scsi: qedf: Print message during bailout conditions
authorSaurav Kashyap <skashyap@marvell.com>
Fri, 23 Aug 2019 09:52:31 +0000 (02:52 -0700)
committerMartin K. Petersen <martin.petersen@oracle.com>
Thu, 29 Aug 2019 22:44:21 +0000 (18:44 -0400)
Print messages during exiting condition to help debugging.

Signed-off-by: Saurav Kashyap <skashyap@marvell.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
drivers/scsi/qedf/qedf_debugfs.c
drivers/scsi/qedf/qedf_els.c
drivers/scsi/qedf/qedf_fip.c
drivers/scsi/qedf/qedf_io.c
drivers/scsi/qedf/qedf_main.c

index d905a30..b88bed9 100644 (file)
@@ -47,13 +47,13 @@ qedf_dbg_host_init(struct qedf_dbg_ctx *qedf,
  * @pf: the pf that is stopping
  **/
 void
-qedf_dbg_host_exit(struct qedf_dbg_ctx *qedf)
+qedf_dbg_host_exit(struct qedf_dbg_ctx *qedf_dbg)
 {
-       QEDF_INFO(qedf, QEDF_LOG_DEBUGFS, "Destroying debugfs host "
+       QEDF_INFO(qedf_dbg, QEDF_LOG_DEBUGFS, "Destroying debugfs host "
                   "entry\n");
        /* remove debugfs  entries of this PF */
-       debugfs_remove_recursive(qedf->bdf_dentry);
-       qedf->bdf_dentry = NULL;
+       debugfs_remove_recursive(qedf_dbg->bdf_dentry);
+       qedf_dbg->bdf_dentry = NULL;
 }
 
 /**
@@ -140,10 +140,10 @@ qedf_dbg_debug_cmd_read(struct file *filp, char __user *buffer, size_t count,
                        loff_t *ppos)
 {
        int cnt;
-       struct qedf_dbg_ctx *qedf =
+       struct qedf_dbg_ctx *qedf_dbg =
                                (struct qedf_dbg_ctx *)filp->private_data;
 
-       QEDF_INFO(qedf, QEDF_LOG_DEBUGFS, "entered\n");
+       QEDF_INFO(qedf_dbg, QEDF_LOG_DEBUGFS, "debug mask=0x%x\n", qedf_debug);
        cnt = sprintf(buffer, "debug mask = 0x%x\n", qedf_debug);
 
        cnt = min_t(int, count, cnt - *ppos);
@@ -158,7 +158,7 @@ qedf_dbg_debug_cmd_write(struct file *filp, const char __user *buffer,
        uint32_t val;
        void *kern_buf;
        int rval;
-       struct qedf_dbg_ctx *qedf =
+       struct qedf_dbg_ctx *qedf_dbg =
            (struct qedf_dbg_ctx *)filp->private_data;
 
        if (!count || *ppos)
@@ -178,7 +178,7 @@ qedf_dbg_debug_cmd_write(struct file *filp, const char __user *buffer,
        else
                qedf_debug = val;
 
-       QEDF_INFO(qedf, QEDF_LOG_DEBUGFS, "Setting debug=0x%x.\n", val);
+       QEDF_INFO(qedf_dbg, QEDF_LOG_DEBUGFS, "Setting debug=0x%x.\n", val);
        return count;
 }
 
index 5996f68..87e169d 100644 (file)
@@ -179,8 +179,11 @@ static void qedf_rrq_compl(struct qedf_els_cb_arg *cb_arg)
 
        orig_io_req = cb_arg->aborted_io_req;
 
-       if (!orig_io_req)
+       if (!orig_io_req) {
+               QEDF_ERR(&qedf->dbg_ctx,
+                        "Original io_req is NULL, rrq_req = %p.\n", rrq_req);
                goto out_free;
+       }
 
        if (rrq_req->event != QEDF_IOREQ_EV_ELS_TMO &&
            rrq_req->event != QEDF_IOREQ_EV_ELS_ERR_DETECT)
@@ -350,8 +353,10 @@ void qedf_restart_rport(struct qedf_rport *fcport)
        u32 port_id;
        unsigned long flags;
 
-       if (!fcport)
+       if (!fcport) {
+               QEDF_ERR(NULL, "fcport is NULL.\n");
                return;
+       }
 
        spin_lock_irqsave(&fcport->rport_lock, flags);
        if (test_bit(QEDF_RPORT_IN_RESET, &fcport->flags) ||
@@ -418,8 +423,11 @@ static void qedf_l2_els_compl(struct qedf_els_cb_arg *cb_arg)
         * If we are flushing the command just free the cb_arg as none of the
         * response data will be valid.
         */
-       if (els_req->event == QEDF_IOREQ_EV_ELS_FLUSH)
+       if (els_req->event == QEDF_IOREQ_EV_ELS_FLUSH) {
+               QEDF_ERR(NULL, "els_req xid=0x%x event is flush.\n",
+                        els_req->xid);
                goto free_arg;
+       }
 
        fcport = els_req->fcport;
        mp_req = &(els_req->mp_req);
@@ -532,8 +540,10 @@ static void qedf_srr_compl(struct qedf_els_cb_arg *cb_arg)
 
        orig_io_req = cb_arg->aborted_io_req;
 
-       if (!orig_io_req)
+       if (!orig_io_req) {
+               QEDF_ERR(NULL, "orig_io_req is NULL.\n");
                goto out_free;
+       }
 
        clear_bit(QEDF_CMD_SRR_SENT, &orig_io_req->flags);
 
@@ -547,8 +557,11 @@ static void qedf_srr_compl(struct qedf_els_cb_arg *cb_arg)
                   orig_io_req, orig_io_req->xid, srr_req->xid, refcount);
 
        /* If a SRR times out, simply free resources */
-       if (srr_req->event == QEDF_IOREQ_EV_ELS_TMO)
+       if (srr_req->event == QEDF_IOREQ_EV_ELS_TMO) {
+               QEDF_ERR(&qedf->dbg_ctx,
+                        "ELS timeout rec_xid=0x%x.\n", srr_req->xid);
                goto out_put;
+       }
 
        /* Normalize response data into struct fc_frame */
        mp_req = &(srr_req->mp_req);
@@ -721,8 +734,11 @@ void qedf_process_seq_cleanup_compl(struct qedf_ctx *qedf,
        cb_arg = io_req->cb_arg;
 
        /* If we timed out just free resources */
-       if (io_req->event == QEDF_IOREQ_EV_ELS_TMO || !cqe)
+       if (io_req->event == QEDF_IOREQ_EV_ELS_TMO || !cqe) {
+               QEDF_ERR(&qedf->dbg_ctx,
+                        "cqe is NULL or timeout event (0x%x)", io_req->event);
                goto free;
+       }
 
        /* Kill the timer we put on the request */
        cancel_delayed_work_sync(&io_req->timeout_work);
@@ -825,8 +841,10 @@ static void qedf_rec_compl(struct qedf_els_cb_arg *cb_arg)
 
        orig_io_req = cb_arg->aborted_io_req;
 
-       if (!orig_io_req)
+       if (!orig_io_req) {
+               QEDF_ERR(NULL, "orig_io_req is NULL.\n");
                goto out_free;
+       }
 
        if (rec_req->event != QEDF_IOREQ_EV_ELS_TMO &&
            rec_req->event != QEDF_IOREQ_EV_ELS_ERR_DETECT)
@@ -838,8 +856,12 @@ static void qedf_rec_compl(struct qedf_els_cb_arg *cb_arg)
                   orig_io_req, orig_io_req->xid, rec_req->xid, refcount);
 
        /* If a REC times out, free resources */
-       if (rec_req->event == QEDF_IOREQ_EV_ELS_TMO)
+       if (rec_req->event == QEDF_IOREQ_EV_ELS_TMO) {
+               QEDF_ERR(&qedf->dbg_ctx,
+                        "Got TMO event, orig_io_req %p orig_io_xid=0x%x.\n",
+                        orig_io_req, orig_io_req->xid);
                goto out_put;
+       }
 
        /* Normalize response data into struct fc_frame */
        mp_req = &(rec_req->mp_req);
index 362d2be..5143d93 100644 (file)
@@ -23,8 +23,11 @@ void qedf_fcoe_send_vlan_req(struct qedf_ctx *qedf)
        int rc = -1;
 
        skb = dev_alloc_skb(sizeof(struct fip_vlan));
-       if (!skb)
+       if (!skb) {
+               QEDF_ERR(&qedf->dbg_ctx,
+                        "Failed to allocate skb.\n");
                return;
+       }
 
        eth_fr = (char *)skb->data;
        vlan = (struct fip_vlan *)eth_fr;
index d881e82..5b42892 100644 (file)
@@ -104,6 +104,8 @@ static void qedf_cmd_timeout(struct work_struct *work)
                qedf_process_seq_cleanup_compl(qedf, NULL, io_req);
                break;
        default:
+               QEDF_INFO(&qedf->dbg_ctx, QEDF_LOG_IO,
+                         "Hit default case, xid=0x%x.\n", io_req->xid);
                break;
        }
 }
@@ -122,8 +124,10 @@ void qedf_cmd_mgr_free(struct qedf_cmd_mgr *cmgr)
        num_ios = max_xid - min_xid + 1;
 
        /* Free fcoe_bdt_ctx structures */
-       if (!cmgr->io_bdt_pool)
+       if (!cmgr->io_bdt_pool) {
+               QEDF_ERR(&qedf->dbg_ctx, "io_bdt_pool is NULL.\n");
                goto free_cmd_pool;
+       }
 
        bd_tbl_sz = QEDF_MAX_BDS_PER_CMD * sizeof(struct scsi_sge);
        for (i = 0; i < num_ios; i++) {
@@ -226,8 +230,11 @@ struct qedf_cmd_mgr *qedf_cmd_mgr_alloc(struct qedf_ctx *qedf)
                io_req->sense_buffer = dma_alloc_coherent(&qedf->pdev->dev,
                    QEDF_SCSI_SENSE_BUFFERSIZE, &io_req->sense_buffer_dma,
                    GFP_KERNEL);
-               if (!io_req->sense_buffer)
+               if (!io_req->sense_buffer) {
+                       QEDF_ERR(&qedf->dbg_ctx,
+                                "Failed to alloc sense buffer.\n");
                        goto mem_err;
+               }
 
                /* Allocate task parameters to pass to f/w init funcions */
                io_req->task_params = kzalloc(sizeof(*io_req->task_params),
@@ -437,8 +444,12 @@ void qedf_release_cmd(struct kref *ref)
        struct qedf_rport *fcport = io_req->fcport;
        unsigned long flags;
 
-       if (io_req->cmd_type == QEDF_SCSI_CMD)
+       if (io_req->cmd_type == QEDF_SCSI_CMD) {
+               QEDF_WARN(&fcport->qedf->dbg_ctx,
+                         "Cmd released called without scsi_done called, io_req %p xid=0x%x.\n",
+                         io_req, io_req->xid);
                WARN_ON(io_req->sc_cmd);
+       }
 
        if (io_req->cmd_type == QEDF_ELS ||
            io_req->cmd_type == QEDF_TASK_MGMT_CMD)
@@ -447,8 +458,10 @@ void qedf_release_cmd(struct kref *ref)
        atomic_inc(&cmd_mgr->free_list_cnt);
        atomic_dec(&fcport->num_active_ios);
        atomic_set(&io_req->state, QEDF_CMD_ST_INACTIVE);
-       if (atomic_read(&fcport->num_active_ios) < 0)
+       if (atomic_read(&fcport->num_active_ios) < 0) {
                QEDF_WARN(&(fcport->qedf->dbg_ctx), "active_ios < 0.\n");
+               WARN_ON(1);
+       }
 
        /* Increment task retry identifier now that the request is released */
        io_req->task_retry_identifier++;
@@ -951,6 +964,9 @@ qedf_queuecommand(struct Scsi_Host *host, struct scsi_cmnd *sc_cmd)
 
        if (test_bit(QEDF_UNLOADING, &qedf->flags) ||
            test_bit(QEDF_DBG_STOP_IO, &qedf->flags)) {
+               QEDF_INFO(&qedf->dbg_ctx, QEDF_LOG_IO,
+                         "Returning DNC as unloading or stop io, flags 0x%lx.\n",
+                         qedf->flags);
                sc_cmd->result = DID_NO_CONNECT << 16;
                sc_cmd->scsi_done(sc_cmd);
                return 0;
@@ -967,6 +983,9 @@ qedf_queuecommand(struct Scsi_Host *host, struct scsi_cmnd *sc_cmd)
 
        rval = fc_remote_port_chkready(rport);
        if (rval) {
+               QEDF_INFO(&qedf->dbg_ctx, QEDF_LOG_IO,
+                         "fc_remote_port_chkready failed=0x%x for port_id=0x%06x.\n",
+                         rval, rport->port_id);
                sc_cmd->result = rval;
                sc_cmd->scsi_done(sc_cmd);
                return 0;
@@ -974,12 +993,14 @@ qedf_queuecommand(struct Scsi_Host *host, struct scsi_cmnd *sc_cmd)
 
        /* Retry command if we are doing a qed drain operation */
        if (test_bit(QEDF_DRAIN_ACTIVE, &qedf->flags)) {
+               QEDF_INFO(&qedf->dbg_ctx, QEDF_LOG_IO, "Drain active.\n");
                rc = SCSI_MLQUEUE_HOST_BUSY;
                goto exit_qcmd;
        }
 
        if (lport->state != LPORT_ST_READY ||
            atomic_read(&qedf->link_state) != QEDF_LINK_UP) {
+               QEDF_INFO(&qedf->dbg_ctx, QEDF_LOG_IO, "Link down.\n");
                rc = SCSI_MLQUEUE_HOST_BUSY;
                goto exit_qcmd;
        }
@@ -1297,8 +1318,10 @@ void qedf_scsi_done(struct qedf_ctx *qedf, struct qedf_ioreq *io_req,
        struct scsi_cmnd *sc_cmd;
        int refcount;
 
-       if (!io_req)
+       if (!io_req) {
+               QEDF_INFO(&qedf->dbg_ctx, QEDF_LOG_IO, "io_req is NULL\n");
                return;
+       }
 
        if (test_and_set_bit(QEDF_CMD_ERR_SCSI_DONE, &io_req->flags)) {
                QEDF_INFO(&qedf->dbg_ctx, QEDF_LOG_IO,
@@ -1414,8 +1437,12 @@ void qedf_process_warning_compl(struct qedf_ctx *qedf, struct fcoe_cqe *cqe,
        u64 err_warn_bit_map;
        u8 err_warn = 0xff;
 
-       if (!cqe)
+       if (!cqe) {
+               QEDF_INFO(&qedf->dbg_ctx, QEDF_LOG_IO,
+                         "cqe is NULL for io_req %p xid=0x%x\n",
+                         io_req, io_req->xid);
                return;
+       }
 
        QEDF_ERR(&(io_req->fcport->qedf->dbg_ctx), "Warning CQE, "
                  "xid=0x%x\n", io_req->xid);
@@ -1477,8 +1504,11 @@ void qedf_process_error_detect(struct qedf_ctx *qedf, struct fcoe_cqe *cqe,
 {
        int rval;
 
-       if (!cqe)
+       if (!cqe) {
+               QEDF_INFO(&qedf->dbg_ctx, QEDF_LOG_IO,
+                         "cqe is NULL for io_req %p\n", io_req);
                return;
+       }
 
        QEDF_ERR(&(io_req->fcport->qedf->dbg_ctx), "Error detection CQE, "
                  "xid=0x%x\n", io_req->xid);
@@ -1543,8 +1573,10 @@ void qedf_flush_active_ios(struct qedf_rport *fcport, int lun)
        int wait_cnt = 100;
        int refcount = 0;
 
-       if (!fcport)
+       if (!fcport) {
+               QEDF_ERR(NULL, "fcport is NULL\n");
                return;
+       }
 
        /* Check that fcport is still offloaded */
        if (!test_bit(QEDF_RPORT_SESSION_READY, &fcport->flags)) {
@@ -1976,6 +2008,10 @@ void qedf_process_abts_compl(struct qedf_ctx *qedf, struct fcoe_cqe *cqe,
        clear_bit(QEDF_CMD_IN_ABORT, &io_req->flags);
 
        if (io_req->sc_cmd) {
+               if (!io_req->return_scsi_cmd_on_abts)
+                       QEDF_INFO(&qedf->dbg_ctx, QEDF_LOG_SCSI_TM,
+                                 "Not call scsi_done for xid=0x%x.\n",
+                                 io_req->xid);
                if (io_req->return_scsi_cmd_on_abts)
                        qedf_scsi_done(qedf, io_req, DID_ERROR);
        }
@@ -2201,6 +2237,10 @@ int qedf_initiate_cleanup(struct qedf_ioreq *io_req,
        }
 
        if (io_req->sc_cmd) {
+               if (!io_req->return_scsi_cmd_on_abts)
+                       QEDF_INFO(&qedf->dbg_ctx, QEDF_LOG_SCSI_TM,
+                                 "Not call scsi_done for xid=0x%x.\n",
+                                 io_req->xid);
                if (io_req->return_scsi_cmd_on_abts)
                        qedf_scsi_done(qedf, io_req, DID_ERROR);
        }
@@ -2241,7 +2281,7 @@ static int qedf_execute_tmf(struct qedf_rport *fcport, struct scsi_cmnd *sc_cmd,
        u16 sqe_idx;
 
        if (!sc_cmd) {
-               QEDF_ERR(&(qedf->dbg_ctx), "invalid arg\n");
+               QEDF_ERR(&qedf->dbg_ctx, "sc_cmd is NULL\n");
                return FAILED;
        }
 
index a42babd..3e245b0 100644 (file)
@@ -128,8 +128,11 @@ static bool qedf_initiate_fipvlan_req(struct qedf_ctx *qedf)
                        return false;
                }
 
-               if (qedf->vlan_id > 0)
+               if (qedf->vlan_id > 0) {
+                       QEDF_INFO(&qedf->dbg_ctx, QEDF_LOG_DISC,
+                                 "vlan = 0x%x already set.\n", qedf->vlan_id);
                        return true;
+               }
 
                QEDF_INFO(&(qedf->dbg_ctx), QEDF_LOG_DISC,
                           "Retry %d.\n", qedf->fipvlan_retries);
@@ -162,6 +165,8 @@ static void qedf_handle_link_update(struct work_struct *work)
                        return;
 
                if (atomic_read(&qedf->link_state) != QEDF_LINK_UP) {
+                       QEDF_INFO(&qedf->dbg_ctx, QEDF_LOG_DISC,
+                                 "Link is down, resetting vlan_id.\n");
                        qedf->vlan_id = 0;
                        return;
                }
@@ -311,8 +316,10 @@ int qedf_send_flogi(struct qedf_ctx *qedf)
 
        lport = qedf->lport;
 
-       if (!lport->tt.elsct_send)
+       if (!lport->tt.elsct_send) {
+               QEDF_ERR(&qedf->dbg_ctx, "tt.elsct_send not set.\n");
                return -EINVAL;
+       }
 
        fp = fc_frame_alloc(lport, sizeof(struct fc_els_flogi));
        if (!fp) {
@@ -2340,12 +2347,14 @@ static void qedf_recv_frame(struct qedf_ctx *qedf,
        fr_dev(fp) = lport;
        fr_sof(fp) = hp->fcoe_sof;
        if (skb_copy_bits(skb, fr_len, &crc_eof, sizeof(crc_eof))) {
+               QEDF_INFO(NULL, QEDF_LOG_LL2, "skb_copy_bits failed.\n");
                kfree_skb(skb);
                return;
        }
        fr_eof(fp) = crc_eof.fcoe_eof;
        fr_crc(fp) = crc_eof.fcoe_crc32;
        if (pskb_trim(skb, fr_len)) {
+               QEDF_INFO(NULL, QEDF_LOG_LL2, "pskb_trim failed.\n");
                kfree_skb(skb);
                return;
        }
@@ -2406,9 +2415,9 @@ static void qedf_recv_frame(struct qedf_ctx *qedf,
         * empty then this is not addressed to our port so simply drop it.
         */
        if (lport->port_id != ntoh24(fh->fh_d_id) && !vn_port) {
-               QEDF_INFO(&(qedf->dbg_ctx), QEDF_LOG_LL2,
-                   "Dropping frame due to destination mismatch: lport->port_id=%x fh->d_id=%x.\n",
-                   lport->port_id, ntoh24(fh->fh_d_id));
+               QEDF_INFO(&qedf->dbg_ctx, QEDF_LOG_LL2,
+                         "Dropping frame due to destination mismatch: lport->port_id=0x%x fh->d_id=0x%x.\n",
+                         lport->port_id, ntoh24(fh->fh_d_id));
                kfree_skb(skb);
                return;
        }
@@ -2417,6 +2426,8 @@ static void qedf_recv_frame(struct qedf_ctx *qedf,
        if ((fh->fh_type == FC_TYPE_BLS) && (f_ctl & FC_FC_SEQ_CTX) &&
            (f_ctl & FC_FC_EX_CTX)) {
                /* Drop incoming ABTS response that has both SEQ/EX CTX set */
+               QEDF_INFO(&qedf->dbg_ctx, QEDF_LOG_LL2,
+                         "Dropping ABTS response as both SEQ/EX CTX set.\n");
                kfree_skb(skb);
                return;
        }
@@ -2560,8 +2571,9 @@ static int qedf_alloc_and_init_sb(struct qedf_ctx *qedf,
            sizeof(struct status_block_e4), &sb_phys, GFP_KERNEL);
 
        if (!sb_virt) {
-               QEDF_ERR(&(qedf->dbg_ctx), "Status block allocation failed "
-                         "for id = %d.\n", sb_id);
+               QEDF_ERR(&qedf->dbg_ctx,
+                        "Status block allocation failed for id = %d.\n",
+                        sb_id);
                return -ENOMEM;
        }
 
@@ -2569,8 +2581,9 @@ static int qedf_alloc_and_init_sb(struct qedf_ctx *qedf,
            sb_id, QED_SB_TYPE_STORAGE);
 
        if (ret) {
-               QEDF_ERR(&(qedf->dbg_ctx), "Status block initialization "
-                         "failed for id = %d.\n", sb_id);
+               QEDF_ERR(&qedf->dbg_ctx,
+                        "Status block initialization failed (0x%x) for id = %d.\n",
+                        ret, sb_id);
                return ret;
        }
 
@@ -2654,13 +2667,18 @@ void qedf_process_cqe(struct qedf_ctx *qedf, struct fcoe_cqe *cqe)
        io_req = &qedf->cmd_mgr->cmds[xid];
 
        /* Completion not for a valid I/O anymore so just return */
-       if (!io_req)
+       if (!io_req) {
+               QEDF_ERR(&qedf->dbg_ctx,
+                        "io_req is NULL for xid=0x%x.\n", xid);
                return;
+       }
 
        fcport = io_req->fcport;
 
        if (fcport == NULL) {
-               QEDF_ERR(&(qedf->dbg_ctx), "fcport is NULL.\n");
+               QEDF_ERR(&qedf->dbg_ctx,
+                        "fcport is NULL for xid=0x%x io_req=%p.\n",
+                        xid, io_req);
                return;
        }
 
@@ -2669,7 +2687,8 @@ void qedf_process_cqe(struct qedf_ctx *qedf, struct fcoe_cqe *cqe)
         * isn't valid and shouldn't be taken. We should just return.
         */
        if (!test_bit(QEDF_RPORT_SESSION_READY, &fcport->flags)) {
-               QEDF_ERR(&(qedf->dbg_ctx), "Session not offloaded yet.\n");
+               QEDF_ERR(&qedf->dbg_ctx,
+                        "Session not offloaded yet, fcport = %p.\n", fcport);
                return;
        }
 
@@ -2881,6 +2900,7 @@ static int qedf_alloc_global_queues(struct qedf_ctx *qedf)
         */
        if (!qedf->p_cpuq) {
                status = 1;
+               QEDF_ERR(&qedf->dbg_ctx, "p_cpuq is NULL.\n");
                goto mem_alloc_failure;
        }
 
@@ -2896,8 +2916,10 @@ static int qedf_alloc_global_queues(struct qedf_ctx *qedf)
 
        /* Allocate DMA coherent buffers for BDQ */
        rc = qedf_alloc_bdq(qedf);
-       if (rc)
+       if (rc) {
+               QEDF_ERR(&qedf->dbg_ctx, "Unable to allocate bdq.\n");
                goto mem_alloc_failure;
+       }
 
        /* Allocate a CQ and an associated PBL for each MSI-X vector */
        for (i = 0; i < qedf->num_queues; i++) {
@@ -3209,6 +3231,7 @@ static int __qedf_probe(struct pci_dev *pdev, int mode)
        qed_params.is_vf = is_vf;
        qedf->cdev = qed_ops->common->probe(pdev, &qed_params);
        if (!qedf->cdev) {
+               QEDF_ERR(&qedf->dbg_ctx, "common probe failed.\n");
                rc = -ENODEV;
                goto err1;
        }
@@ -3277,8 +3300,10 @@ static int __qedf_probe(struct pci_dev *pdev, int mode)
 
        /* Setup interrupts */
        rc = qedf_setup_int(qedf);
-       if (rc)
+       if (rc) {
+               QEDF_ERR(&qedf->dbg_ctx, "Setup interrupts failed.\n");
                goto err3;
+       }
 
        rc = qed_ops->start(qedf->cdev, &qedf->tasks);
        if (rc) {