On Tue, 2015-11-17 at 07:33 -0800, Eric Dumazet wrote: > This is exactly what I suggested in the changelog ;)
In v2 I am adding a patch to move skb_mark_napi_id() out of drivers into net core stack (napi_gro_receive() & napi_gro_frags()) But adding the rcu stuff in napi_napi_del() wont be a trivial patch. Also, mlx4 uses a NAPI for TX completions only, I suspect other drivers might do the same, so we would fill napi_hash[] with napi contexts that would not be used in RX at all. -- To unsubscribe from this list: send the line "unsubscribe netdev" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
