Merge tag 'scsi-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi

Pull first round of SCSI updates from James Bottomley:
 "This patch set is driver updates for qla4xxx, scsi_debug, pm80xx,
  fcoe/libfc, eas2r, lpfc, be2iscsi and megaraid_sas plus some assorted
  bug fixes and cleanups"

* tag 'scsi-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi: (106 commits)
  [SCSI] scsi_error: Escalate to LUN reset if abort fails
  [SCSI] Add 'eh_deadline' to limit SCSI EH runtime
  [SCSI] remove check for 'resetting'
  [SCSI] dc395: Move 'last_reset' into internal host structure
  [SCSI] tmscsim: Move 'last_reset' into host structure
  [SCSI] advansys: Remove 'last_reset' references
  [SCSI] dpt_i2o: return SCSI_MLQUEUE_HOST_BUSY when in reset
  [SCSI] dpt_i2o: Remove DPTI_STATE_IOCTL
  [SCSI] megaraid_sas: Fix synchronization problem between sysPD IO path and AEN path
  [SCSI] lpfc: Fix typo on NULL assignment
  [SCSI] scsi_dh_alua: ALUA handler attach should succeed while TPG is transitioning
  [SCSI] scsi_dh_alua: ALUA check sense should retry device internal reset unit attention
  [SCSI] esas2r: Cleanup snprinf formatting of firmware version
  [SCSI] esas2r: Remove superfluous mask of pcie_cap_reg
  [SCSI] esas2r: Fixes for big-endian platforms
  [SCSI] esas2r: Directly call kernel functions for atomic bit operations
  [SCSI] lpfc 8.3.43: Update lpfc version to driver version 8.3.43
  [SCSI] lpfc 8.3.43: Fixed not processing task management IOCB response status
  [SCSI] lpfc 8.3.43: Fixed spinlock hang.
  [SCSI] lpfc 8.3.43: Fixed invalid Total_Data_Placed value received for els and ct command responses
  ...

1 files changed, 172 insertions, 11 deletions

diff --git a/drivers/scsi/lpfc/lpfc_sli.c b/drivers/scsi/lpfc/lpfc_sli.c
index 612f489..8f580fd 100644
--- a/drivers/scsi/lpfc/lpfc_sli.c
+++ b/drivers/scsi/lpfc/lpfc_sli.c
@@ -71,6 +71,8 @@ static int lpfc_sli4_post_els_sgl_list(struct lpfc_hba *, struct list_head *,
 				       int);
 static void lpfc_sli4_hba_handle_eqe(struct lpfc_hba *, struct lpfc_eqe *,
 			uint32_t);
+static bool lpfc_sli4_mbox_completions_pending(struct lpfc_hba *phba);
+static bool lpfc_sli4_process_missed_mbox_completions(struct lpfc_hba *phba);
 
 static IOCB_t *
 lpfc_get_iocb_from_iocbq(struct lpfc_iocbq *iocbq)
@@ -6566,6 +6568,108 @@ lpfc_mbox_timeout(unsigned long ptr)
 	return;
 }
 
+/**
+ * lpfc_sli4_mbox_completions_pending - check to see if any mailbox completions
+ *                                    are pending
+ * @phba: Pointer to HBA context object.
+ *
+ * This function checks if any mailbox completions are present on the mailbox
+ * completion queue.
+ **/
+bool
+lpfc_sli4_mbox_completions_pending(struct lpfc_hba *phba)
+{
+
+	uint32_t idx;
+	struct lpfc_queue *mcq;
+	struct lpfc_mcqe *mcqe;
+	bool pending_completions = false;
+
+	if (unlikely(!phba) || (phba->sli_rev != LPFC_SLI_REV4))
+		return false;
+
+	/* Check for completions on mailbox completion queue */
+
+	mcq = phba->sli4_hba.mbx_cq;
+	idx = mcq->hba_index;
+	while (bf_get_le32(lpfc_cqe_valid, mcq->qe[idx].cqe)) {
+		mcqe = (struct lpfc_mcqe *)mcq->qe[idx].cqe;
+		if (bf_get_le32(lpfc_trailer_completed, mcqe) &&
+		    (!bf_get_le32(lpfc_trailer_async, mcqe))) {
+			pending_completions = true;
+			break;
+		}
+		idx = (idx + 1) % mcq->entry_count;
+		if (mcq->hba_index == idx)
+			break;
+	}
+	return pending_completions;
+
+}
+
+/**
+ * lpfc_sli4_process_missed_mbox_completions - process mbox completions
+ *					      that were missed.
+ * @phba: Pointer to HBA context object.
+ *
+ * For sli4, it is possible to miss an interrupt. As such mbox completions
+ * maybe missed causing erroneous mailbox timeouts to occur. This function
+ * checks to see if mbox completions are on the mailbox completion queue
+ * and will process all the completions associated with the eq for the
+ * mailbox completion queue.
+ **/
+bool
+lpfc_sli4_process_missed_mbox_completions(struct lpfc_hba *phba)
+{
+
+	uint32_t eqidx;
+	struct lpfc_queue *fpeq = NULL;
+	struct lpfc_eqe *eqe;
+	bool mbox_pending;
+
+	if (unlikely(!phba) || (phba->sli_rev != LPFC_SLI_REV4))
+		return false;
+
+	/* Find the eq associated with the mcq */
+
+	if (phba->sli4_hba.hba_eq)
+		for (eqidx = 0; eqidx < phba->cfg_fcp_io_channel; eqidx++)
+			if (phba->sli4_hba.hba_eq[eqidx]->queue_id ==
+			    phba->sli4_hba.mbx_cq->assoc_qid) {
+				fpeq = phba->sli4_hba.hba_eq[eqidx];
+				break;
+			}
+	if (!fpeq)
+		return false;
+
+	/* Turn off interrupts from this EQ */
+
+	lpfc_sli4_eq_clr_intr(fpeq);
+
+	/* Check to see if a mbox completion is pending */
+
+	mbox_pending = lpfc_sli4_mbox_completions_pending(phba);
+
+	/*
+	 * If a mbox completion is pending, process all the events on EQ
+	 * associated with the mbox completion queue (this could include
+	 * mailbox commands, async events, els commands, receive queue data
+	 * and fcp commands)
+	 */
+
+	if (mbox_pending)
+		while ((eqe = lpfc_sli4_eq_get(fpeq))) {
+			lpfc_sli4_hba_handle_eqe(phba, eqe, eqidx);
+			fpeq->EQ_processed++;
+		}
+
+	/* Always clear and re-arm the EQ */
+
+	lpfc_sli4_eq_release(fpeq, LPFC_QUEUE_REARM);
+
+	return mbox_pending;
+
+}
 
 /**
  * lpfc_mbox_timeout_handler - Worker thread function to handle mailbox timeout
@@ -6583,6 +6687,10 @@ lpfc_mbox_timeout_handler(struct lpfc_hba *phba)
 	struct lpfc_sli *psli = &phba->sli;
 	struct lpfc_sli_ring *pring;
 
+	/* If the mailbox completed, process the completion and return */
+	if (lpfc_sli4_process_missed_mbox_completions(phba))
+		return;
+
 	/* Check the pmbox pointer first.  There is a race condition
 	 * between the mbox timeout handler getting executed in the
 	 * worklist and the mailbox actually completing. When this
@@ -7077,6 +7185,10 @@ lpfc_sli4_async_mbox_block(struct lpfc_hba *phba)
 						1000) + jiffies;
 	spin_unlock_irq(&phba->hbalock);
 
+	/* Make sure the mailbox is really active */
+	if (timeout)
+		lpfc_sli4_process_missed_mbox_completions(phba);
+
 	/* Wait for the outstnading mailbox command to complete */
 	while (phba->sli.mbox_active) {
 		/* Check active mailbox complete status every 2ms */
@@ -8076,6 +8188,7 @@ lpfc_sli4_iocb2wqe(struct lpfc_hba *phba, struct lpfc_iocbq *iocbq,
 		bf_set(wqe_qosd, &wqe->els_req.wqe_com, 1);
 		bf_set(wqe_lenloc, &wqe->els_req.wqe_com, LPFC_WQE_LENLOC_NONE);
 		bf_set(wqe_ebde_cnt, &wqe->els_req.wqe_com, 0);
+		wqe->els_req.max_response_payload_len = total_len - xmit_len;
 		break;
 	case CMD_XMIT_SEQUENCE64_CX:
 		bf_set(wqe_ctxt_tag, &wqe->xmit_sequence.wqe_com,
@@ -8120,8 +8233,10 @@ lpfc_sli4_iocb2wqe(struct lpfc_hba *phba, struct lpfc_iocbq *iocbq,
 		command_type = FCP_COMMAND_DATA_OUT;
 		/* word3 iocb=iotag wqe=payload_offset_len */
 		/* Add the FCP_CMD and FCP_RSP sizes to get the offset */
-		wqe->fcp_iwrite.payload_offset_len =
-			xmit_len + sizeof(struct fcp_rsp);
+		bf_set(payload_offset_len, &wqe->fcp_iwrite,
+		       xmit_len + sizeof(struct fcp_rsp));
+		bf_set(cmd_buff_len, &wqe->fcp_iwrite,
+		       0);
 		/* word4 iocb=parameter wqe=total_xfer_length memcpy */
 		/* word5 iocb=initial_xfer_len wqe=initial_xfer_len memcpy */
 		bf_set(wqe_erp, &wqe->fcp_iwrite.wqe_com,
@@ -8139,8 +8254,10 @@ lpfc_sli4_iocb2wqe(struct lpfc_hba *phba, struct lpfc_iocbq *iocbq,
 	case CMD_FCP_IREAD64_CR:
 		/* word3 iocb=iotag wqe=payload_offset_len */
 		/* Add the FCP_CMD and FCP_RSP sizes to get the offset */
-		wqe->fcp_iread.payload_offset_len =
-			xmit_len + sizeof(struct fcp_rsp);
+		bf_set(payload_offset_len, &wqe->fcp_iread,
+		       xmit_len + sizeof(struct fcp_rsp));
+		bf_set(cmd_buff_len, &wqe->fcp_iread,
+		       0);
 		/* word4 iocb=parameter wqe=total_xfer_length memcpy */
 		/* word5 iocb=initial_xfer_len wqe=initial_xfer_len memcpy */
 		bf_set(wqe_erp, &wqe->fcp_iread.wqe_com,
@@ -8156,8 +8273,13 @@ lpfc_sli4_iocb2wqe(struct lpfc_hba *phba, struct lpfc_iocbq *iocbq,
 		bf_set(wqe_dbde, &wqe->fcp_iread.wqe_com, 1);
 		break;
 	case CMD_FCP_ICMND64_CR:
+		/* word3 iocb=iotag wqe=payload_offset_len */
+		/* Add the FCP_CMD and FCP_RSP sizes to get the offset */
+		bf_set(payload_offset_len, &wqe->fcp_icmd,
+		       xmit_len + sizeof(struct fcp_rsp));
+		bf_set(cmd_buff_len, &wqe->fcp_icmd,
+		       0);
 		/* word3 iocb=IO_TAG wqe=reserved */
-		wqe->fcp_icmd.rsrvd3 = 0;
 		bf_set(wqe_pu, &wqe->fcp_icmd.wqe_com, 0);
 		/* Always open the exchange */
 		bf_set(wqe_xc, &wqe->fcp_icmd.wqe_com, 0);
@@ -8203,6 +8325,7 @@ lpfc_sli4_iocb2wqe(struct lpfc_hba *phba, struct lpfc_iocbq *iocbq,
 		bf_set(wqe_qosd, &wqe->gen_req.wqe_com, 1);
 		bf_set(wqe_lenloc, &wqe->gen_req.wqe_com, LPFC_WQE_LENLOC_NONE);
 		bf_set(wqe_ebde_cnt, &wqe->gen_req.wqe_com, 0);
+		wqe->gen_req.max_response_payload_len = total_len - xmit_len;
 		command_type = OTHER_COMMAND;
 		break;
 	case CMD_XMIT_ELS_RSP64_CX:
@@ -10073,6 +10196,11 @@ lpfc_sli_issue_iocb_wait(struct lpfc_hba *phba,
 		if (iocb_completed) {
 			lpfc_printf_log(phba, KERN_INFO, LOG_SLI,
 					"0331 IOCB wake signaled\n");
+			/* Note: we are not indicating if the IOCB has a success
+			 * status or not - that's for the caller to check.
+			 * IOCB_SUCCESS means just that the command was sent and
+			 * completed. Not that it completed successfully.
+			 * */
 		} else if (timeleft == 0) {
 			lpfc_printf_log(phba, KERN_ERR, LOG_SLI,
 					"0338 IOCB wait timeout error - no "
@@ -11074,8 +11202,11 @@ lpfc_sli4_iocb_param_transfer(struct lpfc_hba *phba,
 			      struct lpfc_iocbq *pIocbOut,
 			      struct lpfc_wcqe_complete *wcqe)
 {
+	int numBdes, i;
 	unsigned long iflags;
-	uint32_t status;
+	uint32_t status, max_response;
+	struct lpfc_dmabuf *dmabuf;
+	struct ulp_bde64 *bpl, bde;
 	size_t offset = offsetof(struct lpfc_iocbq, iocb);
 
 	memcpy((char *)pIocbIn + offset, (char *)pIocbOut + offset,
@@ -11092,7 +11223,36 @@ lpfc_sli4_iocb_param_transfer(struct lpfc_hba *phba,
 			pIocbIn->iocb.un.ulpWord[4] = wcqe->parameter;
 	else {
 		pIocbIn->iocb.un.ulpWord[4] = wcqe->parameter;
-		pIocbIn->iocb.un.genreq64.bdl.bdeSize = wcqe->total_data_placed;
+		switch (pIocbOut->iocb.ulpCommand) {
+		case CMD_ELS_REQUEST64_CR:
+			dmabuf = (struct lpfc_dmabuf *)pIocbOut->context3;
+			bpl  = (struct ulp_bde64 *)dmabuf->virt;
+			bde.tus.w = le32_to_cpu(bpl[1].tus.w);
+			max_response = bde.tus.f.bdeSize;
+			break;
+		case CMD_GEN_REQUEST64_CR:
+			max_response = 0;
+			if (!pIocbOut->context3)
+				break;
+			numBdes = pIocbOut->iocb.un.genreq64.bdl.bdeSize/
+					sizeof(struct ulp_bde64);
+			dmabuf = (struct lpfc_dmabuf *)pIocbOut->context3;
+			bpl = (struct ulp_bde64 *)dmabuf->virt;
+			for (i = 0; i < numBdes; i++) {
+				bde.tus.w = le32_to_cpu(bpl[i].tus.w);
+				if (bde.tus.f.bdeFlags != BUFF_TYPE_BDE_64)
+					max_response += bde.tus.f.bdeSize;
+			}
+			break;
+		default:
+			max_response = wcqe->total_data_placed;
+			break;
+		}
+		if (max_response < wcqe->total_data_placed)
+			pIocbIn->iocb.un.genreq64.bdl.bdeSize = max_response;
+		else
+			pIocbIn->iocb.un.genreq64.bdl.bdeSize =
+				wcqe->total_data_placed;
 	}
 
 	/* Convert BG errors for completion status */
@@ -15098,6 +15258,7 @@ lpfc_sli4_alloc_rpi(struct lpfc_hba *phba)
 	uint16_t max_rpi, rpi_limit;
 	uint16_t rpi_remaining, lrpi = 0;
 	struct lpfc_rpi_hdr *rpi_hdr;
+	unsigned long iflag;
 
 	max_rpi = phba->sli4_hba.max_cfg_param.max_rpi;
 	rpi_limit = phba->sli4_hba.next_rpi;
@@ -15106,7 +15267,7 @@ lpfc_sli4_alloc_rpi(struct lpfc_hba *phba)
 	 * Fetch the next logical rpi.  Because this index is logical,
 	 * the  driver starts at 0 each time.
 	 */
-	spin_lock_irq(&phba->hbalock);
+	spin_lock_irqsave(&phba->hbalock, iflag);
 	rpi = find_next_zero_bit(phba->sli4_hba.rpi_bmask, rpi_limit, 0);
 	if (rpi >= rpi_limit)
 		rpi = LPFC_RPI_ALLOC_ERROR;
@@ -15122,7 +15283,7 @@ lpfc_sli4_alloc_rpi(struct lpfc_hba *phba)
 	 */
 	if ((rpi == LPFC_RPI_ALLOC_ERROR) &&
 	    (phba->sli4_hba.rpi_count >= max_rpi)) {
-		spin_unlock_irq(&phba->hbalock);
+		spin_unlock_irqrestore(&phba->hbalock, iflag);
 		return rpi;
 	}
 
@@ -15131,7 +15292,7 @@ lpfc_sli4_alloc_rpi(struct lpfc_hba *phba)
 	 * extents.
 	 */
 	if (!phba->sli4_hba.rpi_hdrs_in_use) {
-		spin_unlock_irq(&phba->hbalock);
+		spin_unlock_irqrestore(&phba->hbalock, iflag);
 		return rpi;
 	}
 
@@ -15142,7 +15303,7 @@ lpfc_sli4_alloc_rpi(struct lpfc_hba *phba)
 	 * how many are supported max by the device.
 	 */
 	rpi_remaining = phba->sli4_hba.next_rpi - phba->sli4_hba.rpi_count;
-	spin_unlock_irq(&phba->hbalock);
+	spin_unlock_irqrestore(&phba->hbalock, iflag);
 	if (rpi_remaining < LPFC_RPI_LOW_WATER_MARK) {
 		rpi_hdr = lpfc_sli4_create_rpi_hdr(phba);
 		if (!rpi_hdr) {