libfc: Update rport reference counting
Originally libfc would just be initializing the refcount to '1', and
using the disc_mutex to synchronize if and when the final put should be
happening. This has a race condition as the mutex might be delayed,
causing other threads to access an invalid structure. This patch
updates the rport reference counting to increase the reference every
time 'rport_lookup' is called, and decreases the reference
correspondingly. This removes the need to hold 'disc_mutex' when
removing the structure, and avoids the above race condition.
Signed-off-by: Hannes Reinecke <hare@suse.com>
Acked-by: Vasu Dev <vasu.dev@intel.com>
Reviewed-by: Johannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
diff --git a/drivers/scsi/fcoe/fcoe_ctlr.c b/drivers/scsi/fcoe/fcoe_ctlr.c
index 3e83d48..ada4bde 100644
--- a/drivers/scsi/fcoe/fcoe_ctlr.c
+++ b/drivers/scsi/fcoe/fcoe_ctlr.c
@@ -2496,14 +2496,13 @@
struct fcoe_rport *frport;
int ret = -1;
- rcu_read_lock();
rdata = lport->tt.rport_lookup(lport, port_id);
if (rdata) {
frport = fcoe_ctlr_rport(rdata);
memcpy(mac, frport->enode_mac, ETH_ALEN);
ret = 0;
+ kref_put(&rdata->kref, lport->tt.rport_destroy);
}
- rcu_read_unlock();
return ret;
}
@@ -2585,11 +2584,7 @@
fcoe_ctlr_vn_send(fip, FIP_SC_VN_PROBE_REQ, fcoe_all_vn2vn, 0);
return;
}
- mutex_lock(&lport->disc.disc_mutex);
rdata = lport->tt.rport_lookup(lport, new->ids.port_id);
- if (rdata)
- kref_get(&rdata->kref);
- mutex_unlock(&lport->disc.disc_mutex);
if (rdata) {
if (rdata->ids.node_name == new->ids.node_name &&
rdata->ids.port_name == new->ids.port_name) {