cnss2: Release rddm_complete wait during default recovery

Currently Host driver calls cnss_force_collect_rddm API
to trigger RDDM and wait for dump collection. If device
fails to move to RDDM, CNSS driver initiate DEFAULT recovery
(which is recovery without RDDM dump collection) and return
failure to Host driver. Host driver sees error and again try
recovery which results in double recovery.
To avoid this issue, return success to Host driver if
CNSS driver has already initiated the DEFAULT recovery
and release rddm_complete wait event for DEFAULT recovery
cases also.

Change-Id: Ib4b324704d15b40ec98985745b1981522b3b4f2d
CRs-Fixed: 3627928

cnss2/pci.c

@@ -2857,6 +2857,7 @@ int cnss_pci_call_driver_remove(struct cnss_pci_data *pci_priv)
 
 	if (test_bit(CNSS_DRIVER_RECOVERY, &plat_priv->driver_state) &&
 	    test_bit(CNSS_DRIVER_PROBED, &plat_priv->driver_state)) {
+		complete(&plat_priv->rddm_complete);
 		pci_priv->driver_ops->shutdown(pci_priv->pci_dev);
 	} else if (test_bit(CNSS_DRIVER_UNLOADING, &plat_priv->driver_state)) {
 		pci_priv->driver_ops->remove(pci_priv->pci_dev);
@@ -5821,6 +5822,7 @@ int cnss_pci_force_fw_assert_hdlr(struct cnss_pci_data *pci_priv)
 		cnss_pci_dump_debug_reg(pci_priv);
 		cnss_schedule_recovery(&pci_priv->pci_dev->dev,
 				       CNSS_REASON_DEFAULT);
+		ret = 0;
 		goto runtime_pm_put;
 	}