From: Sameeh Jubran sameehj@amazon.com
[ Upstream commit 3921a81c31df6057183aeb7f7d204003bf699d6f ]
This patch fixes two issues with XDP:
1. If the XDP verdict is XDP_ABORTED we break the loop, which results in us handling one buffer per napi cycle instead of the total budget (usually 64). To overcome this simply change the xdp_verdict check to != XDP_PASS. When the verdict is XDP_PASS, the skb is not expected to be NULL.
2. Update the residual budget for XDP_DROP and XDP_ABORTED, since packets are handled in these cases.
Fixes: 548c4940b9f1 ("net: ena: Implement XDP_TX action") Signed-off-by: Sameeh Jubran sameehj@amazon.com Signed-off-by: David S. Miller davem@davemloft.net Signed-off-by: Greg Kroah-Hartman gregkh@linuxfoundation.org --- drivers/net/ethernet/amazon/ena/ena_netdev.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-)
--- a/drivers/net/ethernet/amazon/ena/ena_netdev.c +++ b/drivers/net/ethernet/amazon/ena/ena_netdev.c @@ -1642,11 +1642,9 @@ static int ena_clean_rx_irq(struct ena_r &next_to_clean);
if (unlikely(!skb)) { - if (xdp_verdict == XDP_TX) { + if (xdp_verdict == XDP_TX) ena_free_rx_page(rx_ring, &rx_ring->rx_buffer_info[rx_ring->ena_bufs[0].req_id]); - res_budget--; - } for (i = 0; i < ena_rx_ctx.descs; i++) { rx_ring->free_ids[next_to_clean] = rx_ring->ena_bufs[i].req_id; @@ -1654,8 +1652,10 @@ static int ena_clean_rx_irq(struct ena_r ENA_RX_RING_IDX_NEXT(next_to_clean, rx_ring->ring_size); } - if (xdp_verdict == XDP_TX || xdp_verdict == XDP_DROP) + if (xdp_verdict != XDP_PASS) { + res_budget--; continue; + } break; }