[SCSI] lpfc 8.3.16: FCoE Discovery and Failover Fixes

- Add support for re-reg'ing changed VPI w/o unregister VPI
- Copy WWN and state from old nodelist when target DID change.
- Clean up old nodelist rport and put the nodelist when target DID change.
- Clear the VFI_REGISTERED flag when UNREG_VFI completes.
- Made both checks of port_state against LPFC_FLOGI and LPFC_FDISC
  non-inclusive for ignoring CVL events.
- Added logic to stop retrying of the ongoing PLOGI and FDISC if
  transitioned back to the FCF rediscovery state in reaction to CVL.
- Removed the dependency of scanning of all the available FCF table
  entries for bulding round-robin bitmap.
- Use the lpfc_sli4_fcf_rr_read_fcf_rec() in responding to
  individual New FCF found event.

Signed-off-by: Alex Iannicelli <alex.iannicelli@emulex.com>
Signed-off-by: James Smart <james.smart@emulex.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>

1 files changed, 2 insertions, 5 deletions

diff --git a/drivers/scsi/lpfc/lpfc_sli.c b/drivers/scsi/lpfc/lpfc_sli.c
index 0ce9eb7ca4e..fb8905f893f 100644
--- a/drivers/scsi/lpfc/lpfc_sli.c
+++ b/drivers/scsi/lpfc/lpfc_sli.c
@@ -12295,12 +12295,9 @@ lpfc_sli4_fcf_scan_read_fcf_rec(struct lpfc_hba *phba, uint16_t fcf_index)
                 spin_lock_irq(&phba->hbalock);
                 phba->hba_flag |= FCF_DISC_INPROGRESS;
                 spin_unlock_irq(&phba->hbalock);
-                /* Reset FCF round robin index bmask for new scan */
-                if (fcf_index == LPFC_FCOE_FCF_GET_FIRST) {
-                        memset(phba->fcf.fcf_rr_bmask, 0,
-                               sizeof(*phba->fcf.fcf_rr_bmask));
+                /* Reset eligible FCF count for new scan */
+                if (fcf_index == LPFC_FCOE_FCF_GET_FIRST)
                         phba->fcf.eligible_fcf_cnt = 0;
-                }
                 error = 0;
         }
 fail_fcf_scan: