scsi: lpfc: Protect vport fc_nodes list with an explicit spin lock

In attempt to reduce the amount of unnecessary shost_lock acquisitions in
the lpfc driver, replace shost_lock with an explicit fc_nodes_list_lock
spinlock when accessing vport->fc_nodes lists.  Although vport memory
region is owned by shost->hostdata, it is driver private memory and an
explicit fc_nodes list lock for fc_nodes list mutations is more appropriate
than locking the entire shost.

Signed-off-by: Justin Tee <justin.tee@broadcom.com>
Link: https://lore.kernel.org/r/20240131185112.149731-14-justintee8345@gmail.com
Reviewed-by: Himanshu Madhani <himanshu.madhani@oracle.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

1 files changed, 18 insertions, 17 deletions

diff --git a/drivers/scsi/lpfc/lpfc_attr.c b/drivers/scsi/lpfc/lpfc_attr.c
index 142c90eb210f..023f4f2c62a6 100644
--- a/drivers/scsi/lpfc/lpfc_attr.c
+++ b/drivers/scsi/lpfc/lpfc_attr.c
@@ -344,6 +344,7 @@ lpfc_nvme_info_show(struct device *dev, struct device_attribute *attr,
 	struct lpfc_fc4_ctrl_stat *cstat;
 	uint64_t data1, data2, data3;
 	uint64_t totin, totout, tot;
+	unsigned long iflags;
 	char *statep;
 	int i;
 	int len = 0;
@@ -543,7 +544,7 @@ lpfc_nvme_info_show(struct device *dev, struct device_attribute *attr,
 	if (strlcat(buf, tmp, PAGE_SIZE) >= PAGE_SIZE)
 		goto buffer_done;
 
-	spin_lock_irq(shost->host_lock);
+	spin_lock_irqsave(&vport->fc_nodes_list_lock, iflags);
 
 	list_for_each_entry(ndlp, &vport->fc_nodes, nlp_listp) {
 		nrport = NULL;
@@ -617,7 +618,7 @@ lpfc_nvme_info_show(struct device *dev, struct device_attribute *attr,
 		if (strlcat(buf, tmp, PAGE_SIZE) >= PAGE_SIZE)
 			goto unlock_buf_done;
 	}
-	spin_unlock_irq(shost->host_lock);
+	spin_unlock_irqrestore(&vport->fc_nodes_list_lock, iflags);
 
 	if (!lport)
 		goto buffer_done;
@@ -681,7 +682,7 @@ lpfc_nvme_info_show(struct device *dev, struct device_attribute *attr,
 	goto buffer_done;
 
  unlock_buf_done:
-	spin_unlock_irq(shost->host_lock);
+	spin_unlock_irqrestore(&vport->fc_nodes_list_lock, iflags);
 
  buffer_done:
 	len = strnlen(buf, PAGE_SIZE);
@@ -3765,15 +3766,14 @@ lpfc_nodev_tmo_init(struct lpfc_vport *vport, int val)
 static void
 lpfc_update_rport_devloss_tmo(struct lpfc_vport *vport)
 {
-	struct Scsi_Host  *shost;
 	struct lpfc_nodelist  *ndlp;
+	unsigned long iflags;
 #if (IS_ENABLED(CONFIG_NVME_FC))
 	struct lpfc_nvme_rport *rport;
 	struct nvme_fc_remote_port *remoteport = NULL;
 #endif
 
-	shost = lpfc_shost_from_vport(vport);
-	spin_lock_irq(shost->host_lock);
+	spin_lock_irqsave(&vport->fc_nodes_list_lock, iflags);
 	list_for_each_entry(ndlp, &vport->fc_nodes, nlp_listp) {
 		if (ndlp->rport)
 			ndlp->rport->dev_loss_tmo = vport->cfg_devloss_tmo;
@@ -3788,7 +3788,7 @@ lpfc_update_rport_devloss_tmo(struct lpfc_vport *vport)
 						       vport->cfg_devloss_tmo);
 #endif
 	}
-	spin_unlock_irq(shost->host_lock);
+	spin_unlock_irqrestore(&vport->fc_nodes_list_lock, iflags);
 }
 
 /**
@@ -3974,8 +3974,8 @@ lpfc_vport_param_init(tgt_queue_depth, LPFC_MAX_TGT_QDEPTH,
 static int
 lpfc_tgt_queue_depth_set(struct lpfc_vport *vport, uint val)
 {
-	struct Scsi_Host *shost = lpfc_shost_from_vport(vport);
 	struct lpfc_nodelist *ndlp;
+	unsigned long iflags;
 
 	if (!lpfc_rangecheck(val, LPFC_MIN_TGT_QDEPTH, LPFC_MAX_TGT_QDEPTH))
 		return -EINVAL;
@@ -3983,14 +3983,13 @@ lpfc_tgt_queue_depth_set(struct lpfc_vport *vport, uint val)
 	if (val == vport->cfg_tgt_queue_depth)
 		return 0;
 
-	spin_lock_irq(shost->host_lock);
 	vport->cfg_tgt_queue_depth = val;
 
 	/* Next loop thru nodelist and change cmd_qdepth */
+	spin_lock_irqsave(&vport->fc_nodes_list_lock, iflags);
 	list_for_each_entry(ndlp, &vport->fc_nodes, nlp_listp)
 		ndlp->cmd_qdepth = vport->cfg_tgt_queue_depth;
-
-	spin_unlock_irq(shost->host_lock);
+	spin_unlock_irqrestore(&vport->fc_nodes_list_lock, iflags);
 	return 0;
 }
 
@@ -5236,8 +5235,8 @@ lpfc_vport_param_show(max_scsicmpl_time);
 static int
 lpfc_max_scsicmpl_time_set(struct lpfc_vport *vport, int val)
 {
-	struct Scsi_Host *shost = lpfc_shost_from_vport(vport);
 	struct lpfc_nodelist *ndlp, *next_ndlp;
+	unsigned long iflags;
 
 	if (val == vport->cfg_max_scsicmpl_time)
 		return 0;
@@ -5245,13 +5244,13 @@ lpfc_max_scsicmpl_time_set(struct lpfc_vport *vport, int val)
 		return -EINVAL;
 	vport->cfg_max_scsicmpl_time = val;
 
-	spin_lock_irq(shost->host_lock);
+	spin_lock_irqsave(&vport->fc_nodes_list_lock, iflags);
 	list_for_each_entry_safe(ndlp, next_ndlp, &vport->fc_nodes, nlp_listp) {
 		if (ndlp->nlp_state == NLP_STE_UNUSED_NODE)
 			continue;
 		ndlp->cmd_qdepth = vport->cfg_tgt_queue_depth;
 	}
-	spin_unlock_irq(shost->host_lock);
+	spin_unlock_irqrestore(&vport->fc_nodes_list_lock, iflags);
 	return 0;
 }
 lpfc_vport_param_store(max_scsicmpl_time);
@@ -6853,17 +6852,19 @@ lpfc_get_node_by_target(struct scsi_target *starget)
 	struct Scsi_Host  *shost = dev_to_shost(starget->dev.parent);
 	struct lpfc_vport *vport = (struct lpfc_vport *) shost->hostdata;
 	struct lpfc_nodelist *ndlp;
+	unsigned long iflags;
 
-	spin_lock_irq(shost->host_lock);
+	spin_lock_irqsave(&vport->fc_nodes_list_lock, iflags);
 	/* Search for this, mapped, target ID */
 	list_for_each_entry(ndlp, &vport->fc_nodes, nlp_listp) {
 		if (ndlp->nlp_state == NLP_STE_MAPPED_NODE &&
 		    starget->id == ndlp->nlp_sid) {
-			spin_unlock_irq(shost->host_lock);
+			spin_unlock_irqrestore(&vport->fc_nodes_list_lock,
+					       iflags);
 			return ndlp;
 		}
 	}
-	spin_unlock_irq(shost->host_lock);
+	spin_unlock_irqrestore(&vport->fc_nodes_list_lock, iflags);
 	return NULL;
 }