | MBX_STOP_IOCB)) == MBX_NOT_FINISHED) { mempool_free( pmb, phba->mbox_mem_pool); lpfc_disc_flush_list(phba); psli->ring[(psli->ip_ring)].flag &= ~LPFC_STOP_IOCB_EVENT; psli->ring[(psli->fcp_ring)].flag &= ~LPFC_STOP_IOCB_EVENT; psli->ring[(psli->next_ring)].flag &= ~LPFC_STOP_IOCB_EVENT; phba->hba_state = LPFC_HBA_READY; } } else { mempool_free( pmb, phba->mbox_mem_pool); } return; } static void lpfc_mbx_cmpl_read_sparam(struct lpfc_hba * phba, LPFC_MBOXQ_t * pmb) { struct lpfc_sli *psli = &phba->sli; MAILBOX_t *mb = &pmb->mb; struct lpfc_dmabuf *mp = (struct lpfc_dmabuf *) pmb->context1; /* Check for error */ if (mb->mbxStatus) { /* READ_SPARAM mbox error <mbxStatus> state <hba_state> */ lpfc_printf_log(phba, KERN_ERR, LOG_MBOX, "%d:0319 READ_SPARAM mbxStatus error x%x " "hba state x%x>\n", phba->brd_no, mb->mbxStatus, phba->hba_state); lpfc_linkdown(phba); phba->hba_state = LPFC_HBA_ERROR; goto out; } memcpy((uint8_t *) & phba->fc_sparam, (uint8_t *) mp->virt, sizeof (struct serv_parm)); memcpy((uint8_t *) & phba->fc_nodename, (uint8_t *) & phba->fc_sparam.nodeName, sizeof (struct lpfc_name)); memcpy((uint8_t *) & phba->fc_portname, (uint8_t *) & phba->fc_sparam.portName, sizeof (struct lpfc_name)); lpfc_mbuf_free(phba, mp->virt, mp->phys); kfree(mp); mempool_free( pmb, phba->mbox_mem_pool); return; out: pmb->context1 = NULL; lpfc_mbuf_free(phba, mp->virt, mp->phys); kfree(mp); if (phba->hba_state != LPFC_CLEAR_LA) { lpfc_clear_la(phba, pmb); pmb->mbox_cmpl = lpfc_mbx_cmpl_clear_la; if (lpfc_sli_issue_mbox(phba, pmb, (MBX_NOWAIT | MBX_STOP_IOCB)) == MBX_NOT_FINISHED) { mempool_free( pmb, phba->mbox_mem_pool); lpfc_disc_flush_list(phba); psli->ring[(psli->ip_ring)].flag &= ~LPFC_STOP_IOCB_EVENT; psli->ring[(psli->fcp_ring)].flag &= ~LPFC_STOP_IOCB_EVENT; psli->ring[(psli->next_ring)].flag &= ~LPFC_STOP_IOCB_EVENT; phba->hba_state = LPFC_HBA_READY; } } else { mempool_free( pmb, phba->mbox_mem_pool); } return; } static void lpfc_mbx_process_link_up(struct lpfc_hba *phba, READ_LA_VAR *la) { int i; LPFC_MBOXQ_t *sparam_mbox, *cfglink_mbox; sparam_mbox = mempool_alloc(phba->mbox_mem_pool, GFP_KERNEL); cfglink_mbox = mempool_alloc(phba->mbox_mem_pool, GFP_KERNEL); spin_lock_irq(phba->host->host_lock); switch(la->UlnkSpeed) { case LA_1GHZ_LINK: phba->fc_linkspeed = LA_1GHZ_LINK; break; case LA_2GHZ_LINK: phba->fc_linkspeed = LA_2GHZ_LINK; break; case LA_4GHZ_LINK: phba->fc_linkspeed = LA_4GHZ_LINK; break; default: phba->fc_linkspeed = LA_UNKNW_LINK; break; } phba->fc_topology = la->topology; if (phba->fc_topology == TOPOLOGY_LOOP) { /* Get Loop Map information */ if (la->il) phba->fc_flag |= FC_LBIT; phba->fc_myDID = la->granted_AL_PA; i = la->un.lilpBde64.tus.f.bdeSize; if (i == 0) { phba->alpa_map[0] = 0; } else { if (phba->cfg_log_verbose & LOG_LINK_EVENT) { int numalpa, j, k; union { uint8_t pamap[16]; struct { uint32_t wd1; uint32_t wd2; uint32_t wd3; uint32_t wd4; } pa; } un; numalpa = phba->alpa_map[0]; j = 0; while (j < numalpa) { memset(un.pamap, 0, 16); for (k = 1; j < numalpa; k++) { un.pamap[k - 1] = phba->alpa_map[j + 1]; j++; if (k == 16) break; } /* Link Up Event ALPA map */ lpfc_printf_log(phba, KERN_WARNING, LOG_LINK_EVENT, "%d:1304 Link Up Event " "ALPA map Data: x%x " "x%x x%x x%x\n", phba->brd_no, un.pa.wd1, un.pa.wd2, un.pa.wd3, un.pa.wd4); } } } } else { phba->fc_myDID = phba->fc_pref_DID; phba->fc_flag |= FC_LBIT; } spin_unlock_irq(phba->host->host_lock); lpfc_linkup(phba); if (sparam_mbox) { lpfc_read_sparam(phba, sparam_mbox); sparam_mbox->mbox_cmpl = lpfc_mbx_cmpl_read_sparam; lpfc_sli_issue_mbox(phba, sparam_mbox, (MBX_NOWAIT | MBX_STOP_IOCB)); } if (cfglink_mbox) { phba->hba_state = LPFC_LOCAL_CFG_LINK; lpfc_config_link(phba, cfglink_mbox); cfglink_mbox->mbox_cmpl = lpfc_mbx_cmpl_config_link; lpfc_sli_issue_mbox(phba, cfglink_mbox, (MBX_NOWAIT | MBX_STOP_IOCB)); } } static void lpfc_mbx_issue_link_down(struct lpfc_hba *phba) { uint32_t control; struct lpfc_sli *psli = &phba->sli; lpfc_linkdown(phba); /* turn on Link Attention interrupts - no CLEAR_LA needed */ spin_lock_irq(phba->host->host_lock); psli->sli_flag |= LPFC_PROCESS_LA; control = readl(phba->HCregaddr); control |= HC_LAINT_ENA; writel(control, phba->HCregaddr); readl(phba->HCregaddr); /* flush */ spin_unlock_irq(phba->host->host_lock); } /* * This routine handles processing a READ_LA mailbox * command upon completion. It is setup in the LPFC_MBOXQ * as the completion routine when the command is * handed off to the SLI layer. */ void lpfc_mbx_cmpl_read_la(struct lpfc_hba * phba, LPFC_MBOXQ_t * pmb) { READ_LA_VAR *la; MAILBOX_t *mb = &pmb->mb; struct lpfc_dmabuf *mp = (struct lpfc_dmabuf *) (pmb->context1); /* Check for error */ if (mb->mbxStatus) { lpfc_printf_log(phba, KERN_INFO, LOG_LINK_EVENT, "%d:1307 READ_LA mbox error x%x state x%x\n", phba->brd_no, mb->mbxStatus, phba->hba_state); lpfc_mbx_issue_link_down(phba); phba->hba_state = LPFC_HBA_ERROR; goto lpfc_mbx_cmpl_read_la_free_mbuf; } la = (READ_LA_VAR *) & pmb->mb.un.varReadLA; memcpy(&phba->alpa_map[0], mp->virt, 128); if (((phba->fc_eventTag + 1) < la->eventTag) || (phba->fc_eventTag == la->eventTag)) { phba->fc_stat.LinkMultiEvent++; if (la->attType == AT_LINK_UP) { if (phba->fc_eventTag != 0) lpfc_linkdown(phba); } } phba->fc_eventTag = la->eventTag; if (la->attType == AT_LINK_UP) { phba->fc_stat.LinkUp++; lpfc_printf_log(phba, KERN_ERR, LOG_LINK_EVENT, "%d:1303 Link Up Event x%x received " "Data: x%x x%x x%x x%x\n", phba->brd_no, la->eventTag, phba->fc_eventTag, la->granted_AL_PA, la->UlnkSpeed, phba->alpa_map[0]); lpfc_mbx_process_link_up(phba, la); } else { phba->fc_stat.LinkDown++; lpfc_printf_log(phba, KERN_ERR, LOG_LINK_EVENT, "%d:1305 Link Down Event x%x received " "Data: x%x x%x x%x\n", phba->brd_no, la->eventTag, phba->fc_eventTag, phba->hba_state, phba->fc_flag); lpfc_mbx_issue_link_down(phba); } lpfc_mbx_cmpl_read_la_free_mbuf: lpfc_mbuf_free(phba, mp->virt, mp->phys); kfree(mp); mempool_free(pmb, phba->mbox_mem_pool); return; } /* * This routine handles processing a REG_LOGIN mailbox * command upon completion. It is setup in the LPFC_MBOXQ * as the completion routine when the command is * handed off to the SLI layer. */ void lpfc_mbx_cmpl_reg_login(struct lpfc_hba * phba, LPFC_MBOXQ_t * pmb) { struct lpfc_sli *psli; MAILBOX_t *mb; struct lpfc_dmabuf *mp; struct lpfc_nodelist *ndlp; psli = &phba->sli; mb = &pmb->mb; ndlp = (struct lpfc_nodelist *) pmb->context2; mp = (struct lpfc_dmabuf *) (pmb->context1); pmb->context1 = NULL; /* Good status, call state machine */ lpfc_disc_state_machine(phba, ndlp, pmb, NLP_EVT_CMPL_REG_LOGIN); lpfc_mbuf_free(phba, mp->virt, mp->phys); kfree(mp); mempool_free( pmb, phba->mbox_mem_pool); return; } /* * This routine handles processing a Fabric REG_LOGIN mailbox * command upon completion. It is setup in the LPFC_MBOXQ * as the completion routine when the command is * handed off to the SLI layer. */ void lpfc_mbx_cmpl_fabric_reg_login(struct lpfc_hba * phba, LPFC_MBOXQ_t * pmb) { struct lpfc_sli *psli; MAILBOX_t *mb; struct lpfc_dmabuf *mp; struct lpfc_nodelist *ndlp; struct lpfc_nodelist *ndlp_fdmi; psli = &phba->sli; mb = &pmb->mb; ndlp = (struct lpfc_nodelist *) pmb->context2; mp = (struct lpfc_dmabuf *) (pmb->context1); if (mb->mbxStatus) { lpfc_mbuf_free(phba, mp->virt, mp->phys); kfree(mp); mempool_free( pmb, phba->mbox_mem_pool); mempool_free( ndlp, phba->nlp_mem_pool); /* FLOGI failed, so just use loop map to make discovery list */ lpfc_disc_list_loopmap(phba); /* Start discovery */ lpfc_disc_start(phba); return; } pmb->context1 = NULL; if (ndlp->nlp_rpi != 0) lpfc_findnode_remove_rpi(phba, ndlp->nlp_rpi); ndlp->nlp_rpi = mb->un.varWords[0]; lpfc_addnode_rpi(phba, ndlp, ndlp->nlp_rpi); ndlp->nlp_type |= NLP_FABRIC; ndlp->nlp_state = NLP_STE_UNMAPPED_NODE; lpfc_nlp_list(phba, ndlp, NLP_UNMAPPED_LIST); if (phba->hba_state == LPFC_FABRIC_CFG_LINK) { /* This NPort has been assigned an NPort_ID by the fabric as a * result of the completed fabric login. Issue a State Change * Registration (SCR) ELS request to the fabric controller * (SCR_DID) so that this NPort gets RSCN events from the * fabric. */ lpfc_issue_els_scr(phba, SCR_DID, 0); /* Allocate a new node instance. If the pool is empty, just * start the discovery process and skip the Nameserver login * process. This is attempted again later on. Otherwise, issue * a Port Login (PLOGI) to the NameServer */ if ((ndlp = mempool_alloc(phba->nlp_mem_pool, GFP_KERNEL)) == 0) { lpfc_disc_start(phba); } else { lpfc_nlp_init(phba, ndlp, NameServer_DID); ndlp->nlp_type |= NLP_FABRIC; ndlp->nlp_state = NLP_STE_PLOGI_ISSUE; lpfc_nlp_list(phba, ndlp, NLP_PLOGI_LIST); lpfc_issue_els_plogi(phba, ndlp, 0); if (phba->cfg_fdmi_on) { if ((ndlp_fdmi = mempool_alloc( phba->nlp_mem_pool, GFP_KERNEL))) { lpfc_nlp_init(phba, ndlp_fdmi, FDMI_DID); ndlp_fdmi->nlp_type |= NLP_FABRIC; ndlp_fdmi->nlp_state = NLP_STE_PLOGI_ISSUE; lpfc_issue_els_plogi(phba, ndlp_fdmi, 0); } } } } lpfc_mbuf_free(phba, mp->virt, mp->phys); kfree(mp); mempool_free( pmb, phba->mbox_mem_pool); return; } /* * This routine handles processing a NameServer REG_LOGIN mailbox * command upon completion. It is setup in the LPFC_MBOXQ * as the completion routine when the command is * handed off to the SLI layer. */ void lpfc_mbx_cmpl_ns_reg_login(struct lpfc_hba * phba, LPFC_MBOXQ_t * pmb) { struct lpfc_sli *psli; MAILBOX_t *mb; struct lpfc_dmabuf *mp; struct lpfc_nodelist *ndlp; psli = &phba->sli; mb = &pmb->mb; ndlp = (struct lpfc_nodelist *) pmb->context2; mp = (struct lpfc_dmabuf *) (pmb->context1); if (mb->mbxStatus) { lpfc_mbuf_free(phba, mp->virt, mp->phys); kfree(mp); mempool_free( pmb, phba->mbox_mem_pool); lpfc_nlp_list(phba, ndlp, NLP_NO_LIST); /* RegLogin failed, so just use loop map to make discovery list */ lpfc_disc_list_loopmap(phba); /* Start discovery */ lpfc_disc_start(phba); return; } pmb->context1 = NULL; if (ndlp->nlp_rpi != 0) lpfc_findnode_remove_rpi(phba, ndlp->nlp_rpi); ndlp->nlp_rpi = mb->un.varWords[0]; lpfc_addnode_rpi(phba, ndlp, ndlp->nlp_rpi); ndlp->nlp_type |= NLP_FABRIC; ndlp->nlp_state = NLP_STE_UNMAPPED_NODE; lpfc_nlp_list(phba, ndlp, NLP_UNMAPPED_LIST); if (phba->hba_state < LPFC_HBA_READY) { /* Link up discovery requires Fabrib registration. */ lpfc_ns_cmd(phba, ndlp, SLI_CTNS_RNN_ID); lpfc_ns_cmd(phba, ndlp, SLI_CTNS_RSNN_NN); lpfc_ns_cmd(phba, ndlp, SLI_CTNS_RFT_ID); } phba->fc_ns_retry = 0; /* Good status, issue CT Request to NameServer */ if (lpfc_ns_cmd(phba, ndlp, SLI_CTNS_GID_FT)) { /* Cannot issue NameServer Query, so finish up discovery */ lpfc_disc_start(phba); } lpfc_mbuf_free(phba, mp->virt, mp->phys); kfree(mp); mempool_free( pmb, phba->mbox_mem_pool); return; } static void lpfc_register_remote_port(struct lpfc_hba * phba, struct lpfc_nodelist * ndlp) { struct fc_rport *rport; struct lpfc_rport_data *rdata; struct fc_rport_identifiers rport_ids; /* Remote port has reappeared. Re-register w/ FC transport */ rport_ids.node_name = wwn_to_u64(ndlp->nlp_nodename.u.wwn); rport_ids.port_name = wwn_to_u64(ndlp->nlp_portname.u.wwn); rport_ids.port_id = ndlp->nlp_DID; rport_ids.roles = FC_RPORT_ROLE_UNKNOWN; if (ndlp->nlp_type & NLP_FCP_TARGET) rport_ids.roles |= FC_RPORT_ROLE_FCP_TARGET; if (ndlp->nlp_type & NLP_FCP_INITIATOR) rport_ids.roles |= FC_RPORT_ROLE_FCP_INITIATOR; ndlp->rport = rport = fc_remote_port_add(phba->host, 0, &rport_ids); if (!rport) { dev_printk(KERN_WARNING, &phba->pcidev->dev, "Warning: fc_remote_port_add failed\n"); return; } /* initialize static port data */ rport->maxframe_size = ndlp->nlp_maxframe; rport->supported_classes = ndlp->nlp_class_sup; if ((rport->scsi_target_id != -1) && (rport->scsi_target_id < MAX_FCP_TARGET)) { ndlp->nlp_sid = rport->scsi_target_id; } rdata = rport->dd_data; rdata->pnode = ndlp; return; } int lpfc_nlp_list(struct lpfc_hba * phba, struct lpfc_nodelist * nlp, int list) { enum { none, unmapped, mapped } rport_add = none, rport_del = none; struct lpfc_sli *psli; psli = &phba->sli; /* Sanity check to ensure we are not moving to / from the same list */ if ((nlp->nlp_flag & NLP_LIST_MASK) == list) { if (list != NLP_NO_LIST) return(0); } switch(nlp->nlp_flag & NLP_LIST_MASK) { case NLP_NO_LIST: /* Not on any list */ break; case NLP_UNUSED_LIST: phba->fc_unused_cnt--; list_del(&nlp->nlp_listp); break; case NLP_PLOGI_LIST: phba->fc_plogi_cnt--; list_del(&nlp->nlp_listp); break; case NLP_ADISC_LIST: phba->fc_adisc_cnt--; list_del(&nlp->nlp_listp); break; case NLP_REGLOGIN_LIST: phba->fc_reglogin_cnt--; list_del(&nlp->nlp_listp); break; case NLP_PRLI_LIST: phba->fc_prli_cnt--; list_del(&nlp->nlp_listp); break; case NLP_UNMAPPED_LIST: phba->fc_unmap_cnt--; list_del(&nlp->nlp_listp); spin_lock_irq(phba->host->host_lock); nlp->nlp_flag &= ~NLP_TGT_NO_SCSIID; nlp->nlp_type &= ~NLP_FC_NODE; spin_unlock_irq(phba->host->host_lock); phba->nport_event_cnt++; if (nlp->rport) rport_del = unmapped; break; case NLP_MAPPED_LIST: phba->fc_map_cnt--; list_del(&nlp->nlp_listp); phba->nport_event_cnt++; if (nlp->rport) rport_del = mapped; break; case NLP_NPR_LIST: phba->fc_npr_cnt--; list_del(&nlp->nlp_listp); /* Stop delay tmo if taking node off NPR list */ if ((nlp->nlp_flag & NLP_DELAY_TMO) && (list != NLP_NPR_LIST)) { spin_lock_irq(phba->host->host_lock); nlp->nlp_flag &= ~NLP_DELAY_TMO; spin_unlock_irq(phba->host->host_lock); del_timer_sync(&nlp->nlp_delayfunc); if (!list_empty(&nlp->els_retry_evt.evt_listp)) list_del_init(&nlp->els_retry_evt.evt_listp); } break; } spin_lock_irq(phba->host->host_lock); nlp->nlp_flag &= ~NLP_LIST_MASK; spin_unlock_irq(phba->host->host_lock); /* Add NPort <did> to <num> list */ lpfc_printf_log(phba, KERN_INFO, LOG_NODE, "%d:0904 Add NPort x%x to %d list Data: x%x\n", phba->brd_no, nlp->nlp_DID, list, nlp->nlp_flag); switch(list) { case NLP_NO_LIST: /* No list, just remove it */ lpfc_nlp_remove(phba, nlp); /* as node removed - stop further transport calls */ rport_del = none; break; case NLP_UNUSED_LIST: spin_lock_irq(phba->host->host_lock); nlp->nlp_flag |= list; spin_unlock_irq(phba->host->host_lock); /* Put it at the end of the unused list */ list_add_tail(&nlp->nlp_listp, &phba->fc_unused_list); phba->fc_unused_cnt++; break; case NLP_PLOGI_LIST: spin_lock_irq(phba->host->host_lock); nlp->nlp_flag |= list; spin_unlock_irq(phba->host->host_lock); /* Put it at the end of the plogi list */ list_add_tail(&nlp->nlp_listp, &phba->fc_plogi_list); phba->fc_plogi_cnt++; break; case NLP_ADISC_LIST: spin_lock_irq(phba->host->host_lock); nlp->nlp_flag |= list; spin_unlock_irq(phba->host->host_lock); /* Put it at the end of the adisc list */ list_add_tail(&nlp->nlp_listp, &phba->fc_adisc_list); phba->fc_adisc_cnt++; break; case NLP_REGLOGIN_LIST: spin_lock_irq(phba->host->host_lock); nlp->nlp_flag |= list; spin_unlock_irq(phba->host->host_lock); /* Put it at the end of the reglogin list */ list_add_tail(&nlp->nlp_listp, &phba->fc_reglogin_list); phba->fc_reglogin_cnt++; break; case NLP_PRLI_LIST: spin_lock_irq(phba->host->host_lock); nlp->nlp_flag |= list; spin_unlock_irq(phba->host->host_lock); /* Put it at the end of the prli list */ list_add_tail(&nlp->nlp_listp, &phba->fc_prli_list); phba->fc_prli_cnt++; break; case NLP_UNMAPPED_LIST: rport_add = unmapped; /* ensure all vestiges of "mapped" significance are gone */ nlp->nlp_type &= ~(NLP_FCP_TARGET | NLP_FCP_INITIATOR); spin_lock_irq(phba->host->host_lock); nlp->nlp_flag |= list; spin_unlock_irq(phba->host->host_lock); /* Put it at the end of the unmap list */ list_add_tail(&nlp->nlp_listp, &phba->fc_nlpunmap_list); phba->fc_unmap_cnt++; phba->nport_event_cnt++; /* stop nodev tmo if running */ if (nlp->nlp_flag & NLP_NODEV_TMO) { spin_lock_irq(phba->host->host_lock); nlp->nlp_flag &= ~NLP_NODEV_TMO; spin_unlock_irq(phba->host->host_lock); del_timer_sync(&nlp->nlp_tmofunc); if (!list_empty(&nlp->nodev_timeout_evt.evt_listp)) list_del_init(&nlp->nodev_timeout_evt. evt_listp); } nlp->nlp_type |= NLP_FC_NODE; break; case NLP_MAPPED_LIST: rport_add = mapped; spin_lock_irq(phba->host->host_lock); nlp->nlp_flag |= list; spin_unlock_irq(phba->host->host_lock); /* Put it at the end of the map list */ list_add_tail(&nlp->nlp_listp, &phba->fc_nlpmap_list); phba->fc_map_cnt++; phba->nport_event_cnt++; /* stop nodev tmo if running */ if (nlp->nlp_flag & NLP_NODEV_TMO) { nlp->nlp_flag &= ~NLP_NODEV_TMO; del_timer_sync(&nlp->nlp_tmofunc); if (!list_empty(&nlp->nodev_timeout_evt.evt_listp)) list_del_init(&nlp->nodev_timeout_evt. evt_listp); } break; case NLP_NPR_LIST: spin_lock_irq(phba->host->host_lock); nlp->nlp_flag |= list; spin_unlock_irq(phba->host->host_lock); /* Put it at the end of the npr list */ list_add_tail(&nlp->nlp_listp, &phba->fc_npr_list); phba->fc_npr_cnt++; /* * Sanity check for Fabric entity. * Set nodev_tmo for NPR state, for Fabric use 1 sec. */ if (nlp->nlp_type & NLP_FABRIC) { mod_timer(&nlp->nlp_tmofunc, jiffies + HZ); } else { mod_timer(&nlp->nlp_tmofunc, jiffies + HZ * phba->cfg_nodev_tmo); } spin_lock_irq(phba->host->host_lock); nlp->nlp_flag |= NLP_NODEV_TMO; nlp->nlp_flag &= ~NLP_RCV_PLOGI; spin_unlock_irq(phba->host->host_lock); break; case NLP_JUST_DQ: break; } /* * We make all the calls into the transport after we have * moved the node between lists. This so that we don't * release the lock while in-between lists. */ /* Don't upcall midlayer if we're unloading */ if (!(phba->fc_flag & FC_UNLOADING)) { /* * We revalidate the rport pointer as the "add" function * may have removed the remote port. */ if ((rport_del != none) && nlp->rport) fc_remote_port_block(nlp->rport); if (rport_add != none) { /* * Tell the fc transport about the port, if we haven't * already. If we have, and it's a scsi entity, be * sure to unblock any attached scsi devices */ if (!nlp->rport) lpfc_register_remote_port(phba, nlp); else fc_remote_port_unblock(nlp->rport); /* * if we added to Mapped list, but the remote port * registration failed or assigned a target id outside * our presentable range - move the node to the * Unmapped List */ if ((rport_add == mapped) && ((!nlp->rport) || (nlp->rport->scsi_target_id == -1) || (nlp->rport->scsi_target_id >= MAX_FCP_TARGET))) { nlp->nlp_state = NLP_STE_UNMAPPED_NODE; spin_lock_irq(phba->host->host_lock); nlp->nlp_flag |= NLP_TGT_NO_SCSIID; spin_unlock_irq(phba->host->host_lock); lpfc_nlp_list(phba, nlp, NLP_UNMAPPED_LIST); } } } return (0); } /* * Start / ReStart rescue timer for Discovery / RSCN handling */ void lpfc_set_disctmo(struct lpfc_hba * phba) { uint32_t tmo; tmo = ((phba->fc_ratov * 2) + 1); mod_timer(&phba->fc_disctmo, jiffies + HZ * tmo); spin_lock_irq(phba->host->host_lock); phba->fc_flag |= FC_DISC_TMO; spin_unlock_irq(phba->host->host_lock); /* Start Discovery Timer state <hba_state> */ lpfc_printf_log(phba, KERN_INFO, LOG_DISCOVERY, "%d:0247 Start Discovery Timer state x%x " "Data: x%x x%lx x%x x%x\n", phba->brd_no, phba->hba_state, tmo, (unsigned long)&phba->fc_disctmo, phba->fc_plogi_cnt, phba->fc_adisc_cnt); return; } /* * Cancel rescue timer for Discovery / RSCN handling */ int lpfc_can_disctmo(struct lpfc_hba * phba) { /* Turn off discovery timer if its running */ if (phba->fc_flag & FC_DISC_TMO) { spin_lock_irq(phba->host->host_lock); phba->fc_flag &= ~FC_DISC_TMO; spin_unlock_irq(phba->host->host_lock); del_timer_sync(&phba->fc_disctmo); phba->work_hba_events &= ~WORKER_DISC_TMO; } /* Cancel Discovery Timer state <hba_state> */ lpfc_printf_log(phba, KERN_INFO, LOG_DISCOVERY, "%d:0248 Cancel Discovery Timer state x%x " "Data: x%x x%x x%x\n", phba->brd_no, phba->hba_state, phba->fc_flag, phba->fc_plogi_cnt, phba->fc_adisc_cnt); return (0); } /* * Check specified ring for outstanding IOCB on the SLI queue * Return true if iocb matches the specified nport */ int lpfc_check_sli_ndlp(struct lpfc_hba * phba, struct lpfc_sli_ring * pring, struct lpfc_iocbq * iocb, struct lpfc_nodelist * ndlp) { struct lpfc_sli *psli; IOCB_t *icmd; psli = &phba->sli; icmd = &iocb->iocb; if (pring->ringno == LPFC_ELS_RING) { switch (icmd->ulpCommand) { case CMD_GEN_REQUEST64_CR: if (icmd->ulpContext == (volatile ushort)ndlp->nlp_rpi) return (1); case CMD_ELS_REQUEST64_CR: case CMD_XMIT_ELS_RSP64_CX: if (iocb->context1 == (uint8_t *) ndlp) return (1); } } else if (pring->ringno == psli->ip_ring) { } else if (pring->ringno == psli->fcp_ring) { /* Skip match check if waiting to relogin to FCP target */ if ((ndlp->nlp_type & NLP_FCP_TARGET) && (ndlp->nlp_flag & NLP_DELAY_TMO)) { return (0); } if (icmd->ulpContext == (volatile ushort)ndlp->nlp_rpi) { return (1); } } else if (pring->ringno == psli->next_ring) { } return (0); } /* * Free resources / clean up outstanding I/Os * associated with nlp_rpi in the LPFC_NODELIST entry. */ static int lpfc_no_rpi(struct lpfc_hba * phba, struct lpfc_nodelist * ndlp) { struct lpfc_sli *psli; struct lpfc_sli_ring *pring; struct lpfc_iocbq *iocb, *next_iocb; IOCB_t *icmd; uint32_t rpi, i; /* * Everything that matches on txcmplq will be returned * by firmware with a no rpi error. */ psli = &phba->sli; rpi = ndlp->nlp_rpi; if (rpi) { /* Now process each ring */ for (i = 0; i < psli->num_rings; i++) { pring = &psli->ring[i]; spin_lock_irq(phba->host->host_lock); list_for_each_entry_safe(iocb, next_iocb, &pring->txq, list) { /* * Check to see if iocb matches the nport we are * looking for */ if ((lpfc_check_sli_ndlp (phba, pring, iocb, ndlp))) { /* It matches, so deque and call compl with an error */ list_del(&iocb->list); pring->txq_cnt--; if (iocb->iocb_cmpl) { icmd = &iocb->iocb; icmd->ulpStatus = IOSTAT_LOCAL_REJECT; icmd->un.ulpWord[4] = IOERR_SLI_ABORTED; spin_unlock_irq(phba->host-> host_lock); (iocb->iocb_cmpl) (phba, iocb, iocb); spin_lock_irq(phba->host-> host_lock); } else { list_add_tail(&iocb->list, &phba->lpfc_iocb_list); } } } spin_unlock_irq(phba->host->host_lock); } } return (0); } /* * Free rpi associated with LPFC_NODELIST entry. * This routine is called from lpfc_freenode(), when we are removing * a LPFC_NODELIST entry. It is also called if the driver initiates a * LOGO that completes successfully, and we are waiting to PLOGI back * to the remote NPort. In addition, it is called after we receive * and unsolicated ELS cmd, send back a rsp, the rsp completes and * we are waiting to PLOGI back to the remote NPort. */ int lpfc_unreg_rpi(struct lpfc_hba * phba, struct lpfc_nodelist * ndlp) { LPFC_MBOXQ_t *mbox; int rc; if (ndlp->nlp_rpi) { if ((mbox = mempool_alloc(phba->mbox_mem_pool, GFP_KERNEL))) { lpfc_unreg_login(phba, ndlp->nlp_rpi, mbox); mbox->mbox_cmpl=lpfc_sli_def_mbox_cmpl; rc = lpfc_sli_issue_mbox (phba, mbox, (MBX_NOWAIT | MBX_STOP_IOCB)); if (rc == MBX_NOT_FINISHED) mempool_free( mbox, phba->mbox_mem_pool); } lpfc_findnode_remove_rpi(phba, ndlp->nlp_rpi); lpfc_no_rpi(phba, ndlp); ndlp->nlp_rpi = 0; return 1; } return 0; } /* * Free resources associated with LPFC_NODELIST entry * so it can be freed. */ static int lpfc_freenode(struct lpfc_hba * phba, struct lpfc_nodelist * ndlp) { LPFC_MBOXQ_t *mb; LPFC_MBOXQ_t *nextmb; struct lpfc_dmabuf *mp; struct fc_rport *rport; /* Cleanup node for NPort <nlp_DID> */ lpfc_printf_log(phba, KERN_INFO, LOG_NODE, "%d:0900 Cleanup node for NPort x%x " "Data: x%x x%x x%x\n", phba->brd_no, ndlp->nlp_DID, ndlp->nlp_flag, ndlp->nlp_state, ndlp->nlp_rpi); lpfc_nlp_list(phba, ndlp, NLP_JUST_DQ); /* * if unloading the driver - just leave the remote port in place. * The driver unload will force the attached devices to detach * and flush cache's w/o generating flush errors. */ if ((ndlp->rport) && !(phba->fc_flag & FC_UNLOADING)) { rport = ndlp->rport; ndlp->rport = NULL; fc_remote_port_unblock(rport); fc_remote_port_delete(rport); ndlp->nlp_sid = NLP_NO_SID; } /* cleanup any ndlp on mbox q waiting for reglogin cmpl */ if ((mb = phba->sli.mbox_active)) { if ((mb->mb.mbxCommand == MBX_REG_LOGIN64) && (ndlp == (struct lpfc_nodelist *) mb->context2)) { mb->context2 = NULL; mb->mbox_cmpl = lpfc_sli_def_mbox_cmpl; } } list_for_each_entry_safe(mb, nextmb, &phba->sli.mboxq, list) { if ((mb->mb.mbxCommand == MBX_REG_LOGIN64) && (ndlp == (struct lpfc_nodelist *) mb->context2)) { mp = (struct lpfc_dmabuf *) (mb->context1); if (mp) { lpfc_mbuf_free(phba, mp->virt, mp->phys); kfree(mp); } list_del(&mb->list); mempool_free(mb, phba->mbox_mem_pool); } } lpfc_els_abort(phba,ndlp,0); spin_lock_irq(phba->host->host_lock); ndlp->nlp_flag &= ~(NLP_NODEV_TMO|NLP_DELAY_TMO); spin_unlock_irq(phba->host->host_lock); del_timer_sync(&ndlp->nlp_tmofunc); del_timer_sync(&ndlp->nlp_delayfunc); if (!list_empty(&ndlp->nodev_timeout_evt.evt_listp)) list_del_init(&ndlp->nodev_timeout_evt.evt_listp); if (!list_empty(&ndlp->els_retry_evt.evt_listp)) list_del_init(&ndlp->els_retry_evt.evt_listp); lpfc_unreg_rpi(phba, ndlp); return (0); } /* * Check to see if we can free the nlp back to the freelist. * If we are in the middle of using the nlp in the discovery state * machine, defer the free till we reach the end of the state machine. */ int lpfc_nlp_remove(struct lpfc_hba * phba, struct lpfc_nodelist * ndlp) { if (ndlp->nlp_flag & NLP_NODEV_TMO) { spin_lock_irq(phba->host->host_lock); ndlp->nlp_flag &= ~NLP_NODEV_TMO; spin_unlock_irq(phba->host->host_lock); del_timer_sync(&ndlp->nlp_tmofunc); if (!list_empty(&ndlp->nodev_timeout_evt.evt_listp)) list_del_init(&ndlp->nodev_timeout_evt.evt_listp); } if (ndlp->nlp_flag & NLP_DELAY_TMO) { spin_lock_irq(phba->host->host_lock); ndlp->nlp_flag &= ~NLP_DELAY_TMO; spin_unlock_irq(phba->host->host_lock); del_timer_sync(&ndlp->nlp_delayfunc); if (!list_empty(&ndlp->els_retry_evt.evt_listp)) list_del_init(&ndlp->els_retry_evt.evt_listp); } if (ndlp->nlp_disc_refcnt) { spin_lock_irq(phba->host->host_lock); ndlp->nlp_flag |= NLP_DELAY_REMOVE; spin_unlock_irq(phba->host->host_lock); } else { lpfc_freenode(phba, ndlp); mempool_free( ndlp, phba->nlp_mem_pool); } return(0); } static int lpfc_matchdid(struct lpfc_hba * phba, struct lpfc_nodelist * ndlp, uint32_t did) { D_ID mydid; D_ID ndlpdid; D_ID matchdid; if (did == Bcast_DID) return (0); if (ndlp->nlp_DID == 0) { return (0); } /* First check for Direct match */ if (ndlp->nlp_DID == did) return (1); /* Next check for area/domain identically equals 0 match */ mydid.un.word = phba->fc_myDID; if ((mydid.un.b.domain == 0) && (mydid.un.b.area == 0)) { return (0); } matchdid.un.word = did; ndlpdid.un.word = ndlp->nlp_DID; if (matchdid.un.b.id == ndlpdid.un.b.id) { if ((mydid.un.b.domain == matchdid.un.b.domain) && (mydid.un.b.area == matchdid.un.b.area)) { if ((ndlpdid.un.b.domain == 0) && (ndlpdid.un.b.area == 0)) { if (ndlpdid.un.b.id) return (1); } return (0); } matchdid.un.word = ndlp->nlp_DID; if ((mydid.un.b.domain == ndlpdid.un.b.domain) && (mydid.un.b.area == ndlpdid.un.b.area)) { if ((matchdid.un.b.domain == 0) && (matchdid.un.b.area == 0)) { if (matchdid.un.b.id) return (1); } } } return (0); } /* Search for a nodelist entry on a specific list */ struct lpfc_nodelist * lpfc_findnode_did(struct lpfc_hba * phba, uint32_t order, uint32_t did) { struct lpfc_nodelist *ndlp, *next_ndlp; uint32_t data1; if (order & NLP_SEARCH_UNMAPPED) { list_for_each_entry_safe(ndlp, next_ndlp, &phba->fc_nlpunmap_list, nlp_listp) { if (lpfc_matchdid(phba, ndlp, did)) { data1 = (((uint32_t) ndlp->nlp_state << 24) | ((uint32_t) ndlp->nlp_xri << 16) | ((uint32_t) ndlp->nlp_type << 8) | ((uint32_t) ndlp->nlp_rpi & 0xff)); /* FIND node DID unmapped */ lpfc_printf_log(phba, KERN_INFO, LOG_NODE, "%d:0929 FIND node DID unmapped" " Data: x%p x%x x%x x%x\n", phba->brd_no, ndlp, ndlp->nlp_DID, ndlp->nlp_flag, data1); return (ndlp); } } } if (order & NLP_SEARCH_MAPPED) { list_for_each_entry_safe(ndlp, next_ndlp, &phba->fc_nlpmap_list, nlp_listp) { if (lpfc_matchdid(phba, ndlp, did)) { data1 = (((uint32_t) ndlp->nlp_state << 24) | ((uint32_t) ndlp->nlp_xri << 16) | ((uint32_t) ndlp->nlp_type << 8) | ((uint32_t) ndlp->nlp_rpi & 0xff)); /* FIND node DID mapped */ lpfc_printf_log(phba, KERN_INFO, LOG_NODE, "%d:0930 FIND node DID mapped " "Data: x%p x%x x%x x%x\n", phba->brd_no, ndlp, ndlp->nlp_DID, ndlp->nlp_flag, data1); return (ndlp); } } } if (order & NLP_SEARCH_PLOGI) { list_for_each_entry_safe(ndlp, next_ndlp, &phba->fc_plogi_list, nlp_listp) { if (lpfc_matchdid(phba, ndlp, did)) { data1 = (((uint32_t) ndlp->nlp_state << 24) | ((uint32_t) ndlp->nlp_xri << 16) | ((uint32_t) ndlp->nlp_type << 8) | ((uint32_t) ndlp->nlp_rpi & 0xff)); /* LOG change to PLOGI */ /* FIND node DID plogi */ lpfc_printf_log(phba, KERN_INFO, LOG_NODE, "%d:0908 FIND node DID plogi " "Data: x%p x%x x%x x%x\n", phba->brd_no, ndlp, ndlp->nlp_DID, ndlp->nlp_flag, data1); return (ndlp); } } } if (order & NLP_SEARCH_ADISC) { list_for_each_entry_safe(ndlp, next_ndlp, &phba->fc_adisc_list, nlp_listp) { if (lpfc_matchdid(phba, ndlp, did)) { data1 = (((uint32_t) ndlp->nlp_state << 24) | ((uint32_t) ndlp->nlp_xri << 16) | ((uint32_t) ndlp->nlp_type << 8) | ((uint32_t) ndlp->nlp_rpi & 0xff)); /* LOG change to ADISC */ /* FIND node DID adisc */ lpfc_printf_log(phba, KERN_INFO, LOG_NODE, "%d:0931 FIND node DID adisc " "Data: x%p x%x x%x x%x\n", phba->brd_no, ndlp, ndlp->nlp_DID, ndlp->nlp_flag, data1); return (ndlp); } } } if (order & NLP_SEARCH_REGLOGIN) { list_for_each_entry_safe(ndlp, next_ndlp, &phba->fc_reglogin_list, nlp_listp) { if (lpfc_matchdid(phba, ndlp, did)) { data1 = (((uint32_t) ndlp->nlp_state << 24) | ((uint32_t) ndlp->nlp_xri << 16) | ((uint32_t) ndlp->nlp_type << 8) | ((uint32_t) ndlp->nlp_rpi & 0xff)); /* LOG change to REGLOGIN */ /* FIND node DID reglogin */ lpfc_printf_log(phba, KERN_INFO, LOG_NODE, "%d:0931 FIND node DID reglogin" " Data: x%p x%x x%x x%x\n", phba->brd_no, ndlp, ndlp->nlp_DID, ndlp->nlp_flag, data1); return (ndlp); } } } if (order & NLP_SEARCH_PRLI) { list_for_each_entry_safe(ndlp, next_ndlp, &phba->fc_prli_list, nlp_listp) { if (lpfc_matchdid(phba, ndlp, did)) { data1 = (((uint32_t) ndlp->nlp_state << 24) | ((uint32_t) ndlp->nlp_xri << 16) | ((uint32_t) ndlp->nlp_type << 8) | ((uint32_t) ndlp->nlp_rpi & 0xff)); /* LOG change to PRLI */ /* FIND node DID prli */ lpfc_printf_log(phba, KERN_INFO, LOG_NODE, "%d:0931 FIND node DID prli " "Data: x%p x%x x%x x%x\n", phba->brd_no, ndlp, ndlp->nlp_DID, ndlp->nlp_flag, data1); return (ndlp); } } } if (order & NLP_SEARCH_NPR) { list_for_each_entry_safe(ndlp, next_ndlp, &phba->fc_npr_list, nlp_listp) { if (lpfc_matchdid(phba, ndlp, did)) { data1 = (((uint32_t) ndlp->nlp_state << 24) | ((uint32_t) ndlp->nlp_xri << 16) | ((uint32_t) ndlp->nlp_type << 8) | ((uint32_t) ndlp->nlp_rpi & 0xff)); /* LOG change to NPR */ /* FIND node DID npr */ lpfc_printf_log(phba, KERN_INFO, LOG_NODE, "%d:0931 FIND node DID npr " "Data: x%p x%x x%x x%x\n", phba->brd_no, ndlp, ndlp->nlp_DID, ndlp->nlp_flag, data1); return (ndlp); } } } if (order & NLP_SEARCH_UNUSED) { list_for_each_entry_safe(ndlp, next_ndlp, &phba->fc_adisc_list, nlp_listp) { if (lpfc_matchdid(phba, ndlp, did)) { data1 = (((uint32_t) ndlp->nlp_state << 24) | ((uint32_t) ndlp->nlp_xri << 16) | ((uint32_t) ndlp->nlp_type << 8) | ((uint32_t) ndlp->nlp_rpi & 0xff)); /* LOG change to UNUSED */ /* FIND node DID unused */ lpfc_printf_log(phba, KERN_INFO, LOG_NODE, "%d:0931 FIND node DID unused " "Data: x%p x%x x%x x%x\n", phba->brd_no, ndlp, ndlp->nlp_DID, ndlp->nlp_flag, data1); return (ndlp); } } } /* FIND node did <did> NOT FOUND */ lpfc_printf_log(phba, KERN_INFO, LOG_NODE, "%d:0932 FIND node did x%x NOT FOUND Data: x%x\n", phba->brd_no, did, order); /* no match found */ return NULL; } struct lpfc_nodelist * lpfc_setup_disc_node(struct lpfc_hba * phba, uint32_t did) { struct lpfc_nodelist *ndlp; uint32_t flg; if ((ndlp = lpfc_findnode_did(phba, NLP_SEARCH_ALL, did)) == 0) { if ((phba->hba_state == LPFC_HBA_READY) && ((lpfc_rscn_payload_check(phba, did) == 0))) return NULL; ndlp = (struct lpfc_nodelist *) mempool_alloc(phba->nlp_mem_pool, GFP_KERNEL); if (!ndlp) return NULL; lpfc_nlp_init(phba, ndlp, did); ndlp->nlp_state = NLP_STE_NPR_NODE; lpfc_nlp_list(phba, ndlp, NLP_NPR_LIST); ndlp->nlp_flag |= NLP_NPR_2B_DISC; return ndlp; } if ((phba->hba_state == LPFC_HBA_READY) && (phba->fc_flag & FC_RSCN_MODE)) { if (lpfc_rscn_payload_check(phba, did)) { ndlp->nlp_flag |= NLP_NPR_2B_DISC; } else { ndlp->nlp_flag &= ~NLP_NPR_2B_DISC; ndlp = NULL; } } else { flg = ndlp->nlp_flag & NLP_LIST_MASK; if ((flg == NLP_ADISC_LIST) || (flg == NLP_PLOGI_LIST)) { return NULL; } ndlp->nlp_state = NLP_STE_NPR_NODE; lpfc_nlp_list(phba, ndlp, NLP_NPR_LIST); ndlp->nlp_flag |= NLP_NPR_2B_DISC; } return ndlp; } /* Build a list of nodes to discover based on the loopmap */ void lpfc_disc_list_loopmap(struct lpfc_hba * phba) { int j; uint32_t alpa, index; if (phba->hba_state <= LPFC_LINK_DOWN) { return; } if (phba->fc_topology != TOPOLOGY_LOOP) { return; } /* Check for loop map present or not */ if (phba->alpa_map[0]) { for (j = 1; j <= phba->alpa_map[0]; j++) { alpa = phba->alpa_map[j]; if (((phba->fc_myDID & 0xff) == alpa) || (alpa == 0)) { continue; } lpfc_setup_disc_node(phba, alpa); } } else { /* No alpamap, so try all alpa's */ for (j = 0; j < FC_MAXLOOP; j++) { /* If cfg_scan_down is set, start from highest * ALPA (0xef) to lowest (0x1). */ if (phba->cfg_scan_down) index = j; else index = FC_MAXLOOP - j - 1; alpa = lpfcAlpaArray[index]; if ((phba->fc_myDID & 0xff) == alpa) { continue; } lpfc_setup_disc_node(phba, alpa); } } return; } /* Start Link up / RSCN discovery on NPR list */ void lpfc_disc_start(struct lpfc_hba * phba) { struct lpfc_sli *psli; LPFC_MBOXQ_t *mbox; struct lpfc_nodelist *ndlp, *next_ndlp; uint32_t did_changed, num_sent; uint32_t clear_la_pending; int rc; psli = &phba->sli; if (phba->hba_state <= LPFC_LINK_DOWN) { return; } if (phba->hba_state == LPFC_CLEAR_LA) clear_la_pending = 1; else clear_la_pending = 0; if (phba->hba_state < LPFC_HBA_READY) { phba->hba_state = LPFC_DISC_AUTH; } lpfc_set_disctmo(phba); if (phba->fc_prevDID == phba->fc_myDID) { did_changed = 0; } else { did_changed = 1; } phba->fc_prevDID = phba->fc_myDID; phba->num_disc_nodes = 0; /* Start Discovery state <hba_state> */ lpfc_printf_log(phba, KERN_INFO, LOG_DISCOVERY, "%d:0202 Start Discovery hba state x%x " "Data: x%x x%x x%x\n", phba->brd_no, phba->hba_state, phba->fc_flag, phba->fc_plogi_cnt, phba->fc_adisc_cnt); /* If our did changed, we MUST do PLOGI */ list_for_each_entry_safe(ndlp, next_ndlp, &phba->fc_npr_list, nlp_listp) { if (ndlp->nlp_flag & NLP_NPR_2B_DISC) { if (did_changed) { spin_lock_irq(phba->host->host_lock); ndlp->nlp_flag &= ~NLP_NPR_ADISC; spin_unlock_irq(phba->host->host_lock); } } } /* First do ADISCs - if any */ num_sent = lpfc_els_disc_adisc(phba); if (num_sent) return; if ((phba->hba_state < LPFC_HBA_READY) && (!clear_la_pending)) { /* If we get here, there is nothing to ADISC */ if ((mbox = mempool_alloc(phba->mbox_mem_pool, GFP_KERNEL))) { phba->hba_state = LPFC_CLEAR_LA; lpfc_clear_la(phba, mbox); mbox->mbox_cmpl = lpfc_mbx_cmpl_clear_la; rc = lpfc_sli_issue_mbox(phba, mbox, (MBX_NOWAIT | MBX_STOP_IOCB)); if (rc == MBX_NOT_FINISHED) { mempool_free( mbox, phba->mbox_mem_pool); lpfc_disc_flush_list(phba); psli->ring[(psli->ip_ring)].flag &= ~LPFC_STOP_IOCB_EVENT; psli->ring[(psli->fcp_ring)].flag &= ~LPFC_STOP_IOCB_EVENT; psli->ring[(psli->next_ring)].flag &= ~LPFC_STOP_IOCB_EVENT; phba->hba_state = LPFC_HBA_READY; } } } else { /* Next do PLOGIs - if any */ num_sent = lpfc_els_disc_plogi(phba); if (num_sent) return; if (phba->fc_flag & FC_RSCN_MODE) { /* Check to see if more RSCNs came in while we * were processing this one. */ if ((phba->fc_rscn_id_cnt == 0) && (!(phba->fc_flag & FC_RSCN_DISCOVERY))) { spin_lock_irq(phba->host->host_lock); phba->fc_flag &= ~FC_RSCN_MODE; spin_unlock_irq(phba->host->host_lock); } else lpfc_els_handle_rscn(phba); } } return; } /* * Ignore completion for all IOCBs on tx and txcmpl queue for ELS * ring the match the sppecified nodelist. */ static void lpfc_free_tx(struct lpfc_hba * phba, struct lpfc_nodelist * ndlp) { struct lpfc_sli *psli; IOCB_t *icmd; struct lpfc_iocbq *iocb, *next_iocb; struct lpfc_sli_ring *pring; struct lpfc_dmabuf *mp; psli = &phba->sli; pring = &psli->ring[LPFC_ELS_RING]; /* Error matching iocb on txq or txcmplq * First check the txq. */ list_for_each_entry_safe(iocb, next_iocb, &pring->txq, list) { if (iocb->context1 != ndlp) { continue; } icmd = &iocb->iocb; if ((icmd->ulpCommand == CMD_ELS_REQUEST64_CR) || (icmd->ulpCommand == CMD_XMIT_ELS_RSP64_CX)) { list_del(&iocb->list); pring->txq_cnt--; lpfc_els_free_iocb(phba, iocb); } } /* Next check the txcmplq */ list_for_each_entry_safe(iocb, next_iocb, &pring->txcmplq, list) { if (iocb->context1 != ndlp) { continue; } icmd = &iocb->iocb; if ((icmd->ulpCommand == CMD_ELS_REQUEST64_CR) || (icmd->ulpCommand == CMD_XMIT_ELS_RSP64_CX)) { iocb->iocb_cmpl = NULL; /* context2 = cmd, context2->next = rsp, context3 = bpl */ if (iocb->context2) { /* Free the response IOCB before handling the command. */ mp = (struct lpfc_dmabuf *) (iocb->context2); mp = list_get_first(&mp->list, struct lpfc_dmabuf, list); if (mp) { /* Delay before releasing rsp buffer to * give UNREG mbox a chance to take * effect. */ list_add(&mp->list, &phba->freebufList); } lpfc_mbuf_free(phba, ((struct lpfc_dmabuf *) iocb->context2)->virt, ((struct lpfc_dmabuf *) iocb->context2)->phys); kfree(iocb->context2); } if (iocb->context3) { lpfc_mbuf_free(phba, ((struct lpfc_dmabuf *) iocb->context3)->virt, ((struct lpfc_dmabuf *) iocb->context3)->phys); kfree(iocb->context3); } } } return; } void lpfc_disc_flush_list(struct lpfc_hba * phba) { struct lpfc_nodelist *ndlp, *next_ndlp; if (phba->fc_plogi_cnt) { list_for_each_entry_safe(ndlp, next_ndlp, &phba->fc_plogi_list, nlp_listp) { lpfc_free_tx(phba, ndlp); lpfc_nlp_remove(phba, ndlp); } } if (phba->fc_adisc_cnt) { list_for_each_entry_safe(ndlp, next_ndlp, &phba->fc_adisc_list, nlp_listp) { lpfc_free_tx(phba, ndlp); lpfc_nlp_remove(phba, ndlp); } } return; } /*****************************************************************************/ /* * NAME: lpfc_disc_timeout * * FUNCTION: Fibre Channel driver discovery timeout routine. * * EXECUTION ENVIRONMENT: interrupt only * * CALLED FROM: * Timer function * * RETURNS: * none */ /*****************************************************************************/ void lpfc_disc_timeout(unsigned long ptr) { struct lpfc_hba *phba = (struct lpfc_hba *)ptr; unsigned long flags = 0; if (unlikely(!phba)) return; spin_lock_irqsave(phba->host->host_lock, flags); if (!(phba->work_hba_events & WORKER_DISC_TMO)) { phba->work_hba_events |= WORKER_DISC_TMO; if (phba->work_wait) wake_up(phba->work_wait); } spin_unlock_irqrestore(phba->host->host_lock, flags); return; } static void lpfc_disc_timeout_handler(struct lpfc_hba *phba) { struct lpfc_sli *psli; struct lpfc_nodelist *ndlp; LPFC_MBOXQ_t *clearlambox, *initlinkmbox; int rc, clrlaerr = 0; if (unlikely(!phba)) return; if (!(phba->fc_flag & FC_DISC_TMO)) return; psli = &phba->sli; spin_lock_irq(phba->host->host_lock); phba->fc_flag &= ~FC_DISC_TMO; spin_unlock_irq(phba->host->host_lock); switch (phba->hba_state) { case LPFC_LOCAL_CFG_LINK: /* hba_state is identically LPFC_LOCAL_CFG_LINK while waiting for FAN */ /* FAN timeout */ lpfc_printf_log(phba, KERN_WARNING, LOG_DISCOVERY, "%d:0221 FAN timeout\n", phba->brd_no); /* Forget about FAN, Start discovery by sending a FLOGI * hba_state is identically LPFC_FLOGI while waiting for FLOGI * cmpl */ phba->hba_state = LPFC_FLOGI; lpfc_set_disctmo(phba); lpfc_initial_flogi(phba); break; case LPFC_FLOGI: /* hba_state is identically LPFC_FLOGI while waiting for FLOGI cmpl */ /* Initial FLOGI timeout */ lpfc_printf_log(phba, KERN_ERR, LOG_DISCOVERY, "%d:0222 Initial FLOGI timeout\n", phba->brd_no); /* Assume no Fabric and go on with discovery. * Check for outstanding ELS FLOGI to abort. */ /* FLOGI failed, so just use loop map to make discovery list */ lpfc_disc_list_loopmap(phba); /* Start discovery */ lpfc_disc_start(phba); break; case LPFC_FABRIC_CFG_LINK: /* hba_state is identically LPFC_FABRIC_CFG_LINK while waiting for NameServer login */ lpfc_printf_log(phba, KERN_ERR, LOG_DISCOVERY, "%d:0223 Timeout while waiting for NameServer " "login\n", phba->brd_no); /* Next look for NameServer ndlp */ ndlp = lpfc_findnode_did(phba, NLP_SEARCH_ALL, NameServer_DID); if (ndlp) lpfc_nlp_remove(phba, ndlp); /* Start discovery */ lpfc_disc_start(phba); break; case LPFC_NS_QRY: /* Check for wait for NameServer Rsp timeout */ lpfc_printf_log(phba, KERN_ERR, LOG_DISCOVERY, "%d:0224 NameServer Query timeout " "Data: x%x x%x\n", phba->brd_no, phba->fc_ns_retry, LPFC_MAX_NS_RETRY); ndlp = lpfc_findnode_did(phba, NLP_SEARCH_UNMAPPED, NameServer_DID); if (ndlp) { if (phba->fc_ns_retry < LPFC_MAX_NS_RETRY) { /* Try it one more time */ rc = lpfc_ns_cmd(phba, ndlp, SLI_CTNS_GID_FT); if (rc == 0) break; } phba->fc_ns_retry = 0; } /* Nothing to authenticate, so CLEAR_LA right now */ clearlambox = mempool_alloc(phba->mbox_mem_pool, GFP_KERNEL); if (!clearlambox) { clrlaerr = 1; lpfc_printf_log(phba, KERN_ERR, LOG_DISCOVERY, "%d:0226 Device Discovery " "completion error\n", phba->brd_no); phba->hba_state = LPFC_HBA_ERROR; break; } phba->hba_state = LPFC_CLEAR_LA; lpfc_clear_la(phba, clearlambox); clearlambox->mbox_cmpl = lpfc_mbx_cmpl_clear_la; rc = lpfc_sli_issue_mbox(phba, clearlambox, (MBX_NOWAIT | MBX_STOP_IOCB)); if (rc == MBX_NOT_FINISHED) { mempool_free(clearlambox, phba->mbox_mem_pool); clrlaerr = 1; break; } /* Setup and issue mailbox INITIALIZE LINK command */ initlinkmbox = mempool_alloc(phba->mbox_mem_pool, GFP_KERNEL); if (!initlinkmbox) { lpfc_printf_log(phba, KERN_ERR, LOG_DISCOVERY, "%d:0226 Device Discovery " "completion error\n", phba->brd_no); phba->hba_state = LPFC_HBA_ERROR; break; } lpfc_linkdown(phba); lpfc_init_link(phba, initlinkmbox, phba->cfg_topology, phba->cfg_link_speed); initlinkmbox->mb.un.varInitLnk.lipsr_AL_PA = 0; rc = lpfc_sli_issue_mbox(phba, initlinkmbox, (MBX_NOWAIT | MBX_STOP_IOCB)); if (rc == MBX_NOT_FINISHED) mempool_free(initlinkmbox, phba->mbox_mem_pool); break; case LPFC_DISC_AUTH: /* Node Authentication timeout */ lpfc_printf_log(phba, KERN_ERR, LOG_DISCOVERY, "%d:0227 Node Authentication timeout\n", phba->brd_no); lpfc_disc_flush_list(phba); clearlambox = mempool_alloc(phba->mbox_mem_pool, GFP_KERNEL); if (!clearlambox) { clrlaerr = 1; lpfc_printf_log(phba, KERN_ERR, LOG_DISCOVERY, "%d:0226 Device Discovery " "completion error\n", phba->brd_no); phba->hba_state = LPFC_HBA_ERROR; break; } phba->hba_state = LPFC_CLEAR_LA; lpfc_clear_la(phba, clearlambox); clearlambox->mbox_cmpl = lpfc_mbx_cmpl_clear_la; rc = lpfc_sli_issue_mbox(phba, clearlambox, (MBX_NOWAIT | MBX_STOP_IOCB)); if (rc == MBX_NOT_FINISHED) { mempool_free(clearlambox, phba->mbox_mem_pool); clrlaerr = 1; } break; case LPFC_CLEAR_LA: /* CLEAR LA timeout */ lpfc_printf_log(phba, KERN_ERR, LOG_DISCOVERY, "%d:0228 CLEAR LA timeout\n", phba->brd_no); clrlaerr = 1; break; case LPFC_HBA_READY: if (phba->fc_flag & FC_RSCN_MODE) { lpfc_printf_log(phba, KERN_ERR, LOG_DISCOVERY, "%d:0231 RSCN timeout Data: x%x x%x\n", phba->brd_no, phba->fc_ns_retry, LPFC_MAX_NS_RETRY); /* Cleanup any outstanding ELS commands */ lpfc_els_flush_cmd(phba); lpfc_els_flush_rscn(phba); lpfc_disc_flush_list(phba); } break; } if (clrlaerr) { lpfc_disc_flush_list(phba); psli->ring[(psli->ip_ring)].flag &= ~LPFC_STOP_IOCB_EVENT; psli->ring[(psli->fcp_ring)].flag &= ~LPFC_STOP_IOCB_EVENT; psli->ring[(psli->next_ring)].flag &= ~LPFC_STOP_IOCB_EVENT; phba->hba_state = LPFC_HBA_READY; } return; } static void lpfc_nodev_timeout(unsigned long ptr) { struct lpfc_hba *phba; struct lpfc_nodelist *ndlp; unsigned long iflag; struct lpfc_work_evt *evtp; ndlp = (struct lpfc_nodelist *)ptr; phba = ndlp->nlp_phba; evtp = &ndlp->nodev_timeout_evt; spin_lock_irqsave(phba->host->host_lock, iflag); if (!list_empty(&evtp->evt_listp)) { spin_unlock_irqrestore(phba->host->host_lock, iflag); return; } evtp->evt_arg1 = ndlp; evtp->evt = LPFC_EVT_NODEV_TMO; list_add_tail(&evtp->evt_listp, &phba->work_list); if (phba->work_wait) wake_up(phba->work_wait); spin_unlock_irqrestore(phba->host->host_lock, iflag); return; } /* * This routine handles processing a NameServer REG_LOGIN mailbox * command upon completion. It is setup in the LPFC_MBOXQ * as the completion routine when the command is * handed off to the SLI layer. */ void lpfc_mbx_cmpl_fdmi_reg_login(struct lpfc_hba * phba, LPFC_MBOXQ_t * pmb) { struct lpfc_sli *psli; MAILBOX_t *mb; struct lpfc_dmabuf *mp; struct lpfc_nodelist *ndlp; psli = &phba->sli; mb = &pmb->mb; ndlp = (struct lpfc_nodelist *) pmb->context2; mp = (struct lpfc_dmabuf *) (pmb->context1); pmb->context1 = NULL; if (ndlp->nlp_rpi != 0) lpfc_findnode_remove_rpi(phba, ndlp->nlp_rpi); ndlp->nlp_rpi = mb->un.varWords[0]; lpfc_addnode_rpi(phba, ndlp, ndlp->nlp_rpi); ndlp->nlp_type |= NLP_FABRIC; ndlp->nlp_state = NLP_STE_UNMAPPED_NODE; lpfc_nlp_list(phba, ndlp, NLP_UNMAPPED_LIST); /* Start issuing Fabric-Device Management Interface (FDMI) * command to 0xfffffa (FDMI well known port) */ if (phba->cfg_fdmi_on == 1) { lpfc_fdmi_cmd(phba, ndlp, SLI_MGMT_DHBA); } else { /* * Delay issuing FDMI command if fdmi-on=2 * (supporting RPA/hostnmae) */ mod_timer(&phba->fc_fdmitmo, jiffies + HZ * 60); } lpfc_mbuf_free(phba, mp->virt, mp->phys); kfree(mp); mempool_free( pmb, phba->mbox_mem_pool); return; } /* * This routine looks up the ndlp hash * table for the given RPI. If rpi found * it return the node list pointer * else return 0. */ struct lpfc_nodelist * lpfc_findnode_rpi(struct lpfc_hba * phba, uint16_t rpi) { struct lpfc_nodelist *ret; ret = phba->fc_nlplookup[LPFC_RPI_HASH_FUNC(rpi)]; while ((ret != 0) && (ret->nlp_rpi != rpi)) { ret = ret->nlp_rpi_hash_next; } return ret; } /* * This routine looks up the ndlp hash table for the * given RPI. If rpi found it return the node list * pointer else return 0 after deleting the entry * from hash table. */ struct lpfc_nodelist * lpfc_findnode_remove_rpi(struct lpfc_hba * phba, uint16_t rpi) { struct lpfc_nodelist *ret, *temp;; ret = phba->fc_nlplookup[LPFC_RPI_HASH_FUNC(rpi)]; if (ret == 0) return NULL; if (ret->nlp_rpi == rpi) { phba->fc_nlplookup[LPFC_RPI_HASH_FUNC(rpi)] = ret->nlp_rpi_hash_next; ret->nlp_rpi_hash_next = NULL; return ret; } while ((ret->nlp_rpi_hash_next != 0) && (ret->nlp_rpi_hash_next->nlp_rpi != rpi)) { ret = ret->nlp_rpi_hash_next; } if (ret->nlp_rpi_hash_next != 0) { temp = ret->nlp_rpi_hash_next; ret->nlp_rpi_hash_next = temp->nlp_rpi_hash_next; temp->nlp_rpi_hash_next = NULL; return temp; } else { return NULL; } } /* * This routine adds the node list entry to the * ndlp hash table. */ void lpfc_addnode_rpi(struct lpfc_hba * phba, struct lpfc_nodelist * ndlp, uint16_t rpi) { uint32_t index; index = LPFC_RPI_HASH_FUNC(rpi); ndlp->nlp_rpi_hash_next = phba->fc_nlplookup[index]; phba->fc_nlplookup[index] = ndlp; return; } void lpfc_nlp_init(struct lpfc_hba * phba, struct lpfc_nodelist * ndlp, uint32_t did) { memset(ndlp, 0, sizeof (struct lpfc_nodelist)); INIT_LIST_HEAD(&ndlp->nodev_timeout_evt.evt_listp); INIT_LIST_HEAD(&ndlp->els_retry_evt.evt_listp); init_timer(&ndlp->nlp_tmofunc); ndlp->nlp_tmofunc.function = lpfc_nodev_timeout; ndlp->nlp_tmofunc.data = (unsigned long)ndlp; init_timer(&ndlp->nlp_delayfunc); ndlp->nlp_delayfunc.function = lpfc_els_retry_delay; ndlp->nlp_delayfunc.data = (unsigned long)ndlp; ndlp->nlp_DID = did; ndlp->nlp_phba = phba; ndlp->nlp_sid = NLP_NO_SID; return; }