[SCSI] lpfc 8.2.4 : Rework misplaced reference taking on node structure

Rework misplaced reference taking on node structure

Signed-off-by: James Smart <James.Smart@emulex.com>
Signed-off-by: James Bottomley <James.Bottomley@HansenPartnership.com>

5 files changed, 157 insertions, 58 deletions

diff --git a/drivers/scsi/lpfc/lpfc_ct.c b/drivers/scsi/lpfc/lpfc_ct.c
index 3759ae1dc5e7..92441ce610ed 100644
--- a/drivers/scsi/lpfc/lpfc_ct.c
+++ b/drivers/scsi/lpfc/lpfc_ct.c
@@ -561,7 +561,6 @@ lpfc_cmpl_ct_cmd_gid_ft(struct lpfc_hba *phba, struct lpfc_iocbq *cmdiocb,
         if (vport->load_flag & FC_UNLOADING)
                 goto out;
 
-
         if (lpfc_els_chk_latt(vport) || lpfc_error_lost_link(irsp)) {
                 lpfc_printf_vlog(vport, KERN_INFO, LOG_DISCOVERY,
                                  "0216 Link event during NS query\n");
diff --git a/drivers/scsi/lpfc/lpfc_els.c b/drivers/scsi/lpfc/lpfc_els.c
index 8da6e8be9d83..c6b739dc6bc3 100644
--- a/drivers/scsi/lpfc/lpfc_els.c
+++ b/drivers/scsi/lpfc/lpfc_els.c
@@ -120,12 +120,8 @@ lpfc_prep_els_iocb(struct lpfc_vport *vport, uint8_t expectRsp,
         pcmd = kmalloc(sizeof(struct lpfc_dmabuf), GFP_KERNEL);
         if (pcmd)
                 pcmd->virt = lpfc_mbuf_alloc(phba, MEM_PRI, &pcmd->phys);
-        if (!pcmd || !pcmd->virt) {
-                kfree(pcmd);
-
-                lpfc_sli_release_iocbq(phba, elsiocb);
-                return NULL;
-        }
+        if (!pcmd || !pcmd->virt)
+                goto els_iocb_free_pcmb_exit;
 
         INIT_LIST_HEAD(&pcmd->list);
 
@@ -135,13 +131,8 @@ lpfc_prep_els_iocb(struct lpfc_vport *vport, uint8_t expectRsp,
                 if (prsp)
                         prsp->virt = lpfc_mbuf_alloc(phba, MEM_PRI,
                                                      &prsp->phys);
-                if (!prsp || !prsp->virt) {
-                        kfree(prsp);
-                        lpfc_mbuf_free(phba, pcmd->virt, pcmd->phys);
-                        kfree(pcmd);
-                        lpfc_sli_release_iocbq(phba, elsiocb);
-                        return NULL;
-                }
+                if (!prsp || !prsp->virt)
+                        goto els_iocb_free_prsp_exit;
                 INIT_LIST_HEAD(&prsp->list);
         } else {
                 prsp = NULL;
@@ -152,15 +143,8 @@ lpfc_prep_els_iocb(struct lpfc_vport *vport, uint8_t expectRsp,
         if (pbuflist)
                 pbuflist->virt = lpfc_mbuf_alloc(phba, MEM_PRI,
                                                  &pbuflist->phys);
-        if (!pbuflist || !pbuflist->virt) {
-                lpfc_sli_release_iocbq(phba, elsiocb);
-                lpfc_mbuf_free(phba, pcmd->virt, pcmd->phys);
-                lpfc_mbuf_free(phba, prsp->virt, prsp->phys);
-                kfree(pcmd);
-                kfree(prsp);
-                kfree(pbuflist);
-                return NULL;
-        }
+        if (!pbuflist || !pbuflist->virt)
+                goto els_iocb_free_pbuf_exit;
 
         INIT_LIST_HEAD(&pbuflist->list);
 
@@ -205,7 +189,10 @@ lpfc_prep_els_iocb(struct lpfc_vport *vport, uint8_t expectRsp,
                 bpl->tus.w = le32_to_cpu(bpl->tus.w);
         }
 
+        /* prevent preparing iocb with NULL ndlp reference */
         elsiocb->context1 = lpfc_nlp_get(ndlp);
+        if (!elsiocb->context1)
+                goto els_iocb_free_pbuf_exit;
         elsiocb->context2 = pcmd;
         elsiocb->context3 = pbuflist;
         elsiocb->retry = retry;
@@ -231,8 +218,20 @@ lpfc_prep_els_iocb(struct lpfc_vport *vport, uint8_t expectRsp,
                                  cmdSize);
         }
         return elsiocb;
-}
 
+els_iocb_free_pbuf_exit:
+        lpfc_mbuf_free(phba, prsp->virt, prsp->phys);
+        kfree(pbuflist);
+
+els_iocb_free_prsp_exit:
+        lpfc_mbuf_free(phba, pcmd->virt, pcmd->phys);
+        kfree(prsp);
+
+els_iocb_free_pcmb_exit:
+        kfree(pcmd);
+        lpfc_sli_release_iocbq(phba, elsiocb);
+        return NULL;
+}
 
 static int
 lpfc_issue_fabric_reglogin(struct lpfc_vport *vport)
@@ -513,6 +512,9 @@ lpfc_cmpl_els_flogi(struct lpfc_hba *phba, struct lpfc_iocbq *cmdiocb,
 
         /* Check to see if link went down during discovery */
         if (lpfc_els_chk_latt(vport)) {
+                /* One additional decrement on node reference count to
+                 * trigger the release of the node
+                 */
                 lpfc_nlp_put(ndlp);
                 goto out;
         }
@@ -731,6 +733,9 @@ lpfc_initial_flogi(struct lpfc_vport *vport)
         }
 
         if (lpfc_issue_els_flogi(vport, ndlp, 0)) {
+                /* This decrement of reference count to node shall kick off
+                 * the release of the node.
+                 */
                 lpfc_nlp_put(ndlp);
         }
         return 1;
@@ -754,7 +759,11 @@ lpfc_initial_fdisc(struct lpfc_vport *vport)
                 lpfc_dequeue_node(vport, ndlp);
         }
         if (lpfc_issue_els_fdisc(vport, ndlp, 0)) {
+                /* decrement node reference count to trigger the release of
+                 * the node.
+                 */
                 lpfc_nlp_put(ndlp);
+                return 0;
         }
         return 1;
 }
@@ -1557,6 +1566,9 @@ lpfc_issue_els_scr(struct lpfc_vport *vport, uint32_t nportid, uint8_t retry)
                                      ndlp->nlp_DID, ELS_CMD_SCR);
 
         if (!elsiocb) {
+                /* This will trigger the release of the node just
+                 * allocated
+                 */
                 lpfc_nlp_put(ndlp);
                 return 1;
         }
@@ -1578,10 +1590,17 @@ lpfc_issue_els_scr(struct lpfc_vport *vport, uint32_t nportid, uint8_t retry)
         phba->fc_stat.elsXmitSCR++;
         elsiocb->iocb_cmpl = lpfc_cmpl_els_cmd;
         if (lpfc_sli_issue_iocb(phba, pring, elsiocb, 0) == IOCB_ERROR) {
+                /* The additional lpfc_nlp_put will cause the following
+                 * lpfc_els_free_iocb routine to trigger the rlease of
+                 * the node.
+                 */
                 lpfc_nlp_put(ndlp);
                 lpfc_els_free_iocb(phba, elsiocb);
                 return 1;
         }
+        /* This will cause the callback-function lpfc_cmpl_els_cmd to
+         * trigger the release of node.
+         */
         lpfc_nlp_put(ndlp);
         return 0;
 }
@@ -1613,6 +1632,9 @@ lpfc_issue_els_farpr(struct lpfc_vport *vport, uint32_t nportid, uint8_t retry)
         elsiocb = lpfc_prep_els_iocb(vport, 1, cmdsize, retry, ndlp,
                                      ndlp->nlp_DID, ELS_CMD_RNID);
         if (!elsiocb) {
+                /* This will trigger the release of the node just
+                 * allocated
+                 */
                 lpfc_nlp_put(ndlp);
                 return 1;
         }
@@ -1649,10 +1671,17 @@ lpfc_issue_els_farpr(struct lpfc_vport *vport, uint32_t nportid, uint8_t retry)
         phba->fc_stat.elsXmitFARPR++;
         elsiocb->iocb_cmpl = lpfc_cmpl_els_cmd;
         if (lpfc_sli_issue_iocb(phba, pring, elsiocb, 0) == IOCB_ERROR) {
+                /* The additional lpfc_nlp_put will cause the following
+                 * lpfc_els_free_iocb routine to trigger the release of
+                 * the node.
+                 */
                 lpfc_nlp_put(ndlp);
                 lpfc_els_free_iocb(phba, elsiocb);
                 return 1;
         }
+        /* This will cause the callback-function lpfc_cmpl_els_cmd to
+         * trigger the release of the node.
+         */
         lpfc_nlp_put(ndlp);
         return 0;
 }
@@ -1712,7 +1741,10 @@ lpfc_els_retry_delay(unsigned long ptr)
                 return;
         }
 
-        evtp->evt_arg1  = ndlp;
+        /* We need to hold the node by incrementing the reference
+         * count until the queued work is done
+         */
+        evtp->evt_arg1  = lpfc_nlp_get(ndlp);
         evtp->evt       = LPFC_EVT_ELS_RETRY;
         list_add_tail(&evtp->evt_listp, &phba->work_list);
         if (phba->work_wait)
@@ -2190,6 +2222,11 @@ lpfc_cmpl_els_logo_acc(struct lpfc_hba *phba, struct lpfc_iocbq *cmdiocb,
                          * thread, just unregister the RPI.
                          */
                         lpfc_unreg_rpi(vport, ndlp);
+                } else {
+                        /* Indicate the node has already released, should
+                         * not reference to it from within lpfc_els_free_iocb.
+                         */
+                        cmdiocb->context1 = NULL;
                 }
         }
         lpfc_els_free_iocb(phba, cmdiocb);
@@ -2208,7 +2245,6 @@ lpfc_mbx_cmpl_dflt_rpi(struct lpfc_hba *phba, LPFC_MBOXQ_t *pmb)
         mempool_free(pmb, phba->mbox_mem_pool);
         if (ndlp) {
                 lpfc_nlp_put(ndlp);
-
                 /* This is the end of the default RPI cleanup logic for this
                  * ndlp. If no other discovery threads are using this ndlp.
                  * we should free all resources associated with it.
@@ -2236,11 +2272,13 @@ lpfc_cmpl_els_rsp(struct lpfc_hba *phba, struct lpfc_iocbq *cmdiocb,
         if (cmdiocb->context_un.mbox)
                 mbox = cmdiocb->context_un.mbox;
 
-        /* First determine if this is a LS_RJT cmpl */
+        /* First determine if this is a LS_RJT cmpl. Note, this callback
+         * function can have cmdiocb->contest1 (ndlp) field set to NULL.
+         */
         pcmd = (uint8_t *) (((struct lpfc_dmabuf *) cmdiocb->context2)->virt);
-        if (*((uint32_t *) (pcmd)) == ELS_CMD_LS_RJT) {
-                /* A LS_RJT associated with Default RPI cleanup
-                 * has its own seperate code path.
+        if (ndlp && (*((uint32_t *) (pcmd)) == ELS_CMD_LS_RJT)) {
+                /* A LS_RJT associated with Default RPI cleanup has its own
+                 * seperate code path.
                  */
                 if (!(ndlp->nlp_flag & NLP_RM_DFLT_RPI))
                         ls_rjt = 1;
@@ -2257,8 +2295,14 @@ lpfc_cmpl_els_rsp(struct lpfc_hba *phba, struct lpfc_iocbq *cmdiocb,
                         mempool_free(mbox, phba->mbox_mem_pool);
                 }
                 if (ndlp && (ndlp->nlp_flag & NLP_RM_DFLT_RPI))
-                        if (lpfc_nlp_not_used(ndlp))
+                        if (lpfc_nlp_not_used(ndlp)) {
                                 ndlp = NULL;
+                                /* Indicate the node has already released,
+                                 * should not reference to it from within
+                                 * the routine lpfc_els_free_iocb.
+                                 */
+                                cmdiocb->context1 = NULL;
+                        }
                 goto out;
         }
 
@@ -2302,14 +2346,27 @@ lpfc_cmpl_els_rsp(struct lpfc_hba *phba, struct lpfc_iocbq *cmdiocb,
                                 ndlp->nlp_DID, ndlp->nlp_flag, ndlp->nlp_state,
                                 ndlp->nlp_rpi);
 
-                        if (lpfc_nlp_not_used(ndlp))
+                        if (lpfc_nlp_not_used(ndlp)) {
                                 ndlp = NULL;
+                                /* Indicate node has already been released,
+                                 * should not reference to it from within
+                                 * the routine lpfc_els_free_iocb.
+                                 */
+                                cmdiocb->context1 = NULL;
+                        }
                 } else {
                         /* Do not drop node for lpfc_els_abort'ed ELS cmds */
                         if (!lpfc_error_lost_link(irsp) &&
                             ndlp->nlp_flag & NLP_ACC_REGLOGIN) {
-                                if (lpfc_nlp_not_used(ndlp))
+                                if (lpfc_nlp_not_used(ndlp)) {
                                         ndlp = NULL;
+                                        /* Indicate node has already been
+                                         * released, should not reference
+                                         * to it from within the routine
+                                         * lpfc_els_free_iocb.
+                                         */
+                                        cmdiocb->context1 = NULL;
+                                }
                         }
                 }
                 mp = (struct lpfc_dmabuf *) mbox->context1;
@@ -2331,7 +2388,12 @@ out:
                  * resources.
                  */
                 if (ls_rjt)
-                        lpfc_nlp_not_used(ndlp);
+                        if (lpfc_nlp_not_used(ndlp))
+                                /* Indicate node has already been released,
+                                 * should not reference to it from within
+                                 * the routine lpfc_els_free_iocb.
+                                 */
+                                cmdiocb->context1 = NULL;
         }
 
         lpfc_els_free_iocb(phba, cmdiocb);
@@ -3292,7 +3354,10 @@ lpfc_els_rsp_rps_acc(struct lpfc_hba *phba, LPFC_MBOXQ_t *pmb)
         elsiocb = lpfc_prep_els_iocb(phba->pport, 0, cmdsize,
                                      lpfc_max_els_tries, ndlp,
                                      ndlp->nlp_DID, ELS_CMD_ACC);
+
+        /* Decrement the ndlp reference count from previous mbox command */
         lpfc_nlp_put(ndlp);
+
         if (!elsiocb)
                 return;
 
@@ -3375,11 +3440,13 @@ lpfc_els_rcv_rps(struct lpfc_vport *vport, struct lpfc_iocbq *cmdiocb,
                         mbox->context2 = lpfc_nlp_get(ndlp);
                         mbox->vport = vport;
                         mbox->mbox_cmpl = lpfc_els_rsp_rps_acc;
-                        if (lpfc_sli_issue_mbox (phba, mbox, MBX_NOWAIT)
+                        if (lpfc_sli_issue_mbox(phba, mbox, MBX_NOWAIT)
                                 != MBX_NOT_FINISHED)
                                 /* Mbox completion will send ELS Response */
                                 return 0;
-
+                        /* Decrement reference count used for the failed mbox
+                         * command.
+                         */
                         lpfc_nlp_put(ndlp);
                         mempool_free(mbox, phba->mbox_mem_pool);
                 }
@@ -4284,7 +4351,6 @@ lpfc_cmpl_reg_new_vport(struct lpfc_hba *phba, LPFC_MBOXQ_t *pmb)
         spin_lock_irq(shost->host_lock);
         vport->fc_flag &= ~FC_VPORT_NEEDS_REG_VPI;
         spin_unlock_irq(shost->host_lock);
-        lpfc_nlp_put(ndlp);
 
         if (mb->mbxStatus) {
                 lpfc_printf_vlog(vport, KERN_ERR, LOG_MBOX,
@@ -4317,6 +4383,12 @@ lpfc_cmpl_reg_new_vport(struct lpfc_hba *phba, LPFC_MBOXQ_t *pmb)
                 else
                         lpfc_do_scr_ns_plogi(phba, vport);
         }
+
+        /* Now, we decrement the ndlp reference count held for this
+         * callback function
+         */
+        lpfc_nlp_put(ndlp);
+
         mempool_free(pmb, phba->mbox_mem_pool);
         return;
 }
@@ -4336,26 +4408,29 @@ lpfc_register_new_vport(struct lpfc_hba *phba, struct lpfc_vport *vport,
                 mbox->mbox_cmpl = lpfc_cmpl_reg_new_vport;
                 if (lpfc_sli_issue_mbox(phba, mbox, MBX_NOWAIT)
                     == MBX_NOT_FINISHED) {
+                        /* mailbox command not success, decrement ndlp
+                         * reference count for this command
+                         */
+                        lpfc_nlp_put(ndlp);
                         mempool_free(mbox, phba->mbox_mem_pool);
-                        spin_lock_irq(shost->host_lock);
-                        vport->fc_flag &= ~FC_VPORT_NEEDS_REG_VPI;
-                        spin_unlock_irq(shost->host_lock);
 
-                        lpfc_vport_set_state(vport, FC_VPORT_FAILED);
                         lpfc_printf_vlog(vport, KERN_ERR, LOG_MBOX,
                                 "0253 Register VPI: Can't send mbox\n");
+                        goto mbox_err_exit;
                 }
         } else {
-                lpfc_vport_set_state(vport, FC_VPORT_FAILED);
-
                 lpfc_printf_vlog(vport, KERN_ERR, LOG_MBOX,
                                  "0254 Register VPI: no memory\n");
-
-                spin_lock_irq(shost->host_lock);
-                vport->fc_flag &= ~FC_VPORT_NEEDS_REG_VPI;
-                spin_unlock_irq(shost->host_lock);
-                lpfc_nlp_put(ndlp);
+                goto mbox_err_exit;
         }
+        return;
+
+mbox_err_exit:
+        lpfc_vport_set_state(vport, FC_VPORT_FAILED);
+        spin_lock_irq(shost->host_lock);
+        vport->fc_flag &= ~FC_VPORT_NEEDS_REG_VPI;
+        spin_unlock_irq(shost->host_lock);
+        return;
 }
 
 static void
@@ -4436,7 +4511,8 @@ lpfc_cmpl_els_fdisc(struct lpfc_hba *phba, struct lpfc_iocbq *cmdiocb,
                 else
                         lpfc_do_scr_ns_plogi(phba, vport);
 
-                lpfc_nlp_put(ndlp); /* Free Fabric ndlp for vports */
+                /* Unconditionaly kick off releasing fabric node for vports */
+                lpfc_nlp_put(ndlp);
         }
 
 out:
diff --git a/drivers/scsi/lpfc/lpfc_hbadisc.c b/drivers/scsi/lpfc/lpfc_hbadisc.c
index 644d96012d56..dc042bd97baa 100644
--- a/drivers/scsi/lpfc/lpfc_hbadisc.c
+++ b/drivers/scsi/lpfc/lpfc_hbadisc.c
@@ -149,7 +149,10 @@ lpfc_dev_loss_tmo_callbk(struct fc_rport *rport)
                 return;
 
         spin_lock_irq(&phba->hbalock);
-        evtp->evt_arg1  = ndlp;
+        /* We need to hold the node by incrementing the reference
+         * count until this queued work is done
+         */
+        evtp->evt_arg1  = lpfc_nlp_get(ndlp);
         evtp->evt       = LPFC_EVT_DEV_LOSS;
         list_add_tail(&evtp->evt_listp, &phba->work_list);
         if (phba->work_wait)
@@ -300,12 +303,18 @@ lpfc_work_list_done(struct lpfc_hba *phba)
                         ndlp = (struct lpfc_nodelist *) (evtp->evt_arg1);
                         lpfc_els_retry_delay_handler(ndlp);
                         free_evt = 0; /* evt is part of ndlp */
+                        /* decrement the node reference count held
+                         * for this queued work
+                         */
+                        lpfc_nlp_put(ndlp);
                         break;
                 case LPFC_EVT_DEV_LOSS:
                         ndlp = (struct lpfc_nodelist *)(evtp->evt_arg1);
-                        lpfc_nlp_get(ndlp);
                         lpfc_dev_loss_tmo_handler(ndlp);
                         free_evt = 0;
+                        /* decrement the node reference count held for
+                         * this queued work
+                         */
                         lpfc_nlp_put(ndlp);
                         break;
                 case LPFC_EVT_ONLINE:
@@ -1176,6 +1185,9 @@ lpfc_mbx_cmpl_reg_login(struct lpfc_hba *phba, LPFC_MBOXQ_t *pmb)
         lpfc_mbuf_free(phba, mp->virt, mp->phys);
         kfree(mp);
         mempool_free(pmb, phba->mbox_mem_pool);
+        /* decrement the node reference count held for this callback
+         * function.
+         */
         lpfc_nlp_put(ndlp);
 
         return;
@@ -1363,6 +1375,9 @@ lpfc_mbx_cmpl_ns_reg_login(struct lpfc_hba *phba, LPFC_MBOXQ_t *pmb)
 
         if (mb->mbxStatus) {
 out:
+                /* decrement the node reference count held for this
+                 * callback function.
+                 */
                 lpfc_nlp_put(ndlp);
                 lpfc_mbuf_free(phba, mp->virt, mp->phys);
                 kfree(mp);
@@ -1414,6 +1429,9 @@ out:
                 goto out;
         }
 
+        /* decrement the node reference count held for this
+         * callback function.
+         */
         lpfc_nlp_put(ndlp);
         lpfc_mbuf_free(phba, mp->virt, mp->phys);
         kfree(mp);
@@ -1661,13 +1679,14 @@ void
 lpfc_drop_node(struct lpfc_vport *vport, struct lpfc_nodelist *ndlp)
 {
         /*
-         * Use of lpfc_drop_node and UNUSED list. lpfc_drop_node should
+         * Use of lpfc_drop_node and UNUSED list: lpfc_drop_node should
          * be used if we wish to issue the "last" lpfc_nlp_put() to remove
-         * the ndlp from the vport.  The ndlp resides on the UNUSED list
-         * until ALL other outstanding threads have completed. Thus, if a
-         * ndlp is on the UNUSED list already, we should never do another
-         * lpfc_drop_node() on it.
+         * the ndlp from the vport. The ndlp marked as UNUSED on the list
+         * until ALL other outstanding threads have completed. We check
+         * that the ndlp not already in the UNUSED state before we proceed.
          */
+        if (ndlp->nlp_state == NLP_STE_UNUSED_NODE)
+                return;
         lpfc_nlp_set_state(vport, ndlp, NLP_STE_UNUSED_NODE);
         lpfc_nlp_put(ndlp);
         return;
@@ -2767,7 +2786,9 @@ lpfc_mbx_cmpl_fdmi_reg_login(struct lpfc_hba *phba, LPFC_MBOXQ_t *pmb)
         else
                 mod_timer(&vport->fc_fdmitmo, jiffies + HZ * 60);
 
-                                /* Mailbox took a reference to the node */
+        /* decrement the node reference count held for this callback
+         * function.
+         */
         lpfc_nlp_put(ndlp);
         lpfc_mbuf_free(phba, mp->virt, mp->phys);
         kfree(mp);
diff --git a/drivers/scsi/lpfc/lpfc_nportdisc.c b/drivers/scsi/lpfc/lpfc_nportdisc.c
index 783659aa2102..4a0e3406e37a 100644
--- a/drivers/scsi/lpfc/lpfc_nportdisc.c
+++ b/drivers/scsi/lpfc/lpfc_nportdisc.c
@@ -922,6 +922,9 @@ lpfc_cmpl_plogi_plogi_issue(struct lpfc_vport *vport,
                                            NLP_STE_REG_LOGIN_ISSUE);
                         return ndlp->nlp_state;
                 }
+                /* decrement node reference count to the failed mbox
+                 * command
+                 */
                 lpfc_nlp_put(ndlp);
                 mp = (struct lpfc_dmabuf *) mbox->context1;
                 lpfc_mbuf_free(phba, mp->virt, mp->phys);
diff --git a/drivers/scsi/lpfc/lpfc_sli.c b/drivers/scsi/lpfc/lpfc_sli.c
index 584c5451641f..fdd01e384e36 100644
--- a/drivers/scsi/lpfc/lpfc_sli.c
+++ b/drivers/scsi/lpfc/lpfc_sli.c
@@ -2605,7 +2605,7 @@ lpfc_sli_issue_mbox(struct lpfc_hba *phba, LPFC_MBOXQ_t *pmbox, uint32_t flag)
                                         "1806 Mbox x%x failed. No vport\n",
                                         pmbox->mb.mbxCommand);
                         dump_stack();
-                        return MBXERR_ERROR;
+                        return MBX_NOT_FINISHED;
                 }
         }