[patch 1/2] ice: Use ICE_FW_RECOVERY_MODE state to track recovery mode

[Cyrill Gorcunov]

If card is in recovery mode and we remove it physically then during
ice_remove() we read device register to fetch its state again which
is wrong since the device is no longer present. Fix it marking the
device state with ICE_FW_RECOVERY_MODE bit.

Signed-off-by: Cyrill Gorcunov <gorcunov@xxxxxxxxx>
---
 drivers/net/ethernet/intel/ice/ice.h      |    1 +
 drivers/net/ethernet/intel/ice/ice_main.c |    9 ++++++++-
 2 files changed, 9 insertions(+), 1 deletion(-)

Index: linux-tip.git/drivers/net/ethernet/intel/ice/ice.h
===================================================================
--- linux-tip.git.orig/drivers/net/ethernet/intel/ice/ice.h
+++ linux-tip.git/drivers/net/ethernet/intel/ice/ice.h
@@ -310,6 +310,7 @@ enum ice_pf_state {
 	ICE_PHY_INIT_COMPLETE,
 	ICE_FD_VF_FLUSH_CTX,		/* set at FD Rx IRQ or timeout */
 	ICE_AUX_ERR_PENDING,
+	ICE_FW_RECOVERY_MODE,	/* Firmware in recovery mode */
 	ICE_STATE_NBITS		/* must be last */
 };
 
Index: linux-tip.git/drivers/net/ethernet/intel/ice/ice_main.c
===================================================================
--- linux-tip.git.orig/drivers/net/ethernet/intel/ice/ice_main.c
+++ linux-tip.git/drivers/net/ethernet/intel/ice/ice_main.c
@@ -5121,6 +5121,8 @@ static int ice_probe_recovery_mode(struc
 
 	dev_err(dev, "Firmware recovery mode detected. Limiting functionality. Refer to the Intel(R) Ethernet Adapters and Devices User Guide for details on firmware recovery mode\n");
 
+	set_bit(ICE_FW_RECOVERY_MODE, pf->state);
+
 	INIT_HLIST_HEAD(&pf->aq_wait_list);
 	spin_lock_init(&pf->aq_wait_lock);
 	init_waitqueue_head(&pf->aq_wait_queue);
@@ -5366,7 +5368,12 @@ static void ice_remove(struct pci_dev *p
 		msleep(100);
 	}
 
-	if (ice_is_recovery_mode(&pf->hw)) {
+	/*
+	 * .remove() may be called when adapter is physically
+	 * unplugged so we can't read the fw state but use
+	 * the flag instead.
+	 */
+	if (test_bit(ICE_FW_RECOVERY_MODE, pf->state)) {
 		ice_service_task_stop(pf);
 		scoped_guard(devl, priv_to_devlink(pf)) {
 			ice_deinit_devlink(pf);