scsi: lpfc: Fix EEH support for NVMe I/O

Injecting errors on the PCI slot while the driver is handling NVMe I/O will
cause crashes and hangs.

There are several rather difficult scenarios occurring. The main issue is
that the adapter can report a PCI error before or simultaneously to the PCI
subsystem reporting the error. Both paths have different entry points and
currently there is no interlock between them. Thus multiple teardown paths
are competing and all heck breaks loose.

Complicating things is the NVMs path. To a large degree, I/O was able to be
shutdown for a full FC port on the SCSI stack. But on NVMe, there isn't a
similar call. At best, it works on a per-controller basis, but even at the
controller level, it's a controller "reset" call. All of which means I/O is
still flowing on different CPUs with reset paths expecting hw access
(mailbox commands) to execute properly.

The following modifications are made:

 - A new flag is set in PCI error entrypoints so the driver can track being
   called by that path.

 - An interlock is added in the SLI hw error path and the PCI error path
   such that only one of the paths proceeds with the teardown logic.

 - RPI cleanup is patched such that RPIs are marked unregistered w/o mbx
   cmds in cases of hw error.

 - If entering the SLI port re-init calls, a case where SLI error teardown
   was quick and beat the PCI calls now reporting error, check whether the
   SLI port is still live on the PCI bus.

 - In the PCI reset code to bring the adapter back, recheck the IRQ
   settings. Different checks for SLI3 vs SLI4.

 - In I/O completions, that may be called as part of the cleanup or
   underway just before the hw error, check the state of the adapter.  If
   in error, shortcut handling that would expect further adapter
   completions as the hw error won't be sending them.

 - In routines waiting on I/O completions, which may have been in progress
   prior to the hw error, detect the device is being torn down and abort
   from their waits and just give up. This points to a larger issue in the
   driver on ref-counting for data structures, as it doesn't have
   ref-counting on q and port structures. We'll do this fix for now as it
   would be a major rework to be done differently.

 - Fix the NVMe cleanup to simulate NVMe I/O completions if I/O is being
   failed back due to hw error.

 - In I/O buf allocation, done at the start of new I/Os, check hw state and
   fail if hw error.

Link: https://lore.kernel.org/r/20210910233159.115896-10-jsmart2021@gmail.com
Co-developed-by: Justin Tee <justin.tee@broadcom.com>
Signed-off-by: Justin Tee <justin.tee@broadcom.com>
Signed-off-by: James Smart <jsmart2021@gmail.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

1 files changed, 51 insertions, 2 deletions

diff --git a/drivers/scsi/lpfc/lpfc_nvme.c b/drivers/scsi/lpfc/lpfc_nvme.c
index bd88477f9b82..33266e1b24ab 100644
--- a/drivers/scsi/lpfc/lpfc_nvme.c
+++ b/drivers/scsi/lpfc/lpfc_nvme.c
@@ -937,6 +937,7 @@ lpfc_nvme_io_cmd_wqe_cmpl(struct lpfc_hba *phba, struct lpfc_iocbq *pwqeIn,
 #ifdef CONFIG_SCSI_LPFC_DEBUG_FS
 	int cpu;
 #endif
+	int offline = 0;
 
 	/* Sanity check on return of outstanding command */
 	if (!lpfc_ncmd) {
@@ -1098,11 +1099,12 @@ out_err:
 			nCmd->transferred_length = 0;
 			nCmd->rcv_rsplen = 0;
 			nCmd->status = NVME_SC_INTERNAL;
+			offline = pci_channel_offline(vport->phba->pcidev);
 		}
 	}
 
 	/* pick up SLI4 exhange busy condition */
-	if (bf_get(lpfc_wcqe_c_xb, wcqe))
+	if (bf_get(lpfc_wcqe_c_xb, wcqe) && !offline)
 		lpfc_ncmd->flags |= LPFC_SBUF_XBUSY;
 	else
 		lpfc_ncmd->flags &= ~LPFC_SBUF_XBUSY;
@@ -2169,6 +2171,10 @@ lpfc_nvme_lport_unreg_wait(struct lpfc_vport *vport,
 			abts_nvme = 0;
 			for (i = 0; i < phba->cfg_hdw_queue; i++) {
 				qp = &phba->sli4_hba.hdwq[i];
+				if (!vport || !vport->localport ||
+				    !qp || !qp->io_wq)
+					return;
+
 				pring = qp->io_wq->pring;
 				if (!pring)
 					continue;
@@ -2176,6 +2182,10 @@ lpfc_nvme_lport_unreg_wait(struct lpfc_vport *vport,
 				abts_scsi += qp->abts_scsi_io_bufs;
 				abts_nvme += qp->abts_nvme_io_bufs;
 			}
+			if (!vport || !vport->localport ||
+			    vport->phba->hba_flag & HBA_PCI_ERR)
+				return;
+
 			lpfc_printf_vlog(vport, KERN_ERR, LOG_TRACE_EVENT,
 					 "6176 Lport x%px Localport x%px wait "
 					 "timed out. Pending %d [%d:%d]. "
@@ -2215,6 +2225,8 @@ lpfc_nvme_destroy_localport(struct lpfc_vport *vport)
 		return;
 
 	localport = vport->localport;
+	if (!localport)
+		return;
 	lport = (struct lpfc_nvme_lport *)localport->private;
 
 	lpfc_printf_vlog(vport, KERN_INFO, LOG_NVME,
@@ -2531,7 +2543,8 @@ lpfc_nvme_unregister_port(struct lpfc_vport *vport, struct lpfc_nodelist *ndlp)
 		 * return values is ignored.  The upcall is a courtesy to the
 		 * transport.
 		 */
-		if (vport->load_flag & FC_UNLOADING)
+		if (vport->load_flag & FC_UNLOADING ||
+		    unlikely(vport->phba->hba_flag & HBA_PCI_ERR))
 			(void)nvme_fc_set_remoteport_devloss(remoteport, 0);
 
 		ret = nvme_fc_unregister_remoteport(remoteport);
@@ -2560,6 +2573,42 @@ lpfc_nvme_unregister_port(struct lpfc_vport *vport, struct lpfc_nodelist *ndlp)
 }
 
 /**
+ * lpfc_sli4_nvme_pci_offline_aborted - Fast-path process of NVME xri abort
+ * @phba: pointer to lpfc hba data structure.
+ * @lpfc_ncmd: The nvme job structure for the request being aborted.
+ *
+ * This routine is invoked by the worker thread to process a SLI4 fast-path
+ * NVME aborted xri.  Aborted NVME IO commands are completed to the transport
+ * here.
+ **/
+void
+lpfc_sli4_nvme_pci_offline_aborted(struct lpfc_hba *phba,
+				   struct lpfc_io_buf *lpfc_ncmd)
+{
+	struct nvmefc_fcp_req *nvme_cmd = NULL;
+
+	lpfc_printf_log(phba, KERN_INFO, LOG_NVME_ABTS,
+			"6533 %s nvme_cmd %p tag x%x abort complete and "
+			"xri released\n", __func__,
+			lpfc_ncmd->nvmeCmd,
+			lpfc_ncmd->cur_iocbq.iotag);
+
+	/* Aborted NVME commands are required to not complete
+	 * before the abort exchange command fully completes.
+	 * Once completed, it is available via the put list.
+	 */
+	if (lpfc_ncmd->nvmeCmd) {
+		nvme_cmd = lpfc_ncmd->nvmeCmd;
+		nvme_cmd->transferred_length = 0;
+		nvme_cmd->rcv_rsplen = 0;
+		nvme_cmd->status = NVME_SC_INTERNAL;
+		nvme_cmd->done(nvme_cmd);
+		lpfc_ncmd->nvmeCmd = NULL;
+	}
+	lpfc_release_nvme_buf(phba, lpfc_ncmd);
+}
+
+/**
  * lpfc_sli4_nvme_xri_aborted - Fast-path process of NVME xri abort
  * @phba: pointer to lpfc hba data structure.
  * @axri: pointer to the fcp xri abort wcqe structure.