From: Karol Kolacinski <[email protected]>
The offset verification work task for E822 reschedules itself if the device
driver is resetting. However, it may already be executing and attempting to
access the PHY when a reset has begun. Flush the work to ensure that any
current execution has completed and will see the updated reset state. This
avoids the potential accesses to the PHY state while resetting. It is also
simpler than attempts to cancel the work which could lead to a race between
the clock owner PF and the port PF.
Fixes: 95af1f1c4c9f ("ice: reschedule ice_ptp_wait_for_offset_valid during
reset")
Signed-off-by: Karol Kolacinski <[email protected]>
Reviewed-by: Aleksandr Loktionov <[email protected]>
Signed-off-by: Arkadiusz Kubalewski <[email protected]>
Signed-off-by: Przemyslaw Korba <[email protected]>
Reviewed-by: Maciek Machnikowski <[email protected]>
Signed-off-by: Jacob Keller <[email protected]>
---
drivers/net/ethernet/intel/ice/ice_ptp.c | 3 +++
1 file changed, 3 insertions(+)
diff --git a/drivers/net/ethernet/intel/ice/ice_ptp.c
b/drivers/net/ethernet/intel/ice/ice_ptp.c
index 364f0f389d85..fccea0511eb2 100644
--- a/drivers/net/ethernet/intel/ice/ice_ptp.c
+++ b/drivers/net/ethernet/intel/ice/ice_ptp.c
@@ -3015,6 +3015,9 @@ void ice_ptp_prepare_for_reset(struct ice_pf *pf, enum
ice_reset_req reset_type)
if (reset_type == ICE_RESET_PFR)
return;
+ if (hw->mac_type == ICE_MAC_GENERIC)
+ kthread_flush_work(&ptp->port.ov_work.work);
+
if (ice_pf_src_tmr_owned(pf) && hw->mac_type == ICE_MAC_GENERIC_3K_E825)
ice_ptp_prepare_rebuild_sec(pf, false, reset_type);
--
2.56.0.rc0.395.gd1f3524e15dc
