From 93bcc5f3984bf4f51da1529700aec351872dbfff Mon Sep 17 00:00:00 2001 From: Justin Tee Date: Thu, 12 Sep 2024 16:24:40 -0700 Subject: scsi: lpfc: Add ELS_RSP cmd to the list of WQEs to flush in lpfc_els_flush_cmd() During HBA stress testing, a spam of received PLOGIs exposes a resource recovery bug causing leakage of lpfc_sqlq entries from the global phba->sli4_hba.lpfc_els_sgl_list. The issue is in lpfc_els_flush_cmd(), where the driver attempts to recover outstanding ELS sgls when walking the txcmplq. Only CMD_ELS_REQUEST64_CRs and CMD_GEN_REQUEST64_CRs are added to the abort and cancel lists. A check for CMD_XMIT_ELS_RSP64_WQE is missing in order to recover LS_ACC usages of the phba->sli4_hba.lpfc_els_sgl_list too. Fix by adding CMD_XMIT_ELS_RSP64_WQE as part of the txcmplq walk when adding WQEs to the abort and cancel list in lpfc_els_flush_cmd(). Also, update naming convention from CRs to WQEs. Signed-off-by: Justin Tee Link: https://lore.kernel.org/r/20240912232447.45607-2-justintee8345@gmail.com Signed-off-by: Martin K. Petersen --- drivers/scsi/lpfc/lpfc_els.c | 7 ++++--- 1 file changed, 4 insertions(+), 3 deletions(-) (limited to 'drivers') diff --git a/drivers/scsi/lpfc/lpfc_els.c b/drivers/scsi/lpfc/lpfc_els.c index de0ec945d2f1..7219b1ada1ea 100644 --- a/drivers/scsi/lpfc/lpfc_els.c +++ b/drivers/scsi/lpfc/lpfc_els.c @@ -9658,11 +9658,12 @@ lpfc_els_flush_cmd(struct lpfc_vport *vport) if (piocb->cmd_flag & LPFC_DRIVER_ABORTED && !mbx_tmo_err) continue; - /* On the ELS ring we can have ELS_REQUESTs or - * GEN_REQUESTs waiting for a response. + /* On the ELS ring we can have ELS_REQUESTs, ELS_RSPs, + * or GEN_REQUESTs waiting for a CQE response. */ ulp_command = get_job_cmnd(phba, piocb); - if (ulp_command == CMD_ELS_REQUEST64_CR) { + if (ulp_command == CMD_ELS_REQUEST64_WQE || + ulp_command == CMD_XMIT_ELS_RSP64_WQE) { list_add_tail(&piocb->dlist, &abort_list); /* If the link is down when flushing ELS commands -- cgit v1.2.3 From fc318cac66ac50398f9fc7cbec7b339e6d08a7e6 Mon Sep 17 00:00:00 2001 From: Justin Tee Date: Thu, 12 Sep 2024 16:24:41 -0700 Subject: scsi: lpfc: Update phba link state conditional before sending CMF_SYNC_WQE It's possible for the driver to send a CMF_SYNC_WQE to nonresponsive firmware during reset of the adapter. The phba link_state conditional check is currently a strict == LPFC_LINK_DOWN, which does not cover initialization states before reaching the LPFC_LINK_UP state. Update the phba->link_state conditional to < LPFC_LINK_UP so that all initialization states are covered before allowing sending CMF_SYNC_WQE. Update taking of the hbalock to be during this link_state check as well. Signed-off-by: Justin Tee Link: https://lore.kernel.org/r/20240912232447.45607-3-justintee8345@gmail.com Signed-off-by: Martin K. Petersen --- drivers/scsi/lpfc/lpfc_sli.c | 9 ++++++--- 1 file changed, 6 insertions(+), 3 deletions(-) (limited to 'drivers') diff --git a/drivers/scsi/lpfc/lpfc_sli.c b/drivers/scsi/lpfc/lpfc_sli.c index 332b8d2348e9..bb5fd3322273 100644 --- a/drivers/scsi/lpfc/lpfc_sli.c +++ b/drivers/scsi/lpfc/lpfc_sli.c @@ -1940,12 +1940,15 @@ lpfc_issue_cmf_sync_wqe(struct lpfc_hba *phba, u32 ms, u64 total) atot = atomic_xchg(&phba->cgn_sync_alarm_cnt, 0); wtot = atomic_xchg(&phba->cgn_sync_warn_cnt, 0); + spin_lock_irqsave(&phba->hbalock, iflags); + /* ONLY Managed mode will send the CMF_SYNC_WQE to the HBA */ if (phba->cmf_active_mode != LPFC_CFG_MANAGED || - phba->link_state == LPFC_LINK_DOWN) - return 0; + phba->link_state < LPFC_LINK_UP) { + ret_val = 0; + goto out_unlock; + } - spin_lock_irqsave(&phba->hbalock, iflags); sync_buf = __lpfc_sli_get_iocbq(phba); if (!sync_buf) { lpfc_printf_log(phba, KERN_ERR, LOG_CGN_MGMT, -- cgit v1.2.3 From 05ab4e7846f1103377133c00295a9a910cc6dfc2 Mon Sep 17 00:00:00 2001 From: Justin Tee Date: Thu, 12 Sep 2024 16:24:42 -0700 Subject: scsi: lpfc: Restrict support for 32 byte CDBs to specific HBAs An older generation of HBAs are failing FCP discovery due to usage of an outdated field in FCP command WQEs. Fix by checking the SLI Interface Type register for applicable support of 32 Byte CDB commands, and restore a setting for a WQE path using normal 16 byte CDBs. Fixes: af20bb73ac25 ("scsi: lpfc: Add support for 32 byte CDBs") Cc: stable@vger.kernel.org # v6.10+ Signed-off-by: Justin Tee Link: https://lore.kernel.org/r/20240912232447.45607-4-justintee8345@gmail.com Signed-off-by: Martin K. Petersen --- drivers/scsi/lpfc/lpfc_hw4.h | 3 +++ drivers/scsi/lpfc/lpfc_init.c | 21 ++++++++++++++++++--- drivers/scsi/lpfc/lpfc_scsi.c | 2 +- 3 files changed, 22 insertions(+), 4 deletions(-) (limited to 'drivers') diff --git a/drivers/scsi/lpfc/lpfc_hw4.h b/drivers/scsi/lpfc/lpfc_hw4.h index 500253007b1d..26e1313ebb21 100644 --- a/drivers/scsi/lpfc/lpfc_hw4.h +++ b/drivers/scsi/lpfc/lpfc_hw4.h @@ -4847,6 +4847,7 @@ struct fcp_iwrite64_wqe { #define cmd_buff_len_SHIFT 16 #define cmd_buff_len_MASK 0x00000ffff #define cmd_buff_len_WORD word3 +/* Note: payload_offset_len field depends on ASIC support */ #define payload_offset_len_SHIFT 0 #define payload_offset_len_MASK 0x0000ffff #define payload_offset_len_WORD word3 @@ -4863,6 +4864,7 @@ struct fcp_iread64_wqe { #define cmd_buff_len_SHIFT 16 #define cmd_buff_len_MASK 0x00000ffff #define cmd_buff_len_WORD word3 +/* Note: payload_offset_len field depends on ASIC support */ #define payload_offset_len_SHIFT 0 #define payload_offset_len_MASK 0x0000ffff #define payload_offset_len_WORD word3 @@ -4879,6 +4881,7 @@ struct fcp_icmnd64_wqe { #define cmd_buff_len_SHIFT 16 #define cmd_buff_len_MASK 0x00000ffff #define cmd_buff_len_WORD word3 +/* Note: payload_offset_len field depends on ASIC support */ #define payload_offset_len_SHIFT 0 #define payload_offset_len_MASK 0x0000ffff #define payload_offset_len_WORD word3 diff --git a/drivers/scsi/lpfc/lpfc_init.c b/drivers/scsi/lpfc/lpfc_init.c index e1dfa96c2a55..0c1404dc5f3b 100644 --- a/drivers/scsi/lpfc/lpfc_init.c +++ b/drivers/scsi/lpfc/lpfc_init.c @@ -4699,6 +4699,7 @@ lpfc_create_port(struct lpfc_hba *phba, int instance, struct device *dev) uint64_t wwn; bool use_no_reset_hba = false; int rc; + u8 if_type; if (lpfc_no_hba_reset_cnt) { if (phba->sli_rev < LPFC_SLI_REV4 && @@ -4773,10 +4774,24 @@ lpfc_create_port(struct lpfc_hba *phba, int instance, struct device *dev) shost->max_id = LPFC_MAX_TARGET; shost->max_lun = vport->cfg_max_luns; shost->this_id = -1; - if (phba->sli_rev == LPFC_SLI_REV4) - shost->max_cmd_len = LPFC_FCP_CDB_LEN_32; - else + + /* Set max_cmd_len applicable to ASIC support */ + if (phba->sli_rev == LPFC_SLI_REV4) { + if_type = bf_get(lpfc_sli_intf_if_type, + &phba->sli4_hba.sli_intf); + switch (if_type) { + case LPFC_SLI_INTF_IF_TYPE_2: + fallthrough; + case LPFC_SLI_INTF_IF_TYPE_6: + shost->max_cmd_len = LPFC_FCP_CDB_LEN_32; + break; + default: + shost->max_cmd_len = LPFC_FCP_CDB_LEN; + break; + } + } else { shost->max_cmd_len = LPFC_FCP_CDB_LEN; + } if (phba->sli_rev == LPFC_SLI_REV4) { if (!phba->cfg_fcp_mq_threshold || diff --git a/drivers/scsi/lpfc/lpfc_scsi.c b/drivers/scsi/lpfc/lpfc_scsi.c index 60cd60ebff38..0eaede8275da 100644 --- a/drivers/scsi/lpfc/lpfc_scsi.c +++ b/drivers/scsi/lpfc/lpfc_scsi.c @@ -4760,7 +4760,7 @@ static int lpfc_scsi_prep_cmnd_buf_s4(struct lpfc_vport *vport, /* Word 3 */ bf_set(payload_offset_len, &wqe->fcp_icmd, - sizeof(struct fcp_cmnd32) + sizeof(struct fcp_rsp)); + sizeof(struct fcp_cmnd) + sizeof(struct fcp_rsp)); /* Word 6 */ bf_set(wqe_ctxt_tag, &wqe->generic.wqe_com, -- cgit v1.2.3 From d1a2ef63fc8b3e3dc5b74b7e08636896b32acbdc Mon Sep 17 00:00:00 2001 From: Justin Tee Date: Thu, 12 Sep 2024 16:24:43 -0700 Subject: scsi: lpfc: Fix kref imbalance on fabric ndlps from dev_loss_tmo handler With a FLOGI outstanding and loss of physical link connection to the fabric for the duration of dev_loss_tmo, there is a fabric ndlp kref imbalance that decrements the kref and sets the NLP_IN_RECOV_POST_DEV_LOSS flag at the same time. The issue is that when the FLOGI completion routine executes, the fabric ndlp could already be freed because of the final kref put from the dev_loss_tmo handler. Fix by early returning before the ndlp kref put if the ndlp is deemed a candidate for NLP_IN_RECOV_POST_DEV_LOSS in the FLOGI completion routine. Signed-off-by: Justin Tee Link: https://lore.kernel.org/r/20240912232447.45607-5-justintee8345@gmail.com Signed-off-by: Martin K. Petersen --- drivers/scsi/lpfc/lpfc_hbadisc.c | 10 ++++------ 1 file changed, 4 insertions(+), 6 deletions(-) (limited to 'drivers') diff --git a/drivers/scsi/lpfc/lpfc_hbadisc.c b/drivers/scsi/lpfc/lpfc_hbadisc.c index 35c9181c6608..9241075f72fa 100644 --- a/drivers/scsi/lpfc/lpfc_hbadisc.c +++ b/drivers/scsi/lpfc/lpfc_hbadisc.c @@ -527,6 +527,9 @@ lpfc_dev_loss_tmo_handler(struct lpfc_nodelist *ndlp) * the following lpfc_nlp_put is necessary after fabric node is * recovered. */ + spin_lock_irqsave(&ndlp->lock, iflags); + ndlp->nlp_flag &= ~NLP_IN_DEV_LOSS; + spin_unlock_irqrestore(&ndlp->lock, iflags); if (recovering) { lpfc_printf_vlog(vport, KERN_INFO, LOG_DISCOVERY | LOG_NODE, @@ -539,6 +542,7 @@ lpfc_dev_loss_tmo_handler(struct lpfc_nodelist *ndlp) spin_lock_irqsave(&ndlp->lock, iflags); ndlp->save_flags |= NLP_IN_RECOV_POST_DEV_LOSS; spin_unlock_irqrestore(&ndlp->lock, iflags); + return fcf_inuse; } else if (ndlp->nlp_state == NLP_STE_UNMAPPED_NODE) { /* Fabric node fully recovered before this dev_loss_tmo * queue work is processed. Thus, ignore the @@ -552,15 +556,9 @@ lpfc_dev_loss_tmo_handler(struct lpfc_nodelist *ndlp) ndlp->nlp_DID, kref_read(&ndlp->kref), ndlp, ndlp->nlp_flag, vport->port_state); - spin_lock_irqsave(&ndlp->lock, iflags); - ndlp->nlp_flag &= ~NLP_IN_DEV_LOSS; - spin_unlock_irqrestore(&ndlp->lock, iflags); return fcf_inuse; } - spin_lock_irqsave(&ndlp->lock, iflags); - ndlp->nlp_flag &= ~NLP_IN_DEV_LOSS; - spin_unlock_irqrestore(&ndlp->lock, iflags); lpfc_nlp_put(ndlp); return fcf_inuse; } -- cgit v1.2.3 From 0a3c84f71680684c1d41abb92db05f95c09111e8 Mon Sep 17 00:00:00 2001 From: Justin Tee Date: Thu, 12 Sep 2024 16:24:44 -0700 Subject: scsi: lpfc: Ensure DA_ID handling completion before deleting an NPIV instance Deleting an NPIV instance requires all fabric ndlps to be released before an NPIV's resources can be torn down. Failure to release fabric ndlps beforehand opens kref imbalance race conditions. Fix by forcing the DA_ID to complete synchronously with usage of wait_queue. Signed-off-by: Justin Tee Link: https://lore.kernel.org/r/20240912232447.45607-6-justintee8345@gmail.com Signed-off-by: Martin K. Petersen --- drivers/scsi/lpfc/lpfc_ct.c | 12 ++++++++++++ drivers/scsi/lpfc/lpfc_disc.h | 7 +++++++ drivers/scsi/lpfc/lpfc_vport.c | 43 +++++++++++++++++++++++++++++++++++------- 3 files changed, 55 insertions(+), 7 deletions(-) (limited to 'drivers') diff --git a/drivers/scsi/lpfc/lpfc_ct.c b/drivers/scsi/lpfc/lpfc_ct.c index 2dedd1493e5b..1e5db489a00c 100644 --- a/drivers/scsi/lpfc/lpfc_ct.c +++ b/drivers/scsi/lpfc/lpfc_ct.c @@ -1647,6 +1647,18 @@ lpfc_cmpl_ct(struct lpfc_hba *phba, struct lpfc_iocbq *cmdiocb, } out: + /* If the caller wanted a synchronous DA_ID completion, signal the + * wait obj and clear flag to reset the vport. + */ + if (ndlp->save_flags & NLP_WAIT_FOR_DA_ID) { + if (ndlp->da_id_waitq) + wake_up(ndlp->da_id_waitq); + } + + spin_lock_irq(&ndlp->lock); + ndlp->save_flags &= ~NLP_WAIT_FOR_DA_ID; + spin_unlock_irq(&ndlp->lock); + lpfc_ct_free_iocb(phba, cmdiocb); lpfc_nlp_put(ndlp); return; diff --git a/drivers/scsi/lpfc/lpfc_disc.h b/drivers/scsi/lpfc/lpfc_disc.h index f82615d87c4b..f5ae8cc15820 100644 --- a/drivers/scsi/lpfc/lpfc_disc.h +++ b/drivers/scsi/lpfc/lpfc_disc.h @@ -90,6 +90,8 @@ enum lpfc_nlp_save_flags { NLP_IN_RECOV_POST_DEV_LOSS = 0x1, /* wait for outstanding LOGO to cmpl */ NLP_WAIT_FOR_LOGO = 0x2, + /* wait for outstanding DA_ID to finish */ + NLP_WAIT_FOR_DA_ID = 0x4 }; struct lpfc_nodelist { @@ -159,7 +161,12 @@ struct lpfc_nodelist { uint32_t nvme_fb_size; /* NVME target's supported byte cnt */ #define NVME_FB_BIT_SHIFT 9 /* PRLI Rsp first burst in 512B units. */ uint32_t nlp_defer_did; + + /* These wait objects are NPIV specific. These IOs must complete + * synchronously. + */ wait_queue_head_t *logo_waitq; + wait_queue_head_t *da_id_waitq; }; struct lpfc_node_rrq { diff --git a/drivers/scsi/lpfc/lpfc_vport.c b/drivers/scsi/lpfc/lpfc_vport.c index 4439167a5188..7a4d4d8e2ad5 100644 --- a/drivers/scsi/lpfc/lpfc_vport.c +++ b/drivers/scsi/lpfc/lpfc_vport.c @@ -626,6 +626,7 @@ lpfc_vport_delete(struct fc_vport *fc_vport) struct Scsi_Host *shost = lpfc_shost_from_vport(vport); struct lpfc_hba *phba = vport->phba; int rc; + DECLARE_WAIT_QUEUE_HEAD_ONSTACK(waitq); if (vport->port_type == LPFC_PHYSICAL_PORT) { lpfc_printf_vlog(vport, KERN_ERR, LOG_TRACE_EVENT, @@ -679,21 +680,49 @@ lpfc_vport_delete(struct fc_vport *fc_vport) if (!ndlp) goto skip_logo; + /* Send the DA_ID and Fabric LOGO to cleanup the NPIV fabric entries. */ if (ndlp && ndlp->nlp_state == NLP_STE_UNMAPPED_NODE && phba->link_state >= LPFC_LINK_UP && phba->fc_topology != LPFC_TOPOLOGY_LOOP) { if (vport->cfg_enable_da_id) { - /* Send DA_ID and wait for a completion. */ + /* Send DA_ID and wait for a completion. This is best + * effort. If the DA_ID fails, likely the fabric will + * "leak" NportIDs but at least the driver issued the + * command. + */ + ndlp = lpfc_findnode_did(vport, NameServer_DID); + if (!ndlp) + goto issue_logo; + + spin_lock_irq(&ndlp->lock); + ndlp->da_id_waitq = &waitq; + ndlp->save_flags |= NLP_WAIT_FOR_DA_ID; + spin_unlock_irq(&ndlp->lock); + rc = lpfc_ns_cmd(vport, SLI_CTNS_DA_ID, 0, 0); - if (rc) { - lpfc_printf_log(vport->phba, KERN_WARNING, - LOG_VPORT, - "1829 CT command failed to " - "delete objects on fabric, " - "rc %d\n", rc); + if (!rc) { + wait_event_timeout(waitq, + !(ndlp->save_flags & NLP_WAIT_FOR_DA_ID), + msecs_to_jiffies(phba->fc_ratov * 2000)); } + + lpfc_printf_vlog(vport, KERN_INFO, LOG_VPORT | LOG_ELS, + "1829 DA_ID issue status %d. " + "SFlag x%x NState x%x, NFlag x%x " + "Rpi x%x\n", + rc, ndlp->save_flags, ndlp->nlp_state, + ndlp->nlp_flag, ndlp->nlp_rpi); + + /* Remove the waitq and save_flags. It no + * longer matters if the wake happened. + */ + spin_lock_irq(&ndlp->lock); + ndlp->da_id_waitq = NULL; + ndlp->save_flags &= ~NLP_WAIT_FOR_DA_ID; + spin_unlock_irq(&ndlp->lock); } +issue_logo: /* * If the vpi is not registered, then a valid FDISC doesn't * exist and there is no need for a ELS LOGO. Just cleanup -- cgit v1.2.3 From 1af9af1f8ab38f1285b27581a5e6920ec58296ba Mon Sep 17 00:00:00 2001 From: Justin Tee Date: Thu, 12 Sep 2024 16:24:45 -0700 Subject: scsi: lpfc: Revise TRACE_EVENT log flag severities from KERN_ERR to KERN_WARNING Revise certain log messages marked as KERN_ERR LOG_TRACE_EVENT to KERN_WARNING and use the lpfc_vlog_msg() macro to still log the event. The benefit is that events of interest are still logged and the entire trace buffer is not dumped with extraneous logging information when using default lpfc_log_verbose driver parameter settings. Also, delete the keyword "fail" from such log messages as they aren't really causes for concern. The log messages are more for warnings to a SAN admin about SAN activity. Signed-off-by: Justin Tee Link: https://lore.kernel.org/r/20240912232447.45607-7-justintee8345@gmail.com Signed-off-by: Martin K. Petersen --- drivers/scsi/lpfc/lpfc_ct.c | 10 ++-- drivers/scsi/lpfc/lpfc_els.c | 125 ++++++++++++++++++++----------------------- 2 files changed, 64 insertions(+), 71 deletions(-) (limited to 'drivers') diff --git a/drivers/scsi/lpfc/lpfc_ct.c b/drivers/scsi/lpfc/lpfc_ct.c index 1e5db489a00c..134bc96dd134 100644 --- a/drivers/scsi/lpfc/lpfc_ct.c +++ b/drivers/scsi/lpfc/lpfc_ct.c @@ -1572,8 +1572,8 @@ lpfc_cmpl_ct_cmd_gft_id(struct lpfc_hba *phba, struct lpfc_iocbq *cmdiocb, } } } else - lpfc_printf_vlog(vport, KERN_ERR, LOG_TRACE_EVENT, - "3065 GFT_ID failed x%08x\n", ulp_status); + lpfc_vlog_msg(vport, KERN_WARNING, LOG_DISCOVERY, + "3065 GFT_ID status x%08x\n", ulp_status); out: lpfc_ct_free_iocb(phba, cmdiocb); @@ -2258,7 +2258,7 @@ lpfc_cmpl_ct_disc_fdmi(struct lpfc_hba *phba, struct lpfc_iocbq *cmdiocb, } lpfc_printf_vlog(vport, KERN_INFO, LOG_DISCOVERY, - "0229 FDMI cmd %04x failed, latt = %d " + "0229 FDMI cmd %04x latt = %d " "ulp_status: x%x, rid x%x\n", be16_to_cpu(fdmi_cmd), latt, ulp_status, ulp_word4); @@ -2275,9 +2275,9 @@ lpfc_cmpl_ct_disc_fdmi(struct lpfc_hba *phba, struct lpfc_iocbq *cmdiocb, /* Check for a CT LS_RJT response */ cmd = be16_to_cpu(fdmi_cmd); if (be16_to_cpu(fdmi_rsp) == SLI_CT_RESPONSE_FS_RJT) { - /* FDMI rsp failed */ + /* Log FDMI reject */ lpfc_printf_vlog(vport, KERN_INFO, LOG_DISCOVERY | LOG_ELS, - "0220 FDMI cmd failed FS_RJT Data: x%x", cmd); + "0220 FDMI cmd FS_RJT Data: x%x", cmd); /* Should we fallback to FDMI-2 / FDMI-1 ? */ switch (cmd) { diff --git a/drivers/scsi/lpfc/lpfc_els.c b/drivers/scsi/lpfc/lpfc_els.c index 7219b1ada1ea..d737b897ddd8 100644 --- a/drivers/scsi/lpfc/lpfc_els.c +++ b/drivers/scsi/lpfc/lpfc_els.c @@ -979,7 +979,7 @@ lpfc_cmpl_els_flogi(struct lpfc_hba *phba, struct lpfc_iocbq *cmdiocb, phba->fcoe_cvl_eventtag_attn = phba->fcoe_cvl_eventtag; lpfc_printf_log(phba, KERN_WARNING, LOG_FIP | LOG_ELS, - "2611 FLOGI failed on FCF (x%x), " + "2611 FLOGI FCF (x%x), " "status:x%x/x%x, tmo:x%x, perform " "roundrobin FCF failover\n", phba->fcf.current_rec.fcf_indx, @@ -997,11 +997,11 @@ stop_rr_fcf_flogi: if (!(ulp_status == IOSTAT_LOCAL_REJECT && ((ulp_word4 & IOERR_PARAM_MASK) == IOERR_LOOP_OPEN_FAILURE))) - lpfc_printf_vlog(vport, KERN_ERR, LOG_TRACE_EVENT, - "2858 FLOGI failure Status:x%x/x%x TMO" - ":x%x Data x%lx x%x\n", - ulp_status, ulp_word4, tmo, - phba->hba_flag, phba->fcf.fcf_flag); + lpfc_vlog_msg(vport, KERN_WARNING, LOG_ELS, + "2858 FLOGI Status:x%x/x%x TMO" + ":x%x Data x%lx x%x\n", + ulp_status, ulp_word4, tmo, + phba->hba_flag, phba->fcf.fcf_flag); /* Check for retry */ if (lpfc_els_retry(phba, cmdiocb, rspiocb)) { @@ -1023,7 +1023,7 @@ stop_rr_fcf_flogi: lpfc_nlp_put(ndlp); lpfc_printf_vlog(vport, KERN_WARNING, LOG_ELS, - "0150 FLOGI failure Status:x%x/x%x " + "0150 FLOGI Status:x%x/x%x " "xri x%x TMO:x%x refcnt %d\n", ulp_status, ulp_word4, cmdiocb->sli4_xritag, tmo, kref_read(&ndlp->kref)); @@ -1032,11 +1032,11 @@ stop_rr_fcf_flogi: if (!(ulp_status == IOSTAT_LOCAL_REJECT && ((ulp_word4 & IOERR_PARAM_MASK) == IOERR_LOOP_OPEN_FAILURE))) { - /* FLOGI failure */ - lpfc_printf_vlog(vport, KERN_ERR, LOG_TRACE_EVENT, - "0100 FLOGI failure Status:x%x/x%x " - "TMO:x%x\n", - ulp_status, ulp_word4, tmo); + /* Warn FLOGI status */ + lpfc_vlog_msg(vport, KERN_WARNING, LOG_ELS, + "0100 FLOGI Status:x%x/x%x " + "TMO:x%x\n", + ulp_status, ulp_word4, tmo); goto flogifail; } @@ -1964,16 +1964,16 @@ lpfc_cmpl_els_rrq(struct lpfc_hba *phba, struct lpfc_iocbq *cmdiocb, if (ulp_status) { /* Check for retry */ - /* RRQ failed Don't print the vport to vport rjts */ + /* Warn RRQ status Don't print the vport to vport rjts */ if (ulp_status != IOSTAT_LS_RJT || (((ulp_word4) >> 16 != LSRJT_INVALID_CMD) && ((ulp_word4) >> 16 != LSRJT_UNABLE_TPC)) || (phba)->pport->cfg_log_verbose & LOG_ELS) - lpfc_printf_vlog(vport, KERN_ERR, LOG_TRACE_EVENT, - "2881 RRQ failure DID:%06X Status:" - "x%x/x%x\n", - ndlp->nlp_DID, ulp_status, - ulp_word4); + lpfc_vlog_msg(vport, KERN_WARNING, LOG_ELS, + "2881 RRQ DID:%06X Status:" + "x%x/x%x\n", + ndlp->nlp_DID, ulp_status, + ulp_word4); } lpfc_clr_rrq_active(phba, rrq->xritag, rrq); @@ -2077,16 +2077,16 @@ lpfc_cmpl_els_plogi(struct lpfc_hba *phba, struct lpfc_iocbq *cmdiocb, } goto out; } - /* PLOGI failed Don't print the vport to vport rjts */ + /* Warn PLOGI status Don't print the vport to vport rjts */ if (ulp_status != IOSTAT_LS_RJT || (((ulp_word4) >> 16 != LSRJT_INVALID_CMD) && ((ulp_word4) >> 16 != LSRJT_UNABLE_TPC)) || (phba)->pport->cfg_log_verbose & LOG_ELS) - lpfc_printf_vlog(vport, KERN_ERR, LOG_TRACE_EVENT, - "2753 PLOGI failure DID:%06X " - "Status:x%x/x%x\n", - ndlp->nlp_DID, ulp_status, - ulp_word4); + lpfc_vlog_msg(vport, KERN_WARNING, LOG_ELS, + "2753 PLOGI DID:%06X " + "Status:x%x/x%x\n", + ndlp->nlp_DID, ulp_status, + ulp_word4); /* Do not call DSM for lpfc_els_abort'ed ELS cmds */ if (!lpfc_error_lost_link(vport, ulp_status, ulp_word4)) @@ -2323,7 +2323,6 @@ lpfc_cmpl_els_prli(struct lpfc_hba *phba, struct lpfc_iocbq *cmdiocb, struct lpfc_vport *vport = cmdiocb->vport; struct lpfc_nodelist *ndlp; char *mode; - u32 loglevel; u32 ulp_status; u32 ulp_word4; bool release_node = false; @@ -2372,17 +2371,14 @@ lpfc_cmpl_els_prli(struct lpfc_hba *phba, struct lpfc_iocbq *cmdiocb, * could be expected. */ if (test_bit(FC_FABRIC, &vport->fc_flag) || - vport->cfg_enable_fc4_type != LPFC_ENABLE_BOTH) { - mode = KERN_ERR; - loglevel = LOG_TRACE_EVENT; - } else { + vport->cfg_enable_fc4_type != LPFC_ENABLE_BOTH) + mode = KERN_WARNING; + else mode = KERN_INFO; - loglevel = LOG_ELS; - } - /* PRLI failed */ - lpfc_printf_vlog(vport, mode, loglevel, - "2754 PRLI failure DID:%06X Status:x%x/x%x, " + /* Warn PRLI status */ + lpfc_printf_vlog(vport, mode, LOG_ELS, + "2754 PRLI DID:%06X Status:x%x/x%x, " "data: x%x x%x x%x\n", ndlp->nlp_DID, ulp_status, ulp_word4, ndlp->nlp_state, @@ -2854,11 +2850,11 @@ lpfc_cmpl_els_adisc(struct lpfc_hba *phba, struct lpfc_iocbq *cmdiocb, } goto out; } - /* ADISC failed */ - lpfc_printf_vlog(vport, KERN_ERR, LOG_TRACE_EVENT, - "2755 ADISC failure DID:%06X Status:x%x/x%x\n", - ndlp->nlp_DID, ulp_status, - ulp_word4); + /* Warn ADISC status */ + lpfc_vlog_msg(vport, KERN_WARNING, LOG_ELS, + "2755 ADISC DID:%06X Status:x%x/x%x\n", + ndlp->nlp_DID, ulp_status, + ulp_word4); lpfc_disc_state_machine(vport, ndlp, cmdiocb, NLP_EVT_CMPL_ADISC); @@ -3045,12 +3041,12 @@ lpfc_cmpl_els_logo(struct lpfc_hba *phba, struct lpfc_iocbq *cmdiocb, * discovery. The PLOGI will retry. */ if (ulp_status) { - /* LOGO failed */ - lpfc_printf_vlog(vport, KERN_ERR, LOG_TRACE_EVENT, - "2756 LOGO failure, No Retry DID:%06X " - "Status:x%x/x%x\n", - ndlp->nlp_DID, ulp_status, - ulp_word4); + /* Warn LOGO status */ + lpfc_vlog_msg(vport, KERN_WARNING, LOG_ELS, + "2756 LOGO, No Retry DID:%06X " + "Status:x%x/x%x\n", + ndlp->nlp_DID, ulp_status, + ulp_word4); if (lpfc_error_lost_link(vport, ulp_status, ulp_word4)) skip_recovery = 1; @@ -4837,11 +4833,10 @@ lpfc_els_retry(struct lpfc_hba *phba, struct lpfc_iocbq *cmdiocb, if ((phba->sli3_options & LPFC_SLI3_NPIV_ENABLED) && (cmd == ELS_CMD_FDISC) && (stat.un.b.lsRjtRsnCodeExp == LSEXP_OUT_OF_RESOURCE)){ - lpfc_printf_vlog(vport, KERN_ERR, - LOG_TRACE_EVENT, - "0125 FDISC Failed (x%x). " - "Fabric out of resources\n", - stat.un.lsRjtError); + lpfc_vlog_msg(vport, KERN_WARNING, LOG_ELS, + "0125 FDISC (x%x). " + "Fabric out of resources\n", + stat.un.lsRjtError); lpfc_vport_set_state(vport, FC_VPORT_NO_FABRIC_RSCS); } @@ -4877,11 +4872,10 @@ lpfc_els_retry(struct lpfc_hba *phba, struct lpfc_iocbq *cmdiocb, LSEXP_NOTHING_MORE) { vport->fc_sparam.cmn.bbRcvSizeMsb &= 0xf; retry = 1; - lpfc_printf_vlog(vport, KERN_ERR, - LOG_TRACE_EVENT, - "0820 FLOGI Failed (x%x). " - "BBCredit Not Supported\n", - stat.un.lsRjtError); + lpfc_vlog_msg(vport, KERN_WARNING, LOG_ELS, + "0820 FLOGI (x%x). " + "BBCredit Not Supported\n", + stat.un.lsRjtError); } break; @@ -4891,11 +4885,10 @@ lpfc_els_retry(struct lpfc_hba *phba, struct lpfc_iocbq *cmdiocb, ((stat.un.b.lsRjtRsnCodeExp == LSEXP_INVALID_PNAME) || (stat.un.b.lsRjtRsnCodeExp == LSEXP_INVALID_NPORT_ID)) ) { - lpfc_printf_vlog(vport, KERN_ERR, - LOG_TRACE_EVENT, - "0122 FDISC Failed (x%x). " - "Fabric Detected Bad WWN\n", - stat.un.lsRjtError); + lpfc_vlog_msg(vport, KERN_WARNING, LOG_ELS, + "0122 FDISC (x%x). " + "Fabric Detected Bad WWN\n", + stat.un.lsRjtError); lpfc_vport_set_state(vport, FC_VPORT_FABRIC_REJ_WWN); } @@ -5355,8 +5348,8 @@ lpfc_cmpl_els_rsp(struct lpfc_hba *phba, struct lpfc_iocbq *cmdiocb, u32 ulp_status, ulp_word4, tmo, did, iotag; if (!vport) { - lpfc_printf_log(phba, KERN_ERR, LOG_TRACE_EVENT, - "3177 ELS response failed\n"); + lpfc_printf_log(phba, KERN_WARNING, LOG_ELS, + "3177 null vport in ELS rsp\n"); goto out; } if (cmdiocb->context_un.mbox) @@ -11328,10 +11321,10 @@ lpfc_cmpl_els_fdisc(struct lpfc_hba *phba, struct lpfc_iocbq *cmdiocb, /* Check for retry */ if (lpfc_els_retry(phba, cmdiocb, rspiocb)) goto out; - /* FDISC failed */ - lpfc_printf_vlog(vport, KERN_ERR, LOG_TRACE_EVENT, - "0126 FDISC failed. (x%x/x%x)\n", - ulp_status, ulp_word4); + /* Warn FDISC status */ + lpfc_vlog_msg(vport, KERN_WARNING, LOG_ELS, + "0126 FDISC cmpl status: x%x/x%x)\n", + ulp_status, ulp_word4); goto fdisc_failed; } -- cgit v1.2.3 From eeb85c658e1bceaccb6ca3ffc1796741abd7b687 Mon Sep 17 00:00:00 2001 From: Justin Tee Date: Thu, 12 Sep 2024 16:24:46 -0700 Subject: scsi: lpfc: Support loopback tests with VMID enabled The VMID feature adds an extra application services header to each frame. As such, the loopback test path is updated to accommodate the extra application header. Changes include filling in APPID and WQES bit fields for XMIT_SEQUENCE64 commands, a special loopback source APPID for verifying received loopback data matches what is sent, and increasing ELS WQ size to accommodate the APPID field in loopback test mode. Signed-off-by: Justin Tee Link: https://lore.kernel.org/r/20240912232447.45607-8-justintee8345@gmail.com Signed-off-by: Martin K. Petersen --- drivers/scsi/lpfc/lpfc_bsg.c | 3 +++ drivers/scsi/lpfc/lpfc_hw.h | 21 +++++++++++++++++++++ drivers/scsi/lpfc/lpfc_init.c | 11 +++++++++-- drivers/scsi/lpfc/lpfc_sli.c | 39 +++++++++++++++++++++++++++++++++++++-- 4 files changed, 70 insertions(+), 4 deletions(-) (limited to 'drivers') diff --git a/drivers/scsi/lpfc/lpfc_bsg.c b/drivers/scsi/lpfc/lpfc_bsg.c index 4156419c52c7..8738c1b0e388 100644 --- a/drivers/scsi/lpfc/lpfc_bsg.c +++ b/drivers/scsi/lpfc/lpfc_bsg.c @@ -3208,6 +3208,9 @@ lpfc_bsg_diag_loopback_run(struct bsg_job *job) cmdiocbq->num_bdes = num_bde; cmdiocbq->cmd_flag |= LPFC_IO_LIBDFC; cmdiocbq->cmd_flag |= LPFC_IO_LOOPBACK; + if (phba->cfg_vmid_app_header) + cmdiocbq->cmd_flag |= LPFC_IO_VMID; + cmdiocbq->vport = phba->pport; cmdiocbq->cmd_cmpl = NULL; cmdiocbq->bpl_dmabuf = txbmp; diff --git a/drivers/scsi/lpfc/lpfc_hw.h b/drivers/scsi/lpfc/lpfc_hw.h index 2108b4cb7815..d5c15742f7f2 100644 --- a/drivers/scsi/lpfc/lpfc_hw.h +++ b/drivers/scsi/lpfc/lpfc_hw.h @@ -561,6 +561,27 @@ struct fc_vft_header { #include +/* + * Application Header + */ +struct fc_app_header { + uint32_t dst_app_id; + uint32_t src_app_id; +#define LOOPBACK_SRC_APPID 0x4321 + uint32_t word2; + uint32_t word3; +}; + +/* + * dfctl optional header definition + */ +enum lpfc_fc_dfctl { + LPFC_FC_NO_DEVICE_HEADER, + LPFC_FC_16B_DEVICE_HEADER, + LPFC_FC_32B_DEVICE_HEADER, + LPFC_FC_64B_DEVICE_HEADER, +}; + /* * Extended Link Service LS_COMMAND codes (Payload Word 0) */ diff --git a/drivers/scsi/lpfc/lpfc_init.c b/drivers/scsi/lpfc/lpfc_init.c index 0c1404dc5f3b..48a3dfdd51d3 100644 --- a/drivers/scsi/lpfc/lpfc_init.c +++ b/drivers/scsi/lpfc/lpfc_init.c @@ -10451,6 +10451,7 @@ lpfc_sli4_queue_create(struct lpfc_hba *phba) struct lpfc_vector_map_info *cpup; struct lpfc_vector_map_info *eqcpup; struct lpfc_eq_intr_info *eqi; + u32 wqesize; /* * Create HBA Record arrays. @@ -10670,9 +10671,15 @@ lpfc_sli4_queue_create(struct lpfc_hba *phba) * Create ELS Work Queues */ - /* Create slow-path ELS Work Queue */ + /* + * Create slow-path ELS Work Queue. + * Increase the ELS WQ size when WQEs contain an embedded cdb + */ + wqesize = (phba->fcp_embed_io) ? + LPFC_WQE128_SIZE : phba->sli4_hba.wq_esize; + qdesc = lpfc_sli4_queue_alloc(phba, LPFC_DEFAULT_PAGE_SIZE, - phba->sli4_hba.wq_esize, + wqesize, phba->sli4_hba.wq_ecount, cpu); if (!qdesc) { lpfc_printf_log(phba, KERN_ERR, LOG_TRACE_EVENT, diff --git a/drivers/scsi/lpfc/lpfc_sli.c b/drivers/scsi/lpfc/lpfc_sli.c index bb5fd3322273..a44afb84cd3d 100644 --- a/drivers/scsi/lpfc/lpfc_sli.c +++ b/drivers/scsi/lpfc/lpfc_sli.c @@ -11093,9 +11093,17 @@ __lpfc_sli_prep_xmit_seq64_s4(struct lpfc_iocbq *cmdiocbq, /* Word 9 */ bf_set(wqe_rcvoxid, &wqe->xmit_sequence.wqe_com, ox_id); - /* Word 12 */ - if (cmdiocbq->cmd_flag & (LPFC_IO_LIBDFC | LPFC_IO_LOOPBACK)) + if (cmdiocbq->cmd_flag & (LPFC_IO_LIBDFC | LPFC_IO_LOOPBACK)) { + /* Word 10 */ + if (cmdiocbq->cmd_flag & LPFC_IO_VMID) { + bf_set(wqe_appid, &wqe->xmit_sequence.wqe_com, 1); + bf_set(wqe_wqes, &wqe->xmit_sequence.wqe_com, 1); + wqe->words[31] = LOOPBACK_SRC_APPID; + } + + /* Word 12 */ wqe->xmit_sequence.xmit_len = full_size; + } else wqe->xmit_sequence.xmit_len = wqe->xmit_sequence.bde.tus.f.bdeSize; @@ -18434,6 +18442,7 @@ lpfc_fc_frame_check(struct lpfc_hba *phba, struct fc_frame_header *fc_hdr) { /* make rctl_names static to save stack space */ struct fc_vft_header *fc_vft_hdr; + struct fc_app_header *fc_app_hdr; uint32_t *header = (uint32_t *) fc_hdr; #define FC_RCTL_MDS_DIAGS 0xF4 @@ -18489,6 +18498,32 @@ lpfc_fc_frame_check(struct lpfc_hba *phba, struct fc_frame_header *fc_hdr) goto drop; } + if (unlikely(phba->link_flag == LS_LOOPBACK_MODE && + phba->cfg_vmid_app_header)) { + /* Application header is 16B device header */ + if (fc_hdr->fh_df_ctl & LPFC_FC_16B_DEVICE_HEADER) { + fc_app_hdr = (struct fc_app_header *) (fc_hdr + 1); + if (be32_to_cpu(fc_app_hdr->src_app_id) != + LOOPBACK_SRC_APPID) { + lpfc_printf_log(phba, KERN_WARNING, + LOG_ELS | LOG_LIBDFC, + "1932 Loopback src app id " + "not matched, app_id:x%x\n", + be32_to_cpu(fc_app_hdr->src_app_id)); + + goto drop; + } + } else { + lpfc_printf_log(phba, KERN_WARNING, + LOG_ELS | LOG_LIBDFC, + "1933 Loopback df_ctl bit not set, " + "df_ctl:x%x\n", + fc_hdr->fh_df_ctl); + + goto drop; + } + } + lpfc_printf_log(phba, KERN_INFO, LOG_ELS, "2538 Received frame rctl:x%x, type:x%x, " "frame Data:%08x %08x %08x %08x %08x %08x %08x\n", -- cgit v1.2.3 From b071c1a9099c7bc5fe24089117e7a15e52d4198f Mon Sep 17 00:00:00 2001 From: Justin Tee Date: Thu, 12 Sep 2024 16:24:47 -0700 Subject: scsi: lpfc: Update lpfc version to 14.4.0.5 Update lpfc version to 14.4.0.5 Signed-off-by: Justin Tee Link: https://lore.kernel.org/r/20240912232447.45607-9-justintee8345@gmail.com Signed-off-by: Martin K. Petersen --- drivers/scsi/lpfc/lpfc_version.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'drivers') diff --git a/drivers/scsi/lpfc/lpfc_version.h b/drivers/scsi/lpfc/lpfc_version.h index 2fe0386a1fee..e70f163fab90 100644 --- a/drivers/scsi/lpfc/lpfc_version.h +++ b/drivers/scsi/lpfc/lpfc_version.h @@ -20,7 +20,7 @@ * included with this package. * *******************************************************************/ -#define LPFC_DRIVER_VERSION "14.4.0.4" +#define LPFC_DRIVER_VERSION "14.4.0.5" #define LPFC_DRIVER_NAME "lpfc" /* Used for SLI 2/3 */ -- cgit v1.2.3