ice: Don't reset and rebuild for Tx timeout on PFC enabled queue
When there's a Tx timeout for a queue which belongs to a PFC enabled TC,
then it's not because the queue is hung but because PFC is in action.
In PFC, peer sends a pause frame for a specified period of time when its
buffer threshold is exceeded (due to congestion). Netdev on the other
hand checks if ACK is received within a specified time for a TX packet, if
not, it'll invoke the tx_timeout routine.
Signed-off-by: Avinash JD <avinash.dayanand@intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
diff --git a/drivers/net/ethernet/intel/ice/ice_main.c b/drivers/net/ethernet/intel/ice/ice_main.c
index 44ff4fe..8c792ec 100644
--- a/drivers/net/ethernet/intel/ice/ice_main.c
+++ b/drivers/net/ethernet/intel/ice/ice_main.c
@@ -5114,6 +5114,16 @@ static void ice_tx_timeout(struct net_device *netdev, unsigned int txqueue)
pf->tx_timeout_count++;
+ /* Check if PFC is enabled for the TC to which the queue belongs
+ * to. If yes then Tx timeout is not caused by a hung queue, no
+ * need to reset and rebuild
+ */
+ if (ice_is_pfc_causing_hung_q(pf, txqueue)) {
+ dev_info(ice_pf_to_dev(pf), "Fake Tx hang detected on queue %u, timeout caused by PFC storm\n",
+ txqueue);
+ return;
+ }
+
/* now that we have an index, find the tx_ring struct */
for (i = 0; i < vsi->num_txq; i++)
if (vsi->tx_rings[i] && vsi->tx_rings[i]->desc)
