Ice driver allocates per cpu XDP queues so that redirect path can safely
use smp_processor_id() as an index to the array. At the same time
though, XDP rings are used to pick NAPI context to call napi_schedule()
or set NAPIF_STATE_MISSED. When user reduces queue count, say to 8, and
num_possible_cpus() of underlying platform is 44, then this means queue
vectors with correlated NAPI contexts will carry several XDP queues.
This in turn can result in a broken behavior where NAPI context of
interest will never be scheduled and AF_XDP socket will not process any
traffic.
To fix this issue, use Rx ring to pull out the NAPI context.
Fixes: 2d4238f55697 ("ice: Add support for AF_XDP")
Fixes: 22bf877e528f ("ice: introduce XDP_TX fallback path")
Signed-off-by: Maciej Fijalkowski <maciej.fijalkowski@xxxxxxxxx>
---
drivers/net/ethernet/intel/ice/ice_xsk.c | 8 ++++----
1 file changed, 4 insertions(+), 4 deletions(-)
diff --git a/drivers/net/ethernet/intel/ice/ice_xsk.c b/drivers/net/ethernet/intel/ice/ice_xsk.c
index 49ba8bfdbf04..34d851d3e767 100644
--- a/drivers/net/ethernet/intel/ice/ice_xsk.c
+++ b/drivers/net/ethernet/intel/ice/ice_xsk.c
@@ -353,7 +353,7 @@ int ice_xsk_pool_setup(struct ice_vsi *vsi, struct xsk_buff_pool *pool, u16 qid)
if (if_running) {
ret = ice_qp_ena(vsi, qid);
if (!ret && pool_present)
- napi_schedule(&vsi->xdp_rings[qid]->q_vector->napi);
+ napi_schedule(&vsi->rx_rings[qid]->q_vector->napi);
else if (ret)
netdev_err(vsi->netdev, "ice_qp_ena error = %d\n", ret);
}
@@ -936,7 +936,7 @@ ice_xsk_wakeup(struct net_device *netdev, u32 queue_id,
struct ice_netdev_priv *np = netdev_priv(netdev);
struct ice_q_vector *q_vector;
struct ice_vsi *vsi = np->vsi;
- struct ice_tx_ring *ring;
+ struct ice_rx_ring *ring;
if (test_bit(ICE_VSI_DOWN, vsi->state))
return -ENETDOWN;
@@ -944,13 +944,13 @@ ice_xsk_wakeup(struct net_device *netdev, u32 queue_id,
if (!ice_is_xdp_ena_vsi(vsi))
return -EINVAL;
- if (queue_id >= vsi->num_txq)
+ if (queue_id >= vsi->num_txq || queue_id >= vsi->num_rxq)
return -EINVAL;
if (!vsi->xdp_rings[queue_id]->xsk_pool)
return -EINVAL;
- ring = vsi->xdp_rings[queue_id];
+ ring = vsi->rx_rings[queue_id];
/* The idea here is that if NAPI is running, mark a miss, so
* it will run again. If not, trigger an interrupt and
--
2.27.0
